E Coli 2015

Enterohemorrhagic
Escherichia coli
and Other Shiga Toxin-Producing E. coli
Ecoli TItle pages.indd 1 3/26/15 3:52 PM

Enterohemorrhagic
Escherichia coli
and Other Shiga Toxin-Producing E. coli
EDITED BY
Vanessa Sperandio
Department of Microbiology
University of Texas
Dallas, TX 75390
and
Carolyn J. Hovde
School of Food Science
Idaho INBRE Program
Moscow, ID 83844-3025
Washington, DC
Ecoli TItle pages.indd 2 3/26/15 3:52 PM

Copyright 2015 American Society for Microbiology. ASM Press is a registered trademark of the
American Society for Microbiology. All rights reserved. No part of this publication may be repro-
duced or transmitted in whole or in part or reutilized in any form or by any means, electronic or
mechanical, including photocopying and recording, or by any information storage and retrieval
system, without permission in writing from the publisher.
Disclaimer: To the best of the publishers knowledge, this publication provides information con-
cerning the subject matter covered that is accurate as of the date of publication. The publisher is
not providing legal, medical, or other professional services. Any reference herein to any specific
commercial products, procedures, or services by trade name, trademark, manufacturer, or otherwise
does not constitute or imply endorsement, recommendation, or favored status by the American
Society for Microbiology (ASM). The views and opinions of the author(s) expressed in this publication
do not necessarily state or reflect those of ASM, and they shall not be used to advertise or endorse
any product.
Library of Congress Cataloging-in-Publication Data
Enterohemorrhagic Escherichia coli and other shiga toxin-producing E. coli / edited by Vanessa
Sperandio, Department of Microbiology, University of Texas, Dallas, TX, [and] Carolyn J. Hovde,
School of Food Science, Idaho INBRE Program, Moscow, ID.
pages cm
Includes index.
ISBN 978-1-55581-878-4 (print) -- ISBN 978-1-55581-879-1 (electronic) 1. Escherichia coli infections.
2. Escherichia coli O157:H7. 3. Verocytotoxins. 4. Hemolytic-uremic syndrome. I. Sperandio, Vanessa,
editor. II. Hovde, Carolyn J., editor.
QR201.E82E58 2015
616.926--dc23
2015003716
doi:10.1128/9781555818791
Printed in the United States of America
10 9 8 7 6 5 4 3 2 1
Address editorial correspondence to: ASM Press, 1752 N St., N.W., Washington, DC 20036-2904, USA.
Send orders to: ASM Press, P.O. Box 605, Herndon, VA 20172, USA.
Phone: 800-546-2416; 703-661-1593. Fax: 703-661-1501.
E-mail: books@asmusa.org
Online: http://www.asmscience.org
Contents
Contributorsix
Prefacexv
OVERVIEW
1 Overview and Historical Perspectives3
James B. Kaper and Alison D. OBrien
MICROBIOLOGY
2 Taxonomy Meets Public Health: The Case of Shiga Toxin-Producing
Escherichia coli17
Flemming Scheutz
3 Shiga Toxin (Stx) Classification, Structure, and Function37
Angela R. Melton-Celsa
4 Enterohemorrhagic Escherichia coli Genomics: Past, Present,
and Future55
Shah M. Sadiq, Tracy H. Hazen, David A. Rasko, and Mark Eppinger
PATHOGENESIS
5 Role of Shiga/Vero Toxins in Pathogenesis75
Fumiko Obata and Thomas Obrig
6 The Locus of Enterocyte Effacement and Associated Virulence Factors
of Enterohemorrhagic Escherichia coli97
Mark P. Stevens and Gad M. Frankel
7 Enterohemorrhagic Escherichia coli Adhesins131
Brian D. McWilliams and Alfredo G. Torres
8 Animal Models of Enterohemorrhagic Escherichia coli Infection157
Jennifer M. Ritchie
9 Enterohemorrhagic Escherichia coli Virulence Gene Regulation175
Jay L. Mellies and Emily Lorenzen
INCIDENCE, EPIDEMIOLOGY, AND ECOLOGY

10 Shiga Toxin (Verotoxin)-Producing Escherichia coli in Japan199
Jun Terajima, Sunao Iyoda, Makoto Ohnishi, and Haruo Watanabe
v
vicontents
11 Animal Reservoirs of Shiga Toxin-Producing Escherichia coli211

Anil K. Persad and Jeffrey T. LeJeune
12 Shiga Toxin-Producing Escherichia coli in Fresh Produce:
A Food Safety Dilemma231
Peter Feng
13 Public Health Microbiology of Shiga Toxin-Producing
Escherichia coli245
Alfredo Caprioli, Gaia Scavia, and Stefano Morabito
DIAGNOSIS, DETECTION, AND STRAIN CHARACTERIZATION

14 Detection of Shiga Toxin-Producing Escherichia coli from
Nonhuman Sources and Strain Typing263
Lothar Beutin and Patrick Fach
CLINICAL, PATHOLOGICAL, AND PATHOPHYSIOLOGICAL ASPECTS

15 Shiga Toxin/Verocytotoxin-Producing Escherichia coli Infections:
Practical Clinical Perspectives299
T. Keefe Davis, Nicole C. A. J. van de Kar, and Phillip I. Tarr
16 The Inflammatory Response during Enterohemorrhagic
Escherichia coli Infection321
Jaclyn S. Pearson and Elizabeth L. Hartland
17 New Therapeutic Developments against Shiga Toxin-Producing
Escherichia coli341
Angela R. Melton-Celsa and Alison D. OBrien
HOST DETERMINANTS OF DISEASE AND HOST RESPONSE

18 Risk Factors for Shiga Toxin-Producing Escherichia coli-Associated
Human Diseases361
Marta Rivas, Isabel Chinen, Elizabeth Miliwebsky, and Marcelo Masana
19 Enterohemorrhagic Escherichia coli Pathogenesis and the
Host Response381
Diana Karpman and Anne-lie Sthl
20 The Interplay between the Microbiota and Enterohemorrhagic
Escherichia coli403
Reed Pifer and Vanessa Sperandio
PREVENTION AND CONTROL STRATEGIES

21. Preharvest Food Safety for Escherichia coli O157 and Other
Pathogenic Shiga Toxin-Producing Strains421
Thomas E. Besser, Carrie E. Schmidt, Devendra H. Shah, and Smriti Shringi
contents vii
22 Peri- and Postharvest Factors in the Control of Shiga Toxin-Producing

Escherichia coli in Beef437
Rodney A. Moxley and Gary R. Acuff
23 Veterinary Public Health Approach to Managing Pathogenic
Verocytotoxigenic Escherichia coli in the Agri-Food Chain457
Geraldine Duffy and Evonne McCabe
24 Clinical Studies of Escherichia coli O157:H7 Conjugate Vaccines in
Adults and Young Children477
Shousun Chen Szu and Amina Ahmed
25 Vaccination of Cattle against Escherichia coli O157:H7487
David R. Smith
ESCHERICHIA COLI O104:H4

26 Escherichia coli O104:H4 Pathogenesis: An Enteroaggregative
E. coli/Shiga Toxin-Producing E. coli Explosive Cocktail of
High Virulence505
Fernando Navarro-Garcia
THE WAY FORWARD

27 The Way Forward533
Vanessa Sperandio
Index541
About the Editors553
Contributors
Gary R. Acuff
Department of Animal Science, 2471 TAMU, Texas A&M University,
College Station, TX 77843-2471
Amina Ahmed
Levine Childrens Specialty Center - Pediatric Infectious Disease,
Carolina Medical Centers, Charlotte, NC 28203
Thomas E. Besser
Veterinary Microbiology and Pathology, Washington State University, Pullman,
WA 99164
Lothar Beutin
National Reference Laboratory for Escherichia coli, Department of Biological
Safety, Federal Institute for Risk Assessment (BfR), Diedersdorfer Weg 1,
D-12277 Berlin, Germany
Alfredo Caprioli
EU Reference Laboratory for E. coli, Dipartimento di Sanit Pubblica
Veterinaria e Sicurezza Alimentare, Istituto Superiore di Sanit,
Viale Regina Elena 299, 00161 Rome, Italy
Isabel Chinen
Servicio Fisiopatogenia, Instituto Nacional de Enfermedades Infecciosas
ANLIS Dr. C. G. Malbrn, (1281) Buenos Aires, Argentina
T. Keefe Davis
Division of Nephrology, Department of Pediatrics, Washington University
School of Medicine, St. Louis, MO 63110
Geraldine Duffy
Teagasc Food Research Centre, Ashtown, Dublin 15, Ireland
Mark Eppinger
Department of Biology and South Texas Center for Emerging Infectious
Diseases, University of Texas at San Antonio, San Antonio, TX 78249
Patrick Fach
Food Safety Laboratory, ANSES (French Agency for Food, Environmental
and Occupational Health and Safety), Fr-94706 Maisons-Alfort, France
ix
xCONTRIBUTORS
Peter Feng
Division of Microbiology, U.S. Food and Drug Administration,
College Park, MD 20740-3835
Gad M. Frankel
MRC Centre for Molecular Bacteriology & Infection, Department of Life
Sciences, Imperial College London, London, SW7 2AZ, United Kingdom
Elizabeth L. Hartland
Department of Microbiology and Immunology, University of Melbourne,
Victoria 3010, and Murdoch Childrens Research Institute, Royal Childrens
Hospital, Parkville, Victoria 3052, Australia
Tracy H. Hazen
Institute for Genome Sciences, Department of Microbiology and Immunology,
University of Maryland School of Medicine, Baltimore, MD 21201
Sunao Iyoda
Department of Bacteriology, National Institute of Infectious Diseases, 1-23-1
Toyama, Shinjuku-ku, Tokyo 162-8640, Japan
James B. Kaper
Department of Microbiology & Immunology, University of Maryland School of
Medicine, Baltimore, MD 21122
Diana Karpman
Department of Pediatrics, Clinical Sciences, Lund University, 22185 Lund,
Sweden
Jeffrey T. LeJeune
Food Animal Health Research Program, Ohio Agricultural Research and
Development Center, The Ohio State University, Wooster, OH 4491
Emily Lorenzen
Laboratory of Chemical Biology and Signal Transduction, The Rockefeller
University, 1230 York Avenue, New York, NY 10065
Marcelo Masana
Instituto Tecnologa de Alimentos, Centro de Investigacin de Agroindustria,
Instituto Nacional de Tecnologa Agropecuaria, INTA, (B1708WAB) Morn,
Pcia. De Buenos Aires, Argentina
Evonne McCabe
Teagasc Food Research Centre, Ashtown, Dublin 15, Ireland
Brian D. McWilliams
Department of Microbiology and Immunology, University of Texas
Medical Branch, Galveston, TX 77555
Jay L. Mellies
Department of Biology, Reed College, 3203 SE Woodstock Blvd.,
Portland, OR 97202
CONTRIBUTORS xi
Angela R. Melton-Celsa
Department of Microbiology & Immunology, Uniformed Services University of
the Health Sciences, 4301 Jones Bridge Road, Bethesda, MD 20814
Elizabeth Miliwebsky
Stefano Morabito
Rodney A. Moxley
School of Veterinary Medicine and Biomedical Sciences,
University of Nebraska-Lincoln, Lincoln, NE 68685-0905
Fernando Navarro-Garcia
Department of Cell Biology, Centro de Investigacin y de Estudios
Avanzados del IPN (CINVESTAV-IPN), Mxico DF, Mexico
Fumiko Obata
University of Maryland School of Medicine, 685 W. Baltimore St.,
HSF-1 Suite 380, Baltimore, MD 21201
Alison D. OBrien
Department of Microbiology & Immunology, Uniformed Services
University of the Health Sciences, Bethesda, MD 20814
Thomas Obrig
University of Maryland School of Medicine, 685 W. Baltimore St.,
HSF-1 Suite 380, Baltimore, MD 21201
Makoto Ohnishi
Department of Bacteriology, National Institute of Infectious Diseases,
1-23-1 Toyama, Shinjuku-ku, Tokyo 162-8640, Japan
Jaclyn S. Pearson
Department of Microbiology and Immunology, University of Melbourne,
Victoria 3010, Australia
Anil K. Persad
Food Animal Health Research Program, Ohio Agricultural Research and
Development Center, The Ohio State University, Wooster, OH 4491
Reed Pifer
Department of Microbiology and Department of Biochemistry,
University of Texas Southwestern Medical Center, Dallas, TX 75390
David A. Rasko
Institute for Genome Sciences, Department of Microbiology and Immunology,
University of Maryland School of Medicine, Baltimore, MD 21201
xiiCONTRIBUTORS
Jennifer M. Ritchie
School of Biosciences and Medicine, University of Surrey, Guildford GU27XH,
United Kingdom
Marta Rivas
Shah M. Sadiq
Department of Biology and South Texas Center for Emerging Infectious
Diseases, University of Texas at San Antonio, San Antonio, TX 78249
Gaia Scavia
Flemming Scheutz
WHO Collaborating Centre for Reference and Research on Escherichia and
Klebsiella, Department of Microbiology and Infection Control,
Statens Serum Institut, DK-2300 Copenhagen S, Denmark
Carrie E. Schmidt
Veterinary Microbiology and Pathology, Washington State University,
Pullman, WA 99164
Devendra H. Shah
Pullman, WA 99164
Smriti Shringi
Pullman, WA 99164
David R. Smith
College of Veterinary Medicine, Mississippi State University,
Mississippi State, MS 39762-6100
Vanessa Sperandio
Department of Microbiology and Department of Biochemistry,
University of Texas Southwestern Medical Center, Dallas, TX 75390
Anne-lie Sthl
Department of Pediatrics, Clinical Sciences, Lund University, 22185
Lund, Sweden
Mark P. Stevens
The Roslin Institute & Royal (Dick) School of Veterinary Studies,
University of Edinburgh, Midlothian, EH25 9RG, United Kingdom
Shousun Chen Szu
Eunice Kennedy Shriver National Institute of Child Health & Human
Development, National Institutes of Health, 9000 Rockville Pike,
Bethesda, MD 20892
CONTRIBUTORS xiii
Phillip I. Tarr
Division of Gastroenterology, Hepatology, and Nutrition, Department of
Pediatrics, and Department of Molecular Microbiology,
Washington University School of Medicine, St. Louis, MO 63110
Jun Terajima
Department of Microbiology, National Institute of Health Sciences,
Kamiyoga 1-18-1, Setagaya-ku, Tokyo 158-8501, Japan
Alfredo G. Torres
Department of Pathology and Sealy Center for Vaccine Development,
University of Texas Medical Branch, Galveston, TX 77555
Nicole C. A. J. van de Kar
Division of Nephrology, Department of Pediatrics, Radboud University
Medical Centre, Nijmegen, The Netherlands
Haruo Watanabe
Director, National Institute of Infectious Diseases, 1-23-1 Toyama,
Shinjuku-ku, Tokyo 162-8640, Japan
Preface
This book is an exceptional compilation of our current worldwide understand-

ing of the enterohemorrhagic E. coli (EHEC) and other Shiga toxin-producing
E. coli. It spans diverse topics including microbial pathogenesis, pathophysiology
of the disease, food safety, genetic analysis, veterinary microbiology, epidemiology,
and environmental microbiology. It was compiled as an introduction, review, and
critical overview of the pertinent areas of knowledge and brings the previous edi-
tion (Kaper JB, OBrien AD [ed], Escherichia coli O157:H7 and Other Shiga Toxin-
Producing E. coli Strains, ASM Press, 1998) up to date with the current literature.
The style and content are intended to make this volume of interest and val-
ue as a resource for research scientists, clinicians, students, health profession-
als, policy makers, and those in industry. In addition, we believe the text could
be used for advanced courses in microbiology, food safety, infectious disease, or
microbial pathogenesis.
The book contributors come from many and diverse research disciplines. Its
breadth demonstrates the complexity of the problem of EHEC and Shiga toxin-
producing E. coli. For rapid and timely dissemination, each chapter previously
appeared in Microbiology Spectrum and is available online, but they are assem-
bled here as a convenient hardbound reference volume. The book begins with a
broad overview and historical perspective, followed by eight sections that orga-
nize information into subtopics. The text concludes with the traditional look to
the future in chapter 27, titled The Way Forward, which was not previously
published.
We gratefully acknowledge the outstanding skill and organizational work of all
those at ASM Press who made this book possible. It was a pleasure working with
Greg Payne, Ellie Tupper, Kenneth April, Courtenay Brown, and Cathy Balogh.
Both of us have devoted our professional careers to understanding the EHEC
in hopes of contributing to effective interventions to improve human health. As
editors, we are humbled by the exceptional work done by our colleagues, the
chapter authors. They brought their full and thoughtful expertise and knowl-
edge to their writing and fueled our excitement for the book. We hope that you,
as a reader, will find the topics covered to be relevant and that their depth will
bring new insights to your own work.
Vanessa Sperandio
Carolyn J. Hovde
December 2014
xv
OVERVIEW
Part01.proof.3d 1 Manila Typesetting Company 03/31/15 12:06

Overview and Historical Perspectives
JAMES B. KAPER1 and ALISON D. OBRIEN2

1
The scope of topics covered in this book reflects the broad areas of research
required for the comprehensive study of Shiga toxin-producing Escherichia
coli (STEC) infections. Substantial progress has been made in all of these areas
since the first edition of this book (1). Although this second edition brings the
field up to date in all major areas of research, these pathogens have a long and
complicated history, and understanding this history is valuable for a full un-
derstanding of this field. The purpose of this chapter is to set the stage for this
book by examining the seminal discoveries about STEC biology, epidemiology,
and pathogenesis. In this article, we refer to the cytotoxins of E. coli O157:H7,
E. coli O104:H4, and other E. coli as Shiga toxins (Stxs; formerly called Shiga-
like toxins), hence the nomenclature STEC. However, for reasons described
below, a number of investigators prefer the term verotoxin (VT). We refer the
reader to past discussions of nomenclature (2, 3) for a better understanding
of the historical basis for the dichotomy in nomenclature. Additionally, a re-
cently published multicenter study by Scheutz and colleagues (4) provides clear
guidance on nomenclature for Stx subtypes. Scheutz reviews that typing scheme
in chapter 2 of this volume.
1
Department of Microbiology and Immunology, University of Maryland School of Medicine, Baltimore, MD 21122;
2
Department of Microbiology and Immunology, Uniformed Services University of the Health Sciences, Bethesda, MD
20814.
Enterohemorrhagic Escherichia coli and Other Shiga Toxin-Producing E. coli
Edited by Vanessa Sperandio and Carolyn J. Hovde
2015 American Society for Microbiology, Washington, DC
doi:10.1128/microbiolspec.EHEC-0028-2014.
Chap01.proof.3d 3 Manila Typesetting Company 03/31/15 10:44

4 KAPER AND OBRIEN
Our understanding of what constitutes a acquisition, through horizontal genetic ex-

virulent STEC isolate for humans has evolved, change, of the Stx2a-converting phage. Rec-
as have the organisms themselves. The history ognizing the essential role Stx plays in the
of STEC as an emerging pathogen can be di- development of severe disease caused by
vided into two phases. Phase one of STEC STEC infection, and for chronological reasons,
chronology is a history of the convergence of we begin this tale with the discovery and
two independent laboratory-based research characterization of Stxs.
tracks and two independent epidemiology-
based areas of investigation. One laboratory-
based track focused on Stx, its discovery,
HISTORY
characterization, and relationship to E. coli
cytotoxins, and the other concentrated on
Stxs: History of the Field
enteropathogenic E. coli (EPEC) adherence
characteristics. These studies led to the re- In 1898, Kioshi Shiga (7) provided the defini-
alization that a subset of STEC strains, in- tive description of the agent of epidemic bac-
cluding E. coli O157:H7, shares with EPEC terial dysentery, Shigella dysenteriae type 1
both the pathogenic trait of producing (Shigas bacillus). Five years following that
attaching and effacing (A/E) intestinal lesions discovery, Conradi (8) reported that extracts
and the genes to provoke that lesion. The of Shigas bacillus paralyzed and killed rab-
epidemiology-based areas include the search bits. Similar findings were published inde-
for the cause of the well-described but idio- pendently by Neisser and Shiga (9). The next
pathic hemolytic-uremic syndrome (HUS) nearly 70 years of Stx research led to the clear
and the search for the agent responsible for separation of the endotoxic activity associated
a newly described clinical syndrome called with Shigas bacillus from the activity of the
hemorrhagic colitis. protein Stx; the partial purification of Stx (10);
Phase two of the STEC story started with the discovery that high iron concentrations
the understanding that the backbone of the inhibit Stx synthesis (11); the seminal obser-
E. coli strain that produces Stx does not nec- vation by Bridgwater et al. (12) and Howard
essarily have to have adherence traits sim- (13) that Stx appears to target vascular endo-
ilar to those of EPEC and E. coli O157:H7 to thelium in the brain; and the discovery by
cause a large-scale outbreak of human dis- Vicari et al. that Stx is lethal for certain epi-
ease. Witness the large 2011 STEC outbreak thelial cells in culture (14). Although these
in Germany in which 3,816 cases were re- findings were of interest to toxicologists, none
ported (including 54 deaths); 845 of those of the results proved a direct role for Stx in
cases led to HUS (5). The etiologic agent in the pathogenesis of shigellosis. Only decades
that outbreak was an enteroaggregative E. coli later, with the evaluation of data obtained
O104:H4 strain that had become transduced from infection of volunteers (15) and subse-
with the Stx type 2a (Stx2a)-encoding phage quently of monkeys (16) with S. dysenteriae
(6). Enteroaggregative E. coli (EAEC) strains type 1, was it clear that production of Stx by
that do not make Stxs are well-established the organism exacerbates the severity of the
etiologic agents of diarrhea. The fact that an intestinal and systemic lesions in human sub-
EAEC isolate can be converted from an ex- jects and increases the intestinal pathology in
clusively diarrheagenic agent to one that primate hosts. The ultimate proof of a role for
also causes hemorrhagic colitis and HUS in Stx in shigellosis due to Shigas bacillus was
so many patients affirms the preeminent the establishment of a connection between
role that Stx plays in these syndromes and production of this and related toxins with the
illustrates how readily an E. coli strain can subsequent development of HUS (see below
evolve into a life-threatening pathogen by the or chapters 5 and 19).

CHAPTER 1 Overview and Historical Perspectives 5
In 1972, Keusch and colleagues made the the verotoxin produced by these organisms
significant finding that Stx alone caused fluid was linked epidemiologically to the develop-
accumulation and enteritis in ligated rabbit ment of HUS (see below).
intestinal segments (17). This observation re- The mid to late 1980s heralded the era of
vealed that Stx can contribute to the intestinal the molecular characterization of the genes
phase of bacillary dysentery, i.e., bloody diar- encoding the Stx family members (reviewed
rhea. That this enterotoxic activity of Stx is a in reference 30). In the mid-1980s, it was also
result of the same molecule responsible for its discovered that Stx1 and Stx2 are usually en-
cytotoxic and lethal activities was convinc- coded on lambdoid prophages in E. coli (31
ingly demonstrated by the purification of Stx 35). In contrast, Stx of Shigas bacillus and
to homogeneity (first by Olsnes and Eiklid Stx2e (edema disease toxin) of animal STEC
[18], followed shortly by reports from OBrien were later shown to be chromosomally en-
et al. [19], Brown and colleagues [20], and coded (27, 36). Subsequent genomic analysis
Donohue-Rolfe and coworkers [21]) and the revealed that the genome of E. coli O157:H7 is
subsequent testing of that material for all riddled with prophage regions that not only
three bioactivities (19). encode Stx but also other potential virulence
factors (see chapter 4 of this volume). The
toxin genes themselves show considerable var-
Stxs and Verotoxins Are Different Names
iation that can correlate with epidemiological
for the Same Family of Toxins
significance, with more than 100 Stx variants
With the availability of purified Stx came the so far described (see chapter 2).
capacity to produce monospecific, cytotoxin-
neutralizing rabbit anti-Stx antibodies. OBrien
Intimate Adherence to Mucosal
and colleagues used such sera to ascertain
Epithelium: The Connection
that certain strains of E. coli produce a cyto-
between EPEC and EHEC
toxin that can be neutralized by anti-Stx (22,
23), an observation that explains the original With the genetics and biology of Stxs fairly
Shiga-like toxin nomenclature. The prelimi- well elucidated by the mid to late 1980s, the
nary report of that discovery (23) occurred in focus of research on STEC broadened to ad-
1977, the same year that Konowalchuk and dress the question of how E. coli O157:H7
colleagues found that certain diarrheagenic adheres to epithelial cells. The primary find-
E. coli strains make a cytotoxin that can kill ing that initiated a series of discoveries about
Vero cells (24), hence the name verotoxin. In E. coli O157:H7 adherence mechanisms was
1983, OBrien and colleagues (25) reported that the observation by Tzipori et al. (37) that
a Shiga-like toxin was produced by the E. coli E. coli O157:H7 causes intestinal A/E lesions in
O157:H7 strain that had caused an outbreak of gnotobiotic piglets and that these lesions re-
hemorrhagic colitis in the United States (see semble those produced by Stx-negative EPEC,
below) and that this toxin was the same as the albeit at different sites in the bowel of the
verotoxin shown by Johnson et al. (26) to be animals. The A/E lesion is characterized by
produced by E. coli O157:H7. Thus, 1983 be- intimate adherence of the bacteria to the enter-
came the year when the paths of research on ocyte membrane and effacement of the mi-
Stxs and verotoxins merged. Subsequent ge- crovilli. This observation suggested to Levine
netic studies showed that Stx1 (VT1) differs by (38) that these lesions might be a hallmark
none or only a single amino acid from Shiga of E. coli O157:H7 and related bacteria. He
toxin (27, 28). These studies on the Stx/VT of proposed that the capacity of E. coli O157:H7
E. coli culminated in a pivotal report published and related organisms to evoke A/E lesions,
by Karmali et al. (29) in that same year. In that together with the production of Stxs by these
paper Karmali and colleagues proposed that microbes and the presence of a characteristic

6 KAPER AND OBRIEN
large plasmid, was sufficient to define a new reactivity, a type III secretion apparatus, and
category of virulent enterohemorrhagic E. coli secreted proteins (48). The island was named
(EHEC). With the identification of pathogenic locus of enterocyte effacement (LEE) (48). In
and genetic markers for this newly recognized chapter 6 of this volume, Stevens and Frankel
group of E. coli came the realization that the review the LEE pathogenicity island, intimin,
E. coli O26:H11 serotype, which had long been the type III secretion system, and other asso-
considered a classic EPEC serotype, should be ciated virulence factors of EHEC.
reclassified as an EHEC. This reclassification The discovery and characterization of LEE
was supported by the findings that strains of prompted the realization that the term EHEC
the O26:H11 serotype also produced Stx and represents a subset of STEC since not all
possessed the large plasmid found in O157:H7 STEC strains contain LEE and the large EHEC
(39). Thus, the O26:H11 serotype is one example plasmid (see chapter 2). Although the majority
of an STEC serotype that had been associated of STEC strains associated with human dis-
with diarrheal disease (reviewed in reference ease possess LEE and the large plasmid, some
40) long before the discovery of E. coli O157:H7. strains lacking these factors, most notably
The pathognomonic A/E histopathology O104:H4, have been implicated in human dis-
of EPEC, E. coli O157:H7, and a few additional ease, thereby leading to the use of the more
serotypes of STEC was subsequently shown general term STEC rather than EHEC.
by Knutton et al. (41) to correspond in vitro
to a lesion characterized by bacterial micro-
Hemolytic-Uremic Syndrome
colonies intimately adherent to the surface of
tissue culture cells, with accumulation of cy- HUS, first described in 1955 by Gasser et al.
toskeletal actin under the bacteria. The actin (49) in Switzerland, is defined by a triad of
accumulation was visualized in that study by clinical features that include acute renal fail-
a fluorescent actin stain (FAS) test in which ure, thrombocytopenia, and microangiopathic
fluoresceinated phalloidin was used as a hemolytic anemia. HUS is a leading cause of
probe. The FAS-positive phenotype of EPEC acute renal failure in children, and in some
bound to HEp-2 laryngeal epithelial cells was studies it is the most common cause of renal
used by Jerse et al. (42) to screen EPEC for failure in this age group. A variety of agents,
genes required for this intimate adherence. including drugs, chemicals, toxins, and various
These investigators identified the gene eae microbes, had been proposed as the cause of
(for E. coli attaching and effacing), which was HUS; indeed, before 1983, most nephrologists
also present in O157:H7. The eae gene prod- thought that HUS was a multifactorial disease
uct, appropriately named intimin, was subse- that could result from a number of initiating
quently shown to be required for EPEC to events (50). Because HUS occasionally oc-
cause A/E lesions in gnotobiotic pigs (43). curred in outbreaks, an infectious cause was
Shortly after the discovery of EPEC eae, the sought. The strongest documented linkage
homologous gene was cloned and sequenced between HUS and a microorganism was the
from two strains of E. coli O157:H7 (44, 45). association with S. dysenteriae type 1, but
The intimin of E. coli O157:H7 was also shown numerous microorganisms, including Salmo-
to be necessary but not sufficient to induce nella typhi, Campylobacter jejuni, Yersinia
A/E lesions in vitro and in vivo (46, 47). pseudotuberculosis, Streptococcus pneumoniae,
An additional twist to the similarities in rickettsia-like organisms, coxsackievirus, echo-
pathogenic mechanisms between EPEC and virus, and Epstein-Barr virus, were proposed as
EHEC was the provocative finding of McDaniel the causative agent (reviewed in reference 50).
et al. (48) that eae lies within a pathogenic- Several studies noted that many, if not the
ity island of approximately 35 kb and that majority, of HUS cases were preceded by di-
this island encodes genes for attachment, FAS arrhea. Interestingly, a survey of HUS in South

Africa led Kibel and Barnard in 1968 (51) to from a specific restaurant chain (56). This syn-
speculate that HUS is caused by an entero- drome, called hemorrhagic colitis, was char-
pathogenic strain of E. coli that had acquired a acterized by severe abdominal cramps, grossly
bacteriophage. These authors also raised con- bloody stools, little or no fever, and evidence
cerns that treatment with antibiotics might of colonic mucosal edema, erosion, or hemor-
lead to excessive bacterial destruction and en- rhage (57). E. coli strains of a previously rare
hanced absorption of toxin, with an adverse serotype, O157:H7, were isolated from the
clinical outcome (see chapter 15 of this vol- stools of about half the cases but from none
ume). Additional information on the history of of the healthy controls. Strains of this sero-
HUS and various proposed pathogenic mech- type were subsequently shown to produce
anisms can be found in several reviews (50, 52, Stx (see above). Numerous studies have since
53; chapters 5 and 19 of this volume). confirmed that O157:H7 is an important cause
The key event in the linkage of HUS and of hemorrhagic colitis, nonbloody diarrhea,
STEC was the 1983 report in The Lancet by and HUS in the United States, Canada, the
Karmali et al. (29) that sporadic cases of HUS United Kingdom, and Japan, as reviewed in
were linked to the presence of verotoxin, which other chapters in this volume.
OBrien et al. (25) reported was equivalent to The abrupt appearance of E. coli O157:H7
Stx (see above), and/or E. coli that produced Stx in 1982 raised questions as to whether this
in patients stools. The toxigenic E. coli strains organism had recently emerged as a patho-
characterized by Karmali and colleagues be- gen or had always been present and had sim-
longed to different serogroups, thus ruling out ply been unrecognized. To address this issue,
a single strain as the cause of this disease, investigators at national laboratories in the
and serum collected from several patients con- United States, Canada, and the United King-
tained rising titers of neutralizing antibody dom reviewed their records and E. coli col-
activity against verotoxin. This initial report lections and found archived E. coli O157:H7
was confirmed by a prospective case-control strains recovered before 1982 from the stool
study that linked cases of HUS with isolation of one patient in the United States, one patient
from stool of STEC belonging to at least six O in the United Kingdom, and six patients in
serogroups (O26, O111, O113, O121, O145, and Canada, some of whom had bloody diarrhea
O157) (54). The 1985 article in Journal of In- (reviewed in reference 52). The clinical syn-
fectious Diseases describing the case-control drome of hemorrhagic colitis is so distinctive
study was reprinted by that journal in 2004 that outbreaks are unlikely to have been over-
along with a commentary (55) describing it as looked, although occasional cases of a hem-
one of the landmark papers published in that orrhagic colitis-like syndrome of unknown
journal over the first centennial of its history. etiology were reported in the 1960s and 1970s
The connection between Stx (VT) production (reviewed in references 50 and 52). Thus,
by an E. coli strain and the development of the available evidence indicates that the inci-
HUS after infection with that organism was dence of infections with O157:H7 and other
most recently substantiated by the 2011 O104:H4 STEC strains increased in the 1980s and 1990s.
outbreak, showing that diarrheagenic entero- However, this conclusion is confounded by the
aggregative E. coli could cause HUS after ac- increase in the number of laboratories seeking
quiring the capacity to produce Stx2 (see below). this pathogen (50) and by the 25 to 75% of
patients with O157:H7 infection who present
with nonbloody diarrhea (50, 58), a clinical
Hemorrhagic Colitis
manifestation that may go unrecognized as
In 1982, two outbreaks of a severe bloody di- one of the manifestations of O157:H7 disease.
arrheal syndrome in Oregon and Michigan Studies by Whittam and colleagues using mul-
were linked to the consumption of hamburgers tilocus enzyme electrophoresis demonstrated

8 KAPER AND OBRIEN
the stepwise evolution of STEC O157:H7 from causative agent might be a novel pathogenic
an O55:H7 ancestor (59, 60), a Stx-negative variant of STEC.
serotype of EPEC that had previously been Investigators quickly established that the
only associated with nonbloody diarrhea. O104:H4 strain lacked the LEE pathogenicity
The comprehensive study of the Oregon island present in O157:H7 and other common
and Michigan O157:H7 outbreaks of hemor- EHEC strains. Instead, the strain possessed an
rhagic colitis (in which no cases of HUS were unusual combination of virulence factors that
noted) was published in the March 24, 1983, were typical of EAEC in addition to Stx (64).
issue of New England Journal of Medicine (61). EAEC strains that do not make Stxs are well-
Karmalis study linking verotoxin-producing established etiologic agents of nonbloody di-
E. coli strains of different serogroups to HUS arrhea that can be either acute or persistent
was published in the March 19, 1983, issue of in duration. Disease is seen in both children
Lancet (29), and 1 week later, in the March 26, and adults, in travelers, and in people infected
1983, issue of Lancet, OBrien and colleagues with human immunodeficiency virus in both
(25) reported that a Shiga-like toxin was pro- the developed and developing world (reviewed
duced by the E. coli O157:H7 strains from the in reference 65). The term enteroaggregative
Oregon and Michigan outbreaks of hemor- is derived from the stacked-brick appearance
rhagic colitis and that this toxin was the same of EAEC on intestinal epithelial cells, in which
as the verotoxin previously shown to be pro- large numbers of bacteria closely adhere to
duced by E. coli O157:H7 (26). Thus, March enterocytes in a biofilm. The pathogenesis
1983 was a momentous month in which nu- of EAEC is poorly understood, but a variety
merous laboratory, clinical, and epidemiological of virulence factors have been described, in-
studies on HUS, bloody diarrhea, verotoxins, cluding aggregative adherence fimbriae (AAF)
Shiga toxin, and E. coli came together to estab- that mediate intestinal adherence and induce
lish the field of STEC infections. inflammation, several serine protease auto-
transporters of Enterobacteriaceae (SPATEs)
implicated in mucosal damage and coloniza-
O104:H4
tion, and several other putative adhesins and
In May 2011, a large outbreak of gastroenter- toxins (reviewed in reference 66).
itis and HUS began in Germany, one of the Analysis of the genome sequence of the
largest outbreaks of STEC yet reported. In O104:H4 strain from the Germany outbreak
3 months, 3,816 cases (including 54 deaths) revealed that it closely resembled other EAEC
were reported, of which 845 (22%) were HUS strains, but that it had become transduced
(5). The etiologic agent was identified as with the Stx2a-encoding phage (6). The ge-
Stx-producing E. coli O104:H4, and sprouts nome sequence also revealed the presence of
were identified as the outbreak vehicle (62). an unusual combination of SPATEs and sev-
Sprouts had previously been identified as the eral antibiotic-resistance factors. Other E. coli
vehicle in STEC outbreaks, most notably the O104:H4 strains unrelated to the Germany
1996 outbreak in Sakai City, Japan, where outbreak did not possess Stx. Further discus-
12,680 cases were reported (63). The most sion of the pathogenesis of O104:H4 is pre-
striking clinical and epidemiological finding sented in chapter 26 of this volume.
in the 2011 outbreak was the very high num- Although more than 100 different serotypes
ber of HUS cases (22% of all cases), with 88% of E. coli have been shown to produce Stx, the
of the HUS cases occurring in adults rather majority of such STEC strains are not con-
than in children. In contrast, the incidence of sidered to be pathogens. Several serotypes
HUS in the 1996 Japan outbreak was 1% and such as O26:H11, O111:NM, and O121:H19 con-
all cases were in children. The dramatic fea- tain the LEE and other pathogenicity islands
tures of the 2011 outbreak suggested that the found in O157:H7 and are clearly pathogens.

A few other STEC serotypes, such as O91:H21 pathogenesis, with a particular emphasis on
and O113:H21, lack the LEE but have addi- effects in the renal system. In chapter 6,
tional virulence factors and have been epide- Stevens and Frankel review the LEE patho-
miologically implicated as pathogens (67). The genicity island and virulence factors encoded
different STEC serotypes and their epidemi- therein, as well as other virulence factors en-
ological significance are reviewed in chapter coded outside the LEE. Colonization of the
2. In the case of the German Shiga toxin- intestinal tract is an essential first step in
producing enteroaggregative E. coli O104:H4, STEC pathogenesis, and a variety of potential
the addition of the Stx2a phage to a pathotype adherence factors have been described, as
of E. coli that was already capable of avidly ad- reviewed by McWilliams and Torres in
hering to and damaging the intestinal epithe- chapter 7. Unfortunately, there is no single
lium produced a novel pathogen with profound animal model that reproduces all aspects of
clinical and epidemiological consequences. STEC disease, but Ritchie reviews the various
models available and their advantages and dis-
advantages in chapter 8. The range of envi-
THE PRESENT ronments where STEC can be foundfrom
the farm environment to the human intestine
The current themes and directions of STEC requires numerous regulatory genetic elements
research span an astonishing range of topics. to optimize expression of virulence factors and
Multiple disciplines encompass epidemiology, survival factors. Mellies and Lorenzen describe
animal ecology, food safety, clinical microbiol- the complex regulation of STEC virulence in
ogy, gastroenterology, nephrology, infectious chapter 9.
disease, toxicology, bacterial pathogenesis, cell The incidence, epidemiology, and ecology
biology, and immunology. Topics in this area of STEC are reviewed in the third section.
of research range from farm management of In chapter 10, Terajima et al. review the in-
livestock and manure to clinical management cidence and epidemiology of STEC in Japan,
of end-stage renal disease. This book, edited the site of the largest STEC outbreak reported.
and written by internationally recognized ex- Animals, particularly cattle, serve as the res-
perts in this area, reflects this breadth of topics. ervoir of STEC infections, and Persad and
The first section of the volume describes LeJeune review this critical reservoir in chap-
the microbiology of STEC. In chapter 2, ter 11. Transmission to humans most often
Flemming Scheutz describes the taxonomy of involves consumption of contaminated food
STEC and Stx toxins and relates this infor- items. Initial outbreaks of STEC disease in-
mation to the public health significance of the volved improperly cooked hamburgers, an
different serotypes and toxin subtypes. The issue that was relatively easy to address by in-
discussion of Stx toxin continues in chapter 3 creasing cooking temperatures. However, as
where Angela Melton-Celsa reviews the struc- reviewed by Feng in chapter 12, most out-
ture and function of these toxins. Sadiq and breaks in recent years involved produce that
colleagues take a genomic perspective in is consumed raw. Caprioli et al. review in
chapter 4 to review the history of typing and chapter 13 the epidemiology and other public
genetic analysis from distinguishing STEC health aspects of STEC infection, with a par-
strains using pregenomic methodologies to ticular emphasis on Europe. Methods for de-
the current technology, where the genome se- tecting STEC from nonhuman sources and
quences of multiple strains can be determined strain typing are reviewed by Beutin and Fach
in a single day. in chapter 14.
The pathogenesis of STEC infections is Clinical, pathological, and pathophysiolog-
covered in section two. In chapter 5, Obata ical aspects of human disease are reviewed
and Obrig discuss the role of Stx toxins in in chapters 15 through 17. Tarr and coauthors

10 KAPER AND OBRIEN
review in chapter 15 the clinical features of promise than direct immunization of humans.
STEC infections in humans, including out- In chapter 24, Szu and Ahmed present data
comes and prognosis, and provide insights showing that parenteral O157 lipopolysaccha-
from both gastroenterological and nephrologi- ride conjugate vaccines are safe and immuno-
cal perspectives. The inflammatory response to genic in children and adults. However, there
STEC infection and the virulence factors these are multiple potential problems in vaccinat-
pathogens have evolved to thwart this response ing humans against STEC disease, including
are discussed by Pearson and Hartland in finding a population with a high enough in-
chapter 16. Unfortunately, no ideal therapy is cidence in which to determine vaccine effi-
available for STEC infections, and the use of cacy and identifying an appropriate target
antimicrobials is contraindicated, at least for population to vaccinate once vaccine efficacy
typical EHEC infections, although investigators is established (reviewed in reference 68).
of the German O104:H4 outbreak reported the Success in vaccinating cattle to reduce fecal
benefit of azithromycin treatment to reduce shedding of STEC has been achieved, and
fecal shedding of the organism. This issue, David Smith reviews these studies in chapter
along with novel therapeutic interventions 25.
under study, is reviewed by Melton-Celsa and The emergence of STEC-EAEC O104:H4
OBrien in chapter 17. in 2011 was a landmark development in the
Host determinants of disease and host re- history of STEC infections. The virulence fac-
sponses encompass factors ranging from cul- tors that combined to produce this highly vir-
tural and dietary practices to host genetics and ulent strain are discussed by Navarro-Garcia
immune status to intestinal microbiota, all in chapter 26. Such a development makes
of which can play important roles in STEC one wonder what the future holds for the
infection and outcome. Risk factors for STEC STEC field, and Vanessa Sperandio offers in
infections are discussed by Rivas et al. in chap- the final chapter some speculations on cur-
ter 18. The host response and other aspects of rent questions and future directions for in-
STEC pathogenesis are reviewed by Karpman vestigating this fascinating and ever-changing
and Stahl in chapter 19. With the recognition pathogen.
that regulation of EHEC virulence factors can
be influenced by commensal intestinal bacte-
ACKNOWLEDGMENTS
ria, the interplay between the microbiota and
EHEC can be important, as reviewed by Pifer The opinions or assertions contained herein
and Sperandio in chapter 20. are the private ones of the authors and are
Prevention and control strategies to reduce not to be construed as official or reflecting
or eliminate the risk of STEC infections are the views of the Department of Defense, the
particularly important in the control of STEC Uniformed Services University of the Health
infections. Preharvest and peri- and post- Sciences, or the National Institutes of Health.
harvest food safety factors are reviewed by This work was supported by National Insti-
Besser and colleagues in chapter 21 and by tutes of Health grants R01 DK58957 (JBK),
Moxley and Acuff in chapter 22. A veterinary R37 AI21657 (JBK), and R37 AI020148 (ADO).
public health approach to managing patho- We declare no conflicts of interest with re-
genic STEC in the agri-food chain is discussed gard to the manuscript.
by Duffy and McCabe in chapter 23. Vaccines
have been critical in reducing the disease
CITATION
burden in humans for many infectious dis-
eases, but for STEC infections, vaccines to Kaper JB, OBrien AD. 2014. Overview and
reduce carriage in the bovine reservoir to re- historical perspectives. Microbiol Spectrum
duce transmission to humans may hold more 2(6):EHEC-0028-2014.

REFERENCES 11. Dubos RJ, Geiger JW. 1946. Preparation and

properties of Shiga toxin and toxoid. J Exp Med
1. Kaper JB, OBrien AD (ed). 1998. Escherichia coli 84:143156.
O157:H7 and Other Shiga Toxin-Producing E. coli 12. Bridgwater FA, Morgan RS, Rowson KE,
Strains. ASM Press, Washington, DC. Wright GP. 1955. The neurotoxin of Shigella
2. Calderwood SB, Acheson DWK, Keusch GT, shigae: morphological and functional lesions pro-
Barrett TJ, Griffin PM, Strockbine NA, duced in the central nervous system of rabbits.
Swaminathan B, Kaper JB, Levine MM, Kaplan Br J Exp Pathol 36:447453.
BS, Karch H, OBrien AD, Obrig TG, Takeda Y, 13. Howard JG. 1955. Observations on the intoxica-
Tarr PI, Wachsmuth IK. 1996. Proposed new tion produced in mice and rabbits by the neuro-
nomenclature for SLT (VT) family. ASM News toxin of Shigella shigae. Br J Exp Pathol 36:
62:118119. 439446.
3. Karmali MA, Lingwood CA, Petrie M, Brunton 14. Vicari G, Olitzki AL, Olitzki Z. 1960. The action
J, Gyles C. 1996. Maintaining the existing phe- of the thermolabile toxin of Shigella dysenteriae
notype nomenclatures for E. coli cytotoxins. ASM on cells cultivated in vitro. Br J Exp Pathol 41:179
News 62:167169. 189.
4. Scheutz F, Teel LD, Beutin L, Pierard D, 15. Levine MM, DuPont HL, Formal SB, Hornick
Buvens G, Karch H, Mellmann A, Caprioli A, RB, Takeuchi A, Gangarosa EJ, Snyder MJ,
Tozzoli R, Morabito S, Strockbine NA, Melton- Libonati JP. 1973. Pathogenesis of Shigella dys-
Celsa AR, Sanchez M, Persson S, OBrien AD. enteriae 1 (Shiga) dysentery. J Infect Dis 127:261
2012. Multicenter evaluation of a sequence-based 270.
protocol for subtyping Shiga toxins and stan- 16. Fontaine A, Arondel J, Sansonetti PJ. 1988. Role
dardizing Stx nomenclature. J Clin Microbiol of Shiga toxin in the pathogenesis of bacillary
50:29512963. dysentery, studied by using a Tox- mutant of
5. Frank C, Werber D, Cramer JP, Askar M, Faber Shigella dysenteriae 1. Infect Immun 56:3099
M, an der Heiden M, Bernard H, Fruth A, 3109.
Prager R, Spode A, Wadl M, Zoufaly A, Jordan 17. Keusch GT, Grady GF, Mata LJ, McIver J. 1972.
S, Kemper MJ, Follin P, Muller L, King LA, The pathogenesis of Shigella diarrhea. I. Entero-
Rosner B, Buchholz U, Stark K, Krause G. 2011. toxin production by Shigella dysenteriae I. J Clin
Epidemic profile of Shiga-toxin-producing Esch- Invest 51:12121218.
erichia coli O104:H4 outbreak in Germany. N Engl 18. Olsnes S, Eiklid K. 1980. Isolation and charac-
J Med 365:17711780. terization of Shigella shigae cytotoxin. J Biol Chem
6. Rasko DA, Webster DR, Sahl JW, Bashir A, 255:284289.
Boisen N, Scheutz F, Paxinos EE, Sebra R, 19. OBrien AD, LaVeck GD, Griffin DE, Thompson
Chin CS, Iliopoulos D, Klammer A, Peluso P, MR. 1980. Characterization of Shigella dysenteriae
Lee L, Kislyuk AO, Bullard J, Kasarskis A, 1 (Shiga) toxin purified by anti-Shiga toxin affinity
Wang S, Eid J, Rank D, Redman JC, Steyert chromatography. Infect Immun 30:170179.
SR, Frimodt-Moller J, Struve C, Petersen AM, 20. Brown JE, Griffin DE, Rothman SW, Doctor
Krogfelt KA, Nataro JP, Schadt EE, Waldor BP. 1982. Purification and biological character-
MK. 2011. Origins of the E. coli strain causing ization of Shiga toxin from Shigella dysenteriae 1.
an outbreak of hemolytic-uremic syndrome in Infect Immun 36:9961005.
Germany. N Engl J Med 365:709717. 21. Donohue-Rolfe A, Keusch GT, Edson C,
7. Shiga K. 1898. Ueber den Dysenterie-bacillus Thorley-Lawson D, Jacewicz M. 1984. Patho-
(Bacillus dysenteriae). Zentralbl Baktcriol Orig genesis of Shigella diarrhea. IX. Simplified high
24:913918. yield purification of Shigella toxin and charac-
8. Conradi H. 1903. Uber Iosliche, durch asptische terization of subunit composition and function by
Autolyse erhaltene Giftstoffe vonRuhr- und the use of subunit-specific monoclonal and poly-
Typhus-Bazillen. Dtsch Med Wochenschr 29:26 clonal antibodies. J Exp Med 160:17671781.
28. 22. OBrien AD, LaVeck GD. 1983. Purification and
9. Neisser M, Shiga K. 1903. Ueber freie Receptoren characterization of a Shigella dysenteriae 1-like
von Typhus- und Dysenterie-Bazillen und ueber toxin produced by Escherichia coli. Infect Immun
das Dysenterie Toxin. Dtsch Med Wochenschr 40:675683.
29:6162. 23. OBrien AD, Thompson MR, Cantey JR, Formal
10. Van Heyningen WE, Gladstone GP. 1953. The SB. 1977. Production of a Shigella dysenteriae-like
neurotoxin of Shigella shigae. III. The effect of toxin by pathogenic Escherichia coli, abstr. B-103.
iron on production of the toxin. Br J Exp Pathol Abstr 77th Annu Meet Am Soc Microbiol. Amer-
34:221229. ican Society for Microbiology, Washington, DC.

12 KAPER AND OBRIEN
24. Konowalchuk J, Speirs JI, Stavric S. 1977. Vero a Shiga-like toxin type II variant from Escherichia
response to a cytotoxin of Escherichia coli. Infect coli strain responsible for edema disease of swine.
Immun 18:775779. J Bacteriol 170:42234230.
25. OBrien AO, Lively TA, Chen ME, Rothman 37. Tzipori S, Wachsmuth IK, Chapman C, Birden
SW, Formal SB. 1983. Escherichia coli O157:H7 R, Brittingham J, Jackson C, Hogg J. 1986.
strains associated with haemorrhagic colitis in The pathogenesis of hemorrhagic colitis caused
the United States produce a Shigella dysenteriae 1 by Escherichia coli O157:H7 in gnotobiotic piglets.
(SHIGA) like cytotoxin. Lancet i:702. J Infect Dis 154:712716.
26. Johnson WM, Lior H, Bezanson GS. 1983. Cy- 38. Levine MM. 1987. Escherichia coli that cause di-
totoxic Escherichia coli O157:H7 associated with arrhea: enterotoxigenic, enteropathogenic, entero-
haemorrhagic colitis in Canada. Lancet i:76. invasive, enterohemorrhagic, and enteroadherent.
27. Strockbine NA, Jackson MP, Sung LM, Holmes J Infect Dis 155:377389.
RK, OBrien AD. 1988. Cloning and sequencing of 39. Levine MM, Xu JG, Kaper JB, Lior H, Prado V,
the genes for Shiga toxin from Shigella dysenteriae Tall B, Nataro J, Karch H, Wachsmuth K. 1987.
type 1. J Bacteriol 170:11161122. A DNA probe to identify enterohemorrhagic
28. Takao T, Tanabe T, Hong YM, Shimonishi Y, Escherichia coli of O157:H7 and other serotypes
Kurazono H, Yutsudo T, Sasakawa C, Yoshikawa that cause hemorrhagic colitis and hemolytic
M, Takeda Y. 1988. Identity of molecular structure uremic syndrome. J Infect Dis 156:175182.
of Shiga-like toxin I (VT1) from Escherichia coli 40. Robins-Browne RM. 1987. Traditional entero-
O157:H7 with that of Shiga toxin. Microb Pathog pathogenic Escherichia coli of infantile diarrhea.
5:5769. Rev Infect Dis 9:2853.
29. Karmali MA, Steele BT, Petric M, Lim C. 1983. 41. Knutton S, Baldwin T, Williams PH, McNeish
Sporadic cases of haemolytic-uraemic syndrome AS. 1989. Actin accumulation at sites of bacterial
associated with faecal cytotoxin and cytotoxin- adhesion to tissue culture cells: basis of a new
producing Escherichia coli in stools. Lancet i:619 diagnostic test for enteropathogenic and entero-
620. hemorrhagic Escherichia coli. Infect Immun 57:
30. OBrien AD, Tesh VL, Donohue-Rolfe A, 12901298.
Jackson MP, Olsnes S, Sandvig K, Lindberg AA, 42. Jerse AE, Yu J, Tall BD, Kaper JB. 1990. A ge-
Keusch GT. 1992. Shiga toxin: biochemistry, genetic locus of enteropathogenic Escherichia coli
netics, mode of action, and role in pathogenesis. necessary for the production of attaching and
Curr Top Microbiol Immunol 180:6594. effacing lesions on tissue culture cells. Proc Natl
31. OBrien AD, Marques LR, Kerry CF, Newland Acad Sci USA 87:78397843.
JW, Holmes RK. 1989. Shiga-like toxin converting 43. Tzipori S, Gunzer F, Donnenberg MS, de
phage of enterohemorrhagic Escherichia coli strain Montigny L, Kaper JB, Donohue-Rolfe A. 1995.
933. Microb Pathog 6:381390. The role of the eaeA gene in diarrhea and neu-
32. OBrien AD, Newland JW, Miller SF, Holmes rological complications in a gnotobiotic piglet
RK, Smith HW, Formal SB. 1984. Shiga-like toxin- model of enterohemorrhagic Escherichia coli in-
converting phages from Escherichia coli strains fection. Infect Immun 63:36213627.
that cause hemorrhagic colitis or infantile diar- 44. Beebakhee G, Louie M, De Azavedo J, Brunton
rhea. Science 226:694696. J. 1992. Cloning and nucleotide sequence of the eae
33. Scotland SM, Smith HR, Willshaw GA, Rowe B. gene homologue from enterohemorrhagic Esche-
1983. Vero cytotoxin production in strain of Esche- richia coli serotype O157:H7. FEMS Microbiol Lett
richia coli is determined by genes carried on bac- 70:6368.
teriophage. Lancet ii:216. 45. Yu J, Kaper JB. 1992. Cloning and characteriza-
34. Smith HW, Green P, Parsell Z. 1983. Vero cell tion of the eae gene of enterohaemorrhagic Esch-
toxins in Escherichia coli and related bacteria: erichia coli O157:H7. Mol Microbiol 6:411417.
transfer by phage and conjugation and toxic ac- 46. Donnenberg MS, Tzipori S, McKee ML,
tion in laboratory animals, chickens and pigs. OBrien AD, Alroy J, Kaper JB. 1993. The role of
J Gen Microbiol 129:31213137. the eae gene of enterohemorrhagic Escherichia
35. Strockbine NA, Marques LR, Newland JW, coli in intimate attachment in vitro and in a por-
Smith HW, Holmes RK, OBrien AD. 1986. cine model. J Clin Invest 92:14181424.
Two toxin-converting phages from Escherichia 47. McKee ML, Melton-Celsa AR, Moxley RA,
coli O157:H7 strain 933 encode antigenically dis- Francis DH, OBrien AD. 1995. Enterohemor-
tinct toxins with similar biologic activities. Infect rhagic Escherichia coli O157:H7 requires intimin
Immun 53:135140. to colonize the gnotobiotic pig intestine and to
36. Weinstein DL, Jackson MP, Samuel JE, Holmes adhere to HEp-2 cells. Infect Immun 63:3739
RK, OBrien AD. 1988. Cloning and sequencing of 3744.

48. McDaniel TK, Jarvis KG, Donnenberg MS, Producing E. coli Strains. ASM Press, Washington,
Kaper JB. 1995. A genetic locus of enterocyte DC.
effacement conserved among diverse enterobac- 61. Riley LW, Remis RS, Helgerson SD, McGee HB,
terial pathogens. Proc Natl Acad Sci USA 92:1664 Wells JG, Davis BR, Hebert RJ, Olcott ES,
1668. Johnson LM, Hargrett NT, Blake PA, Cohen
49. Gasser C, Gautier E, Steck A, Siebenmann RE, ML. 1983. Hemorrhagic colitis associated with a
Oechslin R. 1955. [Hemolytic-uremic syndrome: rare Escherichia coli serotype. N Engl J Med
bilateral necrosis of the renal cortex in acute 308:681685.
acquired hemolytic anemia]. Schweiz Med 62. Buchholz U, Bernard H, Werber D, Bohmer
Wochenschr 85:905909 (In German). MM, Remschmidt C, Wilking H, Delere Y,
50. Karmali MA. 1989. Infection by verocytotoxin- an der Heiden M, Adlhoch C, Dreesman J,
producing Escherichia coli. Clin Microbiol Rev Ehlers J, Ethelberg S, Faber M, Frank C, Fricke
2:1538. G, Greiner M, Hohle M, Ivarsson S, Jark U,
51. Kibel MA, Barnard PJ. 1968. The haemolytic- Kirchner M, Koch J, Krause G, Luber P, Rosner
uraemic syndrome: a survey in Southern Africa. B, Stark K, Kuhne M. 2011. German outbreak of
S Afr Med J 42:692698. Escherichia coli O104:H4 associated with sprouts.
52. Griffin PM, Tauxe RV. 1991. The epidemiology of N Engl J Med 365:17631770.
infections caused by Escherichia coli O157:H7, 63. Fukushima H, Hashizume T, Morita Y, Tanaka
other enterohemorrhagic E. coli, and the associ- J, Azuma K, Mizumoto Y, Kaneno M, Matsuura
ated hemolytic uremic syndrome. Epidemiol Rev M, Konma K, Kitani T. 1999. Clinical experiences
13:6098. in Sakai City Hospital during the massive out-
53. Kaplan BS, Trompeter RS, Moake JL. 1992. In- break of enterohemorrhagic Escherichia coli O157
troduction, p xviixxvii. In Kaplan BS, Trompeter infections in Sakai City, 1996. Pediatr Int 41:213
RS, Moake JL (ed), Hemolytic Uremic Syndrome 217.
and Thrombotic Thrombocytopenic Purpura. 64. Scheutz F, Nielsen EM, Frimodt-Moller J,
Marcel Dekker, Inc, New York, NY. Boisen N, Morabito S, Tozzoli R, Nataro JP,
54. Karmali MA, Petric M, Lim C, Fleming PC, Caprioli A. 2011. Characteristics of the entero-
Arbus GS, Lior H. 1985. The association between aggregative Shiga toxin/verotoxin-producing Esch-
idiopathic hemolytic uremic syndrome and in- erichia coli O104:H4 strain causing the outbreak of
fection by verotoxin-producing Escherichia coli. haemolytic uraemic syndrome in Germany, May to
J Infect Dis 151:775782. June 2011. Euro Surveill 16(24):pii=19889.
55. Blaser MJ. 2004. Bacteria and diseases of un- 65. Huang DB, Mohanty A, DuPont HL, Okhuysen
known cause: hemolytic-uremic syndrome. J In- PC, Chiang T. 2006. A review of an emerging
fect Dis 189:552555. enteric pathogen: enteroaggregative Escherichia
56. Centers for Disease Control and Prevention. coli. J Med Microbiol 55:13031311.
1982. Isolation of E. coli O157:H7 from sporadic 66. Estrada-Garcia T, Navarro-Garcia F. 2012.
cases of hemorrhagic colitisUnited States. Enteroaggregative Escherichia coli pathotype: a
MMWR Morb Mortal Wkly Rep 31:580, 585. genetically heterogeneous emerging foodborne
57. Riley LW. 1987. The epidemiologic, clinical, and enteropathogen. FEMS Immunol Med Microbiol
microbiologic features of hemorrhagic colitis. 66:281298.
Annu Rev Microbiol 41:383407. 67. Karmali MA, Mascarenhas M, Shen S, Ziebell
58. Griffin PM. 1995. Escherichia coli O157:H7 and K, Johnson S, Reid-Smith R, Isaac-Renton J,
other enterohemorrhagic Escherichia coli, p 739 Clark C, Rahn K, Kaper JB. 2003. Association
761. In Blaser MJ, Smith PD, Ravdin JI, Greenberg of genomic O island 122 of Escherichia coli EDL
HB, Guerrant RL (ed), Infections of the Gastroin- 933 with verocytotoxin-producing Escherichia coli
testinal Tract. Raven Press, New York, NY. seropathotypes that are linked to epidemic and/or
59. Feng P, Lampel KA, Karch H, Whittam TS. serious disease. J Clin Microbiol 41:49304940.
1998. Genotypic and phenotypic changes in the 68. Tauxe RV. 1998. Public health perspective on
emergence of Escherichia coli O157:H7. J Infect immunoprophylactic strategies for Escherichia
Dis 177:17501753. coli O157:H7: who or what would we immunize,
60. Whittam TS. 1998. Evolution of Escherichia coli p 445452. In Kaper JB, OBrien AD (ed), Esch-
O157:H7 and other Shiga toxin-producing E. coli erichia coli O157:H7 and Other Shiga Toxin-
strains, p 195209. In Kaper JB, OBrien AD (ed), Producing E. coli Strains. ASM Press, Washington,
Escherichia coli O157:H7 and Other Shiga Toxin- DC.

MICROBIOLOGY

Taxonomy Meets Public Health:
The Case of Shiga Toxin-Producing
Escherichia coli
FLEMMING SCHEUTZ1
2
PATHOTYPES AND TAXONOMY
The term enteropathogenic Escherichia coli was originally used to refer to

strains belonging to a limited number of O groups epidemiologically associated
with infantile diarrhea (1). Subsequently, E. coli strains isolated from intestinal
diseases have been grouped into at least six main categories on the basis of
epidemiological evidence, phenotypic traits, clinical features of the disease they
produce, and specific virulence factors. The well-described intestinal patho-
types or categories of diarrheagenic E. coli groups are enteropathogenic E. coli
(EPEC), Shiga toxin-producing E. coli (STEC) or verocytotoxin-producing
E. coli (VTEC) (including enterohemorrhagic E. coli [EHEC]), enterotoxigenic
E. coli (ETEC), enteroaggregative E. coli (EAEC), enteroinvasive E. coli, and
diffusely adherent E. coli. The general definition of an E. coli pathotype as
a group of strains of a single species that cause a common disease using a
common set of virulence factors (2) has been further refined for STEC to help
assess the clinical and public health risks associated with different STEC strains
(3). An empirical classification scheme was used to classify STEC serotypes
into five seropathotypes (A through E) according to the reported association
of serotypes with human intestinal disease, outbreaks, and hemolytic-uremic
1
WHO Collaborating Centre for Reference and Research on Escherichia and Klebsiella, Department of Microbiology and
Infection Control, Statens Serum Institut, DK-2300 Copenhagen S, Denmark.
Enterohemorrhagic Escherichia coli and Other Shiga ToxinProducing E. coli
doi:10.1128/microbiolspec.EHEC-0019-2013
17

18 SCHEUTZ
syndrome (HUS) (3). This classification sys- and Stx2 or VT2 (first described as Shiga-like
tem uses a gradient ranging from seropatho- toxin II, SLTII). The cytotoxin production
type A (high risk) to seropathotypes D and is usually bacteriophage-mediated (912), and
E (minimal risk). This definition has been of the diversity of this toxin family has since
considerable value in cases of human infection become clear.
but is also problematic because the majority The public health significance of STEC was
of isolates from STEC infections are not fully first recognized in 1982, when two outbreaks
characterized and coupled to reliable clinical in the United States affected at least 47 people
information. Although the definition of HUS in Oregon and Michigan. Nine of 12 stool cul-
is distinct, the spectrum of diarrheal disease tures yielded a rare E. coli serotype, O157:H7,
varies considerably and may include a range that was also isolated from a beef patty from a
of symptoms from nonbloody to scanty blood suspected lot of meat in Michigan (13). This
to true hemorrhagic colitis. Additionally, the strain was designated EDL933 and has since
use of A through E adds confusion because been used as the prototype STEC strain by
Shiga toxin subtypes are also named alpha- researchers worldwide. Distinct clinical fea-
betically. Most importantly, the concept of tures of hemorrhagic colitis (HC) included
(sero)-pathotypes collides with the require- abdominal cramps, copious bloody diarrhea
ments of a good taxonomy, which separates described as all blood and no stool, unac-
elements of each group into subgroups that companied by fecal leukocytes, and no fever.
are mutually exclusive, unambiguous, and, Duration of illness was 2 to 9 days, and there
together, include all possibilities. In practice, were no deaths, complications, or sequelae in
a good taxonomy should be simple to apply, any of the cases.
easy to remember, and easy to use. The need In an outbreak of HC in November 1982
to define human pathogenic STEC and to iden- at a Canadian institution for elderly patients,
tify factors of STEC that absolutely predict the sorbitol-negative E. coli O157:H7 was shown
potential to cause human disease is obvious to produce verocytotoxin. Two of six sporadic
in terms of clinical management, supportive cytotoxic O157:H7 strains were associated
or antibiotic treatment, quarantine measure- with HC, and 70% of 78 cytotoxic serotypes
ments, risk assessment, surveillance, and out- isolated from sporadic cases of diarrhea dur-
break investigations and management. This ing 1978 to 1982 were E. coli O26:H11 (14). The
chapter presents a brief history of the concept cytotoxicity of an O26 strain (H30, described
of pathotypes and describes the possible alter- as a verocytotoxin producer by Konowalchuk
natives for categorizing STEC based on phe- et al. [4]), two of the E. coli O157:H7 strains
notypic or molecular typing. from the U.S. cases of HC, and the beef patty
isolate EDL933 from this outbreak could be
neutralized by rabbit antiserum to purified
PATHOTYPES Shiga toxin from S. dysenteriae Type 1, sub-
stantiating the premise that these cytotoxins
First discovered in 1977 (4), verocytotoxin were the same and that they played an im-
was found to be biologically and structurally portant role in E. coli diarrheal diseases (15).
similar to Shiga toxin produced by Shigella In 1983, 11 of 15 sporadic cases of enteropathic
dysenteriae Type 1 (5, 6). It was soon realized HUS were shown to have evidence of in-
that antigenically distinct cytotoxins could be fection with VTEC, indicating an association
found in different E. coli serotypes (7, 8). between sporadic cases of HUS and cytotoxin-
STEC or VTEC strains are characterized by producing E. coli strains (16). Verocytotoxin
their ability to produce either one or both of was proposed as having direct etiological im-
these cytotoxins, referred to as Stx1 or VT1 portance in the pathogenesis of both HUS and
(first described as Shiga-like toxin I, SLTI) HC (17).

CHAPTER 2 Taxonomy Meets Public Health: STEC 19
The term EHEC, coined to refer to strains composed of 12 serotypes that have been im-
such as O157:H7 that manifest the above- plicated in sporadic cases of diarrhea but not
mentioned clinical, epidemiological, and with outbreaks or HUS, and seropathotype E
pathogenic features (18), further defined a is composed of at least 14 animal serotypes
pathogenic subgroup of STEC strains based that have not been implicated in disease in
on their association with disease in humans humans (3).
and their ability to hybridize to a DNA probe This approach has been of considerable
(CVD419) derived from a large plasmid pres- value in defining pathogenic STEC serotypes
ent in most O157:H7 strains and the majority of importance in cases of human infection
of other STEC strains isolated from cases and also for STEC isolates from ruminants.
of HC (19). Many of the other STEC strains However, classification of strains based on
belonged to classical EPEC O groups or sero- the criteria above is also problematic because
types such as O26, O111, O114, O125, O126, and the majority of isolates from STEC infections
O128 (4, 5, 12, 1921), but non-EPEC O groups are not fully serotyped nor characterized for
were also found to produce Stx: O1, O2, O4, the presence of virulence factors. A recent
O5, O6, O18, O45, O50, O68, O91, O103, O113, Belgian study of STEC added 14 serotypes
O121, and O145 (4, 5, 19). Furthermore, Stx to seropathotype C (including four serotypes
production was also described in O groups associated with HUS) and 54 serotypes to
O138, O139, and O141 isolated from weaned seropathotype D, demonstrating how versa-
pigs with edema disease (12). Among the more tile the definition of seropathotypes can be
than 472 STEC/VTEC serotypes, and apart (24). Outbreaks with emerging or new hybrid
from O157:H and O157:H7, those in O groups strains with hitherto unknown virulence fac-
O26, O103, O111, and O145 are most commonly tors may also continuously challenge our un-
isolated from humans worldwide (22). They, derstanding and appreciation of virulence
along with strains that have caused outbreaks, potential. The limitation to relevant sero-
are clearly recognized as pathogens. types may therefore result in the omission or
Consequently, serotypes and their associa- incorrect classification of specific pathotypes.
tion with diseases of varying severity in hu- Strain O104:H4 is such an example, because
mans and with sporadic disease or outbreaks until the May-June 2011 outbreak in Germany
have been used to classify STEC into five this highly virulent strain would have been
seropathotypes (A through E) according to classified as seropathotype D on the basis of
the reported occurrence of serotypes in hu- its sporadic occurrence, its lack of association
man disease, in outbreaks, and in HUS (3). with outbreaks, and its limited association
Seropathotype A consists of O157:H7 and with HUS (25). In fact, other STEC serotypes
O157:NM, considered to be the most virulent. such as O104:H21 and O113:H21 strains lacking
Seropathotype B originally consisted of five the eae gene were responsible for an outbreak
serotypes that were similar to seropathotype and a cluster of three HUS cases in the United
A in causing severe disease and outbreaks but States and Australia, respectively (26, 27).
occurred at lower frequency, but in the United It has been suggested that these unusual
States, seropathotype B has been extended and emerging types should have their own se-
to include 13 STEC serotypes: O26:H11 and ropathotype designation (28). Using data from
NM; O45:H2 and NM; O103:H2, H11, H25, and the European Surveillance System (TESSy
NM; O111:H8 and NM; O121:H19 and H7; data) as provided by the European Centre
and O145:NM (23). Seropathotype C includes for Disease Prevention and Control (ECDC)
serotypes infrequently implicated in sporadic and data available in the European Union
HUS but not typically with outbreaks and in- Summary Report on Trends and Sources of
cludes O5:NM, O91:H21, O104:H21, O113:H21, Zoonoses, Zoonotic Agents and Food-borne
O121:NM, and O165:H25. Seropathotype D is Outbreaks in 2011 (29), the European Food

20 SCHEUTZ
Safety Authority Panel on Biological Hazards or analysis of the public health surveillance
(BIOHAZ Panel) concluded that the sero- data (25). In addition, even though the clin-
pathotype classification by Karmali et al. (3) ical manifestations of non-O157 STEC in-
does not define pathogenic VTEC nor does it fection may differ considerably from those of
provide an exhaustive list of pathogenic sero- O157:H7 (31), STEC O157:H7 has also been
types. Eighty-five percent of 13,545 confirmed isolated from stools of healthy individuals.
cases of human VTEC infection from 2007 to Many studies have tried to correlate the pre-
2010 were not fully serotyped and could not sence of specific virulence factors with dis-
be classified by using the seropathotype con- ease or severity of disease (see discussion in
cept, and about 27% of the cases could not reference 32). The combination of the locus
be assigned to a seropathotype group as they of enterocyte effacement (LEE)-encoded eae
were not listed by Karmali et al. (25). How- gene for intimin and stx2 is significantly more
ever, about half of the isolates with missing frequent in isolates from serotypes found in
H type were from cases of O157 infection humans and is most strongly associated with
(5,610 cases) and would most likely have been disease in humans, particularly with severe
typed as O157:H7 or O157:H and therefore disease (3234). The reverse is true for stx1,
assigned to seropathotype A. This would have which is found more frequently in serotypes
expanded this group to 6,657 cases, or 87% not found in humans (32). Enterohemolysin,
of cases, 78% of fatal cases, 91% of the hos- a plasmid-encoded toxin expressed by the
pitalizations, 91% of the HUS cases, and 95% ehxA gene that readily causes the hemolysis
of the cases with bloody diarrhea (25). of washed sheep erythrocytes and liberates
Reporting of detailed clinical data is often hemoglobin from the red blood cells during
incomplete. Of the reported confirmed VTEC infection, has been linked to severe disease
cases in the European Union between 2007 symptoms (35, 36).
and 2010, the health outcome was reported The taxonomy of EPEC and VTEC is inti-
for 53% of diarrheal cases and 59% of HUS mately intertwined in that many VTEC types
cases (25). Most patients (ca. 64%) presented share specific virulence factors such as path-
with only diarrhea. Clinical information is not ogenicity islands, e.g., LEE (including the eae
obtained according to standardized guidelines gene), O island (OI) 122, and plasmids with
and definitions, and detailed information on EPEC. Non-LEE (nle)-encoded genes have
the individual clinical course of disease is gen- also been found in both EPEC and STEC iso-
erally absent. Although the definition of HUS lates, such as the Espl/NleA effector protein
is distinct, the spectrum of diarrheal disease encoded on a prophage (37). Several studies
varies considerably and may include symp- have indicated that the major difference be-
toms ranging from nonbloody to scanty blood tween certain EPEC and STEC isolates is
to true hemorrhagic colitis. A specific STEC the absence or presence of the bacteriophage
pathotype may even be associated with clini- encoding Stx. However, the presence of LEE
cal presentations ranging from asymptomatic does not seem to be essential for full viru-
carriage to life-threatening HUS and death, lence, as a wide number of LEE-negative
as observed during outbreaks and in person- STEC strains have been associated with spo-
to-person transmission, whereby an index radic cases and small outbreaks of HC and
patient may experience only mild symptoms HUS (38).
whereas secondary cases may develop into se- How are environmental, food, and veteri-
vere complications (30). nary STEC types, which by nature cannot be
The BIOHAZ Panel concluded that isolated from human cases of either HC or
pathogenicity can neither be excluded nor HUS, classified? Animals often carry types
confirmed for a given STEC serogroup or se- that are referred to as EHEC or could be clas-
rotype based on the seropathotype concept sified in the A through E pathotype scheme

without any clinical symptoms. The classifi- and less well resolved (39). E. coli strains are
cation based on the clinical course of disease assigned by MLST to different sequence types
in humans is clearly host associated, and the (STs), and within each ST diverse clusters
term EHEC or the A through E classification can be observed by PFGE. Three distinct
of nonhuman isolates could be misleading. sequence-based methods targeting house-
keeping genes exist for E. coli. As of October
2013, Institut Pasteurs MLST scheme using
PHYLOGENY AND DISEASE dinB, icdA, pabB, polB, putP, trpA, trpB, and
uidA genes lists 599 unique STs (http://www.
An E. coli genome contains between 4,200 pasteur.fr/recherche/genopole/PF8/mlst/
and 5,500 genes, with fewer than 2,000 genes EColi.html). Mark Achtmans MLST scheme
conserved among all strains of the species using adk, fumC, gyrB, icd, mdh, purA, and
(the core genome). Comparison of 61 E. coli recA lists 5,873 isolates belonging to 3,874 STs,
and Shigella spp. sequenced genomes has of which 1,485 are found in 54 ST complexes
shown that the genetic repertoire or pan- and 4,388 isolates belonging to 3,562 unique
genome comprises 15,741 gene families and STs have not been assigned an ST complex
that only 993 (6%) of the families are rep- (mlst.warwick.ac.uk). T. Whittams MLST
resented in every genome (39). The variable scheme, the EcMLST, lists allele sequences
or accessory genes thus make up more than and allele profiles for 679 E. coli strains and
90% of the pan-genome and about 80% of a uses different combinations of 15 housekeep-
typical genome; some of these variable genes ing genes, i.e., internal fragments of the seven
tend to be collocated on genomic islands. aspC, clpX, fadD, icdA, lysP, mdh, and uidA
Continuous gene flux has occurred during genes where the associated sequence type
E. coli divergence, mainly as a result of hori- is defined as st7 or further characterized by
zontal gene transfers and deletions. This ge- using internal fragments of the eight addi-
netic plasticity accelerates the adaptation of tional arcA, aroE, cyaA, dnaG, grpE, mtlD,
E. coli to varied environments and lifestyles, mutS, and rpoS genes where the associated
as it allows multiple gene combinations that sequence type is defined as st15. Some isolates
result in phenotypic diversification and the are only characterized by using internal frag-
emergence of new hypervirulent strains such ments of two of the seven housekeeping
as the hybrid STEC and EAggEC O104:H4-B1- genes, mdh and uidA, where the associated se-
ST678 strain that combine resistance and vir- quence type is defined as st2 (www.shigatox.
ulence genes, which in classical pathogenic net). A comparison study is ongoing to estab-
E. coli strains traditionally have been mutually lish correspondence between the two former
exclusive (40). More than describing STEC MLST schemes.
stains as a separate group of E. coli, STEC There is increasing evidence that within
represents many, if not the majority, of phy- serotype O157:[H7] there are differences in
logenetic lineages of E. coli in general. The the clinical outcome and association with
sequence-based method targeting housekeep- HUS. Molecular methods have identified dif-
ing genes (using distinct sets of genes), multi- ferent genetic lineages of E. coli O157:H7.
locus-sequence typing (MLST), generally has Octamer-based genome scanning and mi-
less discriminatory power than pulsed-field croarray comparative genomic hybridization
gel electrophoresis (PFGE) but has been con- were first to identify three lineages designated
sidered the most reliable method to determine I, II, and I/II (4143). Nucleotide polymor-
the genetic relatedness of epidemiologically phism-derived genotyping and phylogenetic
unrelated isolates. However, when in silico analyses identified eight major STEC O157
MLST is performed on whole-genome se- lineages (44). Seven lineages are typically
quences, many of the strains appear jumbled found in cattle, including one that does not

22 SCHEUTZ
associate with human disease and may be O groups (O26, O111, O118, etc.). These two
evolving away from human virulence and clonal groups differ in their virulence and
two other lineages accounting for a majority global distribution. Although several fully an-
of human disease (44). The 30 sorbitol- notated genomic sequences exist for strains of
fermenting O157 strains belong to a lineage serotype O157:H7, much less is known about
VIII exclusively isolated from humans (44). the genomic composition of EHEC 2. Analy-
In a study of 528 O157 strains primarily from sis of 24 clinical EHEC 2 strains representing
Michigan patients but also including strains serotypes O26:H11, O111:H8/H11, O118:H16,
from Argentina, Australia, Canada, Germany, O153:H11, and O15:H11 from humans and ani-
Japan, and the United Kingdom from 1982 mals by comparative genomic hybridization
to 2006, Manning et al. identified 39 single supports the hypothesis that extensive mod-
nucleotide polymorphism (SNP) genotypes ular shuffling of mobile DNA elements has
that differed at 20% of SNP loci and sepa- occurred among STEC strains, and the gene
rated them into nine distinct clades (45). content variation of phage-related genes in
The outbreak strain TW14359, implicated in a EHEC 2 seems to indicate that EHEC 2 is a
multistate outbreak associated with the con- multiform pathogenic clonal complex, char-
sumption of bagged spinach in North America acterized by substantial intraserotype genetic
in 2006 (46, 47), was shown to be a member variation. The heterogeneous distribution of
of clade 8, which was significantly associated mobile elements is especially seen in O26:H11
with younger age (0 to 18 years) and patients more than in other EHEC 2 serotypes (51).
with HUS, who were seven times more likely Comparative analysis of whole-genome phy-
to be infected with clade 8 strains than pa- logeny and of type III secretion system ef-
tients with strains from clades 1 to 7 combined fectors of 114 LEE+ E. coli isolates shows that
(45). The study revealed substantial genomic attaching and effacing E. coli is divided into
differences between clades, suggesting that five distinct genomic lineages and that the
an emergent subpopulation of the clade 8 lin- LEE+/stx+/bfp genomes are primarily di-
eage has acquired critical factors that con- vided into two genomic lineages, the O157/
tribute to more severe disease. Comparison of O55 EHEC1 and non-O157 EHEC2 (52). Most
the phylogenetically divergent O157:H7 out- importantly, phylogenic relatedness was in-
break strains TW14359 and RIMD0509952, dependent of the presence or absence of
which caused the largest O157:H7 outbreak to stx-encoding phages, highlighting the close
date in Sakai, Japan, showed that these two relation between LEE+ EPEC and STEC lin-
strains vary in their ability to colonize or ini- eages (52). In this study of 138 whole genomes
tiate the disease process (48). Interestingly, of which 114 were LEE+, stx genes were only
most LEE genes, the stx2 genes, and several found in phylogroups B1, including EHEC2,
pO157-encoded genes that promote adher- EPEC2, and unclassified attaching and effac-
ence, including type II secretion genes and ing E. coli, and in E, represented by O157
their effectors stcE and adfO, are upregulated EHEC1 (52). Even less is known about non-
in the spinach outbreak strain, whereas fla- LEE STEC, but whole-genome comparative
gellar and chemotaxis genes are primarily up- analysis of nine non-LEE genomes revealed
regulated in the Sakai strain (48). that phage-encoded genes, including non-
Evolutionary analyses of STEC by multi- LEE-encoded effectors, were absent from
locus enzyme electrophoresis (49) and par- all nine STEC genomes. Several plasmid-
tial sequencing of 13 housekeeping genes encoded virulence factors reportedly identi-
(50) have identified two distantly related fied in LEE-negative STEC isolates were
clonal groups classified as EHEC 1, including identified in only a subset of the nine LEE-
serotype O157:H7 and its inferred ancestor negative isolates, further confirming the di-
O55:H7, and EHEC 2, represented by several versity of this group. Characterization of the

lambdoid stx-encoding phages showed that

although the integrase gene sequence cor-
responded with genomic location, it was not
correlated with stx subtype, highlighting the
mosaic nature of these phages. A wide range
of basal and induced expression of the Shiga
toxin genes, stx1 and stx2, and the Q genes was
observed (53).
TOXIN AND DISEASE
The Shiga toxin family is divided into two

branches, Stx1 (almost identical to Stx from
S. dysenteriae Type 1) and Stx2. Subtypes, de-
noted by Arabic letters that follow the main FIGURE 1 Stx subtypes and variants. Parsimony tree
type name, may exhibit significant differences of 107 variants: nine variants of Stx1a (including
Shiga toxin from S. dysenteriae), four variants of
in biologic activity, including serologic re- Stx1c, one variant of Stx1d, and subtypes of Stx2,
activity, receptor binding, and the capacity to including 21 Stx2a, 16 Stx2b, 18 Stx2c, 18 Stx2d,
be activated by elastase in intestinal mucus. 14 Stx2e, two Stx2f, and four Stx2g variants. Data
Variants have been defined by relatedness of from reference 54 and updated by the author.
sequence within a subtype that differs by one doi:10.1128/microbiolspec.EHEC-0019-2013.f1
or more AAs from the prototype (54). The
phylogenetic relationship of variants has been study of 922 patients with STEC infection
analyzed (5460), but not all variants have found that 81 of 107 patients (76%) with
been examined for all classical phenotypic dif- sorbitol-fermenting O157 had HUS (54). This
ferences, biologic activity, and hybridization particular STEC strain is typically positive for
properties. The variants are designated by both stx2a and eae. Sporadic adult cases of
toxin subtype, O group if the host strain is HUS in France were seen in patients infected
E. coli and generic name of the host bacterium with six different serotypes, all vtx2 (vtx1 and
if the host strain is not E. coli, followed by the eae negative) (67). In Australia, a VTEC O113:
strain name or number from which that toxin H21 strain lacking eae was responsible for a
was described (54). At present, 107 variants cluster of three cases of HUS (27). In Finland,
have been identified: subtypes of Stx1 include O174:H21 (68) and O:rough:K-:H49 (69), both
9 variants of Stx1a (including Shiga toxin from eae negative and vtx2 positive, were isolated
S. dysenteriae), 4 of Stx1c, and 1 of Stx1d, and from two separate cases of HUS. In Germany,
subtypes of Stx2 include 21 variants of Stx2a, a large outbreak in 2011 of a hybrid STEC-
16 of Stx2b, 18 of Stx2c, 18 of Stx2d, 14 of EAggEC strain O104:H4, stx2a and eae nega-
Stx2e, 2 of Stx2f, and 4 of Stx2g (Fig. 1). tive but positive for many EAggEC-associated
Stx subtypes, and maybe specific variants, genes, affected 3,167 patients without HUS (16
are clinically relevant. Stx2a (with or with- deaths) and 908 with HUS (34 deaths) (70).
out Stx2c) seems to be highly associated with In Germany, an association between the
HUS (30, 6164). The combination of eae and activatable subtype vtx2d in eae-negative STEC
stx2 especially has been associated with the strains and HUS has been found (61). The
development of HUS and bloody diarrhea (34, median age of 21 years in patients with stx2d
62, 63, 65, 66), and an unspecified synergism was considerably higher than in other patients
between the adhesin intimin encoded by eae with STEC. Stx2 subtypes Stx2a and Stx2d
and Stx2 has been suggested (32). A German studied in vitro with Vero monkey kidney

24 SCHEUTZ
cells and primary human renal proximal tu- of eae, stx1a, and stx2a, HUS developed in
bule epithelial cells were at least 25 times 62% of the patients. HUS had been previously
more potent than Stx2b and Stx2c. The in vivo found in two of six (33%) patients infected with
potency of Stx2b and Stx2c in mice was sim- STEC. Infections with STEC O157 containing
ilar to that of Stx1, whereas Stx2a and Stx2d the vtx2a toxin profile were associated with
were 40 to 400 times more potent than Stx1 a higher number of HUS cases (24 to 30%),
(71). It has been suggested that disease out- whereas HUS developed in only 1 of 31 (3%)
breaks select for producers of high levels of patients infected with O157:H7 eae + stx2c-
Stx2a among E. coli O157:H7 strains shed by positive strains and did not develop in 93
animals and that Stx1 expression is unlikely to patients with eae + stx1a with (85 patients) or
be significant in human outbreaks (72). Nearly without (8 patients) stx2c (76). The association
all lineage I strains carry stx2a, whereas all between stx2a and severity of disease has also
lineage II strains carry stx2c, and 4 of 14 lin- been demonstrated in two animal models for
eage I/II strains have copies of both stx2a and clinical genotype (CG) strains (carrying stx2a
stx2c (73). Real-time PCR and enzyme-linked with or without Stx2c), which induced more
immunosorbent assay have demonstrated that severe clinical symptoms, earlier and higher
lineage I and I/II strains produce significantly mortality, and more severe histopathologic
more stx2a mRNA and Stx2a than lineage II lesions compared to bovine-biased genotype
strains. However, among lineage I strains sig- (BBG) strains (carrying stx2c only) (77). Puri-
nificantly more Stx2a is also produced by fied Stx2a has also been shown to be more
strains from humans than from cattle. There- potent than Stx2c against primary human
fore, lineage-associated differences among kidney cell lines and in mouse models (71). It is
E. coli O157:H7 strains, such as prophage con- possible that carriage and expression of Stx2a
tent, toxin type, and toxin expression, may alone are sufficient to confer increased viru-
contribute to host isolation bias. However, the lence in animal models and increased expres-
level of Stx2 production alone may also play sion of human disease, but alternatively, these
an important role in the within-lineage asso- phenotypes may result in whole or in part in
ciation of O157:H7 strains with human clini- other genetic factors that are correlated with
cal disease. Indeed, clade 8 strains associated Stx2a.
with HUS overexpress Stx2a when compared Stx2c has occasionally been associated with
to strains from clades 1 to 3, and SNPs, which HUS but with a significantly lower risk (62,
may affect Stx2a expression and could be use- 64). The majority of 210 patients in Japan
ful in the genetic differentiation of highly vir- infected with Stx2c strains presented no or
ulent strains, have been described (74). mild symptoms, except for 3 patients with
Analyses of the 2006 outbreak of O157:H7 bloody diarrhea (78). Also in Japan, 169 strains
clade 8 strains in spinach suggested the pre- carrying only stx2vha (now stx2c) (54) were
sence of stx2a and an stx2c variant (75), but probably less virulent and caused bloody diar-
not all clade 8 strains have both stx2a and rhea less frequently (79). In an Austrian study
stx2c, and none of the strains has only stx2c. of 201 STEC strains collected from patients
The presence and presumable production of and environmental sources, the stx2a and stx2c
the Stx2c variant alone cannot be solely re- alleles were associated with high virulence and
sponsible for the enhanced virulence attrib- the ability to cause HUS, whereas stx2d, stx2e,
uted to the clade 8 lineage. This also is true stx1a, and stx1c occurred in milder or asymp-
for the production of Stx2a, because it was tomatic infections (80). However, caution is
detected in nearly every strain representing warranted for infections by Stx2c O157 strains,
all nine clades (45). in addition to Stx2a O157 strains.
In a Danish outbreak with an O157:H7 strain Other subtypes or variants of Stx1 and Stx2
with a toxin gene subtype profile consisting are primarily associated with a milder course

of disease (6163). Except for the few indi- or a double infection with O157:H7 (eae + vtx1
vidual, unusual cases mentioned below, HUS + vtx2). HUS also developed in one patient
did not develop in a total of 825 Danish pa- infected with O55:H12 vtx1 only who had re-
tients as follows: 426 with vtx1a STEC (313 eae ceived six different antibiotics during the
positive and 113 eae negative), 65 with stx1c acute phase and in one patient with O13,O73:
(5 eae positive and 60 eae negative), 13 stx1d K1:H18 vtx2d (unpublished data). The asso-
(all eae negative), 87 with stx2b (1 eae positive ciation of specific virulence factors with
and 86 eae negative), 12 with stx2e (1 eae HUS should be carefully examined for under-
positive and 11 eae negative), 37 with stx2f lying disease, epidemiologic relation, and an-
(all eae positive), 5 with stx2g (eae negative), tibiotic treatment during the acute phase of
and 180 with seven different combinations of illness.
stx1 and stx2 subtypes, excluding stx2a and STEC producing Stx2e is closely associ-
stx2d (70 eae positive and 110 eae negative) ated with edema disease in swine, and Stx2e-
(unpublished Danish surveillance data). In producing STEC strains are probably not
France, sporadic cases of STEC infection com- pathogens for humans (83). However, Stx2e-
plicated with HUS have been described in producing E. coli strains belonging to O
patients infected with a clone of O103:H2, vir- groups O101 and O9 have sporadically been
ulence type vtx1 and eae (81). Among infec- isolated from patients with diarrhea and
tions with VTEC strains with only the vtx1 HUS. These O groups in STEC strains are not
gene, one case of HUS has been described; a usually associated with edema disease (84).
strain with the serotype O115:H18 (82) and Thus, STEC without both stx2 and eae is,
8 of 169 non-O157 HUS-associated strains rep- with a few exceptions, only sporadically as-
resenting 42 different ST types have been sub- sociated with HUS, and toxin subtyping can
typed as vtx1-only strains (54). Data from the be useful in identifying high-risk or HUSEC
European Surveillance System (TESSy data), strains.
as provided by the ECDC on virulence charac-
teristics of reported confirmed VTEC cases
in 20072010, including all cases, hospitalized Stx-ASSOCIATED
cases only, and HUS cases only, show that BACTERIOPHAGE INSERTION
most HUS cases (89.2%), for which informa-
tion was reported on virulence factors, were In E. coli, the capacity to produce Shiga toxin
either eae, vtx2 positive or eae, vtx1, vtx2 is encoded by genes on bacteriophages, and it
positive and that an additional 5.9% were has been suggested that rather than describing
either vtx2 positive or vtx1, vtx2 positive but the taxonomy of STEC bacteria, the biology
without the eae gene. Only 2.3% were positive and taxonomy of the bacteriophages that are
for stx1 (1.6% eae, stx1 and 0.7% stx1) (25). hosted by certain lineages of E. coli could
None of the 124 reported infections due to serve as the basis for refining this pathogenic
O103:H2, of which all but one were eae, vtx1 group.
positive, caused HUS (25). Surveillance data Several common stx phage insertion sites
rarely include detailed clinical descriptions have been reported in LEE-positive STEC
of individual HUS cases. Follow-up on clinical genomes. These insertion sites include wrbA,
data on Danish patients with unusual or rare yecE, torS/T, sbcB, yehV, argW, ssrA, and prfC
association between HUS and the virulence (85, 86). In LEE-negative strains, additional
profile eae, vtx1 revealed possible complicat- insertion sites are often different and include
ing factors such as antibiotics during the acute patC, yciD, ynfH, serU, mlrA, and yjbM (53)
infection, underlying nephrotic syndrome, (Fig. 2). However, a number of sites have not
association with outbreak cases infected with been determined for some of the most com-
an HUS-associated outbreak O157:H7 strain, mon serotypes like O157, O26, O111, and O103

26 SCHEUTZ
FIGURE 2 Stx bacteriophage insertion sites in LEE-positive STEC include wrbA, yecE, torS/T, sbcB, yehV, argW,
ssrA, and prfC. Data from references 85, 86, and 118. In LEE-negative STEC genomes additional insertion
sites are often different and include patC, yciD, ynfH, serU, mlrA, and yjbM. Data from references 53, 87, and
119. Big circles indicate the preferred bacteriophage integration site.
doi:10.1128/microbiolspec.EHEC-0019-2013.f2
(85). Insertion site occupancy by stx phages 419 STEC O157 strains from geographically and
depends on the host strain and on the avail- temporally diverse cattle- and human-origin
ability of the preferred locus in the host strain. isolates, the presence of Stx2a-associated bac-
For the most part, yehV is occupied by stx1 teriophage sequences detected adjacent to
encoding phages in LEE-positive O157 strains either wrbA or argW and the detection of stx2a
whereas wrbA or argW is preferentially se- were significantly associated with CG com-
lected by the stx2a phages and sbcB by stx2c pared to BBG strains, of which many (42.9%)
phages (87). Phages preferentially use one in- had Stx2a-associated bacteriophage sequences
sertion site, but if this primary insertion site adjacent to wrbA or argW but lacked detectable
is unavailable, then a secondary insertion site stx2a. All 281 CG strains had Stx2a-associated
is selected (88, 89). Molecular epidemiologi- bacteriophage sequences inserted in wrbA or
cal studies using Stx-associated bacteriophage argW and carried stx2a. In contrast, the pre-
insertion sites for strain differentiation have sence of Stx2c-associated bacteriophage se-
shown that specific bacteriophage-associated quences inserted in sbcB and the presence of
genetic factors underlie the differential viru- stx2c were significantly more common in BBG
lence of CG and BBG, where genotypes that compared to CG isolates. All 107 BBG isolates
are significantly overrepresented among hu- had both traits, whereas only 11.8% of CG
man isolates, as compared to cattle isolates, isolates belonged to genotypes with Stx2c-
are referred as CG and genotypes among associated bacteriophage sequences inserted
cattle isolates that show significantly over- in sbcB, and only 7.1% of CG isolates carried
representation or similar representation, as stx2c (87). Deep sequencing of pooled STEC
compared to human isolates, are referred as O157 DNAs from human clinical cases (n = 91)
BBG. STEC O157 can be further classified into and cattle (n = 102) identified 42 genotypes
several genotypes (90). In a study of a panel of that could be tagged by a minimal set of 32

polymorphisms. Phylogenetic trees of these found in phage 21, which does not contain any
genotypes are also divided into clades that stx genes (99), but phage 2851 encoding vtx2c
represent strains of cattle origin, or cattle and contains a Q gene with 96.9% identity to Q21
human origin (91), thus confirming that certain and has frequently been detected in Stx2c-
O157 lineages are more associated with the producing O157 strains, indicating that phages
bovine reservoir and do not turn up in human related to 2851 are associated with Stx2c pro-
disease. duction in O157 strains from different loca-
The lineage-associated differences among tions and time periods (100), although it can be
STEC O157:H7 strains such as prophage confound in other Stx subtypes (101). The QO111:H
tent, toxin type, and toxin expression may was found in a Japanese O111:H isolate, 11128,
contribute to the observed host isolation bias. in a study of O157:H clinical isolates related to
However, the level of Stx2 production alone HUS (86). Other Q genes found in Stx1- and
may also play an important role in the within- Stx2a-producing phages show high identity to
lineage association of STEC O157:H7 strains the Q933, i.e., the Stx1 phage in H19-B, which
with human clinical disease (73). shows 96.5% identity to Q933 (98); the Stx2a-
producing phage, VT2-Sakai, which is identi-
cal to Q933 (102); the Stx1-producing phage,
TOXIN EXPRESSION AND Q GENES VT1-Sakai, which shows 97.4% identity with
Q933 (103); and the Stx1-producing phage in
The biology of the Stx-encoding phages con- strain Morioka, which shows 97% identity to
tributes greatly to the production of Stx, and Q933 (104). Using the above-mentioned Q gene
many research results during the past decade sequences as references, Jacobsen (105) re-
have contradicted the prevailing assumption cently queried 287 E. coli genome sequences of
that phages serve merely as agents for viru- varying quality: Two genome sequences were
lence gene transfer. from Los Alamos National Laboratory and
In a majority of STEC strains, expression of 285 genome sequences were publicly available
stx genes within lambdoid phages is believed online (106 from three sequencing projects:
to be largely under the control of the late Escherichia group, E. coli O104:H4, and E. coli
phage promoter, pR, and the Q antitermina- antibiotic resistance) from the Broad Insti-
tor protein (92). In STEC lysogenic for Stx- tute of Harvard and Massachusetts Institute
converting bacteriophages, expression of the of Technology (www.broadinstitute.org/) and
stx genes is usually repressed and production 179 from National Center for Biotechnology
of Stx is preceded by prophage induction (93, Information (www.ncbi.nlm.nih.gov/). Infor-
94). Variations in the Q gene have been pro- mation about serotype was available for 131
posed to influence the quantitative expres- (45.6%) strains and resulted in data on 57 dif-
sion of Stx (95). The Q gene transcription is ferent serotypes with a bias toward O157:H7
increased under inducing conditions allowing and O104:H4. Jacobsen identified 474 anti-
for increased transcription of the stx genes terminator Q genes in 235 of 287 genome se-
that are downstream of the Q-binding site (92). quences, of which 80 Q genes were found
Inducing factors include nutritional stress, oxi- upstream of the stx operon in 62 genomes of
dative stress, UV radiation, antibiotics (mito- at least 14 different serotypes. On the basis of
mycin C, quinolones, -lactams, etc.), EDTA their amino acid sequence, these 80 Q genes
and other chelating agents, hydrogen peroxide, could be classified into five main groups (I to
heat shock, and quorum sensing (96), but Stx1 V) and further into 10 variants whereby vari-
production can also be induced through low- ants within the main groups are distinguished
iron induction of pStx1 (97). Different Q variants by one amino acid difference. Thirteen Q genes
have been described. Q933 was found in the are classified as variant QI, which is homolog to
Stx2a-producing strain EDL933 (98). Q21 was Q2851 and Q21 and, in accordance with previous

28 SCHEUTZ
findings, primarily found in stx2c phages al- molecular marker to distinguish among Stx2a-
though one stx2a phage was also identified. and Stx2c-encoding phages in O157 strains
Main group QII, represented by six Q genes, is (73).
homolog to QO111:H, where one is character- The great variation in Q genes compared
ized as QIIb in an stx2a phage and five are to stx genes indicates that the stx genes have
characterized as QIIa found in stx2a (one) and been taken up by lambdoid phages on several
stx2d (four) phages, respectively. QIII is the occasions and that there is a high genetic se-
smallest group with only two Q genes in stx2e lection for this gene combination and that all
phages, whereas QIV has as many as 59 Q genes, the Q genes, functionally expressing their lysis
with homologs to Q933W, QH-19B, QMorioka, and genes, can take up the stx genes under certain
QVT1-Sakai. QIV can be divided into six variants, favorable conditions.
none of which is found in stx2c, stx2d, or stx2e
phages, but QIVa is found in stx1a and stx2a,
QIVb in stx1a, QIVc in stx1c, QIVd in stx2a, QIVe in GENOMIC ISLANDS
stx1a, and QIVf in stx2a phages (105).Group V
contained only two sequences that were not Sequencing of STEC O157:H7 EDL933 revealed
used in Jacobsens study. Among all Q variants, O157:H7-specific DNA organized in genomic
23 of 169 sites were conserved sites, and 32 O islands (106), some of which are referred
were functionally conserved. There were 13 to as pathogenicity islands because they can
different Q-vtx combinations based on Q vari- contain a flexible pool of virulence-associated
ants and vtx subtypes, but some strains carried genes. One such pathogenicity island is LEE,
two VT phages, resulting in 17 different Q-vtx which encodes a type III secretion system
genotypes. These phylogenetic analyses of involved in the formation attaching and ef-
the associated stx genes thus revealed six dif- facing lesions. The most important protein is
ferent vtx subtypes, vtx1a, vtx1c, vtx2a, vtx2c, intimin encoded by the gene eae, which is res-
vtx2d, and vtx2e. Q genes associated with the ponsible for both the intimate adhesion of
stx2b and stx2g phages in isolates EH250 bacteria to the intestinal epithelium and the
and 7V, respectively, are apparently associated attaching and effacing lesion. Analysis of the
with another phage in the isolate, indicating variable C-terminal encoding sequence of eae
that Q expression might have a contribution defines at least 29 distinct intimin types (1,
from additional Q genes in non-stx phages (53). 2, 1, 2, 3, 1, 1, , , 1, 2, 3, 4, 5, 1, 3,
The genetic architecture does not appear 1, 2, 1-A, 1-B, 1-C, 2, , , , , , , ) (107
to be the only factor affecting Stx expres- 109) that have been associated with tissue
sion. The inferred phylogeny based on the tropism. Intimin binds to the cell receptor
alignment of stx bacteriophages indicates a Tir, which is translocated by bacteria to the
broad phylogenetic diversity also observed enterocyte through a type III secretion sys-
with whole-genome phylogeny (53). How- tem (110). One SNP, 255 T>A in tir, has been
ever, taken together, the present data seem shown to predict the propensity of STEC O157
to suggest that the primary sequence of Q isolates to cause human clinical disease (111).
may play a major role in the regulation and The overrepresentation of the tir 255 T>A
released amount of Shiga toxin. In O157, DNA T allele in human-derived isolates versus the
sequencing of the genes flanking the stx2c tir 255 T>A A allele suggests that these iso-
gene from a lineage II strain, EC970520, has lates have a higher propensity to cause dis-
revealed that the Q gene is replaced by a pphA ease. The high frequency of bovine isolates
(serine/threonine phosphatase) homolog in with the A allele suggests a possible bovine
this and all other lineage II strains tested with ecological niche for this STEC O157 subset. A
very low similarity to the antiterminator Q of Norwegian study of 167 human and nonhu-
lineage I strains, which appears to be a useful man E. coli O157:H7/NM (nonmotile) isolates

with respect to the 255 T A/T SNP in the association with serious disease. Thus, after
tir gene was able to differentiate STEC O157 acquiring OI-122 modules from STEC O157:
into distinct virulence clades (1 to 3 and 8) and H7 through horizontal transfer, less patho-
found an overrepresentation of the T allele genic non-O157 strains could cross a virulence
among human strains compared to nonhuman threshold, resulting in sufficient pathogenicity
strains, including five of six HUS cases (112). to cause HUS (115).
OI-122 is another 23-kb pathogenicity is-
land with at least six genes with significant
homology to known virulence genes: pagC of TAXONOMY AND PUBLIC HEALTH
Salmonella enterica serovar Typhimurium; sen
of Shigella flexneri; two non-LEE effector (nle) More than 472 O:H serotypes of STEC have
genes nleB and nleE of Citrobacter rodentium; been described (28). Some seropathotypes
and the EHEC factor for adherence gene clus- seem to be more closely related to specific
ter efa1 and efa2 found in STEC O157:H7. A reservoirs and specific diseases, and their
modular arrangement of OI-122 genes based number and characteristics are ever changing.
on their association with each other across However, there seems to be a much closer
HUS-associated non-O157 VTEC strains was relationship between the virulence profile,
proposed: module 1 contains Z4318, pagC, and i.e., stx subtype, additional cocktail genes,
Z4322; module 2 contains Z4323, sen, nleB, and the clinical course of disease than to the
and nleE; and module 3 contains the efa gene serotype itselfeven for O157 (76). Many of
cluster, also referred to as lifA (lymphocyte these genes are acquired by horizontal gene
activation inhibitor) (113, 114). A progressive transfer, and it is therefore possible (and
decrease in the prevalence of OI-122 genes plausible) that certain types of E. coli that
in non-O157 VTEC strains belonging to sero- have never been associated with disease or
pathotype B through E has been suggested (3), outbreaks could acquire mobile genetic ele-
and a Belgian study of 265 STEC isolates also ments such as the Stx encoding bacterio-
observed a progressive decrease in the fre- phages through parallel evolution, which in
quency of complete OI-122 in seropathotypes turn could convert them into virulent strains
A through D, with a concomitant increase capable of causing human infection and out-
in the frequencies of incomplete and absent breaks. Understanding this complex biology of
OI-122. The variable virulence profiles of the host-pathogen interaction is evidently a high-
non-O157 serotypes indicated that complete priority research subject, and a better under-
OI-122 was more frequently present in isolates standing of the real genetic basis behind our
associated with HUS and that the individual classification of seropathotypes is needed. The
genes stx2, eae, espP, as well as the OI-122- presence of LEE and OI-122 in many of the A
associated genes sen, nleB, nleE, and the efa/ and B seropathotypes has been very useful but
lifA gene cluster were significantly more often has also been challenged by the serious O104:
present in non-O157 STEC associated with H4 outbreak in Germany in 2011 as well as
HUS, and that the combined virulence profile by other non-LEE STEC outbreaks and HUS
vtx2-nleE-efa/lifA showed the strongest asso- cases. Furthermore, the genome sequences
ciation with HUS (24). The presence of pu- for some of the common non-O157 O groups,
tative transposases in OI-122 has led to the O26, O103, and O111, have revealed that these
hypothesis that its elements are acquired or pathogenic groups have arisen multiple times
lost in a modular manner. Although pagC, by acquisition of mobile genetic elements
Z4322, sen, nleB, nleE, and efa1/lifA individu- harboring virulence genes. Lambdoid pro-
ally are more prevalent in non-O157 VTEC phages and other integrative elements have
associated with HUS, the simultaneous pre- played a major role in the stepwise evolution
sence of all of these genes strengthens the of pathogenic lineages of STEC, especially the

30 SCHEUTZ
acquisition of type 3 secretion effectors, i.e., might have an HUSEC infection. If vtx2f is
non-LEE-encoding nle genes, and some of diagnosed as unspecified vtx2, an additional
these otherwise unlinked effectors can work 55 (3%) patients would have been added, but
in concert with one another to produce a de- as stx2f requires a different and specific set of
sired effect on the host cell (116, 117). More primers, this would not have been relevant in
whole-genome sequence information on non- most cases. Further refining of the definition
O157 STEC and STEC-related E. coli strains of HUSEC by stx subtyping to include only
is needed to better understand and further stx2a and stx2d as HUSEC would reduce the
categorize the clinically relevant pathotypes number of patients with HUSEC to 11% (224
for the necessary public health response and of 2,062). For specific, mainly socioeconomic
action. reasons, antibiotic treatment of patients in-
In public health the primary goal is to fected with VTEC is also considered accord-
(i) prevent onset of disease, (ii) minimize the ing to specific criteria, which include the
risk of progression of the disease in individu- isolation of identical low-risk types of STEC
als or transmission of illness, and (iii) provide found in two separate stool specimens, no iso-
rehabilitation to prevent the worsening of an lation of HUSEC, and a detailed character-
individuals health. The methods and tech- ization of the STEC strains virulence profile
niques to categorize an STEC strain isolated and serotype. In the first line this includes the
from a patient at the present time require ex- detection of the eae, stx1, stx2, aggR, and aaiC
tensive analyses that are usually not available genes followed by a more detailed second-
in the primary diagnostic line where public line subtyping of the stx genes. This simpli-
health action and management of STEC- fied approach has been generally accepted
infected patients begins. In the first-line di- by the primary diagnostic clinical microbiol-
agnosis, three primary issues need immediate ogy laboratories and public health officers
attention: in Denmark. Though not complete in the
characterization of each STEC isolate, this
Risk of severe disease and prognosis
approach is practical and easy to use in an
Quarantine of infected individuals
operational environment, which is expected
Treatment with antibiotics, especially with
to quickly (i) evaluate the risk of progression
long-term otherwise healthy carriers
of the disease in individuals, (ii) minimize
In Denmark, the observed and well- transmission of HUSEC, (iii) rehabilitate
documented association between Stx2 and individuals to prevent the worsening of an
eae or ability to colonize and persist in the gut, individuals health, and (iv) reduce the socio-
such as the EAEC hybrid strains, has resulted economic impact on their families. However,
in a basic and primary definition of HUSEC such a simple approach should be applied
for first-line public health action: stx2 in a with prudence because each individual case is
background of eae- or aggR-positive E. coli unique, and each patient should be carefully
is HUS-associated and is kept on hold until evaluated with regard to predisposing factors,
stx is subtyped and further characterized. All general clinical condition, contact with other
other STEC strains with the virulence profiles STEC-infected individuals, possible link to an
of vtx1, vtx1 and eae, vtx2, vtx1 and vtx2, re- outbreak, and so forth. The identification of
gardless of serotype, are considered low-risk low-risk STEC does not per se exclude the
STEC. If this HUSEC definition had been ap- risk of progression to severe disease, dehy-
plied, 71% (1,454 of 2,046) of Danish patients dration, or HUS, and patients with bloody
would have been informed that they had a stools and/or affected kidney function must
low-risk STEC infection during the years be carefully monitored.
1983 to 2012. Twenty-nine percent (552 of The versatile integration of mobile viru-
2,046) would have been informed that they lence factors in the stepwise and parallel

evolution of pathogenic lineages of STEC col- 4. Konowalchuk J, Speirs JI, Stavric S. 1977. Vero
lides with the requirements of a good taxon- response to a cytotoxin of Escherichia coli. Infect
Immun 18:775779.
omy and complicates how public health goals 5. OBrien AD, LaVeck GD, Thompson MR,
are met in a timely manner. The intertwine- Formal SB. 1982. Production of Shigella dys-
ment of the traditional taxonomy concept and enteria type 1-like cytotoxin by Escherichia coli.
the association of (sero)-pathotypes to epide- J Infect Dis 146:763769.
miological evidence, phenotypic traits, clinical 6. OBrien AD, La Veck GF. 1983. Purification and
characterization of a Shigella dysenteria 1-like
features of the disease, and specific virulence toxin produced by Escherichia coli. Infect Immun
factors is therefore an ongoing challenge, and 40:675683.
the need to identify factors of STEC that ab- 7. Karmali MA, Petric M, Louie S, Cheung R.
solutely predict the potential to cause human 1986. Antigenic heterogeneity of Escherichia coli
disease is obvious in terms of clinical man- verotoxins. Lancet i:164165.
8. Scotland SM, Smith HR, Rowe B. 1985. Two
agement, supportive or antibiotic treatment,
distinct toxins active on Vero cells from Esche-
quarantine measurements, risk assessment, richia coli O157. Lancet ii:885886.
surveillance, and outbreak investigations and 9. OBrien AD, Newland JW, Miller SF, Holmes
management. RK, Smith HW, Formal SB. 1984. Shiga-like
However, recent sequence information has toxin-converting phages from Escherichia coli
provided advanced analytical tools that will strains that cause hemorrhagic colitis or infantile
diarrhea. Science 226:694696.
reduce the clinical divide between the STEC
10. Scotland SM, Smith HR, Willshaw GA,
types that are associated with severe disease Rowe B. 1983. Vero cytotoxin production in
from the seemingly benign STEC types. strain of Escherichia coli is determined by
genes carried on bacteriophage. Lancet ii:216.
(Letter).
ACKNOWLEDGMENTS 11. Smith HR, Day NP, Scotland SM, Gross RJ,
Rowe B. 1984. Phage-determined production of
I declare no conflicts of interest with regard to Vero cytotoxin in strains of Escherichia coli sero-
the manuscript. group O157. Lancet i:12421243. (Letter).
12. Smith HW, Green P, Parsell Z. 1983. Vero cell
toxins in Escherichia coli and related bacteria:
CITATION transfer by phage and conjugation and toxic ac-
tion in laboratory animals, chickens and pigs.
Scheutz F. 2014. Taxonomy meets public health: J Gen Microbiol 129:31213137.
the case of Shiga toxin-producing Escherichia 13. Riley LW, Remis RS, Helgerson SD, McGee HB,
coli. Microbiol Spectrum 2(3):EHEC-0019-2013. Wells JG, Davis BR, Herbert RJ, Olcott ES,
Johnson LM, Hargret NT, Blake PA, Cohen
ML. 1983. Hemorrhagic colitis associated with a
rare Escherichia coli serotype. N Engl J Med 308:
REFERENCES
681685.
1. Neter E, Westphal O, Lderitz O, Gino RM, 14. Johnson WM, Lior H, Bezanson GS. 1983. Cy-
Gorzynski EA. 1955. Demonstration of antibodies totoxic Escherichia coli O157:H7 associated with
against enteropathogenic Escherichia coli in sera hemorrhagic colitis in Canada. Lancet i:76.
of children of various ages. Pediatrics 16:801807. 15. OBrien AD, Lively TA, Chen ME, Rothman
2. Kaper JB, Nataro JP, Mobley HL. 2004. Path- SW, Formal SB. 1983. Escherichia coli O157:H7
ogenic Escherichia coli. Nat Rev Microbiol 2:123 strains associated with haemorrhagic colitis in
140. the United States produce a Shigella dysenteriae 1
3. Karmali MA, Mascarenhas M, Shen S, Ziebell (Shiga) like cytotoxin. Lancet i:702.
KA, Johnson S, Reid-Smith R, Isaac-Renton J, 16. Karmali MA, Steele BT, Petric M, Lim CB. 1983.
Clark C, Rahn K, Kaper JB. 2003. Association Sporadic cases of haemolytic-uraemic syndrome
of genomic O island 122 of Escherichia coli EDL associated with faecal cytotoxin and cytotoxin-
933 with verocytotoxin-producing Escherichia producing Escherichia coli in stools. Lancet i:619
coli seropathotypes that are linked to epidemic 620.
and/or serious disease. J Clin Microbiol 41:4930 17. Karmali MA, Petric M, Lim C, Fleming PC,
4940. Steele BT. 1983. Escherichia coli cytotoxin

32 SCHEUTZ
haemolytic-uraemic syndrome, and haemorrhagic 2013. The European Union summary report on
colitis. Lancet ii:12991300. trends and sources of zoonoses, zoonotic agents
18. Levine MM, Edelman R. 1984. Enteropathogenic and food-borne outbreaks in 2011. EFSA J 11:3129.
Escherichia coli of classic serotypes associated 1250. 4-9-2013. European Food Safety Authority,
with infant diarrhea: epidemiology and patho- Parma, Italy.
genesis. Epidemiol Rev 6:3151. 30. Kawano K, Okada M, Haga T, Maeda K, Goto Y.
19. Levine MM, Xu J-G, Kaper JB, Lior H, Prado V, 2008. Relationship between pathogenicity for
Nataro JP, Karch H, Wachsmuth IK. 1987. A humans and stx genotype in Shiga toxin-producing
DNA probe to identify enterohemorrhagic Esch- Escherichia coli serotype O157. Eur J Clin Micro-
erichia coli of O157:H7 and other serotypes that biol Infect Dis 27:227232.
cause hemorrhagic colitis and hemolytic uremic 31. Pai CH, Ahmed N, Lior H, Johnson WM, Sims
syndrome. J Infect Dis 156:175182. HV, Woods DE. 1988. Epidemiology of sporadic
20. Wade WG, Thom BT, Evans N. 1979. Cytotoxic diarrhea due to verocytotoxin-producing Esche-
enteropathogenic Escherichia coli. Lancet ii:1235 richia coli: a two-year prospective study. J Infect
1236. Dis 157:10541057.
21. Scotland SM, Day NP, Willshaw GA, Rowe B. 32. Boerlin P, Mcewen SA, Boerlin-Petzold F,
1980. Cytotoxic enteropathogenic Escherichia coli. Wilson JB, Johnson RP, Gyles CL. 1999. Asso-
Lancet 1:90. ciations between virulence factors of Shiga toxin-
22. Scheutz F, Strockbine NA. 2005. Escherichia, p producing Escherichia coli and disease in humans.
607624. In Garrity GM, Brenner DJ, Krieg NR, J Clin Microbiol 37:497503.
Staley JT (ed), Bergeys Manual of Systematic 33. Gyles CL. 2007. Shiga toxin-producing Esche-
Bacteriology, vol 2, part B, The Gammaproteo- richia coli: an overview. J Anim Sci 85:E45E62.
bacteria. Springer, New York, NY. 34. Ethelberg S, Olsen KEP, Scheutz F, Jensen C,
23. Bosilevac JM, Koohmaraie M. 2011. Prevalence Schiellerup P, Engberg J, Petersen AM, Olesen
and characterization of non-O157 Shiga toxin- B, Gerner-Smidt P, Mlbak K. 2004. Virulence
producing Escherichia coli isolates from com- factors for hemolytic uremic syndrome, Denmark.
mercial ground beef in the United States. Appl Emerg Infect Dis 10:842847.
Environ Microbiol 77:21032112. 35. Cookson AL, Bennett J, Thomson-Carter F,
24. Buvens G, Pierard D. 2012. Virulence pro- Attwood GT. 2007. Molecular subtyping and
filing and disease association of verocytotoxin- genetic analysis of the enterohemolysin gene
producing Escherichia coli O157 and non-O157 (ehxA) from Shiga toxin-producing Escherichia
isolates in Belgium. Foodborne Pathog Dis 9:530 coli and atypical enteropathogenic E. coli. Appl
535. Environ Microbiol 73:63606369.
25. EFSA Panel on Biological Hazards (BIOHAZ). 36. Schmidt H, Karch H. 1996. Enterohemolytic phe-
2013. Scientific opinion on VTEC-seropathotype notypes and genotypes of Shiga toxin-producing
and scientific criteria regarding pathogenicity Escherichia coli O111 strains from patients with diar-
assessment. EFSA J 11:3138. 1106. 7-3-2013. Eu- rhea and hemolytic-uremic syndrome. J Clin Micro-
ropean Food Safety Authority, Parma, Italy. biol 34:23642367.
26. Centers for Disease Control and Prevention. 37. Mundy R, Jenkins C, Yu J, Smith H, Frankel G.
1995. Outbreak of acute gastroenteritis attributable 2004. Distribution of espI among clinical entero-
to Escherichia coli serotype O104:H21Helena, haemorrhagic and enteropathogenic Escherichia
Montana. Morb Mortal Wkly Rep 44:501503. coli isolates. J Med Microbiol 53:11451149.
27. Paton AW, Woodrow MC, Doyle RM, Lanser 38. Bettelheim KA. 2007. The non-O157 shiga-toxigenic
JA, Paton JC. 1999. Molecular characterization (verocytotoxigenic) Escherichia coli; under-rated
of a Shiga toxigenic Escherichia coli O113:H21 pathogens. Crit Rev Microbiol 33:6787.
strain lacking eae responsible for a cluster of cases 39. Lukjancenko O, Wassenaar TM, Ussery DW.
of hemolytic-uremic syndrome. J Clin Microbiol 2010. Comparison of 61 sequenced Escherichia
37:33573361. coli genomes. Microb Ecol 60:708720.
28. Mora A, Herrera A, Lopez C, Dahbi G, Mamani 40. Mellmann A, Harmsen D, Cummings CA, Zentz
R, Pita JM, Alonso MP, Llovo J, Bernardez MI, EB, Leopold SR, Rico A, Prior K, Szczepanowski
Blanco JE, Blanco M, Blanco J. 2011. Charac- R, Ji Y, Zhang W, McLaughlin SF, Henkhaus JK,
teristics of the Shiga-toxin-producing entero- Leopold B, Bielaszewska M, Prager R, Brzoska
aggregative Escherichia coli O104:H4 German PM, Moore RL, Guenther S, Rothberg JM,
outbreak strain and of STEC strains isolated in Karch H. 2011. Prospective genomic characteriza-
Spain. Int Microbiol 14:121141. tion of the German enterohemorrhagic Escherichia
29. European Food Safety Authority and European coli O104:H4 outbreak by rapid next generation
Centre for Disease Prevention and Control. sequencing technology. PloS One 6:e22751.

41. Kim J, Nietfeldt J, Benson AK. 1999. Octamer- 51. Abu-Ali GS, Lacher DW, Wick LM, Qi W,
based genome scanning distinguishes a unique Whittam TS. 2009. Genomic diversity of patho-
subpopulation of Escherichia coli O157:H7 strains genic Escherichia coli of the EHEC 2 clonal
in cattle. Proc Natl Acad Sci USA 96:1328813293. complex. BMC Genomics 10:296.
42. Yang Z, Kovar J, Kim J, Nietfeldt J, Smith DR, 52. Hazen TH, Sahl JW, Fraser CM, Donnenberg
Moxley RA, Olson ME, Fey PD, Benson AK. MS, Scheutz F, Rasko DA. 2013. Refining the
2004. Identification of common subpopulations pathovar paradigm via phylogenomics of the
of non-sorbitol-fermenting, beta-glucuronidase- attaching and effacing Escherichia coli. Proc Natl
negative Escherichia coli O157:H7 from bovine Acad Sci USA 110:1281012815.
production environments and human clinical 53. Steyert SR, Sahl JW, Fraser CM, Teel LD,
samples. Appl Environ Microbiol 70:68466854. Scheutz F, Rasko DA. 2012. Comparative ge-
43. Zhang Y, Laing C, Steele M, Ziebell K, Johnson nomics and stx phage characterization of LEE-
R, Benson AK, Taboada E, Gannon VP. 2007. negative Shiga toxin-producing Escherichia coli.
Genome evolution in major Escherichia coli O157: Front Cell Infect Microbiol 2:133.
H7 lineages. BMC Genomics 8:121. 54. Scheutz F, Teel LD, Beutin L, Pierard D,
44. Bono JL, Smith TP, Keen JE, Harhay GP, Buvens G, Karch H, Mellmann A, Caprioli A,
McDaneld TG, Mandrell RE, Jung WK, Besser Tozzoli R, Morabito S, Strockbine NA, Melton-
TE, Gerner-Smidt P, Bielaszewska M, Karch H, Celsa AR, Sanchez M, Persson S, OBrien AD.
Clawson ML. 2012. Phylogeny of Shiga toxin- 2012. Multicenter evaluation of a sequence-based
producing Escherichia coli O157 isolated from protocol for subtyping shiga toxins and stan-
cattle and clinically ill humans. Mol Biol Evol dardizing Stx nomenclature. J Clin Microbiol 50:
29:20472062. 29512963.
45. Manning SD, Motiwala AS, Springman AC, Qi 55. Asakura H, Makino SI, Kobori H, Watarai M,
W, Lacher DW, Ouellette LM, Mladonicky JM, Shirahata T, Ikeda T, Takeshi K. 2001. Phylo-
Somsel P, Rudrik JT, Dietrich SE, Zhang W, genetic diversity and similarity of active sites of
Swaminathan B, Alland D, Whittam TS. 2008. Shiga toxin (Stx) in Shiga toxin-producing Esch-
Variation in virulence among clades of Esche- erichia coli (STEC) isolates from humans and
richia coli O157:H7 associated with disease out- animals. Epidemiol Infect 127:2736.
breaks. Proc Natl Acad Sci USA 105:48684873. 56. Bastian SN, Carle I, Grimont F. 1998. Compari-
46. Centers for Disease Control and Prevention. son of 14 PCR systems for the detection and
2006. Ongoing multistate outbreak of Escherichia subtyping of stx genes in Shiga-toxin-producing
coli serotype O157:H7 infections associated with Escherichia coli. Res Microbiol 149:457472.
consumption of fresh spinachUnited States, 57. De Baets L, Van der Taelen I, De Filette M,
September 2006. Morb Mortal Wkly Rep 55:1045 Pirard D, Allison L, De Greve H, Hernalsteens
1046. JP, Imberechts H. 2004. Genetic typing of shiga
47. Grant J, Wendelboe AM, Wendel A, Jepson B, toxin 2 variants of Escherichia coli by PCR-
Torres P, Smelser C, Rolfs RT. 2008. Spinach- restriction fragment length polymorphism analy-
associated Escherichia coli O157:H7 outbreak, sis. Appl Environ Microbiol 70:63096314.
Utah and New Mexico, 2006. Emerg Infect Dis 58. Iwasa M, Makino S, Asakura H, Kobori H,
14:16331636. Morimoto Y. 1999. Detection of Escherichia
48. Abu-Ali GS, Ouellette LM, Henderson ST, coli O157:H7 from Musca domestica (Diptera:
Whittam TS, Manning SD. 2010. Differences in Muscidae) at a cattle farm in Japan. J Med
adherence and virulence gene expression between Entomol 36:108112.
two outbreak strains of enterohaemorrhagic Esch- 59. Lee JE, Reed J, Shields MS, Spiegel KM, Farrell
erichia coli O157:H7. Microbiology 156:408419. LD, Sheridan PP. 2007. Phylogenetic analysis of
49. Whittam TS. 1998. Evolution of Eschericia coli Shiga toxin 1 and Shiga toxin 2 genes associated
O157:H7 and other Siga toxin-producing E. coli with disease outbreaks. BMC Microbiol 7:109.
strains, p 195209. In Kaper JB, OBrien AD (ed), 60. Pirard D, Muyldermans G, Moriau L, Stevens
Escherichia coli O157:H7 and Other Shiga Toxin- D, Lauwers S. 1998. Identification of new vero-
Producing E. coli. Strains, 1st ed. ASM Press, cytotoxin type 2 variant B-subunit genes in
Washington, DC. human and animal Escherichia coli isolates. J Clin
50. Tarr CL, Large TM, Moeller CL, Lacher DW, Microbiol 36:33173322.
Tarr PI, Acheson DW, Whittam TS. 2002. 61. Bielaszewsk M, Friedrich AW, Aldick T,
Molecular characterization of a serotype O121:H19 Schurk-Bulgrin R, Karch H. 2006. Shiga toxin
clone, a distinct Shiga toxin-producing clone of activatable by intestinal mucus in Escherichia
pathogenic Escherichia coli. Infect Immun 70: coli isolated from humans: predictor for a severe
68536859. clinical outcome. Clin Infect Dis 43:11601167.

34 SCHEUTZ
62. Friedrich AW, Bielaszewsk M, Zhang WL, 72. Baker DR, Moxley RA, Steele MB, LeJeune JT,
Pulz M, Kuczius T, Ammon A, Karch H. 2002. Christopher-Hennings J, Chen DG, Hardwidge
Escherichia coli harboring Shiga toxin 2 gene PR, Francis DH. 2007. Differences in virulence
variants: frequency and association with clinical among Escherichia coli O157:H7 strains isolated
symptoms. J Infect Dis 185:7484. from humans during disease outbreaks and from
63. Persson S, Olsen KEP, Ethelberg S, Scheutz F. healthy cattle. Appl Environ Microbiol 73:7338
2007. Subtyping method for Escherichia coli Shiga 7346.
toxin (verocytotoxin) 2 variants and correlations 73. Zhang Y, Laing C, Zhang Z, Hallewell J, You C,
to clinical manifestations. J Clin Microbiol 45: Ziebell K, Johnson RP, Kropinski AM, Thomas
20202024. JE, Karmali M, Gannon VP. 2010. Lineage and
64. Eklund M, Leino K, Siitonen A. 2002. Clinical host source are both correlated with levels of
Escherichia coli strains carrying stx genes: stx Shiga toxin 2 production by Escherichia coli
variants and stx-positive virulence profiles. J Clin O157:H7 strains. Appl Environ Microbiol 76:474
Microbiol 40:45854593. 482.
65. Pedersen MG, Hansen C, Riise E, Persson S, 74. Neupane M, Abu-Ali GS, Mitra A, Lacher DW,
Olsen KE. 2008. Subtype-specific suppression of Manning SD, Riordan JT. 2011. Shiga toxin 2
Shiga toxin 2 released from Escherichia coli upon overexpression in Escherichia coli O157:H7 strains
exposure to protein synthesis inhibitors. J Clin associated with severe human disease. Microb
Microbiol 46:29872991. Pathog 51:466470.
66. Rivas M, Miliwebsky E, Chinen I, Roldan CD, 75. Uhlich GA, Sinclair JR, Warren NG, Chmielecki
Balbi L, Garcia B, Fiorilli G, Sosa-Estani S, WA, Fratamico P. 2008. Characterization of
Kincaid J, Rangel J, Griffin PM. 2006. Char- Shiga toxin-producing Escherichia coli isolates
acterization and epidemiologic subtyping of associated with two multistate food-borne out-
Shiga toxin-producing Escherichia coli strains breaks that occurred in 2006. Appl Environ
isolated from hemolytic uremic syndrome and Microbiol 74:12681272.
diarrhea cases in Argentina. Foodborne Pathog 76. Sborg B, Lassen SG, Muller L, Jensen T,
Dis 3:8896. Ethelberg S, Molbak K, Scheutz F. 2013. A
67. Bonnet R, Souweine B, Gauthier G, Rich C, verocytotoxin-producing E. coli outbreak with a
Livrelli V, Sirot J, Joly B, Forestier C. 1998. surprisingly high risk of haemolytic uraemic syn-
Non-O157:H7 Stx2-producing Escherichia coli drome, Denmark, SeptemberOctober 2012. Euro
strains associated with sporadic cases of hemo- Surveill 18:ii, 20350.
lytic-uremic syndrome in adults. J Clin Microbiol 77. Shringi S, Garcia A, Lahmers KK, Potter KA,
36:17771780. Muthupalani S, Swennes AG, Hovde CJ, Call
68. Keskimki M, Ikaheimo R, Karkkainen P, DR, Fox JG, Besser TE. 2012. Differential viru-
Scheutz F, Ratiner Y, Puohiniemi R, Siitonen A. lence of clinical and bovine-biased enterohemor-
1997. Shiga toxin-producing Escherichia coli se- rhagic Escherichia coli O157:H7 genotypes in
rotype OX3:H21 as a cause of hemolytic-uremic piglet and Dutch belted rabbit models. Infect
syndrome. Clin Infect Dis 24:12781279. Immun 80:369380.
69. Keskimki M, Ratiner Y, Oinonen S, Leijala 78. Kawano K, Ono H, Iwashita O, Kurogi M,
E, Nurminen M, Saari M, Siitonen A. 1999. Haga T, Maeda K, Goto Y. 2012. stx genotype
Haemolytic-uraemic syndrome caused by vero and molecular epidemiological analyses of Shiga
toxin-producing Escherichia coli serotype Rough: toxin-producing Escherichia coli O157:H7/H- in
K-:H49. Scand J Infect Dis 31:141144. human and cattle isolates. Eur J Clin Microbiol
70. Rasko DA, Webster DR, Sahl JW, Bashir A, Infect Dis 31:119127.
Boisen N, Scheutz F, Paxinos EE, Sebra R, Chin 79. Nishikawa Y, Zhou Z, Hase A, Ogasawara J,
CS, Iliopoulos D, Klammer A, Peluso P, Lee L, Cheasty T, Haruki K. 2000. Relationship of ge-
Kislyuk AO, Bullard J, Kasarskis A, Wang S, Eid netic type of shiga toxin to manifestation of
J, Rank D, Redman JC, Steyert SR, Frimodt- bloody diarrhea due to enterohemorrhagic Esch-
Moller J, Struve C, Petersen AM, Krogfelt erichia coli serogroup O157 isolates in Osaka City,
KA, Nataro JP, Schadt EE, Waldor MK. 2011. Japan. J Clin Microbiol 38:24402442.
Origins of the E. coli strain causing an outbreak of 80. Orth D, Grif K, Khan AB, Naim A, Dierich
hemolytic-uremic syndrome in Germany. N Engl MP, Wurzner R. 2007. The Shiga toxin geno-
J Med 365:709717. type rather than the amount of Shiga toxin or
71. Fuller CA, Pellino CA, Flagler MJ, Strasser JE, the cytotoxicity of Shiga toxin in vitro correlates
Weiss AA. 2011. Shiga toxin subtypes display with the appearance of the hemolytic uremic
dramatic differences in potency. Infect Immun syndrome. Diagn Microbiol Infect Dis 59:235
79:13291337. 242.

81. Mariani-Kurkdjian P, Denamur E, Milon A, 91. Clawson ML, Keen JE, Smith TP, Durso LM,
Picard B, Cave H, Lambert-Zechovsky N, Loirat McDaneld TG, Mandrell RE, Davis MA, Bono
C, Goullet P, Sansonetti PJ, Elion J. 1993. Iden- JL.. 2009. Phylogenetic classification of Esche-
tification of a clone of Escherichia coli O103:H2 as richia coli O157:H7 strains of human and bovine
a potential agent of hemolytic-uremic syndrome origin using a novel set of nucleotide polymor-
in France. J Clin Microbiol 31:296301. phisms. Genome Biol 10:R56.
82. Beutin L, Krause G, Zimmermann S, Kaulfuss S, 92. Brssow H, Canchaya C, Hardt WD. 2004.
Gleier K. 2004. Characterization of Shiga toxin- Phages and the evolution of bacterial pathogens:
producing Escherichia coli strains isolated from from genomic rearrangements to lysogenic con-
human patients in Germany over a 3-year period. version. Microbiol Mol Biol Rev 68:560602.
J Clin Microbiol 42:10991108. 93. Herold S, Karch H, Schmidt H. 2004. Shiga
83. Scheutz F, Ethelberg S. 2008. Nordic meeting on toxin-encoding bacteriophagesgenomes in mo-
detection and surveillance of VTEC infections in tion. Int J Med Microbiol 294:115121.
humans. 1-30. 2008. http://www.ssi.dk/English/ 94. Waldor MK, Friedman DI. 2005. Phage regula-
HealthdataandICT/National%20Reference% tory circuits and virulence gene expression. Curr
20Laboratories/Bacteria/~/media/Indhold/EN% Opin Microbiol 8:459465.
20-%20engelsk/Public%20Health/National% 95. Wagner PL, Neely MN, Zhang XP, Acheson
20Reference%20Laboratories/Nordic%20VTEC DWK, Waldor MK, Friedman DI. 2001. Role for
%20Report.ashx. a phage promoter in Shiga toxin 2 expression
84. Franke S, Harmsen D, Caprioli A, Pirard D, from a pathogenic Escherichia coli strain. J Bac-
Wieler LH, Karch H. 1995. Clonal relatedness of teriol 183:20812085.
Shiga-like toxin-producing Escherichia coli O101 96. Fortier LC, Sekulovic O. 2013. Importance of
strains of human and porcine origin. J Clin prophages to evolution and virulence of bacterial
Microbiol 33:31743178. pathogens. Virulence 4:354365.
85. Ogura Y, Ooka T, Asadulghani, Terajima J, 97. Wagner PL, Livny J, Neely MN, Acheson DW,
Nougayrede JP, Kurokawa K, Tashiro K, Tobe T, Friedman DI, Waldor MK. 2002. Bacteriophage
Nakayama K, Kuhara S, Oswald E, Watanabe H, control of Shiga toxin 1 production and release by
Hayashi T. 2007. Extensive genomic diversity and Escherichia coli. Mol Microbiol 44:957970.
selective conservation of virulence-determinants in 98. Plunkett G, Rose DJ, Durfee TJ, Blattner FR.
enterohemorrhagic Escherichia coli strains of O157 1999. Sequence of Shiga toxin 2 phage 933W from
and non-O157 serotypes. Genome Biol 8:R138. Escherichia coli O157:H7: Shiga toxin as a phage
86. Ogura Y, Ooka T, Iguchi A, Toh H, Asadulghani late-gene product. J Bacteriol 181:17671778.
M, Oshima K, Kodama T, Abe H, Nakayama K, 99. Guo HC, Kainz M, Roberts JW. 1991. Charac-
Kurokawa K, Tobe T, Hattori M, Hayashi T. terization of the late-gene regulatory region of
2009. Comparative genomics reveal the mecha- phage 21. J Bacteriol 173:15541560.
nism of the parallel evolution of O157 and non- 100. Strauch E, Schaudinn C, Beutin L. 2004. First-
O157 enterohemorrhagic Escherichia coli. Proc time isolation and characterization of a bacterio-
Natl Acad Sci USA 106:1793917944. phage encoding the Shiga toxin 2c variant, which
87. Shringi S, Schmidt C, Katherine K, Brayton KA, is globally spread in strains of Escherichia coli
Hancock DD, Besser TE. 2012. Carriage of stx2a O157. Infect Immun 72:70307039.
differentiates clinical and bovine-biased strains 101. Strauch E, Hammerl JA, Konietzny A,
of Escherichia coli O157. PLoS One 7:e51572. Schneiker-Bekel S, Arnold W, Goesmann A,
doi:10.1371/journal.pone.0051572. Puhler A, Beutin L. 2008. Bacteriophage 2851 is a
88. Garcia-Aljaro C, Muniesa M, Jofre J, Blanch prototype phage for dissemination of the Shiga
AR.. 2009. Genotypic and phenotypic diversity toxin variant gene 2c in Escherichia coli O157:H7.
among induced, stx2-carrying bacteriophages Infect Immun 76:54665477.
from environmental Escherichia coli strains. Appl 102. Makino K, Yokoyama K, Kubota Y, Yutsudo
Environ Microbiol 75:329336. CH, Kimura S, Kurokawa K, Ishii K, Hattori M,
89. Serra-Moreno R, Jofre J, Muniesa M. 2007. Tatsuno I, Abe H, Iida T, Yamamoto K, Onishi
Insertion site occupancy by stx2 bacteriophages M, Hayashi T, Yasunaga T, Honda T, Sasakawa
depends on the locus availability of the host strain C, Shinagawa H. 1999. Complete nucleotide
chromosome. J Bacteriol 189:66456654. sequence of the prophage VT2-Sakai carrying
90. Shringi S, Schmidt C, Katherine K, Brayton KA, the verotoxin 2 genes of the enterohemorrhagic
Hancock DD, Besser TE. 2012. Carriage of stx2a Escherichia coli O157:H7 derived from the Sakai
differentiates clinical and bovine-biased strains outbreak. Genes Genet Syst 74:227239.
of Escherichia coli O157. PLoS One 7:e51572. 103. Yokoyama K, Makino K, Kubota Y, Watanabe
doi:10.1371/journal.pone.0051572. M, Kimura S, Yutsudo CH, Kurokawa K, Ishii

36 SCHEUTZ
K, Hattori M, Tatsuno I, Abe H, Yoh M, Iida T, coli infections: translocation, translocation, trans-
Ohnishi M, Hayashi T, Yasunaga T, Honda T, location. Infect Immun 73:25732585.
Sasakawa C, Shinagawa H. 2000. Complete 111. Bono JL, Keen JE, Clawson ML, Durso LM,
nucleotide sequence of the prophage VT1-Sakai Heaton MP, Laegreid WW. 2007. Association of
carrying the Shiga toxin 1 genes of the entero- Escherichia coli O157:H7 tir polymorphisms with
hemorrhagic Escherichia coli O157:H7 strain de- human infection. BMC Infect Dis 7:98.
rived from the Sakai outbreak. Gene 258:127139. 112. Haugum K, Brandal LT, Lobersli I, Kapperud
104. Sato T, Shimizu T, Watarai M, Kobayashi M, G, Lindstedt BA. 2011. Detection of virulent
Kano S, Hamabata T, Takeda Y, Yamasaki S. Escherichia coli O157 strains using multiplex PCR
2003. Genome analysis of a novel Shiga toxin 1 and single base sequencing for SNP character-
(Stx1)-converting phage which is closely related ization. J Appl Microbiol 110:15921600.
to Stx2-converting phages but not to other Stx1- 113. Badea L, Doughty S, Nicholls L, Sloan J,
converting phages. J Bacteriol 185:39663971. Robins-Browne RM, Hartland EL. 2003. Con-
105. Jacobsen A. 2012. M.Sc. thesis. Statens Serum tribution of Efa1/LifA to the adherence of en-
Institut, Technical University of Denmark, Copen- teropathogenic Escherichia coli to epithelial cells.
hagen, Denmark. Microb Pathog 34:205215.
106. Perna NT, Plunkett G 3rd, Burland V, Mau B, 114. Klapproth JA, Scaletsky ICA, McNamara BP,
Glasner JD, Rose DJ, Mayhew GF, Evans PS, Lai LC, Malstrom C, James SP, Donnenberg
Gregor J, Kirkpatrick HA, Posfai G, Hackett J, MS. 2000. A large toxin from pathogenic Esche-
Klink S, Boutin A, Shao Y, Miller L, Grotbeck EJ, richia coli strains that inhibits lymphocyte acti-
Davis NW, Lim A, Dimalanta ET, Potamousis vation. Infect Immun 68:21482155.
KD, Apodaca J, Anantharaman TS, Lin J, Yen G, 115. Wickham ME, Lupp C, Mascarenhas M,
Schwartz DC, Welch RA, Blattner FR. 2001. Vazquez A, Coombes BK, Brown NF, Coburn
Genome sequence of enterohaemorrhagic Esche- BA, Deng W, Puente JL, Karmali MA, Finlay
richia coli O157:H7. Nature 409:529533. BB. 2006. Bacterial genetic determinants of non-
107. Blanco M, Blanco JE, Dahbi G, Alonso MP, Mora O157 STEC outbreaks and hemolytic-uremic syn-
A, Coira MA, Madrid C, Juarez A, Bernardez MI, drome after infection. J Infect Dis 194:819827.
Gonzalez EA, Blanco J. 2006. Identification of 116. Newton HJ, Pearson JS, Badea L, Kelly M,
two new intimin types in atypical enteropatho- Lucas M, Holloway G, Wagstaff KM, Dunstone
genic Escherichia coli. Int Microbiol 9:103110. MA, Sloan J, Whisstock JC, Kaper JB, Robins-
108. Garrido P, Blanco M, Moreno-Paz M, Briones Browne RM, Jans DA, Frankel G, Phillips AD,
C, Dahbi G, Blanco J, Blanco J, Parro V. Coulson BS, Hartland EL. 2010. The type III
2006. STEC-EPEC oligonucleotide microarray: a effectors NleE and NleB from enteropathogenic
new tool for typing genetic variants of the LEE E. coli and OspZ from Shigella block nuclear
pathogenicity island of human and animal Shiga translocation of NF-kappaB p65. PLoS Pathog
toxin-producing Escherichia coli (STEC) and en- 6:e1000898. doi:10.1371/journal.ppat.1000898 [doi].
teropathogenic E. coli (EPEC) strains. Clin Chem 117. Tree JJ, Wolfson EB, Wang D, Roe AJ, Gally
52:192201. DL. 2009. Controlling injection: regulation of type
109. Mora A, Blanco M, Yamamoto D, Dahbi G, III secretion in enterohaemorrhagic Escherichia
Blanco JE, Lopez C, Alonso MP, Vieira MA, coli. Trends Microbiol 17:361370.
Hernandes RT, Abe CM, Piazza RM, Lacher 118. Shaikh N, Tarr PI. 2003. Escherichia coli O157:H7
DW, Elias WP, Gomes TA, Blanco J. 2009. Shiga toxin-encoding bacteriophages: integrations,
HeLa-cell adherence patterns and actin aggrega- excisions, truncations, and evolutionary implica-
tion of enteropathogenic Escherichia coli (EPEC) tions. J Bacteriol 185:35963605.
and Shiga-toxin-producing E. coli (STEC) strains 119. Recktenwald J, Schmidt H. 2002. The nucleo-
carrying different eae and tir alleles. Int Microbiol tide sequence of Shiga toxin (Stx) 2e-encoding
12:243251. phage phiP27 is not related to other Stx phage
110. Garmendia J, Frankel G, Crepin VF. 2005. En- genomes, but the modular genetic structure is
teropathogenic and enterohemorrhagic Escherichia conserved. Infect Immun 70:18961908.

Shiga Toxin (Stx) Classication,
Structure, and Function
ANGELA R. MELTON-CELSA1
3
Shiga toxin (Stx) is one of the most potent biological poisons known. Stx causes
fluid accumulation in rabbit ileal loops; causes renal damage in mice, rabbits,
greyhounds, and baboons; and is lethal to animals upon injection. However,
humans encounter Stx as a consequence of infection with Shigella dysenteriae
type 1 or certain serogroups of Escherichia coli such as O157:H7. There are
two immunologically distinct groups of Stxs, and this review discusses toxin
classification, structure, and function and the virulence associated with Stx-
producing E. coli (STEC).
OVERVIEW
S. dysenteriae and Stx were identified in the 19th century by Neisser and Shiga
(1) and Conradi (2). Approximately 80 years later, the same toxin (now called
Stx1 to distinguish it from the toxin produced by S. dysenteriae) was found in a
group of E. coli isolates. These bacteria caused bloody diarrhea and a serious
sequela, hemolytic-uremic syndrome (HUS), a condition characterized by
thrombocytopenia, hemolytic anemia, and kidney failure (3, 4). Some E. coli
strains were later shown to produce a highly related toxin, Stx2, that has the
1
Department of Microbiology and Immunology, Uniformed Services University of the Health Sciences, Bethesda,
MD 20814.
37

38 MELTON-CELSA
same mode of action as Stx/Stx1 but is im- TYPING AND NOMENCLATURE

munologically distinct. The Stxs (also known
as verotoxins and previously as Shiga-like Although the prototype E. coli Stxs from each
toxins) are a group of bacterial AB5 protein main group, Stx1 and Stx2 (now called Stx1a
toxins of about 70 kDa that inhibit protein and Stx2a for distinction in the nomenclature
synthesis in sensitive eukaryotic cells. Protein from other toxin subtypes [10]), are the most
synthesis is blocked by the Stxs through the common types found in association with dis-
removal of an adenine residue from the 28S ease from their respective groups, subtypes
rRNA of the 60S ribosome. This N-glycosidase of each toxin exist, as listed in Table 1. Toxin
activity of the toxin resides in the A subunit. subtypes were originally only recognized when
The pentamer of identical B subunits mediates differences in biological activity and/or im-
toxin binding to the cellular receptor globo- munoreactivity could be demonstrated. How-
triaosylceramide (Gb3). Additional common- ever, as many new STEC strains were isolated
alities between the Stx groups are that the and the toxin genes from those strains were
subunit genes are encoded in an operon with sequenced, it became difficult to know if any
the A subunit gene 5 to that of the B sub- differences found between the newly isolated
unit, that the stx operon is usually found gene and the prototype stx2a should result
within the sequence for an inducible, lyso- in the designation of another toxin subtype.
genic, lambda-like bacteriophage, and that Therefore, a phylogenetic analysis of stx se-
the toxins use a retrograde pathway to reach quences was undertaken and a PCR typing
the cytoplasm. Differences between the two scheme developed that enables the assignment
toxin groups are that the genes for stx/stx1a of a toxin to a particular subtype (10).
are repressed by Fur when high levels of iron
are present (57) and that E. coli strains
Stx/Stx1 Subtypes
that encode stx2 are epidemiologically linked
to more severe disease than those that carry To date, no variants of Stx as produced
stx1 (8, 9). by Shigella have been described, but Stx is
TABLE 1 Prototype toxins and strains that produce those toxins

Prototype strain used for Linked with serious human disease;
Toxin type(s) determination of stx subtypea difference(s) from prototype toxinb Reference(s)
Stx 3818T Yes 133

Stx1a EDL933 (makes Stx1a and Stx2a) Yes 113
Stx1c DG131/3 No; immunologically distinct 134, 135
Stx1d MHI813 No; immunologically distinct, less potent 136
Stx2a EDL933 (makes Stx1a and Stx2a) Yes 113
Stx2b (originally EH250 No; the B subunit gene was not detected 24
named VT-2d by methods used to detect other
or Stx2d) sx2 B subunit genes
Stx2c 031 Yes, less toxic to Vero cells and mice 137
Stx2d (Stx2dact) C165-02 Yes; more toxic after incubation with 138
elastase, less toxic to Vero cells
Stx2e S1191 No; binds to Gb4, associated with disease 139
in pigs
Stx2f T4/97 No; originally isolated in STEC from 140
pigeons; immunologically distinct
Stx2g 7v No; the stx2g gene is not amplied by 141
primers specic for stx2a
a
More information about the prototypes strains may be found in reference 10.
b
Prototype toxin indicated in bold.

CHAPTER 3 Stx Classication, Structure, and Function 39
occasionally found in Shigella sonnei and type when injected into mice (23). A different
4 S. dystenteriae (11, 12). Only two variants of Stx2a subtype that was originally also named
Stx1a have been identified: Stx1c and Stx1d. Stx2d, now known as Stx2b, is not activatable
Both Stx1c and Stx1a can be distinguished and is associated with mild disease (24, 25).
immunologically from Stx1 (13, 14). Stx1c and Stx2e, Stx2f, and Stx2g are associated with
Stx1d are rarely found in human disease, and animal STEC infection. Of the latter toxins,
when associated with STEC isolated from only Stx2e is associated with disease in the
patients, are linked with a mild disease course animal host; this toxin causes edema disease
(15, 16). of swine, a rare but serious neurological dis-
order that is frequently fatal (26).
Stx2a Subtypes
The first Stx2a toxin variant identified as im- SHIGA TOXIN STRUCTURE
portant for human disease, Stx2c, exhibits re-
duced cytotoxicity on Vero cells and reacts The mature A subunit of Stx/Stx1a consists
differently than Stx2a to some monoclonal of 293 aa whereas the Stx2a A chain is 4
antibodies (17). Another Stx2a variant, Stx2d aa longer at the C terminus. The active-site
(Stx2d activatable), was identified because in- amino acid of the toxin is the glutamic acid at
cubation with elastase from intestinal mucus position 167 (27). A trypsin-sensitive region (aa
increases the Vero cell cytotoxicity of the 248251) allows the A subunit to be cleaved
toxin (18, 19). The activatable Stx2d is associ- asymmetrically into an A1 subunit and A2 pep-
ated with the most serious manifestation of tide held together by a disulfide bridge (Fig. 1).
STEC infection, HUS (20). Both Stx2c and The enzymatic activity of the toxin resides
Stx2d show reduced cytotoxicity for Vero within A1 while the A2 peptide tethers A1
cells due to 2 amino acid differences in the B to the binding moiety, and further, threads
subunit, but Stx2d is as toxic as Stx2a when through the B pentamer. For Stx2a, the A2
injected into animals (21). Moreover, strains peptide, in addition to maintaining holotoxin
that produce Stx2d are highly virulent in a structure, appears to partially block the site 3
streptomycin-treated mouse model of infec- Gb3-binding site (discussed further below)
tion (18, 22). In contrast, Stx2c is reported (28), and for Stx2d, is crucial for the activa-
to show reduced toxicity compared to Stx2 tion phenotype, as the final 2 aa of the A2 are
FIGURE 1 Cartoon representation of the Stx structure. The active-site glutamic acid is indicated as a vertical
blue line, the ribosome interaction region is shown in purple, the protease (furin)- sensitive site is depicted in
green, and the B pentamer as an orange block. The disulde bridge that connects the A1 subunit and the A2
peptide is shown above the protease-sensitive site. A region important for translocation from the ER to the
cytosol is indicated by a bracket. Not to scale. doi:10.1128/microbiolspec.EHEC-0024-2013.f1

40 MELTON-CELSA
cleaved when the toxin is treated with elastase include N75, Y77, Y114, R170, and W203
(29). The pentamer of identical B subunits (W202 in Stx2) (27, 3638). An analysis of
(each subunit equals 69 aa for Stx/Stx1a, 71 truncated A1 fragments of Stx1a in yeast con-
aa for Stx2a) enables the toxin to find target firmed that residues within aa 1239 are re-
cells that express surface Gb3. The crystal quired for full enzymatic activity of the toxin,
structures of the Stx B pentamer (30), Stx (31), whereas the amino acids from 240 to 245 (and
Stx2a (28), Stx2a complexed with adenine perhaps up to 251) are necessary for translo-
(32), and the solution structures of the Stx1a B cation of the A1 from the endoplasmic retic-
pentamer alone (33) or with the trisaccharide ulum (ER) to the cytosol (39). The authors of
moiety from Gb3 (34) or a Gb3 analog (35) the latter study hypothesized that it is the
have been reported. A ribbon diagram of the general structure of the region between amino
crystal structure of Stx is shown in Fig. 2. acid residues 240 and 251 that is recognized
by an ER mechanism that directs proteins
from the ER into the cytosol. To define re-
Genetic Analyses of the Stx A Subunit
gions of Stx1a and the ribosome that interact,
Besides the active-site glutamic acid, colored Gariepys group used yeast two-hybrid and
red in Fig. 2, other A subunit amino acid pull-down experiments and identified three
residues that contribute to the full enzymatic ribosomal proteins that interact with the A1
function of Stx, colored pale blue in Fig. 2, subunit of the toxin and further showed that a
FIGURE 2 Ribbon diagram of the Stx1 crystal structure. The B pentamer is shown in orange and the A2 in
blue. The majority of the A1 is depicted in green except for the region that interacts with the ribosome, which
is shown in purple. The active residue 167 is red, and other active-site side chains are pale blue. The A2 chain
is medium blue, and the B subunits are orange. The structure (1R4Q) was drawn with PyMOL Molecular
Graphics System, Version 1.5.0, Schrdinger, LLC. Figure kindly provided by Dr. James Vergis.

conserved peptide from two of the ribosomal for cholesterol in cell binding by the B subunit
proteins inhibited Stx1a activity in vitro (40). of Stx1a, but they did find that reductions
They also found that the region from R170 to in the cholesterol content of cells resulted in
L233 (purple in Fig. 2) in Stx1a is important a decrease in uptake of the B pentamer (55).
for ribosome interaction (41). More recently, Lingwoods group found that
extraction of cholesterol from adult renal tissue
sections enhanced Stx binding to glomeruli
Stx AND CELL BINDING (59), a result that confirms the colocalization
of Gb3 and cholesterol molecules. However,
As noted previously, the Stxs bind to the gly- it may be difficult to dissect a possible role for
cosphingolipid Gb3, a molecule composed of cholesterol because Gb3 and cholesterol are
a lipid or ceramide component and a trisac- both present in the lipid rafts, and alteration of
charide of [a-gal(14)-b-gal(14)-b-glc] (42 either component may disrupt the integrity
44). Cell lines and mice deficient in Gb3 are of the rafts. Together, these studies paint a
insensitive to the Stxs (4547). The normal picture of a complex interaction of the toxin
cellular function of Gb3 is not known; how- with the host cell receptor and suggest that the
ever, individuals with excess Gb3, a condi- Stxs may interact differently with cells based
tion called Fabrys disease (48), exhibit kidney on the nature of the receptor environment in
disease among other symptoms. When chal- the cell membrane.
lenged with Stx2a, a mouse model of Fabrys The high affinity of the Stxs for Gb3 is
disease exhibits an elevated LD50 (49), perhaps likely due to the presence of at least two and
due to mistargeting of the toxin to multiple up to three Gb3-binding sites per B monomer
sites of the body rather than the kidney or to (two of the binding sites are present between
altered cellular trafficking due to changes in adjacent monomers), as demonstrated by mod-
Gb3 subspecies populations. Stx1a and Stx2a eling studies and the crystal structure of the
interact with globotetraosylceramide (Gb4) in Stx B pentamer complexed with a Gb3 tri-
addition to Gb3, though weakly (50), whereas saccharide analog (35, 60). However, precise
Stx2e binds Gb4 in preference to Gb3 (51, 52). measurements of the binding affinity of the
An early report of a possible protein receptor Stxs for Gb3 are difficult because Gb3 is not
for the Stxs (53) has not been substantiated in soluble. Therefore, soluble forms of the tri-
the literature. saccharide or trisaccharide analogs, or im-
Gb3 clusters within detergent-insoluble por- mobilized versions of the trisaccharide, Gb3,
tions of membranes called lipid rafts, which or Gb3 analogs are used to measure toxin-
also contain cholesterol and the cholera toxin receptor interaction. When the interaction
receptor monosialotetrahexosylganglioside, or between the Stx1a B pentamer and the Gb3
GM1 (54, 55). The interaction of the Stx1a B trisaccharide was assessed in solution, bind-
subunit with HeLa cells causes a 2.5-fold in- ing site 2 (present within each monomer) was
crease in Gb3 present in the lipid raft domains found to be the most highly occupied; site
(56), a result that suggests that cell binding 1 was less engaged, and no interaction with
by the B pentamer may promote stronger or site 3 was detected (34). Another study that
additional toxin-cell interaction by recruiting also examined the interaction of the B penta-
more receptors to the rafts. Several reports in- mer with the Gb3 oligosaccharide component
dicate that the fatty acid chain length of the found that, at low concentrations of ligand,
lipid component of Gb3 and the saturation state only site 2 was occupied though the authors
of the lipid also influence toxin binding (57, 58). of that study acknowledged that Stx binding
The presence of cholesterol in the lipid rafts is likely to be polyvalent at the cell surface
has been reported to increase toxin binding (61). A mutational analysis of the three puta-
(50). Conversely, another group found no role tive Gb3-binding sites in Stx1a confirmed the

42 MELTON-CELSA
role of sites 1 and 2 for toxin-receptor inter- others have shown that incubation of intesti-
action (62). A mutation within site 3 (W34A) nal cells with butyrate increases the sensitivity
reduced binding of the holotoxin to Gb3, al- of intestinal cell lines to the Stxs (72, 73), and
though the mutant toxin had the same overall we further found that expression of Gb3 on
affinity for Gb3 as did Stx1a. Another group intestinal cells is exquisitely sensitive to the
found that Gb3-binding site 2 alone is not removal of that gut metabolite (74). This latter
sufficient to confer high avidity binding of the finding suggests that previous measurements
toxin to Gb3 and that optimal binding to the of Gb3 levels on intestinal tissues may be under-
receptor required all three sites (63). Although estimates if butyrate was removed before Gb3
these latter studies on the number of func- detection. In support of the possible presence
tional Gb3-binding sites per toxin molecule of functional Gb3 receptors on intestinal tis-
were performed with Stx/Stx1a, the presence sue is the finding that incubation of pediatric
and primary importance of site 2 were con- intestinal explants with Stx2a showed cell
firmed for Stx2a (64). In addition, that site 1 extrusion from tissues taken from both ileum
contributes to the Vero cell toxicity pheno- and colon (71). The damage observed in the
type of the Stx2a group is demonstrated by latter study was specific as pre-incubation of
the fact that a single amino acid change in the the toxin with antiserum to Stx2a ameliorated
Stx2d B subunit in the site 1 binding region the tissue injury. Furthermore, both Stx1a and
renders that toxin as potent for Vero cells as Stx2a bind to Paneth cells when overlaid onto
Stx2a (21). Finally, a point in favor of the normal or inflamed pediatric duodenal tissues
polyvalent nature of Stx binding is that Gb3 collected during endoscopy (75). We also found
mimics with higher densities and clustering that toxin overlaid into normal adult tissue binds
of the Gb3 trisaccharide (up to 18 copies) colonic tissue (72). Finally, Malyukova et al. iden-
demonstrate higher affinities for both Stx1a tified both Stx1a and Stx2a in intestinal tissues
and Stx2a (65, 66). (epithelial cells and lamina propria) from O157-
The question of how Stx enters the host infected patients (68), although Gb3 could not
from the site of pathogen colonization in the be found in those intestinal epithelial cells (68).
intestine remains, due to reports that colonic However, it may be that the Gb3 is masked in
tissue lacks Gb3 (67). Macropinocytosis is a such tissues by cholesterol, as described earlier,
possible alternate mechanism for the uptake or that the expression of Gb3 was reduced by
of Stxs into cells that do not express Gb3 (68). removal of butyrate when those samples were
Polarized colonic T84 cells (which lack Gb3) processed.
nonetheless take up the Stx1a B pentamer in Stx-Gb3 interaction leads to uptake of
a way that is partially blocked by inhibitors of the toxin-receptor complex through clathrin-
clathrin-dependent endocytic processes (68). dependent or -independent mechanisms. How-
However, such inhibition of the macropino- ever, the clathrin-dependent process of entry
cytic pathway does not reduce the amount appears to be the most common mode of up-
of B subunit that is transcytosed through take of the toxin-receptor complex (76). One
the monolayer, a finding that suggests that perhaps surprising finding is that the A sub-
no matter which entry mechanism the Stx unit of the Stx is involved in toxin uptake:
uses to enter these cells, the toxin can reach more holotoxin is taken up via the endocytic
the basolateral side. That both Stx1a and Stx2a pathway when the holotoxin binds to a cell as
cross polarized T84 cells without disrupting compared to B pentamer alone (77). The latter
the monolayer (46, 69, 70) or inhibiting pro- study illustrates the importance of confirming
tein synthesis has also been demonstrated (71). findings based on the B pentamer alone with
Therefore, it may be that some Gb3-negative the holotoxin. In clathrin-independent up-
intestinal cells allow systemic delivery of toxin take, the Stx B subunit induces tubular-shaped
in the absence of receptor. However, we and invaginations within the HeLa cell plasma

membrane, the function of which is not clear

(78). Curiously, a mutation in Gb3-binding site
3 of the B subunit prevented formation of the
invaginations (78), although the pentamer still
bound to the cells.
RETROGRADE TRAFFICKING OF THE Stxs
Stx was first shown to use a retrograde path-

way to reach the cytosol more than 20 years
ago in the human epidermal carcinoma cell
line A431, which had been sensitized to the
toxin by butyric acid treatment (79). An out-
line of the retrograde pathway used by the
Stxs is shown in Fig. 3 and summarized
as follows: after binding to Gb3, the toxin-
receptor complex enters early endosomes,
then traffics to the Golgi, and finally the ER
(for a review see reference 80). The protease-
sensitive site of the toxin A subunit is nicked
either within the intestinal mucus (18) or
within the Golgi (81); however, the nicked
toxin retains the AB5 structure because of a
disulfide bond between the A1 and A2 chains.
The disulfide bond is reduced once the toxin
gets into the ER, and only the toxin A1 subunit
enters the cytosol. The Stx receptor, Gb3, is FIGURE 3 An illustration of the retrograde pathway
required for the toxin-receptor complex to for Stxs. The toxin binds to Gb3 within lipid rafts that
move through the retrograde pathway, as toxin contain cholesterol and that complex is internalized
taken up by a nonreceptor-mediated mecha- within an endosome. From the endosome the toxin
trafcs to the Golgi where it is nicked by furin if that
nism or in cells depleted of glycosphingolipids nicking did not occur in the intestine. The nicked
does not reach the Golgi, though the complex toxin moves to the ER where the disulde bridge
does reach endosomes and possibly the ER (56, that keeps the A1 tethered to A2B5 is reduced. The A1
70, 82). Many cellular proteins have now been chain then enters the cytosol and removes an ade-
identified that are involved in the Stx retro- nine residue from the 28S ribosome.
grade pathway (reviewed in references 80, 83).
An unfolded toxin A1 chain exits the ER, ap- cell, discussed further below, and the Stx B pen-
parently by subverting the ER-associated pro- tamer has been shown to enter the retrograde
tein degradation pathway, to reach the target pathway in mesangial and glomerular vascular
ribosomes (84, 85). The A1 then removes an endothelial cells as well (88).
adenine from the alpha-sarcin loop in the 28S
ribosomal subunit. The injured ribosome no
longer associates with elongation factor 1 (86, ACTIONS OF Stx IN TARGET CELLS
87), and protein synthesis is halted. Although
much of the work describing the retrograde Numerous studies show that Stx inhibits pro-
pathway was done in epithelial cells, the pri- tein synthesis in target cells and that active-
mary target cell for the Stxs is the endothelial site mutants of Stx are no longer cytotoxic.

44 MELTON-CELSA
Furthermore, a single molecule of Stx may umbilical vein endothelial cells were equiva-
be sufficient to kill a cell (85). Recently, how- lently sensitive (100). Since the Stxs are
ever, the effects that the toxin has on cell equally enzymatically active and they both
signal transduction and immune modulation bind Gb3, the question remains why different
have begun to be explored (reviewed in ref- cell types exhibit differential sensitivity to
erences 89, 90). Stx-mediated damage to the the two Stx types. However, Gb3 does not
ribosome induces a response in cells called a exist in a homogeneous population within the
ribotoxic stress response, which is both pro- cell, and it may be that the toxins do not bind
inflammatory and proapoptotic (see review in equivalently to the Gb3 in target cells. An in
reference 89). The Stxs are also associated vitro study with soluble forms of the trisac-
with an unfolded protein response that comes charide moiety of Gb3 showed that Stx1a
about through stress on the ER, and the result and Stx2a bind with similar affinity for that
of which may also be apoptotic (reviewed in component of the receptor (64). Furthermore,
references 89, 91). Evidence that Stx and STEC Stx1a and Stx2a demonstrate similar affini-
are associated with apoptosis came from rabbit ties for renally derived Gb3, as measured by
studies in the mid-1980s (92, 93). Since that enzyme-linked immunosorbent assay, though
time, the molecular mechanisms of apoptosis by thin-layer chromatography Stx2a exhibits
due to the Stxs have been elucidated further less binding than Stx1a does (101). Examina-
(see review in reference 91). tions of toxin interaction with immobilized
trisaccharides suggest that Stx1a has a higher
affinity for the carbohydrate than does Stx2a
Stx1a COMPARED TO Stx2a (50, 102). Other experiments that examined
the capacity of the toxins to bind Gb3 analogs
Sx1a is about 10-fold more cytotoxic to Vero showed that Stx1a and Stx2a exhibit differ-
cells than Stx2a; however, the reverse is true ential binding to those analogs in vitro (103),
in mice: Stx1a is 100- to 400-fold less lethal but whether similar analogs exist or act func-
to mice than Stx2a, even though the toxins tionally as receptors in vivo has not been
exhibit equivalent enzymatic activities/ng of demonstrated.
protein in vitro (94, 95). Epidemiological data Membrane cholesterol may also mask a
from human disease indicate a stronger asso- portion of the potential Stx1a-binding Gb3
ciation with severe disease for STEC strains pool in Vero cell membranes (104) and adult
that produce Stx2a than Stx1a alone (9698). renal glomeruli sections, as mentioned previ-
So an important question within the STEC ously (59), and incubation of Stx1a but not
field is how to explain the paradox of the dif- Stx2a with Gb3, cholesterol, and phosphati-
ferential toxicity of the Stxs in vitro as com- dylcholine partially neutralizes the cytotox-
pared to in vivo. Russo et al. recently found icity of Stx1a but not Stx2a (103). Stx2a, in
that Stx2a is more toxic than Stx1a by the oral contrast, is neutralized by human serum am-
route in mice, data that support the hypothesis yloid component P and lipopolysaccharide
that Stx2a is more potent from the gut and (LPS) from O107 or O117 E. coli (105, 106).
not just when injected intraperitoneally (99). That the above compounds neutralize either
One possible explanation for the differential Stx1a or Stx2a specifically, and presumably
toxicity of the toxins is that in contrast to epi- by preventing the toxins from binding to the
thelial cells such as Vero, the endothelial cells cell, supports the hypothesis that the B pen-
targeted in HUS are more sensitive to Stx2a tamers of these toxins bind differently to cells.
than to Stx1a. Obrigs laboratory did find that The Stx1a B pentamer is more stable than
renal microvascular endothelial cells obtained the comparable binding moiety of Stx2a (64),
from human glomeruli are about 1,000-fold and that difference was shown to be at least
more sensitive to Stx2a than to Stx1a, although partially due to a single amino acid residue

within the B subunit (107). A study that ex- Stx ASSOCIATION WITH DISEASE
amined the interaction of Stx1a and Stx2a
with lipid rafts and subsequent intracellular HUS cases were reported in the literature
trafficking in Vero cells showed that although as early as the 1950s, though the cause was
both toxins associated with lipid rafts, Stx2a unknown. An understanding of the origins
was also found in detergent-soluble (non-lipid of infection-associated HUS was also compli-
raft) fractions of the membrane (108). In ad- cated by the fact that some cases of HUS are
dition, some of the internalized Stx2a, but not of genetic origin (see review in reference 112).
Stx1a, colocalized with the cell surface marker However, in 1983, bloody diarrhea and HUS
transferrin, a finding that suggests that Stx2a is were linked to certain serogroups of E. coli,
endocytosed from different areas of the mem- and OBrien et al. found that those strains
brane than Stx1a. However, both toxins localized produced a toxin related to the Stx of S. dys-
to the Golgi after 1 h of incubation with the Vero enteriae (3, 4, 113). The strong association
cells, although there was a suggestion that Stx2a between HUS and Stx2a was underscored in
exited from the Golgi at a slower rate than Stx1a. the largest outbreak of HUS in Germany in
At the level of the immune response to the 2011 in which more than 800 cases of the
toxins, Stx1a and Stx2a elicit differential che- disease were identified (114). The surprise
mokine responses from human umbilical vein from German outbreak was that although the
endothelial cells (109) and a renal tubular implicated strain was an E. coli, unlike typical
epithelial cell line, HK-2 (110). The HK-2 STEC, the epidemic isolate also encoded vir-
cells are more sensitive to Stx1a than to Stx2a, ulence factors associated with enteroaggre-
and Stx2a treatment induces expression of gative E. coli. Previously, enteroaggregative
the macrophage chemoattractants macrophage E. coli had only rarely been associated with
inflammatory protein-1a (MIP-1a) and MIP- HUS, and in each of those cases, the impli-
1b (110). However, in a baboon intoxication cated isolate was found to carry an Stx gene.
model, there was no change in MIP-1b in In animal models, purified Stx in the ab-
response to either Stx1a or Stx2a (111). None- sence of detectable LPS is linked to kidney
theless, in that baboon model many charac- damage and death (99, 115). S. dysenteriae mu-
teristics of HUS are observed, and Stx2a is tants that lack stx caused a milder dysentery
lethal at lower doses than Stx1a is, although in monkeys and produced only a trace of
the kidney injury takes longer to develop in the blood in stool compared to the wild-type iso-
Stx2a-treated animals. late (116). Furthermore, Stx can be detected in
Taken together these studies on the dif- the kidneys of some patients (117119). Animal
ferences between Stx1a and Stx2a indicate models further show that antibodies to the
that the interaction between each toxin and Stxs are protective, and humanized versions
the receptor is complex and influenced by of those antibodies have completed phase I
several factors. Four distinctions between the trials (120, 121).
Stx and Stx2a crystal structures that may con-
tribute to the known biological differences
STEC Pathogenesis
exhibited by these toxins, as discussed above,
are: (i) the Stx active site is partially blocked STEC pathogenesis requires ingestion of the
by the N-terminal region of the A2 peptide; bacterium in contaminated food or water,
(ii) the Stx2a A2 C terminus has a more or- though the organism may occasionally be
dered structure than that of Stx; (iii) Gb3- passed person to person. To cause disease,
binding site 2 has a different conformation in the organism colonizes the intestine, a process
each of the toxins; and (iv) Gb3-binding site 3 that likely includes attachment and prolifera-
is partially blocked by the C-terminal 2 aa of tion. That STEC strains that carry the locus of
the Stx2a A2 peptide (28). enterocyte effacement (LEE) use the adhesin

46 MELTON-CELSA
intimin (encoded by the eae gene) to adhere extrusion (see reviews in references 126, 127).
within the intestine has been shown in mouse Such damaged endothelial cells also become a
and pig models (122, 123). In addition, STEC site for the adhesion of leukocytes, the con-
strains that are eae positive are more likely sequence of which is additional damage to
to cause disease than intimin-negative strains, the kidney (128). The localized prothrombotic
even in non-O157 strains (8, 124). Zumbrun environment in the glomerulus leads to the
et al. recently demonstrated in mice that a reduction of platelets in the serum (throm-
high-fiber diet may influence the development bocytopenia), hemolytic anemia, and renal
of disease after O157 infection (74, 125). The damage, the triad of HUS (for reviews see
influence of the higher fiber diet was 2-fold references 126, 129, 130). Altered complement
and was associated with the production of levels have been found in patients with STEC
butyrate within the intestine: mice on a high- HUS (131, 132), but whether those changes are
fiber diet were colonized to a greater degree part of the etiology of the disease or occur as a
than mice on a low-fiber diet, perhaps as a consequence is not clear (see reference 127 for
consequence of reduced overall levels of com- a review of the possible role for complement
petitive native Escherichia species, and in- in STEC HUS).
creased levels of Gb3 were observed in both
the intestine and kidney.
After colonization with STEC, the innate SUMMARY
proinflammatory response may initially be sup-
pressed in those infected with strains that are The Stxs are potent poisons associated with
LEE+ (see review in reference 90), a conclu- bloody diarrhea and the potentially severe dis-
sion supported by the fact that STEC infection ease, HUS. The Stxs act as ribotoxins that halt
is generally not an inflammatory diarrhea. In protein synthesis within the cell and induce
the intestine, the STEC strains elaborate Stxs, apoptosis, but they can also prompt altered gene
and the patient may develop hemorrhagic or protein expression in epithelial cells, endo-
colitis, the pathogenesis of which is not clearly thelial cells, monocytes, and mesangial cells.
understood. For systemic consequences, the The consequence of exposure to E. coli strains
Stxs enter the host from the intestine and that elaborate one or more Stx can range from
reach target endothelium cells in the kidney asymptomatic infection to bloody diarrhea and,
and, in some cases, the central nervous sys- in some patients, HUS. In the research arena,
tem. The Stxs damage cells directly and cause the Stxs have been exploited beautifully to in-
apoptosis, as described above. The Stxs and terrogate mechanisms of retrograde transport,
likely LPS also elicit cytokines and other im- to localize sites of Gb3 expression, and, more
mune mediators within the kidney and from recently, to determine the possible function of
monocytes (for a review see reference 90). tubular invaginations in membranes.
Endothelial cells within the glomeruli become
damaged and may die and detach from the
ACKNOWLEDGMENTS
membrane. Extrusion of the cell exposes col-
lagen, normally not present within the blood A special thank-you to Alison D. OBrien for
vessel, and platelets become activated. The review of this chapter. This work was sup-
presence of activated platelets encourages fi- ported by R37 AI020148 and 2U54 AI057168.
brin deposition and leads to the development
of a thrombus and promotes a coagulation
CITATION
cascade. Recent studies suggest that Stx treat-
ment of endothelial cells can induce the ex- Melton-Celsa AR. 2014. Shiga toxin (Stx) clas-
pression of factors that may stabilize clot sification, structure, and function. Microbiol
formation, even in the absence of cell death or Spectrum 2(4):EHEC-0024-2013.

REFERENCES 13. Zhang W, Bielaszewska M, Kuczius T, Karch H.

2002. Identification, characterization, and dis-
1. Trofa AF, Ueno-Olsen H, Oiwa R, Yoshikawa tribution of a Shiga toxin 1 gene variant (stx(1c))
M. 1999. Dr. Kiyoshi Shiga: discoverer of the in Escherichia coli strains isolated from humans.
dysentery bacillus. Clin Infect Dis 29:13031306. J Clin Microbiol 40:14411446.
2. Conradi H. 1903. Uber Iosliche, durch asptische 14. Ohmura-Hoshino M, Ho ST, Kurazono H,
Autolyse erhaltene Giftstoffe vonRuhr- und Typhus- Igarashi K, Yamasaki S, Takeda Y. 2003. Genetic
Bazillen. Dtsch Med Wochenschr 29:2628. and immunological analysis of a novel variant of
3. Karmali MA, Steele BT, Petric M, Lim C. 1983. Shiga toxin 1 from bovine Escherichia coli strains
Sporadic cases of haemolytic-uraemic syndrome and development of bead-ELISA to detect the
associated with faecal cytotoxin and cytotoxin- variant toxin. Microbiol Immunol 47:717725.
producing Escherichia coli in stools. Lancet i:619 15. Friedrich AW, Borell J, Bielaszewska M,
620. Fruth A, Tschape H, Karch H. 2003. Shiga toxin
4. OBrien AO, Lively TA, Chen ME, Rothman 1c-producing Escherichia coli strains: pheno-
SW, Formal SB. 1983. Escherichia coli O157:H7 typic and genetic characterization and association
strains associated with haemorrhagic colitis in with human disease. J Clin Microbiol 41:2448
the United States produce a Shigella dysenteriae 1 2453.
(SHIGA) like cytotoxin. Lancet i:702. 16. Kumar A, Taneja N, Kumar Y, Sharma M. 2012.
5. Calderwood SB, Mekalanos JJ. 1987. Iron reg- Detection of Shiga toxin variants among Shiga
ulation of Shiga-like toxin expression in Esche- toxin-forming Escherichia coli isolates from ani-
richia coli is mediated by the fur locus. J Bacteriol mal stool, meat and human stool samples in India.
169:47594764. J Appl Microbiol 113:12081216.
6. Dubos RJ, Geiger JW. 1946. Preparation and 17. Schmitt CK, McKee ML, OBrien AD. 1991. Two
properties of Shiga toxin and toxoid. J Exp Med copies of Shiga-like toxin II-related genes com-
84:143156. mon in enterohemorrhagic Escherichia coli strains
7. OBrien AD, LaVeck GD, Thompson MR, are responsible for the antigenic heterogeneity of
Formal SB. 1982. Production of Shigella dys- the O157:H- strain E32511. Infect Immun 59:1065
enteriae type 1-like cytotoxin by Escherichia coli. 1073.
J Infect Dis 146:763769. 18. Melton-Celsa AR, Darnell SC, OBrien AD. 1996.
8. Luna-Gierke RE, Griffin PM, Gould LH, Herman Activation of Shiga-like toxins by mouse and
K, Bopp CA, Strockbine N, Mody RK. 2014. human intestinal mucus correlates with virulence
Outbreaks of non-O157 Shiga toxin-producing of enterohemorrhagic Escherichia coli O91:H21
Escherichia coli infection: USA. Epidemiol Infect isolates in orally infected, streptomycin-treated
7:111. mice. Infect Immun 64:15691576.
9. Friedrich AW, Bielaszewska M, Zhang WL, 19. Kokai-Kun JF, Melton-Celsa AR, OBrien AD.
Pulz M, Kuczius T, Ammon A, Karch H. 2002. 2000. Elastase in intestinal mucus enhances the
Escherichia coli harboring Shiga toxin 2 gene cytotoxicity of Shiga toxin type 2d. J Biol Chem
variants: frequency and association with clinical 275:37133721.
symptoms. J Infect Dis 185:7484. 20. Bielaszewska M, Friedrich AW, Aldick T,
10. Scheutz F, Teel LD, Beutin L, Pierard D, Schurk-Bulgrin R, Karch H. 2006. Shiga toxin
Buvens G, Karch H, Mellmann A, Caprioli A, activatable by intestinal mucus in Escherichia coli
Tozzoli R, Morabito S, Strockbine NA, Melton- isolated from humans: predictor for a severe
Celsa AR, Sanchez M, Persson S, OBrien AD. clinical outcome. Clin Infect Dis 43:11601167.
2012. Multicenter evaluation of a sequence-based 21. Lindgren SW, Samuel JE, Schmitt CK, OBrien
protocol for subtyping Shiga toxins and stan- AD. 1994. The specific activities of Shiga-like
dardizing Stx nomenclature. J Clin Microbiol toxin type II (SLT-II) and SLT-II-related toxins
50:29512963. of enterohemorrhagic Escherichia coli differ when
11. Gupta SK, Strockbine N, Omondi M, Hise K, measured by Vero cell cytotoxicity but not by
Fair MA, Mintz E. 2007. Emergence of Shiga mouse lethality. Infect Immun 62:623631.
toxin 1 genes within Shigella dysenteriae type 22. Lindgren SW, Melton AR, OBrien AD. 1993.
4 isolates from travelers returning from the Virulence of enterohemorrhagic Escherichia coli
Island of Hispanola. Am J Trop Med Hyg 76: O91:H21 clinical isolates in an orally infected
11631165. mouse model. Infect Immun 61:38323842.
12. Beutin L, Strauch E, Fischer I. 1999. Isolation of 23. Fuller CA, Pellino CA, Flagler MJ, Strasser JE,
Shigella sonnei lysogenic for a bacteriophage en- Weiss AA. 2011. Shiga toxin subtypes display
coding gene for production of Shiga toxin. Lancet dramatic differences in potency. Infect Immun
353:1498. 79:13291337.

48 MELTON-CELSA
24. Pierard D, Muyldermans G, Moriau L, Stevens R, Swanton E, High S, Spooner RA. 2011.
D, Lauwers S. 1998. Identification of new vero- Eeyarestatin 1 interferes with both retrograde and
cytotoxin type 2 variant B-subunit genes in anterograde intracellular trafficking pathways.
human and animal Escherichia coli isolates. J Clin PLoS One 6:e22713.
Microbiol 36:33173322. 37. Deresiewicz RL, Calderwood SB, Robertus JD,
25. Stephan R, Hoelzle LE. 2000. Characterization Collier RJ. 1992. Mutations affecting the activity
of Shiga toxin type 2 variant B-subunit in Esche- of the Shiga-like toxin I A-chain. Biochemistry
richia coli strains from asymptomatic human 31:32723280.
carriers by PCR-RFLP. Lett Appl Microbiol 31: 38. Di R, Kyu E, Shete V, Saidasan H, Kahn PC,
139142. Tumer NE. 2011. Identification of amino acids
26. Moxley RA. 2000. Edema disease. Vet Clin North critical for the cytotoxicity of Shiga toxin 1 and
Am Food Anim Pract 16:175185. 2 in Saccharomyces cerevisiae. Toxicon 57:525
27. Hovde CJ, Calderwood SB, Mekalanos JJ, 539.
Collier RJ. 1988. Evidence that glutamic acid 167 39. LaPointe P, Wei X, Gariepy J. 2005. A role for
is an active-site residue of Shiga-like toxin I. Proc the protease-sensitive loop region of Shiga-like
Natl Acad Sci USA 85:25682572. toxin 1 in the retrotranslocation of its A1 domain
28. Fraser ME, Fujinaga M, Cherney MM, Melton- from the endoplasmic reticulum lumen. J Biol
Celsa AR, Twiddy EM, OBrien AD, James MN. Chem 280:2331023318.
2004. Structure of Shiga toxin type 2 (Stx2) from 40. McCluskey AJ, Poon GM, Bolewska-Pedyczak
Escherichia coli O157:H7. J Biol Chem 279:27511 E, Srikumar T, Jeram SM, Raught B, Gariepy J.
27517. 2008. The catalytic subunit of Shiga-like toxin 1
29. Melton-Celsa AR, Kokai-Kun JF, OBrien AD. interacts with ribosomal stalk proteins and is in-
2002. Activation of Shiga toxin type 2d (Stx2d) by hibited by their conserved C-terminal domain.
elastase involves cleavage of the C-terminal two J Mol Biol 378:375386.
amino acids of the A2 peptide in the context of the 41. McCluskey AJ, Bolewska-Pedyczak E, Jarvik
appropriate B pentamer. Mol Microbiol 43:207215. N, Chen G, Sidhu SS, Gariepy J. 2012. Charged
30. Stein PE, Boodhoo A, Tyrrell GJ, Brunton JL, and hydrophobic surfaces on the a chain of
Read RJ. 1992. Crystal structure of the cell- Shiga-like toxin 1 recognize the C-terminal do-
binding B oligomer of verotoxin-1 from E. coli. main of ribosomal stalk proteins. PLoS One 7:
Nature 355:748750. e31191.
31. Fraser ME, Chernaia MM, Kozlov YV, James 42. Lindberg AA, Brown JE, Stromberg N,
MN. 1994. Crystal structure of the holotoxin from Westling-Ryd M, Schultz JE, Karlsson KA.
Shigella dysenteriae at 2.5 A resolution. Nat Struct 1987. Identification of the carbohydrate receptor
Biol 1:5964. for Shiga toxin produced by Shigella dysenteriae
32. Fraser ME, Cherney MM, Marcato P, Mulvey type 1. J Biol Chem 262:17791785.
GL, Armstrong GD, James MN. 2006. Binding of 43. Lingwood CA, Law H, Richardson S, Petric M,
adenine to Stx2, the protein toxin from Esche- Brunton JL, De Grandis S, Karmali M. 1987.
richia coli O157:H7. Acta Crystallogr Sect F Struct Glycolipid binding of purified and recombinant
Biol Cryst Commun 62:627630. Escherichia coli produced verotoxin in vitro. J Biol
33. Richardson JM, Evans PD, Homans SW, Chem 262:88348839.
Donohue-Rolfe A. 1997. Solution structure of the 44. Waddell T, Head S, Petric M, Cohen A,
carbohydrate-binding B-subunit homopentamer Lingwood C. 1988. Globotriosyl ceramide is spe-
of verotoxin VT-1 from E. coli. Nat Struct Biol cifically recognized by the Escherichia coli vero-
4:190193. cytotoxin 2. Biochem Biophys Res Commun 152:
34. Shimizu H, Field RA, Homans SW, Donohue- 674679.
Rolfe A. 1998. Solution structure of the complex 45. Jacewicz MS, Mobassaleh M, Gross SK,
between the B-subunit homopentamer of verotoxin Balasubramanian KA, Daniel PF, Raghavan S,
VT-1 from Escherichia coli and the trisaccharide McCluer RH, Keusch GT. 1994. Pathogenesis of
moiety of globotriaosylceramide. Biochemistry 37: Shigella diarrhea: XVII. A mammalian cell mem-
1107811082. brane glycolipid, Gb3, is required but not suffi-
35. Ling H, Boodhoo A, Hazes B, Cummings MD, cient to confer sensitivity to Shiga toxin. J Infect
Armstrong GD, Brunton JL, Read RJ. 1998. Dis 169:538546.
Structure of the Shiga-like toxin I B-pentamer 46. Acheson DW, Moore R, De Breucker S,
complexed with an analogue of its receptor Gb3. Lincicome L, Jacewicz M, Skutelsky E, Keusch
Biochemistry 37:17771788. GT. 1996. Translocation of Shiga toxin across
36. Aletrari MO, McKibbin C, Williams H, Pawar polarized intestinal cells in tissue culture. Infect
V, Pietroni P, Lord JM, Flitsch SL, Whitehead Immun 64:32943300.

47. Okuda T, Tokuda N, Numata S, Ito M, Ohta 59. Khan F, Proulx F, Lingwood CA. 2009.
M, Kawamura K, Wiels J, Urano T, Tajima O, Detergent-resistant globotriaosyl ceramide may
Furukawa K. 2006. Targeted disruption of Gb3/ define verotoxin/glomeruli-restricted hemolytic
CD77 synthase gene resulted in the complete de- uremic syndrome pathology. Kidney Int 75:1209
letion of globo-series glycosphingolipids and loss of 1216.
sensitivity to verotoxins. J Biol Chem 281:10230 60. Nyholm PG, Magnusson G, Zheng Z, Norel R,
10235. Binnington-Boyd B, Lingwood CA. 1996. Two
48. Tarabuso AL. 2011. Fabry disease. Skinmed 9:173 distinct binding sites for globotriaosyl ceramide
177. on verotoxins: identification by molecular mod-
49. Cilmi SA, Karalius BJ, Choy W, Smith RN, elling and confirmation using deoxy analogues
Butterton JR. 2006. Fabry disease in mice pro- and a new glycolipid receptor for all verotoxins.
tects against lethal disease caused by Shiga toxin- Chem Biol 3:263275.
expressing enterohemorrhagic Escherichia coli. 61. Thompson GS, Shimizu H, Homans SW,
J Infect Dis 194:11351140. Donohue-Rolfe A. 2000. Localization of the bind-
50. Nakajima H, Kiyokawa N, Katagiri YU, Taguchi ing site for the oligosaccharide moiety of Gb3 on
T, Suzuki T, Sekino T, Mimori K, Ebata T, verotoxin 1 using NMR residual dipolar coupling
Saito M, Nakao H, Takeda T, Fujimoto J. 2001. measurements. Biochemistry 39:1315313156.
Kinetic analysis of binding between Shiga toxin 62. Bast DJ, Banerjee L, Clark C, Read RJ, Brunton
and receptor glycolipid Gb3Cer by surface plas- JL. 1999. The identification of three biologically
mon resonance. J Biol Chem 276:4291542922. relevant globotriaosyl ceramide receptor binding
51. DeGrandis S, Law H, Brunton J, Gyles C, sites on the Verotoxin 1 B subunit. Mol Microbiol
Lingwood CA. 1989. Globotetraosylceramide is 32:953960.
recognized by the pig edema disease toxin. J Biol 63. Soltyk AM, MacKenzie CR, Wolski VM,
Chem 264:1252012525. Hirama T, Kitov PI, Bundle DR, Brunton JL.
52. Samuel JE, Perera LP, Ward S, OBrien AD, 2002. A mutational analysis of the globotriaosyl-
Ginsburg V, Krivan HC. 1990. Comparison of the ceramide-binding sites of verotoxin VT1. J Biol
glycolipid receptor specificities of Shiga-like toxin Chem 277:53515359.
type II and Shiga-like toxin type II variants. Infect 64. Kitova EN, Kitov PI, Paszkiewicz E, Kim J,
Immun 58:611618. Mulvey GL, Armstrong GD, Bundle DR,
53. Devenish J, Gyles C, LaMarre J. 1998. Binding of Klassen JS. 2007. Affinities of Shiga toxins 1 and
Escherichia coli verotoxins to cell surface protein 2 for univalent and oligovalent Pk-trisaccharide
on wild-type and globotriaosylceramide-deficient analogs measured by electrospray ionization mass
Vero cells. Can J Microbiol 44:2834. spectrometry. Glycobiology 17:11271137.
54. Katagiri YU, Mori T, Nakajima H, Katagiri C, 65. Nishikawa K, Matsuoka K, Watanabe M,
Taguchi T, Takeda T, Kiyokawa N, Fujimoto J. Igai K, Hino K, Hatano K, Yamada A, Abe N,
1999. Activation of Src family kinase yes induced Terunuma D, Kuzuhara H, Natori Y. 2005.
by Shiga toxin binding to globotriaosyl ceramide Identification of the optimal structure required
(Gb3/CD77) in low density, detergent-insoluble for a Shiga toxin neutralizer with oriented carbo-
microdomains. J Biol Chem 274:3527835282. hydrates to function in the circulation. J Infect Dis
55. Kovbasnjuk O, Edidin M, Donowitz M. 2001. 191:20972105.
Role of lipid rafts in Shiga toxin 1 interaction 66. Watanabe M, Igai K, Matsuoka K, Miyagawa A,
with the apical surface of Caco-2 cells. J Cell Sci Watanabe T, Yanoshita R, Samejima Y, Terunuma
114:40254031. D, Natori Y, Nishikawa K. 2006. Structural analysis
56. Falguieres T, Romer W, Amessou M, Afonso C, of the interaction between Shiga toxin B subunits
Wolf C, Tabet JC, Lamaze C, Johannes L. 2006. and linear polymers bearing clustered globotriose
Functionally different pools of Shiga toxin re- residues. Infect Immun 74:19841988.
ceptor, globotriaosyl ceramide, in HeLa cells. 67. Holgersson J, Jovall PA, Breimer ME. 1991.
FEBS J 273:52055218. Glycosphingolipids of human large intestine: de-
57. Kiarash A, Boyd B, Lingwood CA. 1994. Glyco- tailed structural characterization with special
sphingolipid receptor function is modified by fatty reference to blood group compounds and bacte-
acid content. Verotoxin 1 and verotoxin 2c prefer- rial receptor structures. J Biochem 110:120131.
entially recognize different globotriaosyl ceramide 68. Malyukova I, Murray KF, Zhu C, Boedeker E,
fatty acid homologues. J Biol Chem 269:1113811146. Kane A, Patterson K, Peterson JR, Donowitz
58. Pellizzari A, Pang H, Lingwood CA. 1992. M, Kovbasnjuk O. 2009. Macropinocytosis in
Binding of verocytotoxin 1 to its receptor is in- Shiga toxin 1 uptake by human intestinal epithe-
fluenced by differences in receptor fatty acid lial cells and transcellular transcytosis. Am J
content. Biochemistry 31:13631370. Physiol Gastrointest Liver Physiol 296:G7892.

50 MELTON-CELSA
69. Hurley BP, Thorpe CM, Acheson DW. 2001. 81. Garred O, van Deurs B, Sandvig K. 1995. Furin-
Shiga toxin translocation across intestinal epi- induced cleavage and activation of Shiga toxin.
thelial cells is enhanced by neutrophil transmi- J Biol Chem 270:1081710821.
gration. Infect Immun 69:61486155. 82. Raa H, Grimmer S, Schwudke D, Bergan J, Walchli
70. Philpott DJ, Ackerley CA, Kiliaan AJ, Karmali S, Skotland T, Shevchenko A, Sandvig K. 2009.
MA, Perdue MH, Sherman PM. 1997. Translo- Glycosphingolipid requirements for endosome-to-
cation of verotoxin-1 across T84 monolayers: Golgi transport of Shiga toxin. Traffic 10:868882.
mechanism of bacterial toxin penetration of epi- 83. Sandvig K, Skotland T, van Deurs B, Klokk TI.
thelium. Am J Physiol 273:G13491358. 2013. Retrograde transport of protein toxins
71. Schuller S, Frankel G, Phillips AD. 2004. Inter- through the Golgi apparatus. Histochem Cell Biol
action of Shiga toxin from Escherichia coli with 140:317326.
human intestinal epithelial cell lines and explants: 84. Spooner RA, Lord JM. 2012. How ricin and
Stx2 induces epithelial damage in organ culture. Shiga toxin reach the cytosol of target cells: retro-
Cell Microbiol 6:289301. translocation from the endoplasmic reticulum.
72. Zumbrun SD, Hanson L, Sinclair JF, Freedy J, Curr Top Microbiol Immunol 357:1940.
Melton-Celsa AR, Rodriguez-Canales J, Hanson 85. Tam PJ, Lingwood CA. 2007. Membrane cyto-
JC, OBrien AD. 2010. Human intestinal tissue solic translocation of verotoxin A1 subunit in
and cultured colonic cells contain globotriaosyl- target cells. Microbiology 153:27002710.
ceramide synthase mRNA and the alternate Shiga 86. Obrig TG, Moran TP, Brown JE. 1987. The
toxin receptor globotetraosylceramide. Infect mode of action of Shiga toxin on peptide elon-
Immun 78:44884499. gation of eukaryotic protein synthesis. Biochem J
73. Jacewicz MS, Acheson DW, Mobassaleh M, 244:287294.
Donohue-Rolfe A, Balasubramanian KA, Keusch 87. Furutani M, Kashiwagi K, Ito K, Endo Y,
GT. 1995. Maturational regulation of globotriao- Igarashi K. 1992. Comparison of the modes of
sylceramide, the Shiga-like toxin 1 receptor, in action of a Vero toxin (a Shiga-like toxin) from
cultured human gut epithelial cells. J Clin Invest Escherichia coli, of ricin, and of alpha-sarcin. Arch
96:13281335. Biochem Biophys 293:140146.
74. Zumbrun SD, Melton-Celsa AR, Smith MA, 88. Warnier M, Romer W, Geelen J, Lesieur J,
Gilbreath JJ, Merrell DS, OBrien AD. 2013. Amessou M, van den Heuvel L, Monnens L,
Dietary choice affects Shiga toxin-producing Johannes L. 2006. Trafficking of Shiga toxin/
Escherichia coli (STEC) O157:H7 colonization and Shiga-like toxin-1 in human glomerular micro-
disease. Proc Natl Acad Sci USA 110:E21262133. vascular endothelial cells and human mesangial
75. Schuller S, Heuschkel R, Torrente F, Kaper JB, cells. Kidney Int 70:20852091.
Phillips AD. 2007. Shiga toxin binding in normal 89. Jandhyala DM, Thorpe CM, Magun B. 2012.
and inflamed human intestinal mucosa. Microbes Ricin and Shiga toxins: effects on host cell signal
Infect 9:3539. transduction. Curr Top Microbiol Immunol 357:
76. Bergan J, Dyve Lingelem AB, Simm R, Skotland 4165.
T, Sandvig K. 2012. Shiga toxins. Toxicon 90. Lee MS, Kim MH, Tesh VL. 2013. Shiga toxins
60:10851107. expressed by human pathogenic bacteria induce
77. Torgersen ML, Lauvrak SU, Sandvig K. 2005. immune responses in host cells. J Microbiol
The A-subunit of surface-bound Shiga toxin 51:724730.
stimulates clathrin-dependent uptake of the toxin. 91. Tesh VL. 2012. The induction of apoptosis by
FEBS J 272:41034113. Shiga toxins and ricin. Curr Top Microbiol
78. Romer W, Berland L, Chambon V, Gaus K, Immunol 357:137178.
Windschiegl B, Tenza D, Aly MR, Fraisier V, 92. Pai CH, Kelly JK, Meyers GL. 1986. Experimental
Florent JC, Perrais D, Lamaze C, Raposo G, infection of infant rabbits with verotoxin-producing
Steinem C, Sens P, Bassereau P, Johannes L. Escherichia coli. Infect Immun 51:1623.
2007. Shiga toxin induces tubular membrane 93. Keenan KP, Sharpnack DD, Collins H, Formal
invaginations for its uptake into cells. Nature SB, OBrien AD. 1986. Morphologic evaluation of
450:670675. the effects of Shiga toxin and E. coli Shiga-like
79. Sandvig K, Garred O, Prydz K, Kozlov JV, toxin on the rabbit intestine. Am J Pathol 125:69
Hansen SH, van Deurs B. 1992. Retrograde 80.
transport of endocytosed Shiga toxin to the en- 94. Tesh VL, Burris JA, Owens JW, Gordon VM,
doplasmic reticulum. Nature 358:510512. Wadolkowski EA, OBrien AD, Samuel JE. 1993.
80. Sandvig K, Bergan J, Dyve AB, Skotland T, Comparison of the relative toxicities of Shiga-like
Torgersen ML. 2010. Endocytosis and retrograde toxins type I and type II for mice. Infect Immun
transport of Shiga toxin. Toxicon 56:11811185. 61:33923402.

95. Smith MJ, Teel LD, Carvalho HM, Melton- 107. Conrady DG, Flagler MJ, Friedmann DR,
Celsa AR, OBrien AD. 2006. Development of a Vander Wielen BD, Kovall RA, Weiss AA, Herr
hybrid Shiga holotoxoid vaccine to elicit heter- AB. 2010. Molecular basis of differential B-
ologous protection against Shiga toxins types 1 pentamer stability of Shiga toxins 1 and 2. PLoS
and 2. Vaccine 24:41224129. One 5:e15153.
96. Soborg B, Lassen SG, Muller L, Jensen T, 108. Tam P, Mahfoud R, Nutikka A, Khine AA,
Ethelberg S, Molbak K, Scheutz F. 2013. A Binnington B, Paroutis P, Lingwood C. 2008.
verocytotoxin-producing E. coli outbreak with a Differential intracellular transport and binding of
surprisingly high risk of haemolytic uraemic syn- verotoxin 1 and verotoxin 2 to globotriaosylceramide-
drome, Denmark, SeptemberOctober 2012. Euro containing lipid assemblies. J Cell Physiol 216:750
Surveill 18:ii, 20350. 763.
97. Boerlin P, McEwen SA, Boerlin-Petzold F, 109. Matussek A, Lauber J, Bergau A, Hansen W,
Wilson JB, Johnson RP, Gyles CL. 1999. Asso- Rohde M, Dittmar KE, Gunzer M, Mengel M,
ciations between virulence factors of Shiga toxin- Gatzlaff P, Hartmann M, Buer J, Gunzer F.
producing Escherichia coli and disease in humans. 2003. Molecular and functional analysis of Shiga
J Clin Microbiol 37:497503. toxin-induced response patterns in human vas-
98. Ostroff SM, Tarr PI, Neill MA, Lewis JH, cular endothelial cells. Blood 102:13231332.
Hargrett-Bean N, Kobayashi JM. 1989. Toxin 110. Lentz EK, Leyva-Illades D, Lee MS, Cherla RP,
genotypes and plasmid profiles as determinants Tesh VL. 2011. Differential response of the
of systemic sequelae in Escherichia coli O157:H7 human renal proximal tubular epithelial cell line
infections. J Infect Dis 160:994998. HK-2 to Shiga toxin types 1 and 2. Infect Immun
99. Russo LM, Melton-Celsa AR, Smith MA, Smith 79:35273540.
MJ, OBrien DA. 2013. Oral intoxication of mice 111. Stearns-Kurosawa DJ, Collins V, Freeman S,
with Shiga toxin type 2a (Stx2a) and protection by Tesh VL, Kurosawa S. 2010. Distinct physiologic
anti-Stx2a monoclonal antibody 11E10. Infect and inflammatory responses elicited in baboons
Immun 82:12131221. after challenge with Shiga toxin type 1 or 2 from
100. Louise CB, Obrig TG. 1995. Specific interaction enterohemorrhagic Escherichia coli. Infect Immun
of Escherichia coli O157:H7-derived Shiga-like 78:24972504.
toxin II with human renal endothelial cells. 112. Bu F, Borsa N, Gianluigi A, Smith RJ. 2012.
J Infect Dis 172:13971401. Familial atypical hemolytic uremic syndrome: a
101. Chark D, Nutikka A, Trusevych N, Kuzmina J, review of its genetic and clinical aspects. Clin Dev
Lingwood C. 2004. Differential carbohydrate Immunol 2012:370426.
epitope recognition of globotriaosyl ceramide by 113. Riley LW, Remis RS, Helgerson SD, McGee
verotoxins and a monoclonal antibody. Eur J HB, Wells JG, Davis BR, Hebert RJ, Olcott ES,
Biochem 271:405417. Johnson LM, Hargrett NT, Blake PA, Cohen
102. Head SC, Karmali MA, Lingwood CA. 1991. ML. 1983. Hemorrhagic colitis associated with a
Preparation of VT1 and VT2 hybrid toxins from rare Escherichia coli serotype. N Engl J Med
their purified dissociated subunits. Evidence for B 308:681685.
subunit modulation of a subunit function. J Biol 114. Frank C, Werber D, Cramer JP, Askar M, Faber
Chem 266:36173621. M, an der Heiden M, Bernard H, Fruth A,
103. Gallegos KM, Conrady DG, Karve SS, Prager R, Spode A, Wadl M, Zoufaly A, Jordan
Gunasekera TS, Herr AB, Weiss AA. 2012. Shiga S, Kemper MJ, Follin P, Muller L, King LA,
toxin binding to glycolipids and glycans. PLoS Rosner B, Buchholz U, Stark K, Krause G. 2011.
One 7:e30368. Epidemic profile of Shiga-toxin-producing Esch-
104. Mahfoud R, Manis A, Binnington B, Ackerley C, erichia coli O104:H4 outbreak in Germany. N Engl
Lingwood CA. 2010. A major fraction of glyco- J Med 365:17711780.
sphingolipids in model and cellular cholesterol- 115. Sauter KA, Melton-Celsa AR, Larkin K, Troxell
containing membranes is undetectable by their ML, OBrien AD, Magun BE. 2008. Mouse
binding proteins. J Biol Chem 285:3604936059. model of hemolytic-uremic syndrome caused by
105. Gamage SD, McGannon CM, Weiss AA. 2004. endotoxin-free Shiga toxin 2 (Stx2) and protec-
Escherichia coli serogroup O107/O117 lipopoly- tion from lethal outcome by anti-Stx2 antibody.
saccharide binds and neutralizes Shiga toxin 2. Infect Immun 76:44694478.
J Bacteriol 186:55065512. 116. Fontaine A, Arondel J, Sansonetti PJ. 1988. Role
106. Kimura T, Tani S, Matsumoto Yi Y, Takeda T. of Shiga toxin in the pathogenesis of bacillary
2001. Serum amyloid P component is the Shiga dysentery, studied by using a Tox- mutant of
toxin 2-neutralizing factor in human blood. J Biol Shigella dysenteriae 1. Infect Immun 56:3099
Chem 276:4157641579. 3109.

52 MELTON-CELSA
117. Chaisri U, Nagata M, Kurazono H, Horie H, of Shiga toxin haemolytic uraemic syndrome.
Tongtawe P, Hayashi H, Watanabe T, Tapchaisri Pediatr Nephrol Jul 11. (Epub ahead of print.)
P, Chongsa-nguan M, Chaicumpa W. 2001. Lo- 128. Zoja C, Buelli S, Morigi M. 2010. Shiga toxin-
calization of Shiga toxins of enterohaemorrhagic associated hemolytic uremic syndrome: patho-
Escherichia coli in kidneys of paediatric and geri- physiology of endothelial dysfunction. Pediatr
atric patients with fatal haemolytic uraemic syn- Nephrol 25:22312240.
drome. Microb Pathog 31:5967. 129. Andreoli SP, Trachtman H, Acheson DW,
118. Uchida H, Kiyokawa N, Horie H, Fujimoto J, Siegler RL, Obrig TG. 2002. Hemolytic uremic
Takeda T. 1999. The detection of Shiga toxins in syndrome: epidemiology, pathophysiology, and
the kidney of a patient with hemolytic uremic therapy. Pediatr Nephrol 17:293298.
syndrome. Pediatr Res 45:133137. 130. Maye CL, Leibowitz CS, Kurosawa S, Stearns-
119. Tazzari PL, Ricci F, Carnicelli D, Caprioli A, Kurosawa DJ. 2012. Shiga toxins and the path-
Tozzi AE, Rizzoni G, Conte R, Brigotti M. ophysiology of hemolytic uremic syndrome in
2004. Flow cytometry detection of Shiga toxins in humans and animals. Toxins (Basel) 4:12611287.
the blood from children with hemolytic uremic 131. Stahl AL, Sartz L, Karpman D. 2011. Comple-
syndrome. Cytometry B Clin Cytom 61:4044. ment activation on platelet-leukocyte complexes
120. Bitzan M, Poole R, Mehran M, Sicard E, and microparticles in enterohemorrhagic Esche-
Brockus C, Thuning-Roberson C, Riviere M. richia coli-induced hemolytic uremic syndrome.
2009. Safety and pharmacokinetics of chimeric Blood 117:55035513.
anti-Shiga toxin 1 and anti-Shiga toxin 2 mono- 132. Thurman JM, Marians R, Emlen W, Wood S,
clonal antibodies in healthy volunteers. Anti- Smith C, Akana H, Holers VM, Lesser M, Kline
microb Agents Chemother 53:30813087. M, Hoffman C, Christen E, Trachtman H. 2009.
121. Dowling TC, Chavaillaz PA, Young DG, Melton- Alternative pathway of complement in children
Celsa A, OBrien A, Thuning-Roberson C, with diarrhea-associated hemolytic uremic syn-
Edelman R, Tacket CO. 2005. Phase 1 safety drome. Clin J Am Soc Nephrol 4:19201924.
and pharmacokinetic study of chimeric murine- 133. Hale TL, Formal SB. 1980. Cytotoxicity of
human monoclonal antibody c alpha Stx2 Shigella dysenteriae 1 for cultured mammalian
administered intravenously to healthy adult vol- cells. Am J Clin Nutr 33:24852490.
unteers. Antimicrob Agents Chemother 49:1808 134. Koch C, Hertwig S, Lurz R, Appel B, Beutin L.
1812. 2001. Isolation of a lysogenic bacteriophage car-
122. McKee ML, Melton-Celsa AR, Moxley RA, rying the stx(1(OX3)) gene, which is closely as-
Francis DH, OBrien AD. 1995. Enterohemor- sociated with Shiga toxin-producing Escherichia
rhagic Escherichia coli O157:H7 requires intimin coli strains from sheep and humans. J Clin
to colonize the gnotobiotic pig intestine and to Microbiol 39:39923998.
adhere to HEp-2 cells. Infect Immun 63:3739 135. Paton AW, Beutin L, Paton JC. 1995. Hetero-
3744. geneity of the amino-acid sequences of Esche-
123. Judge NA, Mason HS, OBrien AD. 2004. Plant richia coli Shiga-like toxin type-I operons. Gene
cell-based intimin vaccine given orally to mice 153:7174.
primed with intimin reduces time of Escherichia 136. Burk C, Dietrich R, Acar G, Moravek M,
coli O157:H7 shedding in feces. Infect Immun Bulte M, Martlbauer E. 2003. Identification and
72:168175. characterization of a new variant of Shiga toxin 1
124. Girardeau JP, Dalmasso A, Bertin Y, Ducrot C, in Escherichia coli ONT:H19 of bovine origin.
Bord S, Livrelli V, Vernozy-Rozand C, Martin J Clin Microbiol 41:21062112.
C. 2005. Association of virulence genotype with 137. Paton AW, Paton JC, Manning PA. 1993. Poly-
phylogenetic background in comparison to dif- merase chain reaction amplification, cloning and
ferent seropathotypes of Shiga toxin-producing sequencing of variant Escherichia coli Shiga-like
Escherichia coli isolates. J Clin Microbiol 43:6098 toxin type II operons. Microb Pathog 15:7782.
6107. 138. Persson S, Olsen KE, Ethelberg S, Scheutz F.
125. Zumbrun SD, Melton-Celsa AR, OBrien AD. 2007. Subtyping method for Escherichia coli Shiga
2014. When a healthy diet turns deadly. Gut toxin (verocytotoxin) 2 variants and correlations
Microbes 5(1):4043. to clinical manifestations. J Clin Microbiol 45:
126. Petruzziello-Pellegrini TN, Moslemi-Naeini M, 20202024.
Marsden PA. 2013. New insights into Shiga toxin- 139. Weinstein DL, Jackson MP, Samuel JE, Holmes
mediated endothelial dysfunction in hemolytic RK, OBrien AD. 1988. Cloning and sequencing of
uremic syndrome. Virulence 4:556563. a Shiga-like toxin type II variant from Escherichia
127. Keir LS, Saleem MA. 2013. Current evidence coli strain responsible for edema disease of swine.
for the role of complement in the pathogenesis J Bacteriol 170:42234230.

140. Schmidt H, Scheef J, Morabito S, Caprioli A, 141. Leung PH, Peiris JS, Ng WW, Robins-Browne
Wieler LH, Karch H. 2000. A new Shiga toxin 2 RM, Bettelheim KA, Yam WC. 2003. A newly
variant (Stx2f) from Escherichia coli isolated discovered verotoxin variant, VT2g, produced by
from pigeons. Appl Environ Microbiol 66:1205 bovine verocytotoxigenic Escherichia coli. Appl
1208. Environ Microbiol 69:75497553.

Enterohemorrhagic Escherichia coli
Genomics: Past, Present, and Future
SHAH M. SADIQ,1 TRACY H. HAZEN,2 DAVID A. RASKO,2 and

MARK EPPINGER1
4
INTRODUCTION
O157:H7 is the most common enterohemorrhagic Escherichia coli (EHEC)

serotype in North America, and it has been the principal causative agent of
numerous food-poisoning outbreaks worldwide (1, 2). Initially E. coli O157:H7
was recognized as a human pathogen in 1982 when it was isolated from 47
persons in two states who had developed bloody diarrhea after consuming ham-
burgers contaminated with this organism (3). Since then, E. coli O157:H7 has
emerged as a major enteric pathogen, capable of causing localized infections
and large outbreaks of gastrointestinal disease (4). Data accumulated from 1982
to 1996 showed that approximately two-thirds of the 3,000 cases of E. coli
infections from 139 recognized outbreaks were associated with the ingestion
of contaminated food products, whereas 22% of the reported cases were from
direct person-to-person transmission and 10% were from drinking water (5).
Surveillance data have demonstrated a high prevalence of E. coli O157:H7 among
cattle and their environment, but a relatively low incidence of human infection.
This supports the potential hypothesis that a subset of E. coli O157:H7 harbored
1
Department of Biology, and South Texas Center for Emerging Infectious Diseases, University of Texas at San Antonio,
San Antonio, TX 78249; 2Institute for Genome Sciences, Department of Microbiology and Immunology, University of
Maryland School of Medicine, Baltimore, MD 21201.
55

56 SADIQ ET AL.
in cattle may be responsible for the majority of

human disease (6). To minimize or eradicate
adverse effects on public health, the E. coli
O157:H7 lineage has been the focus of numer-
ous epidemiological, microbiological, genomic,
forensic, and diagnostic studies. Overall, it is
estimated that E. coli O157:H7 alone causes
more than 76,000 infections and 61 deaths in
humans due to severe complications annually
in the United States (7). Symptoms include a
range of gastrointestinal morbidities, such as
FIGURE 1 Figure depicts the molecular differences
severe abdominal cramping accompanied with
that dene each of the attaching and effacing E. coli.
little or no associated fever and a watery diar- The eae gene encodes the intimin protein on the LEE
rhea that leads to severe bloody diarrhea (8). region; the bfp gene in this case is the presence of
Although many infected individuals remain the bundle-forming pilus operon, and the stx gene
asymptomatic, approximately 15 to 20% of encodes the Shiga toxin. These three features are
classically used to dene the pathotypes, including
people infected with EHEC present severe
EHEC. doi.10.1128/microbiolspec.EHEC-0020-2013.f1
enough symptoms to require hospitalization. In
such severe cases, patients display renal dys-
function known as hemolytic-uremic syn- atypical EPEC, containing only the eae gene.
drome (HUS), hemorrhagic colitis, and central One can quickly identify the potential flaws in
nervous system failure with potentially lethal these molecular definitions, especially since
outcomes (911). each of these virulence factors is encoded on a
The nomenclature surrounding EHEC and mobile element: stx on the phage; eae on the
other Shiga-toxin containing E. coli (STEC) LEE genomic island, and bfp on the EAF plas-
can be confusing; hence, the molecular defi- mid. The ease with which each or many of
nitions for EHEC and STEC that we employ these genes can be lost has not been defined in
in this article are outlined in Fig. 1. The de- many strains, and thus, the lack of any of these
finitions are based on key virulence factors features may prevent the proper categoriza-
of the EHEC, STEC, and enteropathogenic tion of the isolate to the appropriate pathovar.
E. coli (EPEC) pathogenic variants or path- This highlights the crucial need for the de-
ovars (1, 2). We briefly introduce virulence velopment of more stable genetic biomarkers
factors as a way to define the pathogens and from genomic information that is addressed
isolates we are examining. The three virulence later in this article.
factors are the eae gene that encodes intimin It has been noted that recent outbreaks, es-
and is used as a surrogate marker for the locus pecially those associated with green leafy veg-
of enterocyte effacement (LEE region); the etables, have been associated with increased
bfpA gene, used as a marker for the bundle- virulence, as measured by the number of in-
forming pilus operon that has been described dividuals that present in health care facilities.
as essential in the binding to the epithelial Three E. coli O157:H7 outbreaks in 2006 from
cells; and the stx gene (in this case, this means ingested fresh produce, referred to as spinach,
any Shiga toxin genes). Figure 1 demonstrates Taco Bell, and Taco John outbreaks (http://
how these three genes separate the EPEC, www.cdc.gov/ecoli/), captured the attention of
EHEC, and STEC isolates. EHEC isolates con- both the public health and lay communities.
tain both the eae and stx genes, whereas STEC For example, the spinach outbreak caused ill-
isolates contain only the stx gene. The EPEC ness in 199 people from 26 states after they in-
can be separated into two groups: typical gested fresh spinach contaminated with E. coli
EPEC, containing both eae and bfp, and O157:H7 (12). Three deaths were attributed to

CHAPTER 4 EHEC Genomics 57
the spinach outbreak: two elderly women and RESERVOIRS FOR HUMAN HEALTH
a 2-year-old child. Among the ill, 51% were
hospitalized, and in 16% (12) of the cases, in- Cattle are recognized as a main reservoir of
fection progressed to HUS and kidney failure. STEC O157:H7. The significant differences
The high number of patients hospitalized and in host prevalence, transmissibility, and viru-
the high rate of kidney failure suggest that lence phenotypes among strains from bovine
this outbreak was due to a more virulent strain and human sources are of major interest to
of E. coli O157:H7. The isolates for the most the public health community and livestock
recent outbreaks have been examined by ge- industry (16). Genomic analysis revealed di-
nome sequencing as a part of the epidemio- vergence into three lineages: lineage I and
logical and microbiological examination of the lineage I/II strains are commonly associated
outbreak. with human disease, whereas lineage II
The introduction of microbial sequencing strains are overrepresented in the asymp-
has opened up the opportunity for comparative tomatic bovine host reservoir (17). Growing
analysis of many genomes to identify regions evidence suggests that genotypic differences
of the genome that may be associated with the between these lineages, such as polymor-
greater virulence described above. The se- phisms in Shiga toxin subtypes and synergis-
quencing of the E. coli MG1655 isolate was the tically acting virulence factors, are correlated
beginning of E. coli genome sequencing (13); with phenotypic differences in virulence,
this was followed by the sequencing of two host ecology, and epidemiology (18). To as-
EHEC O157:H7 isolates, EDL933 (14) and the sess the genomic plasticity on a genomewide
Sakai isolate (15). The sequencing of these two scale, the whole genome of strain EC869,
isolates and the associated comparative analy- a bovine-associated E. coli O157:H7 isolate,
sis provided significant insights into two key was sequenced. Comparative phylogenomic
points of the evolution of E. coli in general and analysis of this key isolate enabled placement
EHEC specifically: (i) the isolates of EHEC of the bovine lineage II strains within the
were closely related and the analysis could dis- genetically homogeneous E. coli O157:H7
tinguish differences in each group, and (ii) a clade (18). Identification of polymorphic loci
significant amount of diversity within and be- that are anchored both in the chromosomal
tween the EHEC isolates existed, but there backbone and horizontally acquired regions
was >1 Mb of DNA that was unique in the allowed association of bovine genotypes with
EHEC isolates that was not present in the altered virulence phenotypes and host preva-
laboratory-adapted K-12 isolate. This repre- lence. Polymorphic markers are valuable in
sented approximately 20% of the genome that the development of a robust typing system
was unique in the pathogen, providing ample critical for forensic, diagnostic, and epide-
opportunity for functional characterization. miological studies of this emerging human
Continued genome sequencing of E. coli and pathogen.
EHEC specifically has resulted in significant
insights into the evolution of these important
pathogens, but with the advent of new se- TYPING AND GENETIC ANALYSIS OF
quencing technologies, we no longer sequence EHEC OUTBREAKS WITH
one or two prototype isolates; rather, we se- PREGENOMIC METHODOLOGIES
quence collections or outbreaks to define
the molecular markers of these pathogens. Unlike other E. coli serotypes, the O157:H7
This article highlights where we started with lineage is distinguished by its genetically highly
the molecular characterization of E. coli and homogeneous population structure, compara-
EHEC and where we are today with high- ble to clonal microbial species such as Yer-
throughput technologies. sinia pestis (19) or Bacillus anthracis (20). With

58 SADIQ ET AL.
potentially lethal and widespread outbreaks, Multilocus Sequence Typing

a large-scale and in-depth survey of genetic
Sequenced-based methods, such as multilocus
and architectural polymorphisms is a cru-
sequence typing (MLST), have been powerful
cial prerequisite to obtain insights into the
subtyping tools in molecular epidemiology.
natural pathogenome evolution and extent of
These methods have the advantage of being
bacterial disease virulence genotypes. Genetic
easily standardized and automated. MLST
heterogeneity among O157:H7 EHEC strains
was first developed for Neisseria meningitidis
has been established by using a broad panel
in 1998 to overcome the poor reproducibility
of targeted- and whole-genome-based typing
between laboratories applying older molecular
assays to determine diversity and evolution-
typing schemes (47). The principle behind the
ary relationships among EHEC isolates, such
MLST scheme is to identify internal nucleo-
as multilocus sequence typing (21), octamer-
tide sequences of approximately 400 to 500 bp
and PCR-based genome scanning, (22, 23),
in multiple housekeeping genes. Unique se-
phage typing (2426), multiple-locus variable-
quences (alleles) are assigned a random integer
number tandem repeat analysis (27, 28), mi-
number, and a unique combination of alleles
croarrays (29), microarray-based comparative
at each locus, an allelic profile, specifies
genome hybridization (17), nucleotide poly-
the sequence type. In this new era of high-
morphism assays (19, 30, 31), pulsed-field
throughput sequencing, it may be more ratio-
electrophoresis (PFGE) (32), subtractive hy-
nal to use whole-genome sequence data for
bridization (33, 34), and optical mapping (12,
typing. Two MLST schemes exist for E. coli:
35, 36).
E. coli scheme 1, which employs seven genes
(adk, fumC, gyrB, icd, mdh, purA, recA) (48),
Pulsed-Field Gel Electrophoresis and E. coli scheme 2, which employs eight
genes (dinB, icdA, pabB, polB, putP, trpA, trpB,
PFGE has been widely employed as a mo-
uidA) (49). Though these MLST methods are
lecular typing method in epidemiological in-
generally congruent, there are some differ-
vestigations of EHEC (32, 37). According to
ences for some strains.
differences in the XbaI PFGE patterns, EHEC
These MLST systems first determined that
O157:H7 isolates are classified into different
EHEC isolates are genetically highly clonal but
types (types I to V and ND) (38). Alterations
could also be separated into multiple clades,
in PFGE patterns after restriction digest are
suggesting that there are multiple evolutionary
results of genomic rearrangements driven by
paths to generate a fully virulent EHEC isolate
recombination of prophages and mobile ele-
(50).
ments (39, 40). PFGE, though considered a
first-line molecular screening tool, provides
insight into the genome structure and changes
Multiple-Locus Variable Number Tandem
in restriction patterns and potentially iden-
Repeat Analysis
tifies spontaneous genomic rearrangements
or recombination of mobile elements (39, As a result of the poor discriminatory ability
40). In short-term epidemiological studies, it of MLST for E. coli O157:H7, it was decided
may be that the PFGE pattern will be useful to target short tandem repeats, which are
as an inclusion or exclusion criterion when areas of the bacterial genome that evolve
E. coli O157:H7 isolates are examined over a rapidly. Targeting of these elements, which
defined time window in an outbreak situa- often vary in number among different strains
tion. However, even now, we are seeing the of the same species (the definition of a vari-
greater use of whole-genome sequencing for able-number tandem repeat), has success-
the investigation of outbreaks of E. coli (41 fully been used to discriminate between
46). strains of prokaryotes (51). Multiple-locus

variable-number tandem repeat analysis in- screening a diverse strains set. Polymorphic
volves determination of the number of re- OBGS products that were specific to lineage I
peats at multiple loci, thereby providing a or lineage II strains excluding the polymor-
powerful tool for assessing the genetic rela- phisms are not found within prophage, inser-
tionships between bacterial strains of the tion sequences, or plasmid sequences. OBGS
same species. Multiple-locus variable-number first demonstrated that the E. coli O157:H7
tandem repeat analysis has several advantages clonal complex had diverged into two highly
over PFGE because, like MLST, the output related lineages, designated lineages I and II,
is highly objective, making the data amenable that were found to be disproportionately rep-
to automated computer analysis for the rapid resented among human and bovine isolates,
detection of outbreaks and easy to compare respectively (22). Lineage II strains are found
across laboratories. less frequently associated with human disease
due either to inefficient transmission from
bovine sources or to attenuated virulence in
Whole-Genome Mapping
humans and the bovine host (26).
Whole-genome mapping has been used for
detailed genome comparisons to differentiate Lineage-Specic Polymorphism Typing Assay
closely related E. coli O157:H7 strains based on Further analyses led to a refined classification
alterations in the chromosomal architectures system, termed the lineage-specific polymor-
(insertions, deletions, and rearrangements). phism assay (LSPA), that ultimately partitions
The sizing and positioning of lateral acquired E. coli O157:H7 strains into three lineages, I,
genomic regions in EHEC are crucial in as- I/II, and II, according to a PCR-based assay
sessing the Shiga toxin virulence status deter- and polyacrylamide gel separation by testing
mined by Stx allele prevalence and respective the repeat length at six genic and intergenic
chromosomal insertion sites of the convert- chromosomal loci (52). In reference to the
ing prophages (12, 35, 36). Comparative map EDL933 genome, the LSPA markers comprise
analysis reveals valid biological markers to a 9-base insertion in gene Z5935, a 78-base
trace evolution and also assists in genome as- insertion in the yhcG gene, a 9-base deletion in
sembly for molecular epidemiology outbreak the rbsB gene, a 9-base insertion in the rtcB
investigations (36). gene, and an 18-base insertion in the inter-
genic region spanning the arp-iclR genes.
These techniques showed divergence into
Genome Sequence-Based
three different but interlinked lineages; line-
High-Resolution Genotyping
age I and lineage I/II strains are commonly
Octamer-Based Genome Scanning Typing Assay associated with human infections, whereas
The rapid accumulation of whole-genome lineage II strains are overrepresented in the
sequence data for O157:H7 has allowed the asymptomatic bovine host reservoir. Genetic
development of high-resolution subtyping evidence also suggests that distributions of
methods that enable inter- and intraspecies genotypes between these lineages, such as
bacterial genome comparisons. Sequence- polymorphisms in Shiga toxin subtypes and
based phylogenetic assays have determined synergistically acting virulence factors, are cor-
that EHEC strains comprise three highly re- related with phenotypic differences of major
lated but distinct populations with a global biological relevance in virulence, host ecol-
prevalence that differs in genotype and host ogy, and epidemiology. Identification of poly-
ecology. Polymorphisms were identified using morphic loci that are anchored both in the
high-density octamer-based genome scanning chromosomal backbone and in horizontally ac-
(OBGS) analysis by testing multiple OBGS quired regions allowed us to associate bovine
primer combinations in independent reactions genotypes with altered virulence phenotypes

60 SADIQ ET AL.
and host prevalence. Numerous novel lineage 500 clinical E. coli O157 strains. This resulted
II-specific genome signatures, some of which in a refined LSPA-6 lineage classification and
appear to be intimately associated with the further separated EHEC isolates into nine
altered pathogenic potential and niche adap- distinct clades. The frequency and distribution
tation within the bovine rumen and further of Shiga toxin-converting prophages and form
discriminate bovine super shedders, have of clinical disease manifestation could also be
been cataloged (18, 26, 36). elucidated (30, 57). Recently, Manning et al.
(30) used SNP analysis to identify a group
Whole-Genome Sequence Typing/Single (clade 8) of hypervirulent E. coli O157:H7
Nucleotide Polymorphism Typing strains found in the United States. In 2006,
Despite multiple genomes of this lineage hav- food-borne outbreaks involving O157 contam-
ing become available in the genomics era ination of fresh produce (e.g., spinach) were
(14, 15, 18, 36, 53, 54), the biological insight associated with more severe infection, causing
into epidemiology and disease mechanism suf- higher rates of HUS and more frequencies of
fers from a lack of markers for accurate typing hospitalization, demonstrating that increased
and genotype/phenotype association. High- virulence had been acquired (30). In silico
resolution phylogenomic approaches allow the scoring of SNPs was successfully deployed by
dynamics of pathogenome evolution to be fol- Eppinger et al. to investigate the degree of
lowed at a high level of phylogenetic accuracy genetic heterogeneity in stains derived from a
and resolution. Whereas the current molecu- single outbreak of human disease (36) and to
lar markers and typing assays used by public establish biologically relevant markers among
health microbiology laboratories may be ade- strains from clinical settings and the animal
quate for routine surveillance and identifica- reservoir (36). These enriched mutational
tion of E. coli O157:H7, these approaches lack database resources will provide a robust foun-
the polymorphic markers and discriminatory dation to better associate genotypic group pro-
power to study the relatedness of strains of un- files and virulence phenotypes within E. coli
known provenance, which becomes of parti- O157:H7.
cular importance when investigating outbreak
strains that form clonal complexes with only a
few genetic polymorphisms. With the increase HISTORICAL VIEW OF
of next-generation sequencing technologies. EHEC/STEC GENOMICS
whole-genome sequences typing approaches,
such as the discovery of single nucleotide poly- Whereas Helicobacter pylori was the first or-
morphisms (SNPs), have gained popularity. ganism to have multiple isolates sequenced
SNP typing not only provides stable genetic (58, 59), EHEC was the first E. coli pathotype
markers to study evolution but also offers with genomes of multiple isolates available for
greater phylogenetic resolution. SNP discov- comparison (14, 15). Considering the related-
ery approaches have yielded thousands of ness of these isolates, it can be argued as the
high-quality SNPs as critical bases for the de- true beginning of comparative genomics as
velopment of a refined phylogenomic frame- we know it. Irrespective of when comparative
work (36, 55, 56). SNP typing allowed the genomics began, the scope of the analyses has
elucidation of the evolutionary origin and changed over the years. We have transitioned
emergence of the pathogenic O157:H7 lineage, from sequencing prototype isolates from a
and the determination of the genetic relation- group of pathogens to sequencing large num-
ships to the intermediate immobile O157:H() bers of isolates in an effort to understand the
and ancestral EPEC O55:H7 serotypes (56). population structure of the pathogen associ-
A SNP-based clade typing assay that detects ated with human health and to explore the con-
SNPs in 96 loci has been applied to more than cept of genomic epidemiology. This section

briefly covers some of the important publica- were only between two genomes, and thus the
tions on genomics associated with E. coli O157 comparison of two points on a continuum does
isolates. not provide a deep insight into the pathogen.
The Sakai genome, as mentioned above,
represented the start of intrapathotype com-
First View of the Interpathotype
parisons of E. coli, as well as the two most
Comparisons: EDL993 and Sakai Genomes
closely related bacterial genomes to be se-
The first genome of E. coli was published in quenced at this point in history (15). The ge-
1997 by the Blattner group and signaled a nome of Sakai was 859 kb larger than that of the
change in our understanding of model orga- MG1655 genome and included two plasmids, a
nisms (13). It could be argued that the se- large virulence plasmid of 92,721 bp (pO157) and
quencing of the MG1655 genome was the first a cryptic plasmid of 3,306 bp (pOSAK1). Com-
nonpathogen genome to be decoded, as it had parative genomics identified >1,600 genes that
long been known that this isolate could no were Sakai specific when compared to MG1655,
longer colonize humans or cause disease in hu- with 131 of them having a direct correlation
mans. The sequencing of the EDL933 genome to virulence. Interestingly, there were 24 pro-
(14) provided the opportunity to examine two phage in the Sakai genome (18 in EDL933), and
genomes from the same species. These early much of the unique DNA was associated with
comparisons identified that there was >1 Mb horizontal transfer or was suggested to have
of unique genomic material, encoding 1,387 been transferred into the genome.
genes, associated with the EDL933 genome These seminal publications provided the
that was not in the MG1655 genome. It was spark that allowed others to start functionally
assumed that the majority of this DNA, pres- characterizing these genes and the role they
ent in what was termed O-islands, would be play in either survival or pathogenesis in the
associated with the ability to cause disease. human model systems. They also allowed com-
Although it was true that the majority of genes parison of EHEC to other pathotypes of
that had been identified to be associated with E. coli. Two studies examined the pan-genome
human disease were in these O157-specifc re- of E. coli and included numerous EHEC iso-
gions, including Shiga-toxin phage and the lates (61, 62). Each of these studies identified
LEE region, it was clear that not all of these that there was an open genome in E. coli, sug-
regions were directly associated with disease gesting that these organisms continue to col-
and could be considered part of the central lect more DNA. In each case, more sequencing
metabolism of these organisms. It was also would continue to identify new genes in all
noted that the genome of the EDL933 isolate additional divergent isolates. Additionally, each
contained an 400-kb chromosomal inversion study demonstrated that there were pathotype-
when compared to MG1655 genome. It is still specific adaptations that made each pathotype
unclear if this genomic inversion contributes unique and contributed to the niche speciali-
to the pathogenesis of EDL933. One key point zation that each had displayed (61, 62). At this
that was highlighted in this first genomic point, a limited number of genomes were avail-
comparison was the identification of 18 iden- able (<20 complete and draft genomes), and
tifiable prophage regions ranging in size from comparisons were limited to mostly prototype
7 kb to 62 kb that were designated BP-933 isolates. The study by Rasko et al. highlighted
[A-Q]. These phage regions include a number that there were significant genetic differences
of potential secreted effectors (14, 60). Overall, between the EHEC isolates, but this observa-
the EDL933 genome provided a starting point tion could also be attributed to the fact that the
for the study of potentially new genes asso- annotation of different genomes could have a
ciated with pathogenesis; however, one must significant impact on the findings that were
keep in mind that at this point the comparisons gene based (62).

62 SADIQ ET AL.
Additional interesting genomes of the As the cost of sequencing decreased, there

O157 serotype were published by Kulasekara started to be studies that examined large num-
et al. in 2009 (63); they sequenced an isolate, bers of isolates. With EHEC, these large-scale
TW14359, that was associated with HUS. In sequencing projects were often linked to out-
this study they identified an additional 70 kb breaks of food-borne disease in the mid-2000s.
of genetic material that was not present in
either of the reference O157 isolates. These
How Outbreaks Have Shaped the
regions encoded additional putative Type III
Sequencing of EHEC
secreted effectors and a gene for an anaerobic
nitric oxide reductase, norV. It was suggested From the early days of genomics, the majority
that the norV gene could be used as a marker of the focus has been on the EHEC O157:H7
for increased virulence leading to HUS; how- isolates associated with human disease. In re-
ever, screening of large numbers of isolates cent years, with the decrease in the cost of
suggested that the norV gene was associated sequencing we have observed an increase in
with the loss of stx1, but the direct impact on sequencing a significant number of isolates
pathogenesis was not confirmed. This publi- associated with disease within an outbreak
cation is important in that it attempts to meld setting. Some of these studies have attempted
comparative genomics with disease presenta- to associate genetic changes with increases in
tion and gene presence or absence. This can virulence or markers that laboratories could
be considered the beginning of the field we use as identifiers of outbreak strains. The
now know as genomic epidemiology. study by Eppinger et al. describes the se-
Also in 2009, the first non-O157 EHEC ge- quencing of a large number of isolates (36) to
nomes were published when Ogura et al. (53) reconstruct the anatomy of the outbreak. In
published the genomes of isolates from sero- this study the authors sequenced 25 isolates
types O26, O111, and O103. These genomes were and through comparative genomics identi-
compared to the Sakai genome. This study fied mutational and structural markers in
demonstrated that the EHEC genomes were the genome core and mobilome that allowed
routinely larger than the MG1655 genome and distinguishing these outbreak isolates among
contained a diverse array of phage and inte- three almost simultaneous outbreaks that oc-
grative elements, sometimes associated with curred in the United States as well as other
the virulence gene catalog. Most importantly, O157 isolates that have been sequenced with
it demonstrated how different serogroups and other outbreaks and clinical sources as well
lineages of E. coli could evolve into EHEC iso- as from animal reservoirs. Using a panel of
lates and demonstrated that there was a ge- >1,200 SNPs they were able to identify specific
nomic core among similar isolates by comparing lineages of E. coli O157:H7. This type of study
the gene content of 345 orthologous genes in all demonstrates how the genomics of a patho-
E. coli isolates that were sequenced to date. This gen can distill the identification to a very fine
is similar to the presentations in Fig. 2 and Fig. 3 scale. This is possible among the O157 isolates,
in this article; however, we now avoid the prob- since genomically they are all very similar
lems caused by calling genes and curation errors when compared to the other E. coli strains
by using the unannotated whole genome se- (Fig. 2); however, there are significant differ-
quence for comparison (see below). Addition- ences when only the O157 isolates are com-
ally, the study demonstrated that not all features pared (Fig. 3).
of the genomes were the same. For example, the The E. coli outbreak in Europe in 2011 pro-
LEE pathogenicity island was inserted at dif- vided an opportunity to highlight the use of
ferent locations in the genome in these isolates, genomics to examine the bacterial isolates
but it appeared to be functional as all isolates in near real time (4144, 46). Additionally, the
were derived from individuals who were ill. results provided insight into the evolution

FIGURE 2 Phylogeny of E. coli reference isolates, including EHEC. The reference used for the SNP calling is E. coli
HS as a true commensal isolate (62). The tree is maximum likelihood with 100 bootstraps made using RAxML,
as previously described (68). This gure highlights, as does MLST, that when compared to a large and diverse
collection of isolates, EHEC, and specically the O157 and O55 (in red and purple) and other serotypes
(highlighted in yellow), demonstrates a high level of similarity. This high-level of similarity extends within
these groups. Only once these clades are examined in detail can one begin to identify regions that are di-
agnostic in these clades. doi.10.1128/microbiolspec.EHEC-0020-2013.f2
of bacterial pathogens as a species and chal- chromosome of the outbreak-associated iso-

lenged the definitions and boundaries of the late was most similar to EAEC isolate 55989
STEC pathovar. The genome sequencing of (61). There were significant changes in the
the isolates associated with this outbreak genome in that it had acquired the Shiga-toxin
and comparison to isolates that were in the phage, as well as an antimicrobial resistance
database identified that the majority of the plasmid. Neither of these features had been

64 SADIQ ET AL.
FIGURE 3 Phylogeny of O157 E. coli reference isolates demonstrating that there is variation within the O157
clade that is not observed in the larger genomic comparisons. This multi-whole genome alignment contains
4,093,272 bp of sequence. The tree is maximum likelihood with 100 bootstraps made using RAxML, as pre-
viously described (68). The color indicates distinguishable clades within this selection of 50 EHEC isolates. The
additional numbers in parentheses are the clade and LSPA designations, as described in the text.
doi.10.1128/microbiolspec.EHEC-0020-2013.f3
previously identified in EAEC isolates, but both was technically STEC, but the majority of the
potentially contribute to the pathogenicity of gene markers identified pointed toward an
the bacteria. Thus the definition of the isolate EAEC isolate (64). Many names were provided

to this new type of isolate with this collection of Additionally, other virulence factors found on
virulence factors, including entero-aggregative- the O157 genomes, such as fimbrial and non-
haemorrhagic E. coli,(46)but themost appropri- fimbrial adhesins, iron uptake systems, and
ate name proposed would be enteroaggregative, non-LEE effectors, are also thought to be re-
Shiga toxin-producing E. coli O104:H4. The iso- quired for the full virulence of EHEC, but
lates in this outbreak highlighted the issues with their prevalence among non-O157 EHEC iso-
the current nomenclature that is largely based lates, and in some cases among the O157
on the identification of virulence-associated EHEC isolates, has not yet been systematically
genes that are present on mobile elements. analyzed. Global genomic differences (or con-
The genomic studies also highlighted that servation) between O157 and non-O157 EHEC
the determination of the presence or absence strains have recently been addressed in the
of specific genes or gene combinations could study by Hazen et al. (68), which demon-
not really predict the virulence of an isolate strated that the O55 and O157 isolates formed
and that further functional characterization is closely related clades, but the non-O157/non-
required. In addition to a novel collection of O55 EHEC (also known to some as EHEC2)
virulence factors in the European outbreak were divergent. We demonstrate this point
isolates, it was demonstrated that, like EHEC in Fig. 2, using a collection of reference E. coli
isolates, the Shiga toxin genes were activated isolates from each of the pathovars as well
by the exposure to subclinical doses of anti- as a selection of the O157, O55, and non-O157
biotics (42). As time goes on, more detailed EHEC isolates that have been recently se-
studies of the functional virulence of these iso- quenced. This clearly demonstrates that al-
lates will be examined and the true impact of though the virulence factor typing has grouped
the novel combinations of virulence factors these isolates together, pathogenomic analysis
will be identified. suggests they are significantly different.
That there has been a parallel evolution
and independent acquisition of the virulence
GENOMICS-GUIDED EXAMINATION factors in each of the serogroups (69) has been
OF PATHOGENESIS proposed. Hazen et al. (68) recently noted that
there appears to be a pattern of secreted effec-
The study of the evolution of pathogenesis in tors that are common in EHEC strains when
EHEC has mainly focused on the traditional compared to other EAEC isolates. This sug-
virulence factors, such as LEE, Shiga toxin, a gests that there is a complement of effectors
limited number of adhesins, and plasmid en- that are tuned to the EHEC genome content
coded features. Molecular evidence suggests for optimal regulation, interaction, and secre-
that EHEC isolates have acquired the majority tion; however, this does not necessarily have
of their virulence factors through horizontal to be linked to human virulence and further
gene transfer and thereby acquisition of the characterization of this orchestrated program
LEE pathogenicity-associated island and the is required.
Stx genes were two crucial steps in the evo-
lution of EHEC O157 from a commensal an-
Are There Genomic Differences in Recent
cestor (53). The LEE pathogenicity-associated
Isolates That Have Come from Green
island is clearly a mosaic structure, which
Leaf Vegetables Rather Than Most
arose from multiple recombination events
of the Original Isolates That Were
with foreign DNA, as did the large EHEC-
Beef Related in Some Way?
hemolysin plasmid (65, 66). LEE can be found
on chromosomes of EHEC next to tRNA genes Recent Shiga-toxigenic E. coli O157:H7 out-
at different locations (67), suggesting that LEE breaks have been linked to consumption of
has been acquired on more than one occasion. fresh produce (12, 63). It is generally recognized

66 SADIQ ET AL.
that bacterial attachment to vegetal matrices experience with such data is limited. For the
would constitute the first step in contamina- identification of an isolate or an outbreak
tion of fresh produce, but how and when this strain, comparison with genomes of very
occurs are under some debate. Cellular ap- closely related organisms is required. What is
pendages, such as curli fibers, and cellulose, a clear is that a large sequence database com-
constituent of extracellular matrix, have been prising a comprehensive panel of well-chosen
suggested to be involved in E. coli attachment isolates is necessary if genome sequencing is
and persistence in fresh produce. A compar- to be useful in the field. Perhaps the most com-
ative evaluation was conducted on the ability prehensive single-species genome data avail-
of STEC O157:H7 strains EDL933 and 86-24, able thus far are for E. coli. This database
linked to two independent food-borne disease became extremely useful when the outbreak
outbreaks in humans, and their mutants defi- in Europe occurred in 2011 (4143). This out-
cient in curli and/or cellulose expression to break resulted in 3,816 identified STEC in-
colonize and to firmly attach to spinach leaf. fections and 54 deaths (45). It was rapidly
Inoculated spinach leaves were incubated at determined that the etiological agent was
22C, and at 0, 24, and 48 h after incubation E. coli, but the isolate had features of EAEC
loosely and strongly attached E. coli O157:H7 and was Shiga toxin positive but lacked the
populations were determined. Curli-expressing majority of features associated with EHEC
E. coli O157:H7 strains developed stronger (LEE, ehx, etc.), suggesting that this isolate
association with leaf surface, whereas curli- may simply be an EAEC isolate that had ac-
deficient mutants attached to spinach at sig- quired the Shiga toxin phage. Rapid compar-
nificantly (P<0.01) lower numbers. Attachment ative genomics highlighted that this was the
of cellulose-impaire mutants to spinach leaves case and that the isolate had also acquired
was not significantly different from strains an antibiotic resistance plasmid. That we had
that contained curli; however, the relative at- large collections of isolates already sequenced
tachment strength of E. coli O157:H7 to spin- allowed the accurate and rapid determination
ach increased with incubation time for the of the genomic origin of this isolate. In the
curli-expressing strains. Laser scanning con- future these databases of isolate collections
focal microscopy analysis of inoculated leaves will become invaluable in the public health
revealed that curli-expressing E. coli O157:H7 setting in transmitting information from the
was surrounded by extracellular structures sequencer to the physician treatment para-
strongly immunostained with anticurli anti- digm for any patient. We are still far away
bodies. Production of cellulose was not re- from this reality, as the software for these ana-
quired to develop strong attachment to lyses are somewhat custom, but as the bench-
spinach leaf. These results indicate that curli top sequencers become more prevalent in
fibers are essential for strong attachment of clinical laboratories, the methodology must
E. coli O157:H7 to spinach whereas cellulose follow.
is dispensable.
CONCLUSION
PUBLIC HEALTH VIEWWHAT ARE THE
NEXT STEPS? The use of modern techniques like whole-
genome sequence typing can benefit research-
Whole bacterial genome sequencing in medi- ers interested in surveillance (18), forensics
cal microbiology is fast becoming a reality; (70), diagnostics, and epidemiological stud-
however, the challenge of converting the pri- ies of microbial pathogens and pathogenome
mary sequence data into useful clinical or evolution, and holds the potential to discover
public health action remains unmet, because novel E. coli O157:H7 virulence genotypes

that are intimately associated with adverse ACKNOWLEDGMENTS

human disease outcomes. Use of these tech-
D.A.R. and T.H.H. are supported by federal
niques is not limited to E. coli O157:H7 or
funds from the National Institute of Allergy
EHEC as this research can provide principles
and Infectious Diseases, National Institutes
and valuable benchmarking data crucial in
of Health, Department of Health and Human
the modeling of bacterial outbreak-associated
Services under grant number 1U19AI090873
population dynamics. Once data are availa-
and funds from the State of Maryland. M.E. is
ble, the scientific community will be key to
supported by the Department of Biology and
the analysis of the gathered sequence data.
South Texas Center for Emerging Infectious
Application of postgenomic tools to the gen-
Diseases, University of Texas at San Antonio.
erated data can help to understand variability
This review is further based upon work at the
in gene content and activity. These technol-
University of Texas at San Antonio supported
ogies include DNA microarray analysis or
by the Army Research Office of the Depart-
RNAseq for gene profiling and differentia-
ment of Defense under Contract No. W911NF-
tion of strains; optical mapping of chromo-
11-1-0136. We thank Sara S. K. Koenig for
somes for discovery of novel gene insertions
critically reading the manuscript and all re-
and deletions; and phenotypic microarray of
search groups who have produced data dis-
pathogens physiological and metabolic capa-
cussed in this work.
bilities for strain characterization. Virulence
assays in host model systems allow validation of
genetic predictions and association of isolate- CITATION
specific gene content to outbreak-associated
physiological and virulence capabilities. Better Sadiq SM, Hazen TH, Rasko DA, Eppinger
and more effective schemes for epidemio- M. 2014. Enterohemorrhagic Escherichia coli
logical studies can be developed to trace the genomics: past, present, and future. Microbiol
source of the outbreaks, or ultimately link the Spectrum 2(4):EHEC-0020-2013.
source to the infection in outbreak investiga-
tions. The identification of subtle but defined REFERENCES
polymorphisms would allow us to investigate
1. Kaper JB, Nataro JP, Mobley HL. 2004. Path-
the drivers of pathogenome evolution and
ogenic Escherichia coli. Nat Rev Microbiol 2:123
analyze mutations that are correlated with 140.
human disease and severity. This will enable 2. Croxen MA, Law RJ, Scholz R, Keeney KM,
us to track and distinguish separate unrelated Wlodarska M, Finlay BB. 2013. Recent advances
outbreaks and classify isolates as members in understanding enteric pathogenic Escherichia
coli. Clin Microbiol Rev 26:822880.
of an outbreak population, and also to refine
3. Riley LW, Remis RS, Helgerson SD, McGee HB,
standards currently used in outbreak investi- Wells JG, Davis BR, Hebert RJ, Olcott ES,
gations. The potential discovery of novel poly- Johnson LM, Hargrett NT, Blake PA, Cohen
morphisms complements current techniques ML. 1983. Hemorrhagic colitis associated with a
used to classify strains, and is crucial in the rare Escherichia coli serotype. N Engl J Med
308:681685.
development of rapid methods of detecting
4. Mead PS, Griffin PM. 1998. Escherichia coli
and tracking microbial outbreaks. Studies of O157:H7. Lancet 352:12071212.
non-O157 EHEC are currently under way by 5. Karch H, Bielaszewska M, Bitzan M, Schmidt
several independent research teams (68). These H. 1999. Epidemiology and diagnosis of Shiga
data sets will potentiate each other in gathering toxin-producing Escherichia coli infections. Diagn
Microbiol Infect Dis 34:229243.
a more complete and refined picture of the
6. Clawson ML, Keen JE, Smith TP, Durso LM,
E. coli O157:H7 pathogenome evolution and its McDaneld TG, Mandrell RE, Davis MA, Bono
pathogenic potential in relation to the E. coli JL. 2009. Phylogenetic classification of Esche-
species biology. richia coli O157:H7 strains of human and bovine

68 SADIQ ET AL.
origin using a novel set of nucleotide polymor- 17. Zhang Y, Laing C, Steele M, Ziebell K, Johnson
phisms. Genome Biol 10:R56. R, Benson AK, Taboada E, Gannon VP. 2007.
7. Griffin PM, Tauxe RV. 1991. The epidemiology of Genome evolution in major Escherichia coli
infections caused by Escherichia coli O157:H7, O157:H7 lineages. BMC Genomics 8:121.
other enterohemorrhagic E. coli, and the associ- 18. Eppinger M, Mammel MK, Leclerc JE, Ravel
ated hemolytic uremic syndrome. Epidemiol Rev J, Cebula TA. 2011. Genome signatures of Esch-
13:6098. erichia coli O157:H7 isolates from the bovine
8. Griffin PM, Ostroff SM, Tauxe RV, Greene KD, host reservoir. Appl Environ Microbiol 77:2916
Wells JG, Lewis JH, Blake PA. 1988. Illnesses 2925.
associated with Escherichia coli O157:H7 infec- 19. Eppinger M, Mammel MK, LeClerc JE, Ravel J,
tions. A broad clinical spectrum. Ann Intern Med Cebula TA. 2011. Genomic anatomy of Esche-
109:705712. richia coli O157:H7 outbreaks. Proc Natl Acad Sci
9. Riley DG, Gray JT, Loneragan GH, Barling KS, USA 108:2014220147.
Chase CC Jr. 2003. Escherichia coli O157:H7 20. Medini D, Donati C, Tettelin H, Masignani V,
prevalence in fecal samples of cattle from a Rappuoli R. 2005. The microbial pan-genome.
southeastern beef cow-calf herd. J Food Prot Curr Opin Genet Dev 15:589594.
66:17781782. 21. Afset JE, Anderssen E, Bruant G, Harel J,
10. Cimolai N, Morrison BJ, Carter JE. 1992. Risk Wieler L, Bergh K. 2008. Phylogenetic back-
factors for the central nervous system manifes- grounds and virulence profiles of atypical entero-
tations of gastroenteritis-associated hemolytic- pathogenic Escherichia coli strains from a case-
uremic syndrome. Pediatrics 90:616621. control study using multilocus sequence typing
11. Besser RE, Griffin PM, Slutsker L. 1999. Esche- and DNA microarray analysis. J Clin Microbiol
richia coli O157:H7 gastroenteritis and the hemo- 46:22802290.
lytic uremic syndrome: an emerging infectious 22. Kim J, Nietfeldt J, Benson AK. 1999. Octamer-
disease. Annu Rev Med 50:355367. based genome scanning distinguishes a unique
12. Kotewicz ML, Mammel MK, LeClerc JE, subpopulation of Escherichia coli O157:H7
Cebula TA. 2008. Optical mapping and 454 se- strains in cattle. Proc Natl Acad Sci USA 96:
quencing of Escherichia coli O157: H7 isolates 1328813293.
linked to the US 2006 spinach-associated out- 23. Ohnishi M, Terajima J, Kurokawa K, Nakayama
break. Microbiology 154:35183528. K, Murata T, Tamura K, Ogura Y, Watanabe H,
13. Blattner FR, Plunkett G 3rd, Bloch CA, Perna Hayashi T. 2002. Genomic diversity of entero-
NT, Burland V, Riley M, Collado-Vides J, Glasner hemorrhagic Escherichia coli O157 revealed by
JD, Rode CK, Mayhew GF, Gregor J, Davis NW, whole genome PCR scanning. Proc Natl Acad Sci
Kirkpatrick HA, Goeden MA, Rose DJ, Mau B, USA 99:1704317048.
Shao Y. 1997. The complete genome sequence of 24. Ratnam S, March SB, Ahmed R, Bezanson GS,
Escherichia coli K-12. Science 277:14531462. Kasatiya S. 1988. Characterization of Escherichia
14. Perna NT, Plunkett G 3rd, Burland V, Mau B, coli serotype O157:H7. J Clin Microbiol 26:2006
Glasner JD, Rose DJ, Mayhew GF, Evans PS, 2012.
Gregor J, Kirkpatrick HA, Posfai G, Hackett J, 25. Ahmed R, Bopp C, Borczyk A, Kasatiya S. 1987.
Klink S, Boutin A, Shao Y, Miller L, Grotbeck EJ, Phage-typing scheme for Escherichia coli O157:
Davis NW, Lim A, Dimalanta ET, Potamousis H7. J Infect Dis 155:806809.
KD, Apodaca J, Anantharaman TS, Lin J, Yen G, 26. Bono JL, Smith TP, Keen JE, Harhay GP,
Schwartz DC, Welch RA, Blattner FR. 2001. McDaneld TG, Mandrell RE, Jung WK, Besser
Genome sequence of enterohaemorrhagic Esche- TE, Gerner-Smidt P, Bielaszewska M, Karch H,
richia coli O157:H7. Nature 409:529533. Clawson ML. 2012. Phylogeny of Shiga toxin-
15. Hayashi T, Makino K, Ohnishi M, Kurokawa K, producing Escherichia coli O157 isolated from
Ishii K, Yokoyama K, Han CG, Ohtsubo E, cattle and clinically ill humans. Mol Biol Evol
Nakayama K, Murata T, Tanaka M, Tobe T, Iida 29:20472062.
T, Takami H, Honda T, Sasakawa C, Ogasawara 27. Keys C, Kemper S, Keim P. 2005. Highly diverse
N, Yasunaga T, Kuhara S, Shiba T, Hattori M, variable number tandem repeat loci in the E. coli
Shinagawa H. 2001. Complete genome sequence O157:H7 and O55:H7 genomes for high-resolution
of enterohemorrhagic Escherichia coli O157:H7 molecular typing. J Appl Microbiol 98:928940.
and genomic comparison with a laboratory strain 28. Noller AC, McEllistrem MC, Pacheco AG,
K-12. DNA Res 8:1122. Boxrud DJ, Harrison LH. 2003. Multilocus vari-
16. Ferens WA, Hovde CJ. 2011. Escherichia coli able-number tandem repeat analysis distinguishes
O157:H7: animal reservoir and sources of human outbreak and sporadic Escherichia coli O157:H7
infection. Foodborne Pathog Dis 8:465487. isolates. J Clin Microbiol 41:53895397.

29. Jackson SA, Mammel MK, Patel IR, Mays T, 40. Murase T, Yamai S, Watanabe H. 1999. Changes
Albert TJ, LeClerc JE, Cebula TA. 2007. Inter- in pulsed-field gel electrophoresis patterns in
rogating genomic diversity of E. coli O157:H7 clinical isolates of enterohemorrhagic Escherichia
using DNA tiling arrays. Forensic Sci Int 168:183 coli O157:H7 associated with loss of Shiga toxin
199. genes. Curr Microbiol 38:4850.
30. Manning SD, Motiwala AS, Springman AC, 41. Rohde H, Qin J, Cui Y, Li D, Loman NJ,
Qi W, Lacher DW, Ouellette LM, Mladonicky Hentschke M, Chen W, Pu F, Peng Y, Li J, Xi F,
JM, Somsel P, Rudrik JT, Dietrich SE, Zhang Li S, Li Y, Zhang Z, Yang X, Zhao M, Wang P,
W, Swaminathan B, Alland D, Whittam TS. Guan Y, Cen Z, Zhao X, Christner M, Kobbe R,
2008. Variation in virulence among clades of Loos S, Oh J, Yang L, Danchin A, Gao GF,
Escherichia coli O157:H7 associated with disease Song Y, Yang H, Wang J, Xu J, Pallen MJ,
outbreaks. Proc Natl Acad Sci USA 105:48684873. Aepfelbacher M, Yang R. 2011. Open-source ge-
31. Zhang W, Qi W, Albert TJ, Motiwala AS, Alland nomic analysis of Shiga-toxin-producing E. coli
D, Hyytia-Trees EK, Ribot EM, Fields PI, O104:H4. N Engl J Med 365:718724.
Whittam TS, Swaminathan B. 2006. Probing 42. Rasko DA, Webster DR, Sahl JW, Bashir A,
genomic diversity and evolution of Escherichia Boisen N, Scheutz F, Paxinos EE, Sebra R,
coli O157 by single nucleotide polymorphisms. Chin CS, Iliopoulos D, Klammer A, Peluso P,
Genome Res 16:757767. Lee L, Kislyuk AO, Bullard J, Kasarskis A,
32. Gerner-Smidt P, Kincaid J, Kubota K, Hise K, Wang S, Eid J, Rank D, Redman JC, Steyert
Hunter SB, Fair MA, Norton D, Woo-Ming A, SR, Frimodt-Moller J, Struve C, Petersen AM,
Kurzynski T, Sotir MJ, Head M, Holt K, Krogfelt KA, Nataro JP, Schadt EE, Waldor
Swaminathan B. 2005. Molecular surveillance of MK. 2011. Origins of the E. coli strain causing
shiga toxigenic Escherichia coli O157 by PulseNet an outbreak of hemolytic-uremic syndrome in
USA. J Food Prot 68:19261931. Germany. N Engl J Med 365:709717.
33. Steele M, Ziebell K, Zhang Y, Benson A, 43. Mellmann A, Harmsen D, Cummings CA, Zentz
Johnson R, Laing C, Taboada E, Gannon V. EB, Leopold SR, Rico A, Prior K, Szczepanowski
2009. Genomic regions conserved in lineage II R, Ji Y, Zhang W, McLaughlin SF, Henkhaus
Escherichia coli O157:H7 strains. Appl Environ JK, Leopold B, Bielaszewska M, Prager R,
Microbiol 75:32713280. Brzoska PM, Moore RL, Guenther S, Rothberg
34. Steele M, Ziebell K, Zhang Y, Benson A, Konczy JM, Karch H. 2011. Prospective genomic char-
P, Johnson R, Gannon V. 2007. Identification of acterization of the German enterohemorrhagic
Escherichia coli O157:H7 genomic regions con- Escherichia coli O104:H4 outbreak by rapid next
served in strains with a genotype associated with generation sequencing technology. PLoS One 6:
human infection. Appl Environ Microbiol 73:2231. e22751.
35. Kotewicz ML, Jackson SA, LeClerc JE, Cebula 44. Grad YH, Lipsitch M, Feldgarden M, Arachchi
TA. 2007. Optical maps distinguish individual HM, Cerqueira GC, Fitzgerald M, Godfrey P,
strains of Escherichia coli O157:H7. Microbiology Haas BJ, Murphy CI, Russ C, Sykes S, Walker
153:17201733. BJ, Wortman JR, Young S, Zeng Q, Abouelleil
36. Eppinger M, Mammel MK, Leclerc JE, Ravel J, A, Bochicchio J, Chauvin S, Desmet T, Gujja S,
Cebula TA. 2011. Genomic anatomy of Esche- McCowan C, Montmayeur A, Steelman S,
richia coli O157:H7 outbreaks. Proc Natl Acad Sci Frimodt-Moller J, Petersen AM, Struve C,
USA 108:2014220147. Krogfelt KA, Bingen E, Weill FX, Lander ES,
37. Barrett TJ, Lior H, Green JH, Khakhria R, Nusbaum C, Birren BW, Hung DT, Hanage WP.
Wells JG, Bell BP, Greene KD, Lewis J, Griffin 2012. Genomic epidemiology of the Escherichia
PM. 1994. Laboratory investigation of a multistate coli O104:H4 outbreaks in Europe, 2011. Proc Natl
food-borne outbreak of Escherichia coli O157:H7 Acad Sci USA 109:30653070.
by using pulsed-field gel electrophoresis and 45. Frank C, Werber D, Cramer JP, Askar M, Faber
phage typing. J Clin Microbiol 32:30133017. M, an der Heiden M, Bernard H, Fruth A,
38. Izumiya H, Terajima J, Wada A, Inagaki Y, Prager R, Spode A, Wadl M, Zoufaly A, Jordan
Itoh KI, Tamura K, Watanabe H. 1997. Molec- S, Kemper MJ, Follin P, Muller L, King LA,
ular typing of enterohemorrhagic Escherichia coli Rosner B, Buchholz U, Stark K, Krause G. 2011.
O157:H7 isolates in Japan by using pulsed-field gel Epidemic profile of Shiga-toxin-producing Esch-
electrophoresis. J Clin Microbiol 35:16751680. erichia coli O104:H4 outbreak in Germany. N Engl
39. Akiba M, Masuda T, Sameshima T, Katsuda K, J Med 365:17711780.
Nakazawa M. 1999. Molecular typing of Esche- 46. Brzuszkiewicz E, Thurmer A, Schuldes J,
richia coli O157:H7 (H-) isolates from cattle in Leimbach A, Liesegang H, Meyer FD, Boelter
Japan. Epidemiol Infect 122:337341. J, Petersen H, Gottschalk G, Daniel R. 2011.

70 SADIQ ET AL.
Genome sequence analyses of two isolates from and genetic subtypes. Appl Environ Microbiol
the recent Escherichia coli outbreak in Ger- 78:66896703.
many reveal the emergence of a new pathotype: 56. Leopold SR, Magrini V, Holt NJ, Shaikh N,
entero-aggregative-haemorrhagic Escherichia coli Mardis ER, Cagno J, Ogura Y, Iguchi A, Hayashi
(EAHEC). Arch Microbiol 193:883891. T, Mellmann A, Karch H, Besser TE, Sawyer
47. Maiden MC, Bygraves JA, Feil E, Morelli G, SA, Whittam TS, Tarr PI. 2009. A precise re-
Russell JE, Urwin R, Zhang Q, Zhou J, Zurth K, construction of the emergence and constrained
Caugant DA, Feavers IM, Achtman M, Spratt radiations of Escherichia coli O157 portrayed by
BG. 1998. Multilocus sequence typing: a portable backbone concatenomic analysis. Proc Natl Acad
approach to the identification of clones within Sci USA 106:87138718.
populations of pathogenic microorganisms. Proc 57. Abu-Ali GS, Ouellette LM, Henderson ST,
Natl Acad Sci USA 95:31403145. Lacher DW, Riordan JT, Whittam TS, Manning
48. Wirth T, Falush D, Lan R, Colles F, Mensa P, SD. 2010. Increased adherence and expression
Wieler LH, Karch H, Reeves PR, Maiden MC, of virulence genes in a lineage of Escherichia coli
Ochman H, Achtman M. 2006. Sex and virulence O157:H7 commonly associated with human in-
in Escherichia coli: an evolutionary perspective. fections. PLoS One 5:e10167.
Mol Microbiol 60:11361151. 58. Tomb JF, White O, Kerlavage AR, Clayton RA,
49. Jaureguy F, Landraud L, Passet V, Diancourt L, Sutton GG, Fleischmann RD, Ketchum KA,
Frapy E, Guigon G, Carbonnelle E, Lortholary Klenk HP, Gill S, Dougherty BA, Nelson K,
O, Clermont O, Denamur E, Picard B, Nassif X, Quackenbush J, Zhou L, Kirkness EF, Peterson
Brisse S. 2008. Phylogenetic and genomic diver- S, Loftus B, Richardson D, Dodson R, Khalak
sity of human bacteremic Escherichia coli strains. HG, Glodek A, McKenney K, Fitzegerald LM,
BMC Genomics 9:560. Lee N, Adams MD, Hickey EK, Berg DE,
50. Whittam TS, Wachsmuth IK, Wilson RA. 1988. Gocayne JD, Utterback TR, Peterson JD,
Genetic evidence of clonal descent of Escherichia Kelley JM, Cotton MD, Weidman JM, Fujii C,
coli O157:H7 associated with hemorrhagic colitis Bowman C, Watthey L, Wallin E, Hayes WS,
and hemolytic uremic syndrome. J Infect Dis Borodovsky M, Karp PD, Smith HO, Fraser
157:11241133. CM, Venter JC. 1997. The complete genome se-
51. van Belkum A, Scherer S, van Alphen L, quence of the gastric pathogen Helicobacter py-
Verbrugh H. 1998. Short-sequence DNA repeats lori. Nature 388:539547.
in prokaryotic genomes. Microbiol Mol Biol Rev 59. Alm RA, Ling LS, Moir DT, King BL, Brown ED,
62:275293. Doig PC, Smith DR, Noonan B, Guild BC,
52. Yang Z, Kovar J, Kim J, Nietfeldt J, Smith DR, deJonge BL, Carmel G, Tummino PJ, Caruso A,
Moxley RA, Olson ME, Fey PD, Benson AK. Uria-Nickelsen M, Mills DM, Ives C, Gibson R,
2004. Identification of common subpopulations Merberg D, Mills SD, Jiang Q, Taylor DE, Vovis
of non-sorbitol-fermenting, beta-glucuronidase- GF, Trust TJ. 1999. Genomic-sequence compar-
negative Escherichia coli O157:H7 from bovine ison of two unrelated isolates of the human gas-
production environments and human clinical sam- tric pathogen Helicobacter pylori. Nature 397:
ples. Appl Environ Microbiol 70:68466854. 176180.
53. Ogura Y, Ooka T, Iguchi A, Toh H, Asadulghani 60. Tobe T, Beatson SA, Taniguchi H, Abe H, Bailey
M, Oshima K, Kodama T, Abe H, Nakayama K, CM, Fivian A, Younis R, Matthews S, Marches
Kurokawa K, Tobe T, Hattori M, Hayashi T. O, Frankel G, Hayashi T, Pallen MJ. 2006. An
2009. Comparative genomics reveal the mecha- extensive repertoire of type III secretion effectors
nism of the parallel evolution of O157 and non- in Escherichia coli O157 and the role of lambdoid
O157 enterohemorrhagic Escherichia coli. Proc phages in their dissemination. Proc Natl Acad Sci
Natl Acad Sci USA 106:1793917944. USA 103:1494114946.
54. Eppinger M, Daugherty S, Agrawal S, Galens K, 61. Touchon M, Hoede C, Tenaillon O, Barbe V,
Sengamalay N, Sadzewicz L, Tallon L, Cebula Baeriswyl S, Bidet P, Bingen E, Bonacorsi S,
TA, Mammel MK, Feng P, Soderlund R, Tarr Bouchier C, Bouvet O, Calteau A, Chiapello H,
PI, Debroy C, Dudley EG, Fraser CM, Ravel J. Clermont O, Cruveiller S, Danchin A, Diard M,
2013. Whole-genome draft sequences of 26 entero- Dossat C, Karoui ME, Frapy E, Garry L, Ghigo
hemorrhagic Escherichia coli O157:H7 strains. JM, Gilles AM, Johnson J, Le Bouguenec C,
Genome Announc 1:e0013412. Lescat M, Mangenot S, Martinez-Jehanne V,
55. Norman KN, Strockbine NA, Bono JL. 2012. Matic I, Nassif X, Oztas S, Petit MA, Pichon C,
Association of nucleotide polymorphisms within Rouy Z, Ruf CS, Schneider D, Tourret J,
the O-antigen gene cluster of Escherichia coli O26, Vacherie B, Vallenet D, Medigue C, Rocha EP,
O45, O103, O111, O121, and O145 with serogroups Denamur E. 2009. Organised genome dynamics

in the Escherichia coli species results in highly toxin-producing E. coli. Infect Immun 66:2553
diverse adaptive paths. PLoS Genet 5:e1000344. 2561.
62. Rasko DA, Rosovitz MJ, Myers GS, Mongodin 66. Morabito S, Tozzoli R, Oswald E, Caprioli A.
EF, Fricke WF, Gajer P, Crabtree J, Sebaihia M, 2003. A mosaic pathogenicity island made up of
Thomson NR, Chaudhuri R, Henderson IR, the locus of enterocyte effacement and a patho-
Sperandio V, Ravel J. 2008. The pangenome genicity island of Escherichia coli O157:H7 is fre-
structure of Escherichia coli: comparative ge- quently present in attaching and effacing E. coli.
nomic analysis of E. coli commensal and patho- Infect Immun 71:33433348.
genic isolates. J Bacteriol 190:68816893. 67. Wieler LH, McDaniel TK, Whittam TS, Kaper
63. Kulasekara BR, Jacobs M, Zhou Y, Wu Z, JB. 1997. Insertion site of the locus of enterocyte
Sims E, Saenphimmachak C, Rohmer L, Ritchie effacement in enteropathogenic and enterohem-
JM, Radey M, McKevitt M, Freeman TL, orrhagic Escherichia coli differs in relation to the
Hayden H, Haugen E, Gillett W, Fong C, Chang clonal phylogeny of the strains. FEMS Microbiol
J, Beskhlebnaya V, Waldor MK, Samadpour M, Lett 156:4953.
Whittam TS, Kaul R, Brittnacher M, Miller SI. 68. Hazen TH, Sahl JW, Fraser CM, Donnenberg
2009. Analysis of the genome of the Escherichia MS, Scheutz F, Rasko DA. 2013. Refining the
coli O157:H7 2006 spinach-associated outbreak pathovar paradigm via phylogenomics of the
isolate indicates candidate genes that may en- attaching and effacing Escherichia coli. Proc Natl
hance virulence. Infect Immun 77:37133721. Acad Sci USA 110:1281012815.
64. Scheutz F, Nielsen EM, Frimodt-Moller J, Boisen 69. Reid SD, Herbelin CJ, Bumbaugh AC, Selander RK,
N, Morabito S, Tozzoli R, Nataro JP, Caprioli A. Whittam TS. 2000. Parallel evolution of virulence
2011. Characteristics of the enteroaggregative in pathogenic Escherichia coli. Nature 406:6467.
Shiga toxin/verotoxin-producing Escherichia coli 70. Rasko DA, Worsham PL, Abshire TG, Stanley
O104:H4 strain causing the outbreak of haemolytic ST, Bannan JD, Wilson MR, Langham RJ,
uraemic syndrome in Germany, May to June 2011. Decker RS, Jiang L, Read TD, Phillippy AM,
Euro Surveill 16(24):ii, 19889. Salzberg SL, Pop M, Van Ert MN, Kenefic LJ,
65. Boerlin P, Chen S, Colbourne JK, Johnson R, Keim PS, Fraser-Liggett CM, Ravel J. 2011.
De Grandis S, Gyles C. 1998. Evolution of entero- Bacillus anthracis comparative genome analysis in
hemorrhagic Escherichia coli hemolysin plasmids support of the Amerithrax investigation. Proc
and the locus for enterocyte effacement in shiga Natl Acad Sci USA 108:50275032.

PATHOGENESIS

Role of Shiga/Vero Toxins
in Pathogenesis
FUMIKO OBATA1 and TOM OBRIG1

5
ACTIVITIES OF Stx AND LPS IN RENAL DISEASE
Shiga Toxin Actions

It is generally accepted that all actions of Shiga toxin (Stx) depend on its in-
teraction with the receptor, globotriaosylceramide (Gb3), on eukaryotic cells.
Although alternative receptors for Stx have been postulated, no definitive data
have been forthcoming in support. Stx holotoxin is internalized by receptor-
mediated endocytosis, retrograde transported via the Golgi apparatus and pro-
cessed through in the endoplasmic reticulum, and released into the cytoplasm
where it enzymatically inactivates ribosomes and inhibits protein synthesis
(Fig. 1). However, it is important to note that, in addition to Stx holotoxin, the
B-subunit alone can interact with Gb3 in a physiologically meaningful manner
where it activates signal transduction pathways in target cells (Fig. 1) (1). An
additional but unexplained anomaly is the interaction of Stx with eukaryotic
cells in a Gb3-independent manner that leads to induction of cytokines by these
cells (2). As shown in Fig. 1, intracellular responses to Stx are diverse, including
inhibition of protein synthesis, activation of cellular stress responses, and in-
duction of cytokines and chemokines. It is likely that these different schemes
take place in cell-specific activities during Shiga toxin-producing Escherichia
1
University of Maryland School of Medicine, Baltimore, MD 21201.
75

76 OBATA AND OBRIG
FIGURE 1 Schema: Shiga toxin interaction with eukaryotic cells. doi:10.1128/microbiolspec.EHEC-0005-2013.f1
coli (STEC) infections in humans, culminating sensitive to Stx in vitro (7). However, other
in typical hemolytic-uremic syndrome (HUS). cells that make up the human renal glomeru-
As depicted, it is clear that in some cases lus are also sensitive to Stx, including podo-
Stx can result in activation of p38 mitogen- cytes and mesangial cells (8, 9). In addition,
activated protein kinase as well as apoptotic extraglomerular epithelial cell types of the
and necrotic cell death (Fig. 1). The topic of human kidney have been postulated to be tar-
HUS renal disease has been reviewed recently gets of Stx, including proximal tubule and col-
(35). lecting duct cells (8, 10, 11). Cell types in the
blood circulation that may be key to devel-
opment of HUS and that are sensitive to Stx
Cell Types Responsive to Stx
include platelets, neutrophils, and monocytes
The high number of Stx-sensitive cell types (1216).
makes more difficult identification of more In summary, most, if not all, of the cell
important events responsible for HUS. Renal types mentioned may well have a role in
microvascular endothelial cells are generally STEC-related kidney disease and typical HUS.
accepted to be the primary target of Stxs in The relative importance and role of these cell
HUS. Data in support of this concept come types in STEC HUS remain to be determined.
from many sources, most notably autopsy kid- For example, it is not clear which of the renal
ney pathology samples showing swollen and cell types are actually responsible for renal
detached endothelial cells accompanied by failure in STEC HUS, although apoptosis of
thrombi (6). Such human renal microvascular tubules appears to be a common feature (8,
endothelial cells were also shown to be very 17). The relative contributions in HUS disease

CHAPTER 5 Shiga/Vero Toxins in Pathogenesis 77
of renal microvascular coagulation and throm- HUS, a pathway leading to fibrin deposition
bosis (i.e., endothelial cells), imbalance of was revealed (Fig. 2). LPS activation of cells
fluid and electrolytes (i.e., nephron tubules), such as endothelial and renal tubule cells elic-
and altered filtration barrier function (i.e., ited chemokines (monocyte chemotactic pro-
endothelial and podocyte cells) have yet tein 1 [MCP-1], macrophage inflammatory
to be elucidated for typical HUS. If in vitro protein 1 [MIP-1] alpha, RANTES) known as
cell culture studies are pertinent to HUS chemoattractants for monocyte/macrophage
in patients, the sensitivity (50% lethal dose) cells and coactivators of platelets. In this re-
of human renal cells to Stx2 (endothelial, sponse, Stx enhances the effects of but does
0.1 pM > podocyte, 0.5 pM >> proximal tubule, not replace LPS. The response was associated
10 pM) suggests the renal filtration barrier is with renal fibrin deposition (12, 20). In the
at considerable risk (8). murine model, simultaneous neutralization of
these three chemokines inhibited LPS/Stx-
induced monocyte accumulation and fibrin
Inammatory Cells, Chemokines, and
deposition in the kidneys (20). Further, ad-
Renal Thrombosis
ministration of adenosine A2a receptor ago-
A primary feature in the renal pathology of nists to Stx/LPS mice also reduced monocyte
STEC HUS is microvascular coagulation and and fibrin accumulation in the kidneys. As
thrombosis. In humans and in a murine model shown in Fig. 3, adenosine A2a receptor ago-
of HUS, the interaction of Stx and lipopoly- nists act as anti-inflammatory agents in mono-
saccharide (LPS) with circulating cells and re- cytes, platelets, and endothelial cells (21).
sident renal cells appears to have a causal role Taken together, these studies indicate that
in microvascular thrombosis (18, 19). In a se- both LPS and Stx are required for maximal
ries of studies of the Stx/LPS murine model of renal fibrin deposition and that platelets may
FIGURE 2 Proposed pathways of Stx and LPS actions in mice. Data derived from a Stx/LPS murine model of HUS
indicate that LPS is the primary elicitor of brin deposition in kidneys. This pathway requires chemokines and
platelets but is not responsible for renal failure. Stx is responsible for renal failure in this murine model in a
process that involves nonendothelial renal cell types. doi:10.1128/microbiolspec.EHEC-0005-2013.f2

78 OBATA AND OBRIG
FIGURE 3 Anti-inammatory actions of adenosine in HUS. Data derived from an Stx/LPS murine model of HUS
suggest adenosine A2a receptor agonist, i.e., adenosine, effectively blocks the actions of LPS (enhanced by
Stx2) at the level of different renal cell types to prevent platelet activation and coagulation.
be required. Because mice deficient in MCP-1 chemokines are being generated from other
have sharply reduced platelet deposition after cell types such as renal tubules (20). Important
exposure to Stx/LPS, we have suggested that conclusions from the murine HUS model are
this chemokine serves as a coactivator of plate- that LPS, not Stx, is the initial primary elicitor
lets in typical HUS (Keepers TR, unpublished of renal coagulation and thrombosis, but Stx,
data). The primary activators of platelet acti- not LPS, is the lethal agent of STEC.
vation are thrombin or adenosine diphosphate. In the murine Stx/LPS model of HUS,
Our renal gene array analysis of the LPS re- monocyte migration into the kidneys was re-
sponse in mice indicated that LPS strongly stricted to the extraglomerular space in con-
elicited fibrinogen mRNA, the precursor of trast to polymorphonuclear leukocytes (PMN),
fibrin (Obrig T, unpublished data). In addition, which, in addition, migrated into the glomer-
it is noteworthy that selective elimination of uli. The latter may be important in humans
monocytes from mice prior to the above stud- because neutrophilia has been implicated as a
ies had no effect on the ability of Stx/LPS to primary risk factor for HUS disease and in-
elicit renal fibrin deposition, suggesting the creased neutrophil migration into the kidneys

was a key observation in HUS renal biopsies Renal Gene Array Analysis of
(22, 23). In the murine model of HUS, the Murine Responses to Stx2 and LPS
neutrophil chemotactic factors chemokine li-
Much information is now available regarding
gand 1 (CXCL1) keratinocyte-derived chemo-
the biological effects of Stx2 and LPS on kid-
kine (KC) and CXCL2 (MIP-2) were induced
neys in the murine HUS model. The following
in the kidneys by LPS (15). The induction was
is a synopsis of the more pertinent gene mi-
at the transcriptional level and was enhanced
croarray data obtained from temporal studies
by Stx2. Administration of neutralizing anti-
of the murine renal responses to Stx2, LPS,
bodies for these neutrophil chemotactic fac-
or Stx2/LPS (19). On the basis of the total of
tors prevented the movement of neutrophils
both up- and downregulated genes, five times
into the kidneys. It was also demonstrated that
more renal genes responded to LPS than to
vascular cell adhesion molecule 1 (VCAM-1)
Stx2 over the 72-h time course. Response to
was induced in the kidneys simultaneously
LPS was mostly early, whereas Stx2 responses
with CXCL-1 and CXCL-2 in response to
occurred later in the 72-h time course. These
Stx2/LPS in mice (Fig. 4). VCAM-1 is known
results are more meaningful when viewed in
to assist movement of neutrophils across the
the larger picture of HUS where renal failure
endothelium and appeared to exhibit this func-
occurs later in the time course in both mice
tion for neutrophils in the Stx2/LPS murine
and humans. It should be emphasized that
model of HUS. However, the relative impor-
Stx2, rather than LPS, is the lethal factor
tance of renal neutrophils in Stx-induced renal
in the murine HUS model. The gene array
failure has yet to be determined in mice and
data revealed different roles for LPS and Stx2
humans.
in the renal physiological responses. LPS re-
sponses were mostly inflammatory, stress re-
lated, or cell defensive in nature. In contrast,
Stx2 responses were related to cell repair and
involved cell proliferation and differentiation
or cell cycle control genes. An interesting find-
ing was that renal genes downregulated by
Stx2 included membrane transporters, which
appeared to signal a protective survival mode
and slowing of cell metabolism.
The renal genes most upregulated by Stx2
or LPS are depicted in Fig. 5. As expected
from the inflammatory responses described
above, LPS induced a number of chemokine
genes that code for chemotactic factors for
monocytes and neutrophils. These tend to
be immediate response genes, which attract
monocytes and neutrophils into the kidneys
and set the stage for a broad inflammatory
FIGURE 4 Neutrophil-endothelial cell interactions in response in the kidneys. Such LPS immedi-
HUS. In the Stx2/LPS murine model of HUS, analysis ate response genes are mentioned in the lit-
of renal gene activation and neutrophil inltration erature in descriptions of typical HUS, i.e.,
into kidneys demonstrates a concomitant increase in MCP-1, MIP-2alpha, and the murine inter-
PMNs and VCAM-1 expression, suggesting a mecha-
nism of PMN-endothelial association. , Neutrophils
leukin-8 mimic, KC. It was also observed that
in the glomeruli; , VCAM-1 in the glomeruli. interferon-gamma-inducible protein-10 (IP-
doi:10.1128/microbiolspec.EHEC-0005-2013.f4 10) (CXCL10) was induced by LPS and by

80 OBATA AND OBRIG
FIGURE 5 Renal gene activation in the Stx/LPS murine model. Shown are the nine most upregulated genes in
the temporal response of mice to either LPS or Stx2. Gene microarrays were employed to analyze kidney gene
activation over a 72-h response of C57BL/6 mice to 300 g/kg of LPS or 100 ng/kg of Stx2.
Stx2, albeit in early and late parts of the HUS such as lupus nephritis (25, 26). Lipocalin 2
disease time course, respectively. Related to (neutrophil gelatinase-associated lipocalin),
renal coagulation and thrombosis in HUS, an LPS-induced early gene (Fig. 5), is a com-
LPS induced a set of fibrinogen genes late in mon urine biomarker for numerous renal dis-
the time course of the murine model of HUS eases, including STEC-HUS (27).
concomitant with the appearance of fibrin
deposition and coagulation in the renal mi-
How Valid Is the Murine Model of HUS for
crovasculature of HUS (Fig. 5). These data
Translation to the Human Disease?
agree with our observation that LPS is re-
sponsible, in part, for fibrin deposition in the A large volume of data exists for mouse models
Stx2/LPS murine model of HUS (19). Amy- of Stx-HUS (28). The two common experi-
loid protein, which has been reported to be mental approaches for these murine models
a Stx-sensitizing factor in HUS, is induced are either oral infection with STEC or injec-
at the mRNA level by LPS in mice, as shown tion with purified Stx plus or minus LPS (17,
in Fig. 5, as a renal late gene product (24). 19, 29, 30). In virtually all cases these are le-
More recently, complement has been identi- thality models within 4 to 12 days after expo-
fied as a factor that may contribute to renal sure to the agents and are accompanied by
failure in atypical HUS. renal damage. Where examined, these murine
Products of some of the genes shown in models usually exhibit the three hallmarks of
Fig. 5 have been examined by investigators as HUS: thrombocytopenia, hemolytic anemia,
potential biomarkers for diagnostic purposes. and renal failure. However, every animal model
For example, IP-10 has been identified as a has its limitations, and for the murine models
urine biomarker for other kidney diseases of HUS, the renal microvascular endothelial

cells do not express Gb3 and are resistant to Stx of extremities. Most frequently, the hind legs
action. This is important if one believes that the are affected first, followed by the forelegs.
primary target of Stx is the renal microvascular Other symptoms include anorexia, lethargy,
endothelium. Indeed, human renal endothelial ataxic gait, recumbency (the affected animals
cells in vitro are very sensitive to Stx, and the lose strength to hold their body in an up-
pathology of human kidneys in HUS describes right position), convulsions, seizure, coma,
swollen and detached glomerular endothelial and death.
cells. But it is surprising why such human glo- STEC oral administration animal models
merular endothelium is not killed by Stx in are summarized in Table 1. The oral inoc-
HUS kidneys. This suggests either a more in- ulation models of STEC that describe CNS
direct action of Stx in human HUS or domi- symptoms are limited to pig and mouse. Pigs
nant survival activities are activated within the develop edema disease with Stx2e-producing
endothelium after exposure to Stx. An alter- E. coli and present CNS symptoms (Table 2).
native explanation is that the primary target Experimentally, the edema disease-like state is
of Stx in human kidneys is not the endotheli- reproducible with Stx2-producing E. coli that
um, but rather glomerular podocytes and extra- has been isolated from human patients. CNS
glomerular tubules along the nephron. Support symptoms are only seen in Stx2 (both Stx2
for this exists for HUS in mice and humans and Stx2e) producers, but not in non-Stx2
where urine specific gravity changes, chemo- producers. This indicates a strong association
kines are increased in the urine, and bio- of Stx2 with CNS impairment.
markers of damaged podocytes and tubule cells LPS is an outer membrane component of
are detected. gram-negative bacteria and a strong inflam-
Mouse models have been helpful in sepa- mation inducer. The involvement of LPS in
rating the actions of Stx and LPS in HUS. In STEC-associated CNS symptoms was tested
general, and as described above, LPS is the by using LPS nonresponder mouse C3H/HeJ
primary inducer of cytokines and chemokines (29). C3H/HeJ did present CNS symptoms
where Stx enhances the activity of LPS. The when given Stx2-producing E. coli but not
complexity of inflammation in HUS is critical when Stx-nonproducer was inoculated. This
but has yet to be fully delineated in murine again suggests a strong involvement of Stx in
models and in human HUS. The murine CNS symptoms. The difference between LPS-
model mirrors typical HUS of humans as responder mouse (C3H/HeN) and C3H/HeJ
resting platelets are resistant to Stx and re- in CNS symptoms was that C3H/HeN showed
quire preactivation with LPS (19). However, a progressive time course of CNS symptoms
it is most important to reiterate that Stx, whereas C3H/HeJ showed a biphasic re-
not LPS, is responsible for the renal failure sponse in that they developed milder CNS
in typical HUS. In conclusion, the murine re- symptoms and recovered once, but then prog-
sponses to Stx and LPS include most of the ressed to a severe form of CNS impairment.
features of STEC-HUS in humans. This suggests that even though Stx2 may be
the central cause of CNS symptoms, addition
of LPS response may contribute to the prog-
ress of the disease.
ACTIVITIES OF Stx IN CNS DISEASE
To further study the action of Stx2 in CNS
disease, different animals were tested with
CNS Symptoms of Animal Models
purified Stx2. Stx2 injection animal models
In either an oral inoculation of STEC model or with CNS complications are summarized in
purified Stx injection animal model, the most Table 3. Also, LPS involvement or contribu-
common and most frequently reported central tion to Stx2-associated CNS disease was tested
nervous system (CNS) impairment is paralysis in some reports. The reproducible results of

82 OBATA AND OBRIG
TABLE 1 STEC oral administration model with CNS descriptions

Ref. Animal E. coli strain CNSa Histopathologyb,c,d IHCe/TUNELb,c,f Model notes Other assays
31 Pig RCH/86 (Stx2+) Yes HE: CL cap, small inf, ND Gnotobiotic NA

small hrrg, b in (cesarean
sub and cap section
derived)
32 Pig 86-24 (Stx2+) Yes Gross: MO and ND Suckling NA
CL hrrg and necPAS: (colostrum
MO, CL, and Sc, provided)
cap swl nec,
peri deposits
32 Pig 87-23 Stx () No No lesion ND Suckling NA
(colostrum
provided)
33 Pig S1191 (Stx2e+) Yes EM: myo and TUNEL + myo 3-w-o NA
M112 (Stx2e+) cap nec not apop, in MO
mono apop (5/11 pigs)
33 Pig Strain 123 No No lesion ND 3-w-o NA
(nonpathogenic
E. coli)
34 Pig sakai Yes LFB: mye deg, ND Neonatal NA
hrrg, pyk and
prolif cap,
peri ede
35 ICR E32511/HSC ND EM: cap ede in Immuno Sm, MMCh Tracer (i.v.)
mouse (Stx2c+)i CR ctx, mye degHE: EM-DABg: detected in
hrrg and ede in Stx2+ in CR cap and
CR ctx only in ctx pyr and deg mye
CNS symptom deg mye
(+) mice
29 C3H/HeN 86-24, 86BL or Yes ND ND Fasted NA
mouse 134 (Stx2+)
29 C3H/HeN 87-23, 87BL Yes ND ND Fasted NA
mouse (Stx2)
29 C3H/HeJ 86-24, 86BL or Yes ND ND Fasted NA
mouse 134 (Stx2+) (biphasic)
29 C3H/HeJ 87-23, 87BL No ND ND Fasted NA
mouse (Stx2)
36 C57BL/6 N-9 ND HE: inlt, hrrg, Anti-Stx + PCMj NA
mouse (Stx1+/Stx2+) cap with b in hippo
CR ctxLFB:
No deg mye
in hippo
37 IQI EDL931 Yes HE: ede, b in ND Gnotobiotic Brain TNF
mouse (Stx1+/Stx2+) cap, neu deg, increased
cap prolif
55 C57BL/6 Smr N-9 ND ND TUNEL + PCM Serum Stx ,
mouse (Stx1+/Stx2+) hippo neu TNF ,
during CNS IL10 TLC-
symptom (+) anti-PkMab k
brain +
38 IQI O157:H7 strain 6 Yes HE: CR ctx and ND Gnotobiotic ND
mouse (Stx1+/Stx2+) CL neu nec
and slight loss
of Purkinje
(Continued on next page)

TABLE 1 (Continued)
Ref. Animal E. coli strain CNSa Histopathologyb,c,d IHCe/TUNELb,c,f Model notes Other assays
m
56 ICR E32511/HSC Yes ND GFAP , Sm, MMC ISHl Gb3
mouse (Stx2c+) AQP4, synthase
casp3nneu
cer Sc ventral
and MO
dorsal
a
Detailed CNS symptoms are summarized in Table 2.
b
Histopathology analysis keys are Gross, gross observation in nonstained tissue; HE, hematoxylin-esosin stain that stains cytoplasm in
pink and nucleus blue, light microscopic ndings (LM); PAS, periodic acid-Schiff stain that detects polysaccharides, glycoproteins, and
glycolipid, LM; LFB, Luxol fast blue stain that stains myelin in blue, LM; EM, electron microscopic ndings; ND, not described; NA, not
applicable.
c
CNS regions and cell type abbreviations are CR, cerebrum; ctx, cortex; hippo, hippocampus; str, striatum; CL, cerebellum; MO, medulla
oblongata; Sc, spinal cord; cer cervical; tho, thoracic; lum, lumbaris; sub, subarachinoid space; BS, brain stem is used where midbrain,
pons, or medulla oblongata is not specied. Histopathologic feature abbreviations are cap, endothelial cells or capillaries; inf, infarction;
hrrg, hemorrhage; b, brin deposition; nec, necrosis; swl, swelling; peri, perivascular; myo, myocytes; apop, apoptotic; mono,
monocytes; mye, myelin; deg, degeneration; pyk, pyknotic nuclei; prolif, proliferation/hyperplasia; ede, edema; pyr, pyramidal neuron;
int, inltration of blood cells to parenchyma; neu, neuron; Purkinje, Purkinje cells are large neurons in CL.
d
Histopathologic feature abbreviations are cap, endothelial cells or capillaries; inf, infarction; hrrg, hemorrhage; b, brin deposition; nec,
necrosis; swl, swelling; peri, perivascular; myo, myocytes; apop, apoptotic; mono, monocytes; mye, myelin; deg, degeneration; pyk,
pyknotic nuclei; prolif, proliferation/hyperplasia; ede, edema; pyr, pyramidal neuron; int, inltration of blood cells to parenchyma; neu,
neuron; Purkinje, Purkinje cells are large neurons in CL.
e
IHC, immunohistochemistry, immunodetection of the target in the tissue sections.
f
TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling detects DNA fragmentation that is a hallmark of apoptosis.
g
Immuno-EM-DAB, immunodetection of the target with 3,3-diaminobenzidine deposition by electron microscopy.
h
Sm, streptomycin; MMC, mitomycin C.
i
Smr, MMCr.
j
PCM, protein calorie malnutrition.
k
TLC-anti-PkMab (thin layer chromatography with anti-Pk monoclonal antibody detectin).
l
ISH, in situ hybridization.
m
GFAP, glial brillary acidic protein, an astrocyte marker; an increase of GFAP suggests astrogliosis.
n
IHC for activated (cleaved) caspase-3.
hind-leg paralysis and high frequency of is a great possibility that analyzing these ani-
convulsions and seizures with purified Stx mal models may give some clues to define
confirm the central role of the toxin in STEC- the mechanisms of CNS impairment in Stx-
associated CNS disease. Human STEC pa- associated disease.
tients present CNS symptoms that range from
eye involvement (diplopia, hallucinations, and
CNS Histopathology of Animal Models
cortical blindness), behavioral changes (hyper-
activity, distractibility, irritability, and altered In animal models with STEC oral inoculation
sensorium), posturing/coordination difficul- that describe CNS symptoms, most exhibit
ties (poor fine-motor coordination, hemiplegia, defective capillaries (pig [3134], mouse [35,
ataxia, and clumsiness), to severe symptoms 36]). Those capillary lesions are mostly related
such as seizures, dysregulation of breathing, to endothelial cell weakening that appears as
and alteration in consciousness such as coma. hemorrhage, with leaked red blood cells in
Within these varieties of symptoms, ataxia parenchyma. Noncapillary components in the
or hemiparesis resembles Stx-associated ani- parenchyma such as neurons and myelin de-
mal CNS symptoms. Also, it is notable that in fects were seen in some mouse STEC models
human patients, seizures are a frequent ob- (35, 37, 38), but not others (36). In purified
servation. This resemblance between patients Stx2 injection models, similar lesions involv-
and animal models of STEC/Stx suggests there ing capillary/endothelial cells were found in

84 OBATA AND OBRIG
TABLE 2 Observed CNS symptoms in animal modelsa

HL FL
Ref. Animal Model ANOX LTHG para para ATX RCMb CV/TR SZR Coma Death Other
c
46 Baboon Stx1 ND ND ND ND ND ND ND ND ND Yes
67 Baboon Stx1 Yes ND ND ND ND ND ND 3/6 Yes Yes
(50%)
31 Pig STEC Yes Yes Yes ND Yes Yes Yes Yes Yes Yes
71 Pig STEC ND ND ND ND Yes Yes ND ND ND Yes Diarrhea,
then CNS+
32 Pig STEC ND ND Yes Yes ND Yes Yes ND ND Yes Paddling
39 Pig Stx2e Yes ND ND ND Yes ND Yes ND Yes Yes Paddling,
i.v. extensor
rigidity,
dyspnea
40 Pig Sup ND ND ND ND Yes Yes Yes ND Yes Yes Paddling,
Stx2e extensor
i.v. rigidity
33 Pig STEC ND ND ND ND Yes Yes ND ND ND ND
(1/11) (1/11)
41 Rabbit Stx1 Yes Yes Yes ND Yes ND ND ND ND Yes
42 Rabbit Stx1 i.v. Yes Yes Yes Yes ND Yes No ND ND Yes Rufed fur,
rapid
respiration
52 Rabbit Stx2 i.v. ND ND Yes Yes Yes ND Yes ND ND Yes Opisthotonic
(50%) (50%) (33%) (50%) (50%) posture
47 Rabbit Stx2 i.v. Yes Yes Yes Yes ND Yes ND ND ND Yes
and i.t.
68 Rabbit Stx2 i.v. Yes ND Yes Yes ND ND ND ND ND Yes
54 Rabbit Stx2 i.v. Yes ND Yes Yes ND ND ND ND ND Yes Dyspnea
43 Rabbit Stx2 i.v. Yes ND Yes ND Yes ND ND ND ND Yes
(83.3%) (83.3%)
48 Rabbit Stx2 i.v. ND ND Yes ND ND ND Yes ND ND ND
(25%) (25%)
57 Rat Stx2 ND Yes Yes ND ND ND ND Yes ND Yes Crawling
i.c.v.
35 Mouse STEC ND Yes Yes Yes ND ND ND ND ND Death Deformity of
backbone,
loss of
pain
29 Mouse STEC ND ND Yes Yes Yes ND Yes ND Yes Yes Jerky
rhythmic
motion
36 Mouse STEC Yes Yes Yes ND ND ND (Yes)d ND ND Yes Rufed fur,
jerky
rhythmic
motion
37 Mouse STEC Yes Yes Yes ND ND ND ND ND ND Yes
44 Mouse Stx2 i.v. ND ND Yes ND ND ND ND ND ND Yes
44 Mouse Stx2 ND ND ND ND ND ND Yes Yes ND Yes
+LPS i.v.
38 Mouse STEC Yes Yes Yes ND ND ND Yes ND ND Yes Rufed
fur

TABLE 2 (Continued)
HL FL
Ref. Animal Model ANOX LTHG para para ATX RCMb CV/TR SZR Coma Death Other
53 Mouse Stx2 i.p. ND Yes Yes ND Yes ND Yes Yes ND Yes Retain
sense
(pain)
56 Mouse STEC ND ND Yes ND ND ND Yes ND ND Yes Spinal
deformity
a
Abbreviations for CNS symptoms are ANOX, anorexia; LTHG, lethargy; HL para, hind-leg paralysis; FL para, foreleg paralysis; ATX, ataxic
gait; RCM, recumbency, difculty holding body upright by itself; CV/TR, convulsions/tremors; SZR, seizure. Injection route abbreviations:
i.c.v., intracerebroventricular; i.p., intraperitoneal; i.t., intrathecal; i.v., intravenous.
b
Lateral, sternal, or dorsal recumbency; the animal is lying down with leaning on its side, abdomen, or back, having difculty holding its
body upright.
c
ND, not described.
d
Shivering.
pig (39, 40), rabbit (4143), and mouse (44, nick end labeling (TUNEL) stain detects frag-
45). In contrast, other models did not have mented DNA and therefore is often used as
these lesions but rather lesions related to neu- an apoptotic assay. Capillaries (pig [33], rab-
ronal degeneration (baboon [46], rabbit [43, bit [43, 54]), neurons (mouse [55], rabbit [43]),
47, 48], rat [49, 50], mouse [51]) or myelin de- and glial cells (rabbit [43]) have been detected
generation (baboon [46], rabbit [52], rat [49]). as TUNEL positive. Activated caspase-3 tar-
Also, some reports showed normal appear- geted immunohistochemistry has been used
ance of neurons (rabbit [47], striatal neurons; for another marker of apoptotic cells. Neurons
mouse [53] lumbar spinal cord neurons). As all (mouse [56]) and capillaries (rabbit [54]) have
models exhibit similar CNS symptoms such as been detected positive. Another pro-apoptotic
hind-leg paralysis, the difference in histopath- marker, bax, was found increased in rat neu-
ological lesions may be due to involvement of rons (57). Along with electron microscopy ob-
different parts of CNS, different time points servation (rat [49]), some neurons and capillary
in the disease, or species-specific sensitivities. cells (endothelial cells and pericytes) undergo
The mechanism of inducing CNS symptoms apoptosis, but some appear as necrotic (rabbit
may be weakening of endothelial cells/capillary [33]). Careful and detailed information of
composition-caused neurotoxicity or direct ef- which area of the CNS and what types of cells
fect of Stx in neuronal toxicity. The observation in that area present apoptotic features may
of lamellipodia-like processes of glial origin help elucidate these conflicting results.
interrupting synaptic connections at the lum- Aquaporin 4 (AQP4) is mostly expressed
bar spinal cord interneuron to motor neuron in astrocyte foot processes that have a direct
may explain the resulting hind leg paralysis contact with capillaries in the CNS. The re-
(mouse [53]). A similar observation is reported duction of AQP4 suggests that there is alter-
in a rat model of striatum neurons (51). ation in astrocytic foot process, which is
important to strengthen the blood-brain bar-
rier (BBB). AQP4 expression decreased in
CNS Molecular Physiology of
Stx2-injected rat (50) and STEC-infected
Animal Model
mouse (56), while astrocytic activation marker
Molecular marker analysis in STEC or Stx glial fibrillary acidic protein increased. This
animal models suggests possible mechanisms suggests Stx-associated astrocyte activation
for Stx-associated CNS impairment. may participate in weakening the BBB.
The apoptotic nature of Stx-associated An increase in tumor necrosis factor alpha
lesions has been described. Terminal deoxy- in STEC-inoculated mouse (37) and Stx2-
nucleotidyltransferase-mediated dUTP-biotin injected rabbit (43) brain along with serum

86
TABLE 3 Shiga toxin and/or LPS administration model with CNS descriptions
Gb3/Stx Injection Other
Ref. Animala Toxin CNSb binding Histopathologyc,d,e Imaging IHCf routeg assays
Chap05.proof.3d
66 Baboonh Stx1 ND ND EM: mye deg, peri ede, large neu ND ND i.v.
86
and glia deg, cap normal
67 Baboon Stx1 Yes ND ND ND ND i.v.
39 Weaned Stx2e Yes ND Gross: Sc ede, CL ede and hrrg ND ND i.v.
OBATA AND OBRIG
YL pigs HE: CL hrrg but not ede or cap

nec, no lesions in CR, BS, thalamus
40 Weaned Supi Yes ND HE: CL, sub ede, hrrg, cap nec ND ND i.v.
YL pigs Stx2e HE: MB, b nec , peri eos
41 NZW Stx1 Yes TLC-Stx1 over lay HE: BS, Sc, CL cap narrowed, ND ND i.v.
rabbit of CL, BS and Sc (+) peri ede, cap damage and b,
at LacCerj position Purkinje decreased
125
42 NZW Stx1 Yes I-Stx1 tissue HE: cerv Sc hrrg, inf, ede, ND Anti-Stx (+) i.v.
rabbit distribution high in b in cap, lum Sc b cap, pyk cap in Sc cap
cecum, brain, small
intestine, colon, Sc
52 JW Stx2 Yes ND EM: mye deg, axo normal MRI: V3 (24 h), Immuno EM-DABk: i.v. Tracer
rabbit later BS, cc, lateral anti-Stx2 (+) at
amy, CL vermis luminal side
of cap, deg mye
Manila Typesetting Company

47 JW Stx2 Yes ND CR ctx, CL ctx, Sc, pyk neu, ND Anti-Stx2 IHC (+) i.v. Stx2 in
rabbit str neu not affected, inf in cap and sub and i.t. CSF
MB and CL
65 JW Stx2 ND ND ND MRI: CL at 82 h ND i.v. Stx2 in
rabbit CSF
65 JW Stx2 ND ND ND MRI: the rear of CL ND i.t. ND
rabbit (contact w CSF)
at 48 h
68 JW Stx2 Yes ND ND MRI: BS and cerv Sc ND i.v. Baroreex
rabbit dorsal close to death function
69 JW Stx1 ND ND ND MRI: MB, BS and Sc ND i.v. Stx1
rabbit ede, cerv Sc dorsal slightly in
hematoma CSF
03/30/15 17:36
54 JW Stx2 Yes ND HE: inf CL, myo thickening, ND Anti-Stx2 IHC(+) in i.v.
rabbit b, pyk and fragmented, pons cap and myo,
young pons myo nec anti-ssDNA IHC small
Chap05.proof.3d
number cap (+),
IHC caspase-3 and-9
87
small number (+)
43 JW Stx2 Yes Anti-Gb3 (+) HE: inf lum Sc with hrrg and ND TUNEL + in hippo i.v. qRT-PCR:
rabbit in cap lum Sc b cap, str neu pyk, hippos pyr (DG, CA1, CR ctx TNF and
neu apop neu, pons glia, cap IFN
IHC: Ib4 (microglia
activated ) in lum
Sc and thalamus
70 JW Stx2 ND ND ND MRI: Gd leak ND i.v.
rabbit (permeability)
48 DB Stx2 Yes ND HE: neu deg in the BS ND ND i.v.
rabbit
49 SD rat Stx2 ND ND ND ND Anti-Stx IHC (+) peri i.p.
49 SD rat Stx2 ND ND EM: irregular shape neu, ND Anti-Stx2 (+) in neu, i.c.v.
mye deg, hypertrophic axo, immune-EM-DAB:
astro phago mye, neu apop, deg, anti-Stx2 (+) in neu
vacuol, peri astro ede, non-peri bers and astro
astro gliosis, oligo pathologic nucleus

58 SD rat Stx2 ND ND ND ND Stx2 (+) in anterior i.c.v. Str and CR
hippo astro, ctx neu,
ips hippo Stx2(+) cap
astro and neu, NADPH-d/NOS
cont hippo Stx2(+) activity
neuropils
57 SD rat Stx2 Yes Anti-Gb3 ND ND Anti-Stx2 MAP2 i.c.v.
CA1, str neu CA1 neu.Anti-bax
neu (inner ctx, CA1,
subV dorsal str,
hypothalamic peri
50 SD rat Sup ND ND Nissle: neu pyk hypertrophy axo, ND Anti-AQP4 cp i.p.
Stx2 ede ctx, subV CL.
CHAPTER 5 Shiga/Vero Toxins in Pathogenesis

87
03/30/15 17:36
88
TABLE 3 Shiga toxin and/or LPS administration model with CNS descriptions (Continued)
Chap05.proof.3d
35 ICR Stx2 Yes ND ND ND Immune-EM-DAB: i.p. Tracer
88
mouse anti-Stx2 (+) mye
deg, lyso pyr ctx
44 C57BL/6 Stx2 Yes ND HE: pyk oligo astro nuclei, ND ND i.v.
OBATA AND OBRIG
mouse hrrg sub

44 C57BL/6 Stx2 Yes ND HE: hrrg BS sub severe than ND ND i.v.
mouse + LPS Stx2 alone
45 ICR Stx2 ND ND HE: congestion CL and hippo, ND Anti-Stx2 IHC (+) in i.v.
mouse hrrg CL, neu and glia normal RBC and cap CL,
MB and thalamus,
anti-Stx2 IHC ()
in neu, glia
63 C57BL/6 Stx1 Not Anti-Gb3 became ND ND ND Not
a4galt/ and susceptible negative in cap specied
Stx2 to Stx1
53 C57BL/6 Stx2 Yes Anti-Gb3 (+) in neu EM: glia lamellipodia-like foot Immuno-gold EMl: i.p.
mouse mouse and human process interrupts synapse anti-Gb3 and
Sc, human cap at motor neu of lum Sc anti-Stx2
double positive in

motor neu of lum Sc
51 NIH Stx2 ND ND EM: neu, astro and peri ede, i.v. Behavioral
mouse synaptic disruption, oligo defect motor test +
61 Human NA NA DRGm, Stx1 binding NA NA NA NA
(+) neu and cap
61 Rabbit NA NA DRG, Stx1 binding NA NA NA NA
(+) neu and cap
61 Rat NA NA DRG, Stx1 binding NA NA NA NA
(+) neu
62 Human NA NA DRG, anti-Gb3 NA NA NA NA
and Stx1 binding
(+) neu and cap
62 Rabbit NA NA DRG, anti-Gb3 and NA NA NA NA
Stx1 binding (+)
neu and cap
03/30/15 17:36
62 Rat NA NA DRG, anti-Gb3 and NA NA NA NA

Stx1 binding (+) neu
62 Mouse NA NA DRG, anti-Gb3 and NA NA NA NA
Chap05.proof.3d
Stx1 binding (+) neu
60 C57BL/6 NA NA Anti-Gb3 (+) neu at NA NA NA NA
89
mouse olf, CR ctx, str, hippo,
hypothalamus,
CVOs, CL, MO, Sc
V3 ependyma
a
Animal keys: YL, Yorkshire-Landrace; NZW, New Zealand White; JW, Japanese White; DB, Dutch Belted; SD, Sprague-Dawley.
b
Detailed CNS symptoms are summarized in Table 2. ND, not described; NA, not applicable.
c
Histopathology analysis keys are Gross, gross observation in non-stained tissue; HE, hematoxylin-esosin stain that stains cytoplasm in pink and nucleus blue, light microscopic ndings (LM); PAS, periodic
acid-Schiff stain that detects polysaccharides, glycoproteins and glycolipid, LM; LFB, Luxol fast blue stain that stains myelin in blue, LM; EM, electron microscopic ndings.
d
CNS regions and cell type abbreviations are CR, cerebrum; ctx, cortex; hippo, hippocampus; DG, dentate gyrus; str, striatum and other basal ganglia; CL, cerebellum; MO, medulla oblongata; MB, midbrain;
BS, brain stem is used where midbrain, pons, or medulla oblongata are not specied; Sc, spinal cord; cerv, cervical; tho, thoracic; lum, lumbaris; sub, subarachnoid space. Histopathologic feature
abbreviations are cap, endothelial cells or capillaries; inf, infarction; hrrg, hemorrhage; b, brin deposition; nec, necrosis; swl, swelling; peri, perivascular; myo, myocytes; apop, apoptotic; mono, monocytes;
mye, myelin; deg, degeneration; pyk, pyknotic nuclei; prolif, proliferation/hyperplasia; ede, edema; pyr, pyramidal neuron; int, inltration of blood cells to parenchyma; neu, neuron; Purkinje, Purkinje cells
are large neurons in CL; V3, third ventricle; cc, corpus callosum; amy, amygdala; ips, ipsilateral, injection side of brain; cont, contlateral, opposite of injection side; subV, subventricular region; cp, choroid
plexus; CVO, circumventricular organs.
e
Histopathologic feature abbreviations are cap, endothelial cells or capillaries; inf, infarction; hrrg, hemorrhage; b, brin deposition; nec, necrosis; swl, swelling; peri, perivascular; myo, myocytes; apop,
apoptotic, mono, monocytes; mye, myelin;, degeneration; pyk, pyknotic nuclei; prolif, proliferation/hyperplasia; ede, edema; pyr, pyramidal neuron; int, inltration of blood cells to parenchyma; neu,
neuron; Purkinje, Purkinje cells are large neurons in CL; glia, glial cells such as astrocytes, microglia, and oligodendrocytes); eos, eosinophilic globules, deposits); axo, axon, axoplasm; astro, astrocyte; oligo,
oligodendrocyte; phago, phagocytosis; lyso, lysosome; RBC, red blood cells.

f
IHC, immunohistochemistry, immunodetection of the target in the tissue sections.
g
Injection route abbreviations: i.v., intravenous; i.t. (intrathecal, injection from cysterna magna that makes it possible to inject into cerebrospinal uid (CSF); i.p., intraperitoneal; i.c.v., intracerebroventricular
injection that injects solution directly into CNS parenchyma of cerebral cortex/ventricle.
h
Baboon in this chart is Papio c. cynocephalus, or Papio c. Anubis
i
Sup, E. coli culture supernatant.
j
LacCer, lactosylceramide; adding galactose to LacCer completes Gb3.
k
Immuno-EM-DAB: immunodetection of the target with 3,3-diaminobenzidine deposition by electron microscopy.
l
Immuno-gold EM: immunodetection of the target with 5- to-10-nm gold particle allows precise localization as well as double labeling.
m
DRG, dorsal root ganglion, a peripheral nervous system structure consisting of sensory neurons and other cell types.
CHAPTER 5 Shiga/Vero Toxins in Pathogenesis
89
03/30/15 17:36
90 OBATA AND OBRIG
tumor necrosis factor alpha increase in STEC- mouse expression was restricted to neurons
inoculated rabbit (55) suggests Stx-associated (62, 63). Our group reported that throughout
inflammation in the CNS. the mouse CNS, the only nonneuronal cell
Ca2+ imaging and electrophysiological study type to exhibit anti-Gb3 immunoreactivity was
are useful tools to assess direct physiological the third ventricle ependymal cell (61). Studies
action of Stx in fresh brain slices. Our group have suggested, in the nave state, humans and
showed Stx2-associated neuronal glutamate rabbits express Stx receptor in their vessels
release in mouse brain slice (cerebral cortex) as well as neurons, and rodents appear to
indirectly by recording intracellular Ca2+ in express Gb3 mainly in neurons. However, it
astrocyte (53). Recently, it is shown that Stx2 was shown that Stx receptors in the rat
induces depolarization of neurons in the tha- CNS are induced by Stx administration (57).
lamic area of female rat (58). Among different species, the receptor expres-
sion patterns in different regions of CNS, the
cell types, and the amount expressed may be
Receptor Gb3 Expression in Animal
different, but all models present with common
Central and Peripheral Nervous
CNS impairment such as hind-leg paralysis.
Systems (CNS, PNS)
This may be interpreted as expression of Stx
Shiga toxin receptor localization in the animal receptor in endothelial cells is not necessary
nervous system has been described for dif- for toxin to be able to internalize into the CNS
ferent species. There are three ways to local- parenchyma to have an effect.
ize Shiga toxin receptor. First is to perform In 2006, Okuda et al. (64) reported a 4galt
anti-Stx immunodetection in tissues of STEC- knockout mouse that lacks Gb3 synthase
infected or Stx-injected animals (rabbit [42, (alpha 1,4-galactosyltransferase) and therefore
47, 52], rat [49, 59], mouse [35, 36]). Second is produces no Gb3. In this mouse, originally
to incubate a nave tissue section with Stx Gb3-positive vessels lost their anti-Gb3 im-
followed by anti-Stx immunodetection (pig munoreactivity and became Stx resistant. Gb3
[60]). Third is to recognize Gb3 as an Stx re- synthase probe has been applied for an in situ
ceptor with anti-Gb3 immunodetection in hybridization in the mouse (56) and rat (58)
tissues. Detecting anti-Gb3 immunoreaction in CNS. While metabolic pathway enzymes such
the nave tissue provides a basal expression as Gb3 synthase, a glycosyltransferase, add the
level and cell types that would be influenced terminal galactose to complete Gb3, other gly-
by Stx initially in the course of disease. These cosyltransferases in the pathway are unique in
include neurons in the mouse spinal cord (53) each step of glycolipid synthesis, and there
and other regions of CNS (61). In the Stx-ad- are catabolic pathway enzymes as well (see
ministered tissue, it may or may not indicate Fig. 6). All these enzymes participate in deter-
the spontaneous Stx receptor expression but mining the amount of Gb3 in the cell. Measuring
certainly indicates cell types responsive to these Gb3-associated enzymes may provide
Stx. The cell types that are positive in either more insight into Stx receptor regulation.
of the analyses above often include small
vessel endothelial cells (rabbit [42, 43, 47, 52,
Discussion about How Stx Enters
62, 63], mouse [45, 64]), neurons (rat [49, 57,
CNS of Animals
59], mouse [35, 45, 53, 61]), and glial cells (rat
[49, 57, 59], mouse [45, 61]). Miyatake and Purified Stx peripheral injection (intraperito-
colleagues compared the peripheral nervous neal [i.p.] or intravenous [i.v.]) is able to in-
system (dorsal root ganglion) of different duce CNS impairment similar to that of STEC
species with the same method and found that oral infection, suggesting that there is a direct
human and rabbit expressed Stx receptor in effect of Stx on CNS parenchymal cells. The
endothelial cells and neurons, whereas rat and rat model of intraventricular purified Stx2

FIGURE 6 Metabolic and catabolic pathway enzymes for Gb3 synthesis. A part of Gb3 synthesis pathway is
shown. From lactosylceramide (LacCer) to Gb3, alpha 1, 4-galactosyltransferase (EC 2.4.1.228) adds a galactose
to LacCer to produce Gb3. Likewise, UDP-GalNAc: beta 1,3-galactosaminyltransferase (EC 2.4.1.79) works on Gb3
to make Gb4. In the catabolic pathway, beta-hexosaminidase (EC 3.2.1.52) degrades Gb4 to Gb3, and alpha-
galactosidase (EC 3.2.1.22) makes LacCer from Gb3. doi:10.1128/microbiolspec.EHEC-0005-2013.f6
injection in which purified Stx2 is inoculated parenchyma (rabbit [52]), and magnetic reso-
directly into CNS parenchyma also induces nance imaging showed the third ventricle
similar CNS symptoms such as lethargy, hind- area with a bright signal that is an indication
leg weakness, or paralysis (57). These results of leakiness into the fluid in this area. Taken
suggest that Stx released from STEC inter- together, it is reasonable to think that Stx
nalizes into the blood and then transfers to uses blood-CSF barrier penetration as one of
CNS parenchyma and asserts its toxicity. the routes into CNS parenchyma. On the other
The route and CNS region of Stx perme- hand, Stx injected i.p. was detected in the
abilization are of great interest to explain perivascular area in rat (49), and BBB weak-
which part of the CNS is most likely influ- ening was suggested by the reduction of AQP4
enced by Stx. Stx injected by i.v. has been (rat [50], mouse [56], and by tracer horserad-
detected in cerebrospinal fluid (CSF) (rabbit ish peroxidase (i.v.) detection in parenchyma
[47, 65]). This suggests there is translocation (mouse [35]). These results suggest that Stx
of Stx from blood to CSF. A reduction of AQP1 can also use the BBB crossing route to enter
in choroid plexus in rat with Stx (i.p.) suggests the CNS. An important fact to note is that
that there is weakening of the blood-CSF purified Stx by itself, without any other bac-
barrier in this location that may allow Stx to terial component, can enter CNS and assert its
enter CSF from the blood. The ependymal toxicity regardless of differences in receptor-
cells lining at the third ventricle are a border expressing cell types among different species.
between CSF and CNS parenchyma. Our
group showed in mouse CNS that ependymal
ACKNOWLEDGMENTS
cells at the third ventricle are expressing Gb3
in a nave state (61). The tracer horseradish This work was supported by National Insti-
peroxidase that is injected intrathecally into tutes of Health grants 5U01AI075778-05 and
CSF crossed and entered ependymal cells and 1R56AI090144-01A1 to F.O.

92 OBATA AND OBRIG
CITATION 14. Fernandez GC, Rubel C, Dran G, Gomez S,

Isturiz MA, Palermo MS. 2000. Shiga toxin-2
Obata F, Obrig T. 2014. Role of Shiga/vero induces neutrophilia and neutrophil activation in
toxins in pathogenesis. Microbiol Spectrum a murine model of hemolytic uremic syndrome.
Clin Immunol 95:227234.
2(3):EHEC-0005-2013.
15. Roche JK, Keepers TR, Gross LK, Seaner RM,
Obrig TG. 2007. CXCL1/KC and CXCL2/MIP-2
are critical effectors and potential targets for
REFERENCES
therapy of Escherichia coli O157:H7-associated
1. Ohmura M, Yamamoto M, Tomiyama-Miyaji C, renal inflammation. Am J Pathol 170:526537.
Yuki Y, Takeda Y, Kiyono H. 2005. Nontoxic 16. Foster GH, Armstrong CS, Sakiri R, Tesh VL.
Shiga toxin derivatives from Escherichia coli 2000. Shiga toxin-induced tumor necrosis factor
possess adjuvant activity for the augmentation of alpha expression: requirement for toxin enzy-
antigen-specific immune responses via dendritic matic activity and monocyte protein kinase C and
cell activation. Infect Immun 73:40884097. protein tyrosine kinases. Infect Immun 68:5183
2. Tesh VL, Ramegowda B, Samuel JE. 1994. 5189.
Purified Shiga-like toxins induce expression of 17. Eaton KA, Friedman DI, Francis GJ, Tyler JS,
proinflammatory cytokines from murine perito- Young VB, Haeger J, Abu-Ali G, Whittam TS.
neal macrophages. Infect Immun 62:50855094. 2008. Pathogenesis of renal disease due to en-
3. Obrig TG. 2010. Escherichia coli Shiga toxin terohemorrhagic Escherichia coli in germ-free
mechanisms of action in renal disease. Toxins mice. Infect Immun 76:30543063.
(Basel) 2:27692794. 18. Stahl AL, Sartz L, Nelsson A, Bekassy ZD,
4. Obrig TG, Karpman D. 2012. Shiga toxin patho- Karpman D. 2009. Shiga toxin and lipopoly-
genesis: kidney complications and renal failure. saccharide induce platelet-leukocyte aggregates
Curr Top Microbiol Immunol 357:105136. and tissue factor release, a thrombotic mechanism
5. Karpman D, Sartz L, Johnson S. 2010. Patho- in hemolytic uremic syndrome. PLoS One 4:
physiology of typical hemolytic uremic syndrome. e6990.
Semin Thromb Hemost 36:575585. 19. Keepers TR, Psotka MA, Gross LK, Obrig TG.
6. Habib R, Mathieu H, Royer P. 1967. [Hemolytic- 2006. A murine model of HUS: Shiga toxin with
uremic syndrome of infancy: 27 clinical and ana- lipopolysaccharide mimics the renal damage and
tomic observations]. Nephron 4:139172 (In French). physiologic response of human disease. J Am Soc
7. Obrig TG, Louise CB, Lingwood CA, Boyd B, Nephrol 17:34003414.
Barley-Maloney L, Daniel TO. 1993. Endothelial 20. Keepers TR, Gross LK, Obrig TG. 2007. Mono-
heterogeneity in Shiga toxin receptors and re- cyte chemoattractant protein 1, macrophage in-
sponses. J Biol Chem 268:1548415488. flammatory protein 1 alpha, and RANTES recruit
8. Psotka MA, Obata F, Kolling GL, Gross LK, macrophages to the kidney in a mouse model of
Saleem MA, Satchell SC, Mathieson PW, Obrig hemolytic-uremic syndrome. Infect Immun 75:
TG. 2009. Shiga toxin 2 targets the murine renal 12291236.
collecting duct epithelium. Infect Immun 77:959 21. Hasko G, Linden J, Cronstein B, Pacher P. 2008.
969. Adenosine receptors: therapeutic aspects for in-
9. Simon M, Cleary TG, Hernandez JD, Abboud flammatory and immune diseases. Nat Rev Drug
HE. 1998. Shiga toxin 1 elicits diverse biologic re- Discov 7:759770.
sponses in mesangial cells. Kidney Int 54:11171127. 22. Walters MD, Matthei IU, Kay R, Dillon MJ,
10. Shibolet O, Shina A, Rosen S, Cleary TG, Brezis Barratt TM. 1989. The polymorphonuclear
M, Ashkenazi S. 1997. Shiga toxin induces med- leucocyte count in childhood haemolytic uraemic
ullary tubular injury in isolated perfused rat kid- syndrome. Pediatr Nephrol 3:130134.
neys. FEMS Immunol Med Microbiol 18:5560. 23. Inward CD, Howie AJ, Fitzpatrick MM, Rafaat
11. Hughes AK, Stricklett PK, Kohan DE. 1998. F, Milford DV, Taylor CM. 1997. Renal histo-
Cytotoxic effect of Shiga toxin-1 on human prox- pathology in fatal cases of diarrhoea-associated
imal tubule cells. Kidney Int 54:426437. haemolytic uraemic syndrome. Pediatr Nephrol
12. Ghosh SA, Polanowska-Grabowska RK, Fujii 11:556559.
J, Obrig T, Gear AR. 2004. Shiga toxin binds to 24. Griener TP, Strecker JG, Humphries RM,
activated platelets. J Thromb Haemost 2:499506. Mulvey GL, Fuentealba C, Hancock RE,
13. Karpman D, Manea M, Vaziri-Sani F, Stahl AL, Armstrong GD. 2011. Lipopolysaccharide renders
Kristoffersson AC. 2006. Platelet activation in transgenic mice expressing human serum amyloid
hemolytic uremic syndrome. Semin Thromb P component sensitive to Shiga toxin 2. PLoS One
Hemost 32:128145. 6:e21457.

25. Kawachi H, Han GD, Miyauchi N, Hashimoto T, 37. Isogai E, Isogai H, Kimura K, Hayashi S, Kubota
Suzuki K, Shimizu F. 2009. Therapeutic targets T, Fujii N, Takeshi K. 1998. Role of tumor ne-
in the podocyte: findings in anti-slit diaphragm crosis factor alpha in gnotobiotic mice infected
antibody-induced nephropathy. J Nephrol 22: with an Escherichia coli O157:H7 strain. Infect
450456. Immun 66:197202.
26. Das L, Brunner HI. 2009. Biomarkers for renal 38. Taguchi H, Takahashi M, Yamaguchi H,
disease in childhood. Curr Rheumatol Rep 11:218 Osaki T, Komatsu A, Fujioka Y, Kamiya S. 2002.
225. Experimental infection of germ-free mice with
27. Trachtman H, Christen E, Cnaan A, Patrick J, hyper-toxigenic enterohaemorrhagic Escherichia
Mai V, Mishra J, Jain A, Bullington N, Devarajan coli O157:H7, strain 6. J Med Microbiol 51:336
P. 2006. Urinary neutrophil gelatinase-associated 343.
lipocalcin in D+HUS: a novel marker of renal in- 39. MacLeod DL, Gyles CL, Wilcock BP. 1991. Re-
jury. Pediatr Nephrol 21:989994. production of edema disease of swine with puri-
28. Mohawk KL, OBrien AD. 2011. Mouse models of fied Shiga-like toxin-II variant. Vet Pathol 28:66
Escherichia coli O157:H7 infection and Shiga toxin 73.
injection. J Biomed Biotechnol 2011:258185. 40. Gannon VP, Gyles CL, Wilcock BP. 1989. Effects
29. Karpman D, Connell H, Svensson M, Scheutz F, of Escherichia coli Shiga-like toxins (verotoxins)
Alm P, Svanborg C. 1997. The role of lipopoly- in pigs. Can J Vet Re. 53:306312.
saccharide and Shiga-like toxin in a mouse model 41. Zoja C, Corna D, Farina C, Sacchi G, Lingwood
of Escherichia coli O157:H7 infection. J Infect Dis C, Doyle MP, Padhye VV, Abbate M, Remuzzi
175:611620. G. 1992. Verotoxin glycolipid receptors determine
30. Wadolkowski EA, Sung LM, Burris JA, Samuel the localization of microangiopathic process in
JE, OBrien AD. 1990. Acute renal tubular ne- rabbits given verotoxin-1. J Lab Clin Med 120:
crosis and death of mice orally infected with 229238.
Escherichia coli strains that produce Shiga-like 42. Richardson SE, Rotman TA, Jay V, Smith CR,
toxin type II. Infect Immun 58:39593965. Becker LE, Petric M, Olivieri NF, Karmali MA.
31. Tzipori S, Chow CW, Powell HR. 1988. Cerebral 1992. Experimental verocytotoxemia in rabbits.
infection with Escherichia coli O157:H7 in humans Infect Immun 60:41544167.
and gnotobiotic piglets. J Clin Pathol 41:1099 43. Takahashi K, Funata N, Ikuta F, Sato S. 2008.
1103. Neuronal apoptosis and inflammatory responses
32. Dean-Nystrom EA, Pohlenz JF, Moon HW, in the central nervous system of a rabbit treated
OBrien AD. 2000. Escherichia coli O157:H7 with Shiga toxin-2. J Neuroinflammation 5:11.
causes more-severe systemic disease in suckling 44. Sugatani J, Igarashi T, Munakata M, Komiyama
piglets than in colostrum-deprived neonatal Y, Takahashi H, Komiyama N, Maeda T,
piglets. Infect Immun 68:23562358. Takeda T, Miwa M. 2000. Activation of coagu-
33. Matise I, Sirinarumitr T, Bosworth BT, Moon lation in C57BL/6 mice given verotoxin 2 (VT2)
HW. 2000. Vascular ultrastructure and DNA and the effect of co-administration of LPS with
fragmentation in swine infected with Shiga VT2. Thromb Res 100:6172.
toxin-producing Escherichia coli. Vet Pathol 37: 45. Nishikawa K, Matsuoka K, Kita E, Okabe N,
318327. Mizuguchi M, Hino K, Miyazawa S, Yamasaki
34. Shringi S, Garcia A, Lahmers KK, Potter KA, C, Aoki J, Takashima S, Yamakawa Y, Nishijima
Muthupalani S, Swennes AG, Hovde CJ, Call M, Terunuma D, Kuzuhara H, Natori Y. 2002.
DR, Fox JG, Besser TE. 2012. Differential viru- A therapeutic agent with oriented carbohydrates
lence of clinical and bovine-biased enterohemor- for treatment of infections by Shiga toxin-
rhagic Escherichia coli O157:H7 genotypes in producing Escherichia coli O157:H7. Proc Natl
piglet and Dutch belted rabbit models. Infect Acad Sci USA 99:76697674.
Immun 80:369380. 46. Taylor CM, Williams JM, Lote CJ, Howie AJ,
35. Fujii J, Kita T, Yoshida S, Takeda T, Kobayashi Thewles A, Wood JA, Milford DV, Raafat F,
H, Tanaka N, Ohsato K, Mizuguchi Y. 1994. Chant I, Rose PE. 1999. A laboratory model of
Direct evidence of neuron impairment by oral toxin-induced hemolytic uremic syndrome. Kid-
infection with verotoxin-producing Escherichia ney Int 55:13671374.
coli O157:H- in mitomycin-treated mice. Infect 47. Mizuguchi M, Tanaka S, Fujii I, Tanizawa H,
Immun 62:34473453. Suzuki Y, Igarashi T, Yamanaka T, Takeda T,
36. Kurioka T, Yunou Y, Kita E. 1998. Enhancement Miwa M. 1996. Neuronal and vascular pathology
of susceptibility to Shiga toxin-producing Esche- produced by verocytotoxin 2 in the rabbit central
richia coli O157:H7 by protein calorie malnutrition nervous system. Acta Neuropathol (Berl) 91:254
in mice. Infect Immun 66:17261734. 262.

94 OBATA AND OBRIG
48. Garcia A, Marini RP, Catalfamo JL, Knox KA, P, Coulon P, Herrmann AM, Storck W, Kohmann
Schauer DB, Rogers AB, Fox JG. 2008. Intra- D, Muthing J, Pavenstadt H, Kuhlmann T, Karch
venous Shiga toxin 2 promotes enteritis and renal H, Peters G, Budde T, Wiendl H, Pape HC. 2013.
injury characterized by polymorphonuclear leu- Thalamic involvement in patients with neurologic
kocyte infiltration and thrombosis in Dutch Belted impairment due to Shiga toxin 2. Ann Neurol 73:
rabbits. Microbes Infect 10:650656. 419429.
49. Goldstein J, Loidl CF, Creydt VP, Boccoli J, 59. Boccoli J, Loidl CF, Lopez-Costa JJ, Creydt VP,
Ibarra C. 2007. Intracerebroventricular admin- Ibarra C, Goldstein J. 2008. Intracerebroventric-
istration of Shiga toxin type 2 induces striatal ular administration of Shiga toxin type 2 altered
neuronal death and glial alterations: an ultra- the expression levels of neuronal nitric oxide
structural study. Brain Res 1161:106115. synthase and glial fibrillary acidic protein in rat
50. Lucero MS, Mirarchi F, Goldstein J, Silberstein brains. Brain Res 1230:320333.
C. 2012. Intraperitoneal administration of Shiga 60. Winter KR, Stoffregen WC, Dean-Nystrom EA.
toxin 2 induced neuronal alterations and reduced 2004. Shiga toxin binding to isolated porcine
the expression levels of aquaporin 1 and aqua- tissues and peripheral blood leukocytes. Infect
porin 4 in rat brain. Microb Pathog 53:8794. Immun 72:66806684.
51. Tironi-Farinati C, Geoghegan PA, Cangelosi A, 61. Obata F, Obrig T. 2010. Distribution of Gb(3)
Pinto A, Loidl CF, Goldstein J. 2013. A transla- immunoreactivity in the mouse central nervous
tional murine model of sub-lethal intoxication system. Toxins (Basel) 2:19972006.
with Shiga toxin 2 reveals novel ultrastructural 62. Ren J, Utsunomiya I, Taguchi K, Ariga T, Tai T,
findings in the brain striatum. PLoS One 8:e55812. Ihara Y, Miyatake T. 1999. Localization of
52. Fujii J, Kinoshita Y, Kita T, Higure A, Takeda verotoxin receptors in nervous system. Brain Res
T, Tanaka N, Yoshida S. 1996. Magnetic reso- 825:183188.
nance imaging and histopathological study of 63. Utsunomiya I, Ren J, Taguchi K, Ariga T, Tai T,
brain lesions in rabbits given intravenous vero- Ihara Y, Miyatake T. 2001. Immunohistochemi-
toxin 2. Infect Immun 64:50535060. cal detection of verotoxin receptors in nervous
53. Obata F, Tohyama K, Bonev AD, Kolling GL, system. Brain Res Brain Res Protoc 8:99103.
Keepers TR, Gross LK, Nelson MT, Sato S, 64. Okuda T, Tokuda N, Numata S, Ito M, Ohta M,
Obrig TG. 2008. Shiga toxin 2 affects the central Kawamura K, Wiels J, Urano T, Tajima O,
nervous system through receptor globotriaosyl- Furukawa K. 2006. Targeted disruption of Gb3/
ceramide localized to neurons. J Infect Dis 198: CD77 synthase gene resulted in the complete
13981406. deletion of globo-series glycosphingolipids and
54. Mizuguchi M, Sugatani J, Maeda T, Momoi T, loss of sensitivity to verotoxins. J Biol Chem 281:
Arima K, Takashima S, Takeda T, Miwa M. 1023010235.
2001. Cerebrovascular damage in young rabbits 65. Fujii J, Kinoshita Y, Yamada Y, Yutsudo T, Kita
after intravenous administration of Shiga toxin 2. T, Takeda T, Yoshida S. 1998. Neurotoxicity of
Acta Neuropathol (Berl) 102:306312. intrathecal Shiga toxin 2 and protection by in-
55. Kita E, Yunou Y, Kurioka T, Harada H, trathecal injection of anti-Shiga toxin 2 antiserum
Yoshikawa S, Mikasa K, Higashi N. 2000. in rabbits. Microb Pathog 25:139146.
Pathogenic mechanism of mouse brain damage 66. Taylor CM, Williams JM, Lote CJ, Howie AJ,
caused by oral infection with Shiga toxin- Thewles A, Wood JA, Milford DV, Raafat F,
producing Escherichia coli O157:H7. Infect Immun Chant I, Rose PE. 1999. A laboratory model of
68:12071214. toxin-induced hemolytic uremic syndrome. Kid-
56. Amran MY, Fujii J, Suzuki SO, Kolling GL, ney Int 55:13671374.
Villanueva SY, Kainuma M, Kobayashi H, 67. Siegler RL, Pysher TJ, Lou R, Tesh VL, Taylor
Kameyama H, Yoshida S. 2013. Investigation of FB Jr. 2001. Response to Shiga toxin-1, with and
encephalopathy caused by Shiga toxin 2c-pro- without lipopolysaccharide, in a primate model of
ducing Escherichia coli infection in mice. PLoS hemolytic uremic syndrome. Am J Nephrol
One 8:e58959. 21:420425.
57. Tironi-Farinati C, Loidl CF, Boccoli J, Parma Y, 68. Yamada Y, Fujii J, Murasato Y, Nakamura
Fernandez-Miyakawa ME, Goldstein J. 2010. T, Hayashida Y, Kinoshita Y, Yutsudo T,
Intracerebroventricular Shiga toxin 2 increases Matsumoto T, Yoshida S. 1999. Brainstem mech-
the expression of its receptor globotriaosylcera- anisms of autonomic dysfunction in encephalop-
mide and causes dendritic abnormalities. J Neuro- athy-associated Shiga toxin 2 intoxication. Ann
immunol 222:4861. Neurol 45:716723.
58. Meuth SG, Gobel K, Kanyshkova T, Ehling P, 69. Fujii J, Kinoshita Y, Yutsudo T, Taniguchi H,
Ritter MA, Schwindt W, Bielaszewska M, Lebiedz Obrig T, Yoshida SI. 2001. Toxicity of Shiga

toxin 1 in the central nervous system of rabbits. 71. Tzipori S, Gunzer F, Donnenberg MS, de
Infect Immun 69:65456548. Montigny L, Kaper JB, Donohue-Rolfe A. 1995.
70. Fujii J, Kinoshita Y, Matsukawa A, Villanueva The role of the eaeA gene in diarrhea and neu-
SY, Yutsudo T, Yoshida S. 2009. Successful ste- rological complications in a gnotobiotic piglet
roid pulse therapy for brain lesion caused by Shiga model of enterohemorrhagic Escherichia coli in-
toxin 2 in rabbits. Microb Pathog 46:179184. fection. Infect Immun 63:36213627.

The Locus of Enterocyte Effacement
and Associated Virulence Factors of
MARK P. STEVENS1 and GAD M. FRANKEL2

6
INTRODUCTION
Enterohemorrhagic Escherichia coli (EHEC) was first recognized as a cause

of human disease in 1983 and is associated with diarrhea and hemorrhagic
colitis, which may be complicated by life-threatening renal and neurological
sequelae (reviewed in reference 242). EHEC strains are defined by their ability
to produce one or more Shiga toxins (Stx), which mediate the systemic com-
plications of EHEC infections (reviewed in reference 243), and to induce
attaching and effacing (A/E) lesions on intestinal epithelia. The ability of EHEC
to induce such lesions is shared by enteropathogenic E. coli (EPEC), Escherichia
albertii (formerly classified as eae-positive Hafnia alvei), and the murine path-
ogen Citrobacter rodentium. The A/E histopathology was first described in
gnotobiotic piglets challenged with a strain of EHEC serotype O157:H7 (1) but
has subsequently been observed in ruminant reservoirs and diverse animal
models (reviewed in reference 244).
A/E lesions are characterized by intimate bacterial adherence to the api-
cal surface of enterocytes, sometimes on raised pedestals (pseudopodia), and
destruction of nearby microvilli (Fig. 1A and B). Though reports of such lesions
1
The Roslin Institute and Royal (Dick) School of Veterinary Studies, University of Edinburgh, Midlothian, EH25 9RG, United
Kingdom; 2MRC Centre for Molecular Bacteriology and Infection, Department of Life Sciences, Imperial College London,
London, SW7 2AZ, United Kingdom.
97

98 STEVENS AND FRANKEL
during human infection are lacking, Knutton Reasoning that the electron-dense pedestals
et al. described that EPEC strains of many may contain polymerized F-actin, the authors
serogroups were able to form A/E lesions on subsequently reported the use of fluorescein-
human duodenal biopsies cultured ex vivo (2). labeled phallotoxin to stain F-actin at sites of
FIGURE 1 (A) Transmission electron micrograph (TEM) showing A/E lesions induced by EHEC O111:H strain
E45035N in the spiral colon of a neonatal calf (note raised electron-dense pedestals and microvillus effacement
relative to proximal uninfected enterocyte). From reference 233; scale bar = 1 m. (B) TEM of A/E lesions in-
duced by EHEC O157:H7 strain 85-170 12 h after inoculation of a bovine ligated ileal loop (note intimate
adherence but relative absence of elongated pedestals). From reference 84; scale bar = 5 m. (C) Fluorescence
micrograph showing nucleation of F-actin under EHEC O103:H3 strain PMK5 adhering to a HeLa cell (green,
F-actin detected with Oregon green514-phalloidin; red, bacteria stained with rabbit anti-O103 typing serum
detected with anti-rabbit immunoglobulin-Alexa568). From reference 240; scale bar = 5 m.

CHAPTER 6 LEE and Associated Virulence Factors of EHEC 99
adherence, forming the basis of a fluorescent- required for secretion of several EPEC O127:
actin staining (FAS) test for A/E lesion for- H6 proteins required for A/E lesion forma-
mation by fluorescence microscopy (3) (Fig. tion, designated Esps (E. coli secreted pro-
1C). Many other host cell proteins are re- teins) (6). Secretion of Esps was also detected
cruited to A/E lesions, sometimes in a host-, in EHEC O157:H7 and EHEC O26:H11 (8, 9)
pathotype-, serotype-, and time-dependent and found to be escN-dependent (9), sup-
manner, and the molecular mechanisms un- porting the existence of a conserved locus for
derlying lesion formation have been the sub- secretion of proteins involved in A/E lesion
ject of intense study. This article focuses on formation.
the factors required for A/E lesion formation, The LEE of EPEC O127:H6 strain E2348/
their mode of action, and role in persistence, 69 contains 41 open reading frames that are
pathogenesis, and protection during EHEC mostly organized in five polycistronic operons
infection. (7). Soon after the sequencing of the EPEC
O127:H6 strain E2348/69 LEE, the sequence
of the homologous region of the genome of
THE LOCUS OF ENTEROCYTE EFFACEMENT the prototype EHEC O157:H7 strain EDL933
was reported (10), followed by the LEE of
The locus of enterocyte effacement (LEE) is C. rodentium (11) and rabbit EPEC strains
required for A/E lesion formation and is (12, 13). Many other LEE sequences have since
the most intensively studied of all virulence emerged with the advent of high-throughput
factors of attaching and effacing E. coli. The genome sequencing. These typically exhibit
role of LEE-encoded genes in adherence significantly lower %GC content than the flank-
to epithelial cells and nucleation of F-actin ing DNA (ca. 38% compared to the ca. 50%
was first described after screening random genomic mean), are inserted at tRNA loci,
TnphoA mutants of EPEC O127:H6 for mu- and contain remnants of mobile genetic ele-
tants deficient in these processes. This iden- ments, indicating that they are likely to have
tified the E. coli attaching and effacing (eae) been acquired by horizontal transfer. The
gene encoding intimin (4) and other genes cloned LEE of EPEC O127:H6 strain E2348/
required for adherence and actin nucleation 69 is necessary and sufficient to confer the
in a single ca. 35-kb locus that is conserved ability to form pedestals upon a laboratory-
among A/E pathogens, including strains of adapted E. coli K-12 strain (14); however, the
EHEC serotypes O157:H7 and O26:H11 (5). cloned EHEC O157:H7 LEE cannot (15), con-
Sequencing of a limited region of the LEE of sistent with the requirement for non-LEE-
the prototype EPEC O127:H6 strain E2348/69 encoded proteins in pedestal formation by
identified four genes predicted to encode com- EHEC O157:H7, as discussed below.
ponents of a Type III secretion system (T3SS), The genetic organization of the EHEC
based on homology to Yersinia Lcr/Ysc and O157:H7 strain EDL933 LEE is depicted in
Shigella Mxi/Spa proteins (6). These genes Fig. 2. The LEE region in this strain is inserted
were initially designated sepAD (secretion of at the equivalent of 82 min on the E. coli K-12
EPEC proteins) (6), but many were renamed chromosome, proximal to the selC seleno-
esc to conform to the nomenclature of ho- cysteine tRNA locus, and shares the 41 genes
mologous Yersinia T3SS genes upon complete found in EPEC O127:H6 strain E2348/69 in
sequencing of the LEE of E2348/69 (7). Type the same order, but is larger owing to the
III secretion is one of multiple pathways for presence of a cryptic P4 family prophage at
export of proteins in bacteria and, consistent one end (10). Predicted structural components
with a role in this process, the product of the of the T3SS exhibit remarkable conservation
sepB (escN) gene, which encodes a predicted among A/E pathogens; however, greater levels
inner membrane ATPase, was found to be of amino acid sequence divergence exist

FIGURE 2 Genetic organization of LEE of E. coli O157:H7. Open reading frames are represented by thick arrows,
and putative polycistronic operons are designated by thin arrows. Clear arrows represent open reading frames
of unknown function and are designated orf or rorf, depending on the direction of transcription relative to eae.
among components that are predicted interact be present in the genome but not physically
with host cells. Indeed, comparative analyses linked to the LEE, owing to activity of bacte-
have suggested that the LEE comprises units riophages in the capture and transfer of genes
under distinct selection pressures (16). Anal- encoding Type III secreted proteins (18, 25).
ysis of the insertion sites and flanking regions As more genomes are sequenced, it is evident
of the LEE in serogroup O26, O103, and O111 that the nature of lateral transfer of the LEE,
EHEC strains and atypical EPEC has indi- and genes associated with the function of the
cated that while insertions tend to be re- T3SS, is highly variable, and further studies
stricted to the selC, pheU, and pheV tRNA loci, are required to define the consequences of
there is marked variation in the flanking re- such.
gions (17, 18; reviewed in reference 19). For Regulation of the LEE is highly complex,
example, substrates for the LEE-encoded with inputs from numerous global regulators,
T3SS may be encoded in the flanking regions, bacteriophage-encoded regulators, and LEE-
such as Ibe 5 of the LEE in atypical EPEC encoded regulators. Control occurs at tran-
(20), and a conserved cassette encoding scriptional, translational, posttranslational, and
NleE, NleB, and EspL 3 of the LEE4 operon in bacterial population levels (reviewed in refer-
EHEC O26, O103, and O111 (18). In some ence 245). Of particular interest, LEE expres-
strains the lifA/efa-1 gene or a truncated var- sion and the efficiency of pedestal formation
iant thereof may be located 3 of LEE4 as part are enhanced by passage in the mammalian
of a mosaic island comprising the LEE and host (26), and evidence exists that LEE ex-
O-island 122 (18, 21). LifA/Efa1 and the trun- pression is sensitive to host stress-related cat-
cated LifA/Efa1 (Z4332) have recently been echolamine hormones that are detected by
reported to be secreted by the LEE-encoded bacterial adrenergic sensor kinases (reviewed
T3SS (22) and (together with other O-island in reference 27). Such studies reinforce the
122-encoded genes) have been associated challenge of understanding LEE regulation
with isolates that cause epidemic and seri- and functions in cell-based systems devoid of
ous disease (23). In EHEC O103:H2 strain cues from the host. Moreover, it is evident that
RW1374, the LEE, lifA/efa-1, and other viru- the repertoire of regulators is strain-dependent
lence-associated genes form a pathogenicity- and that studies on LEE regulation in single
related island spanning ca. 111 kb at pheV (24), strains are therefore to be interpreted with
whereas in other cases effector genes may caution. For example, lysogeny with Shiga

toxin 2-encoding bacteriophages was recently

reported to repress Type III secretion in
phage-type (PT) 21/28 EHEC O157:H7 strains
that are commonly isolated from humans in the
United Kingdom and to account for a distinct
level of T3SS activity relative to Stx2-minus
PT32 strains (28).
The LEE-Encoded Type III

Secretion System
T3SSs are key virulence factors of Gram-
negative enteric pathogens and the four major
genera of plant pathogenic bacteria and serve
to inject bacterial proteins directly into host FIGURE 3 Schematic representation of the Type III
cells (reviewed in reference 29). A needle com- secretion apparatus showing the predicted spatial
organization of LEE-encoded proteins. Adapted from
plex spans both the inner and outer bacterial reference 30.
membranes, and a subset of Type III secreted doi:10.1128/microbiolspec.EHEC-0007-2013.f3
proteins forms a translocon that interacts
with the eukaryotic cell membrane and me-
diates the delivery of secreted effector pro- escE, orf4, escL, rorf3, escI, orf12, escA, sepQ,
teins into target cells. Effectors of A/E E. coli and escG) with mutations in a further four
then modulate cellular processes to the ben- genes (orf3, rorf6, orf16, and sepL) impairing
efit of the pathogen. Predicted functions have secretion of translocators preferentially (34).
been assigned to LEE-encoded genes on the Real-time imaging has subsequently indicated
basis of homology with components of T3SSs a distinct order in the efficiency of secretion of
in other bacteria (30); a schematic represen- specific effectors in a manner dependent on
tation of the organization of the apparatus intra-bacterial effector concentration, effector-
is shown in Fig. 3. Predicted protein-protein chaperone interactions, and the efficiency of
interactions within the T3SS apparatus have bacterial attachment to target cells (35). Al-
been confirmed experimentally in some cases, though these studies indicated that EspZ
for example, by yeast-2-hybrid analysis (31) is injected into cells early after infection, re-
and analysis of purified needle complexes cent work has indicated that EspZ arrests
(32). Transmission electron microscopy of translocation of effectors from within the
the purified EPEC O127:H6 Type III secretion host cell through an ill-defined translocation
needle complex has indicated that it com- stop activity (36). Indeed, ectopic expression
prises cylindrical inner and outer rings similar of EspZ inside eukaryotic cells renders them
to the flagellar basal body (33), consistent with refractory to actin pedestal formation and
the shared evolution of such systems. prior infection of cultured cells with one A/E
Systematic mutagenesis of the LEE in pathogen prevents superinfection by another
C. rodentium has been used to assess the role (36). Mutants lacking espZ cause elevated
of all 41 genes in regulation of LEE expression, cytotoxicity, probably owing to uncontrolled
the level and hierarchy of Type III secretion, injection of effectors (36, 37). The importance
the ability to form pedestals via the FAS test, of sepL, escD, sepQ, escV, rorf8, escC, escR, and
and virulence in mice (34). Nineteen genes orf4 in EspA secretion by EHEC O157:H7 was
essential for secretion of both translocator and confirmed by analysis of transposon mutants
effector proteins were identified (escR, escS, that exhibit reduced adherence to Caco-2 cells
escT, escU, escC, escJ, escV, escN, escD, escF, in vitro (38).

Mutations affecting components of the length up to 600 nm and are ca. 12 nm in di-
needle complex and translocon were found ameter (33, 50). Such filaments form a physi-
to impair intestinal colonization of calves by cal bridge between the bacteria and host cells
signature-tagged transposon mutagenesis of that is required for the translocation of EspB
EHEC O157:H7 (39) and EHEC O26:H (40). and Tir into host cells (48, 49, 51, 52) (Fig. 4A).
A retrospective transposon-directed insertion- This led to the hypothesis that EspA filaments
site sequencing analysis of the library of EHEC comprise a hollow channel through which
O157:H7 strain EDL933 mutants screened in effector proteins are injected into host cells, a
calves by Dziva et al. identified 51 attenuating notion supported by resolution of the three-
mutations in the LEE, many of which affected dimensional structure, which showed that
predicted needle complex components (41), EPEC O127:H6 EspA subunits polymerized
extending earlier observations on the role of in a helical tube of 120 diameter with 5.6
the LEE genes without further use of cattle. subunits per turn enclosing a central channel
Screening of mutants in complex pools can of 25 diameter (53). Immunogold electron
exaggerate fitness costs; however, EHEC mu- microscopy revealed that Tir is secreted from
tants with defects affecting the escN-encoded the tips of EspA filaments (54) (Fig. 4B and C),
inner membrane ATPase were also found to providing direct evidence that EspA filaments
be highly attenuated when tested in isolation are hollow conduits through which effec-
in infant rabbits (42) and calves (39, 40), con- tors are secreted. Secretion and intracellular
firming a key role for the LEE-encoded T3SS in stability of EspA require the LEE1-encoded
EHEC persistence and pathogenesis. chaperone CesAB (55), and the filaments
Inhibitors of Type III secretion in A/E are elongated by addition of subunits at the
pathogens have been identified (4345). Many growing tip (54). EspA binds directly to the
of these belong to a family of salicylidene needle complex protein EscF, which is re-
acylhydrazides and appear to act in part by quired for EspA filament assembly and effec-
inhibition of transcription of LEE genes in tor translocation (33, 56).
EHEC O157:H7 (45), rather than by inter- It remains unclear precisely how EspA
acting with basal components of the T3SS ap- filaments are connected to the host cell sur-
paratus and interfering with needle complex face and the translocation pore is created. The
assembly as reported for Shigella sp. (46). LEE4-encoded Type III secreted EspB and
Different salicylidene acylhydrazides produce EspD proteins are believed to mediate for-
distinct patterns of LEE repression (45), and mation of the translocation pore on the basis
the targets of such drugs have so far been of homology to the Yersinia YopB and YopD,
mapped to genes outside the LEE (44, 47), their presence in the host cell plasma mem-
providing insights into how LEE expression brane, and ability to form pores in erythrocyte
and T3SS function are controlled. The poten- membranes (48, 5761). EspD is required for
tial prophylactic and therapeutic applications assembly of EspA filaments (48, 58), interacts
of such inhibitors have received relatively lit- directly with EspA and EspB (62), and re-
tle attention. quires the chaperones CesD and CesD2 for
secretion (63). EspB interacts with EspA and
is required for the translocation of effectors,
Translocon Components
including Tir (64). However, EspA filaments
A filamentous extension of the T3SS needle can bind to host cell membranes in the
complex is transiently expressed on the sur- absence of EspB, indicating that they may
face of A/E pathogens in the early stages of interact directly with cellular components
lesion formation and comprises mostly the (64). Indeed, studies using single, double, or
LEE4-encoded protein EspA (48, 49) (Fig. 4). triple mutants of EPEC O127:H6 lacking EspA
EspA filaments in EPEC O127:H6 can vary in filaments, intimin, and bundle-forming pili

FIGURE 4 (A) Scanning electron micrograph showing EspA laments (arrow) of EHEC O26:H11 strain H19
attaching to the surface of an erythrocyte. From reference 241. (B) Transmission electron micrograph of an
EspA lament of wild-type EPEC O127:H6 strain E2348/69 showing Tir issuing from the tip. EspA laments
were immunolabeled with anti-EspA conjugated to 5-nm diameter gold particles, and Tir was detected with
anti-Tir conjugated to 10-nm diameter gold particles. (C) The specicity of Tir staining was conrmed using the
same gold-labeled antibodies but an isogenic tir mutant. Panels B and C from reference 54.
indicate that EspA filaments play a role in sequences exist between serogroup O157 and
intimin-independent adherence to HEp-2 and non-O157 EHEC strains. Indeed, antibodies
Caco-2 cells, as an espA bfpA eae triple mutant induced by immunization of cattle with Type
was less adherent than a bfpA eae double III secreted proteins from EHEC O26:H11,
mutant (65). EspA filaments appear shorter in O103:H2, or O111:NM failed to cross-react
EHEC O157:H7, and their expression is het- with EspA from EHEC O157:H7 (73), leading
erogeneous at the bacterial population level the authors to conclude that cross-serogroup
(66) but is coordinated with expression of protection due to EspA may be limited.
LEE5 in single cells (67). EspA filaments are Although EspB is acknowledged to be a
typically absent from bacteria at the time of key part of the T3SS translocon, it can also
intimate adherence to host cells on pedestals, be detected in the cytoplasm of infected cells
and the basis of loss or disassembly of the (74), where it modulates cellular processes.
translocon after initial attachment remains ill- For example, EspB binds to the host proteins
defined. a-catenin, a1-antitrypsin, and myosin that
It is clear that LEE4-encoded translocon regulate the actin network leading to alter-
components play pivotal roles in intestinal ations in cell morphology (reviewed in refer-
colonization by EHEC, as random and refined ence 75). The myosin-binding domain of EspB
mutants are highly attenuated in cattle (3941, inhibits the interaction of myosins with actin
68) and in mice (69). Moreover, EspA is con- leading to microvillus effacement, and mu-
sidered an important constituent of vaccines tants lacking this domain lack the ability
for control of EHEC in cattle (70, 71). In this to efface enterocytes, suppress phagocytosis,
regard it is noteworthy that EspA filaments or colonize the intestines of mice (76). EspB
of EPEC and EHEC are antigenically dis- was also implicated in the ability of EHEC to
tinct (72) and that variations in EspA primary suppress NF-kB activation and synthesis of

proinflammatory cytokines (77); however, the facilitate bacterial invasion of eukaryotic cells.
extent to which this was dependent on EspB Frankel et al. demonstrated that the carboxyl-
per se, or its role in delivery of other effectors, terminal 280 amino acids of intimin (Int280)
was not elucidated. from EPEC O127:H6 and EHEC O157:H7
could directly bind to HEp-2 cells (91), and
the same region also mediates binding of in-
Intimin
timin from EHEC O26:H to host cells (92).
Intimin is a 94- to 97-kDa outer membrane Furthermore, expression of the carboxyl-
adhesin produced by all EHEC strains and terminal two-thirds of intimin was sufficient
related A/E pathogens and is encoded by to restore adherence of an EHEC O157:H7
the eae gene. Though first identified in EPEC Deae mutant (93). Separate reports indicated
O127:H6 (4), a homolog in EHEC O157:H7 that binding of purified intimin to eukaryotic
that exhibits 83% amino acid sequence iden- cells could only be detected if the cells were
tity was subsequently reported to mediate pre-infected with EPEC or EHEC, indicating
bacterial adherence to cultured cells and in- that a bacterial factor also influences intimin-
testinal colonization in gnotobiotic piglets mediated adherence (94, 95).
(78, 79). Subsequent studies established that Studies to identify the host cell receptor for
EHEC O157:H7 intimin plays a pivotal role intimin initially focused on a 90-kDa mem-
in persistence and pathogenesis in mice, in- brane protein (Hp90) that became tyrosine
fant rabbits, neonatal calves and lambs, and phosphorylated during EPEC infection of ep-
adult cattle and sheep (8084). It has also ithelial cells and localized under adherent
been reported to drive mucosal inflammatory bacteria (96). EPEC intimin was later shown
responses; for example, it induces a T-helper to bind to tyrosine-phosphorylated Hp90 but
cell type 1 (TH1) response characterized not to nonphosphorylated Hp90, and it was
by mucosal thickening and infiltration of hypothesized that the bacteria signal to host
CD4+ T cells during infection of mice with cells to induce phosphorylation of a mem-
C. rodentium (85), and can augment mitogen- brane protein that is subsequently bound by
stimulated proliferation of spleen CD4+ T lym- intimin (95). Kenny et al. later reported Hp90
phocytes and cells from organized lymphoid to be the tyrosine-phosphorylated version
tissues (85, 86). Intimin is a component of of a 78-kDa EPEC O127:H6 protein that is
chimeric and multivalent subunit vaccines translocated into the eukaryotic cell plasma
that control EHEC in experimental models membrane via the LEE-encoded T3SS where
(reviewed in reference 246), and immuniza- it acts as the receptor for intimin (51). Hp90
tion with intimin alone has been reported was thus renamed Tir (for translocated in-
to confer protection against intestinal coloni- timin receptor). Tir was independently dis-
zation by EHEC O157:H7 when delivered by covered in EHEC O26:H as EspE, an 80-kDa
live-attenuated Salmonella sp. to cattle and protein that is translocated into host cells
mice (87, 88) and recombinant plants to where it appears as a 90-kDa tyrosine-
mice (82). Moreover, neonatal piglets suckling phosphorylated protein associated with ad-
dams vaccinated with EHEC O157:H7 intimin herent bacteria (97). EHEC O157:H7 Tir is
are passively protected (89), consistent with also delivered into host cells where it acts as a
the ability of antibodies directed against the receptor for intimin; however, it is not tyro-
carboxyl-terminal domain to inhibit adher- sine phosphorylated (97, 98).
ence of EHEC O157:H7 to HEp-2 cells (90). The three-dimensional structures of the
Intimin shares significant homology with carboxyl-terminal domain of intimin and the
invasin proteins of Yersinia pseudotuberculosis Int280-Tir complex were first resolved for
and Yersinia enterocolitica, the carboxyl ter- the EPEC O127:H6 protein by X-ray crystal-
mini of which bind to b1-chain integrins to lography and nuclear magnetic resonance

(99101). These studies indicate that intimin EPEC clone 1 serotypes O55:H6 and O127:H6,
contains four distinct domains within the whereas Int-g is found in EHEC clone 1 sero-
carboxyl-terminal 380 residues that extend types, including O157:H7 and its progenitor
from the outer membrane. Domains D1, D2, EPEC serotype O55:H7. As the carboxyl-
and D3 belong to the immunoglobulin super- terminal domain of intimin mediates binding
family and are predicted to comprise an to Tir and host cell surfaces, it was anticipated
articulated rod connected to the membrane- that divergence of intimin subtypes may in-
anchored amino-terminal portion by a flexible fluence the avidity and specificity of adher-
linker comprising two glycine residues. The ence. Int-g is required for colonization of the
terminal D4 domain is predicted to be acces- surface and glandular epithelium of the large
sible to the target cell and shares similarity intestine in gnotobiotic piglets (78). However,
with C-type lectins, a family of proteins that expression of Int-a from EPEC O127:H6 in an
recognize cell surface carbohydrates. The EHEC O157:H7 Deae mutant was reported to
most distal immunogloblin-like domain (D3) cause a shift in intestinal tissue tropism, with
and D4 C-type lectin domain form a rigid colonization of the terminal ileum as well as
superdomain that binds Tir (99), principally the surface of the large intestine in a manner
through contacts in a b-hairpin motif at the similar to wild-type EPEC (110). In subse-
tip of the extracellular domain of Tir and an quent studies, precise chromosomal replace-
analogous region of the intimin D4 domain ment of EHEC O157:H7 eae with EPEC O127:
(101). This portion of Tir is critical for intimin H6 eae did not alter tissue tropism in piglets
binding (102104). The crystal structure of the (111), and the basis of the disparity is unclear.
EPEC O127:H6 Int280-Tir complex reveals Support for the role of intimin in mediating
the intimin-binding domain of Tir to be a intestinal tissue tropism derives from in vitro
dimer, with the two parallel a-helices of the organ culture studies using human and por-
Tir extracellular domains aligning to form a cine intestinal explants. Adherence of EHEC
four-helix bundle bound by two intimin mol- O157:H7 to human intestinal mucosa and the
ecules (101). Analysis of the crystal structure formation of A/E lesions are restricted to
of the Tir-binding domain of EHEC O157:H7 follicle-associated epithelium (FAE) overlying
intimin (Int188) at 2.8 resolution suggests ileal Peyers patches (112), whereas Int-a-
that a similar conformation and contacts are expressing EPEC O127:H7 can adhere to hu-
adopted (105), extending earlier predictions man small intestinal explants from a variety of
from a yeast-2-hybrid analysis (106). Although sites (113). Expression of EPEC O127:H6 Int-a
it is unclear if this structure is formed by the in an EHEC O157:H7 Deae mutant without
native proteins in vivo or by A/E pathogens exchanging Tir results in an EHEC strain ca-
with other intimin and Tir sequences, it pre- pable of adhering to explants from various
dicts that the proteins bind in a plane roughly intestinal sites from human (114) and pigs
parallel to the surfaces of the bacteria and (115). A similar extension of tropism for hu-
host cell, consistent with the estimated gap man and porcine explants was observed when
between the host cell plasma membrane and EHEC O157:H7 was engineered to express
intimately attached bacteria of just 100 . Int-b from C. rodentium (116). In reciprocal ex-
Early phylogenetic analysis of eae genes periments EHEC O157:H7 Int-g conferred
identified distinct subtypes, designated Int-a, upon an EPEC O127:H6 Deae mutant a tro-
-b, -g, -d, -, and -v (107109), though further pism for FAE overlying Peyers patches (113).
intimin subtypes are now recognized. These FAE at the recto-anal junction has been
differ in sequence in the carboxyl-terminal reported to be a key site of persistence of
cell-binding domain and are often associated EHEC O157:H7 in cattle (117); however, while
with specific clonal lineages of EPEC and Int-g is known to be required for efficient
EHEC. For example, Int-a is associated with colonization of this site by EHEC O157:H7

(118), it is unclear if different subtypes vary in Sites of adherence of EHEC O157:H7 to

their specificity for this region of the bovine porcine and bovine tissue are also enriched
gut. Indeed, non-O157 EHEC has been re- in b1-chain integrins (122). Consistent with
ported to colonize squamous epithelium in a role for such factors as coreceptors, EPEC
the bovine terminal rectum independently O127:H6 intimin-a can bind to b1-chain in-
of intimin (119), and studies have indicated tegrins expressed on the surface of human
that factors other than those encoded by the lymphocytes or in enzyme-linked immuno-
LEE play a role in adherence of EHEC O157: sorbent assays (124). However, b1-chain in-
H7 to rectal squamous epithelial cells in cul- tegrins were reported to be dispensable for
ture (120). intimin-mediated adherence as inactivation
The ability of intimin to bind to eukaryotic of the b1-chain integrin gene or the addition
cell surfaces in the absence of Tir and influ- of antagonists of integrin function, such as
ence tissue tropism suggests the existence of EDTA or anti-b1 chain antibody, did not affect
a cellular coreceptor(s). Sinclair and OBrien EPEC adherence or pedestal formation (94).
demonstrated that the carboxyl-terminal do- Conversely, Muza-Moons et al. suggested that
main of EHEC O157:H7 Int-g binds in a spe- disruption of tight junctions by EPEC in an
cific and saturable manner to HEp-2 cells EspF-dependent manner leads to redistribu-
with a dissociation constant of 84 nM, con- tion of b1-chain integrins from the basolateral
sistent with the existence of a single host cell to apical surface to promote bacterial adher-
receptor (121). By affinity purification and se- ence (125). Studies using isogenic single and
quencing of peptides derived from a 110-kDa double eae and tir mutants of EHEC O157:H7
intimin-binding HEp-2 cell protein, the re- in calves have indicated that intimin-Tir in-
ceptor was identified as nucleolin, a protein teractions are more significant than intimin-
involved in regulation of cell growth that coreceptor interactions during colonization
can be expressed at the cell surface (121). Cell of the bovine intestines, as mutation of eae in a
surface-localized nucleolin was observed to tir mutant did not exert further attenuation
colocalize with bound purified Int-g and (84).
with EHEC O157:H7 adhering to HEp-2 cells
(121), and with EHEC O157:H7 adhering to
Tir
porcine and bovine intestinal mucosa in vivo
(122). Antinucleolin antibodies partially in- In addition to its role as the translocated re-
hibit EHEC O157:H7 adherence to cultured ceptor for intimin (above), Tir is required
cells (121). Interestingly, surface expression to activate actin assembly and recruit cyto-
of nucleolin is enhanced by Stx2, and it is skeletal proteins at the site of bacterial ad-
believed that this may partially explain the herence. Intimin-Tir interactions are required
ability of Shiga toxin to promote EHEC to cluster Tir to initiate this process (126).
O157:H7 adherence to Hep-2 cells and intes- The mechanism of insertion of Tir into the
tinal colonization in mice (122). Intimin sub- apical plasma membrane of enterocytes is ill-
types a, b, and g bind to nucleolin with equal defined, but phosphorylated intermediates of
affinity, indicating that the distribution of EPEC O127:H6 Tir can be detected in the
nucleolin along the gastrointestinal tract is cytoplasm and a delay exists between Tir in-
unlikely to determine tissue tropism of EHEC jection and intimin-Tir interaction (102, 127),
and EPEC (123). All three intimin subtypes implying that it may first be found in the cy-
bind to nucleolin with a lower avidity than to toplasm prior to insertion. Phosphorylation of
Tir. Furthermore, binding of intimin a, b, or g tyrosine residue 474 in a 12-amino-acid motif
to Tir in vitro blocks the interaction between in the carboxyl-terminal domain of EPEC
intimin and nucleolin (123), suggesting that O127:H7 Tir is required to trigger local actin
Tir and nucleolin compete for intimin binding. assembly through recruitment of the host cell

adaptor protein Nck (102, 126, 128, 129). Such repeats varies in natural isolates and is asso-
phosphorylation requires redundant cellular ciated with the efficiency of actin assembly
tyrosine kinases, including c-Fyn, Abl, Arg, (144, 145). Remarkably, the Arp2/3 complex is
and Etk, and was proposed to allow Nck to recruited to actin pedestals formed in a Tir-
bind a proline-rich domain of the neural and EspFU/TccP-dependent process even in
Wiskott-Aldrich syndrome protein (N-WASP), the absence of N-WASP, implying that re-
which in turn stimulates the actin-nucleating dundant as yet unknown pathways are used to
activity of the cellular Arp2/3 complex (102, ensure actin assembly under adherent bacte-
129, 130). WASP-interacting protein (WIP), ria (146). Indeed, an EPEC O125:H6 strain
which binds Nck and a WASP-homology 1 unable to use the Nck or IRTKS/IRSp53 path-
(WH1) domain of N-WASP, may also act as ways was nevertheless able to form typical
linker between Nck and N-WASP, and such A/E lesions on human intestinal explants cul-
interactions may act in synergy with Nck in- tured ex vivo (147).
teractions with the N-WASP proline-rich do- It is noteworthy that differences in the
main as interference in either interaction does relative importance of Nck-dependent and
not abolish pedestal formation (131). Recent Nck-independent pathways have been de-
studies have indicated that WIP is not essen- tected between immortalized epithelial lines
tial for pedestal formation by typical EPEC in culture and human intestinal explants cul-
(132). tured ex vivo (148) as well as between in-
N-WASP is also required for pedestal for- fection of cultured cells and animals (149).
mation by EHEC O157:H7 (133); however, in During C. rodentium infection in mice or
contrast to EPEC O127:H6, the Tir of EHEC EHEC infection of human intestinal biopsies,
O157:H7 lacks the Y474 residue and is not IRTKS, but not IRSp53, was recruited to sites
tyrosine phosphorylated upon entry into host of bacterial attachment sites, and N-WASP
cells (98, 134). Instead, a conserved Asn-Pro- was recruited to pedestals even in the absence
Tyr (NPY458) motif in a 12-amino-acid motif in of tyrosine residues required for Nck or
the cytoplasmic domain of EHEC O157:H7 Tir IRTKS binding (149). Thus, despite the key
is required for Nck-independent actin assem- role of the Tir:Nck and Tir:IRTKS/IRSp53
bly (135136) through recruitment of insulin pathways in actin polymerization in cultured
receptor tyrosine kinase substrate (IRTKS) epithelial cells, they appear not to be essential
(137) and insulin receptor substrate p53 for N-WASP recruitment or A/E lesion for-
(IRSp53) (138). This motif is also found in mation on explants or in mice. C. rodentium
EPEC O127:H7 Tir as NPY454 and can initiate mutants with Y451 or Y471 substitutions were
actin assembly independently of Nck, albeit outcompeted by the wild-type strain during
inefficiently (139). In EHEC O157:H7 the path- mixed infection of mice (149), suggesting that
way of actin assembly mediated by recruit- such pathways do contribute to persistence.
ment of IRTKS/IRSp53 to Tir is amplified by Further, while an EHEC O157:H7 espFU/tccP
another Type III secreted effector protein mutant exhibited no obvious defect in A/E
termed EspFU or TccP (Tir cytoskeleton lesion formation or early intestinal coloniza-
coupling protein), which contains 47-amino- tion of calves (84), such mutants appear im-
acid proline-rich repeats that bind to IRTKS/ paired in the efficiency of formation and tissue
IRSp53 (138) and the GTPase-binding domain distribution of A/E lesions in gnotobiotic pig-
of N-WASP (140, 141). Structural studies have lets and infant rabbits (150).
indicated that the proline-rich repeats in The Tir:Nck and Tir:IRTKS/IRSp53 path-
EspFU/TccP compete with an auto-inhibitory ways appear conserved in isolates of EPEC
domain of N-WASP for binding of its GTPase- belonging to lineage 1 and EHEC O157:H7,
binding domain, thereby relieving N-WASP respectively (151). Typical isolates of EHEC
auto-inhibition (142, 143). The number of such O157:H7 carry espFU/tccP on prophage

CP-933U/Sp14 but also harbor a related pseu- have evolved. The repertoire of Tir, EspFU/
dogene (z1385/ecs2715) on prophage CP-933 M/ TccP, and EspFM/TccP2 pathways may ex-
Sp4 (tccP2/espFM). Sorbitol-fermenting EHEC plain the recruitment of distinct sets of cyto-
O157:H strains contain an intact copy of tccP2/ skeletal proteins to the pedestals formed by
espFM, which can restore function to an EspFU/ different A/E pathogens as well as distinct
tccP mutant and encodes a protein with near- pedestal morphologies (Fig. 1A and B). Mod-
identical proline-rich repeats but a distinct ulation of the cytoskeleton by A/E pathogens
80-amino acid amino terminus, which shares is summarized in Fig. 5.
42.7% identity with the amino terminus of Tir plays a key role in persistence and
EspFU/TccP (151). Interestingly, non-O157 pathogenesis of EHEC, with tir mutants of
EHEC isolates in general possess a Tir with a EHEC O157:H7 being attenuated in infant
conserved Y474 residue for Nck recruitment rabbits (83), calves (84), and colonization of
and an intact copy of tccP2/espFM (151), sug- the terminal rectum of steers (118). Moreover,
gesting that they are able to use both the Tir is an important component of vaccines for
Tir:Nck and Tir:IRTKS/IRSp53 pathways. The control of EHEC O157:H7 in cattle (70, 71).
same applies to a high proportion of EPEC Despite intensive research into the molecular
lineage 2 strains (152), indicating that varied mechanisms underlying actin assembly during
and independent solutions to form pedestals A/E lesion formation, the advantages to the
FIGURE 5 Diagram summarizing the activities of a subset of EHEC Type III secreted proteins on the cytoskel-
eton. Note (a), the Tir:Nck pathway dependent on phosphorylation of the residue equivalent to tyrosine 474 of
EPEC O127:H6 Tir operates in some non-O157 EHEC but not prototype E. coli O157:H7 strains. Effectors are
represented by circles. Adapted from reference 164. doi:10.1128/microbiolspec.EHEC-0007-2013.f5

pathogen of forming pedestals remain open to activity of reporter fusions have been used.
question. It has been reported that N-WASP is Commonly, candidate non-LEE-encoded effec-
required for efficient translocation of Type III tors have been fused to TEM-1 b-lactamase,
secreted effectors (146), and a role in stabi- enabling translocation to be detected by a
lizing the translocon and/or clustering Tir for shift in the wavelength of fluorescence emit-
interactions with intimin is plausible. In turn, ted by target cells loaded with a fluorescent
stabilization of intimin-Tir interactions may b-lactamase substrate (CCF2/AM) (158). Fu-
help prevent bacterial detachment from cells sions to epitopes (e.g., FLAG) or Bordetella
under flow. Pedestals have also been reported pertussis adenylate cyclase (CyaA) have also
to be mobile on the cell surface (153), and a been employed for this purpose (25).
role in spread between enterocytes and eva- In total, 39 Type III secreted effector
sion of phagocytosis by remodeling of the cell proteins were experimentally validated in
surface cannot be excluded. EHEC O157:H7 (25). It is evident that lamb-
doid bacteriophages have played a key role in
the lateral transfer of such effectors, both in
Other Type III Secreted Effectors
EHEC O157:H7 (25) and non-O157 EHEC
In addition to EspB and Tir, five other effector (18). Prophage-encoded effector genes are fre-
proteins are encoded within the LEE of the quently carried downstream of tail fiber genes
prototype EHEC O157:H7 strains EDL933 and and exhibit a low %GC content relative to the
RIMD 0509952 (EspF, EspG, EspH, EspZ, prophage backbone (18, 25). Analysis of the
and Map) (Fig. 2). As discussed above, the content and insertion site of effector-encoding
boundaries of the LEE are less defined in some prophages in non-O157 EHEC suggests distinct
EHEC strains, with additional effector pro- evolutionary histories (18). Although many
teins encoded in the flanking regions. Upon se- effector-encoding prophages identified by ge-
quencing of the complete genome sequences nome sequencing were initially considered
of the above strains (154, 155), more than 60 defective, it is now evident that many are in-
candidate Type III secreted proteins were ducible and can release their DNA from EHEC
identified on the basis of homology with ef- O157:H7, possibly owing to recombination and
fectors in other pathogens, of which 49 were other interprophage interactions (159). Many
judged to be potentially functional (25). The effectors encoded outside the LEE are desig-
T3SS-dependent secretion of 31 candidate ef- nated Nle (non-LEE-encoded) effectors. Ef-
fectors of EHEC O157:H7 was confirmed by fector proteins of EHEC O157:H7 strain RIMD
analysis of the secreted proteome of an EHEC 0509952 and their activities, where known or
O157:H7 DsepL strain with an isogenic DsepL inferred from studies with homologs in other
DescR double mutant incapable of Type III A/E pathogens, are listed in Table 1. A notable
secretion (25). SepL regulates the hierarchy absence from the genomes of the sequenced
of secretion in EHEC, and mutation of sepL EHEC O157:H7 strains EDL933 and RIMD
causes the secretion of effectors in preference 0509952 is the non-LEE-encoded effector
to translocon components (156). Subsequent cycle-inhibiting factor (Cif), which was first
studies have indicated that the ability of the described in an EHEC O103:H2 strain (160) and
carboxyl-terminal portion of SepL to bind to is found in many non-O157 EHEC. Cif arrests
Tir is sufficient to delay the export of effectors the cell cycle at G1/S and G2/M phases and
while the EspABD translocon components are inhibits other cellular processes by deamida-
secreted (157). tion of ubiquitin or the ubiquitin-like protein
To confirm that candidate effectors iden- NEDD8 that regulates cullin-RING-ubiquitin
tified by bioinformatics or proteomic analysis ligase complexes (reviewed in reference 161).
are translocated into eukaryotic cells in a It is clear that effectors are often multi-
T3SS-dependent manner, assays to detect the functional. Indeed, EspF has been described

TABLE 1 Type III secreted effector proteins of EHEC O157:H7 and their activities, where known or inferred from
homologs in other A/E pathogensa
Interacting
No. of Subcellular partners
Effector allelesb localization or substrates Homologs Functionc
EspB 1 Plasma 1-Antitrypsin, Yersinia YopD Pore-forming translocon

membrane, -catenin, myosin-1c component; microvillus
cytosol effacement; antiphagocytosis;
disruption of adherens junctions
EspF 1 Plasma 14-3-3zeta, ABCF2, EspFU (TccP) Disrupts mitochondrial function,
membrane, actin, Arp2, CK18, nucleolus, tight junctions and
cytosol, N-WASP, proling, intermediate laments.
mitochondria, SNX9, ZO-1/-2 Inactivates NHE3 and SGLT-1;
nucleus activates SNX9 and N-WASP;
inhibits PI3K-dependent
phagocytosis and induces
apoptosis
EspG 1 Cytosol, Golgi Arf1/6, PAK1/2/3, Shigella VirA Disrupts microtubules,
tubulin, GM130, Rab1 tight junctions and paracellular
permeability; sequesters
ADP-ribosylating factor (Arf) to
modulate GTPase signaling;
stimulates p21-activated kinases
(PAKs) to inhibit
endomembrane trafcking;
binds GM130 and inactivates
Rab1 to disrupt Golgi structure
and protein secretion;
induces calpain protease and
necrosis in absence of Tir
EspH 1 Plasma DH-PH Rho guanine None known Inhibits RhoGTPase signaling
membrane, nucleotide exchange and FCR-mediated
pedestal factors (GEFs) phagocytosis; causes cell
detachment via disassembly of
focal adhesions and remodels
brush border; promotes actin
nucleation and pedestal
elongation by recruiting
N-WASP and Arp2/3 via WIP
EspZ 1 Plasma CD98, translocase of None known Inhibits apoptosis and
membrane, inner mitochondrial cytotoxicity; regulates Type III
mitochondria membrane 17b secretion via translocation
(TIM17b) stop activity
Map 1 Mitochondria Na+/H+ exchanger Salmonella GEF for Cdc42 that induces
regulatory factor SopE/SopE2, transient lopodia formation;
(NHERF)-1 and -2, Shigella IpgB causes mitochondrial
Cdc42 dysfunction; inactivates
sodium-D-glucose cotransporter
(SGLT-1) in a manner that may
result in net uid secretion;
disrupts tight junctions,
causing loss of epithelial
barrier integrity

TABLE 1 (Continued)
Interacting
Tir 1 Plasma Intimin, 14-3-3tau, None known Receptor for intimin,

membrane -actinin, cortactin, actin pedestal formation,
CK18, IQGAP1, IRTKS, regulates activities of Map
IRSp53, Nck, PI3K, and EspG
SHP-1, Talin, Vinculin
EspI/NleA 1 Plasma Syntrophin, Sec23/24, None known Inhibits protein export from the
membrane, MALS3, PDZK11, endoplasmic reticulum by
Golgi SNX27, TCOF1, disrupting COPII function;
NHERF-1 and disrupts tight junctions
-2, MAGI-3, SAP97
and -102 PSD-95
EspJ 1 Cytosol, Unknown Pseudomonas Inhibits phagocytosis mediated
mitochondria HopF by complement receptor
3- and Fc-receptor
EspK 1 Cytosol Unknown Salmonella Unknown, inuences intestinal
GogB colonization of calves by
EHEC O157:H7
EspL 4 (1) Pedestal Annexin 2 Shigella OspD Promotes F-actin bundling
activity of annexin 2
EspM 2 Cytosol RhoA Salmonella GEF for RhoA that induces actin
SopE/SopE2, stress bers and modulates
Shigella IpgB pedestal formation; disrupts
tight junctions and
monolayer integrity
EspN 1 Unknown Unknown E. coli CNF1, Unknown
Salmonella
Arizonae
SARI_01330
SARI _01464
EspO 2 Unknown Integrin-linked Shigella OspE Regulates EspM2-mediated
kinase (ILK) RhoA activity and stabilizes focal
adhesions to block cell
detachment
EspR 4 (1) Unknown Unknown Shigella exneri Unknown
SF1757
EspV 1 (1) Cytosol Unknown Pseudomonas Alters cell morphology
AvrA
EspW 1 Unknown Unknown Pseudomonas Unknown
HopPmaA,
HopW1
EspX 7 (1) Unknown Unknown Shigella sonnei Unknown
SSON_0027,
Shigella
dysenteriae
Sd1012_0237
EspY 5 (1) Unknown Unknown Salmonella Unknown
SopD

TABLE 1 Type III secreted effector proteins of EHEC O157:H7 and their activities, where known or inferred
from homologs in other A/E pathogensa (Continued)
Interacting
NleB 3 (1) Cytosol Glyceraldehyde Salmonella SseK Inhibits TNF-induced activation

3-phosphate of NF-B and proinammatory
dehydrogenase responses by transferring
(GAPDH) N-acetyl-D-glucosamine to
GAPDH, thereby disrupting
TRAF2-GAPDH interaction to
suppress TRAF2
polyubiquitination and
NF-B activation
NleC 1 Cytosol, p65 (RelA), p50, c-Rel, Photobacterium Zinc metalloprotease that
nucleus IB, acetyltransferase AIP56 cleaves p65 (RelA), p50,
p300 c-Rel and IB to inhibit
NF-B activation
NleD 1 Cytosol c-Jun N-terminal Pseudomonas Zinc metalloprotease that
kinase (JNK), p38 HopAP1, HopH1 cleaves JNK and MAPK to inhibit
mitogen-activated induction of apoptosis and
protein kinase proinammatory responses
(MAPK)
NleE 1 Cytosol TAB2 and -3 Shigella OspZ Blocks IB degradation to inhibit
activation of NF-B,
proinammatory responses and
neutrophil transepithelial
migration. Uses S-adenosyl-L-
methionine-dependent
methyltransferase activity to
modify Npl4 zinc nger
domains in TAB2 and TAB3,
which regulate NF-B signaling
NleF 1 Unknown Caspase-4, -8, None known Inhibitor of caspase activation
and -9, Tmp21 and apoptosis; binds the
COPI-vesicle receptor Tmp21
involved in Golgi function and
slows protein secretion
NleG/NleI 14 (5) Cytosol UBE2D2 Salmonella E3 ubiquitin ligases analogous
Typhi STY1076 to eukaryotic RING nger and
U-box enzymes
NleH 2 Plasma Bax-inhibitor 1, Shigella OspG Binds Bax-inhibitor 1 to block
membrane, NHERF2, ribosomal apoptosis; sequesters RPS3 to
cytosol, protein S3 (RPS3) inhibit NF-B signaling
endoplasmic
reticulum
TccP/ 2 (1) Pedestal N-WASP, IRSp53, EspF Relieves auto-inhibition of
EspFU IRTKS, cortactin N-WASP to stimulate the Arp2/3
complex to polymerize actin
and form pedestals
a
Adapted from reference 164.
b
Allele number is shown for the Sakai outbreak strain of EHEC O157:H7 (RIMD 0509952). Values in parentheses denote the number of
predicted pseudogenes.
c
References in support of functions may be found in the text and reference 164.

as the Swiss army knife of A/E pathogens other A/E pathogens disrupt these processes
owing to its effect on epithelial barrier func- at multiple levels, as summarized in Fig. 6.
tion, apoptosis, mitochondrial function, cyto- For example, NleB suppresses tumor necrosis
skeletal components, microvillus effacement, factor alpha-mediated NF-kB activation (165,
and other processes (reviewed in reference 166) by transferring N-acetyl-D-glucosamine
162). It is also evident that effectors may act in to the glycolysis enzyme glyceraldehyde-3-
redundant, synergistic, and antagonistic ways. phosphate dehydrogenase, disrupting its abil-
As a consequence, mutation of effector genes ity to interact with the tumor necrosis factor
does not produce attenuation at the same level receptor-associated factor 2 (TRAF2), which
as defects affecting the translocon or needle is required for NF-kB activation (167). In
complex, which result in loss of secretion all EPEC it appears that Tir also impairs TRAF
effectors. The LEE-encoded effectors EspF, function by interacting with the TRAF2 adap-
EspG, EspH, and Map have been reported tor protein and inducing its proteasome-
to be required for full colonization of the in- independent degradation (168). It has also
testines of infant rabbits by EHEC O157:H7, been reported that EPEC Tir interferes with
albeit that the effect was modest and tissue- signaling via Toll-like receptors by binding
specific in some cases (42). Moreover, muta- to the host cell tyrosine phosphatase SHP-1
tions affecting 29 or the 39 effectors of EHEC in a manner dependent on phosphorylated
O157:H7 strain EDL933 were detected by immunoreceptor tyrosine-based inhibition
screening pools of random transposon mu- motifs in Tir (169). The association of Tir with
tants in calves (39, 41); however, the extent SHP-1 facilitates the recruitment of SHP-1 to
of negative selection was often modest, and the adaptor TRAF6, thereby inhibiting the
in the case of map and nleD mutants, could ubioquitination of TRAF6 and the subsequent
not be reproduced when the mutants were induction of proinflammatory cytokines and
screened in isolation (39, 163). Functions of intestinal immunity (169).
Type III secreted proteins of EHEC are con- At another level, NleC inactivates the NF-
sidered below by functional categories and are kB subunits p50 and p65 by means of a zinc-
reviewed in detail elsewhere (164). dependent endopeptidase activity (170173).
Another zinc metalloprotease (NleD) degrades
key kinases (c-Jun N-terminal kinase [JNK]
MODULATION OF INNATE IMMUNITY and mitogen-activated protein kinase p38) in-
volved in activation of activator protein-1,
In recent years it has become clear that EHEC which like NF-kB controls expression of genes
and other A/E pathogens inject multiple involved in the inflammatory response (170).
non-LEE-encoded effectors to suppress pro- A further effector, NleE, interferes with deg-
inflammatory responses, converging on inhi- radation of IkB to prevent translocation of
bition of the activation of host transcription NF-kB to the nucleus (165, 166, 174) and uses
factors including NF-kB and activator protein- S-adenosyl-L-methionine-dependent methyl-
1. NF-kB comprises a dimer of p50 and p65 transferase activity to modify Npl4 zinc finger
subunits that are retained in the cytoplasm by domains in the cellular proteins TAB2 and
a family of inhibitory proteins (IkB). Activa- TAB3, which regulate NF-kB signaling (175).
tion of NF-kB requires phosphorylation of IkB NleH also acts in concert with these effectors,
by the IkB kinase complex, which is subse- binding to a subunit of NF-kB transcriptional
quently ubiquitinated and degraded by the complexes (ribosomal protein S3), preventing
proteasome, allowing p50/p65 to translocate its phosphorylation by the kinase IkB kinase
to the nucleus to bind consensus sequences complex-b and translocation to the nucleus (176,
upstream of genes associated with the in- 177). NleH1 and NleH2 also suppress IkB deg-
flammatory response. Remarkably, EHEC and radation by interfering with the ubiqitination

FIGURE 6 Diagram summarizing the activities of a subset of EHEC Type III secreted proteins on signaling
pathways leading to inammation and apoptosis. Effectors are represented by circles. Adapted from reference
164. doi:10.1128/microbiolspec.EHEC-0007-2013.f6
of phosphorylated IkB, thereby aiding re- EHEC and EPEC are able to disable phago-
tention of NF-kB in the cytoplasm (178). The cytosis by macrophages via the FCg recep-
functional significance of inhibition of pro- tor (FCgR) and complement receptor (CR3),
inflammatory responses through these path- which, respectively, mediate uptake of parti-
ways is evidenced by the fact that mutants cles opsonised with immunoglobulin G or
lacking the effectors above typically elicit ele- complement fragment C3bi (180). Such inhi-
vated proinflammatory cytokine responses and bition is partly dependent on EspJ (180),
are attenuated in animal models, albeit that though its mode of action is ill-defined at the
combination of mutations produces stronger time of writing. The myosin-binding effector
phenotypes owing to functional redundancy. EspB also plays a role in suppressing phago-
One should also caution that subtle differences cytosis of EPEC, possibly by interfering with
may exist in the activities of homologs be- myosin-actin interactions required to form
tween EHEC and other A/E pathogens, as well phagocytic cups and force phagosome closure
as between members of the same family of during uptake (76). EspB may act in concert
effectors, as recently reported for NleH1 and with EspH to inhibit remodeling of the cyto-
NleH2 of EHEC O157:H7 (179). skeleton during phagocytosis, as the ability of
Another level at which Type III secreted EspH to inactivate RhoGTPase signaling has
effectors act in concert to repel host innate been associated with blocking of uptake of
defenses is the inhibition of phagocytosis. EPEC O127:H6 by macrophages (181).

In addition to the effectors above, EspF has of filopodia at attachment sites (127, 190, 191),
been described to play a key role in preventing whereas EspM activates RhoA to form actin
phagocytosis by macrophages and transloca- stress fibers (192, 193). A further WxxxE ef-
tion across microfold-cell-like cells in culture fector (EspT) has a putative GEF activity for
(180, 182184). In the case of EPEC O127:H6 Cdc42 and Rac1 to trigger membrane ruffling,
EspF, this involves inhibition of PI3 kinase- and filopodia formation has been found in
dependent pathways (183, 185), but not FCgR- a minority of EPEC strains, but so far not
or CR3-dependent opsonophagocytosis (181), in EHEC (194). The relevance of the GEF
suggesting that it acts at a distinct level activity of Map and EspM in vivo is not fully
to EspJ. Antiphagocytic activity of EPEC understood, but recent research with Sal-
O127:H6 EspF has been mapped to the amino- monella sp. has indicated that activation of
terminal 101 residues (183), which contain RhoGTPases by SopE triggers the NOD1 sig-
binding sites for sorting nexin 9 (SNX9), actin, naling pathway leading to inflammation, in-
and profilin (186188), and it is possible that dicating that their activity may extend beyond
it acts in concert with EspB and EspH to dis- changes to cell architecture (195). Interest-
rupt the cytoskeletal rearrangements required ingly, A/E pathogens also inject EspH, which
to form and close phagosomes. Paradoxically, has an opposing activity to Map, EspM, and
interactions between EspF and SNX9 have EspT. EspH inactivates cellular Dbl-homology
been reported to remodel the apical plasma and pleckstrin-homology (DH-PH) Rho fam-
membrane and facilitate EPEC invasion in ily GEFs by competitively binding to tandem
epithelial cells (189), indicating the potential DH-PH domains, resulting in inhibition of
for cell type-specific effects. Moreover, dis- activation of RhoA (181). The bacterial GEFs
tinct espF alleles may confer distinct activities Map, EspM, and EspT lack sequence homol-
as comparative studies found EspF of EHEC ogy with DH-PH domains, and recent studies
O157:H7 to be far less effective in inhibiting have shown that EspH does not directly in-
macrophage phagocytosis than the variant in terfere with their activity (196). Rather, EHEC
EPEC O127:H6, possibly owing to weaker in- and EPEC appear to use EspH to inactivate
teractions with SNX9 (184). Conversely, EspF endogenous RhoGEFs while injecting their
from EHEC O157:H7 was more effective than own GEFs to modulate Rho GTPase signaling
the EPEC O127:H6 EspF in preventing uptake to their advantage (196). The interplay be-
by primary epithelial cells from the site of tween these effectors is such that pheno-
EHEC O157:H7 persistence in the bovine types detected by expression of effectors in
terminal rectum (184). isolation are to be interpreted with caution.
For example, expression of EspH alone causes
disassembly of focal adhesions, cell detach-
Modulation of the Cytoskeleton
ment, caspase-3 activation, and cytotoxicity,
The role of Tir and adaptor proteins in for- but its activity is normally offset by the op-
mation of actin pedestals is well established posing activity of EspM2 and, in EPEC, EspT
(above); however, it is becoming clear that (196). Overexpression of EspH has been re-
other effectors are injected that act on the ported to result in elongation of actin pedes-
host cell cytoskeleton (summarized in Fig. 5). tals, and conversely, espH mutation causes
For example, mitochondrial-associated pro- them to shorten (197). It is now understood
tein (Map) and EspM are WxxxE-family ef- that EspH plays a role in recruitment of
fectors homologous to the Shigella IpgB and N-WASP and the Arp2/3 complex to assemble
Salmonella SopE proteins that act as guanine actin at sites of bacterial attachment (198).
nucleotide exchange factors (GEFs) to activate It is possible this involves recruitment of the
GTPases that regulate the actin network. Map WASP-interacting protein WIP, as the WIP-
activates Cdc42 to trigger transient formation binding domain of N-WASP and carboxyl

terminus of Tir are required for the effect the process. EspF triggers apoptosis by dam-
(198). A further effector reported to subvert aging mitochondria and reducing the levels or
actin dynamics is EspV (199). Ectopic expres- function of Abcf2, which suppresses caspase 3
sion of EspV causes nuclear condensation, cell activation (212216). Map also targets mito-
rounding, and formation of dendrite-like pro- chondria, disrupting their membrane poten-
jections; however, it is relatively rarely found tial (127, 217), and may act in concert with
in EHEC and its mode of action is unknown. EspF to trigger cell death via mitochondrial
Actin is not the only component of the damage. In EPEC and some EHEC, Cif in-
cytoskeletion, and evidence is growing that duces a delayed form of apoptosis character-
effectors act on other structures involved ized by activation of caspases, accumulation
in cell archietcture. For example, EspG and of cleaved caspase-3, and exposure of the
EspG2 interact with tubulin, deplete micro- phosphatidylserine on the cell surface (218).
tubules at sites of attachment, and disrupt The basis of this effect is ill-defined but may
the integrity of tight junctions and epithelial be a consequence of Cif-mediated arrest of
paracellular permeability (200204). In a fur- the cell cycle and subversion of the ubiquitin-
ther example of the multifunctional nature of dependent degradation pathway. The rele-
effectors, recent work has also established that vance of such events during infection is open
EspG also interferes with protein secretion, to question. Indeed, one study reported a de-
including release of interleukin-8, by disrupting crease in the number of apoptotic cells in the
endoplasmic reticulum to Golgi transport via ileum and ileal Peyers patches of rabbits fol-
interactions with ADP-ribosylation factors, p21- lowing infection with rabbit enteropathogenic
activated kinases, Rab1 GTPase, and Golgi ma- E. coli (219).
trix protein GM130 (205208). Intermediate NleH effectors block the induction of mul-
filaments are also targets for effectors of A/ E tiple apoptotic pathways, including those
pathogens, and evidence exists that Tir recruits stimulated by staurosporine, brefeldin A, tuni-
the cytokeratins CK8 and CK18 to pedestals to camycin (220), and Clostridium difficile toxin
influence actin accretion (209) in a manner as- B (221). This appears to be due to the ability
sociated with direct binding of Tir to CK18 of NleH1 and NleH2 to interact with the
(209) and 14-3-3tau (210), which acts as an anti-apoptotic protein Bax inhibitor-1 (BI-1),
adaptor protein of CK18. Depletion of either as knockdown of BI-1 abolished their anti-
CK18 or 14-3-3tau impaired pedestal formation apoptotic activity (220). NleH was also ob-
(209, 210), implying a significant role. Spectrin, served to inhibit cleavage of the procaspase-3
which polymerizes under the plasma mem- at sites of attachment of C. rodentium to the
brane to form a scaffold important for cell shape murine gut (220), implying that it would pre-
and membrane integrity, is also recruited to vent infected cells from entering apoptosis in
EPEC O127:H6 and EHEC O157:H7 pedestals, a caspase-3-dependent manner. NleH proteins
and siRNA-mediated knockdown of spectrin have been reported to bind to Na(+)/ H(+)
and spectrin-associated proteins impaired pe- exchanger regulatory factor 2 (NHERF2), and
destal formation (211). this may limit its ability to block apoptosis,
as overexpression of NHERF2 appears to se-
quester NleH and impede its anti-apoptotic
Modulation of Apoptosis
activity (222). In recent studies NleF has
Synergistic and antagonist interactions of also been reported to prevent epithelial cell
Type III secreted effectors influence host cell apoptosis in a manner associated with di-
survival during infection by A/E pathogens, rect binding to caspase-4, -8, and -9 (223).
as summarized in Fig. 6. For example, EspF, NleD-mediated cleavage of JNK may also
Map, and Cif induce apoptosis whereas NleD, help reduce apoptosis as JNK stimulates this
NleF, NleH1, and NleH2 appear to counteract process.

Induction of Diarrhea cytokine synthesis (230) and is nearly identi-

cal to an adhesin of EHEC O111:H termed
It is evident that the LEE-encoded T3SS and
EHEC factor for adherence 1 (Efa1) (231).
selected effectors play key roles in the induc-
LifA/Efa1 is highly conserved in non-O157
tion of enteritis and fluid loss during infection
EHEC and plays a significant role in intestinal
of animals; however, the precise mechanisms
colonization of calves by EHEC strains of
involved in EHEC-induced diarrhea are not
serotypes O5:H, O26:H, and O111:H (232,
well understood. It is likely that a combina-
233). In each of these strains, mutations af-
tion of factors lead to net fluid accumulation
fecting lifA/efa1 impair adherence, but in
in the gut lumen, including a loss of adsorptive
EHEC O5:H and O111:H this could not be
capacity owing to microvillus effacement and
separated from a posttranscriptional reduc-
enterocyte extrusion, disruption of tight junc-
tion in the expression and secretion of EspA
tions, inhibition of water adsorption chan-
and Tir (233). LifA/Efa1 of an EHEC O26:H
nels (aquaporins), and dysregulation of ion
strain influences adherence independently of
and glucose transporters (reviewed in refer-
effects on EspA (232), and the basis of pleo-
ence 224). EspB has been reported to con-
tropic effects of lifA/efa1 mutation in non-
tribute to loss of microvilli (76), and EspF,
O157 EHEC is unknown.
EspG, Map, EspI/ NleA, and EspM disrupt
It is assumed that lymphostatin activity
tight junctions and monolayer integrity (201,
does not require Type III secretion as it is
203, 225227). EspF has been reported
detectable in cell-free lysates and can be
to downregulate the activity of Na(+)/H(+)
transferred to laboratory-adapted E. coli K-12
exchanger 3 (228), which is the main Na
lacking a T3SS (230). Moreover, its ability to
(+)-absorbing isoform in the mammalian small
promote adherence independently of effects
intestine. It is not yet clear how binding of
on Type III secretion and in a manner sensi-
the Na(+)/ H(+) exchanger regulatory factor 2
tive to antibody indicates that a portion of the
by Map, EspI, and NleH affects the activity of
protein may be surface localized (232, 234). It
such ion exchangers. It is evident that EPEC
is evident, however, that some Type III se-
O127:H6 EspF, Map, Tir, and intimin act co-
creted proteins can be secreted from E. coli
operatively to inactivate the sodium-D-glucose
by alternative pathways (e.g., the EPEC serine
cotransporter (SGLT-1), which plays a key
protease EspC) (235). Predicted glycosyl-
role in the uptake of fluid from the lumen
transferase and cysteine protease motifs of
in the normal intestine (229). However, it is
LifA/Efa1 were dispensable for intestinal col-
not clear if the same applies during EHEC
onization of calves by EHEC O26:H (232),
infection, and a lack of tractable models of
and an earlier report claiming roles for such
EHEC-induced diarrhea has hindered our
motifs during C. rodentium infection of mice
understanding of the role of specific effectors
(236) can be explained by the fact that the
in the process.
motif deletions caused truncation.
Most EHEC O157:H7 strains lack lifA/efa1
but contain a truncated variant (z4332) and
LifA/Efa1, Z4332 and ToxB
a full-length pO157-encoded homolog (toxB).
Quantitative analysis of the secreted proteome An extremely high frequency of carriage of
of EPEC O127:H7 by stable isotope labeling lifA/efa1, z4332, and/or toxB exists in LEE-
with amino acids in cell culture-based mass positive strains, with physical linkage of such
spectrometry recently revealed that LifA is a genes to the LEE in many cases. Deng et al.
substrate for secretion by the LEE-encoded recently confirmed that Z4332 and ToxB are
T3SS (22). LifA (lymphostatin) is a 366-kDa also effectors of the T3SS (22). Mutation of
protein that inhibits mitogen-activated lym- either factor has been reported to impair Type
phocyte proliferation and proinflammatory III secretion at a posttranscriptional level

(237, 238) but does not affect a lymphostatin- Tir, and translocon components show promise
like activity of EHEC O157:H7 (239). Intes- in control of EHEC O157 in cattle (reference
tinal colonization of calves or lambs by a 246); however, the extent of cross-protection
Stx-negative E. coli O157:H7 strain was not against other EHEC strains is ill-defined. In-
affected by single or double z4332 or toxB mu- hibitors of Type III secretion have proven
tations (237). As with other Type III secreted useful for chemical genetics to understand
effectors, it is likely LifA/Efa1 and ToxB are LEE function and regulation, but delivery of
multifunctional and their location and activi- such molecules at the required sites, concen-
ties inside host cells require study. trations, and times to ablate T3SS activity
poses a formidable challenge. Given the array
of effectors dedicated to disarming the innate
FUTURE PERSPECTIVES response, future research could consider the
extent to which this mediates persistence in
Research on the LEE-encoded T3SS and the reservoirs and prevents development of ef-
repertoire and functions of secreted effectors fective adaptive immunity, as this may inform
has expanded enormously. Today, it repre- the design of live vaccines.
sents one of the most vibrant areas of study
on E. coli O157:H7 and other EHEC strains,
ACKNOWLEDGMENTS
yet significant challenges remain. It is already
clear that analysis of the activities of effectors The authors gratefully acknowledge the fi-
in isolation cannot account for synergistic nancial support of the Biotechnology and Bi-
or antagonist activities with other effectors, ological Sciences Research Council, Medical
or reproduce the timing and intracellular Research Council and The Wellcome Trust.
concentration of effectors delivered in vivo. We apologize to those authors whose work
Moreover, a relatively small number of pro- could not be described or cited, owing to
totype strains are used, and variation in the constraints of space.
repertoire, sequence, and regulation of effec-
tors between strains may exert distinct phe-
CITATION
notypes. The sources of novel effectors and
regulators of LEE function, and frequency and Stevens MP, Frankel GM. 2014. The locus
impact of acquisition of these on the viru- of enterocyte effacement and associated
lence of EHEC, require further investigation. virulence factors of enterohemorrhagic Esch-
Moreover, research should recognize that erichia coli. Microbiol Spectrum 2(4):EHEC-
EHEC can infect multiple hosts and that ef- 0007-2013.
fectors that appear unimportant in surrogate
rodent- or cell-based assays may play key roles
REFERENCES
in adaptation to humans or reservoir hosts.
Unraveling the activities in effectors in reser- 1. Tzipori S, Wachsmuth IK, Chapman C, Birden
voir hosts is presently hindered by the paucity R, Brittingham J, Jackson C, Hogg J. 1986. The
pathogenesis of hemorrhagic colitis caused by
of reagents to study signaling pathways and
Escherichia coli O157:H7 in gnotobiotic piglets.
detect proteins in ruminant cells relative to J Infect Dis 154:712716.
murine and human systems. 2. Knutton S, Lloyd DR, McNeish AS. 1987. Ad-
While there is much merit in the intellec- hesion of enteropathogenic Escherichia coli to
tual pursuit of the molecular mechanisms human intestinal enterocytes and cultured human
intestinal mucosa. Infect Immun 55:6977.
by which constituents of the T3SS and its
3. Knutton S, Baldwin T, Williams PH, McNeish
effectors act, future research should also con- AS. 1989. Actin accumulation at sites of bac-
sider how we may exploit this knowledge for terial adhesion to tissue culture cells: basis of
practical gain. Vaccines that target intimin, a new diagnostic test for enteropathogenic and

enterohemorrhagic Escherichia coli. Infect Immun coli confers the attaching and effacing phenotype
57:12901298. on E. coli K-12. Mol Microbiol 23:399407.
4. Jerse AE, Yu J, Tall BD, Kaper JB. 1990. A ge- 15. Psfai G, Koob MD, Kirkpatrick HA, Blattner
netic locus of enteropathogenic Escherichia coli FR. 1997. Versatile insertion plasmids for targeted
necessary for the production of attaching and genome manipulations in bacteria: isolation, de-
effacing lesions on tissue culture cells. Proc Natl letion, and rescue of the pathogenicity island
Acad Sci USA 87:78397843. LEE of the Escherichia coli O157:H7 genome.
5. McDaniel TK, Jarvis KG, Donnenberg MS, J Bacteriol 179:44264428.
Kaper JB. 1995. A genetic locus of enterocyte 16. Castillo A, Eguiarte LE, Souza V. 2005. A ge-
effacement conserved among diverse enterobac- nomic population genetics analysis of the patho-
terial pathogens. Proc Natl Acad Sci USA 92:1664 genic enterocyte effacement island in Escherichia
1668. coli: the search for the unit of selection. Proc Natl
6. Jarvis KG, Girn JA, Jerse AE, McDaniel TK, Acad Sci USA 102:15421547.
Donnenberg MS, Kaper JB. 1995. Enteropatho- 17. Mller D, Benz I, Liebchen A, Gallitz I, Karch
genic Escherichia coli contains a putative type H, Schmidt MA. 2009. Comparative analysis
III secretion system necessary for the export of the locus of enterocyte effacement and its
of proteins involved in attaching and effacing le- flanking regions. Infect Immun 77:35013513.
sion formation. Proc Natl Acad Sci USA 92:7996 18. Ogura Y, Ooka T, Iguchi A, Toh H, Asadulghani
8000. M, Oshima K, Kodama T, Abe H, Nakayama K,
7. Elliott SJ, Wainwright LA, McDaniel TK, Kurokawa K, Tobe T, Hattori M, Hayashi T.
Jarvis KG, Deng YK, Lai LC, McNamara BP, 2009. Comparative genomics reveal the mecha-
Donnenberg MS, Kaper JB. 1998. The complete nism of the parallel evolution of O157 and non-
sequence of the locus of enterocyte effacement O157 enterohemorrhagic Escherichia coli. Proc
(LEE) from enteropathogenic Escherichia coli Natl Acad Sci USA 106:1793917944.
E2348/69. Mol Microbiol 28:14. 19. Schmidt MA. 2010. LEEways: tales of EPEC,
8. Ebel F, Deibel C, Kresse AU, Guzman CA, ATEC and EHEC. Cell Microbiol 12:15441552.
Chakraborty T. 1996. Temperature- and medium- 20. Buss C, Mller D, Rter C, Heusipp G, Schmidt
dependent secretion of proteins by Shiga toxin- MA. 2009. Identification and characterization
producing Escherichia coli. Infect Immun 64:4472 of Ibe, a novel type III effector protein of A/E
4479. pathogens targeting human IQGAP1. Cell Micro-
9. Jarvis KG, Kaper JB. 1996. Secretion of extra- biol 11:661677.
cellular proteins by enterohemorrhagic Esche- 21. Konczy P, Ziebell K, Mascarenhas M, Choi
richia coli via a putative type III secretion system. A, Michaud C, Kropinski AM, Whittam TS,
Infect Immun 64:48264829. Wickham M, Finlay B, Karmali MA. 2008.
10. Perna NT, Mayhew GF, Posfai G, Elliott S, Genomic O island 122, locus for enterocyte efface-
Donnenberg MS, Kaper JB, Blattner FR. ment, and the evolution of virulent verocytotoxin-
1998. Molecular evolution of a pathogenicity is- producing Escherichia coli. J Bacteriol 190:5832
land from enterohemorrhagic Escherichia coli 5840.
O157:H7. Infect Immun 66:38103817. 22. Deng W, Yu HB, de Hoog CL, Stoynov N,
11. Deng W, Li Y, Vallance BA, Finlay BB. 2001. Li Y, Foster LJ, Finlay BB. 2012. Quantitative
Locus of enterocyte effacement from Citrobacter proteomic analysis of type III secretome of en-
rodentium: sequence analysis and evidence for teropathogenic Escherichia coli reveals an ex-
horizontal transfer among attaching and effacing panded effector repertoire for attaching/effacing
pathogens. Infect Immun 69:63236335. bacterial pathogens. Mol Cell Proteomics 11:692
12. Tauschek M, Strugnell RA, Robins-Browne 709.
RM. 2002. Characterization and evidence of 23. Karmali MA, Mascarenhas M, Shen S, Ziebell
mobilization of the LEE pathogenicity island of K, Johnson S, Reid-Smith R, Isaac-Renton J,
rabbit-specific strains of enteropathogenic Esch- Clark C, Rahn K, Kaper JB. 2003. Association of
erichia coli. Mol Microbiol 44:15331550. genomic O island 122 of Escherichia coli EDL 933
13. Zhu C, Agin TS, Elliott SJ, Johnson LA, Thate with verocytotoxin-producing Escherichia coli
TE, Kaper JB, Boedeker EC. 2001. Complete seropathotypes that are linked to epidemic and/or
nucleotide sequence and analysis of the locus of serious disease. J Clin Microbiol 41:49304940.
enterocyte effacement from rabbit diarrheagenic 24. Jores J, Wagner S, Rumer L, Eichberg J,
Escherichia coli RDEC-1. Infect Immun 69:2107 Laturnus C, Kirsch P, Schierack P, Tschpe
2115. H, Wieler LH. 2005. Description of a 111-kb
14. McDaniel TK, Kaper JB. 1997. A cloned patho- pathogenicity island (PAI) encoding various vir-
genicity island from enteropathogenic Escherichia ulence features in the enterohemorrhagic E. coli

(EHEC) strain RW1374 (O103:H2) and detec- pathogenicity island. Proc Natl Acad Sci USA
tion of a similar PAI in other EHEC strains of 101:35973602.
serotype O103:H2. Int J Med Microbiol 294:417 35. Mills E, Baruch K, Charpentier X, Kobi S,
425. Rosenshine I. 2008. Real-time analysis of effector
25. Tobe T, Beatson SA, Taniguchi H, Abe H, Bailey translocation by the type III secretion system of
CM, Fivian A, Younis R, Matthews S, Marches enteropathogenic Escherichia coli. Cell Host Mi-
O, Frankel G, Hayashi T, Pallen MJ. 2006. An crobe 3:104113.
extensive repertoire of type III secretion effectors 36. Berger CN, Crepin VF, Baruch K, Mousnier A,
in Escherichia coli O157 and the role of lambdoid Rosenshine I, Frankel G. 2012. EspZ of entero-
phages in their dissemination. Proc Natl Acad Sci pathogenic and enterohemorrhagic Escherichia
USA 103:1494114946. coli regulates type III secretion system protein
26. Brady MJ, Radhakrishnan P, Liu H, Magoun L, translocation. MBio 3:e00317-12.
Murphy KC, Mukherjee J, Donohue-Rolfe A, 37. Shames SR, Deng W, Guttman JA, de Hoog CL,
Tzipori S, Leong JM. 2011. Enhanced actin Li Y, Hardwidge PR, Sham HP, Vallance BA,
pedestal formation by enterohemorrhagic Esche- Foster LJ, Finlay BB. 2010. The pathogenic E.
richia coli O157:H7 adapted to the mammalian coli type III effector EspZ interacts with host
host. Front Microbiol 2:226. CD98 and facilitates host cell prosurvival signal-
27. Hughes DT, Sperandio V. 2008. Inter-kingdom ling. Cell Microbiol 12:13221339.
signalling: communication between bacteria and 38. Tatsuno I, Kimura H, Okutani A, Kanamaru K,
their hosts. Nat Rev Microbiol 6:111120. Abe H, Nagai S, Makino K, Shinagawa H,
28. Xu X, McAteer SP, Tree JJ, Shaw DJ, Wolfson Yoshida M, Sato K, Nakamoto J, Tobe T,
EB, Beatson SA, Roe AJ, Allison LJ, Chase- Sasakawa C. 2000. Isolation and characterization
Topping ME, Mahajan A, Tozzoli R, Woolhouse of mini-Tn5Km2 insertion mutants of entero-
ME, Morabito S, Gally DL. 2012. Lysogeny with hemorrhagic Escherichia coli O157:H7 deficient in
Shiga toxin 2-encoding bacteriophages represses adherence to Caco-2 cells. Infect Immun 68:5943
type III secretion in enterohemorrhagic Esche- 5952.
richia coli. PLoS Pathog 8:e1002672. 39. Dziva F, van Diemen PM, Stevens MP, Smith
29. Bttner D. 2012. Protein export according to AJ, Wallis TS. 2004. Identification of Escherichia
schedule: architecture, assembly, and regulation coli O157:H7 genes influencing colonization of
of type III secretion systems from plant- and the bovine gastrointestinal tract using signature-
animal-pathogenic bacteria. Microbiol Mol Biol tagged mutagenesis. Microbiology 150:36313645.
Rev 76:262310. 40. van Diemen PM, Dziva F, Stevens MP, Wallis
30. Pallen MJ, Beatson SA, Bailey CM. 2005. Bio- TS. 2005. Identification of enterohemorrhagic
informatics analysis of the locus for enterocyte Escherichia coli O26:H genes required for intes-
effacement provides novel insights into type-III tinal colonization in calves. Infect Immun 73:
secretion. BMC Microbiol 5:9. 17351743.
31. Creasey EA, Delahay RM, Daniell SJ, Frankel 41. Eckert SE, Dziva F, Chaudhuri RR, Langridge
G. 2003. Yeast two-hybrid system survey of in- GC, Turner DJ, Pickard DJ, Maskell DJ,
teractions between LEE-encoded proteins of en- Thomson NR, Stevens MP. 2011. Retrospective
teropathogenic Escherichia coli. Microbiology application of transposon-directed insertion site
149:20932106. sequencing to a library of signature-tagged mini-
32. Ogino T, Ohno R, Sekiya K, Kuwae A, Tn5Km2 mutants of Escherichia coli O157:H7
Matsuzawa T, Nonaka T, Fukuda H, Imajoh- screened in cattle. J Bacteriol 193:17711776.
Ohmi S, Abe A. 2006. Assembly of the type III 42. Ritchie JM, Waldor MK. 2005. The locus of
secretion apparatus of enteropathogenic Esche- enterocyte effacement-encoded effector proteins
richia coli. J Bacteriol 188:28012811. all promote enterohemorrhagic Escherichia coli
33. Sekiya K, Ohishi M, Ogino M, Tamano K, pathogenicity in infant rabbits. Infect Immun
Sasakawa C, Abe A. 2001. Supermolecular struc- 73:14661474.
ture of the enteropathogenic Escherichia coli type 43. Gauthier A, Robertson ML, Lowden M, Ibarra
III secretion system and its direct interaction JA, Puente JL, Finlay BB. 2005. Transcriptional
with the EspA-sheath-like structure. Proc Natl inhibitor of virulence factors in enteropathogenic
Acad Sci USA 98:1163811643. Escherichia coli. Antimicrob Agents Chemother
34. Deng W, Puente JL, Gruenheid S, Li Y, Vallance 49:41014109.
BA, Vzquez A, Barba J, Ibarra JA, ODonnell 44. Rasko DA, Moreira CG, Li de R, Reading NC,
P, Metalnikov P, Ashman K, Lee S, Goode D, Ritchie JM, Waldor MK, Williams N, Taussig R,
Pawson T, Finlay BB. 2004. Dissecting viru- Wei S, Roth M, Hughes DT, Huntley JF, Fina
lence: systematic and functional analyses of a MW, Falck JR, Sperandio V. 2008. Targeting

QseC signaling and virulence for antibiotic de- CesAB is an enteropathogenic Escherichia
velopment. Science 321:10781080. coli chaperone for the type-III translocator pro-
45. Tree JJ, Wang D, McInally C, Mahajan A, teins EspA and EspB. Microbiology 149:3639
Layton A, Houghton I, Elofsson M, Stevens MP, 3647.
Gally DL, Roe AJ. 2009. Characterization of the 56. Wilson RK, Shaw RK, Daniell S, Knutton S,
effects of salicylidene acylhydrazide compounds Frankel G. 2001. Role of EscF, a putative needle
on type III secretion in Escherichia coli O157:H7. complex protein, in the type III protein translo-
Infect Immun 77:42094220. cation system of enteropathogenic Escherichia
46. Veenendaal AK, Sundin C, Blocker AJ. 2009. coli. Cell Microbiol 3:753762.
Small-molecule type III secretion system inhibi- 57. Ide T, Laarmann S, Greune L, Schillers H,
tors block assembly of the Shigella type III Oberleithner H, Schmidt MA. 2001. Character-
secreton. J Bacteriol 191:563570. ization of translocation pores inserted into plasma
47. Wang D, Zetterstrm CE, Gabrielsen M, membranes by type III-secreted Esp proteins of
Beckham KS, Tree JJ, Macdonald SE, Byron O, enteropathogenic Escherichia coli. Cell Microbiol
Mitchell TJ, Gally DL, Herzyk P, Mahajan A, 3:669679.
Uvell H, Burchmore R, Smith BO, Elofsson M, 58. Kresse AU, Rohde M, Guzman CA. 1999. The
Roe AJ. 2011. Identification of bacterial target EspD protein of enterohemorrhagic Escherichia
proteins for the salicylidene acylhydrazide class coli is required for the formation of bacterial
of virulence-blocking compounds. J Biol Chem surface appendages and is incorporated in the
286:2992229931. cytoplasmic membranes of target cells. Infect
48. Ebel F, Podzadel T, Rohde M, Kresse AU, Immun 67:48344842.
Kramer S, Deibel C, Guzman CA, Chakraborty 59. Wachter C, Beinke C, Mattes M, Schmidt MA.
T. 1998. Initial binding of Shiga toxin-producing 1999. Insertion of EspD into epithelial target cell
Escherichia coli to host cells and subsequent in- membranes by infecting enteropathogenic Esch-
duction of actin rearrangements depend on fila- erichia coli. Mol Microbiol 31:16951707.
mentous EspA-containing surface appendages. 60. Warawa J, Finlay BB, Kenny B. 1999. Type III
Mol Microbiol 30:147161. secretion-dependent hemolytic activity of en-
49. Knutton S, Rosenshine I, Pallen MJ, Nisan I, teropathogenic Escherichia coli. Infect Immun
Neves BC, Bain C, Wolff C, Dougan G, Frankel 67:55385540.
G. 1998. A novel EspA-associated surface organ- 61. Shaw RK, Daniell S, Ebel F, Frankel G, Knutton
elle of enteropathogenic Escherichia coli involved S. 2001. EspA filament-mediated protein translo-
in protein translocation into epithelial cells. cation into red blood cells. Cell Microbiol 3:213
EMBO J. 17:21662176. 222.
50. Daniell SJ, Takahashi N, Wilson R, Friedberg 62. Luo W, Donnenberg MS. 2011. Interactions and
D, Rosenshine I, Booy FP, Shaw RK, Knutton predicted host membrane topology of the en-
S, Frankel G, Aizawa S. 2001. The filamentous teropathogenic Escherichia coli translocator pro-
type III secretion translocon of enteropathogenic tein EspB. J Bacteriol 193:29722980.
Escherichia coli. Cell Microbiol 3:865871. 63. Neves BC, Mundy R, Petrovska L, Dougan G,
51. Kenny B, DeVinney R, Stein M, Reinscheid DJ, Knutton S, Frankel G. 2003. CesD2 of entero-
Frey EA, Finlay BB. 1997. Enteropathogenic pathogenic Escherichia coli is a second chaperone
E. coli (EPEC) transfers its receptor for inti- for the type III secretion translocator protein
mate adherence into mammalian cells. Cell 91: EspD. Infect Immun 71:21302141.
511520. 64. Hartland EL, Daniell SJ, Delahay RM, Neves
52. Wolff C, Nisan I, Hanski E, Frankel G, BC, Wallis T, Shaw RK, Hale C, Knutton S,
Rosenshine I. 1998. Protein translocation into Frankel G. 2000. The type III protein translo-
host epithelial cells by infecting enteropathogenic cation system of enteropathogenic Escherichia coli
Escherichia coli. Mol Microbiol 28:143155. involves EspA-EspB protein interactions. Mol
53. Daniell SJ, Kocsis E, Morris E, Knutton S, Booy Microbiol 35:14831492.
FP, Frankel G. 2003. 3D structure of EspA fila- 65. Cleary J, Lai LC, Shaw RK, Straatman-
ments from enteropathogenic Escherichia coli. Iwanowska A, Donnenberg MS, Frankel G,
Mol Microbiol 49:301308. Knutton S. 2004. Enteropathogenic Escherichia
54. Crepin VF, Shaw R, Abe CM, Knutton S, coli (EPEC) adhesion to intestinal epithelial cells:
Frankel G. 2005. Polarity of enteropathogenic role of bundle-forming pili (BFP), EspA filaments
Escherichia coli EspA filament assembly and and intimin. Microbiology 150:527538.
protein secretion. J Bacteriol 187:28812889. 66. Roe AJ, Yull H, Naylor SW, Woodward MJ,
55. Creasey EA, Friedberg D, Shaw RK, Umanski T, Smith DG, Gally DL. 2003. Heterogeneous sur-
Knutton K, Rosenshine I, Frankel G. 2003. face expression of EspA translocon filaments

by Escherichia coli O157:H7 is controlled at the 77. Hauf N, Chakraborty T. 2003. Suppression
posttranscriptional level. Infect Immun 71:5900 of NF-kappa B activation and proinflammatory
5909. cytokine expression by Shiga toxin-producing
67. Roe AJ, Naylor SW, Spears KJ, Yull HM, Escherichia coli. J Immunol 170:20742082.
Dransfield TA, Oxford M, McKendrick IJ, 78. Donnenberg MS, Tzipori S, McKee ML,
Porter M, Woodward MJ, Smith DG, Gally OBrien AD, Alroy J, Kaper JB. 1993. The role of
DL. 2004. Co-ordinate single-cell expression of the eae gene of enterohemorrhagic Escherichia
LEE4- and LEE5-encoded proteins of Escherichia coli in intimate attachment in vitro and in a por-
coli O157:H7. Mol Microbiol 54:337352. cine model. J Clin Investig 92:14181424.
68. Naylor SW, Roe AJ, Nart P, Spears K, Smith 79. Yu J, Kaper JB. 1992. Cloning and character-
DG, Low JC, Gally DL. 2005. Escherichia coli ization of the eae gene of enterohaemorrhagic
O157:H7 forms attaching and effacing lesions at Escherichia coli O157:H7. Mol Microbiol 6:411
the terminal rectum of cattle and colonization re- 417.
quires the LEE4 operon. Microbiology 151:2773 80. Cornick NA, Booher SL, Moon HW. 2002.
2781. Intimin facilitates colonization by Escherichia
69. Nagano K, Taguchi K, Hara T, Yokoyama S, coli O157:H7 in adult ruminants. Infect Immun
Kawada K, Mori H. 2003. Adhesion and coloni- 70:27042707.
zation of enterohemorrhagic Escherichia coli 81. Dean-Nystrom EA, Bosworth BT, Moon HW,
O157:H7 in cecum of mice. Microbiol Immunol OBrien AD. 1998. Escherichia coli O157:H7 re-
47:125132. quires intimin for enteropathogenicity in calves.
70. McNeilly TN, Mitchell MC, Rosser T, McAteer Infect Immun 66:45604563.
S, Low JC, Smith DG, Huntley JF, Mahajan A, 82. Judge NA, Mason HS, OBrien AD. 2004. Plant
Gally DL. 2010. Immunization of cattle with a cell-based intimin vaccine given orally to mice
combination of purified intimin-531, EspA and primed with intimin reduces time of Escherichia
Tir significantly reduces shedding of Escherichia coli O157:H7 shedding in feces. Infect Immun
coli O157:H7 following oral challenge. Vaccine 72:168175.
28:14221428. 83. Ritchie JM, Thorpe CM, Rogers AB, Waldor
71. Potter AA, Klashinsky S, Li Y, Frey E, MK. 2003. Critical roles for stx2, eae, and tir in
Townsend H, Rogan D, Erickson G, Hinkley S, enterohemorrhagic Escherichia coli-induced di-
Klopfenstein T, Moxley RA, Smith DR, Finlay arrhea and intestinal inflammation in infant
BB. 2004. Decreased shedding of Escherichia coli rabbits. Infect Immun 71:71297139.
O157:H7 by cattle following vaccination with type 84. Vlisidou I, Dziva F, La Ragione RM, Best A,
III secreted proteins. Vaccine 22:362369. Garmendia J, Hawes P, Monaghan P, Cawthraw
72. Neves BC, Shaw RK, Frankel G, Knutton S. SA, Frankel G, Woodward MJ, Stevens MP.
2003. Polymorphisms within EspA filaments of 2006. Role of intimin-Tir interactions and the Tir-
enteropathogenic and enterohemorrhagic Esche- cytoskeleton coupling protein in the colonization
richia coli. Infect Immun 71:22622265. of calves and lambs by Escherichia coli O157:H7.
73. Asper DJ, Sekirov I, Finlay BB, Rogan D, Potter Infect Immun 74:758764.
AA. 2007. Cross reactivity of enterohemorrhagic 85. Higgins LM, Frankel G, Connerton I, Gonalves
Escherichia coli O157:H7-specific sera with non- NS, Dougan G, MacDonald TT. 1999. Role of
O157 serotypes. Vaccine 25:82628269. bacterial intimin in colonic hyperplasia and in-
74. Taylor KA, OConnell CB, Luther PW, flammation. Science 285:588591.
Donnenberg MS. 1998. The EspB protein of en- 86. Gonalves NS, Hale C, Dougan G, Frankel G,
teropathogenic Escherichia coli is targeted to the MacDonald TT. 2003. Binding of intimin from
cytoplasm of infected HeLa cells. Infect Immun enteropathogenic Escherichia coli to lymphocytes
66:55015507. and its functional consequences. Infect Immun
75. Hamada D, Hamaguchi M, Suzuki KN, Sakata I, 71:29602965.
Yanagihara I. 2010. Cytoskeleton-modulating ef- 87. Khare S, Alali W, Zhang S, Hunter D, Pugh R,
fectors of enteropathogenic and enterohemorrhagic Fang FC, Libby SJ, Adams LG. 2010. Vaccination
Escherichia coli: a case for EspB as an intrinsically with attenuated Salmonella enterica Dublin ex-
less-ordered effector. FEBS J 277:24092415. pressing E. coli O157:H7 outer membrane protein
76. Iizumi Y, Sagara H, Kabe Y, Azuma M, Kume K, intimin induces transient reduction of fecal
Ogawa M, Nagai T, Gillespie PG, Sasakawa C, shedding of E. coli O157:H7 in cattle. BMC Vet Res
Handa H. 2007. The enteropathogenic E. coli 6:35.
effector EspB facilitates microvillus effacing and 88. Oliveira AF, Cardoso SA, Almeida FB, de
antiphagocytosis by inhibiting myosin function. Oliveira LL, Pitondo-Silva A, Soares SG, Hanna
Cell Host Microbe 2:383392. ES. 2012. Oral immunization with attenuated

Salmonella vaccine expressing Escherichia coli which is translocated to the host cell membrane
O157:H7 intimin gamma triggers both systemic but is not tyrosine phosphorylated. Infect Immun
and mucosal humoral immunity in mice. Micro- 67:23892398.
biol Immunol 56:513522. 99. Batchelor M, Prasannan S, Daniell S, Reece S,
89. Dean-Nystrom EA, Gansheroff LJ, Mills M, Connerton I, Bloomberg G, Dougan G, Frankel
Moon HW, OBrien AD. 2002. Vaccination of G, Matthews S. 2000. Structural basis for rec-
pregnant dams with intimin(O157) protects suck- ognition of the translocated intimin receptor (Tir)
ling piglets from Escherichia coli O157:H7 infec- by intimin from enteropathogenic Escherichia
tion. Infect Immun 70:24142418. coli. EMBO J 19:24522464.
90. Gansheroff LJ, Wachtel MR, OBrien AD. 100. Kelly G, Prasannan S, Daniell S, Fleming K,
1999. Decreased adherence of enterohemorrhagic Frankel G, Dougan G, Connerton I, Matthews
Escherichia coli to HEp-2 cells in the presence of S. 1999. Structure of the cell-adhesion fragment of
antibodies that recognize the C-terminal region of intimin from enteropathogenic Escherichia coli.
intimin. Infect Immun 67:64096417. Nat Struct Biol 6:313318.
91. Frankel G, Candy DC, Everest P, Dougan G. 101. Luo Y, Frey EA, Pfuetzner RA, Creagh AL,
1994. Characterization of the C-terminal domains Knoechel DG, Haynes CA, Finlay BB, Strynadka
of intimin-like proteins of enteropathogenic and NC. 2000. Crystal structure of enteropathogenic
enterohemorrhagic Escherichia coli, Citrobacter Escherichia coli intimin-receptor complex. Nature
freundii, and Hafnia alvei. Infect Immun 62:1835 405:10731077.
1842. 102. Kenny B. 1999. Phosphorylation of tyrosine 474
92. Deibel C, Dersch P, Ebel F. 2001. Intimin from of the enteropathogenic Escherichia coli (EPEC)
Shiga toxin-producing Escherichia coli and its Tir receptor molecule is essential for actin nu-
isolated C-terminal domain exhibit different cleating activity and is preceded by additional
binding properties for Tir and a eukaryotic sur- host modifications. Mol Microbiol 31:12291241.
face receptor. Int J Med Microbiol 290:683691. 103. de Grado M, Abe A, Gauthier A, Steele-
93. McKee ML, OBrien AD. 1996. Truncated Mortimer O, DeVinney R, Finlay BB. 1999.
enterohemorrhagic Escherichia coli (EHEC) O157: Identification of the intimin-binding domain of
H7 intimin (EaeA) fusion proteins promote ad- Tir of enteropathogenic Escherichia coli. Cell
herence of EHEC strains to HEp-2 cells. Infect Microbiol 1:717.
Immun 64:22252233. 104. Hartland EL, Batchelor M, Delahay RM,
94. Liu H, Magoun L, Leong JM. 1999. beta1-chain Hale C, Matthews S, Dougan G, Knutton S,
integrins are not essential for intimin-mediated Connerton I, Frankel G. 1999. Binding of intimin
host cell attachment and enteropathogenic Esch- from enteropathogenic Escherichia coli to Tir and
erichia coli-induced actin condensation. Infect to host cells. Mol Microbiol 32:151158.
Immun 67:20452049. 105. Yi Y, Ma Y, Gao F, Mao X, Peng H, Feng Y,
95. Rosenshine I, Ruschkowski S, Stein M, Fan Z, Wang G, Guo G, Yan J, Zeng H, Zou Q,
Reinscheid DJ, Mills SD, Finlay BB. 1996. A Gao GF. 2010. Crystal structure of EHEC intimin:
pathogenic bacterium triggers epithelial signals insights into the complementarity between EPEC
to form a functional bacterial receptor that me- and EHEC. PLoS One 5:e15285.
diates actin pseudopod formation. EMBO J 15: 106. Liu H, Radhakrishnan P, Magoun L, Prabu M,
26132624. Campellone KG, Savage P, He F, Schiffer CA,
96. Rosenshine I, Donnenberg MS, Kaper JB, Leong JM. 2002. Point mutants of EHEC intimin
Finlay BB. 1992. Signal transduction between that diminish Tir recognition and actin pedestal
enteropathogenic Escherichia coli (EPEC) and formation highlight a putative Tir binding pocket.
epithelial cells: EPEC induces tyrosine phospho- Mol Microbiol 45:15571573.
rylation of host cell proteins to initiate cytoskel- 107. Adu-Bobie J, Frankel G, Bain C, Gonalves AG,
etal rearrangement and bacterial uptake. EMBO J Trabulsi LR, Douce G, Knutton S, Dougan G.
11:35513560. 1998. Detection of intimins alpha, beta, gamma,
97. Deibel C, Kramer S, Chakraborty T, Ebel F. and delta, four intimin derivatives expressed by
1998. EspE, a novel secreted protein of attaching attaching and effacing microbial pathogens. J Clin
and effacing bacteria, is directly translocated into Microbiol 36:662668.
infected host cells, where it appears as a tyrosine- 108. Oswald E, Schmidt H, Morabito S, Karch H,
phosphorylated 90 kDa protein. Mol Microbiol Marchs O, Caprioli A. 2000. Typing of intimin
28:463474. genes in human and animal enterohemorrhagic
98. DeVinney R, Stein M, Reinscheid D, Abe A, and enteropathogenic Escherichia coli: character-
Ruschkowski S, Finlay BB. 1999. Enterohemor- ization of a new intimin variant. Infect Immun
rhagic Escherichia coli O157:H7 produces Tir, 68:6471.

109. Zhang WL, Kohler B, Oswald E, Beutin L, cells appears to be mediated by mechanisms dis-
Karch H, Morabito S, Caprioli A, Suerbaum S, tinct from those used by O157. Foodborne Pathog
Schmidt H. 2002. Genetic diversity of intimin Dis 10:375381.
genes of attaching and effacing Escherichia coli 120. Kudva IT, Griffin RW, Krastins B, Sarracino
strains. J Clin Microbiol 40:44864492. DA, Calderwood SB, John M. 2012. Proteins
110. Tzipori S, Gunzer F, Donnenberg MS, de other than the locus of enterocyte effacement-
Montigny L, Kaper JB, Donohue-Rolfe A. 1995. encoded proteins contribute to Escherichia coli
The role of the eaeA gene in diarrhea and neu- O157:H7 adherence to bovine rectoanal junction
rological complications in a gnotobiotic piglet stratified squamous epithelial cells. BMC Micro-
model of enterohemorrhagic Escherichia coli in- biol 12:103.
fection. Infect Immun 63:36213627. 121. Sinclair JF, OBrien AD. 2002. Cell surface-
111. Mallick EM, Brady MJ, Luperchio SA, Vanguri localized nucleolin is a eukaryotic receptor for the
VK, Magoun L, Liu H, Sheppard BJ, Mukherjee adhesin intimin-gamma of enterohemorrhagic
J, Donohue-Rolfe A, Tzipori S, Leong JM, Escherichia coli O157:H7. J Biol Chem 277:2876
Schauer DB. 2012. Allele- and Tir-independent 2885.
functions of intimin in diverse animal infection 122. Sinclair JF, Dean-Nystrom EA, OBrien AD.
models. Front Microbiol 3:11. 2006. The established intimin receptor Tir and the
112. Phillips AD, Navabpour S, Hicks S, Dougan G, putative eucaryotic intimin receptors nucleolin
Wallis T, Frankel G. 2000. Enterohaemorrhagic and beta1 integrin localize at or near the site of
Escherichia coli O157:H7 target Peyers patches enterohemorrhagic Escherichia coli O157:H7 ad-
in humans and cause attaching/effacing lesions herence to enterocytes in vivo. Infect Immun
in both human and bovine intestine. Gut 47:377 74:12551265.
381. 123. Sinclair JF, OBrien AD. 2004. Intimin types
113. Phillips AD, Frankel G. 2000. Intimin-mediated alpha, beta, and gamma bind to nucleolin with
tissue specificity in enteropathogenic Escherichia equivalent affinity but lower avidity than to the
coli interaction with human intestinal organ translocated intimin receptor. J Biol Chem
cultures. J Infect Dis 181:14961500. 279:3375133758.
114. Fitzhenry RJ, Pickard DJ, Hartland EL, Reece 124. Frankel G, Lider O, Hershkoviz R, Mould
S, Dougan G, Phillips AD, Frankel G. 2002. AP, Kachalsky SG, Candy DCA, Cahalon L,
Intimin type influences the site of human intes- Humphries MJ, Dougan G. 1996. The cell-
tinal mucosal colonisation by enterohaemorrhagic binding domain of intimin from enteropathogenic
Escherichia coli O157:H7. Gut 50:180185. Escherichia coli binds to beta1 integrins. J Biol
115. Girard F, Batisson I, Frankel GM, Harel J, Chem 271:2035920364.
Fairbrother JM. 2005. Interaction of entero- 125. Muza-Moons MM, Koutsouris A, Hecht G.
pathogenic and Shiga toxin-producing Esche- 2003. Disruption of cell polarity by enteropatho-
richia coli and porcine intestinal mucosa: role of genic Escherichia coli enables basolateral mem-
intimin and Tir in adherence. Infect Immun 73: brane proteins to migrate apically and to potentiate
60056016. physiological consequences. Infect Immun 71:
116. Mundy R, Schller S, Girard F, Fairbrother JM, 70697078.
Phillips AD, Frankel G. 2007. Functional studies 126. Campellone KG, Rankin S, Pawson T,
of intimin in vivo and ex vivo: implications for Kirschner MW, Tipper DJ, Leong JM. 2004.
host specificity and tissue tropism. Microbiology Clustering of Nck by a 12-residue Tir phospho-
153:959967. peptide is sufficient to trigger localized actin as-
117. Naylor SW, Low JC, Besser TE, Mahajan A, sembly. J Cell Biol 164:407416.
Gunn GJ, Pearce MC, McKendrick IJ, Smith 127. Kenny B, Ellis S, Leard AD, Warawa J, Mellor
DG, Gally DL. 2003. Lymphoid follicle-dense H, Jepson MA. 2002. Co-ordinate regulation of
mucosa at the terminal rectum is the principal distinct host cell signalling pathways by multi-
site of colonization of enterohemorrhagic Esche- functional enteropathogenic Escherichia coli ef-
richia coli O157:H7 in the bovine host. Infect fector molecules. Mol Microbiol 44:10951107.
Immun 71:15051512. 128. Campellone KG, Giese A, Tipper DJ, Leong JM.
118. Sheng H, Lim JY, Knecht HJ, Li J, Hovde CJ. 2002. A tyrosine-phosphorylated 12-amino-acid
2006. Role of Escherichia coli O157:H7 virulence sequence of enteropathogenic Escherichia coli Tir
factors in colonization at the bovine terminal binds the host adaptor protein Nck and is re-
rectal mucosa. Infect Immun 74:46854693. quired for Nck localization to actin pedestals. Mol
119. Kudva IT, Hovde CJ, John M. 2013. Adherence Microbiol 43:12271241.
of non-O157 Shiga toxin-producing Escherichia coli 129. Gruenheid S, DeVinney R, Bladt F, Goosney F,
to bovine recto-anal junction squamous epithelial Gelkop S, Gish GD, Pawson T, Finlay BB. 2001.

Enteropathogenic E. coli Tir binds Nck to initiate 139. Campellone KG, Leong JM. 2005. Nck-
actin pedestal formation in host cells. Nat Cell independent actin assembly is mediated by two
Biol 3:856859. phosphorylated tyrosines within enteropatho-
130. Rohatgi R, Nollau P, Ho H, Kirschner MW, genic Escherichia coli Tir. Mol Microbiol 56:416
Mayer BJ. 2001. Nck and phosphatidylinositol 432.
4,5-bisphosphate synergistically activate actin 140. Campellone KG, Robbins D, Leong JM. 2004.
polymerization through the N-WASP-Arp2/3 EspFU is a translocated EHEC effector that
pathway. J Biol Chem 276:2644826452. interacts with Tir and N-WASP and promotes
131. Lommel S, Benesch S, Rottner K, Franz T, Nck-independent actin assembly. Dev Cell 7:217
Wehland J, Kuhn R. 2001. Actin pedestal 228.
formation by enteropathogenic Escherichia coli 141. Garmendia J, Phillips AD, Carlier MF, Chong Y,
and intracellular motility of Shigellaflexneri are Schller S, Marches O, Dahan S, Oswald E,
abolished in N-WASP-defective cells. EMBO Rep Shaw RK, Knutton S, Frankel G. 2004. TccP is
2:850857. an enterohaemorrhagic Escherichia coli O157:H7
132. Garber JJ, Takeshima F, Antn IM, Oyoshi type III effector protein that couples Tir to the
MK, Lyubimova A, Kapoor A, Shibata T, Chen actin-cytoskeleton. Cell Microbiol 6:11671183.
F, Alt FW, Geha RS, Leong JM, Snapper SB. 142. Cheng HC, Skehan BM, Campellone KG,
2012. Enteropathogenic Escherichia coli and vac- Leong JM, Rosen MK. 2008. Structural mecha-
cinia virus do not require the family of WASP- nism of WASP activation by the enterohaemor-
interacting proteins for pathogen-induced actin rhagic E. coli effector EspF(U). Nature 454:1009
assembly. Infect Immun 80:40714077. 1013.
133. Lommel S, Benesch S, Rohde M, Wehland J, 143. Sallee NA, Rivera GM, Dueber JE, Vasilescu D,
Rottner K. 2004. Enterohaemorrhagic and en- Mullins RD, Mayer BJ, Lim WA. 2008. The
teropathogenic Escherichia coli use different pathogen protein EspF(U) hijacks actin poly-
mechanisms for actin pedestal formation that merization using mimicry and multivalency. Na-
converge on N-WASP. Cell Microbiol 6:243254. ture 454:10051008.
134. DeVinney R, Puente JL, Gauthier A, Goosney 144. Campellone KG, Cheng HC, Robbins D, Siripala
D, Finlay BB. 2001. Enterohaemorrhagic and AD, McGhie EJ, Hayward RD, Welch MD,
enteropathogenic Escherichia coli use a different Rosen MK, Koronakis V, Leong JM. 2008. Re-
Tir-based mechanism for pedestal formation. Mol petitive N-WASP-binding elements of the entero-
Microbiol 41:14451458. hemorrhagic Escherichia coli effector EspF(U)
135. Brady MJ, Campellone KG, Ghildiyal M, Leong synergistically activate actin assembly. PLoS
JM. 2007. Enterohaemorrhagic and entero- Pathog 4:e1000191.
pathogenic Escherichia coli Tir proteins trigger a 145. Garmendia J, Carlier MF, Egile C, Didry D,
common Nck-independent actin assembly path- Frankel G. 2006. Characterization of TccP-
way. Cell Microbiol 9:22422253. mediated N-WASP activation during entero-
136. Campellone KG, Brady MJ, Alamares JG, Rowe haemorrhagic Escherichia coli infection. Cell
DC, Skehan BM, Tipper DJ, Leong JM. 2006. Microbiol 8:14441455.
Enterohaemorrhagic Escherichia coli Tir requires 146. Vingadassalom D, Campellone KG, Brady MJ,
a C-terminal 12-residue peptide to initiate EspF- Skehan B, Battle SE, Robbins D, Kapoor A,
mediated actin assembly and harbours N-terminal Hecht G, Snapper SB, Leong JM. 2010. Entero-
sequences that influence pedestal length. Cell hemorrhagic E. coli requires N-WASP for
Microbiol 8:14881503. efficient type III translocation but not for EspFU-
137. Vingadassalom D, Kazlauskas A, Skehan B, mediated actin pedestal formation. PLoS Pathog
Cheng HC, Magoun L, Robbins D, Rosen MK, 6:e1001056.
Saksela K, Leong JM. 2009. Insulin receptor 147. Bai L, Schller S, Whale A, Mousnier A,
tyrosine kinase substrate links the E. coli O157:H7 Marches O, Wang L, Ooka T, Heuschkel R,
actin assembly effectors Tir and EspF(U) during Torrente F, Kaper JB, Gomes TA, Xu J, Phillips
pedestal formation. Proc Natl Acad Sci USA AD, Frankel G. 2008. Enteropathogenic Esche-
106:67546759. richia coli O125:H6 triggers attaching and effacing
138. Weiss SM, Ladwein M, Schmidt D, Ehinger J, lesions on human intestinal biopsy specimens
Lommel S, Stding K, Beutling U, Disanza A, independently of Nck and TccP/TccP2. Infect
Frank R, Jnsch L, Scita G, Gunzer F, Rottner Immun 76:361368.
K, Stradal TE. 2009. IRSp53 links the entero- 148. Schller S, Chong Y, Lewin J, Kenny B, Frankel
hemorrhagic E. coli effectors Tir and EspFU for G, Phillips AD. 2007. Tir phosphorylation and
actin pedestal formation. Cell Host Microbe Nck/N-WASP recruitment by enteropathogenic
5:244258. and enterohaemorrhagic Escherichia coli during

ex vivo colonization of human intestinal mucosa 157. Wang D, Roe AJ, McAteer S, Shipston MJ,
is different to cell culture models. Cell Microbiol Gally DL. 2008. Hierarchal type III secretion of
9:13521364. translocators and effectors from Escherichia coli
149. Crepin VF, Girard F, Schller S, Phillips AD, O157:H7 requires the carboxy terminus of SepL
Mousnier A, Frankel G. 2010. Dissecting the role that binds to Tir. Mol Microbiol 69:14991512.
of the Tir:Nck and Tir:IRTKS/IRSp53 signalling 158. Charpentier X, Oswald E. 2004. Identification
pathways in vivo. Mol Microbiol 75:308323. of the secretion and translocation domain of
150. Ritchie JM, Brady MJ, Riley KN, Ho TD, the enteropathogenic and enterohemorrhagic
Campellone KG, Herman IM, Donohue-Rolfe Escherichia coli effector Cif, using TEM-1 beta-
A, Tzipori S, Waldor MK, Leong JM. 2008. lactamase as a new fluorescence-based reporter.
EspFU, a type III-translocated effector of actin J Bacteriol 186:54865495.
assembly, fosters epithelial association and late- 159. Asadulghani M, Ogura Y, Ooka T, Itoh T,
stage intestinal colonization by E. coli O157:H7. Sawaguchi A, Iguchi A, Nakayama K, Hayashi
Cell Microbiol 10:836847. T. 2009. The defective prophage pool of Esche-
151. Ogura Y, Ooka T, Whale A, Garmendia J, richia coli O157:prophage-prophage interactions
Beutin L, Tennant S, Krause G, Morabito S, potentiate horizontal transfer of virulence deter-
Chinen I, Tobe T, Abe H, Tozzoli R, Caprioli A, minants. PLoS Pathog 5:e1000408.
Rivas M, Robins-Browne R, Hayashi T, Frankel 160. Marchs O, Ledger TN, Boury M, Ohara M,
G. 2007. TccP2 of O157:H7 and non-O157 entero- Tu X, Goffaux F, Mainil J, Rosenshine I, Sugai
hemorrhagic Escherichia coli (EHEC): challenging M, De Rycke J, Oswald E. 2003. Enteropatho-
the dogma of EHEC-induced actin polymeriza- genic and enterohaemorrhagic Escherichia coli
tion. Infect Immun 75:604612. deliver a novel effector called Cif, which blocks
152. Whale AD, Hernandes RT, Ooka T, Beutin L, cell cycle G2/M transition. Mol Microbiol 50:
Schller S, Garmendia J, Crowther L, Vieira 15531567.
MA, Ogura Y, Krause G, Phillips AD, Gomes TA, 161. Taieb F, Nougayrde JP, Oswald E. 2011. Cycle
Hayashi T, Frankel G. 2007. TccP2-mediated inhibiting factors (cifs): cyclomodulins that usurp
subversion of actin dynamics by EPEC 2a dis- the ubiquitin-dependent degradation pathway of
tinct evolutionary lineage of enteropathogenic host cells. Toxins (Basel) 3:356368.
Escherichia coli. Microbiology 153:17431755. 162. Holmes A, Mhlen S, Roe AJ, Dean P. 2010. The
153. Sanger JM, Chang R, Ashton F, Kaper JB, EspF effector, a bacterial pathogens Swiss army
Sanger JW. 1996. Novel form of actin-based mo- knife. Infect Immun 78:44454453.
tility transports bacteria on the surfaces of in- 163. Marchs O, Wiles S, Dziva F, La Ragione RM,
fected cells. Cell Motil Cytoskeleton 34:279287. Schller S, Best A, Phillips AD, Hartland EL,
154. Perna NT, Plunkett G 3rd, Burland V, Mau B, Woodward MJ, Stevens MP, Frankel G. 2005.
Glasner JD, Rose DJ, Mayhew GF, Evans PS, Characterization of two non-locus of enterocyte
Gregor J, Kirkpatrick HA, Psfai G, Hackett J, effacement-encoded type III-translocated effec-
Klink S, Boutin A, Shao Y, Miller L, Grotbeck EJ, tors, NleC and NleD, in attaching and effacing
Davis NW, Lim A, Dimalanta ET, Potamousis pathogens. Infect Immun 73:84118417.
KD, Apodaca J, Anantharaman TS, Lin J, Yen G, 164. Wong AR, Pearson JS, Bright MD, Munera D,
Schwartz DC, Welch RA, Blattner FR. 2001. Robinson KS, Lee SF, Frankel G, Hartland EL.
Genome sequence of enterohaemorrhagic Esche- 2011. Enteropathogenic and enterohaemorrhagic
richia coli O157:H7. Nature 409:529533. Escherichia coli: even more subversive elements.
155. Hayashi T, Makino K, Ohnishi M, Kurokawa Mol Microbiol 80:14201438.
K, Ishii K, Yokoyama K, Han CG, Ohtsubo E, 165. Nadler C, Baruch K, Kobi S, Mills E, Haviv G,
Nakayama K, Murata T, Tanaka M, Tobe T, Farago M, Alkalay I, Bartfeld S, Meyer TF,
Iida T, Takami H, Honda T, Sasakawa C, Ben-Neriah Y, Rosenshine I. 2010. The type III
Ogasawara N, Yasunaga T, Kuhara S, Shiba T, secretion effector NleE inhibits NF-kappaB acti-
Hattori M, Shinagawa H. 2001. Complete ge- vation. PLoS Pathog 6:e1000743.
nome sequence of enterohemorrhagic Escherichia 166. Newton HJ, Pearson JS, Badea L, Kelly M,
coli O157:H7 and genomic comparison with a Lucas M, Holloway G, Wagstaff KM, Dunstone
laboratory strain K-12. DNA Res 8:1122. MA, Sloan J, Whisstock JC, Kaper JB, Robins-
156. Deng W, Li Y, Hardwidge PR, Frey EA, Browne RM, Jans DA, Frankel G, Phillips AD,
Pfuetzner RA, Lee S, Gruenheid S, Strynakda Coulson BS, Hartland EL. 2010. The type III
NC, Puente JL, Finlay BB. 2005. Regulation of effectors NleE and NleB from enteropathogenic
type III secretion hierarchy of translocators and E. coli and OspZ from Shigella block nuclear
effectors in attaching and effacing bacterial patho- translocation of NF-kappaB p65. PLoS Pathog 6:
gens. Infect Immun 73:21352146. e1000898.

167. Gao X, Wang X, Pham TH, Feuerbacher LA, 178. Royan SV, Jones RM, Koutsouris A, Roxas JL,
Lubos ML, Huang M, Olsen R, Mushegian A, Falzari K, Weflen AW, Kim A, Bellmeyer A,
Slawson C, Hardwidge PR. 2013. NleB, a bacte- Turner JR, Neish AS, Rhee KJ, Viswanathan
rial effector with glycosyltransferase activity, VK, Hecht GA. 2010. Enteropathogenic E. coli
targets GAPDH function to inhibit NF-kB activa- non-LEE encoded effectors NleH1 and NleH2
tion. Cell Host Microbe 13:8799. attenuate NF-kB activation. Mol Microbiol
168. Ruchaud-Sparagano MH, Mhlen S, Dean P, 78:12321245.
Kenny B. 2011. The enteropathogenic E. coli 179. Pham TH, Gao X, Tsai K, Olsen R, Wan F,
(EPEC) Tir effector inhibits NF-kB activity by Hardwidge PR. 2012. Functional differences and
targeting TNFa receptor-associated factors. PLoS interactions between the Escherichia coli type III
Pathog 7:e1002414. secretion system effectors NleH1 and NleH2. In-
169. Yan D, Wang X, Luo L, Cao X, Ge B. 2012. In- fect Immun 80:21332140.
hibition of TLR signaling by a bacterial protein 180. Marchs O, Covarelli V, Dahan S, Cougoule C,
containing immunoreceptor tyrosine-based in- Bhatta P, Frankel G, Caron E. 2008. EspJ of
hibitory motifs. Nat Immunol 13:10631071. enteropathogenic and enterohaemorrhagic Esch-
170. Baruch K, Gur-Arie L, Nadler C, Koby S, erichia coli inhibits opsono-phagocytosis. Cell
Yerushalmi G, Ben-Neriah Y, Yogev O, Microbiol 10:11041115.
Shaulian E, Guttman C, Zarivach R, Rosenshine 181. Dong N, Liu L, Shao F. 2010. A bacterial effector
I. 2011. Metalloprotease type III effectors that targets host DH-PH domain RhoGEFs and an-
specifically cleave JNK and NF-kB. EMBO J tagonizes macrophage phagocytosis. EMBO J 29:
30:221231. 13631376.
171. Mhlen S, Ruchaud-Sparagano MH, Kenny B. 182. Martinez-Argudo I, Sands C, Jepson MA. 2007.
2011. Proteasome-independent degradation of Translocation of enteropathogenic Escherichia
canonical NFkappaB complex components by the coli across an in vitro M cell model is regulated
NleC protein of pathogenic Escherichia coli. J Biol by its type III secretion system. Cell Microbiol
Chem 286:51005107. 9:15381546.
172. Pearson JS, Riedmaier P, Marchs O, Frankel 183. Quitard S, Dean P, Maresca M, Kenny B.
G, Hartland EL. 2011. A type III effector protease 2006. The enteropathogenic Escherichia coli EspF
NleC from enteropathogenic Escherichia coli effector molecule inhibits PI-3 kinase-mediated
targets NF-kB for degradation. Mol Microbiol uptake independently of mitochondrial targeting.
80:219230. Cell Microbiol 8:972981.
173. Yen H, Ooka T, Iguchi A, Hayashi T, Sugimoto 184. Tahoun A, Siszler G, Spears K, McAteer S,
N, Tobe T. 2010. NleC, a type III secretion pro- Tree J, Paxton E, Gillespie TL, Martinez-
tease, compromises NF-kB activation by targeting Argudo I, Jepson MA, Shaw DJ, Koegl M,
p65/RelA. PLoS Pathog 6:e1001231. Haas J, Gally DL, Mahajan A. 2011. Comparative
174. Vossenkmper A, Marchs O, Fairclough PD, analysis of EspF variants in inhibition of Esche-
Warnes G, Stagg AJ, Lindsay JO, Evans PC, richia coli phagocytosis by macrophages and
Luong le A, Croft NM, Naik S, Frankel G, inhibition of E. coli translocation through human-
MacDonald TT. 2010. Inhibition of NF-kB sig- and bovine-derived M cells. Infect Immun 79:
naling in human dendritic cells by the entero- 47164729.
pathogenic Escherichia coli effector protein NleE. 185. Celli J, Olivier M, Finlay BB. 2001. Entero-
J Immunol 185:41184127. pathogenic Escherichia coli mediates antiphago-
175. Zhang L, Ding X, Cui J, Xu H, Chen J, Gong YN, cytosis through the inhibition of PI 3-kinase-
Hu L, Zhou Y, Ge J, Lu Q, Liu L, Chen S, Shao F. dependent pathways. EMBO J 20:12451258.
2011. Cysteine methylation disrupts ubiquitin- 186. Alto NM, Weflen AW, Rardin MJ, Yarar D,
chain sensing in NF-kB activation. Nature 481: Lazar CS, Tonikian R, Koller A, Taylor SS,
204208. Boone C, Sidhu SS, Schmid SL, Hecht GA,
176. Gao X, Wan F, Mateo K, Callegari E, Wang D, Dixon JE. 2007. The type III effector EspF
Deng W, Puente J, Li F, Chaussee MS, Finlay coordinates membrane trafficking by the spatio-
BB, Lenardo MJ, Hardwidge PR. 2009. Bacterial temporal activation of two eukaryotic signaling
effector binding to ribosomal protein s3 subverts pathways. J Cell Biol 178:12651278.
NF-kappaB function. PLoS Pathog 5:e1000708. 187. Marchs O, Batchelor M, Shaw RK, Patel A,
177. Wan F, Weaver A, Gao X, Bern M, Hardwidge Cummings N, Nagai T, Sasakawa C, Carlsson
PR, Lenardo MJ. 2011. IKKb phosphorylation SR, Lundmark R, Cougoule C, Caron E,
regulates RPS3 nuclear translocation and NF-kB Knutton S, Connerton I, Frankel G. 2006. EspF
function during infection with Escherichia coli of enteropathogenic Escherichia coli binds sorting
strain O157:H7. Nat Immunol 12:335343. nexin 9. J Bacteriol 188:31103115.

188. Peralta-Ramrez J, Hernandez JM, Manning- 199. Arbeloa A, Oates CV, Marchs O, Hartland EL,
Cela R, Luna-Muoz J, Garcia-Tovar C, Frankel G. 2011. Enteropathogenic and entero-
Nougayrde JP, Oswald E, Navarro-Garcia F. hemorrhagic Escherichia coli type III secretion
2008. EspF interacts with nucleation-promoting effector EspV induces radical morphological
factors to recruit junctional proteins into pedes- changes in eukaryotic cells. Infect Immun 79:
tals for pedestal maturation and disruption of 10671076.
paracellular permeability. Infect Immun 76:3854 200. Hardwidge PR, Deng W, Vallance BA,
3868. Rodriguez-Escudero I, Cid VJ, Molina M,
189. Weflen AW, Alto NM, Viswanathan VK, Hecht Finlay BB. 2005. Modulation of host cytoskeleton
G. 2010. E. coli secreted protein F promotes function by the enteropathogenic Escherichia coli
EPEC invasion of intestinal epithelial cells via and Citrobacter rodentium effector protein EspG.
an SNX9-dependent mechanism. Cell Microbiol Infect Immun 73:25862594.
12:919929. 201. Matsuzawa T, Kuwae A, Abe A. 2005. Entero-
190. Berger CN, Crepin VF, Jepson MA, Arbeloa A, pathogenic Escherichia coli type III effectors
Frankel G. 2009. The mechanisms used by en- EspG and EspG2 alter epithelial paracellular per-
teropathogenic Escherichia coli to control filopodia meability. Infect Immun 73:62836289.
dynamics. Cell Microbiol 11:309322. 202. Matsuzawa T, Kuwae A, Yoshida S, Sasakawa
191. Huang Z, Sutton SE, Wallenfang AJ, Orchard C, Abe A. 2004. Enteropathogenic Escherichia
RC, Wu X, Feng Y, Chai J, Alto NM. 2009. coli activates the RhoA signaling pathway via the
Structural insights into host GTPase isoform se- stimulation of GEF-H1. EMBO J 23:35703582.
lection by a family of bacterial GEF mimics. Nat 203. Tomson FL, Viswanathan VK, Kanack KJ,
Struct Mol Biol 16:853860. Kanteti RP, Straub KV, Menet M, Kaper JB,
192. Arbeloa A, Bulgin RR, MacKenzie G, Shaw RK, Hecht G. 2005. Enteropathogenic Escherichia
Pallen MJ, Crepin VF, Berger CN, Frankel G. coli EspG disrupts microtubules and in conjunc-
2008. Subversion of actin dynamics by EspM tion with Orf3 enhances perturbation of the tight
effectors of attaching and effacing bacterial junction barrier. Mol Microbiol 56:447464.
pathogens. Cell Microbiol 10:14291441. 204. Shaw RK, Smollett K, Cleary J, Garmendia J,
193. Arbeloa A, Garnett J, Lillington J, Bulgin RR, Straatman-Iwanowska A, Frankel G, Knutton S.
Berger CN, Lea SM, Matthews S, Frankel G. 2005. Enteropathogenic Escherichia coli type III
2010. EspM2 is a RhoA guanine nucleotide ex- effectors EspG and EspG2 disrupt the microtu-
change factor. Cell Microbiol 12:654664. bule network of intestinal epithelial cells. Infect
194. Bulgin R, Arbeloa A, Goulding D, Dougan G, Immun 73:43854390.
Crepin VF, Raymond B, Frankel G. 2009. The 205. Clements A, Smollett K, Lee SF, Hartland EL,
T3SS effector EspT defines a new category of Lowe M, Frankel G. 2011. EspG of enteropatho-
invasive enteropathogenic E. coli (EPEC) which genic and enterohemorrhagic E. coli binds the
form intracellular actin pedestals. PLoS Pathog 5: Golgi matrix protein GM130 and disrupts the
e1000683. Golgi structure and function. Cell Microbiol
195. Keestra AM, Winter MG, Auburger JJ, Frssle 13:14291439.
SP, Xavier MN, Winter SE, Kim A, Poon V, 206. Dong N, Zhu Y, Lu Q, Hu L, Zheng Y, Shao F.
Ravesloot MM, Waldenmaier JF, Tsolis RM, 2012. Structurally distinct bacterial TBC-like
Eigenheer RA, Bumler AJ. 2013. Manipulation GAPs link Arf GTPase to Rab1 inactivation to
of small Rho GTPases is a pathogen-induced counteract host defenses. Cell 150:10291041.
process detected by NOD1. Nature 496:233237. 207. Germane KL, Spiller BW. 2011. Structural and
196. Wong AR, Clements A, Raymond B, Crepin VF, functional studies indicate that the EPEC effector,
Frankel G. 2012. The interplay between the EspG, directly binds p21-activated kinase. Bio-
Escherichia coli Rho guanine nucleotide exchange chemistry 50:917919.
factor effectors and the mammalian RhoGEF in- 208. Selyunin AS, Sutton SE, Weigele BA, Reddick
hibitor EspH. MBio 3:e00250-11. LE, Orchard RC, Bresson SM, Tomchick DR,
197. Tu X, Nisan I, Yona C, Hanski E, Rosenshine I. Alto NM. 2011. The assembly of a GTPase-kinase
2003. EspH, a new cytoskeleton-modulating signalling complex by a bacterial catalytic scaf-
effector of enterohaemorrhagic and enteropath- fold. Nature 469:107111.
ogenic Escherichia coli. Mol Microbiol 47:595606. 209. Batchelor M, Guignot J, Patel A, Cummings N,
198. Wong AR, Raymond B, Collins JW, Crepin VF, Cleary J, Knutton S, Holden DW, Connerton I,
Frankel G. 2012. The enteropathogenic E. coli Frankel G. 2004. Involvement of the intermedi-
effector EspH promotes actin pedestal formation ate filament protein cytokeratin-18 in actin ped-
and elongation via WASP-interacting protein estal formation during EPEC infection. EMBO
(WIP). Cell Microbiol 14:10511070. Rep 5:104110.

210. Patel A, Cummings N, Batchelor M, Hill PJ, NleH inhibits apoptosis induced by Clostridium
Dubois T, Mellits KH, Frankel G, Connerton I. difficile toxin B. Microbiology 156:18151823.
2006. Host protein interactions with entero- 222. Martinez E, Schroeder GN, Berger CN, Lee SF,
pathogenic Escherichia coli (EPEC): 14-3-3tau Robinson KS, Badea L, Simpson N, Hall RA,
binds Tir and has a role in EPEC-induced actin Hartland EL, Crepin VF, Frankel G. 2010.
polymerization. Cell Microbiol 8:5571. Binding to Na(+)/H(+) exchanger regulatory fac-
211. Ruetz TJ, Lin AE, Guttman JA. 2012. Entero- tor 2 (NHERF2) affects trafficking and function of
haemorrhagic Escherichia coli requires the spec- the enteropathogenic Escherichia coli type III
trin cytoskeleton for efficient attachment and secretion system effectors Map, EspI and NleH.
pedestal formation on host cells. Microb Pathog Cell Microbiol 12:17181731.
52:149156. 223. Blasche S, Mrtl M, Steuber H, Siszler G, Nisa
212. Crane JK, McNamara BP, Donnenberg MS. S, Schwarz F, Lavrik I, Gronewold TM, Maskos
2001. Role of EspF in host cell death induced by K, Donnenberg MS, Ullmann D, Uetz P, Kgl M.
enteropathogenic Escherichia coli. Cell Microbiol 2013. The E. coli effector protein NleF is a caspase
3:197211. inhibitor. PLoS One 8:e58937.
213. Nagai T, Abe A, Sasakawa C. 2005. Targeting of 224. Viswanathan VK, Hodges K, Hecht G. 2009.
enteropathogenic Escherichia coli EspF to host Enteric infection meets intestinal function: how
mitochondria is essential for bacterial pathogen- bacterial pathogens cause diarrhoea. Nat Rev
esis: critical role of the 16th leucine residue in Microbiol 7:110119.
EspF. J Biol Chem 280:29983011. 225. Dean P, Kenny B. 2004. Intestinal barrier dys-
214. Nougayrde JP, Donnenberg MS. 2004. En- function by enteropathogenic Escherichia coli is
teropathogenic Escherichia coli EspF is targeted mediated by two effector molecules and a bacte-
to mitochondria and is required to initiate the rial surface protein. Mol Microbiol 54:665675.
mitochondrial death pathway. Cell Microbiol 226. Thanabalasuriar A, Koutsouris A, Weflen A,
6:10971111. Mimee M, Hecht G, Gruenheid S. 2010. The
215. Nougayrde JP, Foster GH, Donnenberg MS. bacterial virulence factor NleA is required for the
2007. Enteropathogenic Escherichia coli effector disruption of intestinal tight junctions by entero-
EspF interacts with host protein Abcf2. Cell pathogenic Escherichia coli. Cell Microbiol 12:31
Microbiol 9:680693. 41.
216. Zhao S, Zhou Y, Wang C, Yang Y, Wu X, Wei Y, 227. Simovitch M, Sason H, Cohen S, Zahavi
Zhu L, Zhao W, Zhang Q, Wan C. 2013. The EE, Melamed-Book N, Weiss A, Aroeti B,
N-terminal domain of EspF induces host cell ap- Rosenshine I. 2010. EspM inhibits pedestal for-
optosis after infection with enterohaemorrhagic mation by enterohaemorrhagic Escherichia coli
Escherichia coli O157:H7. PLoS One 8:e55164. and enteropathogenic E. coli and disrupts the
217. Kenny B, Jepson M. 2000. Targeting of an en- architecture of a polarized epithelial monolayer.
teropathogenic Escherichia coli (EPEC) effector Cell Microbiol 12:489505.
protein to host mitochondria. Cell Microbiol 228. Hodges K, Alto NM, Ramaswamy K, Dudeja
3:417426. PK, Hecht G. 2008. The enteropathogenic Esch-
218. Samba-Louaka A, Nougayrde JP, Watrin C, erichia coli effector protein EspF decreases sodi-
Oswald E, Taieb F. 2009. The enteropathogenic um hydrogen exchanger 3 activity. Cell Microbiol
Escherichia coli effector Cif induces delayed apo- 10:17351745.
ptosis in epithelial cells. Infect Immun 77:5471 229. Dean P, Maresca M, Schller S, Phillips AD,
5477. Kenny B. 2006. Potent diarrheagenic mechanism
219. Heczko U, Carthy CM, OBrien BA, Finlay BB. mediated by the cooperative action of three
2001. Decreased apoptosis in the ileum and ileal enteropathogenic Escherichia coli-injected effec-
Peyers patches: a feature after infection with tor proteins. Proc Natl Acad Sci USA 103:1876
rabbit enteropathogenic Escherichia coli O103. 1881.
Infect Immun 69:45804589. 230. Klapproth JM, Scaletsky IC, McNamara BP,
220. Hemrajani C, Berger CN, Robinson KS, Lai LC, Malstrom C, James SP, Donnenberg
Marchs O, Mousnier A, Frankel G. 2010. NleH MS. 2000. A large toxin from pathogenic Esche-
effectors interact with Bax inhibitor-1 to block richia coli strains that inhibits lymphocyte acti-
apoptosis during enteropathogenic Escherichia vation. Infect Immun 68:21482155.
coli infection. Proc Natl Acad Sci USA 107:3129 231. Nicholls L, Grant TH, Robins-Browne RM.
3134. 2000. Identification of a novel genetic locus that
221. Robinson KS, Mousnier A, Hemrajani C, is required for in vitro adhesion of a clinical iso-
Fairweather N, Berger CN, Frankel G. 2010. late of enterohaemorrhagic Escherichia coli to
The enteropathogenic Escherichia coli effector epithelial cells. Mol Microbiol 35:275288.

232. Deacon V, Dziva F, van Diemen PM, Frankel G, Escherichia coli O157:H7 influences the exp-
Stevens MP. 2010. Efa-1/LifA mediates intestinal ression and secretion of locus of enterocyte
colonization of calves by enterohaemorrhagic effacement-encoded proteins but not intestinal
Escherichia coli O26:H- in a manner independent colonization in calves or sheep. Infect Immun
of glycosyltransferase and cysteine protease mo- 72:54025411.
tifs or effects on type III secretion. Microbiology 238. Tatsuno I, Horie M, Abe H, Miki T, Makino K,
156:25272536. Shinagawa H, Taguchi H, Kamiya S, Hayashi
233. Stevens MP, van Diemen PM, Frankel G, Phil- T, Sasakawa C. 2001. toxB gene on pO157 of
lips AD, Wallis TS. 2002. Efa1 influences colo- enterohemorrhagic Escherichia coli O157:H7 is
nization of the bovine intestine by Shiga toxin- required for full epithelial cell adherence pheno-
producing Escherichia coli serotypes O5 and O111. type. Infect Immun 69:66606669.
Infect Immun 70:51585166. 239. Abu-Median AB, van Diemen PM, Dziva F,
234. Badea L, Doughty S, Nicholls L, Sloan J, Vlisidou I, Wallis TS, Stevens MP. 2006. Func-
Robins-Browne RM, Hartland EL. 2003. Con- tional analysis of lymphostatin homologues in
tribution of Efa1/LifA to the adherence of en- enterohaemorrhagic Escherichia coli. FEMS
teropathogenic Escherichia coli to epithelial cells. Microbiol Lett 258:4349.
Microb Pathog 34:205215. 240. Stevens MP, Marchs O, Campbell J, Huter V,
235. Vidal JE, Navarro-Garca F. 2008. EspC trans- Frankel G, Phillips AD, Oswald E, Wallis TS.
location into epithelial cells by enteropathogenic 2002. Intimin, Tir, and Shiga toxin 1 do not
Escherichia coli requires a concerted participa- influence enteropathogenic responses to Shiga
tion of Type V and III secretion systems. Cell toxin-producing Escherichia coli in bovine ligated
Microbiol 10:19751986. intestinal loops. Infect Immun 70:945952.
236. Klapproth JM, Sasaki M, Sherman M, Babbin 241. Neves BC, Shaw RK, Frankel G, Knutton S.
B, Donnenberg MS, Fernandes PJ, Scaletsky 2003. Polymorphisms within EspA filaments of
IC, Kalman D, Nusrat A, Williams IR. 2005. enteropathogenic and enterohemorrhagic Esche-
Citrobacter rodentium lifA/efa1 is essential for richia coli. Infect Immun 71:22622265.
colonic colonization and crypt cell hyperplasia in 242. Davis TK, van der Kar NCAJ, Tarr PI. Chapter
vivo. Infect Immun 73:14411451. 15, this volume.
237. Stevens MP, Roe AJ, Vlisidou I, van Diemen 243. Obata F, Obrig T. Chapter 5, this volume.
PM, La Ragione RM, Best A, Woodward MJ, 244. Ritchie JM. Chapter 8, this volume.
Gally DL, Wallis TS. 2004. Mutation of toxB 245. Mellies JL, Lorenzen E. Chapter 9, this volume.
and a truncated version of the efa-1 gene in 246. Smith DR. Chapter 25, this volume.

Adhesins
BRIAN D. McWILLIAMS1 and ALFREDO G. TORRES1,2

7
INTRODUCTION
Among the thousands of bacterial species contained within the intestinal gut
flora, it is accepted that each species requires the use of adhesin proteins, or
some combination thereof, that bring the bacteria closer to the epithelia and
allow them to colonize the intestine. In a similar way, enteric pathogens also
require surface-localized adhesins for colonization of the host intestine and
eventual establishment of disease. Enterohemorrhagic Escherichia coli (EHEC)
and, in general, Shiga toxin-producing E. coli (STEC) strains are known to con-
tain a large number of proteins responsible for adhesion and contribute to es-
tablishment, persistence, and tissue tropism observed during infection with
these pathogens. Understanding how these adhesins work is critical to having
a full picture of the pathogenic and pathophysiological process associated with
EHEC. Further, because adhesins play such an important role in virulence, they
are targets for therapeutic intervention. Thus, this review summarizes the cur-
rent knowledge on the adhesive proteins in EHEC, emphasizing up-to-date in-
formation and discussing gaps in knowledge and future directions in the study
of these virulence factors.
1
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, TX 77555; 2Department of
Pathology and Sealy Center for Vaccine Development, University of Texas Medical Branch, Galveston, TX 77555.
131

132 McWILLIAMS AND TORRES
LOCUS OF ENTEROCYTE EFFACEMENT region present on the terminal end of the

operon (9). Other parallel studies verified the
The study of adhesion in E. coli strains that presence of the LEE in a large number of
produce an intimate attachment to the epi- pathogenic E. coli strains, all of which were
thelia dates back to 1990, when a single gene, associated with A/E lesion formation, includ-
eae, was discovered in enteropathogenic E. coli ing the E. coli RDEC-1 strain found in rabbits,
(EPEC) through a transposon-based muta- EHEC O26:H, O15:H, O103:H2, and Citro-
genesis system. The product of this gene was bacter rodentium (912). Further studies iden-
established as necessary for the formation of tified specific gene products within this island
attaching and effacing lesions (A/E lesions) as being required for the colonization of the
(1). A follow-up study indicated that this gene bovine gut (13, 14), and others verified that the
was conserved in EHEC (2), and early in vivo LEE was associated with the development of
studies showed that this genes protein prod- enteritis in other animal models, such as mice
uct, intimin, was important for intimate at- (15), calves (16), sheep (17), and rabbits (18, 19).
tachment and colonization of the intestine in A large portion of the LEE represents the
both piglets and humans (3, 4). Though in- coding region for a type 3 secretion system
timin and other locus of enterocyte effacement (T3SS), a complex structural system that allows
(LEE)-encoded factors linked to pathogenesis translocation of effector proteins out of the bac-
have been extensively reviewed elsewhere (5, terial cell and into the surrounding environ-
6), we briefly describe some historical infor- ment or directly into a host cell. Early studies
mation and focus on recent advances in the of EPEC suggested that this pathogen was ca-
understanding of LEE-encoded Tir and in- pable of protein secretion, that these proteins
timin interactions. were required for full virulence and the for-
The eae gene is part of a larger set of genes mation of A/E lesions (20, 21), and that this
that make up the LEE pathogenicity island. secretion was likely mediated by a T3SS (22).
The LEE was first described as a 35-kb locus The relationship between the T3SS and the
that was conserved in different isolates of LEE was confirmed by the publication of the
EHEC and EPEC but was not present in non- aforementioned sequencing projects (8), which
pathogenic strains of E. coli (7). Two inde- helped identify individual components of the
pendent sequencing projects determined the T3SS and their role in pathogenesis (2325).
DNA sequence of the LEE and described a Since the LEE and its associated effector
locus that was divided into five major operons proteins play such a significant role in the vir-
containing up to 41 genes (8, 9) (Fig. 1). An ulence of EHEC, understanding the factors
additional 11 genes were identified in the that affect the expression of these genes be-
LEE of EHEC that were not found in EPEC came a priority. A number of studies have
and were associated with a prophage coding identified a series of transcriptional regulators
FIGURE 1 Illustration of the EHEC prototype strain EDL933 LEE pathogenicity island. The ve LEE operons are
depicted, specically emphasizing gene-encoded proteins associated with adhesion and A/E lesion formation.
They include the genes encoding the regulatory proteins Ler, GrlA, and GrlR; the translocated receptor protein
Tir; and the adhesin protein intimin. doi:10.1128/microbiolspec.EHEC-0003-2013.f1

CHAPTER 7 EHEC Adhesins 133
that affect the expression of various LEE- specifically curved DNA commonly found in
associated genes and include several proteins, promoter regions (45). The Ler protein origi-
such as QseA (26), H-NS (27), and LEE- nates from a similar protein family as H-NS,
encoded regulators Ler (28), GrlA, and GrlR and the ability of both of these proteins to
(29). In addition, there are specific environ- affect their target promoters requires a ca-
mental conditions that affect LEE expres- pacity to form long oligomers. Recent studies
sion, including pH, osmolarity, Fe(NO3)2, Ca2+, suggest that these two proteins contain simi-
temperature (30), quorum sensing (31, 32), and lar DNA-binding domains and differ only in
HCO3 (30, 33). More generally, LEE expres- the behavior surrounding oligomerization
sion is also modulated by carbon catabolite (46). This difference seems to represent the
repression in vitro (34) and by a combination mechanism by which Ler is able to antagonize
of factors found in spent media from epithelial the binding capacity of H-NS to the target
cell culture (35) or by posttranscriptional reg- promoter regions (38).
ulators like CsrA (36).
Intimin and Tir
H-NS and Ler as Important Regulators
Though H-NS and Ler regulate multiple genes
of EHEC Adhesion
located throughout the chromosome, it is
Of all the factors involved in the regulation of their ability to regulate the proteins intimin
the LEE, two of the most extensively studied and Tir and the long polar fimbriae (discussed
and closely associated with virulence are the below) that highlights their capacity to impact
negative regulator (silencer) H-NS and the EHEC adhesion. As mentioned, the intimin
positive regulator (antisilencer) Ler. The Ler protein was originally identified in EPEC (47),
protein was first identified as a member of and researchers quickly identified a homolog
the LEE1 operon, and it is a transcription fac- in EHEC (2). The EPEC and EHEC intimin
tor that positively regulates LEE2, LEE3, and proteins share 86% homology at the nucleo-
LEE4, and may also contribute to regulation tide level and 83% homology at the protein
of its own transcription (28, 3739). As such, level, with most of the divergence between
Ler and its regulatory activity are required for these two proteins at their C-terminal end,
adherence, A/E lesion formation, and viru- which is the domain responsible for target
lence of EHEC (40). Certain ler mutants have interaction (2, 4). Over the following decade,
also shown modified adhesion capabilities and it became firmly established that intimin was
displayed a novel set of fimbriae on the sur- the primary molecule in both EHEC and
face of the bacterial cell (39). Ler is involved EPEC associated with intimate bacterial in-
in multiple regulatory systems, and its tran- teraction with the epithelia. Cell-culture stud-
scription is mediated by a large number of ies using EHEC confirmed that this protein is
factors, including available sugars (41), quorum- essential for intimate adhesion, and disrup-
sensing associated proteins such as QseA (26), tion of the eae gene abolishes this phenotype
the noncoding RNA DsrA (42), and the uni- (4). Furthermore, intimin has been shown
versal regulator Hfq (43). Ler is also regulated to be essential for A/E lesion formation in
by GrlA, which completes an autoregulatory both gnotobiotic pigs and colostrum-deprived
loop, as Ler is responsible for initial GrlA acti- calves (14, 48, 49). Finally, antibodies against
vation (44). intimin inhibit EHEC adherence in a cell-
Ler imparts its positive regulatory effects by culture model (14, 49, 50). This central role
directly interfering with the negative histone- of intimin in the interaction with the intesti-
like negative regulator H-NS. This protein nal epithelia spurred further research toward
regulates a huge number of genes through- understanding this proteins function and role
out the EHEC chromosome by binding to in virulence.

DNA sequencing and structural studies intimin (57, 58). Upon integration, however,
indicated that intimin is a 95-kDa (939 amino Tir (EHEC) and Tir (EPEC) behave differ-
acids) protein that can be divided into two ently. Tir (EPEC) function is dependent on
functional regions. The N-terminal region (resi- the phosphorylation of tyrosine 474, which is
dues 1 to 550) contains a signal peptide region lacking in EHEC Tir (5961). Further, EHEC
(1 to 39) that is responsible for its interaction Tir is incapable of complementing EPEC
with the bacterial Sec pathway and eventual Tir on account of this amino acid difference
secretion to the bacterial outer membrane, (6062). Since Tir (EHEC) lacks tyrosine
as well as a LysM-type region that allows for 474, EHEC-mediated actin rearrangement
integration into peptidoglycan (51). From re- is instead controlled by a second bacterial
sidues 189 to 550, a b-domain spans the outer translocated virulence factor, EspFU (formerly
membrane of the bacteria and is responsi- TccP) (63, 64). This protein directly interacts
ble for promotion of surface exposure for with EHEC Tir upon its translocation into
host-cell interaction (52). The C-terminal re- the host cell (63), and these two proteins form
gion contains four smaller domains (D1 to a larger complex with two host proteins,
D4), each of which is an independent bacte- IRSp53 and IRTKS (6466). The complex
rial immunoglobulin-like domain that forms recruits and activates the host proteins N-
a surface-exposed rod that binds to Tir and WASP and Arp2/3, leading to the reorgani-
is capped by a C-type lectin domain (53, 54). zation of actin within the host cell and the
The interaction between intimin and its bind- pedestal formation commonly associated with
ing partner Tir primarily occurs between a A/E lesion formation (63, 65, 66).
superdomain formed between domain 3 and Though this interaction between Tir and
domain 4 of intimin and a b-hairpin motif on intimin is likely the primary means by which
Tir (55). EHEC is able to intimately adhere to epithelial
The effort to characterize the Tir and in- cells, other studies attempting to understand
timin interaction originally started as an effort the relationship between these two proteins
to identify what was thought to be a host- revealed that adhesion is a more complex
cell receptor protein mediating EHEC adhe- process than originally realized. Frankel et al.
sion. This receptor was eventually identified used purified EPEC intimin to determine if
in EPEC and called Hp90, whose phospho- and how intimin was interacting with HEp-2
rylation was shown to mediate the actin re- cells in vitro (67). In this scenario, intimin
arrangements visible in A/E lesion formation lacks its translocated binding partner Tir. Des-
and directly interact with intimin (56). Fur- pite this, fluorescence was observed on the
ther studies in EPEC recognized that this surface of these cells, suggesting that intimin
was actually the first (and only) example of a may be capable of interacting with host cells
bacterial protein being translocated into the independent of Tir. Follow-up studies using
host-cell membrane, acting as its own recep- an eae mutant were able to identify a single
tor for host-cell interaction (30). This protein amino acid in the C-terminal 150 residues of
used the T3SS to reach the host-cell cyto- intimin that was responsible for this interac-
plasm and was named Tir for translocated tion (68). Further, through solid-phase binding
intimin receptor (55). In EHEC, the previ- assays, intimin was shown to interact with the
ously discovered protein EspE was shown to host cell-surface-located protein b1 integrin
be a homolog of Tir and was renamed to re- (69). This interaction, however, is not required
flect this identity. Once translocation into for EPEC adhesion, and mutation of the gene
the host cell is complete, Tir integrates it- encoding b1 integrin did not have a negative
self into the host-cell membrane, exposing effect on A/E lesion formation or the attach-
both terminal ends to the cytoplasm and ment of EPEC to the surface of cells in vitro
an extracellular domain for interaction with (70). Finally, studies to map the binding domain

within intimin identified the previously dis- and b1 integrin. In the absence of Stx, EHEC
cussed lectin domain located on the C-terminal adhesion was significantly reduced, suggesting
end of intimin, and removal of this domain com- that the pathogen is capable of modifying host
pletely abrogated Tir-independent binding (71). transcription patterns to increase its chances
Similar studies characterizing Tir-independent of successful adhesion and attachment of ep-
binding in EHEC-derived intimin were done by ithelial cells (76).
Sinclair et al., who used fluorescent microscopy As previously mentioned, intimin contains
to colocalize b1 integrin and intimin (72); a two functional regions: the N-terminal region
separate study further strengthened this con- required for anchoring the protein into the
clusion by abrogating the interaction between bacterial membrane, and the C-terminal region
EHEC and cultured epithelial monolayers by responsible for host-cell and Tir interactions.
adding heparin and heparin sulfate (73). These As a result of this extracellular localization,
molecules directly interact with integrin and this region was shown to be recognized by the
act in a similar fashion as a monoclonal anti- immune system, as antibodies are known to be
body, interfering with the binding of these two developed against intimin during the normal
proteins. course of infection in patients with hemor-
In addition to the intimin-b1 integrin inter- rhagic colitis and hemolytic-uremic syndrome
action, a second Tir-independent interaction (7779). It has been speculated that to avoid re-
was discovered during these initial character- cognition by the immune system, both EPEC-
izations. The first studies looking closely at and EHEC-derived intimin proteins have
Tir-independent binding in EHEC identified shown a high degree of plasticity and subse-
a second binding event occurring between in- quent variability in this exposed C-terminal
timin and nucleolin (74). Still associated with region (67). This variability has allowed allele-
the C-terminal end of intimin, the binding of specific subtyping of the intimin protein.
this lectin domain was calculated to have an Overall, 27 total variants of the intimin pro-
affinity of around 100 nM to host cells, and tein have been identified that are made up of
through affinity purification and sequencing 18 different types and 9 subtypes. They include
steps, nucleolin was identified as a binding a, a2, b1, b2, b3, g1, g2, d, e, e2, e3, e4, z, h, h2, v,
partner mediating this interaction. More de- i, i2, k, l, m, n, y, o, p, r, and j (8087). The most
tailed studies on this interaction showed common clinically relevant subtypes are those
that both EPEC- and EHEC-derived intimin within a-, b-, and g-type. The a-type is found
bound to nucleolin with the same affinity but in multiple EPEC strains, including O127:H6,
that this interaction was weaker than that as is the b-type, with the only exception being
between intimin and Tir (75). These results EHEC O26:H11. The g-type includes EHEC
laid the foundation for a hypothesis indicating O157:H7 and two other EPEC strains, O55:H
that the intimin-nucleolin and the intimin- and O55:H7 (6).
integrin interactions were only secondary Because one unique feature between dif-
binding events during infection and that ferent types of A/E-inducing pathogenic E. coli
intimin-Tir interactions are required to keep strains relates to tissue tropism, it is logical to
the pathogen in place while the T3SS and its propose that these different intimin subtypes
effector proteins were injected into the host. are one of the driving forces behind the di-
Finally, some more recent data further illus- verse tropism and variability in EHEC and
trated the importance of intimins interaction EPEC interaction with host cells. Indeed, early
with both nucleolin and b1 integrin through studies prior to the establishment of the in-
the effects of EHEC-encoded Shiga toxin timin typing system determined that substi-
(Stx). During infection, Stx is internalized via tuting intimin from a g-subtype (O157:H7) to
receptor-mediated endocytosis and is capable a b-type (O127:H6) completely changed the
of modifying the transcription of nucleolin target of infection from the large intestine,

which is commonly seen in EHEC, to both the Intimin-based vaccines have also been shown
small and large intestines, which are more to be effective in rabbits (95) and adult cattle
frequently associated with EPEC infections (96), and infant calves were protected by anti-
(48). Later studies using in vitro organ cultures intimin antibodies obtained from colostrum
supported this hypothesis by showing that (97). Second, a number of different vector sys-
restoring different intimin subtypes into a tems have been used to produce intimin, such
O157:H7 eae mutant strain shifted the adhe- as developing a plant-derived edible version
sion from Peyers patches, common in EHEC, of intimin (98) or using attenuated Salmonella
to the small intestine (88). Further, this study species as the delivery vehicle (99, 100).
validated grouping within the intimin g-type, Ultimately, the ideal vaccine should protect
as EPEC O55:H7 was also associated with against all major intimin subtypes found in
Peyers patches in the large intestine. Other EHEC and EPEC, and it could be used to
studies, using subtypes from a-, b-, and g- protect bovine or porcine herds from coloni-
types, showed shifting tissue tropism in both zation, thus indirectly diminishing the poten-
in vitro organ-culture models and in vivo in- tial exposure of these categories of E. coli
fections in mice (89). strains to humans.
Though strong evidence suggests that the
emergence of these intimin types and subtypes
Long-Polar Fimbriae (Lpf1 and Lpf2)
is a way to predict tissue tropism, these dif-
ferences may have originally been the result Cumulative evidence supports the fact that
of immune system avoidance. As mentioned, although intimin is the primary adhesin in
antibodies against intimin are detectable in EHEC/STEC, it is by no means the only con-
patients with EHEC and EPEC infection, and tributing factor. This became clear when a
these antibodies are generally specific to their novel STEC O113:H21 strain was isolated from
serotype (90, 91). The presence of antibodies infected humans and showed to be lacking
against intimin, combined with the fact that the intimin gene (101, 102). Not only was this
this protein is extracellular, has made anti- strain infectious, it was able to adhere to in
body-mediated protection a promising solu- vitro cultured cells, implicating other proteins
tion to both EPEC and EHEC infections. This in contributing to the adhesive capabilities of
protection likely occurs as a result of antibody- pathogenic E. coli. In addition to this obser-
mediated interference with intimin inter- vation, two different EPEC O55 serotypes,
action with Tir, b1 integrin, or nucleolin. O55:H6 and O55:H7, have been shown to ex-
Initial studies have shown that antibodies did press intimin-a and -g, respectively, which
significantly limit the amount of adherence should result in specific tropism to two differ-
observed on the surface of in vitro cultured ent regions of the intestine. However, it was
epithelial cells (50, 92), and vaccination with observed that these strains are both restricted
recombinant intimin was able to induce spe- to interaction with the follicle-associated en-
cific antibody responses (93). Further, a wide dothelium (FAE), suggesting that perhaps
range of protection studies using many dif- other factors are contributing to tissue tro-
ferent types of host organisms and delivery pism in addition to intimin (103).
systems have afforded a number of interest- Full-genome sequencing and subsequent
ing conclusions. First, multiple studies have analysis of two prototype O157:H7 strains iden-
reinforced the fact that antibody-mediated tified approximately 1.34 Mbp of DNA se-
protection is subtype-specific. For example, quence present in O157:H7 but absent in the
g-intimin-vaccinated mice generated anti- nonpathogenic K-12 strain (104, 105). These
bodies against EHEC infection and displayed a regions are distributed throughout the genome
decreased shedding time (15). Similar results as O islands, and two of these, islands #141
were seen with intimin-a and EPEC (94). and #154, contained regions with homology

to Salmonella enterica serovar Typhimurium. The lpf2 operon contains a duplication of lpfD
The proteins encoded in these regions were called lpfD and lacks the lpfE gene (110). On
predicted to produce the components of the account of these genetic differences, predicted
long polar fimbriae (Lpf), originally charac- structural variations have been proposed and
terized in Salmonella (106) and found to con- may contribute to some small differences in
tribute to the pathogens adherence to the function between Lpf1 and Lpf2. Further, the
human intestine (Table 1). A mutation in the lpf operons are not limited to EHEC, and are
gene encoding the Salmonella major fimbrial seen in a wide range of pathogenic E. coli
subunit (lpfA) showed a significant reduction strains and some nonpathogenic commensal
in adhesion when in vitro tissue-cultured cells strains (109). Similar to intimin, the lpfA gene
were used; specifically, this mutation seemed has been divided into allele groups, of which
to affect the bacterial interaction with Peyers five exist for the lpfA1 gene and three exist for
patches and M cells (107). As a result of the the lpfA2 gene (112, 113). Interestingly, there
homology with the Salmonella lpf operon, the seems to be a correlation between the differ-
O157:H7 regions O-141 and O-154 were named ent Lpf allele group combinations and the
lpf1 and lpf2. subtype of intimin associated with a given
The lpf1 operon consists of five genes with E. coli strain that may be contributing to tissue
functions predicted to be similar to those de- tropism within the intestine (112).
scribed in Salmonella serovar Typhimurium In addition to slight differences in their
(108110). The first gene, lpfA, encodes the genetic makeup, lpf1 and lpf2 are regulated
major fimbrial subunit protein whereas lpfB is in slightly different ways as well. Studies in-
a chaperone predicted to participate in proper dicate that transcription of lpf1 is increased
folding. The third gene, lpfC, is the outer mem- during late log phase in Dulbeccos modified
brane usher protein and contains a stop codon Eagle medium at 37C, pH of 6.5, and in re-
within this predicted reading frame, dividing sponse to salt concentrations (108). The lpf2
the gene into lpfC and lpfC. The lpfD gene is positively affected by iron depletion, sug-
represents the coding region for a minor fim- gesting that the regulatory protein Fur might
brial subunit. Finally, the lpfE gene encodes participate in this activation (114). At least for
for another predicted subunit of the fimbriae. the lpf1 operon, the relationship between the
TABLE 1 Major characteristics of the EHEC mbrial proteins

Primary Cell lines used
EHEC mbria name structural gene Representative strains Ligand to show binding
Long polar mbria 1 lpfA1 O157:H7, O55:H7, Fibronectin, Caco-2, HeLa

O127:H6 collagen IV, laminin
Long polar mbria 2 lpfA2 O157:H7 Unknown Caco-2
Curli csgA O157:H7, O26:H11 Abiotic surfaces, T84
lettuce leaves
E. coli common pilus ecpA O157:H7, O26:H11, Unknown HT-29, HEp-2,
O32:H37, O104:H4 HeLa
F9 mbria z2200 O157:H7, O26:H, O104:H4, Fibronectin EBL, HeLa
O55:H7, O127:H6
E. coli laminin-binding ycbQ O157:H7, O26:H11, Laminin HT-29, HEp-2,
mbria O104:H4 MDBK
Sorbitol-fermenting sfpA O157:H, O157:NM, Unknown Caco-2, HCT-8
mbria protein O165:H25
Type 1 mbria mA O26:H11, O118, O157:H7, Abiotic surfaces REC
O157:H, O55:H
Hemorrhagic hcpA O157:H7 Laminin, bronectin T84, Caco-2, HeLa,
E. coli protein HEp-2, MDBK

environment and its expression is linked to proteins (110). Other studies of EHEC strain
two regulatory proteins: H-NS and Ler. The O113:H21 showed the presence of a single lpf
H-NS and Ler relationship with the lpf1 op- operon located in the same chromosomal po-
eron was originally established by using E. coli sition as Lpf2. This operon contained a short-
O157:H7 hns and ler mutant strains and b- ened lpfD gene (similar to REPEC O15:H)
galactosidase assays in which the lpf1 promo- that was nonduplicated (making it distinct
ter region was fused to a reporter gene (115). from Lpf2 from EHEC O157:H7) and also
The hns mutation showed an increase in lacked an lpfE gene (found in EHEC O157:H7
b-galactosidase activity, and the ler mutant Lpf1) (109). When this operon was deleted
showed a corresponding decrease in activity. from the genome of O113:H21, the bacteria
The regulatory control of these proteins was were significantly less capable of adhering
further established through quantitative real- to CHO-K1 epithelial cells, emphasizing the
time reverse transcription-PCR analysis, primer fact that Lpf proteins are critical for adher-
extension, and electrophoretic mobility shift ence in vitro. Finally, to further define the
assays. These experiments also helped identify contribution of H-NS and Ler regulation to
two j70-type promoters in the lpf1 regulatory the function of EHEC Lpf1 and its adhesion
region and showed that both Ler and H-NS characteristics (117), HeLa cells were exposed
were able to bind those regions (115). Further to EHEC carrying single- or double-knockout
analysis allowed for the precise definition and mutations in hns and ler. It was demonstrated
localization of three individual binding sites that an hns mutation caused an increase in
for H-NS (116). adhesion and a ler mutation resulted in a de-
Once conditions are appropriate for the crease. This reduction in adhesion was also
expression of Lpf, these proteins, particularly observed when anti-Lpf antibodies were
Lpf1 fimbriae, display characteristics similar added during an in vitro adhesion assay with
to those described for Lpf in Salmonella spe- the wild-type strain. Interestingly, when both
cies. Early studies showed that expression of hns and ler genes were deleted, a 6-fold in-
Lpf1 from a plasmid containing the lpf operon crease in adhesion was observed, suggesting
and transformed into an E. coli K-12 strain in- that other positive regulators influencing Lpf1
creased bacterial adhesion to tissue-cultured might exist and have not been described.
cells and also increased the number of mi- Some progress has been made in estab-
crocolonies that were formed in vitro (108). lishing the host binding partner for Lpf1, as
Further, the Lpf fimbriae were visible by it was shown that wild-type EHEC is able to
transmission electron microscopy, and it was interact with the extracellular matrix (ECM)
determined that they displayed a long, fine proteins, specifically fibronectin, laminin, and
structure. In contrast, when the lpf2 genes collagen IV (118). Further, anti-LpfA1 anti-
were cloned into this same strain of K-12, a bodies were added to these interaction assays,
reduction in the adherence phenotype was and it was observed that binding to these
observed (110). Further, a mutation of the proteins decreased. Deletion of lpf1 and lpf2
lpfA2 gene in the wild-type O157:H7 did not reduced binding to the ECM proteins, and
have a significant effect on the adhesive prop- deletions of both hns and ler genes caused
erties in vitro while interacting with HeLa a similar decrease in the binding phenotype.
cells. However, when Caco-2 cells were used, It was also noted that purified recombinant
a reduction in adhesion was visible at early LpfA1 was capable of interacting directly
time points, suggesting that Lpf2 may con- with these ECM proteins, and when T84
tribute to early adhesion whereas Lpf1 may cells were preincubated with ECM proteins
contribute to later steps in the adherence pro- and then incubated with O157:H7 wild-type
cess, perhaps when Ler and H-NS are also strains, an increase in bacterial adherence was
regulating the expression of the LEE-encoded observed.

The first in vivo models to study the role with cellular adhesion independent of tissue
of Lpf during infection used sheep, conven- specificity (120). The presence of these alter-
tional pigs, and gnotobiotic piglets infected native mechanisms has been suggested in an-
with the lpfA1 single or lpfA1/lpfA2 double other study using the rabbit infant model (18).
mutant strains (119). The study sought to dif- This model uses rabbit intragastric infections,
ferentiate stages of colonization in lieu of which are considered to be the most effective
the previously discussed results, suggesting way to reproduce EHEC infection in vivo in
that Lpf1 may play a role early in colonization an animal model. In this study, the lpfA1/lpfA2
(110). The early colonization results, in which double mutant mixed with the wild-type
germ-free pigs were sacrificed 24 h post- strain was inoculated into the rabbits. The mu-
infection, indicated that the lpfA1/lpfA2 dou- tant was outcompeted for colonization in the
ble mutant strain colonized significantly less ileum, cecum, and mid-colon, reinforcing the
in the spiral colon, but no statistical difference fact that Lpf was an important contributing
was detected in the cecum. Over an extended adherence factor (121). Unexpectedly, when
period, the lpf double mutant also shed sig- this double mutant was used in in vitro ad-
nificantly less in sheep starting at 2 weeks hesion assays with colonic epithelial cells
postinfection, and the lpfA1/lpfA2 double mu- (Caco-2), adhesion was observed to increase
tant was not recovered from any intestinal when compared to wild-type strain. Trans-
tissue after sacrifice, in contrast to the wild- mission electron microscopy indicated the
type strain, which was detected at 100 CFU/g presence of a novel structure that appeared
of tissue. A different study used in vitro organ to be curli, a thin afimbrial adhesin factor
cultures from human intestinal tissue and implicated in biofilm formation and induc-
sought to better understand the effects that tion of inflammation (122). When the major
Lpf had on tissue tropism (120). This study structural subunit of curli (CsgA) was deleted
indicated that the wild-type strain bound from the parent O157:H7 strain, in vitro ad-
only to the FAE of Peyers patches but was hesion was not significantly decreased, sug-
unable to interact with the duodenum, ter- gesting that curli may only be expressed under
minal ileum, or the transverse colon. How- conditions in which Lpf expression was de-
ever, when a knockout was introduced into creased or eliminated. This also suggested that
either or both lpfA1 and lpfA2 genes, the there was some sort of common regulatory
strains were able to interact with all of these mechanism between these two surface struc-
intestinal regions except the transverse colon. tures, though changes in the transcription of
This result further emphasized the hypothesis csgA were not observed in the double mutant
that Lpf may be specifically driving adhe- strain (123).
sion to the FAE and working with intimin to
dictate EHEC tissue tropism. This FAE spec-
Curli
ificity is especially relevant with the under-
standing that O157:H7 tends to colonize areas Curli has been recognized as an adherence-
in the intestine where FAE is found, thus related factor capable of mediating interac-
providing a role for Lpf during colonization tions between EHEC and host cells (Table 1),
to the human intestine. A study by Fitzhenry specifically with the ECM proteins laminin
et al. identified expression changes in 124 and fibronectin, along with plasminogen and
genes between the lpf mutants and the wild- major histocompatibility complex class 1 mole-
type strain (120). The expression changes in- cules (124, 125). Curli is formed by thin fibers
volved genes encoding some surface-exposed that aggregate on the surface of the cells
structures (FimG and YagZ) and some tran- and are characterized by their binding to
scriptional regulators, suggesting that EHEC Congo red dye (124, 126). In addition to po-
contains alternative mechanisms associated tentially acting as a compensatory mechanism

for adhesion in the EHEC lpfA double mutant, several regulators, possibly giving EHEC the
curli and Lpf share a number of other char- opportunity to colonize and survive on other
acteristics. Curli and Lpf1 genes are both surfaces until more favorable environmental
negatively regulated by H-NS and contribute conditions are available (121).
to in vitro adhesion of tissue-cultured cells
(122, 127). Curli is transcriptionally influenced
Other Fimbrial Adhesin Proteins
by environmental signals, allowing curli tran-
scription to respond to a number of different ECP (Mat) Fimbriae
stimuli, such as pH, temperature, and nutri- In addition to EHEC Lpf1, Lpf2, and curli, a
ent limitations (128130). Curli exists at the study by Low et al. used genomic data analy-
genetic level as two distinct and divergently sis to predict and analyze the expression of 13
transcribed operons: csgDEFG and csgAB (127, more fimbrial operons, many of which have
131). CsgD is the primary transcriptional acti- proved to be functionally relevant (Table 1)
vator of csgA, and the gene encoding this (140). One of these operons encodes the E. coli
regulator is under direct control of H-NS, YagZ homolog (also known as MatB) that was
RpoS, Crl, MlrA, Rcs, and the two-component later renamed as E. coli common pilus (ECP)
systems OmpR/EnvZ and CpxA/CpxR, IHF, (141). The isolate O18:K1:H7 is an E. coli strain
and SdiA (122, 132136). associated with neonatal meningitis and sep-
In addition to assuming the role of a pri- ticemia (NMEC) and provided the first in-
mary adhesin of EHEC in the absence of Lpf, formation regarding Mat/ECP function. This
the primary function of curli is to mediate fimbria was first identified after attempts to
bacterial interaction with abiotic surfaces and create a strain of NMEC that was completely
protection of the pathogen from antiseptic devoid of fimbriae, and despite a triple-
chemicals. One study indicated that curli can knockout of known fimbrial proteins, some
directly contribute to bacterial attachment to fimbrial structures were still present on the
stainless steel and promotes bacterial resis- surface of these cells (142). The fimbriae were
tance to chlorine (137). Other studies have iden- purified and identified, and their coding re-
tified a large number of other abiotic surfaces gion was verified in a wide range of pathogenic
to which curli mediates binding, including E. coli strains, including EHEC, EPEC, en-
polystyrene, glass, and rubber (138). Perhaps terotoxigenic E. coli, enteroaggregative E. coli,
most important is its role in adhesion of and nonpathogenic E. coli K-12. Despite this,
EHEC to vegetable leaves, making curli of the study was only able to verify Mat/ECP
particular interest in the investigation of re- expression in NMEC strains, and expression
cent outbreaks associated with contaminated was only observed at 20C in rich media
spinach and lettuce. Curli-expressing strains and involved both transcriptional and post-
of pathogenic E. coli have been isolated from transcriptional events. The mat (ecp) mutant
multiple outbreaks worldwide, and a recent strains showed a loss of Mat/ECP on the
study indicates that CsgA is up-regulated surface of their respective strains; these struc-
20-fold during interactions with lettuce leaves tures could be restored in trans via plasmid-
(139). The csgA mutant showed a decrease expressed Mat/ECP (142). Later studies of
in adhesion to leaf tissue during both short- EHEC showed that ECP is also critical for
term and long-term colonization assays. As adherence to HEp-2 and HeLa cells and that
discussed by Lloyd et al., the difference in this adherence phenotype can be abolished
positive regulators may explain the differ- by mutating ecpA (141). The study also assayed
ence in expression patterns for Lpf and curli. for EcpA expression in 176 different strains
Whereas Lpf is the default on adhesion fac- of E. coli, verifying production of the fim-
tor and expressed under conditions mimick- briae in almost 72% of strains tested, which
ing the intestine, curli is strictly controlled by included EHEC O157 and non-O157, as well

as non-EHEC strains. In comparison with pre- ELF and Sfp

vious studies, these strains were grown at Another set of genes analyzed by Low et al.
26C in Dulbeccos modified Eagle medium was associated with adhesion to host cells, but
and showed a lack of expression at 37C and no extensive work was done to confirm their
in rich media. This study also showed that relationship with adherence or their func-
ECP expression increases under low oxygen tional roles (140). One of those proteins, origi-
and high CO2, environmental conditions that nally identified as locus 5, was identified as
are experienced in the intestinal tract. Studies part of the ycbQRST fimbrial-like operon and
of EPEC also verified this environmental de- was renamed as ELF for E. coli laminin-
pendence for the expression of ecpA (143). binding fimbria (147). The ELF fimbria was
Purified EPEC EcpA was able to bind to and purified and shown to interact with the ECM
interact with the surface of HT-29 epithelial protein laminin, and its expression was veri-
cells, reaffirming that this protein is an adhesin fied on the surface of O157:H7 during adher-
(143). ence to epithelial cells (147). Like many of the
previously discussed fimbriae, this protein is
F9 Fimbriae expressed maximally during growth in mini-
Another fimbrial gene cluster analyzed in mal media and in the presence of host cells,
the study by Low et al. was the F9 fimbriae. and mutation of the coding region resulted
This protein had been previously identified as in a loss of adhesion to HEp-2, HT-29, and
a factor associated with tissue colonization MDBK (bovine kidney) cells.
when a series of 59 mutants were obtained by Another fimbria described in the EHEC
transposon mutagenesis in O157:H7 screened pathogroup is the sorbitol-fermenting fimbria
for intestinal colonization in calves (16) and protein (Sfp) that was originally identified in
in a second study using O26:H in cattle (144). an O157:H strain from Germany and showed a
A subsequent study of this fimbria performed high degree of similarity to the pyelonephritis-
by Low et al. as a follow-up analysis of one of associated pili (pap) genes in uropathogenic
four candidate adhesins from their original E. coli (148). Sfp was also identified in EHEC
screen showed expression under in vitro con- O157:NM and O165:H25/NM strains and shown
ditions (145). Deletion of the z2200 gene in to express poorly under nutrient-rich and aer-
O157:H7 significantly reduced the pathogens obic conditions. However, under anaerobic con-
ability to colonize the calf intestine, but the ditions and on media simulating that of the
fimbria was also shown as not essential for colonic environment, induction was increased
bacterial shedding or required for proper tis- and corresponded with an increased level of
sue tropism (145). F9 fimbria also seems to adhesion to Caco-2 and HCT-8 cells (149, 150).
have an affinity for ECM proteins, as it was
shown to interact with fibronectin. Interest- HCP
ingly, when the F9 operon is transformed into Another fimbria recently described is the hem-
an F9-deficient O157 strain, the in vitro ad- orrhagic E. coli pilus (HCP). The fimbria,
hesion decreased significantly (145). A later originally referred to as TFP (type IV pilus) or
study identified CadA as a protein associated simply PpdD (predicted protein D), was orig-
with F9 fimbriae expression when a micro- inally identified in a study that visualized the
array analysis indicated that the F9 fimbriae fimbria on the surface of O157:H7 grown in
are up-regulated in a cadA mutant (146). The minimal media (151). This study also showed
O157 cadA mutant resulted in a hyperadherent that HCP was involved in EHEC interactions
phenotype in the ileum of an infant rabbit, with human colonic cells (T84 and Caco-2),
implying that the F9 fimbriae may also be con- nonintestinal cells (HeLa and HEp-2), and
tributing to tissue-specific adherence during bovine kidney (MDBK) epithelial cells, as dis-
infection. ruption in the hcp coding region caused a

significant decrease in adhesion to these cell Autotransporters

types. HCP was also shown to contribute to
EhaA-D
adhesion when pig and cow intestinal explants
In addition to the different fimbrial proteins,
were used, and it was suggested that HCP in-
there is a second group of surface-exposed
teraction works in concert with other adhesins
structures that contribute to EHEC interac-
and serves as a second binding interaction
tion with host cells. These proteins have the
on the surface of host cells (151). Later studies
capacity to be secreted independently of the
characterized a series of other functions, some
conventional cell translocation machinery and
of which are currently unique to HCP, as com-
are thus referred to as autotransporters. These
pared to other EHEC fimbriae, and include
proteins are listed in Table 2 and generally
involvement in the invasion of host epithe-
contain a signal sequence at the N terminus
lial cells, hemagglutination of erythrocytes, the
of the polypeptide, a passenger domain in the
formation of biofilms, and binding to the ECM
middle of the protein that determines the ac-
proteins laminin and fibronectin (152).
tivity of the protein, and a C-terminal region
that contains a b-barrel transmembrane re-
Type 1 Fimbria
gion that embeds into the outer membrane
The final fimbrial protein discussed here is
and provides the passenger domain the ability
the type 1 fimbria, which has been extensively
to be translocated to the extracellular milieu
studied as a common fimbria in E. coli; how-
(reviewed in reference 158). These proteins
ever, it has also been shown to mediate path-
are often associated with the formation and
ogenic E. coli adhesion to epithelial cells in the
maintenance of biofilms and other virulence
rumen of cattle (153). Early studies identified
traits in pathogens and related commensals
the type 1 fimbria in the EHEC strains O26
(159). One study identified a series of four pu-
and O118 (154) after being genetically traced
tative autotransporter-encoding genes, ehaA
to other EHEC strains, including O157:H7,
D present in EHEC O157:H7 (160). EhaA was
O157:H, and O55 (155). Despite the presence
cloned in trans into an avirulent strain of
of the coding regions in the genome of these
E. coli and expressed under control of an arab-
strains, type 1 fimbria expression was not ob-
inose promoter. Upon expression, EhaA was
served in these EHEC strains. This phenom-
able to induce cellular aggregation as deter-
enon was attributed to the presence of a fim
mined by fluorescence microscopy; further,
switch, a 16-bp region in the regulatory re-
this aggregation was the product of EhaA-
gion of fimA. When absent, the switch is con-
EhaA interactions. This protein is also capable
sidered to be in the off position and will
not allow the expression of fimA, and when
present, is considered to be in the on posi-
TABLE 2 EHEC autotransporters and their functions
tion and expression is allowed to proceed (155,
Autotransporter Primary Binding
156). The fim switch is turned on in the bovine protein function target
EHEC strains O26 and O118, and studies in-
EhaA Biolms EhaA-EhaA
dicate that type 1 fimbria is a contributing fac- EhaB Biolms Collagen,
tor to their virulence in the cattle infection laminin
model (156). When the fim switch from these EspP Protease EspP (ropes)
strains is cloned into the regulatory region of activity, biolms
fimA from O157:H7, expression is restored and Saa Host-cell HEp-2,
adhesion Caco-2
the adhesion to host cells increases (156). The Sab Host-cell adhesion, HEp-2
mutation of the fimA gene in non-O157 STEC biolms
and O128:H2 was shown to affect adherence Cah Autoaggregation, Alfalfa
to abiotic surfaces, such as polystyrene and Ca+ binding, sprouts
glass (157). leaf binding

of inducing the formation of biofilms, but be- espP in O157:H7 abolished adhesion in a calf
cause the protein is physically shorter than model and to primary bovine intestinal epi-
other surface structures, its adhesive capacity thelial cells (164). Addition of exogenous,
can be eliminated by the expression of type 1 purified EspP to the in vitro system caused
fimbria, which is considerably longer than the restoration of the adhesive phenotype of
EhaA. Further, 24 of 50 STEC strains were the espP mutant strain. A later study investi-
tested for the presence of EhaA and showed gated the presence of rope-like fibers during
expression of this protein under growth in growth of both EHEC and EPEC strains (167).
minimal media, perhaps implicating a mech- These fibers were actually made of pure EspP
anism where cell-to-cell adhesion is preferred that had oligomerized into fibers measuring
under conditions where type 1 fimbria is not up to 2 cm in length. The EspP-based ropes
expressed (160). Another study verified a simi- were capable of binding Congo red dye, and it
lar set of characteristics for EhaB; expressing has been proposed that the ropes may par-
this protein in E. coli K-12 conferred an ability tially function in the protection of the bac-
to form biofilms (161). Expression of EhaB terial cells from antibiotics and detergents.
produced an E. coli K-12 that was able to in- E. coli cells were found to interact directly
teract with the ECM proteins collagen and with this EspP-derived rope, possibly using
laminin. When the ehaB gene was disrupted this protein-based structure as a foundation
in EHEC O157:H7, the mutation did not affect for biofilm-mediated interactions between bac-
bacterial adhesion to primary epithelial cells terial cells (167). This protein was capable of
derived from the bovine recto-anal junction interacting with cultured epithelial cells and
or to Caco-2 cells. Interestingly, this protein with fibronectin, and the interaction with
reacted to serum from infected cattle, indi- the cultured cells resulted in a cytopathic, but
cating that it was expressed during infection not cytotoxic, effect. Finally, antibodies against
and is likely to be exposed to the immune EspP were observed in patients with hemolytic-
system (161). uremic syndrome, indicating that this protein
was being expressed during clinical infection
EspP and might play a role in the pathogenesis of
Another important autotransporter protein in EHEC (167).
EHEC is EspP. This protein, a homolog to
EspC from EPEC, is a member of the serine pro- Saa and Sab
tease autotransporters of Enterobacteriaceae Two other autotransporter proteins, Saa and
family of proteins, which, in addition to hav- Sab, are both plasmid-encoded adhesins that
ing adhesive capabilities, also contains prote- have been identified primarily in the EHEC
ase activity (162). The initial reports indicated LEE-negative O113:H21 strain. Saa was the
that this protein cleaves porcine pepsin A, first adhesin protein to be identified in an
human coagulation factor V (from the blood LEE-negative strain; the original studies were
coagulation cascade pathway), and apolipo- done in vitro and showed that the expres-
protein A-1; the purified form of the protein sion in trans resulted in a nearly 10-fold in-
was shown to be cytotoxic to Vero cells (163). crease in bacterial adhesion to HEp-2 cells.
Regarding its pathogenic association, this pro- Conversely, mutation of the saa gene resulted
tein is also capable of cleaving the comple- in a significant reduction in this binding
ment proteins C3/C3b and C5, which may (168). Multiple variants of the saa gene are
help in immunomodulation, and is able to based on the presence or absence of a series
cleave and subsequently inactivate EHEC he- of repeats on the 3 end. Variations of this 3
molysin, which might help in controlling and end had been hypothesized as a mechanism
modulating the degree or timing of its own for bacterial regulation of the adhesive affinity
pathological effect (164, 165). Mutations in and shorter variants as less adhesive than

longer variants (168, 169). This hypothesis promote the long-term adhesion and coloni-
was explored in more depth in a study that zation to a given surface (173).
verified the presence of the gene in 32 dif-
ferent STEC strains and explored whether
Flagella
these strains adhere to HEp-2 and Caco-2
cells (170). Though there were differences in Flagella, proteins associated with bacterial
degrees of adhesion in these strains, the study motility, were first characterized as adherence
could not find a correlation between repeat factors in pathogenic E. coli in a study that
length and adhesive capabilities, nor was any characterized the fliC gene in strains O127:H6,
correlation between Saa expression levels O119:H6, O128:H2, and O127:H40 (174). The
and adhesive properties statistically signifi- in vitro studies found an association among
cant (170). Sab has not been characterized as flagella, adherence, and the formation of micro-
extensively as Saa, but it has been shown to colonies on both HeLa cells and HEp-2 cells
mediate adhesion to HEp-2 cells, and the and also confirmed that these proteins were
mutant strains were less adhesive than the responsible for the E. coli motility phenotype.
wild-type O113 strain (171). Further, the study Flagella were identified as adhesins in EHEC
showed that the protein is present on the when a fliC mutation was found to reduce the
surface of these O113 strains, that these pro- degree of virulence in a chick infection model
teins could interact with HEp-2 cells in vitro, (175). The virulence properties in the intestine
and that mutant strains were less capa- are likely a product of a direct interaction
ble of forming biofilms than their wild-type with the mucus layer, as EHEC flagella have
counterparts (171). been shown to interact directly with Muc2,
and the EPEC flagella have been shown to
Cah interact with Muc1 (176). Interestingly, the
The final autotransporter protein that has EPEC flagellin was capable of interacting with
been associated with EHEC infections is the collagen, laminin, and fibronectin, but this
calcium-binding antigen 43 homolog (Cah). interaction was not observed with EHEC fla-
When expressed in an E. coli K-12 strain, gella (174). The role of flagella as adhesin pro-
this protein caused bacterial autoaggregation teins was further confirmed by a series of
and was also able to bind to Ca+ ions in solu- standard in vitro adhesion studies showing
tion (172). The cah gene is duplicated in EHEC that a mutation in the fliC gene reduced bac-
(cah1 and cah2), and their expression is max- terial adhesion to terminal rectum epithelial
imized under low nutrient conditions; these cells, and complementation of this mutation
proteins were also shown to contribute to the restored the binding phenotype (177). Purified
formation of biofilms when grown in mini- flagella were capable of interacting with these
mal media (172). Further studies indicated cells in vitro and of interfering with bacterial
that this protein was also an important factor binding. Finally, it was shown that the EHEC
in EHEC adherence to alfalfa sprouts and seed flagella are actually down-regulated after con-
coats, a feature of especially high interest tact with the epithelium, suggesting that bind-
when discussing persistence of EHEC in pro- ing may be required only for early stages and
duce (173). These proteins, when provided in provides a means for the T3SS to increase
trans to an E. coli K-12 strain, showed in- transcription of its components and mediate
creased adhesion to these two alfalfa surfaces. a more long-term binding event via the A/E
It was hypothesized that this protein may con- lesion (177). Other studies have indicated that
tribute more to a docking stage of adhesion, flagella are responsible, at least in part, for
where initial contact is made with the binding EHEC binding to leafy greens. Ultrastructural
surface, than to the primary stage of adhe- analysis showed the presence of flagella on the
sion, where interactions of a higher affinity surface of EHEC during this bacterium-plant

interaction, and mutation of the fliC gene IgA (181). Later, in vitro studies using both
negatively affects the binding (152). HCT-8 cells and bovine intestinal epithelial
cells indicated that the chain-like adhesion
pattern varied in size according to which allele
Other Adhesin Proteins
of the eibG gene was present in the genome,
Three final adhesins are not included in any even within a given subtype (182).
of the fimbrial/afimbrial families described Finally, the role of the outer membrane
above, but their contribution to colonization protein A (OmpA) in adherence was first char-
is significant to EHEC pathogenesis. Two of acterized in EHEC during a transposon muta-
these, Iha and EibG, are generally regarded as genesis screen for a hyperadherent phenotype
adhesin proteins, though their adhesive char- (183). This screen identified the transcriptional
acteristics have not been fully demonstrated regulator TdcA as a regulator of OmpA. Fur-
as others discussed in this review. Iha is a ther investigation led to a specific mutation of
homolog of the Vibrio cholerae adhesive pro- ompA, which reduced the hyperadherent phe-
tein IrgA and is present in O157:H7 but absent notype in the tdcA mutant strain. Antibodies
in other strains of EHEC (178). It was origi- specific to OmpA decreased adhesion by 25%
nally identified during a random screen, and during in vitro adhesion assays with Caco-2
mutations in this gene showed decreased ad- and HeLa cells (184). In addition to being pres-
hesive capabilities to HeLa and MDBK cells, ent in EHEC, this protein is commonly present
and subsequently showed increases in bind- in other commensal and pathogenic strains of
ing when added in trans to an otherwise non- E. coli, but it is only known to mediate adhesion
piliated E. coli strain (178). The evidence for to brain microvascular cells by NMEC (185
its contribution to EHEC adhesion was dis- 187). Purified OmpA has been crystallized (188)
played in two studies. The first of these stud- and is known to stimulate dendritic cell mi-
ies showed that transcription increases under gration through polarized epithelial cells (184).
growth conditions simulating short-chain fatty
acid concentrations present in the human gut;
however, transcription was not regulated by CLOSING REMARKS
the presence of iron (179). The other study
showed that a mutation of the EHEC iha gene As evidenced by the wide scope of this article,
resulted in a decrease in colonization in li- a large number of proteins within the EHEC
gated pig intestine, but adherence in vitro proteome contribute to varying degrees to its
seemed to be unaffected (180). Despite these interaction with plant leaf surfaces, cattle and
data, evidence of Iha directly interacting with human intestine, and abiotic surfaces, such as
cells in vivo or in vitro has not been shown, glass and polystyrene. Figure 2 attempts to
and the presence of the protein has not been graphically depict some of these critical in-
verified over the course of a natural infection, teractions, emphasizing the adhesin proteins
so these effects may be of an indirect nature. that have been demonstrated to be important
E. coli immunoglobulin-binding protein in intestinal colonization. The variety of sur-
(EibG) is present in 15% of eae-negative faces capable of interacting with the bacteria
strains and exists as three different subtypes suggests that EHEC possesses a complex regu-
(a, b, and g) (181, 182). Transposon mutagen- latory system to control the expression of mul-
esis experiments identified an eibG mutation tiple adhesins, and the temporal regulation
in a strain that was no longer able to form its adds another layer of complexity in under-
chain-like adhesion pattern. This deletion standing this complex network. Whereas some
also abrogated in vitro adhesion to HEp-2 cells of these systems may be redundancies built
and eliminated its ability to autoagglutinate into the pathogen, others likely have a specific
red blood cells and to interact with IgG and function that is not fully characterized. Future

FIGURE 2 Time line and proposed roles of major mbrial/ambrial adhesin proteins in EHEC. In (A), Lpf is
interacting with the extracellular membrane (ECM) proteins, such as laminin, collagen IV, and bronectin. After
the initial interaction with the intestine is established, EHEC is able to closely attach to the host cell (B) through
an interaction of the surface protein intimin with the translocated receptor protein Tir and nucleolin. In (C), the
interaction between Tir and intimin is established, initiating a host-cell actin rearrangement via participation of
the translocated bacterial protein EspFu and the recruitment of several host proteins. In (D), ECP and HCP are
proposed to interact with the surface of host intestinal cells, perhaps strengthening the colonization through
the formation of microcolonies and/or biolms. Further, curli is shown to contribute in the establishment of
biolms, though an interaction with ECP and HCP is not demonstrated. Finally in (E), dominant curli expression
has been proposed as detrimental for the survival of EHEC in the intestine, suggesting that this surface
structure might become important for the pathogen interaction with abiotic surfaces and during colonization
of the surface of vegetable and plant leaves. doi:10.1128/microbiolspec.EHEC-0003-2013.f2
research will attempt to address these ques- responsibility of the authors and do not nec-
tions through structural and functional stud- essarily represent the official views of NIAID
ies, and it seems probable that new adhesins or NIH.
and their host receptors will be discovered
as well. Finally, as we begin to understand
CITATION
EHEC pathogenesis more thoroughly, adhe-
sion models can become more refined and McWilliams BD, Torres AG. 2014. Enterohem-
can begin to tease out why, when, and how orrhagic Escherichia coli adhesins. Microbiol
this hugely diverse set of proteins are medi- Spectrum 2(3):EHEC-0003-2013.
ating EHEC colonization of the intestine or
persistence in the environment.
REFERENCES
1. Francis DH, Collins JE, Duimstra JR. 1986. In-
ACKNOWLEDGMENTS fection of gnotobiotic pigs with an Escherichia coli
O157:H7 strain associated with an outbreak of
Research work in the AGT laboratory is hemorrhagic colitis. Infect Immun 51:953956.
supported by NIH/NIAID grants AI079154 2. Yu J, Kaper JB. 1992. Cloning and character-
and AI09956001. The contents are solely the ization of the eae gene of enterohaemorrhagic

Escherichia coli O157:H7. Mol Microbiol 6:411 effacement proteins in adult beef cattle following
417. experimental infection. Vet Immunol Immuno-
3. Donnenberg MS, Tzipori S, McKee ML, pathol 118:229238.
OBrien AD, Alroy J, Kaper JB. 1993. The role of 14. Dean-Nystrom EA, Bosworth BT, Moon HW,
the eae gene of enterohemorrhagic Escherichia OBrien AD. 1998. Escherichia coli O157:H7 re-
coli in intimate attachment in vitro and in a por- quires intimin for enteropathogenicity in calves.
cine model. J Clin Invest 92:14181424. Infect Immun 66:45604563.
4. Donnenberg MS, Tacket CO, James SP, 15. Judge NA, Mason HS, OBrien AD. 2004. Plant
Losonsky G, Nataro JP, Wasserman SS, Kaper cell-based intimin vaccine given orally to mice
JB, Levine MM. 1993. Role of the eaeA gene in primed with intimin reduces time of Escherichia
experimental enteropathogenic Escherichia coli coli O157:H7 shedding in feces. Infect Immun
infection. J Clin Invest 92:14121417. 72:168175.
5. Torres AG, Kaper JB. 2002. Pathogenicity 16. Dziva F, van Diemen PM, Stevens MP, Smith
islands of intestinal E. coli. Pathogenicity islands AJ, Wallis TS. 2004. Identification of Escherichia
(PAIs) and the evolution of pathogenic microbes. coli O157:H7 genes influencing colonization of
Curr Top Microbiol Immunol 264:3148. the bovine gastrointestinal tract using signature-
6. Stevens MP, Wallis TS. 29 August 2005, posting tagged mutagenesis. Microbiology 150:36313645.
date Chapter 8.3.2.3, Adhesins of enterohemor- 17. Cornick NA, Booher SL, Moon HW. 2002.
rhagic Escherichia coli. In Bck A, et al. (ed), Intimin facilitates colonization by Escherichia
EcoSalEscherichia coli and Salmonella: Cellular coli O157:H7 in adult ruminants. Infect Immun
and Molecular Biology. ASM Press, Washington, 70:27042707.
DC. doi:10.1128/ecosal.8.3.2.3. 18. Ritchie JM, Thorpe CM, Rogers AB, Waldor
7. McDaniel TK, Jarvis KG, Donnenberg MS, MK. 2003. Critical roles for stx2, eae, and tir in
Kaper JB. 1995. A genetic locus of enterocyte enterohemorrhagic Escherichia coli-induced di-
effacement conserved among diverse enterobac- arrhea and intestinal inflammation in infant
terial pathogens. Proc Natl Acad Sci USA 92:1664 rabbits. Infect Immun 71:71297139.
1668. 19. Abe A, Heczko U, Hegele RG, Brett FB. 1998.
8. Elliott SJ, Wainwright LA, McDaniel TK, Two enteropathogenic Escherichia coli type III
Jarvis KG, Deng YK, Lai LC, McNamara BP, secreted proteins, EspA and EspB, are virulence
Donnenberg MS, Kaper JB. 1998. The complete factors. J Exp Med 188:19071916.
sequence of the locus of enterocyte effacement 20. Ebel F, Deibel C, Kresse AU, Guzman CA,
(LEE) from enteropathogenic Escherichia coli Chakraborty T. 1996. Temperature- and medi-
E2348/69. Mol Microbiol 28:14. um-dependent secretion of proteins by Shiga
9. Perna NT, Mayhew GF, Posfai G, Elliott S, toxin-producing Escherichia coli. Infect Immun
Donnenberg MS, Kaper JB, Blattner FR. 64:44724479.
1998. Molecular evolution of a pathogenicity is- 21. Kenny B, Finlay BB. 1995. Protein secretion by
land from enterohemorrhagic Escherichia coli enteropathogenic Escherichia coli is essential for
O157:H7. Infect Immun 66:38103817. transducing signals to epithelial cells. Proc Natl
10. Zhu C, Agin TS, Elliott SJ, Johnson LA, Thate Acad Sci USA 92:79917995.
TE, Kaper JB, Boedeker EC. 2001. Complete 22. Jarvis KG, Kaper JB. 1996. Secretion of extra-
nucleotide sequence and analysis of the locus of cellular proteins by enterohemorrhagic Esche-
enterocyte effacement from rabbit diarrheagenic richia coli via a putative type III secretion system.
Escherichia coli RDEC-1. Infect Immun 69:2107 Infect Immun 64:48264829.
2115. 23. Hueck CJ. 1998. Type III protein secretion sys-
11. Tauschek M, Strugnell RA, Robins-Browne tems in bacterial pathogens of animals and plants.
RM. 2002. Characterization and evidence of Microbiol Mol Biol Rev 62:379433.
mobilization of the LEE pathogenicity island of 24. Gauthier A, Finlay BB. 2003. Translocated intimin
rabbit-specific strains of enteropathogenic Esch- receptor and its chaperone interact with ATPase of
erichia coli. Mol Microbiol 44:15331550. the type III secretion apparatus of enteropatho-
12. Deng W, Li Y, Vallance BA, Finlay BB. 2001. genic Escherichia coli. J Bacteriol 185:67476755.
Locus of enterocyte effacement from Citrobacter 25. Garmendia J, Frankel G, Crepin VF. 2005. En-
rodentium: sequence analysis and evidence for teropathogenic and enterohemorrhagic Esche-
horizontal transfer among attaching and effacing richia coli infections: translocation, translocation,
pathogens. Infect Immun 69:63236335. translocation. Infect Immun 73:25732585.
13. Bretschneider G, Berberov EM, Moxley RA. 26. Sharp FC, Sperandio V. 2007. QseA directly ac-
2007. Isotype-specific antibody responses against tivates transcription of LEE1 in enterohemorrhagic
Escherichia coli O157:H7 locus of enterocyte Escherichia coli. Infect Immun 75:24322440.

27. Atlung T, Ingmer H. 1997. H-NS: a modulator of 38. Bustamante VH, Santana FJ, Calva E, Puente
environmentally regulated gene expression. Mol JL. 2001. Transcriptional regulation of type III
Microbiol 24:717. secretion genes in enteropathogenic Escherichia
28. Mellies JL, Elliott SJ, Sperandio V, Donnenberg coli: Ler antagonizes H-NS-dependent repression.
MS, Kaper JB. 1999. The Per regulon of entero- Mol Microbiol 39:664678.
pathogenic Escherichia coli: identification of a reg- 39. Elliott SJ, Sperandio V, Giron JA, Shin S,
ulatory cascade and a novel transcriptional activator, Mellies JL, Wainwright L, Hutcheson SW,
the locus of enterocyte effacement (LEE)-encoded McDaniel TK, Kaper JB. 2000. The locus of
regulator (Ler). Mol Microbiol 33:296306. enterocyte effacement (LEE)-encoded regulator
29. Deng W, Puente JL, Gruenheid S, Li Y, Vallance controls expression of both LEE- and non-LEE-
BA, Vazquez A, Barba J, Ibarra JA, ODonnell encoded virulence factors in enteropathogenic
P, Metalnikov P, Ashman K, Lee S, Goode D, and enterohemorrhagic Escherichia coli. Infect
Pawson T, Finlay BB. 2004. Dissecting virulence: Immun 68:61156126.
systematic and functional analyses of a pathoge- 40. Friedberg D, Umanski T, Fang Y, Rosenshine I.
nicity island. Proc Natl Acad Sci USA 101:3597 1999. Hierarchy in the expression of the locus of
3602. enterocyte effacement genes of enteropathogenic
30. Kenny B, Abe A, Stein M, Finlay BB. 1997. En- Escherichia coli. Mol Microbiol 34:941952.
teropathogenic Escherichia coli protein secretion 41. Njoroge JW, Nguyen Y, Curtis MM, Moreira
is induced in response to conditions similar to CG, Sperandio V. 2012. Virulence meets metab-
those in the gastrointestinal tract. Infect Immun olism: Cra and KdpE gene regulation in entero-
65:26062612. hemorrhagic Escherichia coli. MBio 3:e00280-12.
31. Bansal T, Jesudhasan P, Pillai S, Wood TK, 42. Laaberki MH, Janabi N, Oswald E, Repoila F.
Jayaraman A. 2008. Temporal regulation of en- 2006. Concert of regulators to switch on LEE
terohemorrhagic Escherichia coli virulence medi- expression in enterohemorrhagic Escherichia coli
ated by autoinducer-2. Appl Microbiol Biotechnol O157:H7: interplay between Ler, GrlA, HNS and
78:811819. RpoS. Int J Med Microbiol 296:197210.
32. Sperandio V, Li CC, Kaper JB. 2002. Quorum- 43. Kendall MM, Gruber CC, Rasko DA, Hughes
sensing Escherichia coli regulator A: a regulator of DT, Sperandio V. 2011. Hfq virulence regulation
the LysR family involved in the regulation of the in enterohemorrhagic Escherichia coli O157:H7
locus of enterocyte effacement pathogenicity is- strain 86-24. J Bacteriol 193:68436851.
land in enterohemorrhagic E. coli. Infect Immun 44. Barba J, Bustamante VH, Flores-Valdez MA,
70:30853093. Deng W, Finlay BB, Puente JL. 2005. A positive
33. Abe H, Tatsuno I, Tobe T, Okutani A, Sasakawa regulatory loop controls expression of the locus of
C. 2002. Bicarbonate ion stimulates the expres- enterocyte effacement-encoded regulators Ler
sion of locus of enterocyte effacement-encoded and GrlA. J Bacteriol 187:79187930.
genes in enterohemorrhagic Escherichia coli O157: 45. Dorman CJ. 2004. H-NS: a universal regulator
H7. Infect Immun 70:35003509. for a dynamic genome. Nat Rev Microbiol 2:391
34. Yin X, Feng Y, Wheatcroft R, Chambers J, 400.
Gong J, Gyles CL. 2011. Adherence of Escherichia 46. Garcia J, Cordeiro TN, Prieto MJ, Pons M.
coli O157:H7 to epithelial cells in vitro and in pig 2012. Oligomerization and DNA binding of Ler, a
gut loops is affected by bacterial culture condi- master regulator of pathogenicity of enterohem-
tions. Can J Vet Res 75:8188. orrhagic and enteropathogenic Escherichia coli.
35. Bansal T, Kim DN, Slininger T, Wood TK, Nucleic Acids Res 40:1025410262.
Jayaraman A. 2012. Human intestinal epithelial 47. Jerse AE, Yu J, Tall BD, Kaper JB. 1990. A ge-
cell-derived molecule(s) increase enterohemor- netic locus of enteropathogenic Escherichia coli
rhagic Escherichia coli virulence. FEMS Immunol necessary for the production of attaching and
Med Microbiol 66:399410. effacing lesions on tissue culture cells. Proc Natl
36. Bhatt S, Edwards AN, Nguyen HT, Merlin D, Acad Sci USA 87:78397843.
Romeo T, Kalman D. 2009. The RNA binding 48. Tzipori S, Gunzer F, Donnenberg MS, de
protein CsrA is a pleiotropic regulator of the locus Montigny L, Kaper JB, Donohue-Rolfe A. 1995.
of enterocyte effacement pathogenicity island of The role of the eaeA gene in diarrhea and neu-
enteropathogenic Escherichia coli. Infect Immun rological complications in a gnotobiotic piglet
77:35523568. model of enterohemorrhagic Escherichia coli in-
37. Russell RM, Sharp FC, Rasko DA, Sperandio V. fection. Infect Immun 63:36213627.
2007. QseA and GrlR/GrlA regulation of the locus 49. McKee ML, Melton-Celsa AR, Moxley RA,
of enterocyte effacement genes in enterohemor- Francis DH, OBrien AD. 1995. Enterohemor-
rhagic Escherichia coli. J Bacteriol 189:53875392. rhagic Escherichia coli O157:H7 requires intimin

to colonize the gnotobiotic pig intestine and to 61. DeVinney R, Stein M, Reinscheid D, Abe A,
adhere to HEp-2 cells. Infect Immun 63:3739 Ruschkowski S, Finlay BB. 1999. Enterohemor-
3744. rhagic Escherichia coli O157:H7 produces Tir,
50. Gansheroff LJ, Wachtel MR, OBrien AD. which is translocated to the host cell membrane
1999. Decreased adherence of enterohemorrhagic but is not tyrosine phosphorylated. Infect Immun
Escherichia coli to HEp-2 cells in the presence of 67:23892398.
antibodies that recognize the C-terminal region of 62. Kenny B, Warawa J. 2001. Enteropathogenic
intimin. Infect Immun 67:64096417. Escherichia coli (EPEC) Tir receptor molecule
51. Buist G, Steen A, Kok J, Kuipers OP. 2008. LysM, does not undergo full modification when intro-
a widely distributed protein motif for binding to duced into host cells by EPEC-independent
(peptido)glycans. Mol Microbiol 68:838847. mechanisms. Infect Immun 69:14441453.
52. Touze T, Hayward RD, Eswaran J, Leong JM, 63. de Groot JC, Schluter K, Carius Y, Quedenau C,
Koronakis V. 2004. Self-association of EPEC Vingadassalom D, Faix J, Weiss SM, Reichelt J,
intimin mediated by the beta-barrel-containing Standfuss-Gabisch C, Lesser CF, Leong JM,
anchor domain: a role in clustering of the Tir Heinz DW, Bussow K, Stradal TE. 2011. Struc-
receptor. Mol Microbiol 51:7387. tural basis for complex formation between human
53. Kelly G, Prasannan S, Daniell S, Fleming K, IRSp53 and the translocated intimin receptor Tir
Frankel G, Dougan G, Connerton I, Matthews of enterohemorrhagic E. coli. Structure 19:1294
S. 1999. Structure of the cell-adhesion fragment of 1306.
intimin from enteropathogenic Escherichia coli. 64. Garmendia J, Phillips AD, Carlier MF, Chong Y,
Nat Struct Biol 6:313318. Schuller S, Marches O, Dahan S, Oswald E,
54. Batchelor M, Prasannan S, Daniell S, Reece S, Shaw RK, Knutton S, Frankel G. 2004. TccP is
Connerton I, Bloomberg G, Dougan G, Frankel an enterhemorrhagic Escherichia coli O157:H7
G, Matthews S. 2000. Structural basis for rec- type III effector protein that couples Tir to the
ognition of the translocated intimin receptor (Tir) actin cytoskeleton. Cell Microbiol 6:11671183.
by intimin from enteropathogenic Escherichia 65. Weiss SM, Ladwein M, Schmidt D, Ehinger J,
coli. EMBO J 19:24522464. Lommel S, Stading K, Beutling U, Disanza A,
55. Luo Y, Frey EA, Pfuetzner RA, Creagh AL, Frank R, Jansch L, Scita G, Gunzer F, Rottner
Knoechel DG, Haynes CA, Finlay BB, Strynadka K, Stradal TE. 2009. IRSp53 links the entero-
NC. 2000. Crystal structure of enteropathogenic hemorrhagic E. coli effectors Tir and EspFU
Escherichia coli intimin-receptor complex. Nature for actin pedestal formation. Cell Host Microbe
405:10731077. 5:244258.
56. Rosenshine I, Ruschkowski S, Stein M, 66. Vingadassalom D, Kazlauskas A, Skehan B,
Reinscheid DJ, Mills SD, Finlay BB. 1996. A Cheng HC, Magoun L, Robbins D, Rosen MK,
pathogenic bacterium triggers epithelial signals to Saksela K, Leong JM. 2009. Insulin receptor
form a functional bacterial receptor that mediates tyrosine kinase substrate links the E. coli O157:H7
actin pseudopod formation. EMBO J 15:26132624. actin assembly effectors Tir and EspF(U) during
57. Campellone KG, Leong JM. 2003. Tails of two pedestal formation. Proc Natl Acad Sci USA
Tirs: actin pedestal formation by enteropatho- 106:67546759.
genic E. coli and enterohemorrhagic E. coli O157: 67. Frankel G, Candy DC, Everest P, Dougan G.
H7. Curr Opin Microbiol 6:8290. 1994. Characterization of the C-terminal domains
58. Campellone KG, Brady MJ, Alamares JG, Rowe of intimin-like proteins of enteropathogenic and
DC, Skehan BM, Tipper DJ, Leong JM. 2006. enterohemorrhagic Escherichia coli, Citrobacter
Enterohaemorrhagic Escherichia coli Tir requires freundii, and Hafnia alvei. Infect Immun 62:1835
a C-terminal 12-residue peptide to initiate EspF- 1842.
mediated actin assembly and harbours N-terminal 68. Frankel G, Candy DC, Fabiani E, Adu-Bobie J,
sequences that influence pedestal length. Cell Gil S, Novakova M, Phillips AD, Dougan G. 1995.
Microbiol 8:14881503. Molecular characterization of a carboxy-terminal
59. Kenny B. 1999. Phosphorylation of tyrosine 474 eukaryotic-cell-binding domain of intimin from
of the enteropathogenic Escherichia coli (EPEC) enteropathogenic Escherichia coli. Infect Immun
Tir receptor molecule is essential for actin nu- 63:43234328.
cleating activity and is preceded by additional 69. Frankel G, Lider O, Hershkoviz R, Mould
host modifications. Mol Microbiol 31:12291241. AP, Kachalsky SG, Candy DC, Cahalon L,
60. Ismaili A, Philpott DJ, Dytoc MT, Sherman PM. Humphries MJ, Dougan G. 1996. The cell-
1995. Signal transduction responses following binding domain of intimin from enteropathogenic
adhesion of verocytotoxin-producing Escherichia Escherichia coli binds to beta1 integrins. J Biol
coli. Infect Immun 63:33163326. Chem 271:2035920364.

70. Liu H, Magoun L, Leong JM. 1999. beta1-chain and delta, four intimin derivatives expressed by
integrins are not essential for intimin-mediated attaching and effacing microbial pathogens. J Clin
host cell attachment and enteropathogenic Esch- Microbiol 36:662668.
erichia coli-induced actin condensation. Infect 81. Blanco M, Blanco JE, Blanco J, de Carvalho
Immun 67:20452049. VM, Onuma DL, Pestana de Castro AF. 2004.
71. Hartland EL, Batchelor M, Delahay RM, Typing of intimin (eae) genes in attaching and
Hale C, Matthews S, Dougan G, Knutton S, effacing Escherichia coli strains from monkeys.
Connerton I, Frankel G. 1999. Binding of intimin J Clin Microbiol 42:13821383.
from enteropathogenic Escherichia coli to Tir and 82. China B, Goffaux F, Pirson V, Mainil J. 1999.
to host cells. Mol Microbiol 32:151158. Comparison of eae, tir, espA and espB genes of
72. Sinclair JF, Dean-Nystrom EA, OBrien AD. bovine and human attaching and effacing Esche-
2006. The established intimin receptor Tir and the richia coli by multiplex polymerase chain reac-
putative eucaryotic intimin receptors nucleolin tion. FEMS Microbiol Lett 178:177182.
and beta1 integrin localize at or near the site of 83. Oswald E, Schmidt H, Morabito S, Karch H,
enterohemorrhagic Escherichia coli O157:H7 ad- Marches O, Caprioli A. 2000. Typing of intimin
herence to enterocytes in vivo. Infect Immun genes in human and animal enterohemorrhagic
74:12551265. and enteropathogenic Escherichia coli: character-
73. Gu L, Wang H, Guo YL, Zen K. 2008. Heparin ization of a new intimin variant. Infect Immun
blocks the adhesion of E. coli O157:H7 to human 68:6471.
colonic epithelial cells. Biochem Biophys Res 84. Reid SD, Betting DJ, Whittam TS. 1999. Mo-
Commun 369:10611064. lecular detection and identification of intimin
74. Sinclair JF, OBrien AD. 2002. Cell surface- alleles in pathogenic Escherichia coli by multiplex
localized nucleolin is a eukaryotic receptor for the PCR. J Clin Microbiol 37:27192722.
adhesin intimin-gamma of enterohemorrhagic 85. Zhang WL, Kohler B, Oswald E, Beutin L,
Escherichia coli O157:H7. J Biol Chem 277:2876 Karch H, Morabito S, Caprioli A, Suerbaum S,
2885. Schmidt H. 2002. Genetic diversity of intimin
75. Sinclair JF, OBrien AD. 2004. Intimin types genes of attaching and effacing Escherichia coli
alpha, beta, and gamma bind to nucleolin with strains. J Clin Microbiol 40:44864492.
equivalent affinity but lower avidity than to the 86. Ito K, Iida M, Yamazaki M, Moriya K,
translocated intimin receptor. J Biol Chem Moroishi S, Yatsuyanagi J, Kurazono T, Hiruta
279:3375133758. N, Ratchtrachenchai OA. 2007. Intimin types
76. Liu B, Yin X, Feng Y, Chambers JR, Guo A, determined by heteroduplex mobility assay of
Gong J, Zhu J, Gyles CL. 2010. Verotoxin 2 intimin gene (eae)-positive Escherichia coli strains.
enhances adherence of enterohemorrhagic Esch- J Clin Microbiol 45:10381041.
erichia coli O157:H7 to intestinal epithelial cells 87. Torres AG, Zhou X, Kaper JB. 2005. Adherence
and expression of {beta}1-integrin by IPEC-J2 of diarrheagenic Escherichia coli strains to epi-
cells. Appl Environ Microbiol 76:44614468. thelial cells. Infect Immun 73:1829.
77. Jenkins C, Chart H, Smith HR, Hartland EL, 88. Phillips AD, Frankel G. 2000. Intimin-mediated
Batchelor M, Delahay RM, Dougan G, Frankel tissue specificity in enteropathogenic Escherichia
G. 2000. Antibody response of patients infected coli interaction with human intestinal organ
with verocytotoxin-producing Escherichia coli cultures. J Infect Dis 181:14961500.
to protein antigens encoded on the LEE locus. 89. Mundy R, Schuller S, Girard F, Fairbrother JM,
J Med Microbiol 49:97101. Phillips AD, Frankel G. 2007. Functional studies
78. Karpman D, Bekassy ZD, Sjogren AC, Dubois of intimin in vivo and ex vivo: implications for
MS, Karmali MA, Mascarenhas M, Jarvis KG, host specificity and tissue tropism. Microbiology
Gansheroff LJ, OBrien AD, Arbus GS, Kaper 153:959967.
JB. 2002. Antibodies to intimin and Escherichia 90. Ahmed S, Byrd W, Kumar S, Boedeker EC. 2013. A
coli secreted proteins A and B in patients with directed intimin insertion mutant of a rabbit entero-
enterohemorrhagic Escherichia coli infections. pathogenic Escherichia coli (REPEC) is attenuated,
Pediatr Nephrol 17:201211. immunogenic and elicits serogroup specific protec-
79. Li Y, Frey E, Mackenzie AM, Finlay BB. 2000. tion. Vet Immunol Immunopathol 152:146155.
Human response to Escherichia coli O157:H7 in- 91. Guirro M, de Souza RL, Piazza RM, Guth BE.
fection: antibodies to secreted virulence factors. 2013. Antibodies to intimin and Escherichia coli-
Infect Immun 68:50905095. secreted proteins EspA and EspB in sera of
80. Adu-Bobie J, Frankel G, Bain C, Goncalves AG, Brazilian children with hemolytic uremic syndrome
Trabulsi LR, Douce G, Knutton S, Dougan G. and healthy controls. Vet Immunol Immunopathol
1998. Detection of intimins alpha, beta, gamma, 152:121125.

92. McKee ML, OBrien AD. 1996. Truncated entero- producing Escherichia coli of serotype O113:H21.
hemorrhagic Escherichia coli (EHEC) O157:H7 Infect Immun 62:34943505.
intimin (EaeA) fusion proteins promote adher- 102. Paton AW, Woodrow MC, Doyle RM, Lanser
ence of EHEC strains to HEp-2 cells. Infect JA, Paton JC. 1999. Molecular characterization
Immun 64:22252233. of a Shiga toxigenic Escherichia coli O113:H21
93. Dean-Nystrom EA, Gansheroff LJ, Mills M, strain lacking eae responsible for a cluster of cases
Moon HW, OBrien AD. 2002. Vaccination of of hemolytic-uremic syndrome. J Clin Microbiol
pregnant dams with intimin(O157) protects suck- 37:33573361.
ling piglets from Escherichia coli O157:H7 infec- 103. Fitzhenry RJ, Reece S, Trabulsi LR, Heuschkel
tion. Infect Immun 70:24142418. R, Murch S, Thomson M, Frankel G, Phillips
94. Ghaem-Maghami M, Simmons CP, Daniell S, AD. 2002. Tissue tropism of enteropathogenic
Pizza M, Lewis D, Frankel G, Dougan G. 2001. Escherichia coli strains belonging to the O55
Intimin-specific immune responses prevent bac- serogroup. Infect Immun 70:43624368.
terial colonization by the attaching-effacing path- 104. Hayashi T, Makino K, Ohnishi M, Kurokawa K,
ogen Citrobacter rodentium. Infect Immun 69: Ishii K, Yokoyama K, Han CG, Ohtsubo E,
55975605. Nakayama K, Murata T, Tanaka M, Tobe T,
95. Agin TS, Zhu C, Johnson LA, Thate TE, Yang Z, Iida T, Takami H, Honda T, Sasakawa C,
Boedeker EC. 2005. Protection against hemor- Ogasawara N, Yasunaga T, Kuhara S, Shiba T,
rhagic colitis in an animal model by oral immu- Hattori M, Shinagawa H. 2001. Complete ge-
nization with isogeneic rabbit enteropathogenic nome sequence of enterohemorrhagic Escherichia
Escherichia coli attenuated by truncating intimin. coli O157:H7 and genomic comparison with a
Infect Immun 73:66086619. laboratory strain K-12. DNA Res 8:1122.
96. Vilte DA, Larzabal M, Cataldi AA, Mercado EC. 105. Perna NT, Plunkett G III, Burland V, Mau B,
2008. Bovine colostrum contains immunoglobulin Glasner JD, Rose DJ, Mayhew GF, Evans PS,
G antibodies against intimin, EspA, and EspB and Gregor J, Kirkpatrick HA, Posfai G, Hackett J,
inhibits hemolytic activity mediated by the type Klink S, Boutin A, Shao Y, Miller L, Grotbeck EJ,
three secretion system of attaching and effacing Davis NW, Lim A, Dimalanta ET, Potamousis
Escherichia coli. Clin Vaccine Immunol 15:1208 KD, Apodaca J, Anantharaman TS, Lin J, Yen G,
1213. Schwartz DC, Welch RA, Blattner FR. 2001.
97. Rabinovitz BC, Gerhardt E, Tironi FC, Abdala Genome sequence of enterohaemorrhagic Esche-
A, Galarza R, Vilte DA, Ibarra C, Cataldi A, richia coli O157:H7. Nature 409:529533.
Mercado EC. 2012. Vaccination of pregnant cows 106. Baumler AJ, Heffron F. 1995. Identification
with EspA, EspB, gamma-intimin, and Shiga toxin and sequence analysis of lpfABCDE, a putative
2 proteins from Escherichia coli O157:H7 induces fimbrial operon of Salmonella typhimurium. J Bac-
high levels of specific colostral antibodies that teriol 177:20872097.
are transferred to newborn calves. J Dairy Sci 107. Baumler AJ, Tsolis RM, Heffron F. 1996. The lpf
95:33183326. fimbrial operon mediates adhesion of Salmonella
98. Amani J, Mousavi SL, Rafati S, Salmanian AH. typhimurium to murine Peyers patches. Proc
2011. Immunogenicity of a plant-derived edible Natl Acad Sci USA 93:279283.
chimeric EspA, Intimin and Tir of Escherichia coli 108. Torres AG, Giron JA, Perna NT, Burland V,
O157:H7 in mice. Plant Sci 180:620627. Blattner FR, Avelino-Flores F, Kaper JB. 2002.
99. Hur J, Lee JH. 2011. Immune responses to new Identification and characterization of lpfABCC'DE,
vaccine candidates constructed by a live attenu- a fimbrial operon of enterohemorrhagic Esche-
ated Salmonella Typhimurium delivery system richia coli O157:H7. Infect Immun 70:54165427.
expressing Escherichia coli F4, F5, F6, F41 and 109. Doughty S, Sloan J, Bennett-Wood V,
intimin adhesin antigens in a murine model. J Vet Robertson M, Robins-Browne RM, Hartland
Med Sci 73:12651273. EL. 2002. Identification of a novel fimbrial gene
100. Oliveira AF, Cardoso SA, Almeida FB, de cluster related to long polar fimbriae in locus of
Oliveira LL, Pitondo-Silva A, Soares SG, Hanna enterocyte effacement-negative strains of entero-
ES. 2012. Oral immunization with attenuated hemorrhagic Escherichia coli. Infect Immun 70:
Salmonella vaccine expressing Escherichia coli 67616769.
O157:H7 intimin gamma triggers both systemic and 110. Torres AG, Kanack KJ, Tutt CB, Popov V,
mucosal humoral immunity in mice. Microbiol Kaper JB. 2004. Characterization of the second
Immunol 56:513522. long polar (LP) fimbriae of Escherichia coli
101. Dytoc MT, Ismaili A, Philpott DJ, Soni R, O157:H7 and distribution of LP fimbriae in other
Brunton JL, Sherman PM. 1994. Distinct bind- pathogenic E. coli strains. FEMS Microbiol Lett
ing properties of eaeA-negative verocytotoxin- 238:333344.

111. Osek J, Weiner M, Hartland EL. 2003. Preva- Escherichia coli O157:H7 expresses curli and exhibits
lence of the lpfO113 gene cluster among Esche- reduced in vivo colonization. Infect Immun 80:914
richia coli O157 isolates from different sources. 920.
Vet Microbiol 96:259266. 122. Barnhart MM, Chapman MR. 2006. Curli bio-
112. Torres AG, Blanco M, Valenzuela P, Slater TM, genesis and function. Annu Rev Microbiol 60:131
Patel SD, Dahbi G, Lopez C, Barriga XF, Blanco 147.
JE, Gomes TA, Vidal R, Blanco J. 2009. Genes 123. Fitzhenry RJ, Stevens MP, Jenkins C, Wallis TS,
related to long polar fimbriae of pathogenic Esch- Heuschkel R, Murch S, Thomson M, Frankel G,
erichia coli strains as reliable markers to identify Phillips AD. 2003. Human intestinal tissue tro-
virulent isolates. J Clin Microbiol 47:24422451. pism of intimin epsilon O103 Escherichia coli.
113. Galli L, Torres AG, Rivas M. 2010. Identification FEMS Microbiol Lett 218:311316.
of the long polar fimbriae gene variants in the locus 124. Olsen A, Jonsson A, Normark S. 1989. Fibro-
of enterocyte effacement-negative Shiga toxin- nectin binding mediated by a novel class of surface
producing Escherichia coli strains isolated from organelles on Escherichia coli. Nature 338:652
humans and cattle in Argentina. FEMS Microbiol 655.
Lett 308:123129. 125. Olsen A, Wick MJ, Morgelin M, Bjorck L. 1998.
114. Torres AG, Milflores-Flores L, Garcia-Gallegos Curli, fibrous surface proteins of Escherichia coli,
JG, Patel SD, Best A, La Ragione RM, Martinez- interact with major histocompatibility complex
Laguna Y, Woodward MJ. 2007. Environmental class I molecules. Infect Immun 66:944949.
regulation and colonization attributes of the long 126. Collinson SK, Clouthier SC, Doran JL, Banser
polar fimbriae (LPF) of Escherichia coli O157:H7. PA, Kay WW. 1997. Characterization of the
Int J Med Microbiol 297:177185. agfBA fimbrial operon encoding thin aggregative
115. Torres AG, Lopez-Sanchez GN, Milflores- fimbriae of Salmonella enteritidis. Adv Exp Med
Flores L, Patel SD, Rojas-Lopez M, Martinez de Biol 412:247248.
la Pena CF, Arenas-Hernandez MM, Martinez- 127. Romling U, Bian Z, Hammar M, Sierralta WD,
Laguna Y. 2007. Ler and H-NS, regulators con- Normark S. 1998. Curli fibers are highly con-
trolling expression of the long polar fimbriae of served between Salmonella typhimurium and
Escherichia coli O157:H7. J Bacteriol 189:5916 Escherichia coli with respect to operon structure
5928. and regulation. J Bacteriol 180:722731.
116. Rojas-Lopez M, Arenas-Hernandez MM, 128. Dong T, Schellhorn HE. 2009. Global effect of
Medrano-Lopez A, Martinez de la Pena CF, RpoS on gene expression in pathogenic Esche-
Puente JL, Martinez-Laguna Y, Torres AG. richia coli O157:H7 strain EDL933. BMC Genomics
2011. Regulatory control of the Escherichia coli 10:349.
O157:H7 lpf1 operon by H-NS and Ler. J Bacteriol 129. Olsen A, Arnqvist A, Hammar M, Sukupolvi S,
193:16221632. Normark S. 1993. The RpoS sigma factor relieves
117. Torres AG, Slater TM, Patel SD, Popov VL, H-NS-mediated transcriptional repression of csgA,
Arenas-Hernandez MM. 2008. Contribution of the subunit gene of fibronectin-binding curli in
the Ler- and H-NS-regulated long polar fimbriae Escherichia coli. Mol Microbiol 7:523536.
of Escherichia coli O157:H7 during binding to tis- 130. Lee JH, Cho MH, Lee J. 2011. 3-indolylacetonitrile
sue-cultured cells. Infect Immun 76:50625071. decreases Escherichia coli O157:H7 biofilm forma-
118. Farfan MJ, Cantero L, Vidal R, Botkin DJ, Torres tion and Pseudomonas aeruginosa virulence. Envi-
AG. 2011. Long polar fimbriae of enterohemor- ron Microbiol 13:6273.
rhagic Escherichia coli O157:H7 bind to extracellular 131. Hammar M, Arnqvist A, Bian Z, Olsen A,
matrix proteins. Infect Immun 79:37443750. Normark S. 1995. Expression of two csg operons
119. Jordan DM, Cornick N, Torres AG, Dean- is required for production of fibronectin- and
Nystrom EA, Kaper JB, Moon HW. 2004. Long congo red-binding curli polymers in Escherichia
polar fimbriae contribute to colonization by coli K-12. Mol Microbiol 18:661670.
Escherichia coli O157:H7 in vivo. Infect Immun 132. Vidal O, Longin R, Prigent-Combaret C, Dorel
72:61686171. C, Hooreman M, Lejeune P. 1998. Isolation of an
120. Fitzhenry R, Dahan S, Torres AG, Chong Y, Escherichia coli K-12 mutant strain able to form
Heuschkel R, Murch SH, Thomson M, Kaper biofilms on inert surfaces: involvement of a new
JB, Frankel G, Phillips AD. 2006. Long polar ompR allele that increases curli expression. J Bac-
fimbriae and tissue tropism in Escherichia coli teriol 180:24422449.
O157:H7. Microbes Infect 8:17411749. 133. Dorel C, Vidal O, Prigent-Combaret C, Vallet I,
121. Lloyd SJ, Ritchie JM, Rojas-Lopez M, Blumentritt Lejeune P. 1999. Involvement of the Cpx signal
CA, Popov VL, Greenwich JL, Waldor MK, Torres transduction pathway of E. coli in biofilm forma-
AG. 2012. A double, long polar fimbria mutant of tion. FEMS Microbiol Lett 178:169175.

134. Brown PK, Dozois CM, Nickerson CA, Zuppardo Escherichia coli O26:H genes required for intes-
A, Terlonge J, Curtiss R III. 2001. MlrA, a novel tinal colonization in calves. Infect Immun 73:
regulator of curli (AgF) and extracellular matrix 17351743.
synthesis by Escherichia coli and Salmonella 145. Low AS, Dziva F, Torres AG, Martinez JL,
enterica serovar Typhimurium. Mol Microbiol Rosser T, Naylor S, Spears K, Holden N, Mahajan
41:349363. A, Findlay J, Sales J, Smith DG, Low JC, Stevens
135. Saldana Z, Xicohtencatl-Cortes J, Avelino F, MP, Gally DL. 2006. Cloning, expression, and char-
Phillips AD, Kaper JB, Puente JL, Giron JA. acterization of fimbrial operon F9 from entero-
2009. Synergistic role of curli and cellulose in cell hemorrhagic Escherichia coli O157:H7. Infect
adherence and biofilm formation of attaching and Immun 74:22332244.
effacing Escherichia coli and identification of Fis 146. Vazquez-Juarez RC, Kuriakose JA, Rasko DA,
as a negative regulator of curli. Environ Microbiol Ritchie JM, Kendall MM, Slater TM, Sinha M,
11:9921006. Luxon BA, Popov VL, Waldor MK, Sperandio
136. Sharma VK, Bearson SM, Bearson BL. 2010. V, Torres AG. 2008. CadA negatively regulates
Evaluation of the effects of sdiA, a luxR homo- Escherichia coli O157:H7 adherence and intestinal
logue, on adherence and motility of Escherichia colonization. Infect Immun 76:50725081.
coli O157:H7. Microbiology 156:13031312. 147. Samadder P, Xicohtencatl-Cortes J, Saldana Z,
137. Ryu JH, Beuchat LR. 2005. Biofilm formation by Jordan D, Tarr PI, Kaper JB, Giron JA. 2009.
Escherichia coli O157:H7 on stainless steel: effect The Escherichia coli ycbQRST operon encodes
of exopolysaccharide and curli production on its fimbriae with laminin-binding and epithelial cell
resistance to chlorine. Appl Environ Microbiol adherence properties in Shiga-toxigenic E. coli
71:247254. O157:H7. Environ Microbiol 11:18151826.
138. Pawar DM, Rossman ML, Chen J. 2005. Role of 148. Brunder W, Khan AS, Hacker J, Karch H. 2001.
curli fimbriae in mediating the cells of entero- Novel type of fimbriae encoded by the large plas-
haemorrhagic Escherichia coli to attach to abiotic mid of sorbitol-fermenting enterohemorrhagic
surfaces. J Appl Microbiol 99:418425. Escherichia coli O157:H(). Infect Immun 69:4447
139. Fink RC, Black EP, Hou Z, Sugawara M, 4457.
Sadowsky MJ, Diez-Gonzalez F. 2012. Tran- 149. Bielaszewska M, Prager R, Vandivinit L, Musken
scriptional responses of Escherichia coli K-12 and A, Mellmann A, Holt NJ, Tarr PI, Karch H, Zhang
O157:H7 associated with lettuce leaves. Appl En- W. 2009. Detection and characterization of the
viron Microbiol 78:17521764. fimbrial sfp cluster in enterohemorrhagic Esche-
140. Low AS, Holden N, Rosser T, Roe AJ, Constantinidou richia coli O165:H25/NM isolates from humans and
C, Hobman JL, Smith DG, Low JC, Gally DL. 2006. cattle. Appl Environ Microbiol 75:6471.
Analysis of fimbrial gene clusters and their ex- 150. Musken A, Bielaszewska M, Greune L, Schweppe
pression in enterohaemorrhagic Escherichia coli CH, Muthing J, Schmidt H, Schmidt MA, Karch
O157:H7. Environ Microbiol 8:10331047. H, Zhang W. 2008. Anaerobic conditions promote
141. Rendon MA, Saldana Z, Erdem AL, Monteiro- expression of Sfp fimbriae and adherence of sorbi-
Neto V, Vazquez A, Kaper JB, Puente JL, Giron tol-fermenting enterohemorrhagic Escherichia coli
JA. 2007. Commensal and pathogenic Escherichia O157:NM to human intestinal epithelial cells. Appl
coli use a common pilus adherence factor for Environ Microbiol 74:10871093.
epithelial cell colonization. Proc Natl Acad Sci 151. Xicohtencatl-Cortes J, Monteiro-Neto V,
USA 104:1063710642. Ledesma MA, Jordan DM, Francetic O, Kaper
142. Pouttu R, Westerlund-Wikstrom B, Lang H, JB, Puente JL, Giron JA. 2007. Intestinal ad-
Alsti K, Virkola R, Saarela U, Siitonen A, herence associated with type IV pili of entero-
Kalkkinen N, Korhonen TK. 2001. matB, a hemorrhagic Escherichia coli O157:H7. J Clin
common fimbrillin gene of Escherichia coli, ex- Invest 117:35193529.
pressed in a genetically conserved, virulent clonal 152. Xicohtencatl-Cortes J, Monteiro-Neto V, Saldana
group. J Bacteriol 183:47274736. Z, Ledesma MA, Puente JL, Giron JA. 2009. The
143. Saldana Z, Erdem AL, Schuller S, Okeke IN, type 4 pili of enterohemorrhagic Escherichia coli
Lucas M, Sivananthan A, Phillips AD, Kaper JB, O157:H7 are multipurpose structures with patho-
Puente JL, Giron JA. 2009. The Escherichia coli genic attributes. J Bacteriol 191:411421.
common pilus and the bundle-forming pilus act in 153. Galfi P, Neogrady S, Semjen G, Bardocz S,
concert during the formation of localized adher- Pusztai A. 1998. Attachment of different Esche-
ence by enteropathogenic E. coli. J Bacteriol richia coli strains to cultured rumen epithelial
191:34513461. cells. Vet Microbiol 61:191197.
144. van Diemen PM, Dziva F, Stevens MP, Wallis 154. Enami M, Nakasone N, Honma Y, Kakinohana
TS. 2005. Identification of enterohemorrhagic S, Kudaka J, Iwanaga M. 1999. Expression of

type I pili is abolished in verotoxin-producing Bielaszewska M, Karch H. 2011. Enterohaemor-

Escherichia coli O157. FEMS Microbiol Lett 179: rhagic Escherichia coli haemolysin is cleaved and
467472. inactivated by serine protease EspPalpha. Environ
155. Li B, Koch WH, Cebula TA. 1997. Detection Microbiol 13:13271341.
and characterization of the fimA gene of Esch- 166. Dziva F, Mahajan A, Cameron P, Currie C,
erichia coli O157:H7. Mol Cell Probes 11:397 McKendrick IJ, Wallis TS, Smith DG, Stevens
406. MP. 2007. EspP, a Type V-secreted serine pro-
156. Roe AJ, Currie C, Smith DG, Gally DL. 2001. tease of enterohaemorrhagic Escherichia coli
Analysis of type 1 fimbriae expression in vero- O157:H7, influences intestinal colonization of
toxigenic Escherichia coli: a comparison between calves and adherence to bovine primary intesti-
serotypes O157 and O26. Microbiology 147:145 nal epithelial cells. FEMS Microbiol Lett 271:258
152. 264.
157. Cookson AL, Cooley WA, Woodward MJ. 2002. 167. Xicohtencatl-Cortes J, Saldana Z, Deng W,
The role of type 1 and curli fimbriae of Shiga Castaneda E, Freer E, Tarr PI, Finlay BB,
toxin-producing Escherichia coli in adherence Puente JL, Giron JA. 2010. Bacterial macro-
to abiotic surfaces. Int J Med Microbiol 292:195 scopic rope-like fibers with cytopathic and ad-
205. hesive properties. J Biol Chem 285:3233632342.
158. Leyton DL, Sevastsyanovich YR, Browning 168. Paton AW, Srimanote P, Woodrow MC,
DF, Rossiter AE, Wells TJ, Fitzpatrick RE, Paton JC. 2001. Characterization of Saa, a novel
Overduin M, Cunningham AF, Henderson IR. autoagglutinating adhesin produced by locus of
2011. Size and conformation limits to secretion of enterocyte effacement-negative Shiga-toxigenic
disulfide-bonded loops in autotransporter pro- Escherichia coli strains that are virulent for hu-
teins. J Biol Chem 286:4228342291. mans. Infect Immun 69:69997009.
159. Henderson IR, Nataro JP. 2001. Virulence 169. Lucchesi PM, Kruger A, Parma AE. 2006. Dis-
functions of autotransporter proteins. Infect tribution of saa gene variants in verocytotoxigenic
Immun 69:12311243. Escherichia coli isolated from cattle and food. Res
160. Wells TJ, Sherlock O, Rivas L, Mahajan A, Microbiol 157:263266.
Beatson SA, Torpdahl M, Webb RI, Allsopp LP, 170. Toma C, Nakasone N, Miliwebsky E, Higa N,
Gobius KS, Gally DL, Schembri MA. 2008. EhaA Rivas M, Suzuki T. 2008. Differential adherence
is a novel autotransporter protein of enterohem- of Shiga toxin-producing Escherichia coli harbor-
orrhagic Escherichia coli O157:H7 that contrib- ing saa to epithelial cells. Int J Med Microbiol
utes to adhesion and biofilm formation. Environ 298:571578.
Microbiol 10:589604. 171. Herold S, Paton JC, Paton AW. 2009. Sab, a novel
161. Wells TJ, McNeilly TN, Totsika M, Mahajan A, autotransporter of locus of enterocyte effacement-
Gally DL, Schembri MA. 2009. The Escherichia negative shiga-toxigenic Escherichia coli O113:H21,
coli O157:H7 EhaB autotransporter protein binds contributes to adherence and biofilm formation.
to laminin and collagen I and induces a serum IgA Infect Immun 77:32343243.
response in O157:H7 challenged cattle. Environ 172. Torres AG, Perna NT, Burland V, Ruknudin A,
Microbiol 11:18031814. Blattner FR, Kaper JB. 2002. Characterization
162. Peterson JH, Tian P, Ieva R, Dautin N, of Cah, a calcium-binding and heat-extractable
Bernstein HD. 2010. Secretion of a bacterial autotransporter protein of enterohaemorrhagic
virulence factor is driven by the folding of a Escherichia coli. Mol Microbiol 45:951966.
C-terminal segment. Proc Natl Acad Sci USA 173. Torres AG, Jeter C, Langley W, Matthysse AG.
107:1773917744. 2005. Differential binding of Escherichia coli
163. Brunder W, Schmidt H, Karch H. 1997. EspP, O157:H7 to alfalfa, human epithelial cells, and
a novel extracellular serine protease of entero- plastic is mediated by a variety of surface struc-
haemorrhagic Escherichia coli O157:H7 cleaves tures. Appl Environ Microbiol 71:80088015.
human coagulation factor V. Mol Microbiol 24: 174. Giron JA, Torres AG, Freer E, Kaper JB.
767778. 2002. The flagella of enteropathogenic Esche-
164. Orth D, Ehrlenbach S, Brockmeyer J, Khan AB, richia coli mediate adherence to epithelial cells.
Huber G, Karch H, Sarg B, Lindner H, Wurzner Mol Microbiol 44:361379.
R. 2010. EspP, a serine protease of enterohemor- 175. Best A, La Ragione RM, Sayers AR, Woodward
rhagic Escherichia coli, impairs complement acti- MJ. 2005. Role for flagella but not intimin in the
vation by cleaving complement factors C3/C3b persistent infection of the gastrointestinal tissues
and C5. Infect Immun 78:42944301. of specific-pathogen-free chicks by shiga toxin-
165. Brockmeyer J, Aldick T, Soltwisch J, Zhang W, negative Escherichia coli O157:H7. Infect Immun
Tarr PI, Weiss A, Dreisewerd K, Muthing J, 73:18361846.

176. Erdem AL, Avelino F, Xicohtencatl-Cortes J, 182. Merkel V, Ohder B, Bielaszewska M, Zhang W,
Giron JA. 2007. Host protein binding and adhe- Fruth A, Menge C, Borrmann E, Middendorf
sive properties of H6 and H7 flagella of attaching B, Muthing J, Karch H, Mellmann A. 2010.
and effacing Escherichia coli. J Bacteriol 189: Distribution and phylogeny of immunoglobulin-
74267435. binding protein G in Shiga toxin-producing Esch-
177. Mahajan A, Currie CG, Mackie S, Tree J, erichia coli and its association with adherence
McAteer S, McKendrick I, McNeilly TN, Roe phenotypes. Infect Immun 78:36253636.
A, La Ragione RM, Woodward MJ, Gally DL, 183. Torres AG, Kaper JB. 2003. Multiple elements
Smith DG. 2009. An investigation of the expres- controlling adherence of enterohemorrhagic Esch-
sion and adhesin function of H7 flagella in the in- erichia coli O157:H7 to HeLa cells. Infect Immun
teraction of Escherichia coli O157:H7 with bovine 71:49854995.
intestinal epithelium. Cell Microbiol 11:121137. 184. Torres AG, Li Y, Tutt CB, Xin L, Eaves-Pyles T,
178. Tarr PI, Bilge SS, Vary JC Jr, Jelacic S, Habeeb Soong L. 2006. Outer membrane protein A of
RL, Ward TR, Baylor MR, Besser TE. 2000. Escherichia coli O157:H7 stimulates dendritic cell
Iha: a novel Escherichia coli O157:H7 adherence- activation. Infect Immun 74:26762685.
conferring molecule encoded on a recently ac- 185. Prasadarao NV, Wass CA, Weiser JN, Stins MF,
quired chromosomal island of conserved struc- Huang SH, Kim KS. 1996. Outer membrane
ture. Infect Immun 68:14001407. protein A of Escherichia coli contributes to inva-
179. Herold S, Paton JC, Srimanote P, Paton AW. sion of brain microvascular endothelial cells. In-
2009. Differential effects of short-chain fatty acids fect Immun 64:146153.
and iron on expression of iha in Shiga-toxigenic 186. Prasadarao NV. 2002. Identification of Esche-
Escherichia coli. Microbiology 155:35543563. richia coli outer membrane protein A receptor on
180. Yin X, Wheatcroft R, Chambers JR, Liu B, Zhu human brain microvascular endothelial cells. In-
J, Gyles CL. 2009. Contributions of O island 48 to fect Immun 70:45564563.
adherence of enterohemorrhagic Escherichia coli 187. Shin S, Lu G, Cai M, Kim KS. 2005. Escherichia
O157:H7 to epithelial cells in vitro and in ligated coli outer membrane protein A adheres to human
pig ileal loops. Appl Environ Microbiol 75:5779 brain microvascular endothelial cells. Biochem
5786. Biophys Res Commun 330:11991204.
181. Lu Y, Iyoda S, Satou H, Satou H, Itoh K, Saitoh 188. Gu J, Ji X, Qi J, Ma Y, Mao X, Zou Q. 2010.
T, Watanabe H. 2006. A new immunoglobulin- Crystallization and preliminary crystallographic
binding protein, EibG, is responsible for the chain- studies of the C-terminal domain of outer mem-
like adhesion phenotype of locus of enterocyte brane protein A from enterohaemorrhagic Esch-
effacement-negative, shiga toxin-producing Esch- erichia coli. Acta Crystallogr Sect F Struct Biol
erichia coli. Infect Immun 74:57475755. Cryst Commun 66:929931.

Animal Models of Enterohemorrhagic
Escherichia coli Infection
JENNIFER M. RITCHIE1
8
INTRODUCTION
Since the first recognized Escherichia coli O157:H7 outbreak over 3 decades ago
(1), investigators have sought to identify suitable animal hosts that allow study
of enterohemorrhagic E. coli (EHEC)-mediated disease. The value of any animal
infection model ultimately relies on its ability to reproduce the human disease
and enable the mechanistic processes that lead to clinical disease, pathogen
carriage, and transmission to be examined. As yet, no single animal model
mimics the full spectrum of disease caused by EHEC infection. However, since
Moxley and Franciss review in 1998 (2), several advances have been made in
the field, including the generation of a Shiga toxin (Stx)-producing Citrobacter
rodentium-murine model, a human intestine xenograft murine model, and a
renewed interest in the use of rabbit models. This article reviews what is known
about EHEC-mediated disease from human outbreaks and biopsy studies,
and within a historical context, describes the features and limitations of EHEC
infection models that are based on the three most commonly used species
(pigs, rabbits, and mice). Recent new advances are highlighted and discussed in
light of mounting evidence for the need to study the biology and virulence
strategies of EHEC in the context of its niche within the intestine. The reader
is directed elsewhere for excellent reviews on the environmental sources of
1
School of Biosciences and Medicine, University of Surrey, Guildford, United Kingdom.
157

158 RITCHIE
EHEC infection (3), EHEC interactions with the intestinal lumen where it is shed in feces.
the intestinal epithelium (4), the molecular Exactly how Stx crosses the epithelial barrier
basis of pathogenicity (58), and the current remains poorly defined, but it is generally
status of treatment options (9). believed that distribution of the Stx receptor,
globotriaosylceramide (Gb3), determines the
site of tissue damage (16, 17). The presence
OBSERVATIONS FROM of Gb3 in the endothelial cells that line the
HUMAN OUTBREAKS blood vessels (vasculature) and the subse-
quent host cell response to the cellular dam-
Most of what is known about EHEC patho- age it causes explain the resulting clinical
genesis is based on outbreaks caused by E. coli disease (18). Acute renal failure occurs be-
O157:H7 (1, 5, 10). From these outbreaks, cause the highest concentrations of Gb3 are
it is widely accepted that EHEC infection found in this organ (19); however, Gb3 is also
is acquired through the ingestion of feces- detected in the microvasculature of the colon
contaminated food or water or by hand-to- and the cerebellum (2022). Conflicting views
mouth transmission following contact with on whether Gb3 is also expressed on the
infected animals (including humans) or their surface of colonic epithelial cells exist in the
surroundings. Whereas some individuals who literature (2325).
ingest the organism remain asymptomatic, As such, any Stx-producing strain capable
others develop severe abdominal pain and of colonizing the mammalian intestine could
diarrhea typically within 3 to 4 days (11). This potentially cause Stx-mediated disease in hu-
can progress to a bloody diarrhea (or hemor- mans. In reality, E. coli O157:H7 strains are
rhagic colitis), and in approximately 5 to the predominant serotype associated with
15% of infected individuals a potentially fatal EHEC outbreaks in the United Kingdom and
hemolytic-uremic syndrome (HUS) develops elsewhere (2628). The reason behind this
(12). HUS is defined as a clinical syndrome com- selection is not clear but may be related to
prising thrombotic microangiopathy, throm- serotype-specific differences in prevalence,
bocytopenia, and hemolytic anemia that leads persistence, or survival in the environment
to acute kidney injury, for which there is no as well as their inherent virulence. E. coli
effective treatment. The ubiquitousness of the O157:H7 strains usually contain the locus of
causative agent, the severity of the clinical se- enterocyte effacement (LEE) pathogenicity
quelae, and the lack of treatment options and island, a virulence attribute deemed to play
effective preventive measures demand a better a critical role in mediating bacterial attach-
understanding of how this group of organisms ment to the intestinal epithelium (2931).
cause disease. Among non-O157 strains, bacteria belonging
The pathophysiology of disease is largely to serogroups O26, O103, O111, O121, O45,
attributed to the ability of EHEC to colonize and O145 are most commonly reported from
the mammalian intestine and produce Shiga- infected individuals (28). Some serogroups,
like toxin, a family of potent cytotoxins that notably O113, lack the LEE pathogenicity
are able to cross the epithelial barrier and act island, and compared to the LEE-positive
at sites distal to the intestine, inhibiting pro- O157 strains, relatively little is known about
tein synthesis in sensitive target endothelial the factors that are important in mediating
cells. Consistent with this idea, Stx has been their attachment to the epithelial surface (32,
detected on the surface of polymorphonuclear 33).
leukocytes (PMNs) circulating in the blood Epidemiological studies suggest that non-
of some infected individuals and in their feces O157 serogroups cause a similar but somewhat
(13, 14). In contrast, bacteremia has not been less severe range of clinical manifestations
reported (1, 15), and the organism remains in than O157 strains (28, 34, 35). Retrospective

CHAPTER 8 Animal Models of EHEC Infection 159
analyses of patient clinical profiles from a U.S. of diarrhea in volunteers who ingested the
hospital revealed similar incidences of bloody mutant (40). Similarly, deletion of intimin, a
diarrhea and HUS in children presenting 94-kDa outer membrane protein encoded by
with either non-O157 (17 patients) or O157 (33 the gene eaeA, also reduced the severity of
patients) infections (36). In contrast, a larger infection in volunteers (41). In both studies,
study using German national surveillance data however, a few individuals who received the
concluded that other than for the recent O104 mutant strains still developed diarrhea, albeit
outbreak strains, patients presenting with producing stools of reduced volume and fre-
non-O157 infections were half as likely to be quency. Moreover, the numbers of organisms
hospitalized and one-tenth as likely to die recovered from their stools were similar to
compared to those infected with O157 (37). those receiving wild-type organisms, at least
An unexpectedly high percentage of people at around the time of peak organism excretion
developed HUS following the outbreak of Stx- from the host (41). These results indicate that
producing enteroaggregative E. coli O104:H4 factors other than Tir and intimin contribute
that unfolded in northern Germany during to EHEC colonization in the intestine, and
the summer of 2011 (37, 38). In addition, the in this regard, the biological significance of
affected target population consisted of young, Tir-intiminmediated attachment (and the re-
healthy females rather than individuals at sultant formation of the characteristic attach-
the extremes of age, who are more likely to ing and effacing [A/E] lesion) during disease is
be sickened during an E. coli O157:H7 out- not really understood.
break. Thus, clinical outcome may be de-
pendent on the demographic of the affected
Biopsies from Infected Individuals
population as well as the characteristics of
the infectious organism. Predicting which Intestinal damage
strain and host factors led to the progression Relatively few studies report on the intestinal
of a more severe clinical outcome is critical changes that occur in individuals infected
to the development and management of ef- with EHEC, probably due to the risk of worsen-
fective control strategies to prevent the loss of ing the underlying clinical condition. Colo-
human life. noscopy examination of infected individuals,
which allows the full length of the large
intestine to be viewed, found that the most
Volunteer Studies
severe disease occurred in the cecum and as-
Due to the potentially life-threatening se- cending colon (42, 43). Here, tissue damage
quelae associated with EHEC infections, vol- consisting of edema, erythema (redness), hem-
unteer studies using Stx-producing strains are orrhage, and erosion was evident over a
deemed unethical. However, volunteer studies wide area and was associated with a marked
using the closely related diarrheal pathogen narrowing of the luminal space. In some in-
enteropathogenic E. coli (EPEC) have been dividuals, long ulcer-like lesions were also
performed and demonstrate a clear role in noted on the surface of the tissue. Patchy and
pathogenesis for some of the virulence attri- less severe areas of inflammatory damage
butes shared by the two pathotypes of E. coli. were observed in the transverse colon, de-
For example, studies reveal that bacterial ad- scending colon, and sigmoid colon, consistent
herence to the epithelial surface mediated with descriptions from sigmoidoscopy exam-
by Tir-intimin interactions play a major role inations (44, 45).
in the development of diarrhea. Deletion of Histologic examination of the damaged
EspB, a type III secreted protein and required areas revealed destruction of the surface epi-
for targeting of Tir to the host cell membrane thelium and involvement of the lamina pro-
(39), decreased the incidence and severity pria, while the deeper colonic crypts remained

160 RITCHIE
largely intact (42, 43). Focal areas with acute

HISTORICAL PERSPECTIVE ON ANIMAL
inflammation were evident, with neutrophil
MODELS OF EHEC-MEDIATED DISEASE
infiltration of the lamina propria and, in
some cases, the formation of crypt abscesses.
Pigs
Hemorrhage and edema were commonly ob-
served in the lamina propria, as were apo- Nearly 30 years ago, Tzipori and Francis de-
ptotic cells in the surface epithelium and scribed the first use of gnotobiotic piglets to
colonic crypts. Small fibrin/platelet thrombi study the newly emerging group of pathogenic
formed within the mucosal capillaries. It is E. coli strains that included E. coli O157:H7
interesting to note that adherent bacteria (47, 48). Gnotobiotic piglets, which are deliv-
were not observed in the tissue sections, de- ered via cesarean section, maintained in germ-
spite the samples being collected during the free incubators, and artificially reared (aka
acute phase of the disease and only a few days cesarean-derived colostrum-deprived [CDCD]
after the onset of bloody diarrhea (42, 44). piglets), had been used previously to study
These findings indicate that during EHEC other human enteric pathogens, including
infection the integrity of the epithelial barrier cryptosporidium and pathogenic E. coli (49,
is compromised; whether this occurs due to 50). Oral infection of 1-day-old CDCD piglets
the direct action of the bacterium or Stx or with high numbers (1010 CFU) of E. coli
due to the resulting host response to infection O157:H7 caused watery but not bloody diar-
is not clear. rhea in most infected animals by 2 to 4 days
post inoculation. In these early studies, no
Extraintestinal damage other signs of disease were reported. As was
Readers are referred to two reviews describ- found in human infections, histologic abnor-
ing the detailed pathological manifestations malities were detected in the cecum and
of Stx-mediated kidney damage in humans colon, and primarily consisted of destruction
(6, 12). Summarized below are the key path- of the mucosal brush border and inflamma-
ological findings evident in patients who tion (51). These first investigations recognized
present with diarrhea and HUS. Inward and that E. coli O157:H7 caused intestinal lesions
colleagues reported that the renal pathology that were morphologically similar to those
consisted of endothelial swelling and glomer- previously described for EPEC (52), and thus
ular thrombosis with congested rather than EHEC was considered a member of the same
ischemic glomeruli (46). In addition, they pathotype group (53). However, while EHEC
noted the presence of significant numbers of and EPEC apparently shared one common
neutrophils compared to control tissue. Glo- mechanism of bacterial attachment, they ad-
merular capillaries appeared to be the main hered to different sites in the intestine. EPEC
site of damage in the patients. As such, the colonized both the small and large intestine,
site of damage mirrors the distribution of whereas only the large intestine was colonized
Gb3, the Stx receptor, in humans. The highest by EHEC (54).
densities of Gb3 were detected on glomerular Neither the pO157 plasmid nor Stx ap-
endothelial cells, although Gb3 has also been peared to play a role in lesion formation as
detected in renal proximal tubules (6). Con- mutants deleted for these factors still caused
sistent with these findings, Stx was found to intestinal disease (54, 55). Instead, genes
bind to both kidney tubular and glomeruli of encoded in the now well-described LEE
a pediatric patient with HUS, but only the pathogenicity island proved to be critical
tubules of a geriatric patient with HUS (16). for intimate bacterial attachment in vitro
Whether these findings reflect differences in and intestinal lesion formation in vivo (29,
Gb3 expression, the stage of HUS, or some 30). Consistent with findings from volunteer
other factor is not known. studies, deletion of eaeA prevented bacterial

attachment and A/E lesion formation, yet piglets (61). Despite kidney damage, unlike
allowed similar numbers of wild-type and humans, the piglets do not exhibit clinical
mutant cells to be recovered in intestinal signs of kidney failure (e.g., elevated creati-
homogenates obtained from the infected ani- nine levels) (60). As such, while piglets are
mals (56). These findings suggest that the one of the few species that show both intes-
gnotobiotic piglet intestine, like the human tinal and systemic disease following EHEC
intestine, is permissive for EHEC replication infection, they do not fully recapitulate all
and the organism is capable of surviving in facets of the human disease.
this niche. In addition to the inherent biological con-
In addition to intestinal disease, subse- straints associated with the use of gnotobiotic
quent studies reported that some piglets ex- animals (e.g., poorly developed immune sys-
hibit signs of central nervous system (CNS) tems, a lack of normal microflora), most re-
disease, including ataxia (55). Indeed, support searchers lack the specialized facilities needed
for a role for Stx in extraintestinal disease to rear animals under sterile conditions. To
initially came from a study where gnotobiotic overcome some of these limitations, conven-
piglets were given either orally administered tional piglets were assessed for their suscep-
bacteria or intraperitoneally administered tibility to EHEC infection. Naturally born
polymyxin B-treated bacterial cell lysates (57). piglets, which feed directly from the sow and
In both instances, piglets that exhibited CNS either remain with the sow for the duration
disease showed evidence of vascular damage of the study or are subsequently hand-reared
in the cerebellum. The vascular damage bore and housed in individual microbiological iso-
similarities to lesions seen in the brain of a lators, offer an additional advantage as they
patient who died following E. coli O157:H7 enable passive immunization studies to be per-
infection and in piglets with naturally ac- formed (62). Surprisingly, continuously suck-
quired edema disease (57, 58), infections that ling piglets were found to exhibit more severe
are both marked by the presence of Stx. While neurological disease and more quickly suc-
Stx was found to be important in the devel- cumbed to EHEC infection than traditional
opment of extraintestinal disease, intimate CDCD animals (63). The reason for this find-
bacterial attachment to the epithelial surface ing is not known, but it may be related to the
was not required for this to occur (59). amount of Stx produced under the different
While early studies focused on the effect conditions in the intestine. Consistent with
of Stx on the CNS, a significant advance in this idea, Shringi and colleagues found that
the usefulness of this model was made when strains that produced higher amounts of Stx
investigators reported that gnotobiotic pig- caused earlier and more severe signs of CNS
lets also developed renal lesions following disease in conventionally born piglets (64).
oral EHEC infection (5961). Gunzer and Somewhat surprising, no signs of kidney dam-
colleagues were the first to report that the age were found in these animals, perhaps due
kidneys of infected animals exhibited signs to the rapid onset of neurological symptoms.
of endothelial cell damage (60). Damage was The relationships between animal husbandry
characterized by diffuse glomerular endothe- practice, the amount of intestinal Stx, and the
lial swelling and congestion and a narrowing resulting pathology warrant rigorous investi-
of the capillary vessels. In most instances, gation in this species to further define this
morphological signs of thrombotic micro- model system.
angiopathy, which is a hallmark characteristic
of HUS in humans, were also noted. Gunzers
Rabbits
findings were confirmed following retrospec-
tive examination of kidney samples obtained Rabbits have been used both to study the toxic
from multiple EHEC-infected and control effects of Stx and the intestinal biology of the

162 RITCHIE
organism, and as such, the age, breed, and However, Stx2 did not appear to act as a
route of infection affect their response to in- colonization factor because similar numbers
fection. Suckling New Zealand White (NZW) of wild-type and mutant organisms were re-
rabbits are naturally susceptible to oral infec- covered from the colon at various times post-
tion by EHEC and, like piglets, develop diar- infection. In contrast, deletion of the genes
rhea. Farmer and colleagues first reported coding for the bacterial outer membrane ad-
that oral inoculation of high numbers (>109 hesin, intimin (eaeA), and the translocated
CFU) of E. coli O157:H7 caused watery diar- intimin receptor (tir) markedly reduced the
rhea in young (5 to 10 days old) but not older ability of the organism to colonize the rabbit
(20 days old) rabbits. In contrast, the ingestion intestine (31). Thus, in contrast to piglets,
of control (nonpathogenic) E. coli strains EHEC attachment to the epithelial surface is
failed to cause any signs of morbidity in the required for the organism to persist and cause
animals (65, 66). Large numbers of the bac- intestinal disease.
teria could be recovered from the intestine, One of the strengths of the suckling rabbit
suggesting that the organism is able to multi- model is that it allowed, for the first time, a
ply in this environment (65). Subsequent means to study factors that were important in
studies identified the colon as the principal mediating intestinal colonization and disease.
site of EHEC-mediated disease, where histo- Thus, using suckling rabbits we found that
logical abnormalities, including edema, hem- factors other than Tir and intimin influenced
orrhage, the presence of an inflammatory the ability of EHEC to colonize the intes-
infiltrate, and mucosal epithelial apoptosis, tine and cause disease. For example, surface-
were observed (66, 67). In these young ani- exposed structures such as the long polar
mals, the development of diarrhea appears to fimbriae (Lpf), curli, and the O-antigen cap-
be driven by inflammation-mediated changes sule all contributed to the ability of EHEC
in colonic ion transport resulting in decreased to colonize and survive in the intestine (70
Na absorption and increased Cl secretion (68, 72). Indeed, our studies began to reveal the
69). Unlike humans, suckling rabbits do not complexities of regulating colonization factor
develop signs of renal disease; however, the expression in the intestine. While expression
reproduction of diarrhea and colonic inflam- of Lpf contributes to EHEC intestinal coloni-
mation, two key features of EHEC infection in zation, in the absence of these structures,
humans, indicates that they may be a useful the organism appears to overexpress curli, ap-
model for studying the intestinal manifes- pendages more usually associated with ad-
tations of the disease. herence to abiotic surfaces (73). In addition,
The first suckling rabbit studies suggested we found that bacterial factors (other than
a role for Stx in the development of diar- Tir), which are translocated into the host cell
rhea because natural isolates that produced via the organisms type III secretion system,
differing amounts of Stx gave rise to differ- also modulate intestinal colonization (71). For
ing severities of disease in the animals (66, example, EspFu (also known as TccP) appears
67). We later confirmed a role for Stx in to promote the spread of EHEC to new sites in
E. coli O157:H7 pathogenicity using genetically the intestine. Rabbits infected with the espFu
defined isogenic bacterial mutants (31). Three- mutant contained significantly fewer orga-
day-old NZW rabbits infected with an iso- nisms in their ceca and colons at 7 but not
genic mutant that lacked the genes coding 2 days postinfection. Moreover, this mutant
for Stx2 developed a transient diarrhea and covered a significantly smaller area of the in-
largely noninflammatory intestinal pathology. testine than the wild type in the gnotobiotic
Thus, Stx2 appeared to increase the severity piglets (74). However, many of the LEE-
and duration of EHEC-induced diarrhea and encoded effector proteins caused only minor
modulate the host response to infection. changes in the ability of the EHEC to colonize

the intestine, most likely because of functional better adapted to colonizing the rabbit intes-
redundancy between the different proteins tine. Panda and colleagues observed no sig-
(71). nificant differences in the response of weaned
Despite being sensitive to the intestinal NZW or DB rabbits to E. coli O157 infection
manifestations of EHEC infection, suckling (81). However, these authors noted that renal
rabbits do not develop HUS or any other ev- lesions did not appear to be a consistent fea-
idence of renal failure. The reasons for this ture in the infected rabbits. Renal pathology
are not clear but may be related to the species- was detected in only 1 of 11 infected rabbits,
specific distribution of Gb3. Zoja and col- and as such, the reproducibility of this clinical
leagues reported that Gb3 is present in the outcome would need to be improved to enable
tissues of the brain, lung, spinal cord, and in- study of EHEC-induced renal disease or to
testine of weaned NZW rabbits but is missing allow novel therapeutics to be tested.
from the heart, liver, and kidney (75). While While suckling and young (10-week-old)
Gb3 homologs have recently been detected in rabbits are susceptible to oral administration
the kidney cortex of NZW and Dutch Belted of EHEC, older rabbits tend to be more re-
(DB) rabbits, their role in disease and relative calcitrant to colonization by the organism.
affinity for Stx remain unclear (76). In young Two approaches have been used by investi-
rabbits (day 5 of life), Gb3 has been detected gators to circumvent this issue. The rabbit
in lipid extracts obtained from the colon (77). ligated ileal loop assay originally described
Conversely, the appearance of Gb3 appears by De and colleagues has been widely used
to be maturationally regulated in the small to assess the enterotoxicity of pathogenic bac-
intestine, as jejunal microvillous membranes teria, their culture supernatants, or purified
isolated from rabbits less than 16 days of age toxins (82). Likewise, both EHEC strains and
did not bind Stx (78, 79). Together, these ob- purified Stx1 have been injected into ligated
servations may explain the sensitivity of the loops to reproduce the disease found in hu-
rabbit colon to the effects of luminal Stx in mans (2, 83). However, this closed loop
suckling animals, and the predominantly neu- model suffers from many limitations when
rological manifestations of Stx when given intestinal pathogens such as EHEC are stud-
intravenously to adult rabbits (see below). ied, including most notably an interruption
Following a natural outbreak of bloody to normal gut function (e.g., peristalsis).
diarrhea and sudden death in a group of The second approach involves the use of
10-week-old DB rabbits in which an Stx1- a natural rabbit diarrheal pathogen, E. coli
producing E. coli O153 isolate was recovered, strain RDEC-1, which, while naturally lacking
the possibility was raised that this breed of Stx, contains the LEE pathogenicity island
rabbits may be more sensitive to the systemic (84). The histological lesions caused by RDEC-
effects of Stx than NZW rabbits (80). Exper- 1 appeared indistinguishable from those re-
imental infection of 8-week-old weaned DB ported for other A/E pathogens, including
rabbits with the outbreak strain confirmed EHEC (52, 85). To better reproduce the viru-
that it was able to cause diarrhea and renal lence repertoire of EHEC, Sjogren and col-
vascular and glomerular lesions in the animals leagues generated an Stx-producing variant
(76). As part of this study, the authors also of RDEC-1 by infecting RDEC-1 with an
examined the response to infection by an Stx1-converting bacteriophage (86). By com-
E. coli O157:H7 clinical isolate. Rabbits in- paring the disease caused by the two strains,
fected with E. coli O157 lost less weight (mean the authors observed that Stx-producing
SE: 4.6% 1.6 versus 18.1% 4.7) and RDEC-1 induced a more severe infection
shed fewer organisms (range 103 to 105 versus than the parent strain did, causing earlier
105 to 109 CFU/g) than O153-infected rabbits, weight loss and higher rates of mortality in the
perhaps suggesting that the latter strain is animals. While these studies indicate a role for

164 RITCHIE
Stx in the development of severe disease, the Despite this fundamental difference in the
usefulness of this model is somewhat limited pathophysiological response to Stx in mice
by an inability to study whether additional and humans, a rodent model of EHEC-medi-
EHEC virulence factors are important to in- ated disease remains desirable. As such, many
fection. approaches have been taken to render mice
Intravenous administration of purified Stx more susceptible to oral infection or to pres-
causes thrombotic microangiopathy in several ent with a more accurate model of HUS. Be-
organs in rabbits, consistent with the reported cause of the recent publication of an excellent
distribution of Gb3 (75, 87). The location of review on murine models of EHEC and Stx
the resulting damage appears to depend on infection by Mohawk and OBrien (89) to
the type of Stx used because Stx1 primarily which the readers are directed, this article
located to the microvasculature of the brain summarizes only the main features and limi-
and spinal cord, whereas Stx2 localizes to the tations of murine models before recent ad-
microvasculature of the cecum. These differ- vances in the field are discussed.
ences in tissue tropism also likely explain the The natural resistance to EHEC coloniza-
greater lethality of Stx1 than Stx2 in rabbits, tion in mice can be reduced by modifying
although this does not reflect the epidemio- the normal intestinal microflora through pre-
logical data from human outbreaks. treatment with antibiotics, use of germ-free
animals, or via dietary-induced changes (see
Table 1). All lead to increased numbers of
Mice
EHEC in the intestine and the development of
Mice, the workhorse species of infectious dis- varying levels of renal toxic injury and death
ease research, are naturally resistant to colo- in the animals. In addition, the sensitivity
nization by EHEC and fail to develop signs of the host (e.g., lipopolysaccharide [LPS]-
of intestinal disease following oral infection. responder mice, MyD88 knockout mice) or
In contrast, systemic Stx causes acute kidney the virulence of the bacterium (e.g., host-
damage and death in the animals (88). How- passaged organisms, treatment with mitomy-
ever, key differences exist between Stx- cin C) can be manipulated to further increase
mediated renal damage in mice and humans clinical severity. Moreover, to more accurately
(12). Renal damage in mice consists of acute mimic the kidney disease seen during HUS,
tubular necrosis, which is deemed charac- investigators have also tried coadministering
teristic of toxin insult, rather than the pro- endotoxin (LPS) with Stx. In this situation,
thrombotic condition created following LPS acts as an accessory proinflammatory
Stx-mediated endothelial injury in humans. mediator, which primes the host immune re-
HUS is viewed as a thrombotic microangi- sponse and results in the induction of throm-
opathy of the glomerular blood vessels that bocytopenia in the animals. However, the
develops with or without tubular cell death etiology of LPS plus Stx challenge differs from
(12). This difference in pathology likely re- that of HUS and does not reflect what occurs
flects species-specific differences in Gb3 dis- during human infection (12).
tribution. In mice, Gb3 is detected in the An alternative approach that circumvents
microvascular endothelial cells of the brain the problems of establishing EHEC coloni-
cortex and pia mater (membranes surround- zation in mice is the use of the surrogate
ing the brain and spinal cord) and the capil- bacterium Citrobacter rodentium, a natural
lary vascular endothelial cells surrounding murine pathogen that induces colonic hyper-
renal tubules (17). As noted earlier, the highest plasia (90). Both EHEC and C. rodentium
concentrations of Gb3 expression in humans contain the LEE pathogenicity island that
occur in the kidney, specifically on the glo- mediates Tir-intimin-based bacterial attach-
merular endothelium (6). ment to the epithelial surface (91). Because of

TABLE 1 Comparative murine models of EHEC-mediated disease

Model Host Clinical disease Key limitations Reference(s)
Oral infection
Streptomycin-treated CD-1 (5- to 8-week Weight loss, renal tubular Reduced microora, 88, 122
old-male mice) damage, mortality articial colonization,
no intestinal disease
Streptomycin + ICR (3 weeks old) Weight loss, renal tubular Reduced microora, 123
mitomycin C treatment damage, mortality no intestinal disease
Weaned BALB/c (17 to 20 days Mortality No intestinal disease, 124
old); host-passaged rapid clearance of
bacteria the pathogen,
no glomerular disease
LPS responder C3H/HeN and C3H/HeJ Anorexia, rufed fur, No glomerular disease 125
(8 to 16 weeks old) neurological disease
Intact microora BALB/c (6 weeks old) Weight loss, renal tubular No glomerular disease 126
damage, mortality
Streptomycin + MyD88/ C57BL/6 Weight loss, kidney toxic No glomerular disease, 127
MyD88-decient mice (6 to 14 weeks old) insult, mortality host immune system
abnormalities
Germ-free Swiss Webster(3 days Acute renal tubular No natural microora, 128130
to 12 weeks old)/IQI necrosis, limited poorly developed
(8 weeks old) glomerular thrombosis, immune systems
mortality
C. rodentium C57BL/6 (5 to Watery stools, weight Not EHEC 94
8 weeks old) loss, intestinal disease
Injection
Conventional mice + C57BL/6 (8 to Acute renal tubular Nonbacterial, 131
IP injection of OMVa 10 weeks old) necrosis, no thrombosis
Conventional mice + C57Bl/6J (8 to Weight loss, signs of Nonbacterial 132
multiple IP Stx2 10 weeks old) kidney failure, glomerular
endothelial damage,
neutrophilia
Conventional + IP or CD-1 (6 to 8 weeks old) Mild dehydration, No intestinal disease, 133
IV Stx1/2 renal tubular necrosis no glomerular thrombosis
Other
Mice + human colonic CB-17 SCID mice A/E lesion pathology, Incomplete immune 96
xenograft model (6 to 8 weeks old) + some mucosal damage system, no natural
16 maturation microora, nonnatural
organ environment,
no systemic disease
a
Outer membrane vesicles.
this, C. rodentium has been used extensively to does not naturally contain the genes coding
dissect the function and activity of all LEE- for Stx, the principal virulence factor of
encoded proteins as well as many other non- EHEC. To overcome this limitation, Mallick
LEE-encoded proteins that are translocated and colleagues recently developed an Stx-
via the organisms type III secretion system expressing C. rodentium strain by lysogenizing
(92, 93). These studies have done much to the bacterium with an Stx-containing phage
advance our understanding of how trans- (94). The phage taken up by C. rodentium
located bacterial proteins interfere with host was found to carry Stx genes belonging to an
cell processes. However, while C. rodentium Stx2 variant called mucus-activatable Stx2d.
is useful for dissecting the role and function Mucus-activatable Stx2d is a particularly toxic
of type III secreted proteins, C. rodentium subtype of Stx2 that causes high mortality

166 RITCHIE
in mice but is usually found in non-LEE- immune system of the host animal may prove
encoding EHEC strains that are less com- to be significant limitations to the acceptance
monly isolated in human oubreaks than EHEC of this model system.
containing Stx2 or Stx2c (95). Compared to
wild-type C. rodentium, Stx-producing C. ro-
Other Species
dentium causes increased weight loss, severe
destruction of the intestinal mucosa, and Other than pigs, rabbits, and mice, a plethora
renal tubular damage. The development of of animal species have been considered as
this model allows study of robust intestinal model hosts of EHEC-induced disease (see
colonization together with Stx-mediated dis- Table 2). Like mice, ferrets are naturally re-
ease, although it still does not overcome the sistant to oral infection by EHEC and require
different pathophysiological response to Stx pretreatment with antibiotics to establish
seen in mice and humans. In addition, it does moderate numbers of the organism in the in-
not easily allow study of other EHEC factors testine (97). Despite significant weight loss
that are important to infection. following infection, none of the animals
Another recent advance in the develop- showed any signs of intestinal disease, and
ment of murine models has been the con- only around 20% showed signs of renal glo-
struction of human and bovine intestinal merular disease and/or thrombocytopenia.
xenograft models (96). Immature germ-free Greyhounds naturally develop cutaneous and
human and bovine small intestine or colonic renal glomerular vasculopathy, a disease that
fetal tissues were subcutaneously transferred results in the formation of renal glomerular
into SCID mice and the tissue left to develop lesions, which are similar to those seen in
for up to 16 weeks. The authors demonstrate human HUS (98, 99). While this model has
that after maturation, the xenografted tissue not been used extensively since its first de-
displays appropriate organ-specific tissue ar- scription, Raife and colleagues have used it to
chitecture, complete with a range of cell types demonstrate that thrombin-dependent mech-
including goblet cells and microvilli. There- anisms are important in Stx2-mediated path-
after, various EHEC strains were injected into ogenesis. Treatment of greyhounds with
the intestinal lumen of each organ. Com- lepirudin, a thrombin inhibitor, reduced HUS-
pared to the small intestine xenografts, EHEC like pathology and enabled the animals to
strains grew poorly in the lumen of the co- survive beyond the normal course of the ex-
lonic xenografts. Despite this difference, dis- periment (100). Baboons appear to best mimic
tinct A/E lesions and mucosal damage were human HUS symptoms following intravenous
found only on tissue derived from the human injection of Stx (101, 102). In this model, both
colon, and not the human or bovine small Stx1 and Stx2 were able to induce HUS in the
intestine segments. Furthermore, A/E lesion animals, although the dose, timing, and mag-
formation was shown to be dependent on nitude of the response differed (103). Rats
the expression of a functional type III secre- also develop HUS-like symptoms in response
tion system, consistent with what has been to intraperitoneal challenge of filtered EHEC
reported in other model hosts (see above). culture supernatants (104). However, since
While xenograft models allow an opportunity purified Stx preparations were not used in
to study EHEC interactions with the human this model, further investigation is required
intestine without endangering human life, to dissect the contribution of Stx versus
this approach faces several challenges. The other soluble factors. Finally, the nematode
closed nature of the segment, which does Caenorhabditis elegans has beenput forward as
not allow any normal intestinal activity (e.g., a naturally infected and genetic tractable ani-
peristalsis, food passage) to occur, the lack mal host in which to study EHEC infection. As
of a natural microflora, and the incomplete found previously for other enteric pathogens

TABLE 2 Comparative models of EHEC-mediated disease in species other than pigs, rabbits, and mice
Model Host Clinical disease Key limitations Reference(s)
Ferrets Campylobacter-free Weight loss, limited renal Limited natural 97

female (6 weeks old) glomerular pathology microora, no intestinal
disease
Greyhound Bloody diarrhea, Nonbacterial 99, 100
microvascular thrombosis,
HUS
Monkeys Macaca radiata Enteritis with neutrophil High dose required 134
(wild-caught adults) inltration, A/E lesions to establish infection,
seen, moderate tubular no glomerular disease
blood vessel disruption
Caenorhabditis elegans N2 Intestinal colonization, Not a model of 106
fed on EHEC strains A/E lesion formation, EHEC-mediated disease
mortality
Rats + IP injection of Stx Sprague-Dawley Watery diarrhea, kidney Nonbacterial, 104
(adult male) failure, thrombocytopenia, crude bacterial
hemolytic anemia, supernatants were used
leukocytosis
Baboons + Stx injection Renal failure, Nonbacterial, not the 101103
into cephalic vein thrombocytopenia, natural route of infection
glomerular endothelial
injury, inammation,
mortality
Chickens + crop 1 day old Persistent shedding of the No clinical signs 135
cannulation of EHEC organism, A/E lesions
observed
(e.g., EPEC) (105), E. coli O157:H7 infects QseBC acts to recognize and respond to the
and kills C. elegans (106). The bacterium is presence of the host hormones epinephrine
able to multiply and produce A/E lesions in and norepinephrine or to the microbiota-
the digestive tract of the nematodes, and kill- derived quorum-sensing molecule AI-3 to ac-
ing appears to be at least partly dependent tivate flagella expression and motility (108).
on Stx1 (Stx2 was not important in this It is thought that signaling through this path-
model). It remains to be seen whether this way may alert EHEC to its arrival in the colon,
model will lead to new information on the a key site of bacterial attachment. Consistent
pathogenesis and host immune response to with this idea, QseC mutants exhibit a re-
infection; however, the availability of power- duced ability to colonize the colon of infant
ful genetic analysis methods for C. elegans rabbits (109). EHEC also senses fucose, a
may prove invaluable. microbiota-derived sugar, whose presence at
high concentrations in the intestinal lumen
inhibits LEE gene expression until the bac-
ROLE OF THE HOST IN MODULATING terium reaches the epithelial surface (110).
EHEC VIRULENCE Deletion of FusKR, the fucose-sensing two-
component system, reduced EHECs ability
Like other enteric pathogens, EHEC uses to colonize the rabbit intestine compared
many host- and microbiota-derived chemicals to the parental strain. In addition to sugars,
to regulate virulence gene expression when EHEC also appears to sense xanthine oxidase,
in the host (107). To date, only a few of these a host defense molecule that is constitutively
have been described. For example, it is known expressed by intestinal epithelial cells. Physi-
that the two-component system encoded by ological concentrations of xanthine oxidase

168 RITCHIE
upregulate EHEC virulence factor production in the field, none reproduce the full clinical
only as the bacterium comes into close prox- spectrum of illness seen during human infec-
imity to the epithelial surface (111, 112). Finally, tion. The choice of which animal model to
diet-induced changes to butyrate concentra- use must be viewed within the context of
tions in the mouse intestine were found to the scientific question being asked. However,
cause increased mortality in animals following studies focused on understanding the biology
EHEC infection (113). The mechanistic basis of the organism need to be considered within
for this finding is hypothesized to be due to in- a model system that reproduces as much as
creased levels of Gb3 expression in the kidneys possible the natural state of the intestine.
and intestines of mice fed high-fiber diets.
However, mice fed high-fiber diets also con-
ACKNOWLEDGMENTS
tained significantly fewer commensal Esch-
erichia spp. As such, this study highlights some I declare no conflicts of interest with regard to
of the complexities faced when dealing with the manuscript.
intact biological systems, where the host and
its associated microbiota are intimately linked.
CITATION
Over the past decade, there has been a
growing realization that the host and its as- Ritchie JM. 2014. Animal models of entero-
sociated microbiota profoundly influence the hemorrhagic Escherichia coli infection. Micro-
pathologies caused by noninfectious and in- biol Spectrum 2(4):EHEC-0022-2013.
fectious disease agents (e.g., see references 114
to 117). Many factors including diet, stress,
REFERENCES
or the environment alter the physiochemical
conditions in the gut and thereby influence 1. Riley LW, Remis RS, Helgerson SD, McGee HB,
Wells JG, Davis BR, Hebert RJ, Olcott ES,
the composition and activities of the intestinal
Johnson LM, Hargrett NT, Blake PA, Cohen
microbiota (118). In turn, the intestinal mic- ML. 1983. Hemorrhagic colitis associated with a
robiota contributes to numerous biological rare Escherichia coli serotype. New Engl J Med
processes in the host, including metabolic, 308:681685.
nutritional, physiological, and immunological 2. Moxley RA, Francis DH. 1998. Overview of ani-
mal models, p 249260. In Kaper JB, OBrien AD
function. As such, individual differences in
(ed), Escherichia coli O157:H7 and Other Shiga
host microbiota characteristics may explain Toxin-Producing E coli strains. ASM Press,
some of the variation in an individuals re- Washington, DC.
sponse to EHEC during an outbreak. 3. Ferens WA, Hovde CJ. 2011. Escherichia coli
Finally, a recent report suggests that uptake O157:H7: animal reservoir and sources of human
of Stx-encoding genetic material by host cells, infection. Foodborne Pathog Dis 8:465487.
4. Schuller S. 2011. Shiga toxin interaction with
rather than the toxic protein, may contribute human intestinal epithelium. Toxins 3:626639.
to Stx-induced pathology in mice (119, 120). 5. Kaper JB, Nataro JP, Mobley HL. 2004. Path-
These studies raise the question of whether ogenic Escherichia coli. Nat Rev Microbiol 2:123
Stx-coding bacteriophages, which are released 140.
6. Obrig TG. 2010. Escherichia coli Shiga toxin
along with Stx following bacterial lysis in the
mechanisms of action in renal disease. Toxins
intestine, play a role in Stx dissemination (121). 2:27692794.
7. Farfan MJ, Torres AG. 2012. Molecular mecha-
nisms that mediate colonization of Shiga toxin-
CONCLUSIONS producing Escherichia coli strains. Infect Immun
80:903913.
8. Melton-Celsa A, Mohawk K, Teel L, OBrien A.
Pigs, rabbits, and mice are the three most 2012. Pathogenesis of Shiga-toxin producing
common animal species used to study EHEC- Escherichia coli. Curr Top Microbiol Immunol
mediated disease. Despite recent advances 357:67103.

9. Bitzan M. 2009. Treatment options for HUS 21. Jacewicz MS, Acheson DW, Binion DG, West
secondary to Escherichia coli O157:H7. Kidney Int GA, Lincicome LL, Fiocchi C, Keusch GT.
Suppl 112:S62S66. 1999. Responses of human intestinal microvas-
10. Paton JC, Paton AW. 1998. Pathogenesis and cular endothelial cells to Shiga toxins 1 and 2
diagnosis of Shiga toxin-producing Escherichia and pathogenesis of hemorrhagic colitis. Infect
coli infections. Clin Microbiol Rev 11:450479. Immun 67:14391444.
11. Bell BP, Goldoft M, Griffin PM, Davis MA, 22. OLoughlin EV, Robins-Browne RM. 2001. Ef-
Gordon DC, Tarr PI, Bartleson CA, Lewis JH, fect of Shiga toxin and Shiga-like toxins on eu-
Barrett TJ, Wells JG, et al. 1994. A multistate karyotic cells. Microbes Infect 3:493507.
outbreak of Escherichia coli O157:H7-associated 23. Zumbrun SD, Hanson L, Sinclair JF, Freedy J,
bloody diarrhea and hemolytic uremic syndrome Melton-Celsa AR, Rodriguez-Canales J, Hanson
from hamburgers. The Washington experience. JC, OBrien AD. 2010. Human intestinal tissue
JAMA 272:13491353. and cultured colonic cells contain globotriaosyl-
12. Mayer CL, Leibowitz CS, Kurosawa S, Stearns- ceramide synthase mRNA and the alternate Shiga
Kurosawa DJ. 2012. Shiga toxins and the path- toxin receptor globotetraosylceramide. Infect
ophysiology of hemolytic uremic syndrome in Immun 78:44884499.
humans and animals. Toxins 4:12611287. 24. Schuller S, Frankel G, Phillips AD. 2004. Inter-
13. Brigotti M, Caprioli A, Tozzi AE, Tazzari PL, action of Shiga toxin from Escherichia coli with
Ricci F, Conte R, Carnicelli D, Procaccino MA, human intestinal epithelial cell lines and explants:
Minelli F, Ferretti AV, Paglialonga F, Edefonti Stx2 induces epithelial damage in organ culture.
A, Rizzoni G. 2006. Shiga toxins present in the gut Cell Microbiol 6:289301.
and in the polymorphonuclear leukocytes circu- 25. Holgersson J, Stromberg N, Breimer ME. 1988.
lating in the blood of children with hemolytic- Glycolipids of human large intestine: difference in
uremic syndrome. J Clin Microbiol 44:313317. glycolipid expression related to anatomical local-
14. te Loo DM, Monnens LA, van Der Velden TJ, ization, epithelial/non-epithelial tissue and the
Vermeer MA, Preyers F, Demacker PN, van ABO, Le and Se phenotypes of the donors.
Den Heuvel LP, van Hinsbergh VW. 2000. Biochimie 70:15651574.
Binding and transfer of verocytotoxin by poly- 26. Griffin PM, Tauxe RV. 1991. The epidemiology of
morphonuclear leukocytes in hemolytic uremic infections caused by Escherichia coli O157:H7,
syndrome. Blood 95:33963402. other enterohemorrhagic E. coli, and the associ-
15. Johnson WM, Lior H, Bezanson GS. 1983. Cy- ated hemolytic uremic syndrome. Epidemiol Rev
totoxic Escherichia coli O157:H7 associated with 13:6098.
haemorrhagic colitis in Canada. Lancet i:76. 27. Tarr PI, Gordon CA, Chandler WL. 2005. Shiga-
16. Chaisri U, Nagata M, Kurazono H, Horie toxin-producing Escherichia coli and haemolytic
H, Tongtawe P, Hayashi H, Watanabe T, uraemic syndrome. Lancet 365:10731086.
Tapchaisri P, Chongsa-nguan M, Chaicumpa W. 28. Gould LH, Mody RK, Ong KL, Clogher P,
2001. Localization of Shiga toxins of entero- Cronquist AB, Garman KN, Lathrop S, Medus
haemorrhagic Escherichia coli in kidneys of pae- C, Spina NL, Webb TH, White PL, Wymore K,
diatric and geriatric patients with fatal haemolytic Gierke RE, Mahon BE, Griffin PM. 2013.
uraemic syndrome. Microb Pathog 31:5967. Increased recognition of non-O157 Shiga toxin-
17. Okuda T, Tokuda N, Numata S, Ito M, Ohta M, producing Escherichia coli infections in the
Kawamura K, Wiels J, Urano T, Tajima O, United States during 20002010: epidemiologic
Furukawa K. 2006. Targeted disruption of features and comparison with E. coli O157 infec-
Gb3/CD77 synthase gene resulted in the complete tions. Foodborne Pathog Dis 10:453462.
deletion of globo-series glycosphingolipids and 29. Donnenberg MS, Tzipori S, McKee ML,
loss of sensitivity to verotoxins. J Biol Chem 281: OBrien AD, Alroy J, Kaper JB. 1993. The role of
1023010235. the eae gene of enterohemorrhagic Escherichia
18. Ray PE, Liu XH. 2001. Pathogenesis of Shiga coli in intimate attachment in vitro and in a por-
toxin-induced hemolytic uremic syndrome. cine model. J Clin Investig 92:14181424.
Pediatr Nephrol 16:823839. 30. Jerse AE, Yu J, Tall BD, Kaper JB. 1990. A ge-
19. Boyd B, Lingwood C. 1989. Verotoxin receptor netic locus of enteropathogenic Escherichia coli
glycolipid in human renal tissue. Nephron 51:207 necessary for the production of attaching and
210. effacing lesions on tissue culture cells. Proc Natl
20. Ren J, Utsunomiya I, Taguchi K, Ariga T, Tai Acad Sci USA 87:78397843.
T, Ihara Y, Miyatake T. 1999. Localization of 31. Ritchie JM, Thorpe CM, Rogers AB, Waldor
verotoxin receptors in nervous system. Brain Res MK. 2003. Critical roles for stx2, eae, and tir in
825:183188. enterohemorrhagic Escherichia coli-induced di-

170 RITCHIE
arrhea and intestinal inflammation in infant JB, Levine MM. 1993. Role of the eaeA gene in
rabbits. Infect Immun 71:71297139. experimental enteropathogenic Escherichia coli
32. Bonnet R, Souweine B, Gauthier G, Rich C, infection. J Clin Invest 92:14121417.
Livrelli V, Sirot J, Joly B, Forestier C. 1998. Non- 42. Griffin PM, Olmstead LC, Petras RE. 1990.
O157:H7 Stx2-producing Escherichia coli strains Escherichia coli O157:H7-associated colitis. A clin-
associated with sporadic cases of hemolytic- ical and histological study of 11 cases. Gastro-
uremic syndrome in adults. J Clin Microbiol 36: enterology 99:142149.
17771780. 43. Shigeno T, Akamatsu T, Fujimori K, Nakatsuji
33. Newton HJ, Sloan J, Bulach DM, Seemann T, Y, Nagata A. 2002. The clinical significance of
Allison CC, Tauschek M, Robins-Browne RM, colonoscopy in hemorrhagic colitis due to en-
Paton JC, Whittam TS, Paton AW, Hartland terohemorrhagic Escherichia coli O157:H7 infec-
EL. 2009. Shiga toxin-producing Escherichia coli tion. Endoscopy 34:311314.
strains negative for locus of enterocyte efface- 44. Kelly JK, Pai CH, Jadusingh IH, Macinnis
ment. Emerg Infect Dis 15:372380. ML, Shaffer EA, Hershfield NB. 1987. The
34. Hedican EB, Medus C, Besser JM, Juni BA, histopathology of rectosigmoid biopsies from
Koziol B, Taylor C, Smith KE. 2009. Char- adults with bloody diarrhea due to verotoxin-
acteristics of O157 versus non-O157 Shiga toxin- producing Escherichia coli. Am J Clin Pathol
producing Escherichia coli infections in Minnesota, 88:7882.
20002006. Clin Infect Dis 49:358364. 45. Ryan CA, Tauxe RV, Hosek GW, Wells JG,
35. Gerber A, Karch H, Allerberger F, Verweyen Stoesz PA, McFadden HW Jr, Smith PW,
HM, Zimmerhackl LB. 2002. Clinical course and Wright GF, Blake PA. 1986. Escherichia coli
the role of shiga toxin-producing Escherichia coli O157:H7 diarrhea in a nursing home: clinical,
infection in the hemolytic-uremic syndrome in epidemiological, and pathological findings. J In-
pediatric patients, 19972000, in Germany and fect Dis 154:631638.
Austria: a prospective study. J Infect Dis 186:493 46. Inward CD, Howie AJ, Fitzpatrick MM, Rafaat
500. F, Milford DV, Taylor CM. 1997. Renal histo-
36. Hermos CR, Janineh M, Han LL, McAdam AJ. pathology in fatal cases of diarrhoea-associated
2011. Shiga toxin-producing Escherichia coli in haemolytic uraemic syndrome. British Associa-
children: diagnosis and clinical manifestations of tion for Paediatric Nephrology. Pediatr Nephrol
O157:H7 and non-O157:H7 infection. J Clin 11:556559.
Microbiol 49:955959. 47. Tzipori S, Wachsmuth IK, Chapman C, Birden
37. Preussel K, Hohle M, Stark K, Werber D. 2013. R, Brittingham J, Jackson C, Hogg J. 1986. The
Shiga toxin-producing Escherichia coli O157 is pathogenesis of hemorrhagic colitis caused by
more likely to lead to hospitalization and death Escherichia coli O157:H7 in gnotobiotic piglets.
than non-O157 serogroupsExcept O104. PloS J Infect Dis 154:712716.
One 8:e78180. 48. Francis DH, Collins JE, Duimstra JR. 1986. In-
38. Frank C, Werber D, Cramer JP, Askar M, Faber fection of gnotobiotic pigs with an Escherichia coli
M, an der Heiden M, Bernard H, Fruth A, O157:H7 strain associated with an outbreak of
Prager R, Spode A, Wadl M, Zoufaly A, Jordan hemorrhagic colitis. Infect Immun 51:953956.
S, Kemper MJ, Follin P, Muller L, King LA, 49. Tzipori S, McCartney E, Lawson GH, Rowland
Rosner B, Buchholz U, Stark K, Krause G. 2011. AC, Campbell I. 1981. Experimental infection of
Epidemic profile of Shiga-toxin-producing Esch- piglets with cryptosporidium. Res Vet Sci 31:358
erichia coli O104:H4 outbreak in Germany. New 368.
Engl J Med 365:17711780. 50. Tzipori S, Chandler D, Smith M, Makin T,
39. Foubister V, Rosenshine I, Donnenberg MS, Halpin C. 1982. Experimental colibacillosis in
Finlay BB. 1994. The eaeB gene of enteropatho- gnotobiotic piglets exposed to 3 enterotoxigenic
genic Escherichia coli is necessary for signal serotypes. Aust Vet J 59:9395.
transduction in epithelial cells. Infect Immun 51. Tzipori S, Robins-Browne RM, Gonis G, Hayes
62:30383040. J, Withers M, McCartney E. 1985. Entero-
40. Tacket CO, Sztein MB, Losonsky G, Abe A, pathogenic Escherichia coli enteritis: evaluation of
Finlay BB, McNamara BP, Fantry GT, James the gnotobiotic piglet as a model of human in-
SP, Nataro JP, Levine MM, Donnenberg MS. fection. Gut 26:570578.
2000. Role of EspB in experimental human en- 52. Moon HW, Whipp SC, Argenzio RA, Levine
teropathogenic Escherichia coli infection. Infect MM, Giannella RA. 1983. Attaching and effacing
Immun 68:36893695. activities of rabbit and human enteropathogenic
41. Donnenberg MS, Tacket CO, James SP, Escherichia coli in pig and rabbit intestines. Infect
Losonsky G, Nataro JP, Wasserman SS, Kaper Immun 41:13401351.

53. Levine MM, Kaper JB, Black RE, Clements ML. 64. Shringi S, Garcia A, Lahmers KK, Potter KA,
1983. New knowledge on pathogenesis of bacte- Muthupalani S, Swennes AG, Hovde CJ, Call
rial enteric infections as applied to vaccine de- DR, Fox JG, Besser TE. 2012. Differential viru-
velopment. Microbiol Rev 47:510550. lence of clinical and bovine-biased enterohemor-
54. Tzipori S, Gibson R, Montanaro J. 1989. Nature rhagic Escherichia coli O157:H7 genotypes in
and distribution of mucosal lesions associated piglet and Dutch belted rabbit models. Infect
with enteropathogenic and enterohemorrhagic Immun 80:369380.
Escherichia coli in piglets and the role of plasmid- 65. Farmer JJ 3rd, Potter ME, Riley LW, Barrett
mediated factors. Infect Immun 57:11421150. TJ, Blake PA, Bopp CA, Cohen ML, Kaufmann
55. Tzipori S, Karch H, Wachsmuth KI, Robins- A, Morris GK, Remis RS, Thomason BM, Wells
Browne RM, OBrien AD, Lior H, Cohen ML, JG. 1983. Animal models to study Escherichia coli
Smithers J, Levine MM. 1987. Role of a 60- O157:H7 isolated from patients with haemor-
megadalton plasmid and Shiga-like toxins in the rhagic colitis. Lancet i:702703.
pathogenesis of infection caused by enterohem- 66. Potter ME, Kaufmann AF, Thomason BM,
orrhagic Escherichia coli O157:H7 in gnotobiotic Blake PA, Farmer JJ 3rd. 1985. Diarrhea due to
piglets. Infect Immun 55:31173125. Escherichia coli O157:H7 in the infant rabbit.
56. Tzipori S, Gunzer F, Donnenberg MS, de J Infect Dis 152:13411343.
Montigny L, Kaper JB, Donohue-Rolfe A. 1995. 67. Pai CH, Kelly JK, Meyers GL. 1986. Experi-
The role of the eaeA gene in diarrhea and neu- mental infection of infant rabbits with verotoxin-
rological complications in a gnotobiotic piglet producing Escherichia coli. Infect Immun 51:16
model of enterohemorrhagic Escherichia coli in- 23.
fection. Infect Immun 63:36213627. 68. Elliott E, Li Z, Bell C, Stiel D, Buret A, Wallace
57. Tzipori S, Chow CW, Powell HR. 1988. Cerebral J, Brzuszczak I, OLoughlin E. 1994. Modulation
infection with Escherichia coli O157:H7 in humans of host response to Escherichia coli O157:H7 in-
and gnotobiotic piglets. J Clin Pathol 41:1099 fection by anti-CD18 antibody in rabbits. Gastro-
1103. enterology 106:15541561.
58. Francis DH, Moxley RA, Andraos CY. 1989. 69. Li Z, Bell C, Buret A, Robins-Browne R, Stiel D,
Edema disease-like brain lesions in gnotobiotic OLoughlin E. 1993. The effect of enterohemor-
piglets infected with Escherichia coli serotype rhagic Escherichia coli O157:H7 on intestinal
O157:H7. Infect Immun 57:13391342. structure and solute transport in rabbits. Gastro-
59. Dean-Nystrom EA, Melton-Celsa AR, Pohlenz enterology 104:467474.
JF, Moon HW, OBrien AD. 2003. Compara- 70. Lloyd SJ, Ritchie JM, Rojas-Lopez M,
tive pathogenicity of Escherichia coli O157 and Blumentritt CA, Popov VL, Greenwich JL,
intimin-negative non-O157 Shiga toxin-producing Waldor MK, Torres AG. 2012. A double, long
E coli strains in neonatal pigs. Infect Immun polar fimbria mutant of Escherichia coli O157:H7
71:65266533. expresses Curli and exhibits reduced in vivo col-
60. Gunzer F, Hennig-Pauka I, Waldmann KH, onization. Infect Immun 80:914920.
Sandhoff R, Grone HJ, Kreipe HH, Matussek A, 71. Ritchie JM, Waldor MK. 2005. The locus of
Mengel M. 2002. Gnotobiotic piglets develop enterocyte effacement-encoded effector proteins
thrombotic microangiopathy after oral infection all promote enterohemorrhagic Escherichia coli
with enterohemorrhagic Escherichia coli. Am J pathogenicity in infant rabbits. Infect Immun
Clin Pathol 118:364375. 73:14661474.
61. Pohlenz JF, Winter KR, Dean-Nystrom EA. 72. Shifrin Y, Peleg A, Ilan O, Nadler C, Kobi S,
2005. Shiga-toxigenic Escherichia coli-inoculated Baruch K, Yerushalmi G, Berdichevsky T,
neonatal piglets develop kidney lesions that are Altuvia S, Elgrably-Weiss M, Abe C, Knutton
comparable to those in humans with hemolytic- S, Sasakawa C, Ritchie JM, Waldor MK,
uremic syndrome. Infect Immun 73:612616. Rosenshine I. 2008. Transient shielding of int-
62. Dean-Nystrom EA, Gansheroff LJ, Mills M, imin and the type III secretion system of entero-
Moon HW, OBrien AD. 2002. Vaccination of hemorrhagic and enteropathogenic Escherichia
pregnant dams with intimin(O157) protects suck- coli by a group 4 capsule. J Bacteriol 190:5063
ling piglets from Escherichia coli O157:H7 infec- 5074.
tion. Infect Immun 70:24142418. 73. Lloyd SJ, Ritchie JM, Torres AG. 2012. Fim-
63. Dean-Nystrom EA, Pohlenz JF, Moon HW, briation and curliation in Escherichia coli O157:
OBrien AD. 2000. Escherichia coli O157:H7 H7: a paradigm of intestinal and environmental
causes more-severe systemic disease in suckling colonization. Gut Microbes 3:272276.
piglets than in colostrum-deprived neonatal pig- 74. Ritchie JM, Brady MJ, Riley KN, Ho TD,
lets. Infect Immun 68:23562358. Campellone KG, Herman IM, Donohue-Rolfe

172 RITCHIE
A, Tzipori S, Waldor MK, Leong JM. 2008. 84. Agin TS, Cantey JR, Boedeker EC, Wolf MK.
EspFU, a type III-translocated effector of actin 1996. Characterization of the eaeA gene from
assembly, fosters epithelial association and late- rabbit enteropathogenic Escherichia coli strain
stage intestinal colonization by E. coli O157:H7. RDEC-1 and comparison to other eaeA genes
Cell Microbiol 10:836847. from bacteria that cause attaching-effacing le-
75. Zoja C, Corna D, Farina C, Sacchi G, Lingwood sions. FEMS Microbiol Lett 144:249258.
C, Doyle MP, Padhye VV, Abbate M, Remuzzi 85. Cantey JR, Lushbaugh WB, Inman LR. 1981.
G. 1992. Verotoxin glycolipid receptors determine Attachment of bacteria to intestinal epithelial
the localization of microangiopathic process in cells in diarrhea caused by Escherichia coli strain
rabbits given verotoxin-1. J Lab Clin Med 120: RDEC-1 in the rabbit: stages and role of capsule.
229238. J Infect Dis 143:219230.
76. Garcia A, Bosques CJ, Wishnok JS, Feng Y, 86. Sjogren R, Neill R, Rachmilewitz D, Fritz
Karalius BJ, Butterton JR, Schauer DB, Rogers D, Newland J, Sharpnack D, Colleton C,
AB, Fox JG. 2006. Renal injury is a consistent Fondacaro J, Gemski P, Boedeker E. 1994.
finding in Dutch Belted rabbits experimentally Role of Shiga-like toxin I in bacterial enteritis:
infected with enterohemorrhagic Escherichia coli. comparison between isogenic Escherichia coli
J Infect Dis 193:11251134. strains induced in rabbits. Gastroenterology 106:
77. Stone SM, Thorpe CM, Ahluwalia A, Rogers 306317.
AB, Obata F, Vozenilek A, Kolling GL, Kane AV, 87. Richardson SE, Rotman TA, Jay V, Smith CR,
Magun BE, Jandhyala DM. 2012. Shiga toxin Becker LE, Petric M, Olivieri NF, Karmali MA.
2-induced intestinal pathology in infant rabbits is 1992. Experimental verocytotoxemia in rabbits.
A-subunit dependent and responsive to the tyro- Infect Immun 60:41544167.
sine kinase and potential ZAK inhibitor imatinib. 88. Wadolkowski EA, Sung LM, Burris JA, Samuel
Front Cell Infect Microbiol 2:135. JE, OBrien AD. 1990. Acute renal tubular ne-
78. Mobassaleh M, Donohue-Rolfe A, Jacewicz M, crosis and death of mice orally infected with
Grand RJ, Keusch GT. 1988. Pathogenesis of Escherichia coli strains that produce Shiga-like
shigella diarrhea: evidence for a developmentally toxin type II. Infect Immun 58:39593965.
regulated glycolipid receptor for shigella toxin 89. Mohawk KL, OBrien AD. 2011. Mouse models of
involved in the fluid secretory response of rabbit Escherichia coli O157:H7 infection and Shiga toxin
small intestine. J Infect Dis 157:10231031. injection. J Biomed Biotechnol 2011:258185.
79. Mobassaleh M, Gross SK, McCluer RH, 90. Luperchio SA, Schauer DB. 2001. Molecular
Donohue-Rolfe A, Keusch GT. 1989. Quantita- pathogenesis of Citrobacter rodentium and trans-
tion of the rabbit intestinal glycolipid receptor missible murine colonic hyperplasia. Microbes
for Shiga toxin. Further evidence for the devel- Infect 3:333340.
opmental regulation of globotriaosylceramide in 91. Deng W, Li Y, Vallance BA, Finlay BB. 2001.
microvillus membranes. Gastroenterology 97:384 Locus of enterocyte effacement from Citrobacter
391. rodentium: sequence analysis and evidence for
80. Garcia A, Marini RP, Feng Y, Vitsky A, Knox horizontal transfer among attaching and effacing
KA, Taylor NS, Schauer DB, Fox JG. 2002. A pathogens. Infect Immun 69:63236335.
naturally occurring rabbit model of enterohem- 92. Deng W, Puente JL, Gruenheid S, Li Y, Vallance
orrhagic Escherichia coli-induced disease. J Infect BA, Vazquez A, Barba J, Ibarra JA, ODonnell
Dis 186:16821686. P, Metalnikov P, Ashman K, Lee S, Goode D,
81. Panda A, Tatarov I, Melton-Celsa AR, Pawson T, Finlay BB. 2004. Dissecting virulence:
Kolappaswamy K, Kriel EH, Petkov D, systematic and functional analyses of a pathoge-
Coksaygan T, Livio S, McLeod CG, Nataro JP, nicity island. Proc Natl Acad Sci USA 101:3597
OBrien AD, DeTolla LJ. 2010. Escherichia coli 3602.
O157:H7 infection in Dutch belted and New 93. Deng W, de Hoog CL, Yu HB, Li Y, Croxen MA,
Zealand white rabbits. Comp Med 60:3137. Thomas NA, Puente JL, Foster LJ, Finlay BB.
82. De SN, Chatterje DN. 1953. An experimental 2010. A comprehensive proteomic analysis of the
study of the mechanism of action of Vibrio type III secretome of Citrobacter rodentium.
cholerae on the intestinal mucous membrane. J Biol Chem 285:67906800.
J Pathol Bacteriol 66:559562. 94. Mallick EM, McBee ME, Vanguri VK, Melton-
83. Keenan KP, Sharpnack DD, Collins H, Formal Celsa AR, Schlieper K, Karalius BJ, OBrien AD,
SB, OBrien AD. 1986. Morphologic evaluation of Butterton JR, Leong JM, Schauer DB. 2012. A
the effects of Shiga toxin and E coli Shiga-like novel murine infection model for Shiga toxin-
toxin on the rabbit intestine. Am J Pathol 125:69 producing Escherichia coli. J Clin Invest 122:4012
80. 4024.

95. Bielaszewska M, Friedrich AW, Aldick T, Escherichia coli O157:H7 Shiga-like toxin 1 is re-
Schurk-Bulgrin R, Karch H. 2006. Shiga toxin quired for full pathogenicity and activation of
activatable by intestinal mucus in Escherichia coli the p38 mitogen-activated protein kinase path-
isolated from humans: predictor for a severe way in Caenorhabditis elegans. Cell Microbiol 15:
clinical outcome. Clin Infect Dis 43:11601167. 8297.
96. Golan L, Gonen E, Yagel S, Rosenshine I, 107. Hernandez-Doria JD, Sperandio V. 2013. Nu-
Shpigel NY. 2011. Enterohemorrhagic Escherichia trient and chemical sensing by intestinal patho-
coli induce attaching and effacing lesions and gens. Microbes Infect 15:759764.
hemorrhagic colitis in human and bovine intesti- 108. Clarke MB, Hughes DT, Zhu C, Boedeker EC,
nal xenograft models. Dis Model Mech 4:8694. Sperandio V. 2006. The QseC sensor kinase: a
97. Woods JB, Schmitt CK, Darnell SC, Meysick bacterial adrenergic receptor. Proc Natl Acad Sci
KC, OBrien AD. 2002. Ferrets as a model system USA 103:1042010425.
for renal disease secondary to intestinal infection 109. Rasko DA, Moreira CG, Li de R, Reading NC,
with Escherichia coli O157:H7 and other Shiga Ritchie JM, Waldor MK, Williams N, Taussig R,
toxin-producing E. coli. J Infect Dis 185:550554. Wei S, Roth M, Hughes DT, Huntley JF, Fina
98. Cowan LA, Hertzke DM, Fenwick BW, MW, Falck JR, Sperandio V. 2008. Targeting
Andreasen CB. 1997. Clinical and clinicopatho- QseC signaling and virulence for antibiotic de-
logic abnormalities in greyhounds with cutaneous velopment. Science 321:10781080.
and renal glomerular vasculopathy: 18 cases 110. Pacheco AR, Curtis MM, Ritchie JM, Munera
(19921994). J Am Vet Med Assoc 210:789793. D, Waldor MK, Moreira CG, Sperandio V. 2012.
99. Fenwick B, Cowan L. 1998. Canine model of Fucose sensing regulates bacterial intestinal col-
hemolytic-uremic syndrome, p 268277. In Kaper onization. Nature 492:113117.
JB, OBrien AD (ed), Escherichia coli O157:H7 and 111. Crane J. 2013. Role of host xanthine oxidase in
Other Shiga Toxin-Producing E coli strains. ASM infection due to enteropathogenic and Shiga-
Press, Washington, DC. toxigenic Escherichia coli. Gut Microbes 4:388
100. Raife T, Friedman KD, Fenwick B. 2004. 391.
Lepirudin prevents lethal effects of Shiga toxin in 112. Crane JK, Naeher TM, Broome JE, Boedeker
a canine model. Thromb Haemost 2:387393. EC. 2013. Role of host xanthine oxidase in infec-
101. Siegler RL, Obrig TG, Pysher TJ, Tesh VL, tion due to enteropathogenic and Shiga-toxigenic
Denkers ND, Taylor FB. 2003. Response to Shiga Escherichia coli. Infect Immun 81:11291139.
toxin 1 and 2 in a baboon model of hemolytic 113. Zumbrun SD, Melton-Celsa AR, Smith MA,
uremic syndrome. Pediatr Nephrol 18:9296. Gilbreath JJ, Merrell DS, OBrien AD. 2013.
102. Taylor FB Jr, Tesh VL, DeBault L, Li A, Chang Dietary choice affects Shiga toxin-producing
AC, Kosanke SD, Pysher TJ, Siegler RL. 1999. Escherichia coli (STEC) O157:H7 colonization
Characterization of the baboon responses to and disease. Proc Natl Acad Sci USA 110:E2126
Shiga-like toxin: descriptive study of a new pri- E2133.
mate model of toxic responses to Stx-1. Am J 114. Willing BP, Vacharaksa A, Croxen M,
Pathol 154:12851299. Thanachayanont T, Finlay BB. 2011. Altering
103. Stearns-Kurosawa DJ, Collins V, Freeman S, host resistance to infections through microbial
Tesh VL, Kurosawa S. 2010. Distinct physiologic transplantation. PloS One 6:e26988.
and inflammatory responses elicited in baboons 115. Ferreira RB, Gill N, Willing BP, Antunes LC,
after challenge with Shiga toxin type 1 or 2 from Russell SL, Croxen MA, Finlay BB. 2011. The
enterohemorrhagic Escherichia coli. Infect Immun intestinal microbiota plays a role in Salmonella-
78:24972504. induced colitis independent of pathogen coloni-
104. Zotta E, Lago N, Ochoa F, Repetto HA, Ibarra zation. PloS One 6:e20338.
C. 2008. Development of an experimental he- 116. Sekirov I, Finlay BB. 2009. The role of the in-
molytic uremic syndrome in rats. Pediatr Nephrol testinal microbiota in enteric infection. J Physiol
23:559567. 587:41594167.
105. Anyanful A, Dolan-Livengood JM, Lewis T, 117. Wen L, Ley RE, Volchkov PY, Stranges PB,
Sheth S, Dezalia MN, Sherman MA, Kalman Avanesyan L, Stonebraker AC, Hu C, Wong FS,
LV, Benian GM, Kalman D. 2005. Paralysis Szot GL, Bluestone JA, Gordon JI, Chervonsky
and killing of Caenorhabditis elegans by entero- AV. 2008. Innate immunity and intestinal mic-
pathogenic Escherichia coli requires the bac- robiota in the development of Type 1 diabetes.
terial tryptophanase gene. Mol Microbiol 57:988 Nature 455:11091113.
1007. 118. Sekirov I, Russell SL, Antunes LC, Finlay BB.
106. Chou TC, Chiu HC, Kuo CJ, Wu CM, Syu WJ, 2010. Gut microbiota in health and disease.
Chiu WT, Chen CS. 2013. Enterohaemorrhagic Physiol Rev 90:859904.

174 RITCHIE
119. Bentancor LV, Bilen MF, Mejias MP, 2008. Shiga toxin-mediated disease in MyD88-
Fernandez-Brando RJ, Panek CA, Ramos MV, deficient mice infected with Escherichia coli O157:
Fernandez GC, Isturiz M, Ghiringhelli PD, H7. Am J Pathol 173:14281439.
Palermo MS. 2013. Functional capacity of Shiga- 128. Eaton KA, Friedman DI, Francis GJ, Tyler JS,
toxin promoter sequences in eukaryotic cells. Young VB, Haeger J, Abu-Ali G, Whittam TS.
PloS One 8:e57128. 2008. Pathogenesis of renal disease due to
120. Bentancor LV, Mejias MP, Pinto A, Bilen MF, enterohemorrhagic Escherichia coli in germ-free
Meiss R, Rodriguez-Galan MC, Baez N, Pedrotti mice. Infect Immun 76:30543063.
LP, Goldstein J, Ghiringhelli PD, Palermo MS. 129. Tyler JS, Beeri K, Reynolds JL, Alteri CJ,
2013. Promoter sequence of Shiga toxin 2 (Stx2) is Skinner KG, Friedman JH, Eaton KA,
recognized in vivo, leading to production of bio- Friedman DI. 2013. Prophage induction is en-
logically active Stx2. MBio 4:e00501-13. hanced and required for renal disease and le-
121. Lengeling A, Mahajan A, Gally DL. 2013. Bac- thality in an EHEC mouse model. PLoS Pathog 9:
teriophages as pathogens and immune modu- e1003236.
lators? MBio 4:e00868-13. 130. Taguchi H, Takahashi M, Yamaguchi H, Osaki
122. Wadolkowski EA, Burris JA, OBrien AD. 1990. T, Komatsu A, Fujioka Y, Kamiya S. 2002. Ex-
Mouse model for colonization and disease caused perimental infection of germ-free mice with
by enterohemorrhagic Escherichia coli O157:H7. hyper-toxigenic enterohaemorrhagic Escherichia
Infect Immun 58:24382445. coli O157:H7, strain 6. J Med Microbiol 51:336
123. Fujii J, Kita T, Yoshida S, Takeda T, Kobayashi 343.
H, Tanaka N, Ohsato K, Mizuguchi Y. 1994. 131. Kim SH, Lee YH, Lee SH, Lee SR, Huh JW, Kim
Direct evidence of neuron impairment by oral SU, Chang KT. 2011. Mouse model for hemolytic
infection with verotoxin-producing Escherichia uremic syndrome induced by outer membrane
coli O157:H- in mitomycin-treated mice. Infect vesicles of Escherichia coli O157:H7. FEMS
Immun 62:34473453. Immunol Med Microbiol 63:427434.
124. Fernandez-Brando RJ, Miliwebsky E, Mejias 132. Sauter KA, Melton-Celsa AR, Larkin K,
MP, Baschkier A, Panek CA, Abrey-Recalde Troxell ML, OBrien AD, Magun BE. 2008.
MJ, Cabrera G, Ramos MV, Rivas M, Palermo Mouse model of hemolytic-uremic syndrome
MS. 2012. Shiga toxin-producing Escherichia coli caused by endotoxin-free Shiga toxin 2 (Stx2) and
O157 : H7 shows an increased pathogenicity in protection from lethal outcome by anti-Stx2
mice after the passage through the gastrointesti- antibody. Infect Immun 76:44694478.
nal tract of the same host. J Med Microbiol 133. Tesh VL, Burris JA, Owens JW, Gordon VM,
61:852859. Wadolkowski EA, OBrien AD, Samuel JE. 1993.
125. Karpman D, Connell H, Svensson M, Scheutz F, Comparison of the relative toxicities of Shiga-like
Alm P, Svanborg C. 1997. The role of lipopoly- toxins type I and type II for mice. Infect Immun
saccharide and Shiga-like toxin in a mouse model 61:33923402.
of Escherichia coli O157:H7 infection. J Infect Dis 134. Kang G, Pulimood AB, Koshi R, Hull A,
175:611620. Acheson D, Rajan P, Keusch GT, Mathan VI,
126. Mohawk KL, Melton-Celsa AR, Zangari T, Mathan MM. 2001. A monkey model for entero-
Carroll EE, OBrien AD. 2010. Pathogenesis of hemorrhagic Escherichia coli infection. J Infect
Escherichia coli O157:H7 strain 86-24 following Dis 184:206210.
oral infection of BALB/c mice with an intact 135. Beery JT, Doyle MP, Schoeni JL. 1985. Coloni-
commensal flora. Microb Pathog 48:131142. zation of chicken cecae by Escherichia coli asso-
127. Calderon Toledo C, Rogers TJ, Svensson M, ciated with hemorrhagic colitis. Appl Environ
Tati R, Fischer H, Svanborg C, Karpman D. Microbiol 49:310315.

Virulence Gene Regulation
JAY L. MELLIES1 and EMILY LORENZEN2

9
As a species, Escherichia coli is highly successful, adapting to inhabit the lower
intestine of warm-blooded animals. Commensal E. coli, part of the normal biota,
resides harmlessly in the gut, producing vitamin K. However, E. coli also causes
three types of disease in humans: urinary tract infections, sepsis in newborns,
and diarrheal disease. Enterohemorrhagic E. coli (EHEC) plays a prominent
role in the third type of illness. It has been estimated that, for the pan genome
of E. coli, the nonpathogenic and pathogenic strains only contain a core set of
genes comprising approximately 20% of any one genome (1, 2). Much of the
horizontally acquired genetic information is clustered within genomic islands
in pathogens. As for EHEC, this has allowed the organism to not only attach
and colonize the large intestine of humans and other animals, to outcompete
commensal E. coli and other bacteria at the site of infection, but also to cause
serious disease.
Horizontally acquired genetic information in EHEC results in evolution into
a specific pathotypegenotype dictates phenotype. How this genetic informa-
tion is controlled is of equal importance for the success and virulence of the
organism. Indeed, Abu-Ali et al. (3) investigated differences in virulence gene
regulation in two distinct EHEC isolate lineages, clade 8 and clade 2. A clade is a
group of EHEC isolates with one ancestor and all its descendants. Eight clades
of E. coli O157 isolates were defined by single nucleotide polymorphism (SNP)
1
Department of Biology, Reed College, Portland, OR 97202; 2Rockefeller University, New York, NY 10065.
175

176 MELLIES AND LORENZEN
analyses, where clade 8 was a group of hy- systems have been described: a glucose-
pervirulent bacteria compared to the other repressed, or oxidative, system and glutamate-
seven clades (4). By examining multiple and arginine-dependent systems (for review
strains per lineage, the investigators found see reference 6). Evidence suggests that the
increased expression of horizontally acquired glucose-repressed system is used for EHEC
virulence genes in clade 8 versus clade 2. survival in acidic food items (7). For the
Genes expressed to higher levels in clade 8, glutamate-dependent system, glutamate de-
which is associated with a greater number of carboxylases GadA and GadB convert gluta-
cases of E. coli hemorrhagic disease com- mate to g-amino butyric acid, or GABA (810).
pared to those in clade 2, included several The arginine decarboxylase AdiA converts
virulence factors: rpoS, 29 of the 41 locus of arginine to agmatine (11). The Gad system is
enterocyte effacement (LEE)-encoded genes, regulated by a number of environmental con-
gadX involved in acid tolerance, and the ditions and the global regulatory proteins
pO157 plasmid-encoded stcE adhesin and H-NS and CRP and alternate sigma factors
hlyA hemolysin. These data provide evidence RpoS and RpoN (810, 12, 13) (Fig. 1). For
that genetic regulation correlates with EHEC both amino acid-dependent systems, pH ho-
virulence. meostasis is maintained by displacing the
However, our understanding of the regula- a-carboxyl group of the amino acids with a
tory network controlling EHEC pathogenesis proton that is transported from the environ-
remains incomplete. Some of the questions ment to the cytoplasm. Cytoplasmic glutamate
pursued by researchers include the following: and arginine are restored by their respective
What are the molecular signals perceived by
EHEC strains in the human host and within
cattle that allow the bacteria to properly ex-
press virulence traits? What are the signals and
bacterial responses required to pass through
the acid environment of the stomach, to ulti-
mately reside in the large intestine, and to
cause disease in humans but to attach to and
colonize harmlessly the recto-anal junction
(RAJ) of cattle (5). How are expression of
the type III secretion system (T3SS), attach-
ment to host cell surfaces, the effector mole-
cules destined for translocation into host
cells, secretion itself, and the Shiga toxin
leading to serious disease controlled? This
article summarizes our current knowledge of
EHEC virulence gene regulation, indicating
that spatiotemporal control of pathogenesis
in humans and carriage in cattle is coming
to light. FIGURE 1 Control of acid tolerance by EHEC. Resident
ora produce AHL signaling molecules that stimulate
expression of the glutamate-dependent acid resis-
ACID TOLERANCE tance system (14). The GAD system removes excess
protons by exchanging the alpha-carboxyl groups of
glutamate with a proton. The resulting GABA mole-
As for any intestinal pathogen that causes cules are transported out of the cell in exchange for
disease, EHEC must survive the acidic envi- additional glutamate (143).
ronment of the stomach. Three acid resistance doi:10.1128/microbiolspec.EHEC-0004-2013.f1

CHAPTER 9 EHEC Virulence Gene Regulation 177
antiporters, GadC and AdiC. The glutamate-

dependent Gad system provides a high level of
acid tolerance and is necessary for passage
through the stomachs and for colonization of
the RAJ in cattle (7).
In cattle, EHEC passes through the rumen
to eventually colonize the RAJ, and the LuxR
homolog SdiA is involved in this process. Al-
though EHEC does not possess the LuxI
N-acyl homoserine lactone (AHL) synthase,
SdiA perceives the oxo-C6-homoserine lac-
tone produced by other bacteria in the rumen
(Fig. 1). In turn, SdiA increases gadX expres-
sion (14), which is a regulator of the gad genes,
encoding glutamate-dependent acid resis-
tance. Using a cattle model, investigators dem-
onstrated that wild-type EHEC outcompeted
the sdiA deletion strain, which was defective
in colonization of the RAJ over the 6-day as-
say. The finding that an estimated 70 to 80% of FIGURE 2 Regulation of agellar biosynthesis and
cattle herds in the United States carry EHEC is motility. Extracellular signals norepinephrine, epi-
nephrine, and AI-3 stimulate expression of iA
a major step forward in our understanding of
(motility) and hDC (biosynthesis) through the two-
the infection process in the EHEC reservoir. component system QseBC (16). Following induction
of the LEE operons, grlA inhibits expression of iC and
hDC (28). Blunted arrow indicates negative control.
MOTILITY doi:10.1128/microbiolspec.EHEC-0004-2013.f2
Flagellar motility is important in the initial recognizing QseC sensor kinase attenuated
stages of infection for many bacterial patho- EHEC virulence in a rabbit model of infection.
gens. For EHEC, motility and taxis are stim- Collectively, these data clearly demonstrate
ulated through the two-component system, that flagellar motility, and the associated sig-
QseC and QseB, which perceives the AI-3 naling, is necessary for virulence (15, 16).
signal and the interkingdom communication Additional molecules are known to stimu-
molecules epinephrine and norepinephrine late flagellar motility, and data indicate tem-
(1518) (Fig. 2). QseBC signaling activates poral control correlates with the infection
expression of the flhDC master regulators and process. Three classes of promoters are re-
FliA, an alternate sigma factor, ultimately turn- sponsible for transcribing the flagellar operons
ing on a number of operons necessary for (23). The flhDC operon is designated class I
flagellar biosynthesis and motility. Further- and is required for class II promoter activity,
more, motility and biofilm formation are en- which includes the fliA operon. In turn, class II
hanced in the presence of epinephrine and promoters are required for class III expres-
norepinephrine (19). It is important to note sion. Butyrate, a short-chain fatty acid found
that the AI-3 quorum-sensing molecule is dis- in the large intestine, increases expression of
tinct from AI-2, which was initially thought to flhDC (24) (Fig. 2). Regulation of flhDC by
control motility and type III secretion (2022). butyrate is leucine-responsive regulatory pro-
Unlike AI-2, AI-3 is not dependent on LuxS for tein (LRP)-dependent, whereas induction of
synthesis. Deletion of the gene encoding the fliA by this short-chain fatty acid is LRP-
AI-3 and epinephrine- and norepinephrine- independent. This regulation does not occur in

the presence of propionate or acetate. How- expressed only at the appropriate site of in-
ever, propionate and acetate do increase fliC fection. A glycolytic environment, at 0.4% glu-
(class III) expression but not through flhDC. cose, inhibits ler and LEE expression, whereas
Short-chain fatty acids are found in concen- 0.1% glucose mimicking gluconeogenesis con-
trations ranging from 20 to 140 mM in the large ditions enhances their expression (31). The
bowel, and recognition of this signal, along LEE pathogenicity island (PAI) encodes T3SS
with the AI-3 and interkingdom signaling mol- and is organized into five major polycistronic
ecules, likely contributes to EHEC niche rec- operons and a number of bicistronic and mo-
ognition and adaptation. Some evidence also nocistronic genes (32, 33). In gluconeogenesis,
suggests that the flagellum acts as an initial instead of oxidizing glucose, glucose is pro-
adhesin to epithelial cells from the bovine ter- duced from two to three carbon molecules,
minal rectum (25). such as acetate, succinate, or pyruvate. Con-
Once motility is no longer needed, or prior sistently, glucose and glycerol inhibit ler
to its necessity for colonization, the flagella expression, while succinate stimulates ler,
are downregulated, an important step toward mimicking glycolysis and gluconeogenesis,
avoiding immune recognition. One observed respectively (Fig. 3). This regulation occurs
mechanism is that the LuxR homolog SdiA through the proteins KdpE, a response regu-
downregulates FliC expression (26). In another lator that also senses osmotic stress, and Cra,
mechanism, mucin, produced by epithelial tis- both of which bind to ler regulatory DNA.
sues, diminishes expression of flagellar genes.
Mucin on agar plates repressed motility, and
transcriptome analysis and quantitative PCR
confirmed these data (27). In addition, the GrlA
regulator, expressed from the LEE, decreased
flagella biosynthesis by downregulating fliC
and flhDC (28) (Fig. 2). Consistent with flagella
not being needed after surface adherence, Tobe
et al. observed that after 5 h of attachment to
epithelial cells, flagella were downregulated in
a GrlA-dependent manner. Furthermore, in a
neonatal meningitis-causing E. coli strain, the
E. coli common pilus (ECP), found in patho-
gens and nonpathogens alike, is necessary for
attachment and biofilm formation. The regu-
lator of ECP, EcpR, also downregulates flagel-
lar motility after adherence occurs (29). By
downregulating the flagellar master regulators
flhDC by GrlA in EHEC, spatiotemporal regu-
lation controls pathogenesis in coordination FIGURE 3 Inhibition of EHEC effector molecules in the
with the transition from a planktonic to an cattle rumen and human small intestine. Resident
ora in the rumen produce the signaling molecule
adhesive lifestyle. AHL that causes SdiA to fold and inhibit transcrip-
tion of ler (14). Glycolytic conditions in the small in-
testine inhibit the ler transcriptional activators KdpE
CONTINUING THE INTESTINAL JOURNEY and Cra (31). Fucose, a component of cell-surface
glycans, signals through the two-component system
FusKR to inhibit Ler expression (36). The transcrip-
Carbohydrate recognition and metabolism are tional silencer H-NS maintains Ler downregulation
important for EHEC niche adaptation (30) (48). Blunted arrows indicate negative control.
and ensure that EHEC virulence proteins are doi:10.1128/microbiolspec.EHEC-0004-2013.f3

Deletion of kdpE and cra results in ablation of operons 1 through 5 are thus stimulated in
attaching and effacing (A/E) lesion formation the presence of ethanolamine. Though there
by EHEC in vitro(31). are 17 ethanolamine utilization genes, eut,
In the gut, EHEC also perceives and re- necessary for the use of this compound as a
sponds to the sugar fucose, which enhances nitrogen source, the observed signaling in
EHEC colonization (34). In the mammalian EHEC is independent of ethanolamine me-
intestine, fucose is abundant, reaching con- tabolism, but partially dependent on the EutR
centrations of 100 M (35), because it is transcriptional regulator that binds ethanol-
cleaved from host mucin glycoproteins by amine (38) (Fig. 4). Furthermore, there is ev-
fucosidases produced by the commensal bac- idence that ethanolamine confers a growth
terium Bacterioides thetaiotamicron. EHEC advantage to EHEC during the stationary
perceives the fucose signal via the two- phase of growth, when this compound is used
component regulatory pair FusK, the sensor as a nitrogen source in media mimicking bo-
kinase, and FusR, the response regulator (36) vine intestinal contents (39). Ethanolamine
(Fig. 3). Transcriptional activity of the LEE1 is a part of normal physiology of the human
operon-encoded ler is increased in DfusK and bovine large intestine membranes, and
and DfusR strains. Thus these data indicate for EHEC, it provides a growth advantage
that the fucose-sensing, two-component sys- and signals for expression of the T3SS as
tem represses Ler and ultimately T3SS ex- part of ecological niche recognition and ad-
pression. Regulation is direct, as purified FusR aptation. This signaling most likely coincides
binds to ler regulatory sequences, and binding with close proximity or attachment to the
is enhanced when the protein is in the phos- host epithelium.
phorylated state. This carbohydrate signaling
is independent of using fucose as a carbon
source. The importance of fucose signaling in
T3SS EXPRESSION AND ATTACHMENT
virulence is highlighted by the findings that in
an infant rabbit model of infection, fusK and
LEE
fusR deletion mutants are outcompeted by the
wild-type strain (36). In sum, ler and the T3SS Elaboration of the T3SS of EHEC allows for
are repressed by high glucose concentrations, attachment to the intestinal epithelium in
and the presence of mucus-derived fucose, cattle, specifically to the epithelial cells of
while stimulated under gluconeogenesis con- the RAJ (5, 40). Disruption or deletion of
ditions, is mimicked by the presence of mol- genes encoding components of this apparatus
ecules such as succinate, pyruvate, acetate, reduces colonization in the bovine host (41,
and propionate. 42). Furthermore, the translocated intimin
EHEC virulence gene regulation has been receptor (Tir) and intimin proteins, facilitators
correlated further with intestinal physiology. of the tight attachment of the A/E intestinal
Butyrate, found in the large intestine, in- lesions, play a critical role in colonization of a
creases expression of the T3SS (37) (Fig. 4). neonatal calf model of infection (43). In ad-
More recently, ethanolamine, a bacterial and dition to binding to the bacterial-derived Tir
animal cell membrane component that is re- molecule, intimin also binds to integrins and
leased into the lumen of the gut upon normal nucleolin on the host cell surface. The mo-
epithelial cell turnover, has been implicated lecular syringe, the apparatus necessary for
as a molecule important in EHEC virulence altering signaling events in the formation of
gene regulation. Ethanolamine is perceived actin-rich pedestals, ultimately injects 50
by EHEC, leading to increased expression of distinct effector molecules into host cells
the quorum-sensing regulators QseA, QseC, (44). Assembly of the apparatus, expression of
and QseE and Ler (38). In addition, the LEE both LEE and non-LEE encoded effectors,

FIGURE 4 Stimulation of LEE and non-LEE effector molecules required for infection by EHEC. In the large in-
testine of humans and RAJ, the host-produced hormones norepinephrine and epinephrine, EHEC-derived
signaling molecule AI-3, and sulfate and phosphate trigger the two-component sensors QseC,F and/or QseEF
to upregulate ler and espFu transcription (17, 85). In addition, ethanolamine, a cell membrane component,
stimulates ler production through EutR and QseC (38). Gluconeogenic conditions of the large intestine activate
the ler activators KdpE and Cra, and butyrate directly activates LEE transcriptional activity (37, 125). The RNA
chaperone Hfq affects LEE expression through interactions with Ler mRNA but has negative or positive effects
depending on the strain of EHEC, as indicated by a dashed arrow (7173). Finally, nutrient deprivation asso-
ciated with the infection site activates the stringent response leading to production of ppGpp, which promotes
expression of the LEE transcriptional activators pchA and pchB and the non-LEE effector NleA (68, 117). The
asterisk indicates that ler (LEE1) expression is upregulated by several other factors, including temperature, pH,
iron, ammonium, calcium, bicarbonate, and the regulatory proteins IHF, Fis, BipA, PerC, and GadX (previously
reviewed in reference 45). Blunted arrow indicates negative control.
and their translocation must be regulated in a directly silenced by H-NS (4648), and thus,
spatiotemporally correct manner to establish much of the regulatory network described
infection and to avoid immune detection in for the LEE directly involves relieving H-NS-
both humans and cattle. mediated repression. The LEE1-encoded Ler
The LEE PAI, encoding the T3SS, is si- protein, an H-NS homolog, is a key regulator
lenced by H-NS (for review see reference 45). of the LEE and acts as an antisilencer (32). As
Navarre et al. demonstrated that the function most H-NS-controlled genes are regulated in
of H-NS, conserved across the Enterobac- response to environmental adaptation (49),
teriaceae, is to silence horizontally transferred numerous environmental inputs and regula-
genetic elements, indicated by a lower GC or tory proteins control ler gene expression.
higher AT content than the resident genome. To demonstrate the many inputs control-
Along with LEE1, multiple LEE operons are ling ler gene expression, researchers working

with both EHEC and enteropathogenic E. coli light the importance of coordinating expres-
(EPEC) found that Ler is stimulated in sion of LEE genes within the host. RNaseE-
response to environmental signals such as dependent RNA processing occurs at the
temperature, pH, iron, ammonium, calcium, LEE4 operon, resulting in differential control
bicarbonate (5054), and quorum-sensing of expression of the SepL protein. This RNA
signaling (55, 56), as well as the proteins IHF processing is thought to be involved in the
(57), Fis (58), BipA (59), PerC homologs, contact regulation that controls the switch
or Pch (60), GrlRA (61, 62), and GadX (63) from translocator to effector type III secretion
(Fig. 4). The GrlRA regulators constitute a (62, 69) and the regulation of EspADB pro-
complicated feedback loop whereby GrlA teins necessary for forming the filament and
directly activates ler expression, and GrlR pore that ultimately embed in the host cell
acts posttranslationally to reduce the cellular membrane (70). In EHEC strain 86-24, sero-
quantity of the GrlA protein (28, 61, 64, 65). type O157:H7, the RNA chaperone Hfq acts
Once expressed, Ler, in turn, acts as an positively on the expression of ler and LEE
H-NS antisilencer to increase the expression operons 2 through 5 in all phases of growth
of LEE2, LEE3, LEE4, and LEE5 operons (71) (Fig. 4). This study shows that Hfq in-
(Fig. 4). creases the expression of the quorum-sensing,
LEE gene expression is stimulated as two-component system QseBC. QseC also acts
EHEC transitions into stationary phase, con- to increase Ler expression and thus illustrates
trolled by quorum-sensing signaling, and also coordinated regulation. Other researchers,
in response to nutrient deprivation, much though working with strain EDL933 as op-
through direct regulation of ler (21) (Fig. 4). posed to 86-24, observed that the RNA chap-
Nutrient deprivation, the starvation for a erone Hfq negatively affects the expression
number of nutritional requirements such as of the LEE (72, 73). Shakhnovich et al. ob-
amino acids, carbon, nitrogen, or phosphorus, served that Hfq negatively regulates expres-
induces the stringent response (for review see sion of the LEE in a Ler-dependent manner
reference 66). The stringent response pro- and also controls expression of a number
duces the signaling molecule ppGpp, a process of non-LEE effector molecules. The ler, or
dependent on the synthase RelA and SpoT, LEE1 transcript, was a direct target of Hfq
a hydrolase and synthase (67). Expression of regulation. Congruently, Hansen and Kaper
the EHEC LEE and adherence to Caco-2 cells observed that Hfq affects Ler expression
in culture are increased in response to nutri- by negatively affecting the GrlRA regulators
ent deprivation in a RelA-dependent manner, through posttranscriptional control, the sta-
whereby increased pools of ppGpp are created bility of the grlRA transcript, while nega-
(68). The protein DksA, which interacts with tively affecting LEE expression in stationary
RNA polymerase in the transcription complex, phase in a manner independent of grlRA.
is also involved in this regulation. Regulation Consistently with this set of observations, re-
of EspB and Tir, and thus the LEE4 and LEE5 searchers also demonstrated that Hfq nega-
operons, is controlled, in part, by ppGpp in- tively affects LEE expression in the related
teraction with the ler, pchA, and pchB genes. AE pathogen EPEC (73). Thus, LEE genes are
Thus the stringent response rapidly activates apparently regulated by Hfq in an opposite
expression of the T3SS components when manner in the O157:H7 strains 86-24 and
nutrients are limiting and upon transitioning EDL933. Collectively, these regulatory obser-
into stationary phase, conditions EHEC is pre- vations illustrate the intricate and variable
dicted to encounter when entering the lower nature of virulence gene control in related but
intestine. distinct isolates associated with outbreaks of
A number of posttranscriptional control EHEC disease and the evolution of regulatory
mechanisms add to the complexity and high- networks associated with niche adaptation.

Additionally, a number of O-island regula- occurs by perceiving AHLs produced by other

tors contribute to the expression of the LEE. bacteria, leading to alterations in EHEC gene
O-islands are horizontally acquired clusters expression.
of genes that reside in the EHEC genome. AHL molecules are produced by the resi-
The regulators EtrA and EivF are encoded in dent flora in the rumen of the bovine intestine,
a cryptic, second T3SS gene cluster in EHEC, indicated by an Agrobacterium reporter strain,
and negatively regulate the LEE PAI and se- but are not found in the terminal rectum,
cretion (74). Encoded within O-island 51, the the site of EHEC colonization (14). Through
regulator RgdR is a positive regulator of the SdiA sensing of the oxo-C6-AHL signal LEE
LEE and secretion, through the Ler regulatory genes, including ler located in LEE1 and
cascade (75). Thus expression of the T3SS is LEE4 operons, are downregulated (Fig. 3).
controlled not only by LEE-encoded regula- It has been proposed that quorum sensing
tors and regulatory elements that are part of facilitates EHEC escaping the rumen en route
the core E. coli genome but also by genes ac- to the site of colonization, the RAJ in cattle,
quired by horizontal gene transfer. by downregulation and avoiding inappropriate
PerC is an important regulator of the EPEC expression of the LEE, while inducing the gad
LEE, though a perC-like gene does not exist acid tolerances system, as described above
on the pO157 virulence plasmid of EHEC (Fig. 1).
(33, 46, 76). However, in the O157:H7 Sakai Elucidation of the interkingdom signaling
stain, Iyoda and Watanabe (60) identified five systems of EHEC controlling virulence has
chromosomally encoded PerC homologs (pch). revealed how these organisms communicate
Of those identified, pchA, pchB, and pchC, but with members of their own species, resident
not pchD or pchE, were confirmed to regulate bacteria, and the host organism. The QseA
the LEE when expressed in trans. Using quorum-sensing regulator acts directly on
double mutations, the authors demonstrated the expression of the LEE1 operon, specifi-
that, in combination, pchA and pchB, or pchA cally at the P1 promoter (81, 82) (Fig. 4). In-
and pchC resulted in significantly decreased deed, multiple two-component regulators
expression of the LEE, through the global perceive chemical signals and the EHEC-
regulator Ler, and reduction of adherence to derived, aromatic AI-3 molecule to control
HEp-2 epithelial cells in culture (Fig. 4). In- expression of the T3SS. The sensor kinase
terestingly, the pchABC genes of EHEC and QseE responds to sulfate, phosphate, and
perC of EPEC are interchangeable in their host-derived epinephrine to phosphorylate
ability to activate expression of the LEE1 its cognate response regulator QseF that ulti-
operons of both organisms (77). Thus the mately stimulates ler expression (83) (Fig. 4).
EHEC pch genes, with the exception of pchD The sensor kinase QseC perceives the cate-
and pchE, all encoded within phage-like ele- cholamines epinephrine and norepinephrine
ments, are involved in pathogenesis. and the EHEC-produced AI-3 molecule to
Intense investigation has been directed to- increase ler transcription through phospho-
ward cell-to-cell communication or quorum- rylation of KdpE (17), leading to the elabora-
sensing signaling of the LEE and its cognate tion of the type III system. To add emphasis
effector molecules. The best understood of to the importance of quorum-sensing signal-
the autoinducers, AHLs, are not produced by ing controlling virulence, in an infant rabbit
E. coli. Thus EHEC strains do not have the model of infection, deletion or inhibition of
LuxI AHL synthase, though they do possess a qseC in rabbit EPEC severely attenuated vir-
LuxR-type transcriptional regulator SdiA (78, ulence (15, 84), and a qseF deletion mutant
79), and structural studies indicate that SdiA of EHEC does not form A/E lesions (85).
properly folds only in the presence of AHLs Thus the ability to respond to host catechol-
(80). Therefore, regulation through SdiA amines and AI-3 is required for EHEC to

stimulate T3SS expression and to establish both biotic and abiotic, but a comprehensive
infection. understanding of the role of these adhesins in
human disease, carriage in cattle, and survival
and propagation in food does not exist.
Non-LEE
However, the long polar fimbria 1 (Lpf1)
A number of non-T3SS adhesins, either dem- of E. coli O157:H7 is a well-characterized ex-
onstrated to be necessary for or hypothesized ample of one such adhesion. Lpf1 facilitates
to be involved in attachment to host cells, have binding of EHEC to not only epithelial cells
been described (Table 1). Upon sequencing but also extracellular matrix proteins, includ-
the E. coli O157:H7 genome of strain EDL933, ing fibronectin, laminin, and collagen IV (101).
it was noted that this organism has at least 10 Lpf1 is tightly regulated, with maximal ex-
fimbrial gene clusters and 13 regions that en- pression in late exponential phase of growth
code nonfimbrial adhesins, some of which in iron-deprived and slightly acidic envi-
were not found in nonpathogenic E. coli (86, ronments (102) (Table 1). Two lpf loci exist in
87). Extracellular structures include the E. coli E. coli O157:H7, and when one or both are
YcbQ laminin-binding fimbriae (ELF) (88), deleted, colonization in animal infection
two long polar fimbriae (LPF) (89, 90), the F9 models is altered (103, 104) (Table 1). The
fimbriae (91), a type IV pilus called hemor- Lpf1 fimbria is coordinately regulated with the
rhagic coli pilus (HCP) (92), curli (93, 94), LEE because the operon is silenced by H-NS,
OmpA (95), the EHEC factor for adherence while Ler functions as its antisilencer (105).
(Efa1) (96), the IgrA homolog adhesin (Iha) ECP, common to both pathogens and non-
(97), the ECP, the autotransporter protein pathogens, has also been described. Deletion
EhaG (98), and the pO157 virulence plasmid- of the ecp genes resulted in decreased ad-
encoded StcE (99, 100). Surely these mole- herence of both O157:H7 strain EDL933 and
cules contribute to adherence to surfaces, commensal E. coli to human epithelial cells
in culture (106). EcpR, a LuxR-like regulator
TABLE 1 Regulation of non-LEE adhesins and factors containing a helix-turn-helix DNA-binding
involved in EHEC adherence motif, controls expression of the ECP (107).
Adhesin Type Regulation Reference(s) The EcpR protein was shown to bind to a
YcbQ Laminin-binding 88 TTCCT sequence upstream of the ecp operon,
mbriae and deletion of ecpR resulted in decreased
Lpf1 Long-polar H-NS, Ler, late 89, 90, adherence by EHEC. H-NS silences ecp, and
mbriae exponential 101, antisilencing occurs by EcpR, assisted by the
phase, 103105,
low pH 146
protein IHF. Thus the ECP is most likely
F9 Fimbriae 91 involved in colonization by pathogens and
Curli Fimbriae Fis, Hha, 93, 94, nonpathogens alike. Similarly, a trimeric au-
RcsB, zinc, 147149 totransporter protein, EhaG, of EHEC has
heat shock been described (98). As for EcpR, these con-
OmpA Outer membrane Hfq, nitrogen 95,
protein A 150152
served proteins are found in both pathogens
Efa1 Toxin 96, 153 and nonpathogens, and in EHEC EhaG is in-
Iha IgrA homolog Temperature, 97, 154, volved in autoaggregation, biofilm formation,
adhesin iron 155 and adherence to extracellular matrix pro-
ECP E. coli common EcpR, H-NS, 106, 107 teins and colorectal epithelial cells. Not sur-
pilus IHF
EhaG Trimeric H-NS 98
prisingly, expression is, in part, controlled
autotransporter by H-NS (Table 1). Finally, the StcE zinc
protein metalloprotease encoded on the pO157 viru-
StcE Zinc H-NS, Ler 99, 100, lence plasmid enhances pedestal formation
metalloprotease 108 and is predicted to facilitate adherence to

epithelial cells in the host by cleavage of gly- NleA is secreted into the EHEC extracellular
coproteins (99, 108). The stcE gene is coor- milieu (115) and is coordinately regulated with
dinately regulated with the LEE through the LEE. As for the LEE, the regulatory pro-
Ler and H-NS. Clearly there is much work to teins H-NS, Ler, and GrlA control expression
be done to understand the role of non-T3SS of the phage-encoded nleA gene (115117). Ler
adhesins in niche recognition, pathogenesis acts directly, binding to the regulatory region
in general, and how their expression is coor- upstream of the nleA promoter. As with LEE1,
dinated with other virulence factors. nleA is regulated by quorum-sensing signaling,
whereby QseE controls expression through
inhibition of the positive-acting RcsB response
INJECTION OF EFFECTORS regulator (118). Expression of NleA is induced
by osmolarity, response to starvation signals,
Effector molecules slated for injection by the and RecA-dependent DNA-damage signaling,
T3SS into host cells are encoded within the the latter also controlling LEE expression (117,
LEE, O-islands, phage, and other integrative 119) (Fig. 4).
elements located in the genome. Those locat- EspFu (also named TccP) is another non-
ed within the LEE, including Tir, EspF, Map, LEE effector important for the EHEC A/E
EspG, EspH, and EspZ, are of course coordi- histopathology. Unlike EPEC where all genes
nately regulated with the expression of the required for A/E lesion formation are found
secretion apparatus itself. Evidence demon- within the LEE, EspFu is necessary for this
strates that a number of non-LEE-encoded phenotype in EHEC and is encoded within
effectors, including EspJ, NleB, NleE, NleF, phage U (120122). EspFu binds to the GTP-
and NleH, are involved in EHEC colonization binding domain of N-WASP, mimicking the
and survival, though experiments demon- eukaryotic SH2/SH3 adapter protein and
strating these phenotypes were conducted in facilitates actin polymerization, and is subject
a number of different animal infection models to regulation by environmental signals (123).
(109111). Thus further experimentation is The genes espFu and espJ are influenced by
necessary to determine the role of these non- changes in temperature, pH, osmolarity, and
LEE effectors in EHEC disease and carriage, oxygen pressure (124). The regulators KdpE
particularly in colonization of the RAJ in and Cra act to increase expression of espFu
cattle. (125). The QseEF two-component, quorum-
Assuming that the non-LEE effectors are sensing system controls expression of EspFu
important for disease in humans and carriage through AI-3, epinephrine, and norepineph-
in cattle, they should be coordinately ex- rine signaling (85) (Fig. 4). Thus the non-LEE
pressed with LEE-encoded effectors. Investi- effectors NleA and EspFu, important for vir-
gators have begun to unravel their regulation. ulence in EHEC, are coordinately controlled
In particular, several studies have focused on with the LEE by overlapping signaling: nleA
the NleA (also named EspI) effector because expression is Ler-dependent and, in part, in-
of evidence that it plays an important role in duced by the SOS response, while expression
pathogenesis. After translocation into the host of espFu is controlled by the QseA and QseEF
cell, NleA colocalizes with the Golgi appara- quorum-sensing pathways that directly stim-
tus, and has been associated with severe dis- ulate the LEE (Fig. 4).
ease in the Citrobacter rodentium A/E mouse Data are emerging to show that there are
model (112). NleA is conserved in many A/E at least two modes of regulation of the non-
pathogens (113, 114), and evidence indicates LEE effector molecules. The study by Garcia-
that it affects protein trafficking and secretion Angulo et al. (116), addressing whether a
through the rough endoplasmic reticulum by common mechanism for the non-LEE effector
binding to the COPII coat protein (112, 113). molecules exists, is an important investigation

because to date a common regulatory network that stx1 is also regulated by low levels of iron
linking their expression had not been estab- as a mechanism to acquire this micronutrient
lished. Though performing much of their ex- (127129) (Fig. 5). As for induction of other
perimentation on the related A/E pathogen lambdoid phage, Stx production is induced by
EPEC, investigators identified a 13-bp inverted DNA damage and regulated by RecA. Anti-
repeat, with a 5-bp spacer upstream of many biotics that target DNA synthesis, such as
of the genes encoding the non-LEE effector quinolones and mitomycin C, stimulate Stx
molecules. They found in EPEC, by deletion of production (130). Similarly, hydrogen perox-
these sequences and mutagenesis, that the ide induces stx expression, most likely due
repeats were essential for transcription of the to activation of the SOS response (131). In a
nleH1 and nleB2 genes. The authors termed seemingly contradictory study, the reactive
these sequences NRIR, for nle regulatory in- oxygen species NO inhibited Stx phage in-
verted repeat. Transferring this information to duction and Stx production in the presence
EHEC virulence gene regulation, the authors of the DNA-damaging agent mitomycin C
found a number of nle genes in the EDL933 (132). Here the authors claim that iNOS ex-
and Sakai EHEC strains that contain NRIR pressed in enterocytes, simulated by type I
sequences in positions predicted to affect cytokines producing NO, leads to a bacterial
transcription by in silico analysis. These loci NsrR-related stress response, sensitizing and
included, in both strains EDL933 and Sakai, protecting EHEC from DNA damage. None-
nleH1-nleF1, nleB2-nleC, espX, a number of theless, because the only known mechanism
nleG genes, and espFu. The authors concluded
that the NRIR sequences were involved in
coordinated expression of the non-LEE ef-
fector molecules in a second mode of regula-
tion, independent of the LEE regulators Ler
and GrlA. Future work will undoubtedly in-
volve identifying protein components, as well
as additional genetic elements necessary for
this regulatory network, and determining how
these effector molecules are coordinately
regulated with the type III system required
for their translocation into host cells.
TOXIN EXPRESSION AND

PHAGE INDUCTION
As with many of the non-LEE effector mole-

cules and most major bacterial exotoxins, the
FIGURE 5 Stimulation of Shiga toxin expression. Low
EHEC stx1 and stx2 genes are encoded within
iron levels lead to upregulation of stx1 (127, 128).
prophages. The Shiga toxins Stx1 and Stx2 Hydrogen peroxide and antibiotics targeting DNA
are responsible for bloody diarrhea associated synthesis, such as mytomycin C and quinolines, lead
with EHEC infection and the serious compli- to DNA damage and activate RecA to induce Stx1
cation hemolytic-uremic syndrome. Toxin production (130, 144, 145). Sulfate, phosphate,
and epinephrine molecules signal through the two-
production occurs upon phage induction. The
component system QseCF to RecA, thus leading ac-
genes stx1 and stx2 are encoded in separate tivation of Stx2 production. Ethanolamine also up-
prophages, and both are expressed during the regulates stx2 through the transcription factor EutR
phage lytic cycle (126). Initial studies indicated (38). doi:10.1128/microbiolspec.EHEC-0004-2013.f5

of Stx release is by lysing of the bacterium, (136). The expression of ehx is also positively
combined with the now obvious reasons that regulated by GrlA and Ler (64, 137), and thus
DNA-damaging antibiotics lead to increased the enterohemolysin is coordinately regulated
risk of hemolytic-uremic syndrome, there is a with the T3SS.
need for a more complete understanding of
how stx genes are regulated.
More recent data correlate Stx production A REGULATED INFECTION CONTINUUM
with signal molecules associated with specific
regions of the human and bovine intestine. The work of many research groups leads to a
The two-component sensor kinase QseC model of EHEC pathogenesis correlating with
regulates Stx2 expression through the SOS gut physiology, bacterial metabolism along
response, RecA cleavage of the lambda re- the route of infection, and host-microbe and
pressor cI (17) (Fig. 5). Consistently, recA ex- microbe-microbe signaling. For both the
pression was decreased in a qseC mutant. human and bovine hosts, EHEC must pass
QseC responds to the interkingdom signaling through the acidic environment of the stom-
molecules epinephrine and norepinephrine ach and the rumen, respectively. Three acid
and the related AI-3 quorum-sensing mole- tolerance systems assist the bacterium in this
cule produced by E. coli. Several studies in- process. GadX is induced by AHL molecules
dicated that Stx production is controlled by produced by resident flora of the bovine
AI-3. Another two-component pair, QseE, the rumen (14) and perceived by the LuxR ho-
sensor kinase, and QseF, the response regu- molog SdiA (Fig. 1). The LEE is downregu-
lator, affects Stx expression (85, 133). QseF can lated by SdiA in this environment because it
be phosphorylated by either QseE, responding is not the normal site of attachment, allow-
to epinephrine, sulfate, and phosphate, or ing the bacterium to travel toward the RAJ
QseC to stimulate stx2. Ethanolamine, re- in cattle (Fig. 3). Consistently, in a separate
leased from epithelium, also stimulates Stx2 study, data indicated that flagellar motility is
production in EHEC (38). The genetic path- upregulated by acid stress while expression
way necessary for the signaling involves the of Stx remains unchanged and the T3SS is
ethanolamine utilization regulator eutR, as downregulated by acute acid stress (138).
deletion of the gene encoding this protein In the large intestines of humans, condi-
resulted in decreased expression of stx2 by tions mimicking gluconeogenic versus glyco-
microarray (Fig. 5). Thus we begin to see a lytic metabolism are predicted to stimulate
clearer picture of Stx regulation in EHEC, expression of the EHEC T3SS (31, 139)
correlating expression to bacteria-derived (Fig. 4). This observation is also understand-
autoinducers, host-produced signal molecules, able in colonization of the RAJ of cattle,
and stress-related signals, and ultimately the where nutrients, particularly carbohydrates,
genetic pathways necessary for the bacterial are less abundant compared to the rumen.
response. One might posit that EHEC avoids the small
Because of the necessity of iron for colo- intestine in humans by downregulating ex-
nization and pathogenesis within the animal pression of LEE genes in the presence of
host, the EhxA enterohemolysin is also an high-glucose, glycolytic conditions (Fig. 3).
important EHEC virulence factor. EhxA is a It is curious why EPEC, which infects the
member of the RTX family, and its function is small intestine, also downregulates LEE genes
to lyse red blood cells. The ehxCABD operon, under glycolytic conditions, whereas, iden-
found on the pO157 virulence plasmid (134), is tical to EHEC, LEE genes are stimulated
positively regulated by DsrA and RpoS at host under low-glucose conditions. This observa-
body temperature (135). DsrA is a small RNA tion might be explained by the specificity
that acts by allowing translation to proceed of the infection process, that perception of

multiple signals is necessary for proper niche tachment in humans and cattle. These obser-
recognition. vations are consistent with the presumption
Carbohydrate metabolism and signaling are that substantive expression of Stx will occur
clearly linked to virulence, and several studies when colonization has commenced.
have indicated that fucose is important for
EHEC colonization (34, 140). Resident Bac-
terioides spp. cleave fucose from the glycans CONCLUSIONS
found in the intestine, and suppression of Ler,
T3SS function, and A/E lesions ensues from Researchers have made significant progress
the fucose signal perceived by FusKR (36) toward understanding the genetic regulation
(Fig. 3). EHEC can use a number of mucus- controlling EHEC pathogenesis. How EHEC
derived carbohydrates, in particular, mannose, gains safe passage through the acidic envi-
N-acetylglucosamine, and N-acetylglucosamine, ronment of the stomach, how flagellar motility
and galactose catabolism confers a competi- is turned on and turned off, and what sig-
tive advantage against commensal bacteria, nals and metabolites are necessary for out-
most likely assisting safe passage through the competing commensal flora prior to arriving
small intestine (39, 141). In both humans and at the sites of infection for humans and cattle
cattle, EHEC expresses the T3SS in response are now recognized on a basic level. Re-
to membrane-derived ethanolamine, clearly searchers have discovered myriad regulatory
encountered at the sites of infection in both networks and environmental signals that con-
hosts. The catecholamine and AI-3 signals also trol type III secretion, and we are beginning
contribute to elaboration of the secretion ap- to understand how effector molecules slated
paratus (Fig. 4). Why would epinephrine and for translocation into host cells are coordi-
norepinephrine exist here when the adrener- nately regulated with the T3SS itself and how
gic receptors are not located on the apical side the dangerous Shiga toxin is expressed.
of the intestinal lumen, but these molecules Though the LEE-encoded effectors and non-
were detected in the lumen of the intestine LEE NleA and EspFU are controlled by
of mice containing specific pathogen-free Ler and quorum-sensing signaling, the ma-
microbiota (142). It is also plausible that the jority of those proteins targeted for secretion
epinephrine/norepinephrine signal becomes are regulated by different mechanisms. The
more available upon damage to the host epi- NRIR sequences found upstream of many of
thelium, releasing these hormones into the the phage- and O-island-encoded effectors
lumen of the gut, enhancing expression of most likely coordinate their expression, but
the T3SS and firmly establishing infection this has yet to be established experimentally
(Fig. 4). Nonetheless, researchers have estab- for EHEC. Similarly, with the exception of
lished that three important signals, ethanol- Lpf1 and StcE, we do not know if, or how,
amine, epinephrine/norepinephrine, combined expression of most non-LEE adhesins is co-
with the bacteria-derived AI-3, correlated with ordinated with that of the LEE (Table 1). Fi-
the sites of infection. nally, EHEC tissue tropism is not well
A number of signals, including the cate- understood, but the regulatory networks de-
cholamines, also control the expression of scribed herein most likely play an important
stx genes, mediated through two-component, role in niche recognition in human and animal
quorum-sensing proteins QseC and QseE hosts.
(Fig. 5). Membrane-derived ethanolamine
and the well-studied RecA-dependent SOS
ACKNOWLEDGMENTS
response stimulate Stx production. EHEC is
expected to encounter all of these conditions We declare no conflicts of interest with regard
during niche recognition at the sites of at- to the manuscript.

CITATION 11. Gong S, Richard H, Foster JW. 2003. YjdE

(AdiC) is the arginine:agmatine antiporter essen-
Mellies JL, Lorenzen E. 2014. Enterohemor- tial for arginine-dependent acid resistance in
rhagic Escherichia coli virulence gene regulation. Escherichia coli. J Bacteriol 185:44024409.
12. Mitra A, Fay PA, Morgan JK, Vendura KW,
Microbiol Spectrum 2(4):EHEC-0004-2013.
Versaggi SL, Riordan JT. 2012. Sigma factor N,
liaison to an ntrC and rpoS dependent regulatory
pathway controlling acid resistance and the LEE
REFERENCES
in enterohemorrhagic Escherichia coli. PLoS ONE
1. Lukjancenko O, Wassenaar TM, Ussery DW. 7:e46288.
2010. Comparison of 61 sequenced Escherichia 13. Tramonti A, Visca P, De Canio M, Falconi M,
coli genomes. Microb Ecol 60:708720. De Biase D. 2002. Functional characterization
2. Zhaxybayeva O, Doolittle WF. 2011. Lateral gene and regulation of gadX, a gene encoding an AraC/
transfer. Curr Biol 21:R242R246. XylS-like transcriptional activator of the Esche-
3. Abu-Ali GS, Ouellette LM, Henderson ST, richia coli glutamic acid decarboxylase system.
Lacher DW, Riordan JT, Whittam TS, Manning J Bacteriol 184:26032613.
SD. 2010. Increased adherence and expression of 14. Hughes DT, Terekhova DA, Liou L, Hovde CJ,
virulence genes in a lineage of Escherichia coli Sahl JW, Patankar AV, Gonzalez JE, Edrington
O157:H7 commonly associated with human in- TS, Rasko DA, Sperandio V. 2010. Chemical
fections. PLoS ONE 5:e10167. sensing in mammalian host-bacterial commensal
4. Manning SD, Motiwala AS, Springman AC, associations. Proc Natl Acad Sci USA 107:9831
Qi W, Lacher DW, Ouellette LM, Mladonicky 9836.
JM, Somsel P, Rudrik JT, Dietrich SE, Zhang 15. Clarke MB, Hughes DT, Zhu C, Boedeker EC,
W, Swaminathan B, Alland D, Whittam TS. Sperandio V. 2006. The QseC sensor kinase: a
2008. Variation in virulence among clades of bacterial adrenergic receptor. Proc Natl Acad Sci
Escherichia coli O157:H7 associated with disease USA 103:1042010425.
outbreaks. Proc Natl Acad Sci USA 105:4868 16. Clarke MB, Sperandio V. 2005. Transcriptional
4873. regulation of flhDC by QseBC and sigma (FliA) in
5. Naylor SW, Low JC, Besser TE, Mahajan A, enterohaemorrhagic Escherichia coli. Mol
Gunn GJ, Pearce MC, McKendrick IJ, Smith Microbiol 57:17341749.
DG, Gally DL. 2003. Lymphoid follicle-dense 17. Hughes DT, Clarke MB, Yamamoto K, Rasko
mucosa at the terminal rectum is the principal DA, Sperandio V.. 2009. The QseC adrenergic
site of colonization of enterohemorrhagic Esche- signaling cascade in enterohemorrhagic E. coli
richia coli O157:H7 in the bovine host. Infect (EHEC). PLoS Pathog 5:e1000553.
Immun 71:15051512. 18. Sperandio V, Torres AG, Kaper JB. 2002. Quo-
6. Foster JW. 2004. Escherichia coli acid resistance: rum sensing Escherichia coli regulators B and C
tales of an amateur acidophile. Nat Rev Microbiol (QseBC): a novel two-component regulatory sys-
2:898907. tem involved in the regulation of flagella and
7. Price SB, Wright JC, DeGraves FJ, Castanie- motility by quorum sensing in E. coli. Mol
Cornet MP, Foster JW. 2004. Acid resistance Microbiol 43:809821.
systems required for survival of Escherichia coli 19. Bansal T, Englert D, Lee J, Hegde M, Wood TK,
O157:H7 in the bovine gastrointestinal tract and in Jayaraman A. 2007. Differential effects of epi-
apple cider are different. Appl Environ Microbiol nephrine, norepinephrine, and indole on Esche-
70:47924799. richia coli O157:H7 chemotaxis, colonization, and
8. Castanie-Cornet MP, Penfound TA, Smith D, gene expression. Infect Immun 75:45974607.
Elliott JF, Foster JW. 1999. Control of acid re- 20. Kendall MM, Rasko DA, Sperandio V. 2007.
sistance in Escherichia coli. J Bacteriol 181:3525 Global effects of the cell-to-cell signaling mole-
3535. cules autoinducer-2, autoinducer-3, and epi-
9. De Biase D, Tramonti A, Bossa F, Visca P. 1999. nephrine in a luxS mutant of enterohemorrhagic
The response to stationary-phase stress condi- Escherichia coli. Infect Immun 75:48754884.
tions in Escherichia coli: role and regulation of 21. Sperandio V, Torres AG, Jarvis B, Nataro JP,
the glutamic acid decarboxylase system. Mol Kaper JB. 2003. Bacteria-host communication:
Microbiol 32:11981211. the language of hormones. Proc Natl Acad Sci
10. Lin J, Smith MP, Chapin KC, Baik HS, Bennett USA 100:89518956.
GN, Foster JW. 1996. Mechanisms of acid resis- 22. Walters M, Sircili MP, Sperandio V. 2006. AI-3
tance in enterohemorrhagic Escherichia coli. Appl synthesis is not dependent on luxS in Escherichia
Environ Microbiol 62:30943100. coli. J Bacteriol 188:56685681.

23. Kutsukake K, Ohya Y, Iino T. 1990. Transcrip- regulatory cascade and a novel transcriptional
tional analysis of the flagellar regulon of Salmo- activator, the locus of enterocyte effacement
nella typhimurium. J Bacteriol 172:741747. (LEE)-encoded regulator (Ler). Mol Microbiol
24. Tobe T, Nakanishi N, Sugimoto N. 2011. Acti- 33:296306.
vation of motility by sensing short-chain fatty 34. Snider TA, Fabich AJ, Conway T, Clinkenbeard
acids via two steps in a flagellar gene regulatory KD. 2009. E. coli O157:H7 catabolism of intestinal
cascade in enterohemorrhagic Escherichia coli. mucin-derived carbohydrates and colonization.
Infect Immun 79:10161024. Vet Microbiol 136:150154.
25. Mahajan A, Currie CG, Mackie S, Tree J, 35. Jaswal VM, Babbar HS, Mahmood A. 1988.
McAteer S, McKendrick I, McNeilly TN, Roe Changes in sialic acid and fucose contents of
A, La Ragione RM, Woodward MJ, Gally DL, enterocytes across the crypt-villus axis in devel-
Smith DG. 2009. An investigation of the expres- oping rat intestine. Biochem Med Metab Biol
sion and adhesin function of H7 flagella in the 39:105110.
interaction of Escherichia coli O157:H7 with bo- 36. Pacheco AR, Curtis MM, Ritchie JM, Munera
vine intestinal epithelium. Cell Microbiol 11:121 D, Waldor MK, Moreira CG, Sperandio V. 2012.
137. Fucose sensing regulates bacterial intestinal col-
26. Sharma VK, Bearson SM, Bearson BL. 2010. onization. Nature 492:113117.
Evaluation of the effects of sdiA, a luxR homo- 37. Nakanishi N, Tashiro K, Kuhara S, Hayashi T,
logue, on adherence and motility of Escherichia Sugimoto N, Tobe T. 2009. Regulation of viru-
coli O157:H7. Microbiology 156:13031312. lence by butyrate sensing in enterohaemorrhagic
27. Kim JC, Yoon JW, Kim CH, Park MS, Cho SH. Escherichia coli. Microbiology 155:521530.
2012. Repression of flagella motility in entero- 38. Kendall MM, Gruber CC, Parker CT, Sperandio
hemorrhagic Escherichia coli O157:H7 by mucin V. 2012. Ethanolamine controls expression of
components. Biochem Biophys Res Commun genes encoding components involved in inter-
423:789792. kingdom signaling and virulence in enterohemor-
28. Iyoda S, Koizumi N, Satou H, Lu Y, Saitoh T, rhagic Escherichia coli O157:H7. MBio 3:e00050-
Ohnishi M, Watanabe H. 2006. The GrlR-GrlA 12.
regulatory system coordinately controls the ex- 39. Bertin Y, Girardeau JP, Chaucheyras-Durand
pression of flagellar and LEE-encoded type III F, Lyan B, Pujos-Guillot E, Harel J, Martin C.
protein secretion systems in enterohemorrhagic 2011. Enterohaemorrhagic Escherichia coli gains a
Escherichia coli. J Bacteriol 188:56825692. competitive advantage by using ethanolamine as a
29. Lehti TA, Bauchart P, Dobrindt U, Korhonen nitrogen source in the bovine intestinal content.
TK, Westerlund-Wikstrom B. 2012. The fimbri- Environ Microbiol 13:365377.
ae activator MatA switches off motility in Esche- 40. Low JC, McKendrick IJ, McKechnie C, Fenlon
richia coli by repression of the flagellar master D, Naylor SW, Currie C, Smith DG, Allison L,
operon flhDC. Microbiology 158:14441455. Gally DL. 2005. Rectal carriage of enterohemor-
30. Miranda RL, Conway T, Leatham MP, Chang rhagic Escherichia coli O157 in slaughtered cattle.
DE, Norris WE, Allen JH, Stevenson SJ, Laux Appl Environ Microbiol 71:9397.
DC, Cohen PS. 2004. Glycolytic and gluco- 41. Dziva F, van Diemen PM, Stevens MP, Smith
neogenic growth of Escherichia coli O157:H7 AJ, Wallis TS. 2004. Identification of Escherichia
(EDL933) and E. coli K-12 (MG1655) in the mouse coli O157:H7 genes influencing colonization of
intestine. Infect Immun 72:16661676. the bovine gastrointestinal tract using signature-
31. Njoroge JW, Nguyen Y, Curtis MM, Moreira tagged mutagenesis. Microbiology 150:36313645.
CG, Sperandio V. 2012. Virulence meets metab- 42. Naylor SW, Roe AJ, Nart P, Spears K, Smith
olism: Cra and KdpE gene regulation in entero- DG, Low JC, Gally DL. 2005. Escherichia coli
hemorrhagic Escherichia coli. MBio 3:e00280-12. O157:H7 forms attaching and effacing lesions at
32. Elliott SJ, Sperandio V, Giron JA, Shin S, the terminal rectum of cattle and colonization re-
Mellies JL, Wainwright L, Hutcheson SW, quires the LEE4 operon. Microbiology 151:2773
McDaniel TK, Kaper JB. 2000. The locus of 2781.
enterocyte effacement (LEE)-encoded regulator 43. Dean-Nystrom EA, Bosworth BT, Moon HW,
controls expression of both LEE- and non-LEE- OBrien AD. 1998. Escherichia coli O157:H7 re-
encoded virulence factors in enteropathogenic quires intimin for enteropathogenicity in calves.
and enterohemorrhagic Escherichia coli. Infect Infect Immun 66:45604563.
Immun 68:61156126. 44. Tobe T, Beatson SA, Taniguchi H, Abe H, Bailey
33. Mellies JL, Elliott SJ, Sperandio V, Donnenberg CM, Fivian A, Younis R, Matthews S, Marches
MS, Kaper JB. 1999. The Per regulon of entero- O, Frankel G, Hayashi T, Pallen MJ. 2006. An
pathogenic Escherichia coli: identification of a extensive repertoire of type III secretion effectors

in Escherichia coli O157 and the role of lambdoid mechanism in enterohemorrhagic Escherichia coli
phages in their dissemination. Proc Natl Acad Sci O157:H7. J Bacteriol 183:51875197.
USA 103:1494114946. 57. Friedberg D, Umanski T, Fang Y, Rosenshine I.
45. Mellies JL, ABarron AM, Carmona AM. 2007. 1999. Hierarchy in the expression of the locus of
Enteropathogenic and enterohemorrhagic Esche- enterocyte effacement genes of enteropathogenic
richia coli virulence gene regulation. Infect Immun Escherichia coli. Mol Microbiol 34:941952.
75:41994210. 58. Goldberg MD, Johnson M, Hinton JC, Williams
46. Bustamante VH, Santana FJ, Calva E, Puente PH. 2001. Role of the nucleoid-associated protein
JL. 2001. Transcriptional regulation of type III Fis in the regulation of virulence properties of
secretion genes in enteropathogenic Escherichia enteropathogenic Escherichia coli. Mol Microbiol
coli: Ler antagonizes H-NS-dependent repression. 41:549559.
Mol Microbiol 39:664678. 59. Grant AJ, Farris M, Alefounder P, Williams
47. Haack KR, Robinson CL, Miller KJ, Fowlkes PH, Woodward MJ, OConnor CD. 2003. Co-
JW, Mellies JL. 2003. Interaction of Ler at the ordination of pathogenicity island expression by
LEE5 (tir) operon of enteropathogenic Esche- the BipA GTPase in enteropathogenic Escherichia
richia coli. Infect Immun 71:384392. coli (EPEC). Mol Microbiol 48:507521.
48. Umanski T, Rosenshine I, Friedberg D. 2002. 60. Iyoda S, Watanabe H. 2004. Positive effects of
Thermoregulated expression of virulence genes multiple pch genes on expression of the locus of
in enteropathogenic Escherichia coli. Microbiology enterocyte effacement genes and adherence of
148:27352744. enterohaemorrhagic Escherichia coli O157:H7 to
49. Hommais F, Krin E, Laurent-Winter C, HEp-2 cells. Microbiology 150:23572571.
Soutourina O, Malpertuy A, Le Caer JP, 61. Barba J, Bustamante VH, Flores-Valdez MA,
Danchin A, Bertin P. 2001. Large-scale moni- Deng W, Finlay BB, Puente JL. 2005. A positive
toring of pleiotropic regulation of gene expression regulatory loop controls expression of the locus of
by the prokaryotic nucleoid-associated protein, enterocyte effacement-encoded regulators Ler
H-NS. Mol Microbiol 40:2036. and GrlA. J Bacteriol 187:79187930.
50. Abe H, Tatsuno I, Tobe T, Okutani A, Sasakawa 62. Deng W, Puente JL, Gruenheid S, Li Y, Vallance
C. 2002. Bicarbonate ion stimulates the expres- BA, Vazquez A, Barba J, Ibarra JA, ODonnell
sion of locus of enterocyte effacement-encoded P, Metalnikov P, Ashman K, Lee S, Goode D,
genes in enterohemorrhagic Escherichia coli O157: Pawson T, Finlay BB. 2004. Dissecting virulence:
H7. Infect Immun 70:35003509. systematic and functional analyses of a pathoge-
51. Beltrametti F, Kresse AU, Guzman CA. 1999. nicity island. Proc Natl Acad Sci USA 101:3597
Transcriptional regulation of the esp genes of 3602.
enterohemorrhagic Escherichia coli. J Bacteriol 63. Shin S, Castanie-Cornet MP, Foster JW,
181:34093418. Crawford JA, Brinkley C, Kaper JB. 2001. An
52. Ide T, Michgehl S, Knappstein S, Heusipp G, activator of glutamate decarboxylase genes regu-
Schmidt MA. 2003. Differential modulation by lates the expression of enteropathogenic Esche-
Ca2+ of type III secretion of diffusely adhering richia coli virulence genes through control of the
enteropathogenic Escherichia coli. Infect Immun plasmid-encoded regulator, Per. Mol Microbiol
71:17251732. 41:11331150.
53. Kenny B, Abe A, Stein M, Finlay BB. 1997. En- 64. Iyoda S, Honda N, Saitoh T, Shimuta K,
teropathogenic Escherichia coli protein secretion Terajima J, Watanabe H, Ohnishi M. 2011. Co-
is induced in response to conditions similar to ordinate control of the locus of enterocyte ef-
those in the gastrointestinal tract. Infect Immun facement and enterohemolysin genes by multiple
65:26062612. common virulence regulators in enterohemor-
54. Kenny B, Finlay BB. 1995. Protein secretion by rhagic Escherichia coli. Infect Immun 79:4628
enteropathogenic Escherichia coli is essential for 4637.
transducing signals to epithelial cells. Proc Natl 65. Russell RM, Sharp FC, Rasko DA, Sperandio V.
Acad Sci USA 92:79917995. 2007. QseA and GrlR/GrlA regulation of the locus
55. Sperandio V, Mellies JL, Nguyen W, Shin S, of enterocyte effacement genes in enterohemor-
Kaper JB. 1999. Quorum sensing controls ex- rhagic Escherichia coli. J Bacteriol 189:53875392.
pression of the type III secretion gene transcrip- 66. Boutte CC, Crosson S. 2013. Bacterial lifestyle
tion and protein secretion in enterohemorrhagic shapes stringent response activation. Trends
and enteropathogenic Escherichia coli. Proc Natl Microbiol 21:174180.
AcadSci USA 96:1519615201. 67. Hernandez VJ, Bremer H. 1991. Escherichia coli
56. Sperandio V, Torres AG, Giron JA, Kaper JB. ppGpp synthetase II activity requires spoT. J Biol
2001. Quorum sensing is a global regulatory Chem 266:59915999.

68. Nakanishi N, Abe H, Ogura Y, Hayashi T, 78. Ahmer BM.. 2004. Cell-to-cell signalling in
Tashiro K, Kuhara S, Sugimoto N, Tobe T. Escherichia coli and Salmonella enterica. Mol
2006. ppGpp with DksA controls gene expres- Microbiol 52:933945.
sion in the locus of enterocyte effacement (LEE) 79. Michael B, Smith JN, Swift S, Heffron F, Ahmer
pathogenicity island of enterohaemorrhagic BM. 2001. SdiA of Salmonella enterica is a LuxR
Escherichia coli through activation of two viru- homolog that detects mixed microbial communi-
lence regulatory genes. Mol Microbiol 61:194 ties. J Bacteriol 183:57335742.
205. 80. Yao Y, Martinez-Yamout MA, Dickerson TJ,
69. Bhatt S, Edwards AN, Nguyen HT, Merlin D, Brogan AP, Wright PE, Dyson HJ. 2006.
Romeo T, Kalman D. 2009. The RNA binding Structure of the Escherichia coli quorum sensing
protein CsrA is a pleiotropic regulator of the locus protein SdiA: activation of the folding switch by
of enterocyte effacement pathogenicity island of acyl homoserine lactones. J Mol Biol 355:262273.
enteropathogenic Escherichia coli. Infect Immun 81. Sharp FC, Sperandio V. 2007. QseA directly ac-
77:35523568. tivates transcription of LEE1 in enterohemor-
70. Lodato PB, Kaper JB. 2009. Post-transcrip- rhagic Escherichia coli. Infect Immun 75:2432
tional processing of the LEE4 operon in entero- 2440.
haemorrhagic Escherichia coli. Mol Microbiol 82. Sperandio V, Li CC, Kaper JB. 2002. Quorum-
71:273290. sensing Escherichia coli regulator A: a regulator of
71. Kendall MM, Gruber CC, Rasko DA, Hughes the LysR family involved in the regulation of the
DT, Sperandio V. 2011. Hfq virulence regulation locus of enterocyte effacement pathogenicity is-
in enterohemorrhagic Escherichia coli O157:H7 land in enterohemorrhagic E. coli. Infect Immun
strain 86-24. J Bacteriol 193:68436851. 70:30853093.
72. Hansen AM, Kaper JB. 2009. Hfq affects the 83. Reading NC, Rasko D, Torres AG, Sperandio V.
expression of the LEE pathogenicity island in 2010. A transcriptome study of the QseEF two-
enterohaemorrhagic Escherichia coli. Mol component system and the QseG membrane
Microbiol 73:446465. protein in enterohaemorrhagic Escherichia coli
73. Shakhnovich EA, Davis BM, Waldor MK. 2009. O157:H7. Microbiology 156:11671175.
Hfq negatively regulates type III secretion in 84. Rasko DA, Moreira CG, Li de R, Reading NC,
EHEC and several other pathogens. Mol Ritchie JM, Waldor MK, Williams N, Taussig R,
Microbiol 74:347363. Wei S, Roth M, Hughes DT, Huntley JF, Fina
74. Zhang L, Chaudhuri RR, Constantinidou C, MW, Falck JR, Sperandio V. 2008. Targeting
Hobman JL, Patel MD, Jones AC, Sarti D, Roe QseC signaling and virulence for antibiotic de-
AJ, Vlisidou I, Shaw RK, Falciani F, Stevens velopment. Science 321:10781080.
MP, Gally DL, Knutton S, Frankel G, Penn CW, 85. Reading NC, Torres AG, Kendall MM, Hughes
Pallen MJ. 2004. Regulators encoded in the DT, Yamamoto K, Sperandio V. 2007. A novel
Escherichia coli type III secretion system 2 gene two-component signaling system that activates
cluster influence expression of genes within the transcription of an enterohemorrhagic Esche-
locus for enterocyte effacement in enterohemor- richia coli effector involved in remodeling of host
rhagic E. coli O157:H7. Infect Immun 72:7282 actin. J Bacteriol 189:24682476.
7293. 86. Hayashi T, Makino K, Ohnishi M, Kurokawa K,
75. Flockhart AF, Tree JJ, Xu X, Karpiyevich M, Ishii K, Yokoyama K, Han CG, Ohtsubo E,
McAteer SP, Rosenblum R, Shaw DJ, Low CJ, Nakayama K, Murata T, Tanaka M, Tobe T,
Best A, Gannon V, Laing C, Murphy KC, Leong Iida T, Takami H, Honda T, Sasakawa C,
JM, Schneiders T, La Ragione R, Gally DL. 2012. Ogasawara N, Yasunaga T, Kuhara S, Shiba T,
Identification of a novel prophage regulator in Hattori M, Shinagawa H. 2001. Complete ge-
Escherichia coli controlling the expression of type nome sequence of enterohemorrhagic Escherichia
III secretion. Mol Microbiol 83:208223. coli O157:H7 and genomic comparison with a
76. Gomez-Duarte OG, Kaper JB. 1995. A plasmid- laboratory strain K-12. DNA Res 8:1122.
encoded regulatory region activates chromosomal 87. Perna NT, Plunkett G 3rd, Burland V, Mau B,
eaeA expression in enteropathogenic Escherichia Glasner JD, Rose DJ, Mayhew GF, Evans PS,
coli. Infect Immun 63:17671776. Gregor J, Kirkpatrick HA, Posfai G, Hackett J,
77. Porter ME, Mitchell P, Free A, Smith DG, Klink S, Boutin A, Shao Y, Miller L, Grotbeck EJ,
Gally DL. 2005. The LEE1 promoters from Davis NW, Lim A, Dimalanta ET, Potamousis
both enteropathogenic and enterohemorrhagic KD, Apodaca J, Anantharaman TS, Lin J, Yen G,
Escherichia coli can be activated by PerC-like Schwartz DC, Welch RA, Blattner FR. 2001.
proteins from either organism. J Bacteriol 187: Genome sequence of enterohaemorrhagic Esche-
458472. richia coli O157:H7. Nature 409:529533.

88. Samadder P, Xicohtencatl-Cortes J, Saldana Z, 98. Totsika M, Wells TJ, Beloin C, Valle J, Allsopp
Jordan D, Tarr PI, Kaper JB, Giron JA. 2009. LP, King NP, Ghigo JM, Schembri MA.
The Escherichia coli ycbQRST operon encodes 2012. Molecular characterization of the EhaG and
fimbriae with laminin-binding and epithelial cell UpaG trimeric autotransporter proteins from
adherence properties in Shiga-toxigenic E. coli pathogenic Escherichia coli. Appl Environ
O157:H7. Environ Microbiol 11:18151826. Microbiol 78:21792189.
89. Doughty S, Sloan J, Bennett-Wood V, Robertson 99. Grys TE, Siegel MB, Lathem WW, Welch RA.
M, Robins-Browne RM, Hartland EL.. 2002. 2005. The StcE protease contributes to inti-
Identification of a novel fimbrial gene cluster mate adherence of enterohemorrhagic Esche-
related to long polar fimbriae in locus of entero- richia coli O157:H7 to host cells. Infect Immun
cyte effacement-negative strains of entero- 73:12951303.
hemorrhagic Escherichia coli. Infect Immun 70: 100. Lathem WW, Grys TE, Witowski SE, Torres
67616769. AG, Kaper JB, Tarr PI, Welch RA. 2002. StcE,
90. Tatsuno I, Mundy R, Frankel G, Chong Y, a metalloprotease secreted by Escherichia coli
Phillips AD, Torres AG, Kaper JB. 2006. The lpf O157:H7, specifically cleaves C1 esterase inhibitor.
gene cluster for long polar fimbriae is not in- Mol Microbiol 45:277288.
volved in adherence of enteropathogenic Esche- 101. Farfan MJ, Cantero L, Vidal R, Botkin DJ,
richia coli or virulence of Citrobacter rodentium. Torres AG. 2011. Long polar fimbriae of entero-
Infect Immun 74:265272. hemorrhagic Escherichia coli O157:H7 bind to
91. Low AS, Holden N, Rosser T, Roe AJ, extracellular matrix proteins. Infect Immun 79:
Constantinidou C, Hobman JL, Smith DG, Low 37443750.
JC, Gally DL. 2006. Analysis of fimbrial gene 102. Torres AG, Milflores-Flores L, Garcia-Gallegos
clusters and their expression in enterohaemor- JG, Patel SD, Best A, La Ragione RM,
rhagic Escherichia coli O157:H7. Environ Martinez-Laguna Y, Woodward MJ. 2007.
Microbiol 8:10331047. Environmental regulation and colonization attri-
92. Xicohtencatl-Cortes J, Monteiro-Neto V, butes of the long polar fimbriae (LPF) of Esche-
Saldana Z, Ledesma MA, Puente JL, Giron JA. richia coli O157:H7. Int J Med Microbiol 297:177
2009. The type 4 pili of enterohemorrhagic 185.
Escherichia coli O157:H7 are multipurpose struc- 103. Jordan DM, Cornick N, Torres AG, Dean-
tures with pathogenic attributes. J Bacteriol 191 Nystrom EA, Kaper JB, Moon HW. 2004.
:411421. Long polar fimbriae contribute to colonization by
93. Kim SH, Kim YH. 2004. Escherichia coli O157:H7 Escherichia coli O157:H7 in vivo. Infect Immun
adherence to HEp-2 cells is implicated with curli 72:61686171.
expression and outer membrane integrity. J Vet 104. Lloyd SJ, Ritchie JM, Rojas-Lopez M,
Sci 5:119124. Blumentritt CA, Popov VL, Greenwich JL,
94. Saldana Z, Xicohtencatl-Cortes J, Avelino F, MWaldor MK, Torres AG. 2012. A double, long
Phillips AD, Kaper JB, Puente JL, Giron JA. polar fimbria mutant of Escherichia coli O157:H7
2009. Synergistic role of curli and cellulose in cell expresses Curli and exhibits reduced in vivo col-
adherence and biofilm formation of attaching and onization. Infect Immun 80:914920.
effacing Escherichia coli and identification of Fis 105. Torres AG, Lopez-Sanchez GN, Milflores-
as a negative regulator of curli. Environ Microbiol Flores L, Patel SD, Rojas-Lopez M, Martinez de
11:9921006. la Pena CF, Arenas-Hernandez MM, Martinez-
95. Torres AG, Li Y, Tutt CB, Xin L, Eaves-Pyles T, Laguna Y. 2007. Ler and H-NS, regulators con-
Soong L. 2006. Outer membrane protein A of trolling expression of the long polar fimbriae of
Escherichia coli O157:H7 stimulates dendritic cell Escherichia coli O157:H7. J Bacteriol 189:5916
activation. Infect Immun 74:26762685. 5928.
96. Nicholls L, Grant TH, Robins-Browne RM. 106. Rendon MA, Saldana Z, Erdem AL, Monteiro-
2000. Identification of a novel genetic locus that Neto V, Vazquez A, Kaper JB, Puente JL, Giron
is required for in vitro adhesion of a clinical iso- JA. 2007. Commensal and pathogenic Escherichia
late of enterohaemorrhagic Escherichia coli to coli use a common pilus adherence factor for
epithelial cells. Mol Microbiol 35:275288. epithelial cell colonization. Proc Natl Acad Sci
97. Tarr PI, Bilge SS, Vary JCJ, Jelacic S, Habeeb USA 104:1063710642.
RL, Ward TR, Baylor MR, Besser TE. 2000. Iha: 107. Martinez-Santos VI, Medrano-Lopez A,
a novel Escherichia coli O157:H7 adherence-con- Saldana Z, Giron JA, Puente JL. 2012. Tran-
ferring molecule encoded on a recently acquired scriptional regulation of the ecp operon by EcpR,
chromosomal island of conserved structure. Infect IHF, and H-NS in attaching and effacing Esche-
Immun 68:14001407. richia coli. J Bacteriol 194:50205033.

108. Lathem WW, Bergsbaken T, Welch RA. 2004. 117. Schwidder M, Hensel M, Schmidt H. 2011.
Potentiation of C1 esterase inhibitor by StcE, a Regulation of nleA in Shiga toxin-producing
metalloprotease secreted by Escherichia coli Escherichia coli O84:H4 strain 4795/97. J
O157:H7. J Exp Med 199:10771087. Bacteriol 193:832841.
109. Echtenkamp F, Deng W, Wickham ME, 118. Njoroge J, Sperandio V. 2012. Enterohemor-
Vazquez A, Puente JL, Thanabalasuriar A, rhagic Escherichia coli virulence regulation by two
Gruenheid S, Finlay BB, Hardwidge PR. 2008. bacterial adrenergic kinases, QseC and QseE. In-
Characterization of the NleF effector protein fect Immun 80:688703.
from attaching and effacing bacterial pathogens. 119. Mellies JL, Haack KR, Galligan DC. 2007. SOS
FEMS Microbiol Lett 281:98107. regulation of the type III secretion system of
110. Hemrajani C, Marches O, Wiles S, Girard F, enteropathogenic Escherichia coli. J Bacteriol
Dennis A, Dziva F, Best A, Phillips AD, Berger 189:28632872.
CN, Mousnier A, Crepin VF, Kruidenier L, 120. Campellone KG, Robbins D, Leong JM. 2004.
Woodward MJ, Stevens MP, La Ragione RM, EspFU is a translocated EHEC effector that
MacDonald TT, Frankel G. 2008. Role of NleH, a interacts with Tir and N-WASP and promotes
type III secreted effector from attaching and ef- Nck-independent actin assembly. Dev Cell 7:217
facing pathogens, in colonization of the bovine, 228.
ovine, and murine gut. Infect Immun 76:4804 121. Garmendia J, Phillips AD, Carlier MF, Chong Y,
4813. Schuller S, Marches O, Dahan S, Oswald E,
111. Kelly M, Hart E, Mundy R, Marches O, Wiles Shaw RK, Knutton S, Frankel G. 2004. TccP is
S, Badea L, Luck S, Tauschek M, Frankel G, an enterohaemorrhagic Escherichia coli O157:H7
Robins-Browne RM, Hartland EL. 2006. Es- type III effector protein that couples Tir to the
sential role of the type III secretion system ef- actin-cytoskeleton. Cell Microbiol 6:11671183.
fector NleB in colonization of mice by Citrobacter 122. Ritchie JM, Brady MJ, Riley KN, Ho TD,
rodentium. Infect Immun 74:23282337. Campellone KG, Herman IM, Donohue-Rolfe
112. Gruenheid S, Sekirov I, Thomas NA, Deng W, A, Tzipori S, Waldor MK, Leong JM. 2008.
ODonnell P, Goode D, Li Y, Frey EA, Brown EspFU, a type III-translocated effector of actin
NF, Metalnikov P, Pawson T, Ashman K, assembly, fosters epithelial association and late-
Finlay BB. 2004. Identification and character- stage intestinal colonization by E. coli O157:H7.
ization of NleA, a non-LEE-encoded type III Cell Microbiol 10:836847.
translocated virulence factor of enterohaemor- 123. Sallee NA, Rivera GM, Dueber JE, Vasilescu D,
rhagic Escherichia coli O157:H7. Mol Microbiol Mullins RD, Mayer BJ, Lim WA. 2008. The
51:12331249. pathogen protein EspF(U) hijacks actin poly-
113. Creuzburg K, Schmidt H. 2007. Molecular merization using mimicry and multivalency. Na-
characterization and distribution of genes ture 454:10051008.
encoding members of the type III effector nleA 124. Garmendia J, Frankel G, Crepin VF. 2005. En-
family among pathogenic Escherichia coli strains. teropathogenic and enterohemorrhagic Esche-
J Clin Microbiol 45:24982507. richia coli infections: translocation, translocation,
114. Mundy R, Jenkins C, Yu J, Smith H, Frankel translocation. Infect Immun 73:25732585.
G. 2004. Distribution of espI among clini- 125. Njoroge JW, Gruber C, Sperandio V. 2013. The
cal enterohaemorrhagic and enteropathogenic interacting Cra and KdpE regulators are involved
Escherichia coli isolates. J Med Microbiol 53: in the expression of multiple virulence factors in
11451149. enterohemorrhagic Escherichia coli. J Bacteriol
115. Roe AJ, Tysall L, Dransfield T, Wang D, Fraser- 195:24992508.
Pitt D, Mahajan A, Constandinou C, Inglis N, 126. Tyler JS, Mills MJ, Friedman DI. 2004. The
Downing A, Talbot R, Smith DG, Gally DL. operator and early promoter region of the Shiga
2007. Analysis of the expression, regulation and toxin type 2-encoding bacteriophage 933W and
export of NleA-E in Escherichia coli O157:H7. control of toxin expression. J Bacteriol 186:7670
Microbiology 153:13501360. 7679.
116. Garcia-Angulo VA, Martinez-Santos VI, 127. OBrien AD, LaVeck GD, Thompson MR, For-
Villasenor T, Santana FJ, Huerta-Saquero A, mal SB. 1982. Production of Shigella dysenteriae
Martinez LC, Jimenez R, Lara-Ochoa C, Tellez- type 1-like cytotoxin by Escherichia coli. J Infect
Sosa J, Bustamante VH, Puente JL. 2012. A Dis 146:763769.
distinct regulatory sequence is essential for the 128. Wagner PL, Livny J, Neely MN, Acheson DW,
expression of a subset of nle genes in attaching Friedman DI, Waldor MK.. 2002. Bacteriophage
and effacing Escherichia coli. J Bacteriol 194: control of Shiga toxin 1 production and release by
55895603. Escherichia coli. Mol Microbiol 44:957970.

129. Waldor MK, Friedman DI. 2005. Phage regula- Hightower GA, Smith JT, Autieri SM, Leatham
tory circuits and virulence gene expression. Curr MP, Lins JJ, Allen RL, Laux DC, Cohen PS,
Opin Microbiol 8:459465. Conway T. 2008. Comparison of carbon nutrition
130. Zhang X, McDaniel AD, Wolf LE, Keusch GT, for pathogenic and commensal Escherichia coli
Waldor MK, Acheson DW. 2000. Quinolone strains in the mouse intestine. Infect Immun
antibiotics induce Shiga toxin-encoding bacterio- 76:11431152.
phages, toxin production, and death in mice. 141. Kamada N, Kim Y-G, Sham HP, Vallance BA,
J Infect Dis 181:664670. Puente JL, Martens EC, Nez G. 2012. Regu-
131. Los JM, Los M, Wegrzyn A, Wegrzyn G. 2010. lated virulence controls the ability of a pathogen
Hydrogen peroxide-mediated induction of the to compete with the gut microbiota. Science
Shiga toxin-converting lambdoid prophage ST2- 336:13251329.
8624 in Escherichia coli O157:H7. FEMS Immunol 142. Asano Y, Hiramoto T, Nishino R, Aiba Y,
Med Microbiol 58:322329. Kimura T, Yoshihara K, Koga Y, Sudo N. 2012.
132. Vareille M, de Sablet T, Hindre T, Martin C, Critical role of gut microbiota in the production of
Gobert AP. 2007. Nitric oxide inhibits Shiga- biologically active, free catecholamines in the gut
toxin synthesis by enterohemorrhagic Escherichia lumen of mice. Am J Physiol Gastrointest Liver
coli. Proc Natl Acad Sci USA 104:1019910204. Physiol 303:G1288G1295.
133. Reading NC, Rasko DA, Torres AG, Sperandio 143. Hersh BM, Farooq FT, Barstad DN,
V. 2009. The two-component system QseEF and Blankenhorn DL, Slonczewski JL. 1996. A glu-
the membrane protein QseG link adrenergic and tamate-dependent acid resistance gene in Esche-
stress sensing to bacterial pathogenesis. Proc Natl richia coli. J Bacteriol 178:39783981.
Acad Sci USA 106:58895894. 144. McGannon CM, Fuller CA, Weiss AA. 2010.
134. Ogura Y, Ooka T, Iguchi A, Toh H, Asadulghani Different classes of antibiotics differentially in-
M, Oshima K, Kodama T, Abe H, Nakayama K, fluence Shiga toxin production. Antimicrob
Kurokawa K, Tobe T, Hattori M, Hayashi T. Agents Chemother 54:37903798.
2009. Comparative genomics reveal the mecha- 145. Wagner PL, Neely MN, Zhang X, Acheson DW,
nism of the parallel evolution of O157 and non- Waldor MK, Friedman DI. 2001. Role for a
O157 enterohemorrhagic Escherichia coli. Proc phage promoter in Shiga toxin 2 expression from
Natl AcadSci USA 106:1793917944. a pathogenic Escherichia coli strain. J Bacteriol
135. Li H, Granat A, Stewart V, Gillespie JR. 2008. 183:20812085.
RpoS, H-NS, and DsrA influence EHEC hemoly- 146. Torres AG, Slater TM, Patel SD, Popov VL,
sin operon (ehxCABD) transcription in Esche- Arenas-Hernandez MM. 2008. Contribution of
richia coli O157:H7 strain EDL933. FEMS the Ler- and H-NS-regulated long polar fimbriae
Microbiol Lett 285:257262. of Escherichia coli O157:H7 during binding to tis-
136. Gottesman S. 2004. The small RNA regulators of sue-cultured cells. Infect Immun 76:50625071.
Escherichia coli: roles and mechanisms. Annu Rev 147. Carter MQ, Parker CT, Louie JW, Huynh S,
Microbiol 58:303328. Fagerquist CK, Mandrell RE. 2012. RcsB con-
137. Saitoh T, Iyoda S, Yamamoto S, Lu Y, Shimuta tributes to the distinct stress fitness among
K, Ohnishi M, Terajima J, Watanabe H. 2008. Escherichia coli O157:H7 curli variants of the 1993
Transcription of the ehx enterohemolysin gene is hamburger-associated outbreak strains. Appl En-
positively regulated by GrlA, a global regulator viron Microbiol 78:77067719.
encoded within the locus of enterocyte efface- 148. Lim J, Lee KM, Kim SH, Kim Y, Kim SH, Park
ment in enterohemorrhagic Escherichia coli. W, Park S. 2011. YkgM and ZinT proteins are
J Bacteriol 190:48224830. required for maintaining intracellular zinc con-
138. House B, Kus JV, Prayitno N, Mair R, Que L, centration and producing curli in enterohemor-
Chingcuanco F, Gannon V, Cvitkovitch DG, rhagic Escherichia coli (EHEC) O157:H7 under
Barnett Foster D. 2009. Acid-stress-induced zinc deficient conditions. Int J Food Microbiol
changes in enterohaemorrhagic Escherichia coli 149:159170.
O157:H7 virulence. Microbiology 155:29072918. 149. Sharma VK, Bearson BL. 2013. Hha controls
139. Calderon VE, Chang Q, McDermott M, Lytle Escherichia coli O157:H7 biofilm formation by
MB, McKee G, Rodriguez K, Rasko DA, differential regulation of global transcriptional
Sperandio V, Torres AG. 2010. Outbreak caused regulators FlhDC and CsgD. Appl Environ
by cad-negative Shiga toxin-producing Esche- Microbiol 79:23842396.
richia coli O111, Oklahoma. Foodborne Pathog Dis 150. Baev MV, Baev D, Radek AJ, Campbell JW.
7:107109. 2006. Growth of Escherichia coli MG1655 on LB
140. Fabich AJ, Jones SA, Chowdhury FZ, Cernosek medium: monitoring utilization of amino acids,
A, Anderson A, Smalley D, McHargue JW, peptides, and nucleotides with transcriptional

microarrays. Appl Microbiol Biotechnol 71:317 a truncated version of the efa-1 gene in Esche-
322. richia coli O157:H7 influences the expression and
151. Douchin V, Bohn C, Bouloc P. 2006. Down- secretion of locus of enterocyte effacement-en-
regulation of porins by a small RNA bypasses the coded proteins but not intestinal colonization in
essentiality of the regulated intramembrane pro- calves or sheep. Infect Immun 72:54025411.
teolysis protease RseP in Escherichia coli. J Biol 154. Rashid RA, Tabata TA, Oatley MJ, Besser TE,
Chem 281:1225312259. Tarr PI, Moseley SL. 2006. Expression of puta-
152. Udekwu KI, Darfeuille F, Vogel J, Reimegard J, tive virulence factors of Escherichia coli O157:H7
Holmqvist E, Wagner EG. 2005. Hfq-dependent differs in bovine and human infections. Infect
regulation of OmpA synthesis is mediated by an Immun 74:41424148.
antisense RNA. Genes Dev 19:23552366. 155. Rashid RA, Tarr PI, Moseley SL. 2006. Ex-
153. Stevens MP, Roe AJ, Vlisidou I, van Diemen pression of the Escherichia coli IrgA homolog
PM, La Ragione RM, Best A, Woodward MJ, adhesin is regulated by the ferric uptake regula-
Gally DL, Wallis TS. 2004. Mutation of toxB and tion protein. Microb Pathog 41:207217.

INCIDENCE, EPIDEMIOLOGY,
AND ECOLOGY

Shiga Toxin (Verotoxin)-Producing
Escherichia coli in Japan
JUN TERAJIMA,1 SUNAO IYODA,1 MAKOTO OHNISHI,1 and

HARUO WATANABE1
10
A series of outbreaks of infection with Shiga toxin (or verotoxin [VT])-
producing Escherichia coli or enterohemorrhagic E. coli (EHEC) O157:H7
occurred in Japan in 1996, the largest outbreak occurring in primary schools in
Sakai City, Osaka Prefecture, where more than 7,500 cases were reported (1).
Although the reason for the sudden increase in the number of reports of EHEC
isolates in 1996 is not known, the number of reports has grown to more than
3,000 cases per year since 1996 from an average of 105 cases reported each year
during the previous 5-year period (19911995) (2). Despite control measures
instituted since 1996, including designating EHEC infection as a notifiable
disease, and the disease being monitored effectively through nationwide sur-
veillance, the number of reports remains high, around 3,800 cases per year
(Fig. 1). Serogroup O157 predominates over other EHEC serogroups, but iso-
lation frequency of non-O157 EHEC has gone up slightly over the past few years.
Non-O157 EHEC has caused outbreaks where consumption of a raw beef dish
was the source of the infection and some fatal cases were occurred. Laboratory
surveillance consisting of prefectural and municipal public health institutes
(PHIs) and the National Institute of Infectious Diseases has contributed to
finding not only multiprefectural outbreaks but also recognizing sporadic cases
that could have been missed as an outbreak without the aid of molecular
1
Department of Bacteriology I, National Institute of Infectious Diseases, Tokyo, Japan.
199

200 TERAJIMA ET AL.
FIGURE 1 Reported cases of EHEC infection, 1996 to 2012. doi:10.1128/microbiolspec.EHEC-0011-20013.f1
subtyping of EHEC isolates. This short over- surveillance of the population surrounding
view presents recent information on the sur- the outbreak, including checking food-intake
veillance of EHEC infections in Japan. and stool specimens of infected cases and per-
sons engaged in food preparation, revealed
that approximately 35% of EHEC infections
SURVEILLANCE AND REPORTING were asymptomatic and that the proportion
of asymptomatic infection was high among
In 1966, EHEC infection was designated as a the middle-aged group whereas symptomatic
disease requiring mandatory reporting regard- cases were more frequent in young and old
less of source. EHEC infection is one of cat- age groups. The public health importance of
egory III notifiable infectious diseases along symptomatic shedding (3) and asymptomatic
with other bacterial infections caused by shedding (4) in transmission among preschool
Vibrio cholerae O1 or O139, Shigella species, children is well established, but asymptomatic
Salmonella enterica serovar Typhi, and Sal- shedding in adults is not yet well understood
monella serovar Paratyphi A in the National though high rates of culture-positive, asymp-
Epidemiological Surveillance of Infectious Dis- tomatic adults with EHEC O157:H7 infection
eases (NESID) under the Law Concerning the were reported in 1998 (5), and the tendency
Prevention of Infectious Diseases and Medical remains similar with EHEC infection cur-
Care for Patients of Infections (the Infectious rently (6). Apart from NESID, results of char-
Diseases Control Law) enacted in April 1999. acterization (serotypes, VT types, etc.) of EHEC
Under a surveillance system for food poison- isolates at prefectural and municipal PHIs are
ing based on the Food Sanitation Law, an reported to the Infectious Disease Surveillance
EHEC infection is reported as food poisoning Center (IDSC) at the National Institute of In-
by physicians or judged as such by the direc- fectious Diseases. The summary showed that
tor of the health center and reported as such EHEC O157 serogroup is the predominant one,
by the local government to the Ministry of followed by O26, O111, O103, O121, O145, and
Health, Labour and Welfare (MHLW). Active others. However, as seen in the United States

CHAPTER 10 STEC in Japan 201
(7), the percentage of non-EHEC O157 sero- subtyping network based on PFGE analysis,
groups among all EHEC isolates from human PulseNet Japan (12). It participates in the Asia
infection has been growing slightly; the rate of and Pacific regional network, PulseNet Asia
isolation frequency of EHEC O157 has declined Pacific, which constitutes a part of PulseNet
from approximately 70% of all EHEC isolates International (13), and contributed in inves-
in 2000 to less than 60% in 2012, though the tigations of domestic outbreaks (14) and an
situation of EHEC infection in Japan is not international EHEC O157 outbreak that oc-
exactly the same as in continental Europe curred between Japan and the United States
where non-O157 EHEC serotypes are more (15). Although PFGE is still in wide use in
common than infection with O157:H7 EHEC outbreak investigations for its high discrimi-
(8, 9). The number of EHEC isolates reported nation power, there are additional DNA-based
to IDSC is about half the number of EHEC methods for strain analysis that are simpler,
infection cases reported to NESID, because faster to perform, and equivalent to PFGE
only a small proportion of isolates in hospitals in strain discrimination. Among these, multi-
or commercial laboratories are sent to PHIs. locus sequence typing was shown to have in-
sufficient discriminatory power for EHEC
O157:H7 (16, 17). On the other hand, multiple-
LABORATORY SURVEILLANCE locus variable-number tandem-repeat analysis
(MLVA) has proven to be useful for genetic
Stool samples from suspected cases of diar- fingerprinting of EHEC O157:H7 (13, 1821),
rhea should be routinely screened for EHEC O26, and O111 (22) and pathogenic bacteria
by plating on sorbitol-MacConkey agar con- (23). For strain analysis, in addition to MLVA,
taining cefixime and tellurite in addition to an insertion element (IS)-printing system that
conventional E. coli isolation agar. Rhamnose/ is a PCR-based strain-typing method for
sorbose-MacConkey agar containing cefixime EHEC O157 (24) has been applied to outbreak
and tellurite and chromogenic media could investigations. The IS-printing system showed
also be used for O26/O111 isolates as these that it had equivalent capacity to subtype
serogroups are major ones following O157. the isolates in the outbreak. It is based on the
The use of an enrichment culture step or im- variability in genomic location of IS629 among
munomagnetic separation available for E. coli EHEC O157 strains and can produce the re-
O157, O26, O111, O103, and O145 facilitates sults in a much shorter time than other
isolation of these pathogens. It is essential that subtyping methods such as PFGE. The meth-
E. coli isolates be examined for production od has another advantage: if standard pro-
of VT or the presence of VT genes regardless tocol is established and quality assurance is
of serogroup of the isolates since these are achieved, then the results are suitable for
prerequisite results for diagnosis of EHEC creating a database since PCR results can
infection. It could be important to examine be expressed as presence or absence of the
production of VT or the presence of VT genes amplicon for the sample, which is easily dig-
in all isolates on selective agar plates with and itized, and the information can readily be
without cefixime and tellurite to detect vari- shared with other laboratories involved in
ous serogroups of EHEC (10). Molecular sub- the outbreak investigation. Characterization
typing techniques, including pulsed-field gel of EHEC isolates using DNA-based methods
electrophoresis (PFGE), have sufficient sen- could be critical for linking simultaneously
sitivity and discriminatory power for use in occurring sporadic cases that would other-
epidemiological investigations. The establish- wise have been unrecognized as a diffuse
ment of the molecular subtyping network, outbreak (14). It is important to note that a
PulseNet, in the United States (11) encour- prompt epidemiological investigation is nec-
aged us to construct an equivalent molecular essary to confirm that a cluster of genetically

202 TERAJIMA ET AL.
indistinguishable isolates may represent an rate for the age group 15 to 64 years in 2011
outbreak with a common source. Also, it is was due to a large outbreak of EHEC O111
noteworthy that epidemiological links are not in which the age of 21 HUS cases ranged from
established for all the clusters; many remain 15 to 64 years among 34 HUS cases in the
unresolved. outbreak. From 2006 to 2012, 67% of the 605
HUS cases were culture-confirmed by labo-
ratory testing, and the rest of the cases were
HEMOLYTIC-UREMIC SYNDROME (HUS) diagnosed by detecting antilipopolysaccharide
antibody of E. coli in the serum of the patients
The case definition of EHEC infection was or VT in the stool samples of the patients.
partly amended in April 2006 as follows: if VT EHEC O157 was the predominant serogroup,
is detected in feces, or O-antigen agglutinating occupying 85% of all isolates in culture-
antibody or anti-VT antibody is detected in confirmed HUS cases, followed by O111 (5%),
serum of a patient with HUS, the patient O121 (2.1%), O26 (1.7%), O165 (1%), O145 (1%),
should be reported as having EHEC infection. and the rest of the O serogroups, including
From 2006 to 2012, the average annual num- O55, O174, O183, and unknown serogroup
ber of HUS cases (including serodiagnosed samples. Although non-O157 serogroup strains
cases) was 100 and the incidence rate of HUS were isolated in the culture-confirmed HUS
(HUS cases/symptomatic cases) was 3.7% cases, 94% of all EHEC isolates in the culture-
(Fig. 2) (2529). Increased rates of HUS in confirmed HUS cases were either VT2 or
children less than 10 years old and the elderly both VT1 and VT2 producers, which is consist-
(30, 31) are shown in Fig. 2, and a rise in in- ent with epidemiological evidence that VT2-
cidence rate of HUS for the age group older containing EHEC O157:H7 strains are more
than 65 in 2012 reflects an outbreak of EHEC frequently associated with HUS than the
O157 where five HUS cases occurred in a fa- strains containing VT1 (33, 34). Some non-
cility for the elderly (32). A rise in incidence O157 EHEC strains that were also classified
FIGURE 2 Incidence rate of HUS in EHEC infection by age groups in Japan, 2006 to 2012. Incidence rate (%)
was calculated as (number of patients with HUS) (number of symptomatic cases) 100%.

as enteroaggregative E. coli have been isolated 2000 and 2012, there were 12 EHEC out-
in HUS cases due to O104:H4 infection in breaks that had more than 100 culture-
Germany (35, 36) as well as an outbreak due positive cases (Table 1). All 12 outbreaks ap-
to O111:H2 in France (37), but there has been pear to have resulted from consumption of
only one such case reported in Japan, where a contaminated foods, and in some of these
3-year-old boy in Kagoshima Prefecture de- outbreaks, microbiological testing confirmed
veloped HUS, developed encephalopathy, and the implicated foods. These included beef
died due to the infection of VT2-producing products (39); lightly salted cucumber (40);
and enteroaggregative E. coli O86:HNM (38). Koumi-ae consisting of boiled spinach and
steamed chicken meat seasoned with welsh
onion, ginger, and soy sauce (40); boxed meals
OUTBREAKS (43); lettuce (44); school lunch (45); Yukhoe
(raw beef) (6), and Japanese rice cakes (6).
In the 197 outbreaks with more than 10 cul- The infectious dose of EHEC O157 is very low,
ture-positive patients, which were reported probably fewer than 100 organisms (51, 52),
to IDSC from 2000 to 2012 (6, 32, 3950), the and this is a critical factor in the transmission
main mode of transmission of the infection of the EHEC when people consume raw or
was person to person (41%), food borne (29%), lightly cooked foods such as sushi and veg-
and water borne (3%,), and in about one-third etables. Because sushi and raw or lightly
of the outbreaks, the mode of transmission cooked meat are favorite food items in Japan,
of the infection remains unknown. The most outbreaks were associated with consumption
frequent serogroup of EHEC associated with of salmon roe sushi in 1998 (53) and rare
the outbreaks was EHEC O157 (47%), followed roast beef contaminated with EHEC O157 in
by O26 (35%), O111 (8%), O103 (4%), O121 (3%), 2001 (39). In 2011, a large EHEC O111 outbreak
O145 (2%), and others. The predominance of due to consumption of Yukhoe, a Korean dish
EHEC O157 serogroup and other representa- of raw beef and egg yolk, was identified at
tive serogroups that had caused the outbreaks Yakiniku chain restaurants. EHEC O111:H8
is fairly consistent with the laboratory sur- was isolated from 85 of 181 patients (median
veillance results obtained by PHIs for EHEC age 20 years); in 34 of those patients HUS
infection that includes outbreaks and sporadic developed; encephalopathy developed in 21
cases. patients; and 5 patients died (6). EHEC O111:
Among 197 outbreaks, 101 outbreaks have H8 was also isolated from the conserved part
occurred in nursery schools, which may ac- of the original meat preparations distributed
count for the high proportion of person-to- to the chain restaurants. In response to per-
person transmissions of the infection in the sistent food poisonings caused by raw beef,
outbreaks. Interestingly, the most prevalent MHLW revised the standards of beef product
serogroup of EHEC in the outbreaks in nurs- quality for raw-eating and put them into op-
ery schools was O26 (52%), followed by O157 eration in October 2011. Further, after the
(27%), O111 (9%), O103 (4%), O121 (4%), O145 EHEC O157 was detected in the inner part
(3%), and OUT, which may account for rela- of cattle livers, MHLW banned marketing of
tively mild clinical manifestations, including cattle liver for raw-eating. Probably as a con-
asymptomatic cases and, consequently, fre- sequence of these preventive measures, the
quent person-to-person transmission. More incidence of EHEC O157 cases related to con-
O26-associated cases may have occurred; how- sumption of raw meat decreased by almost
ever, such cases would not be detected as half in 1 year from 2011 to 2012 (32).
sporadic due to mild symptoms among adults Some of the outbreaks were associated
and children. Larger outbreaks resulted from with consumption of vegetables. Among four
consumption of contaminated foods. Between outbreaks associated with consumption of

204 TERAJIMA ET AL.
TABLE 1 Characterization of 12 EHEC outbreaks with more than 100 culture-positive cases between 2000
and 2012
Symptomatic Culture Likely mode
Year Prefecture/city Setting (reference) Serotype VT type cases positives of transmission
2001 Chiba P. Patients home (39) O157:H7 VT1 & 2 195 257 Beef productsa
2002 Fukuoka C. Nursery school (40) O157:H- VT2 74 112 Lightly salted
cucumbera
2002 Utsunomiya C. Hospital and home O157:H7 VT1 & 2 123 111 Koumi-aea
for the elderly (40)
2003 Yokohama C. Kindergarten (41) O26:H11 VT1 141 449 Food borne
2004 Ishikawa P. High school (42) O111:H- VT1 & 2 110 103 Food borne
2007 Tokyo M. School refectory (43) O157:H7 VT2 467 204 Food borne
2007 Miyagi P., Restaurant (43) O157:H7 VT1 & 2 314 173 Boxed mealsa
Sendai C.,
and Akita C.
2009 Saga P. Nursery school (44) O26:H11 VT1 ND 133 Lettucea
2010 Mie P. High school (45) O157:H7 VT2 138 164 School luncha
2011 Toyama P. Chain restaurants (6) O111:H8 VT2, VT 181 102 Yukhoe (raw beef)a
O157:H7 VT1, VT2, 38
VT1 & 2
2011 Yamagata P. Festival (6) O157:H7 VT1 & 2 287 189 Japanese rice cakesa
2012 Osaka C. Nursery school (32) O26:H- VT1 68 115 Food borne
a
Conrmed microbiologically; ND, no data.
vegetables in 2011 (6), EHEC O26:H11 was or undercooked foods of bovine origin has
isolated from cabbage in an outbreak and been the most common means of transmitting
EHEC O157:H7 was isolated from pickled egg- EHEC infection. Laboratory investigations
plant and green perilla, green perilla served have played a critical role in the investigation
with grated radish, and cucumber in three of EHEC infections since some cases were
outbreaks, respectively. In 2012, there was a geographically dispersed, and without the aid
large EHEC O157:H7 outbreak due to con- of DNA-based fingerprinting of the isolates,
sumption of a brand of lightly salted vegeta- it was difficult to recognize whole cases
bles from a company in Sapporo, Hokkaido as a diffuse outbreak. For example, in 2009,
(32). EHEC O157:H7 was isolated from the im- there was a diffuse outbreak from a restaurant
plicated product. Because of the products chain that spread all over Japan. Implicated
wide distribution, 169 patients were reported restaurants were providing cubically assem-
from five facilities for the elderly, hotels, res- bled meat, a type of processed meat made by
taurants, and families in Hokkaido and in- mixing and filling different parts of chopped
cluded four cases in different prefectures from meat, that is frozen and cut into cubes to make
which EHEC O157:H7 was isolated. Among the appearance of cube-cut steak. EHEC O157
169 patients, EHEC O157:H7 was isolated from contaminating the process probably survived
73 patients, 8 of whom, mostly elderly, died. in the product, and its intake caused illness.
Initially, some cases were reported as sporadic
cases by different prefectures, but the results
ROUTES OF TRANSMISSION OF of PFGE and MLVA of EHEC O157 isolates
EHEC INFECTIONS of the cases as well as the product showed
identical genotype and strongly suggested that
Consumption of contaminated foods is the these isolates were part of the outbreak (44).
most common source of EHEC infection, Other examples are the EHEC O157 infec-
and because cattle are considered to be major tions due to consumption of raw beef liver in
reservoirs of EHEC, consumption of raw Aichi Prefecture in 2010 (45). There were four

successive food poisoning incidents due to isolates of EHEC O157 (63); however, few
EHEC O157 in Nagoya City, Aichi Prefecture, trace-back studies have been conducted in
and all were related to consumption of raw Japanese outbreaks associated with consump-
beef liver. During the same period, food poison- tion of vegetables.
ings also caused by EHEC O157 occurred in Person-to-person transmission of EHEC is
multiple restaurants serving grilled meat in the most likely mode of transmission in the
Aichi Prefecture. Laboratory investigation of outbreaks reported from nursery schools, and
the cases showed identical genotypes of the prolonged fecal shedding of EHEC that was
isolates, and investigation of the marketing similar to that reported in the literature (64
routes revealed that the apparently diffuse 66) was seen in the outbreaks due to EHEC
incidents were actually the same-sourced food O26 (67), O157 (68), and O103 (69). Because
poisoning cases affecting a wide area. both symptomatic and asymptomatic patients
Prevalence of EHEC strains in beef cattle are usually among members of the staff of the
in Japan between November 2007 and March daycare and family members of the children
2008 was reported to be 8.9% and 0.4% for at daycare, extended stool culture examina-
EHEC O157 and O26, respectively, among tion of all children attending daycare (4) and
2,436 beef cattle reared on 406 farms (54). In family members would facilitate in taking
another study, prevalence of EHEC strains in control measures to prevent further dissemi-
932 healthy dairy cows from 123 farms was nation of EHEC infection.
12%, and 31 different O-serogroups, including
O26 but not O157, were identified (55). Using
stx-PCRs for screening, the same study also CONCLUSIONS
found that the prevalence of the stx gene-
positive samples among the dairy cows was The large outbreak in Sakai City and the
30.4%. Although EHEC O157 was found in subsequent outbreaks that occurred domesti-
retail meat at a low frequency (0.1%) in a na- cally in 1996 illustrate some of the serious
tional food surveillance (56), relatively high public health problems associated with EHEC
prevalence of EHEC in cattle on farms could O157 infections. The government initiated a
account for sporadic cases of EHEC and out- series of preventive control measures against
breaks when people have contact with these the disease, including enacting the new in-
animals at farms (5759). fectious disease control law in 1999. Since
Raw vegetables, such as lettuce (44, 46), then, however, a gradual increase in infection
cabbage (6), and cucumber (6), have been caused by EHEC has been reported, partly be-
implicated in EHEC outbreaks. Pickled veg- cause there have been improvements in meth-
etables (6) and lightly salted vegetables (32, odologies for isolating these organisms and
39, 40), which were fresh as salad, have also more laboratories were practicing enhanced
been implicated in the outbreaks. Trace-back screening for these organisms. Although EHEC
studies of an EHEC O104 outbreak associated O157 has been isolated predominantly in the
with consumption of fenugreek sprouts in EHEC infection, increased rates of non-O157
Germany (60) and France (61) showed that two EHEC cases are being reported with a variety of
outbreaks shared fenugreek seeds imported non-O157 EHEC serotypes. Especially among
from Egypt as the most likely common link the outbreaks between 2000 and 2012, one-half
(62), and studies of a spinach outbreak due to of the outbreaks were due to EHEC O157 and
EHEC O157 in the United States also showed the other half were due to O26, O111, O121,
that environmental samples including river O103, O145, and other serogroups; some serious
water, cattle feces, and wild pig feces from cases were caused by O111. Precautions need to
the field adjacent to the spinach-growing field be taken not only for non-O157 EHEC in addi-
were contaminated with identical pulsotype tion to EHEC O157 but also for emerging types

206 TERAJIMA ET AL.
of Shiga toxin-producing enteroaggregative Verocytotoxin-producing Escherichia coli (Entero-

E. coli, such as O104:H4 in Germany; a fatal case hemorrhagic E. coli) infections, Japan, 1996June
1997. Infec Agen Surv Rep 18:153154.
due to O86:HNM with similar phenotype to 3. Belongia EA, Osterholm MT, Soler JT,
the O104 E. coli was reported in Japan in 1999 Ammend DA, Braun JE, MacDonald KL. 1993.
(38). Nursery schools have become frequent Transmission of Escherichia coli O157:H7 infec-
outbreak settings where EHEC infection was tion in Minnesota child day-care facilities. JAMA
269:883888.
prolonged by person-to-person transmission
4. Gilbert M, Monk C, Wang HL, Diplock K,
involving family members of the child. To re- Landry L. 2008. Screening policies for daycare
duce the risk of outbreaks, it is important to attendees: lessons learned from an outbreak of
improve the sanitary controls in the nursery E. coli O157:H7 in a daycare in Waterloo, Ontario.
schools according to the guideline presented by Can J Public Health 99:281285.
MHLW to prevent EHEC dissemination. 5. Terajima J, Izumiya H, Wada A, Tamura K,
Watanabe H. 1999. Shiga toxin-producing Esch-
Laboratory investigations employing DNA- erichia coli O157:H7 in Japan. Emerg Infect Dis
based methods are critically important com- 5:301302.
ponents in finding and controlling diffuse 6. National Institute of Infectious Diseases and
outbreaks as early as possible while simulta- Tuberculosis and Infectious Diseases Control
neously conducting epidemiological investi- Division, Ministry of Health, Labour and Wel-
fare. 2012. Enterohemorrhagic Escherichia coli
gations of the outbreak. It is also important for infection in Japan as of April 2012. Infec Agen
relevant organizations and bodies to commu- Surv Rep 33:115116.
nicate effectively and work together to control 7. Gould LH, Mody RK, Ong KL, Clogher P,
and prevent EHEC infection. Cronquist AB, Garman KN, Lathrop S, Medus
C, Spina NL, Webb TH, White PL, Wymore K,
Gierke RE, Mahon BE, Griffin PM for the
ACKNOWLEDGMENTS Emerging Infections Program Foodnet Work-
ing Group PM. 2013. Increased recognition of
We thank all investigators of prefectural and non-O157 Shiga toxin-producing Escherichia coli
municipal public health institutes for provid- infections in the United States During 2000
ing us samples and information, and staffs in 2010: Epidemiologic features and comparison
with E. coli O157 infections. Foodborne Pathog Dis
IDSC for data collection and analysis. 10:453460.
We declare no conflicts of interest with 8. Caprioli A, Tozzi AE. 1998. Epidemiology of
regard to the manuscript. Shiga toxin-producing Escherichia coli infections
in continental Europe, p 3848. In Kaper JB,
OBrien AD (ed), Escherichia coli O157:H7 and
CITATION Other Shiga Toxin-Producing E. coli Strains. ASM
Press, Washington, DC.
Terajima J, Iyoda S, Ohnishi M, Watanabe H. 9. Blanco JE, Blanco M, Alonso MP, Mora A,
2014. Shiga toxin (verotoxin)-producing Esche- Dahbi G, Coira MA, Blanco J. 2004. Serotypes,
richia coli in Japan. Microbiol Spectrum 2(5): virulence genes, and intimin types of Shiga toxin
(verotoxin)-producing Escherichia coli isolates
EHEC-0011-2013.
from human patients: prevalence in Lugo, Spain,
from 1992 through 1999. J Clin Microbiol 42:311
319.
REFERENCES
10. Seto K, Taguchi M, Kobayashi K, Kozaki S.
1. National Institute of Infectious Diseases 2007. Biochemical and molecular characteriza-
and Tuberculosis and Infectious Diseases tion of minor serogroups of Shiga toxin-produc-
Control Division, Ministry of Health, Labour ing Escherichia coli isolated from humans in
and Welfare. 1998. Enterohemorrhagic E. coli Osaka Prefecture. J Vet Med Sci 69:12151222.
(verocytotoxin-producing E. coli) infection, 1996 11. Swaminathan B, Barrett TJ, Hunter SB, Tauxe
April 1998. Infec Agen Surv Rep 19:122123. RV. 2001. PulseNet: the molecular subtyping
2. National Institute of Infectious Diseases network for foodborne bacterial disease sur-
and Infectious Diseases Control Division, Min- veillance, United States. Emerg Infect Dis 7:382
istry of Health and Welfare of Japan. 1997. 389.

12. Watanabe H, Terajima J, Izumiya H, Iyoda S, 22. Izumiya H, Pei Y, Terajima J, Ohnishi M,
Tamura K. 2002. [PulseNet Japan: surveillance Hayashi T, Iyoda S, Watanabe H. 2010. New
system for the early detection of diffuse out- system for multilocus variable-number tandem-
break based on the molecular epidemiological repeat analysis of the enterohemorrhagic Esche-
method]. Kansenshogaku Zasshi 76:842848 (In richia coli strains belonging to three major
Japanese). serogroups: O157, O26, and O111. Microbiol Im-
13. Swaminathan B, Gerner-Smidt P, Ng LK, munol 54:569577.
Lukinmaa S, Kam KM, Rolando S, Gutierrez 23. Lindstedt BA. 2005. Multiple-locus variable
EP, Binsztein N. 2006. Building PulseNet Inter- number tandem repeats analysis for genetic fin-
national: an interconnected system of laboratory gerprinting of pathogenic bacteria. Electrophore-
networks to facilitate timely public health resis 26:25672582.
cognition and response to foodborne disease 24. Ooka T, Terajima J, Kusumoto M, Iguchi A,
outbreaks and emerging foodborne diseases. Kurokawa K, Ogura Y, Asadulghani M,
Foodborne Pathog Dis 3:3650. Nakayama K, Murase K, Ohnishi M, Iyoda S,
14. Terajima J, Izumiya H, Iyoda S, Mitobe J, Watanabe H, Hayashi T. 2009. Development of a
Miura M, Watanabe H. 2006. Effectiveness of multiplex PCR-based rapid typing method for
pulsed-field gel electrophoresis for the early de- enterohemorrhagic Escherichia coli O157 strains.
tection of diffuse outbreaks due to Shiga toxin- J Clin Microbiol 47:28882894.
producing Escherichia coli in Japan. Foodborne 25. National Institute of Infectious Diseases and
Pathog Dis 3:6873. Tuberculosis and Infectious Diseases Control
15. Centers for Disease Control and Prevention. Division, Ministry of Health, Labour and Wel-
2005. Escherichia coli O157:H7 infections associ- fare. 2009. Enterohemorrhagic Escherichia coli
ated with ground beef from a U.S. military infection 2006, 2007. Infec Dis Weekly Rep 11:16
installationOkinawa, Japan, February 2004. 22 (In Japanese).
Morbid Mortal Weekly Rep 54:4042. 26. National Institute of Infectious Diseases and
16. Foley SL, Simjee S, Meng J, White DG, Tuberculosis and Infectious Diseases Control
McDermott PF, Zhao S. 2004. Evaluation of Division, Ministry of Health, Labour and Wel-
molecular typing methods for Escherichia coli fare. 2009. Enterohemorrhagic Escherichia coli
O157:H7 isolates from cattle, food, and humans. infection 2008. Infec Dis Weekl. Rep 11:1523 (In
J Food Prot 67:651657. Japanese).
17. Noller AC, McEllistrem MC, Stine OC, Morris 27. National Institute of Infectious Diseases and
JG Jr, Boxrud DJ, Dixon B, Harrison LH. 2003. Tuberculosis and Infectious Diseases Control
Multilocus sequence typing reveals a lack of di- Division, Ministry of Health, Labour and Wel-
versity among Escherichia coli O157:H7 isolates fare. 2010. Enterohemorrhagic Escherichia coli
that are distinct by pulsed-field gel electropho- infection 2009. Infec Dis Weekly Rep 12:1119 (In
resis. J Clin Microbiol 41:675679. Japanese).
18. Hyytia-Trees E, Smole SC, Fields PA, 28. National Institute of Infectious Diseases and
Swaminathan B, Ribot EM. 2006. Second gen- Tuberculosis and Infectious Diseases Control
eration subtyping: a proposed PulseNet protocol Division, Ministry of Health, Labour and Wel-
for multiple-locus variable-number tandem re- fare. 2011. Enterohemorrhagic Escherichia coli
peat analysis of Shiga toxin-producing Escherichia infection 2010. Infec Dis Weekly Rep 13:1120 (In
coli O157 (STEC O157). Foodborne Pathog Dis Japanese).
3:118131. 29. National Institute of Infectious Diseases and
19. Lindstedt BA, Vardund T, Kapperud G. 2004. Tuberculosis and Infectious Diseases Control
Multiple-locus variable-number tandem-repeats Division, Ministry of Health, Labour and Welfare.
analysis of Escherichia coli O157 using PCR 2012. Enterohemorrhagic Escherichia coli infection
multiplexing and multi-colored capillary electro- 2011. Infec Dis Weekly Rep 14:919 (In Japanese).
phoresis. J Microbiol Methods 58:213222. 30. Tarr PI, Gordon CA, Chandler WL. 2005. Shiga-
20. Keys C, Kemper S, Keim P. 2005. Highly diverse toxin-producing Escherichia coli and haemolytic
variable number tandem repeat loci in the E. coli uraemic syndrome. Lancet 365:10731086.
O157:H7 and O55:H7 genomes for high-resolution 31. Boyce TG, Swerdlow DL, Griffin PM. 1995.
molecular typing. J Appl Microbiol 98:928940. Escherichia coli O157:H7 and the hemolytic-
21. Noller AC, McEllistrem MC, Pacheco AG, uremic syndrome. N Engl J Med 333:364368.
Boxrud DJ, Harrison LH. 2003. Multilocus 32. National Institute of Infectious Diseases and
variable-number tandem repeat analysis distin- Tuberculosis and Infectious Diseases Control
guishes outbreak and sporadic Escherichia coli Division, Ministry of Health, Labour and
O157:H7 isolates. J Clin Microbiol 41:53895397. Welfare. 2013. Enterohemorrhagic Escherichia

208 TERAJIMA ET AL.
coli infection in Japan as of April 2013. Infec Agen 42. National Institute of Infectious Diseases and
Surv Rep 34:123124. Tuberculosis and Infectious Diseases Control
33. Ostroff SM, Tarr PI, Neill MA, Lewis JH, Division, Ministry of Health, Labour and Wel-
Hargrett-Bean N, Kobayashi JM. 1989. Toxin fare. 2005. Enterohemorrhagic Escherichia coli
genotypes and plasmid profiles as determinants of infection as of May 2005. Infec Agen Surv Rep
systemic sequelae in Escherichia coli O157:H7 in- 26:137138.
fections. J Infect Dis 160:994998. 43. National Institute of Infectious Diseases and
34. Scotland SM, Willshaw GA, Smith HR, Rowe B. Tuberculosis and Infectious Diseases Control
1987. Properties of strains of Escherichia coli be- Division, Ministry of Health, Labour and Wel-
longing to serogroup O157 with special reference fare. 2008. Enterohemorrhagic Escherichia coli
to production of Vero cytotoxins VT1 and VT2. infection in Japan as of April 2008. Infec Agen
Epidemiol Infect 99:613624. Surv Rep 29:117118.
35. Scheutz F, Nielsen EM, Frimodt-Moller J, 44. National Institute of Infectious Diseases and
Boisen N, Morabito S, Tozzoli R, Nataro JP, Tuberculosis and Infectious Diseases Control
Caprioli A. 2011. Characteristics of the entero- Division, Ministry of Health, Labour and Wel-
aggregative Shiga toxin/verotoxin-producing Esch- fare. 2010. Enterohemorrhagic Escherichia coli
erichia coli O104:H4 strain causing the outbreak of infection in Japan as of May 2010. Infec Agen
haemolytic uraemic syndrome in Germany, May to Surv Rep 31:152153.
June 2011. Euro Surveill 16:510. 45. National Institute of Infectious Diseases and
36. Morabito S, Karch H, Mariani-Kurkdjian P, Tuberculosis and Infectious Diseases Control
Schmidt H, Minelli F, Bingen E, Caprioli A. Division, Ministry of Health, Labour and Wel-
1998. Enteroaggregative, Shiga toxin-producing fare. 2011. Enterohemorrhagic Escherichia coli
Escherichia coli O111:H2 associated with an out- infection in Japan as of April 2011. Infec Agen
break of hemolytic-uremic syndrome. J Clin Micro- Surv Rep 32:125126.
biol 36:840842. 46. National Institute of Infectious Diseases and
37. Frank C, Werber D, Cramer JP, Askar M, Faber Tuberculosis and Infectious Diseases Control
M, an der Heiden M, Bernard H, Fruth A, Division, Ministry of Health, Labour and Wel-
Prager R, Spode A, Wadl M, Zoufaly A, Jordan fare. 2001. Enterohemorrhagic Escherichia coli
S, Kemper MJ, Follin P, Muller L, King LA, infection in Japan as of April 2001. Infec Agen
Rosner B, Buchholz U, Stark K, Krause G, Team Surv Rep 22:135136.
HUSI. 2011. Epidemic profile of Shiga-toxin- 47. National Institute of Infectious Diseases and
producing Escherichia coli O104:H4 outbreak in Tuberculosis and Infectious Diseases Control
Germany. N Engl J Med 365:17711780. Division, Ministry of Health, Labour and Wel-
38. Iyoda S, Tamura K, Itoh K, Izumiya H, Ueno N, fare. 2006. Enterohemorrhagic Escherichia coli
Nagata K, Togo M, Terajima J, Watanabe H. infection in Japan as of May 2006. Infec Agen
2000. Inducible stx2 phages are lysogenized in Surv Rep 27:141142.
the enteroaggregative and other phenotypic Esche- 48. National Institute of Infectious Diseases and
richia coli O86:HNM isolated from patients. Tuberculosis and Infectious Diseases Control
FEMS Microbiol Lett 191:710. Division, Ministry of Health, Labour and Wel-
39. National Institute of Infectious Diseases and fare. 2007. Enterohemorrhagic Escherichia coli
Tuberculosis and Infectious Diseases Control infection in Japan as of April 2007. Infec Agen
Division, Ministry of Health, Labour and Wel- Surv Rep 28:131132.
fare. 2002. Enterohemorrhagic Escherichia coli 49. National Institute of Infectious Diseases and
infection as of April 2002. Infec Agen Surv Rep Tuberculosis and Infectious Diseases Control
23:137138. Division, Ministry of Health, Labour and Wel-
40. National Institute of Infectious Diseases and fare. 2009. Enterohemorrhagic Escherichia coli
Tuberculosis and Infectious Diseases Control infection in Japan as of April 2009. Infec Agen
Division, Ministry of Health, Labour and Wel- Surv Rep 30:119120.
fare. 2003. Enterohemorrhagic Escherichia coli 50. Kato K, Shimoura R, Nashimura K, Yoshifuzi K,
infection as of May 2003. Infec Agen Surv Rep Shiroshita K, Sakurai N, Kodama H, Kuramoto
24:129130. S. 2005. Outbreak of enterohemorrhagic Esche-
41. National Institute of Infectious Diseases and richia coli O111 among high school participants in
Tuberculosis and Infectious Diseases Control excursion to Korea. Jpn J Infect Dis 58:332333.
Division, Ministry of Health, Labour and Wel- 51. Tilden J Jr, Young W, McNamara AM, Custer
fare. 2004. Enterohemorrhagic Escherichia coli C, Boesel B, Lambert-Fair MA, Majkowski J,
infection as of May 2004, Japan. Infec Agen Surv Vugia D, Werner SB, Hollingsworth J, Morris
Rep 25:138139. JG Jr. 1996. A new route of transmission for

Escherichia coli: infection from dry fermented Escherichia coli O104:H4 associated with sprouts.
salami. Am J Public Health 86:11421145. N Engl J Med 365:17631770.
52. Paton AW, Ratcliff RM, Doyle RM, Seymour- 61. Gault G, Weill FX, Mariani-Kurkdjian P,
Murray J, Davos D, Lanser JA, Paton JC. 1996. Jourdan-da Silva N, King L, Aldabe B, Charron
Molecular microbiological investigation of an M, Ong N, Castor C, Mace M, Bingen E, Noel H,
outbreak of hemolytic-uremic syndrome caused Vaillant V, Bone A, Vendrely B, Delmas Y,
by dry fermented sausage contaminated with Combe C, Bercion R, dAndigne E, Desjardin M,
Shiga-like toxin-producing Escherichia coli. J Clin de Valk H, Rolland P. 2011. Outbreak of
Microbiol 34:16221627. haemolytic uraemic syndrome and bloody diar-
53. Terajima J, Izumiya H, Iyoda S, Tamura rhoea due to Escherichia coli O104:H4, south-west
K, Watanabe H. 1999. Detection of a multi- France, June 2011. Euro Surveill 16:13.
prefectual E. coli O157:H7 outbreak caused by 62. European Food Safety Authority. 2011. Tracing
contaminated Ikura-Sushi ingestion. Jpn J Infect seeds, in particular fenugreek (Trigonella foenum-
Dis 52:5253. graecum) seeds, in relation to the Shiga toxin-
54. Sasaki Y, Tsujiyama Y, Kusukawa M, producing E. coli (STEC) O104:H4 2011 outbreaks
Murakami M, Katayama S, Yamada Y. 2011. in Germany and France. Technical Report of
Prevalence and characterization of Shiga toxin- EFSA. European Food Safety Authority, Parma,
producing Escherichia coli O157 and O26 in beef Italy.
farms. Vet Microbiol 150:140145. 63. California Food Emergency Response Team,
55. Kobayashi H, Kanazaki M, Ogawa T, Iyoda S, California Department of Health Services, US
Hara-Kudo Y. 2009. Changing prevalence of Food and Drug Administration. 2007. Investi-
O-serogroups and antimicrobial susceptibility gation of an Escherichia coli O157:H7 outbreak
among STEC strains isolated from healthy dairy associated with Dole pre-packaged spinach. En-
cows over a decade in Japan between 1998 and vironmental Investigation Report. California De-
2007. J Vet Med Sci 71:363366. partment of Public Health, Sacramento, CA.
56. Hara-Kudo Y, Konuma H, Kamata Y, Miyahara 64. Swerdlow DL, Griffin PM. 1997. Duration of
M, Takatori K, Onoue Y, Sugita-Konishi Y, faecal shedding of Escherichia coli O157:H7 among
Ohnishi T. 2012. Prevalence of the main food- children in day-care centres. Lancet 349:745746.
borne pathogens in retail food under the national 65. ODonnell JM, Thornton L, McNamara EB,
food surveillance system in Japan. Food Addit Prendergast T, Igoe D, Cosgrove C. 2002. Out-
Contam Part A Chem Anal Control Expo Risk break of Vero cytotoxin-producing Escherichia
Assess 30:14501458. coli O157 in a child day care facility. Commun Dis
57. Waguri A, Sakuraba M, Sawada Y, Abe K, Public Health 5:5458.
Onishi M, Tanaka J, Kudo Y, Saito K. 2007. 66. Galanis E, Longmore K, Hasselback P, Swann
Enterohemorrhagic Escherichia coli O157 infec- D, Ellis A, Panaro L. 2003. Investigation of an E.
tion presumably caused by contact with infected coli O157:H7 outbreak in Brooks, Alberta, June
cows, Aomori Prefecture, Japan. Jpn J Infect Dis July 2002: the role of occult cases in the spread of
60:321322. infection within a daycare setting. Can Commun
58. Muto T, Matsumoto Y, Yamada M, Ishiguro Y, Dis Rep 29:2128.
Kitazume H, Sasaki K, Toba M. 2008. Outbreaks 67. Kanazawa Y, Ishikawa T, Shimizu K, Inaba S.
of enterohemorrhagic Escherichia coli O157 in- 2007. Enterohemorrhagic Escherichia coli
fections among children with animal contact at a outbreaks in nursery and primary schools. Jpn J
dairy farm in Yokohama City, Japan. Jpn J Infect Infect Dis 60:326327.
Dis 61:161162. 68. Sugiyama A, Iwade Y, Akachi S, Nakano Y,
59. Akiba Y, Kimura T, Takagi M, Akimoto T, Matsuno Y, Yano T, Yamauchi A, Nakayama O,
Mitsui Y, Ogasawara Y, Omichi M. 2005. Out- Sakai H, Yamamoto K, Nagasaka Y, Nakano T,
break of enterohemorrhagic Escherichia coli O121 Ihara T, Kamiya H. 2005. An outbreak of
among school children exposed to cattle in a Shigatoxin-producing Eshcherichia coli O157:H7
ranch for public education on dairy farming. Jpn J in a nursery school in Mie Prefecture. Jpn J Infect
Infect Dis 58:190192. Dis 58:398400.
60. Buchholz U, Bernard H, Werber D, Bohmer 69. Muraoka R, Okazaki M, Fujimoto Y, Jo N,
MM, Remschmidt C, Wilking H, Delere Y, an Yoshida R, Kiyoyama T, Oura Y, Hirakawa K,
der Heiden M, Adlhoch C, Dreesman J, Ehlers Jyukurogi M, Kawano K, Okada M, Shioyama
J, Ethelberg S, Faber M, Frank C, Fricke G, Y, Iryoda K, Wakamatu H, Kawabata N. 2007.
Greiner M, Hohle M, Ivarsson S, Jark U, An enterohemorrhagic Escherichia coli O103
Kirchner M, Koch J, Krause G, Luber P, Rosner outbreak at a nursery school in Miyazaki Prefec-
B, Stark K, Kuhne M. 2011. German outbreak of ture, Japan. Jpn J Infect Dis 60:410411.

Animal Reservoirs of Shiga
Toxin-Producing Escherichia coli
ANIL K. PERSAD1 and JEFFREY T. LEJEUNE1

11
Escherichia coli strains that carry Shiga toxin genes are commonly isolated
from the gastrointestinal tract of a wide variety of animal species (Table 1).
Intestinal carriage of most Shiga toxin-producing E. coli (STEC) strains by
domestic and wild animals has little clinical relevance to either the animal
hosts or humans. Most animals lack receptors for Shiga toxin, and in humans,
the presence of additional virulence factors, in addition to the stx gene, is as-
sociated with disease outcomes (13). However, animals may harbor STEC
strains that are pathogenic to humans. This article focuses on the role of animals
as reservoirs for infection or as spillover hosts. Within the animal, these bacteria
may be resident or transient in the gastrointestinal tract. Determining whether
STEC is resident in flora or transient is not possible during cross-sectional
observational epidemiological studies when only one sample is collected from
an animal and there is no serial testing. Even under experimental conditions it
is difficult to determine if repeated isolation from the feces over time is a result
of replication of the organism in the animal or repeated exposure.
Animals capable of maintaining STEC carriage in the absence of continuous
exposure or those that frequently are reexposed to STEC from environmental
sources can serve as potential sources of interspecies and intraspecies infection.
Cattle are regarded as the natural reservoir of STEC (1), but other ruminant
1
Food Animal Health Research Program, Ohio Agriculture Research and Development Center, The Ohio State University,
Wooster, OH 44691.
211

212 PERSAD AND LEJEUNE
TABLE 1 Animal hosts of Shiga toxin-producing E. coli species such as sheep, goats, and deer may
Common also act as reservoirs. Animals may also be
name Scientic name Reference(s)
categorized as spillover hosts. Similar to re-
Cattle Bos taurus 1, 7, 8, 10, 19, servoirs, these animals are susceptible to col-
2123, 27, onization and may transmit disease; however,
2933
Goats Capra aegagrus 34, 39, 40, 43,
once they are no longer exposed to a source of
hircus 44, 48, 49, 53 STEC, they do not maintain this colonization.
Sheep Ovis aries 1, 35, 39, This inability to maintain STEC colonization
4347 in the absence of exposure is the critical factor
Water buffalo Bubalus bubalis 53, 54, 61
that differentiates these animals from reser-
White-tailed Odocoileus 6264, 6771
deer virginianus
voirs. Epidemiological evidence indicates that
Bison Bison bison 7477 birds, swine, dogs, and horses may be spillover
Elk Cervus canadensis 72, 73, 80 hosts. Dead-end hosts, as the name suggests,
Llamas Lama glama 191 are incapable of transmitting STEC naturally
Alpaca Lama pacos 83, 192 to other animals. In the absence of evidence
Yak Bos grunniens 83
Eland Taurotragus oryx 83 that aquatic species such as finfish and shell-
Antelope Antilope cervicapra 83 fish transmit the organism to other animals,
Mountain Oreamnos 84 they may act as dead-end hosts for STEC, only
goat americanus transmitting STEC when they are consumed
Guanaco Lama guanicoe 79 (46).
Horses Equus ferus caballus 8588, 91
Donkey Equus africanus 84, 89, 90
The factors governing the prevalence and
asinus number of bacteria present in the digestive
Domestic Sus domesticus 1, 92, 9496, tract of animals are poorly understood, even
swine 101, 102 for the best-studied species, bovines. The
Feral swine Sus scrofa 103105
prevalence and magnitude of STEC infection
Chicken Gallus gallus 92, 94, 125, 126
domesticus
in animals is dependent on a complex inter-
Turkeys Meleagris gallopavo 92, 126 action of external and internal conditions:
Pigeon Columba livia 111, 116 the frequency and dose of exposure, the
Starling Sturnus vulgaris 110, 112114 hosts susceptibility to infection, and the du-
Geese Branta canadensis 107, 119 ration of shedding. Moreover, these factors
Turtle dove Streptopelia turtur 112
Barn swallow Hirundo rustica 112 may vary considerably among species and
Dogs Canis lupus 39, 163, 165 even among the same species as a function
familiaris of age, immunity, housing, diet, climate, and
Cats Felis catus 166, 170, 171 sanitation.
Coyote Canis latrans 84
Fox Vulpes vulpes 84
Rabbit Oryctolagus 143, 144
cuniculus
ANIMAL SPECIES OF IMPORTANCE IN THE
Raccoon Procyon lotor 152
Fish and 129132
EPIDEMIOLOGY OF STEC
shellsh
Norway rats Rattus norvegicus 108, 137, 138 Cattle
Ground hog Marmota monax 84
Patagonian Dolichotis 83 Cattle are recognized as a primary reservoir
cavy patagonus for STEC strains, especially the serogroup
Frogs 193 O157 (1). Like humans, cattle are exposed to
Ferretsa Mustela putorius 172 STEC through contaminated food and water
furo
Micea Mus spp. 114, 142, 180 or by exposure to the feces of other animals
a
shedding the organism. The infectious dose
Experimental infections only.
in cattle is estimated to be as low as 300 CFU

CHAPTER 11 Animal Reservoirs of STEC 213
(7). STEC colonization in cattle is usually animals not carried to shows (18). These
asymptomatic due to the absence of vascular animals, on returning to the farm, can then
receptors for Shiga toxins (8). The absence shed STEC, thus exposing other animals to
of these globotriaosylceramide-3 (Gb3) vas- infection or colonization.
cular receptors, especially in the intestinal In the United States, STEC O157 is found
vasculature, means Shiga toxins cannot be on almost all cattle farms, with the orga-
endocytosed and transported to other organs nism being shed intermittently by most ani-
that may be sensitive to Shiga toxins (9). mals (19). STEC is shed mainly through the
The terminal part of the large intestine, the feces of colonized animals; however, Shiga
recto-anal junction (RAJ), is the main site toxin genes have been detected from E. coli
of STEC colonization in cattle (10). The in- strains isolated from the milk of mastitic cows
creased production of factors associated with (20). Although a rare occurrence, milk from
environmental survival among cattle isolates these animals can be a potential source of
of STEC O157, compared to human-origin STEC infection to nursing calves, animals fed
isolates, combined with the low infectious waste milk, and the human population. Most
dose, may provide a selection bias for these milk-borne STEC cases are, however, due
organisms to recolonize cattle and maintain to postmilking contamination and the subse-
the organism in the bovine population (11). quent consumption of these products without
Improper feed storage facilities or poorly de- pasteurization.
signed feeding troughs can result in feed being The prevalence of STEC in cattle popu-
contaminated with the feces of wild or do- lations is highly variable, with peaking and
mestic animals. dropping at unpredictable times. At any spe-
Livestock drinking water contamination cific time, the global prevalence of STEC O157
can occur at its source or at the farm. Surface in cattle has been reported to range from 0 to
water and groundwater sources may be con- 71% (21), and the herd infection rate has been
taminated from effluent runoff from farms reported to be up to 100% in some studies
and urban areas. Leaching from pastures may (22). In the United States, the herd preva-
also result in groundwater contamination (12, lence of STEC may range between 10 and 20%
13). At the farm, improperly designed water (23). The global prevalence of STEC O157 has
troughs can be contaminated by animal feces. been reported to range from 0.2 to 48.8% in
LeJeune et al. (14) showed that 1.3% of 473 dairy cattle and 0.2 to 27.8% in beef animals
water troughs sampled in three U.S. states whereas the global prevalence of non-O157
were positive for STEC O157. STEC has STEC may range from 0.4 to 74% in dairy
also been demonstrated to persist for more cattle and 2.1 to 70.1% for beef animals, re-
than 4 months in contaminated water troughs spectively, as reported in two independent
(15). studies (24, 25).
Other management practices may also Colonized cattle can shed STEC O157 at
affect the incidence of STEC in animal popu- levels as high as 1.1 105 CFU/g feces (26) and
lations. Flushing alleyways with water infor as long as 10 weeks (27). The average
creased the incidence of STEC in animals duration of STEC O157 carriage is 30 days;
compared to other manure removal strategies however, in rare cases animals may be colo-
(16). Animals housed on sawdust were also nized for up to 1 year (28). Animals excreting
found to have a higher incidence of STEC greater than 104 CFU/g feces are termed
than animals housed on sand-based bedding super-shedding animals (29, 30). Longitudi-
(17). Movement of animals to and from farms nal studies, however, indicate super-shedding
also increases the risk of STEC transmission: represents a phase or stage of colonization
Animals carried to animal exhibitions have a of all cattle and can be typically observed in
greater likelihood of contracting STEC than a small fraction of animals in a population

at any given time (31). Nevertheless, it is not (37; CJ Hovde, University of Idaho, personal
debatable that animals excreting these high communication). STEC O157 shedding pat-
levels of STEC are responsible for the major- terns between rectally and orally inoculated
ity of environmental contamination (30, 32). sheep were found to be similar, thus indi-
Calves tend to shed STEC at the lowest levels cating that STEC O157 may be able to ef-
before weaning; however, the highest shed- fectively colonize the terminal rectum (38).
ding is exhibited in the period immediately However, inefficient attachment of large num-
post weaning (33). The shedding of STEC bers of STEC to the RAJ may account for
also tends to be higher in the warmer months, reduced shedding periods compared to those
with peak prevalence being in summer and in cattle.
early fall with a drastic decrease in prevalence Similar to cattle, small ruminants tend to
during the winter months (19). be asymptomatic shedders of STEC. This
trait was demonstrated when the screening
of healthy animals in Berlin reported that
Small Ruminants
66.6% of the sheep and 56.1% of the goats
Small ruminants, particularly sheep and goats, tested were found to be STEC carriers (39).
are important reservoirs of STEC O157 (34). Similar results were obtained in Spain where
Considerable research has focused on the 47% of healthy goats tested positive for
role of sheep in the epidemiology of STEC shedding STEC (40). The asymptomatic fea-
infections; however, there is limited published ture of STEC carriage is possibly also due to
research on the role of goats (34). Although the lack of Shiga toxin vascular receptors in
cattle have been identified as the major res- small ruminants.
ervoir for STEC in the United States, small Many direct-contact human infections are
ruminants play a greater role in the epidemi- attributed to contact with sheep and goats
ology of STEC infections in other countries. at petting zoos and open farms (41, 42).
For example, sheep have been identified as One study investigating the prevalence of
the host of significance in Australia (1) and zoonotic agents on small city farms in south-
have also been recognized as an important ern Germany found that 100% of the sheep
reservoir of STEC O26 in Norway (35). In and 89.3% of the goats tested positive for
addition to STEC serogroups O157 and O26, STEC (43). Small ruminants, especially goats,
sheep have been cited as reservoirs for more generally exhibit inquisitive behavior and
than 100 other serotypes of STEC, including thus may have greater contact with humans,
O115, O128, and O130 (34, 35) increasing the potential for transmission to
Transmission of STEC to small ruminants humans (34). Human infections have also
occurs through the same pathway as in cattle. been linked to the consumption of unpas-
The site of STEC colonization, however, may teurized milk and cheese made from con-
be different. Unlike cattle, tropism for RAJ taminated goat or sheep milk (40, 44).
has not been described for all small ruminants Sheep are the primary reservoir for STEC
(34). Following exposure to STEC O157, in in Australia, with the serotype of importance
some studies few attachment and effacement being O26; however, the risk of human in-
lesions were visible on the intestinal mucosa fection was deemed insignificant due to low
and the entire distal intestine, including the prevalence rates (1, 45). Although the within-
cecum, colon, and rectum, was colonized, not herd prevalence was low, previous research
only the RAJ (34, 36). However, in mature reported that 90% of Australian sheep farms
sheep given a single oral dose of a human had animals testing positive for STEC (46).
clinical isolate of STEC O157:H7, analysis of In Norway, however, the risk of human in-
digesta and intestinal mucosa showed coloni- fection from sheep was much more significant
zation occurs exclusively at the RAJ mucosa since almost 50% of the sheep O26 isolates

had multiple-locus variable number tandem Other Ruminants

repeat analysis profiles similar to that found in
Water Buffalo (Bubalus bubalis)
human clinical cases (35). The importance of
In addition to cattle, sheep, and goats, other
sheep in STEC epidemiology was also dem-
ruminant species have also been identified as
onstrated by Oporto et al. (47), who reported
shedders of STEC. Water buffalo (Bubalus
that greater than 50% of sheep herds in Spain
bubalis) has been identified as an important
had animals shedding non-O157 STEC com-
reservoir of STEC O157 in many countries
pared to 20.7% in dairy cattle and 46% in beef
(61). The water buffalo is reared in many
cattle.
countries because of its ability to serve a dual
In the United States, Jacob et al. reported
purpose, as both a milk and meat producer.
that 11.1% of goat fecal samples collected at
Buffaloes are also able to thrive much better
slaughter had STEC O157 and 14.5% had at
on poor quality forages than the Bos taurus
least one non-O157 STEC serotype (48, 49).
species, thus making them suitable for sub-
The STEC O157 flock prevalence in Spain
sistence farming. There are large commercial
was reported to be 8.7% and individual prev-
meat and milk water buffalo herds in Asia
alence, 7.8% (47). A similarly low STEC O157
and South America, while in Europe water
prevalence of 5.8% was also reported in
buffalo is primarily reared for milk produc-
Scotland (50). Low STEC O157 prevalences
tion. In Bangladesh, STEC colonies were iso-
were also reported in the United Kingdom
lated from 38% of the buffaloes sampled
and Holland, with the prevalence being
before slaughter. Almost half of these isolates
0.1% and 4.0%, respectively (51, 52). Lesser
were identified as being O157 (54). Galiero
developed countries have also reported the
et al. reported an almost a similar prevalence
presence of STEC in their small ruminant
in Italy, with 14.5% of the animals shed-
population. In Vietnam, 100% of the goat
ding O157 (61). In Vietnam, 27% of the buffa-
farms surveyed had animals shedding STEC,
loes screened were found to be positive for
and the within-herd shedding was dramati-
STEC. Serotyping of the isolates, however,
cally higher than that reported elsewhere,
revealed that none of the isolates were O157
with up to 65% of animals shedding STEC
(53).
(53). In Bangladesh, almost 10% of the small
ruminants being slaughtered tested positive
for STEC O157 (54). As evidenced by past Deer
outbreaks, STEC in animals from lesser de- There are an estimated 30 million white-tailed
veloped countries can potentially be a serious deer in the United States (194). The role
threat to food safety, since in those countries of white-tailed deer (Odocoileus virginianus)
there may not be strict hygienic slaughter as a potential reservoir for STEC was first
practices; thus contaminated meat could easily reported in 1999 when almost 2.4% of deer
enter the food chain (55, 56). sampled tested positive (62). The presence of
The shedding of STEC in small ruminants STEC O157:H7 in deer feces was later con-
has been demonstrated to be age and season firmed by Renter et al. (68), who found the
dependent. Younger animals tend to have a STEC prevalence in Nebraska white-tailed
lower prevalence of STEC than older animals deer to be 0.25%. Similarly, low STEC prev-
do (40, 5759). A longitudinal study spanning alences of 0.2% were reported in hunter-
6 months in the United States demonstrated a harvested captive deer in Louisiana (63) and
peak in STEC prevalence during summer (60). 3.3% in farm-raised deer in Ohio (64). Other
This trend was also observed in Italy, where species of deer, including red deer (Cervus
animals screened during the warmer months elaphus), fallow deer (Dama dama), and roe
of the year had a higher prevalence of STEC deer (Capreolus capreolus), have also been
O157 (58). identified as capable of shedding STEC sero-

types (65, 66). Almost 50% of Pennsylvanian morphological characteristics (74). In the
white-tailed deer fecal samples screened United States consumption of bison meat has
tested positive for stx genes; however, only increased, thus increasing the risk of trans-
8% possessed the eae gene, which is neces- mission from bison to humans (75). This risk
sary for colonization of the human intestine was exemplified in 2010 when a multistate
(67). outbreak of STEC O157:H7 was associated
Feral deer are known to share pastures with bison meat consumption. The preva-
with cattle and can also be found in close lence of STEC O157:H7 in bison has been
proximity to many dairy farms. The close as- reported to be as high as 42% (76). STEC
sociation between deer and livestock implies O157 has also been isolated from the car-
that deer can serve to maintain and dissemi- cass of slaughtered bison at a prevalence of
nate STEC between and within cattle herds 1.13% (77). Non-O157 STEC serotypes in-
(68, 69). cluding O45, O103, O111, O113, O121, and O145
Human STEC O157 infections were first have also been isolated from bison carcasses;
associated with venison in 1997 when six however, none of these isolates possessed stx
persons became ill as a result of consuming genes (78).
jerky made from venison (70). Since then, STEC O157 and non-O157 STEC strains
there have been numerous other cases asso- have been isolated from other captive and
ciated with venison, with one of the most re- wild nondomesticated ruminant species. They
cent published reports being an outbreak include llamas, moose, alpacas, antelopes, and
of non-O157 STEC among high school stu- yaks (7984). These animals can transmit
dents that was associated with consumption of STEC to humans directly by contact at petting
venison they had killed and processed. Two zoos or indirectly through fecal deposition
STEC serotypes, O103:H2 and O145:NM, were in water sources, vegetable fields, or recrea-
isolated from the samples analyzed; however, tional areas or on meat. Further research is
the O145:NM serotype was found to be Shiga required to determine the role these animals
toxin negative (71). may have in the epidemiology of human STEC
infections.
Elk (Cervus canadensis)
Similar to deer, elk have also been associated
Equine
with numerous food-borne disease outbreaks
(72). Gilbreath et al. (80) reported that over Published data on the epidemiology of STEC
22% of wild Idaho elk screened were positive carriage in horses are limited. There are
for STEC. A slightly lower prevalence (7%) of also no published case reports describing the
stx genes was detected in fecal pellets col- clinical features of STEC infection in horses.
lected from elk in Colorado (73). In this study The available published data on the preva-
none of the animals were found to shed sero- lence of STEC in horses (8588) and donkeys
group O157, but serogroups O103 and O146 (84, 89, 90) indicate that they are not major
were detected. Interestingly in both studies, reservoirs of STEC and may instead be spill-
the incidence of STEC in the elk feces was over hosts. Only one of 400 horse fecal sam-
found to be higher than in mule deer, which ples screened in Germany was positive for
shared the same grazing grounds. STEC. The serotype isolated was O113: H21
(88). A similarly low prevalence of STEC has
Bison (Bison bison) also been detected in the equine population
Another potentially important animal reser- in the United States. Only one of 242 horse
voir of STEC is the American bison (Bison fecal samples from Ohio tested positive for
bison). This potential is supported by the fact STEC O157:H7. Interestingly, this case shared
that both cattle and bison share similar RAJ housing accommodations with a goat that

also shed STEC O157. The isolates from Though a relatively low prevalence of
both animals had indistinguishable multiple- STEC O157:H7 has been reported, swine
locus variable number tandem repeat analy- have been shown to harbor and shed STEC
sis patterns (87). A similarly low prevalence for up to 2 months post infection (101). Non-
has been reported by Hancock et al., who O157 STEC serotypes have also been isolated
reported that 1% of horses sampled (n = 90) from pigs; however, many of these isolates
tested positive for O157:H7 (85). Screening lack the virulence factors required to cause
of fecal samples from horses located in the human disease (1). Despite a low prevalence
Sacramento Valley revealed a slightly higher of pathogenic STEC serotypes, the poten-
prevalence than that recorded in Ohio. Four tial for human infection from swine exists.
of 156 samples (2.6%) tested positive for the This risk is exemplified by a recent Canadian
Shiga toxin 2 gene (86). Notably, as was seen outbreak of STEC O157:H7 associated with
in Ohio, all the positive horses were also consumption of pork, with infected persons
housed on farms containing ruminants. De- having the identical STEC O157:H7 isolate
spite the low STEC prevalence in horses, to that found in the pork meal served (102)
there are reported human clinical cases asso-
ciated with infection from horse contact
Feral Swine
(91), and one must be aware of this potential
source of infection. Feral swine is another wildlife species that
has been associated with STEC disease in the
human population. There are approximately
Swine
5 million feral swine in the United States, and
Swine can be colonized with various serotypes they can be found in more than 35 states (195).
of STEC, including O157; however, the risk These animals are highly adaptable to varying
of causing human disease is low (12, 92, 93). environmental conditions and can serve as
The prevalence of STEC O157:H7 in domestic a vector for disease between livestock farms
swine has been reported to range from 0 up to and as a source of contamination of vegetable
10%, with the prevalence in the United States production fields.
usually being less than 2% (12, 92, 94, 95). In the United States, feral swine was first
As in humans, STEC O157:H7 can be highly identified as a reservoir for STEC O157:H7
pathogenic in pigs (92). Unlike ruminants, in 2007 in California (103). In that study,
pigs possess stx-sensitive vascular receptors, STEC O157:H7 was isolated from 14.9% of
and edema disease can develop post intestinal the swine specimens tested, and these isolates
colonization with STEC strains producing were found to be indistinguishable from STEC
Stx2e (8, 96). Stx2e is the most frequent sub- O157:H7 isolates obtained from an outbreak
type of Stx2 found in porcine feces (97). The in the human population associated with the
receptor affinity of Stx2e is different from consumption of spinach. Interestingly, all
Stx1a and Stx2a, since the primary recep- cattle, feral swine, and environmental samples
tor targets are not Gb3 receptors but rather from the region where the spinach was culti-
globotetraosylceramide (Gb4) receptors (98, vated had the same STEC isolate O157 (103).
99). Recently weaned pigs are most suscepti- STEC was also detected in feral swine from
ble to edema disease, and the clinical signs Sweden, Switzerland, and Spain. Approxi-
include subcutaneous and submucosal edema, mately 9% of the tonsil samples screened (n =
ataxia, incoordination, stupor, and recum- 153) in Switzerland were positive for STEC
bency (98, 100). While morbidity in the herd O157, but none of the corresponding fecal
may be low, the case fatality rate for edema samples were positive (104). A similar preva-
disease is high, and surviving pigs may have lence of 8% was reported for Spanish feral
neurologic deficits. swine fecal samples (105). The isolates were

serotyped, and 3.3% of the animals were mean they can serve as a mode of transmission
identified as shedding STEC O157:H7 and of STEC between and within farms. This was
5.2% of the animals as shedding non-O157 demonstrated by Williams et al. (114), who
STEC. reported that starlings and cattle on different
The identification of STEC from feral farms had molecularly indistinguishable sub-
swine samples indicates that they can play a types of STEC O157:H7, thus confirming that
role in the epidemiology of STEC infections. starlings were able to transmit STEC to dif-
As such, their ability to potentially contami- ferent farms. Bolstering the role of starlings in
nate vegetable production fields and serve STEC epidemiology is the fact that the num-
as vectors for STEC transmission between ber of starlings per milking dairy cow was also
livestock must be recognized, and measures found to be significantly associated with the
employed to mitigate this risk presence of STEC O157:H7 in bovine fecal
pats (115).
Migratory birds can also interact with
Birds
peridomestic birds such as pigeons and thus
Birds are capable of harboring many bacterial propagate the transmission of STEC. Pigeons
organisms in their gastrointestinal tract and and finches have been identified as two spe-
are capable of acting as spillover hosts for cies that can potentially serve as a source
STEC. Wild birds were first identified as a of human infection since these birds in-
potential source of STEC infection in 1997 habit buildings, parks, and other recreational
(106). Since then, STEC has been isolated areas and are in close association with the
from starlings, pigeons, sparrows, and other human population (111, 116). Fecal deposi-
avian species (106, 107). Many species of wild tions by these birds in areas frequented by
birds can be found in close proximity to live- humans increase the exposure potential to
stock operations and waste disposal landfill STEC.
sites. These birds are attracted to farms since Water fowl, including geese and ducks, are
they can easily obtain a food source from identified as a source of surface water and
animal feed. Nielsen et al. identified that 2% pasture contamination (107, 117, 118) and im-
of the wild bird fecal samples collected in plicated as the source of numerous food-borne
close proximity to farms contained stx genes pathogens (119122). One goose is reportedly
(108). Similar results were also obtained in capable of producing up to 5 pounds of feces
England where 1.5% of wild bird samples had per day, and this can result in mass contami-
the stx1 gene, 7.9% the stx2 gene, and 4.9% nation, since these birds are usually found
the eae gene (109). Similarly, low prevalence in flocks (123). These birds are also able to
rates of STEC O157:H7 in the starling (Sturnus travel large distances per day and can disperse
vulgaris) population in Ohio and other wild pathogens over a wide area. Geese are known
bird species in Scotland and Japan have also to forage within vegetable fields and also in-
been reported (110112). Though the STEC habit ponds and other surface water sources
prevalence levels are reportedly low, the po- used for irrigation (124). The birds can thus
tential of these birds to transmit STEC to contaminate produce when they defecate
other birds and contaminate the environ- within vegetable fields and irrigation water
ment is of serious risk. Studies have shown sources.
that once colonized, a starling may shed STEC carriage has also been reported in
STEC O157 at levels greater than 100 CFU/g domestic poultry. The prevalence of STEC
of feces for up to 13 days post colonization O157:H7 in domestic chicken is relatively low,
(113). ranging from 0 to 1.5%, depending on the
The migratory pattern of birds and the fact geographic location sampled (92, 94, 125, 126).
they can traverse long distances in a single day Interestingly, the prevalence in turkeys was

higher than that in chickens, with up to capable of shedding STEC O157:H7 for up to
7.5% of fecal samples testing positive (92, 126). 11 days post exposure to high doses of STEC
Experimentally colonized chickens have been (109 CFU) and 5 days post exposure to lower
reported to harbor and shed STEC O157:H7 doses of STEC (105 CFU) (139). This shedding
in their feces for periods in excess of 11 ability indicates that while rats may not be
months (127). Pet birds such as canaries long-term reservoirs of STEC, they are cer-
(Serinus canaria domestica) have also been tainly capable of transmitting STEC between
reported to be capable of harboring and and within farms. This transmission potential
shedding STEC (128). was highlighted by Nielsen et al. (108), who
That both wild and domestic birds are found that STEC recovered from Norwegian
able to harbor, transmit, and shed STEC is of rat fecal pellets was identical to that shed
serious concern since they potentially are a by cattle on the same farm. Contaminated
major risk to human health and disease, and rat feces may also be capable of harboring
as such, precautions should be taken to limit STEC O157:H7 for up to 9 months post inoc-
human or animal exposure to the excrement ulation, thus increasing the risk of transmis-
from these birds. sion to other animals (139). Although rodents,
especially rats, are not regarded as having
a major role in the epidemiology of STEC
Fish and Shellsh
(69, 85, 140), their potential to harbor and
Fish and shellfish can be exposed to STEC transmit STEC exists. Unlike for rats, there is
when their aquatic environment becomes little published information describing the
contaminated with mammalian fecal matter. role of mice in the epidemiology of STEC.
These species do not act as reservoirs of Mice have, however, been used as animal
infection or spillover hosts but rather dead- models to study STEC infection in humans
end hosts. Fish residing in close proximity (141, 142).
or downstream of animal livestock facilities
have also been found to be contaminated
Rabbits
with STEC (129). Shellfish, due to their filter
feeding ability, pose a significant risk to Rabbits have been identified as potential ve-
human infection since they can concentrate hicles for the transmission of both O157 and
and retain pathogens (4, 130). Numerous non-O157 STEC (143145). Rabbits have also
studies have reported the recovery of both been used as a possible animal model to
O157 and non-O157 STEC from the carcass study STEC infection in humans, since they
of fish and shellfish offered for sale (131 demonstrate enteric and renal lesions when
135). The detection of STEC in these carcasses challenged with STEC (146). Globally, con-
highlights the potential for human STEC in- sumption of rabbit meat is increasing; over 1
fection through the consumption of under- million tons of rabbit meat are consumed an-
cooked or raw fish and shellfish. nually (147), thus increasing the potential for
food-borne infection. Rabbits are also popular
at petting zoos, and more than 6 million
Rodents
rabbits are kept as pets in United States (148).
Rodents have also been identified as being Similar to dogs and cats, their close associa-
capable of harboring STEC within their gas- tion with humans may lead to the exchange of
trointestinal flora (108, 136138). STEC O157 microbiota between species and thus STEC
and non-O157 STEC have been recovered transmission. Wild rabbits are able to traverse
from Rattus spp. living in urban areas and long distances and may inhabit both urban
on farms (137, 138). Cizek et al. demonstrated and agricultural areas and potentially serve as
that Norway rats (Rattus novegicus) were vectors for the transmission of STEC from

farm environments to the human population Pets

(144).
Pets, especially dogs and cats, are capable of
shedding a diverse range of STEC serotypes in
Raccoons their feces (163166). Interestingly, although
both O157 and non-O157 STEC have been re-
Raccoons are of particular interest since they
covered from dogs, there are no published
can reside in a wide range of habitats, in-
reports indicating that O157 STEC has ever
cluding agricultural, forested, and urban areas.
been recovered from cat feces. Dogs and cats
Raccoons have been identified as a reservoir
have historically had close interaction with
for numerous pathogens, including Salmonel-
humans, with exchange of microbiota result-
la, Leptospira, and Campylobacter species
ing in the possible transmission of STEC
(149151); however, there is only one report of
between species. These animals can be asymp-
STEC being isolated from raccoon feces. This
tomatic shedders of STEC, as demonstrated
animal had been residing within the hay barn
by Beutin et al. (39), who reported that up
of a dairy farm (152) and thus may be a spill-
to 12% of healthy dogs shed STEC in their
over host. Despite an extensive literature
feces. In addition to household dogs, farm
search, no other reports of STEC raccoon
dogs can be a vector for the transmission of
colonization could be found (153, 154).
STEC. These dogs move freely among animals
and humans, thus potentiating the spread of
Insects enterohemorrhagic E. coli (167). Dogs have
also been reported to shed non-O157 STEC
Insects can be important vectors in the
serotypes in their feces (163). Human in-
transmission and dissemination of STEC in
fections due to canine exposure were also
the environment. STEC O157:H7 has been
reported; one outbreak in Sweden resulted in
recovered from houseflies (Musca domestica),
50 cases in humans after they attended a dog
dump flies (Hydrotaea aenescens), and dung
show (168). STEC has also been recovered
beetles (Catharsius molossus) residing on an-
from the feces of wild canids (169).
imal farms and at animal fairs (155158).
A highly virulent strain of STEC O146:NM
Houseflies, in addition to being a mechanical
has been isolated from the feces of an asymp-
vector, may also be involved in bioenhanced
tomatic cat in Argentina (170). In Germany,
transmission (159). Kobayabashi et al. (159)
there is also evidence of a cat and its owner
suggested this additional role because STEC
shedding the same STEC O146:NM serotype
O15H:H7 could be detected within the ali-
(171). In this case, the source of the infection
mentary tract of inoculated houseflies for at
could not be determined, nor which animal
least 3 days post inoculation. The ability of
was the index case.
houseflies to transmit STEC O157:H7 to
animals was demonstrated by Ahmad et al.
(160), who exposed nave calves to houseflies
Animal Models
inoculated with STEC; within 24 hours, STEC
could be recovered from the feces of all eight Animal models have been used to study the
calves in the experiment (160). Houseflies in vivo pathogenesis of STEC. Numerous
have also been demonstrated to transfer STEC animal models have been developed, includ-
onto the surface of vegetable produce (161). ing mice, rats, chickens, rabbits, cows, grey-
Houseflies are able to travel greater than 4 hounds, baboons, and macaques (141, 173178).
miles (162), and given their ability to trans- Although these models do not fully replicate
mit STEC, one has to be cognizant of the role all aspects of the STEC infection in humans,
they may play in the epidemiology of STEC they provide valuable insight into intestinal
infection. colonization, STEC pathogenesis, immune

response, and efficacy of possible treatment by asymptomatic workers. Approximately 12%

regimens (179, 180). of dairy families in a Canadian study tested
Compared to other animal models, the positive for O157 antibodies, and STEC iso-
mouse models are preferred for in vivo STEC lates were recovered from 6% of the fecal
studies because of their small size, low cost, samples, yet none of these positive cases
ease of care and maintenance, availability of could recall any clinical signs associated with
numbers, and varying genetic backgrounds STEC (186). Similarly, 1.1% of farm workers in
(<141). There are at least four mouse mod- Italy were found to be shedding STEC O157
els, including streptomycin-treated, strepto- asymptomatically in their feces. Contamina-
mycin and mitocyin C/ciprofloxacin-treated, tion of meat carcasses at the abattoir during
intragastric-fed but not streptomycin-treated, slaughter is also a possibility. One study re-
and the malnourished mouse models (181). ported that 1.3% of abattoir workers sampled
The two most popular models are, however, were actively shedding STEC in their feces
the streptomycin-treated mouse and axenic (187). Asymptomatic children can also shed
mice. The models have reduced or no gas- STEC serotypes for up to 30 days post de-
trointestinal flora and are also susceptible to tection, whereas adults in the recent German
STEC or enterohemorrhagic E. coli coloniza- O104:H4 outbreak shed the organism for up to
tion (141). The response of mouse models to 13 weeks (182, 188).
STEC exposure is, however, dependent on the Person-to-person or secondary transmis-
method of infection, the strain of STEC, and sion is important in propagation of outbreaks
the type of mouse model used (181). and can account for 15 to 20% of cases within
outbreaks (184, 189). At particular risk of dis-
ease due to secondary transmission are chil-
Humans
dren (1 to 6 years of age) due to close contact,
Although animals are generally regarded as their immature immune systems, reduced per-
the main reservoir of STEC, humans can also sonal hygiene, and prolonged shedding time
be STEC reservoirs and may play a much (184). An analysis of STEC outbreaks occur-
larger role in the epidemiology of STEC in- ring between 1982 and 2006 as a result of
fections than previously thought (182, 183). person-to-person transmission showed that
Asymptomatic infections in the human popu- 45% of these outbreaks occurred due to trans-
lation can result in dissemination STEC and mission at home, 11% at nurseries, and 10% at
further propagation of outbreaks (182). Given recreational water sources (190).
that food contamination is the source of al-
most 40% of STEC outbreaks and almost 30%
are of unknown origin (184), it is possible that CONCLUSION
contamination by asymptomatic humans may
the source of many of these outbreaks. Most warm-blooded animals are capable of
Human STEC infections can present a acting as reservoirs (symptomatic and asymp-
wide spectrum of clinical signs, ranging from tomatic), spillover hosts, or dead-end hosts of
symptomatic infections to severe clinical STEC. Animals are exposed to STEC by direct
syndromes such as hemorrhagic colitis and or indirect contact with the feces of a shed-
possibly hemolytic-uremic syndrome and ding animal. Cattle are recognized as the main
thrombotic thrombocytopenic purpura in ap- reservoir of STEC, however, and other live-
proximately 7% of the cases. Elderly persons, stock species, including goats, sheep, bison,
young children, and immune-compromised horses, pigs, and water buffalo, have been
persons are at greatest risk (185). demonstrated to be capable of harboring
Food and environmental contamination these organisms. Wild birds and animals pose
with STEC can occur as result of shedding a unique risk in their ability to travel long

distances, increasing the dissemination of virulence and cattle colonization. Virulence 4:

STEC in the environment and thus potenti- 366372.
3. Friedrich AW, Bielaszewska M, Zhang WL,
ating its spread. Domestic pets are also capa- Pulz M, Kuczius T, Ammon A, Karch H. 2002.
ble of harboring STEC, and thus serve as a Escherichia coli harboring Shiga toxin 2 gene
source of contamination within the house- variants: frequency and association with clinical
hold. Given the demonstrated ability of STEC symptoms. J Infect Dis 185:7484.
4. Bennani M, Badri S, Baibai T, Oubrim N,
to colonize the gastrointestinal tract of a
Hassar M, Cohen N, Amarouch H. 2011. First
wide variety of animals, it is expected that detection of Shiga toxin-producing Escherichia
numerous other unreported species may also coli in shellfish and coastal environments of Mo-
be potential sources of contamination or rocco. Appl Biochem Biotechnol 165:290299.
transmission of STEC. Recognition of these 5. Manna SK, Das R, Manna C. 2008. Microbio-
potential novel animal sources of transmis- logical quality of finfish and shellfish with special
reference to shiga toxin-producing Escherichia
sion and propagation of STEC is essential coli O157. J Food Sci 73:M283M286.
when conducting epidemiological investiga- 6. Gourmelon M, Montet MP, Lozach S, Le
tions and developing proper risk mitigation Mennec C, Pommepuy M, Beutin L, Vernozy-
strategies. Despite the widespread carriage Rozand C. 2006. First isolation of Shiga toxin 1d
of STEC by a variety of animal species, it is producing Escherichia coli variant strains in
shellfish from coastal areas in France. J Appl
important to consider that the presence of the Microbiol 100:8597.
stx gene alone is not an indication of patho- 7. Besser TE, Richards BL, Rice DH, Hancock DD.
genicity in the human host. Assessment of the 2001. Escherichia coli O157:H7 infection of calves:
complement of virulence factors present in infectious dose and direct contact transmission.
Epidemiol Infect 127:555560.
STEC recovered from animal hosts is there-
8. Pruimboom-Brees IM, Morgan TW, Ackermann
fore important to develop risk models. The MR, Nystrom ED, Samuel JE, Cornick NA,
understanding of why certain pathoserotypes Moon HW. 2000. Cattle lack vascular receptors
or strains of STEC have a predilection for for Escherichia coli O157:H7 Shiga toxins. Proc Natl
different animal species may provide valuable Acad Sci USA 97:1032510329.
insight in the design of interventions to con- 9. Nguyen Y, Sperandio V. 2012. Enterohemor-
rhagic E. coli (EHEC) pathogenesis. Front Cell
trol the organism in the live animals that have Infect Microbiol 2:90.
impacts on human health. 10. Naylor SW, Low JC, Besser TE, Mahajan A,
Gunn GJ, Pearce MC, McKendrick IJ, Smith
DG, Gally DL. 2003. Lymphoid follicle-dense
ACKNOWLEDGMENTS mucosa at the terminal rectum is the principal
site of colonization of enterohemorrhagic Esche-
We declare no conflicts of interest with regard richia coli O157:H7 in the bovine host. Infect
to the manuscript. Immun 71:15051512.
11. Vanaja SK, Springman AC, Besser TE, Whittam
TS, Manning SD. 2010. Differential expression of
CITATION virulence and stress fitness genes between Esch-
erichia coli O157:H7 strains with clinical or bo-
Persad AK, LeJeune JT. 2014. Animal reser- vine-biased genotypes. Appl Environ Microbiol
voirs of Shiga toxin-producing Escherichia 76:6068.
coli. Microbiol Spectrum 2(4):EHEC-0027- 12. Fairbrother JM, Nadeau E. 2006. Escherichia
2014. coli: on-farm contamination of animals. Rev Sci
Tech 25:555569.
13. Gagliardi JV, Karns JS. 2000. Leaching of
REFERENCES Escherichia coli O157:H7 in diverse soils under
various agricultural management practices. Appl
1. Gyles CL. 2007. Shiga toxin-producing Esche- Environ Microbiol 66:877883.
richia coli: an overview. J Anim Sci 85:E45E62. 14. LeJeune JT, Besser TE, Merrill NL, Rice DH,
2. Etcheverria AI, Padola NL. 2013. Shiga toxin- Hancock DD. 2001. Livestock drinking water
producing Escherichia coli: factors involved in microbiology and the factors influencing the

quality of drinking water offered to cattle. J Dairy Escherichia coli from naturally infected cattle.
Sci 84:18561862. Epidemiol Infect 132:6775.
15. LeJeune JT, Besser TE, Hancock DD. 2001. 28. Lim JY, Li J, Sheng H, Besser TE, Potter K,
Cattle water troughs as reservoirs of Escherichia Hovde CJ. 2007. Escherichia coli O157:H7 colo-
coli O157. Appl Environ Microbiol 67:30533057. nization at the rectoanal junction of long-duration
16. Garber L, Wells S, Schroeder-Tucker L, Ferris culture-positive cattle. Appl Environ Microbiol
K. 1999. Factors associated with fecal shedding 73:13801382.
of verotoxin-producing Escherichia coli O157 on 29. Chase-Topping M, Gally D, Low C, Matthews L,
dairy farms. J Food Prot 62:307312. Woolhouse M. 2008. Super-shedding and the
17. Lejeune JT, Kauffman MD. 2005. Effect of sand link between human infection and livestock car-
and sawdust bedding materials on the fecal riage of Escherichia coli O157. Nat Rev Microbiol
prevalence of Escherichia coli O157:H7 in dairy 6:904912.
cows. Appl Environ Microbiol 71:326330. 30. Omisakin F, MacRae M, Ogden ID, Strachan
18. Cernicchiaro N, Pearl DL, Ghimire S, Gyles CL, NJ. 2003. Concentration and prevalence of
Johnson RP, LeJeune JT, Ziebell K, McEwen Escherichia coli O157 in cattle feces at slaughter.
SA. 2009. Risk factors associated with Escherichia Appl Environ Microbiol 69:24442447.
coli O157:H7 in Ontario beef cow-calf operations. 31. Williams ML, Pearl DL, Bishop KE, Lejeune JT.
Prev Vet Med 92:106115. 2013. Use of multiple-locus variable-number tan-
19. Hancock D, Besser T, Lejeune J, Davis M, Rice dem repeat analysis to evaluate Escherichia coli
D. 2001. The control of VTEC in the animal res- O157 subtype distribution and transmission dy-
ervoir. Int J Food Microbiol 66:7178. namics following natural exposure on a closed
20. Lira WM, Macedo C, Marin JM. 2004. The in- beef feedlot facility. Foodborne Pathog Dis 10:827
cidence of Shiga toxin-producing Escherichia coli 834.
in cattle with mastitis in Brazil. J Appl Microbiol 32. Matthews L, Low JC, Gally DL, Pearce MC,
97:861866. Mellor DJ, Heesterbeek JA, Chase-Topping M,
21. Cerqueira AM, Guth BE, Joaquim RM, Naylor SW, Shaw DJ, Reid SW, Gunn GJ,
Andrade JR. 1999. High occurrence of Shiga Woolhouse ME. 2006. Heterogeneous shedding
toxin-producing Escherichia coli (STEC) in of Escherichia coli O157 in cattle and its implica-
healthy cattle in Rio de Janeiro State, Brazil. Vet tions for control. Proc Natl Acad Sci USA 103:547
Microbiol 70:111121. 552.
22. Farrokh C, Jordan K, Auvray F, Glass K, 33. Nielsen EM, Tegtmeier C, Andersen HJ,
Oppegaard H, Raynaud S, Thevenot D, Gronbaek C, Andersen JS. 2002. Influence of
Condron R, De Reu K, Govaris A, Heggum K, age, sex and herd characteristics on the occur-
Heyndrickx M, Hummerjohann J, Lindsay D, rence of verocytotoxin-producing Escherichia coli
Miszczycha S, Moussiegt S, Verstraete K, O157 in Danish dairy farms. Vet Microbiol 88:245
Cerf O. 2013. Review of Shiga-toxin-producing 257.
Escherichia coli (STEC) and their significance in 34. La Ragione RM, Best A, Woodward MJ, Wales
dairy production. Int J Food Microbiol 162:190 AD. 2009. Escherichia coli O157:H7 colonization
212. in small domestic ruminants. FEMS Microbiol Rev
23. Hancock DD, Besser TE, Kinsel ML, Tarr PI, 33:394410.
Rice DH, Paros MG. 1994. The prevalence of 35. Brandal LT, Sekse C, Lindstedt BA, Sunde M,
Escherichia coli O157.H7 in dairy and beef cattle in Lobersli I, Urdahl AM, Kapperud G. 2012.
Washington state. Epidemiol Infect 113:199207. Norwegian sheep are an important reservoir for
24. Hussein HS, Bollinger LM. 2005. Prevalence of human-pathogenic Escherichia coli O26:H11. Appl
Shiga toxin-producing Escherichia coli in beef Environ Microbiol 78:40834091.
cattle. J Food Prot 68:22242241. 36. Grauke LJ, Kudva IT, Yoon JW, Hunt CW,
25. Hussein HS, Sakuma T. 2005. Prevalence of Williams CJ, Hovde CJ. 2002. Gastrointestinal
shiga toxin-producing Escherichia coli in dairy tract location of Escherichia coli O157:H7 in
cattle and their products. J Dairy Sci 88:450465. ruminants. Appl Environ Microbiol 68:2269
26. Fegan N, Vanderlinde P, Higgs G, Desmarchelier 2277.
P. 2004. The prevalence and concentration of 37. Woodward MJ, Best A, Sprigings KA, Pearson
Escherichia coli O157 in faeces of cattle from dif- GR, Skuse AM, Wales A, Hayes CM, Roe JM,
ferent production systems at slaughter. J Appl Low JC, La Ragione RM. 2003. Non-toxigenic
Microbiol 97:362370. Escherichia coli O157:H7 strain NCTC12900 causes
27. Widiasih DA, Ido N, Omoe K, Sugii S, attaching-effacing lesions and eae-dependent per-
Shinagawa K. 2004. Duration and magnitude sistence in weaned sheep. Int J Med Microbiol
of faecal shedding of Shiga toxin-producing 293:299308.

38. Best A, Clifford D, Crudgington B, Cooley WA, 49. Jacob ME, Foster DM, Rogers AT, Balcomb CC,
Nunez A, Carter B, Weyer U, Woodward MJ, Sanderson MW. 2013. Prevalence and related-
La Ragione RM. 2009. Intermittent Escherichia ness of Escherichia coli O157:H7 strains in the
coli O157:H7 colonisation at the terminal rectum feces and on the hides and carcasses of U.S. meat
mucosa of conventionally-reared lambs. Vet Res goats at slaughter. Appl Environ Microbiol
40:9. 79:41544158.
39. Beutin L, Geier D, Steinruck H, Zimmermann 50. Solecki O, MacRae M, Strachan N, Lindstedt
S, Scheutz F. 1993. Prevalence and some prop- BA, Ogden I. 2009. E. coli O157 from sheep in
erties of verotoxin (Shiga-like toxin)-producing northeast Scotland: prevalence, concentration
Escherichia coli in seven different species of shed, and molecular characterization by multi-
healthy domestic animals. J Clin Microbiol 31: locus variable tandem repeat analysis. Foodborne
24832488. Pathog Dis 6:849854.
40. Cortes C, De la Fuente R, Blanco J, Blanco 51. Milnes AS, Stewart I, Clifton-Hadley FA,
M, Blanco JE, Dhabi G, Mora A, Justel P, Davies RH, Newell DG, Sayers AR, Cheasty T,
Contreras A, Sanchez A, Corrales JC, Orden Cassar C, Ridley A, Cook AJ, Evans SJ, Teale
JA. 2005. Serotypes, virulence genes and intimin CJ, Smith RP, McNally A, Toszeghy M, Futter
types of verotoxin-producing Escherichia coli and R, Kay A, Paiba GA. 2008. Intestinal carriage of
enteropathogenic E. coli isolated from healthy verocytotoxigenic Escherichia coli O157, Salmo-
dairy goats in Spain. Vet Microbiol 110:6776. nella, thermophilic Campylobacter and Yersinia
41. Heuvelink AE, van Heerwaarden C, Zwartkruis- enterocolitica, in cattle, sheep and pigs at slaugh-
Nahuis JT, van Oosterom R, Edink K, van ter in Great Britain during 2003. Epidemiol Infect
Duynhoven YT, de Boer E. 2002. Escherichia coli 136:739751.
O157 infection associated with a petting zoo. 52. Heuvelink AE, van den Biggelaar FL, de Boer E,
Epidemiol Infect 129:295302. Herbes RG, Melchers WJ, Huis in 't Veld JH,
42. LeJeune JT, Davis MA. 2004. Outbreaks of Monnens LA. 1998. Isolation and characteriza-
zoonotic enteric disease associated with animal tion of verocytotoxin-producing Escherichia coli
exhibits. J Am Vet Med Assoc 224:14401445. O157 strains from Dutch cattle and sheep. J Clin
43. Schilling AK, Hotzel H, Methner U, Sprague Microbiol 36:878882.
LD, Schmoock G, El-Adawy H, Ehricht R, Wohr 53. Vu-Khac H, Cornick NA. 2008. Prevalence and
AC, Erhard M, Geue L. 2012. Zoonotic agents in genetic profiles of Shiga toxin-producing Esche-
small ruminants kept on city farms in southern richia coli strains isolated from buffaloes, cattle,
Germany. Appl Environ Microbiol 78:37853793. and goats in central Vietnam. Vet Microbiol
44. Espie E, Vaillant V, Mariani-Kurkdjian P, 126:356363.
Grimont F, Martin-Schaller R, De Valk H, 54. Islam MA, Mondol AS, de Boer E, Beumer RR,
Vernozy-Rozand C. 2006. Escherichia coli O157 Zwietering MH, Talukder KA, Heuvelink AE.
outbreak associated with fresh unpasteurized 2008. Prevalence and genetic characterization of
goats cheese. Epidemiol Infect 134:143146. shiga toxin-producing Escherichia coli isolates
45. Duffy LL, Small A, Fegan N. 2010. Concentration from slaughtered animals in Bangladesh. Appl
and prevalence of Escherichia coli O157 and Sal- Environ Microbiol 74:54145421.
monella serotypes in sheep during slaughter at 55. Effler E, Isaacson M, Arntzen L, Heenan R,
two Australian abattoirs. Aust Vet J 88:399404. Canter P, Barrett T, Lee L, Mambo C, Levine
46. Djordjevic SP, Hornitzky MA, Bailey G, Gill P, W, Zaidi A, Griffin PM. 2001. Factors contrib-
Vanselow B, Walker K, Bettelheim KA. 2001. uting to the emergence of Escherichia coli O157 in
Virulence properties and serotypes of Shiga Africa. Emerg Infect Dis 7:812819.
toxin-producing Escherichia coli from healthy 56. Cunin P, Tedjouka E, Germani Y, Ncharre C,
Australian slaughter-age sheep. J Clin Microbiol Bercion R, Morvan J, Martin P. 1999. An epi-
39:20172021. demic of bloody diarrhea: Escherichia coli O157
47. Oporto B, Esteban JI, Aduriz G, Juste RA, emerging in Cameroon? Emerg Infect Dis 5:285.
Hurtado A. 2008. Escherichia coli O157:H7 and 57. Battisti A, Lovari S, Franco A, Di Egidio A,
non-O157 Shiga toxin-producing E. coli in healthy Tozzoli R, Caprioli A, Morabito S. 2006. Prev-
cattle, sheep and swine herds in Northern Spain. alence of Escherichia coli O157 in lambs at
Zoonoses Public Health 55:7381. slaughter in Rome, central Italy. Epidemiol Infect
48. Jacob ME, Foster DM, Rogers AT, Balcomb CC, 134:415419.
Shi X, Nagaraja TG. 2013. Evidence of non-O157 58. Franco A, Lovari S, Cordaro G, Di Matteo P,
Shiga toxin-Producing Escherichia coli in the Sorbara L, Iurescia M, Donati V, Buccella C,
feces of meat goats at a U.S. slaughter plant. Battisti A. 2009. Prevalence and concentration of
J Food Prot 76:16261629. verotoxigenic Escherichia coli O157:H7 in adult

sheep at slaughter from Italy. Zoonoses Public toxin-producing Escherichia coli associated with
Health 56:215220. venison. Emerg Infect Dis 18:279282.
59. Orden JA, Cortes C, Horcajo P, De la Fuente 72. Laidler MR, Tourdjman M, Buser GL, Hostetler
R, Blanco JE, Mora A, Lopez C, Blanco J, T, Repp KK, Leman R, Samadpour M, Keene
Contreras A, Sanchez A, Corrales JC, WE. 2013. Escherichia coli O157:H7 infections
Dominguez-Bernal G. 2008. A longitudinal study associated with consumption of locally grown
of verotoxin-producing Escherichia coli in two strawberries contaminated by deer. Clin Infect Dis
dairy goat herds. Vet Microbiol 132:428434. 57:11291134.
60. Kudva IT, Hatfield PG, Hovde CJ. 1996. Esche- 73. Franklin AB, Vercauteren KC, Maguire H,
richia coli O157:H7 in microbial flora of sheep. Cichon MK, Fischer JW, Lavelle MJ, Powell A,
J Clin Microbiol 34:431433. Root JJ, Scallan E. 2013. Wild ungulates as
61. Galiero G, Conedera G, Alfano D, Caprioli A. disseminators of Shiga toxin-producing Esche-
2005. Isolation of verocytotoxin-producing Esch- richia coli in urban areas. PloS One 8:e81512.
erichia coli O157 from water buffaloes (Bubalus 74. Kudva IT, Stasko JA. 2013. Bison and bovine
bubalis) in southern Italy. Vet Rec 156:382383. rectoanal junctions exhibit similar cellular archi-
62. Sargeant JM, Hafer DJ, Gillespie JR, Oberst tecture and Escherichia coli O157 adherence
RD, Flood SJ. 1999. Prevalence of Escherichia coli patterns. BMC Vet Res 9:266.
O157:H7 in white-tailed deer sharing rangeland 75. Johnson K. 2011. Plains giants have foothold on
with cattle. J Am Vet Med Assoc 215:792794. tables. The New York Times, Jan. 23:A16. http://
63. Dunn JR, Keen JE, Moreland D, Alex T. 2004. www.nytimes.com/2011/01/23/us/23buffalo.
Prevalence of Escherichia coli O157:H7 in white- html?_r=0.
tailed deer from Louisiana. J Wildl Dis 40:361 76. Reinstein S, Fox JT, Shi X, Alam MJ, Nagaraja
365. TG. 2007. Prevalence of Escherichia coli O157:H7
64. French E, Rodriguez-Palacios A, LeJeune JT. in the American bison (Bison bison). J Food Prot
2010. Enteric bacterial pathogens with zoonotic 70:25552560.
potential isolated from farm-raised deer. Foodborne 77. Li Q, Sherwood JS, Logue CM. 2004. The
Pathog Dis 7:10311037. prevalence of Listeria, Salmonella, Escherichia coli
65. Bardiau M, Gregoire F, Muylaert A, Nahayo A, and E. coli O157:H7 on bison carcasses during
Duprez JN, Mainil J, Linden A. 2010. Entero- processing. Food Microbiol 21:791799.
pathogenic (EPEC), enterohaemorragic (EHEC) 78. Magwedere K, Dang HA, Mills EW, Cutter CN,
and verotoxigenic (VTEC) Escherichia coli in wild Roberts EL, Debroy C. 2013. Incidence of Shiga
cervids. J Appl Microbiol 109:22142222. toxin-producing Escherichia coli strains in beef,
66. Sanchez S, Garcia-Sanchez A, Martinez R, pork, chicken, deer, boar, bison, and rabbit retail
Blanco J, Blanco JE, Blanco M, Dahbi G, meat. J Vet Diagn Invest 25:254258.
Mora A, Hermoso de Mendoza J, Alonso JM, 79. Mercado EC, Rodriguez SM, Elizondo AM,
Rey J. 2009. Detection and characterisation of Marcoppido G, Parreo V. 2004. Isolation of
Shiga toxin-producing Escherichia coli other than Shiga toxin-producing Escherichia coli from a
Escherichia coli O157:H7 in wild ruminants. Vet J South American camelid (Lama guanicoe) with
180:384388. diarrhea. J Clin Microbiol 42:48094811.
67. Kistler WM, Mulugeta S, Mauro SA. 2011. De- 80. Gilbreath JJ, Shields MS, Smith RL, Farrell LD,
tection of stx and stx genes in Pennsylvanian Sheridan PP, Spiegel KM. 2009. Shiga toxins,
white-tailed deer. Toxins (Basel) 3:640646. and the genes encoding them, in fecal samples
68. Renter DG, Sargeant JM, Hygnstorm SE, from native Idaho ungulates. Appl Environ Micro-
Hoffman JD, Gillespie JR. 2001. Escherichia coli biol 75:862865.
O157:H7 in free-ranging deer in Nebraska. J Wildl 81. Cid D, Martn-Espada C, Maturrano L, Garca
Dis 37:755760. A, Luna L, Rosadio R. 2012. Diarrheagenic
69. Rice DH, Hancock DD, Besser TE. 2003. Faecal Escherichia coli strains isolated from neonatal
culture of wild animals for Escherichia coli O157: Peruvian alpacas (Vicugnapacos) with diarrhea,
H7. Vet Rec 152:8283. p 223228. In Prez-Cabal MA, Gutirrez JP,
70. Keene WE, Sazie E, Kok J, Rice DH, Hancock Cervantes I, Alcalde MJ (ed), Fibre Production
DD, Balan VK, Zhao T, Doyle MP. 1997. An in South American Camelids and Other Fibre
outbreak of Escherichia coli O157:H7 infections Animals. Wageningen Academic Publishers,
traced to jerky made from deer meat. JAMA Wageningen, The Netherlands.
277:12291231. 82. Mohammed Hamzah A, Mohammed Hussein A,
71. Rounds JM, Rigdon CE, Muhl LJ, Forstner M, Mahmoud Khalef J. 2013. Isolation of Escherichia
Danzeisen GT, Koziol BS, Taylor C, Shaw BT, coli 0157:H7 strain from fecal samples of zoo ani-
Short GL, Smith KE. 2012. Non-O157 Shiga mal. Scientific World Journal 2013:843968.

83. Leotta GA, Deza N, Origlia J, Toma C, Chinen I, 95. Richards HA, Perez-Conesa D, Doane CA,
Miliwebsky E, Iyoda S, Sosa-Estani S, Rivas M. Gillespie BE, Mount JR, Oliver SP, Pangloli P,
2006. Detection and characterization of Shiga Draughon FA. 2006. Genetic characterization of
toxin-producing Escherichia coli in captive non- a diverse Escherichia coli O157:H7 population
domestic mammals. Vet Microbiol 118:151157. from a variety of farm environments. Foodborne
84. Chandran A, Mazumder A. 2013. Prevalence of Pathog Dis 3:259265.
diarrhea-associated virulence genes and genetic 96. Waddell TE, Coomber BL, Gyles CL. 1998. Lo-
diversity in Escherichia coli isolates from fecal calization of potential binding sites for the edema
material of various animal hosts. Appl Environ disease verotoxin (VT2e) in pigs. Can J Vet Res
Microbiol 79:73717380. 62:8186.
85. Hancock DD, Besser TE, Rice DH, Ebel ED, 97. Fratamico PM, Bagi LK, Bush EJ, Solow BT.
Herriott DE, Carpenter LV. 1998. Multiple 2004. Prevalence and characterization of shiga
sources of Escherichia coli O157 in feedlots and toxin-producing Escherichia coli in swine feces
dairy farms in the northwestern USA. Prev Vet recovered in the National Animal Health Moni-
Med 35:1119. toring Systems Swine 2000 study. Appl Environ
86. Larson DG. 2009. Prevalence of Shiga toxin-pro- Microbiol 70:71737178.
ducing Escherichia coli in the Sacramento Valley 98. Imberechts H, De Greve H, Lintermans P. 1992.
equine population. M.S. thesis, California State The pathogenesis of edema disease in pigs. A re-
University, Sacramento. view. Vet Microbiol 31:221233.
87. Lengacher B, Kline TR, Harpster L, Williams 99. Mthing J, Meisen I, Zhang W, Bielaszewska
ML, Lejeune JT. 2010. Low prevalence of Esch- M, Mormann M, Bauerfeind R, Schmidt MA,
erichia coli O157:H7 in horses in Ohio, USA. J Friedrich AW, Karch H. 2012. Promiscuous
Food Prot 73:20892092. Shiga toxin 2e and its intimate relationship to
88. Pichner R, Sander A, Steinruck H, Gareis M. Forssman. Glycobiology 22:849862.
2005. [Occurrence of Salmonella spp. and 100. Meisen I, Rosenbrck R, Galla HJ, Hwel S,
shigatoxin-producing Escherichia coli (STEC) in Kouzel IU, Mormann M, Karch H, Mthing J.
horse faeces and horse meat products]. Berl 2013. Expression of Shiga toxin 2e glycosphingo-
Munch Tierarztl Wochenschr 118:321325 (In lipid receptors of primary porcine brain endo-
German). thelial cells and toxin-mediated breakdown of
89. Pritchard GC, Smith R, Ellis-Iversen J, Cheasty the blood-brain barrier. Glycobiology 23:745759.
T, Willshaw GA. 2009. Verocytotoxigenic Esch- 101. Booher SL, Cornick NA, Moon HW. 2002.
erichia coli O157 in animals on public amenity Persistence of Escherichia coli O157:H7 in ex-
premises in England and Wales, 1997 to 2007. Vet perimentally infected swine. Vet Microbiol 89:
Rec 164:545549. 6981.
90. Momtaz H, Farzan R, Rahimi E, Safarpoor 102. Trotz-Williams LA, Mercer NJ, Walters JM,
Dehkordi F, Souod N. 2012. Molecular charac- Maki AM, Johnson RP. 2012. Pork implicated in
terization of Shiga toxin-producing Escherichia a Shiga toxin-producing Escherichia coli O157:H7
coli isolated from ruminant and donkey raw milk outbreak in Ontario, Canada. Can J Public Health
samples and traditional dairy products in Iran. 103:e622e326.
Scientific World Journal 2012:231341. 103. Jay MT, Cooley M, Carychao D, Wiscomb GW,
91. Chalmers RM, Salmon RL, Willshaw GA, Sweitzer RA, Crawford-Miksza L, Farrar JA,
Cheasty T, Looker N, Davies I, Wray C. 1997. Lau DK, OConnell J, Millington A, Asmundson
Vero-cytotoxin-producing Escherichia coli O157 in RV, Atwill ER, Mandrell RE. 2007. Escherichia
a farmer handling horses. Lancet 349:1816. coli O157:H7 in feral swine near spinach fields and
92. Ferens WA, Hovde CJ. 2011. Escherichia coli cattle, central California coast. Emerg Infect Dis
O157:H7: animal reservoir and sources of human 13:19081911.
infection. Foodborne Pathog Dis 8:465487. 104. Wacheck S, Fredriksson-Ahomaa M, Konig
93. Kaufmann M, Zweifel C, Blanco M, Blanco JE, M, Stolle A, Stephan R. 2010. Wild boars as an
Blanco J, Beutin L, Stephan R. 2006. Escherichia important reservoir for foodborne pathogens.
coli O157 and non-O157 Shiga toxin-producing Foodborne Pathog Dis 7:307312.
Escherichia coli in fecal samples of finished pigs 105. Sanchez S, Martinez R, Garcia A, Vidal D,
at slaughter in Switzerland. J Food Prot 69:260 Blanco J, Blanco M, Blanco JE, Mora A,
266. Herrera-Leon S, Echeita A, Alonso JM, Rey J.
94. Chapman PA, Siddons CA, Gerdan Malo AT, 2010. Detection and characterisation of O157:H7
Harkin MA. 1997. A 1-year study of Escherichia and non-O157 Shiga toxin-producing Esche-
coli O157 in cattle, sheep, pigs and poultry. richia coli in wild boars. Vet Microbiol 143:420
Epidemiol Infect 119:245250. 423.

106. Wallace JS, Cheasty T, Jones K. 1997. Isolation and characterization of Shiga toxin-producing
of vero cytotoxin-producing Escherichia coli Escherichia coli in feral pigeons. Vet Microbiol
O157 from wild birds. J Appl Microbiol 82:399 82:275283.
404. 117. Moriarty EM, Weaver L, Sinton LW, Gilpin B.
107. Pedersen K, Clark L. 2007. A review of Shiga 2012. Survival of Escherichia coli, enterococci and
toxin Escherichia coli and Salmonella enterica in Campylobacter jejuni in Canada goose faeces on
cattle and free-ranging birds: potential association pasture. Zoonoses Public Health 59:490497.
and epidemiological links. HumanWildlife 118. Kirschner AK, Zechmeister TC, Kavka GG,
Conflicts 1:6877. Beiwl C, Herzig A, Mach RL, Farnleitner AH.
108. Nielsen EM, Skov MN, Madsen JJ, Lodal J, 2004. Integral strategy for evaluation of fecal
Jespersen JB, Baggesen DL. 2004. Verocytotoxin- indicator performance in bird-influenced saline
producing Escherichia coli in wild birds and inland waters. Appl Environ Microbiol 70:7396
rodents in close proximity to farms. Appl Environ 7403.
Microbiol 70:69446947. 119. Kullas H, Coles M, Rhyan J, Clark L. 2002.
109. Hughes LA, Bennett M, Coffey P, Elliott J, Prevalence of Escherichia coli serogroups and
Jones TR, Jones RC, Lahuerta-Marin A, human virulence factors in faeces of urban
McNiffe K, Norman D, Williams NJ, Chantrey Canada geese (Branta canadensis). Int J Environ
J. 2009. Risk factors for the occurrence of Esch- Health Res 12:153162.
erichia coli virulence genes eae, stx1 and stx2 in 120. Murinda SE, Nguyen LT, Nam HM, Almeida
wild bird populations. Epidemiol Infect 137:1574 RA, Headrick SJ, Oliver SP. 2004. Detection of
1582. sorbitol-negative and sorbitol-positive Shiga
110. LeJeune J, Homan J, Linz G, Pearl DL. 2008. toxin-producing Escherichia coli, Listeria mono-
Role of the European starling in the transmission cytogenes, Campylobacter jejuni, and Salmo-
of E. coli O157 on dairy farms, p 3134. In Timm nella spp. in dairy farm environmental samples.
RM, Madon MB (ed), Proceedings of the Twenty- Foodborne Pathog Dis 1:97104.
Third Vertebrate Pest Conference. University of 121. Rutledge ME, Siletzky RM, Gu W, Degernes
California, Davis, CA. LA, Moorman CE, DePerno CS, Kathariou S.
111. Foster G, Evans J, Knight HI, Smith AW, 2013. Characterization of Campylobacter from
Gunn GJ, Allison LJ, Synge BA, Pennycott TW. resident Canada geese in an urban environment. J
2006. Analysis of feces samples collected from a Wildl Dis 49:19.
wild-bird garden feeding station in Scotland for 122. Siembieda JL, Miller WA, Byrne BA, Ziccardi
the presence of verocytotoxin-producing Esche- MH, Anderson N, Chouicha N, Sandrock CE,
richia coli O157. Appl Environ Microbiol 72:2265 Johnson CK. 2011. Zoonotic pathogens isolated
2267. from wild animals and environmental samples at
112. Kobayashi H, Kanazaki M, Hata E, Kubo M. two California wildlife hospitals. J Am Vet Med
2009. Prevalence and characteristics of eae- and Assoc 238:773783.
stx-positive strains of Escherichia coli from wild 123. Bedard J, Gauthier G. 1986. Assessment of faecal
birds in the immediate environment of Tokyo output in geese. J Appl Ecol 23:7790.
Bay. Appl Environ Microbiol 75:292295. 124. Ankney CD. 1996. An embarrassment of riches:
113. Kauffman MD, LeJeune J. 2011. European too many geese. J Wildl Managet 60:217223.
starlings (Sturnus vulgaris) challenged with 125. Doyle MP, Schoeni JL. 1987. Isolation of Esche-
Escherichia coli O157 can carry and transmit the richia coli O157:H7 from retail fresh meats and
human pathogen to cattle. Lett Appl Microbiol poultry. Appl Environ Microbiol 53:23942396.
53:596601. 126. Doane CA, Pangloli P, Richards HA, Mount JR,
114. Williams ML, Pearl DL, Lejeune JT. 2011. Golden DA, Draughon FA. 2007. Occurrence of
Multiple-locus variable-nucleotide tandem repeat Escherichia coli O157:H7 in diverse farm envi-
subtype analysis implicates European starlings as ronments. J Food Prot 70:610.
biological vectors for Escherichia coli O157:H7 in 127. Schoeni JL, Doyle MP. 1994. Variable coloniza-
Ohio, USA. J Appl Microbiol 111:982988. tion of chickens perorally inoculated with Esche-
115. Cernicchiaro N, Pearl DL, McEwen SA, richia coli O157:H7 and subsequent contamination
Harpster L, Homan HJ, Linz GM, Lejeune JT. of eggs. Appl Environ Microbiol 60:29582962.
2012. Association of wild bird density and farm 128. Gholami-Ahangaran M, Zia-Jahromi N. 2012.
management factors with the prevalence of E. coli Identification of Shiga toxin and intimin genes in
O157 in dairy herds in Ohio (20072009). Zoo- Escherichia coli detected from canary (Serinus
noses Public Health 59:320329. canaria domestica). Toxicol Ind Health Oct 9.
116. Morabito S, DellOmo G, Agrimi U, Schmidt H, 129. Tuyet DT, Yassibanda S, Nguyen Thi PL,
Karch H, Cheasty T, Caprioli A. 2001. Detection Koyenede MR, Gouali M, Bekondi C, Mazzi J,

Germani Y. 2006. Enteropathogenic Escherichia 141. Mohawk KL, OBrien AD. 2011. Mouse models of
coli o157 in Bangui and NGoila, Central African Escherichia coli O157:H7 infection and Shiga toxin
Republic: a brief report. Am J Trop Med Hyg injection. J Biomed Biotechnol 2011:258185.
75:513515. 142. Wadolkowski EA, Burris JA, OBrien AD. 1990.
130. Guyon R, Dorey F, Collobert JF, Foret J, Mouse model for colonization and disease caused
Goubert C, Mariau V, Malas JP. 2000. Detection by enterohemorrhagic Escherichia coli O157: H7.
of Shiga toxin-producing Escherichia coli O157 Infect Immun 58:24382445.
in shellfish (Crassostrea gigas). Sciences des 143. Pritchard GC, Williamson S, Carson T, Bailey
Aliments 20:457466. JR, Warner L, Willshaw G, Cheasty T. 2001.
131. Surendraraj A, Thampuran N, Joseph TC. 2010. Wild rabbitsa novel vector for verocytotoxi-
Molecular screening, isolation, and characteriza- genic Escherichia coli O157. Vet Rec 149:567.
tion of enterohemorrhagic Escherichia coli O157: 144. Scaife HR, Cowan D, Finney J, Kinghorn-Perry
H7 from retail shrimp. J Food Prot 73:97103. SF, Crook B. 2006. Wild rabbits (Oryctolagus
132. Sanath Kumar H, Otta SK, Karunasagar I, cuniculus) as potential carriers of verocytotoxin-
Karunasagar I. 2001. Detection of Shiga- producing Escherichia coli. Vet Rec 159:175178.
toxigenic Escherichia coli (STEC) in fresh sea- 145. Garcia A, Fox JG. 2003. The rabbit as a new
food and meat marketed in Mangalore, India reservoir host of enterohemorrhagic Escherichia
by PCR. Lett Appl Microbiol 33:334338. coli. Emerg Infect Dis 9:15921597.
133. Thampuran N, Surendraraj A, Surendran PK. 146. Panda A, Tatarov I, Melton-Celsa AR,
2005. Prevalence and characterization of typical Kolappaswamy K, Kriel EH, Petkov D,
and atypical Escherichia coli from fish sold at re- Coksaygan T, Livio S, McLeod CG, Nataro JP,
tail in Cochin, India. J Food Prot 68:22082211. OBrien AD, DeTolla LJ. 2010. Escherichia coli
134. Gupta B, Ghatak S, Gill JPS. 2013. Incidence and O157:H7 infection in Dutch belted and New
virulence properties of E. coli isolated from fresh Zealand white rabbits. Comp Med 60:3137.
fish and ready-to-eat fish products. Vet World 147. Dalle Zotte A, Szendr Z. 2011. The role of rabbit
6:59. meat as functional food. Meat Sci 88:319331.
135. Gourmelon M, Montet MP, Lozach S, Le 148. Cook AJ, McCobb E. 2012. Quantifying the
Mennec C, Pommepuy M, Beutin L, Vernozy- shelter rabbit population: an analysis of Massa-
Rozand C. 2006. First isolation of Shiga toxin 1d chusetts and Rhode Island animal shelters. J Appl
producing Escherichia coli variant strains in Anim WelfSci 15:297312.
shellfish from coastal areas in France. J Appl 149. Compton JA, Baney JA, Donaldson SC, Houser
Microbiol 100:8597. BA, San Julian GJ, Yahner RH, Chmielecki W,
136. Kilonzo C, Li X, Vivas EJ, Jay-Russell MT, Reynolds S, Jayarao BM. 2008. Salmonella in-
Fernandez KL, Atwill ER. 2013. Fecal shedding of fections in the common raccoon (Procyon lotor) in
zoonotic food-borne pathogens by wild rodents in a western Pennsylvania. J Clin Microbiol 46:3084
major agricultural region of the central California 3086.
coast. Appl Environ Microbiol 79:63376344. 150. Lee K, Iwata T, Nakadai A, Kato T, Hayama S,
137. Cizek A, Alexa P, Literak I, Hamrik J, Novak P, Taniguchi T, Hayashidani H. 2011. Prevalence of
Smola J. 1999. Shiga toxin-producing Escherichia Salmonella, Yersinia and Campylobacter spp. in
coli O157 in feedlot cattle and Norwegian rats feral raccoons (Procyonlotor) and masked palm
from a large-scale farm. Lett Appl Microbiol civets (Paguma larvata) in Japan. Zoonoses Public
28:435439. Health 58:424431.
138. Blanco Crivelli X, Rumi MV, Carfagnini JC, 151. Beltrn-Beck B, Garca F, Gortzar C. 2012.
Degregorio O, Bentancor AB. 2012. Synanthropic Raccoons in Europe: disease hazards due to the
rodents as possible reservoirs of shigatoxigenic establishment of an invasive species. Eur J Wildl
Escherichia coli strains. Front Cell Infect Microbiol Res 58:515.
2:134. 152. Shere JA, Bartlett KJ, Kaspar CW. 1998. Lon-
139. Cizek A, Literak I, Scheer P. 2000. Survival of gitudinal study of Escherichia coli O157:H7 dis-
Escherichia coli O157 in faeces of experimentally semination on four dairy farms in Wisconsin.
infected rats and domestic pigeons. Lett Appl Appl Environ Microbiol 64:13901399.
Microbiol 31:349352. 153. Renter DG, Sargeant JM, Oberst RD,
140. Nkogwe C, Raletobana J, Stewart-Johnson A, Samadpour M. 2003. Diversity, frequency,
Suepaul S, Adesiyun A. 2011. Frequency of de- and persistence of Escherichia coli O157 strains
tection of Escherichia coli, Salmonella spp., and from range cattle environments. Appl Environ
Campylobacter spp. in the faeces of wild rats Microbiol 69:542547.
(Rattus spp.) in Trinidad and Tobago. Vet Med Int 154. Jay-Russell M, Atwill E, Cooley M, Carychao D,
2011:686923. Vivas E, Chandler S, Orthmeyer D, Lii X,

Mandrell R. 2010, p 2327. Occurrence of Esch- coli strains isolated from dogs and cats and ge-
erichia coli O157:H7 in wildlife in a major produce netic relationships with isolates from cattle, meat
production region in California. Poster. 110th Gen and humans. Vet Microbiol 156:336342.
Meet Am Soc for Microbiol. American Society for 168. Anonymous. 2011. Dog show sparks new Swedish
Microbiology, Washington, DC. EHEC outbreak, The Local, June 16 http://www.
155. Alam MJ, Zurek L. 2004. Association of Esche- thelocal.se/20110616/34384.
richia coli O157:H7 with houseflies on a cattle 169. Mora A, Lopez C, Dhabi G, Lopez-Beceiro AM,
farm. Appl Environ Microbiol 70:75787580. Fidalgo LE, Diaz EA, Martinez-Carrasco C,
156. Keen JE, Wittum TE, Dunn JR, Bono JL, Durso Mamani R, Herrera A, Blanco JE, Blanco M,
LM. 2006. Shiga-toxigenic Escherichia coli O157 Blanco J. 2012. Seropathotypes, phylogroups,
in agricultural fair livestock, United States. Emerg Stx subtypes, and intimin types of wildlife-
Infect Dis 12:780786. carried, Shiga toxin-producing Escherichia coli
157. Xu J, Liu Q, Jing H, Pang B, Yang J, Zhao G, strains with the same characteristics as human-
Li H. 2003. Isolation of Escherichia coli O157: H7 pathogenic isolates. Appl Environ Microbiol 78:
from dung beetles Catharsius molossus. Microbiol 25782585.
Immunol 47:4549. 170. Rumi MV, Irino K, Deza N, Huguet MJ,
158. Szalanski A, Owens C, McKay T, Steelman C. Bentancor AB. 2012. First isolation in Argentina
2004. Detection of Campylobacter and Escherichia of a highly virulent Shiga toxin-producing Esch-
coli O157: H7 from filth flies by polymerase chain erichia coli O145:NM from a domestic cat. J Infect
reaction. Med Vet Entomol 18:241246. Dev Ctries 6:358363.
159. Kobayashi M, Sasaki T, Saito N, Tamura K, 171. Busch U, Hormansdorfer S, Schranner S, Huber
Suzuki K, Watanabe H, Agui N. 1999. Houseflies: I, Bogner KH, Sing A. 2007. Enterohemorrhagic
not simple mechanical vectors of enterohemor- Escherichia coli excretion by child and her cat.
rhagic Escherichia coli O157:H7. Am J Trop Med Emerg Infect Dis 13:348349.
Hyg 61:625629. 172. Woods JB, Schmitt CK, Darnell SC, Meysick
160. Ahmad A, Nagaraja TG, Zurek L. 2007. Trans- KC, OBrien AD. 2002. Ferrets as a model system
mission of Escherichia coli O157:H7 to cattle by for renal disease secondary to intestinal infection
house flies. Prev Vet Med 80:7481. with Escherichia coli O157:H7 and other Shiga
161. Wasala L, Talley JL, Desilva U, Fletcher J, toxin-producing E. coli. J Infect Dis 185:550554.
Wayadande A. 2013. Transfer of Escherichia coli 173. Zotta E, Lago N, Ochoa F, Repetto HA, Ibarra
O157:H7 to spinach by house flies, Musca domes- C. 2008. Development of an experimental he-
tica (Diptera: Muscidae). Phytopathology 103:373 molytic uremic syndrome in rats. Pediatr Nephrol
380. 23:559567.
162. Pickens L, Morgan N, Hartsock J, Smith J. 1967. 174. Garcia A, Bosques CJ, Wishnok JS, Feng Y,
Dispersal patterns and populations of the house Karalius BJ, Butterton JR, Schauer DB, Rogers
fly affected by sanitation and weather in rural AB, Fox JG. 2006. Renal injury is a consistent
Maryland. J Econ Entomol 60:12501255. finding in Dutch belted rabbits experimentally
163. Roopnarine RR, Ammons D, Rampersad J, infected with enterohemorrhagic Escherichia coli.
Adesiyun AA. 2007. Occurrence and characteriza- J Infect Dis 193:11251134.
tion of verocytotoxigenic Escherichia coli (VTEC) 175. Sueyoshi M, Nakazawa M. 1994. Experimental
strains from dairy farms in Trinidad. Zoonoses infection of young chicks with attaching and ef-
Public Health 54:7885. facing Escherichia coli. Infect Immun 62:4066
164. LeJeune JT, Hancock DD. 2001. Public health 4071.
concerns associated with feeding raw meat diets 176. Tzipori S, Wachsmuth IK, Chapman C, Birden
to dogs. J Am Vet Med Assoc 219:12221225. R, Brittingham J, Jackson C, Hogg J. 1986. The
165. Hogg RA, Holmes JP, Ghebrehewet S, Elders K, pathogenesis of hemorrhagic colitis caused by
Hart J, Whiteside C, Willshaw GA, Cheasty T, Escherichia coli O157:H7 in gnotobiotic piglets. J
Kay A, Lynch K, Pritchard GC. 2009. Proba- Infect Dis 154:712716.
ble zoonotic transmission of verocytotoxigenic 177. Taylor FB Jr, Tesh VL, DeBault L, Li A, Chang
Escherichia coli O 157 by dogs. Vet Rec 164:304 AC, Kosanke SD, Pysher TJ, Siegler RL. 1999.
305. Characterization of the baboon responses to
166. Beutin L. 1999. Escherichia coli as a pathogen in Shiga-like toxin: descriptive study of a new pri-
dogs and cats. Vet Res 30:285298. mate model of toxic responses to Stx-1. Am J
167. Bentancor A, Rumi MV, Carbonari C, Gerhardt Pathol 154:12851299.
E, Larzabal M, Vilte DA, Pistone-Creydt V, 178. Kang G, Pulimood AB, Koshi R, Hull A,
Chinen I, Ibarra C, Cataldi A, Mercado EC. Acheson D, Rajan P, Keusch GT, Mathan VI,
2012. Profile of Shiga toxin-producing Escherichia Mathan MM. 2001. A monkey model for

enterohemorrhagic Escherichia coli infection. J Orrbine E, Spika JS. 1997. Canadian perspectives
Infect Dis 184:206210. on verocytotoxin-producing Escherichia coli in-
179. Mohawk KL, Melton-Celsa AR, Zangari T, fection. J Food Prot 60:14511453.
Carroll EE, OBrien AD. 2010. Pathogenesis of 188. Miliwebsky E, Deza N, Chinen I, Martinez
Escherichia coli O157:H7 strain 86-24 following Espinosa E, Gomez D, Pedroni E, Caprile L,
oral infection of BALB/c mice with an intact Bashckier A, Manfredi E, Leotta G. 2007.
commensal flora. Microb Pathog 48:131142. Prolonged fecal shedding of Shiga toxin-produc-
180. Mayer CL, Leibowitz CS, Kurosawa S, Stearns- ing Escherichia coli among children attending
Kurosawa DJ. 2012. Shiga toxins and the path- day-care centers in Argentina. Rev Argent
ophysiology of hemolytic uremic syndrome in Microbiol 39:9092.
humans and animals. Toxins (Basel) 4:12611287. 189. Rangel JM, Sparling PH, Crowe C, Griffin PM,
181. Melton-Celsa AR, OBrien AD. 2003. Animal Swerdlow DL. 2005. Epidemiology of Escherichia
models for STEC-mediated disease. Methods Mol coli O157:H7 outbreaks, United States, 1982002.
Med 73:291305. Emerg Infect Dis 11:603609.
182. Beutin L, Martin A. 2012. Outbreak of Shiga 190. Pennington H. 2010. Escherichia coli O157. Lancet
toxin-producing Escherichia coli (STEC) O104: 376:14281435.
H4 infection in Germany causes a paradigm shift 191. Mohammed Hamzah A, Mohammed Hussein
with regard to human pathogenicity of STEC A, Mahmoud Khalef J. 2013. Isolation of Esche-
strains. J Food Prot 75:408418. richia coli 0157: H7 strain from fecal samples of
183. Karch H, Tarr PI, Bielaszewska M. 2005. zoo animal. Scientific World Journal 2013.
Enterohaemorrhagic Escherichia coli in human 192. Featherstone CA, Foster AP, Chappell SA,
medicine. Int J Med Microbiol 295:405418. Carson T, Pritchard GC. 2011. Verocytotoxigenic
184. Snedeker KG, Shaw DJ, Locking ME, Prescott Escherichia coli O157 in camelids. Vet Rec 168:
RJ. 2009. Primary and secondary cases in Esch- 194195.
erichia coli O157 outbreaks: a statistical analysis. 193. Dipineto L, Gargiulo A, Russo TP, De Luca
BMC Infect Dis 9:144. Bossa LM, Borrelli L, dOvidio D, Sensale M,
185. Mead PS, Griffin PM. 1998. Escherichia coli O157: Menna LF, Fioretti A. 2010. Survey of Esche-
H7. Lancet 352:12071212. richia coli O157 in captive frogs. J Wildl Dis
186. Wilson JB, Clarke RC, Renwick SA, Rahn K, 46:944946.
Johnson RP, Karmali MA, Lior H, Alves D, 194. Johnson M. 2005. Deer eating away at forests
Gyles CL, Sandhu KS, McEwen SA, Spika JS. nationwide. The Associated Press, January 18,
1996. Vero cytotoxigenic Escherichia coli infec- 2005.
tion in dairy farm families. J Infect Dis 174:1021 195. United States Department of Agriculture. 2011.
1027. Feral swine: damage and disease threat. http://www.
187. Wilson JB, Johnson RP, Clarke RC, Rahn K, aphis.usda.gov/publications/wildlife_damage/
Renwick SA, Alves D, Karmali MA, Michel P, content/printable_version/feral_swine.pdf.

Shiga Toxin-Producing
Escherichia coli in Fresh Produce:
A Food Safety Dilemma
PETER FENG1
12
INTRODUCTION
The worldwide trends for healthier lifestyles to reduce obesity and other com-
plications arising from unhealthy diets have greatly increased the consumption
of fresh fruits and vegetables. This increased demand, coupled with ever busier
consumer lifestyles, also stimulated the growth of a convenience food industry
and popularized the concept of bagged salad vegetables and fruits. It has been
estimated that several millions of bags of fresh produce are sold daily in the
United States. Bagged produce, also referred to as fresh cut or precut, is
often regarded as ready-to-eat (RTE) and consumed without further interven-
tion steps. However, because produce is predominantly cultivated in soil in open
fields, it is susceptible to contamination and can contain high levels of complex
microbial populations, occasionally including bacterial pathogens. As a result,
increases in fresh produce demand and consumption coupled with changes in
production practices have also contributed to increases in incidents of food-
borne illness. In the United States, about 0.7% of the infections in the 1970s
were attributed to fresh produce, but this increased to 6% in the 1990s (1).
Since fresh cut products are often mass produced, broadly distributed, and
marketed worldwide, a single pathogen contamination event can have broadly
impacting consequences, and several large, produce-related outbreaks have
1
Division of Microbiology, U.S. Food and Drug Administration, College Park, MD 20740-3835.
231

232 FENG
occurred in many countries (2, 3). In 2006, ranged from 104 to 107 CFU/g (79). Bagged,
a large multistate outbreak in the United fresh cut produce undergoes processing
States that infected more than 200 persons and is washed multiple times before and
was traced to bagged spinach contaminated after shredding and before bagging and, there-
with Escherichia coli O157:H7 (4). Several fore, would be expected to have lower mi-
months later, another O157:H7 outbreak in crobial contents. However, it was surprising
a fast-food restaurant chain had initially im- that these finished products can also contain
plicated green onions but appeared to have high levels of bacteria. Microbiological sur-
been due to bagged lettuce. At about the veys from several countries showed that total
same time, bagged lettuce was implicated bacterial counts of bagged produce were sim-
in another O157:H7 outbreak that affected ilar to levels observed in plants in the field
three states (5). Likewise, increased consump- and ranged widely from 103 to 107 CFU/g (10).
tion of sprouts caused several outbreaks of Furthermore, data from various countries
Salmonella sp., E. coli O157:H7, and other showed that it was not unusual to have mean
Shigatoxin-producing E. coli (STEC) strains. total microbial counts of 107 CFU/g or higher
STEC serotype O26:H11 strains caused an in both bagged salads and sprouts (1114).
outbreak with alfalfa sprouts and, more re- Recent FDA survey studies of bagged produce
cently, with clover sprouts, and the large in the United States showed similar results
outbreak of O104:H4 in 2011 in the European and wide ranges, and in some cases, total
Union also implicated the consumption of counts as high as 108 CFU/g were observed
sprouts (6). These large produce-related out- in some bags of spinach (15, 16). Analysis of
break incidents worldwide have greatly raised these microflora populations showed that the
concerns about the safety of fresh produce most frequently identified bacteria were Bacil-
and about the microbiological and sanitary lus subtilis, Pseudomonas fluorescens, Pantoea
quality of fresh produce. agglomerans, and Sphingomonas paucimobilis
(16).
MICROBIOLOGICAL QUALITY Coliform and Generic E. coli Populations

OF FRESH PRODUCE
Coliform and E. coli have been used as indirect
indicators of fecal contamination for more
Total Microbial Populations
than 100 years, but there seems to be little
The basic fresh produce production practices correlation between the presence or the levels
have remained essentially unchanged as most of indicator bacteria in food with the presence
produce is still cultivated in soil in a field of pathogens. Still, these indicators continue
and irrigated with available sources of water. to be of use in monitoring general sanitary
This constant exposure of produce plants to conditions. For instance, some produce in-
the environment makes them susceptible to dustries test their products to establish a base-
contamination, which can come from many line indicator level. In subsequent testing, the
sources, including soil, water, compost, ani- detection of any large deviations from baseline
mal wastes, and other environmental sources may be indicative of abnormal lots or process-
(http://www.fda.gov/Food/RecallsOutbreaks ing conditions that warrant investigation.
Emergencies/Outbreaks/ucm235477.htm). Coliform and E. coli are ubiquitous in the
As a result, it is not unusual to find high cultivation environments and, therefore, are
microbial counts on fresh produce in the often found on produce plants. Surveys of
field. Several studies examined the microbial produce in the field showed that coliform and
content of preharvest produce plants in the generic E. coli levels often ranged from 10 to
field and observed that total bacterial counts 104 CFU/g (79), and analogous to the findings

CHAPTER 12 STEC in Fresh Produce 233
with total bacterial counts, high levels of these where high levels were reported, E. coli levels
indicator bacteria can persist in bagged fin- in most bagged produce samples seem to be
ished products as well. low and within acceptable limits (17). These
There are no microbial limits for coliforms findings show that coliforms are too prevalent
and E. coli for bagged produce in the United in produce and their levels are too variable to
States, but guidelines and limits exist in other be able to generate a useful baseline. How-
countries. The Brazilian standard for mini- ever, E. coli does not appear to be part of
mally processed RTE vegetables has a fecal normal flora in fresh produce, and because its
coliform limit of 100 CFU/g (13). The Public presence and levels are usually low, it may be
Health Laboratory Service of the United feasible to establish an E. coli baseline, where
Kingdom has established microbiological qual- spikes in E. coli levels detected may be in-
ity guidelines for RTE foods, which include dicative of unsanitary or abnormal processing
bagged produce, and has set an E. coli limit of conditions.
<20/g as satisfactory, 20 to <100/g as accept- These surveys showed that total bacteria
able, and 100/g as unsatisfactory (17). Indi- and coliforms are prevalent in produce, and
cator bacteria are often enumerated by using interestingly, some studies showed little dif-
the most probable number (MPN) method, ferences in counts between organic versus
and surveys from various countries showed conventional produce (9) or between im-
that, like total counts, coliform levels in ported and domestic produce (8). But one
bagged salad varied greatly, ranging from un- observation that is consistent is that the
detected (<3.0 MPN/g) to 103 or 104 MPN/g levels of total and coliform counts in produce
(10, 15, 16). Similarly, generic E. coli may or vary greatly and can be unpredictable. One
may not be present, but, if found, the preva- study showed that even seemingly identical
lence rate and the levels also varied greatly products, of the same brand and use by
(10, 12, 15, 16). For example, surveys of lettuce dates and tested on the same day, can have as
samples at restaurants and university cafe- much as 2 to 3 log differences in counts (15).
terias in Spain showed that the prevalence Such discrepancies have also been noted in
rate of E. coli varied from 6.6% to 26% RTE produce in other countries (10, 11, 13).
(10, 18). A more startling finding is the study Whether these large variations are due to the
from Brazil that showed that 73% of the produce types, the complexity of produce
minimally processed vegetable salads exam- microflora, seasonal or regional variations, or
ined had exceeded the fecal coliform limit perhaps to inconsistencies or variations in
of 100 CFU/g (13). Likewise, E. coli can be processing parameters remains uncertain.
prevalent in sprouts and at high levels, as a
study from Spain showed that 40% of the
sprout samples tested positive and, in several, PRESENCE OF PATHOGENIC E. coli
the counts were >103 MPN/g (11). Two large IN FRESH PRODUCE
surveys in the United Kingdom tested ap-
proximately 3,000 samples of organic and Considering the huge quantities of fresh pro-
conventional salad vegetables and showed duce that are harvested and consumed daily,
that only 0.5% of the samples exceeded the the frequency of pathogen contamination in
E. coli limit of >100 CFU/g and, therefore, produce is very low. In response to increasing
were unsatisfactory (19, 20). Two surveys of concerns about food-borne outbreaks associ-
bagged lettuce and spinach in the United ated with the consumption of fresh produce,
States showed that E. coli was not detected in the Agricultural Marketing Service of the U.S.
one study (16), and in the other, 16% of the Department of Agriculture initiated the Mi-
samples had E. coli, but all were at levels of crobiological Data Program (MDP) in 2001 to
<10 MPN/g (15). So, with a few exceptions conduct surveillance of fresh produce samples

234 FENG
collected from wholesale distribution centers also had STa or LT. Although not surprising to
across the country. On the average, about find ETEC in fresh produce, ETEC has a fairly
10,000 to 15,000 produce samples of many high infectious dose, which, based on volun-
varieties were tested yearly for the presence teer feeding studies, has been estimated to
of Salmonella sp., enterotoxigenic E. coli be 108 to 1010 cells. As a result, quantifying the
(ETEC), E. coli serotype O157:H7, and other levels of ETEC present in produce would be
STEC strains. MDP analyses showed that more useful than the presence or absence data
many of these pathogens are found in various in assessing the health risks of ETEC in fresh
types of fresh produce, and the program pro- produce.
vided one of the largest, publicly available In the MDP produce testing scheme, a
databases on the presence of pathogens in multiplex PCR assay was used to simulta-
fresh produce. Unfortunately, the program neously screen produce samples for the pre-
was affected by budget cuts and terminated sence of ETEC and STEC (24). Interestingly,
in 2012. The yearly MDP reports are available this assay detected some E. coli strains that
at http://www.ams.usda.gov/AMSv1.0/mdp. carried both ETEC and STEC virulence fac-
tors. Although many of these strains were
only partially identified and characterized,
Enterotoxigenic E. coli
two strains from the 2004 MDP analysis were
ETEC is commonly known as the causative studied in detail. These two E. coli strains,
agent of travelers diarrhea; however, it is also found to carry both Stx1 and ST toxin genes,
an important diarrheal pathogen in infants. had an untypeable O antigen and H52 flagellar
The two trait virulence factors of ETEC are antigen and were isolated from fresh cilan-
the plasmid-encoded, heat-labile (LT) and tro and cantaloupe, respectively. Analysis by
heat-stable (ST) enterotoxins. The two sero- pulsed-field gel electrophoresis showed that
logically distinct LT types are designated LT-I the XbaI profiles of both strains were similar
and LT-II, but the latter is produced mostly by and did not resemble those of selected Stx1-
animal isolates and has not been associated producing STEC or ST-producing ETEC
with illness. There are also two distinct ST strains examined, but they shared >90% XbaI
types; STa is produced by both human and profile similarity to two clinical Ont:H52
porcine ETEC strains, but STb production is strains that had identical phenotypes and
limited mainly to porcine strains (21). traits (24). Furthermore, multilocus sequence
ETEC infections are most often caused by typing of all four strains showed that they
the consumption of contaminated food and had sequence type (ST) 274, which did not fall
water, and so ETEC strains can be found in into any of the defined STEC clonal groups
various foods, including produce and produce and, therefore, appeared to be a unique clone.
used in street-vendor food in Mexico (22). The Stx1 and ST genes reside on phage and
MDP analyses of produce showed that ETEC plasmid, respectively, and can be transferred,
was isolated almost every year from samples so it is uncertain if these Ont:H52 strains in
including lettuce, parsley, cilantro, alfalfa produce were ETEC that was infected by the
sprouts, and spinach. Typically, only a handful Stx1 phage or, conversely, a STEC that ac-
of strains are isolated a year, but in 2009, 11 quired the ST-encoding plasmid.
ETEC isolates were found and 4 came from
cilantro and spinach samples, respectively.
Shiga Toxin-Producing E. coli
Considering that about 2,200 samples of each
were tested yearly, the estimated ETEC prev- STEC is characterized by the production of
alence rate in these products is 0.18% (23). Stx, and there are two main types, designated
Most of these ETEC isolates had STb, which Stx1 and Stx2. Within each are many subtypes,
is associated with porcine strains, but a few and currently, there are three known Stx1

(Stx1a, Stx1c, and Stx1d) and seven known has remained fairly steady. For example, from
Stx2 (Stx2a, Stx2b, Stx2c, Stx2d, Stx2e, Stx2f, 2009 to 2011, 11 to 14 STEC strains were iso-
and Stx2g) subtypes (25). There are estimated lated from spinach yearly, giving a prevalence
to be 300 to 400 known STEC serotypes that rate of 0.5 to 0.6%. But in 2012, there were
can produce any of the Stx or combination 32 STEC isolations, 21 of which came from
of Stx subtypes, but some subtypes seem to spinach (0.95%). This fairly consistent preva-
be found mostly in environmental or animal lence of STEC strains in spinach suggests
strains and have not affected humans (26); that, perhaps, there may be some correlation
thus, not all STEC strains appear to cause between STEC and spinach plants or with
illness. STEC has been found in various forms spinach cultivation practices.
of wildlife (27) and in environmental sources All STEC strains isolated by MDP were
like water and soil and is also common in serotyped by the E. coli Reference Laboratory
farm or agricultural areas (http://www.fda. at Penn State University. However, more than
gov/downloads/Food/FoodSafety/Foodborne 50% of the isolates could only be partially
Illness/UCM235923.pdf). serotyped, when either the O or the H anti-
STEC strains can also be found in meats gens or both could not be identified. Among
and other foods (2729), and fresh produce is those that were serotyped, there were strains
no exception. During 2004 to 2009, European from diverse serogroups, most of which were
Union member states conducted a microbio- unremarkable, with no history of having caused
logical survey of produce and found that of the human illness (Table 1). But there were strains
6,000 samples tested, 11 (0.18%) had STEC from recognized pathogenic serotypes, includ-
(http://www.efsa.europa.eu/en/supporting/ ing O157:H7, O26:H11, O121:H19, and O113:H21,
doc/166e.pdf). Similarly, 10 years of MDP as well as other serotypes like O165:H25 and
analyses showed that STEC strains were iso- O91:H21 (Table 2) (30) that, historically, have
lated from various types of produce in the been implicated with severe human illness
United States, especially, spinach, lettuce, and (31, 32).
cilantro (http://www.ams.usda.gov/AMSv1.0/
mdp). From MDP statistics on the number of
STEC isolations per year in relation to the STECCHALLENGES IN PRODUCE
2,200 samples of each product tested yearly, TESTING AND IN RISK ANALYSIS
it is estimated that STEC was present in 0.5 to
0.6% of the spinach, 0.3 to 0.5% of the cilan- Testing for the presence of any pathogen in
tro, and 0.04 to 0.18% of the lettuce samples. fresh produce is challenging. The heteroge-
These estimates, however, may not reflect the neous distribution of pathogen contamina-
overall trend. For instance, the prevalence tion and the many types and varieties of fresh
rate obtained for STEC in lettuce was based produce present problems to sampling and
on MDP data from the past few years, which sample processing procedures. Furthermore,
had low isolation rates, but there had been fresh produce can contain very high levels
higher numbers of STEC isolations from let- of microflora, which can mask or overwhelm
tuce in other years. These observations sug- the sporadic presence of often very low levels
gest that STEC prevalence may vary from year of pathogens. To complicate matters further,
to year and depends on many factors, includ- produce can contain inhibitors that interfere
ing, perhaps, regional and seasonal variations. with assays; hence, fresh produce samples
Spinach was included in the MDP testing often have to be incubated in broth media
scheme in 2008 in response to the 2006 containing antibiotics or other inhibitory sub-
spinach outbreak with E. coli O157:H7. Since stances to enrich for the target bacteria and
then, this product has accounted for most of to suppress the growth of background flora.
the STEC isolations, and the prevalence rate Typical enrichment steps take 1 to 2 days, so

236 FENG
TABLE 1 Selected STEC strains and serotypes isolated and the susceptibilities of the human host.
from various produce commoditiesa For instance, even generic E. coli strains are
Product No. Serotypesb
deemed as opportunistic pathogens as they
Cantaloupe 3 Ontc:H11; Ont:H52; O88:H38 can cause infections in immunocompromised
Cilantro 18 Ont:H16, H31, H49; O1:H+; hosts and, therefore, cannot be regarded as
O8:H16, H28; O113:H5; being risk free. Risk assessment for STEC is
O139:H1; O153:H21; much more difficult due to the complexity of
O168:H8; others
STEC pathogenesis, plus the group comprises
Coriander 3 Ont:H7; O2:H25; O119:H4 a large diversity of strains, from several hun-
Hot peppers 3 O8:H9; O24:H11; O180:H14 dred serotypes, that can produce any of the
Lettuce 28 Ont:Hnt, O6:H49; O8:H28;
Ont:H2,H8; O136:H16; TABLE 2 Serotype and pathotype of selected produce
O143:H34; O163:H19; O168:H-, STEC strainsa
H8; O181:H49; others Commodity Serotypeb Pathotype(s)c
Parsley 1 Ont:H38
Cherry tomato O8:H19d stx1, stx2, ehxA, subA/B
Spinach 70 O8:H, H28; O11:H15; O21:Hnt; Cilantro O20:H19d stx1, stx2, saa, ehxA
O76:H+; O88:Hnt; O98:H36;
O26:H11 stx1, eae, ehxA
O107:Hnt; O113:H36;
O165:H25 stx1, stx2, ehxA, eae
O130:H11; O159:H19;
Untyped stx1, eae, ehxA
O181:H49; many Ont with
various H types Lettuce O2:H27e stx2, ehxA
O121:H19 stx2, ehxA, eae
Sprouts 3 Ont:H28, Hnt; O36:H14
O157:H7 stx2, eae
(alfalfa)
O157:H7 stx2, eae, ehxA
Tomatoes 3 Ont:Hnt O163:H19d stx2, ehxA, subA/B
Total 132 O165:H25 stx1, stx2, ehxA, eae
a
O174:H21d stx2
Table modied from reference 30.
b
Excludes pathogenic serotypes listed in Table 2. Spinach O157:H7f stx2, eae, ehxA
c
nt, not typeable. O8:H19d stx2, ehxA
O8:H19d stx2, ehxA
these prerequisite produce sample prepara- O26:H11 stx1, eae, ehxA
O82:H8e stx2, ehxA, saa
tion steps before testing are media-, time-, and O82:H8e stx2, ehxA, saa
labor-intensive. O91:H21 stx2, ehxA, saa
Once the produce samples are media- O98:H36g stx1, eae, ehxA
enriched, they are often screened by antibody- O113:H21 stx2, ehxA, saa, subAB
based or PCR or real-time PCR (qPCR) assays O113:H21 stx2, ehxA, saa, subAB
O113:H21 stx2, ehxA, saa, subAB
that are specific for particular targets. Since O116:H21h stx2, ehxA, saa, subAB
STEC strains are characterized solely by the O157:H7 stx2, eae, ehxA
production of Stx, anti-Stx assays or the use O157:H7 stx2, eae, ehxA
of stx-specific PCR or qPCR assays is effective O174:H2d stx2, ehxA, saa
in screening produce enrichments. Although O174:H21d stx1, stx2, ehxA,
saa, subAB
these assays are adequate to determine the a
Table modied from reference 30.
presence of STEC in produce, such data b
Known pathogenic serotypes are shown in bold type.
alone are inadequate for making risk-assess- c
stx1, Shiga toxin 1; stx2, Shiga toxin 2; eae, intimin; ehxA, entero-
ment decisions. hemolysin; saa, STEC agglutinating adhesin; subAB, subtilase cy-
To determine whether a bacterium poses totoxin.
d
Serotype reported to have history of causing HUS (31).
health risk is difficult as it depends on many e
Serotype reported to have history of causing other illnesses (31).
variables such as virulence factors, infectious f
From the spinach outbreak of 2006. Included as reference.
dosage, the pathogenicity of the organism, etc., g
Serotype has no history of causing illness (31), but has eae.
h
but it also depends on the health conditions Serotype reported to have history of causing HUS (32).

many Stx subtypes or combinations of sub- The mechanisms of attachment used by

types. Though some of these subtypes such other eae-negative EHEC strains are less
as Stx1e do not seem to affect humans, and certain. Analysis of the eae-negative O113:H21
infections by STEC strains with Stx1c tend strain that caused an outbreak of HUS in
to range from asymptomatic to mild diarrhea Australia identified the plasmid-borne saa
(33), nevertheless, they may only be regarded gene that encodes for STEC agglutinating
as low or minimal risk, but not without risk. adhesin (Saa) (36). Saa was determined to be
On the other hand, infections by other STEC an adherence protein as a plasmid-cured, saa-
strains can cause severe diseases, such as negative O113:H21 mutant showed reduced
hemorrhagic colitis (HC) that can lead to adherence compared to wild type, and puri-
life-threatening complications like hemolytic- fied Saa protein enhanced adherence to HEp-
uremic syndrome (HUS). In these instances, 2 cells (37). However, studies that examined
the production of Stx alone is deemed to be the distribution of the saa genes found that
insufficient, and an adherence factor that over half of the STEC strains isolated from
enables the pathogen to attach to epithelial healthy cattle were positive for saa (38), and
cells is also needed for severe disease to occur. although it was also found in some clinical
Other general information on pathogenic STEC strains, there was no significant corre-
E. coli can be found in a consumer-oriented lation between the presence of saa and HUS
report titled E. coli: Good, Bad & Deadly (39).
published by the American Academy of Mi- Among the other factors that are often as-
crobiology and is available at http://academy. sociated with pathogenic STEC strains are
asm.org/images/stories/documents/EColi.pdf. the plasmid-encoded enterohemolysin and
The most notable adherence factor among subtilase cytotoxin. The ehxA gene that en-
pathogenic STEC is the intimin protein, encodes for enterohemolysin has been found to
coded by the eae gene that resides on the be very common among STEC strains isolated
locus for enterocyte effacement pathogenicity from all sources, including clinical, environ-
islands. The presence of eae and stx2 has been mental, and foods. The prevalence rate can
found to be a good predictor that the STEC range from 30% in STEC isolated from wild-
strain may cause HC or HUS (34). Hence, the life to as many as 70% of the STEC strains
term enterohemorrhagic E. coli has been used isolated from ground meats (40). However,
by some to designate a subset of STEC that the prevalence of ehxA is not limited to STEC,
comprises pathogenic strains, and most of as analysis of 300 environmental isolates of
these have eae. Among EHEC strains, sero- generic E. coli from surface waters showed
type O157:H7 is the prototypic strain, but that almost all carried and expressed ehxA
others in the serogroup O26, O111, O103, O145, and none were STEC (41). Moreover, the role
to name a few, have also caused severe human of enterohemolysin in STEC pathogenesis
illnesses. Also, other EHEC strains, such as is uncertain. A study of 300 clinical STEC
strains in the serotype O113:H21, O91:H21, isolates showed that 77% had the ehxA gene,
O104:H4, etc., do not have eae but have caused but its presence could not be correlated with
HUS (35, 36), so these pathogens appear to the occurrence of HUS or bloody diarrhea
have other means of attachment. For example, (34). In addition, the sorbitol-fermenting var-
the O104:H4 strain that caused the large HUS iants of O157:H7 (SFO157) do not express
outbreak in the European Union in 2011 did ehxA but are highly pathogenic and have
not have eae, but it was an enteroaggregative caused many outbreaks of HUS in Europe
E. coli strain, and its ability to attach and ag- (42). Despite the uncertainty of its role in
gregate on epithelial cells, coupled with Stx2 pathogenesis, the expression of enterohemo-
production, is postulated to have caused the lysin has become a useful marker in STEC
severity of infections. isolation procedures (see below).

238 FENG
The subtilase cytotoxin encoded by the followed up by other trait-specific assays. The
subAB gene and produced predominantly by key STEC virulence genes tested by almost
eae-negative STEC strains has been determined everyone worldwide are stx and eae, and if the
to be a potent toxin that is even more cyto- samples are positive, they are often tested
toxic to Vero cells than Stx (43). The subAB with assays specific for certain serotypes.
gene has been found in 50% of the STEC While these strategies may seem straightfor-
strains isolated from ground meats (40). An- ward, the selection of target serotypes is
other study showed that 72% and 86% of not. Studies showed that there are regional
the eae-negative STEC strains from patients variations in the STEC serotypes that cause
with diarrhea and healthy sheep, respectively, infections (45); hence, the follow-up serotype-
had subAB (44), so it is very prevalent among specific assays may vary geographically. For
STEC strains. However, other eae-negative example, serotype O157:H7 continues to be
STEC strains, like O91:H21 and O22:H8, have regarded as the most important EHEC path-
been implicated in severe illnesses, but ground ogen, and so it is almost always included in
beef isolates of these serotypes did not have all screening strategies. But other non-O157
subAB (40). So the role of this toxin in the STEC strains are increasingly being impli-
pathogenesis of eae-negative STEC strains cated in food-borne infections worldwide. In
also remains elusive. the United States, strains from serotypes O26,
A recent study characterized 130 STEC O111, O121, O103, O145, and O45 have been
strains isolated from a variety of fresh produce isolated most frequently from clinical infec-
samples over a period of about 10 years. The tions, and so these six, commonly referred to
isolates were tested by PCR for the presence as the big 6, have emerged as being impor-
of eae, ehxA, saa, and subAB genes (30). The tant and, recently, declared as adulterants in
study showed that eae was present in about meats. These strains are equally important
8% of the isolates, and most of these were in fresh produce as strains of O111, O121, and
recognized pathogenic serotypes (Table 2) O145 have caused outbreaks in lettuce and
(30). The other genes were even more prev- O26 strains in alfalfa and clover sprouts.
alent among STEC strains in produce, as However, focusing on the big 6 may be in-
60% of the STEC strains had ehxA, and 35% sufficient in produce testing, as other known
and 32% of the strains had the saa and subAB pathogenic STEC serotypes have been isolated
genes, respectively (30). Thus, considering from fresh produce (30). Hence, for risk as-
the complexities of STEC pathogenesis, the sessment of STEC in produce, it is essential
variety of STEC serotypes that exist, and the that samples found to be positive for key vir-
wide distribution of various virulence and ulence factors by initial screening be further
putative virulence genes among STEC strains tested to obtain additional relevant data to
in produce, it is clear that much more in- determine health risk.
formation beyond the production of Stx or While multiplex assays can generate many
the presence of stx is needed for STEC risk data points in one assay, there are drawbacks
assessment. to using this approach to screen complex,
Fresh produce has a short shelf life, so it is mixed bacterial population samples, as all the
crucial to obtain relevant risk analysis data as gene targets detected may not be within the
quickly as possible. One such testing strategy same bacteria. For example, a survey for stx,
is to use multiplex PCR assays to simulta- eae, and the big 6 serotype-specific genes in
neously screen for a combination of virulence another flora-complex food like ground meats
and putative virulence factor genes plus other showed that it was not uncommon to find
relevant markers directly in produce enrich- bacterial strains that carried each of these
ments. Others use sequential strategies, where gene targets independently (40). But when
samples positive for virulence factor genes are they were all present in the same sample, it

gave positive multiplex results for all three costly, are not very inclusive, and are useful
targets and, thereby, the misleading impres- only for selected serotypes and strains. Also,
sion that a pathogen with those attributes was the high levels and the complex microflora
present in the sample (40). Produce can con- present in produce samples can often mask
tain equally complex, mixed microbial pop- or mimic pathogenic colonies on these media,
ulations; hence, it is imperative that samples thereby interfering with differentiation of
initially screened and found to be positive are STEC colonies. Similarly, to facilitate the iso-
plated onto selective or differential media to lation of selected serotypes from enrichment
isolate the STEC strain and to verify that all broths, specific antibodies have been used in
the relevant genes targeted are within the immunomagnetic separation (IMS) methods
same organism. to capture specific organisms. But the capture
Without a doubt, isolation and confirmation efficiency of IMS can vary greatly, depending
of STEC strains from produce enrichments are on the antibodies used and on the type of
time- and labor-intensive. Fortunately, these produce samples being tested. In addition,
steps are actually easier and straightforward the antibodies used in IMS often target sero-
for O157:H7 strains due to the unique pheno- groups rather than serotypes. For example,
types expressed by this pathogen. Unlike typ- an anti-O157 IMS method is not specific for
ical E. coli, O157:H7 strains do not ferment O157:H7, as it will capture O157 strains re-
sorbitol or express b-glucuronidase activity, gardless of H type. There are many O157
so both traits are used extensively in plating strains that have H antigens other than H7,
agar media to differentiate O157:H7 colonies and they are not STEC nor are they patho-
from others. Presence of the O157 and the genic, but can be found in foods. As a result
H7 antigens has also eased identification, as of these limitations and logistical problems,
simple latex agglutination tests can be used isolating a STEC strain from produce enrich-
to identify O157:H7 strains quickly (21). An ments often entails the laborious process of
example of a method that uses these traits picking and pooling numbers of colonies from
to identify O157:H7 is outlined in the FDA the plate, screening it for Stx or stx by anti-
Bacteriological Analytical Manual (http:// body or PCR assays, and repeating the process
www.fda.gov/Food/FoodScienceResearch/ until a pure STEC strain is isolated for sero-
LaboratoryMethods/ucm2006949.htm). typing and virulence testing. This process is
In contrast, isolation and confirmation of not only labor-intensive and costly, it is time-
the non-O157 STEC strains are much more consuming and does not provide timely data
complex, laborious, and time-consuming and, for making risk-assessment decisions.
undoubtedly, the most problematic step in The availability of good differential plating
STEC testing. As mentioned, there are 300 media would greatly simplify the process of
STEC serotypes that comprise diverse strains STEC isolation from produce enrichments.
and serotypes, and not all appear to affect As mentioned, even though its role in patho-
humans. Those strains that are known path- genesis is uncertain, the majority of STEC
ogens may share common virulence traits strains produce enterohemolysin. Strains of
but do not uniformly exhibit unique pheno- O157:H7 almost always carry ehxA and ex-
types that can be used in differentiation and press enterohemolysin, as do many of the
isolation. As a result, there are no methods other pathogenic STEC strains. Enterohemo-
that can select and isolate all the diverse lysin activity can be visualized as a faint zone
strains existing within the pathogenic STEC of hemolysis around the colonies on washed
group. Many have used various chromogenic blood agar (WBA) medium, which is made up
substrates and combinations of phenotypes to of tryptose blood agar base supplemented
develop plating media for differentiation of with 10 mM of CaCl2 and 5% of defribrinated
STEC strains. However, these media can be sheep blood (46). The WBA medium was later

240 FENG
modified to become the WBMA medium by be regarded as a public health concern. Of

the addition of 0.5 g/ml of mitomycin C, the 130 strains in produce tested in that
which has been shown to induce and enhance study, 11 strains (8%) had eae, and except for
the visualization of enterohemolytic activity a strain of O98:H36 serotype and another
and also increased the detection rate of that was untyped, all the other eae-bearing
STEC strains (47). A similar medium called strains belonged to well-recognized patho-
SHIBAM, which uses heart infusion agar genic serotypes and, therefore, were of safety
base, is described in the FDA Bacteriological concern (Table 2) (30).
Analytical Manual. Although not all STEC There are EHEC strains that do not have
strains express ehxA, the use of WBMA or eae but have caused severe illness, and some
SHIBAM has eased the recognition of STEC of these, like strains of O113:H21 and O91:H21
strains by the zones of hemolysis around the serotype, have been found by MDP in fresh
colonies and, thereby, facilitated the selection spinach samples. Characterization and com-
and isolation of STEC strains from complex, parison of the O113:H21 strains from spinach
mixed-flora food enrichments like produce to strains that caused infections in Australia
samples. showed that although the produce strains had
Once a STEC strain in produce has been a different sequence type (ST223), they had
isolated, characterized, and serotyped, sub- identical traits and phenotypes as the patho-
jectivity remains in making risk-assessment genic strains (48). A more detailed study used
decisions. To illustrate these uncertainties, a microarray to analyze 41 pathogenic E. coli
one study characterized 130 STEC strains markers in 65 O113:H21 strains isolated from
isolated from produce samples over a 10-year food, spinach and environmental sources
period (30). In some cases, risk-assessment worldwide and, showed that they could not be
decisions were straightforward. For example, distinguished from the O113:H21 strains that
some produce samples had O157:H7 strains, caused HUS and therefore, are most likely
and these, without a doubt, are of serious pathogenic as well (49). Produce samples also
safety concern. A few atypical O157:H7 strains contained several strains that belonged to
were also isolated from produce, including serotypes like O2:H27, O82:H8, O116:H21,
strains that did not produce enterohemolysin O174:H21, etc., that have been reported to
or Stx. Both factors are encoded by mobile have been isolated from human infections
genetic elements that can be lost or trans- (31, 32). None of these strains carried eae, but
ferred. The role of enterohemolysin in path- many had ehxA, saa, and subAB (Table 2). In
ogenesis is uncertain and the SFO157 strains spite of the uncertain role of these genes in
do not express enterohemolysin but are highly STEC pathogenesis, the historical association
virulent, so an O157:H7 strain that is entero- of these serotypes with illness should be con-
hemolysin negative should not be regarded sidered when making risk-assessment deci-
lightly. The Stx encoding phage can be lost sions. Excluding the 30 STEC strains in
during culture and isolation, so the possibility produce that fit the criteria mentioned and
that a toxigenic strain exists in the product discussed above, risk assessment on the re-
cannot be ruled out. Hence, in both instances, maining 100 STEC isolates from produce is
the presence of these atypical O157:H7 strains more complex and difficult. These strains did
should also be regarded as being a health not have eae, but 60% had ehxA, 35% had
risk. Similarly, if a STEC strain is found to saa, and 32% had subAB. Moreover, almost
have eae and of a known pathogenic sero- half of these only had partial O and H type
type, it is also a serious health risk; a handful data or could not be serologically typed.
of STEC strains in produce fit these criteria Knowing the serotype of the STEC isolates
(Table 2). However, even if the serotype is is important, but it is difficult to determine
unknown, a STEC strain that had eae should due to the large numbers and the complex

combinations of O and H antigen types that rates of 0.2% and 0.5% for ETEC and
can exist among E. coli strains. The limitations STEC, respectively, have been observed in
of using antibodies for serotyping, however, some years. Screening for STEC in produce
may be remedied in the future with genotypic is straightforward, but there are no effective
assays, as a microarray was able to identify the isolation methods; hence the confirmation
serotype of many untypeable STEC strains process to isolate, characterize, and serotype
isolated from fresh produce (50). In the mean- the strains is time- and labor-intensive. Pro-
time, considering the complexities and uncer- duce contains many different serotypes of
tainties of these putative virulence factors in STEC and many strains that could only be
pathogenesis and the incomplete or lack of partially serotyped or are untyped. Only a
serotype identity, it is almost impossible to handful of the STEC strains isolated from
determine if these STEC strains in produce fresh produce carried the eae gene that en-
are of safety concern. codes for the adherence protein, intimin,
Thus, risk assessment on STEC isolates and can be regarded as a health risk. They
from produce can be made systematically and included O157:H7 strains, a few big 6 sero-
with some rationality in some cases, but most type strains like O121:H19 and O26:H11, but
assessments tend to be inconclusive due to also included other pathogenic serotypes
partial identity or the uncertain role of puta- such as O165:H25. The majority of the STEC
tive virulence factors detected. Even more isolates from produce did not have eae, but
critical is the inability to get timely data some of them were of serotype O113:H21 and
to make risk-assessment decisions on such O91:H21 strains, which historically have been
short-shelf-life products as fresh produce. implicated with severe illness and, therefore,
Occasionally, the ease of isolation and identi- may also be a safety concern. Risk assessment
fication of O157:H7 makes it feasible to obtain for the other STEC strains in produce is more
data in time to decide to withdraw the prod- difficult as many of these strains carried pu-
uct from the market. For most of the other tative virulence factor genes that had uncer-
STEC strains, by the time the strain is isolated, tain roles in pathogenesis, only had partial
characterized, and serotyped, the product has serological identity, or belonged to serotypes
either been consumed or passed its expiration that had no history of causing human ill-
date and has been taken off the market and ness. Thus, the logistical problems associated
discarded. with testing for a complex group of bacteria
like STEC in microflora-complex, short-
shelf-life products like fresh produce have
CONCLUSION greatly limited the ability to obtain timely
and relevant data for making risk-assessment
Fresh produce is constantly exposed to the decisions.
environment and, therefore, can contain com-
plex microflora populations at levels that
ACKNOWLEDGMENT
sometimes exceed 108 CFU/g. It is also com-
mon for produce to contain indicator bacteria, I declare no conflict of interest with regard to
and the levels of coliforms can vary from un- the manuscript.
detected to as high as 104 CFU/g. The levels
of generic E. coli in produce are much lower
CITATION
and, if E. coli is present, the level is usually
<10 CFU/g, but in occasional samples, it may Feng P. 2014. Shiga toxin-producing Esche-
exceed 100 CFU/g. Pathogenic E. coli can be richia coli in fresh producea food safety
found in produce samples but seems to be dilemma. Microbiol Spectrum 2(4):EHEC-
more prevalent in spinach where prevalence 0010-2013.

242 FENG
REFERENCES 13. Frder H, Martins CG, De Souza KL, Landgraf

M, Franco BD, Destro MT. 2007. Minimally
1. Sivapalasingam S, Friedman CR, Cohen L, processed vegetable salads: microbial quality eval-
Tauxe RV. 2004. Fresh produce: a growing cause uation. J Food Prot 70:12771280.
of outbreaks of foodborne illness in the United 14. Hagenmeaier RD, Baker RA. 1998. A survey of
States, 1973 through 1997. J Food Prot 67:2342 the microbial population and ethanol content of
2353. bagged salad. J Food Prot 61:357359.
2. Lynch MF, Tauxe RV, Hedberg CW. 2009. The 15. Valentin-Bon I, Jacobson A, Monday S, Feng P.
growing burden of foodborne outbreaks due to 2008. Microbiological quality of bagged cut spin-
contaminated fresh produce: risks and opportu- ach and lettuce mixes. Appl Environ Microbiol
nities. Epidemiol Infect 137:307315. 74:12401242.
3. Berger CN, Sodha SV, Shaw RK, Griffin PM, 16. Kase JA, Borenstein S, Feng PCH. 2012. Mi-
Pink D, Hand P, Frankel G. 2010. Fresh fruit crobial quality of bagged baby spinach and ro-
and vegetables as vehicles for the transmission of maine lettuceeffects of top vs. bottom sampling.
human pathogens. Environ Microbiol 12:2385 J Food Prot 75:132136.
2397. 17. Sagoo SK, Little CL, Mitchell RT. 2003. Micro-
4. Centers for Disease Control and Prevention. biological quality of open ready-to-eat salad veg-
2006. Ongoing multistate outbreak of Escherichia etables: effectiveness of food hygiene training of
coli serotype O157:H7 infections associated with management. J Food Prot 66:15811586.
consumption of fresh spinachUnited States, 18. Sospedra I, Rubert J, Soriano JM, Maes J.
September 2006. Morb Mortal Wkly Rep 55:1045 2013. Survey of microbial quality of plant-based
1046. foods served in restaurants. Food Control 30:418
5. Doyle MP, Erickson MC. 2008. Summer meeting 422.
2007the problem with fresh produce: an over- 19. Sagoo SK, Little CL, Mitchell RT. 2001. The
view. J Appl Microbiol 105:317330. microbiological examination of ready-to-eat or-
6. Beutin L, Martin A. 2012. Outbreak of Shiga ganic vegetables from retail establishments in the
toxin-producing Escherichia coli (STEC) O104:H4 United Kingdom. Lett Appl Microbiol 33:434439.
infection in Germany causes a paradigm shift 20. Sagoo SK, Little CL, Ward L, Gillespie IA,
with regard to human pathogenicity of STEC Mitchell RT. 2003. Microbiological study of
strains. J Food Prot 75:408418. ready-to-eat salad vegetables from retail estab-
7. Johnston LM, Jaykus LA, Moll D, Martinez lishments uncovers a national outbreak of sal-
MC, Anciso J, Mora B, Moe CL. 2005. A field monellosis. J Food Prot 66:403409.
study of the microbiological quality of fresh pro- 21. Feng P, Strockbine N, Fratamico P. 2012.
duce. J Food Prot 68:18401847. Methods of detection and characterization of
8. Johnston LM, Jaykus LA, Moll D, Anciso J, pathogenic Escherichia coli. In UNESCO-EOLSS
Mora B, Moe CL. 2006. A field study of the mi- Joint Committee (ed), Food Quality and Stan-
crobiological quality of fresh produce of domestic dards, in Encyclopedia of Life Support Systems
and Mexican origin. Int J Food Microbiol 112:83 (EOLSS). Developed under the Auspices of the
95. UNESCO. Eolss Publishers, Oxford, UK. (http://
9. Mukherjee A, Speh D, Jones AT, Buesing KM, www.eolss.net)
Diez-Gonzalez F. 2006. Longitudinal microbio- 22. Lopez-Saucedo C, Cerna JF, Estrada-Garcia T.
logical survey of fresh produce grown by farmers 2010. Non-O157 Shiga toxin-producing Esche-
in the upper Midwest. J Food Prot 69:19281936. richia coli is the most prevalent diarrheagenic
10. Soriano JM, Rico H, Molt JC, Maes J. 2006. E. coli pathotype in street-vended taco dressings
Assessment of the microbiological quality and in Mexico City. Clin Infect Dis 50:450451.
wash treatments of lettuce served in University 23. Feng PCH, Reddy SP. 2014. Prevalence and di-
restaurants. Int J Food Microbiol 58:123128. versity of enterotoxigenic Escherichia coli strains
11. Abadias M, Usall J, Anguera M, Solsona C, in fresh produce. J Food Prot 77:820823.
Vias I. 2008. Microbiological quality of fresh, 24. Monday SR, Keys C, Hansen P, Shen Y,
minimally-processed fruit and vegetables, and Whittam TS, Feng P. 2006. Produce isolates of
sprouts from retail establishments. Int J Food Escherichia coli Ont:H52 serotype that carry both
Microbiol 123:121129. Shiga toxin 1 and StableToxin genes. Appl Environ
12. Caponigro V, Ventura M, Chiancone I, Amato Microbiol 72:30623065.
L, Parente E, Piro F. 2010. Variation of microbial 25. Scheutz F, Teel LD, Beutin L, Pirard D,
load and visual quality of ready-to-eat salads by Buvens G, Karch H, Mellmann A, Caprioli A,
vegetable type, season, processor and retailer. Food Tozzoli R, Morabito S, Strockbine NA, Melton-
Microbiol 27:10711077. Celsa AR, Sanchez M, Persson S, OBrien AD.

2012. Multicenter evaluation of a sequence-based 36. Paton AW, Woodrow MC, Doyle RM, Lanser
protocol for subtyping Shiga toxins and stan- JA, Paton JC. 1999. Molecular characterization
dardizing Stx nomenclature. J Clin Microbiol of a Shiga toxigenic Escherichia coli O113:H21
50:29512963. strain lacking eae responsible for a cluster of cases
26. Martin A, Beutin L. 2011. Characteristics of Shiga of hemolytic-uremic syndrome. J Clin Microbiol
toxin-producing Escherichia coli from meat and 37:33573361.
milk products of different origins and association 37. Paton AW, Srimanote P, Woodrow MC, Paton
with food producing animals as main contami- JC. 2001. Characterization of Saa, a novel auto-
nation sources. Int J Food Microbiol 146:99104. agglutinating adhesin produced by locus of
27. Monaghan A, Byrne B, Fanning S, Sweeney T, enterocyte effacement-negative Shiga-toxigenic
McDowell D, Bolton DJ. 2011. Serotypes and Escherichia coli strains that are virulent for
virulence profiles of non-O157 Shiga toxin- humans. Infect Immun 69:69997009.
producing Escherichia coli isolates from bovine 38. Galli L, Miliwebsky E, Irino K, Leotta G, Rivas
farms. Appl Environ Microbiol 77:86628668. M. 2010. Virulence profile comparison between
28. Mora A, Lpez C, Dhabi G, Lpez-Beceiro AM, LEE-negative Shiga toxin-producing Escherichia
Fidalgo LE, Daz EA, Martnez-Carrasco C, coli (STEC) strains isolated from cattle and
Mamani R, Herrera A, Blanco JE, Blanco M, humans. Vet Microbiol 143:307313.
Blanco J. 2012. Seropathotypes, phylogroups, Stx 39. Jenkins C, Perry NT, Cheasty T, Shaw DJ,
subtypes, and intimin types of wildlife-carried, Frankel G, Dougan G, Gunn GJ, Smith HR,
Shiga toxin-producing Escherichia coli strains Paton AW, Paton JC. 2003. Distribution of
with same characteristics as human-pathogenic the saa gene in strains of Shiga toxin-producing
isolates. Appl Environ Microbiol 78:25782585. Escherichia coli of human and bovine origins.
29. Garca-Aljaro C, Muniesa M, Blanco JE, Blanco J Clin Microbiol 41:17751778.
M, Blanco J, Jofre J, Blanch AR. 2005. Char- 40. Bosilevac JM, Koohmaraie M. 2011. Prevalence
acterization of Shiga toxin-producing Escherichia and characterization of non-O157 Shiga toxin
coli isolated from aquatic environments. FEMS producing Escherichia coli isolated from com-
Microbiol Lett 246:5565. mercial ground beef in the United States. Appl
30. Feng PCH, Reddy S. 2013. Prevalences of Shiga Environ Microbiol 77:21032112.
toxin subtypes and selected other virulence 41. Boczek LA, Johnson CH, Rice EW, Kinkle BK.
factors among Shiga-toxigenic Escherichia coli 2006. The widespread occurrence of the entero-
strains isolated from fresh produce. Appl Environ hemolysin gene ehlyA among environmental
Microbiol 79:69176923. strains of Escherichia coli. FEMS Microbiol Lett
31. Hussein HS. 2007. Prevalence and pathogenicity 254:281284.
of Shiga toxin-producing Escherichia coli in beef 42. Karch H, Bielaszewska M. 2001. Sorbitol-
cattle and their products. J Anim Sci 85(13 Suppl): fermenting Shiga toxin-producing Escherichia
E63E72. coli O157:H- strains: epidemiology, phenotypic
32. Newton HJ, Sloan J, Bulach DM, Seemann T, and molecular characteristic, and microbiological
Allison CC, Tauschek M, Robins-Browne RM, diagnosis. J Clin Microbiol 39:20432049.
Paton JC, Whittam TS, Paton AW, Hartland 43. Paton AW, Beddoe T, Thorpe CM, Whisstock
EL. 2009. Shiga toxin producing Escherichia coli JC, Wilce MC, Rossjohn J, Talbot UM, Paton
strains negative for locus of enterocyte efface- JC. 2006. AB5 subtilase cytotoxin inactivates the
ment. Emerg Infect Dis 15:372380. endoplasmic reticulum chaperone BiP. Nature 443:
33. Friedrich AW, Borell J, Bielaszewska M, Fruth 548552.
A, Tschape H, Karch H. 2003. Shiga toxin 44. Michelacci V, Tozzoli R, Caprioli A, Martnez
1c-producing Escherichia coli strains: phenotypic R, Scheutz F, Grande L, Snchez S, Morabito S.
and genetic characterization and association with 2013. A new pathogenicity island carrying an al-
human disease. J Clin Microbiol 41:22482453. lelic variant of the Subtilase cytotoxin is common
34. Ethelberg S, Olsen KE, Scheutz F, Jensen C, among Shiga toxin producing Escherichia coli of
Schiellerup P, Enberg J, Olesen B, Gerner- human and ovine origin. Clin Microbiol Infect 19:
Smidt P, Mlbak K. 2004. Virulence factors for E149E156.
hemolytic uremic syndrome, Denmark. Emerg 45. Johnson KE, Thorpe CM, Sears CL. 2006. The
Infect Dis 10:842847. emerging clinical importance of non-O157 Shiga
35. Karmali MA, Petric M, Lim C, Fleming PC, toxin-producing Escherichia coli. Clin Infect Dis.
Arbus GS, Lior H. 1985. The association between 43:15871595.
idiopathic hemolytic uremic syndrome and in- 46. Beutin L, Montenegro MA, Orskov I, Prada J,
fection by verocytotoxin-producing Escherichia Zimmerman S, Stephan R. 1989. Close associa-
coli. J Infect Dis 151:775782. tion of verotoxin (Shiga-like toxin) production

244 FENG
with enterohemolysin production in strains 49. Feng PCH, Delannoy S, Lacher DW, dos Santos
of Escherichia coli. J Clin Microbiol 27:2559 LF, Beutin L, Fach P, Rivas M, Hartland EL,
2564. Paton AW, Guth BEC. 2014. Genetic diversity
47. Sugiyama K, Inoue K, Sakazaki R. 2001. Mito- and virulence potential of Shiga toxin-producing
mycin-supplemented washed blood agar for the Escherichia coli O113:H21 strains isolated from
isolation of Shiga toxin-producing Escherichia clinical, environmental, and food sources. Appl
coli other than O157:H7. Lett Appl Microbiol Environ Microbiol 80:47574763.
33:193195. 50. Lacher DW, Gangiredla J, Jackson SA, Elkins
48. Feng PCH, Councell T, Key C, Monday SR. CA, Feng PCH. 2014. A novel microarray design
2011. Virulence characterization of Shigatoxigenic for molecular serotyping of Shiga toxin-producing
Escherichia coli serotypes isolated from wholesale Escherichia coli isolated from fresh produce. Appl
produce. Appl Environ Microbiol 77:343345. Environ Microbiol 80:46774682.

Public Health Microbiology of Shiga
Toxin-Producing Escherichia coli
ALFREDO CAPRIOLI,1 GAIA SCAVIA1 and STEFANO MORABITO1

13
Shiga toxin-producing Escherichia coli (STEC) represents a major issue for
public health because of the capability to cause large outbreaks and the severity
of the associated illnesses (1). STEC strains are the only pathogenic group of
E. coli that has a definite zoonotic origin, with ruminants and, in particular,
cattle being recognized as the major reservoir for human infections (2). Most
human infections are food borne, but the routes of transmission include direct
contact with animals and a wide variety of environment-related exposures
(3). Therefore, STEC public health microbiology spans the fields of medical,
veterinary, food, water, and environmental microbiology, requiring a One
Health perspective (4) and laboratory scientists with the ability to work ef-
fectively across disciplines. Public health microbiology laboratories play a cen-
tral role in the surveillance of STEC infections, as well as in the preparedness
for responding to outbreaks and in providing scientific evidence for the im-
plementation of prevention and control measures. This article reviews in depth
(i) how the integration of surveillance of STEC infections and monitoring of
these pathogens in animal reservoirs and potential food vehicles may contribute
to their control; (ii) the role of reference laboratories; and (iii) the public health
perspectives, including those related to regulatory issues in both the European
Union and the United States.
1
European Union Reference Laboratory for E. coli, Dipartimento di Sanit Pubblica Veterinaria e Sicurezza Alimentare,
Istituto Superiore di Sanit, Rome, Italy.
245

246 CAPRIOLI ET AL.
index.aspx) coordinated by the European

SURVEILLANCE AND MONITORING OF
Centre for Disease Prevention and Control
STEC INFECTIONS
(ECDC). FWD is a passive surveillance sys-
tem, collecting data on STEC infections from
Surveillance of STEC Infections in Humans
the European Union and the European Eco-
In the medical field, dedicated surveillance nomic Area (EEA) countries, based on case
systems of human STEC infections have been definitions that include laboratory-confirmed
developed in most of the industrialized areas cases, probable cases, and possible cases. For
of the world because of the public health laboratory-confirmed cases, collected data in-
importance of these infections. Such systems clude the serotype and the main virulence
are of utmost importance for the prompt re- genes (stx1, stx2, eae) of the infecting strain,
cognition and management of outbreaks and together with the clinical manifestation, in
the implementation of specific strategies to particular, the development of HUS. The data
control the spread of the infections in the on STEC infections are published yearly in the
community. European Union Summary Reports on Trends
Surveillance systems should record not and Sources of Zoonoses, Zoonotic Agents and
only the epidemic outbreaks but also sporadic Food-borne Outbreaks and provide informa-
cases of infection. In particular, cases of se- tion on the trend of STEC infections in the
vere illness, such as bloody diarrhea and the European Union, according to the serogroup
hemolytic-uremic syndrome (HUS), may re- of the infecting strains and the development
veal a wider circulation of STEC strains in a of HUS. Figure 1 shows the number of STEC
community and should be considered as pos- infections reported to the ECDC-FWD sur-
sible syndromic sentinel events of possible veillance system in the period 2007 to 2011
underlying clusters of infections (5). (711). An increasing trend of reported cases
The data sets that should be part of sur- was observed, likely due to a general im-
veillance systems of STEC infections include provement of the surveillance systems in the
laboratory and clinical data, as well as infor- participating countries. This hypothesis is sup-
mation on risk factors and the possible associ- ported by the sharp increase in the cases
ation with other cases of infection. Surveillance caused by STEC O157 and STEC non-O157
systems are usually based on laboratory- other than O104 notified in 2011, when the
confirmed case definitions that are dependent large outbreak of STEC O104:H4 infections
on the methods applied for laboratory diagno- occurred in Germany and other European
sis. Such methods may vary considerably. In countries (12). Such an increase in the re-
many settings, they are limited to the detection porting was likely due to the enhanced at-
of STEC O157, leading to an underestimation tention toward STEC infections raised by
of the incidence of STEC non-O157 infections the outbreak. At the same time, a decrease in
(6). The information on the clinical outcome of the number of cases for which the infecting
the reported cases of STEC infection is very strain was not serotyped was observed, prob-
important, in particular when HUS develops, ably due to the enhanced rate of submission
since it is crucial to estimate the burden of of STEC strains to reference laboratories for
disease due to these pathogens and to define confirmation and typing.
which STEC serotypes are consistently asso- The serogroups other than O157 and O104
ciated with severe disease. most frequently associated with STEC infec-
In Europe, the surveillance of STEC tions in the European Union in the period
infections is embedded in the Food- and 2007 to 2011 are reported in Fig. 2. STEC O26
Waterborne Diseases and Zoonoses (FWD) was the serogroup most frequently reported
surveillance system (http://ecdc.europa.eu/ throughout the period, followed by O103, O91,
en/activities/diseaseprogrammes/fwd/Pages/ O145, and O111.

CHAPTER 13 Public Health Microbiology of STEC 247
FIGURE 1 Number of STEC infections reported in the period 20072011 to the FWD surveillance system co-
ordinated by the ECDC. The 2011 data do not include the cases due to STEC O104, which occurred in the
framework of the large outbreak that occurred in Germany and other European countries. NT, cases with no
information available on the serogroup of the STEC infecting strain.
Unfortunately, not all cases reported in the group 0 to 4 years. The STEC serogroups
ECDC-FWD databases had the complete set consistently associated with HUS were O157,
of data, and the information on the develop- O26, O103, O145, and O111, whereas the syn-
ment of HUS was unknown for 30 to 40% drome was rarely observed among patients
of them (13). Despite this lack of information, with STEC O91 infection.
a number of HUS cases ranging from 103 In the United States, the Centers for Dis-
in 2007 to 277 in 2011 (not including those ease Control and Prevention (CDC) estab-
associated with the STEC O104 outbreak) lished a surveillance network for cases of
were reported. Most cases occurred in chil- STEC infections and other food-borne dis-
dren, with about 60% occurring in the age eases (FoodNet), with the aim of determining
FIGURE 2 Number of STEC infections in the European Union associated with the serogroups other than O157
and O104 most frequently reported in the period 20072011 to the FWD surveillance system coordinated by
the ECDC. doi:10.1128/microbiolspec.EHEC-0014-2013.f2

248 CAPRIOLI ET AL.
the incidence of laboratory-confirmed in- (17) and has pinpointed significant changes in
fections for bacterial pathogens transmitted the prevalence of HUS-associated serogroups
commonly through food and attributing ill- over time. As shown in Fig. 3, STEC O157 was
nesses to specific sources and settings (www. associated with more than 50% of cases in
cdc.gov/foodnet/). FoodNet was established the first decade of the surveillance. After this
in 1995 in collaboration with the U.S. De- first period, infections with STEC O26 and
partment of Agricultures Food Safety and O145 increased, outnumbering those associ-
Inspection Service and the Food and Drug ated with STEC O157 in the decade 1998 to
Administration. It involves 10 state health de- 2007. Finally, the distribution of cases due to
partments, with a surveillance area including STEC O157 and STEC O26 reached a more
15% of the U.S. population (47 million per- balanced relative figure in the current period
sons). Differently from the FWD, FoodNet (unpublished data from the Italian Registry
is an active surveillance system, with public for HUS, 19882007). Since the methods for
health officials routinely communicating with laboratory diagnosis of STEC infection used
the clinical laboratories to identify new cases. in the surveillance (17) did not change over
Cases of HUS are specifically recorded through time, the observed variations in the preva-
a network of pediatric nephrologists and lence of the HUS-associated serogroups likely
infection-control practitioners on the basis reflect a varying exposure of the population to
of clinical diagnosis (14). Infections with the sources and reservoirs of these organisms.
STEC O157 and non-O157 are included in This hypothesis might have important impli-
the FoodNet surveillance system, and inter- cations from a public health perspective and
estingly, similar incidence values (0.97 and seems to be supported by the observation
1.10 per 100,000, respectively) were recorded that patients with STEC O157 infections were
for the two groups in 2011 (15). However, older (median, 32 months) than those with
among the patients with HUS tested with non-O157 infections (median, 22 months) and
appropriate laboratory methods, the preva- showed a more clear summer peak (unpub-
lence of STEC O157 infections was much lished data from the Italian Registry for HUS,
higher than that of STEC non-O157. In gen- 19882007).
eral, among the STEC infections with the Serogroup O26 has been strongly associat-
serogroup identified, the most common were ed with HUS also in other European countries
O157 (47%), O26 (14%), and O103 (11%). (711) and the United States (18). In Europe,
The FoodNet network is pulled alongside most of the STEC O26 pathogens causing
with PulseNet, a network performing stan-
dardized molecular subtyping of STEC and
other food-borne pathogens by pulsed-field gel
electrophoresis to detect case clusters and to
facilitate the early identification of common-
source outbreaks (www.cdc.gov/pulsenet/).
PulseNet is also coordinated by the CDC and
originated as a North American network of
laboratories, then extended to other similar
networks around the world (16).
Surveillance activities should be main-
tained over time to allow evaluating the trends
of STEC infections and understanding the
FIGURE 3 STEC serogroups associated with the
dynamics of the circulation of serotypes and hemolytic-uremic syndrome in Italy, 19882011.
even specific clones. In Italy, a surveillance Data from the Italian Registry for HUS.
system for HUS has been in place since 1988 doi:10.1128/microbiolspec.EHEC-0014-2013.f3

HUS belong to a highly virulent clone, which These serogroups include O26, O103, O111,
seems to have emerged in the mid-1990s (19), and O145 and have been classified in
with the first reports from Germany (20) and the seropathotype B group of the scheme
Italy (21). This O26 clone belongs to a par- proposed by Karmali et al. (23) (see also
ticular multilocus sequence type, ST29, and A Proactive Approach to Food Control:
harbors the stx2a subtype of Shiga toxin gene Which STEC Should Be Considered Patho-
(19). It has been speculated (19) that its evo- genic? below). In the United States, such a
lutionary success might be due to the ability serogroup-specific approach has been ex-
to lose the stx2a-harboring bacteriophage tended to serogroups O121 and O45 that have
(21) without entering the lytic cycle. As a been considered epidemiologically relevant
matter of fact, such an ability might allow the in that country (18).
avoidance of bacterial lysis following stimu- In the European Union, STEC is included
lation to release the stx2a-harboring bacte- among the zoonotic agents for which moni-
riophage in the human gut and might also toring activities are mandatory, according to
account for a prolonged survival in the envi- Directive 2003/99/EC, which obligates the
ronment. Such stx-negative STEC O26 vari- European Union member states to collect
ants also represent targets for lysogenization relevant and possibly comparable data on
by other stx-harboring phages, explaining zoonoses and zoonotic agents. Data are col-
their greater diffusion with respect to other lected by the European Food Safety Authority
lytic clones. (EFSA) and published in the European Union
Summary Report on Trends and Sources of
Zoonoses, Zoonotic Agents and Food-borne
Monitoring of STEC Along the Food Chain
Outbreaks (711) together with data on human
Monitoring the presence of STEC in the ani- STEC infections. EFSA issued recommen-
mal reservoirs and the potential food vehicles dations for the monitoring of STEC in animals
can provide useful information to identify and food (24) that gave priority to STEC O157
which animals and foodstuffs are the main as the STEC serogroup most frequently asso-
sources of human infections. However, data ciated with severe human infections, in par-
on the prevalence of STEC in food and ani- ticular HUS, in Europe. Monitoring should
mals from different investigations could be then be extended to other STEC serogroups
difficult to compare due to differences in (e.g., O26, O103, O91, O145, and O111) indi-
the sampling strategies and analytical meth- cated by the periodic analysis of the data on
ods employed. As International Organization human infections in Europe. EFSA also issued
for Standardization Technical Specification technical specifications for STEC monitoring
(ISO/TS) 13136:2012, the international stan- activities (25). For animals, the sampling of
dard for the detection of STEC in food, was the hide of young cattle and the fleeces of
published only in late 2012, a variety of de- sheep at slaughter was indicated. For food,
tection methods for STEC have been used sampling of commodities that are perceived
in monitoring programs and official control to be sources of STEC infections was sug-
plans, based on two main approaches: (i) the gested, including beef products that could
detection of any STEC strain present in the be eaten with minimal cooking, ready-to-eat
sample, assessed by the production of Stx fermented meats, fresh produce, raw and low-
and/or the presence of stx genes; and (ii) a heat-treated milk, and derived dairy products.
serogroup-specific detection strategy, aimed The monitoring data reported in the last
in most cases at the detection of E. coli O157 published European Union Summary Report
and, more recently, a few other serogroups on Zoonoses, Zoonotic Agents and Food-borne
that have been consistently associated with Outbreaks (11) confirmed that STEC is mainly
HUS and are capable to cause outbreaks. found in ruminants and products thereof,

250 CAPRIOLI ET AL.
meat and raw milk. The proportion of STEC-

Public Health Reference Laboratories
positive samples can vary widely among coun-
tries, but these differences could be due to the Networks of STEC national reference labora-
differences in the sampling strategies and ana- tories (NRLs) have been established in the
lytical methods. The reported prevalence of industrialized countries and operate in dif-
STEC contamination in vegetables and fruits ferent contexts according to their geographic
was very low, but this probably reflects the distribution.
sampling plans adopted so far; these matrices The services provided by STEC NRLs
included numbers of tested samples much should include confirmation, serotyping, and
lower than those of foods of animal origin. molecular typing of suspected E. coli strains
Despite the problems in the standardiza- isolated by front-line clinical microbiology
tion of sampling and detection methods, the laboratories. NRLs should also participate in
European Union Summary Reports on Trends surveillance activities, research, and dissemi-
and Sources of Zoonoses, Zoonotic Agents and nation of information and provide technical
Food-borne Outbreaks represent an important training and reference strains to front-line
example of integrated medical, veterinary, and laboratories.
food data and can provide a valuable contri- In the European Union, the public health
bution to the source attribution of the burden field, differently from the food and veterinary
of STEC infections due the serotypes and sector, does not refer to specific norms regu-
clones in humans and to the evaluation of lating the asset of the networks of reference
cost-effective control measures. laboratories. Nevertheless, a web of STEC
NRLs has been established over time, com-
posed of scientific institutions historically
The Role of Reference Laboratories
collaborating on this topic (26) and, since
The microbiology of STEC infections is par- 2007, connected within the framework of the
ticularly complex because of the difficulties in ECDC-FWD program described earlier. In
distinguishing STEC from the other ubiqui- such a framework, the European E. coli NRLs
tous and generally harmless types of E. coli. provide the data on STEC infections to the
Moreover, the physiologic and genomic duc- ECDC-FWD database. In turn, the ECDC sup-
tility of these microorganisms may favor the ports the STEC NRL network through the
emergence of new pathogenic clones, hinder- standardization of identification and typing
ing the development of reliable schemes for methods, and the provision of reference ma-
their characterization. Therefore, establishing terials (control strains) and EQA, organized to
specific reference laboratories is of utmost im- evaluate the performance of laboratories, to
portance for the prevention and control of identify areas for improvement in laboratory
STEC infections. methods and to ensure that identification and
Reference laboratories, whether established typing of STEC are carried out uniformly and
within a normative framework or not, are that the results provided to the FWD database
usually appointed to provide reference diag- are comparable. The EQA usually includes the
nostics, reference materials, external quality identification of STEC by detection and typing
assessment (EQA), scientific advice, collabo- of the main virulence genes (stx1, stx2, eae)
ration on research and monitoring, and par- and O:H serotyping (www.ecdc.europa.eu/en/
ticipation in alert and response activities. publications/_layouts/forms/Publication_
Reliable STEC reference laboratories are piv- DispForm.aspx?List=4f55ad51-4aed-4d32-
otal for supporting surveillance activities b960-af70113dbb90&ID=1041).
and for enabling preparedness for the threats In the United States, the FoodNet network
caused by emerging pathogenic clones and supports the 650 clinical laboratories that pro-
epidemic outbreaks. vide data on STEC infections with laboratory

protocols and recommendations and con- established EU-RL to create laboratory net-
ducts periodic surveys to understand the cur- works on the specific hazards. The NRLs
rent diagnostic practices and monitor their collaborate with the EU-RL and coordinate
changes over time (27). the official laboratories responsible for food
analyses in their country, also through the
organization of national EQA. All laboratories
Veterinary/Food Reference Laboratories
included in the network at either the national
The role of reference laboratories is particu- or the European Union level must be accred-
larly important in the veterinary/food sector. ited according to the norm EN ISO IEC
Differently from clinical microbiology, where 17025:2005. The final aim of this cascade
the isolation of any STEC strain from a case system is that the official controls conducted
of infection displaying compatible symptoms on any produced or imported foodstuff are
is sufficient to establish an etiologic diagnosis, carried out using the same state-of-art meth-
in food microbiology the need to assign a rank ods and with comparable levels of proficiency
of risk to the STEC isolates for the humans throughout the European Union territory.
health does not benefit from a clear-cut defi- The EU-RL for E. coli was established in
nition of which STEC types have to be 2006, and at present it coordinates a network
considered as a hazard (see A Proactive Ap- of 34 NRLs designated by the European Union
proach to Food Control: Which STEC Should and EEA member states (www.iss.it/vtec).
Be Considered Pathogenic? below). More- The first aim of the EU-RL was the har-
over, most diagnostic tests for the detection of monization of the identification and typing
these organisms in food vehicles and animal methods for STEC strains used by the veteri-
reservoirs have been specifically developed nary/food NRLs with those used within the
for E. coli O157, taking advantage of its par- network of public health NRLs participating
ticular metabolic and antigenic features (24). in the ECDC-FWD surveillance program to
Conversely, the development of assays to dis- allow the comparison of data referring to
tinguish non-O157 STEC from nonpathogenic STEC strains isolated from human infections
E. coli remains challenging. and from food and animal sources. This was
In the food safety field, the duties of ref- achieved by the organization of five EQA
erence laboratories include the development, schemes on STEC strain identification and
evaluation, and validation of the test methods typing, two of which are conducted jointly
for a reliable detection of STEC in foods, the with the ECDC-FWD network. The methods
provision of reference materials, the organi- used in these EQA studies and the results
zation of EQA programs for the laboratories obtained are available at the EU-RL website
involved in food control, and the scientific (www.iss.it/vtec). The results referring to
and technical assistance to the public health the detection of the STEC main virulence
authorities, particularly in risk assessment genes and to the identification of the main
exercises. STEC serogroups are summarized in Fig. 4
In the European Union, a network of E. coli and Fig. 5, respectively.
NRLs operates within the framework of Reg- As far as the methods for the detection of
ulation (EC) 882/2004, which lays out the STEC in food are concerned, the EU-RL de-
official controls on food and also establishes veloped and published on its website several
the nomination of European Union Reference operating procedures destined to the NRL
Laboratories (EU-RLs) to face specific food network (www.iss.it/vtec). Moreover, it coor-
and feed hazards or specific animal diseases. dinated the collaborative development of an
According to the Regulation (EC) 882/2004 international standard for the detection of
prescripts, each European Union member these microorganisms in food upon mandate of
state must designate its own NRL for each the European Committee for Standardization.

252 CAPRIOLI ET AL.
FIGURE 4 External quality assessment organized by the EU-RL for E. coli on the identication of STEC strains by
detection of their main virulence genes by PCR. For each gene, white bars represent the number of laboratories
that obtained correct results for all the strains included in the test and black bars the number of laboratories
that provided incorrect results or did not perform the assay. doi:10.1128/microbiolspec.EHEC-0014-2013.f4
The developed standard is a horizontal meth- step aimed at isolating the STEC strain re-
od based on the real-time PCR screening of sponsible for the positive PCR reactions (see
food enrichment cultures for the presence of also Monitoring of STEC Along the Food
the virulence genes (stx1, stx2, and eae) and Chain above).
genes specific for five STEC serogroups widely The standard has been published by the
involved in severe human infections in Europe: International Organization for Standardiza-
O157, O26, O103, O111, and O145. Samples tion in November 2012 as a technical specifi-
positive for stx genes are submitted to a further cation (ISO/TS 13136:2012). Starting in 2009,
FIGURE 5 External quality assessment organized by the EU-RL for E. coli on the identication of the STEC
serogroups most involved in human disease in Europe. For each serogroup, white bars represent the number
of laboratories that obtained correct results for all the strains included in the test and black bars the number of
laboratories that provided incorrect results or did not perform the assay.

the analytical approach described by ISO/TS presence of the outbreak strains was urgently
13136:2012 was evaluated in five proficiency- required. Because of the activities carried out
testing schemes organized by the EU-RL and in the previous years, the European labora-
conducted on different artificially contami- tory network already had a suitable screening
nated matrices. The reports of such EQA stud- method to exclude the presence of any type
ies are available at the EU-RL website (www. of STEC in food, tested by several rounds of
iss.it/vtec), and a synopsis of the aggregated EQA. An additional standard operating pro-
results is shown in Fig. 6, which shows an cedure specific for detecting the STEC O104:
increasing trend in the number of partici- H4 outbreak strain was developed by the EU-
pating laboratories and in their analytical RL and released through the EU-RL website 3
performances. days after the occurrence of the outbreak was
The existence of networks of reference communicated (28). DNA samples to be used
laboratories reveals its pivotal value during as positive control in the molecular assays for
food safety crises involving different coun- the detection of STEC O104:H4 were also
tries, when testing food with standardized, prepared and distributed to the NRLs in the
rapid, and reliable methods is essential to following days. During the entire period of
provide the competent authorities the data the crisis, the EU-RL provided continuous
needed to plan appropriate control measures scientific and technical support to the Euro-
and to inform consumers correctly. In this pean Union structures involved in the man-
respect, the European Union network of agement of the crisis. The experience of the
E. coli reference laboratories was challenged STEC O104:H4 outbreak confirmed that the
in 2011 by the occurrence of the major out- activities carried out within the networks of
break of STEC O104:H4 infections (12). As reference laboratories can provide an impor-
usually occurs during food safety crises due tant contribution in terms of preparedness to
to food-borne outbreaks, testing food for the face food safety crises.
FIGURE 6 External quality assessment organized by the EU-RL for E. coli on the detection in food of the STEC
serogroups most involved in human disease in Europe, using the real-time PCR-based ISO/TS 13136 method.
For each step of the procedure, white bars represent the number of laboratories that obtained correct results
for all the samples included in the test and black bars the number of laboratories that provided incorrect
results or did not perform the assay. doi:10.1128/microbiolspec.EHEC-0014-2013.f6

254 CAPRIOLI ET AL.
cause severe disease, milder symptoms, or no

PERSPECTIVES IN PUBLIC HEALTH
disease at all.
Since the 1980s, it has been proposed that
A Proactive Approach to Food Control:
the STEC strains involved in HC and HUS,
Which STEC Should Be
also termed enterohemorrhagic E. coli, were
Considered Pathogenic?
restricted to particular serogroups and were
Adequate risk assessments are crucial for se- characterized by the ability to induce a typical
curing the microbiological safety of food, and lesion in the intestine, termed attaching and
the characterization of the microbial hazards effacing (A/E) and governed by the locus of
represents a milestone in such processes. A enterocyte effacement (LEE) pathogenicity
precise definition of the pathogens provides island (29). The first attempt to build a co-
the basis for their detection and allows the herent classification system for pathogenic
assessment of their prevalence and setting STEC was made in the early 2000s, and it
the objectives for the reduction of the associ- is represented by the scheme developed by
ated risk for human health. The character- Karmali and coworkers (23). Such a scheme
ization of STEC as a food-borne microbial was based on the evaluation of the virulence
hazard is complex and, at present, a defini- and serological features of the strains com-
tion of the characteristics associated with bined with their association with severe dis-
STEC pathogenicity remains a matter of disease and epidemic outbreaks. The integrated
cussion. Whether all STEC strains are patho- analysis of such data introduced the concept
genic has been disputed among scientists of seropathotype and led to the construction
and risk assessors for a long time; the dispute of a paradigm allocating STEC strains isolated
has been partially linked to the broad spec- from either human disease or the animal res-
trum of symptoms associated with STEC in- ervoir (Table 1).
fections, with a clinical manifestation ranging The analysis of human cases of STEC in-
from the severe forms of HUS and hemor- fection notified to the ECDC-FWD surveil-
rhagic colitis (HC) to mild diarrhea and lance system between 2007 and 2011 showed
asymptomatic carriage. This variability in the that STEC strains belonging to serogroups
clinical outcome, together with the STEC O157, O26, O111, O103, and O145 were respon-
heterogeneity at the genomic level and the sible for roughly 90% of the cases of HUS
plausible effect of the general health status occurring yearly in the European Union, con-
of the patient on the development of the firming the applicability of Karmalis scheme
disease, makes it difficult to define unambig- (711, 13). A similar analysis performed in the
uously the features and the genetic back- United States led to the same observation,
ground of STEC that might, respectively, with the exception that the most represented
TABLE 1 Classication of STEC serotypes into seropathotypesa

Frequency of involvement Association with
Seropathotype Relative incidence in outbreaksb severe disease Serotypesc
A High Common Yes O157:H7, O157:NM

B Moderate Uncommon Yes O26:H11, O103:H2, O111:NM,
O121:H19, O145:NM
C Low Rare Yes O91:H21, O104:H21,
O113:H21, other
D Low Rare No Multiple
E Nonhuman only NA NA Multiple
a
b
NA, not applicable.
c
NM, nonmotile.

serogroups also included O121 and O45 (15; Union member states (13). A thorough evalu-
www.cdc.gov/foodnet). However, while the ation of the STEC characteristics recorded
seropathotype model is effective in accom- in the ECDC and EFSA databases in the past
modating the strains from human infections, it 5 years led to the acknowledgment that there
presents limitations when a STEC strain iso- is not a combination of genoserotypes that
lated from a potential vehicle of infection has identifies all pathogenic STEC strains, even
to be assigned to a seropathotype or has to be when the field is narrowed to the STEC
generally evaluated for its wherewithal of causing HC or HUS. It was recognized that
representing a hazard. As a matter of fact, some combination of virulence genes, such as
STEC strains possessing the LEE pathoge- particular stx2 gene subtypes (32), together
nicity island are commonly found in animals, with the LEE pathogenicity island, might be
the environment, or food samples, but they associated with HUS (33, 34). However, while
can belong to serotypes different from those this observation seems to be consistent for
indicated in the seropathotype scheme. Simi- the stx2 subtypes, the role of the LEE in the
larly, STEC strains belonging to the serotypes colonization of the human intestinal mucosa
included in groups A and B, those allocating could be taken over by the action of other
the most hazardous STEC strains, can possess adhesion factors, as demonstrated by the
a partial asset of virulence genes, and, theo- STEC O104:H4 outbreak in 2011 (28, 31). More-
retically, they could be considered to be less or over, it was highlighted that patient-associated
not pathogenic at all. Finally, given the high factors, such as age, immune status, and the
genomic plasticity of E. coli, new STEC vari- administration of antibiotic therapy in the
ants may emerge, with characteristics com- early course of infection (13), have an impor-
pletely different from those included in the tant role in the final outcome of the STEC
seropathotype concept (28, 30). infection. Based on these considerations, it
In spite of the limitations in categorizing was recognized that neither the seropatho-
the level of danger associated with STEC from type concept nor the analysis of surveillance
nonhuman sources, the seropathotype con- data allows a certain definition of the patho-
cept raised a large consensus in the scientific genicity of a STEC strain or its level of danger
community and was endorsed by EFSA, which to human health, serving as a proactive tool to
recommended focusing food testing for STEC protect humans health. Such uncertainty also
on the seropathotype A and B groups (24, 25). applies to molecular risk assessments based
However, the massive outbreak caused by on the evaluation of the genetic asset only,
an enteroaggregative STEC O104:H4 strain since there are no specific genetic markers
in 2011 in Germany (12) has permanently that, alone or in combination, can assign un-
questioned the efficacy of the seropathotype ambiguously the status of food-borne micro-
scheme to categorize STEC from food for biological hazard to a STEC strain (13).
the adoption of intervention measures. As a
matter of fact, the STEC O104 outbreak strain
Regulatory Perspectives in the European
was undoubtedly the most pathogenic STEC
Union and the United States
ever described and yet it did not fit the sero-
pathotype A and B groups, being negative for The uncertainties of STEC pathogenicity dis-
the LEE and belonging to a serotype not in- cussed above hindered the development of
cluded in the scheme (28, 31). control measures to limit the presence of
The crisis resulting from the STEC O104 STEC in food and the issuing of specific mi-
outbreak forced the risk managers to recon- crobiological criteria for food safety. Never-
sider the pathogenicity assessment for STEC, theless, some steps toward food-safety rules
and EFSA conducted a new risk assessment have been made, often as a reaction to dra-
exercise at the request of some European matic events such as outbreaks causing deaths

256 CAPRIOLI ET AL.
and with large mass media impact. The first of microbiological criteria for STEC in sprouts
official measure regarding STEC was taken and the methodology to be adopted for the
in the United States at the end of the 20th conformity assessment of this food commod-
century. In October 1994, the U.S. Food ity. The entire process took about 1 year and
Safety and Inspection Service declared E. coli resulted in the issuance of Regulation (EU)
O157:H7 an adulterant in ground beef in re- 209/2013, containing the microbiological cri-
sponse to the multistate outbreak caused in teria for STEC in sprouts and amending
1993 by undercooked hamburgers contami- Regulation (EC) 2073/2005, which lists all
nated with STEC O157 (35). The pronounce- microbiological criteria for the assessment of
ment was extended in 1999 to all nonintact the safety of food and the verification of the
raw beef products (36). In the 1990s, it was a process hygiene criteria in the European
popular opinion in the United States that Union. Regulation (EU) 209/2013 introduced
E. coli O157 was the only STEC serotype for the first time in the European Union leg-
causing human disease, and such a viewpoint islation a specific criterion for STEC regard-
lasted for more than 10 years. However, the ing the presence in sprouts of the five STEC
growing number of infections and outbreaks serogroups included in the seropathotypes A
caused by STEC non-O157 led to the assess- and B of Karmalis scheme (23) plus STEC
ment that E. coli O157 was not the only STEC O104:H4. One of the regulations recitals ex-
representing a hazard (18) and the Food Safety plained that the reasons for the restriction to
and Inspection Service decided on the im- certain STEC groups resided in the observa-
plementation of sampling and testing of beef tion that STEC O157, O26, O103, O111, and
manufacturing trimmings for STEC non-O157 O145 are recognized as causing most of the
starting June 4, 2012 (37). STEC strains be- HUS cases in the European Union, whereas
longing to serogroups O26, O111, O103, O145, STEC O104:H4 caused the large 2011 out-
O121, and O45 were declared as adulterant in break. Such a criterion might appear in op-
these food commodities and included in the position to the conclusions of the newly
sampling plans in addition to E. coli O157. released EFSA opinion on STEC pathogenicity
Reactions included the withdrawal from the assessment (13) (see also A Proactive Ap-
market of the positive batches. proach to Food Control: Which STEC Should
In the European Union, the introduction Be Considered Pathogenic? above). However,
of food-safety rules related to STEC was in- in agreement with such an opinion, the same
cited by the high impact of the STEC O104:H4 regulation recital stated: It cannot be ex-
outbreak (12), linked to the consumption of cluded that other STEC serogroups may be
contaminated sprouts (38). The high number pathogenic to humans as well. In fact, such
of casualties and HUS cases, together with STEC may cause less severe forms of disease
the attention of public opinion and the back- such as diarrhoea and or bloody diarrhoea
lash affecting the trade of vegetable products, or may also cause HUS and therefore repre-
forced the European Commission to take mea- sent a hazard for the consumers health. This
sures against the possibility that other STEC last sentence widens the concept of pathoge-
crises could occur in the European Union. nicity to all STEC and refers to the food
EFSA was asked to perform a risk assessment business operator the choice of releasing on
exercise on the presence of STEC and other the market sprouts positive for the presence of
pathogenic microorganisms in sprouts and STEC that do not fit the microbiological cri-
seeds intended for sprouting (38). At the terion. This approach, apparently contradict-
same time, a technical working group, in- ing the role of the competent authorities in
volving experts from the European Union ensuring food safety, is in agreement with
member states and the EU-RL for E. coli, one of the main principles laid down in the
discussed the issues related to the definition general food law [Regulation (EC) 178/2004]

that assigns to the food business operators 2. Ferens WA, Hovde CJ. 2011. Escherichia coli
the responsibility to place safe food on the O157:H7: animal reservoir and sources of human
infection. Foodborne Pathog Dis 8:465487.
market.
3. Caprioli A, Morabito S, Brugere H, Oswald
In conclusion, securing food safety is a com- E. 2005. Enterohaemorrhagic Escherichia coli:
plex matter that becomes even more compli- emerging issues on virulence and modes of
cated when dealing with STEC. As a matter of transmission. Vet Res 36:289311.
fact, STEC represents one of the most elusive 4. Garcia A, Fox JG, Besser TE. 2010. Zoonotic
enterohemorrhagic Escherichia coli: a One Health
pathogens in terms of phenotypic character-
perspective. ILAR J 51:221232.
istics and genomic arrangement, and a con- 5. Ong KL, Apostal M, Comstock N, Hurd S,
tinuous challenge for the laboratories in Webb TH, Mickelson S, Scheftel J, Smith G,
charge of assessing the food safety by apply- Shiferaw B, Boothe E, Gould LH. 2012. Strate-
ing analytical controls. Furthermore, the food gies for surveillance of pediatric hemolytic uremic
market is a dynamic entity, always chasing syndrome: Foodborne Diseases Active Surveil-
lance Network (FoodNet), 20002007. Clin Infect
the most available and convenient sources of Dis 54(Suppl 5):S424S431.
food commodities capable of satisfying the 6. Lockary VM, Hudson RF, Ball CL. 2007. Shiga
need for cheap food and consumers demand toxin-producing Escherichia coli, Idaho. Emerg
for exotic flavors. This often results in pro- Infect Dis 13:12621264.
viders of food and raw materials from devel- 7. European Food Safety Authority, European
Centre for Disease Prevention and Control.
oping countries (39) introducing, from time to
2009. The Community Summary Report on
time, pathogens such as the STEC O104:H4 trends and sources of zoonoses and zoonotic
that caused the German outbreak in 2011, agents in the European Union in 2007. EFSA
bringing into question the food safety and Journal 7:223.
public health systems. 8. European Food Safety Authority, European
Coping with such complex challenges re- Centre for Disease Prevention and Control.
2010. The Community Summary Report on trends
quires the interplay of the different health and sources of zoonoses, zoonotic agents and
care sectors. In fact, it is crucial that all roles food-borne outbreaks in the European Union in
of the risk assessment, evaluation, and man- 2008. EFSA Journal 8:1496.
agement referring to both the public health 9. European Food Safety Authority, European
and the food and veterinary fields collaborate, Centre for Disease Prevention and Control.
2011. The European Union Summary Report on
enforcing the concept of One Health. trends and sources of zoonoses, zoonotic agents
and food-borne outbreaks in 2009. EFSA Journal
9:2090.
ACKNOWLEDGMENT 10. European Food Safety Authority, European
We declare no conflicts of interest with regard
2012. The European Union Summary Report on
to the manuscript. trends and sources of zoonoses, zoonotic agents
and food-borne outbreaks in 2010. EFSA Journal
10:2597.
CITATION 11. European Food Safety Authority, European
Caprioli A, Scavia G, Morabito S. 2014. Public 2013. The European Union Summary Report on
health microbiology of Shiga toxin-producing trends and sources of zoonoses, zoonotic agents
Escherichia coli. Microbiol Spectrum 2(4): and food-borne outbreaks in 2011. EFSA Journal
EHEC-0014-2013. 11:3129.
12. Buchholz U, Bernard H, Werber D, Bohmer
MM, Remschmidt C, Wilking H, Delere Y,
REFERENCES an der Heiden M, Adlhoch C, Dreesman J,
Ehlers J, Ethelberg S, Faber M, Frank C, Fricke
1. Tarr PI, Gordon CA, Chandler WL. 2005. Shiga- G, Greiner M, Hohle M, Ivarsson S, Jark U,
toxin-producing Escherichia coli and haemolytic Kirchner M, Koch J, Krause G, Luber P, Rosner
uraemic syndrome. Lancet 365:10731086. B, Stark K, Kuhne M. 2011. German outbreak of

258 CAPRIOLI ET AL.
Escherichia coli O104:H4 associated with sprouts. 21. Ricotti GC, Buonomini MI, Merlitti A, Karch H,
N Engl J Med 365:17631770. Luzzi I, Caprioli A. 1994. A fatal case of hemor-
13. European Food Safety Authority, Panel on rhagic colitis, thrombocytopenia, and renal failure
Biological Hazards (BIOHAZ). 2013. Scientific associated with verocytotoxin-producing, non-
Opinion on VTEC-seropathotype and scientific O157 Escherichia coli. Clin Infect Dis 19:815816.
criteria regarding pathogenicity assessment. EFSA 22. Bielaszewska M, Prager R, Kock R, Mellmann
Journal 11(4). A, Zhang W, Tschape H, Tarr PI, Karch H.
14. Gould LH, Demma L, Jones TF, Hurd S, Vugia 2007. Shiga toxin gene loss and transfer in vitro
DJ, Smith K, Shiferaw B, Segler S, Palmer A, and in vivo during enterohemorrhagic Escherichia
Zansky S, Griffin PM. 2009. Hemolytic uremic coli O26 infection in humans. Appl Environ
syndrome and death in persons with Escherichia Microbiol 73:31443150.
coli O157:H7 infection, foodborne diseases active 23. Karmali MA, Mascarenhas M, Shen S, Ziebell
surveillance network sites, 20002006. Clin Infect K, Johnson S, Reid-Smith R, Isaac-Renton J,
Dis 49:14801485. Clark C, Rahn K, Kaper JB. 2003. Association
15. Centers for Disease Control and Prevention. of genomic O island 122 of Escherichia coli EDL
Foodborne Diseases Active Surveillance Network 933 with verocytotoxin-producing Escherichia
(FoodNet): FoodNet Surveillance Report for 2011 coli seropathotypes that are linked to epidemic
(Final Report), 2012. U.S. Department of Health and/or serious disease. J Clin Microbiol 41:4930
and Human Services, Centers for Disease Control 4940.
and Prevention, Atlanta, GA. Available from 24. European Food Safety Authority, Panel on
http://www.cdc.gov/foodnet/PDFs/2011_annual_ Biological Hazards (BIOHAZ). 2007. Scientific
report_508c.pdf. Opinion of the Panel on Biological Hazards on a
16. Swaminathan B, Gerner-Smidt P, Ng LK, request from EFSA on monitoring of verotoxi-
Lukinmaa S, Kam KM, Rolando S, Gutierrez genic Escherichia coli (VTEC) and identification
EP, Binsztein N. 2006. Building PulseNet Inter- of human pathogenic VTEC types. EFSA Journal
national: an interconnected system of laboratory 579:161.
networks to facilitate timely public health recog- 25. European Food Safety Authority, Panel on
nition and response to foodborne disease outbreaks Biological Hazards (BIOHAZ). 2009. Technical
and emerging foodborne diseases. Foodborne specifications for the monitoring and reporting of
Pathog Dis 3:3650. verotoxigenic Escherichia coli (VTEC) on animals
17. Tozzi AE, Caprioli A, Minelli F, Gianviti A, and food (VTEC surveys on animals and food) on
De Petris L, Edefonti A, Montini G, Ferretti A, request of EFSA. EFSA Journal 7:1366.
De Palo T, Gaido M, Rizzoni G. 2003. Shiga 26. Fisher IS. 1999. The Enter-net international sur-
toxin-producing Escherichia coli infections asso- veillance networkhow it works. Euro Surveill
ciated with hemolytic uremic syndrome, Italy, 4:5255.
19882000. Emerg Infect Dis 9:106108. 27. Hoefer D, Hurd S, Medus C, Cronquist A,
18. US Department of Agriculture, Food Safety and Hanna S, Hatch J, Hayes T, Larson K,
Inspection Service. Risk Profile for Pathogenic Nicholson C, Wymore K, Tobin-DAngelo M,
Non-O157 Shiga Toxin-Producing Escherichia coli Strockbine N, Snippes P, Atkinson R, Griffin
(non-O157 STEC). 2012. Available from http://www. PM, Gould LH. 2011. Laboratory practices for the
fsis.usda.gov/wps/wcm/connect/92de038d-c30e- identification of Shiga toxin-producing Esche-
4037-85a6-065c3a709435/Non_O157_STEC_ richia coli in the United States, FoodNet sites,
Risk_Profile_May2012.pdf?MOD=AJPERES. 2007. Foodborne Pathog Dis 8:555560.
19. Bielaszewska M, Mellmann A, Bletz S, Zhang 28. Scheutz F, Nielsen EM, Frimodt-Moller J,
W, Kock R, Kossow A, Prager R, Fruth A, Boisen N, Morabito S, Tozzoli R, Nataro JP,
Orth-Holler D, Marejkova M, Morabito S, Caprioli A. 2011. Characteristics of the entero-
Caprioli A, Pierard D, Smith G, Jenkins C, aggregative Shiga toxin/verotoxin-producing Esch-
Curova K, Karch H. 2013. Enterohemorrhagic erichia coli O104:H4 strain causing the outbreak of
Escherichia coli O26:H11/H: a new virulent haemolytic uraemic syndrome in Germany, May to
clone emerges in Europe. Clin Infect Dis 56:1373 June 2011. Euro Surveill 16(24).
1381. 29. Levine MM. 1987. Escherichia coli that cause di-
20. Zhang WL, Bielaszewska M, Liesegang A, arrhea: enterotoxigenic, enteropathogenic, entero-
Tschape H, Schmidt H, Bitzan M, Karch H. invasive, enterohemorrhagic, and enteroadherent.
2000. Molecular characteristics and epidemio- J Infect Dis 155:377389.
logical significance of Shiga toxin-producing 30. Colic E, Dieperink H, Titlestad K, Tepel M.
Escherichia coli O26 strains. J Clin Microbiol 2011. Management of an acute outbreak of diar-
38:21342140. rhoea-associated haemolytic uraemic syndrome

with early plasma exchange in adults from 35. Barrett TJ, Lior H, Green JH, Khakhria R,
southern Denmark: an observational study. Lan- Wells JG, Bell BP, Greene KD, Lewis J, Griffin
cet 378:10891093. PM. 1994. Laboratory investigation of a multistate
31. Bielaszewska M, Mellmann A, Zhang W, Kock food-borne outbreak of Escherichia coli O157:H7
R, Fruth A, Bauwens A, Peters G, Karch H. 2011. by using pulsed-field gel electrophoresis and
Characterisation of the Escherichia coli strain as- phage typing. J Clin Microbiol 32:30133017.
sociated with an outbreak of haemolytic uraemic 36. US Department of Agriculture, Food Safety and
syndrome in Germany, 2011: a microbiological Inspection Service. 1999. FSIS Policy on Non-
study. Lancet Infect Dis 11:671676. intact Raw Beef Products Contaminated with
32. Scheutz F, Teel LD, Beutin L, Pierard D, E. coli O157:H7. Available from http://www.fsis.
Buvens G, Karch H, Mellmann A, Caprioli A, usda.gov/Oa/background/O157policy.htm.
Tozzoli R, Morabito S, Strockbine NA, Melton- 37. US Department of Agriculture, Food Safety and
Celsa AR, Sanchez M, Persson S, OBrien AD. Inspection Service. 2011. Shiga toxin-producing
2012. Multicenter evaluation of a sequence-based Escherichia coli in certain raw beef products. Fed
protocol for subtyping Shiga toxins and stan- Regist 76:5815758165.
dardizing Stx nomenclature. J Clin Microbiol 38. European Food Safety Authority, Panel on
50:29512963. Biological Hazards (BIOHAZ). 2011. Scientific
33. Friedrich AW, Bielaszewska M, Zhang WL, Opinion on the risk posed by Shiga toxin-
Pulz M, Kuczius T, Ammon A, Karch H. 2002. producing Escherichia coli (STEC) and other
Escherichia coli harboring Shiga toxin 2 gene pathogenic bacteria in seeds and sprouted seeds.
variants: frequency and association with clinical EFSA Journal 9:2424.
symptoms. J Infect Dis 185:7484. 39. European Food Safety Authority. 2011. Tracing
34. Persson S, Olsen KE, Ethelberg S, Scheutz F. seeds, in particular fenugreek (Trigonella foenum-
2007. Subtyping method for Escherichia coli Shiga graecum) seeds, in relation to the Shiga toxin-
toxin (verocytotoxin) 2 variants and correlations producing E. coli (STEC) O104:H4 2011 outbreaks
to clinical manifestations. J Clin Microbiol 45: in Germany and France. Available from http://
20202024. www.efsa.europa.eu/it/supporting/doc/176e.pdf.

DIAGNOSIS, DETECTION,
AND STRAIN CHARACTERIZATION

Detection of Shiga Toxin-Producing
Escherichia coli from Nonhuman
Sources and Strain Typing
LOTHAR BEUTIN1 and PATRICK FACH2

14
INTRODUCTION
Shiga toxin (verotoxin)-producing Escherichia coli (STEC) strains were first

described in 1977 by their ability to cause cytotoxic effects on Vero cells (1).
In the early days, production of Shiga toxin (Stx) by E. coli was thought to be
associated with certain human enteropathogenic E. coli (EPEC) strains (13).
STEC was recognized as a zoonotic agent when the first known outbreaks
occurred in 1982. STEC O157:H7, a rare serotype of E. coli, was isolated from
patients developing hemorrhagic colitis (HC) after they ingested undercooked
beef in restaurant chains (4). STEC O157 was isolated from the incriminated
beef, indicating a possible transmission from a bovine reservoir. In the following
years cattle were identified as a worldwide natural reservoir for STEC O157
and non-O157 strains and as an important source of food contamination (58).
Repeated sampling of cattle revealed that the agent was occasionally present
in the majority of cattle farms in Europe and America (9). However, recent
findings on the epidemiology of enteroaggregative Shiga toxin-producing
E. coli (EAEC-STEC) O104:H4 indicate that not all STEC strains have a zoonotic
origin (10, 11).
1
National Reference Laboratory for Escherichia coli, Department of Biological Safety, Federal Institute for Risk Assessment
(BfR), D-12277 Berlin, Germany; 2Food Safety Laboratory, ANSES (French Agency for Food, Environmental and Occu-
pational Health and Safety), F-94706 Maisons-Alfort, France.
263

264 BEUTIN AND FACH
biotic factors and the physicochemical condi-

STEC in Food, Animals,
tions in general (27). The dissemination of
and the Environment
STEC is also promoted by the occurrence of
With the improvements in STEC detection free stx phage particles in the environment
methods, the search for natural STEC reser- (28). Transfer of stx genes to E. coli was found
voirs and for E. coli serotypes associated with to occur by free Stx phages in food and water,
Stx production intensified. STEC was identi- giving numerous possibilities for the genera-
fied as a frequent colonizer of domestic and tion and spread of STEC strains (29).
wild ruminants such as cattle, sheep, goats, The ecology of STEC in animals and its
and deer (9, 1216). Some nonruminant do- relative stability to physicochemical stress
mestic and wild animals such as pigs, wild conditions have the effect that this agent is
boar, and hares were also identified as im- found in practically all kinds of foodstuffs,
portant sources and natural reservoirs of se- animals, and environments. This also includes
rologically diverse STEC types (12, 15, 16). geographical areas that are not used for agri-
Today, it is known that STEC is present as cultural activities.
a resident of the gut flora or transiently car-
ried by many species of mammals, as well as
Diversity of STEC Types
birds, insects, mollusks, and fish (17). Con-
tamination of food of animal origin occurs After the first outbreaks with Stx-producing
frequently at the critical stages of food pro- O157:H7 became known, the search for
duction, such as milking, slaughtering, and STEC concentrated mainly on this particu-
evisceration. Accordingly, numerous types of lar serotype. However, investigations of ani-
STEC are found in meat and milk products mal reservoirs and human patients infected
from domestic and wildanimals (8, 18). Fecal with STEC revealed an enormous diversity of
shedding of STEC from infected animals E. coli serotypes associated with Stx produc-
results in wide contamination of farmland, tion. A compilation published in 2005 of STEC
water, and fresh produce (19). Wastewater strains isolated from human patients listed
processing plants do not eliminate STEC com- more than 400 serotypes (30). It is likely that
pletely from treated water (20). As a conse- the number of STEC serotypes from patients
quence, STEC may be present in irrigation has increased. A review article from 2007
water, resulting in surface contamination of listed 171 E. coli O groups as associated with
herbal foodstuffs such as vegetables and fruits. Stx production. The STEC strains were from
Effluents from farmland containing STEC con- diseased humans and healthy cattle and sheep
taminate coastal waters and shellfish (21, 22). (31). In summary, with a few exceptions,
Birds and insects play an important role in E. coli strains of virtually all known O sero-
disseminating STEC over long distances, thus groups were already found associated with Stx
allowing the spread of the agent over larger production. In addition, many STEC strains
geographical areas (17). Direct and indirect with untypeable (unknown) O antigens were
transmission of STEC from animal to animal isolated from patients, animals, and food, in-
and from animal to humans was observed for dicating that the serotype diversity of STEC is
O157 and other STEC types (17, 2325). STEC beyond the current serotype scheme, includ-
of various serotypes was found to persist on ing more than 182 O antigens and 53 H anti-
contaminated farmland, pen floors, and the gens (18, 30, 31, 33).
skin of animals, and evidence for horizontal On the other hand, epidemiological data
dissemination of STEC clones and recontam- on STEC serotypes isolated from humans in-
ination of animals was found (26). Once in the dicate that a small number of STEC serotypes
environment, STEC may survive in soil and predominate in patients with severe illness (34
water for weeks and months, depending on 36). Most of these STEC strains belong to the

CHAPTER 14 Detection of STEC from Nonhuman Sources 265
group of enterohemorrhagic E. coli (EHEC) detect virtually all STEC types from any kind
strains and mainly to serogroups O26, O45, of sample (44, 45). Over the last 3 decades, Stx
O103, O111, O121, O145, and O157. EHEC is detection assays were developed. These assays
most frequently implicated in severe disease can be grouped into tissue culture cytotoxicity
such as HC and hemolytic-uremic syndrome assays, immunological assays, and DNA-based
(HUS) (33, 36). In contrast to other STEC methodologies.
strains, EHEC strains were found to be similar Cytotoxicity assays and immunological
to each other for the presence of genes involved assays are phenotypic assays indicating the
in the intestinal attaching and effacing (LEE) synthesis and the relative amount of Stx pro-
mechanism and for non-LEE (nle) effector duced by the bacterium. DNA-based meth-
genes. Accordingly, a molecular risk assessment odologies deal with the molecular detection
(MRA) concept was developed to assess the of stx genes. Shiga toxins comprise a growing
relative pathogenicity of STEC strains accord- family of genes with enormous type diversity
ing to the presence of virulence attributes and (46). The Stx family splits into two major
their frequency in human infections (3740). branches, Stx1 and Stx2, which are immuno-
Other STEC strains lacking LEE and non- logically not cross-reactive and show about
LEE effectors were also found to cause HC and 55% difference in their amino acid sequences
HUS in humans. Among these, STEC strains of (47). The Stx1 family divides into four major
serogroups O91, O104, O113, O128, and O146 subtypes (Stx1, Stx1a, Stx1c, and Stx1d), which
are most frequently mentioned (3336, 41). In further split into a number of genetic variants.
contrast to EHEC, these STEC strains are less The Stx2 family branches into seven major
well defined for the molecular basis of their subtypes (Stx2a, Stx2b, Stx2c, Stx2d, Stx2e,
human pathogenicity (42). However, the out- Stx2f, and Stx2g), which subdivide into a total
break with EAEC-STEC O104:H4 strains in of 93 genetic variants (46). Because of the
Europe showed that effective colonization of enormous type diversity of Stx1 and Stx2
the human intestine together with production family toxins, problems can arise in detecting
of Stx2a is sufficient to cause HC and HUS (11, all toxin subtypes with both serological and
43). Therefore, it is likely that other serotypes genetic assays.
of STEC strains might emerge as severe human
pathogens in the future. As a consequence, the
Cytoxicity Assays for Stx
search for STEC should a priori not be re-
stricted to a limited range of E. coli serotypes. Cytotoxicity tests are regarded as the gold
Accordingly, this report concentrates on both standard for Stx testing as only these tests
STEC detection procedures independent of the indicate the cytotoxic effect on mammalian
serotype and methods specifically developed cells (48). They are commonly performed
for detection of EHEC O157:H7 and major with Vero cells but require neutralization tests
non-O157 EHEC serogroups (O26, O45, O103, with Stx1- and Stx2-specific antisera to con-
O111, O121, and O145) that are associated with firm the specificity of the result (49). HeLa
severe illness in humans worldwide. Methods cells might also be used but were shown to
specifically developed for the detection of be less sensitive to certain Stx variants (50).
EHEC O157:H7 only are not addressed here. Cytotoxicity tests are suitable for demon-
strating the biological activity of Stx or the
expression of the stx genes detected by sero-
STEC DETECTION STRATEGIES logical or DNA-based assays, respectively (51,
52). Comparative studies have shown that
Because of the variety of STEC serotypes and cytotoxicity tests are more sensitive for de-
phenotypes, the identification of Shiga toxins tection of Stx than immunological assays (48).
or the underlying genes is the only way to The detection limits of any test system can

266 BEUTIN AND FACH
be critical if the samples contain low amounts Stx ELISA and an immunochromatographic
of Stx or if the assay fails to detect certain lateral flow system were used to compare
subtypes of Stx1 or Stx2 (51, 53, 54). Cyto- representative strains covering all known toxin
toxicity tests can be used for direct Stx testing types, and all Stx variants, except Stx2f, were
from fecal samples where immunological detectable. Except for Stx2f, the false-negative
methods can fail (53). However, fecal, food, results were assumed to be due to low toxin
and environmental samples may contain production, which was below the detection
mixtures of bacteria that can produce entero- limit of the serological assay (59). Stx ELISA
toxins, cytolysins (hemolysins), or other cyto- and the immunochromatographic flow sys-
toxins different from Stx but equally active tem detected <10 CFU of EHEC (O26, O103,
on Vero or HeLa cells (50). Without Stx- O111, O118, O121, O145, and O157) per 25 g in
neutralization assays, these kinds of samples enrichment cultures grown from artificially
may produce confounding results. contaminated salad samples (59).
On the other hand, test samples containing Detection limits of different serological
glycolipids (Gb3 + Gb4) that bind to the Stx tests were reported as between 0.5 ng/ml to
receptor (B-subunit) can neutralize the cyto- 126 ng/ml, depending on the test system and
toxic effect in cell culture assays (55, 56). For the Stx types analyzed (58, 62, 63). Stx2f is
routine screening of test samples, cytotoxicity genetically most distant from all other Shiga
tests are less frequently used because they are toxins (46). It is therefore possible that the
labor-intensive, need specific equipment and Stx2-specific antibodies used in most immu-
personnel skills, and require 72 h before final nological assays do not react with the Stx2f
conclusive reading (48). variant (59, 60). The combination of a sero-
logical test with a PCR assay (Immuno-PCR)
was shown to decrease the detection limit
Immunological Assays for Stx
for Stx2 from 1 ng/ml to 10 pg/ml (64). PCR-
A large number of immunological assays for ELISA was reported to increase 100-fold the
the detection of Stx have been developed, and sensitivity of a conventional Stx PCR assay
many are available commercially. These sero- (65). A PCR-ELISA technique was success-
logical assays are enzyme-linked immunosor- fully used to screen and isolate STEC from
bent assays (ELISAs), reverse passive latex naturally contaminated dairy products and
agglutination assays, immunochromatographic retail minced-beef samples (66, 67).
lateral flow test systems, and the colony im- Immunological assays, such as the verocyto-
munoblot. Capture of Stx from the sample toxigenic E. coli (VTEC) reverse passive latex
may occur through the P1 glycoprotein, Gb3 agglutination assay, were found sufficiently
receptor, or monoclonal antibodies directed sensitive to identify single Stx-producing colo-
against Stx1 and Stx2 (50). nies from agar plates after treatment with
Some commercially available tests were polymyxin B (68). When larger numbers of
evaluated for their sensitivity and specificity colonies need to be screened for the presence
(51, 53, 54, 5759). A number of serological of STEC, the colony immunoblot offers the
assays gave weak reactions or did not detect possibility of detection and isolation in a one-
some strains expressing Stx variants such step procedure (44, 6972) (Fig. 1). Commer-
as Stx2c, Stx2e, Stx2f, and Stx2g (51, 57, cially available hybridoma cell lines for Stx1
59, 60). Significant differences in the per- (ATCC CRL-1794) and Stx2 (ATCC CRL-1907)
formance of some commercial Stx ELISAs proved to be successful for identification of
were revealed by an interlaboratory study different STEC strains from food and fecal
performed on beef samples inoculated with samples (69, 70, 73). Food inspection laborato-
STEC strains producing different kinds of ries in Germany widely use the colony immu-
Stx (61). noblot since, according to German legislation,

further toxin subtyping. PCR-RFLP was also

helpful in detecting multiple copies of stx2
variant genes, which are frequently present
in certain STEC strains (86, 87). As universal
PCR primers fail to detect some Stx subtypes
such as stx2f, a minimum of one primer pair
for the Stx1 family and two primer pairs for
the Stx2 family were found necessary to cover
all known Shiga toxin variants (46, 88).
Conventional multiplex PCR assays were
developed to detect STEC and were success-
fully tested on bacterial isolates (89). Con-
ventional PCR assays require separation of
amplified DNA molecules by gel electropho-
resis followed by interpretation of the re-
sults from agarose gels. The interpretation of
PCR results can cause problems in assays
FIGURE 1 Identication and isolation of STEC from performed with DNA obtained from mixed
mixed samples of bacteria with the Stx colony
cultures of bacteria. Results were ambiguous
immunoblot. Mauve-stained Stx2-positive STEC col-
onies are detected from a sample containing STEC in testing enrichment cultures from meat sam-
mixed with Stx-negative bacteria. ples with a common primer system for stx1
doi:10.1128/microbiolspec.EHEC-0001-2013.f1 and stx2 (90) (Fig. 2). The same stx primers
were used in a study of STEC in fecal samples
of human healthy volunteers, resulting in
all STEC types present in food must be re- 13.9% stx-PCR-positive samples; however,
garded as potentially hazardous (61). STEC could be detected only in 1 of the 23 stx-
PCR-positive samples (91). New rapid PCR
assays that do not need gel electrophoresis,
DNA-Based Assays for Stx and
such as loop-mediated isothermal amplifica-
Other Virulence Markers of STEC
tion, were used to detect 1 to 20 CFU of
DNA-based assays to detect STEC are rapid EHEC/25 g by detecting serogroup-associated
to perform and have the advantage of being genes in spiked meat and produce samples
independent of the availability of specific Stx and could be an alternative to conventional
antisera and cell culture facilities. Universal PCR assays (92).
Stx1 and Stx2 gene probes (74, 75) were de- Real-time PCR assays overcome the gel
veloped and found useful for simultaneous electrophoresis step and have the advantage
detection and isolation of STEC in colony that the PCR is combined with gene probe
hybridization assays with different kinds of hybridization, thus ensuring a quantitative,
samples (50, 7680). For general detection of more specific, and very rapid detection assay.
STEC from all types of samples, common stx In the last decade, real-time PCR systems
PCR assays covering the different stx variant for universal STEC screening became more
genes were developed (67, 8184). Consider- important for examining clinical, food, and
able differences in specificities were found environmental samples (9397). Two sets of
when some PCR and PCR-restriction frag- primers were successfully used for real-time
ment length polymorphism (RFLP) methods PCR using SYBR Green to detect different
were compared (85). PCR-RFLP protocols use variants of Stx1 and Stx2 genes (98). In an
specific differences in the restriction endo- evaluation study on a large number of STEC
nuclease patterns of stx PCR products for and non-STEC strains, two degenerated

268 BEUTIN AND FACH
O121, O145, and O157 (1 to 10 CFU/25 g) in

spiked salad samples and in sprouted seeds
naturally contaminated with STEC O104:H4
(102). A large number of tests based on real-
time PCR amplifications have been developed,
and many are available commercially. How-
ever, a comparison of different sets of com-
mercialized real-time PCR assays for the
detection of STEC revealed significant differ-
ences in their specificity and sensitivity values
(103).
Most promising with real-time PCR sys-
tems is the simultaneous detection of the
characteristic virulence markers and pheno-
typic traits of EHEC strains, permitting their
rapid identification from mixed cultures in
a multiplex assay. Two reference methods
based on real-time PCR detection of the most
important EHEC types of strains from food
FIGURE 2 Suitability of conventional PCRs for detec- have been developed. One has been stan-
tion of stx genes in enriched samples from minced dardized at European (Commission for Euro-
meat. The same panel of enrichment cultures (#84
pean Normalization, CEN) and international
92) was tested with two PCR systems using the MK1/
MK2 primer system (A) and the Lin-up/Lin-down (International Organization for Standardiza-
primer system (B) (90). Primers MK1/MK2 generate tion, ISO) levels (104). The other is an official
230-bp-long PCR products (arrow), and a ladder of method used in the United States for meat
nonspecic bands of different sizes is visible in al- products (MLG 5B.01) (105). Both methods
most all meat samples. (B) Lin-up/Lin-down gener-
rely on the detection of the most important
ates 900-bp size PCR products (arrow), and only two
STEC-positive meat samples (#88+92) give corre- EHEC types. STEC strains are considered
sponding PCR products. M, molecular weight stan- highly pathogenic when they combine the
dard, K, positive STEC control; arrow, position of presence of stx1 or stx2 and eae genes and
the specic PCR product. belong to one of the serotypes O157:H7,
O26:H11, O145:H28, O111:H8, and O103:H2
(big five) or their nonmotile variants (104,
primer pairs and two labeled gene probes 106). The method used in the United States
were found sufficient to detect all stx1 and includes the same panel of EHEC serotypes
stx2 variant genes except for stx2f (99). The and, in addition, EHEC of serogroups O45
same primers and probes were found to detect and O121 (big seven group) (105). Interest-
STEC from enrichment cultures of ground ingly, EHEC O45:H2 strains are isolated from
beef samples inoculated with 2 to 20 CFU of patients in North America but are not yet
bacteria/25 g (100). Alternative real-time PCR notified as disease agents in Europe. In con-
detection systems using common stx1 and stx2 trast, EHEC O121:H19 strains were repeatedly
primer sets and probes have been successfully isolated from patients with HUS on the
tested on ground beef samples artificially con- European and North American continents
taminated with EHEC (1 to 10 CFU/25 g) (3436, 107, 108).
belonging to the most prominent serogroups The European and U.S. EHEC screening
(101). Two sets of PCR primers and TaqMan methods first investigate the presence of stx
minor groove binder probes were successfully (97) and eae (109) genes. In the case of a
employed to detect EHEC O26, O111, O118, positive result, the presence of genes encoding

the lipopolysaccharide (LPS) of the above- described above. As a consequence of such

mentioned EHEC serogroups is tested (88, restrictions, there is always a risk that other
104, 105). A list of the real-time PCR detectors serotypes of STEC that might be highly path-
used in the European CEN/ISO method was ogenic for humans as well are not taken into
published by Stephan et al. (88). Both proce- account. For example, a considerable number
dures were published and became effective of clinical isolates of STEC strains different
in 2012. from the big five and big seven groups
This sequential procedure can provide four were isolated from patients with HUS in
different results. When stx and eae genes are Germany and are listed in the HUSEC collec-
not detected, one concludes that STEC strains tion (108). A number of eae -and stx-positive
are absent. When an stx gene is detected but STEC strain types serologically different from
no eae gene, one presumes the presence of the big seven group were recently found to
an STEC strain with no or low risk for public be similar to the big seven EHEC strains
health. In both cases, no further action is according to their nle gene profiles (40).
taken. If stx and eae genes are detected, the These emerging EHEC strains belong to
sample contains STEC strains potentially path- serotypes that were previously shown to be
ogenic to humans. In that case, the presence associated with HC and HUS in humans
of genes coding for the above-mentioned (O5:H, O15:H2, O55:H7, O103:H11, O103:H25,
EHEC serogroups is investigated. If one of O118:H16, O123, O165:H25, O172:H25, and
these serogroups is detected, the test por- O177:H) (34, 116119).
tion may contain STEC highly pathogenic to The outbreak in Europe with EAEC-STEC
humans, and a confirmation is required by O104:H4 in summer 2011 caused a paradigm
isolating the supposed EHEC strain from the shift with regard to human pathogenicity
sample. Bacteria from samples positive for stx, of STEC strains (10, 11). The presence of the
eae, and one or more of the investigated LPS- eae-negative EAEC-STEC O104:H4 strain in a
encoding genes are isolated using serotype- food matrix would not have been considered
specific immunomagnetic assays to confirm as dangerous according to the criteria based
the presence of these markers in the same on the European Union and U.S. guidelines
strain (104, 105). The single colonies thus described above (11). It is therefore important
isolated are investigated with the previously to adopt such procedures for the current
described genetic markers to confirm the situation by including new significant gene
presence of all pathogenicity markers in the targets such as aggR for identification of
same strain. EAEC-STEC strains (11, 102). Recently, PCR
The isolation of the EHEC strain is im- tests specific for EAEC-STEC O104:H4 (120)
portant, as samples from feces, food, and were stipulated in the new regulation on
water may contain mixtures of eae-negative sprouts and seeds published in Europe in
STEC strains, stx-negative strains belonging March 2013.
to the EHEC serogroups, and eae-positive Severe disease such as bloody diarrhea
EPEC strains not producing Stx (110114). cannot be attributed simply to the stx geno-
Without doubt, the adaptation of these or al- type or certain serotypes of STEC (36, 121
ternative (115) real-time PCR protocols will be 123) and is not only restricted to infections
very helpful for detecting most EHEC strains with typical EHEC strains. Besides O104:H4,
pathogenic to humans from food samples. other types of LEE (eae)-negative STEC
However, it is still not completely clear strains can play a role as severe pathogens
which other factors contribute to the human of humans (42). Only half of human STEC
pathogenicity of STEC strains not belong- infections in the European Union were caused
ing to the big five or big seven EHEC by big five strains (107). This indicates that
groups that are targeted by sequential assays searching for STEC potentially pathogenic to

270 BEUTIN AND FACH
humans in animals, food, and the environ- enrichment of non-O157 STEC strains from
ment should not be limited to the set of stx- food and other samples (129, 132). As a con-
and LEE (eae)-positive STEC strains. sequence, numerous enrichment media and
growth conditions for STEC from food, fecal,
and environmental samples were developed
Cultural Enrichment of STEC
(129, 133, 134) (Table 1). However, most of
from Nonhuman Samples
these have not been evaluated for their rela-
STEC is generally present in relatively low tive efficacy (135).
numbers in environmental and food matrices Testing growth of STEC strains in different
and in feces of ruminant animals that are the types of commonly used enrichment media
natural hosts for these strains. For STEC screen- revealed strain-specific differences, depending
ing, it is therefore important to include cultural on incubation temperature and media supple-
enrichment steps that allow growth of STEC ments (136). Enrichment media containing
to detectable numbers by either method. En- novobiocin that are used in standard protocols
richment in liquid media by sample dilution for enrichment of STEC O157 strains (130, 131)
(generally 1:10) reduces possible inhibitors were found to inhibit growth of some non-
of STEC growth that might be present in the O157 STEC (136) and freeze-injured STEC
sample. A second enrichment step of plating O157 strains (61, 88, 137). A preenrichment
aliquots from liquid enrichment cultures grown step in a nonselective medium is therefore
on solid agar plates and investigating DNA generally recommended for growing stressed
prepared from the lawn of bacteria can enhance STEC from food and other samples (88, 129,
the sensitivity of STEC detection (99, 102). 131).
For detection of Stx by serological assays Enrichment of STEC from suspected food
including the colony immunoblot, media con- samples is also important to avoid false-positive
taining substances that enhance production results for STEC. Real-time PCR showed high
of Stx such as mitomycin C are employed (50, CT values (>35), indicating the presence of a
53, 59, 72). Mitomycin C can boost drastically very low number of STEC, in 7.5% of meat
Stx production and toxin release by bacteria and 16.0% of cheese samples after cultural en-
(124, 125). Certain STEC strains do not pro- richment (138). Very high CT values after en-
duce detectable quantities of Stx without in- richment may indicate that nonenrichable stx
duction (52, 125). Apart from mitomycin C, sequences are present that might correspond
which is a highly toxic substance, different to stx phages or nonculturable STEC strains
classes of low-toxicity antibiotics were found (138, 139) (Fig. 3).
to induce the production of Stx (126128).
These antibiotics may present alternatives as TABLE 1 Different types of enrichment media
toxin enhancers for detection of Stx in phe- for STECa
notypic assays (125). Tryptose soy broth (TSB)
The addition of Stx enhancers is not Buffered peptone water (BPW)
needed for STEC detection by DNA-based Gram-negative broth (GN)
methods. Nevertheless, microbial enrichment E. coli broth (EC)
Mineral modied glutamate broth (MMGB)
of test samples is always recommended and Brilliant green bile broth (BGBB)
is a critical step in any protocol for detection Lauryl sulfate tryptose broth (LSTB)
of STEC from all kinds of samples (88, 129). MacConkey broth (MCB)
Ideally, enrichment favors growth of STEC Enterohemorrhagic E. coli broth (EHEC)
and disfavors growth of other organisms pres- Brain heart infusion broth (BHI)
a
ent in the sample. However, apart from The following substances are used (singly or in combination) as
supplements to increase selectivity: novobiocin, vancomycin,
protocols developed for STEC O157:H7 (130, cexime, acriavine, cefsulodin, cexime, K-tellurite, and bile salts.
131), there is no standard method for specific Adapted from reference 134.

Apart from the composition of the en- Vegetables such as ready-to-eat salads may
richment medium, the incubation time for contain about 107 CFU/g of total aerobic
enrichment (generally 6 to 24 h) and the bacteria, and bacterial growth continues at
temperature (between 35 and 41.5C) are retail storage temperatures of 6 to 10C (102,
critical. STEC O157 was found to grow best 141). More than 90% of the total microbial
between 30 and 42C (140). The incubation counts in ready-to-eat salads are represented
temperature of 41.5 1C, as recommended by by Pseudomonas and Enterobacteriaceae (142,
ISO 16654 (130), might hamper the recovery 143). As these bacteria are generally inhibited
of injured bacteria (88). The composition of by a growth temperature of 44C, a pre-
the growth medium together with the incu- enrichment step at 37C for 6 h in liquid
bation temperature (37C versus 42C) has a medium (buffered peptone water or brilliant
significant influence and may cause multiple green bile broth) (Table 1) followed by a sec-
effects on the growth behavior of STEC be- ond enrichment step by growth on selective
longing to different serotypes (136). agar plates incubated at 44C for 18 to 22 h
The chosen temperature can be very was found suitable to recover an initial inoc-
important for cultural enrichment of STEC ulum of 1 to 10 CFU of EHEC/25 g from
from samples where high numbers of a bac- ready-to-eat salad samples (102). The choice
terial flora from the environment are present. of the best enrichment media, enrichment
time, and temperature may vary according
to the sample matrix (food, environment, or
feces),which has to be investigated for STEC;
no general recommendation can be given at
the present state.
ISOLATION OF STEC FROM

Stx-POSITIVE SAMPLES
The isolation and further typing of STEC from

incriminated samples are required to confirm
positive results from Stx screening and to
FIGURE 3 Importance of cultural enrichment for the uncover chains of infection in STEC outbreaks
detection of viable STEC in meat samples by con- and sporadic cases (50). Detecting Stx pro-
ventional and real-time PCR. Agarose gel showing duction or the underlying genes in any kind
Stx2-specic PCR products in enrichment cultures. of sample without attempting to isolate the
The PCR was performed with common stx2 PCR
primers LP43/LP44 as described in reference 86. The STEC strain is incomplete and should be con-
corresponding CT values from real-time PCR are in- sidered as presumptive (50, 132).
dicated below. (A) Lanes: m, molecular weight stan- In most samples, STEC is present in low
dard; c, Stx-negative E. coli control strain; +c, Stx2- numbers, and isolation may involve testing
positive E. coli control strain. (B) Sample containing
large numbers of colonies (132). Thus isolation
nonampliable stx genes. PCR and real-time PCR are
performed after 6, 8, and 18 h of enrichment. Stx
of STEC is the most laborious part of STEC
gene-specic signals are decreasing following pro- screening as the majority of STEC strains
longed enrichment with both conventional and real- show phenotypic properties very similar to
time PCR. (C) Sample containing ampliable stx those of commensal E. coli and other Entero-
genes. PCR and real-time PCR are performed after 6,
bacteriaceae, which can be present in any kind
8, and 18 h of enrichment. Stx gene-specic signals
are increasing following prolonged enrichment with of sample.
both conventional and real-time PCR. The most straightforward way to isolate
doi:10.1128/microbiolspec.EHEC-0001-2013.f3 STEC from a sample is direct screening for

272 BEUTIN AND FACH
Stx-positive colonies through the Stx colony proposed for isolation of STEC from positively
immunoblot or colony hybridization (50). tested samples (104). The USDA/FSIS method
Kits for DNA hybridization and a Stx colony uses a chromogenic, selective indicator medi-
immunoblot kit are commercially available, um (modified Rainbow Agar) for the detection
and these methods for STEC detection are of major EHEC-type strains after IMS and a
officially recommended for food inspec- serological slide agglutination test for identi-
tion laboratories in Germany (61). However, fication, followed by PCR detection of relevant
as these procedures are labor-intensive and genes for confirmation (105).
time-consuming, they reach their limits when A number of indicator media and selective
larger numbers of samples have to be inves- media have been developed for detection
tigated in time. and selection of STEC strains (102, 129, 146,
Because many countries concentrate only 150). Most of these media were specifically
on a limited number of EHEC serogroups developed for the detection of non-sorbitol-
(O157, O103, O111, O26, and O145), commer- fermenting STEC O157:H7 strains. Addition
cially available immunomagnetic separation of antibiotics such as novobiocin, vancomy-
(IMS) assays are used for enrichment of sus- cin, cefixime, and cefsulodin to inhibit Gram-
pected colonies before cultural detection (100, positive organisms, plus Proteus and Pseudo-
104, 105, 137, 144). Studies have shown that monas species, was shown to increase the
the IMS enrichment increases the isolation selectivity of growth media for STEC strains
rate of the targeted bacteria (45, 50, 131). IMS (129, 145, 147). However, the efficacy was
beads directed to other EHEC serogroups found to depend largely on the type and the
can be prepared with commercially available concentration of the antibiotics used (129, 136,
reagents if the appropriate antisera for coating 146, 152). Lower recovery of plated STEC
are available (101). However, the use of IMS is strains was found with some commercial se-
limited to the targeted E. coli serogroups, and lective media when compared to nonselective
serologically O-rough derivatives of these brain heart infusion agar (146). Which addi-
STEC strains will not be captured specifically tives are used for the selectivity of the medium
by O-antigen-directed IMS beads. also depends on the background microflora,
Independent of the enrichment procedure which may vary between different sample
used, STEC colonies must be grown on solid types.
agar plates to obtain pure cultures for storage Chromogenic media such as modified
and further characterization. Solid media for Rainbow Agar, Chromagar-STEC, and non-
STEC isolation can be divided into nonselec- commercial formulas are widely used to iden-
tive media (such as brain heart infusion agar) tify STEC strains by observing colony growth,
(145, 146), indicator media for certain pheno- morphology, and color (100, 102, 105, 150, 151,
typic properties of STEC strains (such as en- 153). These media commonly contain tellurite,
terohemolysin agar) (57), and selective media which is used for counterselection to suppress
which frequently contain tellurite or antibi- growth of accompanying non-STEC bacte-
otics (such as CT-SMAC) (50). Very fre- ria. Tellurite is highly selective for growing
quently, indicator media are modified by the non-sorbitol-fermenting E. coli O157 and cer-
addition of selective agents or antibiotics to tain other EHEC strains belonging to the big
increase their selectivity (101, 147151). seven EHEC serogroups. Resistance to tel-
Nonselective media are suggested for iso- lurite is associated with a cluster of genes
lation of presumptive EHEC according to encoded by the ter operon (102, 154156).
the CEN/ISO method. After IMS enrichment, However, about half of the EHEC O103:H2
a minimum of 50 colonies on nonselective strains, all sorbitol-fermenting EHEC O157:[H7]
medium (nutrient agar) followed by stx PCR strains, and single representatives belonging to
investigation of pooled and single colonies is other EHEC serogroups (O121:H19, O145:H28)

were found to lack ter genes and showed no very rarely in STEC strains (160162). Alpha-
growth on media containing tellurite (102, 155). hemolysin is produced in the exponential
The maximum concentrations of tellurite per- growth phase, and lysis zones appear after 3 to
mitting growth were found to be significantly 4 h of incubation of inoculated plates at 37C.
different when STEC strains were compared, The lysis zones are generally larger compared
and spontaneous recombinants of EHEC O157 to those of enterohemolysin. The detection
strains developing sensitivity to tellurite were of low quantities of enterohemolysin-positive
found (146, 157). Most STEC strains not be- (Ehly+) STEC from mixed cultures of bacteria
longing to the big seven group were found to can be improved by plating dilutions of bac-
be negative for ter genes and were inhibited by terial enrichment cultures. Single Ehly+ colo-
tellurite in growth media (102, 156). nies indicating STEC can thus be easily
On the other hand, many Enterobacteria- detected in a lawn of morphologically differ-
ceae from the environment that are found in ent bacteria (57) (Fig. 4).
vegetables and surface water grow on media Enterohemolysin agar is commercially
containing tellurite (Beutin L, unpublished available, and its selectivity can be improved
data). By screening a large number of E. coli by the addition of antibiotics (57, 147, 149)
strains, 12% of apathogenic E. coli (#150) and (Fig. 5). About 90% of EHEC and 40% of
12.5% of EPEC (#287) were found to carry STEC strains carry the ehxA/ehlyA genes,
ter genes and showed growth on media con- which are rarely found in non-STEC strains
taining tellurite (102). In summary, tellurite (40, 163). However, some of the important
supplementation of growth media proved to EHEC types, such as sorbitol-fermenting O157
be suitable for selection of the most important and O104:H4, do not express or do not carry
EHEC strains, but absence of growth on these genes for ehxA/ehlyA, respectively (43, 164).
media does not indicate that EHEC is absent On the other hand, ehxA/ehlyA genes are
from the sample. frequent in derivatives of EHEC that have
Selective agar plates were found to grow lost the stx genes and in some EPEC O26:H11
generally lower portions of inoculated STEC strains as well (165167). In STEC from hu-
when compared with unselective brain heart man patients, the presence of ehxA genes
infusion agar (146). The choice of a nonse- and an enterohemolytic phenotype was found
lective indicator medium, such as enterohemo- more associated with strains possessing the
lysin (washed sheep blood) agar, may thus eae gene (96.2%) than with eae-negative STEC
serve as a supplement or alternative to selec- strains (65.2%) (34, 168). Enterohemolysin
tive media used in isolation of STEC. Entero- genes were found in 37.9 to 55% in STEC from
hemolysin (also called EHEC-hemolysin) is food specimens (18, 86, 87, 169171). More-
encoded by genes (ehxA/ehlyA) located on over, significant differences for the presence
large plasmids present in EHEC and STEC of ehlyA genes were found between STEC
strains (158, 159). In contrast to E. coli alpha- strains according to their animal food origin
hemolysin, enterohemolysin is only visible on (52, 172). ehly/ehxA genes were also found
blood agar plates containing washed sheep in more than 92% of STEC strains isolated
blood erythrocytes (57, 160). Hemolysis zones from the environment of dairy farms (26) and
caused by enterohemolysin are generally were significantly associated with STEC
narrow and turbid and appear after longer from food derived from cattle, wild boar, and
incubation, generally overnight, of bacteria red deer (18). As some E. coli strains produce
plated on enterohemolysin agar (57). They are enterohemolysin without being STEC, all
easily discerned from hemolysis zones caused isolates of presumptive STEC taken from
by alpha-hemolysin, which is frequently found enterohemolysin agar have to be confirmed
in uropathogenic E. coli and enterotoxigenic for Stx production or the presence of stx
E. coli from animals, but not in EHEC and genes.

274 BEUTIN AND FACH
performance, and interpretation; availability;

and costs (173).
Many typing methods are not generally
applicable to all strains of the STEC group
but were found useful for the identification
and characterization of certain STEC strains,
subtypes, and subclones within a given sero-
type or serogroup. Biotyping was found im-
portant for rapid screening of clinically and
epidemiologically important subtypes among
strains belonging to the same O:H type such
as O157:[H7] and O26:[H11] (166, 174176).
Bacteriophage lysotyping was used for sub-
typing clinically and epidemiologically im-
portant STEC O157:H7 strains, but its use
is restricted to a few laboratories and is not
successfully employed on non-O157 STEC
FIGURE 4 Stool sample of a human patient with HUS
infected with EHEC O157:H7 grown on enterohemo- strains (50, 177, 178). Resistance to certain
lysin agar (57). Morphological differences between antimicrobials such as sulfamethoxazole, tet-
the hemolysis zones on enterohemolysin agar facili- racycline, and streptomycin is frequent in
tate the detection of enterohemolytic, ehlyA-positive many STEC and EHEC strains (179). How-
STEC strains. Plating of 10-fold dilutions from stool
ever, an extended antimicrobial resistance
enrichment cultures grown in tryptic soy broth for
detection of enterohemolytic, presumptive STEC typing can be very helpful for the identifica-
colonies. Four morphologically different types of tion and characterization of subclones within
bacteria designated from A to D were detected: STEC strains and types (43, 50, 118, 180). Most
A, alpha-hemolysin producing, Stx-negative E. coli; STEC strains carrying plasmids and plasmid
B, hemolysin-negative Enterobacteriaceae; C, entero-
profiles can be heterogeneous within a given
hemolytic (ehlyA)-positive STEC O157:H7; D, Pseudo-
monas aeruginosa. serotype (181, 182). Analysis of plasmid pro-
doi:10.1128/microbiolspec.EHEC-0001-2013.f4 files can nevertheless be helpful for describing
clinically important subclones among certain
STEC types (183).
STEC STRAIN TYPING METHODS Multi-locus sequence typing (MLST) has
replaced multi-locus enzyme electrophoresis
Bacterial subtyping includes microbiologi- and is now widely used for analysis of phylo-
cal and molecular methods that are useful genetic relationships among E. coli strains (173,
for differentiation below the species level. 184). However, MLST can be also employed
Microbiological methods such as serotyping, for characterization of epidemiologically and
biotyping, colicin typing, phage lysotyping, pathogenetically important subtypes among
and antibiotic susceptibility typing are widely serotypes and clones of E. coli strains. Con-
used for E. coli and other bacterial species. ventional and molecular serotyping was found
A growing number of molecular methods are to provide useful information for outbreak in-
described for typing of STEC strains (50, 173), vestigations, for defining seropathotypes, and
but only some have gained broader accep- for all kinds of epidemiological studies on re-
tance. An overview of genotyping methods servoirs and spread of certain STEC types. The
for STEC was presented recently, and the complete O:H serotype of STEC strains can be
most promising molecular methods for STEC used as an indicator for the clonal type of
genotyping were compared for their dis- certain STEC strains, as shown by multi-locus
criminatory power; ease of standardization, enzyme electrophoresis and MLST analysis

FIGURE 5 Enhancement of selectivity for isolation of STEC from bovine fecal samples on enterohemolysin
agar supplemented with vancomycin. (A) Growth of enterohemolytic STEC colonies from a sample of bovine
feces on enterohemolysin agar supplemented with 8 mg/liter of vancomycin. The gram-positive ora is sup-
pressed. (B) Growth of enterohemolytic STEC colonies from the same sample of bovine feces on nonselective
enterohemolysin agar. The abundant gram-positive ora overgrows STEC present in the sample.
(118, 185, 186). In a similar way, other molec- concentrate on the identification of the most
ular typing methods such as pulsed-field gel important EHEC serogroups. A large panel
electrophoresis (PFGE) gave more meaningful of E. coli antisera for O and H serotyping
results when strains belonging to the same is commercially available and can be used to
serotypes were compared instead of randomly screen diagnostically interesting types. Com-
sampled STEC isolates (187, 188). plete serotyping by agglutination of O and H
antigens is laborious and time-consuming, as
slide agglutination tests with live bacteria have
Phenotypical and Molecular
to be confirmed by agglutination in tubes with
O and H Serotyping
heat-treated (O antigen detection) and form-
A serotyping scheme of E. coli O (LPS) and aldehyde-inactivated motile cultures of bac-
H (flagella) antigens was established in the teria (H antigen detection) (189). A number
1940s by Fritz Kauffmann and further ex- of serological tests for detecting E. coli O157
tended by the International Escherichia and and H7 antigens and conjugated antibodies
Klebsiella Centre in Copenhagen, Denmark, in directed to important EHEC serogroups are
collaboration with national E. coli reference commercially available to design and perform
laboratories (189). At present, a panel of E. coli rapid serological assays.
O antigens (O1 to O181) and H antigens (H1 However, a considerable number of STEC
to H56) is described (190), and four additional and EHEC isolates from clinical, food, and
O antigens (O182 to O186) are currently being other sources are not typeable for the O and H
investigated in some reference laboratories antigens, as O-rough (spontaneously aggluti-
(18, 191, 192). Complete serotyping of E. coli nating) and nonmotile (NM) strains are fre-
O and H antigens is performed in a few lab- quently isolated. About 18% (n=107) of 593
oratories that participate in quality assur- STEC strains from different types of food
ance tests (193). STEC belongs to more than were untypeable (O-rough or unknown O
400 serotypes (see Diversity of STEC Types type) for their O antigens using O antisera
above), and many diagnostic laboratories from O1 to O186 (18). In a study of 677 STEC

276 BEUTIN AND FACH
strains from human patients, 221 (32.6%) were O113, O121, O128, and O145), targeting the wzx
found to be NM and thus not serotypeable and wzy genes of the rfb locus. An alternative
for their H antigens (34). The NM strains molecular O serotyping method based on se-
frequently include representatives of the big quence variations of the gnd gene was pres-
seven EHEC O groups (30, 34, 164). ented by Gilmour and coworkers (211). The
Most STEC strains showing an NM phe- gnd gene is located close to the O antigen
notype were nevertheless found positive for cluster. By sequencing this gene, they dem-
the fliC gene, which encodes most of the fla- onstrated the presence of unique alleles for
gellar antigens in E. coli (194). In those cases, each of the examined STEC serogroups.
H serotyping can be substituted or accompa- Furthermore, this approach differentiates be-
nied by molecular typing of the fliC gene in tween STEC and non-STEC strains of sero-
E. coli. A number of PCR and RFLP methods groups O157, O26, O55, O6, and O117 (https://
were described for fliC genotyping (195197). www.corefacility.ca/ecoli_typer/). Recently
In some cases, fliC genotyping was found to Norman et al. (212) identified single nucleo-
be more specific for detection of genetic sub- tide polymorphisms in the O-antigen operon
types among serologically cross-reacting H of serogroups O26, O45, O103, O111, O121, and
antigens and for possible detection of new O145 that could be used to differentiate be-
flagellar types in E. coli (198200). The fliC tween strains of these serogroups that con-
genes representing 43 H antigens in E. coli tain STEC-associated virulence factors and
have been sequenced, and the sequence in- those that do not. Their study supports the
formation can be used for confirmation of idea that differences in the genetic sequence
H serotyping and fliC typing results (194). of the O-antigen operon correspond well with
Numerous molecular serotyping methods differences in the virulence gene profiles and
have been developed as alternatives to tradi- provide evidence of separate clustering for
tional serotyping and identify most often the STEC and non-STEC strains. These findings
genes associated with the somatic (O) and may open the development of new detection
flagellar (H) antigens or detect particular se- tests and thus represent a significant improve-
rotype-associated sequences in housekeeping ment in the identification of EHEC strains. An-
or virulence genes. These methods have the other recent approach described by Delannoy
advantage of also being applicable on O-rough et al. (120, 213) is based on the genetic diversity
strains that cannot be serotyped phenotypi- of the clustered, regularly interspaced, short
cally. Molecular tools developed for the de- palindromic repeats (CRISPR) regions of EHEC
termination of the O antigen are mainly based to design simplex real-time PCR assays for each
on the detection of the rfbE (also named per), of the big seven EHEC serogroups and for
wzx, and wzy genes (95, 97, 201206). Ballmer STEC O104:H4. The identification of EHEC
and coworkers (207) developed a serotyping serotype-specific CRISPR sequences was found
microarray to identify the 24 most epidemio- to be more specific than the mere identification
logically relevant serogroups based on oligo- of O antigen gene sequences as it is used in
nucleotides designed on the wzx and wzy current PCR protocols for the detection of
genes. In addition, this array tube-based assay EHEC strains (104, 105).
targeted 47 different H antigens based on the
identification of variations on the fliC gene
PFGE and Other Genotyping Methods
that code for the flagellar monomer (207,
208). Lin and coworkers (209, 210) fabricated PFGE of total bacterial DNA digested with
a TaqMan real-time multiplex PCR assay infrequently cutting restriction enzymes pro-
and a microbead-based suspension array that duces DNA patterns that allow comparison of
identify the 10 most clinically relevant STEC strains at a genomic level. PFGE is considered
serogroups (O157, O26, O45, O91, O103, O111, the gold standard of subtyping methods be-

cause of its discriminatory power and its use- food chain (217, 218, 226229). PFGE analysis
fulness in epidemiological investigations (173). of STEC strains belonging to the same sero-
PFGE is widely used for typing of EHEC O157: types was also helpful to show that wild and
H7, and the electronic submission of DNA domestic animals form a common reservoir
fingerprints stored in databases allows world- for certain STEC and EHEC strains that con-
wide survey of these pathogens (214, 215). taminate the respective food products and are
Less frequently, PFGE is also used for finally found in human patients (87, 166, 216).
typing of non-O157 EHEC strains. Most data The ability of Stx phages to integrate at dis-
are available from non-O157 EHEC big five tinct sites in chromosomes of E. coli (230, 231)
strains, but other STEC types also were in- was employed as a molecular typing method
vestigated (15, 87, 187, 216219). A number of for STEC O157:H7 strains with regard to their
studies were performed on E. coli O26 strains virulence for humans. Sixteen genotypes of
from patients, food, and animals (166, 176, 182, STEC O157:H7 strains were defined on the
220, 221). PFGE typing was compared with basis of Stx type and occupation of phage in-
other genotyping methods such as variable tegration sites in the chromosome of STEC
number of tandem-repeats analysis (MLVA) O157:H7 (232). The majority (95%) of STEC
(166). IS621 multiplex PCR-based fingerprint- O157:H7 clinical isolates from humans clus-
ing was found to be less discriminatory than tered into genotypes 1, 2, and 3 in contrast
PFGE in a study on O26 isolates from different to only 51.3% of the bovine STEC O157:H7
sources, but epidemiologically related strains strains (232). Human-associated STEC O157:
were identified with both methods (221). H7 genotype strains were characterized by in-
MLST, PFGE, and MLVA were found suit- creased expression of EHEC virulence genes
able to discriminate between major clusters of (pO157 and LEE), which could explain their
O26:[H11] EPEC and EHEC strains (166, 222). prevalence in disease (233). Human STEC
PFGE was used for typing STEC and other O157:H7 strains are characterized by the pre-
E. coli strains belonging to serogroup O103 sence of Stx2a phages whereas Stx2c phages
and compared with other typing methods are present in typical bovine O157:H7 strains,
such as P gene profiling and ribotyping. Clonal which could also explain the virulence dif-
types were reported to be best defined by a ferences observed (234).
combination of PFGE and P gene profiles
(177). Stx phage PCR-RFLP was found to be
Subtyping of STEC- and EHEC-Related
less discriminative than PFGE for typing a
Virulence Genes
series of STEC strains isolated in Japan, in-
cluding O157, O26, O111, and others (223). Virulence markers such as bacteriophage-
PFGE and MLST with seven housekeeping encoded stx genes, the LEE pathogenicity is-
genes were compared in an analysis on 54 land-located eae (intimin) genes, and some
E. coli O103 strains from different sources. non-LEE effector genes (nle) such as nleA
PFGE was very useful for identification of were shown to be genetically very heteroge-
genetically closely related subgroups among neous, splitting in various subtypes (46, 235
strains belonging to the same MLST type, such 239). Genotyping of eae, stx, and some nle
as Stx2-producing EHEC O103:H2 strains genes was thus employed for typing STEC
from patients with HUS (224). The suitability strains, to describe STEC clones, for defining
of PFGE for identification and subtyping clin- pathotypes, and for investigating animal host
ically important STEC strains was also dem- reservoirs. Subtyping of these genes is per-
onstrated with serogroup O145 strains (225). formed with real-time and conventional
Moreover, PFGE was successfully employed specific PCRs, restriction endonuclease di-
to verify transmission of STEC strains between gestion of relevant PCR products, and nucle-
animals, from animals to humans, and in the otide sequencing.

278 BEUTIN AND FACH
Subtyping of stx genes below the Stx1 and isolated from the farm environment (255,
Stx2 level becomes increasingly important 256).
for describing potentially human pathogenic Some, but not all, subtypes of Stx1 and Stx2
STEC strains. Certain subtypes of Stx1 and are associated with strains carrying the LEE-
Stx2 are related to the virulence of STEC for encoded eae genes involved in intestinal ad-
humans and are frequently associated with hesion of EPEC and EHEC strains. About 35
EHEC strains (46, 240242). Some genotypes variants of the eae gene have been described
of Stx1 and Stx2 are typical for certain sero- in E. coli (257) and in other Enterobacteriaceae
types of STEC strains. Significant relation- (246, 258). The big seven EHEC strains
ships among the food-producing animal host express different genetic variants of the eae
species, STEC serotypes, and stx genotypes gene, which can be analyzed by PCR, PCR-
could be established (18). In detail, Stx2e, as RFLP, multiplex real-time PCR, and nucleo-
the key factor in porcine edema disease prin- tide sequencing (109, 144, 236, 259, 260).
ciple, is frequently found in STEC from pigs, Typing of eae genes was found to be useful for
wild boar, and their meat products but rarely characterizing EHEC from food and animals
in other animal species. Most likely, Stx2e is and for discriminating between EPEC and
not associated with human pathogenicity (18, EHEC strains belonging to serogroups O103,
52, 243, 244). Stx2f is associated with certain O111, O145, and O157 (113, 144, 172, 216, 220,
serotypes of STEC strains from birds and with 224, 225). Typing of eae genes in combination
strains of Escherichia albertii (245, 246). STEC with analysis of nle genes can also be useful
producing Stx2f was found positive for LEE for discriminating between EPEC and EHEC
genes (encoding the attaching and effacing derivative strains that have lost their stx genes
phenotype) and was isolated as an agent of and for characterizing new types of emerging
nonbloody diarrhea in humans (241, 247). EHEC strains (34, 163, 169, 261, 262).
Stx1c and Stx2b are frequent among STEC
strains from sheep, goats, red deer, and their
food products (18, 216, 248250). Stx1c and RISK ASSESSMENT OF STEC FOR HUMANS
Stx2b are more associated with nonbloody
diarrhea in humans but not with HC and HUS With the growing number of different STEC
(34, 241, 242). Stx1a, Stx2a, Stx2c, and Stx2d serotypes isolated from different sources, at-
were found to be significantly associated with tempts to classify these strains according to
STEC from bovines and food of bovine origin their virulence for humans have increased.
(18, 144, 251, 252). Stx2a is highly associated The term EHEC was coined in 1987 for those
with STEC strains causing HUS in humans, STEC strains that cause HC and HUS, and a
such as O157:H7 and O104:H4 (43, 234, 253). DNA probe derived from the O157 virulence
Stx2a, Stx2c, and Stx2d genes show little plasmid (pO157) was developed for typing
differences in their genotypes (46), and PCR- purposes (263). Certain genetic variants of
RFLP typing was found useful for detect- Stx, such as Stx1a and Stx2a, were found more
ing multiple types of these genes, which are associated with HC and HUS in patients than
frequently present in single STEC strains others (46). However, the mere presence of
(46, 86). A close association of certain sero- stx1 or stx2 genes in an STEC strain does not
types, stx subtypes, and the species of the itself indicate that it causes severe disease in
food-producing animals was found in a study humans.
of 597 STEC strains from different food cat- The relation between Stx type and human
egories (18). Accordingly, multiple types of pathogenicity of an STEC strain becomes
stx variants are associated with STEC iso- statistically clear if the presence of other vir-
lated from food (18, 86, 87, 254). A similar ulence genes encoded by virulence plasmids
diversity of Stx types was also found in STEC (ehlyA) and the LEE-encoded eae genes are

considered (11, 37, 264, 265). Efficient adhe- outbreaks, and the presence of genomic
sion of an STEC strain in the human intestine islands, such as OI-122 and OI-71 (39, 270).
seems to be a prerequisite for its capacity to Bugarel and coworkers (40, 163, 271) studied
cause severe disease such as HC and HUS. the distribution of coding sequences found
The LEE-encoded attaching and effacing sys- on genomic islands OI-71 and OI-122 in a
tem is a typical colonization mechanism of the collection of more than 500 EHEC, EPEC,
big seven EHEC strains. On the other hand, STEC, and nonpathogenic E. coli strains. Both
the outbreak with EAEC O104:H4 has dem- genomic islands code for Nle type III effectors
onstrated that the LEE system can be substi- that are potentially involved in STEC viru-
tuted by other effective intestinal adherence lence. The results revealed an association be-
mechanisms such as those present in EAEC tween the distribution of nle genes harbored
strains (11, 43). A number of other types of on OI-71 (nleF, nleH1-2, and nleA) and OI-122
fimbrial and afimbrial adhesions, which can (ent/sen/espL2, nleB, and nleE) with the big
play a role in intestinal adhesion of STEC seven EHEC serotypes and with emerging
strains, have been described (266). Further EHEC serologically different from the big
analysis of STEC strains from human patients seven that had been already associated with
will show whether the presence of one or severe illness in humans.
more of these adhesins is specifically associ- EHEC strains show either one of two pre-
ated with severe illness. dominant profiles, which differ according to
A growing number of genes located on their composition of nle genes. EHEC O157:
virulence plasmids of STEC and on genomic [H7], O111:[H8], O26:[H11], and O121:H19 gen-
islands (LEE, nle genes) were associated with erally have the complete set of the six nle
strains causing disease in humans (42). Elu- genes located on islands OI-71 and OI-122, as
cidation of genomic sequences of major EHEC found with emerging EHEC serotype of O5:
strains allowed further research on the dis- NM, O103:H25, and O55:H7 strains. On the
tribution of various effector genes located other hand, EHEC O103:H2 and O145:[H28]
on prophages and genomic islands in strains strains showed an altered profile composed
belonging to the big seven EHEC types of OI-122 nle genes ent/sen/espL2, nleB, nleE,
(267269). It thus became possible to establish and the OI-71-located nleH1-2 gene. nleF
an MRA approach (39) to define STEC ac- and nleA genes (OI-71) were not detected in
cording to its pathogenicity for humans. O103:H2 and O145:[H28] strains using the
The first MRA approach was established by PCR assays developed in this study. How-
Karmali and coworkers in 2003 (37). STEC ever, the possibility that these strains carry
strains were classified as seropathotypes A genetic variants of these nle genes, as previ-
to E according to their incidence, frequency ously reported for nleA, cannot be excluded
in outbreaks, and association with HC and (235).
HUS. Seropathotypes A and B comprise the The big seven EHEC and some of the
EHEC big five plus O121 strains whereas emerging EHEC types are characterized by
seropathotype C includes LEE-negative STEC four variants of the eae genes present in these
strains that were already associated with HC strains (163, 169, 271). The eae-gamma var-
and HUS (O91:H21, O113:H21, O104:H21, and iant is characteristic for EHEC O157:H7,
others). Strains of seropathotype D encompass O55:H7, and O145:H28. The eae-epsilon vari-
all STEC strains that were already involved ant is a hallmark of EHEC O103:H2, O45:H2,
in cases of diarrhea but not HC and HUS. and O121:H19 strains (163, 236). The eae-
Finally, strains of seropathotype E have not theta variant is associated with EHEC strain
been associated with human infections. The O111:[H8] and with the emerging EHEC O103:
MRA approach was adapted in further stud- H25 strain that caused an HUS outbreak in
ies revealing links between severe illness, Norway (272). Finally, the eae-beta variant is

280
Chap14.proof.3d
280
BEUTIN AND FACH
TABLE 2 Serotypes of STEC isolated from human patientsa

O1:H O8:H21 O25:K2:H2 O52:H23 O83:H1 O103:H21 O114:H? O127 O148:H28 O172:H
O1:H1 O8:H25 O25:H14 O52:H25 O84:H O103:H25 O115:H10 O128:H O150:H O172:H?
O1:H2 O9ab:H O26:H O54:H21 O84:H2 O103:HNT O115:H18 O128:H2 O150:H8 O173:H2
O1:H7 O9:H7 O26:H2 O55:H O84:H20 O104:H O116:H O128:H7 O150:H10 O174:H
O1:H20 O9:H21 O26:H8 O55:H6 O85:H O104:H2 O116:H4 O128:H8 O152:H4 O174:H2

O1:HNT O11:H O26:H11 O55:H7 O85:H10 O104:H7 O116:H10 O128:H10 O153:H2 O174:H8
O2:H O11:H2 O26:H12 O55:H9 O85:H23 O104:H16 O116:H19 O128:H12 O153:H11 O174:H21
O2:H1 O11:H8 O26:H32 O55:H10 O86:H O104:H21 O117:H O128:H25 O153:H12 O175:H16
O2:K1:H2 O11:H49 O26:H46 O55:H19 O86:H10 O105ac:H18 O117:H4 O128:H31 O153:H21 O176:H
O2:H5 O12:H O27:H O55:H? O86:H40 O105:H19 O117:H7 O128:H45 O153:H25 O177:H
O2:H6 O14:H O27:H30 O60:H O87:H6 O105:H20 O117:K1:H7 O129:H O153:H30 O177:H11
O2:H7 O15:H O28ab:H O64:H25 O88:H O106 O117:H8 O130:H11 O153:H33 O178:H7
O2:H11 O15:H2 O28:H25 O65:H16 O88:H25 O107:H27 O117:H19 O131:H4 O154:H O179:H8
O2:H27 O15:H8 O28:H35 O68:H O89:H O109:H2 O117:H28 O132:H O154:H4 O181:H15
O2:H29 O15:H27 O30:H2 O69:H O90:H O109:H16 O118:H O133:H O154:H19/20 O181:H49
O2:H44 O16:H O30:H21 O69:H11 O91:H O110:H O118:H2 O133:H53 O156:H
O3:H10 O16:H6 O30:H41 O70:H11 O91:H4 O110:H19 O118:H12 O134:H25 O156:H4
O4:H O16:H21 O37:H41 O71:H O91:H10 O110:H28 O118:H16 O137:H6 O156:H7
O4:H5 O17:H18 O38:H21 O73:H34 O91:H14 O111:H O118:H30 O137:H41 O156:H25
O4:H10 O17:H41 O38:H26 O74 O91:H15 O111:H2 O119:H O138:H2 O156:H27
03/30/15 17:58
O4:H40 O18:H O39:H4 O75:H O91:H21 O111:H7 O119:H5 O141:H O156:HNT
O5:H O18:H7 O39:H8 O75:H1 O91:H40 O111:H8 O119:H6 O141:H2 O157:H7
O5:H16 O18:H12 O39:H28 O75:H5 O91:HNT O111:H11 O119:H25 O141:H8 O157:NM
O6:H O18:H15 O40:H2 O75:H8 O92:H3 O111:H21 O120:H19 O142 O160:H?
O6:H1 O18:H? O40:H8 O76:H7 O92:H11 O111:H30 O121:H O143:H O161:H
Chap14.proof.3d
O6:H2 O20:H O41:H2 O76:H19 O95:H O111:H34 O121:H8 O144:H O162:H4
O96:H10 O163:H
281
O6:H4 O20:H7 O41:H26 O77:H O111:H40 O121:H11 O145:H
O6:H12 O20:H19 O44 O77:H4 O98:H O111:H49 O121:H19 O145:H4 O163:H19
O6:H28 O21:H5 O45:H O77:H7 O98:H8 O111:H? O123:H19 O145:H8 O163:H25
O6:H29 O21:H8 O45:H2 O77:H18 O100:H25 O112ab:H2 O123:H49 O145:H16 O165:H
O6:H31 O21:H? O45:H7 O77:H41 O100:H32 O112:H19 O124:H O145:H25 O165:H10
O6:H34 O22:H O46:H2 O78:H O101:H O112:H21 O125:H O145:H26 O165:H19
O6:H49 O22:H1 O46:H31 O79:H7 O101:H9 O113:H2 O125:H8 O145:H28 O165:H21
O7:H4 O22:H5 O46:H38 O79:H14 O102:H6 O113:H4 O125:H? O145:H46 O165:H25
O7:H8 O22:H8 O48:H21 O79:H23 O103:H O113:H5 O126:H O145:HNT O166:H12
O8:H O22:H16 O49:H O80:H O103:H2 O113:H7 O126:H2 O146:H O166:H15
O8:H2 O22:H40 O49:H10 O81:H? O103:H4 O113:H21 O126:H8 L O160:H28
O8:H9 O23:H7 O50:H O82:H O103:H6 O113:H32 O126:H11 O146:H11 O168:H
O8:H11 O23:H16 O50:H7 O82:H5 O103:H7 O113:H53 O126:H20 O146:H14 O169:H
O8:H14 O23:H21 O51:H49 O82:H8 O103:H11 O114:H4 O126:H21 O146:H21 O171:H
O8:H19 O25:H O52:H19 O83:H O103:H18 O114:H48 O126:H27 O146:H28 O171:H2
a
Serotypes in bold were identied as food-borne STEC in Germany (593 strains isolated between 2005 and 2009) (18) and correspond to 19.5% of the 375 STEC serotypes from humans listed in this table.

CHAPTER 14 Detection of STEC from Nonhuman Sources
281
03/30/15 17:58
282 BEUTIN AND FACH
present in EHEC and EPEC O26:H11 (167) strains that constitute the major part of STEC
and in emerging EHEC strains of serotype isolated from animals, food, and the environ-
O5:NM and O118:H16. ment. Nevertheless, 16 of the 41 strains of the
In a given STEC strain, the presence of ent/ HUSEC collection are LEE-negative, pointing
sen/espL2, nleB, nleE, and nleH1-2 effectors, out that at least some of these STEC strains
when associated with one of the four genetic are able to cause severe disease in humans
variants of the eae geneeae-beta, eae- (108). Some serogroups of these LEE-negative
gamma, eae-epsilon, or eae-thetaconstitutes STEC strains (O91, O63, O113, O128, and O146)
the characteristic signature of typical and accounted for 21.4 to 25.7% of serotypeable,
emerging EHEC strains that have been iden- non-O157 STEC strains isolated from human
tified over the past few years in Europe. De- patients in the European Union (107). Most of
tection of these genetic markers constitutes an these strains are frequently found in food of
innovative DNA-based approach in the MRA different origins and can thus be transmitted
of STEC strains. as food-borne infections. Food-borne, LEE-
As these nle genes and eae variants can be negative STEC strains were found to group
shared by Stx-negative EPEC strains as well in the same MLST clusters as LEE-negative
(271), identification of genetic markers allow- strains that were isolated from patients with
ing a more targeted screening of EHEC strains HUS (108, 123). Serotyping of 593 food-borne
in complex food samples is still challenging, STEC strains isolated in Germany between
and genetic targets that may support such an 2005 and 2009 revealed 73 serotypes that
approach still have to be clarified. Monitoring were already found associated with human
EHEC in foods requires, in particular, selec- infections (18, 50) (Table 2). A future chal-
tion of genetic markers able to discriminate lenge will be to establish MRA concepts that
clearly EHEC from EPEC strains. A recent enable a better definition of human patho-
study pointed out that genes carried by ge- genic and nonpathogenic STEC of the diverse
nomic O island 57 (OI-57) may be associated group of LEE-negative STEC strains.
with increased virulence of STEC strains in
humans (273). Open reading frames (ORFs)
inside OI-57 such as Z2097, Z2098, Z2121,
ACKNOWLEDGMENT
and Z2149 seem to be associated preferen- We declare no conflicts of interest with regard
tially with the EHEC strains (273). Recently, to the manuscript.
Delannoy and coworkers (32) identified two
ORFs on OI-57 called Z2098 and Z2099 as
suitable markers to discriminate between CITATION
EHEC and EPEC strains. Z2098 and Z2099 Beutin L, Fach P. 2014. Detection of Shiga
show a higher specificity for the big seven toxin-producing Escherichia coli from nonhu-
EHEC serotype strains than does the eae gene man sources and strain typing. Microbiol
and the other nle genes that were investigated Spectrum 2(3):EHEC-0001-2013.
so far. Accordingly, Z2098 and Z2099 were
rarely found in EPEC (10% and 12%, respec-
tively), STEC (2% and 15%), and apathogenic REFERENCES
E. coli (1% each), in contrast to EHEC (87% 1. Konowalchuk J, Speirs JI, Stavric S. 1977. Vero
and 91% positive) and stx-negative EHEC response to a cytotoxin of Escherichia coli. Infect
derivative strains. Immun 18:775779.
2. Wade WG, Thom BT, Evans N. 1979. Cytotoxic
Apart from the big seven EHEC and
enteropathogenic Escherichia coli. Lancet ii:1235
some emerging EHEC serotypes, there is no 1236.
MRA approach to classify the broad num- 3. Wilson MW, Bettelheim KA. 1980. Cytotoxic
ber of LEE-negative and nle-negative STEC Escherichia coli serotypes. Lancet i:201.

4. Riley LW, Remis RS, Helgerson SD, McGee HB, Characteristics of the Shiga-toxin-producing entero-
Wells JG, Davis BR, Hebert RJ, Olcott ES, aggregative Escherichia coli O104:H4 German out-
Johnson LM, Hargrett NT, Blake PA, Cohen break strain and of STEC strains isolated in Spain.
ML. 1983. Hemorrhagic colitis associated with a Int Microbiol 14:121141.
rare Escherichia coli serotype. N Engl J Med 308: 16. Asakura H, Makino S, Shirahata T, Tsukamoto
681685. T, Kurazono H, Ikeda T, Takeshi K. 1998. De-
5. Orskov F, Orskov I, Villar JA. 1987. Cattle as tection and genetical characterization of Shiga
reservoir of verotoxin-producing Escherichia coli toxin-producing Escherichia coli from wild deer.
O157:H7. Lancet ii:276. Microbiol Immunol 42:815822.
6. Martin ML, Shipman LD, Wells JG, Potter ME, 17. Ferens WA, Hovde CJ. 2011. Escherichia coli
Hedberg K, Wachsmuth IK, Tauxe RV, Davis O157:H7: animal reservoir and sources of human
JP, Arnoldi J, Tilleli J. 1986. Isolation of Esche- infection. Foodborne Pathog Dis 8:465487.
richia coli O157:H7 from dairy cattle associated 18. Martin A, Beutin L. 2011. Characteristics of Shiga
with two cases of haemolytic uraemic syndrome. toxin-producing Escherichia coli from meat and
Lancet ii:1043. milk products of different origins and association
7. Borczyk AA, Karmali MA, Lior H, Duncan LM. with food producing animals as main contami-
1987. Bovine reservoir for verotoxin-producing nation sources. Int J Food Microbiol 146:99104.
Escherichia coli O157:H7. Lancet i:98. 19. Franz E, van Bruggen AH. 2008. Ecology of
8. Mathusa EC, Chen Y, Enache E, Hontz L. 2010. E. coli O157:H7 and Salmonella enterica in the
Non-O157 Shiga toxin-producing Escherichia coli primary vegetable production chain. Crit Rev
in foods. J Food Prot 73:17211736. Microbiol 34:143161.
9. Blanco J, Blanco M, Blanco J, Mora A, Alonso 20. Loukiadis E, Kerouredan M, Beutin L, Oswald
MP, Gonzalez EA, Bernardez MI. 2001. Epide- E, Brugere H. 2006. Characterization of Shiga
miology of verocytotoxigenic Escherichia coli toxin gene (stx)-positive and intimin gene (eae)-
(VTEC) in ruminants, p 113148. In Duffy G, positive Escherichia coli isolates from wastewater
Garvey P, McDowell DA (ed), Verocytotoxigenic E. of slaughterhouses in France. Appl Environ Micro-
coli. Food & Nutrition Press, Inc, Trumbull, CT. biol 72:32453251.
10. Karch H, Denamur E, Dobrindt U, Finlay BB, 21. Gourmelon M, Montet MP, Lozach S, Le
Hengge R, Johannes L, Ron EZ, Tonjum T, Mennec C, Pommepuy M, Beutin L, Vernozy-
Sansonetti PJ, Vicente M. 2012. The enemy Rozand C. 2006. First isolation of Shiga toxin 1d
within us: lessons from the 2011 European Esch- producing Escherichia coli variant strains in
erichia coli O104:H4 outbreak. EMBO Mol Med shellfish from coastal areas in France. J Appl
4:841848. Microbiol 100:8597.
11. Beutin L, Martin A. 2012. Outbreak of Shiga 22. Bennani M, Badri S, Baibai T, Oubrim N,
toxin-producing Escherichia coli (STEC) O104:H4 Hassar M, Cohen N, Amarouch H. 2011. First
infection in Germany causes a paradigm shift detection of Shiga toxin-producing Escherichia
with regard to human pathogenicity of STEC coli in shellfish and coastal environments of Mo-
strains. J Food Prot 75:408418. rocco. Appl Biochem Biotechnol 165:290299.
12. Sanchez S, Garcia-Sanchez A, Martinez R, 23. Kauffman MD, LeJeune J. 2011. European star-
Blanco J, Blanco JE, Blanco M, Dahbi G, Mora lings (Sturnus vulgaris) challenged with Esche-
A, Hermoso DM, Alonso JM, Rey J. 2009. richia coli O157 can carry and transmit the human
Detection and characterisation of Shiga toxin- pathogen to cattle. Lett Appl Microbiol 53:596601.
producing Escherichia coli other than Escherichia 24. Ihekweazu C, Carroll K, Adak B, Smith G,
coli O157:H7 in wild ruminants. Vet J 180:384 Pritchard GC, Gillespie IA, Verlander NQ,
388. Harvey-Vince L, Reacher M, Edeghere O,
13. Beutin L, Geier D, Steinruck H, Zimmermann S, Sultan B, Cooper R, Morgan G, Kinross PT,
Scheutz F. 1993. Prevalence and some properties Boxall NS, Iversen A, Bickler G. 2011. Large
of verotoxin (Shiga-like toxin)-producing Esche- outbreak of verocytotoxin-producing Escherichia
richia coli in seven different species of healthy coli O157 infection in visitors to a petting farm in
domestic animals. J Clin Microbiol 31:24832488. South East England, 2009. Epidemiol Infect114.
14. Caprioli A, Morabito S, Brugre H, Oswald E. 25. Henderson H. 2008. Direct and indirect zoonotic
2005. EnterohaemorrhagicEscherichia coli: emerg- transmission of Shiga toxin-producing Escherichia
ing issues on virulence and modes of transmis- coli. J Am Vet Med Assoc 232:848859.
sion. Vet Res 36:289311. 26. Fremaux B, Raynaud S, Beutin L, Rozand CV.
15. Mora A, Herrera A, Lopez C, Dahbi G, Mamani 2006. Dissemination and persistence of Shiga
R, Pita JM, Alonso MP, Llovo J, Bernardez toxin-producing Escherichia coli (STEC) strains
MI, Blanco JE, Blanco M, Blanco J. 2011. on French dairy farms. Vet Microbiol 117:180191.

284 BEUTIN AND FACH
27. Fremaux B, Prigent-Combaret C, Vernozy- seropathotypes that are linked to epidemic and/or
Rozand C. 2008. Long-term survival of Shiga serious disease. J Clin Microbiol 41:49304940.
toxin-producing Escherichia coli in cattle efflu- 38. Konczy P, Ziebell K, Mascarenhas M, Choi
ents and environment: an updated review. Vet A, Michaud C, Kropinski AM, Whittam TS,
Microbiol 132:118. Wickham M, Finlay B, Karmali MA. 2008.
28. Muniesa M, Serra-Moreno R, Jofre J. 2004. Genomic O island 122, locus for enterocyte efface-
Free Shiga toxin bacteriophages isolated from ment, and the evolution of virulent verocytotoxin-
sewage showed diversity although the stx genes producing Escherichia coli. J Bacteriol 190:5832
appeared conserved. Environ Microbiol 6:716 5840.
725. 39. Coombes BK, Wickham ME, Mascarenhas M,
29. Imamovic L, Jofre J, Schmidt H, Serra-Moreno Gruenheid S, Finlay BB, Karmali MA. 2008.
R, Muniesa M. 2009. Phage-mediated Shiga toxin Molecular analysis as an aid to assess the public
2 gene transfer in food and water. Appl Environ health risk of non-O157 Shiga toxin-producing
Microbiol 75:17641768. Escherichia coli strains. Appl Environ Microbiol
30. Scheutz F, Strockbine NA. 2005. Genus I. 74:21532160.
Escherichia, p 607624. In Garrity GM, Brenner 40. Bugarel M, Beutin L, Martin A, Gill A, Fach P.
DJ, Krieg NR, Staley JT (ed), Bergeys Manual of 2010. Micro-array for the identification of Shiga
Systematic Bacteriology, 2nd ed. Springer, New toxin-producing Escherichia coli (STEC) sero-
York, NY. pathotypes associated with hemorrhagic colitis
31. Bettelheim KA. 2007. The non-O157 Shiga- and hemolytic uremic syndrome in humans. Int J
toxigenic (verocytotoxigenic) Escherichia coli; Food Microbiol 142:318329.
under-rated pathogens. Crit Rev Microbiol 33: 41. Centers for Disease Control and Prevention.
6787. 1995. Outbreak of acute gastroenteritis attributable
32. Delannoy S, Beutin L, Fach P. 2013. Towards a to Escherichia coli serotype O104:H21Helena,
molecular definition of enterohemorrhagic Esch- Montana, 1994. Morb Mortal Wkly Rep 44:501
erichia coli (EHEC): detection of genes located on 503.
O island 57 as markers to distinguish EHEC from 42. Kaper JB, Nataro JP, Mobley HL. 2004. Path-
closely related enteropathogenic E. coli strains. ogenic Escherichia coli. Nat Rev Microbiol 2:123
J Clin Microbiol 51:10831088. 140.
33. European Centre for Disease Prevention and 43. Bielaszewska M, Mellmann A, Zhang W, Kock
Control and European Food Safety Authority. R, Fruth A, Bauwens A, Peters G, Karch H. 2011.
2011. Shiga toxin/verotoxin-producing Escherichia Characterisation of the Escherichia coli strain as-
coli in humans, food and animals in the EU/EEA, sociated with an outbreak of haemolytic uraemic
with special reference to the German outbreak syndrome in Germany, 2011: a microbiological
strain STEC O104. European Centre for Disease study. Lancet Infect Dis 11:671676.
Prevention and Control, Stockholm, Sweden, and 44. Bettelheim KA, Beutin L. 2003. Rapid laboratory
European Food Safety Authority, Parma, Italy. identification and characterization of verocyto-
34. Beutin L, Krause G, Zimmermann S, Kaulfuss S, toxigenic (Shiga toxin producing) Escherichia coli
Gleier K. 2004. Characterization of Shiga toxin- (VTEC/STEC). J Appl Microbiol 95:205217.
producing Escherichia coli strains isolated from 45. Paton JC, Paton AW. 2003. Methods for detec-
human patients in Germany over a 3-year period. tion of STEC in humans. An overview. Methods
J Clin Microbiol 42:10991108. Mol Med 73:926.
35. Kappeli U, Hachler H, Giezendanner N, Cheasty 46. Scheutz F, Teel LD, Beutin L, Pierard D,
T, Stephan R. 2010. Shiga toxin-producing Esch- Buvens G, Karch H, Mellmann A, Caprioli A,
erichia coli O157 associated with human infections Tozzoli R, Morabito S, Strockbine NA, Melton-
in Switzerland, 20002009. Epidemiol Infect 28: Celsa AR, Sanchez M, Persson S, OBrien AD.
18. 2012. Multicenter evaluation of a sequence-based
36. Brooks JT, Sowers EG, Wells JG, Greene KD, protocol for subtyping shiga toxins and stan-
Griffin PM, Hoekstra RM, Strockbine NA. dardizing stx nomenclature. J Clin Microbiol
2005. Non-O157 Shiga toxin-producing Esche- 50:29512963.
richia coli infections in the United States, 1983 47. Muthing J, Schweppe CH, Karch H, Friedrich
2002. J Infect Dis 192:14221429. AW. 2009. Shiga toxins, glycosphingolipid diver-
37. Karmali MA, Mascarenhas M, Shen S, Ziebell sity, and endothelial cell injury. Thromb Haemost
K, Johnson S, Reid-Smith R, Isaac-Renton J, 101:252264.
Clark C, Rahn K, Kaper JB. 2003. Association 48. Paton JC, Paton AW. 1998. Pathogenesis and
of genomic O island 122 of Escherichia coli EDL diagnosis of Shiga toxin-producing Escherichia
933 with verocytotoxin-producing Escherichia coli coli infections. Clin Microbiol Rev 11:450479.

49. Smith HR, Scotland SM. 1988. Vero cytotoxin- In Abuelzein E (ed), Trends in Immunolabelled
producing strains of Escherichia coli. J Med and Related Techniques, 1st ed. InTech, Rijeka,
Microbiol 26:7785. Croatia.
50. Strockbine NA, Wells JG, Bopp CA, Barrett TJ. 60. Feng PC, Jinneman K, Scheutz F, Monday SR.
1998. Overview of detection and subtyping 2011. Specificity of PCR and serological assays
methods, p 331356. In Kaper JB, OBrien AD in the detection of Escherichia coli Shiga toxin
(ed), Escherichia coli O157:H7 and Other Shiga subtypes. Appl Environ Microbiol 77:66996702.
Toxin-Producing E. coli Strains. American Society 61. Beutin L, Martin A, Krause G, Steege K, Haby
for Microbiology, Washington, DC. S, Pries K, Albrecht N, Miko A, Jahn S. 2010.
51. Beutin L, Steinruck H, Krause G, Steege K, Ergebnisse, Schlussfolgerungen und Empfehlungen
Haby S, Hultsch G, Appel B. 2007. Comparative aus zwei Ringversuchen zum Nachweis und zur
evaluation of the RidascreenVerotoxin enzyme Isolierung von Shiga (Vero) Toxin bildenden
immunoassay for detection of Shiga-toxin pro- Escherichia coli (STEC) aus Hackfleischproben. J
ducing strains of Escherichia coli (STEC) from Verbrauch Lebensm 5:2134.
food and other sources. J Appl Microbiol 102:630 62. Parma YR, Chacana PA, Lucchesi PM, Roge A,
639. Granobles Velandia CV, Kruger A, Parma AE,
52. Beutin L, Kruger U, Krause G, Miko A, Martin Fernandez-Miyakawa ME. 2012. Detection of
A, Strauch E. 2008. Evaluation of major types of Shiga toxin-producing Escherichia coli by sand-
Shiga toxin 2e producing Escherichia coli present wich enzyme-linked immunosorbent assay using
in food, pigs and in the environment as potential chicken egg yolk IgY antibodies. Front Cell Infect
pathogens for humans. Appl Environ Microbiol Microbiol 2:84.
74:48064816. 63. Kehl SC. 2002. Role of the laboratory in the di-
53. Staples M, Jennison AV, Graham RM, Smith agnosis of enterohemorrhagic Escherichia coli
HV. 2012. Evaluation of the Meridian Premier infections. J Clin Microbiol 40:27112715.
EHEC assay as an indicator of Shiga toxin pre- 64. Zhang W, Bielaszewska M, Pulz M, Becker K,
sence in direct faecal specimens. Diagn Microbiol Friedrich AW, Karch H, Kuczius T. 2008. A new
Infect Dis 73:322325. immuno-PCR assay for the detection of low
54. Segura-Alvarez M, Richter H, Conraths FJ, concentrations of Shiga toxin 2 and its variants.
Geue L. 2003. Evaluation of enzyme-linked im- J Clin Microbiol 46:12921297.
munosorbent assays and a PCR test for detection 65. Ge B, Zhao S, Hall R, Meng J. 2002. A PCR-
of shiga toxins for Shiga toxin-producing Esche- ELISA for detecting Shiga toxin-producing
richia coli in cattle herds. J Clin Microbiol 41: Escherichia coli. Microbes Infect 4:285290.
57605763. 66. Auvray F, Lecureuil C, Tache J, Leclerc V,
55. Watarai S, Tana, Inoue K, Kushi Y, Isogai E, Deperrois V, Lombard B. 2007. Detection,
Yokota K, Naka K, Oguma K, Kodama H. 2001. isolation and characterization of Shiga toxin-
Inhibition of Vero cell cytotoxic activity in Esch- producing Escherichia coli in retail-minced beef
erichia coli O157:H7 lysates by globotriaosylcera- using PCR-based techniques, immunoassays and
mide, Gb3, from bovine milk. Biosci Biotechnol colony hybridization. Lett Appl Microbiol 45:646
Biochem 65:414419. 651.
56. Gallegos KM, Conrady DG, Karve SS, 67. Fach P, Perelle S, Dilasser F, Grout J. 2001.
Gunasekera TS, Herr AB, Weiss AA. 2012. Shiga Comparison between a PCR-ELISA test and the
toxin binding to glycolipids and glycans. PLoS vero cell assay for detecting Shiga toxin-producing
One 7:e30368. Escherichia coli in dairy products and character-
57. Beutin L, Zimmermann S, Gleier K. 1996. ization of virulence traits of the isolated strains.
Rapid detection and isolation of Shiga-like toxin J Appl Microbiol 90:809818.
(verocytotoxin)-producing Escherichia coli by 68. Beutin L, Zimmermann S, Gleier K. 2002.
direct testing of individual enterohemolytic colo- Evaluation of the VTEC-screen Seiken test
nies from washed sheep blood agar plates in the for detection of different types of Shiga toxin
VTEC-RPLA assay. J Clin Microbiol 34:28122814. (verotoxin)-producing Escherichia coli (STEC) in
58. Karmali MA, Petric M, Bielaszewska M. 1999. human stool samples. Diagn Microbiol Infect Dis
Evaluation of a microplate latex agglutination 42:18.
method (Verotox-F assay) for detecting and char- 69. Klie H, Timm M, Richter H, Gallien P, Perlberg
acterizing verotoxins (Shiga toxins) in Escherichia KW, Steinruck H. 1997. Detection and occur-
coli. J Clin Microbiol 37:396399. rence of verotoxin-forming and/or shigatoxin
59. Burgos Y, Beutin L. 2012. Evaluation of an producing Escherichia coli (VTEC and/or STEC)
immuno-chromatographic detection system for in milk. Berl Munch Tierarztl Wochenschr 110:
Shiga toxins and the E. coli O157 antigen, p 2940. 337341 (In German).

286 BEUTIN AND FACH
70. Richter H, Klie H, Timm M, Gallien P, 81. Lin Z, Kurazono H, Yamasaki S, Takeda Y. 1993.
Steinruck H, Perlberg KW, Protz D. 1997. Detection of various variant verotoxin genes in
Verotoxin-producing E. coli (VTEC) in feces Escherichia coli by polymerase chain reaction.
from cattle slaughtered in Germany. Berl Munch Microbiol Immunol 37:543548.
Tierarztl Wochenschr 110:121127 (In German). 82. Karch H, Meyer T. 1989. Single primer pair
71. Desrosiers A, Fairbrother JM, Johnson RP, for amplifying segments of distinct Shiga-like-
Desautels C, Letellier A, Quessy S. 2001. Phe- toxin genes by polymerase chain reaction. J Clin
notypic and genotypic characterization of Esche- Microbiol 27:27512757.
richia coli verotoxin-producing isolates from 83. Paton AW, Paton JC. 1998. Detection and char-
humans and pigs. J Food Prot 64:19041911. acterization of Shiga toxigenic Escherichia coli by
72. Hull AE, Acheson DW, Echeverria P, Donohue- using multiplex PCR assays for stx1, stx2, eaeA,
Rolfe A, Keusch GT. 1993. Mitomycin immuno- enterohemorrhagic E. coli hlyA, rfbO111, and
blot colony assay for detection of Shiga-like rfbO157. J Clin Microbiol 36:598602.
toxin-producing Escherichia coli in fecal samples: 84. Gannon VP, King RK, Kim JY, Thomas EJ.
comparison with DNA probes. J Clin Microbiol 1992. Rapid and sensitive method for detection of
31:11671172. Shiga-like toxin-producing Escherichia coli in
73. Timm M, Klie H, Richter H, Perlberg KW. ground beef using the polymerase chain reaction.
1996. A method for specific isolation of verotoxin- Appl Environ Microbiol 58:38093815.
producing Escherichia coli colonies. Berl Munch 85. Ziebell KA, Read SC, Johnson RP, Gyles CL.
Tierarztl Wochenschr 109:270272 (In German). 2002. Evaluation of PCR and PCR-RFLP proto-
74. Thomas A, Smith HR, Willshaw GA, Rowe B. cols for identifying Shiga toxins. Res Microbiol
1991. Non-radioactively labelled polynucleotide 153:289300.
and oligonucleotide DNA probes, for selectively 86. Beutin L, Miko A, Krause G, Pries K, Haby S,
detecting Escherichia coli strains producing Vero Steege K, Albrecht N. 2007. Identification of
cytotoxins VT1, VT2 and VT2 variant. Mol Cell human-pathogenic strains of Shiga toxin-producing
Probes 5:129135. Escherichia coli from food by a combination of
75. Karch H, Meyer T. 1989. Evaluation of oligonu- serotyping and molecular typing of Shiga toxin
cleotide probes for identification of Shiga-like- genes. Appl Environ Microbiol 73:47694775.
toxin-producing Escherichia coli. J Clin Microbiol 87. Miko A, Pries K, Haby S, Steege K, Albrecht N,
27:11801186. Krause G, Beutin L. 2009. Assessment of Shiga
76. Tokhi AM, Peiris JS, Scotland SM, Willshaw toxin-producing Escherichia coli isolates from
GA, Smith HR, Cheasty T. 1993. A longitudinal wildlife meat as potential pathogens for humans.
study of Vero cytotoxin producing Escherichia coli Appl Environ Microbiol 75:64626470.
in cattle calves in Sri Lanka. Epidemiol Infect 88. Stephan R, Zweifel C, Fach P, Morabito S,
110:197208. Beutin L. 2011. Shiga toxin-producing Esche-
77. Todd EC, Szabo RA, MacKenzie JM, richia coli in food, p 229239. In Hoorfar J
Martin A, Rahn K, Gyles C, Gao A, Alves D, Yee (ed), Rapid Detection, Characterization, and Enu-
AJ. 1999. Application of a DNA hybridization- meration of Foodborne Pathogens. ASM Press,
hydrophobic-grid membrane filter method for Washington, DC.
detection and isolation of verotoxigenic Esche- 89. Wang G, Clark CG, Rodgers FG. 2002. Detection
richia coli. Appl Environ Microbiol 65:47754780. in Escherichia coli of the genes encoding the
78. Montenegro MA, Bulte M, Trumpf T, Aleksic S, major virulence factors, the genes defining the
Reuter G, Bulling E, Helmuth R. 1990. Detection O157:H7 serotype, and components of the type 2
and characterization of fecal verotoxin-producing Shiga toxin family by multiplex PCR. J Clin
Escherichia coli from healthy cattle. J Clin Microbiol 40:36133619.
Microbiol 28:14171421. 90. Peitz R, Weber H, Gleier K, Zimmermann S,
79. Dorn CR, Scotland SM, Smith HR, Willshaw Beutin L. 2000. Detection of enterohemorrhagic
GA, Rowe B. 1989. Properties of Vero cytotoxin- Escherichia coli (EHEC) in raw meat and raw
producing Escherichia coli of human and animal meat sausages. Fleischwirtschaft 3:7174.
origin belonging to serotypes other than O157:H7. 91. Urdahl AM, Solheim HT, Vold L, Hasseltvedt V,
Epidemiol Infect 103:8395. Wasteson Y. 2013. Shiga toxin-encoding genes
80. Kobayashi H, Shimada J, Nakazawa M, (stx genes) in human faecal samples. APMIS
Morozumi T, Pohjanvirta T, Pelkonen S, 121:202210.
Yamamoto K. 2001. Prevalence and character- 92. Wang F, Jiang L, Yang Q, Prinyawiwatkul W,
istics of shiga toxin-producing Escherichia coli Ge B. 2012. Rapid and specific detection of
from healthy cattle in Japan. Appl Environ Escherichia coli serogroups O26, O45, O103, O111,
Microbiol 67:484489. O121, O145, and O157 in ground beef, beef trim,

and produce by loop-mediated isothermal ampli- O157 strains and the aggregative EHEC O104:H4
fication. Appl Environ Microbiol 78:27272736. strain from ready-to-eat vegetables. Int J Food
93. Reischl U, Youssef MT, Kilwinski J, Lehn N, Microbiol 152:1930.
Zhang WL, Karch H, Strockbine NA. 2002. 103. Chui L, Couturier MR, Chiu T, Wang G, Olson
Real-time fluorescence PCR assays for detection AB, McDonald RR, Antonishyn NA, Horsman
and characterization of Shiga toxin, intimin, G, Gilmour MW. 2010. Comparison of Shiga
and enterohemolysin genes from Shiga toxin- toxin-producing Escherichia coli detection meth-
producing Escherichia coli. J Clin Microbiol ods using clinical stool samples. J Mol. Diagn
40:25552565. 12:469475.
94. Sekse C, Solberg A, Petersen A, Rudi K, 104. International Organization for Standardization
Wasteson YB. 2005. Detection and quantification (ISO). 2012. ISO/TS 13136:2012, Microbiology of
of Shiga toxin-encoding genes in sheep faeces by food and animal feedReal-time polymerase chain
real-time PCR. Mol Cell Probes 19:363370. reaction (PCR)-based method for the detection of
95. Stefan A, Scaramagli S, Bergami R, Mazzini C, food-borne pathogensHorizontal method for the
Barbanera M, Perelle S, Fach P. 2007. Real-time detection of Shiga toxin-producing Escherichia coli
PCR and enzyme-linked fluorescent assay meth- (STEC) and the determination of O157, O111, O26,
ods for detecting Shiga-toxin-producing Esche- O103 and O145 serogroups. (ISO/TS 13136:2012).
richia coli in mincemeat samples. Can J Microbiol 11-7-2012. International Organization for Stan-
53:337342. dardization, ISO Central Secretariat, Geneva,
96. Perelle S, Dilasser F, Grout J, Fach P. 2007. Switzerland.
Screening food raw materials for the presence of 105. USDA Working Group. 2011. Detection and
the worlds most frequent clinical cases of Shiga isolation of non-O157 Shiga-toxin producing
toxin-encoding Escherichia coli O26, O103, O111, Escherichia coli strains (STEC) from meat prod-
O145 and O157. Int J Food Microbiol 113:284288. ucts, p 116. In Laboratory QA/QC Division (ed),
97. Perelle S, Dilasser F, Grout J, Fach P. 2004. Laboratory Guidebook. U.S. Department of Agri-
Detection by 5-nuclease PCR of Shiga-toxin culture, Laboratory QA/QC Division, Athens, GA.
producing Escherichia coli O26, O55, O91, O103, 106. European Food Safety Authority. 2012. Manual
O111, O113, O145 and O157:H7, associated with the for Reporting on Zoonoses, Zoonotic Agents and
worlds most frequent clinical cases. Mol Cell Antimicrobial Resistance in the Framework of Di-
Probes 18:185192. rective 2003/99/EC and of Some Other Pathogenic
98. Chassagne L, Pradel N, Robin F, Livrelli V, Microbiological Agents for Information Derived
Bonnet R, Delmas J. 2009. Detection of stx1, stx2, from the Year 2011. European Food Safety Au-
and eae genes of enterohemorrhagic Escherichia coli thority, Parma, Italy.
using SYBR Green in a real-time polymerase chain 107. European Food Safety Authority. 2012. The
reaction. Diagn Microbiol Infect Dis 64:98101. European Union summary report on trends and
99. Beutin L, Jahn S, Fach F. 2009. Evaluation of the sources of zoonoses, zoonotic agents and food-
GeneDisc real-time PCR system for detection borne outbreaks in 2010. EFSA J 10:2597.
of enterohaemorrhagic Escherichia coli (EHEC) 108. Mellmann A, Bielaszewska M, Kock R, Friedrich
O26, O103, O111, O145 and O157 strains accord- AW, Fruth A, Middendorf B, Harmsen D,
ing to their virulence markers and their O- and Schmidt MA, Karch H. 2008. Analysis of collec-
H-antigen-associated genes. J Appl Microbiol tion of hemolytic uremic syndrome-associated
106:11221132. enterohemorrhagic Escherichia coli. Emerg Infect
100. Fratamico PM, Bagi LK. 2012. Detection of Shiga Dis 14:12871290.
toxin-producing Escherichia coli in ground beef 109. Nielsen EM, Andersen MT. 2003. Detection
using the GeneDisc real-time PCR system. Front and characterization of verocytotoxin-producing
Cell Infect Microbiol 2:152. Escherichia coli by automated 5 nuclease PCR
101. Fratamico PM, Bagi LK, Cray WC Jr, Narang assay. J Clin Microbiol 41:28842893.
N, Yan X, Medina M, Liu Y. 2011. Detection by 110. Shelton DR, Karns JS, Higgins JA, Van Kessel
multiplex real-time polymerase chain reaction JA, Perdue ML, Belt KT, Russell-Anelli J,
assays and isolation of Shiga toxin-producing Debroy C. 2006. Impact of microbial diversity on
Escherichia coli serogroups O26, O45, O103, O111, rapid detection of enterohemorrhagic Escherichia
O121, and O145 in ground beef. Foodborne Pathog coli in surface waters. FEMS Microbiol Lett
Dis 8:601607. 261:95101.
102. Tzschoppe M, Martin A, Beutin L. 2012. A rapid 111. Krause G, Zimmermann S, Beutin L. 2005. In-
procedure for the detection and isolation of vestigation of domestic animals and pets as a
enterohaemorrhagic Escherichia coli (EHEC) reservoir for intimin (eae) gene positive Esche-
serogroup O26, O103, O111, O118, O121, O145 and richia coli types. Vet Microbiol 106:8795.

288 BEUTIN AND FACH
112. Kozub-Witkowski E, Krause G, Frankel G, enterocyte effacement-negative Shiga-toxigenic

Kramer D, Appel B, Beutin L. 2008. Serotypes Escherichia coli strains that are virulent for
and virutypes of enteropathogenic and entero- humans. Infect Immun 69:69997009.
haemorrhagic Escherichia coli strains from stool 123. Hauser E, Mellmann A, Semmler T, Stoeber H,
samples of children with diarrhoea in Germany. Wieler LH, Karch H, Kuebler N, Fruth A,
J Appl Microbiol 104:403410. Harmsen D, Weniger T, Tietze E, Schmidt H.
113. Feng PC, Keys C, Lacher D, Monday SR, 2013. Phylogenetic and molecular analysis of
Shelton D, Rozand C, Rivas M, Whittam T. food-borne Shiga toxin-producing Escherichia
2010. Prevalence, characterization and clonal coli. Appl Environ Microbiol 79:27312740.
analysis of Escherichia coli O157: non-H7 sero- 124. Shimizu T, Ohta Y, Noda M. 2009. Shiga toxin 2
types that carry eae alleles. FEMS Microbiol Lett is specifically released from bacterial cells by two
308:6267. different mechanisms. Infect Immun 77:2813
114. Alonso MZ, Padola NL, Parma AE, Lucchesi 2823.
PM. 2011. Enteropathogenic Escherichia coli 125. Rocha LB, Piazza RM. 2007. Production of Shiga
contamination at different stages of the chicken toxin by Shiga toxin-expressing Escherichia coli
slaughtering process. Poult Sci 90:26382641. (STEC) in broth media: from divergence to defi-
115. Anklam KS, Kanankege KS, Gonzales TK, nition. Lett Appl Microbiol 45:411417.
Kaspar CW, Dopfer D. 2012. Rapid and reliable 126. McGannon CM, Fuller CA, Weiss AA. 2010.
detection of Shiga toxin-producing Escherichia Different classes of antibiotics differentially influ-
coli by real-time multiplex PCR. J Food Prot ence shiga toxin production. Antimicrob Agents
75:643650. Chemother 54:37903798.
116. Sekse C, OSullivan K, Granum PE, Rorvik LM, 127. Zhang X, McDaniel AD, Wolf LE, Keusch GT,
Wasteson Y, Jorgensen HJ. 2009. An outbreak Waldor MK, Acheson DW. 2000. Quinolone
of Escherichia coli O103. Int J Food Microbiol antibiotics induce Shiga toxin-encoding bacterio-
133:259264. phages, toxin production, and death in mice.
117. Iguchi A, Iyoda S, Ohnishi M. 2012. Molecular J Infect Dis 181:664670.
characterization reveals three distinct clonal 128. Bielaszewska M, Idelevich EA, Zhang W,
groups among clinical shiga toxin-producing Bauwens A, Schaumburg F, Mellmann A,
Escherichia coli strains of serogroup O103. J Clin Peters G, Karch H. 2012. Effects of antibiotics on
Microbiol 50:28942900. Shiga toxin 2 production and bacteriophage induc-
118. Maidhof H, Guerra B, Abbas S, Elsheikha HM, tion by epidemic Escherichia coli O104:H4 strain.
Whittam TS, Beutin L. 2002. A multiresistant Antimicrob Agents Chemother 56:32773282.
clone of Shiga toxin-producing Escherichia coli 129. Hussein HS, Bollinger LM. 2008. Influence of
O118:[H16] is spread in cattle and humans over selective media on successful detection of Shiga
different European countries. Appl Environ toxin-producing Escherichia coli in food, fecal,
Microbiol 68:58345842. and environmental samples. Foodborne Pathog
119. Bielaszewska M, Prager R, Vandivinit L, Dis 5:227244.
Musken A, Mellmann A, Holt NJ, Tarr PI, 130. International Organization for Standardization.
Karch H, Zhang W. 2009. Detection and char- 2009. ISO 16654: Microbiology of food and animal
acterization of the fimbrial sfp cluster in entero- feeding stuffshorizontal method for the detection
hemorrhagic Escherichia coli O165:H25/NM of Escherichia coli O157. International Organiza-
isolates from humans and cattle. Appl Environ tion for Standardization, Geneva, Switzerland.
Microbiol 75:6471. 131. de Boer E, Heuvelink AE. 2000. Methods for the
120. Delannoy S, Beutin L, Burgos YK, Fach P. 2012. detection and isolation of Shiga toxin-producing
Specific detection of enteroaggregative hemor- Escherichia coli. Symp Ser Soc Appl Microbiol
rhagic Escherichia coli O104:H4 strains using the 29:133S143S.
CRISPR locus as target for a diagnostic real-time 132. World Health Organization. 1998. Zoonotic non-
PCR. J Clin Microbiol 50:34853492. O157 Shiga toxin-producing Escherichia coli
121. Jelacic JK, Damrow T, Chen GS, Jelacic S, (STEC). Report of a WHO Scientific Working
Bielaszewska M, Ciol M, Carvalho HM, Melton- Group Meeting, Berlin, Germany, 2326 June
Celsa AR, OBrien AD, Tarr PI. 2003. Shiga 1998. WHO/CSR/APH/98.8, 130. 1998. World
toxin-producing Escherichia coli in Montana: Health Organization, Geneva, Switzerland.
bacterial genotypes and clinical profiles. J Infect 133. Vimont A, Vernozy-Rozand C, Delignette-
Dis 188:719729. Muller ML. 2006. Isolation of E. coli O157:H7 and
122. Paton AW, Srimanote P, Woodrow MC, Paton non-O157 STEC in different matrices: review of
JC. 2001. Characterization of Saa, a novel auto- the most commonly used enrichment protocols.
agglutinating adhesin produced by locus of Lett Appl Microbiol 42:102108.

134. Bolton DJ, OSullivan J, Duffy G, Baylis CL, 145. Hussein HS, Bollinger LM, Hall MR. 2008.
Tozzoli R, Wasteson Y, Lofdahl S (ed). 2007. Growth and enrichment medium for detection
Methods for Detection and Molecular Charac- and isolation of Shiga toxin-producing Escherichia
terisation of Pathogenic Escherichia coli. Ashtown coli in cattle feces. J Food Prot 71:927933.
Food Research Centre, Ashtown, Ireland. 146. Gill A, Martinez-Perez A, McIlwham S, Blais B.
135. Vimont A, Vernozy-Rozand C, Montet MP, 2012. Development of a method for the detection
Lazizzera C, Bavai C, Delignette-Muller ML. of verotoxin-producing Escherichia coli in food.
2006. Modeling and predicting the simultaneous J Food Prot 75:827837.
growth of Escherichia coli O157:H7 and ground 147. Lehmacher A, Meier H, Aleksic S, Bockemuhl J.
beef background microflora for various enrich- 1998. Detection of hemolysin variants of Shiga
ment protocols. Appl Environ Microbiol 72:261 toxin-producing Escherichia coli by PCR and
268. culture on vancomycin-cefixime-cefsulodin blood
136. Baylis CL. 2008. Growth of pure cultures of agar. Appl Environ Microbiol 64:24492453.
verocytotoxin-producing Escherichia coli in a range 148. Lin A, Nguyen L, Clotilde LM, Kase JA, Son
of enrichment media. J Appl Microbiol 105:1259 I, Lauzon CR. 2012. Isolation of Shiga toxin-
1265. producing Escherichia coli from fresh produce
137. Kanki M, Seto K, Harada T, Yonog S, Kumeda using STEC heart infusion washed blood agar
Y. 2011. Comparison of four enrichment broths with mitomycin-C. J Food Prot 75:20282030.
for the detection of non-O157 Shiga-toxin- 149. Sugiyama K, Inoue K, Sakazaki R. 2001.
producing Escherichia coli O91, O103, O111, O119, Mitomycin-supplemented washed blood agar for
O121, O145 and O165 from pure culture and food the isolation of Shiga toxin-producing Esche-
samples. Lett Appl Microbiol 53:167173. richia coli other than O157:H7. Lett Appl Microbiol
138. Auvray F, Lecureuil C, Dilasser F, Tache J, 33:193195.
Derzelle S. 2009. Development of a real-time 150. Posse B, De Zutter L, Heyndrickx M, Herman
PCR assay with an internal amplification control L. 2008. Novel differential and confirmation
for the screening of Shiga toxin-producing Esch- plating media for Shiga toxin-producing Esche-
erichia coli in foods. Lett Appl Microbiol 48:554 richia coli serotypes O26, O103, O111, O145 and
559. sorbitol-positive and -negative O157. FEMS
139. Jahn S, Weber H, Beutin L. 2008. Comparison of Microbiol Lett 282:124131.
enzyme immunoassay and quantitiver real time 151. Posse B, De Zutter L, Heyndrickx M, Herman
PCR as proof of Shigatoxin producing Escherichia L. 2008. Quantitative isolation efficiency of O26,
coli (STEC) in mincemeat. J Verbrauch Lebensm O103, O111, O145 and O157 STEC serotypes from
3:385395 (In German). artificially contaminated food and cattle faeces
140. Doyle MP, Schoeni JL. 1984. Survival and samples using a new isolation protocol. J Appl Mi-
growth characteristics of Escherichia coli associ- crobiol 105:227235.
ated with hemorrhagic colitis. Appl Environ 152. Vimont A, Delignette-Muller ML, Vernozy-
Microbiol 48:855856. Rozand C. 2007. Supplementation of enrichment
141. Lack WK, Becker B, Holzapfel WH. 1996. broths by novobiocin for detecting Shiga toxin-
Hygienischer Status frischer vorverpackter producing Escherichia coli from food: a contro-
Mischsalate im Jahr 1995. Arch Lebensmittelhyg versial use. Lett Appl Microbiol 44:326331.
47:129152. 153. Gouali M, Ruckly C, Carle I, Lejay-Collin M,
142. Klepzig I, Teufel P, Schott W, Hildebrandt G. Weill FX. 2013. Evaluation of CHROMagar STEC
1999. Auswirkungen einer Unterbrechung der and STEC O104 chromogenic agar media for
Khlkette auf die mikrobiologische Beschaffenheit detection of Shiga toxin-producing Escherichia coli
von vorzerkleinerten Mischsalaten. Arch Lebens- in stool specimens. J Clin Microbiol 51:894900.
mittelhyg 50:95104. 154. Taylor DE, Rooker M, Keelan M, Ng LK,
143. Sagoo SK, Little CL, Mitchell RT. 2003. Micro- Martin I, Perna NT, Burland NT, Blattner FR.
biological quality of open ready-to-eat salad 2002. Genomic variability of O islands encoding
vegetables: effectiveness of food hygiene training tellurite resistance in enterohemorrhagic Esche-
of management. J Food Prot 66:15811586. richia coli O157:H7 isolates. J Bacteriol 184:4690
144. Madic J, Vingadassalon N, Peytavin de Garam 4698.
C, Marault M, Scheutz F, Brugre H, Jamet 155. Bielaszewska M, Tarr PI, Karch H, Zhang W,
E, Auvray F. 2011. Detection of Shiga toxin- Mathys W. 2005. Phenotypic and molecular
producing Escherichia coli (STEC) O26:H11, analysis of tellurite resistance among entero-
O103:H2, O111:H8, O145:H28 and O157:H7 in raw- hemorrhagic Escherichia coli O157:H7 and sorbi-
milk cheeses by using multiplex real-time PCR. tol-fermenting O157:NM clinical isolates. J Clin
Appl Environ Microbiol 77:20352041. Microbiol 43:452454.

290 BEUTIN AND FACH
156. Orth D, Grif K, Dierich MP, Wurzner R. 2007. 167. Bugarel M, Beutin L, Scheutz F, Loukiadis E,
Variability in tellurite resistance and the ter gene Fach P. 2011. Identification of genetic markers for
cluster among Shiga toxin-producing Escherichia differentiation of Shiga toxin-producing, entero-
coli isolated from humans, animals and food. Res pathogenic and avirulent strains of Escherichia
Microbiol 158:105111. coli O26. Appl Environ Microbiol 77:22752281.
157. Bielaszewska M, Middendorf B, Tarr PI, Zhang 168. Gyles C, Johnson R, Gao A, Ziebell K, Pierard
W, Prager R, Aldick T, Dobrindt U, Karch H, D, Aleksic S, Boerlin P. 1998. Association of
Mellmann A. 2011. Chromosomal instability in enterohemorrhagic Escherichia coli hemolysin
enterohaemorrhagic Escherichia coli O157:H7: with serotypes of shiga-like-toxin-producing
impact on adherence, tellurite resistance and Escherichia coli of human and bovine origins. Appl
colony phenotype. Mol Microbiol 79:10241044. Environ Microbiol 64:41344141.
158. Schmidt H, Karch H, Beutin L. 1994. The large- 169. Aidar-Ugrinovich L, Blanco J, Blanco M, Blanco
sized plasmids of enterohemorrhagic Escherichia JE, Leomil L, Dahbi G, Mora A, Onuma DL,
coli O157 strains encode hemolysins which are pre- Silveira WD, Pestana de Castro AF. 2007.
sumably members of the E. coli alpha-hemolysin Serotypes, virulence genes, and intimin types of
family. FEMS Microbiol Lett 117:189196. Shiga toxin-producing Escherichia coli (STEC)
159. Brunder W, Schmidt H, Frosch M, Karch H. and enteropathogenic E. coli (EPEC) isolated
1999. The large plasmids of Shiga-toxin-producing from calves in Sao Paulo, Brazil. Int J Food
Escherichia coli (STEC) are highly variable genetic Microbiol 115:297306.
elements. Microbiology 145(Pt 5):10051014. 170. Mora A, Blanco M, Blanco JE, Dahbi G, Lopez
160. Beutin L, Montenegro MA, Orskov I, Prada J, C, Justel P, Alonso MP, Echeita A, Bernardez
Zimmermann S, Stephan R. 1989. Close associ- MI, Gonzalez EA, Blanco J. 2007. Serotypes,
ation of verotoxin (Shiga-like toxin) production virulence genes and intimin types of Shiga toxin
with enterohemolysin production in strains of (verocytotoxin)-producing Escherichia coli iso-
Escherichia coli. J Clin Microbiol 27:25592564. lates from minced beef in Lugo (Spain) from 1995
161. Burgos Y, Beutin L. 2010. Common origin of through 2003. BMC Microbiol 7:13.
plasmid encoded alpha-hemolysin genes in 171. Pradel N, Livrelli V, Champs C, Palcoux JB,
Escherichia coli. BMC Microbiol 10:193. Reynaud A, Scheutz F, Sirot J, Joly B, Forestier
162. Prada J, Baljer G, De Rycke J, Steinruck H, C. 2000. Prevalence and characterization of Shiga
Zimmermann S, Stephan R, Beutin L. 1991. toxin-producing Escherichia coli isolated from
Characteristics of alpha-hemolytic strains of cattle, food, and children during a one-year pro-
Escherichia coli isolated from dogs with gastro- spective study in France. J Clin Microbiol
enteritis. Vet Microbiol 29:5973. 38:10231031.
163. Bugarel M, Beutin L, Fach P. 2010. Low-density 172. Kaufmann M, Zweifel C, Blanco M, Blanco JE,
macroarray targeting non-locus of enterocyte ef- Blanco J, Beutin L, Stephan RB. 2006. Esche-
facement effectors (nle genes) and major viru- richia coli O157 and non-O157 Shiga toxin-
lence factors of Shiga toxin-producing Escherichia producing Escherichia coli in fecal samples of
coli (STEC): a new approach for molecular risk finished pigs at slaughter in Switzerland. J Food
assessment of STEC isolates. Appl Environ Prot 69:260266.
Microbiol 76:203211. 173. Karama M, Gyles CL. 2010. Methods for
164. Karch H, Bielaszewska M. 2001. Sorbitol- genotyping verotoxin-producing Escherichia coli.
fermenting Shiga toxin-producing Escherichia coli Zoonoses Public Health 57:447462.
O157:H() strains: epidemiology, phenotypic and 174. Karch H, Bohm H, Schmidt H, Gunzer F,
molecular characteristics, and microbiological Aleksic S, Heesemann J. 1993. Clonal structure
diagnosis. J Clin Microbiol 39:20432049. and pathogenicity of Shiga-like toxin-producing,
165. Bielaszewska M, Kock R, Friedrich AW, von sorbitol-fermenting Escherichia coli O157:H.
Eiff C, Zimmerhackl LB, Karch H, Mellmann A. J Clin Microbiol 31:12001205.
2007. Shiga toxin-mediated hemolytic uremic 175. Feng P, Lampel KA, Karch H, Whittam TS.
syndrome: time to change the diagnostic para- 1998. Genotypic and phenotypic changes in the
digm? PLoS One 2:e1024. emergence of Escherichia coli O157:H7. J Infect
166. Miko A, Lindstedt BA, Brandal LT, Lobersli I, Dis 177:17501753.
Beutin L. 2010. Evaluation of multiple-locus 176. Murinda SE, Batson SD, Nguyen LT, Gillespie
variable number of tandem-repeats analysis BE, Oliver SP. 2004. Phenotypic and genetic
(MLVA) as a method for identification of clonal markers for serotype-specific detection of Shiga
groups among enteropathogenic, enterohaemor- toxin-producing Escherichia coli O26 strains
rhagic and avirulent Escherichia coli O26 strains. from North America. Foodborne Pathog Dis 1:125
FEMS Microbiol Lett 303:137146. 135.

177. Prager R, Liesegang A, Voigt W, Rabsch W, 188. Keskimaki M, Eklund M, Pesonen H,

Fruth A, Tschape H. 2002. Clonal diversity Heiskanen T, Siitonen A. 2001. EPEC, EAEC and
of Shiga toxin-producing Escherichia coli STEC in stool specimens: prevalence and molec-
O103:H2/H(-) in Germany. Infect Genet Evol ular epidemiology of isolates. Diagn Microbiol
1:265275. Infect Dis 40:151156.
178. Willshaw GA, Smith HR, Cheasty T, OBrien SJ. 189. Orskov F, Orskov I. 1984. Serotyping of Esche-
2001. Use of strain typing to provide evidence richia coli, p 43112. In Bergan T (ed), Methods in
for specific interventions in the transmission of Microbiology. Academic Press, London, United
VTEC O157 infections. Int J Food Microbiol Kingdom.
66:3946. 190. Scheutz F, Cheasty T, Woodward D, Smith HR.
179. Schroeder CM, Meng J, Zhao S, Debroy C, 2004. Designation of O174 and O175 to temporary
Torcolini J, Zhao C, McDermott PF, Wagner O groups OX3 and OX7, and six new E. coli O
DD, Walker RD, White DG. 2002. Antimicrobial groups that include verocytotoxin-producing E.
resistance of Escherichia coli O26, O103, O111, coli (VTEC): O176, O177, O178, O179, O180 and
O128, and O145 from animals and humans. Emerg O181. APMIS 112:569584.
Infect Dis 8:14091414. 191. Blanco M, Padola NL, Krger A, Sanz ME,
180. Ziebell K, Johnson RP, Kropinski AM, Reid- Blanco JE, Gonzalez EA, Dahbi G, Mora A,
Smith R, Ahmed R, Gannon VP, Gilmour M, Bernardez MI, Etcheverria AI, Arroyo GH,
Boerlin P. 2011. Gene cluster conferring strepto- Lucchesi PMA, Parma AE, Blanco J. 2004.
mycin, sulfonamide, and tetracycline resistance in Virulence genes and intimin types of Shiga-toxin-
Escherichia coli O157:H7 phage types 23, 45, and producing Escherichia coli isolated from cattle
67. Appl Environ Microbiol 77:19001903. and beef products in Argentina. Int Microbiol
181. Souza MR, Klassen G, Toni FD, Rigo LU, 7:269276.
Henkes C, Pigatto CP, Dalagassa CB, Fadel- 192. Much P, Pichler J, Kasper SS, Allerberger F.
Picheth CM. 2010. Biochemical properties, 2009. Foodborne outbreaks, Austria 2007. Wien
enterohaemolysin production and plasmid car- Klin Wochenschr 121:7785.
riage of Shiga toxin-producing Escherichia coli 193. Scheutz F. 2012. Technical report: external quality
strains. Mem Inst Oswaldo Cruz 105:318321. assurance scheme for typing of verocytotoxin-pro-
182. Zhang WL, Bielaszewska M, Liesegang A, ducing E. coli (VTEC). 128. 1-4-2012. Stockholm,
Tschape H, Schmidt H, Bitzan M, Karch H. European Centre for Disease Prevention and
2000. Molecular characteristics and epidemio- Control.
logical significance of Shiga toxin-producing 194. Wang L, Rothemund D, Curd H, Reeves PR.
Escherichia coli O26 strains. J Clin Microbiol 2003. Species-wide variation in the Escherichia
38:21342140. coli flagellin (H-antigen) gene. J Bacteriol
183. Pradel N, Boukhors K, Bertin Y, Forestier C, 185:29362943.
Martin C, Livrelli V. 2001. Heterogeneity of Shiga 195. Machado J, Grimont F, Grimont PA. 2000.
toxin-producing Escherichia coli strains isolated Identification of Escherichia coli flagellar types by
from hemolytic-uremic syndrome patients, cattle, restriction of the amplified fliC gene. Res
and food samples in central France. Appl Environ Microbiol 151:535546.
Microbiol 67:24602468. 196. Prager R, Strutz U, Fruth A, Tschape H. 2003.
184. Chaudhuri RR, Henderson IR. 2012. The evo- Subtyping of pathogenic Escherichia coli strains
lution of the Escherichia coli phylogeny. Infect using flagellar (H)-antigens: serotyping versus
Genet Evol 12:214226. fliC polymorphisms. Int J Med Microbiol 292:
185. Whittam TS, Wolfe ML, Wachsmuth IK, 477486.
Orskov F, Orskov I, Wilson RA. 1993. Clonal 197. Fields PI, Blom K, Hughes HJ, Helsel LO,
relationships among Escherichia coli strains that Feng P, Swaminathan B. 1997. Molecular char-
cause hemorrhagic colitis and infantile diarrhea. acterization of the gene encoding H antigen in
Infect Immun 61:16191629. Escherichia coli and development of a PCR-re-
186. Abu-Ali GS, Lacher DW, Wick LM, Qi W, striction fragment length polymorphism test for
Whittam TS. 2009. Genomic diversity of patho- identification of E. coli O157:H7 and O157:NM.
genic Escherichia coli of the EHEC 2 clonal J Clin Microbiol 35:10661070.
complex. BMC Genomics 10:296. 198. Beutin L, Strauch E. 2007. Identification of se-
187. Nielsen EM, Scheutz F, Torpdahl M. 2006. quence diversity in the Escherichia coli fliC genes
Continuous surveillance of Shiga toxin-producing encoding flagellar types H8 and H40 and its use
Escherichia coli infections by pulsed-field gel in typing of Shiga toxin-producing E. coli O8, O22,
electrophoresis shows that most infections are O111, O174, and O179 strains. J Clin Microbiol
sporadic. Foodborne Pathog Dis 3:8187. 45:333339.

292 BEUTIN AND FACH
199. Ayala CO, Ramos Moreno AC, Martinez MB, based suspension array. J Microbiol Methods
Burgos YK, Pestana de Castro AF, Bando SY. 87:105110.
2012. Determination of flagellar types by PCR- 210. Lin A, Sultan O, Lau HK, Wong E, Hartman G,
RFLP analysis of enteropathogenic Escherichia Lauzon CR. 2011. O serogroup specific real time
coli (EPEC) and Shiga toxin-producing E. coli PCR assays for the detection and identification of
(STEC) strains isolated from animals in Sao nine clinically relevant non-O157 STECs. Food
Paulo, Brazil. Res Vet Sci 92:1823. Microbiol 28:478483.
200. Wang L, Rothemund D, Curd H, Reeves PR. 211. Gilmour MW, Olson AB, Andrysiak AK, Ng LK,
2000. Sequence diversity of the Escherichia coli Chui L. 2007. Sequence-based typing of genetic
H7 fliC Genes: implication for a DNA-based typ- targets encoded outside of the O-antigen gene
ing scheme for E. coli O157:H7. J Clin Microbiol cluster is indicative of Shiga toxin-producing
38:17861790. Escherichia coli serogroup lineages. J Med
201. Desmarchelier PM, Bilge SS, Fegan N, Mills L, Microbiol 56:620628.
Vary JC Jr, Tarr PI. 1998. A PCR specific for 212. Norman KN, Strockbine NA, Bono JL. 2012.
Escherichia coli O157 based on the rfb locus Association of nucleotide polymorphisms within
encoding O157 lipopolysaccharide. J Clin the O-antigen gene cluster of Escherichia coli O26,
Microbiol 36:18011804. O45, O103, O111, O121, and O145 with serogroups
202. Fratamico PM, Debroy C, Miyamoto T, Liu Y. and genetic subtypes. Appl Environ Microbiol 78:
2009. PCR detection of enterohemorrhagic 66896703.
Escherichia coli O145 in food by targeting genes in 213. Delannoy S, Beutin L, Fach P. 2012. Use of
the E. coli O145 O-antigen gene cluster and the clustered regularly interspaced short palindromic
shiga toxin 1 and shiga toxin 2 genes. Foodborne repeat sequence polymorphisms for specific de-
Pathog Dis 6:605611. tection of enterohemorrhagic Escherichia coli
203. OHanlon KA, Catarame TM, Duffy G, Blair IS, strains of serotypes O26:H11, O45:H2, O103:H2,
and McDowell DA. 2004. RAPID detection and O111:H8, O121:H19, O145:H28, and O157:H7 by
quantification of E. coli O157/O26/O111 in minced real-time PCR. J Clin Microbiol 50:40354040.
beef by real-time PCR. J Appl Microbiol 96:1013 214. Gerner-Smidt P, Hise K, Kincaid J, Hunter
1023. S, Rolando S, Hyytia-Trees E, Ribot EM,
204. Paton AW, Paton JC. 1999. Molecular charac- Swaminathan B. 2006. PulseNet USA: a five-year
terization of the locus encoding biosynthesis of update. Foodborne Pathog Dis 3:919.
the lipopolysaccharide O antigen of Escherichia 215. Gerner-Smidt P, Scheutz F. 2006. Standardized
coli serotype O113. Infect Immun 67:59305937. pulsed-field gel electrophoresis of Shiga toxin-
205. Perelle S, Dilasser F, Grout J, Fach P. 2005. producing Escherichia coli: the PulseNet Europe
Detection of Escherichia coli serogroup O103 by feasibility study. Foodborne Pathog Dis 3:7480.
real-time polymerase chain reaction. J Appl 216. Mora A, Lopez C, Dhabi G, Lopez-Beceiro AM,
Microbiol 98:11621168. Fidalgo LE, Diaz EA, Martinez-Carrasco C,
206. Valadez AM, Debroy C, Dudley E, Cutter CN. Mamani R, Herrera A, Blanco JE, Blanco M,
2011. Multiplex PCR detection of Shiga toxin- Blanco J. 2012. Seropathotypes, phylogroups,
producing Escherichia coli strains belonging to Stx subtypes, and intimin types of wildlife-
serogroups O157, O103, O91, O113, O145, O111, and carried, Shiga toxin-producing Escherichia coli
O26 experimentally inoculated in beef carcass strains with the same characteristics as human-
swabs, beef trim, and ground beef. J Food Prot pathogenic isolates. Appl Environ Microbiol 78:
74:228239. 25782585.
207. Ballmer K, Korczak BM, Kuhnert P, Slickers 217. Beutin L, Geier D, Zimmermann S, Aleksic S,
P, Ehricht R, Hachler H. 2007. Fast DNA Gillespie HA, Whittam TS. 1997. Epidemiological
serotyping of Escherichia coli by use of an oligo- relatedness and clonal types of natural popula-
nucleotide microarray. J Clin Microbiol 45:370 tions of Escherichia coli strains producing Shiga
379. toxins in separate populations of cattle and sheep.
208. Anjum MF, Mafura M, Slickers P, Ballmer K, Appl Environ Microbiol 63:21752180.
Kuhnert P, Woodward MJ, Ehricht R. 2007. 218. Thomas KM, McCann MS, Collery MM, Logan
Pathotyping Escherichia coli by using miniatur- A, Whyte P, McDowell DA, Duffy G. 2012.
ized DNA microarrays. Appl Environ Microbiol Tracking verocytotoxigenic Escherichia coli O157,
73:56925697. O26, O111, O103 and O145 in Irish cattle. Int
209. Lin A, Nguyen L, Lee T, Clotilde LM, Kase JA, J Food Microbiol 153:288296.
Son I, Carter JM, Lauzon CR. 2011. Rapid O 219. Welinder-Olsson C, Badenfors M, Cheasty T,
serogroup identification of the ten most clini- Kjellin E, Kaijser B. 2002. Genetic profiling
cally relevant STECs by Luminex microbead- of enterohemorrhagic Escherichia coli strains in

relation to clonality and clinical signs of infection. richia coli O111 infections associated with a cor-
J Clin Microbiol 40:959964. rectional facility diaryColorado 2010. Morb
220. Vaz TMI, Irino K, Nishimura LS, Cergole- Mortal Wkly Rep 61:149152.
Novella MC, Guth BEC. 2006. Genetic hetero- 230. Serra-Moreno R, Jofre J, Muniesa M. 2007.
geneity of Shiga toxin-producing Escherichia coli Insertion site occupancy by stx2 bacteriophages
strains isolated in Sao Paulo, Brazil, from 1976 depends on the locus availability of the host strain
through 2003, as revealed by pulsed-field gel chromosome. J Bacteriol 189:66456654.
electrophoresis. J Clin Microbiol 44:798804. 231. Shaikh N, Tarr PI. 2003. Escherichia coli O157:H7
221. Mainil JG, Bardiau M, Ooka T, Ogura Y, Shiga toxin-encoding bacteriophages: integra-
Murase K, Etoh Y, Ichihara S, Horikawa K, tions, excisions, truncations, and evolutionary
Buvens G, Pierard D, Itoh T, Hayashi T. 2011. implications. J Bacteriol 185:35963605.
Typing of O26 enterohaemorrhagic and entero- 232. Besser TE, Shaikh N, Holt NJ, Tarr PI, Konkel
pathogenic Escherichia coli isolated from humans ME, Malik-Kale P, Walsh CW, Whittam TS,
and cattle with IS621 multiplex PCR-based fin- Bono JL. 2007. Greater diversity of Shiga toxin-
gerprinting. J Appl Microbiol 111:773786. encoding bacteriophage insertion sites among
222. Leomil L, Pestana de Castro AF, Krause G, Escherichia coli O157:H7 isolates from cattle than
Schmidt H, Beutin L. 2005. Characterization of in those from humans. Appl Environ Microbiol
two major groups of diarrheagenic Escherichia 73:671679.
coli O26 strains which are globally spread in 233. Vanaja SK, Springman AC, Besser TE, Whittam
human patients and domestic animals of different TS, Manning SD. 2010. Differential expression of
species. FEMS Microbiol Lett 249:335342. virulence and stress fitness genes between Esch-
223. Sugimoto N, Shima K, Hinenoya A, Asakura M, erichia coli O157:H7 strains with clinical or bo-
Matsuhisa A, Watanabe H, Yamasaki S. 2011. vine-biased genotypes. Appl Environ Microbiol
Evaluation of a PCR-restriction fragment length 76:6068.
polymorphism (PCR-RFLP) assay for molecular 234. Shringi S, Schmidt C, Katherine K, Brayton KA,
epidemiological study of Shiga toxin-producing Hancock DD, Besser TE. 2012. Carriage of stx2a
Escherichia coli. J Vet Med Sci 73:859867. differentiates clinical and bovine-biased strains of
224. Beutin L, Kaulfuss S, Herold S, Oswald E, Escherichia coli O157. PLoS One 7:e51572.
Schmidt H. 2005. Genetic analysis of entero- 235. Creuzburg K, Schmidt H. 2007. Molecular
pathogenic and enterohemorrhagic Escherichia coli characterization and distribution of genes encoding
serogroup O103 strains by molecular typing of members of the type III effector nleA family
virulence and housekeeping genes and pulsed-field among pathogenic Escherichia coli strains. J Clin
gel electrophoresis. J Clin Microbiol 43:15521563. Microbiol 45:24982507.
225. Sonntag AK, Prager R, Bielaszewska M, Zhang 236. Oswald E, Schmidt H, Morabito S, Karch H,
W, Fruth A, Tschape H, Karch H. 2004. Phe- Marches O, Caprioli A. 2000. Typing of intimin
notypic and genotypic analyses of enterohemor- genes in human and animal enterohemorrhagic
rhagic Escherichia coli O145 strains from patients and enteropathogenic Escherichia coli: character-
in Germany. J Clin Microbiol 42:954962. ization of a new intimin variant. Infect Immun
226. Louie M, Read S, Louie L, Ziebell K, Rahn K, 68:6471.
Borczyk A, Lior H. 1999. Molecular typing 237. Jores J, Rumer L, Wieler LH. 2004. Impact of
methods to investigate transmission of Esche- the locus of enterocyte effacement pathogenicity
richia coli O157:H7 from cattle to humans. island on the evolution of pathogenic Escherichia
Epidemiol Infect 123:1724. coli. Int J Med Microbiol 294:103113.
227. Beutin L, Bulte M, Weber A, Zimmermann S, 238. Mora A, Blanco M, Yamamoto D, Dahbi G,
Gleier K. 2000. Investigation of human infec- Blanco JE, Lopez C, Alonso MP, Vieira MA,
tions with verocytotoxin-producing strains of Hernandes RT, Abe CM, Piazza RM, Lacher
Escherichia coli (VTEC) belonging to serogroup DW, Elias WP, Gomes TA, Blanco J. 2009.
O118 with evidence for zoonotic transmission. HeLa-cell adherence patterns and actin aggrega-
Epidemiol Infect 125:4754. tion of enteropathogenic Escherichia coli (EPEC)
228. Thomas KM, McCann MS, Collery MM, and Shiga-toxin-producing E. coli (STEC) strains
Moschonas G, Whyte P, McDowell DA, Duffy carrying different eae and tir alleles. Int Microbiol
G. 2013. Transfer of verocytotoxigenic Escherichia 12:243251.
coli O157, O26, O111, O103 and O145 from fleece to 239. Creuzburg K, Heeren S, Lis CM, Kranz M,
carcass during sheep slaughter in an Irish export Hensel M, Schmidt H. 2011. Genetic background
abattoir. Food Microbiol 34:3845. and mobility of variants of the gene nleA in
229. Centers for Disease Control and Prevention. attaching and effacing Escherichia coli. Appl En-
2012. Outbreak of Shiga toxin-producing Esche- viron Microbiol 77:87058713.

294 BEUTIN AND FACH
240. Persson S, Olsen KE, Ethelberg S, Scheutz F. lymphatic tissue of free-ranging deer. Epidemiol
2007. Subtyping method for Escherichia coli shiga Infect 28:19.
toxin (verocytotoxin) 2 variants and correlations 251. Kumar A, Taneja N, Kumar Y, Sharma M. 2012.
to clinical manifestations. J Clin Microbiol 45: Detection of Shiga toxin variants among Shiga
20202024. toxin-forming Escherichia coli isolates from ani-
241. Friedrich AW, Bielaszewska M, Zhang WL, mal stool, meat and human stool samples in India.
Pulz M, Kuczius T, Ammon A, Karch H. 2002. J Appl Microbiol 113:12081216.
Escherichia coli harboring Shiga toxin 2 gene 252. Masana MO, DAstek BA, Palladino PM,
variants: frequency and association with clinical Galli L, del Castillo LL, Carbonari C, Leotta
symptoms. J Infect Dis 185:7484. GA, Vilacoba E, Irino K, Rivas M. 2011. Geno-
242. Friedrich AW, Borell J, Bielaszewska M, Fruth typic characterization of non-O157 Shiga toxin-
A, Tschape H, Karch H. 2003. Shiga toxin 1c- producing Escherichia coli in beef abattoirs of
producing Escherichia coli strains: phenotypic Argentina. J Food Prot 74:20082017.
and genetic characterization and association 253. DAstek BA, del Castillo LL, Miliwebsky E,
with human disease. J Clin Microbiol 41:2448 Carbonari C, Palladino PM, Deza N, Chinen I,
2453. Manfredi E, Leotta GA, Masana MO, Rivas M.
243. Fratamico PM, Bagi LK, Bush EJ, Solow BT. 2012. Subtyping of Escherichia coli O157:H7 strains
2004. Prevalence and characterization of shiga isolated from human infections and healthy
toxin-producing Escherichia coli in swine feces cattle in Argentina. Foodborne Pathog Dis 9:457
recovered in the National Animal Health Moni- 464.
toring Systems Swine 2000 study. Appl Environ 254. Slanec T, Fruth A, Creuzburg K, Schmidt H.
Microbiol 70:71737178. 2009. Molecular analysis of virulence profiles
244. Muthing J, Meisen I, Zhang W, Bielaszewska and Shiga toxin genes in food-borne Shiga toxin-
M, Mormann M, Bauerfeind R, Schmidt MA, producing Escherichia coli. Appl Environ Micro-
Friedrich AW, Karch H. 2012. Promiscuous biol 75:61876197.
Shiga toxin 2e and its intimate relationship to 255. Vernozy-Rozand C, Montet MP, Bertin Y,
Forssman. Glycobiology 22:849862. Trably F, Girardeau JP, Martin C, Livrelli V,
245. Morabito S, DellOmo G, Agrimi U, Schmidt H, Beutin L. 2004. Serotyping, stx2 subtyping, and
Karch H, Cheasty T, Caprioli A. 2001. Detection characterization of the locus of enterocyte ef-
and characterization of Shiga toxin-producing facement island of Shiga toxin-producing Esche-
Escherichia coli in feral pigeons. Vet Microbiol richia coli and E. coli O157:H7 strains isolated
82:275283. from the environment in France. Appl Environ
246. Ooka T, Seto K, Kawano K, Kobayashi H, Etoh Microbiol 70:25562559.
Y, Ichihara S, Kaneko A, Isobe J, Yamaguchi K, 256. Polifroni R, Etcheverria AI, Sanz ME, Cepeda
Horikawa K, Gomes TA, Linden A, Bardiau M, RE, Kruger A, Lucchesi PM, Fernandez D,
Mainil JG, Beutin L, Ogura Y, Hayashi T. 2012. Parma AE, Padola NL. 2012. Molecular charac-
Clinical significance of Escherichia albertii. Emerg terization of Shiga toxin-producing Escherichia
Infect Dis 18:488492. coli isolated from the environment of a dairy farm.
247. Prager R, Fruth A, Siewert U, Strutz U, Curr Microbiol 65:337343.
Tschape H. 2009. Escherichia coli encoding Shiga 257. Wasilenko JL, Fratamico PM, Narang N,
toxin 2f as an emerging human pathogen. Int J Tillman GE, Ladely S, Simmons M, Cray WC Jr.
Med Microbiol 299:343353. 2012. Influence of primer sequences and DNA
248. Hofer E, Cernela N, Stephan R. 2012. Shiga toxin extraction method on detection of non-O157
subtypes associated with Shiga toxin-producing Shiga toxin-producing Escherichia coli in ground
Escherichia coli strains isolated from red deer, roe beef by real-time PCR targeting the eae, stx, and
deer, chamois, and ibex. Foodborne Pathog Dis serogroup-specific genes. J Food Prot 75:1939
9:792795. 1950.
249. Brett KN, Ramachandran V, Hornitzky MA, 258. Deng W, Puente JL, Gruenheid S, Li Y, Vallance
Bettelheim KA, Walker MJ, Djordjevic SP. BA, Vazquez A, Barba J, Ibarra JA, ODonnell
2003. stx1c Is the most common Shiga toxin 1 P, Metalnikov P, Ashman K, Lee S, Goode D,
subtype among Shiga toxin-producing Escherichia Pawson T, Finlay BB. 2004. Dissecting virulence:
coli isolates from sheep but not among isolates systematic and functional analyses of a pathoge-
from cattle. J Clin Microbiol 41:926936. nicity island. Proc Natl Acad Sci USA 101:3597
250. Eggert M, Stuber E, Heurich M, Fredriksson- 3602.
Ahomaa M, Burgos Y, Beutin L, Martlbauer E. 259. Tarr CL, Whittam TS. 2002. Molecular evolution
2012. Detection and characterization of Shiga of the intimin gene in O111 clones of pathogenic
toxin-producing Escherichia coli in faeces and Escherichia coli. J Bacteriol 184:479487.

260. Zhang WL, Kohler B, Oswald E, Beutin L, Genome sequence of enterohaemorrhagic Esche-
Karch H, Morabito S, Caprioli A, Suerbaum S, richia coli O157:H7. Nature 409:529533.
Schmidt H. 2002. Genetic diversity of intimin 268. Ogura Y, Ooka T, Iguchi A, Toh H, Asadulghani
genes of attaching and effacing Escherichia coli M, Oshima K, Kodama T, Abe H, Nakayama K,
strains. J Clin Microbiol 40:44864492. Kurokawa K, Tobe T, Hattori M, Hayashi T.
261. Jores J, Zehmke K, Eichberg J, Rumer L, 2009. Comparative genomics reveal the mecha-
Wieler LH. 2003. Description of a novel intimin nism of the parallel evolution of O157 and non-
variant (type zeta) in the bovine O84:NM O157 enterohemorrhagic Escherichia coli. Proc
verotoxin-producing Escherichia coli strain 537/ Natl Acad. Sci USA 106:1793917944.
89 and the diagnostic value of intimin typing. Exp 269. Ogura Y, Ooka T, Asadulghani, Terajima J,
Biol Med (Maywood) 228:370376. Nougayrede JP, Kurokawa K, Tashiro K,
262. Creuzburg K, Middendorf B, Mellmann A, Tobe T, Nakayama K, Kuhara S, Oswald E,
Martaler T, Holz C, Fruth A, Karch H, Schmidt Watanabe H, Hayashi T. 2007. Extensive
H. 2011. Evolutionary analysis and distribution of genomic diversity and selective conservation
type III effector genes in pathogenic Escherichia of virulence-determinants in enterohemorrhagic
coli from human, animal and food sources. Envi- Escherichia coli strains of O157 and non-O157
ron Microbiol 13:439452. serotypes. Genome Biol 8:R138.
263. Levine MM. 1987. Escherichia coli that cause di- 270. Wickham ME, Lupp C, Mascarenhas M,
arrhea: enterotoxigenic, enteropathogenic, entero- Vazquez A, Coombes BK, Brown NF, Coburn
invasive, enterohemorrhagic, and enteroadherent. BA, Deng W, Puente JL, Karmali MA, Finlay
J Infect Dis 155:377389. BB. 2006. Bacterial genetic determinants of
264. Boerlin P, McEwen SA, Boerlin-Petzold F, non-O157 STEC outbreaks and hemolytic-uremic
Wilson JB, Johnson RP, Gyles CL. 1999. Asso- syndrome after infection. J Infect Dis 194:819
ciations between virulence factors of Shiga toxin- 827.
producing Escherichia coli and disease in humans. 271. Bugarel M, Martin A, Fach P, Beutin L. 2011.
J Clin Microbiol 37:497503. Virulence gene profiling of enterohemorrhagic
265. Ethelberg S, Olsen KE, Scheutz F, Jensen C, (EHEC) and enteropathogenic (EPEC) Esche-
Schiellerup P, Enberg J, Petersen AM, Olesen richia coli strains: a basis for molecular risk as-
B, Gerner-Smidt P, Molbak K. 2004. Virulence sessment of typical and atypical EPEC strains.
factors for hemolytic uremic syndrome, Denmark. BMC Microbiol 11:142.
Emerg Infect Dis 10:842847. 272. Schimmer B, Nygard K, Eriksen HM, Lassen J,
266. Farfan MJ, Torres AG. 2012. Molecular mecha- Lindstedt BA, Brandal LT, Kapperud G,
nisms that mediate colonization of Shiga toxin- Aavitsland P. 2008. Outbreak of haemolytic
producing Escherichia coli strains. Infect Immun uraemic syndrome in Norway caused by stx2-
80:903913. positive Escherichia coli O103:H25 traced to cured
267. Perna NT, Plunkett G III, Burland V, Mau B, mutton sausages. BMC Infect Dis 8:41.
Glasner JD, Rose DJ, Mayhew GF, Evans PS, 273. Imamovic L, Tozzoli R, Michelacci V, Minelli F,
Gregor J, Kirkpatrick HA, Posfai G, Hackett J, Marziano ML, Caprioli A, Morabito S. 2010.
Klink S, Boutin A, Shao Y, Miller L, Grotbeck EJ, OI-57, a genomic island of Escherichia coli O157,
Davis NW, Lim A, Dimalanta ET, Potamousis is present in other seropathotypes of Shiga toxin-
KD, Apodaca J, Anantharaman TS, Lin J, Yen G, producing E. coli associated with severe human
Schwartz DC, Welch RA, Blattner FR. 2001. disease. Infect Immun 78:46974704.

CLINICAL, PATHOLOGICAL,
AND PATHOPHYSIOLOGICAL ASPECTS

Shiga Toxin/Verocytotoxin-Producing
Escherichia coli Infections: Practical
Clinical Perspectives
T. KEEFE DAVIS,1 NICOLE C. A. J. VAN DE KAR,2 and PHILLIP I. TARR3

15
INTRODUCTION
Shiga toxin (Stx)-producing Escherichia coli (STEC) cause illness with a spec-
trum of severity ranging from mild (even asymptomatic) carriage to life-
threatening disease (13). STEC infections are relatively uncommon; in the
United States, extrapolation of data from FoodNet (4) to a nationwide pop-
ulation that exceeds 300,000,000 indicates there are fewer than 4,000 diag-
nosed cases of E. coli O157:H7 infection per annum. E. coli O157:H7 remains
the near-exclusive cause of hemolytic-uremic syndrome (HUS) throughout
most of the world, and the single serotype on which most data have been
generated. Therefore, we emphasize this particular pathogen in this article.
The European Food Safety Authority and the European Centre for Disease
Prevention and Control report similar epidemiology: 4,000 confirmed infec-
tions caused by Stx-producing E. coli strains (mostly belonging to the O157
serogroup) in 27 European Union member states. The number of reported in-
fections attributed to E. coli strains that produce Shiga toxins has increased
since 2008 (5).
1
Division of Nephrology, Department of Pediatrics, Washington University School of Medicine, St. Louis, MO 63110;
2
Division of Nephrology, Department of Pediatrics, Radboud University Medical Centre, Nijmegen, The Netherlands;
3
Division of Gastroenterology, Hepatology, and Nutrition, Department of Pediatrics, and Department of Molecular
Microbiology, Washington University School of Medicine, St. Louis, MO 63110.
299

300 DAVIS ET AL.
Despite their low overall incidence, human Enterohemorrhagic E. coli (EHEC) is an-
infections are medically and epidemiologically other term for STEC/VTEC strains that cause
actionable. The rarity with which Shiga toxin- human disease. This term is also problematic,
producing E. coli infections occur, barriers because it implies that the diarrhea stools
to timely microbial diagnosis, consequences contain visible blood, which is sometimes not
of missed diagnoses, and the many difficulties the case in E. coli O157:H7 infections, and
in attempts to generate high-quality evidence frequently not the case in infections caused
on which to justify treatments pose challenges by STEC/VTEC strains belonging to other
for clinicians and public health systems. In serotypes. Furthermore, a small subset of pa-
this review, we focus on clinical aspects (i) tients with HUS will have no diarrhea, but
early in illness; (ii) in the intermediate stage their stool will nevertheless contain E. coli
of illness as HUS evolves in the approximately O157:H7 (8, 9).
15 to 20% of infected children in whom We also have preferences for clinical de-
this complication occurs; and (iii) during the scriptors related to STEC/VTEC infection. In
HUS phase and its aftermath. When data lieu of the time-honored term hemorrhagic
exist, we cite the appropriate literature, but colitis (coined in the first outbreak report)
in other circumstances, we rely on our cu- (10), we prefer the more encompassing con-
mulative experience, as noted. cept of bloody diarrhea. For HUS, we urge a
urinalysis-independent, stringent case defini-
tion, consisting of the simultaneous presence
DEFINITIONS of nonimmune hemolytic anemia (hematocrit/
packed cell volume <30% with smear evidence
This field of study has been vexed by multiple of hemolysis and a negative Coombs test),
nomenclature issues. In this review and in thrombocytopenia (platelet count <150,000
our own papers and practice, we describe the mm3), and azotemia (creatinine > upper limit
toxins produced by enteric pathogens that of normal for age) (11). There is much hazard
cause HUS as Shiga toxins (Stxs). This term is and little benefit to be gained from using
synonymous with verocytotoxins (VT), named less stringent clinical definitions of this com-
for the toxic effect of these proteins on Vero plication of STEC/VTEC infections. For ex-
cells, as originally described by Konawalchuk ample, reliance on an abnormal urinalysis to
et al. (6) and identified as the key phenotype define HUS, especially if the serum creatinine
of these pathogens by Karmali et al. (7). E. coli is normal, risks consideration of incorrect di-
strains that produce Stx are termed Stx- agnoses such as urinary tract infections, es-
producing E. coli (STEC), which is synony- pecially as the possibility of contamination
mous with VT-producing E. coli (VTEC). with fecal material is high in the setting of
However, in this text we use the term STEC/ diarrhea. Moreover, these widely available
VTEC when describing bacteria that produce blood tests enable physicians to relate their
Stx/VT. This combined term reflects our cur- patients course to those described in many
rent misgivings about the term Stx to refer to other studies during the past 3 decades from
the cardinal virulence trait of these patho- multiple countries (1231).
gens. These misgivings are rooted in a prac-
tical matter: many physicians, on learning that
their patient is infected with a Shiga toxin- HISTORY
producing organism, assume erroneously that
the laboratory is describing an infection with HUS was first described in the mid-1950s by
Shigella sp. (4). Such misconceptions are, in Gasser et al. (32). In that series of 10 fatal
our experience, preludes to the inappropriate illnesses, cases 3 and 4 had Brechdurchfall,
administration of antibiotics. which is vomiting plus diarrhea (case 3 had

CHAPTER 15 Practical Clinical Perspectives 301
these signs throughout the entire illness, to the laboratory and inoculated on sorbitol-
while the diarrhea of case 4 occurred only MacConkey agar on receipt), the micro-
preterminally). None of the other clinical biologist can often inform the clinician of a
courses suggested enteric illnesses, and nota- presumptive positive or negative for this se-
bly none of the reports used the terms bloody rotype within 18 to 24 h. That simple piece of
diarrhea or dysentery. However, we have information provides valuable clarity to the
found earlier papers describing cases in which management of patients with acute diarrhea.
diarrhea or dysentery preceded renal failure However, because a proportion of STEC/
and death within a time frame that closely VTEC infections are caused by non-O157:H7
resembles that of E. coli-related HUS (3337). STEC/VTEC, there is considerable merit to
E. coli O157:H7 and the closely related also determining their presence, but it is much
pathogen E. coli O157:H were estimated to more imperative to exert the greatest effort
have split from the same progenitor about to confirm or refute the presence of E. coli
7,000 years ago (38). This common ancestor O157:H7 in a stool culture. One important
acquired a gene encoding Stx2/VT2 before exception to this statement exists: sorbitol-
that. However, it was not until the 1970s when fermenting E. coli O157:H is as virulent, or
E. coli that had been isolated from food was possibly more virulent, than E. coli O157:H7
reported to produce Stx/VT (6). This pheno- (50), and this clone remains endemic in Ger-
type preceded by several years the first de- many. These organisms, which are closely re-
scription of STEC/VTEC strains as causes of lated to E. coli O157:H7 (38), are not detected
disease in 1983. In that year, near-simultaneous by sorbitol-MacConkey agar screening.
publications introduced these pathogens to the E. coli O157:H7 is best detected in stool
medical community: Riley et al. described by using sorbitol-MacConkey agar with or
a hamburger-associated outbreak of E. coli without cefixime-tellurite (51), because un-
O157:H7 infections (10), and Karmali et al. like most commensal E. coli and non-O157:H7
linked fecal STEC/VTEC to HUS (7). Also STEC/VTEC strains, E. coli O157:H7 does not
in 1983, additional investigators demonstrated ferment sorbitol after overnight incubation.
the production of Stx/VT by E. coli O157:H7 Hence, the presence of a colorless colony
(3942). on sorbitol-MacConkey agar that agglutinates
with an appropriate serologic reagent enables
the microbiologist to make a confident and
DISTINCTION BETWEEN STEC/VTEC timely presumptive diagnosis. For inexplica-
BELONGING TO SEROTYPE O157:H7 ble reasons, E. coli O157:H7 is more easily
AND THOSE BELONGING TO detected by sorbitol-MacConkey agar plating
ALL OTHER SEROTYPES than by toxin testing of broth cultures of stool
(1, 16, 5256). Because of the greater sensitiv-
There are important practical reasons to dif- ity of agar plating, the critical importance of
ferentiate STEC/VTEC strains that express making a diagnosis of E. coli O157:H7 infection
the O (somatic) 157 and flagellar (H) 7 antigens as rapidly as possible, and the recognition that
from all other serotypes, which we collectively a small subset of non-O157:H7 STEC/VTEC
term non-O157:H7 STEC/VTEC. Most com- infections can be severe, we agree with the
pellingly, E. coli O157:H7 is the STEC/VTEC guidance of the Centers for Disease Control
that remains to this day the near-exclusive and Prevention that advises the simultaneous
cause of postdiarrheal HUS (8, 12, 16, 21, 43 testing for E. coli O157:H7 (on agar plates)
49). This serotype is also the one most strongly and non-O157:H7 STEC/VTEC (using, in most
associated with outbreaks (though most in- cases, a toxin enzyme immunoassay [EIA])
fections are sporadic). If stool specimens are (57). We strongly disagree with detection
handled adroitly (i.e., immediately transported algorithms that assume that the EIA can be

302 DAVIS ET AL.
used as a screen with sorbitol-MacConkey wide study of 83 patients with HUS (12),
agar plating only for positives. Such protocols 70 patients had stool cultures in which bac-
underdetect E. coli O157:H7 and delay an- teria grew. These specimens were obtained a
swering an important question: is the patient median of 8 days after illness onset. Of the
infected with E. coli O157:H7 or not? 30 STEC/VTEC strains identified in these 30
Though non-O157:H7 STEC/VTEC can specimens, 25 were E. coli O157. Of the five
cause HUS, the likelihood that any non- patients whose stools yielded non-O157:H7
O157:H7 STEC/VTEC infection will result in STEC/VTEC, four had serologic testing in
serious kidney injury is extremely low. As convalescence and three of them had anti-
noted above, E. coli O157:H7 is the over- body evidence of recent exposure to the O157
whelming cause of postdiarrheal HUS (8, 12, lipopolysaccharide antigen. Hence, an EIA
16, 21, 4349). If the stool of a patient with (or, increasingly, PCR) (58, 59) to detect an
HUS does not contain E. coli O157:H7, the STEC/VTEC strain, when added to sorbitol
most likely explanation is that the specimen MacConkey agar screening, is likely to lead
was first cultured for this organism at a point to the diagnosis of cases that are at much
in illness when the pathogen had been elimi- lower risk of HUS developing. Nonetheless,
nated (44). making a diagnosis in at least a subset of
Table 1 summarizes the E. coli O157:H7/ these cases is worthwhile, if for no other
non-O157:H7 STEC/VTEC acuity pyramid de- reason than to provide etiologic clarity to
rived from several defined populations (the an illness that is usually of greater severity
HUS studies are globally distributed, but the than most gastroenteritides. However, data
non-HUS cohorts are from the United States). do not exist to calculate the overall value or
As illness severity increases, inferred from cost of such policies, but, again, sorbitol-
the setting of acquisition of culture (large MacConkey agar screening is most crucial to
geographic region to emergency facility to include when performing stool cultures for
cohorts with HUS), the ratio of non-O157:H7 bacterial pathogens.
STEC/VTEC to E. coli O157:H7 diminishes.
Findings from several of these studies are
particularly worth mentioning. Of 229 Con- CLINICAL MANAGEMENT OF STEC/VTEC
necticut patients with infections caused by INFECTIONS: PRE-HUS PHASE
non-O157:H7 STEC/VTEC who were studied
over a decade-long interval (48), HUS devel- Our approach to STEC/VTEC infections is
oped in only one, while HUS developed in 45 based on three overarching principles:
of the 434 patients in this series who were
1. Early detection is critical: time is not on
infected with E. coli O157. In a United States-
your side when treating STEC/VTEC
infections.
TABLE 1 Acuity pyramida 2. Early and vigorous volume expansion is
Setting and Non-O157:H7 associated with avoidance of the most
study years E. coli O157:H7 STEC/VTEC
severe renal injury.
Wide geographic Montana: 38% Montana: 62% 3. Highly strategic test selection avoids
areas (19982009) Connecticut: Connecticut: generating misleading and potentially
(48, 141) 42% 58%
Pediatric emergency 71% 29%
harmful results.
facilities (19912005) Early identification (and hospitalization) of
(1, 16, 142)
HUS (19842010) 9599% 15%
infected patients is critical because it lowers
(8, 12, 16, 21, 4349) the risk of secondary cases (60), avoids diag-
a
Frequency of recovery of E. coli O157:H7 and non-O157:H7 nostic misadventures (for example, we have
STEC/VTEC, according to the setting of acquisition of specimen. seen patients started on steroids because of

presumed fulminant ulcerative colitis), and Antibiotics were first suggested as poten-
facilitates the commencement of intravenous tially increasing the risk of HUS developing
volume expansion. Several lines of evidence in the initial report linking STEC/VTEC in-
suggest that renal perfusion is threatened fection to this disorder (7). No credible evi-
prior to and during HUS and that diminished dence has emerged since then that supports
kidney blood flow increases the likelihood of the concept that antibiotics administered to
severe renal injury. First, there is evidence of children or adults early in illness reduce the
prothrombotic abnormalities in the infected likelihood of HUS subsequently developing.
host before HUS (and even if HUS does not In fact, extensive data from multiple studies,
ensue). Factor 1.2 (the prothrombin activation including more than 1,000 patients infected in
peptide), D-dimers, plasminogen activator in- epidemics and sporadically, demonstrate that
hibitor, and platelet-activating factor are each antibiotics are at best neutral and quite likely
elevated during E. coli O157:H7 infections, increase the risk of HUS developing (Table 2).
and von Willebrands factor is also sheared (in- Indeed, the largest risk is demonstrated in
dicating flow-related rheological stress, prob- the studies with the most robust data: large
ably caused by nascent thrombi) (13, 17, 61). cohorts, interviewed prospectively, with ex-
These prothrombotic abnormalities, which tensive analysis of timing of administration of
are demonstrated at a point in illness when antibiotics, and representing infections with
the blood counts are normal, probably pro- multiple different strains.
duce some degree of renal ischemia, even We also urge against the use of narcotics
before there is smear evidence of microangio- and antidiarrheal agents in patients with in-
pathy. Second, if HUS develops, dehydration fections that could be caused by STEC/VTEC
at presentation (manifest as elevated hemo- because of their association with higher rates
globin) is associated with less favorable short- of HUS or neurological sequelae (67, 68).
term (6264) and long-term (65) outcomes. We also do not endorse nonsteroidal anti-
Third, intravenous volume expansion early in inflammatory agents, because, in our experi-
illness, starting as soon as possible after pre- ence, they have no value and because of
sentation, is associated with less severe (i.e., their nephrotoxic potential (69), which might
nonanuric) HUS, if HUS ensues (20, 21). be exacerbated in the dehydrated state (70).
The details of our fluid management pro-
tocols are provided in reference 66. We
Early in HUS Prognostic Factors
have slightly changed our recommendations
(articulated in reference 66) to now suggest HUS occurs in 15 to 20% of children who are
complete blood counts every 12 h until there culture positive for E. coli O157:H7. Several
is assurance that the hemoglobin is falling indicators apparent early in the course of HUS
with volume expansion, because we have are associated with a severe course of HUS. A
not found other indices of circulating blood combination of hypocalcemia (2 mmol/liter)
volume (BUN:creatinine ratio, skin turgor, plus oliguria (urine output <0.4 ml kg1 h1
or vital signs) to be reliable in this setting for 24 h) within 48 h of hospitalization had
(authors personal experience). We aim for a the highest predictive value for negative
decrement in the hemoglobin of 0.5 g/dl per outcomes (death, need for dialysis, hyper-
each 12-h period over the first 1 or 2 days. tension requiring therapy, or central nervous
It can be difficult to accurately assess host system sequelae at discharge) (71). Multiple
volume status, and there is a risk of overload seizures, coma, retinal hemorrhages, hyper-
with vigorous volume expansion, so we stress kalemia (>7.5 mmol/liter), acidosis (bicar-
the importance of assiduous monitoring of bonate <8 mmol/liter), or a diastolic blood
infected children in centers that are adept at pressure >90 mm Hg are also suggested to be
pediatric care (20, 66). early indicators of poor outcome in HUS (72).

304
TABLE 2 Summary of antibiotic experience in multiple case control studies of children and adultsa
Chap15.proof.3d
HUS rate in group
Year performed, Ages of HUS rate in group not receiving
304
Study setting patients Predominant antibiotics given Details and comments receiving antibiotics antibiotics
DAVIS ET AL.
Carter (143) 1985 Outbreak 1667 yr Amoxicillin, tetracycline Timing not specied. Outbreak Does not specifyb Does not specifyb
analysis, Canada characterized by two phases:
primary, contaminate food;
secondary, person-to-person
transmission. Antibiotic therapy
within the 2 days before food
exposure (primary phase) did not
have increased risk of HUS
developing. However, those on
antibiotics during the outbreak
(secondary phase) had a 10.3
relative risk of HUS developing.
Pavia et al. (144) 1988 Outbreak, 639 yr Predominantly Timing not specied. 5/8 (63%) 0/7 (0%)
case-control trimethoprim-sulfamethoxazole Comment: All antimicrobial
study, Utah agents were begun with 72 h
after onset of diarrhea.

Proulx (145) 19891990 5 4 yr Trimethoprim-sulfamethoxazole Yes 2/22 (8%) 4/25 (16%)
Randomized controlled (average) (1)
trial, Canada, antibiotics
administered late
in illness
Bell et al. (67) 1993 Outbreak, <16 yr Trimethoprim-sulfamethoxazole Yes 8/50 (16%) 28/218 (13%)
retrospective cohort, (62%), ampicillin or amoxicillin
Washington State (26%), cephalosporins (12%),
metronidazole (8%)
Wong et al. (146) 19971999 <10 yr Trimethoprim-sulfamethoxazole Yes 5/9 (56%) 5/62 (8%)
(superseded by Multistrain, prospective (2/5), -lactams (3/5)
reference 31 cohort study, four states
described below)
03/30/15 23:44
HUS rate in group HUS rate in group
Year performed, Ages of receiving antibiot- not receiving
Study setting patients Predominant antibiotics given Details and comments ics antibiotics
Dundas et al. 1996 Outbreak, 18 mo to 94 yr; Ciprooxacin Timing not specied. 8/14 (57%) treated 26/104 (25%)
(147) retrospective cohort Mean = 63 Comment: HUS developed in with antibiotics in
study, Scotland 8 (57%) of the 14 patients who the 4 wk before
received any antibiotic in the 4 wk illness onset
Chap15.proof.3d
prior to the outbreak. 7/15 (47%) treated
305
HUS developed in 7 (47%) of with antibiotics
15 cases treated with ciprooxacin within 4 days after
4 days after symptom onset illness onset
compared to 25% of the 104 cases
that did not receive antibiotic
treatment (the difference was not
statistically signicant).
Wong et al. (31) 19972006 <10 yr Trimethoprim-sulfamethoxazole Yes 9/25 (36%) 27/234 (12%)
(extended cohort Multistrain, prospective (9/25), -lactams (9/25),
analysis of cohort study, ve states metronidazole (3/25),
reference 146 azithromycin (4/25)
described above)
Smith et al. (148). 19962002 <20 yr -lactams (22% case, 4% Timing partly specied. 27/63c (43%) 38/125 (30%)c
Multistrain, age control), sulfonamides (14% case, Subjects received antibiotics in
matched, case-case 24% control), metronidazole two specic periods: within the
comparison (6% case, 2% control) rst 3 days after diarrhea onset
and in the rst 7 days after

diarrhea onset.
Cimolai et al. (68) 19841989 Multistrain, Age range Agents not specied, but were Timing not specied. 14.3%d 4.4%d
sporadic cases, not reported. characterized as appropriate Duration of antibiotic use
retrospective cohort HUS cohort: if antimicrobial was recognized termed short if 24 h or
study, British Columbia, mean = 49 mo to be effective in the treatment prolonged if >24 h.
Canada Gastroenteritis of shigellosis and if isolate was
cohort: mean susceptible in vitro testing.
= 83 mo
a
Modied from reference 93 with permission.
b
Risk ratio in lieu of HUS rate was provided, and was 8.5 (95% condence interval 2.727.5) in favor of antibiotics associated with HUS development.
c
Results report the exposure of antibiotics within the rst 7 days.
d
Results limited to appropriate antibiotics administered for short terms.
CHAPTER 15 Practical Clinical Perspectives
305
03/30/15 23:44
306 DAVIS ET AL.
An elevated polymorphonuclear leukocyte count develops allows the provision of nutrition with-
in diarrhea-associated HUS is also a risk factor out exacerbating the above complications.
for poor outcome (73, 74). Unfortunately, these In our institutions, peritoneal dialysis is the
risk factors and biomarkers were measured at most commonly used modality although in-
different points of illness, often after poor out- termittent hemodialysis and continuous renal
comes are becoming self-evident. Nonetheless, replacement therapies are equally effective.
the greatest determinant of short- (75) and long- From our perspective, there several reasons
term (62, 7687) outcome of E. coli-related acute to use peritoneal dialysis in HUS, including
kidney injury remains oligoanuria. avoiding unnecessary care in the intensive
care unit (decreasing cost) and allowing direct
access to peritoneal fluid, which is helpful if
HUS with and without Oligoanuria
the possibility of bowel perforation is raised.
There are two categories of HUS: oligoanuric If necessary, home renal replacement can be
and nonoligoanuric. We emphasize the im- used in the event of delayed recovery.
portance of averting oligoanuria to the extent Renal replacement therapies, i.e., perito-
possible, because of the repeated associations neal and hemodialysis, are the chief support-
between chronic renal sequelae and presence ive modalities in oligoanuric HUS. A review of
and duration of oligoanuria during HUS (62, these interventions is beyond the scope of this
7687). In reality, oligoanuric HUS is equiva- chapter. However, some complications of di-
lent to anuric HUS (though there can be a day alysis seem to be relatively frequent during
of oliguria before renal shutdown is com- HUS. First, as we recently reviewed (93), in-
plete). Anuria probably reflects acute tubular fectious complications of peritoneal dialysis
necrosis. The mechanism of acute tubular are common. Catheter malfunction, probably
necrosis in STEC/VTEC HUS is not com- related to bowel wall and mesenteric edema,
pletely established but could represent either also complicates peritoneal dialysis during
the effect of Shiga toxin on renal tubules (88 HUS. Catheter failure typically presents when
91) or ischemia secondary to thrombotic oc- a catheter infuses dialysate but does not
clusion of the renal vasculature (92). In view drain. Catheter malfunction often obligates
of the abundant evidence of prothrombotic surgical replacement or conversion to hemo-
activation before azotemia ensues, and in con- dialysis. For hemodialysis, we recommend a
sideration of examples of diminished renal dual-lumen catheter of age-appropriate size,
blood flow preceding anuric renal response in preferably in the internal jugular vein. The
many other clinical situations, we have tended authors generally use regional citrate anti-
to favor the occlusive/ischemic mechanism as coagulation of the extracorporeal circuit, but
the cause of anuria in HUS. As noted above, systemic heparin anticoagulation can also be
oligoanuric HUS has categorically worse short- used. Invasive procedures, such as peritoneal
and long-term implications for patients. dialysis catheter and central vascular line
HUS that requires dialysis occurs in up to placements, can be performed safely without
71% of patients, according to a summary of excessive bleeding during the thrombocyto-
HUS series over the past 4 years (Table 3). penia of HUS; platelet transfusions are rarely
The median length of stay after the case defi- necessary (94, 95).
nition of HUS is attained is 12 days for patients
with oligoanuria versus 6 days for patients
Fluid and Electrolyte Abnormalities
with nonoligoanuric renal failure (20). Dialysis
during HUS
should be instituted soon after anuria onset
to prevent cardiopulmonary overload, avoid There are numerous electrolyte disturbances
electrolyte disturbances, and treat hyperten- associated with acute HUS. Hyponatremia,
sion. Early initiation of dialysis if anuria hyperkalemia or hypokalemia, hypocalcemia,

TABLE 3 Severity of HUS in series identied in PubMed published between 2009 and 2013, using search terms
hemolytic-uremic syndrome AND children, on September 4, 2013a
Year of cases Age Dialysis
(reference) Site group rate Fatalities Comments
19972006 (31) Washington, N = 36, 31% 0 Many of these patients were well
Oregon, Idaho, <10 yr hydrated (i.e., a subset were
Wyoming, and among those in a single center
Missouri series [20] at the onset of HUS,
which could account for the low
dialysis rate).
20072008 (21) California, N = 50 68% 0
Washington, Missouri, <18 yr
Ohio, Wisconsin,
Arkansas, Indiana,
Glasgow, New Mexico,
Tennessee
19942010 (149) Alberta, Canada N = 124 43% 2% This case series employed a case
<18 yr denition of HUS that did not
obligate azotemia. Hence, the low
dialysis rate might reect patients
who would not have been
considered to have had HUS in
other series, and demonstrates
another reason to avoid urinalysis-
dependent denitions of HUS.
19982008 (150) Buenos Aires, N = 365 43% (Not reported) 94% of patients underwent
Argentina Ages not peritoneal dialysis;
reported 24% peritonitis rate
19952011 (62) Buenos Aires, N = 137 52% (Not reported) Better hydration during the
Argentina Ages not prodromal phase was associated
reported with lower frequencies of
oligoanuria and need for dialysis.
2011 (30) Hamburg, Germany N = 90 71% 1.1% Outbreak of HUS caused by E. coli
<18 yr O104:H4. Outcomes and severity
resembled those of E. coli
O157:H7 HUS.
a
Only articles with diarrhea associated HUS and dialysis rates are included.
Hypertension
and hypoalbuminemia are common. Notably,
however, these abnormalities usually by them- Hypertension during and after acute HUS is
selves do not obligate dialysis if urine is still common, with up to 70% of patients affected
being produced (96). Fatalities in the absence (99). The mechanism of HUS-induced hy-
of anuria are exceptionally rare, and recent pertension is multifactorial and likely related
retrospective data suggesting the value of to volume overload and to endothelial and
early renal replacement therapy if children vascular injury. Renovascular hypertension
are >10% overloaded probably do not apply has been suggested to play a role, but plasma
to the still urinating child with HUS (97). renin activity during HUS is below age-
Hyperuricemia is common, as is elevated lac- appropriate norms (100) (but renal vein renin
tate, and could be related to diminished renal concentrations have not been determined in
flow, impaired clearance, and increased pro- this situation). We have had excellent success
duction (98). using calcium channel blockers in acute HUS.

308 DAVIS ET AL.
If hypertension is present at discharge, we served by platelet transfusions. It is also

use angiotensin-converting enzyme inhibition concerning that most HUS-related strokes are
even if the creatinine has not yet normalized, thrombotic and not hemorrhagic (102, 103).
provided the serum creatinine concentration Fibrinogen turnover is not increased in HUS
is falling and the patient is not on dialysis. as it is in classic consumptive coagulopathies
(101). We therefore recommend against re-
questing disseminated intravascular coagula-
Hematologic Complications during HUS
tion laboratory tests as they are not likely to
Almost all patients with HUS require eryth- provide relevant information.
rocyte transfusions because of hemolysis.
The basis for the hemolysis is presumably
Neurologic Complications during HUS
physical shearing as red cells course through
small vessels in which fibrin thrombi are HUS can be associated with a variety of bona
abundant. Transfusion requirements appear fide neurologic lesions, and signs and symp-
independent of the severity of the renal injury toms of central nervous system dysfunction
(authors personal observations). We use car- have been reported in 20 to 50% of cases.
diopulmonary compromise or tachycardia as HUS has an apparent predilection for causing
an indication for transfusion, rather than an basal ganglia lesions (104), but every struc-
arbitrary hemoglobin concentration, though ture of the brain can be affected (105, 106).
in reality it is common to factor in the rate of The most common serious neurologic com-
hemolysis, time of day, vascular access, and plications of HUS are coma, convulsions, and
point in illness. The transfusion requirement strokes. These complications are usually poor
can continue several days beyond resolution prognostic signs but are rarely by them-
of thrombocytopenia and return of creatinine selves lethal. Possible mechanisms for these
to normal (authors personal observations). complications include endothelial injury with
Several additional elements should be con- microthrombotic formation and hypoxia. Neu-
sidered when pondering the need for eryth- rological dysfunction may be further exacer-
rocyte transfusion. First, erythrocyte life span bated by hyponatremia, hypertension, and
is short in HUS, ranging between 8 and 24 uremia. Although involvement of the central
days (101), and the fibrin debris presumably nervous system might portend a poor prog-
recedes on a day-by-day basis. Hence, trans- nosis, it must be appreciated that neurologic
fused red cells might last longer if transfusion recovery can be delayed for weeks or months,
can be delayed until needed. Second, we try even after exceptionally severe HUS (107
to use an entire unit at each transfusion. 110). Indeed, in comprehensive studies of
Third, we have frequently noted hypertension survivors of HUS who were infected during
immediately following transfusion, so anti- the 2011 outbreak of E. coli O104:H4 in-
hypertensive medications should be readily fections, children and adults usually made
available. After transfusion needs abate, we complete neurologic recoveries even after
usually do not provide iron to correct the re- exceptionally severe neurologic abnormalities
sidual anemia because the total body iron is during the acute phase (30, 111, 112).
not low; reticulocyte counts might be helpful Patients infected with STEC/VTEC often
in this situation. are irritable, lethargic, and jittery, and we do
We also rarely transfuse platelets because not treat these signs. It is possible that the
the underlying process leading to thrombo- around-the-clock defecations during the pre-
cytopenia is most likely entrapment of plate- HUS phase contribute to their occurrence.
lets in thrombi, and thrombocytes have short The use of sedatives to prevent patient move-
circulating half-lives in HUS (101). Also, HUS ment is not recommended, because the mental
is a thrombotic process, which is not well status of patients with HUS is often altered

and such sedation might confound clinical HUS (86). In the same paper, a meta-analysis
assessment. Acetaminophen and, if necessary, of 2,372 patients with a minimum of 1 year
fentanyl are our preferred analgesics. Mor- follow-up estimated 8%, 6%, and 1.8% of
phine should be avoided because of the neu- patients who have recovered from HUS will
rotoxic effects of its metabolites, which are have a glomerular filtration rate (per 1.73 m2)
cleared by the kidneys (113, 114). of 6080, 3059, and 529 ml min1. In this
Additional nonnephrologic complications same group, 10% had hypertension and 15%
of HUS are summarized in Table 4. had proteinuria. Meta-regression analysis in-
dicated the severity of illness and the pres-
ence of central nervous system symptoms
Chronic Renal-Related Sequelae of HUS
were associated with worse outcomes. Full
The precise risk of long-term consequences renal recovery was not achieved if dialysis
of HUS is difficult to gauge. A meta-analysis exceeded 4 weeks, but we have seen patients
of 49 papers including 3,476 patients from 18 who have made nearly complete recoveries
countries estimated incidences of death at 9% after such prolonged anuria. There are pro-
and end-stage renal disease at 3%, but most of found selection biases in computing chronic
these two outcomes occurred during acute sequelae rates if the cohorts are limited
TABLE 4 Selected nonnephrologic, nonhematologic complications of HUS

Complication Comments Reference(s)
Pancreatitis Do not pursue mild (chemical) pancreatitis by extensive investigation or 151

withholding oral intake. Hyperlipasemia and hyperamylasemia could be related
to intestinal and not biliary injury, emesis, and diminished renal clearance.
Diabetes mellitus Insulin dependence can be transient during acute HUS, or persist following 152157
renal improvement, or rarely present in the convalescent phase.
Intestinal perforation These complications are often difcult to identify. Acidosis that fails to resolve 158
and necrosis with dialysis suggests a severe intestinal complication warranting a laparotomy.
Biliary lithiasis This is usually apparent in the several weeks after HUS resolves and is manifest 152, 153
as right upper quadrant pain and rarely biliary obstruction. This is probably
caused by massive hemolysis and subsequent pigment load in the biliary
system during HUS.
Irritable bowel This can occur after STEC/VTEC infections, as with other bacterial enteric 159
syndrome infections. Its prognosis is good, often resolving within a year.
Elevated These might reect liver injury or, possibly, arise from hemolysis. There is 160
transaminases rarely actionable liver injury during HUS.
Bowel obstruction Post-HUS strictures can occur in the small or large bowel. One of the authors 161, 162
is aware of a postinfectious stricture occurring in an adult in whom HUS did not
develop, but this complication usually is manifest in the several weeks or
months after HUS resolves. These are best detected by contrast studies
(small bowel follow-through or barium enema studies).
Cardiac ischemia This complication is unusual but can complicate HUS. Troponin determination 9, 163
and/or myocarditis and cardiac ultrasound might be helpful. We have noted late in illness,
i.e., as HUS resolves, complications in elderly patients with HUS.
Retinal hemorrhages Ocular abnormalities are rarely sought in young children, so frequency might 164, 165
be higher than has been appreciated. Long-term complications included
decreased visual acuity and optic nerve atrophy.
Acute respiratory Pulmonary hemorrhage is a poor prognostic factor. 9, 64
distress syndrome and
pulmonary hemorrhage
Sudden death This complication is rare. Case reports occurring during the acute phase 163, 166,
of illness. 167

310 DAVIS ET AL.
to those patients who are still returning to in functional ADAMTS13, the enzyme that
follow-up years after HUS. It is also important catalyzes shearing of von Willebrands factor
to note that it is not clear how clinically rel- into less thrombogenic forms, inhibitors of
evant many of these sequelae (such as mi- this enzyme (i.e., an antibody), or circulating
croalbuminuria) are, and in our experience, ultra-large von Willebrands factor multimers,
chronic renal failure later in life after nor- which suggests a lesion that might be treated
malization of the serum creatinine is exceed- with plasma therapies (124126). Analyses of
ingly unusual. However, it is important to the E. coli O104:H4 outbreak provided no evi-
note that a recurring set of data suggests that dence of the value of eculizumab and plasma
presence and duration of oligoanuria are major exchange (127129). Risks of plasma exchange
predictors of chronic renal sequelae (62, 76 include complications of catheter insertion,
87), reinforcing the need to avoid this com- hypocalcemia, and exposure to blood products
plication during acute HUS, if at all possible. (130). It is important to remember that patients
with HUS often deteriorate after they are ini-
tially diagnosed, that the median number of
Use of Plasmapheresis and Eculizumab
days of anuria (among those whose urine out-
Eculizumab administration and therapeutic put ceases) is 8 (20), and that gradual sponta-
plasma exchange during the large German neous resolution of the microangiopathy and
outbreak in 2011 ignited debate about using recovery of renal and neurologic function are
these modalities in E. coli-related HUS. the rule and not the exceptions. Creative
Eculizumab prevents formation of the mem- treatments offer no benefit to standard, assid-
brane attack complex by inhibiting C5 func- uous, intensive care monitoring but do carry
tion, and its use has been prompted by a letter risks of adverse events. Such interventions
to the New England Journal of Medicine should not be conducted outside the context
(115). However, HUS in each of the patients of controlled trials, and only if sufficient data
described in that letter was already resolv- exist to suggest a state of clinical equipoise as to
ing (decreasing lactate dehydrogenase and/or their potential value. Therapeutic plasma ex-
rising platelet counts) when eculizumab was change and anticomplement therapies do not
started. Several in vitro and animal experi- meet this standard in E. coli-related HUS.
ments suggest activation of complement after
exposure to Stx/VT (116118), but these ex-
Clinical Pitfalls in HUS
periments employed STEC/VTEC concentra-
tions several orders of magnitude higher than Postdiarrheal HUS overwhelmingly occurs
the levels that have ever been documented in on a tightly choreographed trajectory (131):
humans (119). In contrast, a primate model of diarrhea (usually painful) evolves into grossly
lethal STEC/VTEC challenge demonstrated bloody diarrhea (80 to 85% of the time). All
no evidence of complement activation (120). three criteria for HUS are usually met be-
Finding evidence of alternate complement tween the 5th and 13th day of illness, with
pathway activation in children with HUS is not the first day of diarrhea assigned to be the
evidence of a pathogenic role for this branch of first day of illness. However, opportunities
the innate immune system (121), as complement for Type 1 and 2 diagnostic errors arise be-
is often activated in multisystem organ injury, cause many different microangiopathic dis-
such as trauma (122). Finally, there are poten- orders have at least some laboratory and
tially deleterious effects of inhibiting com- historic elements in common with E. coli-re-
plement during HUS, most notably sepsis (123). lated HUS, and microbial diagnosis of E. coli-
There is similarly no justification for thera- related HUS is often elusive.
peutic plasma exchange in E. coli-related HUS. Type 1 diagnostic errors (i.e., falsely as-
There is no credible evidence of deficiency suming a patient has E. coli-related HUS when

the patients microangiopathy is caused by note that microbiology testing early in illness
another process) generally occur in patients has the highest yield for STEC/VTEC and that
with aberrant prodromes to renal failure many children with HUS are culture negative
(minimal or no diarrhea, or chronic diarrhea), at the time of presentation with HUS (44).
or laboratory values that are inconsistent with We strive to increase the microbiologic yield
a diagnosis of E. coli-related HUS. This prob- by performing a rectal swab culture on ad-
lem is compounded by look-alike disorders mission of all children with HUS, attempting
that do not have highly stereotypical presen- to take possession of specimens the earliest
tations so are more difficult to recognize. in illness (usually agar plates), and seeking
Type 2 diagnostic errors (i.e., incorrectly as- STEC/VTEC in these specimens in our own
suming a patient does not have E. coli-related centers microbiology laboratories. Serology,
HUS) generally relate to lost microbiologic i.e., seeking evidence of circulating immuno-
diagnostic opportunities. To avoid Type 1 globulins to the O157 lipopolysaccharide
errors, we search for other etiologies of micro- (or to the lipopolysaccharide of several other
angiopathy if there is a persistent documented serogroups), can also be used to assign etiol-
fever in a health care setting; exceptionally ogy to cases of HUS in patients in whom a
long (>10 days) or short (<5 days) prodromal pathogen has not been recovered from the
illnesses; prominent respiratory symptoms or stool (138, 139). Antibodies to VT/Stx are less
findings; hypotension/shock; family history frequently sought, but newer enzyme immu-
(especially of distant past episodes) (132); the noassays might offer greater ease of perfor-
patient is under 6 months of age, uses specific mance (140). However, serologic testing is
medications (e.g., oral contraceptives, cyclo- not widely available. Also, absence of diar-
sporine), is pregnant; or there are discor- rhea does not exclude the possibility of an
dances between renal injury (severe) and STEC/VTEC infection (8, 9), and finding
hematologic abnormalities (minimal anemia such a pathogen can avert a much more
or thrombocytopenia). The gastrointestinal extensive evaluation and therapeutic mis-
symptoms that accompany thrombotic throm- adventures. Therefore, in any atypical pre-
bocytopenic purpura (133) or atypical HUS sentation of a microangiopathic disorder we
caused by complement regulatory proteins nevertheless exclude, to the best our ability,
(134, 135) rarely resemble those of the enteric an etiologic agent by either culture or sero-
prodrome of STEC/VTEC-related HUS, and logic investigation.
for this reason, we do not routinely seek other
disorders if the presentation is typical. Ad-
junctive tests, which should be requested SUMMARY
and interpreted with circumspection, include
chest X rays, blood and urine cultures (and E. coli O157:H7 remains the most exceptional
testing for Shiga toxin production if an E. coli pathogen among the STEC/VTEC group, in
bacterium is isolated), assays for ADAMTS13 view of its enduring association with HUS
activity, and complement regulatory protein worldwide, its leading frequency in case series
gene sequencing (135137). Tests that are not of STEC/VTEC infections (compared to any
helpful diagnostically and often misleading other serotype), and its ability to cause epi-
include serum C3, C4, and total hemolytic com- demics as well as sporadic cases. Agar plating
plement. In our experience with HUS, Type 2 of all incoming stools is the best way to detect
diagnostic errors are more common than this pathogen. Early in illness, aggressive vol-
Type 1 errors, and are often based on the ume expansion is associated with reduced
misconception that failing to find evidence of renal injury. Specific therapies directed at this
an STEC/VTEC infection proves that such an pathogen or its products are either harmful
etiologic agent is absent. It is important to (antibiotics) or unlikely to work (the toxemia

312 DAVIS ET AL.
is short lived). Therapeutic plasma exchange 4. Gould LH, Mody RK, Ong KL, Clogher P,
and complement inhibition are not justified by Cronquist AB, Garman KN, Lathrop S, Medus
C, Spina NL, Webb TH, White PL, Wymore K,
credible data. Gierke RE, Mahon BE, Griffin PM, Emerging
Infections Program Foodnet Working Group.
2013. Increased recognition of non-O157 Shiga
ACKNOWLEDGMENTS toxin-producing Escherichia coli infections in the
United States during 20002010: epidemiologic
Cepheid has paid P. I. Tarr an honorarium for
features and comparison with E. coli O157 infec-
a lecture at its headquarters on this topic. tions. Foodborne Pathog Dis 10:453460.
Some work in his laboratory is supported by a 5. European Food Safety Authority. 2012. The
grant to another investigator at Washington European Union Summary Report on Trends and
University from Alexion Corporation (manu- Sources of Zoonoses, Zoonotic Agents and Food-
facturers of eculizumab). N.C. A. J. van de Kar borne Outbreaks in 2010. EFSA Journal 10:161
188.
is a member of the International Advisory 6. Konowalchuk J, Speirs JI, Stavric S. 1977. Vero
Board of aHUS, Alexion Corporation. response to a cytotoxin of Escherichia coli. Infect
We are grateful to families, patients, and Immun 18:775779.
collaborators for teaching us much about gut 7. Karmali MA, Steele BT, Petric M, Lim C. 1983.
infections and HUS over the past3 decades. Sporadic cases of haemolytic-uraemic syndrome
associated with faecal cytotoxin and cytotoxin-
We thank Ariana Jasarevic for expert man- producing Escherichia coli in stools. Lancet 1:619
uscript assistance, Alexander Weymann for 620.
translating original literature from German 8. Miceli S, Jure MA, de Saab OA, de Castillo
to English, and Vikas Dharnidharka for pro- MC, Rojas S, de Holgado AP, de Nader OM.
1999. A clinical and bacteriological study of chil-
viding helpful comments during the creation
dren suffering from haemolytic uraemic syn-
of this manuscript. drome in Tucuman, Argentina. Jpn J Infect Dis
52:3337.
9. Brandt JR, Fouser LS, Watkins SL, Zelikovic I,
CITATION Tarr PI, Nazar-Stewart V, Avner ED. 1994.
Davis TK, Van De Kar NCAJ, Tarr PI. Escherichia coli O 157:H7-associated hemolytic-
uremic syndrome after ingestion of contaminated
2014. Shiga toxin/verocytotoxin-producing hamburgers. J Pediatr 125:519526.
Escherichia coli infections: practical clinical 10. Riley LW, Remis RS, Helgerson SD, McGee HB,
perspectives. Microbiol Spectrum 2(4):EHEC- Wells JG, Davis BR, Hebert RJ, Olcott ES,
0025-2014. Johnson LM, Hargrett NT, Blake PA, Cohen
ML. 1983. Hemorrhagic colitis associated with a
rare Escherichia coli serotype. N Engl J Med
REFERENCES 308:681685.
11. Meites S, Buffone GJ. 1989. Pediatric Clinical
1. Denno DM, Shaikh N, Stapp JR, Qin X, Hutter Chemistry: Reference (Normal) Values. AACC
CM, Hoffman V, Mooney JC, Wood KM, Press, Washington, DC.
Stevens HJ, Jones R, Tarr PI, Klein EJ. 2012. 12. Banatvala N, Griffin PM, Greene KD, Barrett
Diarrhea etiology in a pediatric emergency de- TJ, Bibb WF, Green JH, Wells JG, Hemolytic
partment: a case control study. Clin Infect Dis Uremic Syndrome Study Collaborators. 2001.
55:897904. The United States National Prospective Hemo-
2. Watanabe H, Terajima J, Izumiya H, Wada A, lytic Uremic Syndrome Study: microbiologic,
Tamura K. 1999. Molecular analysis of entero- serologic, clinical, and epidemiologic findings.
hemorrhagic Escherichia coli isolates in Japan and J Infect Dis 183:10631070.
its application to epidemiological investigation. 13. Tsai HM, Chandler WL, Sarode R, Hoffman R,
Pediatr Int 41:202208. Jelacic S, Habeeb RL, Watkins SL, Wong CS,
3. Stephan R, Untermann F. 1999. Virulence factors Williams GD, Tarr PI. 2001. von Willebrand
and phenotypical traits of verotoxin-producing factor and von Willebrand factor-cleaving metal-
Escherichia coli strains isolated from asymptom- loprotease activity in Escherichia coli O157:H7-
atic human carriers. J Clin Microbiol 37:1570 associated hemolytic uremic syndrome. Pediatr
1572. Res 49:653659.

14. Jelacic S, Wobbe CL, Boster DR, Ciol MA, 24. Bielaszewska M, Middendorf B, Kock R,
Watkins SL, Tarr PI, Stapleton AE. 2002. Friedrich AW, Fruth A, Karch H, Schmidt
ABO and P1 blood group antigen expression and MA, Mellmann A. 2008. Shiga toxin-negative
stx genotype and outcome of childhood Esche- attaching and effacing Escherichia coli: distinct
richia coli O157:H7 infections. J Infect Dis 185: clinical associations with bacterial phylogeny and
214219. virulence traits and inferred in-host pathogen
15. Cornick NA, Jelacic S, Ciol MA, Tarr PI. 2002. evolution. Clin Infect Dis 47:208217.
Escherichia coli O157:H7 infections: discordance 25. Bielaszewska M, Kock R, Friedrich AW, von
between filterable fecal shiga toxin and disease Eiff C, Zimmerhackl LB, Karch H, Mellmann
outcome. J Infect Dis 186:5763. A. 2007. Shiga toxin-mediated hemolytic uremic
16. Klein EJ, Stapp JR, Clausen CR, Boster DR, syndrome: time to change the diagnostic para-
Wells JG, Qin X, Swerdlow DL, Tarr PI. 2002. digm? PLoS One 2:e1024.
Shiga toxin-producing Escherichia coli in children 26. Friedrich AW, Zhang W, Bielaszewska M,
with diarrhea: a prospective point-of-care study. Mellmann A, Kock R, Fruth A, Tschape H,
J Pediatr 141:172177. Karch H. 2007. Prevalence, virulence profiles, and
17. Smith JM, Jones F, Ciol MA, Jelacic S, clinical significance of Shiga toxin-negative vari-
Boster DR, Watkins SL, Williams GD, Tarr PI, ants of enterohemorrhagic Escherichia coli O157
Henderson WR Jr. 2002. Platelet-activating infection in humans. Clin Infect Dis 45:3945.
factor and Escherichia coli O157:H7 infections. 27. Bielaszewska M, Friedrich AW, Aldick T,
Pediatr Nephrol 17:10471052. Schurk-Bulgrin R, Karch H. 2006. Shiga toxin
18. Thayu M, Chandler WL, Jelacic S, Gordon CA, activatable by intestinal mucus in Escherichia coli
Rosenthal GL, Tarr PI. 2003. Cardiac ischemia isolated from humans: predictor for a severe
during hemolytic uremic syndrome. Pediatr Nephrol clinical outcome. Clin Infect Dis 43:11601167.
18:286289. 28. Sonntag AK, Prager R, Bielaszewska M, Zhang
19. Jelacic JK, Damrow T, Chen GS, Jelacic S, W, Fruth A, Tschape H, Karch H. 2004. Phe-
Bielaszewska M, Ciol M, Carvalho HM, Melton- notypic and genotypic analyses of enterohemor-
Celsa AR, OBrien AD, Tarr PI. 2003. Shiga rhagic Escherichia coli O145 strains from patients
toxin-producing Escherichia coli in Montana: bac- in Germany. J Clin Microbiol 42:954962.
terial genotypes and clinical profiles. J Infect Dis 29. Ikeda K, Ida O, Kimoto K, Takatorige T,
188:719729. Nakanishi N, Tatara K. 1999. Effect of early
20. Ake JA, Jelacic S, Ciol MA, Watkins SL, fosfomycin treatment on prevention of hemolytic
Murray KF, Christie DL, Klein EJ, Tarr PI. uremic syndrome accompanying Escherichia coli
2005. Relative nephroprotection during Esche- O157:H7 infection. Clin Nephrol 52:357362.
richia coli O157:H7 infections: association with 30. Loos S, Ahlenstiel T, Kranz B, Staude H, Pape L,
intravenous volume expansion. Pediatrics 115: Hartel C, Vester U, Buchtala L, Benz K, Hoppe
e673e680. B, Beringer O, Krause M, Muller D, Pohl M,
21. Hickey CA, Beattie TJ, Cowieson J, Miyashita Lemke J, Hillebrand G, Kreuzer M, Konig J,
Y, Strife CF, Frem JC, Peterson JM, Butani L, Wigger M, Konrad M, Haffner D, Oh J, Kemper
Jones DP, Havens PL, Patel HP, Wong CS, MJ. 2012. An outbreak of Shiga toxin-producing
Andreoli SP, Rothbaum RJ, Beck AM, Tarr PI. Escherichia coli O104:H4 hemolytic uremic syn-
2011. Early volume expansion during diarrhea and drome in Germany: presentation and short-term
relative nephroprotection during subsequent he- outcome in children. Clin Infect Dis 55:753759.
molytic uremic syndrome. Arch Pediatr Adolesc 31. Wong CS, Mooney JC, Brandt JR, Staples AO,
Med 165:884889. Jelacic S, Boster DR, Watkins SL, Tarr PI. 2012.
22. Zimmerhackl LB, Rosales A, Hofer J, Riedl M, Risk factors for the hemolytic uremic syndrome
Jungraithmayr T, Mellmann A, Bielaszewska in children infected with Escherichia coli O157:H7:
M, Karch H. 2010. Enterohemorrhagic Esche- a multivariable analysis. Clin Infect Dis 55:3341.
richia coli O26:H11-associated Hemolytic uremic 32. Gasser C, Gautier E, Steck A, Siebenmann RE,
syndrome: bacteriology and clinical presentation. Oechslin R. 1955. [Hemolytic-uremic syndrome:
Semin Thromb Hemost 36:586593. bilateral necrosis of the renal cortex in acute
23. Gerber A, Karch H, Allerberger F, Verweyen acquired hemolytic anemia]. Schweiz Med
HM, Zimmerhackl LB. 2002. Clinical course and Wochenschr 85:905909.
the role of shiga toxin-producing Escherichia coli 33. Fahr T. 1925. Kreislaufstrungen in der Niere,
infection in the hemolytic-uremic syndrome in p. 121155. In Fahr T, Gruber G, Koch M,
pediatric patients, 19972000, in Germany and Lubarsch O, Stoerk O (ed.), Harnorgane Mnnliche
Austria: a prospective study. J Infect Dis 186:493 Geschlechtsorgane, vol. 6. Springer, Vienna,
500. Austria.

314 DAVIS ET AL.
34. Bamforth J. 1923. A case of symmetrical cortical miologic subtyping of Shiga toxin-producing
necrosis of the kidneys occurring in an adult man. Escherichia coli strains isolated from hemolytic
J Pathol Bacteriol 26:4045. uremic syndrome and diarrhea cases in Argentina.
35. Campbell AC, Henderson JL. 1949. Symmetrical Foodborne Pathog Dis 3:8896.
cortical necrosis of the kidneys in infancy and 47. Pollock KG, Young D, Beattie TJ, Todd WT.
childhood. Arch Dis Child 24:269285, illust. 2008. Clinical surveillance of thrombotic micro-
36. Dunn JS, Montgomery GL. 1941. Acute nec- angiopathies in Scotland, 2003-2005. Epidemiol
rotising glomerulonephritis. J Pathol Bacteriol Infect 136:115121.
52:116. 48. Hadler JL, Clogher P, Hurd S, Phan Q,
37. Hensley WJ. 1952. Haemolytic anaemia in acute Mandour M, Bemis K, Marcus R. 2011. Ten-year
glomerulonephritis. Australas Ann Med 1:180 trends and risk factors for non-O157 Shiga toxin-
185. producing Escherichia coli found through Shiga
38. Leopold SR, Magrini V, Holt NJ, Shaikh N, toxin testing, Connecticut, 20002009. Clin Infect
Mardis ER, Cagno J, Ogura Y, Iguchi A, Dis 53:269276.
Hayashi T, Mellmann A, Karch H, Besser TE, 49. Mody RK, Luna-Gierke RE, Jones TF,
Sawyer SA, Whittam TS, Tarr PI. 2009. A pre- Comstock N, Hurd S, Scheftel J, Lathrop S,
cise reconstruction of the emergence and con- Smith G, Palmer A, Strockbine N, Talkington D,
strained radiations of Escherichia coli O157 Mahon BE, Hoekstra RM, Griffin PM. 2012.
portrayed by backbone concatenomic analysis. Infections in pediatric postdiarrheal hemolytic
Proc Natl Acad Sci USA 106:87138718. uremic syndrome: factors associated with identi-
39. OBrien AO, Lively TA, Chen ME, Rothman fying shiga toxin-producing Escherichia coli. Arch
SW, Formal SB. 1983. Escherichia coli O157:H7 Pediatr Adolesc Med 166:902909.
strains associated with haemorrhagic colitis in the 50. Werber D, Bielaszewska M, Frank C, Stark K,
United States produce a Shigella dysenteriae 1 Karch H. 2011. Watch out for the even eviler
(SHIGA) like cytotoxin. Lancet 1:702. cousin-sorbitol-fermenting E coli O157. Lancet
40. OBrien AD, LaVeck GD, Thompson MR, 377:298299.
Formal SB. 1982. Production of Shigella dysen- 51. Chapman PA, Siddons CA. 1996. A comparison
teriae type 1-like cytotoxin by Escherichia coli. of immunomagnetic separation and direct cul-
J Infect Dis 146:763769. ture for the isolation of verocytotoxin-producing
41. Johnson WM, Lior H, Bezanson GS. 1983. Cy- Escherichia coli O157 from cases of bloody diar-
totoxic Escherichia coli O157:H7 associated with rhoea, non-bloody diarrhoea and asymptomatic
haemorrhagic colitis in Canada. Lancet 1:76. contacts. J Med Microbiol 44:267271.
42. OBrien AD, LaVeck GD. 1983. Purification and 52. Fey PD, Wickert RS, Rupp ME, Safranek TJ,
characterization of a Shigella dysenteriae 1-like Hinrichs SH. 2000. Prevalence of non-O157:H7
toxin produced by Escherichia coli. Infect Immun shiga toxin-producing Escherichia coli in diarrheal
40:675683. stool samples from Nebraska. Emerg Infect Dis
43. Neill MA, Tarr PI, Clausen CR, Christie DL, 6:530533.
Hickman RO. 1987. Escherichia coli O157:H7 as 53. Carroll KC, Adamson K, Korgenski K, Croft A,
the predominant pathogen associated with the Hankemeier R, Daly J, Park CH. 2003. Com-
hemolytic uremic syndrome: a prospective study parison of a commercial reversed passive latex
in the Pacific Northwest. Pediatrics 80:3740. agglutination assay to an enzyme immunoassay
44. Tarr PI, Neill MA, Clausen CR, Watkins SL, for the detection of Shiga toxin-producing Esch-
Christie DL, Hickman RO. 1990. Escherichia coli erichia coli. Eur J Clin Microbiol Infect Dis 22:
O157:H7 and the hemolytic uremic syndrome: 689692.
importance of early cultures in establishing the 54. Manning SD, Madera RT, Schneider W,
etiology. J Infect Dis 162:553556. Dietrich SE, Khalife W, Brown W, Whittam
45. Rowe PC, Orrbine E, Lior H, Wells GA, Yetisir TS, Somsel P, Rudrik JT. 2007. Surveillance for
E, Clulow M, McLaine PN. 1998. Risk of hemo- Shiga toxin-producing Escherichia coli, Michigan,
lytic uremic syndrome after sporadic Escherichia 20012005. Emerg Infect Dis 13:318321.
coli O157:H7 infection: results of a Canadian 55. Teel LD, Daly JA, Jerris RC, Maul D, Svanas G,
collaborative study. Investigators of the Cana- OBrien AD, Park CH. 2007. Rapid detection of
dian Pediatric Kidney Disease Research Center. Shiga toxin-producing Escherichia coli by optical
J Pediatr 132:777782. immunoassay. J Clin Microbiol 45:33773380.
46. Rivas M, Miliwebsky E, Chinen I, Roldan CD, 56. Park CH, Kim HJ, Hixon DL, Bubert A. 2003.
Balbi L, Garcia B, Fiorilli G, Sosa-Estani S, Evaluation of the duopath verotoxin test for de-
Kincaid J, Rangel J, Griffin PM, Case-Control tection of shiga toxins in cultures of human stools.
Study Group. 2006. Characterization and epide- J Clin Microbiol 41:26502653.

57. Gould LH, Bopp C, Strockbine N, Atkinson R, factors for the development of Escherichia coli
Baselski V, Body B, Carey R, Crandall C, Hurd O157:H7-associated hemolytic uremic syndrome.
S, Kaplan R, Neill M, Shea S, Somsel P, Tobin- Clin Nephrol 42:8589.
DAngelo M, Griffin PM, Gerner-Smidt P, 69. Misurac JM, Knoderer CA, Leiser JD, Nailescu
Centers for Disease Control and Prevention C, Wilson AC, Andreoli SP. 2013. Nonsteroidal
(CDC). 2009. Recommendations for diagnosis of anti-inflammatory drugs are an important cause
shiga toxinproducing Escherichia coli infections of acute kidney injury in children. J Pediatr
by clinical laboratories. MMWR Recomm Rep 162:11531159, 1159 e1.
58:114. 70. John CM, Shukla R, Jones CA. 2007. Using
58. Grys TE, Sloan LM, Rosenblatt JE, Patel R. NSAID in volume depleted children can precipi-
2009. Rapid and sensitive detection of Shiga tate acute renal failure. Arch Dis Child 92:524
toxin-producing Escherichia coli from non- 526.
enriched stool specimens by real-time PCR in 71. Havens PL, ORourke PP, Hahn J, Higgins J,
comparison to enzyme immunoassay and culture. Walker AM. 1988. Laboratory and clinical varia-
J Clin Microbiol 47:20082012. bles to predict outcome in hemolytic-uremic syn-
59. Vallieres E, Saint-Jean M, Rallu F. 2013. Com- drome. Am J Dis Child 142:961964.
parison of three different methods for detection 72. Vitacco M, Sanchez Avalos J, Gianantonio CA.
of Shiga toxin-producing Escherichia coli in a 1973. Heparin therapy in the hemolytic-uremic
tertiary pediatric care center. J Clin Microbiol syndrome. J Pediatr 83:271275.
51:481486. 73. Walters MD, Matthei IU, Kay R, Dillon MJ,
60. Werber D, Mason BW, Evans MR, Salmon RL. Barratt TM. 1989. The polymorphonuclear
2008. Preventing household transmission of Shiga leucocyte count in childhood haemolytic uraemic
toxin-producing Escherichia coli O157 infection: syndrome. Pediatr Nephrol 3:130134.
promptly separating siblings might be the key. 74. Milford DV, Staten J, MacGreggor I, Dawes J,
Clin Infect Dis 46:11891196. Taylor CM, Hill FG. 1991. Prognostic markers
61. Chandler WL, Jelacic S, Boster DR, Ciol MA, in diarrhoea-associated haemolytic-uraemic syn-
Williams GD, Watkins SL, Igarashi T, Tarr PI. drome: initial neutrophil count, human neutro-
2002. Prothrombotic coagulation abnormalities phil elastase and von Willebrand factor antigen.
preceding the hemolytic-uremic syndrome. N Engl Nephrol Dial Transplant 6:232237.
J Med 346:2332. 75. Balestracci A, Martin SM, Toledo I, Alvarado
62. Balestracci A, Martin SM, Toledo I, Alvarado C, C, Wainsztein RE. 2014. Laboratory predictors
Wainsztein RE. 2012. Dehydration at admission of acute dialysis in hemolytic uremic syndrome.
increased the need for dialysis in hemolytic uremic Pediatr Int 56:234239.
syndrome children. Pediatr Nephrol 27:14071410. 76. Dolislager D, Tune B. 1978. The hemolytic-
63. Coad NA, Marshall T, Rowe B, Taylor CM. 1991. uremic syndrome: spectrum of severity and sig-
Changes in the postenteropathic form of the nificance of prodrome. Am J Dis Child 132:5558.
hemolytic uremic syndrome in children. Clin 77. Gianantonio CA, Vitacco M, Mendilaharzu F,
Nephrol 35:1016. Gallo GE, Sojo ET. 1973. The hemolytic-uremic
64. Oakes RS, Siegler RL, McReynolds MA, syndrome. Nephron 11:174192.
Pysher T, Pavia AT. 2006. Predictors of fatality 78. Gianantonio CA, Vitacco M, Mendilaharzu F,
in postdiarrheal hemolytic uremic syndrome. Pe- Gallo G. 1968. The hemolytic-uremic syndrome.
diatrics 117:16561662. Renal status of 76 patients at long-term follow-up.
65. Ojeda JM, Kohout I, Cuestas E. 2013. Dehy- J Pediatr 72:757765.
dration upon admission is a risk factor for in- 79. Tonshoff B, Sammet A, Sanden I, Mehls O,
complete recovery of renal function in children Waldherr R, Scharer K. 1994. Outcome and
with haemolytic uremic syndrome. Nefrologia 33: prognostic determinants in the hemolytic uremic
372376. syndrome of children. Nephron 68:6370.
66. Holtz LR, Neill MA, Tarr PI. 2009. Acute bloody 80. Siegler RL, Pavia AT, Christofferson RD,
diarrhea: a medical emergency for patients of all Milligan MK. 1994. A 20-year population-based
ages. Gastroenterology 136:18871898. study of postdiarrheal hemolytic uremic syn-
67. Bell BP, Griffin PM, Lozano P, Christie DL, drome in Utah. Pediatrics 94:3540.
Kobayashi JM, Tarr PI. 1997. Predictors of he- 81. Mizusawa Y, Pitcher LA, Burke JR, Falk MC,
molytic uremic syndrome in children during a Mizushima W. 1996. Survey of haemolytic-
large outbreak of Escherichia coli O157:H7 infec- uraemic syndrome in Queensland 19791995. Med
tions. Pediatrics 100:E12. J Aust 165:188191.
68. Cimolai N, Basalyga S, Mah DG, Morrison BJ, 82. Spizzirri FD, Rahman RC, Bibiloni N, Ruscasso
Carter JE. 1994. A continuing assessment of risk JD, Amoreo OR. 1997. Childhood hemolytic

316 DAVIS ET AL.
uremic syndrome in Argentina: long-term follow- transfusions in children with post-diarrheal he-
up and prognostic features. Pediatr Nephrol 11: molytic uremic syndrome. Pediatr Nephrol 28:
156160. 919925.
83. Mencia Bartolome S, Martinez de Azagra A, 96. Schulman SL, Kaplan BS. 1996. Management of
de Vicente Aymat A, Monleon Luque M, patients with hemolytic uremic syndrome dem-
Casado Flores J. 1999. [Uremic hemolytic syn- onstrating severe azotemia but not anuria. Pediatr
drome. Analysis of 43 cases]. An Esp Pediatr Nephrol 10:671674.
50:467470 (In Spanish). 97. Sutherland SM, Zappitelli M, Alexander
84. Huseman D, Gellermann J, Vollmer I, Ohde I, SR, Chua AN, Brophy PD, Bunchman TE,
Devaux S, Ehrich JH, Filler G. 1999. Long-term Hackbarth R, Somers MJ, Baum M, Symons
prognosis of hemolytic uremic syndrome and JM, Flores FX, Benfield M, Askenazi D, Chand
effective renal plasma flow. Pediatr Nephrol 13: D, Fortenberry JD, Mahan JD, McBryde K,
672677. Blowey D, Goldstein SL. 2010. Fluid overload
85. Loirat C. 2001. [Post-diarrhea hemolytic-uremic and mortality in children receiving continuous
syndrome: clinical aspects]. Arch Pediatr 8 Suppl renal replacement therapy: the prospective pedi-
4:776s784s (In French). atric continuous renal replacement therapy reg-
86. Garg AX, Suri RS, Barrowman N, Rehman F, istry. Am J Kidney Dis 55:316325.
Matsell D, Rosas-Arellano MP, Salvadori M, 98. Kaplan BS, Thomson PD. 1976. Hyperuricemia
Haynes RB, Clark WF. 2003. Long-term renal in the hemolytic-uremic syndrome. Am J Dis
prognosis of diarrhea-associated hemolytic ure- Child 130:854856.
mic syndrome: a systematic review, meta-analysis, 99. Robson WL, Leung AK, Brant R. 1993. Rela-
and meta-regression. JAMA 290:13601370. tionship of the recovery in the glomerular filtration
87. Oakes RS, Kirkham JK, Nelson RD, Siegler RL. rate to the duration of anuria in diarrhea-associated
2008. Duration of oliguria and anuria as predic- hemolytic uremic syndrome. Am J Nephrol 13:194
tors of chronic renal-related sequelae in post- 197.
diarrheal hemolytic uremic syndrome. Pediatr 100. Proesmans W, VanCauter A, Thijs L, Lijnen P.
Nephrol 23:13031308. 1994. Plasma renin activity in haemolytic uraemic
88. Nestoridi E, Kushak RI, Duguerre D, Grabowski syndrome. Pediatr Nephrol 8:444446.
EF, Ingelfinger JR. 2005. Up-regulation of tissue 101. Katz J, Krawitz S, Sacks PV, Levin SE,
factor activity on human proximal tubular epi- Thomson P, Levin J, Metz J. 1973. Platelet,
thelial cells in response to Shiga toxin. Kidney Int erythrocyte, and fibrinogen kinetics in the
67:22542266. hemolytic-uremic syndrome of infancy. J Pediatr
89. Hughes AK, Stricklett PK, Kohan DE. 1998. 83:739748.
Cytotoxic effect of Shiga toxin-1 on human prox- 102. Nakahata T, Tanaka H, Tateyama T, Ueda T,
imal tubule cells. Kidney Int 54:426437. Suzuki K, Osari S, Kasai M, Waga S. 2001.
90. Taguchi T, Uchida H, Kiyokawa N, Mori T, Thrombotic stroke in a child with diarrhea-
Sato N, Horie H, Takeda T, Fujimoto J. 1998. associated hemolytic-uremic syndrome with a
Verotoxins induce apoptosis in human renal tu- good recovery. Tohoku J Exp Med 193:7377.
bular epithelium derived cells. Kidney Int 53: 103. Trevathan E, Dooling EC. 1987. Large throm-
16811688. botic strokes in hemolytic-uremic syndrome.
91. Kaneko K, Kiyokawa N, Ohtomo Y, Nagaoka R, J Pediatr 111:863866.
Yamashiro Y, Taguchi T, Mori T, Fujimoto J, 104. Steinborn M, Leiz S, Rudisser K, Griebel M,
Takeda T. 2001. Apoptosis of renal tubular cells Harder T, Hahn H. 2004. CT and MRI in
in Shiga-toxin-mediated hemolytic uremic syn- haemolytic uraemic syndrome with central ner-
drome. Nephron 87:182185. vous system involvement: distribution of lesions
92. Abuelo JG. 2007. Normotensive ischemic acute and prognostic value of imaging findings. Pediatr
renal failure. N Engl J Med 357:797805. Radiol 34:805810.
93. Davis TK, McKee R, Schnadower D, Tarr PI. 105. Nathanson S, Kwon T, Elmaleh M, Charbit M,
2013. Treatment of Shiga toxinProducing Esch- Launay EA, Harambat J, Brun M, Ranchin B,
erichia coli infections. Infect Dis Clin North Am Bandin F, Cloarec S, Bourdat-Michel G,
27:577597. Pietrement C, Champion G, Ulinski T,
94. Weil BR, Andreoli SP, Billmire DF. 2010. Deschenes G. 2010. Acute neurological involve-
Bleeding risk for surgical dialysis procedures in ment in diarrhea-associated hemolytic uremic
children with hemolytic uremic syndrome. syndrome. Clin J Am Soc Nephrol 5:12181228.
Pediatr Nephrol 25:16931698. 106. Weissenborn K, Donnerstag F, Kielstein JT,
95. Balestracci A, Martin SM, Toledo I, Alvarado Heeren M, Worthmann H, Hecker H, Schmitt
C, Wainsztein RE. 2013. Impact of platelet R, Schiffer M, Pasedag T, Schuppner R, Tryc

AB, Raab P, Hartmann H, Ding XQ, Hafer C, active role of complement in hemolytic uremic
Menne J, Schmidt BM, Bultmann E, Haller H, syndrome. J Immunol 182:63946400.
Dengler R, Lanfermann H, Giesemann AM. 117. Morigi M, Galbusera M, Gastoldi S, Locatelli M,
2012. Neurologic manifestations of E coli infec- Buelli S, Pezzotta A, Pagani C, Noris M, Gobbi
tion-induced hemolytic-uremic syndrome in M, Stravalaci M, Rottoli D, Tedesco F, Remuzzi
adults. Neurology 79:14661473. G, Zoja C. 2011. Alternative pathway activation of
107. Signorini E, Lucchi S, Mastrangelo M, Rapuzzi complement by Shiga toxin promotes exuberant
S, Edefonti A, Fossali E. 2000. Central nervous C3a formation that triggers microvascular throm-
system involvement in a child with hemolytic bosis. J Immunol 187:172180.
uremic syndrome. Pediatr Nephrol 14:990992. 118. Stahl AL, Sartz L, Karpman D. 2011. Comple-
108. Steele BT, Murphy N, Chuang SH, McGreal ment activation on platelet-leukocyte complexes
D, Arbus GS. 1983. Recovery from prolonged and microparticles in enterohemorrhagic Esche-
coma in hemolytic uremic syndrome. J Pediatr richia coli-induced hemolytic uremic syndrome.
102:402404. Blood 117:55035513.
109. Steinberg A, Ish-Horowitcz M, el-Peleg O, Mor 119. Lopez EL, Contrini MM, Glatstein E, Ayala SG,
J, Branski D. 1986. Stroke in a patient with he- Santoro R, Ezcurra G, Teplitz E, Matsumoto Y,
molytic-uremic syndrome with a good outcome. Sato H, Sakai K, Katsuura Y, Hoshide S, Morita
Brain Dev 8:7072. T, Harning R, Brookman S. 2012. An epidemio-
110. Kahn SI, Tolkan SR, Kothari O, Garella S. 1982. logic surveillance of Shiga-like toxin-producing
Spontaneous recovery of the hemolytic uremic Escherichia coli infection in Argentinean children:
syndrome with prolonged renal and neurological risk factors and serum Shiga-like toxin 2 values.
manifestations. Nephron 32:188191. Pediatr Infect Dis J 31:2024.
111. Magnus T, Rother J, Simova O, Meier-Cillien 120. Lee BC, Mayer CL, Leibowitz CS, Stearns-
M, Repenthin J, Moller F, Gbadamosi J, Kurosawa DJ, Kurosawa S. 2013. Quiescent
Panzer U, Wengenroth M, Hagel C, Kluge S, complement in nonhuman primates during E coli
Stahl RK, Wegscheider K, Urban P, Eckert B, Shiga toxin-induced hemolytic uremic syndrome
Glatzel M, Fiehler J, Gerloff C. 2012. The neu- and thrombotic microangiopathy. Blood 122:803
rological syndrome in adults during the 2011 806.
northern German E. coli serotype O104:H4 out- 121. Thurman JM, Marians R, Emlen W, Wood S,
break. Brain 135:18501859. Smith C, Akana H, Holers VM, Lesser M, Kline
112. Braune SA, Wichmann D, von Heinz MC, M, Hoffman C, Christen E, Trachtman H. 2009.
Nierhaus A, Becker H, Meyer TN, Meyer GP, Alternative pathway of complement in children
Muller-Schulz M, Fricke J, de Weerth A, with diarrhea-associated hemolytic uremic syn-
Hoepker WW, Fiehler J, Magnus T, Gerloff C, drome. Clin J Am Soc Nephrol 4:19201924.
Panzer U, Stahl RA, Wegscheider K, Kluge S. 122. Ganter MT, Brohi K, Cohen MJ, Shaffer LA,
2013. Clinical features of critically ill patients with Walsh MC, Stahl GL, Pittet JF. 2007. Role of the
Shiga toxin-induced hemolytic uremic syndrome. alternative pathway in the early complement ac-
Crit Care Med 41:17021710. tivation following major trauma. Shock 28:2934.
113. Murphy EJ. 2005. Acute pain management 123. Rmer S, Schaumburg R, Karch H, Schwindt
pharmacology for the patient with concurrent W, Waltenberger J, Lebiedz P. 2012. Septic
renal or hepatic disease. Anaesth Intensive Care shock and ARDS after Eculizumab application
33:311322. in Shiga-toxin associated HUS during the out-
114. Niscola P, Scaramucci L, Vischini G, Giovannini break of Shiga toxin-producing enterohaemor-
M, Ferrannini M, Massa P, Tatangelo P, Galletti rhagic E. coli O104:H4. Medizinische Klinik,
M, Palumbo R. 2010. The use of major analgesics Intensivmedizin und Notfallmedizin 107:329330.
in patients with renal dysfunction. Curr Drug 124. Tarr PI. 2009. Shiga toxin-associated hemolytic
Targets 11:752758. uremic syndrome and thrombotic thrombocyto-
115. Lapeyraque AL, Malina M, Fremeaux-Bacchi V, penic purpura: distinct mechanisms of patho-
Boppel T, Kirschfink M, Oualha M, Proulx F, genesis. Kidney Int Suppl :S29S32.
Clermont MJ, Le Deist F, Niaudet P, Schaefer 125. Colic E, Dieperink H, Titlestad K, Tepel M.
F. 2011. Eculizumab in severe Shiga-toxin- 2011. Management of an acute outbreak of diar-
associated HUS. N Engl J Med 364:25612563. rhoea-associated haemolytic uraemic syndrome
116. Orth D, Khan AB, Naim A, Grif K, Brockmeyer with early plasma exchange in adults from
J, Karch H, Joannidis M, Clark SJ, Day AJ, southern Denmark: an observational study. Lan-
Fidanzi S, Stoiber H, Dierich MP, Zimmerhackl cet 378:10891093.
LB, Wurzner R. 2009. Shiga toxin activates 126. Veyradier A, Obert B, Haddad E, Cloarec S,
complement and binds factor H: evidence for an Nivet H, Foulard M, Lesure F, Delattre P,

318 DAVIS ET AL.
Lakhdari M, Meyer D, Girma JP, Loirat C. mutations and clinical characteristics. Pediatr
2003. Severe deficiency of the specific von Nephrol 27:12831291.
Willebrand factor-cleaving protease (ADAMTS 135. Noris M, Caprioli J, Bresin E, Mossali C,
13) activity in a subgroup of children with atypical Pianetti G, Gamba S, Daina E, Fenili C,
hemolytic uremic syndrome. J Pediatr 142:310 Castelletti F, Sorosina A, Piras R, Donadelli R,
317. Maranta R, van der Meer I, Conway EM, Zipfel
127. Kielstein JT, Beutel G, Fleig S, Steinhoff J, PF, Goodship TH, Remuzzi G. 2010. Relative
Meyer TN, Hafer C, Kuhlmann U, Bramstedt J, role of genetic complement abnormalities in
Panzer U, Vischedyk M, Busch V, Ries W, sporadic and familial aHUS and their impact
Mitzner S, Mees S, Stracke S, Nurnberger J, on clinical phenotype. Clin J Am Soc Nephrol
Gerke P, Wiesner M, Sucke B, Abu-Tair M, 5:18441859.
Kribben A, Klause N, Schindler R, Merkel F, 136. Westra D, Wetzels JF, Volokhina EB, van den
Schnatter S, Dorresteijn EM, Samuelsson O, Heuvel LP, van de Kar NC. 2012. A new era in
Brunkhorst R, Collaborators of the DGfN the diagnosis and treatment of atypical hae-
STEC-HUS registry. 2012. Best supportive care molytic uraemic syndrome. Neth J Med 70:121
and therapeutic plasma exchange with or with- 129.
out eculizumab in Shiga-toxin-producing E. coli 137. Ariceta G, Besbas N, Johnson S, Karpman D,
O104:H4 induced haemolytic-uraemic syndrome: Landau D, Licht C, Loirat C, Pecoraro C, Taylor
an analysis of the German STEC-HUS registry. CM, Van de Kar N, Vandewalle J, Zimmerhackl
Nephrol Dial Transplant 27:38073815. LB, European Paediatric Study Group for HUS.
128. Samuelsson O, Follin P, Rundgren M, Rylander 2009. Guideline for the investigation and initial
C, Selga D, Stahl A. 2012. [The HUS epidemic in therapy of diarrhea-negative hemolytic uremic
the summer of 2011 was severe. German and syndrome. Pediatr Nephrol 24:687696.
Swedish experiences of the EHEC outbreak]. 138. Espie E, Grimont F, Mariani-Kurkdjian P,
Lakartidningen 109:12301234 (In Swedish). Bouvet P, Haeghebaert S, Filliol I, Loirat C,
129. Nitschke M, Sayk F, Hartel C, Roseland RT, Decludt B, Minh NN, Vaillant V, de Valk H.
Hauswaldt S, Steinhoff J, Fellermann K, Derad 2008. Surveillance of hemolytic uremic syndrome
I, Wellhoner P, Buning J, Tiemer B, Katalinic A, in children less than 15 years of age, a system
Rupp J, Lehnert H, Solbach W, Knobloch JK. to monitor O157 and non-O157 Shiga toxin-
2012. Association between azithromycin therapy producing Escherichia coli infections in France,
and duration of bacterial shedding among patients 1996-2006. Pediatr Infect Dis J 27:595601.
with Shiga toxin-producing enteroaggregative 139. Chart H, Cheasty T. 2008. Human infections
Escherichia coli O104:H4. JAMA 307:10461052. with verocytotoxin-producing Escherichia coli
130. Som S, Deford CC, Kaiser ML, Terrell DR, O15710 years of E. coli O157 serodiagnosis.
Kremer Hovinga JA, Lammle B, George JN, J Med Microbiol 57:13891393.
Vesely SK. 2012. Decreasing frequency of plasma 140. Fernandez-Brando RJ, Bentancor LV, Mejias
exchange complications in patients treated for MP, Ramos MV, Exeni A, Exeni C, Laso Mdel C,
thrombotic thrombocytopenic purpura-hemolytic Exeni R, Isturiz MA, Palermo MS. 2011. Anti-
uremic syndrome, 1996 to 2011. Transfusion 52: body response to Shiga toxins in Argentinean
25252532; quiz 2524. children with enteropathic hemolytic uremic syn-
131. Havens PL, Hoffman GM, Shith KJ. 1989. The drome at acute and long-term follow-up periods.
homogeneous nature of the hemolytic uremic PLoS One 6:e19136.
syndrome. Clin Pediatr (Phila) 28:482483. 141. Close RM, Ejidokun OO, Verlander NQ, Fraser
132. Kaplan BS, Chesney RW, Drummond KN. 1975. G, Meltzer M, Rehman Y, Muir P, Ninis N,
Hemolytic uremic syndrome in families. N Engl J Stuart JM. 2011. Early diagnosis model for
Med 292:10901093. meningitis supports public health decision mak-
133. Karpac CA, Li X, Terrell DR, Kremer Hovinga ing. J Infect 63:3238.
JA, Lammle B, Vesely SK, George JN. 2008. 142. Bokete TN, OCallahan CM, Clausen CR, Tang
Sporadic bloody diarrhoea-associated thrombotic NM, Tran N, Moseley SL, Fritsche TR, Tarr PI.
thrombocytopenic purpura-haemolytic uraemic 1993. Shiga-like toxin-producing Escherichia coli
syndrome: an adult and paediatric comparison. Br in Seattle children: a prospective study. Gastro-
J Haematol 141:696707. enterology 105:17241731.
134. Geerdink LM, Westra D, van Wijk JA, 143. Carter AO, Borczyk AA, Carlson JA, Harvey B,
Dorresteijn EM, Lilien MR, Davin JC, Komhoff Hockin JC, Karmali MA, Krishnan C, Korn DA,
M, Van Hoeck K, van der Vlugt A, van den Lior H. 1987. A severe outbreak of Escherichia
Heuvel LP, van de Kar NC. 2012. Atypical he- coli O157:H7-associated hemorrhagic colitis in a
molytic uremic syndrome in children: complement nursing home. N Engl J Med 317:14961500.

144. Pavia AT, Nichols CR, Green DP, Tauxe RV, 154. Suri RS, Mahon JL, Clark WF, Moist LM,
Mottice S, Greene KD, Wells JG, Siegler RL, Salvadori M, Garg AX. 2009. Relationship be-
Brewer ED, Hannon D. 1990. Hemolytic-uremic tween Escherichia coli O157:H7 and diabetes
syndrome during an outbreak of Escherichia coli mellitus. Kidney Int Suppl :S44S46.
O157:H7 infections in institutions for mentally 155. US Census Bureau. 2010. 2010 Census Population
retarded persons: clinical and epidemiologic ob- and Housing Tables (CPH-Ts). http://quickfacts.
servations. J Pediatr 116:544551. census.gov/qfd/states/00000.html. Accessed on
145. Proulx F, Turgeon JP, Delage G, Lafleur L, June 18, 2014.
Chicoine L. 1992. Randomized, controlled trial of 156. Suri RS, Clark WF, Barrowman N, Mahon JL,
antibiotic therapy for Escherichia coli O157:H7 Thiessen-Philbrook HR, Rosas-Arellano MP,
enteritis. J Pediatr 121:299303. Zarnke K, Garland JS, Garg AX. 2005. Diabetes
146. Wong CS, Jelacic S, Habeeb RL, Watkins SL, during diarrhea-associated hemolytic uremic syn-
Tarr PI. 2000. The risk of the hemolytic-uremic drome: a systematic review and meta-analysis.
syndrome after antibiotic treatment of Escherichia Diabetes Care 28:25562562.
coli O157:H7 infections. N Engl J Med 342:1930 157. Andreoli SP, Bergstein JM. 1982. Development
1936. of insulin-dependent diabetes mellitus during the
147. Dundas S, Todd WT, Stewart AI, Murdoch PS, hemolytic-uremic syndrome. J Pediatr 100:541
Chaudhuri AK, Hutchinson SJ. 2001. The cen- 545.
tral Scotland Escherichia coli O157:H7 outbreak: 158. Tapper D, Tarr P, Avner E, Brandt J,
risk factors for the hemolytic uremic syndrome Waldhausen J. 1995. Lessons learned in the
and death among hospitalized patients. Clin Infect management of hemolytic uremic syndrome in
Dis 33:923931. children. J Pediatr Surg 30:158163.
148. Smith KE, Wilker PR, Reiter PL, Hedican EB, 159. Dundas S, Todd WT. 2000. Clinical presentation,
Bender JB, Hedberg CW. 2012. Antibiotic treat- complications and treatment of infection with
ment of Escherichia coli O157 infection and the verocytotoxin-producing Escherichia coli. Chal-
risk of hemolytic uremic syndrome, Minnesota. lenges for the clinician. Symp Ser Soc Appl
Pediatr Infect Dis J 31:3741. Microbiol :24S30S.
149. Grisaru S, Morgunov MA, Samuel SM, Midgley 160. Whitington PF, Friedman AL, Chesney RW.
JP, Wade AW, Tee JB, Hamiwka LA. 2011. 1979. Gastrointestinal disease in the hemolytic-
Acute renal replacement therapy in children with uremic syndrome. Gastroenterology 76:728733.
diarrhea-associated hemolytic uremic syndrome: 161. Yates RS, Osterholm RK. 1980. Hemolytic-
a single center 16 years of experience. Int J uremic syndrome colitis. J Clin Gastroenterol
Nephrol 2011:930539. 2:359363.
150. Adragna M, Balestracci A, Garcia Chervo L, 162. Sawaf H, Sharp MJ, Youn KJ, Jewell PA,
Steinbrun S, Delgado N, Briones L. 2012. Acute Rabbani A. 1978. Ischemic colitis and stricture after
dialysis-associated peritonitis in children with D+ hemolytic-uremic syndrome. Pediatrics 61:315317.
hemolytic uremic syndrome. Pediatr Nephrol 163. Abu-Arafeh I, Gray E, Youngson G, Auchterlonie
27:637642. I, Russell G. 1995. Myocarditis and haemolytic
151. Grodinsky S, Telmesani A, Robson WL, Fick G, uraemic syndrome. Arch Dis Child 72:4647.
Scott RB. 1990. Gastrointestinal manifestations 164. Sturm V, Menke MN, Landau K, Laube GF,
of hemolytic uremic syndrome: recognition of Neuhaus TJ. 2010. Ocular involvement in pae-
pancreatitis. J Pediatr Gastroenterol Nutr 11:518 diatric haemolytic uraemic syndrome. Acta
524. Ophthalmol 88:804807.
152. Brandt JR, Joseph MW, Fouser LS, Tarr PI, 165. Siegler RL, Brewer ED, Swartz M. 1988.
Zelikovic I, McDonald RA, Avner ED, McAfee Ocular involvement in hemolytic-uremic syn-
NG, Watkins SL. 1998. Cholelithiasis following drome. J Pediatr 112:594597.
Escherichia coli O157:H7-associated hemolytic 166. Manton N, Smith NM, Byard RW. 2000. Unex-
uremic syndrome. Pediatr Nephrol 12:222225. pected childhood death due to hemolytic uremic
153. Schweighofer S Jr, Primack WA, Slovis TL, syndrome. Am J Forensic Med Pathol 21:9092.
Fleischmann LE, Slovis TL, Hight DW. 1980. 167. Robson WL, Leung AK, Montgomery MD. 1991.
Cholelithiasis associated with the hemolytic- Causes of death in hemolytic uremic syndrome.
uremic syndrome. Am J Dis Child 134:622. Child Nephrol Urol 11:228233.

The Inammatory Response
during Enterohemorrhagic
Escherichia coli Infection
JACLYN S. PEARSON1 and ELIZABETH L. HARTLAND1,2

16
INTRODUCTION
The mammalian host is equipped with two major types of immune response,
innate and adaptive, that are essential for effective control and elimination of
infectious agents. The innate immune system is the first line of host defense
against invading microbial pathogens and is promptly activated by the recog-
nition of pathogen-associated molecular patterns, such as lipopolysaccharide
(LPS), flagellin, peptidoglycan, and CpG DNA (1). Pathogen-associated molec-
ular patterns are recognized by specialized germline-encoded pattern recogni-
tion receptors (PRRs) expressed by immune cells. To date, a number of PRR
families have been described, including the Toll-like receptors (TLRs), retinoic
acid-inducible gene-I-like receptors, and nucleotide-binding oligomerization
domain (NOD)-like receptors (NLRs) (2, 3). The first and most significant
consequence of PRR-mediated pathogen recognition in the host is the rapid
production of proinflammatory cytokines that stimulates the innate immune
response (2, 3). In addition, the innate immune response directs the develop-
ment of the more specific and long-term adaptive response to a particular
pathogen, mediated by B and T cells.
1
Department of Microbiology and Immunology, University of Melbourne at the Peter Doherty Institute for Infection and
Immunity, Victoria 3000, Australia; 2Murdoch Children's Research Institute, Royal Children's Hospital, Parkville, Victoria
3052, Australia.
321

322 PEARSON AND HARTLAND
The primary site of infection with the or, in the case of extracellular pathogens,
attaching and effacing (A/E) pathogens, en- the release of neutrophil extracellular traps.
terohemorrhagic and enteropathogenic Esch- Transmigration of neutrophils from the vas-
erichia coli (EHEC/EPEC), is the epithelium culature to the lamina propria and into
lining the mucosal surface of the gastrointes- epithelial tissues occurs in response to chemo-
tinal tract (4). Not only do these gastrointes- attractive factors released by either resident
tinal epithelial cells play a pivotal role in ion sentinel leukocytes or the epithelium itself
transport, fluid uptake, and secretion that upon recognition of a pathogen (1719). IL-
are critical to the homeostatic state of the 8 is a potent chemoattractant for neutro-
digestive system, they also coordinate the ex- phils, and a number of early studies showed
pression and upregulation of specific anti- that EHEC (20), EPEC (21), and other bacte-
microbial products in response to infection, rial enteric pathogens stimulate epithelium-
including cytokines with proinflammatory derived IL-8 expression during infection
(tumor necrosis factor [TNF] and interleukin-1 (2224). Furthermore, increased IL-8 levels,
[IL-1]) and chemoattractant (IL-8, macro- along with neopterin and IL-10, are markers
phage inflammatory protein 1-alpha [MIP1-a], for increased risk of hemolytic-uremic syn-
monocyte chemoattractant protein-1 [MCP-1]) drome (HUS) in EHEC-infected children.
functions (5). In vitro, the initial and most Despite this, it is not entirely clear to what
potent activation of inflammation by EHEC/ extent epithelium-derived IL-8 contributes
EPEC occurs by TLR5 recognition of flagellin to mucosal defense and the inflammatory re-
and, to a lesser extent, TLR4 recognition of sponse generated against A/E pathogens.
LPS (68). Infection studies conducted in
gnotobiotic piglets show EHEC and EPEC
induce extensive inflammatory cell infiltra- NF-B AND MAPK SIGNALING
tion in the lamina propria as well as trans-
migration of inflammatory cells across the The most crucial factor for initiation of IL-8
intestinal epithelium into the intestinal lumen gene expression in the host is activation of
(911). Depletion of neutrophils in mice in- the key transcriptional regulator of innate
fected with the mouse A/E pathogen Citro- immune signaling, NF-kB (nuclear factor-kB)
bacter rodentium results in elevated bacterial (25). NF-kB proteins are transcription factors
loads in the liver and spleen, suggesting an that control gene expression during inflam-
important role for neutrophils in controlling mation and are activated rapidly in response
infection with A/E pathogens (12). Similarly, to various stimuli, including pathogens, stress
patients with EHEC infection consistently signals, and proinflammatory cytokines such
have significantly increased numbers of leuko- as TNF and IL-1b (26). There are five mem-
cytes in their feces, more so than with Sal- bers of the NF-kB/Rel family of proteins:
monella or Shigella infections (13), whereas p65 (RelA), p50 (NF-kB1), p52 (NF-kB2),
antibodies against the leukocyte adhesion c-Rel, and RelB, which all share an N-terminal
molecule CD18 reduce clinical symptoms Rel homology domain that mediates DNA
and pathology of EHEC infections in rabbits binding, dimerization, and nuclear trans-
(14). location (27, 28). The p65, c-Rel, and RelB
Neutrophils are a type of polymorphonu- subunits harbor an additional C-terminal
clear (PMN) leukocyte that play a key regu- transactivation domain, which strongly acti-
latory role in acute inflammation (1518). vates transcription from NF-kB-binding sites
They represent the first leukocytes recruited in target genes. The p50 and p52 subunits lack
to the site of infection with a pathogen and the transactivation domain but still bind to
can mediate killing by phagocytosis, degran- NF-kB consensus sites and act as transcrip-
ulation (release of antibacterial proteins), tional repressors (29).

CHAPTER 16 The Inammatory Response during EHEC Infection 323
NF-kB activation is generally categorized substrates, a process that can be reversed by

as canonical, noncanonical, or alternative, de- cellular protein phosphatases (37).
pending on the stimuli. Canonical NF-kB sig-
naling is representative of the general scheme
of NF-kB signaling predominantly considered PROINFLAMMATORY RESPONSES
in this article, and is triggered by TLR ligands, DURING EHEC INFECTION
proinflammatory cytokines, pathogens, and
engagement of the T-cell receptor by antigen EHEC produces a number of factors that
(30). Upon ligand recognition, the cognate potentially upregulate inflammatory cytokine
receptor such as TNFR1, IL-1R, or PRRs production by intestinal epithelial cells. For
triggers signaling events that result in activa- example, early studies showed that purified
tion of the IkB kinase (IKK) complex, com- Shiga toxin (Stx) could stimulate low-level
prising IKKb (IKK2), IKKa (IKK1), and production of IL-8 by cultured colon cancer
IKKg (NEMO). Activated IKK in turn phos- cell lines (3840). However, subsequent stud-
phorylates the inhibitory protein of NF-kB ies using primary human colonic epithelial
(IkB), followed by subsequent ubiquitylation cells showed that TLR5 recognition of H7
and degradation of IkB by the host cell flagellin was the major factor inducing IL-8
proteasome (31). In a resting cell IkB is bound production during EHEC infection, and not
to p50/p65 dimers, which are released upon Stx signaling through the Gb3 receptor (7, 41).
proteosomal degradation of IkB and trans- Indeed, in vitro the initial and most potent
ported into the nucleus through the nuclear activation of inflammation by EHEC/EPEC
pore complex where they can promote ex- occurs by TLR5 recognition of flagellin and,
pression of multiple cytokine genes, including to a lesser extent, TLR4 recognition of LPS
IL8 (26). (68). Culture supernatants from the proto-
In addition to NF-kB, the IL-8 promoter type EPEC strain E2348/69, but not from an
region contains binding sites for several other EPEC flagellin (fliC) mutant, induced signifi-
transcription factors, including NF-IL-6, AP-1, cant IL-8 production from cultured colonic
and AP-3 (32). Several studies have demon- epithelial cells (42). In contrast, EHEC O157:
strated AP-1-dependent IL-8 production dur- H7 lacking Stx potently activates p38 and
ing EHEC and EPEC infection in cultured ERK1/2 MAPKs and NF-kB, and induced
intestinal epithelial cells (20, 33, 34). AP-1 significant production of IL-8 by human in-
activation is regulated by mitogen-activated testinal epithelial cells (41). Flagellin/TLR5
protein kinases (MAPK), which are highly signaling thus seems important to induce in-
conserved host proteins that play a central flammatory signaling in vitro. In vivo studies
role in a number of cell responses, including using EHEC O157:H7-infected rabbits showed
regulation of cytokine expression, stress re- a marked increase in neutrophil infiltration in
sponses, and cytoskeletal reorganization (35, the colonic epithelium resulting in diarrhea
36). The MAPKs comprise three subfamilies and disruption of solute transport, none of
of serine/threonine kinases, including the ex- which was dependent on Stx expression (43,
tracellular signal-related protein kinases ERK 44). Although TLR5 is not generally located
and two stress-activated protein kinases, on the apical surface of enterocytes (45), A/E
p38 and JNK (c-Jun amino-terminal kinase). pathogens may gain access to basolateral re-
These kinases regulate cellular processes ceptors by compromising the epithelial bar-
through phosphorylation of target protein rier and cell polarity (46, 47). TLR5 is also
substrates, including other protein kinases, expressed by lamina propria dendritic cells,
phospholipases, transcription factors, and cy- which may sample luminal flagellin (48).
toskeletal proteins. Phosphorylation by MAPKs In addition, some EHEC strains appear to
acts as an on/off switch for the activity of target have a predilection for the follicle-associated

epithelium of the gastrointestinal tract that confirming the importance of EHEC flagellin
expresses TLR5 (49, 50). Hence it is likely that in eliciting an inflammatory response. NF-kB
flagellin is sensed during EHEC infection. and AP-1 activation were increased in the
Despite the fact that Stx alone induces rel- presence of purified HCP and after 3 h of in-
atively low levels of proinflammatory cytokine fection with EHEC O157:H7 strain EDL933 in
production, some studies have suggested that cultured intestinal epithelial cells (54).
the loss of barrier function induced by neu- Overall, the initial induction of proinflam-
trophil transmigration allows increased trans- matory cytokines in the intestinal mucosa
location of Stx from the lumen into the during EHEC infection appears to result from
circulation (51), promoting a systemic cyto- activation of MAPK and NF-kB signaling by
kine response. Indeed, patients with HUS a variety of factors. This in turn drives the
have increased levels of circulating proin- increased production of cytokines by intes-
flammatory cytokines, including IL-8, TNF, tinal epithelial cells and subsequent PMN
IL-6, IL-1b, and IL-10. Although no studies transmigration. Intestinal tissue damage is
to date have been able to detect circulating sustained as a result of the large number of
Stx in patients with toxin-producing EHEC infiltrating PMNs releasing proteolytic en-
infections, others have demonstrated Stx zymes, reactive oxygen intermediates, and
bound to circulating neutrophils in patients phospholipid derivatives, contributing to di-
with HUS (52). A recent study in nonhuman arrheal disease and the release of luminal
primates used purified Stx1 and Stx2 to eval- contents (including Stx) into the circulation.
uate their relative contribution to inflamma-
tory cytokine production. Both toxins induced
a significant increase in expression of IL-8 C. rodentium AND IMMUNE
(CXCL8), MCP-1 (CCL2), and MIP1-a (CCL3) RESPONSES IN MICE
in kidney tissue (53), whereas TNF and IL-
12p35 levels were comparably low. Urine Studying the in vivo inflammatory response
analysis indicated a substantial increase in the intensively in the human host is not a viable
production of IL-6 and vascular endothelial option for EHEC and EPEC infection. There-
growth factor in response to Stx1 compared fore, animal models of infection have been
to Stx2 48 h after challenge, which was con- invaluable in understanding the type of im-
sistent with an increase in leukocyte infiltra- mune response elicited, including which cy-
tion in the kidneys. Although both toxins were tokines and chemokines are released, and how
chemotactic in the kidneys, Stx1 induced a they might contribute to clearance or patho-
more rapid inflammatory response and time genesis. Rabbits, calves, and lambs are the
to euthanasia compared with Stx2 (53). primary animals used to test the character-
The type IV pilus, hemorrhagic coli pilus istics of infection and pathology with EHEC
(HCP), of EHEC O157:H7 is another factor strains, whereas the closely related mouse
capable of inducing an inflammatory response pathogen C. rodentium is also used as a model
in vitro. Production of IL-8 and TNF is sig- of EHEC and EPEC colonization. Besides the
nificantly increased at the basolateral surface advantage of infecting a small laboratory ani-
of polarized T84 and HT-29 and in non- mal, C. rodentium shares most key virulence
polarized HeLa cells during infection with factors of EHEC and EPEC, including a type
HCP-expressing EHEC strains (54). Deletion III secretion system (T3SS) and effector pro-
of the HCP major pilin subunit hcpA sig- teins encoded by the locus of enterocyte
nificantly reduced IL-8 production in cul- effacement (LEE) pathogenicity island and
tured intestinal epithelial cells during EHEC the non-LEE effectors encoded on various
O157:H7 strain EDL933 infection, but not to genomic islands and integrative elements.
the same extent as a fliC/hcpA double mutant, The key virulence factors that C. rodentium

lacks compared to EHEC are genes encoding C. rodentium-infected mice and produces pa-
Stx and flagellin. Despite these limitations, thology similar to that seen in mouse models
C. rodentium infection of mice has allowed of inflammatory bowel disease (IBD) (62
two critically important research questions in 64). Th17 cells are a subset of T-helper cells
the field to be partially addressed, namely, that have an innate function and secrete large
(i) does innate immunity provide protection amounts of the proinflammatory cytokines
against A/E pathogens or does it just induce IL-17 and IL-22. C. rodentium infection in-
tissue damage and (ii) what are the relative duces a Th17 response within 2 weeks of
contributions of innate and adaptive immunity infection, and this is needed for clearance
in controlling infection with A/E pathogens. of the pathogen (64). The induction of this
Infection with C. rodentium causes colitis response requires NOD1/NOD2 signaling and
characterized by inflammatory cell infiltra- the pro-inflammatory cytokine IL-6 (63). IL-
tion, hyperplasia of crypt cells, loss of goblet 6-deficient mice develop severe mucosal ulcer-
cells, and significant intestinal barrier dis- ations and increased crypt cell apoptosis,
ruption (55, 56). Both innate and adaptive suggesting a protective role for the cytokine
responses contribute to the control of C. ro- during gut infection (65). Although it is not
dentium infection, dissemination, and elimi- clear if all these immune and inflammatory
nation. For instance, mice lacking B cells, T factors are required for mucosal defense dur-
cells, or both have greater pathogen loads ing human infection with EHEC, some studies
in their colonic and peripheral tissues, and have highlighted the importance of IL-6 in
despite a severe disease phenotype in these clearance of EPEC in children (66, 67). In
mice, a significant percentage of mice survive, addition, a chronic Th17 response is strongly
suggesting more than an adaptive response is associated with the development of IBD (68).
involved (57, 58). For example, in the absence Studies have suggested that TNF plays an
of neutrophils, mice infected with C. roden- important role in clearance of EPEC in chil-
tium also have a high bacterial load, which dren (67) and control of bacterial load in an-
highlights the importance of the inflammatory imal models of infection (69, 70). TNF is
response in the early stages of infection (12). produced rapidly in response to local or sys-
Activation of NF-kB has been demonstrated in temic tissue damage and is a potent activator
vivo during C. rodentium infection (59); how- of macrophages and neutrophils. Failure to
ever, unlike EHEC and EPEC, C. rodentium control TNF production can lead to severe
is nonflagellated and does not activate TLR5 tissue damage and organ-specific tissue pa-
(60). Therefore, it is likely that activation is thology as a result of chronic activation of
either T3SS-dependent and/or a result of immune cells and inflammatory responses.
TLR4 recognition of LPS. In the intestine, increased TNF levels are
Mice with an intact adaptive system but highly associated with tissue pathology and
deficient in mast cells display increased co- disease, including Crohns disease and IBD
lonic inflammation and increased production (71). Mice that are TNF deficient (TNFRp55/)
of proinflammatory cytokines (61). These mice demonstrate increased tissue pathology during
also experience systemic infection, and most C. rodentium infection, including pronounced
die within 4 to 7 days of C. rodentium infection hyperplasia, increased T-cell infiltrate, and
(61), which suggests that mast cells play a higher levels of mucosal IFN-g and IL-12
role in control and clearance of C. rodentium production (69). However, TNFRp55/ mice
rather than regulating inflammation. are not compromised in their ability to clear
A strong Th1/Th17 response, characterized infection compared to wild-type C57BL/6
by an increase in the expression of IL-1b, mice, and despite the compensatory in-
TNF, IL-12, gamma interferon (IFNg), IL-17, crease in cytokine production, they experi-
and IL-22, is elicited in the colon of ence significantly higher bacterial loads than

wild-type mice (69). Overall, it is plausible proinflammatory cytokines IL-6 and IFNg
that the increased bacterial burden may be during infection, suggesting that the bacteria
a major contributing factor to increased tis- are mediating severe tissue pathology rather
sue pathology in the absence of TNF and that than a dysregulated immune response and
TNF plays a major role in limiting bacterial that IL-1R signaling regulates susceptibility
replication. to C. rodentium infection (78). The impor-
TLRs have been strongly implicated in tance of IL-1R signaling is supported by
mediating inflammation during C. rodentium work showing that mice deficient for the in-
infection. For instance, TLR4, although not flammasome components Nlrc4, Nlrp3, and
protective, mediates chemokine induction and caspase-1 that lead to IL-1b and IL-18 produc-
subsequent neutrophil and macrophage re- tion were highly susceptible to C. rodentium
cruitment and is partially responsible for tis- and had exacerbated intestinal inflammation
sue pathology during C. rodentium infection and increased bacterial load after day 10 of in-
(72). Conversely, TLR2 is not critical in me- fection (79). Similar observations were made
diating proinflammatory responses associated in IL-1b- and IL-18-deficient mice. Despite
with C. rodentium-induced colitis but helps the fact that Nlrc4-deficient mice were highly
maintain mucosal integrity during C. roden- susceptible to C. rodentium infection, only
tium infection (65). MyD88 is a key signaling Nlrp3 and not Nlrc4 induced caspase-1 activa-
adaptor protein that is shared among TLRs tion (79).
(except TLR3), IL-18R, and IL-1R and is es-
sential in the control of mucosal infection by a
number of pathogens (7376). MyD88 activa- INHIBITION OF INFLAMMATORY
tion initiates NF-kB activation and, in the in- SIGNALING BY A/E PATHOGENS
testine, plays an important role in maintaining
tissue homeostasis (77). During C. rodentium Many of the early studies on EHEC/EPEC-
infection MyD88 protects against bacteremia mediated inflammation focused on the proin-
and severe pathology in C57BL/6 mice (12), flammatory response characterized by marked
and MyD88-deficient mice experience ele- infiltration and transmigration of PMNs in the
vated bacterial load in the colon and periph- intestinal epithelium. However, some 10 years
eral tissues, which correlates with a decrease ago, researchers made the remarkable obser-
in neutrophil infiltration into the colonic tis- vation that EHEC and EPEC strains could
sue. MyD88-deficient mice also experience inhibit NF-kB and MAPK activation as well
severe colonic ulceration and bleeding, result- as IkB degradation and the production of pro-
ing in high levels of morbidity and mortality. inflammatory cytokines, including IL-8 and
This severe disease phenotype is due to im- IL-6, early in the course of infection (80,
paired epithelial barrier function and de- 81). Furthermore, this inhibitory mechanism
fective cellular proliferation, followed by an was dependent on the presence of a func-
inability to repair mucosal damage. tional T3SS, revealing for the first time that
MyD88 association with the IL-1R but not EHEC/EPEC suppressed host inflammatory
the IL-18R affords protection against in- responses in a type III-dependent manner
creased mortality and pathology during in- (80). This suggested that EHEC/EPEC had
fection with C. rodentium (78). IL-1R-deficient evolved specialized mechanisms to dampen
mice infected with C. rodentium have dis- the early inflammatory response during infec-
ease similar to that seen in MyD88-deficient tion, possibly to persist for longer periods
mice, but they do not experience higher bac- before the overall immune response would
terial loads or an inability to recruit neutro- inevitably clear the bacteria.
phils and mediate tissue damage repair. The The early discovery of the LEE pathoge-
mice are unable to induce production of nicity island and its association with A/E

lesion formation led to the intensive study LEE-encoded effectors over the past 3 years
of LEE-encoded effectors and their involve- has revealed that A/E pathogens inject mul-
ment in cytoskeletal reorganization and dis- tiple effector proteins into the host cell that
ruption of tight junctions in the epithelium specifically target innate immune factors.
(82). However, the activity of these effec- These effectors have highly specific and di-
tors could not explain how A/E pathogens verse functions that interfere with a range of
suppressed the production of inflammatory innate signaling pathways, including NF-kB
cytokines. Characterization of several non- and MAPK activation (8386) (Fig. 1).
FIGURE 1 Inammatory signaling pathways stimulated and inhibited by EHEC. EHEC products such as agellin
and LPS stimulate TLR signaling and the production of cytokines by epithelial cells during infection. At the
same time, T3SS effector proteins injected by the LEE-encoded translocon inhibit inammatory signaling at
different points in the various pathways. The signaling factors and T3SS effector proteins, and their mecha-
nisms of action, are described in detail in the main text. TGN, trans-Golgi network; ER, endoplasmic reticulum;
Ub, ubiquitin; P, phosphorylation. doi:10.1128/microbiolspec.EHEC-0012-2013.f1

infection, NleE is sufficient to block TNF-

EFFECTORS THAT INHIBIT
induced IkB degradation and nuclear translo-
INFLAMMATORY SIGNALING
cation of activated NF-kB (p65) in epithelial
cells when it is expressed ectopically (84,
The Cysteine Methyltransferase, NleE
85). Furthermore, NleE-dependent inhibition
NleE was one of the first non-LEE-encoded of NF-kB activation contributes significantly
effectors of EHEC implicated in the inhibi- to the low levels of IL8 expression and IL-8
tion of NF-kB signaling (84, 85). This 27-kDa production in EPEC-infected epithelial cells
translocated effector is highly conserved ac- (84, 85). A concurrent study also demon-
ross A/E pathogens and was first discovered strated the ability of NleE to inhibit nuclear
as a type III secreted effector of C. rodentium translocation of activated NF-kB and proin-
(87, 88). In EHEC O157:H7 strain EDL933, flammatory cytokine production in human
NleE is located on the virulence-associated dendritic cells (DCs) (94). IL-8, TNF, and IL-6
O-island (OI) 122, whereas in EPEC O127:H6 expression and production were suppressed
E2348/69, NleE is encoded on genomic in the presence of EPEC-delivered NleE in
pathogenicity island integrative element six human monocyte DCs and Peyers patch DCs.
(IE6) (89, 90). In both cases, NleE is encoded The follicle-associated epithelium is rich in
directly downstream of the type III effector DCs and forms the interface between the lu-
NleB1. Homology at the amino acid level be- minal contents of the intestinal tract and the
tween NleE from EHEC and EPEC is high gut-associated lymphoid tissue. Therefore it
(99% identity) and also between EHEC/EPEC seems plausible that enteric pathogens such as
and C. rodentium (85% identity). Further- EHEC, EPEC, and Shigella sp. would evolve a
more, all Shigella species carry a strong ho- mechanism to evade immune detection at this
molog of NleE known as OspZ, which share junction.
about 74% identity with NleE from EHEC/ The precise mechanism by which NleE in-
EPEC (91). Overall, the high level of similarity hibits NF-kB activation was described recently
between the NleE/OspZ effectors suggests (95). NleE possesses a unique S-adenosyl-L-
that their function is conserved across the methionine-dependent methyltransferase ac-
species. tivity that modifies a cysteine residue in the
Initial in vivo studies revealed no signifi- zinc finger domain of the signaling adaptor
cant colonization defect for a C. rodentium proteins TAB2 and TAB3 (TAK1-binding pro-
DnleE mutant compared to the wild type teins 2 and 3) (95). TAB2/3 contain zinc finger
strain, although the DnleE mutant was out- domains, which bind K63-linked polyubiquitin
competed by the wild type in a competitive chains on the target TNF receptor-associated
infection (88). A subsequent study, however, proteins (TRAF) 2/6 following activation of
argued that NleE was essential for full colo- TNF or TLR/IL-1R signaling. The binding of
nization in mice and that NleE contributed to TAB2/3 to ubiquitinated TRAF allows TAK1 to
disease pathology during C. rodentium infec- form a complex with IKK and subsequently
tion (90). In addition, NleE is consistently as- phosphorylate IKKb, which leads to the deg-
sociated with the virulence profile of O157 and radation of IkB and activation of NF-kB. The
non-O157 EHEC and EPEC strains (92, 93). modification by NleE abolishes ubiquitin-chain
More recent in vitro studies have shown binding of the zinc finger domains of TAB2/3
that NleE from EHEC, EPEC, rabbit specific and thereby disrupts NF-kB signaling in host
EPEC (REPEC), and C. rodentium and full- cells. The activity of NleE depends on a con-
length OspZ from Shigella species inhibit the served 6-amino-acid motif, 209IDSYMK214,
activation of NF-kB in cultured epithelial cells within the C-terminal region that is essential
(84, 85). Despite the translocation of numer- for its ability to block NF-kB activation and
ous type III effectors into the host cell during modify TAB2/3 (85, 95).

histidines or the glutamate within the metal-

The Metalloprotease Effectors,
loprotease motif of NleC renders the protein
NleC and NleD
inactive (9699, 102), a phenomenon that
The non-LEE effectors NleC and NleD are po- had been previously observed for other zinc
tent inhibitors of two key innate signaling metalloproteases (103). Although two inde-
networks through their proteolytic activities. pendent research groups concurrently identi-
While NleC targets NF-kB Rel proteins for fied specific cleavage sites of recombinant p65
degradation, NleD cleaves the MAPK en- by NleC, they were not consistent with each
zymes JNK and p38 (9699). NleC was other (96, 102). Both groups identified the
first identified as an effector secreted by site within the N-terminal Rel homology do-
the LEE-encoded T3SS of C. rodentium (87). main of p65. However, whereas Baruch et al.
The 37-kDa effector is conserved across (96) put the cleavage site between residue C38
A/E pathogens, encoded on OI 36 in EHEC and E39 by N-terminal sequencing, Yen et al.
O157:H7 EDL933 and on prophage four (PP4) (102) proposed the cleavage site to be between
in EPEC O127:H6 E2348/69 (89, 100). EHEC residue P10 and A11. Either way, since the Rel
and EPEC NleC share 100% amino acid se- homology domain of NF-kB proteins is re-
quence similarity and are 95% similar to NleC quired for binding to NF-kB consensus, di-
from C. rodentium. NleD is a 30-kDa effec- merization, and nuclear localization (104),
tor that is also conserved across A/E patho- cleavage by NleC would render p65 unable to
gens and is also encoded on OI 36 in EHEC bind target genes and activate transcription of
O157:H7 EDL933 and on PP4 in EPEC inflammatory cytokines.
O127:H6 E2348/69. EPEC and EHEC NleD To date, there is conflicting evidence on
share 99% amino acid similarity and 77% whether NleC degrades other NF-kB proteins
with NleD from C. rodentium. in addition to p65. Whereas one study showed
Initial studies on NleC and NleD revealed clear degradation of p50 with NleC expressed
that both effectors were translocated into host ectopically or delivered by the T3SS (98), an-
cells by the EPEC T3SS. However, when other group stated NleC could not cleave p50
tested in calves or lambs, DnleC and DnleD directly (102). Similarly, while IkB degradation
deletion mutants of EHEC O157:H7 exhibited was observed upon ectopic expression of NleC
wild-type virulence (101). Similarly, neither (97), another study found that reduced levels
nleC nor nleD mutants of C. rodentium were of IkB in EPEC infected cells could not be
attenuated in mice (88), although mice in- attributed directly to NleC (102). Yet another
fected with an nleC deletion mutant showed study suggested that NleC cleaves the acetyl-
increased pathology, suggesting that NleC transferase p300, a transcriptional coactiva-
assists in reducing the severity of colitis dur- tor of many host cell genes, including p65
ing C. rodentium infection (99). (105, 106). NleC delivered by the T3SS bound
The study by Marchs et al. (2005) (101) to endogenous p300 and reduced host cell
first identified putative zinc metalloprotease nuclear levels of p300 in a metalloprotease-
motifs, HEXXH, within both NleC and NleD dependent manner (106). Furthermore, over-
although no further research on the signifi- expression of p300 in host cells resulted in a
cance of the motif for protein function was significant increase in IL-8 production during
conducted until recently. In fact, NleC de- wild-type EPEC infection, suggesting that NleC-
grades the NF-kB subunit, p65, and this dependent cleavage of p300 assists suppression
function is dependent on the zinc metallo- of IL-8 during infection (106). In addition, NleC
protease motif, 183HEIIH (9699, 102). Deg- was observed to target MAPK signaling by in-
radation is direct, as recombinant His6-NleC hibiting phosphorylation of p38 during EPEC
cleaves both p65 and p50 in epithelial cell infection, although this was independent of the
lysates (98). Furthermore, mutation of the zinc metalloprotease motif (99).

Given the diversity and complexity of in-

The Glycosyl Transferase, NleB
nate immune signaling pathways, it is per-
haps not surprising that A/E pathogens have The non-LEE-encoded effector B (NleB)
evolved mechanisms to target specific path- was first identified as a secreted effector of
ways such as NF-kB signaling at multiple C. rodentium and is conserved across all A/E
levels. Although there is apparent redundancy pathogens (87). EHEC and EPEC carry two
in the functions of NleE and NleC, secreted copies of nleB (nleB1 and nleB2) whereas
IL-8 levels are significantly higher upon in- C. rodentium carries only one that is most
fection with a double nleC/nleE mutant than similar to NleB1. NleB1 from EPEC and EHEC
single nleE or nleC mutants, suggesting that share a high level of amino acid identity
NleC and NleE act synergistically to inhibit (98%) with each other and about 89% amino
IL-8 production (96, 98, 99, 102, 106). acid identity and 96% similarity with NleB
NleD-mediated proteolytic degradation of from C. rodentium. In EHEC O157:H7, the
the MAPKs p38 and JNK also depends on the nleB1 gene is located on genomic OI 122, which
141
HELLH motif of NleD (96). JNK cleavage is consistently present in Stx-producing EHEC
by wild-type EPEC is detected as early as strains that cause outbreaks and severe disease
30 min post infection with the kinase com- (107). NleB2 is located on a separate pathoge-
pletely degraded by 2.5 h post infection. NleD nicity island in both EHEC and EPEC strains
directly cleaves JNK and p38 within a con- and has not yet been implicated in disease
served activation loop present in both signal- epidemiology of either pathotype. NleB2 is
ing kinases. ERK, however, is unaffected by highly related to NleB1 from both EHEC and
the metalloprotease, which is likely explained EPEC strains and shares 60 to 62% amino acid
by the fact that it does not contain the acti- identity and 80% similarity.
vation loop present in p38 and JNK. The Early work in vivo using a signature-tagged
specific cleavage site of JNK as determined by mutagenesis approach revealed that nleB is
N-terminal sequencing occurs after residue essential for full colonization of mice by
P184 within a TPY motif (96). Phosphoryl- C. rodentium, thereby establishing the effector
ation of T183 and Y185 in this motif is required as an important virulence determinant of A/E
for JNK activation; therefore, NleD selectively pathogens (88). Furthermore, a study on the
cleaves the signaling kinase at its site of acti- transmissibility of C. rodentium showed that
vation. As with NleC, mutation of the gluta- NleB contributes significantly to the fitness of
mate within the metalloprotease motif of NleD the pathogen (90). Strong homologs of NleB
renders NleD inactive, and furthermore, al- (termed SseK1, SseK2, and SseK3) exist in a
though not sufficient, NleD contributes to the number of Salmonella spp., where their func-
relief of IL-8 suppression in wild-type EPEC tion and contribution to virulence are still
infection of cultured epithelial cells (96). under investigation (108).
Overall, NleC and NleD are zinc-containing Recent studies in cell signaling have shown
metalloproteases that provide yet another that transient expression of NleB1 in cultured
mechanism by which EHEC and EPEC con- epithelial cells inhibits IkB degradation and
trol the host inflammatory response during NF-kB activation in response to TNF but not
infection. Although the degradation of p65, IL-1b stimulation (85, 96). This suggested that
JNK, and p38 may appear extraneous given the point at which NleB1 interrupted NF-kB
the efficiency of NleE alone to inhibit IL-8 signaling was downstream of TNF-R1 but up-
production, NleC and NleD clearly amplify stream of IKK and TAK1, where the TNF and
this modulatory effect, demonstrating that IL-1 pathways converge. Despite compelling
EHEC and EPEC are highly evolved bacteria evidence for NleB-mediated inhibition of NF-
that specifically target signaling networks at kB activation from TNF-R1, subsequent work
multiple levels. showed that the effector had no significant

role in blocking production of the NF-kB- to the Shigella effector OspG, a Ser/Thr pro-
dependent cytokine, IL-8, during EPEC in- tein kinase that prevents ubiquitylation and
fection (J. S. Pearson, unpublished data). subsequent degradation of phospho-IkBa and
A recent study revealed that NleB of downstream activation of NF-kB (111). Alanine
C. rodentium is a member of the GT-8 family replacement of the conserved lysine residue
of glycosyltransferase proteins and exhibits at position 53 abolishes the kinase activity of
GlcNAc transferase activity (109). NleB1 and OspG (83). NleH1 and NleH2 are also auto-
NleB2 contain a catalytic motif (DxD) that is phosphorylated Ser/Thr protein kinases, but
conserved across all A/E pathogens and the their kinase activity is independent of their
SseK proteins from Salmonella spp. The host interaction with RPS3 (110). NleH1 but not
protein GAPDH was proposed as the target NleH2 reduces the nuclear abundance of
for NleB binding and O-GlcNAc modification. RPS3 with no effect on other NF-kB signaling
Coincidently, the study proposed a role for factors, and in fact, the N-terminal region
GAPDH binding and promoting TRAF2 ubiq- of NleH1 is sufficient to perform this func-
uitylation and subsequent NF-kB activation, tion (83). The mechanism by which NleH1
and thus concluded that GlcNAcylation of executes this through inhibition of IKKb-
GAPDH by NleB hindered TRAF2-mediated mediated phosphorylation of RPS3 at serine
NF-kB activation (109). In this instance, mod- residue 209 was demonstrated during EHEC
ification of GAPDH may explain previous O157:H7 infection in vitro and in vivo (112).
observations that NleB could block TNF- NleH2 does not inhibit RPS3 nuclear trans-
induced IkB degradation and p65 nuclear location, but it does increase expression from
translocation (84, 85); however, this study did an AP-1-dependent luciferase reporter; hence
not examine the cytokine production that it may affect a different signaling pathway
would implicate NleB in dampening the in- (83). Another study suggested that NleH1 and
flammatory response during infection. This NleH2 suppress TNF-induced IkBa degrada-
work did, however, show that a C. rodentium tion in cultured epithelial cells by interfering
nleB mutant complemented with a catalytic with phospho-IkBa ubiquitylation, which is
mutant of NleB (NleBAAA) was attenuated to dependent on conserved lysine residues, K159
similar levels as the nleB mutant, indicating and K169, in NleH1 and NleH2, respectively,
that the DxD motif contributes to the viru- that are implicated in kinase activity (86).
lence of A/E pathogens (109). A number of studies conducted with
animal models have tried to dissect the con-
tribution of NleH to pathogenesis and in-
NleH and the Inhibition of NF-B Signaling
flammatory responses in the host; however,
NleH1 and NleH2 are non-LEE-encoded effec- they have been somewhat inconsistent. C. ro-
tors of A/E pathogens that also contribute to dentium carries only one copy of NleH and is
suppression of NF-kB activation. The effectors functionally most similar to NleH1 of EHEC/
share 84% amino acid identity and are located EPEC (83). Initial work showed that an nleH
on separate genomic islands in both EHEC mutant was attenuated early in infection
and EPEC strains. Studies to date have been (6 days post infection) in C57BL/6 mice, but
varied in defining the mechanism by which not at later time points (10 days post infection)
the NleH effectors interfere with NF-kB sig- (113), plus the mutant strain was cleared more
naling. Initial work showed that EHEC nleH1 rapidly than the wild-type strain (114), sug-
and nleH2 bind to a recently discovered non- gesting the effector plays a role in persistence
Rel NF-kB subunit, ribosomal protein S3 of the pathogen. As for inflammation, one
(RPS3), a K homology domain protein that study showed that wild-type C. rodentium
regulates NF-kB dependent transcription (83, infection in C57BL/6 mice induces higher
110). NleH has significant sequence similarity transcription of TNF mRNA in mouse tissues

than an nleH mutant at 14 days post infection signaling (119) these motifs are critical nega-
(114), whereas a more recent study showed tive regulators of eukaryotic immunoreceptor
the opposite at 7 days post infection (115), signaling pathways (120, 121). Tir was re-
suggesting an anti-inflammatory role for NleH quired for the inhibition of proinflammatory
in vivo earlier in infection. The former study cytokine (TNF and IL-6) expression during
also showed that NleH induced an increase EPEC and C. rodentium infection in vitro and
in NF-kB activity in the colonic mucosa of in vivo, respectively (119). However, the es-
C57BL/6 mice at later time points; however, sential role of Tir in intimate adherence and
there was no overall significant tissue damage hence effector translocation makes it difficult
or increase in T-cell infiltrate in the colons to attribute this property directly to Tir, par-
of mice infected with either wild type or an ticularly as the background strain used was
nleH mutant of C. rodentium over 14 days of JPN15, which has been cured of the plasmid
infection (114). Another study suggested that that encodes bundle-forming pili and already
KC (a functional homolog of human IL-8) adheres less efficiently to cells than wild-type
secretion in streptomycin-treated C57BL/6 EPEC. Likewise, while deletion of tir en-
mice increases upon infection with an EPEC hanced NF-kB, Erk, JNK, and p38 MAPK ac-
nleH1nleH2 double mutant, suggesting that tivation during EPEC infection of cultured
NleH contributes to the inhibition of inflam- epithelial cells (119), this was not distin-
mation in EPEC-infected mice (86). However, guished from the role of Tir in adherence and
it is difficult to reconcile this result with the the translocation of other effectors. Never-
fact that EPEC colonization of mice does not theless, this study provided evidence that Tir
lead to a productive infection (116). binds directly to SHP-1 (119), a phosphatase
Despite intense interest in the mechanism that suppresses cellular immune responses
of action and function of NleH, neither NleH1 by dephosphorylation of NF-kB and MAPKs
nor NleH2 inhibits NF-kB activity to the same (122). Downregulation or mutagenesis of SHP-1
extent as NleE or NleC in vitro (85). Given the increased TNF and IL-6 production in EPEC-
possible additional role of NleH in cell death infected cells, and recruitment of SHP-1 by
signaling (117), the mechanism of action of Tir was essential for its inhibitory function.
NleH and its contribution to EPEC/EHEC Finally, Tir also enhanced SHP-1 association
pathogenesis warrant further investigation. with TRAF6 by an unknown mechanism,
which in turn prevented TRAF6 ubiquitylation
and subsequent signaling cascades.
Another Role for the Translocated
Intimin Receptor (Tir)?
Proinammatory T3SS Effectors
Two studies have identified a role for Tir in
inhibiting host innate signaling mechanisms. Evidence for the proinflammatory activity of
Ruchaud-Sparagano et al. (2012) (118) showed some T3SS effector proteins comes primarily
that transient expression of Tir from EPEC in from the study of genomic island deletion
HeLa cells could inhibit TNF-induced NF-kB mutants. An EPEC mutant lacking genomic
activation by binding and degrading the cyto- islands PP4 and IE6 (encoding seven effectors
plasmic TRAF2 in a proteasome-independent including NleE, NleC, and NleD) induced p65
manner. Consequently, a slight yet significant nuclear translocation and pronounced IL-8
increase in IL-8 production in HeLa cells in- secretion independently of TNF stimulation
fected with an EPEC tir deletion mutant was (85, 98). Nuclear translocation of p65 induced
observed (118). A subsequent study showed by the PP4/IE6 double mutant was greater
that Tir shares sequence similarity with host than for a T3SS mutant (escN), suggesting that
cellular immunoreceptor tyrosine-based in- other translocated effectors stimulated NF-kB
hibition motifs (ITIMs) and inhibits TLR activity (85). Recent work on the WxxxE

effector, EspT, supports the idea that some CONCLUSIONS

effectors directly activate NF-kB signaling
(123). EspT induced IL-8 and IL-1b secretion, Given that the gastrointestinal tract is the
which required the small GTPase Rac1 (123). largest inflammatory organ in the mammalian
A similar phenomenon was evident during body, it is not surprising that EHEC/EPEC
Salmonella sp. infection where activation of and other enteric bacterial pathogens in-
Rac1 and Cdc42 by the WxxxE effector, SopE, cluding Salmonella sp. and Shigella sp. have
also induced NF-kB activation. NF-kB activa- evolved numerous mechanisms for diverting
tion by Rac1 occurred via NOD1 and RIPK2 or arresting inflammatory processes in order
signaling and was also applicable to the sens- to persist and cause disease. Although there
ing of peptidoglycan by NOD1 (124). This is a certain level of redundancy between the
study concluded that NOD1 senses microbial effectors and their functions, each one plays
infection by monitoring the activation state a contributing role in counteracting the in-
of small GTPases, such as Rac1 and Cdc42. flammatory responses of the host. This di-
Because the WxxxE effectors have known versity of effector protein function exemplifies
roles in cytoskeletal rearrangement and bac- how A/E pathogens have evolved to evade
terial invasion, this illustrates the point that numerous host defense mechanisms, espe-
the activity of an effector protein on a host cially innate immunity. Current research
cell protein may inadvertently stimulate an efforts to determine effector function are re-
inflammatory response. vealing more novel enzymatic functions and
targets for effectors and at the same time re-
fining our understanding of host-pathogen
The pO157-Encoded SteC Protease from
interactions.
EHEC Impairs Neutrophil Function
In addition to the type III effectors, a number
ACKNOWLEDGMENT
of other EHEC-related virulence factors are
believed to counteract host inflammatory We declare no conflicts of interest with regard
processes. StcE (secreted protease of Cl- to the manuscript.
esterase inhibitor) is a type II secreted pro-
tease encoded by the pO157 virulence plasmid
CITATION
of EHEC strains that cleaves the protein
backbone of mucin-type glycoproteins (125). Pearson JS, Hartland EL. 2014. The inflam-
Mucins are found on almost all hematopoietic matory response during enterohemorrhagic
cells, including neutrophils, and play critical Escherichia coli infection. Microbiol Spectrum
roles in cellular interactions within the im- 2(4):EHEC-0012-2013.
mune system. CD43 and CD45 are neutrophil-
specific mucins (126, 127) specifically targeted
REFERENCES
by StcE for proteolysis, which leads to in-
creased neutrophil adhesion, impaired mi- 1. Akira S, Uematsu S, Takeuchi O. 2006. Pathogen
recognition and innate immunity. Cell 124:783
gratory capacity, and an enhanced oxidative
801.
burst (128). The impaired neutrophil function 2. ONeill LA, Golenbock D, Bowie AG. 2013. The
may contribute to increased tissue damage history of Toll-like receptorsredefining innate
and inflammation or, conversely, may act to immunity. Nat Rev Immunol 13:453460.
dampen the inflammatory response due to a 3. Latz E, Xiao TS, Stutz A. 2013. Activation
and regulation of the inflammasomes. Nat Rev
lack of neutrophil mobility. Indeed, neutro-
Immunol 13:397411.
phils isolated from children with HUS are 4. Robins-Browne RM, Hartland EL. 2002. Esch-
more adherent to the vascular endothelium erichia coli as a cause of diarrhea. J Gastroenterol
(129). Hepatol 17:467475.

5. Kagnoff MF, Eckmann L. 1997. Epithelial cells as 18. Sadik CD, Kim ND, Luster AD. 2011. Neutrophils
sensors for microbial infection. J Clin Invest cascading their way to inflammation. Trends
100:610. Immunol 32:452460.
6. Badea L, Beatson SA, Kaparakis M, Ferrero RL, 19. Ley K, Laudanna C, Cybulsky MI, Nourshargh
Hartland EL. 2009. Secretion of flagellin by the S. 2007. Getting to the site of inflammation: the
LEE- encoded type III secretion system of en- leukocyte adhesion cascade updated. Nat Rev
teropathogenic Escherichia coli. BMC Microbiol Immunol 7:678689.
9:30. 20. Dahan S, Busuttil V, Imbert V, Peyron
7. Miyamoto Y, Iimura M, Kaper JB, Torres AG, J-F, Rampal P, Czerucka D. 2002. Entero-
Kagnoff MF. 2006. Role of Shiga toxin versus H7 hemorrhagic Escherichia coli infection induces
flagellin in enterohaemorrhagic Escherichia coli interleukin-8 production via activation of mito-
signalling of human colon epithelium in vivo. Cell gen-activated protein kinases and the transcrip-
Microbiol 8:869879. tion factors NF-kB and AP-1 in T84 cells. Infect
8. Schuller S, Lucas M, Kaper JB, Giron JA, Immun 70:23042310.
Phillips AD. 2009. The ex vivo response of hu- 21. Savkovic S, Koutsouris A, Hecht G. 1997. Acti-
man intestinal mucosa to enteropathogenic Esch- vation of NF-kappaB in intestinal epithelial cells
erichia coli infection. Cell Microbiol 11:521530. by enteropathogenic Escherichia coli. Am J
9. Moon HW, Whipp SC, Argenzio RA, Levine Physiol 273:11601167.
MM, Giannella RA. 1983. Attaching and effacing 22. Jung HC, Eckmann L, Yang SK, Panja A, Fierer
activities of rabbit and human enteropathogenic J, Morzycka-Wroblewska E, Kagnoff MF. 1995.
Escherichia coli in pig and rabbit intestines. Infect A distinct array of proinflammatory cytokines is
Immun 41:13401351. expressed in human colon epithelial cells in re-
10. Tzipori S, Gibson R, Montanaro J. 1989. Nature sponse to bacterial invasion. J Clin Invest 95:5565.
and distribution of mucosal lesions associated 23. McCormick BA, Colgan SP, Delp-Archer C,
with enteropathogenic and enterohemorrhagic Miller SI, Madara JL. 1993. Salmonella typhi-
Escherichia coli in piglets and the role of plasmid- murium attachment to human intestinal epithelial
mediated factors. Infect Immun 57:11421150. monolayers: transcellular signalling to subepithe-
11. Tzipori S, Robins-Browne RM, Gonis G, Hayes lial neutrophils. J Cell Biol 123:895907.
J, Withers M, McCartney E. 1985. Entero- 24. Philpott DJ, Yamaoka S, Israel A, Sansonetti
pathogenic Escherichia coli enteritis: evaluation of PJ. 2000. Invasive Shigella flexneri activates NF-
the gnotobiotic piglet as a model of human in- kB through a lipopolysaccharide-dependent innate
fection. Gut 26:570578. intracellular response and leads to IL-8 expression
12. Lebeis SL, Bommarius B, Parkos CA, Sherman in epithelial cells. J Immunol 165:903914.
MA, Kalman D. 2007. TLR signaling mediated 25. Mukaida N, Mahe Y, Matsushima K. 1990. Co-
by MyD88 is required for a protective innate operative interaction of nuclear factor-kappa B-
immune response by neutrophils to Citrobacter and cis-regulatory enhancer binding protein-like
rodentium. J Immunol 179:566577. factor binding elements in activating the inter-
13. Slutsker L, Ries A, Greene KD, Wells JG, leukin-8 gene by pro-inflammatory cytokines.
Hutwagner L, Griffin PM. 1997. Escherichia coli J Biol Chem 265:2112821133.
O157:H7 diarrhea in the United States: clinical 26. Li Q,Verma IM. 2002. NF-kB regulation in the
and epidemiologic features. Ann Intern Med immune system. Nat Rev Immunol 2:725734.
126:505513. 27. Ghosh S, May MJ, Kopp EB. 1998. NF-kB and
14. Elliott E, Li Z, Bell C, Stiel D, Buret AG, Wallace Rel proteins: evolutionarily conserved mediators
J, Brzuszczak I, OLoughlin EV. 1994. Modula- of immune responses. Annu Rev Immunol 16:225
tion of host response to Escherichia coli O157:H7 260.
infection by anti-CD18 antibody in rabbits. Gas- 28. Verma IM, Stevenson JK, Schwarz EM, Van
troenterology 106:15541561. Antwerp D, Miyamoto S. 1995. Rel/NF-kappa B/
15. Amulic B, Cazalet C, Hayes GL, Metzler KD, I kappa B family: intimate tales of association and
Zychlinsky A. 2012. Neutrophil function: from dissociation. Genes Dev 9:27232735.
mechanisms to disease. Ann Rev Immunol 30: 29. May MJ, Ghosh S. 1998. Signal transduction
459489. through NF-kB. Immunol Today 19:8088.
16. Brinkmann V, Reichard U, Goosmann C, Fauler 30. Schulze-Luehrmann J, Ghosh S. 2006. Antigen-
B, Uhlemann Y, Weiss DS, Weinrauch Y, receptor signaling to nuclear factor kB. Immunity
Zychlinsky A. 2004. Neutrophil extracellular 25:701715.
traps kill bacteria. Science 303:15321535. 31. Perkins N. 2007. Integrating cell-signaling path-
17. Phillipson M, Kubes P. 2011. The neutrophil in ways with NF-kB and IKK function. Nat Rev Mol
vascular inflammation. Nat Med 17:13811390. Cell Biol 8:4962.

32. Mukaida N, Okamoto S, Ishikawa Y, 44. Li Z, Bell C, Buret AG, Robins-Browne RM,
Matsushima K. 1994. Molecular mechanism of Stiel D, OLoughlin EV. 1993. The effect of
interleukin-8 gene expression. J Leuk Biol 56: enterohemorrhagic Escherichia coli O157:H7 on
554558. intestinal structure and solute transport in
33. Czerucka D, Dahan S, Mograbi B, Rossi B, rabbits. Gastroenterology 104:467474.
Rampal P. 2001. Implication of mitogen-activated 45. Gewirtz AT, Navas TA, Lyons S, Godowski PJ,
protein kinases in T84 cell responses to entero- Madara JL. 2001. Cutting edge: bacterial flagellin
pathogenic Esherichia coli infection. Infect Immun activates basolaterally expressed TLR5 to in-
69:12981305. duce epithelial proinflammatory gene expression.
34. de Grado M, Rosenberger CM, Gauthier A, J Immunol 167:18821885.
Vallance BA, Finlay BB. 2001. Enteropathogenic 46. McNamara BP, Koutsouris A, O'Connell CB,
Escherichia coli infection induces expression of Nougayrede JP, Donnenberg MS, Hecht G.
the early growth response factor by activating 2001. Translocated EspF protein from entero-
mitogen-activated protein kinase cascades in ep- pathogenic Escherichia coli disrupts host intesti-
ithelial cells. Infect Immun 69:62176224. nal barrier function. J Clin Invest 107:621629.
35. Davis RJ. 1993. The mitogen-activated protein 47. Muza-Moons MM, Koutsouris A, Hecht G.
kinase signal transduction pathway. J Biol Chem 2003. Disruption of cell polarity by enteropatho-
268:1455314556. genic Escherichia coli enables basolateral mem-
36. Davis RJ. 2000. Signal transduction by the JNK brane proteins to migrate apically and to potentiate
group of MAP kinases. Cell 103:239252. physiological consequences. Infect Immun 71:
37. Johnson GL, Lapadat R. 2002. Mitogen-activat- 70697078.
ed protein kinase pathways mediated by ERK, 48. Kinnebrew MA, Buffie CG, Diehl GE, Zenewicz
JNK, and p38 protein kinases. Science 298:1911 LA, Leiner I, Hohl TM, Flavell RA, Littman DR,
1912. Pamer EG. 2012. Interleukin 23 production by
38. Thorpe CM, Hurley BP, Lincicome LL, intestinal CD103(+)CD11b(+) dendritic cells in
Jacewicz MS, Keusch GT, Acheson DWK. 1999. response to bacterial flagellin enhances mucosal
Shiga toxins stimulate secretion of interleukin- innate immune defense. Immunity 36:276287.
8 from intestinal epithelial cells. Infect Immun 49. Phillips A, Frankel G. 2000. Intimin-mediated
67:59855993. tissue specificity in enteropathogenic Escherichia
39. Thorpe CM, Smith WE, Hurley BP, Acheson coli interaction with human intestinal organ
DWK. 2001. Shiga toxins induce, superinduce, cultures. J Infect Dis 181:14961500.
and stabilize a variety of C-X-C chemokine 50. Cashman SB, Morgan JG. 2009. Transcriptional
mRNAs in intestinal epithelial cells, resulting in analysis of Toll-like receptors expression in M
increased chemokine expression. Infect Immun cells. Mol Immunol 47:365372.
69:61406147. 51. Hurley BP, Thorpe CM, Acheson DWK. 2001.
40. Yamasaki C, Natori Y, Zeng XT, Ohmura M, Neutrophil translocation across intestinal epithe-
Yamasaki S, Takeda Y. 1999. Induction of cyto- lial cells is enhanced by neutrophil transmigra-
kines in a human colon epithelial cell line by tion. Infect Immun 69:61486155.
Shiga toxin 1 (Stx1) and Stx2 but not by non-toxic 52. Te Loo DM, Hinsbergh VW, Heuvell LP,
mutant Stx1 which lacks N-glycosidase activity. Monnens LA. 2001. Detection of verotoxin bound
FEBS Lett 442:231234. to circulationg polymorphonuclear leukocytes of
41. Berin MC, Darfeuille-Michaud A, Egan LJ, patients with hemolytic uremic syndrome. J Am
Miyamoto Y, Kagnoff MF. 2002. Role of EHEC Soc Nephrol 12:800806.
O157:H7 virulence factors in the activation of in- 53. Stearns-Kurosawa DJ, Oh S-Y, Cherla RP, Lee
testinal epithelial cell NF-kB and MAP kinase M-S, Tesh VL, Papin J, Henderson J, Kurosawa
pathways and the upregulated expression of in- S. 2013. Distinct renal pathology and a chemo-
terleukin 8. Cell Microbiol 4:635648. tactic phenotype after enterohemorrhagic Esche-
42. Zhou X, Girn JA, Torres AG, Crawford JA, richia coli Shiga toxins in non-human primate
Negrete E, Vogel SN, Kaper JB. 2003. Flagellin models of hemolytic uremic syndrome. Am J
of enteropathogenic Escherichia coli stimulates Pathol 182:12271238.
interleukin-8 production in T84 cells. Infect 54. Ledesma MA, Ochoa SA, Cruz A, Rocha-
Immun 71:21202129. Ramirez LM, Mas-Oliva J, Eslava CA, Giron JA,
43. Elliott E, Li Z, Bell C, Stiel D, Buret A, Wallace Xicohtencatl-Cortes J. 2010. The hemorrhagic
J, Brzuszczak I, OLoughlin E. 1994. Modulation coli pilus (HCP) of Escherichia coli O157:H7 is
of host response to Escherichia coli O157:H7 in- an inducer of proinflammatory cytokine secre-
fection by anti-CD18 antibody in rabbits. Gastro- tion in intestinal epithelial cells. PLoS One 5:
enterology 106:15541561. e12127.

55. Luperchio SA, Schauer DB. 2001. Molecular BA. 2008. Toll-like receptor 2 plays a critical role
pathogenesis of Citrobacter rodentium and trans- in maintaining mucosal integrity during Citro-
missible murine colonic hyperplasia. Microb In- bacter rodentium-induced colitis. Cell Microbiol
fect 3:333340. 10:388403.
56. Ma C, Wickham ME, Guttman JA, Deng 66. Eckmann L. 2006. Animal models of inflamma-
W, Walker J, Madsen KL. 2006. Citrobacter tory bowel disease: lessons from enteric infec-
rodentium infection causes both mitochondrial tions. Ann New York Acad Sci 1072:2838.
dysfunction and intestinal epithelial barrier dis- 67. Long KZ, Rosado JL, Santos JI, Haas M, Al
ruption in vivo: role of mitochondrial associated Mamun A, DuPont HL, Nanthakumar NN,
protein (Map). Cell Microbiol 8:16691686. Estrada-Garcia T. 2010. Associations between
57. Maaser C, Housley MP, Iimura M, Smith JR, mucosal innate and adaptive immune responses
Vallance BA, Finlay BB, Schreiber JR, Varki and resolution of diarrheal pathogen infections.
NM, Kagnoff MF, Eckmann L. 2004. Clearance Infect Immun 78:12211228.
of Citrobacter rodentium requires B cells but not 68. Sarra M, Pallone F, Macdonald TT, Monteleone
secretory immunoglobulin A (IgA) or IgM anti- G. 2010. IL-23/IL-17 axis in IBD. Inflamm Bowel
bodies. Infect Immun 72:33153324. Dis 16:18081813.
58. Simmons CP, Clare S, Ghaem-Maghami M, 69. Gonalves NS, Ghaem-Maghami M, Monteleone
Uren TK, Rankin J, Huett A, Goldin R, Lewis G, Frankel G, Dougan G, Lewis DJ, Simmons
DJ, MacDonald TT, Strugnell RA, Frankel G, CP, MacDonald TT. 2001. Critical role for
Dougan G. 2003. Central role for B lymphocytes tumour necrosis factor alpha in controlling the
and CD4+ T cells in immunity to infection by number of lumenal pathogenic bacteria and im-
the attaching and effacing pathogen Citrobacter munopathology in infectious colitis. Infect Immun
rodentium. Infect Immun 71:50775086. 69:66516659.
59. Wang Y, Xiang GS, Kourouma F, Umar S. 2006. 70. Ramirez K, Huerta R, Oswald E, Garcia-Tovar
Citrobacter rodentium-induced NF-kB activation C, Hernandez JM, Navarro-Garcia F. 2005. Role
in hyperproliferating colonic epithelia: role of of EspA and intimin in expression of proinflam-
p65 (Ser536) phosphorylation. Br J Pharmacol matory cytokines from enterocytes and lympho-
148:814824. cytes by rabbit enteropathogenic Escherichia
60. Khan MA, Bouzari S, Ma C, Rosenberger CM, coli-infected rabbits. Infect Immun 73:103113.
Bergstrom KSB, Gibson DL, Steiner TS, 71. Kollias G, Douni E, Kassiotis G, Kontoyiannis D.
Vallance BA. 2008. Flagellin-dependent and 1999. On the role of tumor necrosis factor and
-independent inflammatory responses. Infect receptors in models of multiorgan failure, rheu-
Immun 76:14101422. matoid arthritis, multiple sclerosis and inflam-
61. Wei OL, Hillard A, Kalman D, Sherman MA. matory bowel disease. Immunol Rev 169:175194.
2005. Mast cells limit systemic bacterial dis- 72. Khan MA, Ma C, Knodler LA, Valdez Y,
semenation but not colitis in response to Citro- Rosenberger CM, Deng W, Finlay BB, Vallance
bacter rodentium. Infect Immun 73:19781985. BA. 2006. Toll-like receptor 4 contributes to co-
62. Higgins LM, Frankel G, Douce G, Dougan G, litis development but not to host defense during
MacDonald TT. 1999. Citrobacter rodentium in- Citrobacter rodentium infection in mice. Infect
fection in mice elicits a mucosal Th1 cytokine Immun 74:25222536.
response and lesions similar to those in murine 73. Hawn TR, Smith KD, Aderem A, Skerrett SJ.
inflammatory bowel disease. Infect Immun 2006. Myeloid differentiation primary response
67:30313039. gene (88)- and toll-like receptor 2-deficient mice
63. Geddes K, Rubino SJ, Magalhaes JG, Streutker are susceptible to infection with aerosolized
C, Le Bourhis L, Cho JH, Robertson SJ, Kim CJ, Legionella pneumophila. J Infect Dis 193:1693
Kaul R, Philpott DJ, Girardin SE. 2011. Identi- 1702.
fication of an innate T helper type 17 response to 74. Skerrett SJ, Liggitt HD, Hajjar AM, Wilson CB.
intestinal bacterial pathogens. Nat Med 17:837 2004. Cutting edge: myeloid differentiation factor
844. 88 is essential for pulmonary host defense against
64. Zheng Y, Valdez PA, Danilenko DM, Hu Y, Sa Pseudomonas aeruginosa but not Staphylococcus
SM, Gong Q, Abbas AR, Modrusan Z, Ghilardi aureus. J Immunol 172:33773381.
N, de Sauvage FJ, Ouyang W. 2008. Interleukin- 75. Watson RO, Novik V, Hofreuter D, Lara-Tejero
22 mediates early host defense against attaching M, Galan JE. 2007. A MyD88-deficient mouse
and effacing bacterial pathogens. Nat Med model reveals a role for Nramp1 in Campylobacter
14:282289. jejuni infection. Infect Immun 75:19942003.
65. Gibson DL, Ma A, Rosenberger CM, Bergstrom 76. Weiss DS, Takeda K, Akira S, Zychlinsky A,
KSB, Valdez Y, Huang JT, Khan MA, Vallance Moreno E. 2005. MyD88, but not Toll-like

receptors 4 and 2, is required for efficient clear- systematic and functional analyses of a pathoge-
ance of Brucella abortus. Infect Immun 73:5137 nicity island. Proc Natl Acad Sci USA 101:3597
5143. 3602.
77. Akira S, Takeda K. 2004. Toll-like receptor sig- 88. Kelly M, Hart E, Mundy R, Marchs O, Wiles S,
nalling. Nat Rev Immunol 4:499511. Badea L, Luck S, Tauschek M, Frankel G,
78. Lebeis SL, Powell KR, Merlin D, Sherman MA, Robins-Browne RM, Hartland EL. 2006. Es-
Kalman D. 2009. Interleukin-1 receptor signaling sential role of the type III secretion system ef-
protects mice from lethal intestinal damage fector NleB in colonization of mice by Citrobacter
caused by the attaching and effacing pathogen rodentium. Infect Immun 74:23282337.
Citrobacter rodentium. Infect Immun 77:604614. 89. Iguchi A, Thomson NR, Ogura Y, Saunders D,
79. Liu Z, Zaki MH, Vogel P, Gurung P, Finlay BB, Ooka T, Henderson IR, Harris D, Asadulghani
Deng W, Lamkanfi M, Kanneganti TD. 2012. M, Kurokawa K, Dean P, Kenny B, Quail MA,
Role of inflammasomes in host defense against Thurston S, Dougan G, Hayashi T, Parkhill J,
Citrobacter rodentium infection. J Biol Chem Frankel G. 2009. Complete genome sequence
287:1695516964. and comparative genome analysis of enteropatho-
80. Hauf N, Charkraborty T. 2003. Suppression genic Escherichia coli O127:H6 strain E2348/69.
of NF-kappaB activation and proinflammatory J Bacteriol 191:347.
cytokine expression by Shiga toxin-producing 90. Wickham ME, Lupp C, Vazquez A,
Escherichia coli. J Immunol 170:20742082. Marscarenhas M, Coburn B, Coombes BK,
81. Ruchaud-Sparagano M, Maresca M, Kenny B. Karmali MA, Puente JL, Deng W, Finlay BB.
2007. Enteropathogenic Escherichia coli (EPEC) 2007. Citrobacter rodentium virulence in mice
inactivate innate immune responses prior to associates with bacterial load and the type III
compromising epithelial barrier function. Cell effector NleE. Microb Infect 9:400407.
Microbiol 9:19091921. 91. Zurawski DV, Mumy KL, Badea L, Prentice JA,
82. Wong ARC, Pearson JS, Bright MD, Munera D, Hartland EL, McCormick BA, Maurelli AT.
Robinson KS, Lee SF, Frankel G, Hartland EL. 2008. The NleE/OspZ family of effector proteins
2011. Enteropathogenic and enterohaemorrhagic is required for polymorphonuclear transepithelial
Escherichia coli: even more subversive elements. migration, a characteristic shared by entero-
Mol Microbiol 80:14201438. pathogenic Escherichia coli and Shigella flexneri
83. Gao X, Wan F, Mateo K, Callegari E, Wang D, infections. Infect Immun 76:369379.
Deng W, Puente J, Li F, Chaussee MS, Finlay 92. Bugarel M, Martin A, Fach P, Beutin L. 2011.
BB, Lenardo MJ, Hardwidge PR. 2009. Bacterial Virulence gene profiling of enterohemorrhagic
effector binding to ribosomal protein S3 subverts (EHEC) and enteropathogenic (EPEC) Esche-
NF-kB function. PLoS Pathog 5:118. richia coli strains: a basis for molecular risk as-
84. Nadler C, Baruch K, Kobi S, Mills E, Haviv G, sessment of typical and atypical EPEC strains.
Farago M, Alkalay I, Bartfeld S, Meyer T, Ben- BMC Microbiol 11:142152.
Neriah Y, Rosenshine I. 2010. The type III se- 93. Buvens G, Pirard D. 2012. Virulence profiling
cretion effector NleE inhibits NF-kB activation. and disease association of verotoxin-producing
PLoS Pathog 6:111. Escherichia coli O157 and non-O157 isolates in
85. Newton HJ, Pearson JS, Badea L, Kelly M, Belgium. Foodborne Path Dis 9:16.
Lucas M, Holloway G, Wagstaff KM, Dunstone 94. Vossenkmper A, Marches O, Fairclough PD,
MA, Sloan J, Whisstock J, Kaper JB, Robins- Warnes G, Stagg AJ, Lindsay JO, Evans PC,
Browne RM, Jans DA, Frankel G, Philips AD, Luong IA, Croft NM, Naik S, Frankel G,
Coulson BS, Hartland EL. 2010. The type III MacDonald TT. 2010. Inhibition of NF-kB sig-
effectors NleE and NleB from enteropathogenic naling in human dentritic cells by the entero-
E. coli and OspZ from Shigella block nuclear pathogenic Escherichia coli protein NleE. J
translocation of NF-kB p65. PLoS Pathog 6:116. Immunol 185:41184127.
86. Royan SV, Jones RM, Koutsouris A, Roxas JL, 95. Zhang L, Ding X, Cui J, Xu H, Chen J, Gong Y-
Falzari K, Weflen AW, Kim A, Bellmeyer A, N, Hu L, Zhou Y, Ge J, Lu Q, Liu L, Chen S,
Turner JR, Neish AS, Rhee K-J, Viswanathan Shao F. 2011. Cysteine methylation disrupts
VK, Hecht GA. 2010. Enteropathogenic E. coli non- ubiquitin-chain sensing in NF-kB activation. Na-
LEE encoded effectors NleH1 and NleH2 attenuate ture 481:204208.
NF-kB activation. Mol Microbiol 78:12321245. 96. Baruch K, Gur-Arie L, Nadler C, Koby S,
87. Deng W, Puente JL, Gruenheid S, Li Y, Vallance Yerushalmi G, Ben-Neriah Y, Yogev O, Shaulian
BA, Vazquez A, Barba J, Ibarra JA, ODonnell E, Guttman C, Zarivach R, Rosenshine I. 2011.
P, Metalnikov P, Ashman K, Lee S, Goode D, Metalloprotease type III effectors that specifically
Pawson T, Finlay BB. 2004. Dissecting virulence: cleave JNK and NF-kB. EMBO J 30:221231.

97. Mhlen S, Ruchaud-Sparagano M-H, Kenny B. seropathotypes that are linked to epidemic
2011. Proteasome-independent degradation of and/or serious disease. J Clin Microbiol 41:4930
canonical NFkB complex components by the 4940.
NleC protein of pathogenic Escherichia coli. J Biol 108. Brown NF, Coombes BK, Bishop JL, Wickham
Chem 286:51005107. ME, Lowden MJ, Gal-Mor O, Goode DL, Boyle
98. Pearson JS, Riedmaier P, Marches O, Frankel EC, Sanderson KL, Finlay BB. 2011. Salmonella
G, Hartland EL. 2011. A type III effector protease phage ST64B encodes a member of the SseK/
NleC from enteropathogenic Escherichia coli NleB effector family. PLoS One 6:e17824.
targets NF-kB for degradation. Mol Microbiol 109. Gao X, Wang X, Pham TH, Feuerbacher LA,
80:219230. Lubos M-L, Huang M, Olsen R, Mushegian A,
99. Sham HP, Shames SR, Croxen MA, Ma C, Chan Slawson C, Hardwidge PR. 2013. NleB, a bacte-
JM, Khan MA, Wickham M, Deng W, Finlay rial effector with glycosyltransferase activity,
BB, Vallance BA. 2011. Attaching and effacing targets GADPH function to inhibit NF-kB activa-
bacterial effector NleC suppresses epithelial in- tion. Cell Host Microb 13:8799.
flammatory responses by inhibiting NF-kB and 110. Wan F, Anderson D, Barnitz R, Snow A, Bidere
p38 mitogen-activated protein kinase activation. N, Zheng L, Hegde V, Lam L, Staudt L, Levens
Infect Immun 79:35523562. D, Deutsch W, Lenardo M. 2007. Ribosomal
100. Perna NT, Plunkett G, Burland V, Mau B, protein S3: a KH domain subunit in NF-kB com-
Glasner JD, Rose DJ, Mayhew GF, Evans PS, plexes that mediates selective gene regulation.
Gregor J, Kirkpatrick HA, Posfai G, Hackett J, Cell 131:927939.
Klink S, Boutin A, Shao Y, Miller L, Grotbeck EJ, 111. Kim DW, Lenzen G, Page AL, Legrain P,
Davis NW, Lim A, Dimalanta ET, Potamousis Sansonetti PJ, Parsot C. 2005. The Shigella
KD, Apodaca J, Anantharaman TS, Lin J, Yen G, flexneri effector OspG interferes with innate
Schwartz DC, Welch RA, Blattner FR. 2001. immune responses by tagretting ubiquitin-
Genome sequence of enterohaemorrhagic Esche- conjugating enzymes. Proc Natl Acad Sci USA
richia coli O157:H7. Nature 409:529533. 102:1404614051.
101. Marchs O, Wiles S, Dziva F, La Ragione RM, 112. Wan F, Weaver A, Gao X, Bern M, Hardwidge
Schller S, Best A, Phillips AD, Hartland EL, PR, Lenardo MJ. 2011. IKKb phosphorylation
Woodward MJ, Stevens MP, Frankel G. 2005. regulates RPS3 nuclear translocation and NF-kB
Characterization of two non-locus of enterocyte function during infection with Escherichia coli
effacement-encoded type III-translocated effectors, strain O157:H7. Nat Immunol 12:335343.
NleC and NleD, in attaching and effacing patho- 113. Garcia-Angulo VA, Deng W, Thomas NA,
gens. Infect Immun 73:84118417. Finlay BB, Puente JL. 2008. Regulation of ex-
102. Yen H, Ooka T, Iguchi A, Hayashi T, Sugimoto pression and secretion of NleH, a new non-
N, Tobe T. 2010. NleC, a type III secretion pro- locus of enterocyte effacement-encoded effector
tease, compromises NF-kappaB activation by in Citrobacter rodentium. J Bacteriol 190:2388
targeting p65/RelA. PLoS Pathog 6:e1001231. 2399.
103. Jongeneel CV, Bouvier J, Bairoch A. 1989. 114. Hemrajani C, Marches O, Wiles S, Girard F,
A unique signature identifies a family of zinc- Dennis A, Dziva F, Best A, Phillips AD, Berger
dependent metallopeptidases. FEBS Lett. 242:211 CN, Mousnier A, Crepin VF, Kruidenier L,
214. Woodward MJ, Stevens MP, La Ragione RM,
104. Gilmore TD, Wolenski FS. 2012. NF-kB: where MacDonald TT, Frankel G. 2008. Role of NleH, a
did it come from and why? Immunol Rev 246:14 type III secreted effector from attaching and ef-
35. facing pathogens, in colonization of the bovine,
105. Hayden MS, Ghosh S. 2008. Shared principles in ovine, and murine gut. Infect Immun 76:4804
NF-kB signaling. Cell 132:344362. 4813.
106. Shames SR, Bhavsar AP, Croxen MA, Law RJ, 115. Pham TH, Gao X, Tsai K, Olsen R, Wan F,
Mak SHC, Deng W, Li Y, Bidshari R, de Hoog Hardwidge PR. 2012. Functional differences and
CL, Foster LJ, Finlay BB. 2011. The pathogenic interactiions between the E. coli type III secretion
Escherichia coli type III secreted protease NleC system effectors NleH1 and NleH2. Infect Immun
degrades the host acetyltransferase p300. Cell 80:21332140.
Microbiol 13:15421557. 116. Mundy R, Girard F, FitzGerald AJ, Frankel G.
107. Karmali MA, Mascarenhas M, Shen S, Ziebell 2006. Comparison of colonization dynamics and
K, Johnson S, Reid-Smith R, Isaac-Renton J, pathology of mice infected with enteropathogenic
Clark C, Rahn K, Kaper JB. 2003. Association Escherichia coli, enterohaemorrhagic E. coli and
of genomic O island 122 of Escherichia coli EDL Citrobacter rodentium. FEMS Microbiol Lett
933 with verocytotoxin-producing Escherichia coli 265:12632.

117. Hemrajani C, Berger CN, Robinson KS, 123. Raymond B, Crepin VF, Collins JW, Frankel G.
Marchs O, Mousnier A, Frankel G. 2010. NleH 2011. The WxxxE effector EspT triggers expres-
effectors interact with Bax inhibitor-1 to block sion of immune mediators in an Erk/JNK and
apoptosis during enteropathogenic Escherichia NF-kappaB-dependent manner. Cell Microbiol
coli infection. Proc Natl Acad Sci USA 107:3129 13:18811893.
3134. 124. Keestra AM, Winter MG, Auburger JJ, Frassle
118. Ruchaud-Sparagano M-H, Mhlen S, Dean P, SP, Xavier MN, Winter SE, Kim A, Poon V,
Kenny B. 2011. The enteropathogenic E. coli Ravesloot MM, Waldenmaier JF, Tsolis RM,
(EPEC) Tir effector inhibits NF-kB activity by Eigenheer RA, Baumler AJ. 2013. Manipulation
targeting TNFa receptor-associated factors. PLoS of small Rho GTPases is a pathogen-induced
Pathog 7:114. process detected by NOD1. Nature 496:233237.
119. Yan D, Wang X, Luo L, Cao X, Ge B. 2012. In- 125. Lathem WW, Grys TE, Witowski SE, Torres
hibition of TLR signaling by a bacterial protein AG, Kaper JB, Tarr PI, Welch RA. 2002. StcE, a
containing immunoreceptor tyrosine-based in- metalloprotease secreted by Escherichia coli O157:
hibitory motifs. Nat Immunol 13:10631071. H7, specifically cleaves Cl esterase inhibitor. Mol
120. Barrow AD, Trowsdale J. 2006. You say ITAM Microbiol. 45:277288.
and I say ITIM, lets call the whole thing off: the 126. Hermiston ML, Xu Z, Weiss A. 2003. CD45: a
ambiguity of immunoreceptor signaling. Eur J critical regulator of signaling thresholds in im-
Immunol 36:16461653. mune cells. Annu Rev Immunol 21:107137.
121. Daeron M, Jaeger S, Du Pasquier L, Vivier E. 127. Ostberg J, Barth RK, Frelinger JG. 1998. The
2008. Immunoreceptor tyrosine-based inhibition Roman god Janus: a paradigm for the function of
motifs: a quest in the past and future. Immunol CD43. Immunol Today 19:546550.
Rev 224:1143. 128. Szabady RL, Lokuta MA, Walters KB,
122. Nandan D, Lo R, Reiner NE. 1999. Activa- Huttenlocher A, Welch RA. 2009. Modulation of
tion of phosphotyrosine phosphatase activity neutrophil function by a secreted mucinase of
attenuates mitogen-activated protein kinase sig- Escherichia coli O157:H7. PLoS Pathog 5:e1000320.
naling and inhibits c-FOS and nitric oxide syn- 129. Forsyth KD, Simpson AC, Fitzpatrick MM,
thase expression in macrophages infected with Barratt TM, Levinsky RJ. 1989. Neutrophil-
Leishmania donovani. Infect Immun 67:4055 mediated endothelial injury in haemolytic uremic
4063. syndrome. Lancet 2:411414.

New Therapeutic Developments
against Shiga Toxin-Producing
Escherichia coli
ANGELA R. MELTON-CELSA1 and ALISON D. OBRIEN1

17
BACKGROUND
Shiga toxin (Stx)-producing Escherichia coli (STEC) colonizes the intestine and
causes hemorrhagic colitis. STEC encodes a variety of colonization factors, but
a significant subset of STEC, the enterohemorrhagic E. coli (EHEC) strains,
have the locus of enterocyte effacement (LEE), the products of which allow
the bacteria to intimately adhere to and form attaching and effacing lesions
on intestinal tissue. The O157:H7 strains, which are responsible for the major-
ity of large outbreaks due to STEC infection, are members of the EHEC
group. All STEC strains make one or more Stxs; these pathogens may pro-
duce two immunologically distinct but highly similar Stxs, Stx1 and Stx2.
These toxins are briefly described in the section on therapeutics targeted to
the Stxs.
Some individuals infected with STEC manifest a serious sequela called he-
molytic uremic syndrome (HUS), a thrombotic microangiopathy defined by
the presence of hemolytic anemia, thrombocytopenia, and renal failure. The
initial insult leading to the development of the thrombotic microangiopathy
is damage by the Stx(s) to vascular endothelial cells that express the toxin
1
Department of Microbiology and Immunology, Uniformed Services University of the Health Sciences, Bethesda,
MD 20814.
341

342 MELTON-CELSA AND OBRIEN
receptor, globotriaosylceramide (Gb3). The of toxin at one time as the bacteria diea
Stx-mediated injury to endothelial cells ini- result that would be bad for a patient; (iii) in
tiates a cascade of events that lead to the some STEC strains, the Stxs are either chro-
activation of platelets and the formation of mosomally encoded or are associated with
thrombi in the small vessels of the kidney and defective bacteriophages, and thus antibi-
sometimes the central nervous system. Pre- otic treatment may not alter Stx expression;
vention of HUS is important as HUS can lead (iv) antibiotics may be used only in the most
to death or long-term consequences such as ill patients, a fact that may skew the apparent
hypertension and renal disease (13). Because risk for HUS development; and (v) published
STEC strains are intestinal bacterial patho- studies on the risk of HUS associated with
gens, the inclination by physicians is to treat antibiotic treatment used different treatments
with antibiotics to eliminate the organisms administered at different stages in the disease
from the gut. However, several studies show process. In addition, although several studies
that treatment of STEC-infected patients with demonstrate an increased risk for HUS after
certain antibiotics may lead to an increase in treatment with antibiotics (8, 9), others do not
HUS (antibiotic use is discussed further in the find an increased risk (10). Conversely, several
next section). Therefore, the focus for thera- studies indicate that the use of b-lactam anti-
peutics against STEC and HUS has been to biotics, particularly in the first 2 to 3 days of
find (i) compounds that act at the level of the illness, and perhaps associated with age <13
bacterium but do not cause an increase in Stx years, is correlated with increased HUS risk
production; (ii) receptor mimics or other (11). In contrast, many people were treated
molecules that alter trafficking of the toxin or with fosfomycin during a large O157:H7 out-
the cellular response to the toxin; (iii) anti- break in Japan without an apparent increase
bodies directed against the Stxs; or (iv) ther- in HUS, and perhaps even a protective effect
apies to prevent or treat the HUS disease if the antibiotic was given within the first 2
process. The therapies discussed in this article days of illness (12). However, the low overall
are listed in Table 1. HUS rate in the Sakai outbreak (13) suggests
that the causative strain may (fortunately)
have had reduced virulence compared with
THE DIFFICULTIES WITH ANTIBIOTIC USE O157 strains from other outbreaks, and, as
FOR STEC INFECTIONS such, would make that outbreak a poor plat-
form on which to make generalized decisions
In the United States, antibiotics are not re- about treatment with antibiotics.
commended for treatment of STEC infections During the O104:H4 outbreak in Germany
because of the increased risk for the devel- in 2011 caused by an unusual enteroaggrega-
opment of HUS (4, 5). However, even though tive E. coli (EAEC) strain that makes Stx2,
antibiotics are contraindicated for those with antibiotics were used in many patients, some
STEC infection, a recent study at FoodNet of whom were given up to three antibiotics at
sites found that antibiotics are commonly used various times after disease onset (14). Because
in those with proven O157 infection (6). The of the vastly different protocols used to treat
issue of the use of antibiotics to treat STEC patients during the outbreak in Germany, it
infections is confounded by a number of fac- is difficult to make generalized conclusions
tors: (i) some, but not all, antibiotics at sub- about antibiotic use based on that outbreak;
lethal doses increase the expression of Stx however, overall there was not definitive evi-
by inducing the lysogenic phage that encodes dence of a benefit due to antibiotic treatment.
the toxin genes (see diagram in Fig. 1) (7); Although one surprising study asserted cip-
(ii) treatment of STEC with a lethal dose of rofloxacin treatment reduced the HUS in-
antibiotics may cause release of a large bolus cidence in patients, the number of treated

CHAPTER 17 New Therapeutic Developments against STEC 343
patients was small (n = 5), and in two of those COMPOUNDS DIRECTED TOWARD STEC
patients HUS did develop (15). Furthermore,
in several patients from the O104:H4 out- Pyocin
break HUS was reported to develop after
A novel strategy to kill O157 strains based on a
ciprofloxacin or metronidazole treatment (16,
modified pyocin (a bactericidal protein, simi-
17). We consider the use of ciprofloxacin to
lar to bacteriophage tail fibers, made by Pseu-
be especially dangerous because of the evi-
domonas aeruginosa that usually targets other
dence that ciprofloxacin increases Stx ex-
P. aeruginosa cells) that consists of the R2 tail
pression in the EAEC O104:H4 strain (18) and
fiber fused to a phage spike protein specific
STEC O157 strains. Finally, one unique aspect
for O157 was found to kill O157 strains without
of the outbreak in Germany was the use of
increasing expression of Stx (26). Subsequent
azithromycin to reduce shedding of O104:H4
tests demonstrated that a slightly modified
in patients who appeared to be long-term
version of the pyocin, AvR2-V10.3, given 3 h
carriers (19).
after infection and once daily for the next
Animal studies also give contradictory in-
2 days to infant rabbits infected with O157:H7
formation about the use of antibiotics for the
strain EDL933 reduced colonization and pre-
treatment of STEC infections. In two studies
vented diarrhea at the highest dose of pyocin
fosfomycin reduced colonization and mortal-
administered (27). When the modified pyocin
ity from STEC infection in mice (20, 21), but
was given to EDL933-infected infant rabbits
another study did not show a protective effect
that exhibited diarrhea, the amount of loose
by that antibiotic (22). In a mouse model in
stool decreased relative to the control ani-
which the mice are starved for protein calories,
mals given buffer alone and the numbers of
the use of trimethoprim-sulfamethoxazole was
EDL933 organisms decreased in the intestines
detrimental when given between days 3 and
and stools. One problem with the pyocin ap-
5 post infection, whereas the same antibiotic
proach is that it is specific for serogroup, so
as well as ampicillin and fosfomycin were
other pyocins would have to be developed for
protective when given anywhere from days
non-O157 STEC strains.
1 to 5 after infection (23). Nevertheless, we
should note that the use of ampicillin to
treat an infection with an organism (Shigella
A Small Molecule and a Divalent Cation
dysenteriae or the Stx2+ EAEC O104:H4, re-
spectively) resistant to that antibiotic ap- Another alternative strategy to antibiotics di-
pears to be detrimental to humans and mice rected toward the bacterium uses a small
(24, 25). molecule, LED209, which inhibits the activity
Overall, the current data do not support the of the QseC sensor that is involved in regu-
use of antibiotics for the treatment of STEC lating some of the proteins involved in EHEC
infection in humans, and the majority of the adherence and, indirectly, Stx2 expression.
evidence indicates that antibiotic treatment is Although LED209 inhibited the formation of
potentially detrimental. The empirical use of attaching and effacing lesions by EHEC on
antibiotics could be dangerous, since some HeLA cells, the compound was not effective
of these treatments could increase the risk of in reducing colonization or disease in infant
HUS. Furthermore, since there is no clear rabbits (28). The reason for LED209 failing
demonstrable benefit to the use of antibiotics to protect in the rabbits may be that the
in STEC-infected patients, and because anti- concentration of the compound was not high
biotic use itself may pose possible risk to the enough at the site of EHEC colonization in
patient (allergy or other risks), we believe the gut.
antibiotics should not be used to treat STEC Similar to LED209, treatment of STEC in-
infections. fection with the divalent cation zinc reduces

344
Chap17.proof.3d
344
TABLE 1 Therapies directed against STEC, the Stx receptor, Stx function, or the cellular response to Stx
Model used to test
MELTON-CELSA AND OBRIEN
Targeta Therapeutic Function function or efcacy Reference(s)
STEC or STEC Pyocin (AvR2-V10.3) Kills O157 In vitro growth; infant rabbits 26, 27
pathogenesis LED209 Reduces attaching and effacing lesions HeLa cells; infant rabbits 28
on HeLa cells but did not reduce
colonization or disease in rabbits
Zinc Reduces stx transcription; HeLa cells; rabbit ileal loops (using rabbit 29
reduces adherence enteropathogenic E. coli transduced with
phage carrying stx1 or stx2)
Receptor (Gb3) C-9 Receptor generation; HRTEC; Stx2-intoxicated rats 36, 37
synthesis inhibitor glucosylceramide synthase

Receptor analogs (B) SYNSORB Pk Receptor analog Phase I and II trials 39, 41
STARFISH Receptor analog Vero cells; Stx1-intoxicated mice 42, 43
(does not protect Stx2-intoxicated mice)
Daisy Receptor analog Vero cells; Stx1- and Stx2-intoxicated mice; 43
B2F1-infected, streptomycin-treated mice
SUPER TWIG (1)6 Receptor analog Vero cells; Stx1- and Stx2-intoxicated mice; 44
O157-infected, protein-calorie-decient mice
Gb3 polymers Receptor analog Mice 45
TVP (also known Receptor analog, binds Stx2 Vero cells; O157-infected, protein-calorie-decient 46, 49, 50
as Ac-PPP-tet) but not Stx1 mice, rabbit ileal loops; baboon
MMA-tet Receptor analog Vero cells; O157-infected, protein-calorie-decient mice 51
Probiotic that displays Receptor analog Vero cells; streptomycin-treated mice infected with 47, 5254
Gal1-4Gal1-4Glc- an Stx2 or Stx2dact producer
HuSAP Binds Stx2 but not Stx1 Stx1- or Stx2-intoxicated mice 5759
(only protects Stx2-intoxicated animals)
03/30/15 18:03
Model used to test
Targeta Therapeutic Function function or efcacy Reference(s)
Antitoxin antibodies Polyclonal anti-Stx2 Neutralize Stx2 Gnotobiotic pigs 62
(B, U, T, or E) cStx1 (human/mouse Block Stx1 binding (B subunit) Vero cells; Stx1-intoxicated mice; phase I and II trials 64, 68, 70
chimeric version of 13C4)
cStx2 (anti-A subunit, Alters intracellular trafcking and Vero cells; streptomycin-treated mice; 63, 6870
Chap17.proof.3d
human/mouse chimeric inhibits enzymatic function phase I and II trials
345
of 11E10)
Anti-Stx1 (5-5B) Neutralize Stx1 (B subunit) Ramos cells 71
Urtoxazumab (also called Neutralize Stx2 (B subunit) Human renal adenocarcinoma cells; 7375
TMA-15; humanized version Stx2-intoxicated mice; B2F1-infected
of VTm1.1) streptomycin-treated mice, phase I trial
Anti-Stx1 monoclonals Neutralize Stx1 (most are anti-B subunit) HeLa cells; Stx1-intoxicated mice 76
Anti-Stx2 (5C12) Neutralize Stx2 (A subunit) HeLa cells; Stx2-intoxicated mice; B2F1-infected 77, 78
streptomycin-treated mice; O157-infected
gnotobiotic piglets
Toxin transport (U or T) Exo2 Blocks transport at the level of the Vero cells 82, 83
early endosome/trans-Golgi interface
Retro-2cycl Toxin transport HeLa cells 85, 86
Chloroquine Toxin transport HEp-2 cells 87
Small molecules Toxin transport HeLa cells 88
Eeyarestatin 1 Intracellular trafcking HeLa cells 90
Nitrobenzyl-thioinosine Stx1 trafcking (retention in early Human renal cortical epithelial cells 91
endosomes)

Manganese Protects HeLa cells and BALB/c mice from HeLa cells; Stx1-intoxicated BALB/c mice 92, 93
Stx1-mediated lethality; blocks StxB1 (does not protect Stx2-intoxicated mice)
trafcking; does not protect against Stx2
Toxin processing (N) Furin inhibitor Cleavage of A subunit to A1 and A2 HEp-2 cells 94
Toxin function (E) Small molecule Enzymatic activity Cell-free reporter assay 89
Cellular or host response Imatinib Ribotoxic stress response inhibitor HCT-8 cells; infant rabbit 98
to toxin (ribotoxic stress Ouabain Apoptosis inhibition Rat proximal tubule cells (RPTC) 99
response or apoptosis)
Complement factor C5 Eculizumab Anticomplement factor C5 STEC-infected patients 101, 105, 106
a
Step in Fig. 2 at which therapeutic acts.
CHAPTER 17 New Therapeutic Developments against STEC
345
03/30/15 18:03
FIGURE 1 Induction of the stx-encoding phage by antibiotics such as ciprooxacin. Subinhibitory concentra-
tions of some antibiotics induce the lysogenic stx phage to enter the lytic cycle, and as consequence, stx-
bacteriophage are made and released. Furthermore, 10- to 100-fold more toxin is made and released from
the cell. doi:10.1128/microbiolspec.EHEC-0013-2013.f1
adherence to a HeLa cell monolayer, though binding, uptake, trafficking, or function are
through a different mechanism, perhaps due strong contenders for treatment of STEC
to decreased expression of E. coli-secreted infections. Stx1 and Stx2 share about 56%
proteins EspA and EspB (29). In addition to a homology at the amino acid level but are im-
reduction in adherence to cells, zinc treatment munologically distinct. Structurally, the toxins
also appears to decrease expression of the stx consist of five identical B subunits and a single
genes. The potential therapeutic effect of zinc A subunit. The B pentamer binds to the cel-
was tested in a rabbit ileal loop model, in lular receptor, Gb3. The A subunit contains
which rabbit enteropathogenic E. coli strains the enzymatic function of the toxin and re-
transduced with a bacteriophage that encodes moves an adenine residue from the 28S rRNA,
stx1 or stx2 were inoculated into the loop in an action that destroys ribosome function.
the presence or absence of zinc. The presence After binding to Gb3, the toxin receptor com-
of 1 mM of zinc reduced the amount of toxin plex is taken up by clathrin-dependent and
found in the loops, decreased adherence, and -independent mechanisms. The toxin then
lessened histological damage. The next avenue traffics in a retrograde direction from the en-
to explore for zinc is to determine if the cation dosome to the Golgi apparatus to the endo-
could prevent disease or mortality in an oral plasmic reticulum (ER). The A subunit of the
infection model. toxin can be nicked within the Golgi body
such that the enzymatic function (within A1)
is separated by a disulfide bond from the A2
THERAPEUTICS THAT INTERFERE peptide that links A1 to the B pentamer. The
WITH TOXIN BINDING, UPTAKE, nicked toxin traffics next to the ER where the
TRAFFICKING, OR FUNCTION disulfide bond between A1 and A2 is reduced.
The A1 subunit then enters the cytoplasm and
Stx1 and Stx2 are the key STEC virulence targets the ribosome. The damage to the ri-
factors that lead to the development of HUS. bosome halts protein synthesis and can lead
Therefore, therapeutics that impede toxin to a ribotoxic stress response and apoptosis

(see review in reference 30). A model of Stx Stx2a (34). Stx2d has two amino acid differ-
trafficking is shown in Fig. 2. ences from Stx2a and Stx2c in the A subunit,
Stx1 and Stx2 are the major toxin types and those two changes contribute to the ca-
produced by STEC that cause human disease. pacity of the toxin to exhibit increased toxicity
There are subtypes of both Stx1 and Stx2, when treated with elastase from intestinal
however, and toxin nomenclature was re- mucus (35), a phenotype indicated with the
cently updated so that the prototype Stxs are designation Stx2dact.
now designated Stx1a and Stx2a. The less
specific Stx1 and Stx2 designations are used
Receptor Synthesis Inhibitor: C-9
when the toxin subtype is unknown (31). We
use the more general designations for the One step at which to stop Stx is at the level of
toxins elsewhere in this review for simplicity. the toxin receptor. Various laboratories have
Besides the prototypic Stx2a, two Stx2 sub- tried to protect cells or animals from Stxs with
types are associated with HUS, Stx2c and compounds that prevent Gb3 synthesis or that
Stx2d (32, 33). Stx2c and Stx2d have two bind to the toxin to prevent toxin-receptor
amino acid differences in the B subunit as interaction. For example, a molecule called
compared to Stx2a, and those changes are C-9 was used in human renal tubular epithe-
responsible for reduced cytotoxicity on Vero lial cells (HRTEC) to prevent the conversion
cells, though Stx2d is just as toxic to mice as of ceramide to glucosylceramide, an early step
FIGURE 2 Cellular trafcking of Stx and points in the pathway where therapeutics function. Therapeutics can
interfere with Stx action at several points as it trafcs into and through the cell. The steps in toxin trafcking
are briey diagramed above and outlined here. The toxin rst binds (B) to the receptor Gb3. The toxin/Gb3
complex is taken up (U) by both clathrin-dependent and -independent mechanisms, and then trafcs (T) from
the early endosome to the late endosome to the trans-Golgi apparatus. Within the Golgi the toxin A subunit is
nicked (N), but the toxin remains intact due to a disulde bond between the A1 and A2 subunits. The nicked
toxin continues to trafc (T) along the retrograde pathway to the endoplasmic reticulum. The disulde bond in
the A subunit is reduced (R) within the endoplasmic reticulum and the A1 subunit enters the cytoplasm where it
exerts its enzymatic (E) attack and depurinates the ribosome. The action of the toxin within the cell can lead to
a ribotoxic stress response (RSR) and apoptosis (APOP). doi:10.1128/microbiolspec.EHEC-0013-2013.f2

in Gb3 synthesis. HRTEC pretreated with C-9 Stx1 but not Stx2 (43). Because of the lack of
for 24 h exhibited reduced levels of Gb3 protective efficacy by STARFISH in the Stx2-
and decreased sensitivity to Stx2 (36). In rats injection model, a similar but modified diva-
treated with C-9 for 2 days before intoxication lent trisaccharide inhibitor called Daisy was
and for 4 days post intoxication, about 50% synthesized (43). Daisy protects Vero cells
protection from injection with bacterial super- and mice from both Stx1 and Stx2, and in a
natants that contained Stx2 was observed (37). streptomycin-treated mouse model, prevented
Treated rats also showed lower rises in serum death due to infection with O91:H21 STEC
creatinine and urea and reduced renal tubular strain B2F1 in 50% of the animals (43). The
injury. SUPER TWIG (1)6 identified by Nishikawas
group carries six trisaccharides and blocks
the binding of both Stx1 and Stx2 to Vero cells,
Receptor Analogs
protects mice from Stx2 when administered
The first drug tested in humans intended for together with the toxin, and prevents death of
the treatment of HUS was SYNSORB Pk, O157:H7-infected mice in a protein-calorie-
a silicon dioxide compound that contains the deficient animal model when given intra-
trisaccharide component of Gb3 (38, 39). venously after infection (44). However, one
The theory behind the use of SYNSORB Pk is point to note is that SUPER TWIG (1)12, de-
that the compound would bind up free Stx(s) veloped in the latter study, neutralized as well
within the intestines of infected patients and as SUPER TWIG (1)6 in vitro but did not
prevent that toxin from binding to the func- protect in vivo, a finding that demonstrates
tional receptor so that the toxin could not act the importance of testing potential therapeu-
either locally or systemically. SYNSORB Pk tics in an animal model system. To find a re-
was shown to neutralize both Stx1 and Stx2 on ceptor analog that could be given orally, the
human renal adenocarcinoma cells (40). Al- Nishikawa group developed Gb3 polymers
though SYNSORB Pk was tolerated well in the that bind to the Stxs with higher affinity than
phase I trial (41), the double-blind placebo- SUPER TWIG (1)6, and those polymers, when
controlled trial suggested no difference be- given by gavage twice daily in the protein-
tween treated and placebo groups (39). The calorie-deficient mouse model on days 3 to 5
reason for the lack of efficacy by SYNSORB Pk after infection, protect mice from death (45).
in the latter study may be that the patients The strong neutralization effect observed
were enrolled after HUS diagnosis, whereas with SUPER TWIG and the Gb3 polymers is
the best time to neutralize toxin to prevent due to the fact that the B pentamer of the Stxs
HUS is most likely before development of this contains multiple Gb3-binding sites, so re-
serious sequela. ceptor mimics that display multiple copies of
A number of other Stx receptor analogs in the Gb3 trisaccharide bind the toxin tightly. A
addition to SYNSORB Pk have been devel- similar approach as described above was used
oped, such as STARFISH (42), Daisy (43), by another group to display the receptor sugar
SUPER TWIG (1)6 (44), Gb3 polymers (45), moiety on chitosan, and they similarly found
and Ac-PPPtet (46), as well as a probiotic that that the analog neutralizes in vitro and in vivo
displays an Stx binder on its surface (47). (48).
STARFISH is a five-armed molecule with According to Nishikawa, however, a draw-
two receptor mimics at the end of each arm; back to the multiple trisaccharide display ap-
the compound appears to bind two B penta- proach is the complexity of synthesis (46, 49);
mers at the same time and neutralizes both therefore, a tetravalent peptide library was
Stx1 and Stx2 in vitro. However, STARFISH screened to find molecules that bind the Stx2
has a higher avidity for Stx1 than Stx2 and pentamer (49). In that latter screen, tetrava-
was only able to protect mice injected with lent peptides were identified that form a

complex with Stx2 and neutralize the toxin on inhibition by Stx1, a finding that may indicate
Vero cells but do not prevent the toxin/pep- that MMA-tet somehow prevents the toxin
tide moiety from binding to the cell. Rather, A1 subunit from reaching the cytoplasm from
the peptide-bound Stx2 was unable to reach the ER.
the ER. Furthermore, O157-infected protein- Finally, a novel modification of the recep-
calorie-deficient mice gavaged after infection tor analog approach to bind up Stx specifically
with the peptides were protected from death; within the intestine was designed by Patons
the protection appeared to be dose dependent group: the core trisaccharide from Gb3, Gala1-
and treatment had to be started by day 3 post 4Galb1-4Glc-, was displayed on the lipopoly-
infection. Further testing of the optimized saccharide (LPS) structure of a commensal
tetrapeptide, Ac-PPP-tet (later renamed TVP), E. coli to create a probiotic (52). The con-
demonstrated that the compound inhibits structed strain, CWG308:pJCP-Gb3, neutral-
Stx2-mediated fluid accumulation in rabbit izes Stx1, Stx2, Stx2c, and Stx2d on Vero
ileal loops (46). However, the tetravalent pep- cells. In addition, streptomycin-treated mice
tide only protects when administered orally are protected from infection with either an
and not intravenously in the protein-calorie- Stx2- or Stx2d-producer when fed the Gala1-
deficient mouse model, so TVP efficacy was 4Galb1-4Glc-probiotic twice daily. What was
tested in the baboon model of intoxication not clear in that study was whether the mice
(50). Baboons given Stx2 and TVP simulta- would continue to do well once the probiotic
neously were protected from death, renal in- was discontinued. However, the probiotic
jury, and thrombocytopenia, but not anemia. with a receptor-analog approach might be pos-
TVP also rescued 75% of animals that received sible as long as the infecting strain did not
the drug 24 h post intoxication and a supple- persist for long-term colonization. Addition-
mental dose on days 2, 3, and 4. Untreated ally, it may be necessary to monitor shedding
but intoxicated baboons died by day 6. The of the infecting strain over time. CWG308:
authors of that study suggest that the advan- pJCP-Gb3 was found to be effective even
tage of TVP compared to other Stx binders is when killed prior to administration; however,
that the compound is cell-permeable. How- the formaldehyde-treated probiotic had to be
ever, it is unclear how TVP would rescue in- administered three times daily for complete
toxicated cells since the proposed mechanism protection (53). To develop a probiotic that
for action is for the Stx2-TVP complex to might be able to be used in humans, a K-12
traffic differently within the cell than Stx2 strain that expressed the same Gala1-4Galb1-
alone, and indeed, in vitro, cells are only pro- 4Glc- epitope on its LPS was created (54).
tected if the TVP is added to cells at the same In that strain, additional mutations were in-
time as Stx2. In a recent study, the Nishikawa corporated so that no antibiotic resistance
group identified another tetravalent peptide, markers were needed for plasmid mainte-
MMA-tet, that inhibits both Stx1 and Stx2 on nance. The K-12 probiotic that displays Gala1-
Vero cells when it is added at the same time 4Galb1-4Glc- neutralizes both in vitro and in
as the toxins, and when administered orally, vivo (54).
protects protein-calorie-deficient mice infec-
ted with STEC strain N-9 (51). The mecha-
Human Serum Amyloid
nism whereby MMA-tet protects cells and
Component P (HuSAP)
animals is unclear: MMA-tet forms a complex
with the Stx1 B subunit but does not appear Reports that there is an Stx2-neutralizing
to alter binding or trafficking of the MMA- component in normal human serum that was
tetStxB1 through the cell, at least as far as not immunoglobulin were published in 1993
the ER. However, MMA-tet pretreatment (55, 56). Later, HuSAP was identified as the
of Vero cells did prevent protein synthesis protein from plasma that can neutralize Stx2,

but not Stx1 (57). HuSAP administered intra- protective in mice (67). The epitopes for both
venously protects BALB/c mice given Stx2 1 h antibodies have been mapped (65, 66); of note,
later (58). Another group showed similar re- although 11E10 neutralizes Stx2 in a cell-
sults, though they administered the HuSAP free protein synthesis assay, the antibody also
twice daily intraperitoneally and the Stx2 by alters the intracellular localization of the toxin
the subcutaneous route (59). This group then such that the toxin-antibody complex does
challenged transgenic C57Bl/6 mice that ex- not reach the cytoplasm (65). Both 11E10 and
press HuSAP with two 50% lethal doses of 13C4 were transformed into human/mouse
Stx2. The HuSAP-transgenic mice survived chimeras by genetic techniques. The hu-
nearly twice as long as the control mice manized versions of anti-Stx1 and anti-Stx2
in that study. The reason that the HuSAP- neutralize the toxins on Vero cells and in ei-
injected BALB/c mice survived longer than ther a mouse intoxication model (Stx1) or the
the HuSAP-transgenic mice was hypothesized streptomycin-treated mouse model of infec-
to be because the circulating levels of HuSAP tion with strain B2F1 (68). The humanized
were higher in the BALB/c mice injected versions of the antibodies (caStx1 for 13C4
with HuSAP than in the transgenic animals. and caStx2 for 11E10) were evaluated in phase
If, however, the HuSAP-transgenic mice were I (69, 70) and II trials. The preliminary results
injected with a combination of LPS and Stx2, of the phase II trial indicate that the anti-
the animals were not protected, even though bodies were safe and well tolerated in sick
the LPS did not prevent HuSAP-Stx2 in- children infected with STEC.
teraction or the neutralization by HuSAP of Takedas group developed monoclonal
Stx2 for Vero cells in vitro (60). Exactly how antibodies that neutralize Stx1 or Stx2 by
HuSAP binds Stx2 is not clear; Marcato et al. preventing receptor binding (71, 72). The anti-
reported that the interaction requires both Stx2 monoclonal was later humanized and
toxin A and B subunits and cannot be com- named TMA-15. TMA-15 demonstrates Vero
peted with Daisy (61). The latter finding sug- cell neutralization and protective efficacy in
gests that HuSAP does not bind within the animal models (73, 74). Antibody TMA-15 was
Gb3-binding sites on the B pentamer. renamed urtoxazumab and evaluated in safety
trials in healthy adults and STEC-infected
children (75). Urtoxazumab was well tolerated
ANTIBODIES TO THE Stxs in STEC-infected children, but no efficacy
data are yet published.
Because the Stxs are the primary factors re- Finally, Tziporis group generated fully
sponsible for the development of HUS, neu- humanized neutralizing monoclonal anti-Stx1
tralization of the toxins is a critical therapeutic and -Stx2 antibodies in transgenic mice (76,
approach for STEC infections. Polyclonal an- 77). The anti-Stx2 monoclonal antibody, 5C12,
tiserum against Stx2 protects gnotobiotic pig- was developed further, and shown to protect
lets from infection by O157:H7 strain 86-24 piglets and streptomycin-treated mice from
(62). Furthermore, monoclonal antibodies to STEC strains that produce Stx2 or Stx2dact,
the toxins are protective in animal models, and respectively (77, 78). Similarly to 11E10, anti-
two groups have taken such therapeutics into body 5C12 neutralizes Stx2 by altering its in-
phase II safety trials, as detailed below. tracellular trafficking pattern and also has the
Monoclonal antibodies specific for the Stx1 capacity to prevent protein synthesis inhibi-
B subunit (13C4) and the Stx2 A subunit tion by Stx2 in a cell-free assay (79, 80). The
(11E10) were developed in the OBrien labo- Tzipori group also evaluated isotype variants
ratory in the mid to late 1980s (63, 64). Those and Fab and F(ab)2 fragments of 5C12 in in
antibodies neutralize the toxins on Vero cells vitro and in vivo models of efficacy and found
(65, 66), and the Stx2 monoclonal antibody is that all of the isotype variants demonstrate in

vitro and in vivo neutralization but the Fab (87). Since chloroquine does not inhibit the
and F(ab)2 were only efficacious in the in enzymatic activity of Stx nor degrade the
vitro assay (81). One point of note from the toxin in vitro, the drug likely acts by inhibiting
latter study is that the IgG4 variant of 5C12 the toxin from exiting the ER.
was the least protective of the isotype variants Further small-molecule screens have iden-
in vitro, but was as protective in vivo as the tified other compounds that inhibit Stx and
best in vitro neutralizer (the IgG3 variant). ricin transport within Vero cells (88) or the
These latter results indicate once again that in enzymatic activity of both toxins (which have
vitro and in vivo neutralization data do not the same mode of action) (89). Another small
always correlate. molecule, eeyarestatin I, which interferes with
intracellular trafficking among cellular com-
partments, was shown to cause a lag in protein
Nonantibody Inhibitors of Toxin
synthesis inhibition in HeLa cells by Stx (90).
Trafcking or the Cellular Response to Stx
Finally, nitrobenzylthioinosine was shown in
Brefeldin A (BFA), a fungal toxin that disrupts 2002 to inhibit the trafficking of Stx1, trapping
the Golgi apparatus but also causes tabulation the toxin within the early endosome (91);
of early endosomes, is a known inhibitor of however, no additional information on the use
Stx transport; in fact, the protective effect of of this compound to protect cells or animals
BFA against the Stxs helped define the path- has been published.
way by which Stx is transported through the A recent report suggests that manganese
cell. Because BFA is a global inhibitor of pro- inhibits intracellular trafficking of the B sub-
tein transport within the cell, it is not included unit of Stx (92), a finding that may or may not
in Table 1. Another Golgi disruptor, Exo2, a extend to the holotoxin. However, manganese
small molecule that does not inhibit cholera treatment increased the viability of HeLa cells
toxin trafficking, prevents Stx from reaching incubated with Stx1 in that study and pro-
the Golgi apparatus (82). Unfortunately, cell tected BALB/c mice injected with 500 ng of
disruptors such as BFA and Exo2 are toxic to Stx1 and injected with daily doses of manga-
cells. Another drawback to molecules such as nese (at least 10 mg/kg daily injections were
BFA and Exo2 is that treated cells recover required). Whether such levels of manganese
over time, such that incubation of the cells would be reasonably achievable in a patient
with the inhibitor and the toxin for longer is not clear. In addition, manganese fails to
periods reduces the effectiveness of the ther- protect HeLa cells from Stx2 intoxication
apeutic. Exo2 was recently derivatized to (93), a finding that makes the potential use of
generate an inhibitor with lower toxicity that manganese less likely as Stx2 is more com-
still interferes with Stx trafficking (83). An- monly associated with HUS development than
other small molecule, Retro-2, blocks Stx from Stx1.
reaching the trans-Golgi network (84, 85) and
is effective against ricin, a toxin that traffics
in the same manner as the Stxs. Recently, a A COMPOUND THAT INTERFERES
derivative of Retro-2, Retro-2cycl, was identi- WITH TOXIN PROCESSING
fied. Retro-2cycl is 100-fold more active on
HeLa cells than Retro-2 (86). However, the Because the A subunit must be cleaved by a
drawback to Retro-2cycl and derivatives is that furin- or trypsin-like protease so that the A1 or
the cells needed to be pretreated with the enzymatically active moiety of the toxin may
drug to observe protection from the toxin (84, be separated from the holotoxin, furin inhib-
86). Another compound known to protect itors were tested in a cell-based assay for
cells from Stx, chloroquine, was also recently the capacity to protect HEp-2 cells from Stx
shown to permit the toxin to reach the ER (94). Although one of the inhibitors showed

moderate inhibition activity against Stx, the Lapeyraque and colleagues reported in 2011
inhibition was overcome at higher concen- the use of eculizumab (an antibody directed
trations of Stx. Furthermore, since the toxin against complement protein C5 used to treat
can be cleaved by enzymes in intestinal mucus atypical HUS, a disease that may arise due to
(35), as well as by intracellular proteases complement dysregulation) in three children
(95), and because Stx mutated at the trypsin- with HUS due to STEC infection (100). Al-
sensitive site can still be cleaved (96, 97), we though the sick children exhibited improve-
suspect that it will be difficult to protect ments in platelet counts and reductions in
animals from Stx with protease inhibitors. lactate dehydrogenase, plasma exchange and
hemodialysis were used concurrently. In ad-
dition, no children in the small study received
Therapies That Interfere with the
just antibody infusion alone. Eculizumab was
Cellular Response to Stx
also used to treat many patients with HUS
Inhibitors of mitogen-activated protein kinase during the O104:H4 outbreak in Germany
pathways were used to assess potential proin 2011. Although there was no evidence of
tective efficacy against the Stx2-mediated efficacy for eculizumab in treating HUS from
ribotoxic stress response in HCT-8 cells or for the outbreak in Germany, the cohorts are
oral gavage of Stx2 into infant rabbits (98). difficult to compare because the patients re-
The authors observed modest protective ca- ceived other concurrent treatments, such as
pacity by imatinib in HCT-8 cells and against plasma exchange, antibiotic therapy, immu-
some aspects of Stx2 intoxication in the ani- noglobulin G immunoadsorption, and dialy-
mals, specifically heterophil infiltration into sis (14, 101105). In addition, many of the
the intestine. That study indicates that the HUS patients given eculizumab were quite
infant rabbit model of Stx2 gavage may ill and had neurological complications. There-
prove useful for the evaluation of therapeutics. fore, a randomized controlled trial is neces-
However, such a model would be expensive, sary to answer the question about the efficacy
and determining effective therapeutic doses of eculizumab for the treatment of Stx-
and the timing for those doses might prove associated HUS.
challenging.
A recent paper suggests that an inhibitor of
apoptosis, ouabain, protects rat renal proximal CONCLUSION
tubules from Stx-mediated cell death (99).
Kidneys from ouabain-treated mice that were Since the first outbreak of STEC infection in
given about four 50% lethal doses of Stx2 the United States in 1982, much research
showed reduced podocyte depletion as com- has focused on ways to neutralize the action
pared to phosphate-buffered salinetreated of the Stxs. At this time, only SYNSORB Pk,
mice. However, no data were reported on urtoxazumab, and the Shiga monoclonal an-
whether the ouabain could protect the mice tibodies caStx1 and caStx2 have been tested
from Stx2-mediated lethality. in phase I and II trials intended to treat or
prevent HUS. Efforts to prevent STEC infec-
tion, such as elimination from the food supply
TREATMENT ONCE HUS IS DIAGNOSED and proper food handling, are important as
well as we try to stop the development of the
What do you do for patients who already have potentially deadly HUS. Lastly, consideration
HUS? We leave discussion of specific clinical should be given to the development of a vac-
interventions such as dialysis and apheresis cine against the Stxs, minimally for those in
for medical experts. In terms of a therapeutic research or clinical labs who are exposed to
that may interfere in the etiology of HUS, STEC.

ACKNOWLEDGMENT control study of Escherichia coli O157:H7 infection

in the United States. J Infect Dis 177:962966.
We declare no conflicts of interest with regard 10. Panos GZ, Betsi GI, Falagas ME. 2006. Sys-
to the manuscript. tematic review: are antibiotics detrimental or
beneficial for the treatment of patients with
Escherichia coli O157:H7 infection? Aliment
CITATION Pharmacol Ther 24:731742.
11. Smith KE, PWilke PR, Reiter PL, Hedican EB,
Melton-Celsa AR, OBrien AD. 2014. New Bender JB, Hedberg CW. 2012. Antibiotic treat-
therapeutic developments against Shiga toxin- ment of Escherichia coli O157 infection and the
producing Escherichia coli. Microbiol Spec- risk of hemolytic uremic syndrome, Minnesota.
Pediatr Infect Dis J 31:3741.
trum 2(5):EHEC-0013-2013.
12. Ikeda K, Ida O, Kimot K, Takatorige T,
Nakanishi N, Tatar K. 1999. Effect of early
REFERENCES fosfomycin treatment on prevention of hemolytic
uremic syndrome accompanying Escherichia coli
1. Rosales A, Hofer J, Zimmerhackl LB, O157:H7 infection. Clin Nephrol 52:357362.
Jungraithmayr TC, Riedl M, Giner T, Strasak A, 13. Fukushima H, Hashizume T, Morita Y, Tanaka
Orth-Holler D, Wurzner R, Karch H. 2012. Need J, Azuma K, Mizumoto Y, Kaneno M, Matsuura
for long-term follow-up in enterohemorrhagic M, Konma K, Kitani T. 1999. Clinical experiences
Escherichia coli-associated hemolytic uremic syn- in Sakai City Hospital during the massive out-
drome due to late-emerging sequelae. Clin Infect break of enterohemorrhagic Escherichia coli O157
Dis 54:14131421. infections in Sakai City, 1996. Pediatr Int 41:213
2. Palermo MS, Exeni RA, Fernandez GC. 2009. 217.
Hemolytic uremic syndrome: pathogenesis and 14. Menne J, Nitschke M, Stingele R, Abu-Tair M,
update of interventions. Expert Rev Anti Infect Beneke J, Bramstedt J, Bremer JP, Brunkhorst
Ther 7:697707. R, Busch V, Dengler R, Deuschl G, Fellermann
3. Spinale JM, Ruebner RL, Copelovitch L, Kaplan K, Fickenscher H, Gerigk C, Goettsche A,
BS. 2013. Long-term outcomes of Shiga toxin Greeve J, Hafer C, Hagenmuller F, Haller H,
hemolytic uremic syndrome. Pediatr Nephrol Herget-Rosenthal S, Hertenstein B, Hofmann
28:20972105. C, Lang M, Kielstein JT, Klostermeier UC,
4. Thielman NM, Guerrant RL. 2004. Clinical Knobloch J, Kuehbacher M, Kunzendorf U,
practice. Acute infectious diarrhea. N Engl J Med Lehnert H, Manns MP, Menne TF, Meyer TN,
350:3847. Michael C, Munte T, Neumann-Grutzeck C,
5. Molbak K, Mead PS, Griffin PM. 2002. Antimi- Nuernberger J, Pavenstaedt H, Ramazan L,
crobial therapy in patients with Escherichia coli Renders L, Repenthin J, Ries W, Rohr A,
O157:H7 infection. JAMA 288:10141016. LRump LC, Samuelsson O, Sayk F, Schmidt
6. Nelson JM, Griffin PM, Jones TF, Smith KE, BM, Schnatter S, Schocklmann H, Schreiber S,
Scallan E. 2011. Antimicrobial and antimotility von Seydewitz CU, Steinhoff J, Stracke S,
agent use in persons with Shiga toxin-producing Suerbaum S, van de Loo A, Vischedyk M,
Escherichia coli O157 infection in FoodNet sites. Weissenborn K, Wellhoner P, Wiesner M,
Clin Infect Dis 52:11301132. Zeissig S, Buning J, Schiffer M, Kuehbacher T.
7. Karch H, Goroncy-Bermes P, Opferkuch W, 2012. Validation of treatment strategies for
Kroll H-P, OBrien A. 1985. Subinhibitory con- enterohaemorrhagic Escherichia coli O104:H4
centrations of antibiotics modulate amount of induced haemolytic uraemic syndrome: case-
Shiga-like toxin produced by Escherichia coli, control study. BMJ 345:e4565.
p 239245. In Adam D, Hahn H, Opferkuch W 15. Geerdes-Fenge HF, Lobermann M, Nurnberg
(ed), The Influence of Antibiotics on the Host- M, Fritzsche C, Koball S, Henschel J, Hohn R,
Parasite Relationship II. SpringerVerlag, Berlin, Schober HC, Mitzner S, Podbielski A, Reisinger
Germany. EC. 2013. Ciprofloxacin reduces the risk of
8. Wong CS, Jelacic S, Habeeb RL, Watkins SL, hemolytic uremic syndrome in patients with
Tarr PI. 2000. The risk of the hemolytic-uremic Escherichia coli O104:H4-associated diarrhea. In-
syndrome after antibiotic treatment of Escherichia fection 41:669673.
coli O157:H7 infections. N Engl J Med 342:1930 16. Binks S, Regan K, Richenberg J, Chevassut T.
1936. 2012. Microbes without frontiers: severe haemolytic-
9. Slutsker L, Ries AA, Malone K, Wells JG, uraemic syndrome due to E coli O104:H4. BMJ Case
Greene KD, Griffin PM. 1998. A nationwide case- Rep 2012.

17. Colic E, Dieperink H, Titlestad K, Tepel M. 27. Ritchie JM, Greenwich JL, Davis BM, Bronson
2011. Management of an acute outbreak of diar- RT, Gebhart D, Williams SR, Martin D,
rhoea-associated haemolytic uraemic syndrome Scholl D, Waldor MK. 2011. An Escherichia coli
with early plasma exchange in adults from south- O157-specific engineered pyocin prevents and
ern Denmark: an observational study. Lancet ameliorates infection by E. coli O157:H7 in an
378:10891093. animal model of diarrheal disease. Antimicrob
18. Bielaszewska M, Idelevich EA, Zhang W, Agents Chemother 55:54695474.
Bauwens A, Schaumburg F, Mellmann A, 28. Rasko DA, Moreira CG, Li de R, Reading NC,
Peters G, Karch H. 2012. Effects of antibiotics Ritchie JM, Waldor MK, Williams N, Taussig R,
on Shiga toxin 2 production and bacteriophage Wei S, Roth M, Hughes DT, Huntley JF, Fina
induction by epidemic Escherichia coli O104:H4 MW, Falck JR, Sperandio V. 2008. Targeting
strain. Antimicrob Agents Chemother 56:3277 QseC signaling and virulence for antibiotic de-
3282. velopment. Science 321:10781080.
19. Nitschke M, Sayk F, Hartel C, Roseland RT, 29. Crane JK, Byrd IW, Boedeker EC. 2011. Viru-
Hauswaldt S, Steinhoff J, Fellermann K, Derad lence inhibition by zinc in Shiga-toxigenic Esch-
I, Wellhoner P, Buning J, Tiemer B, Katalinic A, erichia coli. Infect Immun 79:16961705.
Rupp J, Lehnert H, Solbach W, Knobloch JK. 30. Tesh VL. 2012. The induction of apoptosis
2012. Association between azithromycin therapy by Shiga toxins and ricin. Curr Top Microbiol
and duration of bacterial shedding among patients Immunol 357:137178.
with Shiga toxin-producing enteroaggregative 31. Scheut F, Teel LD, Beutin L, Pierard D, Buvens
Escherichia coli O104:H4. JAMA 307:10461052. G, Karch H, Mellmann A, Caprioli A, Tozzoli R,
20. Isogai E, Isogai H, Hayashi S, Kubota T, Kimura Morabito S, Strockbine NA, Melton-Celsa
K, Fujii N, Ohtani T, Sato K. 2000. Effect of AR, Sanchez M, Persson S, OBrien AD. 2012.
antibiotics, levofloxacin and fosfomycin, on a Multicenter evaluation of a sequence-based pro-
mouse model with Escherichia coli O157 infection. tocol for subtyping Shiga toxins and standard-
Microbiol Immunol 44:8995. izing Stx nomenclature. J Clin Microbiol 50:
21. Zhang X, McDaniel AD, Wolf LE, Keusch GT, 29512963.
Waldor MK, Acheson DW. 2000. Quinolone 32. Bielaszewska M, Friedrich AW, Aldick T,
antibiotics induce Shiga toxin-encoding bacterio- Schurk-Bulgrin R, Karch H. 2006. Shiga toxin
phages, toxin production, and death in mice. activatable by intestinal mucus in Escherichia
J Infect Dis 181:664670. coli isolated from humans: predictor for a severe
22. Yoshimura K, Fujii J, Taniguchi H, Yoshida S. clinical outcome. Clin Infect Dis 43:11601167.
1999. Chemotherapy for enterohemorrhagic Esch- 33. Friedrich AW, Bielaszewska M, Zhang WL,
erichia coli O157:H infection in a mouse model. Pulz M, Kuczius T, Ammon A, Karch H. 2002.
FEMS Immunol Med Microbiol 26:101108. Escherichia coli harboring Shiga toxin 2 gene
23. Kurioka T, Yunou Y, Harada H, Kita E. 1999. variants: frequency and association with clinical
Efficacy of antibiotic therapy for infection with symptoms. J Infect Dis 185:7484.
Shiga-like toxin-producing Escherichia coli O157: 34. Lindgren SW, Samuel JE, Schmitt CK, OBrien
H7 in mice with protein-calorie malnutrition. Eur AD. 1994. The specific activities of Shiga-like
J Clin Microbiol Infect Dis 18:561571. toxin type II (SLT-II) and SLT-II-related toxins
24. Zangari T, Melton-Celsa AR, Panda A, Boisen of enterohemorrhagic Escherichia coli differ when
N, Smith MA, Taratov I, De Tolla LJ, Nataro measured by Vero cell cytotoxicity but not by
JP, OBrien AD. 2013. Virulence of the Shiga mouse lethality. Infect Immun 62:623631.
toxin type 2-expressing Escherichia coli O104:H4 35. Melton-Celsa AR, Darnell SC, OBrien AD. 1996.
German outbreak isolate in two animal models. Activation of Shiga-like toxins by mouse and
Infect Immun 81:15621574. human intestinal mucus correlates with virulence
25. Al-Qarawi S, Fontaine RE, Al-Qahtani MS. 1995. of enterohemorrhagic Escherichia coli O91:H21
An outbreak of hemolytic uremic syndrome as- isolates in orally infected, streptomycin-treated
sociated with antibiotic treatment of hospital mice. Infect Immun 64:15691576.
inpatients for dysentery. Emerg Infect Dis 1:138 36. Silberstein C, Copeland DP, Chiang W-L,
140. Repetto HA, Ibarra C. 2008. A glucosylceramide
26. Scholl D, Cooley M, Williams SR, Gebhart D, synthase inhibitor prevents the cytotoxic effects
Martin D, Bates A, Mandrell R. 2009. An of Shiga toxin-2 on human renal tubular epithelial
engineered R-type pyocin is a highly specific and cells. J Epithel Biol Pharmacol 1:7175.
sensitive bactericidal agent for the food-borne 37. Silberstein C, Lucero MS, Zotta E, Copeland
pathogen Escherichia coli O157:H7. Antimicrob DP, Lingyun L, Repetto HA, Ibarra C. 2011. A
Agents Chemother 53:30743080. glucosylceramide synthase inhibitor protects rats

against the cytotoxic effects of Shiga toxin 2. 47. Paton AW, Morona R, Paton JC. 2001. Neu-
Pediatr Res 69:39394. tralization of Shiga toxins Stx1, Stx2c, and Stx2e
38. Armstrong GD, Fodor E, Vanmaele R. 1991. In- by recombinant bacteria expressing mimics of
vestigation of Shiga-like toxin binding to chemi- globotriose and globotetraose. Infect Immun
cally synthesized oligosaccharide sequences. 69:19671970.
J Infect Dis 164:11601167. 48. Li X, Wu P, Cheng S, Lv X. 2012. Synthesis
39. Trachtman H, Cnaan A, Christen E, Gibbs K, and assessment of globotriose-chitosan conjugate,
Zhao S, Acheson DW, Weiss R, Kaskel FJ, a novel inhibitor of Shiga toxins produced by
Spitzer A, Hirschman GH. 2003. Effect of Escherichia coli. J Med Chem 55:27022710.
an oral Shiga toxin-binding agent on diarrhea- 49. Nishikawa K, Watanabe M, Kita E, Igai K,
associated hemolytic uremic syndrome in chil- Omata K, Yaffe MB, Natori Y. 2006. A multi-
dren: a randomized controlled trial. JAMA 290: valent peptide library approach identifies a novel
13371344. Shiga toxin inhibitor that induces aberrant cellu-
40. Takeda T, Yoshino K, Adachi E, Sato Y, lar transport of the toxin. FASEB J 20:25972599.
Yamagata K. 1999. In vitro assessment of a 50. Stearns-Kurosawa DJ, Collins V, Freeman S,
chemically synthesized Shiga toxin receptor an- Debord D, Nishikawa K, Oh SY, Leibowitz CS,
alog attached to chromosorb P (Synsorb Pk) as a Kurosawa S. 2011. Rescue from lethal Shiga toxin
specific absorbing agent of Shiga toxin 1 and 2. 2-induced renal failure with a cell-permeable
Microbiol Immunol 43:331337. peptide. Pediatr Nephrol 26:20312039.
41. Armstrong GD, Rowe PC, Goodyer P, Orrbine 51. Tsutsuki K, Watanabe-Takahashi M, Takenaka
E, Klassen TP, Wells G, MacKenzie A, Lior H, Y, Kita E, Nishikawa K. 2013. Identification of a
Blanchard C, Auclair F, et al. 1995. A phase I peptide-based neutralizer that potently inhibits
study of chemically synthesized verotoxin (Shiga- both Shiga toxins 1 and 2 by targeting specific
like toxin) Pk-trisaccharide receptors attached receptor-binding regions. Infect Immun 81:2133
to chromosorb for preventing hemolytic-uremic 2138.
syndrome. J Infect Dis 171:10421045. 52. Paton AW, Morona R, Paton JC. 2000. A new
42. Kitov PI, Sadowska JM, Mulvey G, Armstrong biological agent for treatment of Shiga toxigenic
GD, Ling H, Pannu NS, Read RJ, Bundle DR. Escherichia coli infections and dysentery in
2000. Shiga-like toxins are neutralized by tailored humans. Nat Med 6:265270.
multivalent carbohydrate ligands. Nature 403:66 53. Paton JC, Rogers TJ, Morona R, Paton AW.
672. 2001. Oral administration of formaldehyde-killed
43. Mulve GL, Marcato P, Kitov PI, Sadowska J, recombinant bacteria expressing a mimic of the
Bundle DR, Armstrong GD. 2003. Assessment in Shiga toxin receptor protects mice from fatal
mice of the therapeutic potential of tailored, challenge with Shiga-toxigenic Escherichia coli.
multivalent Shiga toxin carbohydrate ligands. Infect Immun 69:13891393.
J Infect Dis 187:640649. 54. Pinyon RA, Paton JC, Paton AW, JBotten JA,
44. Nishikawa K, Matsuoka K, Kita E, Okabe N, Morona R. 2004. Refinement of a therapeutic
Mizuguchi M, Hino K, Miyazawa S, Yamasaki Shiga toxin-binding probiotic for human trials.
C, Aoki J, Takashima S, Yamakawa Y, Nishijima J Infect Dis 189:15471555.
M, Terunuma D, Kuzuhara H, Natori Y. 2002. A 55. Bitzan M, Klemt M, Steffens R, Muller-Wiefel
therapeutic agent with oriented carbohydrates for DE. 1993. Differences in verotoxin neutralizing
treatment of infections by Shiga toxin-producing activity of therapeutic immunoglobulins and sera
Escherichia coli O157:H7. Proc Natl Acad Sci USA from healthy controls. Infection 21:140145.
99:76697674. 56. Bitzan M, Ludwig K, Klemt M, Konig H,
45. Watanabe M, Matsuoka K, Kita E, Igai Buren J, Muller-Wiefel DE. 1993. The role of
K, Higashi N, Miyagawa A, Watanabe T, Escherichia coli O157 infections in the classical
Yanoshita R, Samejima Y, Terunuma D, Natori (enteropathic) haemolytic uraemic syndrome: re-
Y, Nishikawa K. 2004. Oral therapeutic agents sults of a Central European, multicentre study.
with highly clustered globotriose for treatment Epidemiol Infect 110:183196.
of Shiga toxigenic Escherichia coli infections. 57. Kimura T, Tani S, Matsumoto Yi Y, Takeda T.
J Infect Dis 189:360368. 2001. Serum amyloid P component is the Shiga
46. Watanabe-Takahashi M, Sato T, Dohi T, toxin 2-neutralizing factor in human blood. J Biol
Noguchi N, Kano F, Murata M, Hamabata T, Chem 276:4157641579.
Natori Y, Nishikawa K. 2010. An orally applica- 58. Kimura T, Tani S, Motoki M, Matsumoto Y.
ble Shiga toxin neutralizer functions in the in- 2003. Role of Shiga toxin 2 (Stx2)-binding
testine to inhibit the intracellular transport of the protein, human serum amyloid P component
toxin. Infect Immun 78:177183. (HuSAP), in Shiga toxin-producing Escherichia

coli infections: assumption from in vitro and in and protection from lethal outcome by anti-Stx2
vivo study using HuSAP and anti-Stx2 humanized antibody. Infect Immun 76:44694478.
monoclonal antibody TMA-15. Biochem Biophys 68. Edwards AC, Melton-Celsa AR, Arbuthnott K,
Res Commun 305:10571060. Stinson JR, Schmitt CK, Wong HC, OBrien AD.
59. Armstrong GD, Mulvey GL, Marcato P, Griener 1998. Vero cell neutralization and mouse protec-
TP, Kahan MC, Tennent GA, Sabin CA, Chart tive efficacy of humanized monoclonal antibodies
H, Pepys MB. 2006. Human serum amyloid P against Escherichia coli toxins Stx1 and Stx2,
component protects against Escherichia coli p 388392. In Kaper JB, OBrien AD (ed), Esch-
O157:H7 Shiga toxin 2 in vivo: therapeutic impli- erichia coli O157:H7 and Other Shiga Toxin-
cations for hemolytic-uremic syndrome. J Infect Producing E. coli Strains. ASM Press, Washington,
Dis 193:11201124. DC.
60. Griener TP, Strecker JG, Humphries RM, 69. Dowling TC, Chavaillaz PA, Young DG,
Mulvey GL, Fuentealba C, Hancock RE, Melton-Celsa A, OBrien A, Thuning-Roberson
Armstrong GD. 2011. Lipopolysaccharide renders C, Edelman R, Tacket CO. 2005. Phase 1 safety
transgenic mice expressing human serum amyloid and pharmacokinetic study of chimeric murine-
P component sensitive to Shiga toxin 2. PLoS One human monoclonal antibody c alpha Stx2 admin-
6:e21457. istered intravenously to healthy adult volunteers.
61. Marcato P, Vander Helm K, Mulvey GL, Antimicrob Agents Chemother 49:18081812.
Armstrong GD. 2003. Serum amyloid P compo- 70. Bitzan M, Poole R, Mehran M, Sicard E,
nent binding to Shiga toxin 2 requires both a Brockus C, Thuning-Roberson C, Riviere M.
subunit and B pentamer. Infect Immun 71:607 2009. Safety and pharmacokinetics of chimeric
6078. anti-Shiga toxin 1 and anti-Shiga toxin 2 mono-
62. Donohue-Rolfe A, Kondova I, Mukherjee J, clonal antibodies in healthy volunteers. Anti-
Chios K, Hutto D, Tzipori S. 1999. Antibody- microb Agents Chemother 53:30813087.
based protection of gnotobiotic piglets infected 71. Nakao H, Kataoka C, Kiyokawa N, Fujimoto J,
with Escherichia coli O157:H7 against systemic Yamasaki S, Takeda T. 2002. Monoclonal anti-
complications associated with Shiga toxin 2. In- body to Shiga toxin 1, which blocks receptor
fect Immun 67:36453648. binding and neutralizes cytotoxicity. Microbiol
63. Perera LP, Marques LR, OBrien AD. 1988. Immunol 46:777780.
Isolation and characterization of monoclonal an- 72. Nakao H, Kiyokawa N, Fujimoto J, Yamasaki
tibodies to Shiga-like toxin II of enterohemor- S, Takeda T. 1999. Monoclonal antibody to
rhagic Escherichia coli and use of the monoclonal Shiga toxin 2 which blocks receptor binding and
antibodies in a colony enzyme-linked immuno- neutralizes cytotoxicity. Infect Immun 67:5717
sorbent assay. J Clin Microbiol 26:21272131. 5722.
64. Strockbine NA, Marques LR, Holmes RK, 73. Kimura T, Co MS, Vasquez M, Wei S, Xu H,
OBrien AD. 1985. Characterization of mono- Tani S, Sakai Y, Kawamura T, Matsumoto Y,
clonal antibodies against Shiga-like toxin from Nakao H, Takeda T. 2002. Development of
Escherichia coli. Infect Immun 50:69700. humanized monoclonal antibody TMA-15 which
65. Smith MJ, Melton-Celsa AR, Sinclair JF, neutralizes Shiga toxin 2. Hybrid Hybridomics
Carvalho HM, Robinson CM, OBrien AD. 2009. 21:161168.
Monoclonal antibody 11E10, which neutralizes 74. Yamagami S, Motoki M, Kimura T, Izumi H,
shiga toxin type 2 (Stx2), recognizes three regions Takeda T, Katsuura Y, Matsumoto Y. 2001.
on the Stx2 A subunit, blocks the enzymatic ac- Efficacy of postinfection treatment with anti-
tion of the toxin in vitro, and alters the overall Shiga toxin (Stx) 2 humanized monoclonal anti-
cellular distribution of the toxin. Infect Immun body TMA-15 in mice lethally challenged with
77:27302740. Stx-producing Escherichia coli. J Infect Dis
66. Smith MJ, Carvalho HM, Melton-Celsa AR, 184:738742.
OBrien AD. 2006. The 13C4 monoclonal anti- 75. Lopez EL, Contrini MM, Glatstein E, Gonzalez
body that neutralizes Shiga toxin type 1 (Stx1) Ayala S, Santoro R, Allende D, Ezcurra G,
recognizes three regions on the Stx1 B subunit Teplitz E, Koyama T, Matsumoto Y, Sato H,
and prevents Stx1 from binding to its eukaryotic Sakai K, Hoshide S, Komoriya K, Morita T,
receptor globotriaosylceramide. Infect Immun Harning R, Brookman S. 2010. Safety and phar-
74:69926998. macokinetics of urtoxazumab, a humanized mono-
67. Sauter KA, Melton-Celsa AR, Larkin K, clonal antibody, against Shiga-like toxin 2 in
Troxell ML, OBrien AD, Magun BE. 2008. healthy adults and in pediatric patients infected
Mouse model of hemolytic-uremic syndrome with Shiga-like toxin-producing Escherichia coli.
caused by endotoxin-free Shiga toxin 2 (Stx2) Antimicrob Agents Chemother 54:239243.

76. Mukherjee J, Chios K, Fishwild D, Hudson D, of Retro-2 with enhanced efficacy against Shiga
ODonnell S, Rich SM, Donohue-Rolfe A, toxin. J Med Chem 56:34043413.
Tzipori S. 2002. Production and characterization 87. Dyve Lingelem AB, Bergan J, Sandvig K. 2012.
of protective human antibodies against Shiga Inhibitors of intravesicular acidification protect
toxin 1. Infect Immun 70:58965899. against Shiga toxin in a pH-independent manner.
77. Mukherjee J, Chios K, Fishwild D, Hudson Traffic 13:443454.
D, ODonnell S, Rich SM, Donohue-Rolfe A, 88. Saenz JB, Doggett TA, Haslam DB. 2007.
Tzipori S. 2002. Human Stx2-specific monoclo- Identification and characterization of small mol-
nal antibodies prevent systemic complications of ecules that inhibit intracellular toxin transport.
Escherichia coli O157:H7 infection. Infect Immun Infect Immun 75:45524561.
70:612619. 89. Wahome PG, Bai Y, Neal LM, Robertus JD,
78. Sheoran AS, Chapman S, Singh P, Donohue- Mantis NJ. 2010. Identification of small-molecule
Rolfe A, Tzipori S. 2003. Stx2-specific human inhibitors of ricin and Shiga toxin using a cell-based
monoclonal antibodies protect mice against lethal high-throughput screen. Toxicon 56:313323.
infection with Escherichia coli expressing Stx2 90. Aletrari MO, McKibbin C, Williams H, Pawar
variants. Infect Immun 71:31253130. V, Pietroni P, Lord JM, Flitsch SL, Whitehead
79. Krautz-Peterson G, Chapman-Bonofiglio S, R, Swanton E, High S, Spooner RA. 2011.
Boisvert K, Feng H, Herman IM, Tzipori S, Eeyarestatin 1 interferes with both retrograde and
Sheoran AS. 2008. Intracellular neutralization of anterograde intracellular trafficking pathways.
Shiga toxin 2 by an a subunit-specific human PLoS One 6:e22713.
monoclonal antibody. Infect Immun 76:19311939. 91. Sekino T, Kiyokawa N, Taguchi T, Ohmi K,
80. Jeong KI, Chapman-Bonofiglio S, Singh P, Lee Nakajima H, Suzuki T, Furukawa S, Nakao H,
J, Tzipori S, Sheoran AS. 2010. In vitro and in Takeda T, Fujimoto J. 2002. Inhibition of Shiga
vivo protective efficacies of antibodies that neu- toxin cytotoxicity in human renal cortical epi-
tralize the RNA N-glycosidase activity of Shiga thelial cells by nitrobenzylthioinosine. J Infect Dis
toxin 2. BMC Immunol 11:16. 185:785796.
81. Akiyoshi DE, Sheoran AS, Rich CM, Richard L, 92. Mukhopadhyay S, Linstedt AD. 2012. Manga-
Chapman-Bonofiglio S, Tzipori S. 2010. Evalu- nese blocks intracellular trafficking of Shiga toxin
ation of Fab and F(ab)2 fragments and isotype and protects against Shiga toxicosis. Science 335:
variants of a recombinant human monoclonal 332335.
antibody against Shiga toxin 2. Infect Immun 93. Mukhopadhyay S, Redler B, Linstedt AD. 2013.
78:13761382. Shiga toxin-binding site for host cell receptor
82. Spooner RA, Watson P, Smith DC, Boal F, GPP130 reveals unexpected divergence in toxin-
Amessou M, Johannes L, Clarkson GJ, Lord trafficking mechanisms. Mol Biol Cell 24:23112318.
JM, Stephens DJ, Roberts LM. 2008. The se- 94. Becker GL, Lu Y, Hardes K, Strehlow B,
cretion inhibitor Exo2 perturbs trafficking of Levesque C, Lindberg I, Sandvig K, Bakowsky
Shiga toxin between endosomes and the trans- U, Day R, Garten W, Steinmetzer T. 2012.
Golgi network. Biochem J 414:471484. Highly potent inhibitors of proprotein convertase
83. Guetzoyan LJ, Spooner RA, Boal F, Stephens furin as potential drugs for treatment of infectious
DJ, Lord JM, Roberts LM, Clarkson GJ. 2010. diseases. J Biol Chem 287:2199222003.
Fine tuning Exo2, a small molecule inhibitor of 95. Garred O, van Deurs B, Sandvig K. 1995. Furin-
secretion and retrograde trafficking pathways in induced cleavage and activation of Shiga toxin.
mammalian cells. Mol Biosyst 6:20302038. J Biol Chem 270:1081710821.
84. Stechmann B, Bai SK, Gobbo E, Lopez R, Merer 96. Garred O, Dubinina E, Holm PK, Olsnes S, van
G, Pinchard S, Panigai L, Tenza D, Raposo G, Deurs B, Kozlov JV, Sandvig K. 1995. Role of
Beaumelle B, Sauvaire D, Gillet D, Johannes L, processing and intracellular transport for optimal
Barbier J. 2010. Inhibition of retrograde trans- toxicity of Shiga toxin and toxin mutants. Exp Cell
port protects mice from lethal ricin challenge. Res 218:3949.
Cell 141:231242. 97. Burgess BJ, Roberts LM. 1993. Proteolytic
85. Park JG, Kahn JN, Tumer NE, Pang YP. 2012. cleavage at arginine residues within the hydro-
Chemical structure of Retro-2, a compound that philic disulphide loop of the Escherichia coli
protects cells against ribosome-inactivating pro- Shiga-like toxin I A subunit is not essential for
teins. Sci Rep 2:631. cytotoxicity. Mol Microbiol 10:171179.
86. Noel R, Gupta N, Pons V, Goudet A, Garcia- 98. Stone SM, Thorpe CM, Ahluwalia A, Rogers
Castillo MD, Michau A, Martinez J, Buisson AB, Obata F, Vozenilek A, Kolling GL, Kan AV,
DA, Johannes L, Gillet D, Barbier J, Cintrat JC. Magun BE, Jandhyala DM. 2012. Shiga toxin
2013. N-methyl dihydroquinazolinones derivatives 2-induced intestinal pathology in infant rabbits is

A-subunit dependent and responsive to the tyro- 103. Loos S, Ahlenstiel T, Kranz B, Staude H, Pape L,
sine kinase and potential ZAK inhibitor imatinib. Hartel C, Vester U, Buchtala L, Benz K, Hoppe
Front Cell Infect Microbiol 2:135. B, Beringer O, Krause M, Muller D, Pohl M,
99. Burlaka I, Liu XL, Rebetz J, Arvidsson I, Yang Lemke J, Hillebrand G, Kreuzer M, Konig J,
L, Brismar H, Karpman D, Aperia A. 2013. Wigger M, Konrad M, Haffner D, Oh J, Kemper
Ouabain protects against Shiga toxin-triggered MJ. 2012. An outbreak of Shiga toxin-producing
apoptosis by reversing the imbalance between Escherichia coli O104:H4 hemolytic uremic syn-
Bax and Bcl-xL. J Am Soc Nephrol 24:14131423. drome in Germany: presentation and short-
100. Lapeyraque AL, Malina M, Fremeaux-Bacchi V, term outcome in children. Clin Infect Dis 55:753
Boppel T, Kirschfink M, Oualha M, Proulx F, 759.
Clermont MJ, Le Deist F, Niaudet P, Schaefer 104. Hauswaldt S, Nitschke M, Sayk F, Solbach W,
F. 2011. Eculizumab in severe Shiga-toxin- Knobloch JK. 2013. Lessons learned from
associated HUS. N Engl J Med 364:25612563. uutbreaks of Shiga toxin producing Escherichia
101. Greinache A, Friesecke S, Abel P, Dressel A, coli. Curr Infect Dis Rep 15:9.
Stracke S, Fiene M, Ernst F, Selleng K, 105. Kielstein JT, Beutel G, Fleig S, Steinhoff J,
Weissenborn K, Schmidt BM, Schiffer M, Meyer TN, Hafer C, Kuhlmann U, Bramstedt J,
Felix SB, Lerch MM, Kielstein JT, Mayerle J. Panzer U, Vischedyk M, Busch V, Ries W,
2011. Treatment of severe neurological deficits Mitzner S, Mees S, Stracke S, Nurnberger J,
with IgG depletion through immunoadsorption in Gerke P, Wiesner M, Sucke B, Abu-Tair M,
patients with Escherichia coli O104:H4-associated Kribben A, Klause N, Schindler R, Merkel F,
haemolytic uraemic syndrome: a prospective trial. Schnatter S, Dorresteijn EM, Samuelsson O,
Lancet 378:11661173. Brunkhorst R. 2012. Best supportive care and
102. Ullrich S, Bremer P, Neumann-Grutzeck C, Otto therapeutic plasma exchange with or without
H, Ruther C, von Seydewitz CU, Meyer GP, eculizumab in Shiga-toxin-producing E. coli
Ahmadi-Simab K, Rother J, Hogan B, Schwenk O104:H4 induced haemolytic-uraemic syndrome:
W, Fischbach R, Caselitz J, Puttfarcken J, an analysis of the German STEC-HUS registry.
Huggett S, Tiedeken P, Pober J, Kirkiles-Smith Nephrol Dial Transplant 27:38073815.
NC, Hagenmuller F. 2013. Symptoms and clinical 106. Trachtman H, Austin C, Lewinski M, Stahl RA.
course of EHEC O104 infection in hospitalized 2012. Renal and neurological involvement in
patients: a prospective single center study. PLoS typical Shiga toxin-associated HUS. Nat Rev
One 8:e55278. Nephrol 8:658669.

HOST DETERMINANTS OF DISEASE
AND HOST RESPONSE

Risk Factors for Shiga
Toxin-Producing Escherichia coli-
Associated Human Diseases
MARTA RIVAS,1 ISABEL CHINEN,1 ELIZABETH MILIWEBSKY,1 and

MARCELO MASANA2
18
INTRODUCTION
Shiga toxin-producing Escherichia coli (STEC) strains emerged in the late 1970s
or early 1980s as highly significant zoonotic threats to public health. In 1982,
two outbreaks of severe bloody diarrhea, related to a previously rare serotype of
E. coli, O157:H7, were reported in the United States (1).
At present, we know that STEC strains are an important cause of morbidity
and mortality, with associated loss of life years and diminished health-related
quality of life. The clinical manifestations of infection range from symptom-
free carriage to nonbloody diarrhea, hemorrhagic colitis (HC), and hemolytic-
uremic syndrome (HUS) (2). The linkage between STEC infection and the
development of HUS was established by Karmali and colleagues in 1983 to 1985
(3, 4).
HUS is a systemic thrombotic microangiopathy caused by different etiologies
and mechanisms, involving acute kidney failure that may result in death or end-
stage renal disease (ESRD), a serious chronic condition that reduces life ex-
pectancy. Patients with ESRD are initially treated with peritoneal dialysis or
1
Instituto Nacional de Enfermedades Infecciosas, ANLIS Dr. C. G. Malbrn, (1281) Buenos Aires, Argentina; 2Instituto
Tecnologa de Alimentos, Centro de Investigacin de Agroindustria, Instituto Nacional de Tecnologa Agropecuaria,
(B1708WAB) Morn, Pcia. de Buenos Aires, Argentina.
361

362 RIVAS ET AL.
hemodialysis and may later be eligible for with outbreaks. Seropathotypes C (O91:H21,
kidney transplantation (5). The cascade lead- O104:H21 and O113:H21, among others) and
ing from gastrointestinal infection to renal D have a low incidence in human illness and
impairment is complex, the production of are rarely associated with outbreaks. Finally,
Shiga toxin 1, Shiga toxin 2, and/or their seropathotype E is composed of many sero-
subtypes (Stx1a, Stx1c, Stx2a, Stx2b, Stx2c, types that have not been implicated in human
Stx2dactivatable, and Stx2f) being the primary diseases.
virulence trait responsible for human disease.
However, a mosaic of different virulence
traits, comprising several adhesins and other SURVEILLANCE AND DISEASE TRENDS
toxins that may play a role in pathogenesis,
has also been described (2). Surveillance practices vary considerably among
Cattle have been recognized as the main countries, and therefore caution is required
reservoir for STEC for more than 30 years; when comparing STEC incidence rates among
however, several different surveys have dem- countries.
onstrated that STEC strains occurred in the In the United States, E. coli O157:H7 infec-
gastrointestinal tracts of other domestic ani- tion became nationally notifiable in 1995. Since
mals such as sheep, goats, water buffalos, pigs, 2000, all STEC infections that cause human
dogs, and cats (6). Humans usually become illness are notifiable to the Nationally Notifi-
infected by eating undercooked beef products, able Diseases Surveillance System (NNDSS)
but secondary sources, including leafy green in the United States. In 2011, the Centers
vegetables, apple cider, and dairy products for Disease Control and Prevention (CDC)
that have been contaminated with manure, Emerging Infections Program analyzed the
are also vehicles for food-borne infection (7). data gathered from the Foodborne Diseases
Infections have also been caused by drinking Active Surveillance Network (FoodNet). A
or swimming in contaminated water, person- total of 521 laboratory-confirmed cases of
to-person transmission, or contact with in- STEC non-O157 and 463 of STEC O157 infec-
fected animals (6, 8). tions were identified, with incidence rates
STEC isolates that cause human infections of 1.10 and 0.97 per 100,000 persons, respec-
belong to a large number of O:H serotypes, tively. Moreover, FoodNet ascertained 96
and O157:H7 is the most prevalent serotype HUS cases, including 93 (97%) postdiarrheal
associated with large outbreaks and sporadic HUS cases in 2010. The population under
cases of HC and HUS in many countries (9). surveillance was 47,505,580, which represents
In 2003, Karmali et al. (10) proposed clas- 15.2% of the total U.S. population. According to
sifying STEC serotypes into five seropatho- CDC, illnesses caused by non-O157 STEC sero-
types (A to E) based on their reported groups tended to be less severe than those
frequencies in human illness and their known caused by E. coli O157:H7 because they re-
associations with outbreaks and severe out- quired less hospitalization (18% versus 43.4%),
comes, including HC and HUS. Seropatho- the death rate was lower (0.19% versus 0.43%),
type A (O157:H7 and O157:NM), considered and HUS developed in fewer patients (1.7%
the most virulent, is associated with the versus 6.3%) (11).
highest incidence in human disease and is In Canada, STEC infection has been clas-
often involved in outbreaks. Seropathotype sified as a notifiable disease since 1990.
B (O26:H11 and NM; O45:H2 and NM; C-EnterNet is a national integrated enteric
O103:H2, H11, H25, and NM; O111:H8 and pathogen surveillance system that collects
NM; O121:H19 and H7; and O145:NM) is as- information on both cases and source of ex-
sociated with severe human disease, but at a posure in two sentinel sites, Ontario (since
lower frequency, and is uncommonly involved 2005) and British Columbia (since 2010).

CHAPTER 18 Risk Factors for STEC-Associated Human Diseases 363
In 2010, the incidence of illnesses caused by was established in 1999, and all clinical labo-
STEC was 2.2 and 2.9 per 100,000 persons in ratories must report and send isolates to the
each site, respectively. The national incidence Reference Laboratory. In Uruguay, reports of
rate was 2.3 per 100,000 persons in 2008 (12). HUS are not mandatory, and only a few cases
In 2010, the European Centre for Disease are recognized each year (16). In Paraguay,
Prevention and Control reported 3,710 con- reporting of HUS has been mandatory since
firmed cases of STEC infection, with an inci- 2005, and the estimated annual incidence is
dence of 0.96 per 100,000 persons. The annual 0.6 cases per 100,000 children under 5 years
Community Summary Report on the Europe- old (8).
an Union gave an incidence of all STEC in- In Argentina, STEC-associated illnesses
fections in Austria, Belgium, Finland, and Italy are a serious public health concern. Data on
as 1 case per 100,000 or less; and in Germany, human STEC infections are gathered through
the United Kingdom, the Netherlands, Den- different strategies: (i) reporting of clinical
mark, Sweden, and Ireland, as 1.2, 1.8, 2.9, 3.2, HUS cases to the National Health Surveillance
3.6, and 4.4 cases per 100,000 persons, re- System (in Argentina the system is named
spectively (13). Sistema Nacional de Vigilancia de la Salud
In Australia, information on the incidence [SNVS]); reports, which have been mandatory
of STEC infections and HUS is obtained from since 2000, must be immediate and individu-
the Australian NNDSS, and has been manda- alized; (ii) the Sentinel Surveillance System
tory in all jurisdictions since 2000, except through 25 HUS sentinel units; (iii) the labo-
Queensland and Western Australia, where ratory-based surveillance system through the
the incidence became notifiable in 2001. For National Diarrheal and Foodborne Pathogens
the 11-year period from 2000 to 2010, the Network; and (iv) the Molecular Surveillance
overall annual rate was 0.4 cases per 100,000 through the PulseNet of Latin America and
persons, and the annual rate of notification Caribbean.
for HUS was 0.07 cases per 100,000 persons, Postdiarrheal HUS is endemic, with the
while neighboring New Zealand reported a highest rate of pediatric cases globally. Over
STEC infection rate of 3.3 cases per 100,000 the last 10 years, approximately 400 HUS
per year (14). cases were reported annually. The incidence
In Latin America, STEC infections are en- ranged from 10 to 17 cases per 100,000 chil-
demic and contribute to the burden of acute dren less than 5 years of age, and lethality was
diarrheal syndrome in children less than 5 between 1 and 4% (Fig. 1).
years of age, being responsible for 2% of total Between 2004 and 2010, a total of 1,245
cases of acute diarrhea, and in a few studies O157 and non-O157 STEC strains, isolated
correspond to 20 to 30% of bloody diarrhea from HUS (597) and bloody (335) and non-
(8). bloody (167) diarrhea cases, healthy carriers
Important differences exist in the inci- (74), and unspecified pathologies (72), were
dence of STEC infections and HUS in South confirmed by the laboratory-based surveil-
America. A regional network for surveillance lance system and the HUS sentinel units. Mul-
purposes is still nonexistent, and data are re- tiple serotypes were identified, but O157:H7
stricted to only a few countries. HUS is en- was the predominant (>70%), and O145:NM
demic in some countries of the southern cone (13.6%) was the second most important sero-
region, and reporting is mandatory only in type identified. Among the STEC O157 strains,
Argentina, Bolivia, Chile, and Paraguay. the stx2a/stx2c/eae/ehxA genotype prevailed
In Brazil, STEC infections are important (>80%). For the non-O157 strains, the geno-
public health issues, at least in some regions, types were more diverse, but the full viru-
but in general, the incidence is relatively low lent stx2a/eae/ehxA genotype was prevalent
(15). In Chile, a National Surveillance System (>60%). In Argentina, outbreaks are identified

364 RIVAS ET AL.
FIGURE 1 Number of HUS cases, incidence rates, and percentages of lethality in Argentina, 20022011.
through the surveillance system of HUS and past 10 years, is identical to the most prevalent
STEC-associated diseases. The definition of pattern in Sweden (named SMI-H) and to the
an outbreak used for this analysis is two or most common type, named 047, in human
more linked cases. Pulsed-field gel electro- infections in the United States (17). Among
phoresis (PFGE) and phage typing are used the non-O157 strains, the PFGE patterns
to establish the clonal relatedness of the iso- are more diverse and two patterns, named
lates. In the period 2004 to 2010, a total of 12 AREXSX01.0006 (O145) and AREXPX01.0008
outbreaks of bloody diarrhea and HUS cases (O113), are prevalent.
associated with O157 and non-O157 STEC
strains occurred in kindergartens, families,
and the community. The outbreak size ranged COST OF STEC-ASSOCIATED DISEASES
from 2 to 32 cases, and two patients with HUS
died. Person-to-person transmission was the Severe illnesses with long-term sequelae caused
main route identified. by STEC have a social and economic cost to the
As a part of PulseNet Latin America and community and the health system. However, it
Caribbean, national databases were created is difficult to compare data from cost of illness
for O157 and non-O157 E. coli, including studies among countries because of differences
strains isolated since 1988 from different in definitions of costs, methodologies used, and
sources. Among O157 strains, two patterns, income distribution.
named AREXHX01.011 and AREXHX01.022, From FoodNet data (20052008), non-
are prevalent, representing around 13% of O157 STEC strains are estimated to cause
the database (Fig. 2a and 2b). Pattern 011 and 168,698 illnesses each year, and E. coli
other related patterns, with 95% similarity, O157:H7, 96,534 cases in the United States,
are part of the hypervirulent clone described with more than 3,600 hospitalizations and
in different countries (Fig. 2c). Pattern 011, 30 deaths (18). Frenzen et al. (19) estimated
which has been prevalent in Argentina in the the total annual cost of illness (COI) due

FIGURE 2 (a) Top ten XbaI-PFGE patterns associated with human STEC O157 strains in Argentina. (b) Den-
drogram of the top ten XbaI-PFGE patterns. (c) Dendrogram of AREHXHX01.0011 pattern and other related
patterns. doi:10.1128/microbiolspec.EHEC-0002-2013.f2
to STEC O157 at USD$405 million in 2003, non-health care costs of 231 HUS cases at-
Buzby and Roberts (20) updated this estimate tending the Hospital Nacional de Pediatra
to USD$459 million in 2007, and Scharff Dr. Juan P. Garrahan in Buenos Aires City,
et al. (21) suggested that STEC O157 infections in the period 1987 to 2003. The total annual
cost about USD$990 million in 2009 to U.S. cost for HUS cases was approximately USD$2
residents. million.
In the Netherlands, Havelaar et al. (22)
described the burden of disease associated
with STEC O157 at the population level RISK FACTORS
using the public health indicator Disability-
Adjusted Life Years (DALYs), and showed Some determinants of the pathogen and its
that mortality due to HUS, ESRD, and dialysis reservoir, the host, and cultural and dietary
due to ESRD constitute the main determi- behaviors have been described as risk factors
nants. Tariq et al. (23) evaluated the societal for acquiring an STEC-associated disease.
impact of STEC O157 infection using a com-
bination of DALYs and COI, including direct
Pathogen Factors
health care costs and indirect non-health
care costs. Total annual COI due to STEC Several determinants have been described as
O157 infection for the Dutch society was es- risk factors that may play a role in the out-
timated at 9.1 million. The authors concluded come of an STEC infection, such as the initial
that, compared to other food-borne patho- bacterial inoculum, the amount and type of
gens, STEC O157 infections result in a rela- Stx produced, and the serotype of the infect-
tively low burden and low annual costs at the ing strain; the ability of E. coli to horizontally
societal level, but the burden and costs per acquire specific genetic elements known as
case are high. pathogenicity islands (PAI), and stx genes
In Australia, McPherson et al. (24) esti- from free bacteriophages in the environment
mated the cost of STEC infections at approx- and in the mammalian hosts; and the improved
imately AUD$2.6 million each year. adaptation of the bacteria to human hosts (26).
In Argentina, Caletti et al. (25) evaluated Among STEC strains, E. coli O157:H7 has
the direct health care costs and indirect become a significant food-borne pathogen,

366 RIVAS ET AL.
exhibiting some characteristic features, such outbreak in the United States in 2006 (30),
as low infectious dose (100 to 500 orga- several studies were performed in the north-
nisms) and acid tolerance that certainly ern hemisphere countries to assess the fre-
favors their transmission to humans by the quency of this clade in human disease and
food chain. In addition to STEC O157, only a cattle. However, no data were available from
restricted number of STEC serotypes (mainly the southern hemisphere countries.
those of seropathotype B) are associated with As notable differences were observed in
outbreaks and HUS. Moreover, Stx types dif- the prevalence and severity of human diseases
fer in their biological activity and associa- caused by O157 isolates in Argentina (high)
tion with disease. There is evidence of a and Australia (low), Mellor et al. (31) com-
linkage of Stx2a (formerly Stx2) with a higher pared human and bovine O157 isolates from
risk of severe human disease. The presence of both countries. The locus-specific polymor-
both stx2a and eae genes in an STEC isolate phism analysis genotyping revealed that line-
is considered to be a predictor of HUS (27). age I/II (LI/II) E. coli O157 isolates were
However, it has been shown that highly the most prevalent in Argentina (88%) and
pathogenic strains producing Stx2dactivatable Australia (88%). Argentinean LI/II isolates
are eae negative (28). were shown to belong to clades 4 (30%) and
Besides Stx and lipopolysaccharide, other 8 (65%) while Australian LI/II isolates were
putative virulence factors, including adhesins, identified as clades 6 (15%), 7 (80%), and
other toxins and proteases are required to 8 (2%). In Argentina, clade 8 isolates domi-
develop disease. Extensive evidence suggests nated in both cattle (50%) and humans (80%);
that a major pathogen determinant is the pres- meanwhile, in Australia clade 7 dominated in
ence of specific PAIs. A number of PAIs, in- both cattle (70%) and humans (90%).
cluding the locus of enterocyte effacement
(LEE), play a major role in enhancing the
Host Factors
ability of some serotypes to cause severe hu-
man disease. In addition to the proteins en- Several host factors influence the risk of ac-
coded in the LEE, the type III secretion quiring STEC infection, including age, im-
system also secretes other effectors encoded munity, health status, the use of antibiotics
outside the LEE. At least three of these non- and antimotility agents, stress, and genetic
LEE-encoded effectors have been linked to factors.
non-O157 STEC strains that cause HUS (29). The highest age-specific frequency of HUS
As the population of STEC O157 strains is in infants and young children. It declines
increased in frequency and spread geograph- with increasing age and increases again in the
ically, it has genetically diversified. Iso- elderly, probably due to changes in the im-
lates of STEC O157 from clinical and bovine munity status. Gastric acidity is an important
sources have been shown to be genotypically initial host barrier to ingested pathogens. Its
diverse by different methods, including PFGE, protective role against E. coli O157:H7 infec-
octomer-based genome scanning, and multi- tion has been suggested because individuals
locus variable number of tandem repeats with low gastric acidity are at a significantly
analysis. Studies of prophage and prophage higher risk for HUS developing than those
remnants in STEC O157 strains have indicated with normal physiological gastric function.
that genotypic diversity is largely attributable The possible role of stress as a risk factor
to bacteriophage-related insertions, deletions, for severe disease, and host genetic factors
and duplications of variable sizes of DNA that may influence host-pathogen interac-
fragments (30). tions, including the innate immune response
After the description of the hypervirulent to infection and the nature of the toxin-cell
clade of O157 associated with the raw spinach interaction, have been described. The genes

that regulate the gut colonization by E. coli In the animal production environment, the
O157:H7 may be modulated by hormone-like prevalence of other STEC serogroups, mainly
soluble factors produced by other bacteria in a those named big six (O26, O45, O103, O111,
process known as quorum sensing. The quo- O121, O145), is less well known. Recent stud-
rum-sensing pathway could be activated by ies have shown that only a small fraction of
host stress hormones such as epinephrine and strains from those O groups isolated from
norepinephrine (32). cattle carry Stx genes (46, 47).
Risk of infection could be increased in
spring and summer, as most reports state that
Behavior and Cultural Factors
during warmer months there is a higher
Since the emergence of STEC, case-control prevalence of STEC in cattle (48). The risk of
and population-based studies, varying in sizes colonization and shedding through the use
and rigor, have been conducted to examine of different cattle diets is a complex issue,
risk factors in associated infections. thoroughly revised lately (49). Attention has
In the 1990s, studies to evaluate risk factors been given to finishing diets with an increased
were focused on sporadic cases of E. coli O157 ratio of grains that could enhance STEC
infection. The consumption of undercooked O157 shedding. Inclusion of orange peel (50)
hamburgers and meat, eating in restaurants or or a dietary shift to forage has been proposed
fast-food establishments, living or working on to decrease STEC O157 shedding. However,
or visiting a cattle farm, drinking untreated much less is known about the effect of diets on
surface water, swimming in contaminated the ecology of other STEC serotypes.
water, contact with animal feces, and con- The emergence of super-shedding bovines
sumption of raw milk were the main risk is a relevant risk for STEC O157 contamination
factors identified. of the beef supply chain. Super-shedders have
In a literature review of several articles been linked to the diversity of STEC O157
published in the past decade, it was observed prevalence in cattle populations (51) and to the
that several studies were conducted to learn increasing spread of hide contamination in
more about the risk factors of both non-O157 feedlot cattle (52). These studies and others
and STEC O157 infections. The main findings (53) have shown that by controlling super-
of some of these studies are summarized in shedding bovines there would be a high impact
Table 1. on preventing STEC O157 infection in humans.
STEC strains are able to survive some
months in the environment, feces slurries, and
Cattle Management Factors
cattle manure (5456). A recent and system-
The role of cattle as a source of human in- atic review (57) concluded that there is a com-
fection has been extensively studied and re- plex relationship among animal reservoirs,
viewed, mostly in reference to STEC O157 pathogens, and the environment leading to the
(42). Prevalence of STEC O157 in cattle feces contamination of fresh produce from the en-
and hides is highly variable (43), dependent vironment. Manure and fecal contamination
on region, farm and cattle type, age, and sea- of irrigation water were the most important
son, among other factors. The degree of herd media for STEC presence in fresh produce
infection is uneven, with herds usually being at preharvest. The importance of minimizing
colonized by a low number of predominant these sources of contamination has been high-
strains (44). In Argentina, a similar pattern lighted by STEC outbreaks with fresh produce
was found at the abattoirs, where STEC O157 such as spinach (58).
was detected in approximately 15% of arriving Mapping studies have given epidemio-
lots, with an average prevalence in feces of logical evidence of the significance of envi-
4.1% (45). ronmental contamination for STEC human

TABLE 1 Risk factors for STEC infections identied by case-control studies and population-based studies, 2000-2012
368
Year of Advice to change behaviors

study Country Study design and size Risk factors identied Conclusions and reduce health risks Reference
Chap18.proof.3d
20002001 France Matched case-control study to Eating undercooked ground Adequate cooking of ground Proper cooking of ground beef 33
evaluate risk factors for sporadic beef; contact with a person beef may reduce the incidence (National Public Health
368
HUS cases in children. 105 cases; with diarrhea; drinking well of STEC infection. Campaign, 2006).
RIVAS ET AL.
196 controls. water during the summer period Assiduous attention to hygienic Revision of safety control
measures to prevent the spread measures at all levels of the
of STEC within families and ground beef food chain.
child-care facilities has the Specic education program for
potential to further reduce HUS professionals involved in the
episodes in childhood. meat industry.
20012002 Argentina Matched case-control study to Eating undercooked beef at Meat-related dietary habits and Apply effective safe practices at 34
evaluate risk factors for sporadic any place; contact with a young animal exposures were linked all stages of the food chain
STEC infections in children child with diarrhea; attending a to illness. (industry, government, and
enrolled in two sites, Mendoza day care center or kindergarten; Person-to-person spread was an consumers).
(urban and semirural area) and living in or visiting a place with important mode of transmission. Ensure that beef is
Buenos Aires (urban area). farm animals; contact with farm Some risk factors were specic well cooked at home and
150 cases; 299 controls. animals and cattle manure; by location. outside home.
having nonparental income Eating a wider variety of fruits Establish an educational
and vegetables and washing program to avoid risks of dietary

hands after handling raw beef, habits and behaviors,
especially with soap and water, and recommendations to
were protective factors. protect people through
hand washing.
20012003 Germany Matched case-control study to Contact with small ruminants Risk factors were age specic. Modify the upper pH limit (5.6) 35
evaluate risk factors for sporadic and consuming raw milk In children, food-borne of raw spreadable sausage
STEC infections in different age (children <3 years); transmission played a lesser role because STEC survives under
groups. playing in a sandbox (children in both acquiring STEC infection acidic conditions (pH 4.0).
202 cases; 202 controls. <3 years and aged 39 years); and developing HUS.
swimming in nonpublic Consumption of lamb meat and
swimming pools (children aged raw spreadable sausages was
39 years); consuming lamb identied as risk factor for the
meat and raw fermented rst time.
spreadable sausages
(cases 10 years or older)
03/30/15 18:05
19972006 Finland Population-based study applying Increased occurrence and Socioeconomic factors, Applying good hygiene in 36
a statistical model to distinguish incidence: proportion of like low-level income and animal and slaughter process
between risk factors for beef cattle to human population; education, were important for and food retail.
occurrence and incidence of proportion of population with acquiring STEC infections. Consumer education.
Chap18.proof.3d
STEC diseases. higher education related to Ecological factors such as the Proper food handling.
369
131 cases. consumption habits of relation between population and Up-to-date national legislation
undercooked meat. density of beef cattle were and regulation.
Increased incidence: important for the incidence of
proportion of fresh water; the disease.
number of cultivated farms;
proportion of low-income
households with children
20032007 Australia Case-control study to evaluate STEC O157: Consumption of Risk factors were serogroup Design educational programs for 37
risk factors in sporadic O157 and hamburgers; eating at specic. people who live or work with
non-O157 STEC infections. restaurants; having occupational Hamburgers and ground beef animals or raw meat, as well as
213 cases; 304 controls. exposure to red meat. have not been implicated in for those who enjoy camping.
Non-O157 STEC (O111, O26, outbreaks of STEC and have not Recommend that consumers
O103, O113, O172): occupational been previously considered a and food handlers cook
exposure to animals; source of infection. hamburgers thoroughly.
consumption of sliced processed
chicken meat and sliced corned
beef; eating out at a catered

event; camping in the bush
20002009 United Population-based study to O157 and O103 STEC: eating Some STEC serogroups such as Continue population-level 38
States compare risk factors in patients pink hamburger; handling raw O103 seem to have an monitoring of the epidemiology
infected with O157 and ground beef. epidemiological and exposure of STEC to determine
non-O157 STEC. Non-O157 STEC (O111, O103, prole similar to O157, and they longer-term trends and
392 patients. O26): international travel likely occupy a similar ecologic opportunities for control.
in the 7 days prior to symptoms. niche.
Non-O157 STEC (O111, O103, Other serogroups are quite
O26, O45): consumption of different (O45) and may not be
untreated surface water able to be managed through
identical measures to control
O157.
CHAPTER 18 Risk Factors for STEC-Associated Human Diseases

369
03/30/15 18:05
TABLE 1 Risk factors for STEC infections identied by case-control studies and population-based studies, 2000-2012 (Continued)
370

Chap18.proof.3d
20022009 Argentina Case-control study to identify Eating food prepared outside Protective effects of a diet that Plan strategies for local 39
risks factors for sporadic STEC home includes vegetables. prevention to diminish the
370
infections in children aged up to incidence of HUS in the region
RIVAS ET AL.
6 from the Central Eastern area. under study.

63 cases; 374 controls
2007 Argentina Epidemiological survey to Eating commercially prepared Differences in the frequency of Improve educational programs 40
evaluate risk factors for STEC precooked or homemade hamburger consumption were to enhance personal hygiene,
infections in different hamburgers, exposure to water observed among children from adequate meat handling and
socioeconomic groups. of swimming pools; no hand different socioeconomic strata. cooking techniques,
883 students aged 1012 years wash after going to the Children from high and medium and maintenance of safe
from elementary public schools toilet or before eating food strata attending private recreational water.
of an urban area of Buenos Aires swimming pools and children
Province. from low stratum attending
public swimming pools were
at risk.
20032012 Argentina Epidemiological survey to Eating undercooked ground Food, contact with rural workers, Improve the surveillance system, 41
evaluate risk factors for sporadic beef, sausages, barbecue, and environment were risk considering the particular
HUS cases and STEC infections in and unpasteurized milk, factors for children. conditions of the region.

urban and semirural area of contact with farm animals; Establish a health education
Ro Negro Province. poor hygiene in food and father program sustained over time.
42 cases. with rural activities, and poor
hygiene of work clothes
03/30/15 18:05
infection in rural areas (5961). In Argentina, the emergence of outbreaks with different
Tanaro et al. (62) have shown a degree of epidemiological characteristics and for the
contamination of surface waters in rural areas widespread epidemics are (i) the develop-
of a cattle-producing region of Entre Rios ment of new food processing technologies and
Province. foods, (ii) the more centralized and rapid food
Armstrong et al. (63) suggested, as a prob- distribution systems, (iii) the changes in con-
able cause for the emergence of STEC O157, sumer preferences and behaviors, and (iv) the
the changes in modern livestock and food considerable increase in the volume of food
processing industries, characterized by the products traded internationally. Additionally,
concentration, homogenization, and increased the enormous increase in global travel allows
scale of operations. Within the abattoir, the individuals to be infected in one country and
hide-removal operation is a critical point for to become ill on their return to their country
carcass contamination (53). In addition, lairage of origin (70).
areas have been identified as key for STEC This epidemiological change was also in-
dissemination as most isolates from carcasses fluenced by the genetic variation and re-
were traceable to the lairage environment lentless evolution of the O157 pathogen
rather than to the original feedlot (64). population (71). Mellmann et al. (72) remarked
Studies conducted in Argentina showed that bacterial evolution is an ongoing process
that 11 STEC strains (ten O157:H7 and one that undoubtedly will lead to the emergence
O178:H19) isolated in abattoirs had similar of other successful pathogenic clones of E. coli
phage type-XbaI-PFGE pattern-stx genotype in the future.
profile as those responsible for 19 HUS cases Since the emergence of STEC as a food-
in the same period (45, 65). For STEC O157 it borne pathogen in 1982 (1), large outbreaks of
was possible to link 12% of reported human gastrointestinal disease, involving numerous
infections to the bovine reservoir (66). In the persons and associated with different sources
same study, it was estimated that, at slaughter, and vehicles of transmission, have been de-
ca. 38,000 bovines would carry STEC O157 in scribed worldwide.
their feces per each clinical case of HUS and Each outbreak showed different aspects of
bloody and nonbloody diarrhea cases reported. complexity, from the detection of the source
Risk factors identified in cattle production and the vehicles of transmission to the need
and at the abattoir can be integrated as com- for the development of control strategies to
ponents of risk assessment models that allow avoid the occurrence of new cases. These
estimates of the risk of STEC infection for the large outbreaks were of great concern and
population from consumption of a specific challenge for the public health system, com-
meat product (67). They are also useful tools pelling the food regulatory and health agen-
to evaluate the outcome of mitigation strate- cies and the food industry to establish
gies in cattle production and the meat indus- improved guidelines to control and prevent
try (67, 68). Risk assessment studies have also new incidents.
been conducted to estimate the risk of STEC Large outbreaks have been commonly no-
in ground beef hamburgers in Argentina (69). tified in industrialized countries like the
United States (73), Japan (74), Germany (75),
Canada (76), among others. In these countries,
NEW SCENARIO AND LEARNED LESSONS the advanced epidemiological system has con-
FROM OUTBREAKS tributed to a better investigation and under-
standing of these events. Furthermore, small
The epidemiological profile of food-borne dis- indoor outbreaks were described worldwide,
eases has changed dramatically in recent mainly in families and child-care centers with
decades. Some of the contributing factors for a lower frequency (77, 78).

372 RIVAS ET AL.
In this context, it is interesting to point because the product was exported to Canada
out the lessons learned about risk factors, and Mexico. During the epidemiological sur-
treatment, diagnosis, and advances in health vey, most of the patients (95%) reported
and food regulations from some emblematic consuming uncooked fresh spinach during the
outbreaks (Table 2). 10 days before the onset of the illness (58).
The large outbreak of E. coli O157:H7 in- Women (71%) were the most affected, prob-
fection that occurred in four western states of ably reflecting that women are more likely to
the United States in February 1993 was the consume fresh vegetables. Moreover, the out-
first event with a great population and media break was related to the practice of con-
impact. For the first time, people were mas- suming prewashed, bagged leafy greens, as
sively informed about good practices for food the general public assumed food handling
handling regarding this pathogen. The out- in farms and in processing facilities to be
break investigation allowed us to know about safe. The sequencing of the genome of the
the low infectious dose of this organism, its TW14359 strain, responsible of this outbreak,
capacity to survive for a long time (more than helped identify the genetic factors that en-
2 months) in frozen storage, and the temper- hanced the ability of this strain to cause such a
ature and time combinations (68C, 15 sec) high number of HUS cases (82).
that ensure a safe hamburger cooking con- Manning et al. (30) published a comparison
dition. These findings were used as strong among outbreaks with different severity of
arguments to enforce a zero tolerance policy the reported illness. The 1993 outbreak in
for this microorganism in processed food western North America and the large 1996
and for the need for a pronounced decrease outbreak in Japan had low rates of hospitali-
in the contamination of raw ground beef (73, zation and HUS, whereas the 2006 North
79, 80). American spinach outbreak had high rates of
At present, the Sakai outbreak of E. coli both hospitalization (50%) and HUS (10%).
O157 infection in July 1996 was the largest Single nucleotide polymorphism genotyping
outbreak ever experienced, because of the revealed the genetic variability among patho-
number of people affected (74). As described genic strains associated with clinical infec-
by Fukushima et al. (81), lunch foods con- tion. Their results support the hypothesis
taminated with E. coli O157 and supplied to that the clade 8 lineage has recently acquired
elementary schools by a centralized distribu- novel factors that contribute to the enhanced
tion system were the cause of this massive virulence.
outbreak. It is interesting to note the clinical The massive outbreak of bloody diarrhea
experiences in the treatment received by the and HUS that occurred in 2011 in Germany
affected children. Almost all patients were and other 13 European countries, the United
treated with antibiotics, fosfomycin and lacto- States, and Canada was a challenge for clini-
bacilli, from the early days of the illness. An cians and microbiologists given its atypical
evaluation of the outcomes revealed that these presentation. The seriousness of the illness
treatments were more than satisfactory com- and the fatalities, coupled with the lack of a
pared to previous reports. However, the fa- definitive source of the causative agent, cre-
vorable outcome could not be only attributed ated a highly negative impact on the popula-
to the effectiveness of antibiotics. Other fac- tion and gained the front page of newspapers
tors, like age and racial differences, might also around the world (83).
be responsible. This unprecedented outbreak affected
The spinach outbreak in the United States mainly adults (90%), predominantly women,
in July 2006 highlighted the importance of and resulted in an unusually high number of
fresh produce as a vehicle in STEC infec- HUS cases (n=855). The augmented adher-
tions and also the role of international trade ence of the strain to the intestinal epithelium

facilitating systemic absorption of Stx could CONCLUSIONS

explain the high progression to HUS. The
outbreak provided important new insight into Differences in the frequency and the severity
novel antibiotic strategies in the treatment of of STEC-associated human diseases are ob-
HUS in adults and for decolonization of long- served from country to country. The more
term STEC carriers (84). reliable information is provided by developed
The characterization of the O104:H4 out- countries as better enteric pathogen surveil-
break strain revealed an unusual combination lance systems are in place. Because of the se-
of pathogenic features typical of enteroaggre- verity and the long-term sequelae of STEC-
gative E. coli combined with the capacity to associated illnesses, they have a high social
produce Stx. Additionally, isolates displayed and economic cost for both the affected fam-
an extended-spectrum b-lactamase pheno- ilies and the health system. In Argentina, be-
type, carrying plasmid-borne blaCTX-M-15 cause of the number of HUS cases reported
and blaTEM-1 genes (85). DNA sequencing each year, those social and economic costs are
data rapidly revealed that outbreak strain particularly significant.
was a new hybrid of two types of pathogenic The risks for acquiring an STEC infection
E. coli. Up to date, nine O104:H4 isolates have are associated with several determinants of
been sequenced (https://github.com/ehec- the pathogen and its reservoir and with bio-
outbreak-crowdsourced/BGI-data-analysis/ logical and cultural factors of the host. The
wiki; http://www.bgisequence.com/eu/index. best knowledge about risk factors was ob-
php?cID=194). Different methodologies for tained from case-control and population-
detection and characterization of the O104:H4 based studies. Main risk factors identified in
strain in clinical and food samples were de- earlier studies were dietary behaviors related
veloped in a short time (85, 86). Possible to beef consumption, but at present they in-
mitigation options for safe consumption of clude a wider range of foods, such as fresh
raw vegetables were advised after a fast-track produce or sprouts. Other risky behaviors
assessment of the consumer exposure to identified have been connected to environ-
STEC through this type of food (87). mental sources, as living in, working in, or
A restaurant cohort study and the trace- visiting rural areas; swimming and camping in
back and trace-forward data analysis of the recreational areas; and being in contact with
Task Force EHEC contributed to the identi- farm animals. Risk factors for STEC infection
fication of fenugreek seeds as sources of have also been identified in cattle manage-
transmission. This study also contributed to ment and at the abattoir, such as the effect
confirm the implications of international trade of finishing diets, the existence of super-
in food-borne outbreaks (88). shedders in a herd, the cross-contamination
One of the most important lessons from the in lairage areas, and the hide-removal step
O104 outbreak is the successful cooperation at the abattoir, among the most important
among health and food networks and agen- ones. Another important risk factor is person-
cies. This type of work could in a short time to-person transmission, especially for young
produce valuable epidemiological and micro- children. In Argentina, beef is a traditional
biological information, essential for develop- component of diet, with an average consump-
ing public health measures to improve the tion of ca. 62.5 kg per capita per year. This
management of future outbreak situations high rate of consumption, and particularly
(75). This outbreak emphasized the impor- some meat-related dietary habits, could be
tance of common alert and surveillance sys- risk factors for STEC illness in our population.
tems for the early detection of international A new risk scenario has emerged in the last
outbreaks and for a better assessment of their decades due to the bacterial evolution that
spread. gave rise to the emergence of hypervirulent

374
Chap18.proof.3d
374
RIVAS ET AL.
TABLE 2 Features of major food-borne outbreaks associated with Shiga toxin-producing E. coli
No. of No. of No. of No. of Associated
Outbreak cases hosp. HUS deaths pathogen Impact Vehicle Risk factors Lessons learned
a a a a
United States, 501 151 45 3 STEC O157:H7 First large outbreak Burger Centralized productionImportance of:
4 western states (stx1a/stx2a) associated with beef and distribution The low infectious dose.
19921993 product. system. The survival in frozen storage.
Different restaurants of Inadequate cooking. The safe cooking conditions.
the same food chain PFGE as subtyping technique
involved. for molecular surveillance.
High number of people The mandatory notication of
affected. E. coli O157 infections.
Population and media Good hygienic practices in food
impact. handling.
Updating beef-related regulations.
Japan, 12,680b 398c 121b 3b STEC O157:H7 Massive and Radish sprouts, Centralized distribution Treatment of bloody diarrhea
Sakai City (stx1a/stx2a) widespread outbreak. among others. of school lunch food. with fosfomycin and lactobacilli

1996 Children in elementary Raw vegetables. from the early days of the illness.
schools in different Advances in food regulations.
districts affected. Food monitoring programs.
Specic sanitization control
measures in catering facilities.
Hygiene guidelines for cooking
practices at home.
United States, 205d 95d 29d 2d STEC O157:H7 Large outbreak with Fresh spinach. Consumption of Recognition of fresh produce as
26 states 2006 (stx2a) high number of prewashed, bagged vehicles of transmission.
hospitalizations and fresh produce. Genome sequence and new
HUS cases. Emergence of virulence determinants
hypervirulent E. coli recognized.
O157 strain. Advances in the knowledge of
International trade. hypervirulent clades.
03/30/15 18:05
No. of No. of No. of No. of Associated
Outbreak cases hosp. HUS deaths pathogen Impact Vehicle Risk factors Lessons learned
e f e e
Germany and 3,842 ND 855 53 STEC O104:H4 Large outbreak with: Fenugreek Eating salads. New methodologies for STEC
other 13 (stx2a) High proportion of Sprouts/seeds. Emergence of O104 detection in food and
European adults affected, mainly hypervirulent hybrid humans.
Chap18.proof.3d
countries, women. EAEC-STEC O104 strain, Sequencing by new-generation
United States, Unusually high number with enhanced technologies.
375
and Canada of HUS in adults. colonization and New therapeutic approaches for
2011 Patients with severe long-term shedding. adults: azithromycin treatment for
neurological symptoms HUS and decolonization of
followed by death. long-term STEC carriers.
Cooperation and teamwork of
networks.
New vision: Emergence of a new
E. coli. pathotype.
Any STEC could be potentially
pathogenic for humans.
Open-source data release,
and prompt crowd-sourced
analyses.
a
Data from reference 79.
b
c
d
e

f
ND, no data.
CHAPTER 18 Risk Factors for STEC-Associated Human Diseases
375
03/30/15 18:05
376 RIVAS ET AL.
O157 clones with a worldwide distribution, 4. Karmali MA, Petric M, Lim C, Fleming PC,
and other STEC strains with unusual combi- Arbus GS, Lior H. 1985. The association between
hemolytic uremic syndrome and infection by
nations of pathogenic features, such as the verotoxin-producing Escherichia coli. J Infect Dis
O104:H4 strain. The epidemiological changes 151:775782.
were also influenced by the increase in cen- 5. Palermo MS, Exeni RA, Fernndez GC. 2009.
tralized food production and distribution Hemolytic uremic syndrome: pathogenesis and
update of interventions. Expert Rev Anti Infect
systems and the growth in the volume of in-
Ther 7:697707.
ternational trade of food ingredients.
6. Caprioli A, Morabito S, Brugre H, Oswald
The learned lessons from large and em- E. 2005. Enterohaemorrhagic Escherichia coli:
blematic outbreaks could be summarized as emerging issues on virulence and modes of trans-
(i) the advances in the knowledge of virulence mission. Vet Res 36:289311.
determinants of new pathogenic strains; (ii) 7. Erickson MC, Doyle MP. 2007. Food as vehicle
for transmission of Shiga toxin-producing Esche-
the recognition of new vehicles of infection;
richia coli. J Food Prot 70:24262449.
(iii) the development of new methodologies 8. Guth BEC, Prado V, Rivas M. 2010. Shiga toxin-
for STEC detection in foods and humans; producing Escherichia coli, p 6583. In Torres
(iv) the improvement of food regulations and AG (ed), Pathogenic Escherichia coli in Latin
hygiene guidelines; (v) the new therapeutic America. Betham Science Publishers Ltd., Sharjah,
approaches in the treatment of STEC-infected United Arab Emirates.
9. Mora A, Blanco M, Blanco JE, Dahbi G, Lpez
patients, especially HUS in adults; (vi) the
C, Justel P, Alonso MP, Echeita A, Bernrdez
establishment of continuous educational pro- MI, Gonzlez EA, Blanco J. 2007. Serotypes,
grams for food consumers; and (vii) the en- virulence genes and intimin types of Shiga toxin
hanced cooperation and teamwork of regional (verocytotoxin)-producing Escherichia coli iso-
and international networks. lates from minced beef in Lugo (Spain) from 1995
through 2003. BMC Microbiol 1:713.
10. Karmali MA, Mascarenhas M, Shen S, Ziebell
ACKNOWLEDGMENT K, Johnson S, Reid-Smith R, Isaac-Renton J,
Clark C, Rahn K, Kaper JB. 2003. Association of
We declare no conflicts of interest with regard genomic O island 122 of Escherichia coli EDL 933
to the manuscript. with verocytotoxin-producing Escherichia coli
seropathotypes that are linked to epidemic and/or
serious disease. J Clin Microbiol 41:49304940.
CITATION 11. Centers for Disease Control and Prevention.
2012. Foodborne Diseases Active Surveillance
Rivas M, Chinen I, Miliwebsky E, Masana M. Network (FoodNet): FoodNet Surveillance Report
2014. Risk factors for Shiga toxin-producing for 2011 (Final Report). Atlanta, Georgia: U.S.
Escherichia coli-associated human diseases. Department of Health and Human Services,
Centers for Disease Control and Prevention,
Microbiol Spectrum 2(5):EHEC-0002-2013. Atlanta, GA. http://www.cdc.gov/foodnet/PDFs/
2011_annual_report_508c.pdf. Accessed 25 June
2014.
REFERENCES
12. Government of Canada, National Integrated
1. Riley LW, Remis RS, Helgerson SD, McGee HB, Enteric Pathogen SurveillanceSystem (C-
Wells JG, Davis BR, Hebert RJ, Olcott ES, EnterNet). 2012. 2011 Short Report. Guelph, ON:
Johnson LM, Hargrett NT, Blake PA, Cohen Public Health Agency of Canada, Guelph, ON.
ML. 1983. Hemorrhagic colitis associated with a http://www.phac-aspc.gc.ca/c-enternet/publications-
rare Escherichia coli O157:H7 serotype. N Engl J eng.php#a2. Accessed 17 April 2013.
Med 308:681685. 13. European Centre for Disease Prevention and
2. Gyles CL. 2007. Shiga toxin-producing Esche- Control. 2013. Annual epidemiological report 2012.
richia coli: an overview. J Anim Sci 85:E45E62. Reporting on 2010 surveillance data and 2011 epi-
3. Karmali MA, Steele BT, Petric M, Lim C. 1983. demic intelligence data. ECDC, Stockholm, Sweden.
Sporadic cases of hemolytic uremic syndrome http://ecdc.europa.eu/en/Publications/Annual-
associated with fecal cytotoxin and cytotoxin Epidemiological-Report-2012.pdf. Accessed 17
producing Escherichia coli. Lancet i:619620. April 2013.

14. Vally H, Hall G, Dyda A, Raupach J, Knope K, 26. Beutin L. 2006. Emerging enterohaemorrhagic
Combs B, Desmarchelier P. 2012. Epidemiology Escherichia coli, causes and effects of the rise of
of Shiga toxin-producing Escherichia coli in a human pathogen. J Vet Med B Infect Dis Vet
Australia, 20002010. BMC Public Health 12:63 Public Health 53:299305.
74. 27. Boerlin P, McEwen SA, Boerlin-Petzold F,
15. Guth BEC, Picheth CF, Gomes TAT. 2010. Wilson JB, Johnson RP, Gyles CL. 1999. Asso-
Escherichia coli situation in Brazil, p 162178. In ciations between virulence factors of Shiga toxin-
Torres AG (ed), Pathogenic Escherichia coli in producing Escherichia coli and disease in humans.
Latin America. Betham Science Publishers Ltd., J Clin Microbiol 37:497503.
Sharjah, United Arab Emirates. 28. Bielaszewska M, Friedrich AW, Aldick T,
16. Varela G, Gmez-Duarte OG, Ochoa T. 2010. Schurk-Bulgrin R, Karch H. 2006. Shiga toxin
Diarrheigenic Escherichia coli in children from activatable by intestinal mucus in Escherichia
Uruguay, Colombia and Per, p 209222. In coli isolated from humans: predictor for a severe
Torres AG (ed), Pathogenic Escherichia coli in clinical outcome. Clin Infect Dis 43:11601167.
Latin America. Betham Science Publishers Ltd., 29. Coombes BK, Wickham ME, Mascarenhas M,
Sharjah, United Arab Emirates. Gruenheid S, Finlay BB, Karmali MA. 2008.
17. Lfdahl S. 2008. How global is VTEC?, p 6567. Molecular analysis as an aid to assess the public
Epidemiology and Transmission of VTEC and health risk of non-O157 Shiga toxin-producing
Other Pathogenic Escherichia coli. Swedish In- Escherichia coli strains. Appl Environ Microbiol
stitute for Infectious Disease Control, Stockholm, 74:21532160.
Sweden. 30. Manning SD, Motiwala AS, Springman AC, Qi
18. Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV, W, Lacher DW, Ouellette LM, Mladonicky JM,
Widdowson MA, Roy SL, Jones JL, Griffin PM. Somsel P, Rudrik JT, Dietrich SE, Zhang W,
2011. Foodborne illness acquired in the United Swaminathan B, Alland D, Whittam TS. 2008.
Statesmajor pathogens. Emerg Infect Dis 17: Variation in virulence among clades of Esche-
715. richia coli O157:H7 associated with disease out-
19. Frenzen PD, Drake A, Angulo FJ, the Emerging breaks. Proc Natl Acad Sci USA 105:48684873.
Infections Program Foodnet Working Group. 31. Mellor GE, Sim EM, Barlow RS, DAstek BA,
2005. Economic cost of illness due to Escherichia Galli L, Chinen I, Rivas M, Gobius KS. 2012.
coli O157 infections in the United States. J Food Phylogenetically related Argentinean and Aus-
Prot 68:26232630. tralian Escherichia coli O157 isolates are distin-
20. Buzby JC, Roberts T. 2009. The economics guished by virulence clades and alternative Shiga
of enteric infections: human foodborne disease toxin 1 and 2 prophages. Appl Environ Microbiol
costs. Gastroenterology 136:18511862. 78:47244731.
21. Scharff RL, McDowell J, Medeiros L. 2009. 32. Clarke MB, Sperandio V. 2005. Events at the
Economic cost of foodborne illness in Ohio. host-microbial interface of the gastrointestinal
J Food Prot 72:128136. tract III. Cell-to-cell signalling among microbial
22. Havelaar AH, van Duynhoven YTHP, Nauta flora, host, and pathogens: there is a whole lot of
MJ, Bouwknegt M, Heuvelink AE, de Wit talking going on. Am J Physiol Gastrointest Liver
GA, Nieuwenhuizen MGM, van de Kar Physiol 288:G1105G1109.
NCAJ. 2004. Disease burden in The Netherlands 33. Vaillant V, Espi E, de Valk H, Durr U,
due to infections with Shiga toxin-producing Barataud D, Bouvet P, Grimont F, Desenclos
Escherichia coli O157. Epidemiol Infect 132:467 JC. 2009. Undercooked ground beef and person-
484. to-person transmission as major risk factors for
23. Tariq L, Haagsma J, Havelaar A. 2011. Cost of sporadic hemolytic uremic syndrome related to
illness and disease burden in The Netherlands Shiga-toxin producing Escherichia coli infections
due to infections with Shiga toxinproducing in children in France. Pediatr Infect Dis J 28:650
Escherichia coli O157. J Food Prot 74:545552. 653.
24. McPherson M, Kirk MD, Raupach J, Combs B, 34. Rivas M, Sosa-Estani S, Rangel J, Caletti MG,
Butler JRG. 2011. Economic costs of Shiga toxin- Valls P, Roldn CD, Balbi L, Marsano de
producing Escherichia coli infection in Australia. Mollar MC, Amoedo D, Miliwebsky E, Chinen I,
Foodborne Pathog Dis 8:5562. Hoekstra RM, Mead P, Griffin PM. 2008. Risk
25. Caletti MG, Petetta D, Jaitt M, Casaliba S, factors for sporadic Shiga toxin-producing Esch-
Gimenez A. 2006. Evaluacin de costos directos e erichia coli infections in children, Argentina.
indirectos del tratamiento del Sndrome Urmico Emerg Infect Dis 14:763771.
Hemoltico en sus distintas etapas evolutivas. 35. Werber D, Behnke SC, Fruth A, Merle R,
Medicina (Buenos Aires) 66(Suppl):2226. Menzler S, Glaser S, Kreienbrock L, Prager R,

378 RIVAS ET AL.
Tschape H, Roggentin P. 2007. Shiga toxin- 46. Kalchayanand N, Arthur TM, Bosilevac JM,
producing Escherichia coli infection in Germany: Wells JE, Wheeler TL. 2013. Chromogenic agar
different risk factors for different age groups. Am medium for detection and isolation of Escherichia
J Epidemiol 165:425434. coli serogroups O26, O45, O103, O121 and O145
36. Jalava K, Ollgren J, Eklund M, Sitonen A, from fresh beef and cattle feces. J Food Prot
Kuusi M. 2011. Agricultural, socioeconomic and 76:192199.
environmental variables as risks for human vero- 47. Thomas KM, McCann MS, Collery MM,
toxigenic Escherichia coli (VTEC) infection in Logan A, Whyte P, McDowell DA, Duffy G.
Finland. BMC Infect Dis 11:275. 2012. Tracking verotoxigenic Escherichia coli
37. McPherson M, Lalor K, Combs B, Raupach J, O157, O26, O111, O103 and O145 in Irish cattle. Int
Stafford R, Kirk MD. 2009. Serogroup-specific J Food Microbiol 153:288296.
risk factors for Shiga toxinproducing Escherichia 48. Barkocy-Gallagher GA, Arthur TM, Rivera-
coli infection in Australia. Clin Infect Dis 49:249 Betancourt M, Nou X, Shackelford SD, Wheeler
256. TL, Koohmaraie M. 2003. Seasonal prevalence of
38. Hadler JL, Clogher P, Hurd S, Phan Q, Shiga toxin-producing Escherichia coli including
Mandour M, Bemis K, Marcus R. 2011. Ten-year O157 and non-O157 serotypes, and Salmonella in
trends and risk factors for non-O157 Shiga toxin- commercial beef processing plants. J Food Prot
producing Escherichia coli found through Shiga 66:19781986.
toxin testing, Connecticut, 20002009. Clin Infect 49. Callaway TR, Carr MA, Edrington TS, Anderson
Dis 53:269276. RC, Nisbet DJ. 2009. Diet, Escherichia coli
39. Rivero MA, Passucci JA, Rodriguez EM, O157:H7, and cattle: a review after 10 years. Curr
Signorini ML, Tarabla HD, Parma AE. 2011. Issues Mol Biol 11:6779.
Factors associated with sporadic verotoxigenic 50. Callaway TR, Carroll JA, Arthington JD,
Escherichia coli infection in children with diar- Edrington TS, Rossman ML, Carr MA, Krueger
rhea from the Central Eastern Area of Argentina. NA, Ricke SC, Crandall P, Nisbet DJ. 2011.
Foodborne Pathog Dis 8:901906. Escherichia coli O157:H7 populations in ruminants
40. Bentancor AB, Ameal LA, Calvino MF, can be reduced by orange peel product feeding.
Martinez MC, Miccio L, Osvaldo J, Degregorio J Food Prot 74:19171921.
OJ. 2012. Risk factors for Shiga toxin-producing 51. Matthews L, McKendrick IJ, Ternent H,
Escherichia coli infections in preadolescent school- Gunn GJ, Synge B, Woolhouse ME. 2006.
children in Buenos Aires, Argentina. J Infect Dev Super-shedding cattle and the transmission dy-
Ctries 6:378386. namics of Escherichia coli O157. Epidemiol Infect
41. Di Pietro S, Stafforini G, Cifone N, Alvarez ML, 134:131142.
Arellano O, Manzini S, Haritchabalet K, Nbile 52. Arthur TM, Keen JE, Bosilevac JM, Brichta-
M, Chinen I, Miliwebsky E, Rivas M, Larrieu E. Harhay DM, Kalchayanand N, Shackelford
2013. Epidemiologa del sndrome urmico SD, Wheeler TL, Nou X, Koohmaraie M. 2009.
hemoltico, Viedma, Provincia de Ro Negro Longitudinal study of Escherichia coli O157:H7
20032012. Rev Arg Zoonosis III:1115. in a beef cattle feedlot and role of high-level
42. Ferens WA, Hovde CJ. 2011. Escherichia coli shedders in hide contamination. Appl Environ
O157:H7: animal reservoir and sources of human Microbiol 75:65156523.
infection. Foodborne Pathog Dis 8:465487. 53. Arthur TM, Brichta-Harhay DM, Bosilevac JM,
43. Rhoades JR, Duffy G, Koutsoumanis K. 2009. Kalchayanand N, Shackelford SD, Wheeler TL,
Prevalence and concentration of verocytotoxigenic Koohmaraie M. 2010. Super shedding of Esche-
Escherichia coli, Salmonella enterica and Listeria richia coli O157:H7 by cattle and the impact on
monocytogenes in the beef production chain: a beef contamination. Meat Sci 86:3237.
review. Food Microbiol 4:357376. 54. Fukushima H, Hoshina K, Gomyoda M. 1999.
44. Renter DG, Sargeant JM, Oberst RD, Samadpour Long-term survival of Shiga toxin-producing
M. 2003. Diversity, frequency, and persistence of Escherichia coli O26, O111, and O157 in bovine
Escherichia coli O157 strains from range cattle en- feces. Appl Environ Microbiol 65:51775181.
vironments. Appl Environ Microbiol 69:542547. 55. Fremaux B, Prigent-Combaret C, Delignette-
45. Masana MO, Leotta GA, Del Castillo LL, Muller ML, Dothal M, Vernozy-Rozand C.
DAstek BA, Palladino PM, Galli L, Vilacoba E, 2007. Persistence of Shiga toxin-producing Esche-
Carbonari C, Rodrguez HR, Rivas M. 2010. richia coli O26 in cow slurry. Lett Appl Microbiol
Prevalence, characterization, and genotypic anal- 45:5561.
ysis of Escherichia coli O157:H7/NM from select- 56. Fremaux B, Delignette-Muller ML, Prigent-
ed beef exporting abattoirs of Argentina. J Food Combaret C, Gleizal A, Vernozy-Rozand C.
Prot 73:649656. 2007. Growth and survival of non-O157:H7 Shiga

toxin-producing Escherichia coli in cow manure. 67. Smith BA, Fazil A, Lammerding AM. 2013.
J Appl Microbiol 102:8999. A risk assessment model for Escherichia coli
57. Park S, Szonyi B, Gautam R, Nightingale K, O157:H7 in ground beef and beef cuts in Canada:
Anciso J, Ivanek R. 2012. Risk factors for mi- evaluating the effects of interventions. Food
crobial contamination in fruits and vegetables at Control 29:364381.
the preharvest level: a systematic review. J Food 68. Hurd HS, Malladi S. 2012. An outcomes model to
Prot 11:20552081. evaluate risks and benefits of Escherichia coli
58. Centers for Disease Control and Prevention. vaccination in beef cattle. Foodborne Pathog Dis
2006. Ongoing multistate outbreak of Escherichia 10:952961.
coli serotype O157:H7 infections associated with 69. Signorini M, Tarabla H. 2009. Quantitative risk
consumption of fresh spinachUnited States, assessment for verocytotoxigenic Escherichia coli
September 2006. Morbid Mortal Wkly Rep 55: in ground beef hamburgers in Argentina. Int J
10451046. Food Microbiol 132:153161.
59. Frank C, Kapfhammer S, Werber D, Stark K, 70. World Health Organization. 1997. Prevention and
Held L. 2008. Cattle density and Shiga toxin- control of enterohaemorrhagic Escherichia coli
producing Escherichia coli infection in Germany: (EHEC) infections. Report of a WHO Consultation.
increased risk for most but not all serotypes. World Health Organization, Geneva, Switzerland.
Vector Borne Zoonotic Dis 8:635643. 71. Robins-Browne RM. 2005. The relentless evo-
60. Strachan NJC, Dunn GM, Locking ME, Reid lution of pathogenic Escherichia coli. Clin Infect
TMS, Ogden ID. 2006. Escherichia coli O157: Dis 41:793794.
burger bug or environmental pathogen? Int J 72. Mellmann A, Bielaszewska M, Zimmerhackl
Food Microbiol 112:129137. LB, Prager R, Harmsen D, Tschpe H, Karch
61. Kistemann T, Zimmer S, Vagsholm I, Andersson H. 2005. Enterohemorrhagic Escherichia coli in
Y. 2004. GIS-supported investigation of human human infection: in vivo evolution of a bacterial
EHEC and cattle VTEC O157 infections in Sweden: pathogen. Clin Infect Dis 41:785792.
geographical distribution, spatial variation and pos- 73. Centers for Disease Control and Prevention.
sible risk factors. Epidemiol Infect 132:495505. 1993. Multistate outbreak of Escherichia coli O157:
62. Tanaro JD, Leotta GA, Lound LH, Galli L, H7 infections from hamburgerswestern United
Piaggio MC, Carbonari CC, Araujo S, Rivas M. States, 19921993. Morbid Mortal Wkly Rep
2010. Escherichia coli O157 in bovine feces and 42:258263.
surface water streams in a beef cattle farm of 74. Michino H, Araki K, Minami S, Takaya S, Sakai
Argentina. Foodborne Pathog Dis 7:475477. N, Miyazaki M, Ono A, Yanagawa H. 1999.
63. Armstrong GL, Hollingsworth J, Morris JG Jr. Massive outbreak of Escherichia coli O157:H7
1996. Emerging foodborne pathogens: Escherichia infection in schoolchildren in Sakai City, Japan,
coli O157:H7 as a model of entry of a new path- associated with consumption of white radish
ogen into the food supply of the developed world. sprouts. Am J Epidemiol 150:787796.
Epidemiol Rev 18:2951. 75. STEC Workshop Reporting Group. 2012.
64. Arthur TM, Bosilevac JM, Brichta-Harhay DM, Experiences from the Shiga toxin-producing
Kalchayanand N, King DA, Shackelford SD, Escherichia coli O104:H4 outbreak in Germany
Wheeler TL, Koohmaraie M. 2008. Source and research needs in the field, Berlin, 2829
tracking of Escherichia coli O157:H7 and Salmo- November 2011. Euro Surveill 17(7):pii=20091.
nella contamination in the lairage environment http://www.eurosurveillance.org/ViewArticle.
of commercial U.S. beef processing plants and aspx?ArticleId=20091. Accessed 24 April 2013.
identification of an effective intervention. J Food 76. Matsell DG, White CT. 2009. An outbreak of
Prot 71:17521760. diarrhea-associated childhood hemolytic uremic
65. Masana MO, DAstek BA, Palladino PM, Galli syndrome: the Walkerton epidemic. Kidney Int
L, Del Castillo LL, Carbonari C, Leotta GA, 75:S35S37.
Vilacoba E, Irino K, Rivas M. 2011. Geno- 77. Snedeker KG, Shaw DJ, Locking ME, Prescott
typic characterization of non-O157 Shiga toxin- RJ. 2009. Primary and secondary cases in Esch-
producing Escherichia coli in beef abattoirs of erichia coli O157 outbreaks: a statistical analysis.
Argentina. J Food Prot 74:20082017. BMC Infect Dis 9:144.
66. DAstek BA, Del Castillo LL, Miliwebsky E, 78. Miliwebsky E, Deza N, Carbonari C, DAstek B,
Carbonari C, Palladino PM, Deza N, Chinen I, Baschkier A, Zolezzi G, Manfredi E, Chinen I,
Manfredi E, Leotta GA, Masana MO, Rivas M. Rivas M. Outbreaks of Shiga toxin-producing
2012. Subtyping of Escherichia coli O157:H7 strains Escherichia coli in Argentina, 20072011, abstr.
isolated from human infections and healthy cattle P-017, p 107. Abstr. VTEC 2012, Amsterdam, The
in Argentina. Foodborne Pathog Dis 9:457464. Netherlands, 69 May 2012.

380 RIVAS ET AL.
79. Bell BP, Goldoft M, Griffin PM, Davis MA, causing haemolytic syndrome (HUS) in Germany
Gordon DC, Tarr PI, Bartleson CA, Lewis JH, and France. J Infect Dev Ctries 5:437440.
Barrett TJ, Wells JG, Baron R, Kobayashi JA. 84. Hauswaldt S, Nitschke M, Sayk F, Solbach W,
1994. A multistate outbreak of Escherichia coli Knobloch JKM. 2013. Lessons learned from
O157:H7-associated bloody diarrhea and hemo- outbreaks of Shiga toxin-producing Escherichia
lytic uremic syndrome from hamburgers. The coli. Curr Infect Dis Rep 15:49.
Washington experience. JAMA 272:13491353. 85. Bielaszewska M, Mellmann A, Zhang W, Kck
80. Tuttle J, Gomez T, Doyle MP, Wells JG, Zhao R, Fruth A, Bauwens A, Peters G, Karch H. 2011.
T, Tauxe RV, Griffin PM. 1999. Lessons from a Characterisation of the Escherichia coli strain as-
large outbreak of Escherichia coli O157:H7 infec- sociated with an outbreak of haemolytic uraemic
tions: insights into the infectious dose and method syndrome in Germany, 2011: a microbiological
of widespread contamination of hamburger pat- study. Lancet Infect Dis 11:671676.
ties. Epidemiol Infect 122:185192. 86. Scheutz F, Mller Nielsen E, Frimodt-Mller
81. Fukushima I, Hashizume T, Morita Y, Tanaka J, Boisen N, Morabito S, Tozzoli R, Nataro JP,
J, Azuma K, Mizumoto Y, Kaneno M, Matsuura Caprioli A. 2011. Characteristics of the entero-
M, Konma K, Kitani T. 1999. Clinical experiences aggregative Shiga toxin/verotoxin-producing Es-
in Sakai City Hospital during the massive out- cherichia coli O104:H4 strain causing the outbreak
break of enterohemorrhagic Escherichia coli O157 of haemolytic uraemic syndrome in Germany,
infections in Sakai City, 1996. Pediatr Int 41:213 May to June 2011. Euro Surveill 16(24):pii=19889.
217. http://www.eurosurveillance.org/ViewArticle.
82. Kulasekara BR, Jacobs M, Zhou Y, Wu Z, aspx?ArticleId=19889.
Sims E, Saenphimmachak C, Rohmer L, Ritchie 87. European Food Safety Authority (EFSA). 2011.
JM, Radey M, McKevitt M, Freeman TL, Scientific Report of EFSA: Urgent advice on the
Hayden H, Haugen E, Gillett W, Fong C, Chang public health risk of Shiga toxin-producing Esche-
J, Beskhlebnaya V, Waldor MK, Samadpour M, richia coli in fresh vegetables. EFSA J 9:22742324.
Whittam TS, Kaul R, Brittnacher M, Miller 88. Weiser AA, Gross S, Schielke A, Wigger JF,
SI. 2009. Analysis of the genome of the Esche- Ernert A, Adolphs J, Fetsch A, Mller-Graf C,
richia coli O157:H7 2006 spinach-associated out- Ksbohrer A, Mosbach-Schulz O, Appel B,
break isolate indicates candidate genes that Greiner M. 2013. Trace-back and trace-forward
may enhance virulence. Infect Immun 77:3713 tools developed ad hoc and used during the STEC
3721. O104:H4 outbreak 2011 in Germany and generic
83. Rubino S, Cappuccinelli P, Kelvin DJ. 2011. concepts for future outbreak situations. Foodborne
Escherichia coli (STEC) serotype O104 outbreak Pathog Dis 10:263269.

Pathogenesis and the Host Response
DIANA KARPMAN1 and ANNE-LIE STHL1

19
INTRODUCTION
Bacterial exotoxins may cause damage to host cells by defined mechanisms.

Depending on the presence of the globotriaosylceramide (Gb3) receptor, Shiga
toxin may bind to cells and induce the ribotoxic stress response and apoptosis
(1, 2). The toxin can also induce a proinflammatory response in cells, an effect
that may be dissociated from ribosome inactivation and can even occur in cells
lacking protein synthesis machinery. Bacterial lipopolysaccharide (LPS) induces
a host response by binding to Toll-like receptor 4 (TLR4) and activating specific
intracellular pathways. The activation of proinflammatory pathways, if exces-
sive, promotes damage to the host. This article addresses enterohemorrhagic
Escherichia coli (EHEC) pathogenesis and the host response, examining the
innate and adaptive immune responses to the bacteria and virulence factors and
how they affect the process of colonization, transfer and transport of virulence
factors in the circulation, activation of thrombosis and inflammation, and spe-
cific end-organ damage to the kidney and the brain.
1
Department of Pediatrics, Clinical Sciences, Lund University, Lund, Sweden.
381

382 KARPMAN AND STHL
EHEC COLONIZATION AND THE HOST Stx receptor and thus increased susceptibility
RESPONSE IN THE INTESTINE to the toxin (13).
Short-chain fatty acids also affect the ex-
During intestinal colonization EHEC strains pression of antimicrobial peptides, cathelicidin
encounter chemical, mechanical, and biolo- and defensins (14), thus creating a hostile en-
gical barriers (3). Chemical encounters include vironment for bacterial colonization, as shown
saliva containing mucins and enzymes, acid by using the A/E-forming pathogen Citrobacter
stress in the stomach, bile secretion in the rodentium (15). In a mouse model inoculated
small intestine, and antimicrobial peptides with EHEC, cathelicidin protected wild-type
throughout the intestine. The mechanical mice from EHEC colonization, infection, and
barrier consists of the mucus layer, and the subsequent renal injury in comparison to
biological barrier includes the intestinal mi- knockout mice (3). The effect occurred at the
croflora. These encounters, and the innate and intestinal level, as cathelicidin-sufficient mice
acquired immune response to the pathogen, were not protected from the effects of injected
attempt to eliminate the bacteria. EHEC Stx. In addition to enhanced bacterial survival,
strains must overcome these barriers to colo- cathelicidin-deficient mice had a thinner co-
nize. After ingestion, strains survive stomach lonic mucus layer and thus a defective intesti-
acidity by expressing acid-resistant systems nal mechanical barrier. Although antimicrobial
(46). Interestingly, the response to acid stress peptides protect the host from bacterial colo-
is not only a survival strategy but was also nization, colonic pathogens may actually
found to activate certain EHEC properties as- downregulate antimicrobial peptides, as dem-
sociated with enhanced motility and cell ad- onstrated in shigellosis (16), to the advantage of
hesion, but not to affect the expression of Shiga the bacteria.
toxin (Stx) (4, 6). From the stomach the bac- The interaction with the intestinal com-
teria pass to the small intestine where contact mensal microflora also activates commu-
with bile may further promote migration (6). nication between bacteria known as quorum
Initial binding is assumed to occur at sensing and between bacteria and host
the follicle-associated epithelium of Peyers hormones. The interaction of bacteria with
patches and villi of the terminal ileum (7, 8). host catecholamines promotes virulence by
Bacteria may be taken up by intestinal M cells activating bacterial motility (flagellar synthe-
and transferred to underlying macrophages sis), formation of A/E lesions, and increased
where they can survive and produce Stx (9). expression of Stx (17, 18). Presumably an in-
In the small and large intestine the bacteria crease in catecholamines in the circulation
come in contact with short-chain fatty acids, and the local intestinal environment could
acetate, propionate, and butyrate, secreted by occur during hemorrhagic colitis.
the intestinal flora as fermentation products of The bacteria release Stx into the intestinal
dietary carbohydrates (10). Low concentra- lumen and onto enterocytes. Toxin was dem-
tions of butyrate were shown to upregulate onstrated inside intestinal cells from a patient
EHEC virulence genes involved in motility with EHEC infection, indicating that it can be
and formation of attaching and effacing (A/E) taken up by these cells (19). These cells may
lesions (11). This effect of butyrate was abro- express the Gb3 receptor to which the toxin
gated by deletion of the bacterial lrp gene binds (13, 20). The toxin may, however, un-
encoding the leucine-responsive regulatory dergo endocytosis by macropinocytosis (19).
protein. High concentrations of short-chain Furthermore, the toxin binds to intestinal
fatty acids trigger expression of the bacterial Paneth cells (21). Alternatively, excess per-
iha gene conferring colonic adherence (12). meability of the mucosal barrier may allow Stx
Furthermore, butyrate treatment of colon cells transport from the lumen in a paracellular
in vitro resulted in enhanced expression of the fashion (22) although this remains a specula-

CHAPTER 19 EHEC Pathogenesis and the Host Response 383
tion. Thus the precise manner by which Stx occur simultaneously (36). Stx also induced
binds to intestinal cells and is internalized or the expression of tumor necrosis factor alpha
transported to the endothelium in the in vivo (TNF-a) and IL-6 from murine peritoneal
setting, is as yet, unclear. Stx leads to apoptosis macrophages (37). In addition to Stx, long-
of epithelial cells in vitro in human (2325) and polar fimbriae expressed by EHEC were re-
mouse intestinal cells (26). Apoptosis was cently shown to induce a proinflammatory
demonstrated in the human intestine affected response in T84 cells in a study showing that
by EHEC infection, in a rabbit in vivo model the NF-kB pathway was activated (38).
(27), and in the murine EHEC model. The EHEC could, however, suppress the intes-
major virulence factor associated with this tinal epithelial cytokine response to Stx (39).
effect was determined to be Stx (28). The ap- Likewise, EHEC, and Stx in particular, were
optosis-inducing effect initiates an inflamma- shown to inhibit gamma interferon-mediated
tory response and leukocyte influx of primarily epithelial cell activation (40); both these effects
phagocytes. Neutrophil migration toward the could mitigate the host response and thus po-
intestinal lumen occurs simultaneously and tentially promote bacterial colonization.
enhances Stx translocation from the lumen via The importance of LPS in EHEC infection
enterocytes in vitro (29, 30). High leukocyte was demonstrated in a murine model using
counts were demonstrated in feces of patients C3H/HeJ LPS-hyporesponsive mice in com-
with E. coli O157:H7 infection (31), suggesting parison to wild-type C3H/HeN mice. C3H/
that a similar process may occur in vivo. HeN mice developed earlier and simultaneous
The intestinal inflammatory response is a systemic and neurologic symptoms whereas
paramount feature of host resistance to in- C3H/HeJ mice exhibited a biphasic course of
fection. The initial interaction between the disease, first developing systemic symptoms
intestinal mucosa and bacterial virulence and later severe neurologic symptoms (41).
factors may promote bacterial clearance by The discrepancy between LPS-responders and
inducing an appropriate degree of inflamma- nonresponders was assumed to be related to
tion. Invasive enteropathogenic strains induce the initial intestinal inflammatory response.
an excessive release of chemokines from in- LPS may induce a mucosal immune response
testinal cell lines in comparison to noninvasive during the initial phase of disease, which could
strains, including E. coli O157:H7 (32). All the facilitate bacterial clearance. A reduced initial
same, in vitro studies have demonstrated re- response, due to lack of response to LPS,
lease of interleukin-8 (IL-8) by T84 intestinal would promote more severe disease as clear-
cells stimulated with EHEC, thus promoting ance of bacteria from the intestine would be
inflammatory influx. Stx alone can also induce delayed, enabling bacterial proliferation and
secretion of IL-8 and other C-X-C chemokines extended toxin release intestinally and sys-
in the gut (3335). Stx-induced IL-8 expres- temically, thus explaining the biphasic pro-
sion was associated with induction of mito- longed course of disease in C3H/HeJ mice.
gen-activated protein (MAP) kinase pathways Pathogen-associated molecular patterns, or
Jun N-terminal protein kinase and stress- PAMPs, are specific pathogen-associated mol-
activated protein kinase and p38 in intestinal ecules recognized by cell receptors, such as
epithelial cells (23). Thus the apparently dis- TLRs, transmembrane receptors within the
parate proinflammatory and apoptotic path- innate immune system. Stimulation of TLRs
ways may converge via induction of host triggers a downstream cellular signal that
stress-activated MAP kinases. Stx induces the results in the production and release of cyto-
ribotoxic stress response, thereby inhibiting kines and chemokines. TLRs recruit intracel-
protein synthesis, but may, via eIF4E phos- lular adaptor proteins. MyD88 is an adaptor
phorylation, promote translation of inflam- molecule common for all TLRs except for
matory mediators, so that both processes TLR3. LPS binds to the MD2-TLR4 receptor

complex and initiates a signal cascade via cord sera of uninfected women in areas of
MyD88-dependent or MyD88-independent EPEC endemicity (56). Patients may likewise
pathways (42). TLR signaling depends on four develop serum antibodies against Stx2 and
adaptor proteinsMyD88, TIRAP (also known Stx1 (57), and a lesser antibody response to Stx
as MAL), TRIF (also known as TICAM1), and and LPS was even detected in asymptomatic
TRAM (also known as TICAM2)that recruit household contacts (58).
downstream signaling components (43). To what degree anti-Stx and anti-LPS an-
The importance of TLR4, TRIF, and tibodies exert a protective effect against in-
MyD88 for the pathogenesis of EHEC infec- fection is unclear. However, antibodies against
tion was demonstrated in wild-type and common EPEC and EHEC antigens may
knockout mice infected with E. coli O157:H7 have a protective effect and thus explain the
(both Stx2-producing and non-producing) low prevalence of EHEC infections in areas
(44). Only Stx2-producing mice developed where EPEC is endemic (59). Colostrum from
symptoms, and the most severe symptoms Brazilian mothers in areas of EPEC ende-
and pathology were demonstrated in MyD88- micity was shown to inhibit adherence of
deficient mice. These mice also had the EHEC strains to HEp-2 cells (52). The pro-
highest bacterial burden, suggesting that the tective effect of cross-immunity between
immune response at the intestinal mucosa EPEC and EHEC was demonstrated in a
was essential for bacterial elimination. Even mouse model in which mice were first inoc-
TLR4-knockout mice exhibited more severe ulated with EPEC, followed by inoculation
disease than wild-type mice did. In contrast, with EHEC (60). Mice prechallenged with
an in vitro study showed that Stx uses TLR4 EPEC developed antibodies against common
on intestinal cells for cellular uptake and antigens and were protected from symptoms
transport (45); thus lack of TLR4 should and pathology (intestinal and renal) caused by
have been protective in vivo, which was not EHEC infection.
the case.
SHIGA TOXIN AND LPS INTERACTIONS

THE ACQUIRED IMMUNE RESPONSE TO
WITH BLOOD CELLS
EHEC INFECTION
Neutrophils
An acquired immune response develops after
EHEC infection. In developing countries, en- Neutrophils are a prominent component of the
teropathogenic E. coli (EPEC) is a similar acute inflammatory response, and levels of
pathogen capable of causing diarrhea. In sim- circulating neutrophils rise during an infec-
ilarity to EHEC, it has a type-3 secretion sys- tious process. In HUS, neutrophil counts are
tem and can thus form A/E lesions on the usually elevated at presentation (61), and the
intestinal epithelium, but EPEC does not pro- degree of neutrophilia at the onset of HUS is
duce Stx. Antibodies against antigens common predictive of outcome (62). Neutrophils in
to both EPEC and EHEC strains, such as E. HUS are activated and degranulated, thus re-
coli-secreted proteins A and B as well as leasing proteases (63, 64), as demonstrated by
intimin, have been found in human serum, sa- the presence of high levels of neutrophil elas-
liva, colostrum, and breast-milk (4652). tase in patient sera (65, 66) and a higher ca-
EHEC-infected patients were also found to pacity to adhere to cultured human endothelial
have serum and saliva antibodies against the cells and degrade fibronectin (67). Degranu-
infecting strains LPS (53, 54) although the lation and activation of neutrophils correlated
antibody levels decreased over time (55). Anti- with poor prognosis (68, 69). Moreover, pa-
LPS antibodies were also detected in umbilical tient neutrophils were also delayed in their

apoptotic program with an increased life span regulating innate and adaptive immune re-
(63, 70). sponses. In HUS, monocytes are differentiated
Augmented levels of circulating apoptotic toward an inflammatory phenotype with an
and necrotic leukocytes were found in increased population of cells with reduced
patients with E. coli O104:H4-associated CD14 and enhanced CD16 membrane ex-
HUS (71). Plasma levels of leukocyte-derived pression (81). In addition, monocytes from
microparticles were elevated as was binding patients with HUS exhibited reduced circu-
of platelet microparticles to leukocytes. Inhi- latory expression of function-related proteins
bition of complement had only a moderate such as CD11b, CD64, CD62L, and CX3CR,
impact on microparticle levels while it in- possibly due to monocyte infiltration of tissue
creased the amount of apoptotic and necrotic lesions (81, 82).
leukocytes in the circulation (71). Patients with HUS were found to have Stx2
Stx2 may promote neutrophilia in mice by on monocytes and platelet-monocyte aggre-
triggering the release of cells of myeloid line- gates (78). Monocytes express small amounts
age from the bone marrow or by accelerated of the Gb3 receptor that is slightly different
proliferation of polymorphonuclear leukocyte from the Gb3 lipoforms present on endothelial
(PMN) progenitors (72). Whether Stx binds to cells (83). The number of Gb3 receptors on
neutrophils, or not, is debatable because the cells monocytes can be increased by LPS binding,
lack the Gb3 receptor but may bind to TLR4, see thus leading to activation of the cells and en-
note added in proof (7375). Studies on circu- hanced binding of Stx (83). Stx is not cyto-
lating neutrophils taken during the acute phase toxic for monocytes but triggers a variety
of HUS detected Stx bound to the surface of of proinflammatory events, including in vitro
neutrophils (7679) and to platelet-neutrophil synthesis and secretion of proinflammatory
aggregates (78) although the toxin apparently cytokines and chemokines such as IL-6, IL-8,
does not exert a cytotoxic effect on these cells. TNF-a, and IL-1b (83).
In vitro studies have shown that Stx2 ac- Monocytes may indirectly contribute to the
tivates neutrophils, predominantly those in thrombotic process occurring during HUS.
complex with platelets (78). Moreover, Stx Stimulation of macrophage-like cells of the
may transfer from the neutrophil surface to monocytic cell line THP-1 with Stx2 induced
endothelial cells, implying that neutrophils the release of macrophage-derived chemo-
could serve as a carrier for the toxin in the kine, RANTES, and IL-8, an effect that was
circulation. In mice, Stx2 was shown to impair further enhanced in the presence of LPS,
neutrophil migration (72). In addition, StcE, a leading to platelet activation and aggregation
metalloprotease released from E. coli O157:H7, (84). Incubation of monocytes with Stx1 or
increased the neutrophil oxidative burst and Stx2 induced expression of tissue factor on
cell adhesion leading to impaired migration their surfaces (78, 85), which was further en-
(80). Taken together, activation and degranu- hanced upon coincubation with LPS and when
lation of neutrophils in HUS and impaired monocytes were in complex with platelets
migration leading to increased tissue de- (78), and thus presumably involved in the
struction at sites of neutrophil infiltration can prothrombotic process occurring during HUS.
be explained by the interaction between Monocyte-derived microparticles express-
neutrophils and Stx as well as StcE. ing tissue factor were detected in patients
with HUS (78). Stx2 induced the release of
tissue factor-expressing microparticles from
Monocytes
monocytes in vitro. Monocyte-derived micro-
Monocytes are a critical effector component particles can deliver and transfer tissue factor
in the innate immune response by present- to platelets (86) and neutrophils (87) and
ing antigens and producing cytokines, thus thus induce a prothrombotic surface. Taken

together, Stx can bind to and activate mono- tion and can influence both innate and adaptive
cytes and induce the formation of platelet- immunity (100). Activated platelets interact
monocyte aggregates and the release of with both neutrophils (101) and monocytes
tissue factor-expressing microparticles with (102) and release chemokines such as platelet
prothrombotic properties. However, the trans- factor-4, macrophage inflammatory protein
fer of Stx from monocytes to endothelial cells (MIP), RANTES, IL-8, b-thromboglobulin, and
probably does not occur in vivo (88), as both monocyte chemoattractant protein (MCP) from
cells express receptors with similar affinities their a-granules, which potentiate the inflam-
(83). matory process (103).
Both human and murine platelets express
TLR4 and other TLR receptors (98, 104, 105).
Platelets
Platelet TLRs mediate LPS-induced throm-
Low platelet counts and formation of renal bocytopenia and TNF-a production by
thrombi are characteristic of HUS, suggesting leukocytes (105109). Recent evidence sug-
that platelet activation is involved in the path- gests that platelets may bridge the innate and
ogenesis. Thrombocytopenia is assumed to adaptive immune system by expressing
be related to consumption of platelets in immunostimulatory proteins such as CD40L
microthrombi and may be caused by the direct and thereby stimulate CD8+ T-cell induction
effects of Stx on platelets or by Stx-induced and adaptive immunity (110). Resting platelets
endothelial cell damage leading to second- must be primed (with LPS or other activators)
ary platelet activation and the formation of before interaction with Stx (78, 111). LPS binds
microthrombi (89, 90). During the acute to platelets via a receptor complex of TLR4
phase of HUS platelets are degranulated and CD62 and activates them as shown by the
(91), with reduced intracellular levels of b- expression of CD40L, activated GPIIb/IIIa
thromboglobulin and impaired aggregating receptor, and fibrinogen binding (98). GPIIb/
response (92). Platelet-derived microparticles IIIa and CD40L both play a central role in
are increased, indicating platelet activation thrombotic diseases. In addition, CD40L can
(78, 93), and plasma from patients with HUS interact with CD40 on endothelial cells, trig-
induced aggregation of normal platelets (94). gering an inflammatory response leading to
Mice inoculated with E. coli O157:H7 mimicked local or systemic release of MCP-1, VCAM,
the human disease and developed thrombocy- and ICAM (112114).
topenia, which was also demonstrated in mice
injected with Stx2 and LPS (44, 95).
Platelets bind Stx through Gb3 and an al- LPS INTERACTION WITH BLOOD CELLS
ternative glycosphingolipid receptor, termed
band 0.03 (96, 97). Gb3 expression on resting Recognition of LPS by innate immune cells is
platelets is very low (97), and the distribution vital for host defense against gram-negative
in humans is quite heterogeneous (96). Stx bacteria. LPS, the major component of the
is assumed to bind primarily to activated outer membrane of gram-negative bacteria,
platelets (78, 97), and Stx circulates bound to circulates bound to platelets during acute
platelets, in addition to leukocytes, during HUS (98). Platelets from mice inoculated
HUS (98). Upon binding to platelets, Stx is intraperitoneally with LPS showed surface-
rapidly internalized, leading to further acti- bound LPS and exhibited increased CD40L
vation, aggregation, structural changes en- expression, suggesting that LPS activates
hancing the surface area, and increased platelets in the circulation in the murine
fibrinogen-binding capacity (99). model (98). Patients develop an antibody re-
In addition to their role in thrombosis and sponse to strain-specific LPS (115). The con-
hemostasis, platelets are involved in inflamma- centration of LPS-binding protein, which

binds LPS in plasma and transfers it to cell most of their tissue factor through interaction
surfaces, is increased in the plasma of patients with tissue factor-bearing microparticles from
with HUS compared to those with uncom- monocytes (126, 127).
plicated EHEC diarrhea, suggesting that the
acute-phase response to LPS is associated
with disease severity (116). THE THROMBOTIC PROCESS IN HUS
Neutrophils and monocytes express TLR4
and respond to LPS stimulation by releasing The prothrombotic condition in HUS has been
proinflammatory cytokines (117). At the onset primarily ascribed to damage of the microvas-
of HUS, neutrophils exhibited higher levels of cular endothelium. When vascular injury
TLR4 mRNA and TLR4 protein expression occurs, platelets are recruited to the damaged
(118). TLR4 expression was correlated to site in a multistep process that involves the
increased circulating TNF-a levels. No dif- interaction of specific platelet cell-surface re-
ferences were noted in TLR4 receptor ex- ceptors with subendothelial matrix proteins
pression on monocytes at the onset of HUS such as von Willebrand factor (VWF), collagen,
(118), indicating different regulation of TLR4 and fibronectin. The first event in this process
expression on neutrophils and monocytes. is binding of platelet glycoprotein 1ba receptor
(GP1ba) to the A1 domain of VWF after which a
conformational change in the platelet integrin
TISSUE FACTOR receptor GPIIb/IIIa occurs, allowing binding
to both VWF and fibrinogen, inducing aggre-
During the acute phase of HUS, patients have gation. After initial tethering steps, platelets
high plasma levels of tissue factor and tissue become activated and firmly adhere to the
factor pathway inhibitor (119) and circulating vessel wall and form a clot. Adherent and ac-
platelet-leukocyte aggregates expressing tis- tivated platelets release potent platelet agonists
sue factor and tissue factor-bearing platelet such as thrombin, ADP, and thromboxane A2
microparticles (78). In vitro studies showed from their intracellular granules, promoting
that Stx, in cooperation with LPS, induced further platelet activation and aggregation and
aggregate formation between platelets and resulting in rapid growth of the thrombus.
leukocytes, leading to release of platelet-de- In HUS, platelets are activated (128) and
rived microparticles with surface-bound tis- degranulated (91, 92), and platelet-derived
sue factor (78). factors such as b-thromboglobulin and platelet
Tissue factor-positive microparticles may factor-4 are elevated (129) as is VWF, which
contribute to thrombosis. Tissue factor is a may be secreted from both platelets and en-
transmembrane glycoprotein receptor (120) dothelial cells (130, 131). VWF mediates plate-
for coagulation factor VII (121). By acting as a let adhesion to activated endothelial cells in
cofactor for factor VII, tissue factor promotes response to Stx (132). In addition, Stx delayed
proteolysis and activation of factor VIIa, the cleavage of VWF-platelet strings by the
followed by formation of the prothrombinase metalloprotease ADAMTS13 on activated en-
complex (122) and conversion of prothrombin dothelial cells (133), thus potentiating throm-
into thrombin, resulting in thrombus forma- bus growth in the presence of larger VWF
tion and further platelet activation. Expres- multimers. Functional blockade of receptors or
sion of tissue factor on platelets is debated. adhesive proteins, such GPIIb/IIIa, P-selectin,
Platelets contain tissue factor mRNA (123), or VWF, was associated with a marked re-
which can be spliced into mature mRNA upon duction of thrombi on endothelial cells (132).
platelet activation, leading to minimal pro- Both Stx and LPS activate platelets, especially
tein expression and procoagulant activity under high shear stress, and costimulation
(124, 125). However, platelets seem to acquire with both factors simultaneously induced an

additive effect on the formation of platelet- of neutrophils correlated to patient renal

leukocyte aggregates expressing tissue factor dysfunction (68). Mice that were coinjected
(78, 9799). In addition, platelets may be acti- with Stx2 and LPS developed neutrophilia and
vated indirectly by additional factors such as monocytosis with markers of leukocyte acti-
chemokines and cytokines released by Stx- vation (95). Neutrophil and macrophage ac-
stimulated monocytes (84) or endothelial cells cumulation was demonstrated in the murine
(134). Thus, platelet activation and thrombus kidneys (147, 148). Likewise, kidneys from
formation may be caused directly by Stx and/or rabbits injected with Stx2 showed PMN
LPS or by the release of cytokines, indicating infiltrates correlating with levels of IL-8 (149).
that platelet activation and inflammation are In vitro studies showed that stimulation of
correlated events. glomerular endothelial cells with Stx2 en-
In parallel with the recruitment of platelets, hanced leukocyte adherence and migration
the blood coagulation cascade is activated at under perfusion, effects mediated by MCP-1,
the site of vessel injury (135). In HUS, there is IL-8, and fractalkine (CX3CL1) secreted from
no consumption of coagulation factors, but el- the cells (150, 151) and enhanced by TNF-a
evated levels of prothrombin fragment 1 + 2 (152). Furthermore, Stx borne on leukocytes
(136138), tissue plasminogen activator, tissue migrated through endothelial cells and in-
plasminogen activator inhibitor-1 (139), and D- duced their release of IL-8 and MCP-1 (153).
dimers have been found, even before HUS In addition to leukocytes, platelets have
develops, indicating enhanced thrombotic ca- also been demonstrated in the glomerular
pacity and impaired fibrinolysis. microthrombi characteristic of human throm-
botic microangiopathy (reviewed in reference
154), in the glomeruli of primates treated with
THE SYSTEMIC AND RENAL Stx (155) and of mice injected with Stx2 and
HOST RESPONSE LPS (95). Low platelet counts during HUS are
correlated to poor recovery of renal function
The precise mechanism by which Stx and (62, 156). Stx and LPS induce human micro-
other EHEC virulence factors, such as LPS, vascular endothelial cells to secrete factors that
reach the kidney is, as yet, unknown. Although activate platelets (134). Although platelets are
Stx and O157LPS have been shown to circu- activated and secrete microparticles during
late bound to blood cells, as described above, HUS, and the role of platelets in the renal
the manner by which they transfer to resident thrombotic events is evident, their contribu-
target organ cells in the kidney, brain, or other tion to the proinflammatory events occurring
organs has not been elucidated in the in vivo in the kidney has not been thoroughly inves-
setting. However, the toxin was detected in tigated. For example, degranulation of platelet
glomeruli and tubuli of pediatric and geriatric alpha granules upon activation will release
patients with HUS (140, 141), indicating that it microbicidal proteins, CC-chemokines, and
reaches the kidney. The extensive renal injury CXC-chemokines (103, 154). The binding of
associated with renal cell apoptosis occurring specific ligands and release of potent platelet
during HUS (142) activates a variety of host components could theoretically promote the
responses, including the influx of leukocytes inflammatory state in the kidney.
and the release of cytokines, both of which Proinflammatory and prothrombotic re-
could enhance the tissue damage. sponses have been documented in HUS. In-
Leukocytosis was associated with the de- flammatory and prothrombotic mediators,
velopment of HUS (143, 144) and with a poor including cytokines, chemokines, soluble ad-
outcome (145), as described above. Renal in- hesion molecules, growth factors, cytokine
flux of neutrophils was associated with in- receptors, tissue factor, and acute-phase re-
creased mortality (61, 146). The functionality sponse proteins, are elevated in patients with

EHEC-associated infection and HUS (116, 119, during HUS (164) may contribute to inflam-
138, 157174) and could be correlated, in cer- mation, thrombosis, and end-organ damage.
tain studies, to the progression of renal dam- A recent study using baboons treated with
age (reviewed in reference 89). Increased Stx exhibited high renal mRNA levels of IL-8,
levels of chemoattractants, such as IL-8 (69), MCP-1, and MIP-1a in contrast to a modest
granulocyte colony-stimulating factor (159), effect on TNF-a. These baboons had elevated
and MCP-1 (157), could explain leukocyte in- urinary IL-6, IL-8, MCP-1, and VEGF (155).
flux. Stx and TNF-a act in synergy to cause Similarly, a previous study in baboons showed
cytotoxic effects on human endothelial cells that Stx infusion led to urinary secretion of
(175177). Elevated TNF-a may enhance the TNF-a and IL-6 (187). Stimulation of proximal
expression of the Stx receptor Gb3 on endo- tubular cells with Stx increased expression of
thelial cells (178). Enhanced Gb3 expression IL-6 mRNA and protein (185). A similar effect
was also demonstrated on cultured endothe- was noted in stimulated glomerular endothe-
lial cells stimulated with LPS and IL-1 (175, lial cells, particularly when cells were
177, 179, 180), thus sensitizing the cells to Stx. costimulated with LPS (188). Patients with
Animal models have further addressed the HUS exhibit very high IL-6 levels in serum
importance of inflammatory and chemotactic and urine (164). Elevated glomerular and tu-
pathways for the pathogenesis of HUS. Gno- bular levels of IL-6 may promote renal injury.
tobiotic mice inoculated with E. coli O157:H7 In addition to its role in immune regulation
exhibited TNF-a, IL-1, and IL-6 in the kidney and inflammation, IL-6 may respond to glo-
(181). Certain mice were also treated with merular injury by promoting mesangial pro-
TNF-a, which led to enhanced kidney dam- liferation (189).
age. TNF inhibition, with the protease inhib- The chemokine receptor CCR1 was shown
itor Nafamostat mesilate, reduced target- to be involved in neutrophil and monocyte
organ damage. Stx induced TNF synthesis in infiltrates in the kidney and in host survival
the mouse kidney while increasing renal sen- after Stx2 treatment of CCR1-deficient mice
sitivity to the toxic effects of TNF (182). In- (190). A slower increase in plasma TNF-a and
terestingly, TNF had a protective effect when IL-6 was also noted in the CCR1/ mice
administered to mice before Stx1 (183) but compared to wild-type mice. Similarly, injec-
exacerbated disease when given after Stx1. tion of Stx2 and LPS in a mouse model
These results are in contrast to a previous markedly increased renal expression of the
study using neutralizing anti-TNF-a antibody chemoattractants CXCL1 and CXCL2, affect-
before administration of Stx1, as well as TNF- ing neutrophil influx (147). In the same mouse
a knockout mice, which could not demon- model, macrophage influx was also demon-
strate a role for TNF-a in Stx-induced toxicity strated and associated with expression of
(184). In vitro results indicated that Stx in- monocyte chemoattractants: MCP-1/CCL2,
duced increased secretion of TNF-a from MIP-1/CCL3, and RANTES (CCL5) (148).
human renal proximal tubule cells (185) and Taken together, the studies described provide
that Stx2-induced TNF-a expression was di- evidence for chemokine- and cytokine-medi-
minished by blocking the p38 pathway (186). ated inflammation that underscores the im-
HUS patients exhibited high levels of TNF- portance of the inflammatory response in the
a in the circulation (163, 166, 168, 170) and in development of HUS.
the urine, which did not correlate to levels in The CXCR4/CXCR7/stromal cell-derived
the blood, indicating local synthesis in the factor 1 (SDF-1) pathway, involved in renal
urinary tract (164). TNF-a is a proinflamma- homeostasis, was activated by Stx2 (174), thus
tory cytokine mediating a cytotoxic effect on enhancing endothelial activation and renal
tumor cells, inflammation, and microvascular damage. The results were confirmed by using
coagulation. The high urinary levels occurring human microvascular endothelial cells in vitro

and showed that low concentrations of Stx, triggered by multiple inflammatory mediators
which minimally affect protein synthesis, ac- (202). In EHEC-infected mice astrocytes were
tivated the CXCR4/CXCR7/SDF-1 pathway activated in the medulla oblongata and spinal
(134, 174). The importance of this finding was cord (203). EHEC-infected mice exhibited
established in patients showing elevated TNF-a in the brain, and intraperitoneal treat-
plasma levels of SDF-1 and thus suggested an ment with TNF-a given before and after
effect on the glomerular vasculature. EHEC inoculation worsened the neurological
Tissue factor is a potent prothrombotic symptoms. The TNF inhibitor, Nafamostat
mediator and was found to be elevated in the mesilate, modified these responses and de-
plasma of HUS patients (119, 173), on circu- creased cytokines in the brain, suggesting a
lating platelet-derived microparticles (78), role for TNF-a in the neurological manifes-
and demonstrated in the kidney (191). Stx in- tations (181). Furthermore, intravenous injec-
duced tissue factor activity in glomerular en- tion of Stx2 in rabbits induced brain edema
dothelial cells and proximal tubular cells, an that could be reversed by anti-inflammatory
effect enhanced in the presence of TNF-a steroid treatment (204).
(192, 193). In vitro experiments using human brain
endothelial cells have shown that TNF-a and
IL-1b enhanced Stx toxicity (205). TNF in-
THE HOST RESPONSE IN THE BRAIN creases the Gb3 receptor on human brain en-
dothelium (206). The TNF effect was mitigated
The central nervous system (CNS) is also a by inhibition of p38 MAP kinase (207). Simi-
target organ during HUS, and as many as 48 larly, TNF-a amplified the inflammatory re-
to- 66% of patients developed severe neuro- sponse to Stx1 and LPS-treated rat astrocytes,
logical manifestations during the E. coli O104: enhancing PMN chemotaxis and cytotoxicity
H4 epidemic in Germany is the spring of 2011 (208). Inflammatory mediators released from
(194, 195). Shiga toxin may bind to the Gb3 the stimulated astrocytes affected endothelial
receptor on neurons and endothelial cells permeability and increased binding of PMNs
in the human CNS (196). The induction of and platelets to the endothelial cells (209).
apoptosis (197, 198) and the inflammatory re- Thus in vivo and in vitro evidence points to a
sponse may promote CNS injury. In a study considerable inflammatory response in the
conducted on pediatric patients with HUS CNS affecting the endothelium and astrocytes
and encephalopathy, versus patients with HUS as well as the blood-brain barrier.
without CNS symptoms or with EHEC-asso-
ciated colitis without HUS, serum inflamma-
tory mediators IL-6, soluble TNF receptor 1, THE EFFECTS OF
and tissue inhibitor of metalloproteinase-1 COMPLEMENT ACTIVATION
were correlated to the presence of en-
cephalopathy (199). Increased protein in the Complement activation via the alternative
cerebrospinal fluid (200) also indicates an pathway occurs during EHEC-associated
enhanced inflammatory state. HUS. This has been documented by hypo-
EHEC strains were used in animal models complementemia (low C3) and elevated plas-
to reproduce neurological affection (41, 201) ma levels of degradation products such as
showing endothelial and neuronal damage, complement factors Bb, C3a, and soluble C5b-9
and Stx was detected in the brain cortex and during the acute phase of disease; this did not
spinal cord (201). The blood-brain barrier is correlate with the severity of renal injury or
generally affected by damage to cerebral cap- occurrence of later complications and de-
illary endothelial cells, perivascular pericytes, creased upon recovery (210213). However,
and possibly even astrocytes. Astrogliosis is circulating C3a and soluble C5b-9 may activate

platelets (214, 215). One patient was also shown clearance. Thus a certain degree of comple-
to have C3 deposits on circulating platelet- ment activation is protective at the mucosal
monocyte aggregates (213). Patient platelet- surface. However, prolonged and amplified
and monocyte-derived microparticles were activation will promote platelet activation and
shown to be coated with C3 and C9 deposits endothelial damage in HUS (219, 221).
during the acute phase of HUS but not after
recovery (213). These findings suggest that the
extensive endothelial and blood cell activation SUMMARY
occurring during HUS will lead to secondary
complement activation. A direct effect of Stx The host responds to EHEC infections and,
and/or O157LPS was demonstrated when more specifically, to Shiga toxin and EHEC-
these were incubated in vitro with whole derived LPS by activation of a large variety of
blood, leading to the formation of leukocyte- cells, including blood cells, intestinal and renal
platelet aggregates and the release of platelet- epithelial cells, renal and cerebral endothe-
and monocyte-derived microparticles, both lial cells, and astrocytes. Multiple chemokines
with C3 and C9 deposits (213). and cytokines, as well as stress hormones
Complement could also be activated by and antimicrobial peptides, are secreted. They
dysfunctional inhibition. Stx was shown to contribute to the inflammation and interact
activate the alternative pathway in vitro and to with bacteria or virulence factors in an at-
bind to cell-binding domains of the major tempt to clear the infection. The inflammatory
soluble complement inhibitor, factor H, thus responses may, however, have adverse effects
compromising the inhibitory effect and pro- on the host, thus worsening the outcome.
moting complement activation (216). These In certain instances the bacterial virulence
studies were performed with very high con- factors may hijack host responses to their own
centrations of Stx2 and may thus not reflect advantage. The release of tissue factor and
the in vivo situation. Also, as the toxin does thrombin promotes thrombosis, and the
not circulate in free form, but is rather cell- complement system is activated with second-
bound in the circulation, the interaction with ary effects on the endothelium, tubular cells,
factor H may not occur in vivo (90). Factor H and platelets. The inflammation and throm-
prevents complement activation via the alter- bosis arising explain some of the features of
native pathway on cell surfaces. Factor H HUS. It is thus essential to understand not
binding to proximal tubular cells was im- only the cellular signaling pathways used by
paired when cells were exposed to protein EHEC virulence factors but also the pattern of
overload, resembling the situation during inflammatory and thrombotic responses aris-
renal injury (217). Thus complement activa- ing from the infection.
tion may also occur as a nonspecific conse- Treatments aimed at reducing injury to
quence of tubular protein overload. host organs should primarily attempt to neu-
One study addressed activation of the tralize Shiga toxin. However, as patients may
mannan-binding lectin (MBL) pathway in manifest profound organ damage, treatments
EHEC-induced infection and HUS. MBL de- aimed at reducing the bacterial load, the in-
ficiency may predispose to infection, but no flammatory and thrombotic response, and
correlation was found between EHEC infec- the cell injury may be beneficial if instituted
tion and MBL deficiency (218). The main early in the course of infection. The specific
function of the complement system is to dis- pathways leading to host cell damage should
pose of foreign cells, such as bacterial patho- therefore be determined, so that future treat-
gens, by opsonization and cytolysis (219). ments can effectively abrogate the inflam-
Complement is active in the colon (220), and matory and thrombotic complications of this
its primary effect there is to promote bacterial infection.

ACKNOWLEDGMENTS 7. Phillips AD, Navabpour S, Hicks S, Dougan G,

Wallis T, Frankel G. 2000. Enterohaemorrhagic
Diana Karpmans research is supported by Escherichia coli O157:H7 target Peyers patches in
grants from The Swedish Research Council humans and cause attaching/effacing lesions in
both human and bovine intestine. Gut 47:377381.
(K2013-64X-14008), The Torsten Sderberg
8. Chong Y, Fitzhenry R, Heuschkel R, Torrente
Foundation, Crown Princess Lovisas Society F, Frankel G, Phillips AD. 2007. Human intesti-
for Child Care, and The Konung Gustaf V:s nal tissue tropism in Escherichia coli O157:H7
80-rsfond. initial colonization of terminal ileum and Peyers
We declare no conflicts of interest with patches and minimal colonic adhesion ex vivo.
Microbiology 153:794802.
regard to the manuscript.
9. Etienne-Mesmin L, Chassaing B, Sauvanet P,
Denizot J, Blanquet-Diot S, Darfeuille-Michaud
CITATION A, Pradel N, Livrelli V. 2011. Interactions with M
cells and macrophages as key steps in the patho-
Karpman D, Sthl A. 2014. Enterohemor- genesis of enterohemorrhagic Escherichia coli in-
rhagic Escherichia coli pathogenesis and the fections. PLoS One 6:e23594.
10. Miller TL, Wolin MJ. 1996. Pathways of acetate,
host response. Microbiol Spectrum 2(5): propionate, and butyrate formation by the human
EHEC-0009-2013. fecal microbial flora. Appl Environ Microbiol 62:
15891592.
11. Nakanishi N, Tashiro K, Kuhara S, Hayashi T,
NOTE ADDED IN PROOF Sugimoto N, Tobe T. 2009. Regulation of viru-
lence by butyrate sensing in enterohaemorrhagic
Shiga toxin may bind to TLR4 on neutrophils. Escherichia coli. Microbiology 155:521530.
(Brigotti M, Carnicelli D, Arfilli V, Tamassia 12. Herold S, Paton JC, Srimanote P, Paton AW.
N, Borsetti F, Fabbri E, Tazzari PL, Ricci F, 2009. Differential effects of short-chain fatty acids
Pagliaro P, Spisni E, Cassatella MA. 2013. and iron on expression of iha in Shiga-toxigenic
Identification of TLR4 as the receptor that Escherichia coli. Microbiology 155:35543563.
13. Jacewicz MS, Acheson DW, Mobassaleh M,
recognizes Shiga toxins in human neutrophils.
Donohue-Rolfe A, Balasubramanian KA, Keusch
J Immunol 191:47484758.) GT. 1995. Maturational regulation of globotriao-
sylceramide, the Shiga-like toxin 1 receptor, in
cultured human gut epithelial cells. J Clin Invest
REFERENCES 96:13281335.
1. Tesh VL. 2012. The induction of apoptosis by 14. Schauber J, Svanholm C, Termen S, Iffland K,
Shiga toxins and ricin. Curr Top Microbiol Im- Menzel T, Scheppach W, Melcher R, Agerberth
munol 357:137178. B, Luhrs H, Gudmundsson GH. 2003. Expres-
2. Jandhyala DM, Thorpe CM, Magun B. 2012. sion of the cathelicidin LL-37 is modulated by
Ricin and Shiga toxins: effects on host cell signal short chain fatty acids in colonocytes: relevance of
transduction. Curr Top Microbiol Immunol 357: signalling pathways. Gut 52:735741.
4165. 15. Iimura M, Gallo RL, Hase K, Miyamoto Y,
3. Chromek M, Arvidsson I, Karpman D. 2012. Eckmann L, Kagnoff MF. 2005. Cathelicidin
The antimicrobial peptide cathelicidin protects mediates innate intestinal defense against colo-
mice from Escherichia coli O157:H7-mediated dis- nization with epithelial adherent bacterial path-
ease. PLoS One 7:e46476. ogens. J Immunol 174:49014907.
4. House B, Kus JV, Prayitno N, Mair R, Que L, 16. Islam D, Bandholtz L, Nilsson J, Wigzell H,
Chingcuanco F, Gannon V, Cvitkovitch DG, Christensson B, Agerberth B, Gudmundsson G.
Barnett Foster D. 2009. Acid-stress-induced 2001. Downregulation of bactericidal peptides in
changes in enterohaemorrhagic Escherichia coli enteric infections: a novel immune escape mech-
O157 : H7 virulence. Microbiology 155:29072918. anism with bacterial DNA as a potential regulator.
5. Foster JW. 2004. Escherichia coli acid resistance: Nat Med 7:180185.
tales of an amateur acidophile. Nat Rev Microbiol 17. Pacheco AR, Sperandio V. 2009. Inter-kingdom
2:898907. signaling: chemical language between bacteria
6. Barnett Foster D. 2013. Modulation of the entero- and host. Curr Opin Microbiol 12:192198.
hemorrhagic E. coli virulence program through the 18. Hughes DT, Clarke MB, Yamamoto K, Rasko
human gastrointestinal tract. Virulence 4:315323. DA, Sperandio V. 2009. The QseC adrenergic

signaling cascade in enterohemorrhagic E. coli 29. Hurley BP, Jacewicz M, Thorpe CM, Lincicome
(EHEC). PLoS Pathog 5:e1000553. LL, King AJ, Keusch GT, Acheson DW. 1999.
19. Malyukova I, Murray KF, Zhu C, Boedeker E, Shiga toxins 1 and 2 translocate differently across
Kane A, Patterson K, Peterson JR, Donowitz polarized intestinal epithelial cells. Infect Immun
M, Kovbasnjuk O. 2009. Macropinocytosis in 67:66706677.
Shiga toxin 1 uptake by human intestinal epi- 30. Hurley BP, Thorpe CM, Acheson DW. 2001.
thelial cells and transcellular transcytosis. Am Shiga toxin translocation across intestinal epi-
J Physiol Gastrointest Liver Physiol 296:G78 thelial cells is enhanced by neutrophil transmi-
G92. gration. Infect Immun 69:61486155.
20. Zumbrun SD, Hanson L, Sinclair JF, Freedy J, 31. Slutsker L, Ries AA, Greene KD, Wells JG,
Melton-Celsa AR, Rodriguez-Canales J, Hanson Hutwagner L, Griffin PM. 1997. Escherichia coli
JC, OBrien AD. 2010. Human intestinal tissue O157:H7 diarrhea in the United States: clinical
and cultured colonic cells contain globotriaosyl- and epidemiologic features. Ann Intern Med 126:
ceramide synthase mRNA and the alternate 505513.
Shiga toxin receptor globotetraosylceramide. In- 32. Jung HC, Eckmann L, Yang SK, Panja A, Fierer
fect Immun 78:44884499. J, Morzycka-Wroblewska E, Kagnoff MF. 1995.
21. Schller S, Heuschkel R, Torrente F, Kaper JB, A distinct array of proinflammatory cytokines is
Phillips AD. 2007. Shiga toxin binding in normal expressed in human colon epithelial cells in re-
and inflamed human intestinal mucosa. Microbes sponse to bacterial invasion. J Clin Invest 95:55
Infect 9:3539. 65.
22. Bell CJ, Elliott EJ, Wallace JL, Redmond 33. Thorpe CM, Hurley BP, Lincicome LL,
DM, Payne J, Li Z, OLoughlin EV. 2000. Do Jacewicz MS, Keusch GT, Acheson DW. 1999.
eicosanoids cause colonic dysfunction in experi- Shiga toxins stimulate secretion of interleukin-
mental E. coli O157:H7 (EHEC) infection? Gut 8 from intestinal epithelial cells. Infect Immun
46:806812. 67:59855993.
23. Smith WE, Kane AV, Campbell ST, Acheson 34. Thorpe CM, Smith WE, Hurley BP, Acheson
DW, Cochran BH, Thorpe CM. 2003. Shiga DW. 2001. Shiga toxins induce, superinduce, and
toxin 1 triggers a ribotoxic stress response leading stabilize a variety of C-X-C chemokine mRNAs in
to p38 and JNK activation and induction of apo- intestinal epithelial cells, resulting in increased
ptosis in intestinal epithelial cells. Infect Immun chemokine expression. Infect Immun 69:6140
71:14971504. 6147.
24. Schller S, Frankel G, Phillips AD. 2004. Inter- 35. Yamasaki C, Natori Y, Zeng XT, Ohmura M,
action of Shiga toxin from Escherichia coli with Yamasaki S, Takeda Y, Natori Y. 1999. Induction
human intestinal epithelial cell lines and explants: of cytokines in a human colon epithelial cell line
Stx2 induces epithelial damage in organ culture. by Shiga toxin 1 (Stx1) and Stx2 but not by non-
Cell Microbiol 6:289301. toxic mutant Stx1 which lacks N-glycosidase ac-
25. Barnett Foster D, Abul-Milh M, Huesca M, tivity. FEBS Lett 442:231234.
Lingwood CA. 2000. Enterohemorrhagic Esche- 36. Colpoys WE, Cochran BH, Carducci TM,
richia coli induces apoptosis which augments Thorpe CM. 2005. Shiga toxins activate transla-
bacterial binding and phosphatidylethanolamine tional regulation pathways in intestinal epithelial
exposure on the plasma membrane outer leaflet. cells. Cell Signal 17:891899.
Infect Immun 68:31083115. 37. Tesh VL, Ramegowda B, Samuel JE. 1994.
26. Kashiwamura M, Kurohane K, Tanikawa T, Purified Shiga-like toxins induce expression of
Deguchi A, Miyamoto D, Imai Y. 2009. Shiga proinflammatory cytokines from murine peri-
toxin kills epithelial cells isolated from distal but toneal macrophages. Infect Immun 62:5085
not proximal part of mouse colon. Biol Pharm Bull 5094.
32:16141617. 38. Farfan MJ, Cantero L, Vergara A, Vidal R,
27. Keenan KP, Sharpnack DD, Collins H, Formal Torres AG. 2013. The long polar fimbriae of
SB, OBrien AD. 1986. Morphologic evaluation STEC O157:H7 induce expression of pro-inflam-
of the effects of Shiga toxin and E. coli Shiga-like matory markers by intestinal epithelial cells. Vet
toxin on the rabbit intestine. Am J Pathol 125:69 Immunol Immunopathol 152:126131.
80. 39. Bellmeyer A, Cotton C, Kanteti R, Koutsouris
28. Bkssy ZD, Calderon Toledo C, Leoj G, A, Viswanathan VK, Hecht G. 2009. Entero-
Kristoffersson A, Leopold SR, Perez MT, hemorrhagic Escherichia coli suppresses inflam-
Karpman D. 2011. Intestinal damage in entero- matory response to cytokines and its own toxin.
hemorrhagic Escherichia coli infection. Pediatr Am J Physiol Gastrointest Liver Physiol 297:G576
Nephrol 26:20592071. G581.

40. Ho NK, Ossa JC, Silphaduang U, Johnson Human colostrum contains IgA antibodies reac-
R, Johnson-Henry KC, Sherman PM. 2012. tive to enteropathogenic Escherichia coli viru-
Enterohemorrhagic Escherichia coli O157:H7 lence-associated proteins: intimin, BfpA, EspA,
Shiga toxins inhibit gamma interferon-mediated and EspB. J Pediatr Gastroenterol Nutr 27:166
cellular activation. Infect Immun 80:23072315. 171.
41. Karpman D, Connell H, Svensson M, Scheutz F, 52. Palmeira P, Carbonare SB, Amaral JA, Tino-
Alm P, Svanborg C. 1997. The role of lipopoly- De-Franco M, Carneiro-Sampaio MM. 2005.
saccharide and Shiga-like toxin in a mouse model Colostrum from healthy Brazilian women inhibits
of Escherichia coli O157:H7 infection. J Infect Dis adhesion and contains IgA antibodies reactive
175:611620. with Shiga toxin-producing Escherichia coli. Eur J
42. Takeda K, Akira S. 2004. TLR signaling path- Pediatr 164:3743.
ways. Semin Immunol 16:39. 53. Bitzan M, Moebius E, Ludwig K, Mller-Wiefel
43. Moresco EM, LaVine D, Beutler B. 2011. Toll- DE, Heesemann J, Karch H. 1991. High inci-
like receptors. Curr Biol 21:R488R493. dence of serum antibodies to Escherichia coli O157
44. Calderon Toledo C, Rogers TJ, Svensson M, lipopolysaccharide in children with hemolytic-
Tati R, Fischer H, Svanborg C, Karpman D. uremic syndrome. J Pediatr 119:380385.
2008. Shiga toxin-mediated disease in MyD88- 54. Ludwig K, Grabhorn E, Bitzan M, Bobrowski C,
deficient mice infected with Escherichia coli Kemper MJ, Sobottka I, Laufs R, Karch H,
O157:H7. Am J Pathol 173:14281439. Mller-Wiefel DE. 2002. Saliva IgM and IgA are
45. Torgersen ML, Engedal N, Pedersen AM, a sensitive indicator of the humoral immune re-
Husebye H, Espevik T, Sandvig K. 2011. Toll-like sponse to Escherichia coli O157 lipopolysaccharide
receptor 4 facilitates binding of Shiga toxin to colon in children with enteropathic hemolytic uremic
carcinoma and primary umbilical vein endothelial syndrome. Pediatr Res 52:307313.
cells. FEMS Immunol Med Microbiol 61:6375. 55. Ludwig K, Bitzan M, Bobrowski C, Mller-
46. Carbonare CB, Carbonare SB, Carneiro- Wiefel DE. 2002. Escherichia coli O157 fails to
Sampaio MM. 2003. Early acquisition of serum induce a long-lasting lipopolysaccharide-specific,
and saliva antibodies reactive to enteropathogenic measurable humoral immune response in chil-
Escherichia coli virulence-associated proteins by dren with hemolytic-uremic syndrome. J Infect
infants living in an endemic area. Pediatr Allergy Dis 186:566569.
Immunol 14:222228. 56. Palmeira P, Yu Ito L, Arslanian C, Carneiro-
47. Parissi-Crivelli A, Parissi-Crivelli JM, Giron JA. Sampaio MM. 2007. Passive immunity acquisi-
2000. Recognition of enteropathogenic Esche- tion of maternal anti-enterohemorrhagic Esche-
richia coli virulence determinants by human co- richia coli (EHEC) O157:H7 IgG antibodies by the
lostrum and serum antibodies. J Clin Microbiol newborn. Eur J Pediatr 166:413419.
38:26962700. 57. Ludwig K, Karmali MA, Sarkim V, Bobrowski
48. Karpman D, Bkssy ZD, Sjgren AC, Dubois C, Petric M, Karch H, Muller-Wiefel DE. 2001.
MS, Karmali MA, Mascarenhas M, Jarvis KG, Antibody response to Shiga toxins Stx2 and Stx1
Gansheroff LJ, OBrien AD, Arbus GS, Kaper in children with enteropathic hemolytic-uremic
JB. 2002. Antibodies to intimin and Escherichia syndrome. J Clin Microbiol 39:22722279.
coli secreted proteins A and B in patients with 58. Ludwig K, Sarkim V, Bitzan M, Karmali MA,
enterohemorrhagic Escherichia coli infections. Bobrowski C, Ruder H, Laufs R, Sobottka I,
Pediatr Nephrol 17:201211. Petric M, Karch H, Muller-Wiefel DE. 2002.
49. Sjgren AC, Kaper JB, Caprioli A, Karpman D. Shiga toxin-producing Escherichia coli infection
2004. Enzyme-linked immunosorbent assay for and antibodies against Stx2 and Stx1 in household
detection of Shiga toxin-producing Escherichia contacts of children with enteropathic hemolytic-
coli infection by antibodies to Escherichia coli uremic syndrome. J Clin Microbiol 40:17731782.
secreted protein B in children with hemolytic 59. Martinez MB, Taddei CR, Ruiz-Tagle A,
uremic syndrome. Eur J Clin Microbiol Infect Dis Trabulsi LR, Giron JA. 1999. Antibody response
23:208211. of children with enteropathogenic Escherichia coli
50. Noguera-Obenza M, Ochoa TJ, Gomez HF, infection to the bundle-forming pilus and locus of
Guerrero ML, Herrera-Insua I, Morrow AL, enterocyte effacement-encoded virulence deter-
Ruiz-Palacios G, Pickering LK, Guzman CA, minants. J Infect Dis 179:269274.
Cleary TG. 2003. Human milk secretory anti- 60. Calderon Toledo C, Arvidsson I, Karpman D.
bodies against attaching and effacing Escherichia 2011. Cross-reactive protection against entero-
coli antigens. Emerg Infect Dis 9:545551. hemorrhagic Escherichia coli infection by en-
51. Loureiro I, Frankel G, Adu-Bobie J, Dougan G, teropathogenic E. coli in a mouse model. Infect
Trabulsi LR, Carneiro-Sampaio MM. 1998. Immun 79:22242233.

61. Walters MD, Matthei IU, Kay R, Dillon MJ, 2-induced neutrophilia. Clin Exp Immunol 146:
Barratt TM. 1989. The polymorphonuclear 7684.
leucocyte count in childhood haemolytic uraemic 73. Fukuda MN, Dell A, Oates JE, Wu P, Klock JC,
syndrome. Pediatr Nephrol 3:130134. Fukuda M. 1985. Structures of glycosphingolipids
62. Robson WL, Fick GH, Wilson PC. 1988. Prog- isolated from human granulocytes. The pres-
nostic factors in typical postdiarrhea hemolytic- ence of a series of linear poly-N-acetyllactosa-
uremic syndrome. Child Nephrol Urol 9:203207. minylceramide and its significance in glycolipids
63. Fernandez GC, Gomez SA, Rubel CJ, Bentancor of whole blood cells. J Biol Chem 260:10671082.
LV, Barrionuevo P, Alduncin M, Grimoldi I, 74. Arfilli V, Carnicelli D, Rocchi L, Ricci F,
Exeni R, Isturiz MA, Palermo MS. 2005. Im- Pagliaro P, Tazzari PL, Brigotti M. 2010. Shiga
paired neutrophils in children with the typical toxin 1 and ricin A chain bind to human polymor-
form of hemolytic uremic syndrome. Pediatr phonuclear leucocytes through a common re-
Nephrol 20:13061314. ceptor. Biochem J 432:173180.
64. Milford D, Taylor CM, Rafaat F, Halloran E, 75. Geelen JM, van der Velden TJ, Te Loo DM,
Dawes J. 1989. Neutrophil elastases and haemo- Boerman OC, van den Heuvel LP, Monnens LA.
lytic uraemic syndrome. Lancet 2:1153. 2007. Lack of specific binding of Shiga-like toxin
65. Fitzpatrick MM, Shah V, Filler G, Dillon MJ, (verocytotoxin) and non-specific interaction of
Barratt TM. 1992. Neutrophil activation in the Shiga-like toxin 2 antibody with human poly-
haemolytic uraemic syndrome: free and com- morphonuclear leucocytes. Nephrol Dial Trans-
plexed elastase in plasma. Pediatr Nephrol 6:50 plant 22:749755.
53. 76. Te Loo DM, van Hinsbergh VW, van den
66. Hughes DA, Smith GC, Davidson JE, Murphy Heuvel LP, Monnens LA. 2001. Detection of
AV, Beattie TJ. 1996. The neutrophil oxidative verocytotoxin bound to circulating polymorpho-
burst in diarrhoea-associated haemolytic uraemic nuclear leukocytes of patients with hemolytic
syndrome. Pediatr Nephrol 10:445447. uremic syndrome. J Am Soc Nephrol 12:800806.
67. Forsyth KD, Simpson AC, Fitzpatrick MM, 77. Tazzari PL, Ricci F, Carnicelli D, Caprioli A,
Barratt TM, Levinsky RJ. 1989. Neutrophil- Tozzi AE, Rizzoni G, Conte R, Brigotti M. 2004.
mediated endothelial injury in haemolytic urae- Flow cytometry detection of Shiga toxins in the
mic syndrome. Lancet 2:411414. blood from children with hemolytic uremic syn-
68. Fernandez GC, Gomez SA, Ramos MV, drome. Cytometry B Clin Cytom 61:4044.
Bentancor LV, Fernandez-Brando RJ, Landoni 78. Sthl AL, Sartz L, Nelsson A, Bkssy ZD,
VI, Lopez L, Ramirez F, Diaz M, Alduncin Karpman D. 2009. Shiga toxin and lipopolysac-
M, Grimoldi I, Exeni R, Isturiz MA, Palermo charide induce platelet-leukocyte aggregates and
MS. 2007. The functional state of neutrophils tissue factor release, a thrombotic mechanism in
correlates with the severity of renal dysfunction hemolytic uremic syndrome. PLoS One 4:e6990.
in children with hemolytic uremic syndrome. 79. Brigotti M, Tazzari PL, Ravanelli E, Carnicelli
Pediatr Res 61:123128. D, Rocchi L, Arfilli V, Scavia G, Minelli F,
69. Fitzpatrick MM, Shah V, Trompeter RS, Dillon Ricci F, Pagliaro P, Ferretti AV, Pecoraro C,
MJ, Barratt TM. 1992. Interleukin-8 and Paglialonga F, Edefonti A, Procaccino MA,
polymorphoneutrophil leucocyte activation in Tozzi AE, Caprioli A. 2011. Clinical relevance of
hemolytic uremic syndrome of childhood. Kidney Shiga toxin concentrations in the blood of
Int 42:951956. patients with hemolytic uremic syndrome. Pediatr
70. Liu J, He T, He Y, Zhang Z, Akahoshi T, Kondo Infect Dis J 30:486490.
H, Zhong S. 2002. Prolongation of functional life- 80. Szabady RL, Lokuta MA, Walters KB,
span of neutrophils by recombinant verotoxin 2. Huttenlocher A, Welch RA. 2009. Modulation
Chin Med J (Engl) 115:900903. of neutrophil function by a secreted mucinase
71. Ge S, Hertel B, Emden SH, Beneke J, Menne J, of Escherichia coli O157:H7. PLoS Pathog 5:
Haller H, von Vietinghoff S. 2012. Microparticle e1000320.
generation and leucocyte death in Shiga toxin- 81. Fernandez GC, Ramos MV, Gomez SA, Dran GI,
mediated HUS. Nephrol Dial Transplant 27:2768 Exeni R, Alduncin M, Grimoldi I, Vallejo G,
2775. Elias-Costa C, Isturiz MA, Palermo MS. 2005.
72. Fernandez GC, Lopez MF, Gomez SA, Ramos Differential expression of function-related anti-
MV, Bentancor LV, Fernandez-Brando RJ, gens on blood monocytes in children with he-
Landoni VI, Dran GI, Meiss R, Isturiz MA, molytic uremic syndrome. J Leukoc Biol 78:853
Palermo MS. 2006. Relevance of neutrophils in 861.
the murine model of haemolytic uraemic syn- 82. Ramos MV, Fernandez GC, Patey N, Schierloh
drome: mechanisms involved in Shiga toxin type P, Exeni R, Grimoldi I, Vallejo G, Elias-Costa C,

Del Carmen Sasiain M, Trachtman H, cytopenic purpura and hemolytic uremic syn-
Combadiere C, Proulx F, Palermo MS. 2007. drome. Thromb Haemost 75:427431.
Involvement of the fractalkine pathway in the 94. Walters MD, Levin M, Smith C, Nokes TJ,
pathogenesis of childhood hemolytic uremic Hardisty RM, Dillon MJ, Barratt TM. 1988.
syndrome. Blood 109:24382445. Intravascular platelet activation in the hemolytic
83. van Setten PA, Monnens LA, Verstraten RG, uremic syndrome. Kidney Int 33:107115.
van den Heuvel LP, van Hinsbergh VW. 1996. 95. Keepers TR, Psotka MA, Gross LK, Obrig TG.
Effects of verocytotoxin-1 on nonadherent human 2006. A murine model of HUS: Shiga toxin with
monocytes: binding characteristics, protein syn- lipopolysaccharide mimics the renal damage and
thesis, and induction of cytokine release. Blood physiologic response of human disease. J Am Soc
88:174183. Nephrol 17:34043414.
84. Guessous F, Marcinkiewicz M, Polanowska- 96. Cooling LL, Walker KE, Gille T, Koerner TA.
Grabowska R, Keepers TR, Obrig T, Gear AR. 1998. Shiga toxin binds human platelets via
2005. Shiga toxin 2 and lipopolysaccharide cause globotriaosylceramide (Pk antigen) and a novel
monocytic THP-1 cells to release factors which platelet glycosphingolipid. Infect Immun 66:4355
activate platelet function. Thromb Haemost 94: 4366.
10191027. 97. Ghosh SA, Polanowska-Grabowska RK, Fujii J,
85. Murata K, Higuchi T, Takada K, Oida K, Horie Obrig T, Gear AR. 2004. Shiga toxin binds to ac-
S, Ishii H. 2006. Verotoxin-1 stimulation of tivated platelets. J Thromb Haemost 2:499506.
macrophage-like THP-1 cells up-regulates tissue 98. Sthl AL, Svensson M, Morgelin M, Svanborg
factor expression through activation of c-Yes ty- C, Tarr PI, Mooney JC, Watkins SL, Johnson R,
rosine kinase: possible signal transduction in tis- Karpman D. 2006. Lipopolysaccharide from
sue factor up-regulation. Biochim Biophys Acta enterohemorrhagic Escherichia coli binds to
1762:835843. platelets through TLR4 and CD62 and is detected
86. Del Conde I, Shrimpton CN, Thiagarajan P, on circulating platelets in patients with hemolytic
Lopez JA. 2005. Tissue-factor-bearing micro- uremic syndrome. Blood 108:167176.
vesicles arise from lipid rafts and fuse with acti- 99. Karpman D, Papadopoulou D, Nilsson K,
vated platelets to initiate coagulation. Blood Sjgren AC, Mikaelsson C, Lethagen S. 2001.
106:16041611. Platelet activation by Shiga toxin and circulatory
87. Egorina EM, Sovershaev MA, Olsen JO, factors as a pathogenetic mechanism in the he-
Osterud B. 2008. Granulocytes do not express molytic uremic syndrome. Blood 97:31003108.
but acquire monocyte-derived tissue factor in 100. Semple JW, Italiano JE Jr, Freedman J. 2011.
whole blood: evidence for a direct transfer. Blood Platelets and the immune continuum. Nat Rev
111:12081216. Immunol 11:264274.
88. Geelen JM, van der Velden TJ, van den Heuvel 101. Moore KL, Patel KD, Bruehl RE, Li F, Johnson
LP, Monnens LA. 2007. Interactions of Shiga-like DA, Lichenstein HS, Cummings RD, Bainton
toxin with human peripheral blood monocytes. DF, McEver RP. 1995. P-selectin glycoprotein
Pediatr Nephrol 22:11811187. ligand-1 mediates rolling of human neutrophils on
89. Proulx F, Seidman EG, Karpman D. 2001. P-selectin. J Cell Biol 128:661671.
Pathogenesis of Shiga toxin-associated hemolytic 102. Michelson AD, Barnard MR, Krueger LA, Valeri
uremic syndrome. Pediatr Res 50:163171. CR, Furman MI. 2001. Circulating monocyte-
90. Zoja C, Buelli S, Morigi M. 2010. Shiga toxin- platelet aggregates are a more sensitive marker of
associated hemolytic uremic syndrome: patho- in vivo platelet activation than platelet surface P-
physiology of endothelial dysfunction. Pediatr selectin: studies in baboons, human coronary in-
Nephrol 25:22312240. tervention, and human acute myocardial infarc-
91. Fong JS, Kaplan BS. 1982. Impairment of tion. Circulation 104:15331537.
platelet aggregation in hemolytic uremic syn- 103. Gear AR, Camerini D. 2003. Platelet chemokines
drome: evidence for platelet exhaustion. Blood and chemokine receptors: linking hemostasis,
60:564570. inflammation, and host defense. Microcirculation
92. Sassetti B, Vizcarguenaga MI, Zanaro NL, Silva 10:335350.
MV, Kordich L, Florentini L, Diaz M, Vitacco 104. Shiraki R, Inoue N, Kawasaki S, Takei A,
M, Sanchez Avalos JC. 1999. Hemolytic uremic Kadotani M, Ohnishi Y, Ejiri J, Kobayashi S,
syndrome in children: platelet aggregation and Hirata K, Kawashima S, Yokoyama M. 2004.
membrane glycoproteins. J Pediatr Hematol Expression of Toll-like receptors on human
Oncol 21:123128. platelets. Thromb Res 113:379385.
93. Galli M, Grassi A, Barbui T. 1996. Platelet- 105. Andonegui G, Kerfoot SM, McNagny K, Ebbert
derived microvesicles in thrombotic thrombo- KV, Patel KD, Kubes P. 2005. Platelets express

functional Toll-like receptor-4. Blood 106:2417 lipopolysaccharide binding protein in children

2423. with Escherichia coli O157:H7 hemorrhagic colitis
106. Aslam R, Speck ER, Kim M, Crow AR, Bang and hemolytic uremic syndrome. Clin Diagn Lab
KW, Nestel FP, Ni H, Lazarus AH, Freedman J, Immunol 6:773.
Semple JW. 2006. Platelet Toll-like receptor 117. Jerala R. 2007. Structural biology of the LPS
expression modulates lipopolysaccharide-induced recognition. Int J Med Microbiol 297:353363.
thrombocytopenia and tumor necrosis factor- 118. Valles PG, Melechuck S, Gonzalez A, Manucha
alpha production in vivo. Blood 107:637641. W, Bocanegra V, Valles R. 2012. Toll-like re-
107. Cicala C, Santacroce C, Itoh H, Douglas GJ, ceptor 4 expression on circulating leucocytes
Page CP. 1997. A study on rat platelet respon- in hemolytic uremic syndrome. Pediatr Nephrol
siveness following intravenous endotoxin admin- 27:407415.
istration. Life Sci 60:PL31PL38. 119. Kamitsuji H, Nonami K, Murakami T, Ishikawa
108. Jayachandran M, Brunn GJ, Karnicki K, Miller N, Nakayama A, Umeki Y. 2000. Elevated tissue
RS, Owen WG, Miller VM. 2007. In vivo effects factor circulating levels in children with hemo-
of lipopolysaccharide and TLR4 on platelet pro- lytic uremic syndrome caused by verotoxin-
duction and activity: implications for thrombotic producing E. coli. Clin Nephrol 53:319324.
risk. J Appl Physiol 102:429433. 120. Edgington TS, Mackman N, Brand K, Ruf W.
109. Scott T, Owens MD. 2008. Thrombocytes re- 1991. The structural biology of expression and
spond to lipopolysaccharide through Toll-like function of tissue factor. Thromb Haemost 66:
receptor-4, and MAP kinase and NF-kappaB 6779.
pathways leading to expression of interleukin-6 121. Rao LV, Rapaport SI, Bajaj SP. 1986. Activation
and cyclooxygenase-2 with production of pros- of human factor VII in the initiation of tissue
taglandin E2. Mol Immunol 45:10011008. factor-dependent coagulation. Blood 68:685691.
110. Elzey BD, Tian J, Jensen RJ, Swanson AK, Lees 122. Monroe DM, Hoffman M. 2006. What does it
JR, Lentz SR, Stein CS, Nieswandt B, Wang Y, take to make the perfect clot? Arterioscler Thromb
Davidson BL, Ratliff TL. 2003. Platelet-mediated Vasc Biol 26:4148.
modulation of adaptive immunity. A communi- 123. Camera M, Frigerio M, Toschi V, Brambilla M,
cation link between innate and adaptive immune Rossi F, Cottell DC, Maderna P, Parolari A,
compartments. Immunity 19:919. Bonzi R, De Vincenti O, Tremoli E. 2003.
111. Viisoreanu D, Polanowska-Grabowska R, Platelet activation induces cell-surface immuno-
Suttitanamongkol S, Obrig TG, Gear AR. 2000. reactive tissue factor expression, which is modu-
Human platelet aggregation is not altered by lated differently by antiplatelet drugs. Arterioscler
Shiga toxins 1 or 2. Thromb Res 98:403410. Thromb Vasc Biol 23:16901696.
112. Gawaz M, Neumann FJ, Dickfeld T, Koch W, 124. Schwertz H, Tolley ND, Foulks JM, Denis MM,
Laugwitz KL, Adelsberger H, Langenbrink K, Risenmay BW, Buerke M, Tilley RE, Rondina
Page S, Neumeier D, Schomig A, Brand K. 1998. MT, Harris EM, Kraiss LW, Mackman N,
Activated platelets induce monocyte chemotactic Zimmerman GA, Weyrich AS. 2006. Signal-
protein-1 secretion and surface expression of independent splicing of tissue factor pre-mRNA
tercellular adhesion molecule-1 on endothelial modulates the thrombogenicity of human plate-
cells. Circulation 98:11641171. lets. J Exp Med 203:24332440.
113. Henn V, Slupsky JR, Grafe M, Anagnostopoulos 125. Panes O, Matus V, Saez CG, Quiroga T, Pereira
I, Forster R, Muller-Berghaus G, Kroczek RA. J, Mezzano D. 2007. Human platelets synthesize
1998. CD40 ligand on activated platelets triggers and express functional tissue factor. Blood 109:
an inflammatory reaction of endothelial cells. 52425250.
Nature 391:591594. 126. del Conde I, Nabi F, Tonda R, Thiagarajan P,
114. Hollenbaugh D, Mischel-Petty N, Edwards CP, Lopez JA, Kleiman NS. 2005. Effect of P-selectin
Simon JC, Denfeld RW, Kiener PA, Aruffo A. on phosphatidylserine exposure and surface-
1995. Expression of functional CD40 by vascular dependent thrombin generation on monocytes.
endothelial cells. J Exp Med 182:3340. Arterioscler Thromb Vasc Biol 25:10651070.
115. Caprioli A, Luzzi I, Rosmini F, Resti C, Edefonti 127. sterud B, Olsen JO. 2013. Human platelets do
A, Perfumo F, Farina C, Goglio A, Gianviti A, not express tissue factor. Thromb Res 132:112115.
Rizzoni G. 1994. Community-wide outbreak of 128. Bolande RP, Kaplan BS. 1985. Experimental
hemolytic-uremic syndrome associated with non- studies on the hemolytic-uremic syndrome.
O157 verocytotoxin-producing Escherichia coli. Nephron 39:228236.
J Infect Dis 169:208211. 129. Appiani AC, Edefonti A, Bettinelli A, Cossu
116. Proulx F, Seidman E, Mariscalco MM, Lee K, MM, Paracchini ML, Rossi E. 1982. The rela-
Caroll S. 1999. Increased circulating levels of tionship between plasma levels of the factor VIII

complex and platelet release products (beta- P, Chongsa-nguan M, Chaicumpa W. 2001. Lo-
thromboglobulin and platelet factor 4) in children calization of Shiga toxins of enterohaemorrhagic
with the hemolytic-uremic syndrome. Clin Escherichia coli in kidneys of paediatric and geri-
Nephrol 17:195199. atric patients with fatal haemolytic uraemic syn-
130. van de Kar NC, van Hinsbergh VW, Brommer drome. Microb Pathog 31:5967.
EJ, Monnens LA. 1994. The fibrinolytic system in 141. Uchida H, Kiyokawa N, Horie H, Fujimoto J,
the hemolytic uremic syndrome: in vivo and in Takeda T. 1999. The detection of Shiga toxins in
vitro studies. Pediatr Res 36:257264. the kidney of a patient with hemolytic uremic
131. Tsai HM, Chandler WL, Sarode R, Hoffman R, syndrome. Pediatr Res 45:133137.
Jelacic S, Habeeb RL, Watkins SL, Wong CS, 142. Karpman D, Hkansson A, Perez MT, Isaksson
Williams GD, Tarr PI. 2001. von Willebrand C, Carlemalm E, Caprioli A, Svanborg C. 1998.
factor and von Willebrand factor-cleaving metal- Apoptosis of renal cortical cells in the hemolytic-
loprotease activity in Escherichia coli O157:H7- uremic syndrome: in vivo and in vitro studies.
associated hemolytic uremic syndrome. Pediatr Infect Immun 66:636644.
Res 49:653659. 143. Buteau C, Proulx F, Chaibou M, Raymond D,
132. Morigi M, Galbusera M, Binda E, Imberti B, Clermont MJ, Mariscalco MM, Lebel MH,
Gastoldi S, Remuzzi A, Zoja C, Remuzzi G. 2001. Seidman E. 2000. Leukocytosis in children with
Verotoxin-1-induced up-regulation of adhesive Escherichia coli O157:H7 enteritis developing the
molecules renders microvascular endothelial cells hemolytic-uremic syndrome. Pediatr Infect Dis J
thrombogenic at high shear stress. Blood 98:1828 19:642647.
1835. 144. Salzman MB, Ettenger RB, Cherry JD. 1991.
133. Nolasco LH, Turner NA, Bernardo A, Tao Z, Leukocytosis in hemolytic-uremic syndrome.
Cleary TG, Dong JF, Moake JL. 2005. Hemo- Pediatr Infect Dis J 10:470471.
lytic uremic syndrome-associated Shiga toxins 145. Coad NA, Marshall T, Rowe B, Taylor CM. 1991.
promote endothelial-cell secretion and impair Changes in the postenteropathic form of the he-
ADAMTS13 cleavage of unusually large von molytic uremic syndrome in children. Clin
Willebrand factor multimers. Blood 106:4199 Nephrol 35:1016.
4209. 146. Inward CD, Howie AJ, Fitzpatrick MM, Rafaat
134. Guessous F, Marcinkiewicz M, Polanowska- F, Milford DV, Taylor CM. 1997. Renal histo-
Grabowska R, Kongkhum S, Heatherly D, Obrig pathology in fatal cases of diarrhoea-associated
T, Gear AR. 2005. Shiga toxin 2 and lipopoly- haemolytic uraemic syndrome. British Associa-
saccharide induce human microvascular endo- tion for Paediatric Nephrology. Pediatr Nephrol
thelial cells to release chemokines and factors that 11:556559.
stimulate platelet function. Infect Immun 73: 147. Roche JK, Keepers TR, Gross LK, Seaner RM,
83068316. Obrig TG. 2007. CXCL1/KC and CXCL2/MIP-2
135. Furie B, Furie BC. 2008. Mechanisms of are critical effectors and potential targets for
thrombus formation. N Engl J Med 359:938949. therapy of Escherichia coli O157:H7-associated
136. Nevard CH, Jurd KM, Lane DA, Philippou H, renal inflammation. Am J Pathol 170:526537.
Haycock GB, Hunt BJ. 1997. Activation of coag- 148. Keepers TR, Gross LK, Obrig TG. 2007. Mono-
ulation and fibrinolysis in childhood diarrhoea- cyte chemoattractant protein 1, macrophage in-
associated haemolytic uraemic syndrome. Thromb flammatory protein 1 alpha, and RANTES recruit
Haemost 78:14501455. macrophages to the kidney in a mouse model of
137. Van Geet C, Proesmans W, Arnout J, Vermylen hemolytic-uremic syndrome. Infect Immun 75:
J, Declerck PJ. 1998. Activation of both coagu- 12291236.
lation and fibrinolysis in childhood hemolytic 149. Garcia A, Marini RP, Catalfamo JL, Knox KA,
uremic syndrome. Kidney Int 54:13241330. Schauer DB, Rogers AB, Fox JG. 2008. Intra-
138. Chandler WL, Jelacic S, Boster DR, Ciol MA, venous Shiga toxin 2 promotes enteritis and renal
Williams GD, Watkins SL, Igarashi T, Tarr PI. injury characterized by polymorphonuclear leu-
2002. Prothrombotic coagulation abnormalities kocyte infiltration and thrombosis in Dutch Belt-
preceding the hemolytic-uremic syndrome. N ed rabbits. Microbes Infect 10:650656.
Engl J Med 346:2332. 150. Zoja C, Angioletti S, Donadelli R, Zanchi C,
139. Bergstein JM, Riley M, Bang NU. 1992. Role of Tomasoni S, Binda E, Imberti B, te Loo M,
plasminogen-activator inhibitor type 1 in the Monnens L, Remuzzi G, Morigi M. 2002. Shiga
pathogenesis and outcome of the hemolytic ure- toxin-2 triggers endothelial leukocyte adhesion
mic syndrome. N Engl J Med 327:755759. and transmigration via NF-kappaB dependent up-
140. Chaisri U, Nagata M, Kurazono H, Horie H, regulation of IL-8 and MCP-1. Kidney Int 62:846
Tongtawe P, Hayashi H, Watanabe T, Tapchaisri 856.

151. Zanchi C, Zoja C, Morigi M, Valsecchi F, Liu lytic uremic syndrome. Am J Kidney Dis 36:687
XY, Rottoli D, Locatelli M, Buelli S, Pezzotta A, 694.
Mapelli P, Geelen J, Remuzzi G, Hawiger J. 161. Proulx F, Turgeon JP, Litalien C, Mariscalco
2008. Fractalkine and CX3CR1 mediate leukocyte MM, Robitaille P, Seidman E. 1998. Inflamma-
capture by endothelium in response to Shiga tory mediators in Escherichia coli O157:H7 hem-
toxin. J Immunol 181:14601469. orrhagic colitis and hemolytic-uremic syndrome.
152. Morigi M, Micheletti G, Figliuzzi M, Imberti B, Pediatr Infect Dis J 17:899904.
Karmali MA, Remuzzi A, Remuzzi G, Zoja C. 162. Litalien C, Proulx F, Mariscalco MM, Robitaille
1995. Verotoxin-1 promotes leukocyte adhesion to P, Turgeon JP, Orrbine E, Rowe PC, McLaine
cultured endothelial cells under physiologic flow PN, Seidman E. 1999. Circulating inflammatory
conditions. Blood 86:45534558. cytokine levels in hemolytic uremic syndrome.
153. Brigotti M, Caprioli A, Tozzi AE, Tazzari PL, Pediatr Nephrol 13:840845.
Ricci F, Conte R, Carnicelli D, Procaccino MA, 163. van de Kar NC, Sauerwein RW, Demacker PN,
Minelli F, Ferretti AV, Paglialonga F, Edefonti Grau GE, van Hinsbergh VW, Monnens LA.
A, Rizzoni G. 2006. Shiga toxins present in the 1995. Plasma cytokine levels in hemolytic uremic
gut and in the polymorphonuclear leukocytes syndrome. Nephron 71:309313.
circulating in the blood of children with hemo- 164. Karpman D, Andreasson A, Thysell H, Kaplan
lytic-uremic syndrome. J Clin Microbiol 44:313 BS, Svanborg C. 1995. Cytokines in childhood
317. hemolytic uremic syndrome and thrombotic
154. Karpman D, Manea M, Vaziri-Sani F, Sthl AL, thrombocytopenic purpura. Pediatr Nephrol 9:
Kristoffersson AC. 2006. Platelet activation in 694699.
hemolytic uremic syndrome. Semin Thromb 165. Proulx F, Litalien C, Turgeon JP, Mariscalco
Hemost 32:128145. MM, Seidman E. 2000. Circulating levels of
155. Stearns-Kurosawa DJ, Oh SY, Cherla RP, Lee transforming growth factor-beta1 and lympho-
MS, Tesh VL, Papin J, Henderson J, Kurosawa kines among children with hemolytic uremic
S. 2013. Distinct renal pathology and a chemo- syndrome. Am J Kidney Dis 35:2934.
tactic phenotype after enterohemorrhagic Esche- 166. Inward CD, Varagunam M, Adu D, Milford
richia coli Shiga toxins in non-human primate DV, Taylor CM. 1997. Cytokines in haemolytic
models of hemolytic uremic syndrome. Am J uraemic syndrome associated with verocytotoxin-
Pathol 182:12271238. producing Escherichia coli infection. Arch Dis
156. Lopez EL, Devoto S, Fayad A, Canepa C, Child 77:145147.
Morrow AL, Cleary TG. 1992. Association be- 167. Inward CD, Pall AA, Adu D, Milford DV, Taylor
tween severity of gastrointestinal prodrome and CM. 1995. Soluble circulating cell adhesion mol-
long-term prognosis in classic hemolytic-uremic ecules in haemolytic uraemic syndrome. Pediatr
syndrome. J Pediatr 120:210215. Nephrol 9:574578.
157. van Setten PA, van Hinsbergh VW, van den 168. Murata A, Shimazu T, Yamamoto T, Taenaka
Heuvel LP, Preyers F, Dijkman HB, Assmann N, Nagayama K, Honda T, Sugimoto H, Monden
KJ, van der Velden TJ, Monnens LA. 1998. M, Matsuura N, Okada S. 1998. Profiles of cir-
Monocyte chemoattractant protein-1 and inter- culating inflammatory- and anti-inflammatory
leukin-8 levels in urine and serum of patents with cytokines in patients with hemolytic uremic syn-
hemolytic uremic syndrome. Pediatr Res 43:759 drome due to E. coli O157 infection. Cytokine
767. 10:544548.
158. Decaluwe H, Harrison LM, Mariscalco MM, 169. Yamamoto T, Nagayama K, Satomura K, Honda
Gendrel D, Bohuon C, Tesh VL, Proulx F. 2006. T, Okada S. 2000. Increased serum IL-10 and
Procalcitonin in children with Escherichia coli endothelin levels in hemolytic uremic syndrome
O157:H7 associated hemolytic uremic syndrome. caused by Escherichia coli O157. Nephron 84:326
Pediatr Res 59:579583. 332.
159. Proulx F, Toledano B, Phan V, Clermont MJ, 170. Lopez EL, Contrini MM, Devoto S, de Rosa MF,
Mariscalco MM, Seidman EG. 2002. Circulating Grana MG, Genero MH, Canepa C, Gomez HF,
granulocyte colony-stimulating factor, C-X-C, and Cleary TG. 1995. Tumor necrosis factor concen-
C-C chemokines in children with Escherichia coli trations in hemolytic uremic syndrome patients
O157:H7 associated hemolytic uremic syndrome. and children with bloody diarrhea in Argentina.
Pediatr Res 52:928934. Pediatr Infect Dis J 14:594598.
160. Masri C, Proulx F, Toledano B, Clermont MJ, 171. Nevard CH, Blann AD, Jurd KM, Haycock GB,
Mariscalco MM, Seidman EG, Carcillo J. 2000. Hunt BJ. 1999. Markers of endothelial cell acti-
Soluble Fas and soluble Fas-ligand in children vation and injury in childhood haemolytic urae-
with Escherichia coli O157:H7-associated hemo- mic syndrome. Pediatr Nephrol 13:487492.

172. Caletti MG, Balestracci A, Roy AH. 2010. Levels 182. Harel Y, Silva M, Giroir B, Weinberg A, Cleary
of urinary transforming growth factor beta-1 in TB, Beutler B. 1993. A reporter transgene
children with D+ hemolytic uremic syndrome. indicates renal-specific induction of tumor ne-
Pediatr Nephrol 25:11771180. crosis factor (TNF) by Shiga-like toxin. Possible
173. Bhowmik D. 2001. Elevated tissue factor levels in involvement of TNF in hemolytic uremic syn-
children with hemolytic uremic syndrome. Clin drome. J Clin Invest 92:21102116.
Nephrol 55:262. 183. Lentz EK, Cherla RP, Jaspers V, Weeks BR,
174. Petruzziello-Pellegrini TN, Yuen DA, Page AV, Tesh VL. 2010. Role of tumor necrosis factor
Patel S, Soltyk AM, Matouk CC, Wong DK, alpha in disease using a mouse model of Shiga
Turgeon PJ, Fish JE, Ho JJ, Steer BM, Khajoee toxin-mediated renal damage. Infect Immun 78:
V, Tigdi J, Lee WL, Motto DG, Advani A, 36893699.
Gilbert RE, Karumanchi SA, Robinson LA, Tarr 184. Wolski VM, Soltyk AM, Brunton JL. 2002.
PI, Liles WC, Brunton JL, Marsden PA. 2012. Tumour necrosis factor alpha is not an essential
The CXCR4/CXCR7/SDF-1 pathway contributes component of verotoxin 1-induced toxicity in
to the pathogenesis of Shiga toxin-associated mice. Microb Pathog 32:263271.
hemolytic uremic syndrome in humans and mice. 185. Hughes AK, Stricklett PK, Kohan DE. 1998.
J Clin Invest 122:759776. Shiga toxin-1 regulation of cytokine production by
175. Louise CB, Obrig TG. 1991. Shiga toxin-associ- human proximal tubule cells. Kidney Int 54:1093
ated hemolytic-uremic syndrome: combined cy- 1106.
totoxic effects of Shiga toxin, interleukin-1 beta, 186. Nakamura A, Johns EJ, Imaizumi A, Yanagawa
and tumor necrosis factor alpha on human vas- Y, Kohsaka T. 2001. Activation of beta(2)-
cular endothelial cells in vitro. Infect Immun adrenoceptor prevents shiga toxin 2-induced
59:41734179. TNF-alpha gene transcription. J Am Soc Nephrol
176. Keusch GT, Acheson DW, Aaldering L, Erban J, 12:22882299.
Jacewicz MS. 1996. Comparison of the effects 187. Taylor FB Jr, Tesh VL, DeBault L, Li A, Chang
of Shiga-like toxin 1 on cytokine- and butyrate- AC, Kosanke SD, Pysher TJ, Siegler RL. 1999.
treated human umbilical and saphenous vein en- Characterization of the baboon responses to
dothelial cells. J Infect Dis 173:11641170. Shiga-like toxin: descriptive study of a new pri-
177. van Setten PA, van Hinsbergh VW, van der mate model of toxic responses to Stx-1. Am J
Velden TJ, van de Kar NC, Vermeer M, Mahan Pathol 154:12851299.
JD, Assmann KJ, van den Heuvel LP, Monnens 188. Hughes AK, Stricklett PK, Kohan DE. 2001. Shiga
LA. 1997. Effects of TNF alpha on verocytotoxin toxin-1 regulation of cytokine production by human
cytotoxicity in purified human glomerular micro- glomerular epithelial cells. Nephron 88:1423.
vascular endothelial cells. Kidney Int 51:12451256. 189. Akira S, Hirano T, Taga T, Kishimoto T. 1990.
178. van de Kar NC, Monnens LA, Karmali MA, van Biology of multifunctional cytokines: IL 6 and
Hinsbergh VW. 1992. Tumor necrosis factor and related molecules (IL 1 and TNF). FASEB J 4:
interleukin-1 induce expression of the verocyto- 28602867.
toxin receptor globotriaosylceramide on human 190. Ramos MV, Auvynet C, Poupel L, Rodero M,
endothelial cells: implications for the pathogene- Mejias MP, Panek CA, Vanzulli S, Combadiere
sis of the hemolytic uremic syndrome. Blood C, Palermo M. 2012. Chemokine receptor CCR1
80:27552764. disruption limits renal damage in a murine model
179. Louise CB, Obrig TG. 1992. Shiga toxin-associated of hemolytic uremic syndrome. Am J Pathol 180:
hemolytic uremic syndrome: combined cytotoxic 10401048.
effects of shiga toxin and lipopolysaccharide (en- 191. Karpman D, Sartz L, Johnson S. 2010. Patho-
dotoxin) on human vascular endothelial cells in physiology of typical hemolytic uremic syndrome.
vitro. Infect Immun 60:15361543. Semin Thromb Hemost 36:575585.
180. Kaye SA, Louise CB, Boyd B, Lingwood CA, 192. Nestoridi E, Kushak RI, Duguerre D, Grabowski
Obrig TG. 1993. Shiga toxin-associated hemolytic EF, Ingelfinger JR. 2005. Up-regulation of tissue
uremic syndrome: interleukin-1 beta enhance- factor activity on human proximal tubular epi-
ment of Shiga toxin cytotoxicity toward human thelial cells in response to Shiga toxin. Kidney Int
vascular endothelial cells in vitro. Infect Immun 67:22542266.
61:38863891. 193. Nestoridi E, Tsukurov O, Kushak RI, Ingelfinger
181. Isogai E, Isogai H, Kimura K, Hayashi S, Kubota JR, Grabowski EF. 2005. Shiga toxin enhances
T, Fujii N, Takeshi K. 1998. Role of tumor ne- functional tissue factor on human glomerular en-
crosis factor alpha in gnotobiotic mice infected dothelial cells: implications for the pathophysiol-
with an Escherichia coli O157:H7 strain. Infect ogy of hemolytic uremic syndrome. J Thromb
Immun 66:197202. Haemost 3:752762.

194. Braune SA, Wichmann D, von Heinz MC, 204. Fujii J, Kinoshita Y, Matsukawa A, Villanueva
Nierhaus A, Becker H, Meyer TN, Meyer GP, SY, Yutsudo T, Yoshida S. 2009. Successful
Muller-Schulz M, Fricke J, de Weerth A, steroid pulse therapy for brain lesion caused by
Hoepker WW, Fiehler J, Magnus T, Gerloff C, Shiga toxin 2 in rabbits. Microb Pathog 46:179
Panzer U, Stahl RA, Wegscheider K, Kluge S. 184.
2013. Clinical features of critically ill patients with 205. Ramegowda B, Samuel JE, Tesh VL. 1999.
Shiga toxin-induced hemolytic uremic syndrome. Interaction of Shiga toxins with human brain
Crit Care Med 41:17021710. microvascular endothelial cells: cytokines as sen-
195. Magnus T, Rother J, Simova O, Meier-Cillien sitizing agents. J Infect Dis 180:12051213.
M, Repenthin J, Moller F, Gbadamosi J, Panzer 206. Eisenhauer PB, Chaturvedi P, Fine RE, Ritchie
U, Wengenroth M, Hagel C, Kluge S, Stahl RK, AJ, Pober JS, Cleary TG, Newburg DS. 2001.
Wegscheider K, Urban P, Eckert B, Glatzel M, Tumor necrosis factor alpha increases human
Fiehler J, Gerloff C. 2012. The neurological cerebral endothelial cell Gb3 and sensitivity to
syndrome in adults during the 2011 northern Shiga toxin. Infect Immun 69:18891894.
German E. coli serotype O104:H4 outbreak. Brain 207. Stricklett PK, Hughes AK, Kohan DE. 2005. In-
135:18501859. hibition of p38 mitogen-activated protein kinase
196. Obata F, Tohyama K, Bonev AD, Kolling GL, ameliorates cytokine up-regulated shigatoxin-1
Keepers TR, Gross LK, Nelson MT, Sato S, toxicity in human brain microvascular endothelial
Obrig TG. 2008. Shiga toxin 2 affects the central cells. J Infect Dis 191:461471.
nervous system through receptor globotriaosyl- 208. Landoni VI, de Campos-Nebel M, Schierloh P,
ceramide localized to neurons. J Infect Dis 198: Calatayud C, Fernandez GC, Ramos MV, Rearte
13981406. B, Palermo MS, Isturiz MA. 2010. Shiga toxin
197. Ergonul Z, Hughes AK, Kohan DE. 2003. In- 1-induced inflammatory response in lipopolysac-
duction of apoptosis of human brain microvas- charide-sensitized astrocytes is mediated by en-
cular endothelial cells by shiga toxin 1. J Infect Dis dogenous tumor necrosis factor alpha. Infect
187:154158. Immun 78:11931201.
198. Fujii J, Wood K, Matsuda F, Carneiro-Filho BA, 209. Landoni VI, Schierloh P, de Campos Nebel M,
Schlegel KH, Yutsudo T, Binnington-Boyd B, Fernandez GC, Calatayud C, Lapponi MJ,
Lingwood CA, Obata F, Kim KS, Yoshida S, Isturiz MA. 2012. Shiga toxin 1 induces on lipo-
Obrig T. 2008. Shiga toxin 2 causes apoptosis in polysaccharide-treated astrocytes the release of
human brain microvascular endothelial cells via tumor necrosis factor-alpha that alter brain-like
C/EBP homologous protein. Infect Immun endothelium integrity. PLoS Pathog 8:e1002632.
76:36793689. 210. Monnens L, Molenaar J, Lambert PH, Proesmans
199. Shiraishi M, Ichiyama T, Matsushige T, Iwaki T, W, van Munster P. 1980. The complement system
Iyoda K, Fukuda K, Makata H, Matsubara T, in hemolytic-uremic syndrome in childhood. Clin
Furukawa S. 2008. Soluble tumor necrosis factor Nephrol 13:168171.
receptor 1 and tissue inhibitor of metalloproteinase- 211. Robson WL, Leung AK, Fick GH, McKenna AI.
1 in hemolytic uremic syndrome with encephalop- 1992. Hypocomplementemia and leukocytosis in
athy. J Neuroimmunol 196:147152. diarrhea-associated hemolytic uremic syndrome.
200. Sheth KJ, Swick HM, Haworth N. 1986. Neu- Nephron 62:296299.
rological involvement in hemolytic-uremic syn- 212. Thurman JM, Marians R, Emlen W, Wood S,
drome. Ann Neurol 19:9093. Smith C, Akana H, Holers VM, Lesser M, Kline
201. Fujii J, Kita T, Yoshida S, Takeda T, Kobayashi M, Hoffman C, Christen E, Trachtman H. 2009.
H, Tanaka N, Ohsato K, Mizuguchi Y. 1994. Alternative pathway of complement in children
Direct evidence of neuron impairment by oral with diarrhea-associated hemolytic uremic syn-
infection with verotoxin-producing Escherichia drome. Clin J Am Soc Nephrol 4:19201924.
coli O157:H- in mitomycin-treated mice. Infect 213. Sthl AL, Sartz L, Karpman D. 2011. Comple-
Immun 62:34473453. ment activation on platelet-leukocyte complexes
202. Sofroniew MV, Vinters HV. 2010. Astrocytes: and microparticles in enterohemorrhagic Esche-
biology and pathology. Acta Neuropathol 119:7 richia coli-induced hemolytic uremic syndrome.
35. Blood 117:55035513.
203. Amran MY, Fujii J, Suzuki SO, Kolling GL, 214. Polley MJ, Nachman R. 1978. The human com-
Villanueva SY, Kainuma M, Kobayashi H, plement system in thrombin-mediated platelet
Kameyama H, Yoshida S. 2013. Investigation function. J Exp Med 147:17131726.
of encephalopathy caused by Shiga toxin 2c- 215. Polley MJ, Nachman RL. 1983. Human platelet
producing Escherichia coli infection in mice. PLoS activation by C3a and C3a des-arg. J Exp Med
One 8:e58959. 158:603615.

216. Orth D, Khan AB, Naim A, Grif K, Brockmeyer lectin in children with Escherichia coli O157:H7
J, Karch H, Joannidis M, Clark SJ, Day AJ, haemmorrhagic colitis and haemolytic uraemic
Fidanzi S, Stoiber H, Dierich MP, Zimmerhackl syndrome. Clin Exp Immunol 133:360363.
LB, Wurzner R. 2009. Shiga toxin activates 219. Karpman D, Tati R. 2013. Complement activation
complement and binds factor H: evidence for an in thrombotic microangiopathy. Hamostaseologie
active role of complement in hemolytic uremic 33:96104.
syndrome. J Immunol 182:63946400. 220. Ueki T, Mizuno M, Uesu T, Kiso T, Nasu J,
217. Buelli S, Abbate M, Morigi M, Moioli D, Zanchi Inaba T, Kihara Y, Matsuoka Y, Okada H, Fujita
C, Noris M, Zoja C, Pusey CD, Zipfel PF, T, Tsuji T. 1996. Distribution of activated com-
Remuzzi G. 2009. Protein load impairs factor H plement, C3b, and its degraded fragments, iC3b/
binding promoting complement-dependent dys- C3dg, in the colonic mucosa of ulcerative colitis
function of proximal tubular cells. Kidney Int (UC). Clin Exp Immunol 104:286292.
75:10501059. 221. Noris M, Mescia F, Remuzzi G. 2012. STEC-HUS,
218. Proulx F, Wagner E, Toledano B, Decaluwe H, atypical HUS and TTP are all diseases of comple-
Seidman EG, Rivard GE. 2003. Mannan-binding ment activation. Nat Rev Nephrol 8:622633.

The Interplay between the
Microbiota and Enterohemorrhagic
Escherichia coli
REED PIFER1,2 and VANESSA SPERANDIO1,2

20
INTRODUCTION
The human gastrointestinal (GI) tract harbors trillions of bacterial cells be-
longing to more than 1,000 species (1), and the number of bacterial cells within
the GI tract is 10 times higher than the number of human cells within our bodies
(2). The GI microbiota plays essential roles in human nutrition, physiology,
development, immunity, and behavior, with disruption of the structure and
balance of this community leading to dysbiosis and disease (35). This funda-
mental association between host and bacteria relies on chemical signaling
and nutrient availability and exchange. It is also clear that this important bal-
ance between host and microbiota can be severely disrupted by environmental
stimuli. Two of the most common insults on the microbiota that induce
dysbiosis are antibiotic treatment and infectious diseases. Both insults can
lead to several disease states ranging from autism, to inflammatory bowel dis-
ease, to inflammatory bowel syndrome (IBS). It is also noteworthy that stress
exacerbates these syndromes (3).
Broad-spectrum antibiotics alter the microbiome by reducing diversity and
shifting community composition (6). Although most of the microbiota return
after treatment is ceased, some members of this community are lost indefinitely
1
Department of Microbiology, The University of Texas Southwestern Medical Center, Dallas, TX 75390; 2Department of
Biochemistry, The University of Texas Southwestern Medical Center, Dallas, TX 75390.
403

404 PIFER AND SPERANDIO
(6). These community shifts cause changes the number of anaerobes and increasing the
in the metabolic profiles of the intestine, de- numbers of Gammaproteobacteria (18). C.
creasing concentrations of amino acids and rodentium is a murine pathogen that models
short-chain fatty acids (SCFAs) and increasing the enteric infection of the human pathogen
oligosaccharide levels, suggesting that micro- enterohemorrhagic Escherichia coli (EHEC).
bial fermentation of carbohydrates, a funda- The most compelling evidence for the in-
mental feature of the microbial flora, was volvement of the microbiota in disease states
disrupted (711). Depletion of SCFAs through comes from IBS, with the chief support com-
antibiotic treatment has critical implications ing from data on postinfection IBS, whereby
in intestinal health and immunity. SCFAs are IBS ensues following an episode of bac-
rapidly absorbed in the colon and provide a teriological gastroenteritis (3). The highest
preferred energy source to enterocytes, regu- supported incidence of postinfection IBS is
lating cell proliferation, differentiation, and related to the Walkerton outbreak, when post-
apoptosis (12, 13), as well as many aspects of infection IBS developed in the majority of
intestinal immunity (12, 13). Butyrate is the individuals after an EHEC infection (20).
SCFA with the strongest effect on cell cycle
and plays an anti-inflammatory role in the gut.
These anti-inflammatory properties are ben- MICROBIOTA/EHEC QUORUM-SENSING
eficial because they prevent overinflamma- SIGNALING INTERACTIONS WITHIN
tion and carcinogenesis (14). Butyrate acts by THE INTESTINE
inducing prostaglandin E2 (15). It is also worth
mentioning that in Crohns disease (an in- EHEC colonizes the human colon, where it
flammatory bowel disease pattern) there are has to interact and successfully compete with
decreased levels of prostaglandin E2 and bu- the microbiota to find a colonization niche.
tyrate (16), again suggesting that disruption The first appreciation of EHEC-microbiota
of the metabolome by dysbiosis has im- interactions on EHEC pathogenesis came
portant consequences in GI tract health. An- from the observation that EHEC employs
other fundamental impact of dysbyosis in quorum-sensing (QS) signaling to regulate ex-
the metabolome is disruption of carbohydrate pression of its virulence genes (21). QS is a
fermentation. Primary fermenters such as bacterial cell-to-cell signaling mechanism
Bacteroides spp. are the gateway for the en- through which bacterial cells assess the den-
trance of carbohydrates in the network of sity of their population. These bacteria secrete
syntrophic links in the microbiota, with Bac- hormone-like compounds, usually referred to
teriodes spp., a prominent glycophagic species, as autoinducers. When these autoinducers
degrading complex carbohydrates into mono- reach a threshold concentration, they interact
saccharides that can be consumed by other with transcriptional regulators to drive bac-
members of the gut microbiota (17). terial gene expression (22). Because EHEC
The second important insult that causes infection requires a low infectious dose, esti-
dysbiosis is infection by an invading pathogen. mated to be between 50 and 100 CFU, it was
It is known that invading enteric patho- deemed unlikely that EHEC was responding
gens such as Salmonella spp. and Citrobacter to self-produced signals upon infection of
rodentium cause inflammation within the gut the host, and it was proposed that EHEC was
that in turn diminishes the overall numbers sensing autoinducers produced by the GI
of bacteria in the microbiota, acting as a microbiota (21). The identity of this QS signal
competition advantage to the pathogen (18, was initially a matter of debate. QS was first
19). Additionally, infection with C. rodentium described in the regulation of biolumines-
also causes significant changes in the struc- cence in Vibrio fischeri and Vibrio harveyi (23,
ture of the microbial community, decreasing 24). The luciferase operon in V. fischeri is

CHAPTER 20 The Interplay between Microbiota and STEC 405
regulated by two proteins, LuxI, which is re- (AI-2) system, present in both Gram-positive
sponsible for the production of the acyl- and Gram-negative bacteria (including EHEC),
homoserine-lactone (AHL) autoinducer, and suggested that other QS systems, evolved to
LuxR, which is activated by this autoinducer promote interspecies communication within
to increase transcription of the luciferase op- bacterial populations, may be employed by
eron (25, 26). Since this initial description, EHEC to regulate its virulence repertoire in
homologs of LuxR-LuxI have been identified the GI tract (21, 33).
in other gram-negative bacteria, and in all of AI-2 was originally identified in V. harveyi
these LuxR-LuxI systems, the bacteria pro- as an inducer of cell density-dependent bio-
duce an AHL, which binds to the LuxR pro- luminescence. AI-2 is synthesized from S-
tein and regulates the transcription of several adenosylmethionine in three enzymatic steps,
genes involved in a variety of phenotypes (27 the last of which is catalyzed by LuxS pro-
29). E. coli has a LuxR homolog, SdiA (30), but ducing 4,5-dihydroxy-2,3-pentanedione that
does not have a luxI gene and does not pro- spontaneously cyclizes to form (2S,4S)-2-
duce AHLs (31, 32). Because of these features methyl-2,3,3,4-tetrahydroxytetrahydrofuran
it was considered unlikely for many years (AI-2) (Fig. 1) (34). LuxS is highly conserved,
that EHEC had any functional QS systems. with homologs found in a wide variety of
However, the discovery of the autoinducer-2 pathogenic and commensal bacteria within
FIGURE 1 Structure of known quorum-sensing ligands. N-hexanoyl-l-homoserine lactone (C6-HSL) (A) and N-
(3-Oxo-octanoyl)-l-homoserine lactone (3-oxo-C8-HSL) (B) stabilize SdiA, which can suppress T3S. (2S,4S)-2-
methyl-2,3,3,4-tetrahydroxytetrahydrofuran (AI-2) (C) appears to have a surprisingly modest effect on viru-
lence. QseC responds to host-derived epinephrine (D) and is antagonized by synthetic LED209 (E), perhaps
yielding clues to the identity of AI-3. Indole (F) is a tryptophan-derived metabolite that inuences motility
and type III secretion. Mucin degradation releases fucose (G), which activates the FusKR two-component system
to downregulate T3S. SCFAs, including acetate (H) and butyrate (I), induce motility of EHEC, while only butyrate
induces T3S via Lrp activity. doi:10.1128/microbiolspec.EHEC-0015-2013.f1

Bacteroidetes, Firmicutes, and Proteobacteria. kinase (HK) QseC, sensing the bacterial signal
Human stool samples and in vitro propagated AI-3 and the host signals epinephrine and/or
commensal bacteria from healthy individuals NE (41). Both epinephrine and NE are present
are capable of inducing luminescence from in the GI tract. NE is synthesized within the
V. harveyi, suggesting that AI-2 is produced adrenergic neurons within the enteric nervous
within the GI tract (35). EHEC encodes a system (42). Epinephrine is synthesized in
LuxS homolog, and preconditioned media the central nervous system and in the adrenal
from this strain are capable of inducing medulla; it acts in a systemic manner after
luminescence from V. harveyi in a luxS- being released into the bloodstream, thereby
dependent manner (36), suggesting that reaching the intestine (43). Both hormones
EHEC is capable of producing AI-2. Super- modulate intestinal smooth muscle contrac-
natants of EHEC grown to late exponential tion, submucosal blood flow, and chloride
phase contain a signal capable of inducing and potassium secretion in the intestine
transcription of many EHEC virulence genes, (44). Epinephrine and NE are recognized by
including the locus of enterocyte effacement adrenergic receptors in mammalian cells;
(LEE) that encodes a type III secretion system Freddolino et al. (45) reported that the ligand-
(T3SS) essential for EHEC to attach to and binding sites for epinephrine and NE in a
efface enterocytes and promote disease (37, human adrenergic receptor are similar. Ex-
38), the flagellum regulon, and expression tensive evidence indicates that both epineph-
of Shiga toxin (35, 39). Purified and synthetic rine and NE are recognized by the same
AI-2 was unable to govern LEE and flagella receptors and play important biological roles
expression (40), suggesting the existence of an in the human GI tract.
additional class of autoinducer (AI-3). The The AI-3/epinephrine/NE interkingdom
identity of AI-3 is not yet known, but it is signaling cascade activates expression of vir-
methanol soluble (35), likely tyrosine derived ulence genes in EHEC (39, 40, 46, 47). The
(36), and shares signaling mechanisms with host hormones epinephrine/NE are specifi-
host-derived epinephrine and norepinephrine cally sensed by two HKs: QseC and QseE,
(Fig. 1). AI-3 signal is produced by a variety of which are the first bacterial adrenergic re-
enterobacterial species, including pathogenic ceptors identified (46, 48). QseE is down-
strains and normal flora such as Enterobacter stream of QseC in this signaling cascade, given
cloacae (36), and is also present in the stool of that transcription of qseE is activated through
humans (35). The reliance of a functional luxS QseC (49). In addition to sensing these host
gene in the presence of AI-3 in EHEC was not hormones, QseC also senses the bacterial sig-
due to LuxS being involved in the synthesis of nal AI-3 (46). The QseC sensor is the first ex-
AI-3, but to a metabolic shift that occurs in a ample of a receptor for both a bacterial and a
luxS mutant (LuxS is involved in the central host signal, integrating bacterial-host signaling
methyl cycle in bacterial cells) that leads to at the biochemical level. QseE, however, does
decreased AI-3 production (36). not sense AI-3, thereby discriminating be-
tween host- and bacteria-derived signals (49).
Upon sensing their respective signals, QseC
Interkingdom Chemical Signaling in
and QseE activate virulence gene expression
EHEC-Host Interactions
and pathogenesis in vitro and in vivo in EHEC
The AI-3 QS signaling system intersects with (46, 48, 50). QseC and QseB constitute a cog-
the host adrenergic signaling system (epi- nate two-component system, and the qseBC
nephrine/norepinephrine [NE] hormones) genes are cotranscribed in an operon that is
(40). In fact, this intersection occurs at a bio- also autoregulated (51). QseC transfers its
chemical level, with the same bacterial re- phosphate to three response regulators (RRs):
ceptor, the membrane-bound histidine sensor its cognate RR QseB and the noncognate QseF

and KdpE RRs (52). QseE only transfers its dependent on QseF (52). KdpE functions as an
phosphate to its cognate RR QseF (53). The activator of LEE1 (LEE1 encodes the Ler acti-
concerted action of these RRs activates the vator of all LEE genes [55]) expression (52) by
EHEC virulence repertoire (Fig. 2). direct binding upstream of the distal LEE1
QseC facilitates both phosphorylation and promoter (56) (Fig. 2). An important note is
dephosphorylation of QseB. Unphosphorylated that the QseC signaling cascade can be inhib-
QseB binds upstream of the flhDC promoter ited by the antivirulence drug LED209, pre-
(FlhDC operons are the master regulators of venting activation of the EHEC virulence
the flagella regulon) and inhibits transcription repertoire (50).
of these master activators of motility, whereas
phosphorylated QseB induces expression (52,
Nutrient Signaling in EHEC-Microbiota
54). QseF indirectly stimulates the expression
Interactions and Virulence Regulation
of espFU/tccP (49), a critical non-LEE-encoded
effector required for actin recruitment at the The mammalian GI tract harbors trillions of
site of pedestal formation. Activation of QseC indigenous bacteria whose coexistence relies
induces Shiga toxin expression (50) and this is on the ability of each member to use one or a
FIGURE 2 The QseC signaling cascade. QseC senses AI-3 epinephrine and norepinephrine and phosphorylates
QseB, QseF, and KdpE. QseE senses epinephrine and phosphorylates QseF. QseB activates agella expression.
QseF indirectly promotes Shiga toxin and EspFu expression. KdpE together with Cra activate LEE gene ex-
pression. Both QseBC and QseEF repress fusKR expression. FusK senses fucose and phosphorylates FusR that
represses the LEE. doi:10.1128/microbiolspec.EHEC-0015-2013.f2

few limiting resources (1). Invading pathogens Cra was found to directly interact with the RR
have to compete with the microbiota for these KdpE, previously found to positively regulate
resources to establish colonization. These LEE1 (52). Interestingly, KdpE-dependent
pathogens tend to be aggressive and greedy in LEE regulation was also found to be in effect
search of a colonization niche, and achieve only in low-glucose conditions, and KdpE
this purpose by precisely coordinating ex- binding in vitro to the LEE1 promoter was
pression of an arsenal of virulence genes. diminished by phosphorylation (56). Presum-
Linking carbon and nitrogen metabolism to ably, under high-glucose conditions, KdpE is
the precise coordination of virulence expres- phosphorylated by its cognate sensor kinase,
sion is a key step in the adaptation of patho- KdpD, which is activated by the glucose-
gens toward recognizing suitable sites for sensitive IIANtr phosphotransfer system (57).
colonization, and a means to tip the scale in These data suggest that Cra and KdpE act in
the tug-of-war between pathogens and the concert to induce T3S by inducing ler ex-
microbiota. pression and that this control is only active
EHEC is no stranger to this struggle for under glucose-limiting conditions, such as
nutrients. In addition to purpose-built signal- those found within the colon.
ing molecules like the autoinducers, the in- While a gradient of diet-derived glucose
terplay between the nutrient requirements may provide an indicator of progress through
of normal flora and EHEC is important in the length of the GI tract, spatial regulation of
determining virulence. Glucose and other virulence factor expression likely requires
monomeric dietary sugars are ideal carbon EHEC to distinguish luminal from host mem-
sources for EHEC, but these molecules are brane proximal environments (Fig. 3). A thick
scarce in the lower GI tract. The proximal layer of goblet cell-derived mucus partitions
small intestine is the site of host absorption of the bacteria-laden GI lumen from the host
simple sugars, whereas the distal small intes- enterocytes. The mucous layer is composed of
tine houses a vast population of commensal a dense matrix of cross-linked mucin proteins
bacteria scavenging for free sugars. As a result, heavily decorated with O-linked glycans.
the colonic home of EHEC is a gluconeogenic Within the colon, Muc2 represents the dom-
environment. Njoroge and colleagues uncov- inant species of mucin linked to glycans
ered the importance of glucose availability composed of GalNAc, NANA, GlcNAc, galac-
in regulating T3SS by EHEC (56). Stem- tose, fucose (Fig. 1), and mannose mono-
ming from the observation that high-glucose saccharides, listed in terms of decreasing
growth media suppressed type III secretion relative abundance (58). Members of the
(T3S) while low-glucose conditions induced normal flora of the colon express mucolytic
LEE expression, the authors uncovered a role enzymes to harvest carbon from this barrier
for the catabolite repressor/activator protein and from dietary polysaccharides. Bacteroides
(Cra) in ler regulation. Indeed, a cra-deficient thetaiotaomicron, a well-studied commensal,
mutant of EHEC exhibited diminished attach- is capable of metabolizing pectins, starches,
ing and effacing (A/E) lesion formation, LEE1 fructan, alpha-glucans, and a number of gly-
promoter activity, transcript levels, and EspA cans originating from host tissues (59). In
secretion under low-glucose conditions. How- contrast, E. coli is an organism adapted for
ever, under glucose-rich conditions no effect exploitation of monosaccharides and tricar-
was seen. A Cra-binding site was identified boxylic acid cycle intermediates liberated
upstream of the distal LEE1 promoter, and from complex carbohydrates by these com-
binding to this site in an electrophoretic mo- mensals (61, 62). It is not surprising that
bility shift assay could be prevented by the mucin-derived carbon-source sensing has
inclusion of the glycolytic intermediates fruc- been adapted by EHEC to regulate virulence
tose-1-phosphate or fructose-1,6-bisphosphate. mechanisms.

FIGURE 3 EHEC gastrointestinal colonization. doi:10.1128/microbiolspec.EHEC-0015-2013.f3
Pacheco et al. describe a novel two- compared with glucose; however, no differ-
component system that enables regulation of ence is seen in fusK-deficient EHEC. Simi-
virulence gene expression and carbon-source larly, mucin was found to diminish LEE1
choice by EHEC upon sensing fucose (62). expression when EHEC was grown in the
The fusKR operon is within the genomic presence of B. thetaiotaomicron, which ex-
O-island 20, found only in O55:H7 descendant presses fucosidases capable of liberating fu-
strains and within C. rodentium. FusK is a cose from mucin. The effect was not seen with
transmembrane HK that autophosphoryl- cocultures grown in the presence of fucose
ates at a histidine residue and transfers the rather than mucin. Altogether, these data
phospho-group to an aspartate of FusR to suggest that fucose serves as a signal to down-
regulate DNA binding. The authors observe regulate T3S via FusKR when EHEC is in the
that the FusKR two-component system func- lumen.
tions as a repressor of T3S. Knockout of either As the mucous barrier limits the approach
fusK or fusR results in increased A/E lesion of mucolytic commensals toward the epithe-
formation in a cell culture infection model, lial surface, fucose liberation occurs in the
increased LEE transcript levels, and higher colonic lumen, but not in close contact with
levels of EspB in culture supernatants com- the intestinal epithelium. LEE expression
pared to wild type. Given that FusK shares within the lumen is not appropriate for
sequence homology with UhpB, a glucose-6- EHEC; therefore, a luminal signal to decrease
phosphate sensor kinase, the authors explored T3S would serve to ensure efficient resource
whether FusK was responsive to sugar mono- expenditure. While in the lumen, fucose-FusK
mers and found that fucose, but not other signaling satisfies this need. However, upon
sugars, was sufficient to induce autophos- reaching the epithelium, FusKR would be
phorylation of the kinase. Expression of LEE1 detrimental to EHEC as it would inhibit es-
is diminished in wild-type EHEC grown tablishment of pedestals. EHEC has used the
with fucose as the sole carbon source when interkingdom signaling systems of QseBC and

QseEF to alleviate this issue while at the epi- of Gammaproteobacteria upon C. rodentium
thelial surface. QseB directly binds to and in- infection has also been previously reported by
hibits transcription of the fusKR operon, while Lupp et al. (18), further suggesting that com-
QseF likely induces expression of a fusKR petition for carbon sources with the microbi-
repressor (62). Additionally, EHEC also uses ota plays an important role in EHEC clearance
ethanolamine, a vast source of nitrogen within from the GI tract.
the intestine that is present within mem- Fermentation of starches within the colon
branes, to regulate LEE and Shiga toxin ex- is primarily mediated by Firmicutes, such as
pression (63). Faecalibacterium prausnitzii and Eubacterium
Intriguingly, FusKR was also found to rectale, and also Bacteroides spp. to produce
downregulate a putative fucose transporter SCFAs. The importance of the microbiota to
protein encoded near fusKR. Deletion of this SCFA production is clear from the very low
major facilitator superfamily protein locus, level of these metabolites in germfree animals
z0461, dramatically diminishes the expression (67). Acetate, propionate, and butyrate (Fig. 1)
of genes involved in fucose utilization and dominate the SCFA population and contribute
delays growth in media containing fucose broadly to host physiology (68). SCFAs are
as the sole carbon source. On the surface, present in the colon in millimolar concentra-
this observation represents a paradox in that tions and exist in much lower concentrations
EHEC senses fucose, a prime carbon source, in the upper GI tract. Thus, they represent an
and yet diminishes its own capacity to use excellent indicator of arrival within the large
fucose. This can perhaps be explained by ob- intestine in much the same way that abundant
servations concerning carbon-source prefer- glucose is indicative of localization within
ence in physiological settings of bovine small the small intestine. As such, EHEC uses these
intestine contents (64) and a murine infection molecules as a cue to govern expression of
model (65). Though fucose utilization con- virulence genes.
tributes to EHEC growth within bovine small Butyrate, but not acetate or propionate, is
intestinal contents, deficiencies within this capable of inducing T3S from the Sakai strain
pathway are not as dramatic as those seen in of EHEC (69); 20 mM butyrate increases ad-
galactose or mannose utilization pathways hesion to Caco-2 cells 10-fold over control and
(64). Analysis of mouse colonization of mu- facilitates the formation of microcolonies (70).
tants of EDL933 or MG1655 demonstrates Leucine is able to induce LEE4 and LEE5
that fucose is used by both virulent and avir- protein expression similarly to butyrate (69).
ulent strains of E. coli, whereas galactose and Both leucine- and butyrate-driven expression
mannose are primarily the forte of EHEC and of the LEE is dependent on expression of the
dispensable for MG1655 (65). Therefore, leucine-responsive protein (Lrp) transcription
downregulation of fucose utilization systems factor. Given the similarities in the structure
may be a mechanism of EHEC to avoid car- of leucine and butyrate and the observation
bon-source overlap with commensal E. coli that the M124R leucine-insensitive mutation
and thus avoid direct competition. This pos- to Lrp eliminates butyrate-induced LEE ex-
sibility is further corroborated by the work of pression, it is likely that butyrate is directly
Kamada et al. (66); these authors observed sensed by Lrp. In addition to requiring Lrp,
that expression of virulence genes by the butyrate-induced T3S requires Ler and the ler
EHEC surrogate murine model C. rodentium activator PchA. In fact, the promoter of pchA
is activated at the onset of infection, and cannot be substituted without abolishing the
at later time points, C. rodentium infection butyrate effect, implying a cascade of events:
backfires, triggering a bloom of Gamma- butyrate binds Lrp, which then activates
proteobacteria that effectively compete with transcription of pchA, which in turn activates
C. rodentium for carbon sources. This bloom transcription of the LEE1-encoded ler.

Acetate, propionate, and butyrate increase The interplay of microbiota and host diet
motility of the Sakai strain (70). Synthesis of likely creates a multidimensional gradient of
the flagellin subunit, FliC, is increased, and metabolites within the GI tract of mammals.
the frequency of flagellated bacteria increases EHEC adapted to colonization of the lower
upon exposure to these SCFAs. Two mecha- intestinal tract uses these gradients to regulate
nisms of induction are in play: activation of expression of complex, metabolically expen-
the class 1 flhDC promoter by butyrate and in- sive virulence systems such as the LEE-
duction of class 2 flgN by acetate, propionate, encoded T3SS to maximize fitness within an
and butyrate. Butyrate-induced flhDC expres- animal host. Under such a model, localization
sion requires lrp, whereas lrp is dispensable within the colon is indicated by increased
for FlgN accumulation. concentrations of fermentation products, such
SCFAs heavily influence the colonic epithe- as butyrate, and increasingly gluconeogenic
lium by serving as an energy source for host use, conditions, which promote expression of the
promoting nutrient absorption, and regulating LEE and, thus, attachment as is appropriate. A
cell differentiation. Zumbrun and colleagues luminal-epithelial axis is created within the
recently demonstrated that butyrate increases colon by the presence of luminal sugars lib-
globotriaosylceramide expression on human erated from carbohydrates, such as fucose
colonic epithelial cells (71). Mice fed a high- derived from mucin, that suppress LEE ex-
fiber diet, which increases the concentration of pression until the epithelium is within reach.
butyrate within the intestinal lumen, express
higher globotriaosylceramide levels within the
intestinal epithelium and in kidneys. When MICROBIOTA COMPOSITION AND
challenged with EHEC, mice fed a high-fiber SUSCEPTIBILITY TO EHEC INFECTIONS
diet experience a higher pathogenic burden,
significant weight loss, and decreased survival Within the scope of the complex associations
compared to animals fed a low-fiber diet. In- between EHEC and different members of the
terestingly, a high-fiber diet diminishes the GI microbiota, a burning question is to how
frequency of Escherichia spp. within the gut much microbiota differential composition
while increasing total flora levels. These results contributes to host resistance to EHEC in-
suggest that a high-fiber diet, via increased bu- fections? EHEC infections can vary in their
tyrate levels, may increase susceptibility of degree of severity, ranging from watery diar-
animals to Shiga toxin released from EHEC by rhea, to severe bloody diarrhea, to hemolytic-
promoting the expression of the receptor nec- uremic syndrome. It has been reported that
essary for toxicity. Additionally, a high-fiber diet antibiotic treatment that alters the micro-
may reduce competition faced by EHEC during bial composition of the GI microbiota may
infection by reducing levels of normal E. coli increase susceptibility to GI infection by
flora while simultaneously increasing the prev- C. rodentium (73). It has also been reported
alence of other species that may benefit EHEC. that microbiota transplantation from suscep-
Another important aspect of the role of tible mice to mice resistant to C. rodentium
SCFAs and the outcome of EHEC infections infection also increased susceptibility of re-
stems from the observation that probiotic sistant mice to this pathogen. In the opposite
strains of Bifidobacterium longum that pro- experiment, transplantation of microbiota
duce high levels of acetate effectively prevent from resistant mice to susceptible mice was
Shiga toxin translocation through the intesti- protective (74). Additionally, the combina-
nal epithelium. The authors propose that action of certain microbiota members with dif-
etate produced by these probiotic bacteria ferent metabolites can result in different
improves intestinal defense and protects the outcomes on EHEC virulence gene expres-
host against lethal infection (72). sion. Fucose released from the mucus by

B. thetaiotaomicron decreases LEE gene ex- formation are also necessary for EHEC colo-
pression (62), while under gluconeogenic nization of the recto-anal junction (RAJ) of
conditions in the absence of fucose, B. theta- cattle, facilitating shedding of this pathogen
iotaomicron promotes LEE gene expression in the environment (83). In addition to the
(56). It has also been reported that expression LEE, EHEC uses the glutamate decarboxylase
of Shiga toxin is decreased in the presence (gad) acid resistance system to survive passage
of B. thetaiotaomicron in Leedle and Hesplee through the acidic stomachs of these animals
medium (75). to reach its site of colonization, the RAJ (84).
SdiA is a regulator that senses AHL QS sig-
naling molecules and aids EHEC survival and
DIFFERENTIAL SIGNALING SYSTEMS colonization of the bovine GI tract. AHLs are
GOVERNING HOST ASSOCIATIONS prominent within cattle rumen but absent in
the other sections of the GI tract. Through
The main environmental reservoir of EHEC SdiA, transcription of the LEE genes is de-
is ruminants, and it is estimated that 70 to creased by rumen AHLs, while transcription
80% of cattle herds in the United States are of the gad acid-resistant system is increased.
colonized with EHEC (7681). EHEC is an Expression of the LEE in the rumen would be
example of a bacterium that behaves as a an unnecessary energy burden for EHEC in
commensal or a pathogen, depending on its this GI compartment. However, in prepara-
host. EHEC is a commensal in the GI tract of tion for the acidic distal stomachs, the EHEC
adult cattle but is a human pathogen (82). gad is activated in the rumen. SdiA-AHL sig-
EHEC colonizes the large intestine of humans, naling aids EHEC in gauging these environ-
forming A/E lesion, thought to be largely re- ments and modulates gene expression toward
sponsible for promoting disease (82). The adaptation to a commensal lifestyle in cattle.
genes for A/E lesion formation are encoded Consequently, an sdiA mutant is deficient for
within the LEE (82). The LEE and A/E lesion cattle colonization (85, 86) (Fig. 4).
FIGURE 4 EHEC cattle colonization. Within the rumen EHEC senses AHLs through SdiA to decrease LEE ex-
pression and increase gad expression. Within the RAJ, in the absence of AHLs, LEE expression is promoted.

AHLs are synthesized from S-adenosyl- (89). This, combined with the lack of endog-
methionine and an acylated acyl carrier pro- enous AHLs from E. coli, implies that SdiA
tein by AHL synthases, such as LuxI of likely functions as a sensor for an array of
V. fischeri or RhlI of Pseudomonas aeruginosa foreign AHLs found within the surroundings
(87). AHLs are sensed by a class of unstable of E. coli. In support of this notion, AHLs have
transcription factors exemplified by LuxR been found within the rumen contents of
of V. fischeri. LuxR homologs consist of an cattle fed both grain and forage diets (86, 90,
N-terminal autoinducer binding domain and a 91). However, chemical analysis of lower GI
C-terminal helix-turn-helix motif. AHL bind- tract contents from cattle and nonruminant
ing promotes LuxR folding and dimeriza- animals has not revealed the presence of
tion to allow for binding to a target sequence AHLs, which may be due to inadequate sen-
within a promoter to regulate gene expression. sitivity of the methods used, alkaline instability
Acyl chain length and structure provide speci- of AHLs, or potentially the existence of exotic
ficity for AHL sensing by different species. E. coli homoserine lactones or non-AHL small mol-
does not express any known AHL synthase but ecules that may influence SdiA function (40).
does encode for a luxR homolog, sdiA. These reports highlight how different
Initial studies aimed at understanding the chemical signaling systems can be employed
role of sdiA in virulence of EHEC observed by bacteria to adapt to either pathogenic or
that overexpression of SdiA diminished pro- commensal lifestyles in different hosts and
duction of LEE4 and LEE5 encoded proteins, that this signaling system aids this human
as well as decreasing flagellin expression and pathogen to adapt to a commensal lifestyle in
soft agar motility (88). Like other LuxR homo- cattle, its main reservoir.
logs, SdiA tends to fold poorly in the absence
of exogenous AHLs and accumulates in in-
clusion bodies. Overexpression of SdiA likely CONCLUDING REMARKS
results in a small, transient population of sol-
uble molecules that is capable of affecting The increasing knowledge of the essential
transcription. Indeed, transcriptional analysis roles of the microbiota in human health is
of sdiA-deficient EHEC reveals that the glu- opening many avenues of research to under-
tamate-dependent acid resistance (gad) genes stand differential host susceptibility to infec-
are activated by SdiA independently of exog- tious diseases. These studies are fundamental
enous AHL activity, although AHLs enhance for enteric pathogens, which inhabit a com-
this activation (85). In contrast, SdiA regula- plex and dynamic environment. It is fascinat-
tion of the LEE T3SS is AHL dependent. SdiA- ing to envision how such few EHEC organisms
deficient EHEC has no change in LEE ex- (circa 50 CFU) efficiently manage to establish
pression relative to wild type in the absence of themselves in the host and cause disease. It is
AHL. Exogenous oxo-C6-homoserine lactone becoming clear that EHEC is very crafty at
(Fig. 1) diminishes LEE1 transcription and reading many cues provided by both the
EspA secretion in an SdiA-dependent manner. host and the microbiota and rapidly adapting
SdiA stabilized by exogenous AHLs directly its virulence program toward successful host
binds to the LEE1 promoter to repress ler infection. We are also at the tip of an iceberg in
transcription and diminishes T3S through this determining how different microbiota entero-
action (85). types may determine the severity of EHEC
Nuclear magnetic resonance structural disease. One should also take into consider-
analysis of SdiA of E. coli has demonstrated ation that differences in diets and antibiotic
that the protein is capable of fruitfully inter- regimens, which cause shifts in the composi-
acting with at least three additional AHLs: tion of the GI microbial flora, may also influ-
C8-HSL, oxo-C8-HSL (Fig. 1), and C6-HSL ence the outcome of EHEC disease.

ACKNOWLEDGMENT interactions in a humanized microbiome mouse

model. Mol Syst Biol 4:157.
We declare no conflicts of interest with regard 10. Woodmansey EJ, McMurdo ME, Macfarlane
to the manuscript. GT, Macfarlane S. 2004. Comparison of compo-
sitions and metabolic activities of fecal micro-
biotas in young adults and in antibiotic-treated
CITATION and non-antibiotic-treated elderly subjects. Appl
Environ Microbiol 70:61136122.
Pifer R, Sperandio V. 2014. The interplay be- 11. Hoverstad T, Carlstedt-Duke B, Lingaas E,
tween the microbiota and enterohemorrhagic Midtvedt T, Norin KE, Saxerholt H, Steinbakk
M. 1986. Influence of ampicillin, clindamycin,
Escherichia coli. Microbiol Spectrum 2(5):
and metronidazole on faecal excretion of short-
EHEC-0015-2013. chain fatty acids in healthy subjects. Scand J
Gastroenterol 21:621626.
12. Millard AL, Mertes PM, Ittelet D, Villard F,
REFERENCES Jeannesson P, Bernard J. 2002. Butyrate affects
1. Gill SR, Pop M, Deboy RT, Eckburg PB, differentiation, maturation and function of human
Turnbaugh PJ, Samuel BS, Gordon JI, Relman monocyte-derived dendritic cells and macro-
DA, Fraser-Liggett CM, Nelson KE. 2006. phages. Clin Exp Immunol 130:245255.
Metagenomic analysis of the human distal gut 13. Hossain Z, Konishi M, Hosokawa M,
microbiome. Science 312:13551359. Takahashi K. 2006. Effect of polyunsaturated
2. Hooper LV, Gordon JI. 2001. Commensal host- fatty acid-enriched phosphatidylcholine and phos-
bacterial relationships in the gut. Science 292: phatidylserine on butyrate-induced growth inhi-
11151118. bition, differentiation and apoptosis in Caco-2
3. Grenham S, Clarke G, Cryan JF, Dinan TG. cells. Cell Biochem Funct 24:159165.
2011. Brain-gut-microbe communication in health 14. Hamer HM, Jonkers D, Venema K, Vanhoutvin
and disease. Front Physiol 2:94. S, Troost FJ, Brummer RJ. 2008. Review article:
4. Gordon JI, Klaenhammer TR. 2011. A rendez- the role of butyrate on colonic function. Aliment
vous with our microbes. Proc Natl Acad Sci USA Pharmacol Ther 27:104119.
108 Suppl 1:45134515. 15. Usami M, Kishimoto K, Ohata A, Miyoshi M,
5. Gonzalez A, Stombaugh J, Lozupone C, Aoyama M, Fueda Y, Kotani J. 2008. Butyrate
Turnbaugh PJ, Gordon JI, Knight R. 2011. The and trichostatin A attenuate nuclear factor kappaB
mind-body-microbial continuum. Dialogues Clin activation and tumor necrosis factor alpha secre-
Neurosci 13:5562. tion and increase prostaglandin E2 secretion in
6. Dethlefsen L, Relman DA. 2011. Incomplete re- human peripheral blood mononuclear cells. Nutr
covery and individualized responses of the human Res 28:321328.
distal gut microbiota to repeated antibiotic per- 16. Jansson J, Willing B, Lucio M, Fekete A,
turbation. Proc Natl Acad Sci USA 108 Suppl Dicksved J, Halfvarson J, Tysk C, Schmitt-
1:45544561. Kopplin P. 2009. Metabolomics reveals metabolic
7. Romick-Rosendale LE, Goodpaster AM, biomarkers of Crohns disease. PLoS One 4:e6386.
Hanwright PJ, Patel NB, Wheeler ET, Chona DL, 17. Fischbach MA, Sonnenburg JL. 2011. Eating
Kennedy MA. 2009. NMR-based metabonomics for two: how metabolism establishes interspecies
analysis of mouse urine and fecal extracts following interactions in the gut. Cell Host Microbe 10:336
oral treatment with the broad-spectrum antibiotic 347.
enrofloxacin (Baytril). Magn Reson Chem 47 Suppl 18. Lupp C, Robertson ML, Wickham ME,
1:S36S46. Sekirov I, Champion OL, Gaynor EC, Finlay BB.
8. Yap IK, Li JV, Saric J, Martin FP, Davies H, 2007. Host-mediated inflammation disrupts the
Wang Y, Wilson ID, Nicholson JK, Utzinger J, intestinal microbiota and promotes the over-
Marchesi JR, Holmes E. 2008. Metabonomic and growth of Enterobacteriaceae. Cell Host Microbe
microbiological analysis of the dynamic effect of 2:204.
vancomycin-induced gut microbiota modification 19. Stecher B, Robbiani R, Walker AW, Westendorf
in the mouse. J Proteome Res 7:37183728. AM, Barthel M, Kremer M, Chaffron S,
9. Martin FP, Wang Y, Sprenger N, Yap IK, Macpherson AJ, Buer J, Parkhill J, Dougan G,
Lundstedt T, Lek P, Rezzi S, Ramadan Z, van von Mering C, Hardt WD. 2007. Salmonella
Bladeren P, Fay LB, Kochhar S, Lindon JC, enterica serovar Typhimurium exploits inflam-
Holmes E, Nicholson JK. 2008. Probiotic mod- mation to compete with the intestinal microbiota.
ulation of symbiotic gut microbial-host metabolic PLoS Biol 5:21772189.

20. Dunlop SP, Jenkins D, Neal KR, Spiller RC. 33. Surett MG, Bassler BL. 1998. Quorum sensing
2003. Relative importance of enterochromaffin in Escherichia coli and Salmonella typhimurium.
cell hyperplasia, anxiety, and depression in Proc Natl Acad Sci USA 95:70467050.
postinfectious IBS. Gastroenterology 125:1651 34. Schauder S, Shokat K, Surette MG, Bassler BL.
1659. 2001. The LuxS family of bacterial autoinducers:
21. Sperandio V, Mellies JL, Nguyen W, Shin S, biosynthesis of a novel quorum-sensing signal
Kaper JB. 1999. Quorum sensing controls molecule. MolMicrobiol 41:463476.
expression of the type III secretion gene tran- 35. Sperandio V, Torres AG, Jarvis B, Nataro JP,
scription and protein secretion in entero- Kaper JB. 2003. Bacteria-host communication:
hemorrhagic and enteropathogenic Escherichia the language of hormones. Proc Natl Acad Sci
coli. Proc Natl Acad Sci USA 96:1519615201. USA 100:89518956.
22. Fuqua WC, Winans SC, Greenberg EP. 1994. 36. Walters M, Sircili MP, Sperandio V. 2006. AI-3
Quorum sensing in bacteria: the LuxR-LuxI synthesis is not dependent on luxS in Escherichia
family of cell density-responsive transcriptional coli. JBacteriol 188:56685681.
regulators. J Bacteriol 176:269275. 37. McDaniel TK, Jarvis KG, Donnenberg MS,
23. Nealson KH, Platt T, Hastings JW. 1970. Cel- Kaper JB. 1995. A genetic locus of enterocyte
lular control of the synthesis and activity of the effacement conserved among diverse enterobac-
bacterial luminescent system. J Bacteriol 104:313 terial pathogens. Proc Natl AcadSci USA 92:1664
322. 1668.
24. Nealson KH, Hastings JW. 1979. Bacterial bio- 38. Jarvis KG, Giron JA, Jerse AE, McDaniel TK,
luminescence: its control and ecological signifi- Donnenberg MS, Kaper JB. 1995. Enteropatho-
cance. Microbiol Rev 43:496518. genic Escherichia coli contains a putative type
25. Engebrecht J, Nealson K, Silverman M. 1983. III secretion system necessary for the export
Bacterial bioluminescence: isolation and genetic of proteins involved in attaching and effacing le-
analysis of functions from Vibrio fischeri. Cell sion formation. Proc Natl Acad Sci USA 92:7996
32:773781. 8000.
26. Engebrecht J, Silverman M. 1984. Identification 39. Sperandio V, Torres AG, Giron JA, Kaper JB.
of genes and gene products necessary for bacterial 2001. Quorum sensing is a global regulatory mech-
bioluminescence. Proc Natl Acad Sci USA 81: anism in enterohemorrhagic Escherichia coli
41544158. O157:H7. J Bacteriol 183:51875197.
27. Parsek MR, Greenberg EP. 2000. Acyl- 40. Swearingen MC, Sabag-Daigle A, Ahmer BM.
homoserine lactone quorum sensing in gram- 2013. Are these acyl-homoserine lactones within
negative bacteria: a signaling mechanism involved mammalian intestines? J Bacteriol 195:173179.
in associations with higher organisms. Proc Natl 41. Clarke MB, Hughes DT, Zhu C, Boedeker EC,
Acad Sci USA 97:87898793. Sperandio V. 2006. The QseC sensor kinase: a
28. Davies DG, Parsek MR, Pearson JP, Iglewski bacterial adrenergic receptor. Proc Natl Acad Sci
BH, Costerton JW, Greenberg EP. 1998. The USA 103:1042010425.
involvement of cell-to-cell signals in the devel- 42. Furness JB. 2000. Types of neurons in the en-
opment of a bacterial biofilm. Science 280:295 teric nervous system. J Auton Nerv Syst 81:8796.
298. 43. Purves D, Fitzpatrick D, Williams SM, McNamara
29. de Kievit TR, Iglewski BH. 2000. Bacterial JO, Augustine GJ, Katz LC, LaMantia A. 2001.
quorum sensing in pathogenic relationships. In- Neuroscience, 2nd ed. Sinauer Associates, Inc.,
fect Immun 68:48394849. Sunderland, MA.
30. Wang XD, de Boer PA, Rothfield LI. 1991. A 44. Horger S, Schultheiss G, Diener M. 1998.
factor that positively regulates cell division by Segment-specific effects of epinephrine on ion
activating transcription of the major cluster of transport in the colon of the rat. Am J Physiol 275:
essential cell division genes of Escherichia coli. G1367G1376.
Embo J 10:33633372. 45. Fredollino PL, Kalani MY, Vaidihi N, Floriano
31. Swift S, Lynch MJ, Fish L, Kirke DF, Tomas WB, Hall SE, Trabanino RJ, Kam VW, Goddard
JM, Stewart GS, Williams P. 1999. Quorum WA. 2004. Predicted 3D structure for the human
sensing-dependent regulation and blockade of beta 2 adrenergic receptor and its binding site for
exoprotease production in Aeromonas hydrophila. agonists and antagonists. Proc Natl Acad Sci USA
Infect Immun 67:51925199. 101:27362741.
32. Michael B, Smith JN, Swift S, Heffron F, Ahmer 46. Clarke MB, Hughes DT, Zhu C, Boedeker EC,
BM. 2001. SdiA of Salmonella enterica is a LuxR Sperandio V. 2006. The QseC sensor kinase: a
homolog that detects mixed microbial communi- bacterial adrenergic receptor. Proc Natl Acad Sci
ties. J Bacteriol 183:57335742. USA 103:1042010425.

47. Walters M, Sircili MP, Sperandio V. 2006. AI-3 Studies on the insoluble glycoprotein complex
synthesis is not dependent on luxS in Escherichia from human colon. Identification of reduction-
coli. J Bacteriol 188:56685681. insensitive MUC2 oligomers and C-terminal
48. Reading NC, Rasko DA, Torres AG, Sperandio cleavage. J BiolChem 274:1582815836.
V. 2009. The two-component system QseEF and 59. Martens EC, Lowe EC, Chiang H, Pudlo NA,
the membrane protein QseG link adrenergic and Wu M, McNulty NP, Abbott DW, Henrissat B,
stress sensing to bacterial pathogenesis. Proc Natl Gilbert HJ, Bolam DN, Gordon JI. 2011. Rec-
Acad Sci USA 106:58895894. ognition and degradation of plant cell wall poly-
49. Reading NC, Torres AG, Kendall MM, Hughes saccharides by two human gut symbionts. PLoS
DT, Yamamoto K, Sperandio V. 2007. A novel Biol 9:e1001221.
two-component signaling system that activates 60. Chang DE, Smalley DJ, Tucker DL, Leatham
transcription of an enterohemorrhagic Esche- MP, Norris WE, Stevenson SJ, Anderson AB,
richia coli effector involved in remodeling of host Grissom JE, Laux DC, Cohen PS, Conway T.
actin. J Bacteriol 189:24682476. 2004. Carbon nutrition of Escherichia coli in the
50. Rasko DA, Moreira CG, Li de R, Reading NC, mouse intestine. Proc Natl Acad Sci USA 101:
Ritchie JM, Waldor MK, Williams N, Taussig R, 74277432.
Wei S, Roth M, Hughes DT, Huntley JF, Fina 61. Miranda RL, Conway T, Leatham MP,
MW, Falck JR, Sperandio V. 2008. Targeting Chang DE, Norris WE, Allen JH, Stevenson
QseC signaling and virulence for antibiotic de- SJ, Laux DC, Cohen PS. 2004. Glycolytic
velopment. Science 321:10781080. and gluconeogenic growth of Escherichia coli
51. Clarke MB, Sperandio V. 2005. Transcriptional O157:H7 (EDL933) and E. coli K-12 (MG1655)
autoregulation by quorum sensing E. coli regu- in the mouse intestine. Infect Immun 72:1666
lators B and C (QseBC) in enterohemorrhagic 1676.
E. coli (EHEC). Mol Microbiol 58:441455. 62. Pacheco AR, Curtis MM, Ritchie JM, Munera
52. Hughes DT, Clarke MB, Yamamoto K, Rasko D, Waldor MK, Moreira CG, Sperandio V. 2012.
DA, Sperandio V. 2009. The QseC adrenergic Fucose sensing regulates bacterial intestinal col-
signaling cascade in enterohemorrhagic E. coli onization. Nature 492:113117.
(EHEC). PLoS Pathog 5:e1000553. 63. Kendall MM, Gruber CC, Parker CT, Sperandio
53. Yamamoto K, Hirao K, Oshima T, Aiba H, V. 2012. Ethanolamine controls expression of
Utsumi R, Ishihama A. 2005. Functional char- genes encoding components involved in inter-
acterization in vitro of all two-component signal kingdom signaling and virulence in enterohem-
transduction systems from Escherichia coli. J Biol orrhagic Escherichia coli O157:H7. MBio 3:
Chem 280:14481456. e00050-12.
54. Clarke MB, Sperandio V. 2005. Transcriptional 64. Bertin L, Grilli S, Spagni A, Fava F. 2013. Inno-
regulation of flhDC by QseBC and sigma (FliA) in vative two-stage anaerobic process for effective
enterohaemorrhagic Escherichia coli. Mol Micro- codigestion of cheese whey and cattle manure.
biol 57:17341749. Bioresour Technol 128:779783.
55. Mellies JL, Elliott SJ, Sperandio V, Donnenberg 65. Fabich AJ, Jones SA, Chowdhury FZ, Cernosek
MS, Kaper JB. 1999. The Per regulon of entero- A, Anderson A, Smalley D, McHargue JW,
pathogenic Escherichia coli: identification of a Hightower GA, Smith JT, Autieri SM, Leatham
regulatory cascade and a novel transcriptional MP, Lins JJ, Allen RL, Laux DC, Cohen PS,
activator, the locus of enterocyte effacement Conway T. 2008. Comparison of carbon nutrition
(LEE)-encoded regulator (Ler). Mol Microbiol for pathogenic and commensal Escherichia coli
33:296306. strains in the mouse intestine. Infect Immun
56. Njoroge JW, Nguyen Y, Curtis MM, Moreira 76:11431152.
CG, Sperandio V. 2012. Virulence meets metab- 66. Kamada N, Kim YG, Sham HP, Vallance BA,
olism: Cra and KdpE gene regulation in entero- Puente JL, Martens EC, Nunez G. 2012. Regu-
hemorrhagic Escherichia coli. MBio 3:e00280 lated virulence controls the ability of a pathogen
00212. to compete with the gut microbiota. Science
57. Luttmann D, Heermann R, Zimmer B, Hillmann 336:13251329.
A, Rampp IS, Jung K, Gorke B. 2009. Stimula- 67. Hoverstad T, Midtvedt T. 1986. Short-chain fatty
tion of the potassium sensor KdpD kinase activity acids in germfree mice and rats. J Nutr 116:1772
by interaction with the phosphotransferase pro- 1776.
tein IIA(Ntr) in Escherichia coli. Mol Microbiol 68. Topping DL, Clifton PM. 2001. Short-chain
72:978994. fatty acids and human colonic function: roles of
58. Herrmann A, Davies JR, Lindell G, Martensson resistant starch and nonstarch polysaccharides.
S, Packer NH, Swallow DM, Carlstedt I. 1999. PhysiolRev 81:10311064.

69. Nakanishi N, Tashiro K, Kuhara S, Hayashi T, 81. Nguyen Y, Sperandio V. 2012. Enterohemor-
Sugimoto N, Tobe T. 2009. Regulation of viru- rhagic E. coli (EHEC) pathogenesis. Front Cell
lence by butyrate sensing in enterohaemorrhagic Infect Microbiol 2:90.
Escherichia coli. Microbiology 155:521530. 82. Kaper JB, Nataro JP, Mobley HL. 2004. Path-
70. Tobe T, Nakanishi N, Sugimoto N. 2011. Acti- ogenic Escherichia coli. Nat Rev Microbiol 2:123
vation of motility by sensing short-chain fatty 140.
acids via two steps in a flagellar gene regulatory 83. Sheng H, Lim JY, Knecht HJ, Li J, Hovde CJ.
cascade in enterohemorrhagic Escherichia coli. 2006. Role of Escherichia coli O157:H7 virulence
Infect Immun 79:10161024. factors in colonization at the bovine terminal
71. Zumbrun SD, Melton-Celsa AR, Smith MA, rectal mucosa. Infect Immun 74:46854693.
Gilbreath JJ, Merrell DS, OBrien AD. 2013. 84. Price SB, Wright JC, DeGraves FJ, Castanie-
Dietary choice affects Shiga toxin-producing Cornet MP, Foster JW. 2004. Acid resistance
Escherichia coli (STEC) O157:H7 colonization and systems required for survival of Escherichia coli
disease. Proc Natl Acad Sci USA 110:E2126E2133. O157:H7 in the bovine gastrointestinal tract and in
72. Fukuda S, Toh H, Hase K, Oshima K, Nakanishi apple cider are different. Appl Environ Microbiol
Y, Yoshimura K, Tobe T, Clarke JM, Topping 70:47924799.
DL, Suzuki T, Taylor TD, Itoh K, Kikuchi J, 85. Hughes DT, Terekhova DA, Liou L, Hovde CJ,
Morita H, Hattori M, Ohno H. 2011. Bifido- Sahl JW, Patankar AV, Gonzalez JE, Edrington
bacteria can protect from enteropathogenic in- TS, Rasko DA, Sperandio V. 2010. Chemical
fection through production of acetate. Nature sensing in mammalian host-bacterial commensal
469:543547. associations. Proc Natl Acad Sci USA 107:9831
73. Wlodarska M, Willing B, Keeney KM, 9836.
Menendez A, Bergstrom KS, Gill N, Russell SL, 86. Sheng H, Nguyen Y, Hovde CJ, Sperandio V.
Vallance BA, Finlay BB. 2011. Antibiotic treat- 2013. SdiA aids enterohemorrhagic Escherichi coli
ment alters the colonic mucus layer and predisposes carriage by cattle fed a forage or grain diet. Infect
the host to exacerbated Citrobacter rodentium- Immun 81:34723478.
induced colitis. Infect Immun 79:15361545. 87. Parsek MR, Val DL, Hanzelka BL, Cronan JE
74. Willing BP, Vacharaksa A, Croxen M, Jr, Greenberg EP. 1999. Acyl homoserine-lactone
Thanachayanont T, Finlay BB. 2011. Altering quorum-sensing signal generation. Proc Natl
host resistance to infections through microbial AcadSci USA 96:43604365.
transplantation. PLoS One 6:e26988. 88. Kanamaru K, Kanamaru K, Tatsuno I, Tobe T,
75. de Sablet T, Chassard C, Bernalier-Donadille A, Sasakawa C. 2000. SdiA, an Escherichia coli ho-
Vareille M, Gobert AP, Martin C. 2009. Human mologue of quorum-sensing regulators, controls
microbiota-secreted factors inhibit shiga toxin the expression of virulence factors in entero-
synthesis by enterohemorrhagic Escherichia coli haemorrhagic Escherichia coli O157:H7. Mol Mi-
O157:H7. Infect Immun 77:783790. crobiol 38:805816.
76. Brown CA, Harmon BG, Zhao T, Doyle MP. 89. Yao Y, Martinez-Yamout MA, Dickerson TJ,
1997. Experimental Escherichia coli O157:H7 car- ABrogan AP, Wright PE, Dyson HJ. 2006.
riage in calves. Appl Environ Microbiol 63:2732. Structure of the Escherichia coli quorum sensing
77. Cray WC Jr, Moon HW. 1995. Experimental protein SdiA: activation of the folding switch
infection of calves and adult cattle with Esche- by acyl homoserine lactones. J Mol Biol 355:262
richia coli O157:H7. Appl Environ Microbiol 273.
61:15861590. 90. Edrington TS, Farrow RL, Sperandio V, Hughes
78. Dean-Nystrom EA, Bosworth BT, Cray WC Jr, DT, Lawrence TE, Callaway TR, Anderson RC,
Moon HW. 1997. Pathogenicity of I O157:H7 in anNisbet DJ. 2009. Acyl-homoserine-lactone
the intestines of neonatal calves. Infect Immun autoinducer in the gastrointestinal [corrected]
65:18421848. tract of feedlot cattle and correlation to season, E.
79. Woodward MJ, Gavier-Widen D, McLaren IM, coli O157:H7 prevalence, and diet. Curr Microbiol
Wray C, Sozmen M, Pearson GR. 1999. Infection 58:227232.
of gnotobiotic calves with Escherichia coli O157: 91. Erickson DL, Nsereko VL, Morgavi DP,
H7 strain A84. Vet Rec 144:466470. Selinger LB, Rode LM, Beauchemin KA.
80. Wray C, McLaren IM, Randall LP, Pearson GR. 2002. Evidence of quorum sensing in the rumen
2000. Natural and experimental infection of ecosystem: detection of N-acyl homoserine lac-
normal cattle with Escherichia coli O157. Vet Rec tone autoinducers in ruminal contents. Can J
147:6568. Microbiol 48:374378.

PREVENTION AND
CONTROL STRATEGIES

Preharvest Food Safety for
Escherichia coli O157 and
Other Pathogenic Shiga
Toxin-Producing Strains
THOMAS E. BESSER,1 CARRIE E. SCHMIDT,1

DEVENDRA H. SHAH,1 and SMRITI SHRINGI1
21
INTRODUCTION
Upton Sinclairs novel, The Jungle, which described horrific conditions in his-
torical Chicago meat packing plants, engendered numerous reforms and reg-
ulations of the industry, including the Pure Food and Drug Act and the Meat
Inspection Act of 1906, which in turn led to vast improvements in the sanitary
conditions under which meat and meat products were handled. The massive and
highly publicized 1993 outbreak of Escherichia coli O157 associated with Jack in
the Box had a similar broad impact for the microbiological safety of food, in-
cluding the classification of this pathogen as an adulterant in ground beef, and
led to the implementation of the formal Pathogen Reduction and Hazard Analysis
and Critical Control Point Program for this bacterium and other food-borne
agents in meat processing plants. These changes were credited with significant
reduction in the incidence of human infection with E. coli O157 in the United
States over the subsequent several years; however, this trend did not continue, and
in recent years the incidence of disease due to E. coli O157 has remained stub-
bornly stable. Incidence of disease caused by non-O157 Shiga toxin-producing
E. coli (STEC) has paradoxically steadily increased, although this trend is un-
doubtedly due in part to increased use of more efficient diagnostic procedures.
1
Veterinary Microbiology and Pathology, Washington State University, Pullman, WA 99164.
421

422 BESSER ET AL.
The continued occurrence of disease out- of beef origin. However, in the absence of
breaks from E. coli O157 and other pathogenic such a highly effective decontamination step,
STEC strains linked to ground beef indicates further reductions in meat-borne exposures
the limitations of postprocessing interventions to E. coli O157 may require interventions that
to completely eliminate risk of human expo- reduce the degree of contamination of cattle
sure through contaminated meat and meat sent to slaughter. Over the years, it has be-
products. New evidence of disease linked to come clear that apart from ground beef, there
other sources, including contaminated produce, are numerous vehicles for E. coli O157 that
water, and other environmental exposures in- can result in human exposure, including fresh
cluding direct animal contacts, indicates that produce, drinking and recreational water,
this group of pathogens has a more complex direct contacts with animals, and other envi-
ecology than may have been previously recog- ronmental sources and reservoirs. This com-
nized. This article addresses some of the data plex ecology of E. coli O157 likely contributes
supporting this complexity to explain why hu- to seasonal infection pressure on cattle as
man disease incidence is not declining, dis- well, and needs to be addressed in order to
cusses the implications of the different genetic develop highly effective methods to reduce
lineages of E. coli O157 on sources and severity cattle infections with E. coli O157 and other
of human infection, and reviews the benefits pathogenic STEC strains.
and limitations of control measures directed
toward reducing the prevalence and shedding
How Do Foods of Bovine Origin Become
level of E. coli O157 and other pathogenic STEC
Contaminated with E. coli O157?
strains by cattle, otherwise known as preharvest
food safety in cattle production. There is a strong correlation between E. coli
O157 prevalence in the feces and on the hair
coats of cattle entering slaughter plants and
Twenty Years after the Jack in the Box
carcass contamination during processing (1).
Outbreak, Why Is E. coli O157
Recent studies have begun to characterize
Still a Problem?
the level of pathogen reduction in cattle feces
Why has the incidence of human infection that may be necessary to significantly reduce
with E. coli O157 and other STEC pathogens the hide and carcass contamination during
remained stubbornly steady despite the im- processing. Woerner et al. showed that fecal
plementation of stringent regulations and pen prevalence exceeding 20% was associated
large investments in improved equipment and with hide contamination prevalence of 80%
processing methods in meat packaging plants? or more (2). Similarly, Arthur et al. deter-
One important factor is seasonal variation, or mined that slaughter cattle from feedlot pens
the marked increase in the numbers of cattle with more than 20% positive fecal pats had
shedding E. coli O157 in their feces accompa- both higher hide contamination rates (25.5%)
nied by increased contamination of hair coats and higher carcass contamination at pre-
(hides) during summer months. This seasonal evisceration (14.3%), post-evisceration (2.9%),
variation results in increased contamination and post-final intervention (0.7%) stages (3).
pressure, potentially overwhelming the con- Comparative figures for slaughter cattle from
trol measures that are otherwise effective in feedlot pens with <20% positive fecal pat sam-
preventing meat contamination during the ples were lower hide contamination (5%) and
rest of the year. The effects of higher con- carcass contamination 6.3%, 0% and 0% at
tamination of cattle that overwhelm the con- pre-evisceration, post-evisceration, and post-
trol measures could be mitigated, at least in final intervention stages, respectively (3).
part, by adding a final decontamination step Overall, these data suggest that 20% fecal pat
such as gamma irradiation for meat products prevalence may be a functional threshold or

CHAPTER 21 Preharvest Food Safety 423
marker for predicting groups of feedlot cattle clear disconnect between scientific data and
having increased risk of hide or carcass con- the popular support for an idea.
tamination. Management practices that con- Unfortunately, with the exception of cattle
sistently result in fecal pat prevalence of less vaccination against E. coli O157, the efforts
than 20% may therefore be required to accom- to identify cattle management practices that
plish further progress in preharvest food safety. consistently result in significant reductions in
the frequency of cattle infection with E. coli
O157 have largely failed (reviewed in refer-
Can Live Cattle Be Managed to Reduce
ences 2023).
or Prevent E. coli O157 Infection?
Heavily contaminated cattle entering meat
processing plants can apparently overwhelm VACCINATION OF CATTLE AGAINST
the best sanitary procedures in practice; E. COLI O157: A RAY OF LIGHT
therefore, preharvest interventions in cattle
rearing, management and husbandry, trans- Although certain interventions [for example,
port, and lairage that can effectively reduce probiotics (20, 22)] show some promise for
the frequency of cattle infection with E. coli preharvest food safety against E. coli O157, vac-
O157 offer the potential to reduce human cines have been the most effective interventions
exposures. In the last 2 decades, the devel- documented to date. Currently, two commer-
opment of preharvest interventions has re- cial vaccines against E. coli O157 in cattle have
mained a major focus of the food safety been developed and are available in at least
research in the United States. The early em- some locations: a type III secretion system
phasis of preharvest food safety research was (T3SS) protein-based (Bioniche Life Sciences
based on the hypothesis that the emergence Inc., Belleville, Ontario, Canada) and a sidero-
of E. coli O157 disease in humans resulted phore receptor and porin (SRP) protein-based
from relatively recent changes in cattle man- (Epitopix, LLC, Wilmar, Minnesota) vaccine.
agement practices that favored this pathogen.
Examples of such management practices in-
Cattle Vaccine Mechanisms
cluded increased grain components in cattle
feeds (4, 5), the use of antimicrobial drugs These vaccines target different mechanisms to
and other growth-promoting feed additives induce immunity against E. coli O157 in cattle;
(611), increased intensity of cattle production, T3SS proteins play important roles in bacterial
rearing larger herds and increased confine- adherence to the bovine intestinal epithelium,
ment (1216), and the adoption of new meth- whereas SRP proteins are important for iron
ods of manure handling and disposal on farms acquisition and survival of bacteria within the
(1719). Unfortunately, however, each of these host. The products of T3SS genes such as eae
attractive hypotheses has since either been and tir (intimin and Tir), encoded within the
refuted or shown to have only minor influence locus of enterocyte effacement (LEE), play key
on cattle infection with E. coli O157, as de- roles in the colonization of bovine intestines
scribed in several comprehensive recent re- by E. coli O157 (2530). Translocation of Tir
views (2023). The hypothesis that E. coli and other effector proteins into host cells re-
O157 infection of cattle results from high-grain quires the T3SS-secreted EspA protein, which
diets and that feeding hay to the cattle would forms filaments connecting the bacteria to the
eliminate the problem merits particular note; host cell surface, as well as EspB and EspD,
while it has not been supported by subsequent which are thought to form a membrane pore
research (5, 24), it is still frequently cited as if [reviewed by Frankel et al. (31) and Caron et al.
true in the news media responses to each new (32)]. The T3SS protein-based vaccine strategy
E. coli O157 disease outbreak, demonstrating a results in induction of mucosal antibodies

424 BESSER ET AL.
capable of blocking adherence and subsequent model to compare distributions of E. coli O157
colonization of the bovine intestinal mucosa fecal shedding prevalence between cattle vac-
by E. coli O157. Under low-iron conditions, cinated with T3SS protein vaccine and non-
bacteria produce a high-affinity iron transport vaccinated cattle.
system (e.g., SRP proteins) to bring the re- The model outputs included distribu-
quired nutrient inside the bacterial cell (33). tions of fecal pen prevalence of E. coli O157
The SRP protein-based vaccine results in in- among vaccinated and nonvaccinated sum-
duction of antibodies that bind to SRP located mer-fed cattle and nonvaccinated winter-fed
on the outer membrane of the bacterial cell, cattle. One of the outcomes of this model
subsequently blocking iron transport into the was a reduction in the percentage of high-
cell, compromising the bacterial cell iron ac- prevalence pens among immunized cattle fed
quisition. Blocking iron transport by anti-SRP in the summer. In this model, approximately
antibodies renders the bacteria at a selective 58% of pens of nonvaccinated, summer-fed
disadvantage in a mixed microbial environ- cattle showed fecal prevalence of >20%. In
ment, resulting in reduced colonization. These contrast, when summer-fed cattle were vac-
approaches were recognized over a decade cinated with the T3SS protein-based vaccine,
ago, resulting in a number of subsequent vac- the percentage of pens with >20% fecal prev-
cine trials using purified T3SS or SRP protein- alence was reduced to approximately 30%.
based vaccines. Although vaccines targeting These results suggest that vaccination as an
T3SS proteins and SRP function via two intervention in cattle prior to slaughter may
entirely different mechanisms, recent meta- roughly halve the number of pens with fecal
analysis studies suggest that both vaccines are prevalence of >20%, a significant improve-
efficacious at reducing the proportion of cul- ment but still leaving 30% of pens with fecal
ture-positive animals (3437). prevalence >20%. As already discussed in
this article, according to Arthur et al. (3) and
Woener et al. (2), pens with >20% fecal prev-
Efcacy of Vaccination
alence contribute significantly to hide and
Although the effectiveness of current vaccines carcass contamination. In turn, hide and car-
in terms of reduced carcass contamination cass contamination can compromise apparent
and ultimately reduced human illnesses is un- vaccine efficacy due to cross-contamination
known, if 20% fecal prevalence is considered of hides during transport to harvest (38) or
as a functional threshold marker for signifi- cross-contamination of carcasses during pro-
cantly reduced hide and carcass contamina- cessing (39). On the basis of a postulated
tion, then the current vaccine efficacy would threshold effect involving vaccine-induced
have to effectively reduce pen prevalence to reductions in shedding density (reductions in
<20%. Ideally, the precise efficacy of each vac- the numbers of animals with fecal shedding
cine can be calculated; however, significant exceeding 103 CFU/g E. coli O157, also known
variation in the efficacy of current vaccines is as super-shedders), Matthews et al. recently
reported in different trials, and recent meta- proposed that cattle vaccination would in
analyses of multiple vaccine studies suggest fact produce substantially greater reductions
that the efficacy of current vaccines is largely in human disease caused by E. coli O157 than
uncertain (36, 37). Consequently, Vogstad predicted based solely on effects on cattle
et al. simulated the uncertainty about vaccine shedding prevalence (40). Overall, it is still
efficacy using a log-normal distribution and questionable whether the current vaccines
estimated that the mean efficacy of current would provide sufficient efficacy to accom-
T3SS protein-based vaccine is approximately plish the goal of controlling or reducing
58% (36). Using this vaccine efficacy, the postharvest E. coli O157 contamination of
authors developed a stochastic simulation cattle-derived food products.

Is It Practical To Vaccinate Cattle? with both currently available products could

have synergistic effects and result in signifi-
Recently, Withee et al. (41) combined quanti-
cantly improved efficacy; however, no pub-
tative risk assessment and marginal economic
lished studies in the literature address this
analysis to estimate the cost-benefit ratio of
possibility. Alternatively, new vaccines may be
the hypothetical O157:H7 vaccine to prevent
developed with improved efficacy. Dziva et al.
human food-borne illness. These authors de-
(27) showed that in addition to the genes en-
termined that vaccinating the entire U.S. herd
coded on LEE-T3SS, E. coli O157 colonization
would be an effective intervention for pre-
in cattle is mediated by numerous other cell
venting E. coli O157 illness in humans; how-
surface structures, including fimbriae, outer
ever, the true efficiency of vaccination will
membrane proteins, O antigens, and other bac-
primarily depend on three factors: (i) overall
terial proteins. These authors have identified
efficacy of the vaccine, (ii) herd coverage of
a novel fimbrial locus (z2199z2206; ecs2114
immunity, and (iii) the cost of vaccine per
ecs2107/locus 8) required for intestinal colo-
unit. For example, the authors estimated that
nization in calves, and demonstrated that a
if the vaccine efficacy and coverage for herd
deletion mutant is rapidly outcompeted by the
immunity were assumed at 100% and the
parent strain in coinfection studies (27). For
vaccine cost was assumed to be $3.00 per unit,
another example, Torres et al. (44) described
then vaccination will optimally prevent ap-
two chromosomal operons (lpf1 and lpf2) in
proximately 21,000 human illnesses each year
E. coli O157 closely related to the long polar
(41), or one-third to one-fifth of the annual
fimbrial (lpf) operon of Salmonella enterica
burden of disease as estimated by the CDC
serovar Typhimurium that have been associ-
(42, 43). This level of control would require
ated with the appearance of long fimbriae
vaccinating 22 million cattle intended for
that enhance colonization in animal models
slaughter each year at a total cost of $66
(reviewed in reference 45). Finally, in studies
million. In this scenario, the total benefits
that used bovine terminal rectal primary epi-
expected to accrue as a result of preventing
thelial cells and bovine intestinal tissue ex-
21,000 human illnesses would be $131 million
plants, the H7 flagellum acted as an adhesin to
(21,000 forgone cases times $6,256 per case).
bovine intestinal epithelium and contributed
In contrast, if the vaccine efficacy was assumed
to initiation of intestinal colonization (46, 47).
at 50% (close to the estimated efficacy of cur-
A following study showed that immunization
rent vaccines) and required herd coverage for
of cattle with H7 flagellin reduced coloniza-
immunity was assumed at 100%, then a $4.00
tion rates and delayed peak bacterial shed-
per unit cost of vaccination will optimally
ding following subsequent oral challenge
produce approximately 5,000 forgone illnesses
with E. coli O157 (48). Based on these data,
(41). Therefore, even the moderate efficacy of
incorporation of one or more of these anti-
current vaccines is predicted to prevent sev-
gens, perhaps in combination with antigens
eral thousand food-borne illnesses each year;
used in the currently available vaccines, may
however, there is still clearly significant room
further enhance vaccine efficacy.
for the improvement of the efficacy of current
vaccines and vaccination strategies.
CATTLE INFECTION WITH E. COLI O157
AS AN ECOLOGICAL PROBLEM
Possible Future Directions for
Vaccine Development
As microbiological methods were developed
Given that two current vaccines provide pro- to efficiently detect E. coli O157 and other
tection by completely unrelated mechanisms, pathogenic STEC strains in cattle feces and
it is possible that simultaneous vaccination environmental samples, and as more epide-

426 BESSER ET AL.
miological studies in cattle herds were com- agent, as well as nonmaintenance (incidental
pleted, several observations with profound or amplifying) host populations that do not
implications for preharvest food safety were harbor the microorganism indefinitely, but
made. These included (i) the ubiquitous pre- aid in the dissemination and amplification of
sence of E. coli O157 and other pathogenic the pathogen. Haydon et al. explain that the
STEC strains on cattle farms during the sum- number of maintenance host populations is
mer (49) but its relative absence during the generally limited, whereas the number of non-
winter, (ii) the similarity in prevalence of maintenance host populations may be unlim-
infection among cattle raised under drasti- ited (50). These definitions may be useful in
cally different management conditions rang- considering the role(s) such populations may
ing from dispersed distribution of animals play in the seasonal occurrence of E. coli O157
on pastures to housing in a highly concen- on farms and in understanding how these
trated fashion in feedlots, (iii) the transient populations may serve as targets for prehar-
nature of STEC colonization of individual vest control of these bacteria.
animals, typically lasting one to a few weeks,
(iv) the sporadic occurrence of herd out-
Cattle as Reservoirs
breaks of high prevalence E. coli O157 fecal
shedding that present all the hallmarks of Many human outbreaks with E. coli O157 have
food- or waterborne transmission, and (v) the been associated with the consumption of
detection of E. coli O157 fecal shedding in a contaminated foods of bovine origin or with
very wide range of other mammalian and direct contact with cattle or farms where in-
avian species. Basically, these observations are fected cattle are raised (51, 52). Cattle are
inconsistent with the widely held idea that the sole animal host known to demonstrate
cattle are the central sustaining reservoir for site-specific intestinal colonization with this
E. coli O157 and instead support the idea that agent, at the recto-anal junction (RAJ). RAJ
cattle are just one more mammalian host pe- colonization among cattle has been observed
riodically infected with this agent following on several dairy and beef farms without
oral exposures, albeit a host with particular resulting in a detectable illness in these ani-
significance for human exposure due to its use mals (53). Nearly all cattle herds, including
for producing human foods. The following both beef and dairy types, may be colonized.
sections of this article explore what is known As discussed previously in this article, fecal
of the ecology of this agent. shedding is associated with hide contamina-
tion, which has been demonstrated as a main
source of meat contamination at slaughter
Reservoirs of E. coli O157
(1, 54); thus research has been directed to the
on Cattle Farms
identification of preslaughter interventions
The study of reservoirs is complex, and a va- that can decrease RAJ colonization and fecal
riety of reservoir models exist for different shedding of these bacteria. Vaccination (dis-
pathogens. Much work has gone into identi- cussed above) of cattle may be promising to
fying the reservoir for E. coli O157 to formu- accomplish this goal; however, identifying
late strategies for controlling this pathogen ways to reduce or eliminate the source of
on farms in pursuit of preharvest food safety. cattle infection is equally important.
The clearest type of reservoir is a biological While cattle are likely an important part of
reservoir, a site or host where the agent can the reservoir for E. coli O157 on farms, several
always be found and serves as a source of the pieces of evidence have raised questions on
infection for target populations. Complex re- whether cattle are truly a maintenance pop-
servoirs may include maintenance host population for this pathogen. First, cattle typically
ulations that persistently harbor the infectious shed E. coli O157 only transiently during

summer months, and levels and prevalence of naturally occurring E. coli O157 infection of
cattle shedding cease or decrease drastically cattle on two feedlots in southern Alberta,
during winter months (55, 56). Second, a E. coli O157 was cultured from only 0.8% of
single strain of E. coli O157 frequently predo- the fecal pats, but 12% of the water troughs
minates on individual farms over periods sampled were found positive. E. coli O157 was
of multiple years, despite essentially disap- also cultured from 1.7% of feed bunk feed
pearing from cattle populations each winter. samples but not from fresh total mixed rations
This tendency is particularly interesting on (60). Culture-positive water troughs occurred
large feedlots that go through multiple animal seasonally: 35% of water troughs sampled
population turnovers annually, with incoming during the summer on one feedlot were
cattle originating from many diverse sources culture-positive for E. coli O157, compared to
(57). These data suggest that farms may con- 0% sampled during the winter (60). This
tain other noncattle maintenance hosts (reser- seasonal variation clearly parallels the sea-
voirs) of E. coli O157 and also cast doubt on sonality of cattle infection on farms and also
whether the cattle themselves make up the raises the question of whether the water
true maintenance host population. Recently, contamination is the source of, or results from,
it has been experimentally demonstrated that the cattle infection.
the seasonal differences in E. coli O157 shed-
ding by cattle are not due to intrinsic factors
Water as a Reservoir
within the animals. Cattle given identical
challenge doses of E. coli O157 shed the agent As described above, water is one of the most
in similar amounts and for similar durations, commonly contaminated materials on cattle
regardless of the season of exposure (58); this farms. In culture-positive water troughs E. coli
is the outcome predicted of an amplifying host O157 is consistently detected more frequently
population, where the source is the key factor in sediments than in the water column (59,
in duration and level of bacterial colonization 61). Water trough sediment consists of feed
in cattle. If cattle are simply an amplifying and fecal material admixed with numerous
host population, it seems clear that identifi- bacteria and protozoa, with rare metazoan
cation of the true maintenance reservoir(s) of species (nematodes and rotifers). Viable E. coli
E. coli O157 is critical to the development of O157 in the sediment layers of water troughs
truly preventive systems for management of can persist for greater than 245 days (61). One
E. coli O157 on cattle farms. hypothesis is that ambient temperature during
the summer is more permissible for growth
of bacteria and, therefore, may result in in-
Survivability of E. coli O157
creased bacterial populations in water troughs
in the Environment
during the summer compared to the winter
E. coli O157 is surprisingly persistent in envi- season (62). While increased ambient tem-
ronmental sites, documented to survive in perature likely plays a role in proliferation of
ovine manure for 21 months (19). The envi- bacteria, there may be other factors that in-
ronments (bedding materials and water) of fluence seasonal variation and overall survival
experimentally infected steers maintain de- of these bacteria in water troughs. For exam-
tectable viable E. coli O157 for at least 14 ple, mean coliform counts were significantly
weeks after inoculation of the cattle (59). In- higher in water troughs that were cleaned at
terestingly, in this study, E. coli O157 was least every 2 months compared to those that
cultured from the bedding and water even were cleaned less frequently (63). Addition-
during weeks when it was not possible to really, use of chlorinated or hyperchlorinated
cover E. coli O157 from cattle fecal samples water in trough microcosms failed to elimi-
(59). In a longitudinal, year-long study of nate E. coli O157 (61). These data suggest that

428 BESSER ET AL.
there are likely additional factors other than Salmonella serovar Newport, these bacteria
ambient temperature that contribute to the can be detected in nematode progeny for at
survival and proliferation of E. coli O157 in least the subsequent two generations (71).
water troughs. Another bactivorous, free-living nematode,
Diploscapter spp., has been demonstrated to
migrate rapidly toward colonies of E. coli
Role of Protozoa
O157 and to shed viable bacterial cells for at
While the relationship between E. coli O157 least a day after exposure (72). Free-living
and protozoa has not yet been clarified, nematodes protect themselves in scarcity of
LeJeune et al. demonstrated a significant in- food or harsh environmental conditions by
crease in the quantity of free-living protozoa forming arrested-development larvae (dauer)
within water trough sediment in the winter stages, however, it has not yet been deter-
compared to the summer (63). Other studies mined whether dauer stages can harbor food-
have demonstrated grazing of bacteria by pro- borne pathogens and subsequently act as a
tozoa collected from soil, lakes, and streams. source of contamination or as a reservoir for
While many bactivorous protozoa will feed on these pathogens.
any available food, preferential grazing for
bacteria also occurs. For example, E. coli O157
Role of Flies
containing Stx2a-encoding bacteriophage are
relatively resistant to grazing by Tetrahymena Many different families of flies are present on
sp. (64, 65). Survival of E. coli O157 within the cattle farms. Flies mostly multiply during the
food vacuoles and excretory vacuoles of pro- spring and are in constant contact with cattle
tozoa isolated from dairy lagoon wastewater and feed during summer and early autumn
suggests that protozoa may be vehicles for months. Many flies lay their eggs in cattle
dissemination of the bacterium to crops (66). feces, which hatch into larvae (maggots) that
Many free-living protozoa form cysts under feed on manure before maturing into pupae
stressful conditions such as temperature or within a week (73). Pupae contain a hard du-
salinity changes and food deprivation, and rable shell that allows them to survive under
these cysts can persist in the environment for harsh conditions; most flies survive in this
decades. While several bacterial genera in- stage over the winter (74, 75). Adult flies that
cluding Legionella, Mycobacterium, and Lis- emerge from pupae typically survive for only a
teria spp. have been shown to survive within few weeks. Because of flies close interactions
such protozoan cysts (6770), research is with cattle on farms, some investigators have
needed to determine whether this may also be studied flies as a component of the reservoir
true for E. coli O157. of E. coli O157. These bacteria can be cultured
from adult houseflies found in feed bunks
and cattle feed storage sheds during summer
Role of Environmental Invertebrates
months (76). In an E. coli O157 outbreak at a
Apart from protozoa, invertebrate organisms nursery school in Japan, the strains of E. coli
such as nematodes and rotifers that have the O157 isolated from patients matched those
potential for harboring E. coli O157 also in- detected in houseflies collected from within
habit water troughs and soils on cattle farms. the school (77), and the possibility that the
Research has shown that E. coli O157 can flies were acting as mechanical vectors able
amplify and persist for 5 days or more within to disseminate bacteria to food and eating
one such free-living nematode, Caenorhabditis utensils was considered. Subsequent research
elegans (71). The association of C. elegans with suggested that flies may be more than just
Salmonella spp. has been more thoroughly mechanical vectors. After oral infection of
investigated; when C. elegans is exposed to adult houseflies with E. coli O157, bacteria

were identified in the alimentary canals of suggested to play a role in dissemination of

30% of these flies up to 3 days postinfection. this agent to fresh produce that resulted in a
Orally infected flies with actively proliferating large human outbreak of disease (86).
E. coli O157 on their mouthparts demonstrate
cellular lesions similar to the attaching and
The Need for a Better Understanding
effacing lesions seen in the colonized RAJ of
of the Ecology and Reservoir
cattle (78).
Structure of E. coli O157
As mentioned previously, the reservoir for
Role of Birds
E. coli O157 is very complex. Based on the
E. coli O157 has been cultured from wild birds descriptions by Haydon et al. of complex res-
on cattle farms in many investigations. Birds, ervoirs (50), there is likely one or more
much like flies, may be seen as a general maintenance host populations that could
nuisance on farms and may act to contami- include role(s) for organisms such as proto-
nate cattle feeds and water sources, as well zoa, invertebrates, or flies on cattle farms.
as disseminate bacteria within and between Presence of a maintenance host population
farms. A surveillance study determined that outside cattle is suggested by the fact that
3% of European starlings and 4% of the cattle although swine and poultry, like cattle, are
study population were culture-positive for readily colonized with E. coli O157 in experi-
E. coli O157. In addition, these birds frequently mental settings, contamination of pork or
visited the same farms on daily feeding forays poultry meats with this agent is relatively rare
but returned nightly to share a communal (87, 88). One possible explanation for this low
roost with birds that visited other farms, prevalence may be that swine and poultry are
providing a potential method for pathogen typically reared in confinement in the United
dissemination (79). Poultry are readily exper- States, which may shield them from exposure
imentally colonized with E. coli O157 (80), but to environmental sources of E. coli O157 in-
contamination of poultry products is very fection. If so, this suggests that management
rare and human infection with E. coli O157 systems to reduce cattle exposure to envi-
resulting from contaminated poultry has rarely ronmental sources of E. coli O157 may be re-
been documented. quired to reduce their prevalence of infection.
It is also possible that the bacteria can
survive without hosts in soil or water envi-
Role of Mammals
ronments during the winter, amplifying each
E. coli O157 fecal shedding has been detected spring (as ambient temperatures increase)
in many different domestic animal species, to levels that are infectious to cattle. Several
including dogs, cats, horses, and sheep. Colo- vertebrates, including birds, cattle, and other
nization of the ovine RAJ has been demon- mammals, likely act at least as nonmain-
strated but seems to occur less efficiently tenance host populations that aid in dissemi-
(81). Colonization in wildlife including feral nation and amplification of these bacteria,
swine, deer, raccoons, opossums, and rats has especially during the summer months. More
also been reported. Deer have been frequently research leading to a better understanding
documented to shed E. coli O157, and human of the complex reservoirs of E. coli O157
infections have been traced to contaminated may lead to improved targeting of these bac-
venison (82, 83). Swine are readily experi- teria and improved preharvest control on
mentally colonized with E. coli O157, but the cattle farms along with better strategies to
prevalence of natural infection is very low reduce environmental and non-beef-product-
(84, 85). In contrast, feral swine have been related exposures contributing to human
demonstrated to shed E. coli O157 and were infection.

430 BESSER ET AL.
E. COLI O157 GENOTYPES, HUMAN associated with clinical genotypes in the

DISEASE, AND PREHARVEST FOOD SAFETY United States may simply be the result of
relatively higher virulence of clinical genotype
Various genotyping methods including multi- strains. This possibility has two important
locus enzyme electrophoresis (89, 90), octamer- implications for preharvest food safety: First,
based genome scanning (91, 92), whole-genome the virulence differences among E. coli
PCR scanning (93), pulsed-field gel electro- O157 genotypes suggest the possibility or
phoresis(94), Shiga toxin-associated bacterio- likelihood that these genotypes may respond
phage insertion, typing (95), lineage-specific differently to preharvest food safety inter-
polymorphism assay, (96), comparative ge- ventions due to other intrinsic biological dif-
nomic hybridization, (97, 98), optical mapping ferences associated with their genotypes, and
(99), and single nucleotide polymorphism typ- second, that in evaluating the efficacy of pre-
ing (100, 101) have been used to decipher harvest food safety interventions it is impor-
the population structure of E. coli O157 (102). tant to demonstrate specific reductions of
These studies revealed that bacteriophages play clinical genotypes, rather than assuming that
an important role in establishing the genetic any prevalence or shedding reductions in-
diversity among E. coli O157 isolates and that clude clinical genotypes. Recent studies
certain specific genetic lineages of E. coli O157 have shown that different lineages of E. coli
are associated with most human disease. These O157 may differ in their ability to persist on
strongly disease-associated genotypes have cattle farms through various seasons, cattle
been termed clinical genotypes whereas other diets, and animal husbandry practices. Vanaja
lineages, less frequently isolated from humans et al. (116) demonstrated that certain cattle-
with illness compatible with E. coli O157 in- associated genotypes expressed gene reper-
fection, have been termed bovine-biased geno- toires expected to improve their resistance
types (91, 92, 96, 103107). In general, the to adverse environmental conditions in com-
various genotyping methods are concordant parison to genotypes more commonly associ-
in their identification of clinical genotypes of ated with clinical disease. Some genotypes
E. coli O157 (108, 109). Populations of E. coli of E. coli O157 are more resistant to stress
O157 in different geographical regions differ factors such as heat and starvation compared
significantly in the relative frequency of par- to other genotypes (117). It is similarly possi-
ticular genotypes in different countries, and ble that different lineages of E. coli O157 may
generally clinical genotypes are more frequent respond differently to preharvest control
in cattle populations in countries with higher measures such as vaccines, probiotics, bacte-
incidences of hemolytic-uremic syndrome, a riophage treatments, or animal husbandry
severe form of illness associated with E. coli interventions. Therefore, further studies are
O157 infection (110115). On the other hand, at required to (i) specifically target bacterial
least some genotypes isolated from clinical ill- genetic factors that are responsible for the
ness in humans are not represented in cattle, differential response of different lineages of
indicating the presence of non-cattle-associated E. coli O157 to various preharvest control
reservoirs or sources of human infection (101). measures, and (ii) to confirm that any pre-
Given the similar prevalence of cattle in- harvest control measures put into practice
fection with clinical and bovine-biased line- are effective against clinical genotypes. These
ages in the United States, it seems likely that studies will aid in identifying tools to im-
people in this country are similarly exposed prove the current preharvest food safety
to both clinical and bovine-biased genotypes measures or formulate new better ways to
of E. coli O157 via ground beef, other cattle- reduce prevalence and shedding of E. coli
origin meats, and cattle environments. There- O157 on cattle farms with consistent and re-
fore, the preponderance of human disease liable results.

CONCLUSION REFERENCES
1. Elder RO, Keen JE, Siragusa GR, Barkocy-
Preharvest food safety for E. coli O157 is the Gallagher GA, Koohmaraie M, Laegreid WW.
term used for management systems that re- 2000. Correlation of enterohemorrhagic Esche-
duce the prevalence and/or magnitude of richia coli O157 prevalence in feces, hides, and
shedding of this agent by cattle populations to carcasses of beef cattle during processing. Proc
Natl Acad Sci USA 97:29993003.
reduce the risk of contamination of cattle-
2. Woerner DR, Ransom JR, Sofos JN, Dewell GA,
derived food products and subsequent human Smith GC, Salman MD, Belk KE. 2006. Deter-
exposures. Decades of research have provided mining the prevalence of Escherichia coli O157 in
a better understanding of the epidemiology cattle and beef from the feedlot to the cooler.
and ecology of E. coli O157 on cattle farms, J Food Prot 69:28242827.
but only limited progress on preharvest food 3. Arthur TM, Keen JE, Bosilevac JM, Brichta-
Harhay DM, Kalchayanand N, Shackelford SD,
safety goals has been made. Although several Wheeler TL, Nou X, Koohmaraie M. 2009.
interventions (certain feed ingredients, pro- Longitudinal study of Escherichia coli O157:H7
biotics, and vaccines) have been identified in a beef cattle feedlot and role of high-level
with statistically significant impacts on cattle shedders in hide contamination. Appl Environ
shedding of E. coli O157, the impact of these Microbiol 75:65156523.
4. Grauke LJ, Wynia SA, Sheng HQ, Yoon JW,
potential interventions remains insufficient
Williams CJ, Hunt CW, Hovde CJ. 2003. Acid
due to their limited efficacy, practical diffi- resistance of Escherichia coli O157:H7 from the
culties with their implementation, or incon- gastrointestinal tract of cattle fed hay or grain. Vet
sistency in their results, leading to limited Microbiol 95:211225.
uptakes by producers. To date, the promise 5. Hovde CJ, Austin PR, Cloud KA, Williams CJ,
of the preharvest food safety approach to re- Hunt CW. 1999. Effect of cattle diet on Esche-
richia coli O157:H7 acid resistance. Appl Environ
ducing human infection with E. coli O157 has Microbiol 65:32333235.
not been fulfilled. A more holistic approach, 6. Jacob ME, Fox JT, Narayanan SK, Drouillard
with complex ecology and genetics of this JS, Renter DG, Nagaraja TG. 2008. Effects of
bacterium in mind, is needed toward identi- feeding wet corn distillers grains with solubles
fying true maintenance host populations and with or without monensin and tylosin on the
prevalence and antimicrobial susceptibilities of
developing strategies to control E. coli O157
fecal foodborne pathogenic and commensal bac-
and other pathogenic STEC strains in these teria in feedlot cattle. J Anim Sci 86:11821190.
maintenance host populations. 7. McAllister TA, Bach SJ, Stanford K, Callaway
TR. 2006. Shedding of Escherichia coli O157:H7
by cattle fed diets containing monensin or tylosin.
ACKNOWLEDGMENT J Food Prot 69:20752083.
8. Swyers KL, Carlson BA, Nightingale KK, Belk
We acknowledge support provided by the
KE, Archibeque SL. 2011. Naturally colonized
Agriculture and Food Research Initiative of beef cattle populations fed combinations of yeast
the USDA National Institute of Agriculture culture and an ionophore in finishing diets con-
grants 2009-04248 and 2010-04487. taining dried distillers grains with solubles had
We declare no conflicts of interest with similar fecal shedding of Escherichia coli O157:H7.
J Food Prot 74:912918.
regard to the manuscript.
9. Edrington TS, Callaway TR, Bischoff KM,
Genovese KJ, Elder RO, Anderson RC, Nisbet
DJ. 2003. Effect of feeding the ionophores
CITATION
monensin and laidlomycin propionate and the
Besser TE, Schmidt CE, Shah DH, Shringi S. antimicrobial bambermycin to sheep experimen-
2014. Preharvest food safety for Escherichia tally infected with E. coli O157:H7 and Salmonella
typhimurium. J Anim Sci 81:553560.
coli O157 and other pathogenic Shiga toxin- 10. Reinstein S, Fox JT, Shi X, Alam MJ, Renter
producing strains. Microbiol Spectrum 2(5): DG, Nagaraja TG. 2009. Prevalence of Esche-
EHEC-0021-2013. richia coli O157:H7 in organically and naturally

432 BESSER ET AL.
raised beef cattle. Appl Environ Microbiol 75: 23. Jacob ME, Callaway TR, Nagaraja TG. 2009.
54215423. Dietary interactions and interventions affecting
11. LeJeune JT, Christie NP. 2004. Microbiological Escherichia coli O157 colonization and shedding in
quality of ground beef from conventionally-reared cattle. Foodborne Pathog Dis 6:785792.
cattle and raised without antibiotics label 24. Callaway TR, Carr MA, Edrington TS, Anderson
claims. J Food Prot 67:14331437. RC, Nisbet DJ. 2009. Diet, Escherichia coli
12. Renter DG, Checkley SL, Campbell J, King R. O157:H7, and cattle: a review after 10 years. Curr
2004. Shiga toxin-producing Escherichia coli in Issues Mol Biol 11:6779.
the feces of Alberta feedlot cattle. Can J Vet Res 25. Dean-Nystrom EA, Bosworth BT, Moon HW,
68:150153. OBrien AD. 1998. Escherichia coli O157:H7 re-
13. Renter DG, Sargeant JM, Hungerford LL. 2004. quires intimin for enteropathogenicity in calves.
Distribution of Escherichia coli O157:H7 within Infect Immun 66:45604563.
and among cattle operations in pasture-based 26. Cornick NA, Booher SL, Moon HW. 2002.
agricultural areas. Am J Vet Res 65:13671376. Intimin facilitates colonization by Escherichia
14. Renter DG, Sargeant JM, Oberst RD, Samadpour coli O157:H7 in adult ruminants. Infect Immun
M. 2003. Diversity, frequency, and persistence 70:27042707.
of Escherichia coli O157 strains from range cattle 27. Dziva F, van Diemen PM, Stevens MP, Smith
environments. Appl Environ Microbiol 69:542547. AJ, Wallis TS. 2004. Identification of Escherichia
15. Hancock DD, Besser TE, Kinsel ML, Tarr PI, coli O157:H7 genes influencing colonization of
Rice DH, Paros MG. 1994. The prevalence of the bovine gastrointestinal tract using signature-
Escherichia coli O157. H7 in dairy and beef cattle tagged mutagenesis. Microbiology 150:36313645.
in Washington State. Epidemiol Infect 113:199 28. Naylor SW, Roe AJ, Nart P, Spears K, Smith
207. DG, Low JC, Gally DL. 2005. Escherichia coli
16. Hutchison ML, Walters LD, Avery SM, Munro O157:H7 forms attaching and effacing lesions at the
F, Moore A. 2005. Analyses of livestock produc- terminal rectum of cattle and colonization requires
tion, waste storage, and pathogen levels and prev- the LEE4 operon. Microbiology 151:27732781.
alences in farm manures. Appl Environ Microbiol 29. Stevens MP, Roe AJ, Vlisidou I, van Diemen
71:12311236. PM, La Ragione RM, Best A, Woodward MJ,
17. Herriott DE, Hancock DD, Ebel ED, Carpenter Gally DL, Wallis TS. 2004. Mutation of toxB and
LV, Rice DH, Besser TE. 1998. Association of a truncated version of the efa-1 gene in Esche-
herd management factors with colonization of richia coli O157:H7 influences the expression and
dairy cattle by Shiga toxin-positive Escherichia secretion of locus of enterocyte effacement-
coli O157. J Food Prot 61:802807. encoded proteins but not intestinal colonization
18. Ravva SV, Sarreal CZ, Duffy B, Stanker LH. in calves or sheep. Infect Immun 72:54025411.
2006. Survival of Escherichia coli O157:H7 in 30. Sheng H, Lim JY, Knecht HJ, Li J, Hovde CJ.
wastewater from dairy lagoons. J Appl Microbiol 2006. Role of Escherichia coli O157:H7 virulence
101:891902. factors in colonization at the bovine terminal rec-
19. Kudva IT, Blanch K, Hovde CJ. 1998. Analysis of tal mucosa. Infect Immun 74:46854693.
Escherichia coli O157:H7 survival in ovine or bo- 31. Frankel G, Phillips AD, Rosenshine I, Dougan
vine manure and manure slurry. Appl Environ G, Kaper JB, Knutton S. 1998. Enteropathogenic
Microbiol 64:31663174. and enterohaemorrhagic Escherichia coli: more
20. Berry ED, Wells JE. 2010. Escherichia coli O157: subversive elements. Mol Microbiol 30:911921.
H7: recent advances in research on occurrence, 32. Caron E, Crepin VF, Simpson N, Knutton S,
transmission, and control in cattle and the pro- Garmendia J, Frankel G. 2006. Subversion of
duction environment. Adv Food Nutri Res 60:67 actin dynamics by EPEC and EHEC. Curr Opin
117. Microbiol 9:4045.
21. Food Safety and Inspection Service. 2010. Pre- 33. Neilands JB. 1995. Siderophores: structure and
harvest Management Controls and Intervention function of microbial iron transport compounds.
options for Reducing Escherichia coli O157:H7 Shed- J Biol Chem 270:2672326726.
ding in Cattle. U.S. Department of Agriculture, Food 34. Varela NP, Dick P, Wilson J. 2013. Assessing the
Safety and Inspection Service, Washington, DC existing information on the efficacy of bovine
http://www.fsis.usda.gov/shared/PDF/Reducing_ vaccination against Escherichia coli O157:H7a
Ecoli_Shedding_In_Cattle_0510.pdf?redirecthttp= systematic review and meta-analysis. Zoonoses
true. Public Health 60:253268.
22. LeJeune JT, Wetzel AN. 2007. Preharvest con- 35. Vogstad AR, Moxley RA, Erickson GE,
trol of Escherichia coli O157 in cattle. J Anim Sci Klopfenstein TJ, Smith DR. 2013. Assessment of
85:E73E80. heterogeneity of efficacy of a three-dose regimen

of a type III secreted protein vaccine for reducing 46. Mahajan A, Currie CG, Mackie S, Tree J,
STEC O157 in feces of feedlot cattle. Foodborne McAteer S, McKendrick I, McNeilly TN, Roe A,
Pathog Dis 10:678683. La Ragione RM, Woodward MJ, Gally DL,
36. Vogstad AR, Moxley RA, Erickson GE, Smith DG. 2009. An investigation of the expres-
Klopfenstein TJ, Smith DR. 2013. Stochastic sion and adhesin function of H7 flagella in the in-
simulation model comparing distributions of teraction of Escherichia coli O157 : H7 with bovine
STEC O157 faecal shedding prevalence between intestinal epithelium. Cell Microbiol 11:121137.
cattle vaccinated with type Iii secreted protein 47. Erdem AL, Avelino F, Xicohtencatl-Cortes J,
vaccines and non-vaccinated cattle. Zoonoses Giron JA. 2007. Host protein binding and adhe-
Public Health 61:283289. sive properties of H6 and H7 flagella of attaching
37. Snedeker KG, Campbell M, Sargeant JM. 2012. and effacing Escherichia coli. J Bacteriol 189:
A systematic review of vaccinations to reduce the 74267435.
shedding of Escherichia coli O157 in the faeces of 48. McNeilly TN, Naylor SW, Mahajan A, Mitchell
domestic ruminants. Zoonoses Public Health 59: MC, McAteer S, Deane D, Smith DG, Low JC,
126138. Gally DL, Huntley JF. 2008. Escherichia coli
38. Arthur TM, Bosilevac JM, Brichta-Harhay DM, O157:H7 colonization in cattle following systemic
Guerini MN, Kalchayanand N, Shackelford SD, and mu cosal immunization with purified H7
Wheeler TL, Koohmaraie M. 2007. Transporta- flagellin. Infect Immun 76:25942602.
tion and lairage environment effects on preva- 49. Barkocy-Gallagher GA, Arthur TM, Rivera-
lence, numbers, and diversity of Escherichia coli Betancourt M, Nou X, Shackelford SD, Wheeler
O157:H7 on hides and carcasses of beef cattle at TL, Koohmaraie M. 2003. Seasonal prevalence of
processing. J Food Prot 70:280286. Shiga toxin-producing Escherichia coli, including
39. Jordan D, McEwen SA, Lammerding AM, O157:H7 and non-O157 serotypes, and Salmonella
McNab WB, Wilson JB. 1999. A simulation model in commercial beef processing plants. J Food Prot
for studying the role of pre-slaughter factors on the 66:19781986.
exposure of beef carcasses to human microbial 50. Haydon DT, Cleaveland S, Taylor LH, Laurenson
hazards. Prev Vet Med 41:3754. MK. 2002. Identifying reservoirs of infection: a
40. Matthews L, Reeve R, Gally DL, Low JC, conceptual and practical challenge. Emerg Infect
Woolhouse ME, McAteer SP, Locking ME, Dis 8:14681473.
Chase-Topping ME, Haydon DT, Allison 51. Armstrong GL, Hollingsworth J, Morris JG Jr.
LJ, Hanson MF, Gunn GJ, Reid SW. 2013. 1996. Emerging foodborne pathogens: Escherichia
Predicting the public health benefit of vaccinating coli O157:H7 as a model of entry of a new path-
cattle against Escherichia coli O157. Proc Natl ogen into the food supply of the developed world.
Acad Sci U S A 110:1626516270. Epidemiol Rev 18:2951.
41. Withee J, Williams M, Disney T, Schlosser W, 52. Sargeant JM, Amezcua MR, Rajic A, Waddell L.
Bauer N, Ebel E. 2009. Streamlined analysis for 2007. Pre-harvest interventions to reduce the
evaluating the use of preharvest interventions shedding of E. coli O157 in the faeces of weaned
intended to prevent Escherichia coli O157:H7 ill- domestic ruminants: a systematic review. Zoono-
ness in humans. Foodborne Pathog Dis 6:817825. ses Public Health 54:260277.
42. Mead PS, Slutsker L, Dietz V, McCaig LF, 53. Lim JY, Li J, Sheng H, Besser TE, Potter K,
Bresee JS, Shapiro C, Griffin PM, Tauxe RV. Hovde CJ. 2007. Escherichia coli O157:H7 colo-
1999. Food-related illness and death in the United nization at the rectoanal junction of long-duration
States. Emerg Infect Dis 5:607625. culture-positive cattle. Appl Environ Microbiol 73:
43. Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV, 13801382.
Widdowson MA, Roy SL, Jones JL, Griffin PM. 54. Jacob ME, Renter DG, Nagaraja TG. 2010.
2011. Foodborne illness acquired in the United Animal- and truckload-level associations between
Statesmajor pathogens. Emerg Infect Dis 17:715. Escherichia coli O157:H7 in feces and on hides at
44. Torres AG, Kanack KJ, Tutt CB, Popov V, harvest and contamination of preevisceration beef
Kaper JB. 2004. Characterization of the second carcasses. J Food Prot 73:10301037.
long polar (LP) fimbriae of Escherichia coli O157: 55. Hancock DD, Besser TE, Rice DH, Herriott DE,
H7 and distribution of LP fimbriae in other Tarr PI. 1997. A longitudinal study of Escherichia
pathogenic E. coli strains. FEMS Microbiol Lett coli O157 in fourteen cattle herds. Epidemiol Infect
238:333344. 118:193195.
45. Lloyd SJ, Ritchie JM, Torres AG. 2012. Fim- 56. Hancock DD, Rice DH, Thomas LA, Dargatz
briation and curliation in Escherichia coli O157: DA, Besser TE. 1997. Epidemiology of Escherichia
H7: a paradigm of intestinal and environmental coli O157:H7 in feedlot cattle. J. Food Protect
colonization. Gut Microbes 3:272276. 60:462465.

434 BESSER ET AL.
57. LeJeune JT, Besser TE, Rice DH, Berg JL, 69. Pushkareva VI, Ermolaeva SA. 2010. Listeria
Stilborn RP, Hancock DD. 2004. Longitudinal monocytogenes virulence factor Listeriolysin O
study of fecal shedding of Escherichia coli O157:H7 favors bacterial growth in co-culture with the
in feedlot cattle: predominance and persistence of ciliate Tetrahymena pyriformis, causes protozoan
specific clonal types despite massive cattle popu- encystment and promotes bacterial survival inside
lation turnover. Appl Environ Microbiol 70:377384. cysts. BMC Microbiol 10:26.
58. Hovde CJ, Sheng H, Baker K, Deobald C, Davis 70. El-Etr SH, Margolis JJ, Monack D, Robison RA,
MA, Minnich SA, Besser TE. 2012. Experimental Cohen M, Moore E, Rasley A. 2009. Francisella
evaluation of the basis of seasonal vbariation in tularensis type A strains cause the rapid encyst-
bovine shedding of STEC O157, abstr P206. 8th ment of Acanthamoeba castellanii and survive in
International Symposium on Shiga Toxin-Pro- amoebal cysts for three weeks postinfection. Appl
ducing Escherichia coli Infections, Amsterdam, Environ Microbiol 75:74887500.
The Netherlands. 71. Kenney SJ, Anderson GL, Williams PL,
59. Davis MA, Cloud-Hansen KA, Carpenter J, Millner PD, Beuchat LR. 2005. Persistence of
Hovde CJ. 2005. Escherichia coli O157:H7 in en- Escherichia coli O157:H7, Salmonella Newport,
vironments of culture-positive cattle. Appl Envi- and Salmonella Poona in the gut of a free-living
ron Microbiol 71:68166822. nematode, Caenorhabditis elegans, and transmis-
60. Van Donkersgoed J, Berg J, Potter A, Hancock sion to progeny and uninfected nematodes. Int J
D, Besser T, Rice D, LeJeune J, Klashinsky S. Food Microbiol 101:227236.
2001. Environmental sources and transmission of 72. Gibbs DS, Anderson GL, Beuchat LR, Carta LK,
Escherichia coli O157 in feedlot cattle. Can Vet J Williams PL. 2005. Potential role of Diploscapter
42:714720. sp. strain LKC25, a bacterivorous nematode from
61. LeJeune JT, Besser TE, Hancock DD. 2001. soil, as a vector of food-borne pathogenic bacteria
Cattle water troughs as reservoirs of Escherichia to preharvest fruits and vegetables. Appl Environ
coli O157. Appl Environ Microbiol 67:30533057. Microbiol 71:24332437.
62. Gautam R, Bani-Yaghoub M, Neill WH, Dopfer 73. Smith T. 1908. The housefly as an agent in the
D, Kaspar C, Ivanek R. 2011. Modeling the effect dissemination of infectious disease. Am J Public
of seasonal variation in ambient temperature on Hygiene 18:312324.
the transmission dynamics of a pathogen with a 74. West LS. 1951. The Housefly, Its Natural History,
free-living stage: example of Escherichia coli O157: Medical Importance, and Control. Comstock Pub
H7 in a dairy herd. Prev Vet Med 102:1021. Co, Ithaca, NY.
63. LeJeune JT, Besser TE, Merrill NL, Rice DH, 75. DeBartolo A. 1986. Buzz off! the housefly has
Hancock DD. 2001. Livestock drinking water made a pest of himself for 25 million years. Chicago
microbiology and the factors influencing the Tribune. http://articles.chicagotribune.com/1986-
quality of drinking water offered to cattle. J Dairy 06-05/features/8602090713_1_maggots-labor-day-
Sci 84:18561862. picnic-flies
64. Steinberg KM, Levin BR. 2007. Grazing protozoa 76. Alam MJ, Zurek L. 2004. Association of Esche-
and the evolution of the Escherichia coli O157:H7 richia coli O157:H7 with houseflies on a cattle
Shiga toxin-encoding prophage. Proc Biol Sci 274: farm. Appl Environ Microbiol 70:75787580.
19211929. 77. Wada A. 1997. [Molecular analysis of entero-
65. Lainhart W, Stolfa G, Koudelka GB. 2009. Shiga hemorrhagic Escherichia coli O157:H7 isolates in
toxin as a bacterial defense against a eukaryotic Japan 1996 using pulsed-field gel electrophore-
predator, Tetrahymena thermophila. J Bacteriol sis]. Nippon Rinsho 55:665670 (In Japanese).
191:51165122. 78. Kobayashi M, Sasaki T, Saito N, Tamura K,
66. Ravva SV, Sarreal CZ, Mandrell RE. 2010. Suzuki K, Watanabe H, Agui N. 1999. Houseflies:
Identification of protozoa in dairy lagoon waste- not simple mechanical vectors of enterohemor-
water that consume Escherichia coli O157:H7 rhagic Escherichia coli O157:H7. Am J Trop Med
preferentially. PLoS One 5:e15671. Hyg 61:625629.
67. Steinert M, Birkness K, White E, Fields B, 79. LeJeune JT, Homan J, Linz G, Pearl DL. 2008.
Quinn F. 1998. Mycobacterium avium bacilli grow Role of the European starling in the transmission of
saprozoically in coculture with Acanthamoeba E. coli O157 on dairy farms, p 3134. In Timm RM,
polyphaga and survive within cyst walls. Appl Madon MB (ed), Proceedings of the 23rd Vertebrate
Environ Microbiol 64:22562261. Pest Conference, University of California, Davis, CA.
68. Kilvington S, Price J. 1990. Survival of Legionella 80. Stavric S, Buchanan B, Gleeson TM. 1993.
pneumophila within cysts of Acanthamoeba Intestinal colonization of young chicks with
polyphaga following chlorine exposure. J Appl Escherichia coli O157:H7 and other verotoxin-
Bacteriol 68:519525. producing serotypes. J Appl Bacteriol 74:557563.

81. Best A, Clifford D, Crudgington B, Cooley WA, negative, beta-glucuronidase-negative enterohem-

Nunez A, Carter B, Weyer U, Woodward MJ, orrhagic Escherichia coli O157. J Bacteriol 183:
La Ragione RM. 2009. Intermittent Escherichia 68856897.
coli O157:H7 colonisation at the terminal rectum 93. Ohnishi M, Terajima J, Kurokawa K, Nakayama
mucosa of conventionally-reared lambs. Vet Res K, Murata T, Tamura K, Ogura Y, Watanabe H,
40:9. Hayashi T. 2002. Genomic diversity of entero-
82. Keene WE, Sazie E, Kok J, Rice DH, Hancock hemorrhagic Escherichia coli O157 revealed by
DD, Balan VK, Zhao T, Doyle MP. 1997. An whole genome PCR scanning. Proc Natl Acad Sci
outbreak of Escherichia coli O157:H7 infections USA 99:1704317048.
traced to jerky made from deer meat. JAMA 94. Kudva IT, Evans PS, Perna NT, Barrett TJ,
277:12291231. DeCastro GJ, Ausubel FM, Blattner FR,
83. Rabatsky-Ehr T, Dingman D, Marcus R, Calderwood SB. 2002. Polymorphic amplified
Howard R, Kinney A, Mshar P. 2002. Deer meat typing sequences provide a novel approach to
as the source for a sporadic case of Escherichia Escherichia coli O157:H7 strain typing. J Clin
coli O157:H7 infection, Connecticut. Emerg Infect Microbiol 40:11521159.
Dis 8:525527. 95. Shaikh N, Tarr PI. 2003. Escherichia coli O157:H7
84. Chapman PA, Siddons CA, Gerdan Malo AT, Shiga toxin-encoding bacteriophages: integrations,
Harkin MA. 1997. A 1-year study of Escherichia excisions, truncations, and evolutionary implica-
coli O157 in cattle, sheep, pigs and poultry. tions. J Bacteriol 185:35963605.
Epidemiol Infect 119:245250. 96. Yang Z, Kovar J, Kim J, Nietfeldt J, Smith DR,
85. Booher SL, Cornick NA, Moon HW. 2002. Moxley RA, Olson ME, Fey PD, Benson AK.
Persistence of Escherichia coli O157:H7 in exper- 2004. Identification of common subpopulations
imentally infected swine. Vet Microbiol 89:6981. of non-sorbitol-fermenting, beta-glucuronidase-
86. Jay MT, Cooley M, Carychao D, Wiscomb GW, negative Escherichia coli O157:H7 from bovine
Sweitzer RA, Crawford-Miksza L, Farrar JA, production environments and human clinical sam-
Lau DK, OConnell J, Millington A, Asmundson ples. Appl Environ Microbiol 70:68466854.
RV, Atwill ER, Mandrell RE. 2007. Escherichia 97. Wick LM, Qi W, Lacher DW, Whittam TS.
coli O157:H7 in feral swine near spinach fields and 2005. Evolution of genomic content in the step-
cattle, central California coast. Emerg Infect Dis wise emergence of Escherichia coli O157:H7. J
13:19081911. Bacteriol 187:17831791.
87. Levine P, Rose B, Green S, Ransom G, Hill W. 98. Zhang Y, Laing C, Steele M, Ziebell K, Johnson
2001. Pathogen testing of ready-to-eat meat and R, Benson AK, Taboada E, Gannon VP. 2007.
poultry products collected at federally inspected Genome evolution in major Escherichia coli O157:
establishments in the United States, 1990 to 1999. H7 lineages. BMC Genomics 8:121.
J Food Prot 64:11881193. 99. Kotewicz ML, Mammel MK, LeClerc JE,
88. Tutenel AV, Pierard D, Van Hoof J, Cornelis M, Cebula TA. 2008. Optical mapping and 454 se-
De Zutter L. 2003. Isolation and molecular quencing of Escherichia coli O157:H7 isolates
characterization of Escherichia coli O157 isolated linked to the US 2006 spinach-associated out-
from cattle, pigs and chickens at slaughter. Int J break. Microbiology 154:35183528.
Food Microbiol 84:6369. 100. Manning SD, Motiwala AS, Springman AC, Qi
89. Whittam TS, Wolfe ML, Wachsmuth IK, W, Lacher DW, Ouellette LM, Mladonicky
Orskov F, Orskov I, Wilson RA. 1993. Clonal JM, Somsel P, Rudrik JT, Dietrich SE, Zhang
relationships among Escherichia coli strains that W, Swaminathan B, Alland D, Whittam TS.
cause hemorrhagic colitis and infantile diarrhea. 2008. Variation in virulence among clades of
Infect Immun 61:16191629. Escherichia coli O157:H7 associated with disease
90. Feng P, Lampel KA, Karch H, Whittam TS. outbreaks. Proc Natl Acad Sci USA 105:4868
1998. Genotypic and phenotypic changes in the 4873.
emergence of Escherichia coli O157:H7. J Infect 101. Bono JL, Smith TP, Keen JE, Harhay GP,
Dis 177:17501753. McDaneld TG, Mandrell RE, Jung WK, Besser
91. Kim J, Nietfeldt J, Benson AK. 1999. Octamer- TE, Gerner-Smidt P, Bielaszewska M, Karch H,
based genome scanning distinguishes a unique Clawson ML. 2012. Phylogeny of Shiga toxin-
subpopulation of Escherichia coli O157:H7 strains producing Escherichia coli O157 isolated from
in cattle. Proc Natl Acad Sci USA 96:1328813293. cattle and clinically ill humans. Mol Biol Evol
92. Kim J, Nietfeldt J, Ju J, Wise J, Fegan N, 29:20472062.
Desmarchelier P, Benson AK. 2001. Ancestral 102. Karama M, Gyles CL. 2010. Methods for
divergence, genome diversification, and phylo- genotyping verotoxin-producing Escherichia coli.
geographic variation in subpopulations of sorbitol- Zoonoses Public Health 57:447462.

436 BESSER ET AL.
103. Chapman PA, Siddons CA. 1994. A comparison RM, Rivas M. 2008. Characterisation of Shiga
of strains of Escherichia coli O157 from humans toxin-producing Escherichia coli O157 strains iso-
and cattle in Sheffield, United Kingdom. J Infect lated from humans in Argentina, Australia and
Dis 170:251253. New Zealand. BMC Microbiol 8:46.
104. Besser TE, Shaikh N, Holt NJ, Tarr PI, Konkel 112. Mellor GE, Sim EM, Barlow RS, DAstek BA,
ME, Malik-Kale P, Walsh CW, Whittam TS, Galli L, Chinen I, Rivas M, Gobius KS. 2012.
Bono JL. 2007. Greater diversity of Shiga toxin- Phylogenetically related Argentinean and Aus-
encoding bacteriophage insertion sites among tralian Escherichia coli O157 isolates are distin-
Escherichia coli O157:H7 isolates from cattle than guished by virulence clades and alternative Shiga
in those from humans. Appl Environ Microbiol toxin 1 and 2 prophages. Appl Environ Microbiol
73:671679. 78:47244731.
105. Lejeune JT, Abedon ST, Takemura K, Christie 113. Mellor GE, Besser TE, Davis MA, Beavis B,
NP, Sreevatsan S. 2004. Human Escherichia coli Jung W, Smith HV, Jennison AV, Doyle
O157:H7 genetic marker in isolates of bovine or- CJ, Chandry PS, Gobius KS, Fegan N. 2013.
igin. Emerg Infect Dis 10:14821485. Multilocus genotype analysis of Escherichia coli
106. Bono JL, Keen JE, Clawson ML, Durso LM, O157 isolates from Australia and the United States
Heaton MP, Laegreid WW. 2007. Association of provides evidence of geographic divergence. Appl
Escherichia coli O157:H7 tir polymorphisms with Environ Microbiol 79:50505058.
human infection. BMC Infect Dis 7:98. 114. Franz E, van Hoek AH, van der Wal FJ, de Boer
107. Clawson ML, Keen JE, Smith TP, Durso LM, A, Zwartkruis-Nahuis A, van der Zwaluw K,
McDaneld TG, Mandrell RE, Davis MA, Bono Aarts HJ, Heuvelink AE. 2012. Genetic fea-
JL. 2009. Phylogenetic classification of Esche- tures differentiating bovine, food, and human
richia coli O157:H7 strains of human and bovine isolates of shiga toxin-producing Escherichia coli
origin using a novel set of nucleotide polymor- O157 in The Netherlands. J Clin Microbiol 50:
phisms. Genome Biol 10:R56. 772780.
108. Whitworth J, Zhang Y, Bono J, Pleydell E, 115. Lee K, French NP, Hara-Kudo Y, Iyoda S,
French N, Besser T. 2010. Diverse genetic markers Kobayashi H, Sugita-Konishi Y, Tsubone H,
concordantly identify bovine origin Escherichia coli Kumagai S. 2011. Multivariate analyses revealed
O157 genotypes underrepresented in human dis- distinctive features differentiating human and
ease. Appl Environ Microbiol 76:361365. cattle isolates of Shiga toxin-producing Esche-
109. Laing CR, Buchanan C, Taboada EN, Zhang Y, richia coli O157 in Japan. J Clin Microbiol 49:
Karmali MA, Thomas JE, Gannon VP. 2009. 14951500.
In silico genomic analyses reveal three distinct 116. Vanaja SK, Springman AC, Besser TE, Whittam
lineages of Escherichia coli O157:H7, one of TS, Manning SD. 2010. Differential expression
which is associated with hyper-virulence. BMC of virulence and stress fitness genes between
Genomics 10:287. Escherichia coli O157:H7 strains with clinical or
110. Whitworth JH, Fegan N, Keller J, Gobius KS, bovine-biased genotypes. Appl Environ Microbiol
Bono JL, Call DR, Hancock DD, Besser TE. 76:6068.
2008. International comparison of clinical, bo- 117. Lee K, French NP, Jones G, Hara-Kudo Y,
vine, and environmental Escherichia coli O157 Iyoda S, Kobayashi H, Sugita-Konishi Y,
isolates on the basis of Shiga toxin-encoding Tsubone H, Kumagai S. 2012. Variation in stress
bacteriophage insertion site genotypes. Appl resistance patterns among stx genotypes and
Environ Microbiol 74:74477450. genetic lineages of shiga toxin-producing Esche-
111. Leotta GA, Miliwebsky ES, Chinen I, Espinosa richia coli O157. Appl Environ Microbiol 78:3361
EM, Azzopardi K, Tennant SM, Robins-Browne 3368.

Peri- and Postharvest Factors in the
Control of Shiga Toxin-Producing
Escherichia coli in Beef
RODNEY A. MOXLEY1 and GARY R. ACUFF2

22
ATTRIBUTION OF BEEF AS A SOURCE OF STEC-ASSOCIATED ILLNESS
Based on outbreak data acquired from 1998 to 2008 by the U.S. Centers for
Disease Control and Prevention (CDC), known food-borne etiological agents
were estimated to have caused 4,589 outbreaks, 9,638,301 cases, 57,462 hos-
pitalizations, and 1,451 deaths per year in the United States (1). From these data,
it was estimated that 482,199 cases (5.0%), 2,650 hospitalizations (0.03%), and
51 deaths (0.0005%) were attributable to bacteria consumed from beef. Of the
4,589 outbreaks, 103 (2.2%) and 3 (0.065%) were further attributable to beef-
acquired Shiga toxin-producing Escherichia coli (STEC) O157 and non-O157
strains. Of the outbreaks caused by STEC O157 and non-O157 STEC strains for
all food commodities combined, 103 of 186 (55.3%) and 3 of 6 (50%), respec-
tively, were attributable to beef. On the basis of data acquired from the
CDC from 2000 to 2008, of 9,388,075 cases of domestically acquired food-borne
illness in the United States caused by 31 major pathogens, 63,153 (0.67%) were
due to STEC O157 and 112,752 (1.20%) were due to non-O157 STEC infection
(2). In the same study, of 35,796 hospitalizations, 2,138 (5.97%) were due to
STEC O157 and 271 (0.76%) were due to non-O157 STEC; of 861 deaths, 20
(2.32%) were due to STEC O157 and 0 were due to non-O157 STEC infection.
1
School of Veterinary Medicine and Biomedical Sciences, University of Nebraska-Lincoln, Lincoln, NE 68583-0905;
2
Department of Animal Science, Texas A&M University, College Station, TX 77843-2471.
437

438 MOXLEY AND ACUFF
PERIHARVEST STEC PREVALENCE material. Since the 1970s the hide has been
recognized as the primary source of bacterial
Ruminant Intestines and Feces Are the contamination of carcass surfaces, contami-
Major Reservoir of STEC nation that occurs during the process of hide
removal at slaughter (13). Many studies sup-
STEC strains are part of the normal intestinal
port this hypothesis and have led to the con-
flora of healthy cattle and sheep and, conse-
clusion that the hide is the major source of
quently, are shed in their feces (35). E. coli
carcass contamination by E. coli O157:H7 and
O157:H7 is the prototype of STEC; however, it
non-O157 STEC (1418). Carcass surfaces
is only one of the more than 435 serotypes and
coming in contact with droplets aerosolized
120 O serogroups of STEC known to colonize
from hides during removal were shown to
the intestines of cattle (3, 4). More than 470
contain higher counts of aerobic bacteria
different STEC serotypes have been isolated
and Enterobacteriaceae and also a higher
from humans, with most of these serotypes
prevalence of pathogens (19). E. coli O157:H7
identified in cattle, beef, or both, and more
has been isolated from up to 23% of air
than 100 were associated with human disease
samples from hide removal areas, in contrast
(4, 6). In North America, cattle are a major re-
to 0% for air samples from evisceration
servoir of STEC; however, in countries such
and fabrication areas at different plants (19).
as Australia, sheep are of greater significance
Aerosolization of droplets from hides occurs
(4). Although many different serotypes of
especially through the use of equipment and
STEC are carried and subclinically shed by
processes that remove the hide with consid-
healthy cattle of all ages, a subset naturally
erable force, such as mechanical and hydraulic
causes diarrheal disease in young calves (7).
hide pullers.
Experimentally, young calves and, especially,
Fecal, hide, and carcass prevalence of
neonates develop more extensive intestinal
E. coli O157 is directly correlated, and most
colonization with STEC than do older ani-
contamination is thought to occur from ani-
mals. This increased susceptibility may be
mals within the same lots (16, 20). In one
due to lack of immunity and to a reduced rate
study, the frequency of E. coli O157:H7 or
of epithelial cell turnover compared to that
O157:NM in feces and on hides within groups
of the older animal (8, 9). Animals that shed
of fed cattle from single sources (lots) pre-
higher concentrations of STEC in their feces
sented for slaughter at meat processing
pose a greater risk for transmission to other
plants in the midwestern United States was
animals around them, and also a greater risk
determined, as was the frequency of carcass
of meat contamination during the slaughter-
contamination during processing from cattle
ing process (10, 11). Colonization of the recto-
within the same lots (16). In that study, E. coli
anal junction in cattle of different ages with
O157 prevalence was 28% in feces and 11%
E. coli O157:H7 has been associated with
on hides. Carcass samples were taken at three
higher-level shedding and increased risk of
points during processing: preevisceration, post-
transmission, the so-called super-shedder
evisceration before antimicrobial intervention,
state (12).
and postprocessing after carcasses entered the
cooler. Of 30 lots sampled, 87% had at least
one E. coli O157-positive preevisceration sam-
Fecal, Hide, and Carcass Prevalence
ple, 57% of lots were positive postevisceration,
Is Correlated, and Hides Are the
and 17% had positive postprocessing sam-
Major Vehicles of Contamination
ples. Prevalence of E. coli O157 in the three
of Carcass Surfaces
postprocessing samples was 43%, 18%, and
The exposed surface of the hide and hair 2%, respectively. Reduction in carcass preva-
of cattle accumulates dust, dirt, and fecal lence from preevisceration to postprocessing

CHAPTER 22 Peri- and Postharvest Factors in STEC Control in Beef 439
suggested that sanitary procedures were ef- and non-O157 STEC in feces and hides. The
fective within the processing plants. highest E. coli O157:H7 prevalence in feces
Most E. coli O157 strains that contaminate was detected in the summer, and the highest
carcasses originate from animals within the on hides was from spring through fall. In
same lot going to slaughter. Evidence for this contrast, non-O157 STEC prevalence in feces
was provided in a study that compared pulsed- was lower in the summer and higher in the
field gel electrophoresis (PFGE) patterns of spring and fall and also peaked in the fall on
isolates from feces, hides, and carcasses (20). hides. The efficiency in the recovery of non-
Approximately 68% of E. coli O157 isolates O157 STEC could have influenced these re-
from carcasses had the same PFGE pattern as sults. The prevalence of Shiga toxin gene
those from feces and hides. On individual (stx)-positive bacteria as detected by PCR
carcasses, isolates recovered before eviscera- was, in descending order, 96.6% on carcasses
tion matched 65.3% of those recovered after prior to evisceration, 92.0% on hides, 34.3%
evisceration. Also, on individual carcasses, in feces, and 16.2% on carcasses after inter-
66.7% of the isolates recovered in the cooler ventions. The approximate concentration of
matched those recovered before evisceration. non-O157 STEC and stx-positive cells on
PFGE genotyping confirmed that the majority postintervention carcasses was 3.0 cells per
of E. coli O157 found on the carcass is the re- 100 cm2 for only 4% of carcasses. Pathogen
sult of preevisceration contamination. Hence, prevalence on hides may reflect several
the data indicated the need to apply additional sources of contamination, such as soils, feces
in-plant intervention strategies aimed at pre- from other animals, and possibly lairage. How-
venting direct contamination of the carcasses ever, these results further confirmed that
early in processing. hides were the major source of contamina-
In a study in which beef carcass sponge tion for beef carcasses and that postharvest
samples were collected at four large process- interventions used by the beef industry were
ing plants in the United States, 53.9% of effective.
preevisceration samples and 8.3% of post- The hide-level prevalence of STEC has
processing samples were positive for non- been shown to increase as a result of com-
O157 STEC (21). Altogether, 361 non-O157 mingling cattle, e.g., in pens, sales barns,
STEC isolates were recovered, belonging to trucks, and lairage at the abattoir, in several
41 different O serogroups. O serogroups that studies (2224). The effect of commingling
previously had been associated with human on hide-level prevalence during lairage at
disease accounted for 49% of the isolates. the abattoir was demonstrated experimentally
The significant decrease in prevalence of with nonpathogenic bacteria carrying antibi-
STEC detected from preevisceration to post- otic-resistant markers (viz., E. coli K-12 and
processing was attributed to the various in- Pseudomonas fluorescens) (22). At the abattoir,
terventions in place, viz., steam vacuum, hot the initial prevalence of animals positive for
water, organic acids, and steam pasteurization. the hide marker (11.1%) inoculated at un-
In another study of fed beef cattle har- loading increased to 100% on hides before
vested at three midwestern beef processing skinning and to 88.8% on skinned carcasses.
plants, the prevalence of E. coli O157:H7 in In addition, another marker inoculated on
samples was 5.9% in feces, 60.6% on hides, environmental surfaces in lairage pens, races,
26.7% on dehided carcass surfaces prior to the and the stunning box was detected on 83.3%
preevisceration wash, and 1.2% on carcasses of hides before skinning and 88.8% of skinned
sampled at chilling (postintervention) at con- carcasses. These results demonstrated that
centrations of approximately <3.0 cells per both the livestock market process and the
100 cm2 (15). Somewhat different results were unloading-to-skinning process at abattoirs can
found for the prevalence of E. coli O157:H7 facilitate the extensive spread of microbial

contamination on hides, not just within, but mechanically tenderized or reconstructed

also between, batches of animals. products (32). Based on STEC isolates sub-
The significant role of cross-contamination mitted to the CDC from 1983 to 2002, six O
of hides at the abattoir has also been demon- groups comprising 13 serotypes were identi-
strated by characterization of the phage and fied as the cause of 71% of non-O157 STEC
verocytotoxin (Shiga toxin) types of isolates disease in the United States (6). The six O
(24). The majority of cattle (84%) were found serogroups, O26, O111, O103, O121, O45, and
to have subtypes of STEC O157 on their hides O145, were responsible for 22, 16, 12, 8, 7, and
that had not been found previously in any 5% of cases, respectively. Serotypes included
animal from the farm of origin, strongly sug- O26:H11 or nonmotile (NM); O45:H2 or NM;
gesting that contamination occurred once O103:H2, H11, H25, or NM; O111:H8 or NM;
animals had left the farm of origin. Several O121:H19 or H7; and O145:NM. On September
variables and factors were found to be 20, 2011, the USDA-FSIS declared O26, O45,
strongly associated with cross-contamination O103, O111, O121, and O145 adulterants in
of cattle hides at the univariate level: com- certain raw beef products (33).
mercial transport to slaughter; transport with In response to the need for uniform de-
other animals; use of a crush (restraining crate tection methods for STEC in meat products,
used when reading ear tags); line automation; the USDA-FSIS published preferred methods
and increasing slaughterhouse throughput. in the Microbiology Laboratory Guidebook
Studies on the effects of transportation on (MLG). The most recent MLG specifies dif-
hide contamination with STEC O157 have ferent protocols for O157:H7 and non-O157
yielded variable results. Although transport STEC that involve screening of samples by
stress may possibly lead to immunosuppres- real-time PCR or lateral flow device (in the
sion (25), and transport and fasting are be- case of O157:H7), and if positive, culture pro-
lieved to increase fecal shedding of STEC cedures that involve selective enrichment
O157 (23, 26, 27), other studies have found broth, immunomagnetic separation, chromo-
that transport of cattle had no influence on genic agar plating, agglutination testing, and
shedding (2830). Stanford et al. (31) con- subsequent confirmation steps. The past 2
cluded that transportation did not affect prev- decades of research have resulted in the
alence of hide contamination with E. coli O157, development of effective reagents and meth-
and the feedlot pen prevalence had a greater odologies for detection, isolation, and iden-
effect on hide contamination at the slaughter tification of STEC O157:H7. In contrast,
plant than transportation factors, including immunological reagents have only recently
temperature-humidity index, loading density, become commercially available for non-O157
and duration of transport. STEC, and although DNA- and culture-based
protocols have been developed, investigators
have reported difficulty in their detection
POSTHARVEST STEC PREVALENCE (34, 35). Part of the problem is the lack of
biochemical differences between non-O157
The United States Department of Agriculture STEC and other E. coli strains, in contrast to
(USDA), Food Safety and Inspection Service the relatively unique clone of O157:H7 that
(FSIS) in October 1994 declared E. coli was responsible for most outbreaks of illness
O157:H7 an adulterant in raw ground beef, worldwide. In addition, six times as many
and began a sampling program to test for this serogroups are being targeted, and within
organism in raw ground beef prepared in some of these serogroups the organisms are
federally inspected plants and retail stores. biochemically diverse.
In January 1999, the FSIS expanded that According to the most recent FSIS MLG
declaration to include nonintact beef, such as for detection of E. coli O157:H7 from meat,

confirmation of the sample as positive re- of pathogenic STEC strains are not within the
quires cultural isolation of the organism, bio- top six serogroups. Clearly, improved meth-
chemical identification of the isolate as E. coli, odologies are needed for the accurate testing
and on this isolate, serological or genetic of meat samples for non-O157 STEC.
(PCR) detection of O157, and detection of at
least one of the following: Shiga toxin pro-
duction, stx, or the H7 gene. Similarly, ac- PERI- AND POSTHARVEST INTERVENTIONS
cording to the most recent MLG for detection
of non-O157 STEC, confirmation of a sample USDA-FSIS Directive 7120.1 (http://www.fsis.
as positive requires cultural isolation of the usda.gov/OPPDE/rdad/FSISDirectives/7120.1.
organism, biochemical identification of the pdf) provides a list of Safe and Suitable
isolate as E. coli, and on this isolate, serological Ingredients Used in the Production of Meat,
(agglutination test) and genetic (PCR) detec- Poultry, and Egg Products. Table 1 includes
tion of O26, O45, O103, O111, O121, or O145, those antimicrobials approved for use in
and genetic (PCR) detection of both the stx beef, including hides, carcasses, primals, sub-
and intimin (eae) genes. primals, cuts, ground beef, sausages, cooked
Bosilevac and Koohmaraie conducted a product, ready-to-eat, and other products.
large-scale study to determine the prevalence Most of these antimicrobials are chemicals,
and virulence gene characteristics of non- but they also include biologicals (e.g., bacte-
O157 STEC in commercial ground beef sam- riophage for use on hides, and food-grade
ples in the United States (36). A total of 4,133 bacteria such as Lactobacillus sp. and other
samples were cultured; of these samples, stx genera). Although approved for use, efficacy of
was detected in 1,006 (24.3%) and STEC in these antimicrobials is not a requirement for
300 (7.3%). A total of 338 unique STEC inclusion in this list. Many studies testing
isolates that belonged to 99 different serotypes the efficacy of different antimicrobials against
were obtained from these samples; the most STEC on hides, carcasses, parts, and products
frequent serotype identified was O113:H21. have been published; however, they would
Only six isolates qualified as FSIS-defined need to be approved for use by the USDA-
adulterants; four were O103:H2, one was FSIS before they could be implemented in
O26:H11, and one was O26:H21. Only four plants. Several hide and carcass interventions
other isolates that were deemed pathogenic discussed below were developed to reduce
STEC were detected, based on a PCR screen STEC contamination of beef carcasses and
for a number of virulence genes, and these processed beef. Reductions in the prevalence
isolates came from enrichments that were of E. coli O157:H7 on hides are directly cor-
negative for the intimin gene (eae). The FSIS related with lower carcass prevalence rates
MLG for non-O157 STEC involves classifying (14, 37).
a sample as negative and not subject to further
testing if it screens negative by real-time PCR
Hide Interventions
for eae. Of the six isolates that classified as
adulterants, a number of them (although O It has been surmised for some time that con-
group, stx and eae positive) lacked other vir- tamination on the hide of cattle was the pri-
ulence factors associated with severe disease. mary source of carcass contamination with
The authors noted that, narrowly focusing on enteric pathogens; therefore, numerous re-
only the described top six STEC serogroups ports investigate the possibility of cleaning the
poses the problem of identifying numerous hide before removal. Byrne et al. (38) reported
isolates of little pathogenic concern while in a study conducted in Ireland that washing
missing other significant pathogenic STEC cattle with a power-hose for 3 min would
serogroups, especially since nearly one-third remove all visible fecal contamination and

TABLE 1 USDA-FSIS table of safe and suitable antimicrobials for beefa

Substance Product
Acetic acid Dried and fermented sausages

Aqueous mixture of peroxyacetic acid, hydrogen Use in process water used for washing, rinsing, or cooling
peroxide, acetic acid, sulfuric acid (optional), and whole or cut meat including carcasses, parts, trim,
1-hydroxyethylidine-1, 1-diphosphonic acid (HEDP) and organs
Aqueous solution of sodium diacetate, lactic acid, pectin, Cooked meat products
and acetic acid
Aqueous solution of hydrochloric acid, phosphoric acid, Raw and ready-to-eat (RTE) poultry carcasses, parts, trim,
and lactic acid and organs, and beef products
Aqueous solution of sodium octanoate or octanoic acid Fresh meat primals and subprimals and cuts
and either glycerin and/or propylene glycol and/or
a polysorbate surface active agent
Aqueous solution of sulfuric acid and sodium sulfate In form of spray, wash or dip on surface of meat and
poultry products
Blend of citric acid and sorbic acid To reduce microbial load of purge trapped inside soaker
pads in packages of raw whole muscle cuts of meat
Blend of lactic acid, citric acid, and potassium hydroxide Beef carcasses, heads, and organs including unskinned
livers, tongues, tails, primal cuts, subprimal cuts,
and trimmings
Blend of salt, sodium acetate, lemon extract, Ground beef, cooked, cured, comminuted sausages
and grapefruit extract (e.g., bologna), and RTE whole muscle meat products;
beef steaks
Blend of salt, lactic acid, sodium diacetate, Various nonstandardized RTE meat products and
and mono- and diglycerides standardized meat poultry products that permit use
of any safe and suitable antimicrobial agent
Mixture of hops beta acids, egg white lysozyme, In dressing used in refrigerated meat salads
and cultured skim milk
Mixture of maltodextrin, cultured dextrose, sodium In salads, sauces, and dressings to which fully cooked
diacetate, egg white lysozyme, and nisin preparation meat will be added
Acidied sodium chlorite Meat carcasses, parts and organs; processed,
comminuted, or formed meat products (including RTE)
Ammonium hydroxide Beef carcasses (in hot boxes and holding coolers)
and boneless beef trimmings
Anhydrous ammonia Ground beef
Bacteriophage preparation (E. coli O157:H7 targeted) On hides of live animals in holding pens prior to
slaughter
Bacteriophage preparation (E. coli O157:H7 targeted) Red meat parts and trim prior to grinding
Calcium hypochlorite Red meat carcasses down to a quarter of a carcass;
in water used in meat processing; beef primals
Chlorine dioxide Red meat carcasses down to a quarter of a carcass;
in water used in meat processing; beef primals
Chlorine gas Beef trimmings prior to grinding and beef subprimals;
bologna in edible casing; fully cooked meat products in
impermeable and permeable prestuck casings; separated
beef heads and associated offal products (e.g., hearts,
livers, tails, tongues); in brine to cool fully cooked RTE
meat products: sausages and similar products in
natural casings
Citric acid In meat products (e.g., beef injected with cultured
substrates) and RTE meat products (e.g., hot dogs and

TABLE 1 (Continued)
Substance Product
luncheon meat). Cultured substrates are not intended for

use in infant formula or foods
Cultured substrates that are produced by the In enhanced meat and poultry products (e.g., beef
fermentation of natural food sources such as dairy injected with a solution) and RTE meat products
sources, fruit- and vegetable-based sources, and others; (e.g., hot dogs)
substrate is fermented to organic acids by individual
microorganisms including Streptococcus thermophilus,
Bacillus coagulans, Lactobacillus acidophilus, and others
Cultured sugar (derived from corn, cane, or beets) In enhanced meat products (e.g., beef injected with
a solution) and RTE meat products (e.g., hot dogs)
Cultured sugar and vinegar (derived from corn, cane, For use in water applied to beef hides, carcasses, heads,
or beets) trim, parts, and organs
1,3-dibromo-5,5-dimethylhydantoin (DBDMH) In casings and on cooked (RTE) meat products
Egg white lysozyme Red meat carcasses down to a quarter of a carcass
Electrolytically generated hypochlorous acid In water used in meat processing
Beef primals
Meat carcasses, parts, trim, and organs
An aqueous solution of citric and hydrochloric acids To adjust the acidity in various meat products
A blend of citric acid, hydrochloric acid, In casings and on cooked (RTE) meat products
and phosphoric acid
Hops beta acids In water or ice used for processing meat products;
in water or ice, used as either spray or dip, for meat
(hides on or off)
Hypobromous acid Livestock carcasses prior to fabrication (i.e., pre- and
postchill), offal, and variety meats; beef subprimals and
trimmings; beef heads and tongues
Lactic acid RTE cooked sausages (e.g., frankfurters, bologna, etc.)
and cooked, cured whole muscle products;
nonstandardized comminuted meat products (e.g., beef
patties), ground beef, and raw whole muscle beef cuts
Lactic acid bacteria mixture consisting of Lactobacillus Beef carcasses and parts
acidophilus (NP35, NP51), Lactobacillus lactis (NP7),
and Pediococcus acidilactici (NP3)
Lactoferrin Fresh cuts of meat, nonstandardized RTE comminuted
meat products and standardized RTE comminuted meat
products that permit the use of any safe and suitable
antimicrobial agent
Lauramide arginine ethyl ester (LAE), silicon dioxide, Fresh cuts of meat, nonstandardized RTE comminuted
and rened sea salt meat products and standardized RTE comminuted meat
products that permit the use of any safe and suitable
antimicrobial agent
LAE dissolved at specied concentrations in either RTE meat products; ground beef
propylene glycol, glycerin, or water to which may be
added a polysorbate surface active agent
LAE Cooked, RTE meat products containing sauces;
meat soups; in casings and on cooked (RTE) meat
Nisin preparation Frankfurters and other similar cooked meat sausages
Blend of encapsulated nisin preparation, rosemary extract Cooked (RTE) meat sausages and cured meat products
and salt
Blend of nisin preparation, rosemary extract, salt, Cooked (RTE) meat sausages and cured meat products
maltodextrin, and cultured dextrose

TABLE 1 USDA-FSIS table of safe and suitable antimicrobials for beefa (Continued)
Substance Product
Blend of nisin preparation, rosemary extract, salt, As part of a carcass wash applied prechill
and sodium diacetate
Organic acids (i.e., lactic, acetic, and citric acid) All meat products
Ozone Meat carcasses, parts, trim, and organs
Peroxyacetic acid, octanoic acid, acetic acid, hydrogen Process water for washing, rinsing, cooling, or otherwise
peroxide, peroxyoctanoic acid, and HEDP for processing meat carcasses, parts, trim, and organs
Mixture of peroxyacetic acid, hydrogen peroxide, Process water for washing, rinsing, cooling, or otherwise
acetic acid, and HEDP for processing meat carcasses, parts, trim, and organs
Combination of two aqueous mixtures (FCN 323 and Red meat carcasses, parts, and trim
FCN 880) of peroxyacetic (peracetic) acid, hydrogen
peroxide, acetic acid, and stabilizer HEDP
Aqueous mixture of peroxyacetic acid, Water or ice for washing, rinsing, cooling, or processing
hydrogen peroxide, acetic acid, HEDP, and sulfuric acid whole or cut meat including carcasses, parts, trim,
and organs
Mixture of peroxyacetic acid, hydrogen peroxide, In process water or ice for washing, rinsing, storing,
acetic acid, and HEDP or cooling of processed and preformed meat products
Aqueous mixture of peroxyacetic acid, In process water used for washing, rinsing,
hydrogen peroxide, acetic acid, and HEDP cooling or otherwise processing meat carcasses, parts,
trim, and organs
Aqueous mixture of peroxyacetic acid, In process water or ice used for washing, rinsing, cooling
hydrogen peroxide, acetic acid, and HEDP or processing whole or cut meat including parts, trim,
and organs
Aqueous mixture of peroxyacetic acid, Red meat carcasses, parts, trim, and organs
hydrogen peroxide, HEDP, and optionally sulfuric acid
Aqueous mixture of peroxyacetic acid, Use as a spray, rinse, dip, chiller water, or scald water
hydrogen peroxide, HEDP, dipicolinic acid, for raw meat carcasses, parts, trim, and organs
and sulfuric acid
Mixture of peroxyacetic acid, hydrogen peroxide, acetic Various meat products which permit addition of
acid, HEDP, and water antimicrobial agents, e.g., hot dogs
Potassium diacetate Various RTE meat products, e.g., hot dogs
Solution of water, lactic acid, propionic acid, Raw comminuted beef
and acidic calcium sulfate
Solution of water, acidic calcium sulfate, and lactic acid Raw whole muscle beef cuts and cooked roast beef
and similar cooked beef products (e.g., corned beef,
pastrami, etc.)
Solution of water, acidic calcium sulfate, lactic acid, Beef jerky
and sodium phosphate
Solution of water, acidic calcium sulfate, lactic acid, Meat sausages including those with standards of identity
and sodium which permit the use of antimicrobial agents
Skim milk or dextrose cultured with Propionibacterium RTE meat products that permit the use of any safe and
freudenreichii subsp. Shermanii suitable antimicrobial agent
Sodium benzoate Nonstandardized and standardized comminuted meat
and poultry products which permit ingredients of
this type
Sodium citrate buffered with citric acid RTE meat products that permit the use of any safe and
suitable antimicrobial agent
Sodium diacetate, sodium propionate, Red meat carcasses down to quarter of carcass; water
and sodium benzoate used in meat processing; beef primals

TABLE 1 (Continued)
Substance Product
Sodium hypochlorite Component of marinades used for raw meat products;

raw beef carcasses, subprimals, and trimmings;
RTE meat products
Sodium metasilicate RTE meat, where antimicrobials are permitted
Sodium propionate/propionic acid RTE meat and poultry, where antimicrobials are
permitted
a
For a complete listing, which includes other meats, poultry and eggs, and antimicrobials targeting other specic organisms, please refer
to USDA-FSIS Directive 7120.1 Safe and Suitable Ingredients Used in the Production of Meat, Poultry, and Egg Products, http://www.fsis.
usda.gov/OPPDE/rdad/FSISDirectives/7120.1.pdf.
reduce the presence of E. coli O157:H7 inoc- that the greatest reductions in coliform counts
ulated onto the hide. However, it was re- on the hide sections resulted from treatment
ported that the wash did not significantly with 1% cetylpyridinium chloride (CPC), 2%
reduce the numbers of E. coli O157:H7 trans- LA, and 3% hydrogen peroxide (4.5, 4.1, and
ferred from the hide to the carcass during 3.9 log10 CFU per 100 cm2, respectively). In
slaughter/dressing procedures. However, Nou an investigation of several different potential
et al. (39) demonstrated that prevalence of hide sanitizer treatments, Carlson et al. (44)
E. coli O157:H7 could be significantly reduced reported that 2.4% potassium cyanate, 6.2%
on carcasses by removing bacterial contami- sodium sulfide, and 1.5% sodium hydroxide
nation on the hide before removal. followed by high-pressure washing with 0.02%
Several studies evaluated the effectiveness chlorinated water caused the greatest reduc-
of incorporating sanitizers into hide washes to tions in numbers of E. coli O157:H7, achieving
reduce the potential spread of pathogens from reductions ranging from 4.8 to 5.1 log10 CFU
the hide to carcass surfaces. Mies et al. (40) per cm2.
investigated the implementation of a com- Bosilevac et al. (37) tested a water wash
mercial cattle hide wash system that evaluated plus CPC treatment as a hide intervention
water washes, 0.5% lactic acid (LA), and 50 when applied to cattle in the holding pens of a
ppm chlorine and reported that bacterial num- commercial processing plant. Cattle were
bers actually increased after the treatments. washed with water the day before harvest,
Bosilevac et al. (41) reported that ozonated and before stunning, were sprayed with 1%
and electrolyzed oxidizing water treatments CPC. Hides and carcasses after hide removal
reduced Enterobacteriaceae counts (EBC) by but before evisceration were sampled to de-
3.4 and 4.3 log10 CFU per 100 cm2, respec- termine aerobic plate counts (APC), EBC,
tively. Small et al. (42) compared hide decon- and E. coli O157 prevalence. The prevalence of
tamination treatments and found that the E. coli O157 on hides was reduced by 18%
greatest bacterial reduction (2.3 log10 CFU per (from 56 to 34%) and that on carcasses prior
cm2) could be attained by clipping the hair to evisceration by 20% (from 23 to 3%).
from the hide and then singeing with a On preevisceration carcasses, APC were de-
handheld blowtorch. creased by approximately 77,000 CFU per
A number of studies investigated the effi- 100 cm2 and EBC by approximately 1,150 CFU
cacy of different hide sanitization treatments, per 100 cm2. It was concluded that this treat-
some of which have been commercialized, for ment has great potential and deserves further
reducing E. coli O157:H7 numbers. Baird et al. evaluation.
(43) inoculated beef hide sections with bovine One of the more novel approaches to pre-
fecal slurries and treated them with various vent contamination on the hide from reaching
potential wash sanitizers. The authors reported the carcass surface during slaughter/dressing

was published by Antic et al. (45).The Serbian Castillo et al. (49) found that a chemical
investigators coated cattle hides with a solu- dehairing process significantly reduced APC,
tion of food-grade resin (shellac) in ethanol, TCC, and E. coli, as well as E. coli O157:H7
hypothesizing that the immobilization of bac- and Salmonella enterica serovar Typhimurium
teria on the hide would reduce transmission on artificially inoculated hide pieces. Pieces
of bacteria to the carcass. The shellac-based of hide (4 cm2) were contaminated with bo-
treatment was reported to have successfully vine feces containing both rifampicin-resistant
reduced hide-level E. coli O157:H7 prevalence E. coli O157:H7 and serovar Typhimurium
by 3.7 log10 CFU per cm2. to yield approximately 5.0 log10 CFU/cm2 of
Another novel approach to reduce STEC each pathogen, or with noninoculated feces,
contamination on the hide is the application of which produced an approximate final APC of
bacteriophage (46). One product has received 6.0 log10 CFU/cm2 and a coliform and E. coli
approval in the United States for use on the count of 5.0 log10 CFU/cm2. Counts of path-
hides of live cattle in the holding pens 1 to 4 h ogens, APC, coliforms, and E. coli were con-
before slaughter and hide removal, and spe- ducted before and after chemical dehairing.
cifically targets E. coli O157:H7. Chemical dehairing significantly reduced sero-
var Typhimurium and E. coli O157:H7 popu-
lations from 5.1 to 5.3 log10 CFU/cm2 to <0.5
Chemical Dehairing
log10 CFU/cm2, and reduced APC, coliforms,
Chemical dehairing is a process patented by and E. coli counts by 3.4, 3.9, and >4.3 log10
Bowling and Clayton (47) that involves treat- CFU/cm2, respectively. The authors conclud-
ment of the hide with a sodium sulfide solu- ed that since the hide is a major source of
tion, followed by a hydrogen peroxide solution fecal contamination of beef carcass surfaces,
and water washing. Treatment with sodium chemical dehairing may be beneficial in re-
sulfide solution dissolves and removes hair ducing overall contamination of carcasses.
and extraneous matter from the skin surface, Nou et al. (39) tested the efficacy of chem-
and hydrogen peroxide neutralizes the pH. ical dehairing on reducing the prevalence of
Additional steps, e.g., water rinse prior to the E. coli O157:H7 and other bacteria on the sur-
sodium sulfide or additional neutralization faces of preeviscerated beef carcasses from
steps, may be involved, depending on the which hides had been removed. Hides were
protocol. The first studies evaluating chemical sampled immediately after stunning, before
dehairing as a hide intervention were con- exsanguination or any antimicrobial inter-
ducted by Schnell et al. (48). This study in- vention, to confirm that bacterial loads were
volved 10 grain-fed steers or heifers to be not significantly different between the control
dehaired and 10 controls that were slaughtered and treatment groups. Carcasses were sam-
and dressed without dehairing. Excised hide pled immediately after hide removal and be-
samples from conventional and dehaired car- fore evisceration. Total APC and EBC on hides
casses were analyzed for APC, total coliform in both control and treatment groups were
counts (TCC), and E. coli biotype I counts. not significantly different. Preevisceration car-
Dehairing reduced the amount of visible con- casses processed after chemical dehairing
tamination on beef carcasses, but dehaired had approximately 2 logs lower APC and EBC
cattle had significantly higher TCC (P<0.05) compared with those processed by conven-
and no significant difference in APC or E. coli tional procedures (P<0.0001). In addition,
counts from that of conventionally slaughtered the prevalence of E. coli O157:H7 was lower
cattle. The authors hypothesized that the lack on chemically dehaired (1%) than control
of reduction in APC and E. coli could have (conventionally processed) preevisceration
been the result of aerosol, human, and equip- carcasses (50%, P<0.05). The data indicated
ment contamination in the facility. that chemical dehairing of cattle hides is an

effective intervention to reduce the incidence decontamination of hides also significantly

of hide-to-carcass contamination with E. coli reduces contamination of carcasses during
O157:H7 and potentially other pathogens. processing.
However, although chemical dehairing was The efficacy of hypobromous acid (HOBr)
found to be an effective hide intervention, as a hide intervention was studied (51). Hides
the industry has not considered it feasible to after removal from carcasses at a beef pro-
implement (18). cessing plant were sprayed with 220 or
500 ppm of HOBr. HOBr at a concentration
of 220 ppm significantly reduced bacterial
Alternatives to Chemical
counts (APCs, total coliforms, and E. coli) on
Dehairing and CPC
hides by 2.2 log CFU/100 cm2 (51). HOBr at a
Hide interventions proven to significantly re- concentration of 500 ppm significantly re-
duce carcass contamination in processing duced these bacterial counts by 3.3, 3.7, and
facilities include chemical dehairing, CPC 3.8 log CFU per 100 cm2, respectively, dem-
washing, and a 65C sodium hydroxide wash onstrating a dose effect with HOBr. It was
followed by a water rinse (50). Because CPC is concluded that a HOBr wash would reduce
not yet approved for use in beef processing the pathogen prevalence and concentrations
plants, Bosilevac et al. (50) tested other of spoilage bacteria on hides and decrease the
chemicals and antimicrobial compounds on risk of carcass contamination.
hides that are approved for use in beef pro-
cessing plants, albeit for carcass and boneless
beef trim decontamination. In vitro experi-
CARCASS, PRIMAL, SUBPRIMAL,
ments were conducted on cattle hides to
AND TRIM SURFACES
evaluate 4% trisodium phosphate, 4% phos-
phoric acid, 1.6% sodium hydroxide, and 4%
Water Rinsing
chlorofoam (chlorinated alkaline detergent
containing 1,200 ppm free chlorine at pH 7.0) Empey and Scott (52) reported that washing
as washes. A rinse step, consisting of water or carcasses with cold water reduces their bac-
acidified chlorine, was used following all wash terial populations (13). However, Bell (53)
treatments. These wash treatments reduced found that cold water carcass washing was
hide coliform counts by 1.5 to 2.5 log10 CFU ineffective in removing microbial contamina-
per 100 cm2, and acidified (pH 7.0) chlorine tion and tended to bring about a posterior to
rinses (200 or 500 ppm) further reduced anterior redistribution of microbial contami-
coliforms by 10 to 100 CFU per 100 cm2. Re- nation, resulting in increased counts at fore-
moval of excess liquid by vacuuming treated quarter sites. Patterson (54) found that beef
areas reduced the microbial load by ap- carcasses treated with a steam and hot water
proximately 10 CFU per 100 cm2. An online spray (80 to 96C) for 2 min had significantly
hide-wash cabinet that delivered a sodium reduced bacterial numbers compared to un-
hydroxide wash and a chlorinated (1 ppm) treated carcasses.
water rinse reduced APC and EBC on hides by Hot water treatments of beef carcasses to
2.1 and 3.4 log10 CFU per 100 cm2, respec- reduce E. coli O157:H7 and other bacteria
tively, and reduced the prevalence of E. coli applied through a model carcass spray cabinet
O157 on hides from 44 to 17%. This hide were tested by Castillo et al. (55). Paired hot
washing procedure further resulted in a carcass surface regions with varying fat
reduction of APC and EBC on carcasses characteristics were inoculated with bovine
before evisceration by 0.8 log10 CFU/100 cm2, feces containing 106 CFU bacteria per g. Car-
and E. coli O157 prevalence from 17 to 2%. cass surfaces then were exposed to a warm
These results provided further evidence that water wash followed or not by a hot (95C)

water spray, which raised the carcass surface Castillo et al. (61) evaluated a steam-vacu-
temperature to 82 to 85C in 1 to 2 sec. um system designed to be a spot-cleaning
The effect of time between the application of method for removal of fecal contamination on
feces and water treatment was also evaluated. the surface of carcasses and subsequent re-
Warm water wash followed by hot water duction of E. coli contamination of hot beef
spray provided mean log reductions per cm2 carcasses. The efficacy of accompanying treat-
of 3.7, 2.9, 3.3, and 3.3 of E. coli O157:H7, APC, ments, which included hot (95C) or warm
coliform, and thermotolerant coliform counts, (55C) water, 2% LA spray, or combinations of
respectively. Carcass surface region, but not both methods, was also assessed. In this study,
an increase in time (30 min) before treatment, 0.025 g of bovine fecal material was used as a
affected the efficacy of hot water treatments. vehicle to deliver contaminating bacteria to a
This study also resulted in the conclusion that 5-cm2 area on three specific regions of hot
coliform counts may be used to verify the ef- carcass surfaces, viz., the outside round, bris-
ficacy of hot water interventions used as ket, and clod. These regions were removed
critical control points in a hazard analysis from the rest of the carcass before inoculation,
critical control point (HACCP) system. but the hot carcass temperature was main-
tained by placing the meat in insulated
containers. It was reported that all treatments
Steam and Steam Vacuuming
significantly reduced the numbers of APC,
The use of steam-vacuum systems instead EBC, total coliforms, thermotolerant coli-
of knife trimming for physical removal of forms, and E. coli on beef carcass surfaces.
small areas of fecal contamination from beef However, steam vacuuming alone resulted in
carcasses was approved by the FSIS in 1996 significantly smaller reductions than those ob-
(56). Steam-vacuum systems deliver 82 to tained by a combination of steam vacuuming
88C water via spray nozzles at the carcass with any subsequent sanitizing treatment.
surface while the vacuum removes any loose Steam vacuuming reduced the number of
material. Commercial steam-vacuum systems different indicator organisms tested by ca. 3.0
have been reported to reduce total bacterial log cycles; however, it was also observed that
populations and populations of E. coli O157:H7 the treatment spread the bacterial contami-
on beef carcass surfaces by 3.0 and 5.5 log10 nation to areas of the carcass surface adjacent
CFU/cm2, respectively (57, 58). Steam vac- to the contaminated sites. This relocated
uuming has been implemented in most beef contamination was most effectively reduced
processing plants in the United States at var- by treating the area with a combination of hot
ious stages in the slaughter or dressing process water followed by LA.
(18). Knife trimming and steam vacuuming The application of steam to carcass sur-
of visible contamination in localized areas of faces was shown by Dorsa et al. (57) to be
the carcass surface have been reported to be effective for carcass decontamination, and
useful for pathogen reduction. In addition, these authors, with Frigoscandia, subsequently
application of steam vacuuming is commonly developed commercial cabinets for application
applied to areas of the carcass surface believed of steam, calling the treatment steam pas-
to be hot spots (e.g., hide removal pattern teurization. Phebus et al. (62) designed an
lines). These techniques, however, cannot experimental steam pasteurization chamber
be used efficiently for the entire carcass (59) for laboratory testing and reported reductions
and are intended for spot treatment only. Al- of E. coli O157:H7 and certain other bacterial
though the technology has been reported to be pathogens by 3.4 to 3.7 log cycles on hot beef
successful in reducing carcass contamina- carcass surfaces. However, steam pasteuriza-
tion (57, 58, 60), a report by Castillo et al. (61) tion alone showed no greater reductions than
questioned overall effectiveness. other treatments such as knife trimming or

steam vacuuming. Nutsch et al. (63) conducted to contact the bacterial cells. If bacteria are
evaluations of commercial steam pasteuriza- hidden in small knife cuts or under tissue and
tion application in a beef processing plant. the organic acid cannot contact the cell, the
Carcasses were treated with a preliminary desired antibacterial effect is unlikely.
water wash, followed by passing through air Pittman et al. (72) tested the efficacy of LA
blowers to eliminate excessive humidity that as an initial and secondary subprimal inter-
would favor steam condensation. The car- vention. Sections of chilled beef subprimals
casses were then passed, treated with steam (beef round peeled knuckle and beef brisket
within a chamber, followed by transfer to flats) having 100 cm2 of exposed lean surface
another section of the cabinet where cold were inoculated with E. coli O157:H7, non-
water was applied. After treatment, it was re- O157 STEC, or nonpathogenic (biotype I)
ported that APCs were reduced on carcasses E. coli, the last as surrogates for E. coli
from initial counts of 2.1 to 2.2 log10 CFU/cm2 O157:H7. After 30 min at 4C to allow for
to 0.6 to 0.8 log10 CFU/cm2. Counts of E. coli bacterial attachment, sections were sprayed
were also reduced from original counts of 0.6 with LA in a custom-built spray cabinet. Treat-
to 1.5 log10 CFU/cm2 to undetectable levels ments were applied at 1 of 16 combinations
after 6 or 8 sec of steam treatment. of two LA concentrations (2.0 or 5.0%), two
LA temperatures (22 or 48C), two pressures
(1.03 or 4.83 bar), and two flow rates (0.22 or
Organic Acids and
6.22 liters per min). Sections were allowed to
Miscellaneous Sanitizers
drip for 10 sec and were then vacuum pack-
Various studies evaluated the efficacy of or- aged, sealed, and stored at 4C until bacteria
ganic acids for sanitizing whole carcass sides were enumerated or given a second LA
(13, 60, 64), and over time LA became the treatment 24 h later. The initial application of
most commonly used organic acid for carcass LA spray reduced E. coli surrogate, E. coli
decontamination in commercial practice. O157:H7, and non-O157 STEC counts from
Many processors implemented LA washes on 6.0 CFU per cm2 to 3.6, 4.4, and 4.4 log CFU
preevisceration carcasses and final carcass per cm2, respectively. The second application
sprays before carcasses entered the cooler. further reduced E. coli surrogate, E. coli O157:
Although the most common application of H7, and non-O157 STEC counts to 2.6, 3.2, and
LA is currently on hot carcass surfaces, 3.6 log CFU per cm2, respectively. LA sprays
Castillo et al. (65) reported that the treatment were effective as both the initial and second-
is also effective, although to a lesser degree, ary treatments on beef subprimals in reducing
on chilled carcass surfaces. Kotula and pathogenic and nonpathogenic E. coli, as well
Thelappurate (66) reported that APC and as naturally occurring microflora.
E. coli counts increased more rapidly on un- Fouladkhah et al. (73) tested whether the
treated steaks than on steaks treated with six non-O157 STEC serogroups currently clas-
acetic acid or LA. sified as adulterants in beef by the USDA-FSIS
A concern related to spraying beef car- (O26, O45, O103, O111, O121, and O145) could
casses with organic acid is the reported re- be reduced on beef trimming pieces by LA
sistance of E. coli O157:H7 to low pH (6769). treatments previously shown to be effective
However, several studies indicated that lactic against E. coli O157:H7. Beef trimming sam-
or acetic acid sprays, when applied at 55C, ples weighing approximately 100 g were in-
can effectively reduce levels of Salmonella sp. oculated with approximately 1,000 CFU per
and E. coli O157:H7 (65, 70, 71). Successful cm2 via micropipette and, after time was
reduction of bacteria on meat surfaces by allowed for bacterial attachment (10 min at
using organic acids or other sanitizers re- 4C), were immersed for 30 sec in 5% LA
quires that the sanitizing solution be allowed solutions at 25 or 55C. Treatments resulted in

reductions of 0.5 to 0.9 (25C LA) and 1.0 to none of the treatments would result in their
1.4 (55C LA) log10 CFU per cm2 (P<0.05) for total elimination.
E. coli O157:H7 and non-O157 STEC. It was In an attempt to identify effective and
concluded that the LA treatment used against inexpensive antimicrobial interventions that
E. coli O157:H7 on beef trimmings should also could be used in very small meat plants (i.e.,
be effective against the six non-O157 STEC <10 employees and generating average annual
serogroups. revenue of $2.5 million or less), the relative
Kalchayanand et al. (74) inoculated prerigor effectiveness of eight antimicrobial compounds
beef flanks to determine if antimicrobial inter- (acetic acid, citric acid, LA, peroxyacetic acid,
ventions currently used by the meat industry ASC, chlorine dioxide, sodium hypochlorite,
have a similar effect in reducing non-O157 and aqueous ozone) was tested (75). These
STEC serogroups O26, O45, O103, O111, O121, compounds were applied to beef plate piece
and O145 compared to E. coli O157:H7. The surfaces that had been inoculated with fecal
surfaces of the beef flanks were inoculated slurry containing a pathogen cocktail of STEC
with a high (5 104 CFU per cm2) or low (5 O157:H7, serovar Typhimurium, Campylobac-
101 CFU per cm2) concentration of bacteria ter coli, and Campylobacter jejuni with small,
and given 15 min at room temperature to allow handheld spraying equipment. The relative
for bacterial attachment. After inoculation, antimicrobial effectiveness from greatest to
flanks were subjected to a 15-sec spray treat- least was as follows: organic acids, peroxyacetic
ment with one of the following FSIS-approved acid, chlorinated compounds, and aqueous
treatments using a model spray cabinet with ozone. A 2% LA rinse provided 3.5- to 6.4-log
three oscillating spray nozzles at 60 cycles CFU/cm2 reductions across all four bacterial
per min: acidified sodium chlorite (ASC) populations studied.
(1,000 ppm), peroxyacetic acid (200 ppm), LA The reduction of E. coli O157:H7 and sero-
(4%), or hot water (85C). Surviving bacterial var Typhimurium on beef carcass surfaces
concentrations were determined within 10 min through application of ASC solutions was in-
after treatment or after 48 h storage at 2 to vestigated and reported by Castillo et al. (61).
4C. Against both high- and low-level inocu- When phosphoric acid was used to acidify
lation, hot water, LA, peroxyacetic acid, and sodium chlorite, the resulting ASC solution
ASC ranked in this order as the most to least reduced populations of both pathogens by 3.8
effective. While hot water and LA (reductions to 3.9 log cycles. However, when ASC solu-
of 3.2 to 4.2 and 1.6 to 2.7 log CFU per cm2, tions were prepared through acidification
respectively, for nonchilled specimens) were with citric acid, reductions were reported to
effective against all STEC strains (P 0.05), range from 4.5 to 4.6 log cycles.
peroxyacetic acid was not effective against Carvacrol is a monoterpenoid phenol pres-
O111, and ASC was not effective against O26, ent as an essential oil in Origanum vulgare
O111, and O145. Bacterial reductions with ASC (oregano) and several other plants and has
were increased by storage at 4C for 48 h; been shown to have activity against E. coli
hence, it was concluded that this compound O157:H7 in edible apple films (76). A study
might be a long-acting microbial inhibitor and was conducted to determine whether carva-
suitable as a prepackaged meat intervention. crol has activity against E. coli O157 on cattle
Storage at 4C provided little additional re- hides and beef carcass cuts (77). In this study,
duction of pathogen levels with the other carvacrol was sprayed onto hides and beef
treatments. The levels of reduction of non- carcass cuts at concentrations of 0, 10, 20, and
O157 STEC achieved by these antimicrobial 30 mg per ml. These surfaces were then in-
interventions were comparable to that of oculated with E. coli O157 at a concentration
E. coli O157:H7; however, that low levels of of 5 to 6 log10 CFU per cm2, with 10 min
pathogens were still detectable indicated that allowed for bacterial attachment. After this

time, the hide and carcass cut surfaces were types of meat surfaces to which they are ap-
swabbed and cultured, and surviving E. coli plied during investigations, or that the results
O157 concentrations were determined. E. coli may be a factor of the surface microflora
O157 concentrations were reduced on carcass composition as influenced by prior antimi-
cuts and hides treated with carvacrol at 30 mg crobial treatments. King et al. (80) reported
per ml by approximately 1.4 and 1.6 log10 CFU that peroxyacetic acid concentrations up to
per cm2, respectively (P<0.05). It was con- 600 ppm were ineffective for antimicrobial
cluded that carvacrol has the potential to treatment of chilled inoculated beef carcass
control E. coli O157 on bovine hide and carcass surfaces.
cuts, but should be further studied. Elramady et al. (81), using in vitro broth
culture and cattle hide model experiments,
studied the efficacy of chitosan acetate (CA),
Bacteriophage
sodium dodecyl sulfate (SDS), and LA as
A bacteriophage product has been approved individual treatments and in combination
for use on red meat trim and parts intended to against E. coli O157:H7. CA is a naturally oc-
be ground (47). This product was reported to curring substance with demonstrated anti-
eliminate 95 to 100% of E. coli O157:H7 when bacterial properties, and SDS is a food additive
sprayed on the surfaces of beef (47). that has been shown to enhance the antibac-
terial effects of organic acids, hence the reason
for being tested in this study. CA as a treat-
Combination Treatments
ment consisted of 1% chitosan in 1% acetic
Reports have indicated that a combination of acid solution, and the SDS treatment consisted
treatments, also known as a multiple hurdle of a 1% or 2% solution. LA was applied as a 1%
treatment (14, 55, 62), may be required during solution, and the CA-SDS treatment consisted
processing to reduce pathogen contamination. of 1% chitosan in 1% acetic acid mixed with 1%
Numerous carcass decontamination methods SDS. LA-SDS treatments were tested in two
have been investigated alone and in combi- different concentrations, viz., 1% LA mixed
nation for their ability to reduce pathogens on with 1% SDS, and 1% LA mixed with 2% SDS.
meat; however, results and conclusions are In the in vitro broth experiments, all treat-
varied and often contradictory. For example, ments resulted in a significant reduction in
Gill and Landers (78) reported that spraying survival of E. coli O157:H7. However, only 1%
beef carcasses with 2% LA, steam-vacuuming, LA plus 1% SDS and 1% LA plus 2% SDS
or trimming was ineffective, and that only treatments resulted in significant reductions
steam or hot water treatments substantially of E. coli O157:H7 on hides. These treat-
reduced bacterial contamination. The same ments resulted in 4.6 and 4.7 log10 CFU per
authors indicated that a 200-ppm peroxyacetic cm2 reductions, respectively, compared to
acid carcass spray was likely also ineffective, phosphate buffer control treated hides, which
but efficacy was difficult to determine due to a had 6.0 log10 CFU per cm2 surviving cells. The
subsequent steam treatment. antibacterial efficacy of 1% LA was signifi-
Gill and Badoni (79) evaluated the effects of cantly enhanced when combined with 1%
0.02% peroxyacetic acid, 0.16% ASC, 2% LA, SDS. A low-concentration LA-SDS combina-
and 4% LA on chilled beef surfaces and de- tion treatment as a wash may potentially re-
termined that peroxyacetic acid and ASC produce the risk of E. coli O157:H7 contamination
duced a negligible effect on coliforms or on hides.
E. coli, and both treatments were less effective Signorini and Tarabla (82) used a stochas-
than 4% LA. The authors surmised that eval- tic simulation model to assess the effects of
uation of antimicrobial treatments may pro- measures implemented in the agri-food chain
duce inconsistent results due to different to reduce the contamination of ground beef

with STEC. A published risk assessment HACCP SYSTEMS AND VALIDATION

model developed in Argentina was used as
baseline scenario. Control measures assessed All U.S. establishments producing raw beef
were based on a reduction in herd preva- products are required to use HACCP to con-
lence of infection due to vaccination, reduc- trol contamination with food-borne pathogens
tion in opportunity for cross-contamination in such as the STEC strains that have been de-
the slaughterhouses by the introduction of clared adulterants by the USDA-FSIA. To
an online hide-wash cabinet, and control of meet these requirements, slaughter establish-
storage temperature in slaughterhouses, retail ments have implemented a variety of carcass
stores, and home. Additionally, the increase of decontamination procedures such as critical
feedlot production was modeled. Simulations control points (CCPs) that are essential to
suggested that the greatest potential impact the safety of beef. In the last quarter of 2002,
was associated with hide-wash cabinet and the USDA-FSIS issued a notice reminding
vaccination, measures aimed to reduce the slaughter establishments that all CCPs must
STEC prevalence and concentration in the be validated to ensure they can successfully
cattle hides at the beginning of the food chain. prevent, eliminate, or reduce STEC O157:H7.
Control of storage temperature was not ef- The regulatory agency indicated that until
fective if cross-contamination of the carcasses establishments have collected data to dem-
with the pathogen was not prevented or re- onstrate that CCPs function properly under
duced. An increased production (fattening) of actual in-plant conditions, the effectiveness of
cattle in feedlots may raise the risk of STEC the CCP would be considered theoretical and
infection and its sequelae. This information not validated. The FSIS also noted that many
can be used as a basis for measures of risk establishments have not validated CCPs based
management. on actual in-plant conditions.
In some cases, the combination of treat- Microbiological testing can play a unique
ments has not afforded any greater efficacy role in verification and validation activities.
than the component treatments applied indi- Detection of food-borne pathogens is not
vidually. For example, Bosilevac et al. (un- considered to be an effective tool for moni-
published data) reported that E. coli O157 toring CCPs within a slaughter or processing
prevalence was reduced by 81% and 35% by HACCP plan. Pathogens are often absent from
hot water and LA treatments, respectively, but carcass surfaces and, when present, their un-
only 79% by both treatments combined. even distribution makes it difficult to obtain a
Numerous options are available for de- truly representative sample. In contrast, mi-
contaminating meat surfaces; however, none crobiological testing can be applied within a
of the approved interventions are capable of HACCP plan to validate and verify the effec-
eliminating the presence of pathogens. Fre- tiveness of carcass decontamination procedures.
quently, processors in the meat industry use Because of the difficulty in consistently
several redundant intervention technologies finding and documenting reductions of levels
in an attempt to decrease risk of pathogen of enteric pathogens on carcass surfaces, an
contamination, but a guarantee of complete ideal solution can be the use of nonpathogenic
absence of bacterial contamination is not surrogate bacteria that are capable of indi-
possible (15, 16). Proper end-user handling of cating the probable reduction of pathogens.
meat products is required for assurance of Surrogate bacteria are required to have very
safety. Research continues to seek novel inter- similar growth and resistance characteristics
ventions in an attempt to assist meat pro- to the pathogens of concern. After known
cessors minimize or eliminate enteric origin amount of the surrogate is inoculated to a
pathogens, as well as unknown and emerging carcass surface, the effectiveness of a CCP can
food-borne pathogens. be validated by comparing surviving levels of

the surrogate on the carcass surface following Serotypes, virulence genes, and intimin types of
the processing step. Shiga toxin (verotoxin)-producing Escherichia
coli isolates from cattle in Spain and identifica-
Although verification and validation of tion of a new intimin variant gene (eae-xi). J Clin
HACCP systems may initially seem intimi- Microbiol 42:645651.
dating, careful thought and planning can make 4. Gyles CL. 2007. Shiga toxin-producing Esche-
the process logical, reasonable, and extremely richia coli: an overview. J Anim Sci 85:E45E62.
helpful. Assistance is available through many 5. Moxley RA. 2004. Escherichia coli O157:H7: an
update on intestinal colonization and virulence
tools, such as rapid microbiological tests, ex-
mechanisms. Anim Health Res Rev 5:1533.
tensive publication of research results in the 6. Brooks JT, Sowers EG, Wells JG, Greene KD,
scientific literature, and numerous HACCP Griffin PM, Hoekstra RM, Strockbine NA.
experts. The human tendency is to find a 2005. Non-O157 Shiga toxin-producing Esche-
single tool that works and use it to excess; richia coli infections in the United States, 1983
2002. J Infect Dis 192:14221429.
however, successful verification and valida-
7. Moxley RA, Smith DR. 2010. Attaching-effacing
tion will most likely be attained through the Escherichia coli infections in cattle. Vet Clin North
utilization of as many of the tools as possi- Am Food Anim Pract 26:2956.
ble. Regular challenging of the validity of a 8. Cray WC, Moon HW. 1995. Experimental infec-
HACCP system through verification will only tion of calves and adult cattle with Escherichia
serve to strengthen confidence in the ability of coli O157:H7. Appl Environ Microbiol 61:1586
1590.
the process to control hazards.
9. Dean-Nystrom EA, Bosworth BT, Cray WC,
Moon HW. 1997. Pathogenicity of Escherichia
coli O157:H7 in the intestines of neonatal calves.
ACKNOWLEDGMENTS Infect Immun 65:18421848.
This project was supported by Agriculture 10. Matthews L, McKendrick IJ, Ternent H, Gunn
GJ, Synge B, Woolhouse MEJ. 2006. Super-
and Food Research Initiative Competitive
shedding cattle and the transmission dynamics of
Grant no. 2012-68003-30155 from the USDA Escherichia coli O157. Epidemiol Infect 134:131
National Institute of Food and Agriculture. 142.
We declare no conflicts of interest with 11. Omisakin F, MacRae M, Ogden ID, Strachan
regard to the manuscript. NJC. 2003. Concentration and prevalence of
Escherichia coli O157 in cattle feces at slaughter.
Appl Environ Microbiol 69:24442447.
CITATION 12. Cobbold RN, Hancock DD, Rice DH, Berg
J, Stilborn R, Hovde CJ, Besser TE. 2007.
Moxley RA, Acuff GR. 2014. Peri- and post- Rectoanal junction colonization of feedlot cattle
harvest factors in the control of Shiga toxin- by Escherichia coli O157:H7 and its association
producing Escherichia coli in beef. Microbiol with supershedders and excretion dynamics. Appl
Environ Microbiol 73:15631568.
Spectrum 2(4):EHEC-0017-2013.
13. Dickson JS, Anderson ME. 1992. Microbiological
decontamination of food animal carcasses by
washing and sanitizing systems: a review. J Food
REFERENCES Prot 55:133140.
1. Painter JA, Hoekstra RM, Ayers T, Tauxe RV, 14. Arthur TM, Bosilevac JM, Nou X, Shackelford
Braden CR, Angulo FJ, Griffin PM. 2013. Attri- SD, Wheeler TL, Kent MP, Jaroni D, Pauling B,
bution of foodborne illnesses, hospitalizations, Allen DM, Koohmaraie M. 2004. Escherichia
and deaths to food commodities by using out- coli O157 prevalence and enumeration of aerobic
break data, United States, 19982008. Emerg bacteria, Enterobacteriaceae, and Escherichia
Infect Dis 19:407415. coli O157 at various steps in commercial beef
2. Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV, processing plants. J Food Prot 67:658665.
Widdowson MA, Roy SL, Jones JL, Griffin PM. 15. Barkocy-Gallagher GA, Arthur TM, Rivera-
2011. Foodborne illness acquired in the United Betancourt M, Nou X, Shackelford SD, Wheeler
Statesmajor pathogens. Emerg Infect Dis 17:715. TL, Koohmaraie M. 2003. Seasonal prevalence of
3. Blanco M, Blanco JE, Mora A, Dahbi G, Alonso Shiga toxin-producing Escherichia coli, including
MP, Gonzlez, Bernrdez MI, Blanco J. 2004. O157:H7 and non-O157 serotypes, and Salmonella

in commercial beef processing plants. J Food Prot 26. Arthur TM, Bosilevac JM, Brichta-Harhay DM,
66:19781986. Guerini MN, Kalchayanand N, Shackelford SD,
16. Elder RO, Keen JE, Siragusa GR, Barkocy- Wheeler TL, Koohmaraie M. 2007. Transporta-
Gallagher GA, Koohmaraie M, Laegreid WW. tion and lairage environment effects on preva-
2000. Correlation of enterohemorrhagic Esche- lence, numbers and diversity of Escherichia coli
richia coli O157 prevalence in feces, hides, and O157:H7 on hides and carcasses of beef cattle at
carcasses of beef cattle during processing. Proc processing. J Food Prot 70:280286.
Natl Acad Sci USA 97:29993003. 27. Bach SJ, McAllister TA, Mears GJ, Schwartzkopf-
17. Keen JE, Elder RO. 2002. Isolation of Shiga- Genswein KS. 2004. Long-haul transport and lack
toxigenic Escherichia coli O157 from hide surfaces of preconditioning increases fecal shedding of
and the oral cavity of finished beef feedlot cattle. Escherichia coli O157:H7 by calves. J Food Prot 67:
J Am Vet Med Assoc 220:756763. 672678.
18. Koohmaraie M, Arthur TM, Bosilevac JM, 28. Barham AR, Barman BL, Johnson AK, Allen
Guerini M, Shackelford SD, Wheeler TM. 2005. DM, Blanton JR Jr, Miller MF. 2002. Effects on
Post-harvest interventions to reduce/eliminate the transportation of beef cattle from the feedyard
pathogens in beef. Meat Sci 71:7991. to the packing plant on prevalence levels of
19. Schmidt JW, Arthur TM, Bosilevac JM, Escherichia coli O157 and Salmonella spp. J Food
Kalchayanand N, Wheeler TL. 2012. Detection Prot 65:280283.
of Escherichia coli O157:H7 and Salmonella 29. Fegan N, Higgs G, Duffy LL, Barlow RS. 2009.
enterica in air and droplets at three U.S. com- The effects of transport and lairage on counts of
mercial beef processing plants. J Food Prot Escherichia coli O157 in the feces and on the hides
75:22132218. of individual cattle. Foodborne Pathog Dis 6:1113
20. Barkocy-Gallagher GA, Arthur TM, Siragusa 1120.
GR, Keen JE, Elder RO, Laegreid WW, 30. Minihan D, Whyte P, OMahony M, Clegg T,
Koohmaraie M. 2001. Genotypic analyses of Collins JD. 2003. An investigation on the effect of
Escherichia coli O157:H7 and O157 nonmotile transport and lairage on the faecal shedding
isolates recovered from beef cattle and carcasses prevalence of Escherichia coli O157 in cattle.
at processing plants in the Midwestern states of J Med Vet Ser B 50:378382.
the United States. Appl Environ Microbiol 31. Stanford K, Bryan M, Peters J, Gonzlez LA,
67:38103818. Stephens TP, Schwartzkopf-Genswein. 2011.
21. Arthur TM, Barkocy-Gallagher GA, Rivera- Effects of long- or short-haul transportation of
Betancourt M, Koohmaraie M. 2002. Prevalence slaughter heifers and cattle liner microclimate
and characterization of non-O157 Shiga toxin- on hide contamination with Escherichia coli O157.
producing Escherichia coli on carcasses in com- J Food Prot 74:16051610.
mercial beef cattle processing plants. Appl Envi- 32. Federal Register. 1999. Beef products con-
ron Microbiol 68:48474852. taminated with Escherichia coli O157:H7. U.S.
22. Collis VJ, Reid CA, Hutchison ML, Davies MH, Department of Agriculture, Food Safety and In-
Wheeler KPA, Small A, Buncic S. 2004. Spread spection Service, Code of Federal Regulations.
of marker bacteria from the hides of cattle in a January 19, 1999. Fed Regist 64:28032805.
simulated livestock market and at an abattoir. J 33. Federal Register. 2011. Shiga toxin-producing
Food Prot 67:23972402. Escherichia coli in certain raw beef products. U.S.
23. Dewell GA, Simpson CA, Dewell RD, Hyatt DR, Department of Agriculture, Food Safety and
Belk KE, Scanga JA, Morley PS, Grandin T, Inspection Service, Code of Federal Regula-
Smith GC, Dargatz DA, Wagner BA, Salman tions. September 20, 2011. Fed Regist 76:58157
MD. 2008. Impact of transportation and lairage 58165.
on hide contamination with Escherichia coli O157 34. Brusa V, Aliverti V, Aliverti F, Ortega EE, de la
in finished beef cattle. J Food Prot 71:11141118. Torre JH, Linares LH, Sanz ME, Etcheverria
24. Mather AE, Reid SW, McEwen SA, Ternent AI, Padola NL, Galli L, Peral Garcia P, Copes J,
HE, Reid-Smith RJ, Boerlin P, Taylor DJ, Steele Leotta GA. 2013. Shiga toxin-producing Esche-
WB, Gunn GJ, Mellor DJ. 2008. Factors associ- richia coli in beef retail markets from Argentina.
ated with cross-contamination of hides of Scottish Front Cell Infect Microbiol 2:16.
cattle by Escherichia coli O157. Appl Environ 35. Kalchayanand N, Arthur TM, Bosilevac JM,
Microbiol 74:63136319. Wells JE, Wheeler TL. 2013. Chromogenic agar
25. Stanger KJ, Ketheesan N, Parker AJ, Colman medium for detection and isolation of Escherichia
CJ, Lazzaroni SM, Fitzpatrick LA. 2005. The coli serogroups O26, O45, O103, O111, O121, and
effect of transportation on the immune status of O145 from fresh beef and cattle feces. J Food Prot
Bos indicus steers. J Anim Sci 83:26322636. 76:192199.

36. Bosilevac JM, Koohmaraie M. 2011. Prevalence 46. Sillankorva SM, Oliveira H, Azeredo J. 2012.
and characterization of non-O157 Shiga toxin- Bacteriophages and their role in food safety. Int J
producing Escherichia coli isolates from com- Microbiol 2012:863945.
mercial ground beef in the United States. Appl 47. Bowling RA, Clayton RP. 1992. Method for
Environ Microbiol 77:21032112. dehairing animals. U.S. patent 5,149,295.
37. Bosilevac JM, Arthur TM, Wheeler TL, 48. Schnell TD, Sofos JN, Littlefield VG, Morgan
Shackelford SD, Rossman M, Reagan JO, JB, Gorman BM, Clayton RP, Smith GC. 1995.
Koohmaraie M. 2004. Prevalence of Escherichia Effects of postexanguination dehairing on the
coli O157 and levels of aerobic bacteria and En- microbial load and visual cleanliness of beef
terobacteriaceae are reduced when hides are carcasses. J Food Prot 58:12971302.
washed and treated with cetylpyridinium chlo- 49. Castillo A, Dickson JS, Clayton RP, Lucia LM,
ride at a commercial beef processing plant. J Food Acuff GR. 1998. Chemical dehairing of bovine
Prot 67:646650. skin to reduce pathogenic bacteria and bacteria of
38. Byrne CM, Bolton DJ, Sheridan JJ, McDowell fecal origin. J Food Prot 61:623625.
DA, Blair IS. 2000. The effects of preslaughter 50. Bosilevac JM, Nou X, Osborn MS, Allen DM,
washing on the reduction of Escherichia coli Koohmaraie M. 2005. Development and evalua-
O157:H7 transfer from cattle hides to carcasses tion of an on-line hide decontamination proce-
during slaughter. Lett Appl Microbiol 30:142 dure for use in a commercial beef processing
145. plant. J Food Prot 68:265272.
39. Nou X, Rivera-Betancourt M, Bosilevac JM, 51. Schmidt JW, Wang R, Kalchayanand N,
Wheeler TL, Shackelford SD, Gwartney BL, Wheeler TL, Koohmaraie. 2012. Efficacy of
Reagan JO, Koohmaraie M. 2003. Effect of hypobromous acid as a hide-on carcass antimi-
chemical dehairing on the prevalence of Esche- crobial intervention. J Food Prot 75:955958.
richia coli O157:H7 and the levels of aerobic bac- 52. Empey WA, Scott WJ. 1939. Investigation of
teria and Enterobacteriaceae on carcasses in a chilled beef. Part 1. Microbial contamination ac-
commercial beef processing plant. J Food Prot quired in the meat works. Council for Sci. and
66:20052009. Indus. Res. Bull. No. 126, Melbourne, Australia.
40. Mies PD, Covington BR, Harris KB, Lucia LM, 53. Bell RG. 1997. Distribution and sources of mi-
Acuff GR, Savell JW. 2004. Decontamination of crobial contamination on beef carcasses. J Appl
cattle hides prior to slaughter using washes with Microbiol 82:292300.
and without antimicrobial agents. J Food Prot 54. Patterson JT. 1969. Hygiene in meat process-
67:579582. ing plants. Rec Agric Res Minist Agric NI 18:85
41. Bosilevac JM, Shackelford SD, Brichta DM, 87.
Koohmaraie M. 2005. Efficacy of ozonated and 55. Castillo A, Lucia LM, Goodson KJ, Savell JW,
electrolyzed oxidative waters to decontaminate Acuff GR. 1998. Use of hot water for beef carcass
hides of cattle before slaughter. J Food Prot decontamination. J Food Prot 61:1925.
68:13931398. 56. Federal Register. 1996. Notice of policy change;
42. Small A, Wells-Burr B, Buncic S. 2005. An achieving the zero tolerance performance stan-
evaluation of selected methods for the decon- dard for beef carcasses by knife trimming and
tamination of cattle hides prior to skinning. Meat vacuuming with hot water or steam; use of
Sci 69:263268. acceptable carcass interventions for reducing
43. Baird BE, Lucia LM, Acuff GR, Harris KB, carcass contamination without prior agency ap-
Savell JW. 2006. Beef hide antimicrobial inter- proval. U.S. Department of Agriculture, Food
ventions as a means of reducing bacterial con- Safety and Inspection Service. Fed Regist 61:
tamination. Meat Sci 73:245248. 1502415027.
44. Carlson BA, Ruby J, Smith GC, Sofos JN, 57. Dorsa WJ, Cutter CN, Siragusa GR. 1996. Ef-
Bellinger GR, Warren-Serna W, Centrella B, fectiveness of a steam-vacuum sanitizer for re-
Bowling RA, Belk KE. 2008. Comparison of ducing Escherichia coli O157:H7 inoculated to beef
antimicrobial efficacy of multiple beef hide de- carcass surface tissue. Lett Appl Microbiol 23:61
contamination strategies to reduce levels of 63.
Escherichia coli O157:H7 and Salmonella. J Food 58. Dorsa WJ, Cutter CN, Siragusa GR,
Prot 71:22232227. Koohmaraie M. 1996. Microbial decontamination
45. Antic D, Blagojevic B, Ducic M, Mitrovic R, of beef and sheep carcasses by steam, hot water
Nastasijevic I, Buncic S. 2010. Treatment of spray washes, and a steam-vacuum sanitizer.
cattle hides with shellac-in-ethanol solution to J Food Prot 59:127135.
reduce bacterial transferabilitya preliminary 59. Dorsa WJ, Cutter CN, Siragusa GR. 1997. Effects
study. Meat Sci 85:7781. of acetic acid, lactic acid and trisodium phosphate

on the microflora of refrigerated beef carcass 72. Pittman CI, Geornaras I, Woerner DR, Night-
surface tissue inoculated with Escherichia coli ingale KK, Sofos JN, Goodridge L, Belk KE.
O157:H7, Listeria innocua, and Clostridium sporo- 2012. Evaluation of lactic acid as an initial and
genes. J Food Prot 60:619624. secondary subprimal intervention for Escherichia
60. Dorsa WJ. 1997. New and established carcass coli O157:H7, non-O157 Shiga toxin-producing E.
decontamination procedures commonly used in coli, and a nonpathogenic E. coli surrogate for E.
the beef-processing industry. J Food Prot 60: coli O157:H7. J Food Prot 75:17011708.
11461151. 73. Fouladkhah A, Geornaras I, Yang H, Belk KE,
61. Castillo A, Lucia LM, Goodson KJ, Savell JW, Nightingale KK, Woerner DR, Smith GC, Sofos
Acuff GR. 1999. Decontamination of beef carcass JN. 2012. Sensitivity of Shiga toxin-producing
surface tissue by steam vacuuming alone and Escherichia coli, multidrug-resistant Salmonella,
combined with hot water and lactic acid sprays. and antibiotic-susceptible Salmonella to lactic
J Food Prot 62:146151. acid on inoculated beef trimmings. J Food Prot
62. Phebus RK, Nutsch AL, Schafer DE, Wilson RC, 75:17511758.
Riemann MJ, Leising JD, Kastner CL, Wolf JR, 74. Kalchayanand N, Arthur TM, Bosilevac JM,
Prasai RK. 1997. Comparison of steam pasteuri- Schmidt JW, Wang R, Shackelford SD, Wheeler
zation and other methods for reduction of path- TL. 2012. Evaluation of commonly used antimi-
ogens on surfaces of freshly slaughtered beef. crobial interventions for fresh beef inoculated
J Food Prot 60:476484. with Shiga toxin-producing Escherichia coli
63. Nutsch AL, Phebus RK, Riemann MJ, Schafer serotypes O26, O45, O103, O111, O121, O145, and
DE, Boyer JE, Wilson CR, Leising JD, Kastner O157:H7. J Food Prot 75:12071212.
CL. 1997. Evaluation of a steam pasteurization 75. Yoder SF, Henning WR, Mills EW, Doores S,
process in a commercial beef processing facility. Ostiguy N, Cutter CN. 2012. Investigation of
J Food Prot 60:485492. chemical rinses suitable for very small meat
64. Siragusa GR. 1995. The effectiveness of carcass plants to reduce pathogens on beef surfaces. J
decontamination systems for controlling the pre- Food Prot 75:1421.
sence of pathogens on the surfaces of meat animal 76. Du WX, Olsen CW, Avena-Bustillos RJ,
carcasses. J Food Saf 15:229238. McHugh TH, Levin CE, Friedman M. 2008.
65. Castillo A, Lucia LM, Roberson DB, Stevenson Storage stability and antibacterial activity against
TH, Mercado I, Acuff GR. 2001. Lactic acid Escherichia coli O157:H7 of carvacrol in edible
sprays reduce bacterial pathogens on cold beef apple films made by two different casting meth-
carcass surfaces and in subsequently produced ods. J Agric. Food Chem 56:30823088.
ground beef. J Food Prot 64:5862. 77. McDonnell MJ, Rivas L, Burgess CM, Fanning
66. Kotula KL, Thelappurate R. 1994. Microbiolog- S, Duffy G. 2012. Evaluation of carvacrol for the
ical and sensory attributes of retail cuts of beef control of Escherichia coli O157 on cattle hide and
treated with acetic and lactic solutions. J Food carcass cuts. Foodborne Pathog Dis 9:10491052.
Prot 57:665670. 78. Gill CO, Landers C. 2003. Microbiological effects
67. Cheng CM, Kaspar CW. 1998. Growth and of carcass decontaminating treatments at four
processing conditions affecting acid tolerance in beef packing plants. Meat Sci 65:10051011.
Escherichia coli O157:H7. Food Microbiol 15:157 79. Gill CO, Badoni M. 2004. Effects of peroxyacetic
166. acid, acidified sodium chlorite or lactic acid so-
68. Padhye NV, Doyle MP. 1992. Escherichia lutions on the microflora of chilled beef carcasses.
coli O157:H7: epidemiology, pathogenesis, and Int J Food Microbiol 91:4350.
methods for detection in foods. J Food Prot 55: 80. King DA, Lucia LM, Castillo A, Acuff GR,
555565. Harris KB, Savell JW. 2005. Evaluation of
69. Wang G, Doyle MP. 1998. Heat shock response peroxyacetic acid as a post-chilling intervention
enhances acid tolerance of Escherichia coli O157: for control of Escherichia coli O157:H7 and Sal-
H7. Lett Appl Microbiol 26:3134. monella Typhimurium on beef carcass surfaces.
70. Castillo A, Lucia LM, Goodson KJ, Savell JW, Meat Sci 69:401407.
Acuff GR. 1998. Comparison of water wash, 81. Elramady MG, Aly SS, Rossitto PV, Crook JA,
trimming, and combined hot water and lactic acid Cullor JS. 2013. Synergistic effects of lactic acid
treatments for reducing bacteria of fecal origin on and sodium dodecyl sulfate to decontaminate
beef carcasses. J Food Prot 61:823828. Escherichia coli O157:H7 on cattle hide sections.
71. Hardin MD, Acuff GR, Lucia LM, Oman JS, Foodborne Pathog Dis 10:661663.
Savell JW. 1995. Comparison of methods for 82. Signorini ML, Tarabla HD. 2010. Interventions
contamination removal from beef carcass sur- to reduce verocytotoxigenic Escherichia coli in
faces. J Food Prot 58:368374. ground beef in Argentina. Prev Vet Med 94:3642.

Veterinary Public Health Approach
to Managing Pathogenic
Verocytotoxigenic Escherichia coli
in the Agri-Food Chain
GERALDINE DUFFY1 and EVONNE McCABE1

23
INTRODUCTION
It is well documented that animals and, in particular, ruminants can carry a

range of potentially harmful pathogens, including verocytoxigenic Escherichia
coli (VTEC), in their gastrointestinal tract. VTEC can reportedly survive for
several months in the environment, in feces, and in soil, which allows for the
recycling of VTEC among food animals and wildlife and prolonged environ-
mental contamination. VTEC contamination of fresh produce may arise from
irrigation water, manure or compost applied to soil as a fertilizer, and feces of
wildlife or farmed animals. Table 1 summarizes some of the diverse VTEC
outbreaks over the past few years (20062013). Of significant interest is that
apart from the newly emerging vehicles of infection, serogroups other than O157
are increasingly causing outbreaks, many of which have severe outcomes with
cases of hemolytic-uremic syndrome (HUS) and fatalities. While outbreaks are
important and gain notoriety, the contribution of sporadic cases of human
VTEC infection cannot be ignored. The data show that although foods of animal
origin such as meat and dairy products and fresh produce such as salads and
vegetables are well recognized important vectors of infection, there have also
been VTEC outbreaks related to direct contact with fecal matter through rec-
reational activities including visiting petting zoos, attending agricultural fairs,
1
Teagasc Food Research Centre, Ashtown, Dublin 15, Ireland.
457

458 DUFFY AND McCABE
TABLE 1 Selected VTEC outbreaks (20062013) highlighting different vehicles of infection and serogroups
No. of No. of No. of Likely source or
Country Reference(s) Year Serotype cases HUS cases deaths mode of transmission
United Kingdom Launders et al., 2013 (142) 2013 O157 19 0 0 Watercress

Denmark Soborg et al., 2013 (143) 2012 O157 13 8 0 Ground beef
United States Slayton et al., 2013 (144) 2012 O157 58 3 0 Romaine lettuce
Germany Karch et al., 2012 (10) 2011 O104 4,321 852 32 Sprouts
United States McCollum et al., 2010 (104) 2010 O157 41 1 0 Cheese Gouda
United States Taylor et al., 2013 (145) 2010 O145 33 3 0 Romaine lettuce
Netherlands Greenland et al., 2009 (146) 2009 O157 20 0 0 Steak tartare
United States Neil et al., 2012 (147) 2009 O157 72 10 0 Cookie dough
United States CDC, 2010 (148) 2008 O157 99 1 0 Ground beef
Ireland OSullivan et al., 2008 (89) 2008 O157 148 NAa NA Private wells
drinking water
Belgium De Schrijver et al., 2008 O26 & 12 3 0 Homemade
2008 (107) O145 ice cream, fecal
material from farm
Netherlands/ Friesema et al., 2008 (149) 2007 O157 50 0 0 Lettuce
Iceland
Denmark Ethelberg et al., 2009 (150) 2007 O26 20 0 0 Fermented beef
sausage
United States Wendel et al., 2009 (151); 2006 O157 199 31 3 Fresh spinach and
Jay et al., 2007 (69) feral pigs
a
Data not available.
and swimming in contaminated water (1). human illness. The diversity of serogroups
With so many routes of potential transmission, and virulence factors among human-disease-
it is clear that the management of this patho- causing VTEC strains poses considerable chal-
gen requires a multidisciplinary approach with lenges in assessing the risk posed by VTEC
cooperation among the disciplines of human recovered from the agri-food chain. The
medicine, animal and veterinary sciences, and dominant pathogenic VTEC serotype remains
environmental and food scientists. A veterinary E. coli O157:H7 (2, 3) and has the strongest
public health approach to managing VTEC association with HUS worldwide (4). How-
focuses on collation of data on prevalence of ever, a common pattern being observed in
VTEC in different sources, the importance of the European Union and the United States (5)
each source as a reservoir of human pathogen is the increasing number of reported cases
VTEC, and risk factors for transmission to now attributed to non-O157 serotypes. Ap-
humans. proximately half of all confirmed VTEC cases
This article provides an overview of the in the European Union are now associated
risks for transmission of human pathogenic with serogroup O157 (6), and in the United
VTEC from the various routes of transmission, States in the period 2000 to 2010, O157
including animals, the farm environment, and cases decreased from 2.17 to 0.95/100,000
production of milk and meat, and highlights whereas non-O157 cases increased from 0.12
general management approaches to reduce the to 0.95/100,000 (5). Of the non-O157 cases, six
risk of human exposure from such sources. VTEC serogroups are most commonly isolated
from patients: O26, O103, O145, O111, O21, and
O45. However, it should be noted that not
HUMAN PATHOGENIC VTEC all strains belonging to these serogroups
will cause severe illness and that other non-
VTEC strains are a heterogeneous group of O157 VTEC serogroups also cause illness.
E. coli serogroups, and not all will cause Pathogenic VTEC strains are categorized as

CHAPTER 23 VTEC and Veterinary Public Health 459
enterohemorrhagic E. coli, and usually in in an accumulation of synergistic virulence

these strains, a large outer membrane protein factors. This illustrates that pathotypes of
(9497 kDa) called intimin mediates the inti- E. coli can have common characteristics that
mate contact between the bacterium and the overlap and that they do and have the poten-
enterocyte cytoplasmic membrane (attach- tial to evolve rather than stand as a fixed ar-
ment) and the destruction of the enterocyte chetype (12).
microvilli (effacement). The genetic deter- Although no single marker or combination
minants for this (eae, tir, esc, and sep genes) of markers defines pathogenic VTEC asso-
are grouped together on the chromosome, ciated with human disease (13), the presence
forming a pathogenicity island called LEE, of vt2- and eae-positive strains is associated
for locus of enterocyte effacement (7). The eae with a high risk of more serious illness and
gene encodes for intimin, which is responsi- an increased risk of HUS (14), but other vir-
ble for the intimate attachment, and at ulence gene combinations and/or serotypes
least five different forms (a, b, g, d, e) have may also be associated with serious disease
been reported for VTEC strains (8). The tir in humans, including HUS. Patient-associated
gene codes for a type III-secreted translo- factors such as age, immune status, and the
cated intimin receptor (Tir protein). The esp antibiotic therapy administered may also in-
genes code for type III-secreted translocated fluence the likelihood and severity of disease.
Esp proteins. There are, however, reports The concept of seropathotype has evolved,
of LEE-negative E. coli O157 clones causing which ranks the potential for a particular
illness in humans (9), indicating that such VTEC to cause serious human illness on the
strains express alternative adherence factors basis of both serotype and the presence of
that allow them to colonize the intestinal particular virulence genes. A recent European
tract. Food Safety Authority Scientific Opinion (15)
In 2011, an outbreak strain that contained a proposed that any isolates of VTEC sero-
combination of vt genes and virulence fac- groups O157, O26, O103, O145, O111, or O104
tors normally seen in enteroaggregative E. coli that also have eae (intimin) or aaiC (secreted
(EAEC) emerged as the principal protagonist protein of EAEC) plus aggR (plasmid-encoded
in a major outbreak in Germany. The outbreak regulator) genes should be considered as
resulted in 4,321 confirmed cases, including presenting a potentially high risk for diarrhea
852 cases of HUS, with 54 deaths reported in and HUS. For any other VTEC serogroups
14 European Union member states, the United with the same combination of virulence
States, and Canada (10). This highly unusual genes, the potential risk is regarded as high for
hybrid organism combined the vt genes and diarrhea, but currently unknown for HUS.
the adhesion mechanisms of an EAEC (11) While for any serogroups not having this gene
instead of the Tir/intimin and type III secre- combination, the currently available data do
tion and the system usually seen in entero- not allow any inference regarding potential
hemorrhagic E. coli. Moreover, the strain risks. This concept allows food business
also demonstrated extended spectrum beta- operators in Europe to make an assessment of
lactamase phenotype, thus further underlining the risk posed by a VTEC isolate recovered
the enhanced virulence of the strain and the from a food, animal, or environmental sample.
severity of the outbreak, aided by the specific Nonetheless, to support stakeholders and food
combination of enhanced adhesion, survival business operators, more information is ur-
fitness, vt2 production, and antibiotic resis- gently needed on what constitutes a human
tance as a result of the high genome plasticity virulent VTEC, and this will remain an active
of this E. coli pathogen. This outbreak clone area of research, and rapidly evolving tech-
was a clear example of gene acquisition by niques such as whole genome sequencing will
means of lateral gene transfer that resulted progress this area.

CARRIAGE OF VTEC IN ANIMALS higher in calves than in adult cattle, and while
detected all year-round, carriage rates are
VTEC strains have been isolated from a vari- subject to seasonal effects, with higher rates
ety of animal carriers. including cattle, sheep, reported in spring and summer. Other factors
horses, deer, goats, dogs, geese, pigs, and wild thought to influence carriage rates are the
birds. Recent studies on prevalence of VTEC ages of animals and farming and husbandry
in ruminant animals and meats are reviewed practices (32).
in Table 2. Moreover, feces from cattle, sheep, The number of VTEC (CFU g1) pathogens
goats, deer, and rabbits have all been linked to being shed in the feces of individual animals
cases of infection (1620). Transmission to is considered important in the context of
humans can occur as a result of direct contact hide, environmental, and subsequent carcass
with VTEC-contaminated fecal material, from contamination. The typical pattern of shed-
handling or petting animals, or by exposure to ding in a herd is sporadic, with intense periods
fecally contaminated mud or vegetation dur- of shedding interspersed with periods of non-
ing recreational activities. shedding (33). Ogden et al. have also reported
that concentrations of E. coli O157 being
shed in the feces of positive cattle were
Bovine Carriage of VTEC
highest during summer months (34). The
Cattle are recognized as a principal reservoir phenomenon of super-shedding animals
for VTEC (21). VTEC is generally a transient (those shedding >104 CFU/g feces) is thought
member of the intestinal microflora and only to be a significant contributor in the dissemi-
rarely causes disease in young, weakened nation of O157 VTEC within and between
calves. Although cattle have been shown to herds and within abattoirs (3537). However,
harbor this pathogen on occasion in their quantitative data are few relative to preva-
rumen, VTEC is found more frequently in the lence data as the routine detection methods
distal portion of the bovine gastrointestinal generally employed in surveys are designed to
tract, with the rectal-anal junction identified yield data only on presence of the pathogen
as the predominant colonization site (22, 23). and not on the numbers present. As a result,
Rhoades et al. extensively reviewed the prev- data on concentrations of VTEC and on
alence of VTEC in the beef chain and the fecal occurrence of super-shedders are limited.
prevalence of E. coli O157 in cattle and showed Thomas et al. (38) examined feces and hide for
it varied from 0 to 48.8% (2427). In the concentrations of six VTEC serogroups and
United States, it is reported that most bovine showed that the vast majority of samples had
animals have been exposed to E. coli O157 by counts below the limit of detection of the
the time of weaning (28), and the reported count method and samples with detectable
prevalence in cattle ranges from 10 to 28% counts ranged from 60 to 100 CFU/cm2 on
(29). In the European Union, studies have hide and 100 to 1300 CFU g1 in feces. An
shown that VTEC prevalence in feces ranged abattoir study in the United Kingdom found
in individual animals from 2.1 to 13.5%, in that 70% of E. coli O157-positive animals shed
herds from 6.1 to 16.2%, and in slaughter <100 CFU g1 of feces, but in some individuals
batches from 13 to 20.2%, and for O157 this concentrations could be as high as 106 CFU g1
ranged from 0.3 to 2.3%, 1.5 to 13.7%, and 5.5 of feces (39). These authors also showed that
to 20.2%, respectively (6). 9% of the animals shedding E. coli O157:H7 at
Shedding of VTEC by cattle is generally slaughter produced over 96% of the total
intermittent, with herd members remaining O157:H7 fecal load for the group. It has also
negative for months, with only a proportion been hypothesized that high-level carriage
sporadically becoming positive for a few of these microorganisms is a consequence of
weeks at a time (30, 31). Carriage rates are intestinal colonization whereas low levels

TABLE 2 Selected VTEC prevalence studies (20072013) showing pathogenic VTEC prevalence in ruminant
animals (cattle and sheep)
Sample type
(n = sample no.) Serotype(s) Country Prevalence (%) Reference
Cattle
Carcass (300) O157 Spain 14.7% Ramoneda et al., 2013 (152)
Feces (1,145)a O157 United States 19.7%, Dargatz et al., 2013 (153)
Feces (1145) O45, O103, O121, United States 13.8%, 9.9%, 9.3%, Dargatz et al., 2013
O145, O26, O111 5.5%, 1.1%, 0.5%
Feces (250) O157 Mexico 5.2% Narvaez-Bravo et al., 2013 (154)
Hides (250) O157 Mexico 11.7% Narvaez-Bravo et al., 2013
Carcass (250) O157 Mexico 0.8% Narvaez-Bravo et al., 2013
Feces (301) O157 Ireland 2.7% Thomas et al., 2012 (38)
Feces (402) O26, O103, O145 Ireland 1.5%, 8.5%, 0.7% Thomas et al., 2012
Hide (301) O157 Ireland 18.9% Thomas et al., 2012
Hide (402) O26, O103, O145 Ireland 6%, 27.1%, 2.5% Thomas et al., 2012
Carcass (301) O157 Ireland 0.7% Thomas et al., 2012
Carcass (402) O26, O103, O145 Ireland 0.5%, 5.5%, 0.7% Thomas et al., 2012
Feces (399) O26, O103, O111 O145 Belgium 6%, 6%, 6%, 6% Joris et al., 2011 (47)
Carcass (474) O157 Denmark 3.4% Breum et al., 2010 (155)
Carcass (1,622) O157 Argentina 2.6% Masana et al 2010 (156)
Feces (1,622) O157 Argentina 4.1% Masana et al., 2010
Ear (446) O157 Sweden 12% Boqvist et al., 2009 (157)
Feces (1,758) O157 Sweden 3.4% Boqvist et al., 2009
Sheep
Fleece (500) O157 Ireland 0.8% Thomas et al., 2013 (61)
Fleece (500) O26, O103, O145 Ireland 1.0%, 16.8%, 0.2% Thomas et al., 2013
Carcass (500) O157 Ireland 0.6% Thomas et al., 2013
Carcass (500) O26, O103, O145 Ireland 0.4%, 13.6%, 0 Thomas et al., 2013
Feces (17,550)b O103 Sweden 0.7% Sekse et al., 2013 (58)
Feces (492) O157 Sweden 1.8% Soderlund et al., 2012 (51)
Ear (105) O157 Sweden 1.9% Soderlund et al., 2012
Feces (1,082) O157 Scotland 3.4% Evans et al., 2011 (56)
Feces (1,082) O26, O103, O145 Scotland 3.4%, 5.2%, 2.3% Evans et al., 2011
Feces (491)c O26 Norway 17.9% Sekse et al., 2011 (52)
Feces (533) O157 Italy 7.1% Franco et al., 2009 (54)
Fleece (400) O157 Ireland 5.75% Lenahan et al., 2007 (60)
Carcass (400) O157 Ireland 1.5% Lenahan et al., 2007
a
Individual fecal swabs were collected from cattle approximately 60 days after their arrival in the feedlot and were pooled for evaluation
(153).
b
The investigation of fecal samples from 585 sheep ocks resulted in 1,222 E. coli O103 isolates that were analyzed for the presence of
eae and stx genes; the study documented a low prevalence (0.7%) of isolates potentially pathogenic to humans (58).
c
This number represents the number of ocks that were tested in the study using fecal samples to determine the prevalence; in total, 491
ocks were tested and E. coli O26 was detected in 17.9% of the ocks (52).
within individual animals may be a result of O157 versus non-O157 VTEC serogroups.
environmental exposure with no significant Whereas >90% of E. coli O157 isolates recov-
colonization (32, 40). Horizontal transmission ered from the beef chain in Ireland possessed
between animals may be facilitated by con- vt and eae in combination, <10% of non-
taminated water and feed, with water troughs O157 VTEC serogroups fell into this category
potentially playing a role in the ecology of (38). In a Belgian study on O26, O103, O145,
VTEC on the farm (4146). and O111 in cattle feces, about 6% of samples
Recent studies have examined the human were positive and about 50% of isolates had
virulence potential of different bovine VTEC key human virulence genes (47, 48). A further
isolates, and an interesting picture is seen for study showed that for E. coli O157 genotypes

associated with human illness a minor sub- and O145 (with eae and vt) at level <1% on
population was in the bovine reservoir. (47, sheep fleece and carcases (61). For O26, iso-
48) This shows that while VTEC isolates in lates with vt and eae represented about 50% of
beef play a role in human illness, a risk assess- all isolates of the serogroups recovered, while
ment of their virulence potential is essential. for O103 it was only about 0.01% of all isolates.
As for bovine isolates, these data highlight
the need to risk assess any recovered VTEC
Carriage of VTEC in Sheep and Goats
regardless of serogroup for key virulence
The presence of naturally occurring VTEC genes (61).
has been widely reported in sheep, and studies
indicate that sheep feces and sheep meat are
Carriage of VTEC in Pigs
important reservoirs of human pathogenic
VTEC (4955). Seasonal prevalence of VTEC Although pigs are also known carriers of O157
in sheep has been reported (54, 56), with VTEC, prevalence rates are relatively low,
incidences peaking in summer months in a ranging from 0.4 to 2.1% (6265). In pigs,
similar trend to that in cattle and human in- VTEC can cause edema disease and may also
fections (45). Much of the data on VTEC be involved in postweaning diarrhea syn-
colonization of small ruminants relate to dromes. In general, it is VTEC serogroups
E. coli O157. When colonized, small ruminants O138, O139, and O141 that are implicated in
generally show no clinical symptoms of illness illness in pigs, and these have not been widely
and reinfection occurs frequently, although in linked to human illness. The presence of
young unweaned lambs or kids, scouring or E. coli O157 in pigs has been reported in
diarrhea may occur. Some animals, particu- the United States (62), Norway, and the
larly those that are persistently colonized, can Netherlands (66, 67). In a study carried out in
excrete exceptionally high numbers of E. coli Ireland, the prevalence of E. coli O157:H7 in
O157 (>10,000 CFU/g) in their feces (54, 57). pigs was found to be very low, and only four
There is some evidence of higher shedding isolates were recovered from 1,710 pigs ex-
of O157 in adult sheep and hoggets than in amined. However, three of these four con-
lambs (56). tained the genes vtx2, eae, and hlyA, indicating
E. coli O103 is one of the most common their potential to cause illness in humans
non-O157 VTEC serotypes isolated from (68). While pork meat and products have
human cases in Europe, and a severe human generally not been implicated in human ill-
outbreak was reported in Norway in 2006 that ness, in California in 2006, feral swine were
was caused by E. coli O103:H25 (58). This was implicated in contamination of agricultural
due to fermented sausage mainly consisting fields and surface waterways with E. coli O157:
of mutton, which was shown to be the food H7, which caused a large outbreak linked to
source of the outbreak (59). However, only bagged spinach. However, E. coli O157 isolates
eae+vt-E .coli O103:H25 was ever isolated from were also recovered from cattle on a nearby
food and sheep, but it was proposed that the ranch that had a similar subtype to the swine
infection was caused by VTEC and that re- isolates as determined by using multilocus
covered isolates probably had lost their vt variable tandem repeat analysis and pulsed-
genes during laboratory cultivation. In studies field gel electrophoresis (PFGE), suggesting
carried out in Ireland, the profile of E. coli that not only had swine-to-swine transmission
O157:H7 recovered from sheep showed that taken place but, in addition, that interspecies
29/33 (87%) of isolates contained vt2, eae, and transmission between cattle and swine had
hlyA genes, whereas only five isolates (15%) occurred, facilitated through a common source
contained the vtx1 gene (60, 61). Thomas et al. of exposure such as water or soil (69). More
showed recovery of pathogenic O26, O103, data on the human virulence assessment of

non-O157 serogroups isolated from pigs would slurries are mixtures that include manure,
provide important information on the likely urine, and leftover feed that are held in a tank
future importance of their role as a vector of or pit, generally under anaerobic conditions.
human VTEC infection. Such waste is a valuable source of soil nu-
trients and is generally spread back to the land
as fertilizer. If such waste is contaminated
ANIMAL WASTE AS A SOURCE with VTEC, it poses a reservoir of contami-
OF VTEC FOR HUMANS nation for animals, water supplies, crops, and
fresh produce.
As animal waste can carry VTEC, there is a In experiments on the survival of E. coli
risk of direct infection of humans from contact O157:H7 in feces spread outdoors on grass
with livestock and/or waste in the farm en- under ambient weather conditions, the path-
vironment. A number of cases of human in- ogen could be recovered directly from feces
fection have been associated with visits to on the grass for 50 days, and when feces were
farms, agricultural fairs, and petting farms. no longer visible on the grass, enrichment
Visitors to farms are likely to be at a higher techniques showed the organism was recov-
risk of infection than farming families who erable from the underlying soil for a further
may possess a degree of acquired immunity 49 days (71). A similar study in the United
(70). Awareness by farmers and farming fam- Kingdom (72) reported the survival of E. coli
ilies of the potential dangers of VTEC and the O157 in cattle feces for >50 days but reported
steps required to protect themselves, and much shorter survival times in cattle slurry in
any visitors, from VTEC infection is essential. which it fell to undetectable levels within
Full consideration of hygiene management is 10 days. However, McGee et al. reported that
needed at open farms/fairs to reduce risks E. coli O157:H7 was recoverable from bovine
for visitors. Washing hands after contact slurry for up to 3 months (73). Semenov et al.
with animals, especially before eating and showed the influence of aerobic and anaerobic
drinking, is a simple control measure; there- conditions on the survival of E. coli O157:H7
fore, open farms/fairs should be adequately in cattle manure and slurry (74). The data
equipped with hand-washing facilities, and showed that E. coli O157:H7 survived signifi-
provide designated dining areas that are seg- cantly longer under anaerobic than under
regated from animal holding areas. Farmers aerobic conditions, with survival ranging from
should be aware that farmyard surfaces such approximately 2 weeks for aerobic manure
as gates and stiles pose a direct risk as E. coli and slurry to more than 6 months for anaer-
O157:H7 has the potential to persist for long obic manure at 16C. The importance of
periods on these surfaces (70). In addition, changes in microbial community and chemical
adequate guidance and supervision before and composition of manure and slurry was high-
during farm visits are a key factor in reducing lighted as affecting the survival of E. coli
the risk to children under 5 years of age who O157:H7 in different oxygen conditions.
come into direct contact with animals. In further field experiments to determine
the survival times of pathogens in livestock
manures during storage and following land
Survival of VTEC in Animal Waste
application, VTEC survived for less than 1
Animal waste in the environment can arise month in solid manure heaps where tempe-
from feces directly excreted into the envi- ratures greater than 55C were obtained.
ronment by grazing or wild animals and Following spreading of manure to land, E. coli
from animal waste collected and stored as O157 generally survived in the soil for up to 1
slurry or manure. Manures are feces that may month after application to both sandy arable
have undergone some period of storage, and and clay loam grassland soils (75).

When left on the ground surface, patho- housing with good floor drainage, and prac-
gens such as VTEC are exposed to the full ticing a closed herd policy (81, 82).
force of fluctuating environmental conditions Manure can be subjected to active treat-
that has been shown to increase the inactiva- ment processes including composting, heat
tion rate as compared to the immediate in- drying, and digestion. In general, these treat-
corporation into the soil (75, 76). It has been ments are according to legislative guidance,
found that E. coli O157 can survive on surface which are primarily devised to achieve effi-
vegetation for 6 weeks but persisted when cient and sustainable nutrient use. But any
injected subsurface (77). VTEC may then be treatment will also have implications for the
transported over land or below the surface pathogen/VTEC load in the final output. The
flow to surface and ground waters. Less work regulations generally specify the storage re-
has been focused on survival of non-O157 quirements and spreading restrictions to pre-
VTEC; however, a study that investigated the vent diffuse and point-source contamination
survival of E. coli O26, O111, and O157 in bo- of adjacent surface and ground waters.
vine feces (78) stored at 5, 15, and 25C dem- Composting is a useful way of managing
onstrated that all three pathogens survived at solid animal waste. Under controlled condi-
5 and 25C for 1 to 4 weeks and at 15C for 1 to tions of aeration, moisture, particle size, and
8 weeks when inoculated at a low concen- carbon-nitrogen ratio of the combustible ma-
tration. At high concentrations, O26, O111, and terial, composting temperatures of 55 to 65C
O157 survived at 25C for 3 to 12 weeks, at can be reached, which should be sufficient
15C for 1 to 18 weeks, and at 5C for 2 to 14 to inactivate VTEC. Lung et al. showed that
weeks, respectively. E. coli O157:H7 inoculated at a level of log10
In a study by Bolton et al. soil was ino- 7.00 CFU g1 was nondetectable after 72 h
culated with O26:H11, O113:H4, ONT:H4, in composted manure (83). Research by
O2:H27, O116:H28, O6:H8, ONT:H27, O119: Shepherd et al. showed that compost heaps
H5, O145:H27, O20:H19, O174:H21, O168:H8, covered with finished compost maintained
O136:H2, and O86:H21; the samples were temperatures under the physical covering that
stored at 10C for up to 201 days and showed were around 7 to 15.5C higher than in an un-
D values (time required to achieve a 90% covered heap, resulting in a faster reduction of
or 1 log reduction) ranging from 50.26 to E. coli O157:H7 than was observed in heaps
75.60 days in sandy loam and 31.60 to 48.25 covered with fresh straw or left uncovered
days in clay-based soils (79). (84). This validated recommendations of the
U.S. Environmental Protection Agency for
covering fresh compost (84).
Control in Animal Waste
Treatments particularly suited to slurries
Such long-term survival of E. coli O157:H7 in include anaerobic digestion typically at 30 to
manure and slurry emphasizes the need for 35C with 12 days of retention for pig slurry or
appropriate farm waste management to curtail 20 days for cattle and poultry slurry. The ad-
environmental spread of this bacterium and dition of lime (quick lime or slaked lime to
the risk of human exposure. raise the pH to 12 for at least 2 h) should also
General measures to control the spread of result in significant pathogen reduction (85).
pathogens including VTEC on the farm are
outlined in farm quality assurance and good
VTEC in Water
agricultural practices schemes (80). General
measures to control the spread and persis- Surface water and unprotected groundwater
tence of VTEC on the farm include good man- may become contaminated with VTEC from
agement and hygiene, clean and dry bedding, livestock effluent or fecal contamination from
appropriate stocking rates, well-ventilated humans, livestock, wild animals, and domestic

pets. Water for both drinking and recreational water trough sediments for at least 4 months
swimming has been implicated in a number of and appears to multiply there, especially in
VTEC outbreaks (86). The susceptibility of warm weather. On many farms, troughs are
drinking water from private water supplies in seldom cleaned so that thick sediments accu-
rural areas to VTEC contamination has been mulate and remain a long-term potential source
highlighted, with a private water supply being for continual recycling of VTEC among a herd.
considered as one that is not provided by a Farmers should clean water troughs frequently
water company and includes wells, boreholes, to prevent the accumulation of sediments (42).
springs, streams, rivers, lakes, or ponds. Such Water trough design and location are also
private water supplies are common in both important factors in reducing the possibility
Ireland and Scotland, where rates of human of direct fecal contamination. Water troughs
VTEC illness are high, and in both countries should be positioned away from feed troughs/
private water can be considered a potentially feed passageways, as contamination of water
important risk factor for VTEC human in- with feed can provide a nutrient-rich substrate
fection (8789). A large E. coli O157 outbreak for bacterial growth and survival at the bottom
involving a private group water scheme was of the trough (42). Water troughs should not be
reported in a rural area of Ireland, involving 18 located in shaded areas (42), as direct sunlight
cases of infection, including 2 HUS cases (46). has a bactericidal effect.
A study by Tanaro et al. showed that on a beef
farm in Argentina, non-O157 isolates with
identical PFGE profiles were recovered from VTEC IN FOODS OF ANIMAL ORIGIN
cattle rectal swabs, and water streams on the
farm highlighted the circulation of VTEC Food of animal origin such as milk and dairy
strains among the animals, the environment, products may be contaminated with VTEC
and the water course (90). Other studies have during milking, and the meat carcass may be
shown that the vulnerability of groundwater contaminated during slaughter and dressing
supplies is influenced by soil types, with operations.
groundwater being most at risk where the
subsoils are absent or thin and in areas of
VTEC in Milk
karstic limestone, where surface streams sink
underground at swallow holes (91, 92). Although the public health risks associated
Studies have showed that E. coli O157:H7 with consumption of raw milk and raw milk
survived for 13 weeks in lake water held at 15 products have been well documented, such
C (93), while other studies have reported products continue to be consumed and pro-
survival in farm water for up to 14 days at moted by pro-raw milk advocates because of
temperatures of <15C (94), 70 days at 5C, perceived organoleptic and health benefits
and 40 days at 21C (95). In addition to tem- over pasteurized products (98, 99). However,
perature, the indigenous microflora (93), a number of serious outbreaks of E. coli O157
scarcity of nutrients (96), and protozoan pre- have been associated with the consumption
dations also influence survival. Recent passage of raw milk (100, 101), raw milk cheese (102),
of VTEC through the bovine gastrointestinal and unpasteurized Gouda cheese (103, 104).
tract may enhance survival of O157 in water Other outbreaks were linked to butter made
(97). from raw milk and to commercial ice creams,
Contaminated water troughs and livestock where cross-contamination was identified
drinking water supplies have a role in the as a possible source (105). Allerberger et al.
transmission of VTEC among animals (33), reported two cases of HUS due to E. coli O26
and large numbers of animals can be infected linked to the consumption of raw cows milk
over a short period (42). VTEC can survive in (106). However, pasteurized products have

also been implicated in outbreaks, and E. coli products such as cheese depend on factors
O145 and O26 caused VTEC infection with such as the moisture content, pH, and com-
five HUS cases among consumers of ice cream petitive microflora. Thus the additional hurdles
produced and sold in September 2007 at a imposed during the hard cheese manufactur-
farm in Belgium. The ice cream was made ing process, including low water activity and
from pasteurized milk and most likely was pH that develop during the curing process, and
contaminated by one of the food handlers the changes in the competing microflora, re-
(107). An outbreak of E. coli O111 associated duce the survival and growth potential of the
with a correctional facility dairy in Colorado pathogen. Peng et al. showed survival of non-
showed inmates employed at the dairy might O157 following the production and ripening of
have acquired VTEC O111 infection on the job semi-hard raw milk cheese (113). Miszczycha
or transported contaminated clothing or other et al. examined the growth and survival of four
items into the main correctional facility and VTEC serotypes (O157:H7, O26:H11, O103:H2,
kitchen, thereby exposing other inmates (108). and O145:H28) in raw-milk cheeses manufac-
Contamination of raw milk can occur tured and ripened according to five techno-
either during milking or after milking from on- logical schemes: blue-type cheese, uncooked
farm environmental contamination. A study by pressed cheese with long ripening and short
Murphy et al. examined milk from 97 Irish ripening steps, cooked cheese, and lactic
dairy cattle farms and isolated E. coli O157:H7 cheese (114). VTEC grew 2 to 3 log10 CFU g1 in
from 12% of samples on these farms (109). A the blue-type cheese and the two uncooked
subsequent study by Murphy et al. examined pressed cheeses during the first 24 h of cheese
Irish raw milk specifically destined to be used making, but levels then progressively de-
in raw milk cheese and detected VTEC in raw creased in cheeses that were ripened for more
milk from sheep and goats in addition to cows than 6 months. In the cooked and lactic
(110). Data collated from European Union cheeses, VTEC did not grow but was detect-
member states in 2011 (111) showed 5.3% able at the end of ripening and storage. Inter-
(3/57) of raw milk samples in Germany con- estingly, a serotype effect was observed in all
tained VTEC and 2.6% (1/39) raw milk cheeses, with O157:H7 growing slower and less
samples in Belgium contained E. coli O157. persistently than the other serotypes, indicat-
Among five reporting member states, 1.8% ing that the risk posed by non-O157 serotypes
(36/2,045) of all dairy samples were positive may be higher and the design of safety mea-
for VTEC and 0.1% for E. coli O157. Between sures should take this into account.
2008 and 2011, an Italian study that monitored Several studies reported the ability of
for the prevalence of E. coli O157:H7 in raw E. coli O157:H7 to survive and to grow during
milk (60,907 samples) sold by self-service storage of fermented dairy products. E. coli
vending machines (n = 1,239) in seven Italian O157:H7 inoculated into commercial products
regions found just 24/60,907 (0.04%) of sam- could be recovered for up to 12 days from
ples were positive for E. coli O157:H7. yogurt, 28 days from sour cream, and 32 to
VTEC may reportedly behave differently in 35 days from buttermilk (115, 116).
raw and pasteurized milk and in derived
products. Growth of VTEC may be slower in
VTEC and Meat
raw milk and products because of the pre-
sence of lactoperoxidase enzymes, natural in- Since the first confirmed case in 1982, beef-
hibitors for many pathogenic bacteria in associated VTEC O157 outbreaks were widely
raw milk, and may also be due to the higher reported (3, 117119). Between 2007 and 2009,
numbers of competitive microflora in such of the 57 verified food-borne outbreaks of
products (112). The behavior and the poten- VTEC in the European Union, 8were linked to
tial for survival/growth of VTEC in dairy the consumption of bovine meat or bovine

meat products. During the years 2010 and For the abattoir, knowledge from the farm
2011, of the 16 outbreaks reported with strong of the VTEC prevalence in the slaughter batch
evidence, 2 were linked to the consumption of could allow risk management of animals
bovine meat or bovine meat products (6). coming in for slaughter, including logistic
The hide and fleece of animals presented slaughter, but currently there is insufficient in-
for slaughter are recognized as important formation on VTEC at the farm level to im-
sources of VTEC contamination for the car- plement this approach (6, 127).
cass. Reported prevalence of E. coli O157 on The risk of fecal contamination on the
hide varies from 4.7% (120), 17.6 % (38), to carcass at slaughter can be reduced by specific
75.7% for feedlot cattle (121). On sheep fleece procedures, including rodding (a technique
the reported prevalence of E. coli O157 was used to separate the esophagus from the tra-
5.75 (60) and <1 % (61). Studies looking at an chea and diaphragm). Bagging and tying of the
extended range of VTEC serogroups (O26, bung can also help prevent contamination of
O111, O145, O103) recovered <1% on bovine the carcass. Removal of hides should be car-
hide and on sheep fleece recovered O26 at 1%, ried out in a manner that avoids contact be-
with neither O103, O145, nor O111 recovered tween the hide and the carcass. This can be
(38, 61). achieved by a number of measures, including
Transmission of E. coli O157:H7 and other the use of hide pulling equipment and using
VTEC serotypes can occur rapidly in groups clean equipment (immersion of knives in
of animals, with contamination of the bovine water at 82C) for the dehiding operation.
pens and hides occurring in less than 24 h The reported prevalence of E. coli O157 on
(122). Thus conditions in transport from farm prechill carcasses (postevisceration) is low
to factory and in lairage have a significant (3%) (27). There is a low prevalence (<0.5%)
impact on the level of VTEC contamination of other clinically significant VTEC sero-
on hide or fleece of animals presented for groups (O145, O111, O103, and O26) on prechill
slaughter. The risk of introducing VTEC to a beef and sheep carcasses (38, 61). The simi-
slaughter batch is increased by mixing animals larity of PFGE profiles for VTEC isolates in a
from different farms (123, 124). The length of study by Thomas et al. indicated the origin of
time animals are in transit to the abattoir and VTEC on a carcass could be its own hide or
held in lairage also increases the risk of having the hides or fleece of other animals being
VTEC-positive hide samples at slaughter, slaughtered on the same day (38, 61). VTEC
compared with cattle transported a shorter that may also be present in the air of the
distance (125, 126). E. coli O157 has been slaughter house, particularly at the hide re-
shown to survive on the hides of live cattle moval area, may be an underestimated source
for approximately 9 days. Efforts to reduce of carcass contamination (128). The use of
the level of hide or fleece soiling are thus carcass antimicrobial interventions is com-
warranted for control of VTEC. Farmers can monplace in the U.S. beef industry (129) and
employ suitable animal husbandry facilities include the use of organic acids, acidified so-
and practices, such as bedding quality, stock- dium chlorite, trisodium phosphate, activated
ing density, and feeding regimen, that give lactoferrin, chlorine, and chlorine dioxide and
clean animals for slaughter although a clean hot water. Studies by Kalchayanand et al. in-
animal does not guarantee absence of VTEC. dicated that the reductions in non-O157 VTEC
Livestock transporters can also play an im- by the commonly used antimicrobial interven-
portant role in the cleanliness and dryness of tions were at least as great as for O157 (130).
animals on arrival at the abattoir by ensuring Such antimicrobial interventions for beef car-
that trucks are thoroughly washed and dis- casses are not currently in use in the European
infected between loads and by not overloading Union beef industry sector. This is mainly due
the animals. to customer preferences and associated costs.

The survival and concentrations of E. coli managing public health risk. Whole genome
O157:H7 and other VTEC serogroups on fresh sequencing approaches are now helping iden-
meat during distribution can be affected by tify the genetic differences between human
the storage temperatures (131, 132), the pack- pathogenic and nonpathogenic VTEC, but
aging environment (133), and the competitive more research is needed into the interactions
microflora (134). The processing of beef cuts between virulence factors and the human or
into ground beef can lead to transfer of path- animal host. Much of the knowledge to date
ogen from the surface of beef into the center has been obtained from in vitro observations,
of the product, and beef-processing equip- which do not necessarily reflect what is hap-
ment and knives or needles used to cut into pening in vivo within a specific host. It is also
or inject whole muscle (in tenderizing beef) likely that the microbiota plays a crucial part
can play a role in this spread of contamina- in disease dynamics and in host-pathogen and
tion (135137). Studies on thermal inactivation host-commensal and commensal-pathogen
of different single serotypes of VTEC O26:H11, interactions. For example, signals from the
O45:H2, O103:H2, O104:H4, O111:H, O121:H19, commensal flora may affect expression of
O145:NM, and O157:H7 in wafers of ground virulence determinants in pathogens (140).
beef showed that cooking times and temper- Therefore, the challenges that lie ahead are to
atures effective for inactivating a serotype improve the understanding of the ecology of
O157:H7 strain of E. coli in ground beef were the pathogen at sites of colonization and in
equally effective against the seven non-O157:H7 key contamination in the agri-food environ-
strains investigated (135). ment that will underpin strategies for the
VTEC can reportedly survive in ready-to- prevention of transmission to target this di-
eat, dry, and semidry fermented meats, posing verse group of pathogens.
a particular risk in these commodities (138). In addition, among members of a serotype
This has led to a recommendation that pro- there may be significant strain differences that
cessing protocols achieve a log10 5.0 CFU g1 warrant further research and that need to be
reduction in numbers of E. coli O157:H7. Such considered for effective management prac-
reductions may be achieved by additional ther- tices. Data are now building on the phenotypic
mal processing, and a number of published characteristics and behavior of non-O157
studies outline heating steps that can be in- VTEC in the food chain that indicate that
troduced into the manufacturing process after they behave in a similar way to O157 and that
fermentation to achieve the required decline current general interventions implemented at
in pathogen numbers (139). In the absence of a pre- or postharvest stages of the food chain
thermal processing step, an extended fer- will yield similar reductions regardless of
mentation or maturation period may prove serogroup (130, 135, 141). However, there are
effective in limiting pathogen numbers. some recent studies that indicate that this
may not always be case, i.e., in semi-hard raw-
milk cheese, non-O157 VTEC persisted longer
CONCLUSION than O157 (114). There is an urgent need for
robust studies with more serogroups and
It is clear that there are multiple routes for more strains to validate survival kinetics of
transmission of human pathogenic VTEC diverse VTEC strains and to assess the impact
along the farm-to-fork chain. Emerging chal- of food control measures to ensure that dif-
lenges are new vehicles of infection and a ferences are true serogroup differences and
changing profile of the VTEC serogroups that not related to an interstrain impact. There
are causing human illness. A lack of clarity on is a need for better epidemiological link-
what constitutes a human pathogenic VTEC age of strains from animal, farm environ-
poses additional challenges in assessing and ment, and foods and humans and for source

attribution of VTEC, in particular, for emerg- and Food-borne Outbreaks in 2011. EFSA Journal
ing serogroups. 11(4):3129.
7. Kaper JB. 1998. Enterohemorrhagic Escherichia
This is challenging, but with so many po- coli. Curr Opin Microbiol 1:103108.
tential vehicles and routes of infection, some 8. Law D. 2000. Virulence factors of Escherichia coli
risk ranking at a country or regional level of O157 and other Shiga toxin-producing E. coli. J
the pathways and products, based on the rel- Appl Microbiol 88:729745.
evant regional epidemiological and practices, 9. Paton AW, Paton JC. 1999. Direct detection of
Shiga toxigenic Escherichia coli strains belonging
would be useful in effectively directing and
to serogroups O111, O157, and O113 by multiplex
focusing veterinary public health risk man- PCR. J Clin Microbiol 37:33623365.
agement approaches and resources. 10. Karch H, Denamur E, Dobrindt U, Finlay BB,
Hengge R, Johannes L, Ron EZ, Tonjum T,
Sansonetti PJ, Vicente M. 2012. The enemy
ACKNOWLEDGMENT within us: lessons from the 2011 European Esch-
erichia coli O104:H4 outbreak. EMBO Mol Med
We declare no conflicts of interest with regard 4:841848.
to the manuscript. 11. Bielaszewska M, Mellmann A, Zhang W, Kock
R, Fruth A, Bauwens A, Peters G, Karch H. 2011.
Characterisation of the Escherichia coli strain as-
CITATION sociated with an outbreak of haemolytic uraemic
syndrome in Germany, 2011: a microbiological
Duffy G, McCabe E. 2014. Veterinary public study. Lancet Infect Dis 11:671676.
health approach to managing pathogenic 12. Rasko DA, Webster DR, Sahl JW, Bashir A,
verocytotoxigenic Escherichia coli in the agri- Boisen N, Scheutz F, Paxinos EE, Sebra R, Chin
food chain. Microbiol Spectrum 2(5):EHEC- CS, Iliopoulos D, Klammer A, Peluso P, Lee L,
0023-2013. Kislyuk AO, Bullard J, Kasarskis A, Wang S, Eid
J, Rank D, Redman JC, Steyert SR, Frimodt-
Moller J, Struve C, Petersen AM, Krogfelt KA,
Nataro JP, Schadt EE, Waldor MK. 2011.
REFERENCES
Origins of the E. coli strain causing an outbreak of
1. Fairbrother JM, Nadeau E. 2006. Escherichia hemolytic-uremic syndrome in Germany. N Engl
coli: on-farm contamination of animals. Rev Sci J Med 365:709717.
Tech 25:555569. 13. Beutin L, Martin A. 2012. Outbreak of Shiga
2. Adak GK, Long SM, OBrien SJ. 2002. Trends toxin-producing Escherichia coli (STEC) O104:H4
in indigenous foodborne disease and deaths, infection in Germany causes a paradigm shift
England and Wales: 1992 to 2000. Gut 51:832 with regard to human pathogenicity of STEC
841. strains. J Food Prot 75:408418.
3. Riley LW, Remis RS, Helgerson SD, McGee HB, 14. Ethelberg S, Olsen KE, Scheutz F, Jensen C,
Wells JG, Davis BR, Hebert RJ, Olcott ES, Schiellerup P, Enberg J, Petersen AM, Olesen
Johnson LM, Hargrett NT, Blake PA, Cohen B, Gerner-Smidt P, Molbak K. 2004. Virulence
ML. 1983. Hemorrhagic colitis associated with a factors for hemolytic uremic syndrome, Denmark.
rare Escherichia coli serotype. N Engl J Med Emerg Infect Dis 10:842847.
308:681685. 15. European Food Safety Authority, EFSA Panel
4. Tarr PI, Gordon CA, Chandler WL. 2005. Shiga- on Biological Hazards (BIOHAZ). 2013. Scien-
toxin-producing Escherichia coli and haemolytic tific Opinion on VTEC-seropathotype and scien-
uraemic syndrome. Lancet 365:10731086. tific criteria regarding pathogenicity assessment.
5. Centers for Disease Control and Prevention. EFSA Journal 11(4):3138.
2012. Foodborne Diseases Active Surveillance 16. Bettelheim KA. 2003. Non-O157 verotoxin-
Network (FoodNet): FoodNet Surveillance Report producing Escherichia coli: a problem, paradox,
for 2011 (Final Report). U.S. Department of Health and paradigm. Exp Biol Med (Maywood) 228:333
and Human Services, Centers for Disease Control 344.
and Prevention, Atlanta, GA. 17. Beutin L, Geier D, Steinruck H, Zimmermann S,
6. European Food Safety Authority, European Scheutz F. 1993. Prevalence and some properties
Centre for Disease Prevention and Control. of verotoxin (Shiga-like toxin)-producing Esche-
2013. The European Union Summary Report on richia coli in seven different species of healthy
Trends and Sources of Zoonoses, Zoonotic Agents domestic animals. J Clin Microbiol 31:24832488.

18. Blanco M, Blanco JE, Mora A, Dahbi G, Alonso A, Rodriguez A, Rey J, Alonso JM, Usera MA.
MP, Gonzalez EA, Bernardez MI, Blanco J. 2003. Verotoxin-producing Escherichia coli in
2004. Serotypes, virulence genes, and intimin Spain: prevalence, serotypes, and virulence genes
types of Shiga toxin (verotoxin)-producing Esch- of O157:H7 and non-O157 VTEC in ruminants,
erichia coli isolates from cattle in Spain and ident- raw beef products, and humans. Exp Biol Med
ification of a new intimin variant gene (eae-xi). (Maywood) 228:345351.
J Clin Microbiol 42:645651. 31. Monaghan A, Byrne B, Fanning S, Sweeney T,
19. Sachdeva S, Ahmad G, Malhotra P, Mukherjee McDowell D, Bolton DJ. 2011. Serotypes and
P, Chauhan VS. 2004. Comparison of immuno- virulence profiles of non-O157 Shiga toxin-
genicities of recombinant Plasmodium vivax producing Escherichia coli isolates from bovine
merozoite surface protein 1 19- and 42-kiloDalton farms. Appl Environ Microbiol 77:86628668.
fragments expressed in Escherichia coli. Infect 32. Naylor SW, Gally DL, Low JC. 2005. Entero-
Immun 72:57755782. haemorrhagic E. coli in veterinary medicine. Int J
20. Scaife HR, Cowan D, Finney J, Kinghorn-Perry Med Microbiol 295:419441.
SF, Crook B. 2006. Wild rabbits (Oryctolagus 33. Hancock DD, Besser TE, Rice DH, Ebel ED,
cuniculus) as potential carriers of verocytotoxin- Herriott DE, Carpenter LV. 1998. Multiple
producing Escherichia coli. Vet Rec 159:175178. sources of Escherichia coli O157 in feedlots and
21. Caprioli A, Morabito S, Brugere H, Oswald dairy farms in the northwestern USA. Prev Vet
E. 2005. Enterohaemorrhagic Escherichia coli: Med 35:1119.
emerging issues on virulence and modes of 34. Ogden ID, MacRae M, Strachan NJ. 2004. Is the
transmission. Vet Res 36:289311. prevalence and shedding concentrations of E. coli
22. Naylor SW, Low JC, Besser TE, Mahajan A, O157 in beef cattle in Scotland seasonal? FEMS
Gunn GJ, Pearce MC, McKendrick IJ, Smith Microbiol Lett 233:297300.
DG, Gally DL. 2003. Lymphoid follicle-dense 35. Matthews L, McKendrick IJ, Ternent H, Gunn
mucosa at the terminal rectum is the principal GJ, Synge B, Woolhouse ME. 2006. Super-
site of colonization of enterohemorrhagic Esche- shedding cattle and the transmission dynamics
richia coli O157:H7 in the bovine host. Infect of Escherichia coli O157. Epidemiol Infect 134:
Immun 71:15051512. 131142.
23. Arthur TM, Brichta-Harhay DM, Bosilevac JM, 36. Matthews L, Low JC, Gally DL, Pearce MC,
Kalchayanand N, Shackelford SD, Wheeler TL, Mellor DJ, Heesterbeek JA, Chase-Topping M,
Koohmaraie M. 2010. Super shedding of Esche- Naylor SW, Shaw DJ, Reid SW, Gunn GJ,
richia coli O157:H7 by cattle and the impact on Woolhouse ME. 2006. Heterogeneous shedding
beef carcass contamination. Meat Sci 86:3237. of Escherichia coli O157 in cattle and its implica-
24. Hussein HS, Bollinger LM. 2005. Prevalence of tions for control. Proc Natl Acad Sci USA 103:547
Shiga toxin-producing Escherichia coli in beef. 552.
Meat Sci 71:676689. 37. Duffy G, Cummins E, Nally P, OB S, Butler F.
25. Hussein HS, Bollinger LM. 2005. Prevalence of 2006. A review of quantitative microbial risk as-
Shiga toxin-producing Escherichia coli in beef sessment in the management of Escherichia coli
cattle. J Food Prot 68:22242241. O157:H7 on beef. Meat Sci 74:7688.
26. Naylor SW, Roe AJ, Nart P, Spears K, Smith 38. Thomas KM, McCann MS, Collery MM,
DG, Low JC, Gally DL. 2005. Escherichia coli Logan A, Whyte P, McDowell DA, Duffy G.
O157:H7 forms attaching and effacing lesions at the 2012. Tracking verocytotoxigenic Escherichia coli
terminal rectum of cattle and colonization requires O157, O26, O111, O103 and O145 in Irish cattle. Int
the LEE4 operon. Microbiology 151:27732781. J Food Microbiol 153:288296.
27. Rhoades JR, Duffy G, Koutsoumanis K. 2009. 39. Omisakin F, MacRae M, Ogden ID, Strachan
Prevalence and concentration of verocytotoxi- NJ. 2003. Concentration and prevalence of Esche-
genic Escherichia coli, Salmonella enterica and richia coli O157 in cattle feces at slaughter. Appl
Listeria monocytogenes in the beef production Environ Microbiol 69:24442447.
chain: a review. Food Microbiol 26:357376. 40. Low JC, McKendrick IJ, McKechnie C, Fenlon
28. Laegreid WW, Elder RO, Keen JE. 1999. Prev- D, Naylor SW, Currie C, Smith DG, Allison L,
alence of Escherichia coli O157:H7 in range beef Gally DL. 2005. Rectal carriage of enterohemor-
calves at weaning. Epidemiol Infect 123:291298. rhagic Escherichia coli O157 in slaughtered cattle.
29. Karmali MA, Gannon V, Sargeant JM. 2010. Appl Environ Microbiol 71:9397.
Verocytotoxin-producing Escherichia coli (VTEC). 41. LeJeune JT, Besser TE, Hancock DD. 2001.
Vet Microbiol 140:360370. Cattle water troughs as reservoirs of Esche-
30. Blanco J, Blanco M, Blanco JE, Mora A, richia coli O157. Appl Environ Microbiol 67:
Gonzalez EA, Bernardez MI, Alonso MP, Coira 30533057.

42. LeJeune JT, Besser TE, Merrill NL, Rice DH, coli O26 in Norwegian sheep flocks. Appl Environ
Hancock DD. 2001. Livestock drinking water Microbiol 77:49494958.
microbiology and the factors influencing the 53. Sanchez S, Martinez R, Garcia A, Blanco J,
quality of drinking water offered to cattle. J Dairy Blanco JE, Blanco M, Dahbi G, Lopez C, Mora
Sci 84:18561862. A, Rey J, Alonso JM. 2009. Longitudinal study of
43. Lejeune JT, Besser TE, Rice DH, Hancock DD. Shiga toxin-producing Escherichia coli shedding
2001. Methods for the isolation of water-borne in sheep feces: persistence of specific clones in
Escherichia coli O157. Lett Appl Microbiol 32:316 sheep flocks. Appl Environ Microbiol 75:1769
320. 1773.
44. Sargeant JM, Sanderson MW, Smith RA, 54. Franco A, Lovari S, Cordaro G, Di Matteo P,
Griffin DD. 2003. Escherichia coli O157 in feedlot Sorbara L, Iurescia M, Donati V, Buccella C,
cattle feces and water in four major feeder-cattle Battisti A. 2009. Prevalence and concentration of
states in the USA. Prev Vet Med 61:127135. verotoxigenic Escherichia coli O157:H7 in adult
45. Carroll AM, Gibson A, McNamara EB. 2005. sheep at slaughter from Italy. Zoonoses Public
Laboratory-based surveillance of human verocyto- Health 56:215220.
toxigenic Escherichia coli infection in the Repub- 55. Brandal LT, Sekse C, Lindstedt BA, Sunde M,
lic of Ireland, 20022004. J Med Microbiol 54: Lobersli I, Urdahl AM, Kapperud G. 2012.
11631169. Norwegian sheep are an important reservoir for
46. Mannix M, OConnell N, McNamara E, human-pathogenic Escherichia coli O26:H11. Appl
Fitzgerald A, Prendiville T, Norris T, Greally T, Environ Microbiol 78:40834091.
Fitzgerald R, Whyte D, Barron D, Monaghan R, 56. Evans J, Knight H, McKendrick IJ, Stevenson
Whelan E, Carroll A, Curtin A, Collins C, Quinn H, Varo Barbudo A, Gunn GJ, Low JC. 2011.
J, ODea F, ORiordan M, Buckley J, McCarthy Prevalence of Escherichia coli O157:H7 and sero-
J, Mc Keown P. 2005. Large E. coli O157 out- groups O26, O103, O111 and O145 in sheep pre-
break in Ireland, October-November 2005. Euro sented for slaughter in Scotland. J Med Microbiol
Surveill 10:E051222 051223. 60:653660.
47. Joris MA, Pierard D, De Zutter L. 2011. Occur- 57. Ogden ID, MacRae M, Strachan NJ. 2005.
rence and virulence patterns of E. coli O26, O103, Concentration and prevalence of Escherichia coli
O111 and O145 in slaughter cattle. Vet Microbiol O157 in sheep faeces at pasture in Scotland. J Appl
151:418421. Microbiol 98:646651.
48. Franz E, van Hoek AH, van der Wal FJ, de Boer 58. Sekse C, Sunde M, Hopp P, Bruheim T, Cudjoe
A, Zwartkruis-Nahuis A, van der Zwaluw K, KS, Kvitle B, Urdahl AM. 2013. Occurrence of
Aarts HJ, Heuvelink AE. 2012. Genetic features potentially human-pathogenic Escherichia coli
differentiating bovine, food, and human isolates of O103 in Norwegian sheep. Appl Environ Microbiol
shiga toxin-producing Escherichia coli O157 in 79:75027509.
The Netherlands. J Clin Microbiol 50:772780. 59. Schimmer B, Nygard K, Eriksen HM, Lassen J,
49. Blanco M, Blanco JE, Mora A, Rey J, Alonso Lindstedt BA, Brandal LT, Kapperud G,
JM, Hermoso M, Hermoso J, Alonso MP, Aavitsland P. 2008. Outbreak of haemolytic
Dahbi G, Gonzalez EA, Bernardez MI, Blanco J. uraemic syndrome in Norway caused by stx2-
2003. Serotypes, virulence genes, and intimin positive Escherichia coli O103:H25 traced to cured
types of Shiga toxin (verotoxin)-producing Esch- mutton sausages. BMC Infect Dis 8:41.
erichia coli isolates from healthy sheep in Spain. 60. Lenahan M, OBrien S, Kinsella K, Sweeney T,
J Clin Microbiol 41:13511356. Sheridan JJ. 2007. Prevalence and molecular
50. Rey J, Blanco JE, Blanco M, Mora A, Dahbi G, characterization of Escherichia coli O157:H7 on
Alonso JM, Hermoso M, Hermoso J, Alonso Irish lamb carcasses, fleece and in faeces samples.
MP, Usera MA, Gonzalez EA, Bernardez MI, J Appl Microbiol 103:24012409.
Blanco J. 2003. Serotypes, phage types and vir- 61. Thomas KM, McCann MS, Collery MM,
ulence genes of shiga-producing Escherichia coli Moschonas G, Whyte P, McDowell DA, Duffy
isolated from sheep in Spain. Vet Microbiol 94: G. 2013. Transfer of verocytotoxigenic Escherichia
4756. coli O157, O26, O111, O103 and O145 from fleece to
51. Soderlund R, Hedenstrom I, Nilsson A, Eriksson carcass during sheep slaughter in an Irish export
E, Aspan A. 2012. Genetically similar strains of abattoir. Food Microbiol 34:3845.
Escherichia coli O157:H7 isolated from sheep, 62. Feder I, Wallace FM, Gray JT, Fratamico P,
cattle and human patients. BMC Vet Res 8:200. Fedorka-Cray PJ, Pearce RA, Call JE, Perrine
52. Sekse C, Sunde M, Lindstedt BA, Hopp P, R, Luchansky JB. 2003. Isolation of Escherichia
Bruheim T, Cudjoe KS, Kvitle B, Urdahl AM. coli O157:H7 from intact colon fecal samples of
2011. Potentially human-pathogenic Escherichia swine. Emerg Infect Dis 9:380383.

63. Chapman PA, Cerdan Malo AT, Ellin M, Ashton coli O157:H7 and Salmonella enterica serovar
R, Harkin. 2001. Escherichia coli O157 in cattle Typhimurium in Luria-Bertani broth, farm-
and sheep at slaughter, on beef and lamb yard manure and slurry. J Environ Manage 92:
carcasses and in raw beef and lamb products in 780787.
South Yorkshire, UK. Int J Food Microbiol 75. Nicholson FA, Groves SJ, Chambers BJ. 2005.
64:139150. Pathogen survival during livestock manure stor-
64. Callaway TR, Anderson RC, Tellez G, Rosario age and following land application. Bioresour
C, Nava GM, Eslava C, Blanco MA, Quiroz Technol 96:135143.
MA, Olguin A, Herradora M, Edrington TS, 76. Hutchison ML, Walters LD, Moore A, Crookes
Genovese KJ, Harvey RB, Nisbet DJ. 2004. KM, Avery SM. 2004. Effect of length of time
Prevalence of Escherichia coli O157 in cattle and before incorporation on survival of pathogenic
swine in central Mexico. J Food Prot 67:2274 bacteria present in livestock wastes applied to
2276. agricultural soil. Appl Environ Microbiol 70:5111
65. Chapman PA, Siddons CA, Gerdan Malo AT, 5118.
Harkin MA. 1997. A 1-year study of Escherichia 77. Avery SM, Moore A, Hutchison ML. 2004. Fate
coli O157 in cattle, sheep, pigs and poultry. of Escherichia coli originating from livestock
Epidemiol Infect 119:245250. faeces deposited directly onto pasture. Lett Appl
66. Eriksson E, Nerbrink E, Borch E, Aspan A, Microbiol 38:355359.
Gunnarsson A. 2003. Verocytotoxin-producing 78. Fukushima H, Hoshina K, Gomyoda M.
Escherichia coli O157:H7 in the Swedish pig pop- 1999. Long-term survival of shiga toxin-producing
ulation. Vet Rec 152:712717. Escherichia coli O26, O111, and O157 in bovine
67. Heuvelink AE, Zwartkruis-Nahuis JT, van den feces. Appl Environ Microbiol 65:51775181.
Biggelaar FL, van Leeuwen WJ, de Boer E. 1999. 79. Bolton DJ, Monaghan A, Byrne B, Fanning S,
Isolation and characterization of verocytotoxin- Sweeney T, McDowell DA. 2011. Incidence and
producing Escherichia coli O157 from slaughter survival of non-O157 verocytotoxigenic Esche-
pigs and poultry. Int J Food Microbiol 52:6775. richia coli in soil. J Appl Microbiol 111:484490.
68. Lenahan M, OBrien SB, Byrne C, Ryan M, 80. EC No 853/2004. 2004. Regulation (EC) No 853/
Kennedy CA, McNamara EB, Fanning S, 2004 of the European Parliament and of the
Sheridan JJ, Sweeney T. 2009. Molecular char- Council of 29 April 2004, laying down specific
acterization of Irish E. coli O157:H7 isolates of hygiene rules for food of animal origin. Official
human, bovine, ovine and porcine origin. J Appl Journal of the European Union L 139 of 30 April
Microbiol 107:13401349. 2004.
69. Jay MT, Cooley M, Carychao D, Wiscomb GW, 81. Lyons NA, Smith RP, Rushton J. 2013. Cost-
Sweitzer RA, Crawford-Miksza L, Farrar JA, effectiveness of farm interventions for reducing
Lau DK, OConnell J, Millington A, Asmundson the prevalence of VTEC O157 on UK dairy farms.
RV, Atwill ER, Mandrell RE. 2007. Escherichia Epidemiol Infect 141:19051919.
coli O157:H7 in feral swine near spinach fields and 82. Ellis-Iversen J, Smith RP, Van Winden S, Paiba
cattle, central California coast. Emerg Infect Dis GA, Watson E, Snow LC, Cook AJ. 2008. Farm
13:19081911. practices to control E. coli O157 in young cattlea
70. Williams AP, Avery LM, Killham K, Jones DL. randomised controlled trial. Vet Res 39:3.
2005. Persistence of Escherichia coli O157 on 83. Lung AJ, Lin CM, Kim JM, Marshall MR,
farm surfaces under different environmental con- Nordstedt R, Thompson NP, Wei CI. 2001.
ditions. J Appl Microbiol 98:10751083. Destruction of Escherichia coli O157:H7 and Sal-
71. Bolton DJ, Byrne CM, Sheridan JJ, McDowell monella enteritidis in cow manure composting.
DA, Blair IS. 1999. The survival characteristics of J Food Prot 64:13091314.
a non-toxigenic strain of Escherichia coli O157:H7. 84. Shepherd MW Jr, Kim J, Jiang X, Doyle
J Appl Microbiol 86:407411. MP, Erickson MC. 2011. Evaluation of physical
72. Maule A. 2000. Survival of verocytotoxigenic coverings used to control Escherichia coli O157:H7
Escherichia coli O157 in soil, water and on sur- at the compost heap surface. Appl Environ Micro-
faces. Symp Ser Soc Appl Microbiol71S78S. biol 77:50445049.
73. McGee P, Bolton DJ, Sheridan JJ, Earley B, 85. Bujoczek G, Reiners RS, Olaszkiewicz JA. 2001.
Leonard N. 2001. The survival of Escherichia coli Abiotic factors affecting inactivation of pathogens
O157:H7 in slurry from cattle fed different diets. in sludge. Water Sci Technol 44:7984.
Lett Appl Microbiol 32:152155. 86. Locking M, Allison L, Rae L, Pollock K, Hanson
74. Semenov AV, van Overbeek L, Termorshuizen M. 2006. VTEC infections and livestock-related
AJ, van Bruggen AH. 2011. Influence of aerobic exposures in Scotland, 2004. Euro Surveill 11:
and anaerobic conditions on survival of Escherichia E060223 060224.

87. Money P, Kelly AF, Gould SW, Denholm- review and meta-analysis of the effects of pas-
Price J, Threlfall EJ, Fielder MD. 2010. Cattle, teurization on milk vitamins, and evidence for
weather and water: mapping Escherichia coli raw milk consumption and other health-related
O157:H7 infections in humans in England and outcomes. J Food Prot 74:18141832.
Scotland. Environ Microbiol 12:26332644. 99. Jay-Russell MT. 2010. Raw (unpasteurized) milk:
88. Health Protection Surveillance Centre (HPSC). are health-conscious consumers making an un-
2005. Report of the HPSC Sub-Committee on healthy choice? Clin Infect Dis 51:14181419.
Verotoxigenic E. coli. http://www.hpsc.ie/A-Z/ 100. Denny J, Bhat M, Eckmann K. 2008. Outbreak
Gastroenteric/VTEC/Guidance/Reportofthe of Escherichia coli O157:H7 associated with
HPSCSub-CommitteeonVerotoxigenicEcoli/. raw milk consumption in the Pacific Northwest.
89. OSullivan MB, Garvey P, ORiordan M, Coughlan Foodborne Pathog Dis 5:321328.
H, McKeown P, Brennan A, McNamara E. 2008. 101. Guh A, Phan Q, Randall N, Purviance K,
Increase in VTEC cases in the south of Ireland: link Milardo E, Kinney S, Mshar P, Kasacek W,
to private wells? Euro Surveill 13. Cartter M. 2010. Outbreak of Escherichia coli
90. Tanaro JD, Galli L, Lound LH, Leotta GA, O157 associated with raw milk, Connecticut,
Piaggio MC, Carbonari CC, Irino K, Rivas M. 2008. Clin Infect Dis 51:14111417.
2012. Non-O157:H7 Shiga toxin-producing Esch- 102. Gaulin C, Levac E, Ramsay D, Dion R, Ismail J,
erichia coli in bovine rectums and surface water Gingras S, Lacroix C. 2012. Escherichia coli
streams on a beef cattle farm in Argentina. O157:H7 outbreak linked to raw milk cheese in
Foodborne Pathog Dis 9:878884. Quebec, Canada: use of exact probability calcula-
91. Department of the Environment and Local tion and casecase study approaches to foodborne
Government, Environmental Protection Agen- outbreak investigation. J Food Prot 75:812818.
cy, Geological Survey of Ireland (DELG, EPA, 103. Honish L, Predy G, Hislop N, Chui L, Kowalewska-
GSI). 1999. Groundwater Protection Schemes Re- Grochowska K, Trottier L, Kreplin C, Zazulak I.
port. GSI, Dublin, Ireland. 2005. An outbreak of E. coli O157:H7 hemorrhagic
92. Food Safety Authority of Ireland. 2010. Report colitis associated with unpasteurized gouda cheese.
of the Scientific Committee of the Food Safety Can J Public Health 96:182184.
Authority of Ireland. The Prevention of Vero- 104. McCollum JT, Williams NJ, Beam SW, Cosgrove
cytotoxigenic Escherichia coli (VTEC) Infection: A S, Ettestad PJ, Ghosh TS, Kimura AC, Nguyen L,
Shared Responsibility, 2nd Ed. FSAI, Dublin, Stroika SG, Vogt RL, Watkins AK, Weiss JR,
Ireland. Williams IT, Cronquist AB. 2012. Multistate out-
93. Wang G, Doyle MP. 1998. Heat shock response break of Escherichia coli O157:H7 infections asso-
enhances acid tolerance of Escherichia coli ciated with in-store sampling of an aged raw-milk
O157:H7. Lett Appl Microbiol 26:3134. Gouda cheese, 2010. J Food Prot 75:17591765.
94. McGee P, Bolton DJ, Sheridan JJ, Earley B, 105. Rangel JM, Sparling PH, Crowe C, Griffin PM,
Kelly G, Leonard N. 2002. Survival of Escherichia Swerdlow DL. 2005. Epidemiology of Escherichia
coli O157:H7 in farm water: its role as a vector in coli O157:H7 outbreaks, United States, 19822002.
the transmission of the organism within herds. Emerg Infect Dis 11:603609.
J Appl Microbiol 93:706713. 106. Allerberger F, Friedrich AW, Grif K, Dierich
95. Rice EW, Johnson CH, Wild DK, Reasoner DJ. MP, Dornbusch HJ, Mache CJ, Nachbaur E,
1992. Survival of Escherichia coli O157:H7 in Freilinger M, Rieck P, Wagner M, Caprioli A,
drinking water associated with a waterborne Karch H, Zimmerhackl LB. 2003. Hemolytic-
disease outbreak of hemorrhagic colitis. Lett Appl uremic syndrome associated with enterohemor-
Microbiol 15:3840. rhagic Escherichia coli O26:H infection and
96. Fremaux B, Prigent-Combaret C, Delignette- consumption of unpasteurized cows milk. Int
Muller ML, Mallen B, Dothal M, Gleizal A, J Infect Dis 7:4245.
Vernozy-Rozand C. 2008. Persistence of Shiga 107. De Schrijver K, Buvens G, Posse B, Van den
toxin-producing Escherichia coli O26 in various Branden D, Oosterlynck O, De Zutter L, Eilers K,
manure-amended soil types. J Appl Microbiol Pierard D, Dierick K, Van Damme-Lombaerts
104:296304. R, Lauwers C, Jacobs R. 2008. Outbreak of
97. Scott L, McGee P, Sheridan JJ, Earley B, verocytotoxin-producing E. coli O145 and O26 in-
Leonard N. 2006. A comparison of the survival in fections associated with the consumption of ice
feces and water of Escherichia coli O157:H7 grown cream produced at a farm, Belgium, 2007. Euro
under laboratory conditions or obtained from Surveill 13.
cattle feces. J Food Prot 69:611. 108. Centers for Disease Control and Prevention.
98. Macdonald LE, Brett J, Kelton D, Majowicz SE, 2012. Outbreak of Shiga toxin-producing Esche-
Snedeker K, Sargeant JM. 2011. A systematic richia coli O111 infections associated with a

correctional facility dairyColorado, 2010. Morb 119. King LA, Mailles A, Mariani-Kurkdjian P,
Mortal Wkly Rep 61:149152. Vernozy-Rozand C, Montet MP, Grimont F,
109. Murphy M, Carroll A, Whyte P, OMahony M, Pihier N, Devalk H, Perret F, Bingen E, Espie E,
Anderson W, McNamara E, Fanning S. 2005. Vaillant V. 2009. Community-wide outbreak of
Prevalence and characterization of Escherichia Escherichia coli O157:H7 associated with con-
coli O26 and O111 in retail minced beef in Ireland. sumption of frozen beef burgers. Epidemiol Infect
Foodborne Pathog Dis 2:357360. 137:889896.
110. Murphy M, Buckley JF, Whyte P, OMahony 120. Elder RO, Keen JE, Siragusa GR, Barkocy-
M, Anderson W, Wall PG, Fanning S. 2007. Gallagher GA, Koohmaraie M, Laegreid WW.
Surveillance of dairy production holdings sup- 2000. Correlation of enterohemorrhagic Esche-
plying raw milk to the farmhouse cheese sector richia coli O157 prevalence in feces, hides, and
for Escherichia coli O157, O26 and O111. Zoonoses carcasses of beef cattle during processing. Proc
Public Health 54:358365. Natl Acad Sci USA 97:29993003.
111. European Food Safety Authority, European 121. Arthur TM, Bosilevac JM, Nou X, Shackelford
Centre for Disease Prevention and Control. SD, Wheeler TL, Kent MP, Jaroni D, Pauling B,
2012. The European Union Summary Report on Allen DM, Koohmaraie M. 2004. Escherichia coli
Trends and Sources of Zoonoses, Zoonotic Agents O157 prevalence and enumeration of aerobic
and Food-borne Outbreaks in 2010. EFSA Journal bacteria, Enterobacteriaceae, and Escherichia
10(3):2597. coli O157 at various steps in commercial beef
112. Schvartzman MS, Maffre A, Tenenhaus-Aziza processing plants. J Food Prot 67:658665.
F, Sanaa M, Butler F, Jordan K. 2011. Modelling 122. McGee P, Scott L, Sheridan JJ, Earley B,
the fate of Listeria monocytogenes during manu- Leonard N. 2004. Horizontal transmission of
facture and ripening of smeared cheese made Escherichia coli O157:H7 during cattle housing.
with pasteurised or raw milk. Int J Food Micro- J Food Prot 67:26512656.
biol 145 Suppl 1:S31S38. 123. Arthur TM, Bosilevac JM, Brichta-Harhay DM,
113. Peng S, Hoffmann W, Bockelmann W, Kalchayanand N, King DA, Shackelford SD,
Hummerjohann J, Stephan R, Hammer P. 2013. Wheeler TL, Koohmaraie M. 2008. Source
Fate of Shiga toxin-producing and generic Esch- tracking of Escherichia coli O157:H7 and Salmo-
erichia coli during production and ripening of nella contamination in the lairage environment
semihard raw milk cheese. J Dairy Sci 96:815 at commercial U.S. beef processing plants and
823. identification of an effective intervention. J Food
114. Miszczycha SD, Perrin F, Ganet S, Jamet E, Prot 71:17521760.
Tenenhaus-Aziza F, Montel MC, Thevenot- 124. Arthur TM, Bosilevac JM, Nou X, Shackelford
Sergentet D. 2013. Behavior of different Shiga SD, Wheeler TL, Koohmaraie M. 2007. Com-
toxin-producing Escherichia coli serotypes in parison of the molecular genotypes of Escherichia
various experimentally contaminated raw-milk coli O157:H7 from the hides of beef cattle in dif-
cheeses. Appl Environ Microbiol 79:150158. ferent regions of North America. J Food Prot
115. Dineen SS, Takeuchi K, Soudah JE, Boor KJ. 70:16221626.
1998. Persistence of Escherichia coli O157:H7 in 125. Dewell GA, Simpson CA, Dewell RD, Hyatt DR,
dairy fermentation systems. J Food Prot 61:1602 Belk KE, Scanga JA, Morley PS, Grandin T,
1608. Smith GC, Dargatz DA, Wagner BA, Salman
116. McIngvale SC, Chen XQ, McKillip JL, MD. 2008. Impact of transportation and lairage
Drake MA. 2000. Survival of Escherichia coli on hide contamination with Escherichia coli O157
O157:H7 in buttermilk as affected by contamina- in finished beef cattle. J Food Prot 71:11141118.
tion point and storage temperature. J Food Prot 126. Jacob ME, Renter DG, Nagaraja TG. 2010. An-
63:441444. imal- and truckload-level associations between
117. Currie A, MacDonald J, Ellis A, Siushansian J, Escherichia coli O157:H7 in feces and on hides at
Chui L, Charlebois M, Peermohamed M, harvest and contamination of preevisceration beef
Everett D, Fehr M, Ng LK. 2007. Outbreak of carcasses. J Food Prot 73:10301037.
Escherichia coli 0157:H7 infections associated 127. European Food Safety Authority, EFSA Panel
with consumption of beef donair. J Food Prot on Biological Hazards. 2013. Scientific Opinion
70:14831488. on the public health hazards to be covered by
118. Ethelberg S, Lisby M, Vestergaard LS, Enemark inspection of meat (bovine animals). EFSA Jour-
HL, Olsen KE, Stensvold CR, Nielsen HV, nal 11(6):3266.
Porsbo LJ, Plesner AM, Molbak K. 2009. A 128. Schmidt JW, Arthur TM, Bosilevac JM,
foodborne outbreak of Cryptosporidium hominis Kalchayanand N, Wheeler TL. 2012. Detection
infection. Epidemiol Infect 137:348356. of Escherichia coli O157:H7 and Salmonella

enterica in air and droplets at three U.S. com- of Escherichia coli O157:H7 in pepperoni. Appl
mercial beef processing plants. J Food Prot 75: Environ Microbiol 66:17261729.
22132218. 140. Lupp C, Robertson ML, Wickham ME, Sekirov
129. Koohmaraie M, Arthur TM, Bosilevac JM, I, Champion OL, Gaynor EC, Finlay BB. 2007.
Guerini M, Shackelford SD, Wheeler TL. 2005. Host-mediated inflammation disrupts the intes-
Post-harvest interventions to reduce/eliminate tinal microbiota and promotes the overgrowth of
pathogens in beef. Meat Sci 71:7991. Enterobacteriaceae. Cell Host Microbe 2:119129.
130. Kalchayanand N, Arthur TM, Bosilevac JM, 141. Vasan A, Leong WM, Ingham SC, Ingham BH.
Schmidt JW, Wang R, Shackelford SD, Wheeler 2013. Thermal tolerance characteristics of non-
TL. 2012. Evaluation of commonly used antimi- O157 Shiga toxigenic strains of Escherichia coli
crobial interventions for fresh beef inoculated (STEC) in a beef broth model system are similar
with Shiga toxin-producing Escherichia coli sero- to those of O157:H7 STEC. J Food Prot 76:1120
types O26, O45, O103, O111, O121, O145, and 1128.
O157:H7. J Food Prot 75:12071212. 142. Launders N, Byrne L, Adams N, Glen K,
131. Jacob R, Porto-Fett AC, Call JE, Luchansky JB. Jenkins C, Tubin-Delic D, Locking M, Williams
2009. Fate of surface-inoculated Escherichia coli C, Morgan D. 2013. Outbreak of Shiga toxin-
O157:H7, Listeria monocytogenes, and Salmonella producing E. coli O157 associated with consump-
typhimurium on kippered beef during extended tion of watercress, United Kingdom, August to
storage at refrigeration and abusive temperatures. September 2013. Euro Surveill 18.
J Food Prot 72:403407. 143. Soborg B, Lassen SG, Muller L, Jensen T,
132. Adler JM, Geornaras I, Belk KE, Smith GC, Ethelberg S, Molbak K, Scheutz F. 2013. A
Sofos JN. 2012. Thermal inactivation of Esche- verocytotoxin-producing E. coli outbreak with a
richia coli O157:H7 inoculated at different depths surprisingly high risk of haemolytic uraemic
of non-intact blade-tenderized beef steaks. J Food syndrome, Denmark, September-October 2012.
Sci 77:M108M114. Euro Surveill 18.
133. Kudra LL, Sebranek JG, Dickson JS, Mendonca 144. Slayton RB, Turabelidze G, Bennett SD,
AF, Larson EM, Jackson-Davis AL, Lu Z. 2011. Schwensohn CA, Yaffee AQ, Khan F, Butler C,
Effects of vacuum or modified atmosphere pack- Trees E, Ayers TL, Davis ML, Laufer AS,
aging in combination with irradiation for control Gladbach S, Williams I, Gieraltowski LB. 2013.
of Escherichia coli O157:H7 in ground beef patties. Outbreak of Shiga toxin-producing Escherichia
J Food Prot 74:20182023. coli (STEC) O157:H7 associated with romaine
134. Vold L, Holck A, Wasteson Y, Nissen H. 2000. lettuce consumption, 2011. PLoS One 8:e55300.
High levels of background flora inhibits growth 145. Taylor EV, Nguyen TA, Machesky KD, Koch E,
of Escherichia coli O157:H7 in ground beef. Int J Sotir MJ, Bohm SR, Folster JP, Bokanyi R,
Food Microbiol 56:219225. Kupper A, Bidol SA, Emanuel A, Arends KD,
135. Luchansky JB, Porto-Fett AC, Shoyer BA, Johnson SA, Dunn J, Stroika S, Patel MK,
Phillips J, Eblen D, Evans P, Bauer N. 2013. Williams I. 2013. Multistate outbreak of Esche-
Thermal inactivation of a single strain each of richia coli O145 infections associated with ro-
serotype O26:H11, O45:H2, O103:H2, O104:H4, maine lettuce consumption, 2010. J Food Prot
O111:H(-), O121:H19, O145:NM, and O157:H7 cells 76:939944.
of Shiga toxin-producing Escherichia coli in 146. Greenland K, de Jager C, Heuvelink A, van der
wafers of ground beef. J Food Prot 76:14341437. Zwaluw K, Heck M, Notermans D, van Pelt W,
136. Flores RA, Tamplin ML. 2002. Distribution Friesema I. 2009. Nationwide outbreak of STEC
patterns of Escherichia coli O157:H7 in ground O157 infection in the Netherlands, December
beef produced by a laboratory-scale grinder. 2008-January 2009: continuous risk of consum-
J Food Prot 65:18941902. ing raw beef products. Euro Surveill 14. pii: 19129.
137. Flores RA. 2004. Distribution of Escherichia coli 147. Neil KP, Biggerstaff G, MacDonald JK, Trees E,
O157:H7 in beef processed in a table-top bowl Medus C, Musser KA, Stroika SG, Zink D, Sotir
cutter. J Food Prot 67:246251. MJ. 2012. A novel vehicle for transmission of
138. Riordan DC, Duffy G, Sheridan J, Eblen BS, Escherichia coli O157:H7 to humans: multistate
Whiting RC, Blair IS, McDowell DA. 1998. outbreak of E. coli O157:H7 infections associated
Survival of Escherichia coli O157:H7 during with consumption of ready-to-bake commercial
the manufacture of pepperoni. J Food Prot 61: prepackaged cookie doughUnited States, 2009.
146151. Clin Infect Dis 54:511518.
139. Riordan DC, Duffy G, Sheridan JJ, Whiting RC, 148. Centers for Disease Control and Prevention.
Blair IS, McDowell DA. 2000. Effects of acid 2010. Two multistate outbreaks of Shiga toxin
adaptation, product pH, and heating on survival producing Escherichia coli infections linked to

beef from a single slaughter facilityUnited standards in bovine carcasses from an industrial
States, 2008. Morb Mortal Wkly Rep 59:557 beef slaughter plant. Lett Appl Microbiol 56:408
560. 413.
149. Friesema I, Sigmundsdottir G, van der Zwaluw 153. Dargatz DA, Bai J, Lubbers BV, Kopral CA, An
K, Heuvelink A, Schimmer B, de Jager C, Rump B, Anderson GA. 2013. Prevalence of Escherichia
B, Briem H, Hardardottir H, Atladottir A, coli O-types and Shiga toxin genes in fecal sam-
Gudmundsdottir E, van Pelt W. 2008. An in- ples from feedlot cattle. Foodborne Pathog Dis
ternational outbreak of Shiga toxin-producing 10:392396.
Escherichia coli O157 infection due to lettuce, 154. Narvaez-Bravo C, Miller MF, Jackson T, Jackson
September-October 2007. Euro Surveill 13. pii: S, Rodas-Gonzalez A, Pond K, Echeverry A,
19065. Brashears MM. 2013. Salmonella and Escherichia
150. Ethelberg S, Smith B, Torpdahl M, Lisby M, coli O157:H7 prevalence in cattle and on carcasses
Boel J, Jensen T, Nielsen EM, Molbak K. 2009. in a vertically integrated feedlot and harvest plant in
Outbreak of non-O157 Shiga toxin-producing Mexico. J Food Prot 76:786795.
Escherichia coli infection from consumption of 155. Breum SO, Boel J. 2010. Prevalence of Esche-
beef sausage. Clin Infect Dis 48:e78e81. richia coli O157 and verocytotoxin producing
151. Wendel AM, Johnson DH, Sharapov U, Grant J, E. coli (VTEC) on Danish beef carcasses. Int J
Archer JR, Monson T, Koschmann C, Davis JP. Food Microbiol 141:9096.
2009. Multistate outbreak of Escherichia coli 156. Masana MO, Leotta GA, Del Castillo LL,
O157:H7 infection associated with consumption of DAstek BA, Palladino PM, Galli L, Vilacoba E,
packaged spinach, August-September 2006: the Carbonari C, Rodriguez HR, Rivas M. 2010.
Wisconsin investigation. Clin Infect Dis 48:1079 Prevalence, characterization, and genotypic anal-
1086. ysis of Escherichia coli O157:H7/NM from select-
152. Ramoneda M, Foncuberta M, Simon M, Sabate ed beef exporting abattoirs of Argentina. J Food
S, Ferrer MD, Herrera S, Landa B, Muste N, Prot 73:649656.
Marti R, Trabado V, Carbonell O, Vila M, 157. Boqvist S, Aspan A, Eriksson E. 2009. Preva-
Espelt M, Ramirez B, Duran J. 2013. Prevalence lence of verotoxigenic Escherichia coli O157:H7 in
of verotoxigenic Escherichia coli O157 (VTEC fecal and ear samples from slaughtered cattle in
O157) and compliance with microbiological safety Sweden. J Food Prot 72:17091712.

Clinical Studies of Escherichia coli
O157:H7 Conjugate Vaccines in
Adults and Young Children
SHOUSUN CHEN SZU1 and AMINA AHMED2

24
INTRODUCTION
Shiga toxin (Stx)-producing Escherichia coli (STEC) is a food-borne pathogen

that can lead to complications such as hemorrhagic colitis and hemolytic-
uremic syndrome (HUS), serious sequelae. In the United States, the most
common E. coli serotype causing outbreaks is O157:H7, although non-O157
serotypes also cause the same disease, but in much fewer cases. The highest
incidence rate is among children of preschool age (1, 2).
Prevention of E. coli O157 infection has been difficult because of the broad
spectrum of contaminated sources, ranging from food such as beef, milk, pro-
duce, and fruits, to nonfood origins such as pool water and petting zoo animals
(3, 4). Chemical or antimicrobial interventions are difficult to apply and have
shown limited effectiveness (5). Since cattle are the major animal reservoir
for E. coli O157, cattle vaccines to reduce carriage of E. coli O157 also were
studied to a great extent and reached moderate success (6, 7).
An ideal E. coli O157 vaccine for humans should be safe and immunogenic in
children and elicit bactericidal antibody that kills the inoculum upon contact.
The infectious dose found in E. coli O157 outbreaks was usually low, around
1
Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health,
Bethesda, MD 20892; 2Levine Childrens Specialty CenterPediatric Infectious Disease, Carolina Medical Centers,
Charlotte, NC 28203.
477

478 SZU AND AHMED
102 CFU (8). High levels of serum lipopoly- METHODS AND RESULTS
saccharide (LPS) antibodies are detected
following environmental exposure or symp- Investigational LPS-Based Vaccines
tomatic infection with E. coli O157 (9, 10).
The O-antigen of E. coli O157 consists of a
Although the protective role of these anti-
linear copolymer of the tetrasaccharide re-
bodies is unknown, evidence suggests that
peating unit, a-D-GalpNAc(1-2)-a-D-PerpNAc-
antibodies to the LPS of other enteric patho-
(1-3)-a-L-Fucp-(1-4)-b-D-Glcp-(1-3), where
gens confer immunity by lysing the pathogens
PerpNAc represents perosamine, 4-amino-
in the intestine (11). Convalescence from
4,6-dideoxy-D-mannose (17). The O-antigen
shigellosis, for example, provides LPS-specific
can be extracted from LPS and detoxified by
resistance to infection, and vaccination with
acetic acid (yielding O-specific polysaccharide
a Shigella sonnei LPS-based conjugate is effi-
[OSP]) or by anhydrous hydrazine to remove
cacious in preventing infection (12). In clin-
the O-acyl-linked lipid chains (yielding
ical trials of a Salmonella enterica serovar
DeALPS) (13). Both OSP and DeALPS are
Paratyphi A LPS-based conjugate vaccine,
not immunogenic without conjugation to a
high levels of immunoglobulin G (IgG) anti-
carrier protein. Three investigational vaccines
LPS with bactericidal activities were also in-
had been prepared by covalently linking to the
duced in adults and preschool children (13).
carrier protein rEPA and designated as O157
Because of the similarities between E. coli
OSP-rEPA1, OSP-rEPA2, and DeALPS-rEPA.
O157 and these gram-negative enteropatho-
The carrier protein rEPA was chosen because
gens, vaccine-induced LPS antibodies to E. coli
it has been demonstrated to be clinically safe
O157 may confer protection (14, 15).
and served effectively in several polysaccha-
Based on this postulation, our goal is to
ride conjugates (21). It also has the advantage
develop vaccines that elicit serum bactericidal
of not overloading the already crowded rou-
IgG in young children. Several investigational
tine vaccines with additional doses of diph-
polysaccharide conjugate vaccines were pre-
theria or tetanus toxoids.
pared at the National Institutes of Health. The
The conjugates passed preclinical immu-
vaccines were composed of detoxified LPS for
nogenicity tests in mice and followed the
E. coli O157, covalently linked to the carrier
safety requirements of the Code of Federal
protein rEPA, a recombinant exoprotein of
Regulations and were approved by the U.S.
Pseudomonas aeruginosa. Here we review the
Food and Drug Administration as investiga-
clinical studies of O157-rEPA investigational
tional vaccines. Each 0.5-ml dose contained
vaccines conducted in adults and children 2 to
25 mg of polysaccharide and an approximately
5 years old (1619).
equal amount of rEPA. This dosage followed
One of the major virulence factors in en-
established studies on conjugate vaccines for
terohemorrhagic E. coli (EHEC) is the Stx
noncapsular polysaccharides and is slightly
secreted by both the O157 and non-O157
higher than the licensed Haemophilus in-
serotypes. Therefore, an ideal human vac-
fluenzae type b, pnemococcal, or meningo-
cine is one that elicits neutralizing antibody
coccal conjugate vaccines (12, 13).
against Stx. Passive immunization with neu-
tralizing antibody remains the only effective
therapy for many other toxin-mediated dis-
Clinical Studies
eases (20). However, up until the present,
the only data supporting the development of Both the phase I and phase II clinical studies
Stx-based vaccines for both active and passive of O157 conjugate vaccines were conducted
immunization are in preclinical stage. The at the Carolina Medical Centers, Charlotte,
Stx-based vaccine is reviewed briefly in this NC. Briefly, in phase I, 87 healthy adults
article. were injected once with one of the three

CHAPTER 24 Clinical Studies of O157:H7 Conjugate Vaccines 479
conjugates (18). After safety and immunoge- injection, with 82% having greater than a
nicity were demonstrated, the conjugate that 4-fold rise (Fig. 1). The escalation of anti-
elicited the highest antibody was chosen for bodies shortly after immunization is impor-
the phase II study where 49 children, 2 to tant since the vaccine could be considered as a
5 years old, were recruited and divided into useful control measure during outbreaks be-
two groups receiving one or two doses of fore the source of contamination is identified
the vaccine (19). and to block primary or secondary trans-
The E. coli O157 conjugate vaccines were missions (29). The antibody levels continued
safe for all ages. There were no fever cases to rise 4 weeks after the injection, and the
except for one child who had a temperature geometric mean (GM) in the recipients of
of 38.2C 72 h after the second dose was ad- OSP-rEPA was slightly higher than that in-
ministered. The local reactions were all mild duced by the conjugate prepared with hydra-
and subsided within 24 h. In phase I, serum zine-treated LPS, DeALPS-rEPA. At 6 months,
assays, including lactate dehydrogenase or the levels of anti-LPS IgG waned to 33
alkaline phosphatase, serum bilirubin, and ELISA units (EU) for all three conjugates,
indirect bilirubin, were performed 1 week with 97% of volunteers continuing to have
after injection to evaluate liver function. Six greater than a 10-fold rise than the pre-
(7%) had asymptomatic elevations (up to injection levels.
35% above the normal range) in one or more To a lesser degree than the IgG response,
serum assays that returned to normal within O157-rEPA conjugates also induced increases
5 weeks. There were no significant differ- in serum anti-LPS IgM and IgA levels. Inter-
ences in serum transaminase levels between estingly, there is no correlation between the
pre- and postvaccination in children in the serum IgG and IgM or IgA antibody titers.
trial. The highest incidence of HUS caused by
E. coli O157 infection occurred in children
under 6 years of age (3032). We chose this
Antibody Responses
target age group, children 2 to 5 years old, to
The serum anti-LPS IgG responses and bac- study the safety and immunogenicity of our
tericidal titers in vaccinees were used as the O157-rEPA conjugate. Children had very low
end-point markers for vaccine evaluation. The anti-LPS IgG preinjection levels; however,
responses were examined before and 1, 4, and within 1 week of injection, 81% responded
26 weeks after immunization for adults and 1, with a greater than 4-fold rise in their serum
6, 10, and 26 weeks after the first injection for anti-LPS IgG levels (Fig. 2). The antibody
children.
All adults had low levels of preimmune
anti-LPS IgG (measured in enzyme-linked
immunosorbent assay [ELISA] units), and this
level was approximately two times higher
than those detected in young children (vide
infra) (10, 18). The higher background level in
adults could be a result of longer environ-
mental exposure to cross-reactive organisms
containing perosamine residue in their LPS,
such as Citrobacter species, Yersinia entero-
FIGURE 1 Serum anti-Vi IgG response in healthy
colitica, Salmonella urbana, Pseudomonas
adults receiving one injection of OSP-rEPA (striped
maltophilia, and Brucella melitensis (2228). bars), DeALPS-rEPA1 (dotted bars), or DeALPS-rEPA2
All three conjugates elicited a significant (solid bars); n = 29 in each group.
rise of anti-LPS IgG in just 1 week after the doi:10.1128/microbiolspec.EHEC-0016-2013.f1

480 SZU AND AHMED
representative serum samples from children

before and 26 weeks after the first injection.
After the sera were treated with 2-mercapto-
ethanol to inactivate IgM function, we ob-
served that there was a direct correlation
between the level of IgG anti-LPS and the
bactericidal titers (R2 = 0.78).
There is a possibility for the O157 LPS-
based vaccines to protect against other path-
ogens that have cross-reactive LPS (22, 24).
FIGURE 2 Serum anti-Vi IgG response in children 2 to For instance, the LPS of Vibrio cholerae O:1
5 years old receiving one injection (dotted bars) or a constitutes a monosaccharide repeat of pero-
booster dose 6 weeks later (solid bars); n = 25 in each
samine, coinciding with one of the four sugars
group. No statistical difference between the groups
at all periods. in E. coli O157 O-antigen. It has been reported
doi:10.1128/microbiolspec.EHEC-0016-2013.f2 that LPS antibodies against V. cholerae O:1
cross-react with E. coli O157 (22). We also
observed some low-level cross-reaction in
levels continued to rise; 6 weeks after the first sera from children injected with O157-rEPA
injection, the GM reached 11.36 EU, with 98% with V. cholerae O:1 serotype Inaba, but not
having >10-fold increase compared with the with serotype Ogawa. Interestingly, there
preinjection levels (one child had a 6-fold is no correlation between the anti-LPS titers
rise). At all time intervals, the postinjection to E. coli O157 and those to V. cholerae
GM of anti-LPS IgG is significantly higher (R2<0.2) (19).
than the GM of the preinjection level.
Children who received a second injection
TABLE 1 Reciprocal bactericidal activity of serum
of O157-rEPA at week 6 had an increase of LPS antibodies elicited in 2- to 5-year-old children
antibodies measured 4 weeks later. However, injected with E. coli O157-rEPA conjugatesa
at 26 weeks there was no difference in anti- Bactericidal titerb
LPS IgG levels between the groups receiving IgG titer IgM titer Whole Serum
one or two injections. The lack of a booster Volunteer (IgG ELISA (IgM ELISA serum treated
response in this age group was also observed no. units) units) with 2-MEc
in other polysaccharide conjugate vaccines, ECO 161 100 100 1:1280 1:640
such as Salmonella serovar Paratyphi A and 023 33.63 6.81 1:2560 1:1,280
033 24.44 7.82 1:320 1:160
Shigella flexneri type 2a (12, 13). At 26 weeks,
035 17.21 7.72 1:320 1:320
all children in the study except one continued 041 14.23 3.76 1:160 1:160
to have >4-fold higher anti-LPS IgG than their 043 11.28 3.12 1:320 1:80
preinjection levels (Fig. 2). 047 12.98 18.94 1:320 1:80
The serum bactericidal assay has been a 048 21.87 12.25 1:160 1:160
054 18.55 16.00 1:320 1:160
reliable and reproducible functional bioassay
055 21.60 4.61 1:160 1:80
for gram-negative organisms such as Salmo- a
Sera collected from children 42 days after the rst injection of
nella serovar Typhimurium and Neisseria the conjugate vaccine.
b
meningitidis group C and serves as a good sur- Titers are expressed as the inverse of dilution giving 50% killing.
rogate for protection (33, 34). The assay is The anti-LPS IgG titers are calculated using a reference serum
randomly assigned 100 EU for IgG (ECO 161). Similarly, a separate
mediated by antibody- and complement- reference serum assigned 100 EU for IgM (ECO 110). Correlation
induced lysis of the bacterial cells, mimicking coefcient for IgG vs. serum, R2 = 0.75; for IgG vs. serum treated
the killing of the inoculums in vivo. In Table 1 with 2-mercarptoethanol (2-ME), R2 = 0.78. All bactericidal titers
compared with IgM, no correlation (R2<0.10).
we list the bactericidal titers and corre- c
Sera treated with 50 mM 2-ME. Reference serum (ECO 161) was
sponding levels of anti-LPS IgG and IgM in from the phase I study in adults injected with the OSP-rEPA.

The most essential virulence factor of and conjugates synthesized with O-antigen
STEC is Stx, in particular type 2. Its role as were shown to be safe and immunogenic and
both a prophylactic and a therapeutic antigen elicited bactericidal antibodies in young chil-
has also been observed in animal models and dren. Shigella sp. and E. coli are closely related
in epidemiology findings (9, 35). The major in genetics and pathogenicity (14, 15). Evi-
hindrance in development of an Stx2-based dence from S. sonnei and S. flexneri type 2b
vaccine is the difficulty in producing a suffi- efficacy trials showed that OSP conjugate
cient amount for vaccine preparation. In a vaccines, based on the same construct, are
proof-of-principle test, we conjugated OSP efficacious against similar disease even during
with the nontoxic recombinant B subunit of an outbreak (29).
Stx1 (Stx1B) by two methods, by direct at- We have noticed an age-dependent anti-
tachment or by linker adipic dihydrazide. LPS IgG response between pre- and postin-
Weaning mice injected with either conjugate jection sera. The higher background level of
elicited bactericidal antibodies to E. coli O157 anti-LPS in adults compared to that in chil-
and neutralizing antibodies to holotoxin Stx1 dren was also noticed by others (10). These
in vitro (Table 2). However, there was no preexisting LPS antibodies could come from
observed cross-neutralization against Stx2 prolonged exposure to other gram-negative
(36). Mutants with various promoters for organisms containing cross-reactive LPS, and
Stx2B fragments have been constructed and may in turn explain the age-related incidence
offered future potential in this approach (un- rate of E. coli O157 (1, 2, 30, 31, 32). Adults also
published data). responded with about 4-fold higher anti-LPS
IgG than children at all time intervals. How-
ever, at 26 weeks, the level of antibodies in
DISCUSSION AND FUTURE DEVELOPMENT children remained about 12 times higher than
the adults preimmune level, implying that the
The simple thesis of this review is to demon- children after vaccination had elevated im-
strate the possibility that, similar to polio, munity to E. coli O157.
typhoid fever, and cholera, parenterally ad- There are advantages to using LPS as the
ministered vaccines can protect against orally vaccine source: the O-specific antigen is sta-
transmitted diseases such as infections with ble, its purity and chemical composition can
E. coli O157. The outermost carbohydrate be identified unambiguously, the polysaccha-
moiety of E. coli O157 is the O-antigen of LPS, ride can be produced in large quantity and is
suitable for vaccine production, the detoxifi-
cation procedures are well established, the
TABLE 2 Neutralization titers of Stx1 in sera from
mice injected with E. coli O157 OSP conjugated residual endotoxicity level can be validated to
with Stx1B meet the requirements of the regulatory guide-
Titer to Stx1a lines, and, most of all, serum LPS antibodies
Immunogen n GM (2575%) Titer to Stx2 elicited by the conjugates demonstrated bac-
E. coli O157 LPS 5 <10 <10 tericidal activity against E. coli O157 and can
OSP-AH-Stx1Bb 10 8,040 <10 be adopted as a functional bioassay for po-
(6,40015,250) tency test.
OSP-Stx1Bc 10 14,400 <10
However, there are limitations of LPS-
(12,25018,600)
a
based vaccines. For example, the induced LPS
Neutralization of Stx1 or Stx2 was measured by using HeLa cell
monolayers incubated with dilutions of 100 pg of toxin/ml of antibodies do not neutralize Stx, the major
serum. Sera from mice immunized with saline or LPS alone virulence factor of EHEC. This shortcoming
showed titers of <10. The titers are the highest serum dilutions to limits its usefulness in prophylaxis and treat-
yield 50% neutralization. 14,440 vs. 8,040, P<0.03.
b
O-polysaccharide conjugated with Shiga toxin 1B with a linker. ment during an outbreak, especially for non-
c
O-polysaccharide conjugate with Shiga toxin 1B without a linker. O157 outbreaks. In one attempt to compensate

482 SZU AND AHMED
for this shortcoming, we conjugated OSP with showed monoclonal antitoxin with human-
the B subunit of Stx1. Mice injected with OSP- ized, chimeric, or human monoclonal anti-
Stx1B elicited bactericidal antibodies with bodies produced in transgenic mice was
high neutralization titers against Stx1 (36). successful in animal models and offered high
Since Stx type 2 is the most common type prospect (4850). A recent Stx challenge study
found in EHEC outbreaks, an ideal vaccine showed that administration of Stx2A or Stx2B
would include Stx2 B-subunit or nontoxic human monoclonal antibodies could signifi-
recombinant Stx2 mutants as part of the vac- cantly reduce the Stx accumulation in kidney,
cine component (3739). Another obstacle accompanied by a short-term elevation of
that human vaccine development faces, either Stx-antibody complex in liver during clear-
LPS- or Stx-based vaccine, is the planning of ance (51). With these plausible results, clinical
an efficacy trial to demonstrate the effective- demonstration of passive immunization with
ness. Most EHEC cases occur in outbreaks at these or similar Stx-neutralizing antibodies is
no particular locations or regions, and because much anticipated.
of the unpredictable nature of the disease
epidemiology, designing an efficacy trial for
ACKNOWLEDGMENTS
future human vaccines bears inherent diffi-
culty. Research performed in this review was
LPS that enables gram-negative organisms funded by the Intramural Research of Eunice
to escape complement fixation was considered Kennedy Shriver National Institute of Child
as one of the virulence factors for E. coli O157 Health and Human Development, National
(40). There are other virulence and attach- Institutes of Health, and by Carolinas Medical
ment factors such as adhesin intimin, Tir, and Center, Charlotte, NC. The authors have no
EspA proteins, and some showed various conflict of interests or financial obligations
degrees of protection against E. coli O157 in with the subject matter or materials discussed
animal models (4144). Attempts to include in the article. No writing assistance was used
these protein antigens as chimeric recombi- in the production of this article.
nant proteins in transgenic plants have also
reached some preclinical success (43, 44). In
CITATION
the future development of E. coli O157 human
vaccine, including such virulence factors ei- Szu SC, Ahmed A. 2014. Clinical studies of
ther as the carrier protein for an OSP conju- Escherichia coli O157:H7 conjugate vaccines in
gate or as separate components in a combined adults and young children. Microbiol Spec-
formulation, could be beneficial. Concurrent trum 2(4):EHEC-0016-2013.
immunization with multiple antigens may
generate synergistic protection, broaden the
REFERENCES
coverage to the non-O157 STEC serotypes,
and facilitate both preventive and therapeutic 1. Centers for Disease Control and Prevention.
treatments (45). 2006. Ongoing multistate outbreak of Escherichia
coli serotype O157:H7 infections associated with
Although plasmaphoresis is a common
consumption of fresh spinachUnited States,
emergency intervention for patients with HUS, September 2006. Morb Mortal Wkly Rep 55:1045
the plasma used was not enriched with specific 1046.
Stx neutralization antibodies. Other nonspe- 2. Page AV, Liles WC. 2013. Enterohemorrhagic
cific measures aimed at lowering systemic Stx Escherichia coli infections and hemolytic uremic
syndrome. Med Clin N Am 97:681695.
levels in patients include immunoadsorption,
3. Kassenborg HD, Hedberg CW, Hoekstra M,
IgG replacement activated charcoal absorp- Evans MC, Chin AE, Marcus R, Vugia
tion, or kidney dialysis, and their effectiveness DJ, Smith K, Ahuja SD, Slutsker L, Griffin
remains controversial (46, 47). Several reviews PM, Emerging Infections Program FoodNet

Working Group. 2004. Farm visits and under- serovar Paratyphi A O-specific polysaccharide-
cooked hamburgers as major risk factors for tetanus toxoid conjugates in adults, teenagers, and
sporadic Escherichia coli O157:H7 infection: data 2- to 4-year-old children in Vietnam. Infect
from a case-control study in 5 FoodNet sites. Clin Immun 68:15291534.
Infect Dis 38 Suppl 3:S271S278. 14. Hayashi T, Makino K, Ohnishi M, Kurokawa K,
4. Goode B, OReilly C, Dunn J, Fullerton K, Ishii K, Yokoyama K, Han CG, Ohtsubo E,
Smith S, Ghneim G, Keen J, Durso L, Davies M, Nakayama K, Murata T, Tanaka M, Tobe T,
Montgomery S. 2009. Outbreak of Escherichia Iida T, Takami H, Honda T, Sasakawa C,
coli O157:H7 infections after petting zoo visits, Ogasawara N, Yasunaga T, Kuhara S, Shiba T,
North Carolina State Fair, October-November Hattori M, Shinagawa H. 2001. Complete ge-
2004. Arch Pediatr Adolesc Med 163:4248. nome sequence of enterohemorrhagic Escherichia
5. Durak MZ, Churey JJ, Worobo RW. 2012. Ef- coli O157:H7 and genomic comparison with a
ficacy of UV, acidified sodium hypochlorite, and laboratory strain K-12. DNA Res 8:1122. (Erra-
mild heat for decontamination of surface and in- tum, 8:96.)
filtrated Escherichia coli O157:H7 on green onions 15. Zhang Y, Lin K. 2012. A phylogenomic analysis of
and baby spinach. J Food Prot 75:11981206. Escherichia coli/Shigella group: implications of
6. Sheng H, Lim JY, Knecht HJ, Li J, Hovde CJ. genomic features associated with pathogenicity
2006. Role of Escherichia coli O157:H7 virulence and ecological adaptation. BMC Evol Biol 12:174.
factors in colonization at the bovine terminal 16. Konadu E, Parke JC Jr, Donohue-Rolfe A,
rectal mucosa. Infect Immun 74:46854693. Calderwood SB, Robbins JB, Szu SC. 1998. Syn-
7. Varela NP, Dick P, Wilson J. 2012. Assessing the thesis and immunologic properties of O-specific
existing information on the efficacy of bovine polysaccharide-protein conjugate vaccines for
vaccination against Escherichia coli O157:H7a prevention and treatment of infections with Esch-
systematic review and meta-analysis. Zoonoses erichia coli O157 and other causes of hemolytic-
Public Health 60:253268. uremic syndrome, p 419424. In Kaper JB,
8. Teunis P, Takumi K, Shinagawa K. 2004. Dose OBrien AD (ed), Escherichia coli O157:H7 and
response for infection by Escherichia coli O157:H7 Other Shiga Toxin-Producing E. coli Strains. ASM
from outbreak data. Risk Anal 24:401407. Press, Washington, DC.
9. Reymond D, Johnson RP, Karmali MA, Petric 17. Konadu E, Robbins JB, Shiloach J, Bryla DA,
M, Winkler M, Johnson S, Rahn K, Renwick S, Szu SC. 1994. Preparation, characterization, and
Wilson J, Clarke RC, Spika J. 1996. Neutralizing immunological properties in mice of Escherichia
antibodies to Escherichia coli Vero cytotoxin 1 and coli O157 O-specific polysaccharide-protein con-
antibodies to O157 lipopolysaccharide in healthy jugate vaccines. Infect Immun 62:50485054.
farm family members and urban residents. J Clin 18. Konadu EY, Parke JC Jr, Tran HT, Bryla DA,
Microbiol 34:20532057. Robbins JB, Szu SC. 1998. Investigational vaccine
10. Navarro A, Eslava C, Hernandez U, Navarro- for Escherichia coli O157: phase 1 study of O157 O-
Henze JL, Aviles M, Garcia-de la Torre G, specific polysaccharide-Pseudomonas aeruginosa
Cravioto A. 2003. Antibody responses to Esche- recombinant exoprotein A conjugates in adults.
richia coli O157 and other lipopolysaccharides in J Infect Dis 177:383387.
healthy children and adults. Clin Diagn Lab 19. Ahmed A, Li J, Shiloach Y, Robbins JB, Szu SC.
Immunol. 10:797801. 2006. Safety and immunogenicity of Escherichia
11. Robbins JB, Schneerson R, Szu SC. 1995. Per- coli O157 O-specific polysaccharide conjugate
spective: hypothesis: serum IgG antibody is suf- vaccine in 2-5-year-old children. J Infect Dis
ficient to confer protection against infectious 193:515521.
diseases by inactivating the inoculum (Review). 20. Chow SK, Casadevall A. 2012. Monoclonal anti-
J Infect Dis 171:13871398. bodies and toxinsa perspective on function and
12. Passwell JH, Ashkenzi S, Banet-Levi Y, Ramon- isotype. Toxins (Basel) 4:430454.
Saraf R, Farzam N, Lerner-Geva L, Even-Nir H, 21. Szu SC, Taylor DN, Trofa AC, Clements JD,
Yerushalmi B, Chu C, Shiloach J, Robbins JB, Shiloach J, Sadoff JC, Bryla DA, Robbins JB.
Schneerson R, Israeli Shigella Study Group. 1994. Laboratory and preliminary clinical char-
2010. Age-related efficacy of Shigella O-specific acterization of Vi capsular polysaccharide-protein
polysaccharide conjugates in 1-4-year-old Israeli conjugate vaccines. Infect Immun 62:44404444.
children. Vaccine 28:22312235. 22. Chart H, Rowe B. 1993. Antibody cross-reactions
13. Konadu EY, Lin FY, H VA, Thuy NT, Van Bay with lipopolysaccharide from E. coli O157 after
P, Thanh TC, Khiem HB, Trach DD, Karpas AB, cholera vaccination. Lancet 341:1282.
Li J, Bryla DA, Robbins JB, Szu SC. 2000. Phase 23. Nichiuchi Y, Doe M, Hotta H, Kobayashi K.
1 and phase 2 studies of Salmonella enterica 2000. Structure and serologic properties of

484 SZU AND AHMED
O-specific polysaccharide from Citrobacter A, C, W-135 and Y measured using human versus
freundii possessing cross-reactivity with Esche- rabbit serum as the complement source. Vaccine
richia coli O157:H7. FEMS Immunol Med Micro- 30:2934.
biol 28:163171. 34. Watson DC, Robbins JB, Szu SC. 1992. Protec-
24. Chart H, Cheasty T, Cope D, Gross RJ, Rowe tion of mice against Salmonella typhimurium with
B. 1991. The serological relationship between an O-specific polysaccharide-protein conjugate
Yersinia enterocolitica O9 and Escherichia coli vaccine. Infect Immun 60:46794686.
O157 using sera from patients with yersiniosis and 35. Mohawk KL, Melton-Celsa AR, Robinson CM,
haemolytic uraemic syndrome. Epidemiol Infect OBrien AD. 2010. Neutralizing antibodies to
107:349356. Shiga toxin type 2 (Stx2) reduce colonization of
25. DiFabio JL, Perry MB, Bundle DR. 1987. Anal- mice by Stx2-expressing Escherichia coli O157:H7.
ysis of the lipopolysaccharide of Pseudomonas Vaccine 28:47774785.
maltophilia 555. Biochem Cell Biol 65:968977. 36. Konadu E, Donohue-Rolfe A, Calderwood SB,
26. Samuel G, Hogbin JP, Wang L, Reeves PR. Pozsgay V, Shiloach J, Robbins JB, Szu SC.
2004. Relationships of the Escherichia coli O157, 1999. Syntheses and immunologic properties of
O111, and O55 O-antigen gene clusters with those Escherichia coli O157 O-specific polysaccharide
of Salmonella enterica and Citrobacter freundii, and Shiga toxin 1 B subunit conjugates in mice.
which express identical O antigens. J Bacteriol Infect Immun 67:61916193.
186:65366543. 37. Marcato P, Griener TP, Mulvey GL, Armstrong
27. Bettelheim KA, Evangelidis H, Pearce JL, GD. 2005. Recombinant Shiga toxin B-subunit-
Sowers E, Strockbine NA. 1993. Isolation of a keyhole limpet hemocyanin conjugate vaccine
Citrobacter freundii strain which carries the protects mice from Shigatoxemia. Infect Immun
Escherichia coli O157 antigen. J Clin Microbiol 73:65236529.
31:760761. 38. Perera LP, Samuel JE, Holmes RK, OBrien AD.
28. Vinogradov E, Conlan JW, Perry MB. 2000. 1991. Identification of three amino acid residues
Serological cross-reaction between the lipopoly- in the B subunit of Shiga toxin and Shiga-like
saccharide O-polysaccharide antigens of Esche- toxin type II that are essential for holotoxin ac-
richia coli O157:H7 and strains of Citrobacter tivity. J Bacteriol 173:11511160.
freundii and Citrobacter sedlakii. FEMS Microbiol 39. Suhan ML, Hovde CJ. 1998. Disruption of an
Lett 190:157161. internal membrane-spanning region in Shiga
29. Cohen D, Ashkenazi S, Green MS, Gdalevich toxin 1 reduces cytotoxicity. Infect Immun 66:
M, Robin G, Slepon R, Yavzori M, Orr N, 52525259.
Block C, Ashkenazi I, Shemer J, Taylor DN, 40. Miyashita A, Iyoda S, Ishii K, Hamamoto H,
Hale TL, Sadoff JC, Pavliakova D, Schneerson Sekimizu K, Kaito C. 2012. Lipopolysaccharide
R, Robbins JB. 1997. Double-blind vaccine- O-antigen of enterohemorrhagic Escherichia coli
controlled randomised efficacy trial of an inves- O157:H7 is required for killing both insects and
tigational Shigella sonnei conjugate vaccine in mammals. FEMS Microbiol Lett 333:5968.
young adults. Lancet 349:155159. 41. Bergan J, Dyve Lingelem AB, Simm R, Skotland
30. Espi E, Grimont F, Mariani-Kurkdjian P, T, Sandvig K. 2012. Shiga toxins. Toxicon 60:
Bouvet P, Haeghebaert S, Filliol I, Loirat C, 10851107.
Decludt B, Minh NN, Vaillant V, de Valk H. 42. Melton-Celsa A, Mohawk K, Teel L, OBrien A.
2008. Surveillance of hemolytic uremic syndrome 2012. Pathogenesis of Shiga-toxin producing
in children less than 15 years of age, a system to Escherichia coli. Curr Top Microbiol Immunol
monitor O157 and non-O157 Shiga toxin-producing 357:67103.
Escherichia coli infections in France, 19962006. 43. Amani J, Mousavi SL, Rafati S, Salmanian AH.
Pediatr Infect Dis J 27:595601. 2011. Immunogenicity of a plant-derived edible
31. Eklund M, Nuorti JP, Ruutu P, Siitonen A. chimeric EspA, Intimin and Tir of Escherichia coli
2005. Shigatoxigenic Escherichia coli (STEC) in- O157:H7 in mice. Plant Sci 180:620627.
fections in Finland during 19982002: a popula- 44. Judge NA, Mason HS, OBrien AD. 2004. Plant
tion-based surveillance study. Epidemiol Infect cell-based intimin vaccine given orally to mice
133:845852. primed with intimin reduces time of Escherichia
32. Proctor ME, Davis JP. 2000. Escherichia coli coli O157:H7 shedding in feces. Infect Immun
O157:H7 infections in Wisconsin, 19921999. 72:168175.
WMJ 99:3237. 45. Zangari T, Melton-Celsa AR, Panda A, Boisen
33. Gill CJ, Ram S, Welsch JA, Detora L, Anemona N, Smith MA, Taratov I, De Tolla LJ, Nataro
A. 2011. Correlation between serum bactericidal JP, OBrien AD. 2013. Virulence of the Shiga
activity against Neisseria meningitidis serogroups toxin type 2-expressing Escherichia coli O104:H4

German outbreak isolate in two animal models. Schnatter S, Dorresteijn EM, Samuelsson O,
Infect Immun 81:15621574. Brunkhorst R; Collaborators of the DGfN
46. Menne J, Nitschke M, Stingele R, Abu-Tair M, STEC-HUS registry. 2012. Best supportive care
Beneke J, Bramstedt J, Bremer JP, Brunkhorst and therapeutic plasma exchange with or with-
R, Busch V, Dengler R, Deuschl G, Fellermann out eculizumab in Shiga-toxin-producing E. coli
K, Fickenscher H, Gerigk C, Goettsche A, O104:H4 induced haemolytic-uraemic syndrome:
Greeve J, Hafer C, Hagenmller F, Haller H, an analysis of the German STEC-HUS registry.
Herget-Rosenthal S, Hertenstein B, Hofmann Nephrol Dial Transplant 27:38073815.
C, Lang M, Kielstein JT, Klostermeier UC, 48. Tzipori S, Sheoran A, Akiyoshi D, Donohue-
Knobloch J, Kuehbacher M, Kunzendorf U, Rolfe A, Trachtman H. 2004. Antibody therapy
Lehnert H, Manns MP, Menne TF, Meyer TN, in the management of Shiga toxin-induced he-
Michael C, Mnte T, Neumann-Grutzeck C, molytic uremic syndrome. Clin Microbiol Rev 17:
Nuernberger J, Pavenstaedt H, Ramazan L, 926941.
Renders L, Repenthin J, Ries W, Rohr A, Rump 49. Sheoran AS, Chapman-Bonofiglio S, Harvey
LC, Samuelsson O, Sayk F, Schmidt BM, BR, Mukherjee J, Georgiou G, Donohue-Rolfe
Schnatter S, Schcklmann H, Schreiber S, von A, Tzipori S. 2005. Human antibody against
Seydewitz CU, Steinhoff J, Stracke S, Shiga toxin 2 administered to piglets after the
Suerbaum S, van de Loo A, Vischedyk M, onset of diarrhea due to Escherichia coli O157:H7
Weissenborn K, Wellhner P, Wiesner M, prevents fatal systemic complications. Infect
Zeissig S, Bning J, Schiffer M, Kuehbacher T; Immun 73:46074613.
EHEC-HUS consortium. 2012. Validation of 50. He X, McMahon S, Skinner C, Merrill P,
treatment strategies for enterohaemorrhagic Scotcher MC, Stanker LH. 2013. Development
Escherichia coli O104:H4 induced haemolytic and characterization of monoclonal antibodies
uraemic syndrome: case-control study. BMJ 345: against Shiga toxin 2 and their application for
e4565. toxin detection in milk. J Immunol Methods
47. Kielstein JT, Beutel G, Fleig S, Steinhoff J, 389:1828.
Meyer TN, Hafer C, Kuhlmann U, Bramstedt J, 51. Sheoran A, Jeong KI, Mukherjee J, Wiffin A,
Panzer U, Vischedyk M, Busch V, Ries W, Singh P, Tzipori S. 2012. Biodistribution and
Mitzner S, Mees S, Stracke S, Nrnberger J, elimination kinetics of systemic Stx2 by the Stx2A
Gerke P, Wiesner M, Sucke B, Abu-Tair M, and Stx2B subunit-specific human monoclonal
Kribben A, Klause N, Schindler R, Merkel F, antibodies in mice. BMC Immunol 13:27.

Vaccination of Cattle against
Escherichia coli O157:H7
DAVID R. SMITH1
25
INTRODUCTION
Human infection with Shiga toxin-producing Escherichia coli O157:H7 (STEC

O157) is relatively rare, but the consequences can be serious, especially in the
very young and the elderly. Outcomes associated with STEC O157 infection
include hemorrhagic colitis, renal failure, and death (15). In 2012, the overall
laboratory-confirmed annual incidence of STEC O157 in the United States was
1.1 cases per 100,000 population (6). However, the incidence in children less
than 5 years of age was 4.7 cases per 100,000 population (6).
Infection from STEC O157 occurs directly or indirectly via fecal-oral trans-
mission (7). People are exposed to STEC O157 through a variety of sources,
including direct contact with human or animal feces and indirect contact via
contaminated food, water, or soil (8). The primary route of transmission of
STEC O157 is contaminated food (9, 10); however, large outbreaks have been
associated with contamination of municipal water supplies (1114). Important
environmental hazards for human exposure to STEC include daycare facilities,
nursing homes, children playing with a sick friend, swimming pools, contami-
nated food and water, and direct exposure to animal environments such as
farms, petting zoos, or livestock exhibitions (9, 10, 15). Approximately one-third
1
College of Veterinary Medicine, Mississippi State University, Mississippi State, MS 39762-6100.
487

488 SMITH
of human infections are attributed to con- subsequent rates of carcass contamination (31,
sumption of ground or nonintact beef (16). 32). Finally, since 1998 in the United States,
Some of the earliest and most notorious human incidence of STEC O157 has decreased
outbreaks of STEC O157 infection were asso- (6), largely because of interventions taken
ciated with the consumption of undercooked in abattoirs to reduce the flow of STEC O157
ground beef sandwiches, resulting in the in- from live cattle into the beef supply (33,
fection being commonly known as ham- 34). The decrease in incidence since 1998
burger disease (2, 1719). is greater than the proportion of illnesses
STEC has been recovered from many ani- attributable to contaminated beef, suggest-
mal species, but ruminants are particularly ing that decreasing the bacterial flow from
prone to colonization (20). Of ruminants, beef prevented secondary cases of person-to-
cattle populations are widely recognized as person STEC O157 infection. Unfortunately,
an important reservoir of STEC O157 for the incidence of STEC O157 infection has
human exposure in the United States (8, 21). not changed meaningfully or statistically com-
A variety of vehicles, other than food, have pared to the average annual incidence during
been important in the fecal-oral transmission 20062008, suggesting that additional actions,
of STEC strains to humans, including fomites for example, at the preharvest level, are nec-
such as dust (22) and water (1114) and vec- essary to further reduce rates of STEC O157
tors such as flies (2327). Other animals, be- illness (6).
sides cattle, have caused important STEC
outbreaks in humans because they served as
vehicles for fruit or vegetable crop contami- PREHARVEST ECOLOGY OF STEC O157
nation. For example, a large STEC O157 out-
break in the United States and Canada was Cattle are colonized by STEC O157 primarily
due to consumption of spinach that was con- at the terminal rectum (35, 36). Colonization
taminated in the field by feces from feral pigs by STEC O157 requires attachment to intes-
that had contact with cattle pastures (28). In tinal epithelium and induces attaching and
Oregon, deer were the source of feces that effacing lesions. Following STEC O157 infec-
contaminated strawberries with STEC O157, tion in cattle of all ages, inflammation and
resulting in one death and at least 14 illnesses innate and adaptive immune responses occur
(29). (37), supporting the contention that STEC
Circumstantial evidence supports the con- O157 is a bovine pathogen (37, 38). However,
tention that cattle are the primary reser- this latter point remains controversial because
voir for human exposure to STEC in North infection does not result in clinically observ-
America. First, there is strong correlation able signs of illness in adult cattle (39, 40). In
between seasonal variability in incidence of any case, not all cattle shedding STEC in their
human STEC O157 illness, prevalence of feces are currently colonized; some may be
ground beef contamination with STEC O157, shedding ingested organisms that are simply
and prevalence of STEC O157 shedding passing through the intestinal tract (41, 42).
by cattle in feedlots, all greater in summer The duration of infection in cattle is variable
months than winter months (30). This rela- but short-lived, approximating a month (41,
tionship may indicate that STEC O157, origi- 4345). In field settings, reinfection is com-
nating in or on cattle, contaminates ground mon (44).
beef to eventually become the source for Prevalence of STEC O157 carriage by
subsequent human infection (30). In addition, feedlot cattle varies widely within and across
there is a correlation between the preva- seasons and is affected by both incidence and
lence of carriage of STEC O157 in feces or on duration of shedding (44, 46, 47). The prob-
hides of live cattle entering the abattoir and ability of cattle carrying STEC depends on

CHAPTER 25 Vaccination of Cattle against E. coli O157:H7 489
both gut and environmental conditions that the greater number of organisms being shed,
change over time. As with all E. coli strains, cattle designated as super-shedders may
conditions of the bovine gut that favor STEC have an important effect on environmental
O157 may increase colonization and duration contamination and subsequent transmission
of shedding. Factors of the environment that within cattle production settings (53) or in
favor STEC O157 survival or opportunities for lairage (54). The relevance of super-shedding
fecal-oral transmission increase the incidence to STEC O157 control is not clear. Super-
of exposure. This is because pathogenic and shedding of STEC O157 in feces does not ap-
commensal E. coli strains have two principal pear to be a persistent state, and we do not yet
habitats: a primary habitat in the lower in- understand if super-shedding is a character-
testine of warm-blooded animals and a sec- istic of certain cattle or merely a stage of in-
ondary habitat in water, sediment, and soil fection that cattle transition through following
(48). The suitability of the primary habitat is infection. It has been observed that detection
influenced by factors such as physical charac- of super-shedding cattle is temporally corre-
teristics (e.g., pH); the hosts diet, immune lated with periods of high prevalence, and
system, and physiological state; and interac- super-shedder cattle appear to be a subset of
tions with other microorganisms in the same fecal-culture-positive individuals within the
region. The suitability of the secondary habitat population (42, 55). Super-shedding may not
is also complex and dependent on physical be necessary or sufficient for STEC O157
factors, climatic and meteorological factors, transmission, even in closed (all-in, all-out
nutrients, and interactions with other micro- type) feeding systems (56). Rather than super-
organisms within the ecosystem. In contrast shedding cattle driving transmission of STEC
to the primary habitat, which is uniformly to other cattle, super-shedding may be an
warm, approximately 37C, and nutrient rich, outcome of environmental conditions that
the secondary habitat may have extremes in favor ingestion of the organism (47). When
temperatures and is typically nutrient defi- those conditions favor new host infections,
cient (48). Environmental conditions that then some cattle may become colonized and
favor survival and fecal-oral transmission have transiently shed large numbers of organisms,
been associated with greater rates of exposure and because of favorable conditions for trans-
and shedding in feedlot cattle (46, 47). mission, the duration of detectable shedding
Transmission heterogeneity, or super- may be prolonged (44).
spreading, is the phenomenon of a minority To reduce the prevalence of STEC O157
of infected individuals being responsible for carriage by cattle, efforts have been attempted
transmitting the majority of new infections to make either the primary or secondary
(49, 50). At a given point in time, STEC O157- habitat less favorable to STEC O157 survival
infected cattle shed the organism at varying or growth (5759). To date, efforts to make
concentrations in feces (42, 51, 52). Therefore, the cattle environment less hospitable to
some cattle may contribute vastly more STEC STEC O157, for example, by scraping pen
organisms into the environment, and possibly surfaces or cleaning water tanks, have not
to other cattle, than others. Cattle that shed effectively reduced STEC O157 carriage by
STEC at greater than 103 or 104 CFU/g of cattle (6062). However, several strategies
feces, or cattle that are culture-positive for for modifying the gut environment, including
prolonged periods, have variously been de- the use of vaccines; chemicals, such as so-
fined by the term super-shedder (42, 51). dium chlorate or antibiotics; and competing
It has been proposed that super-shedding microorganisms, such as some strains of
status is indicative of cattle colonized by Lactobacillus, have effectively reduced the
STEC rather than cattle experiencing simple probability of cattle shedding STEC O157 in
pass-through of organisms (42). Because of feces (6366).

490 SMITH
VACCINATION OF CATTLE whether the comparison uses odds (i.e., odds

AGAINST STEC O157 ratio) or probability (i.e., relative risk), a value
of 1 indicates no difference from the treat-
The objective of immunizing cattle against ment. The further the value is from 1, toward
STEC O157 is to make the gut unfavorable 0 or infinity, the larger the measure of asso-
for colonization, thereby reducing duration ciation. If the study is not a case-control study
of carriage and minimizing shedding of the design, then odds ratio can be converted to
pathogen into the cattle environment (58). In relative risk after adjustment for marginal
theory, the benefit of vaccination within dis- probabilities for disease and exposure (73).
crete populations (e.g., pens or herds of cattle) In studies with measures of fecal concentra-
is reduced fecal-oral transmission within cat- tion, the measure of association may be ex-
tle environments, less contamination of cattle pressed as the change in concentration due
hides, and fewer pathogens carried into the to vaccine treatment, which is often described
abattoir at harvest. For vaccination to be as a logarithmic (base 10) reduction (74) and
useful as a preharvest intervention, the bene- sometimes reported as a percentage (e.g., a
fits must not be undone during subsequent decrease from 10,000 CFU/g of feces to
management practices, such as transportation 1,000 CFU/g of feces is a decrease of 1 log(10)
to the abattoir (67) or during holding in lairage in CFU/g of feces and may be expressed as a
(32, 68, 69). Preharvest interventions such as 90% reduction in shedding concentration).
vaccination are not likely to be adopted widely
by cattle producers until they are sufficiently
Vaccine Challenge Studies
valued in the marketplace to offset the cost of
implementation. STEC O157 colonizes bovine intestinal epi-
Some candidate vaccines against STEC thelial cells by a type III secreted protein
O157 have been tested in animal challenge (TTSP) system. Components of the TTSP
studies or under field conditions of natural system include:
exposure. These vaccines either have unde-
Intimin, an outer membrane bacterial
fined antigen targets in the form of bacterial
receptor
extracts or are directed against specific anti-
Translocated intimin receptor (Tir), a re-
gens that function to enable bacterial coloni-
ceptor injected into the host epithelial cell
zation or survival. Unfortunately, because of
membrane
serotype specificity, vaccines targeting STEC
EspA, an injection filament for delivering
O157 may offer poor cross-protection against
Tir to the host cell membrane
other STEC strains (70).
EspB/EspD, which form a pore in the host
In randomized controlled studies, the
cell membrane (7, 40, 75, 76)
strength of effect of a vaccine is often ex-
pressed as vaccine efficacy, a form of attrib- The H7 flagellin is also believed to function
utable fraction that measures the percentage in STEC O157:H7 colonization (7779). For
of cases prevented by vaccination (71). Vac- some STEC non-O157 serotypes, the entero-
cine efficacy is calculated as 1 minus relative hemorrhagic E. coli factor for adherence (efa-1)
risk (72). In this case, relative risk is the is important for colonization of bovine intest-
probability of vaccinated cattle to carry STEC ines, and STEC O157 carries a truncated form
O157 divided by the probability of nonvacci- of the gene (80).
nated cattle to carry the organism. The odds Vaccines targeting various STEC O15-
ratio is the statistical measure of association specific antigens have been tested in animal
often reported from vaccine field studies be- challenge studies. Several studies have dem-
cause logistic regression is a commonly used onstrated immune response against the anti-
method to analyze the data. Regardless of gens but variable results regarding protection

against STEC O157 infection. Suckling pigs O157 (74). Calves injected twice subcutane-
whose dams were vaccinated with an intimin ously with an inactivated, whole-cell envelope
vaccine were protected from colonization or vaccine (STEC O157 bacterial ghosts) dem-
microscopic evidence of intestinal damage onstrated an antibody response and shed
following oral challenge with 106 CFU of a fewer STEC O157 post challenge (88). Vacci-
Shiga toxin-negative strain of EHEC O157:H7 nation of pregnant cows with intimin, EspA,
(81). Calves vaccinated with EspA developed EspB, and Shiga toxin 2 within 2 months
antigen-specific antibody titers but failed to be of calving produced elevated serum and co-
protected against colonization with STEC lostral antibodies against intimin and EspB
O157 following challenge (82). Similarly, sub- and a moderate increase in EspA antibodies
unit vaccines targeting polypeptides of intimin (89). Calves fed the dams colostrum had sig-
or efa-1 elicited humoral responses in 2-week- nificantly increased serum immunoglobulin
old calves following intramuscular priming G titers against intimin and EspB, but not
and intranasal booster doses, but the vaccine EspA (89).
products failed to prevent shedding after Siderophore receptor and porin (SRP)
STEC O157 or STEC O26 challenge (80). In vaccines are targeted against bacterial cell
the same study, a formalin-inactivated STEC membrane proteins used by gram-negative
O157 bacterin administered intramuscularly bacteria for iron transport in conditions of
with subsequent intranasal booster doses also low iron supply (90). By limiting its uptake of
failed to reduce shedding in challenged calves iron, STEC O157 is placed at a competitive
(80). Two-month-old calves vaccinated intra- disadvantage relative to other gut microbiota
muscularly with H7 flagellin had reduced (91). In a study of beef calves orally inoculated
rates of colonization and delayed peak bacte- with STEC O157, the SRP vaccine reduced
rial shedding following oral challenge with fecal prevalence and bacterial concentration
STEC O157, but the calves did not show a to a level that approached statistical signifi-
reduction in total bacterial shedding (83). cance (90).
However, a vaccine prepared with intimin,
EspA, and Tir did reduce STEC O157 coloni-
Vaccine Field Studies
zation and bacterial counts in calves orally
inoculated with STEC O157 (84). Also, lambs The outcomes of experimental challenge stud-
that had been vaccinated with intimin, EspA, ies may not predict the efficacy of a STEC
and EspB shed fewer bacteria in feces than O157 vaccine as it is used under field condi-
placebo-treated controls did following an tions because factors affecting rates of trans-
oral challenge with STEC O157 (85). Six- to mission, sources of pathogens, and dose-loads
8-month-old calves injected intramuscularly of exposure are complex and temporally dy-
with a vaccine product containing intimin and namic in cattle production settings (44, 46,
EspB proteins developed an antibody re- 47). Only a few STEC O157 vaccine products
sponse against the proteins and shed fewer have been evaluated for efficacy in the con-
STEC O157 bacteria in the first 13 days post ditions of natural STEC O157 exposure within
challenge (86). Calves vaccinated with a bac- cattle production systems. An uncharacterized
terial supernatant with TTSP had reduced bacterial extract did not reduce STEC O157
probability, magnitude, and duration of shed- carriage in feedlot cattle (92). Another un-
ding of STEC O157 following challenge (87). characterized STEC O157 vaccine, adminis-
In a follow-up study, calves receiving the same tered to pregnant beef cattle during the last
vaccine product were 21% less likely to shed trimester of gestation, significantly increased
STEC O157 in the feces and shed at a 1.4 log(10) antibody titers in the dam and subsequently
lower fecal concentration 3 to 6 days after the calf, but the study had insufficient power
experimental challenge with 109 CFU of STEC to evaluate efficacy at preventing shedding of

492 SMITH
STEC O157 by the calves (93). Calves suckling postvaccination period except for the last day
cows that had been vaccinated against SRP of the study (91). In a trial testing a 2-ml,
antigens had significantly greater antibody three-dose SRP vaccine regimen against
titers against STEC O157 SRP at branding (i.e., placebo-treated cattle, the vaccine was 85%
30 to 60 days of age), but neither the passively effective in reducing the probability of de-
acquired antibodies nor active immunization tecting STEC O157 in feces and reduced
significantly prevented STEC O157 shedding STEC O157 concentration 1.7 logs compared
by the calves at feedlot entry (94). to controls 56 days after the last dose of vac-
Two vaccine products, one targeting TTSP, cine (91). In a vaccine trial conducted in a
the other SRP, have been tested extensively in commercial feedlot, the SRP vaccine demon-
dry-lot beef feedlots under conditions typical strated 53% vaccine efficacy in reducing STEC
of the Central Plains regions of the United O157 prevalence and 73% efficacy in reducing
States and Canada. These products were the the prevalence of high shedders, defined as
subject of several systematic reviews and cattle shedding >104 CFU/g of feces (98). In
meta-analyses that found sufficient evidence that study, pens of cattle receiving vaccination
to conclude that both vaccines effectively had significantly reduced feed efficiency and
reduce the probability of feedlot cattle to rate of gain, which may represent an addi-
shed STEC O157 in feces (63, 95). One meta- tional cost of the intervention (98).
analysis of fecal shedding found the overall Vaccinating feedlot cattle with a TTSP
odds ratios (and 95% confidence intervals) for vaccine product failed to be efficacious in a
detecting STEC O157 in the feces of vacci- large initial vaccine field trial (99). However,
nated cattle relative to nonvaccinated cattle to the vaccine product was reformulated and
be 0.38 (0.290.51) and 0.42 (0.200.61) for efficacy improved (99). Vaccine efficacy of a
TTSP and SRP vaccines, respectively (63). three-dose regimen of TTSP vaccine to reduce
Given the overall fecal shedding prevalence of the probability of feedlot cattle shedding
15% observed in the TTSP studies (63), the STEC O157 has ranged from 43 to 73% in
odds ratio of 0.38 converts to a relative risk of several randomized controlled trials (87,
0.42 and vaccine efficacy of 0.58 (96). Another 100102). In addition, the vaccine was 92%
meta-analysis looked at all outcomes and and 98% effective in reducing the probability
reported that two doses of TTSP vaccine had of colonization of the terminal rectum when
odds ratios of 0.49 (0.400.60) for preharvest two- (103) or three-dose (104) regimens, re-
outcomes and 0.45 (0.340.60) for preharvest spectively, were used. Two doses of the same
and at-harvest outcomes combined (95). vaccine product significantly reduced carriage
Details from individual studies provide of STEC O157 by feedlot cattle (103, 105, 106),
additional information about the efficacy of and it appears that two doses of vaccine may
STEC O157 vaccine products, although some be sufficient to induce an effective immune
details, such as antigen concentrations, have response (95). However, three doses of vac-
not always been reported. Using steers cine were more effective than two doses in
screened to be negative for STEC O157 car- trials with direct comparisons (100, 107). This
riage before the study start, researchers found vaccine does not appear to affect growth
that steers vaccinated twice with either 2 or performance (104, 107) or carcass quality (104,
3 ml of SRP vaccine were 14 and 47% less 106, 107).
likely than placebo-treated steers to have The duration of immunity after vaccination
STEC O157 detected in either feces or recto- is unknown because the evaluation period
anal mucosa swab samples, respectively (97). in feedlot studies has been relatively short,
Feedlot cattle receiving a 2-ml, two-dose typically with postvaccination observation
SRP vaccine regimen did not differ from periods of between 60 and 100 days (63,
controls in STEC O157 carriage over the 108, 109). Increasing or decreasing immunity

would be evident as a statistical interaction evaluate group-level effects of vaccinating

between vaccine treatment and time elapsing cattle against STEC O157. There is evidence
since vaccination on the probability of cattle that fecal-oral transmission of STEC O157 is
carrying the organism. This interaction has reduced within pens of vaccinated cattle.
not been reported. Even though vaccine effi- Herd immunity was demonstrated in a longi-
cacy appears to persist sufficiently long tudinal STEC O157 vaccine study as nonvac-
enough for cattle on finishing diets, duration cinated cattle housed with vaccinated cattle
of immunity remains an important unmet area were less likely to shed STEC O157 compared
of investigation for beef and dairy young-stock to cattle penned in the same feedyard where
and breeding cattle (109). none of the cattle received vaccine (107).
Cattle are typically managed as groups Vaccinated cattle housed together in large
(e.g., pens or herds of cattle), which are fed commercial feedyard pens were less likely to
and housed together. Similarly, cattle man- have oral exposure to STEC O157 compared
agement practices such as vaccination are to nonvaccinated cattle housed together in
usually applied to the group, partly for ease pens in the same feedyards, based on culturing
of management and to provide protection to ropes hung on feedbunk rails for cattle to
the group rather than simply the individual. chew (103). Culture of STEC O157 from ropes
The ability of groups to resist infection, or to is correlated to fecal shedding prevalence
limit the extent of infection within the group, (112), and more directly measures opportuni-
is termed herd immunity (110). Herd immu- ties for oral exposure (113).
nity is a function of individual resistance to The value of considering the effects of
infection and the dynamics of transmission group-level vaccination when designing a
within the group (110, 111). Individuals lacking STEC O157 cattle vaccination program was
immunity may be protected from infection demonstrated by the greater efficacy in re-
because of group-level factors; for example, ducing hide contamination when all cattle
the majority of individuals with immunity in a region of a feedyard were vaccinated
change the likelihood of exposure to those compared to the efficacy observed when vac-
without (110). cinated and unvaccinated cattle were com-
The probability of cattle carrying STEC mingled within pens (106). Efficacy against
O157 in the gut or on their hides is affected hide contamination is important because
by group-level factors. For example, the dis- the hides of cattle are the primary source
tribution of fecal prevalence of STEC O157 of STEC O157 carcass contamination (32, 69,
within pens of feedlot cattle tends to be 114, 115). It was hypothesized that vaccination
greater or lesser than expected by binomial of all cattle within a region of a feedyard,
distribution around the mean (46), suggesting or the entire feedyard, would result in a
that, at a given point in time, cattle within greater reduction in the load of organisms
pens behave similarly with respect to STEC deposited by cattle into the environment
O157 shedding (i.e., most cattle shedding or and less subsequent contamination of hides
most not). Factors explaining the probability than when vaccinated cattle are commingled
of cattle shedding the agent or having evi- in pens of nonvaccinated cattle (106). This
dence of oral exposure are associated with finding illustrates that the goal of a cattle
characteristics of the pen environment that vaccination program against STEC O157 is
either favor survival of the organism (e.g., to reduce environmental pathogen load to
warm or wet) or increase opportunities for minimize ingestion of the organism or hide
ingestion (e.g., mud or dust), indicating that contamination, and this may be accomplished
sometimes the pen environment favors fecal- most effectively by administering the vac-
oral transmission and sometimes it does not cine to all cattle within a production system
(44, 46, 47, 112). Therefore, it is important to (106).

494 SMITH
Whatever efficacy a vaccine may have be- research questions, and predict the usefulness
fore harvest, it can be undone by events of intervention strategies (117).
occurring during subsequent stages of the From a public health policy perspective,
food system, such as cross-contamination of one might compare the cost of human illness
cattle hides with STEC O157 during trans- to the cost of a preharvest intervention. If the
portation or while cattle are in lairage (32, marginal costs of vaccinating cattle were
67, 116). However, the efficacy of preharvest equivalent to the marginal benefit to public
interventions has persisted into the abattoir. health, then as the cost of a vaccine inter-
In a randomized clinical trial to test a STEC vention increased, fewer cattle would be vac-
O157 cattle vaccine, there was a significant cinated, and as a result, fewer human illnesses
increase in the prevalence of hide contami- would be prevented. Similarly, the number of
nation between the time immediately before cattle that must receive an intervention to
loading at the feedyard versus just before prevent a single human illness increases as the
hide removal in the abattoir. However, vacci- effectiveness of the product decreases (16).
nation treatments had equal efficacy for re- From a beef industry perspective, preharvest
ducing hide contamination in the feedyard interventions might be valued on the basis
and at the abattoir. The preservation of vac- of how cattle carrying STEC O157 into the
cine efficacy into the abattoir may have been abattoir affect subsequent food safety costs.
the result of efforts to load cattle by treatment For example, because an important source of
groups into clean trucks for transportation STEC O157 carcass contamination is the hide
to the abattoir (106). Therefore, to preserve (32), and fecal shedding prevalence above 20%
vaccine efficacy, it may be necessary to devise has been associated with higher prevalence
methods for cattle handling so that preharvest levels of hide contamination (118), postharvest
benefits are retained post harvest. sectors of the beef industry might benefit from
preharvest interventions that supply cattle at
harvest with less hide contamination and re-
Modeling STEC O157 Vaccine Usefulness
duced, less variable, fecal shedding prevalence
Ultimately, the reasons for vaccinating cattle that does not overwhelm subsequent post-
against STEC O157 are to (i) benefit public harvest interventions.
health by preventing human STEC O157 in- Quantitative or qualitative models have
fection and (ii) reduce costs to the beef been used to investigate the value of vacci-
industry due to recalls, lost product value, nating cattle and other methods of interven-
and liability. There is value in preventing tion. Many models predict benefit to both
human illness from direct contact with cattle public health and the beef industry from
or their environments, but this is a less com- vaccinating cattle against STEC O157. For
mon source of human illness compared to example, a model simulating ground beef
infections acquired through contaminated contamination in Argentina predicted that
food, including beef, milk, and vegetable crops vaccinating cattle and online hide washing
(9, 10). The primary value of vaccinating live would have the greatest impacts on reducing
cattle is the benefit to the postharvest sectors STEC O157 prevalence and concentration in
of the food system and the consumers of food ground beef product and the resulting num-
products. An intervention is not likely to be bers of human infections, hemolytic-uremic
used if the costs of the intervention exceed syndrome, and STEC O157-associated mortal-
the benefits to the food industry or public ities per ground beef meal (119). A stochastic
health. Mathematical models provide a con- simulation model based on U.S. beef produc-
ceptual framework for understanding patho- tion systems and risk for infection through
gen transmission dynamics. Models can help consumption of ground beef also concluded
identify knowledge gaps, give insight into new that vaccination of cattle would have a strong

impact on decreasing the number of human STEC O157 by making the gut environment
STEC O157 illnesses, the number of contami- unfavorable to colonization. However, in con-
nated beef production lots, the likelihood of ditions of natural exposure, efficacy afforded
STEC O157 detection by regulatory testing, by vaccination depends on how the products
and the probability of outbreaks due to ground are used to control environmental transmis-
beef servings from the same lot (120). A sim- sion within groups of cattle or throughout the
ulation model was used to investigate infec- production system (106). Preharvest benefits
tion transmission in pastured cattle systems. from vaccination may be nullified unless steps
The modelers concluded that vaccine efficacy are taken to prevent cross-contamination of
of 60% would be particularly effective in re- cattle or beef product throughout the food
ducing levels of infection in a herd (121). system (68). Although cattle vaccines against
Stochastic simulation of the distribution of STEC O157 have gained either full or prelim-
pen-level fecal shedding prevalence in U.S. inary regulatory approval in Canada and the
commercial beef feedyards predicted that United States, it is not yet clear if they will be
vaccination of summer-fed cattle with a 58% widely adopted by cattle feeders because there
effective product would eliminate pens of is not yet an economic signal to indicate that
highest prevalence, resulting in a prevalence cattle vaccinated against STEC O157 are val-
distribution similar to what is typically ob- ued over other cattle.
served in winter-fed cattle. This model
showed that a major effect of vaccination is
ACKNOWLEDGMENT
reduced variability in shedding prevalence
(122). The opinions of experts were used in a I declare no conflicts of interest with regard to
best-worse scaling evaluation to gain consen- the manuscript.
sus on the effectiveness and practicality of
on-farm methods to reduce human exposure
CITATION
to STEC O157 (123). Intervention methods
were evaluated for effectiveness and practi- Smith DR. 2014. Vaccination of cattle against
cality. By this process, vaccination of cattle Escherichia coli O157:H7. Microbiol Spectrum
was considered the most effective, and hand 2(4):EHEC-0006-2013.
washing the most practical, method to reduce
human exposure to STEC (123).
REFERENCES
1. Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV,
CONCLUSION Widdowson MA, Roy SL, Jones JL, Griffin PM.
2011. Foodborne illness acquired in the United
Statesmajor pathogens. Emerg Infect Dis 17:715.
Ideally, preharvest interventions against 2. Riley LW, Remis RS, Helgerson SD, McGee HB,
STEC O157 should be Wells JG, Davis BR, Hebert RJ, Olcott ES,
Johnson LM, Hargrett NT, Blake PA, Cohen
Efficaciouscattle are less likely to carry ML. 1983. Hemorrhagic colitis associated with a
the organism because of the intervention rare Escherichia coli serotype. N Engl J Med
Usefulable to be practically applied 308:681685.
within the beef production system 3. Wells JG, Davis BR, Wachsmuth IK, Riley LW,
Remis RS, Sokolow R, Morris GK. 1983. Labo-
Economicaladdsufficientvalueto theprod-
ratory investigation of hemorrhagic colitis out-
uct to offset the cost of the intervention breaks associated with a rare Escherichia coli
serotype. J Clin Microbiol 18:512520.
A number of STEC O157 antigens are being
4. Karmali MA, Petric M, Lim C, Fleming PC,
investigated as potential vaccine targets. Some Steele BT. 1983. Escherichia coli cytotoxin,
vaccine products have demonstrated efficacy haemolytic-uraemic syndrome, and haemorrhagic
to reduce the prevalence of cattle carrying colitis. Lancet ii:12991300.

496 SMITH
5. Karmali MA, Steele BT, Petric M, Lim C. 1983. Wright GF, Blake PA. 1986. Escherichia coli O157:
Sporadic cases of haemolytic-uraemic syndrome H7 diarrhea in a nursing home: clinical, epide-
associated with faecal cytotoxin and cytotoxin- miological, and pathological findings. J Infect Dis
producing Escherichia coli in stools. Lancet i:619 154:631638.
620. 19. Slutsker L, Ries AA, Maloney K, Wells JG,
6. Centers for Disease Control and Prevention. Greene KD, Griffin PM. 1998. A nationwide case-
2013. Incidence and trends of infection with control study of Escherichia coli O157:H7 infection
pathogens transmitted commonly through food in the United States. J Infect Dis 177:962966.
Foodborne Diseases Active Surveillance Network, 20. Beutin L, Geier D, Steinruck H, Zimmermann
10 U.S. sites, 19962012. Morb Mortal Wkly Rep S, Scheutz F. 1993. Prevalence and some prop-
62:283287. erties of verotoxin (Shiga-like toxin)-producing
7. Nataro JP, Kaper JB. 1998. Diarrheagenic Esch- Escherichia coli in seven different species of
erichia coli. Clin Microbiol Rev 11:142201. healthy domestic animals. J Clin Microbiol 31:
8. Sargeant JM, Smith DR. 2003. The epidemiology 24832488.
of Escherichia coli O157:H7, p 131141. In Torrence 21. Karmali MA, Gannon V, Sargeant JM. 2010.
ME, Isaacson RE (ed), Microbial Food Safety in Verocytotoxin-producing Escherichia coli (VTEC).
Animal Agriculture: Current Topics. Iowa State Vet Microbiol 140:360370.
University Press, Ames, IA. 22. Varma JK, Greene KD, Reller ME, DeLong SM,
9. Rangel JM, Sparling PH, Crowe C, Griffin PM, Trottier J, Nowicki SF, DiOrio M, Koch EM,
Swerdlow DL. 2005. Epidemiology of Escherichia Bannerman TL, York ST, Lambert-Fair MA,
coli O157:H7 outbreaks, United States, 19822002. Wells JG, Mead PS. 2003. An outbreak of Esch-
Emerg Infect Dis 11:603609. erichia coli O157 infection following exposure to a
10. Sparling PH. 1998. Escherichia coli O157:H7 contaminated building. JAMA 290:27092712.
outbreaks in the United States, 19821996. J Am 23. Alam MJ, Zurek L. 2004. Association of Esche-
Vet Med Assoc 213:17331733. richia coli O157:H7 with houseflies on a cattle
11. Swerdlow DL, Woodruff BA, Brady RC, Griffin farm. Appl Environ Microbiol 70:75787580.
PM, Tippen S, Donnell HD Jr, Geldreich E, 24. Hancock DD, Besser TE, Rice DH, Ebel ED,
Payne BJ, Meyer A Jr, Wells JG. 1992. A wa- Herriott DE, Carpenter LV. 1998. Multiple
terborne outbreak in Missouri of Escherichia sources of Escherichia coli O157 in feedlots and
coli O157:H7 associated with bloody diarrhea dairy farms in the northwestern USA. Prev Vet
and death. Ann Intern Med 117:812819. Med 35:1119.
12. Kondro W. 2000. E. coli outbreak deaths spark 25. Janisiewicz WJ, Conway WS, Brown MW,
judicial inquiry in Canada. Lancet 355:2058. Sapers GM, Fratamico P, Buchanan RL. 1999.
13. Kondro W. 2000. Canada reacts to water con- Fate of Escherichia coli O157:H7 on fresh-cut
tamination. Lancet 355:2228. apple tissue and its potential for transmission by
14. Centers for Disease Control and Prevention. fruit flies. Appl Environ Microbiol 65:15.
1999. Outbreak of Escherichia coli O157:H7 and 26. Kobayashi M, Sasaki T, Saito N. 1999. House-
Campylobacter among attendees of the Washington flies: not simple mechanical vectors of entero-
County FairNew York, 1999. Morb Mortal Wkly hemorrhagic Escherichia coli O157:H7. Am J Trop
Rep 48:803805. Med Hyg 61:625629.
15. Feng P. 1995. Escherichia coli serotype O157:H7: 27. Moriya K, Fujibayashi T, Yoshihara T, Matsuda
novel vehicles of infection and emergence of A, Sumi N, Umezaki N, Kurahashi H, Agui N,
phenotypic variants. Emerg Infect Dis 1:4752. Wada A, Watanabe H. 1999. Verotoxin-producing
16. Withee J, Williams M, Schlosser W, Bauer N, Escherichia coli O157:H7 carried by the housefly in
Ebel E. 2009. Streamlined analysis for evaluating Japan. Med Vet Entomol 13:214216.
the use of preharvest interventions intended to 28. Jay MT, Cooley M, Carychao D, Wiscomb GW,
prevent Escherichia coli O157:H7 illness in hu- Sweitzer RA, Crawford-Miksza L, Farrar JA,
mans. Foodborne Pathog Dis 6:817825. Lau DK, OConnell J, Millington A, Asmundson
17. Kassenborg HD, Hedberg CW, Hoekstra M, RV, Atwill ER, Mandrell RE. 2007. Escherichia
Evans MC, Chin AE, Marcus R, Vugia DJ, coli O157:H7 in feral swine near spinach fields and
Smith K, Ahuja SD, Slutsker L, Griffin PM. 2004. cattle, central California coast. Emerg Infect Dis
Farm visits and undercooked hamburgers as major 13:19081911.
risk factors for sporadic Escherichia coli O157:H7 29. Oregon Health Authority. 2012. Strawberries,
infection: data from a case-control study in 5 deer and other investigations. CD Summary 61(13).
FoodNet sites. Clin Infect Dis 38 Suppl 3:S271S278. https://public.health.oregon.gov/DiseasesConditions/
18. Ryan CA, Tauxe RV, Hosek GW, Wells JG, CommunicableDisease/CDSummaryNewsletter/
Stoesz PA, McFadden HW Jr, Smith PW, Documents/2012/ohd6113.pdf

30. Williams MS, Withee JL, Ebel ED, Bauer NE, 40. Moxley RA. 2004. Escherichia coli O157:H7: an
Scholosser WD, Disney WT, Smith DR, Moxley update on intestinal colonization and virulence
RA. 2010. Determining relationships between the mechanisms. Anim Health Res Rev 5:1533.
seasonal occurrence of Escherichia coli O157:H7 in 41. Rice DH, Sheng HQ, Wynia SA, Hovde CJ.
live cattle, ground beef, and humans. Foodborne 2003. Rectoanal mucosal swab culture is more
Pathog Dis 7:18. sensitive than fecal culture and distinguishes Esch-
31. Elder RO, Keen JE, Siragusa GR, Barkocy- erichia coli O157:H7-colonized cattle and those
Gallagher GA, Koohmaraie M, Laegreid WW. transiently shedding the same organism. J Clin
2000. Correlation of enterohemorrhagic Esche- Microbiol 41:49244929.
richia coli O157 prevalence in feces, hides, and 42. Naylor SW, Gally DL, Low JC. 2005. Entero-
carcasses of beef cattle during processing. Proc haemorrhagic E. coli in veterinary medicine. Int J
Natl Acad Sci USA 97:29993003. Med Microbiol 295:419441.
32. Arthur TM, Bosilevac JM, Nou X, Shackelford 43. Besser TE, Hancock DD, Pritchett LC, McRae
SD, Wheeler TL, Kent MP, Jaroni D, Pauling B, EM, Rice DH, Tarr PI. 1997. Duration of detec-
Allen DM, Koohmaraie M. 2004. Escherichia tion of fecal excretion of Escherichia coli O157:H7
coli O157 prevalence and enumeration of aerobic in cattle. J Infect Dis 175:726729.
bacteria, Enterobacteriaceae, and Escherichia coli 44. Khaitsa ML, Smith DR, Stoner JA, Parkhurst
O157 at various steps in commercial beef pro- AM, Hinkley S, Klopfenstein TJ, Moxley RA.
cessing plants. J Food Prot 67:658665. 2003. Incidence, duration, and prevalence of
33. Brichta-Harhay DM, Guerini MN, Arthur TM, Escherichia coli O157:H7 fecal shedding by feedlot
Bosilevac JM, Kalchayanand N, Shackelford cattle during the finishing period. J Food Prot
SD, Wheeler TL, Koohmaraie M. 2008. Salmo- 66:19721977.
nella and Escherichia coli O157:H7 contamination 45. Sanderson MW, Besser TE, Gay JM, Gay CC,
on hides and carcasses of cull cattle presented Hancock DD. 1999. Fecal Escherichia coli O157:
for slaughter in the United States: an evaluation H7 shedding patterns of orally inoculated calves.
of prevalence and bacterial loads by immuno- Vet Microbiol 69:199205.
magnetic separation and direct plating methods. 46. Smith DR, Blackford MP, Younts SM, Moxley
Appl Environ Microbiol 74:62896297. RA, Gray JT, Hungerford LL, Milton CT,
34. Centers for Disease Prevention and Control. Klopfenstein TJ. 2001. Ecological relationships
2006. Preliminary FoodNet data on the incidence between the prevalence of cattle shedding Esch-
of infection with pathogens transmitted com- erichia coli O157:H7 and characteristics of the
monly through food10 States, United States, cattle or conditions of the feedlot pen. J Food Prot
2005. Morbid Mortal Wkly Rep 55:392395. 64:18991903.
35. Naylor SW, Low JC, Besser TE, Mahajan A, 47. Smith DR, Moxley RA, Clowser SL, Folmer JD,
Gunn GJ, Pearce MC, McKendrick IJ, Smith Hinkley S, Erickson GE, Klopfenstein TJ. 2005.
DG, Gally DL. 2003. Lymphoid follicle-dense Use of rope devices to describe and explain the
mucosa at the terminal rectum is the principal feedlot ecology of Escherichia coli O157:H7 by
site of colonization of enterohemorrhagic Esche- time and place. Foodborne Pathog Dis 2:5060.
richia coli O157:H7 in the bovine host. Infect 48. Savageau MA. 1983. Escherichia coli habitats, cell
Immun 71:15051512. types, and molecular mechanisms of gene control.
36. Grauke LJ, Kudva IT, Yoon JW, Hunt CW, The American Naturalist 122:732744.
Williams CJ, Hovde CJ. 2002. Gastrointestinal 49. Galvani AP, May RM. 2005. Epidemiology:
tract location of Escherichia coli O157:H7 in rumi- dimensions of superspreading. Nature 438:293
nants. Appl Environ Microbiol 68:22692277. 295.
37. Moxley RA, Smith DR. 2010. Attaching-effacing 50. Lloyd-Smith JO, Schreiber SJ, Kopp PE, Getz
Escherichia coli infections in cattle. Vet Clin North WM. 2005. Superspreading and the effect of in-
Am Food Anim Pract 26:2956. dividual variation on disease emergence. Nature
38. Phillips AD, Navabpour S, Hicks S, Dougan G, 438:355359.
Wallis T, Frankel G. 2000. Enterohaemorrhagic 51. Chase-Topping ME, McKendrick IJ, Pearce MC,
Escherichia coli O157:H7 target Peyers patches MacDonald P, Matthews L, Halliday J, Allison L,
in humans and cause attaching/effacing lesions Fenlon D, Low JC, Gunn G, Woolhouse MEJ.
in both human and bovine intestine. Gut 47:377 2007. Risk factors for the presence of high-level
381. shedders of Escherichia coli O157 on Scottish farms.
39. Baehler AA, Moxley RA. 2000. Escherichia coli J Clin Microbiol 45:15941603.
O157:H7 induces attaching-effacing lesions in 52. Chase-Topping M, Gally D, Low C, Matthews
large intestinal mucosal explants from adult cat- L, Woolhouse M. 2008. Super-shedding and
tle. FEMS Microbiol Lett 185:239242. the link between human infection and livestock

498 SMITH
carriage of Escherichia coli O157. Nat Rev 63. Snedeker KG, Campbell M, Sargeant JM. 2012.
Microbiol 6:904912. A systematic review of vaccinations to reduce the
53. Matthews L, Low JC, Gally DL, Pearce MC, shedding of Escherichia coli O157 in the faeces of
Mellor DJ, Heesterbeek JA, Chase-Topping M, domestic ruminants. Zoonoses Public Health
Naylor SW, Shaw DJ, Reid SW, Gunn GJ, 59:126138.
Woolhouse ME. 2006. Heterogeneous shedding 64. Brashears MM, Galyean ML, Loneragan GH,
of Escherichia coli O157 in cattle and its implica- Mann JE, Killinger-Mann K. 2003. Prevalence of
tions for control. Proc Natl Acad Sci USA 103:547 Escherichia coli O157:H7 and performance by
552. beef feedlot cattle given Lactobacillus direct-fed
54. Arthur TM, Brichta-Harhay DM, Bosilevac JM, microbials. J Food Prot 66:748754.
Kalchayanand N, Shackelford SD, Wheeler TL, 65. Peterson RE, Klopfenstein TJ, Erickson GE,
Koohmaraie M. 2010. Super shedding of Esche- Folmer J, Hinkley S, Moxley RA, Smith DR.
richia coli O157:H7 by cattle and the impact on 2007. Effect of Lactobacillus acidophilus strain
beef carcass contamination. Meat Sci 86:3237. NP51 on Escherichia coli O157:H7 fecal shedding
55. Cobbold RN, Hancock DD, Rice DH, Berg and finishing performance in beef feedlot cattle.
J, Stilborn R, Hovde CJ, Besser TE. 2007. J Food Prot 70:287291.
Rectoanal junction colonization of feedlot cattle 66. Loneragan GH, Brashears MM. 2005. Pre-
by Escherichia coli O157:H7 and its association harvest interventions to reduce carriage of E. coli
with supershedders and excretion dynamics. Appl O157 by harvest-ready feedlot cattle. Meat Sci
Environ Microbiol 73:15631568. 71:7278.
56. Williams ML, Pearl DL, Bishop KE, Lejeune JT. 67. Miller MF, Loneragan GH, Harris DD, Adams
2013. Use of multiple-locus variable-number tan- KD, Brooks JC, Brashears MM. 2008. Environ-
dem repeat analysis to evaluate Escherichia coli mental dust exposure as a factor contributing to
O157 subtype distribution and transmission dy- an increase in Escherichia coli O157:H7 and Sal-
namics following natural exposure on a closed monella populations on cattle hides in feedyards.
beef feedlot facility. Foodborne Pathog Dis 10:827 J Food Prot 71:20782081.
834. 68. Arthur TM, Bosilevac JM, Brichta-Harhay DM,
57. Callaway TR, Carr MA, Edrington TS, Anderson Guerini MN, Kalchayanand N, Shackelford SD,
RC, Nisbet DJ. 2009. Diet, Escherichia coli Wheeler TL, Koohmaraie M. 2007. Transporta-
O157:H7, and cattle: a review after 10 years. Curr tion and lairage environment effects on preva-
Issues Mol Biol 11:6779. lence, numbers, and diversity of Escherichia coli
58. Smith DR, Vogstad AR. 2012. Vaccination as a O157:H7 on hides and carcasses of beef cattle at
method of E. coli O157:H7 reduction in feedlot processing. J Food Prot 70:280286.
cattle, p 133142. In Callaway TR, Edrington TS 69. Arthur TM, Bosilevac JM, Brichta-Harhay DM,
(ed), On Farm Strategies to Control Foodborne Kalchayanand N, King DA, Shackelford SD,
Pathogens. Nova Science Publishers Inc, Wheeler TL, Koohmaraie M. 2008. Source
Hauppauge, NY. tracking of Escherichia coli O157:H7 and Salmo-
59. Berry ED, Wells JE. 2010. Escherichia coli O157: nella contamination in the lairage environment at
H7: recent advances in research on occurrence, commercial U.S. beef processing plants and
transmission, and control in cattle and the produc- identification of an effective intervention. J Food
tion envrionment. Adv Food Nutr Res 60:67118. Prot 71:17521760.
60. Smith DR, Klopfenstein T, Moxley RA, Milton 70. Asper DJ, Sekirov I, Finlay BB, Rogan D, Potter
CT, Hungerford LL, Gray JT. 2002. An evalua- AA. 2007. Cross reactivity of enterohemorrhagic
tion of three methods to clean feedlot water Escherichia coli O157:H7-specific sera with non-
tanks. The Bovine Practitioner 36:14. O157 serotypes. Vaccine 25:82628269.
61. Folmer J, Macken C, Moxley R, Smith D, 71. Dohoo IR, Martin SW, Stryhn H. 2003. Measures
Brashears M, Hinkley S, Erickson G, of association, p 121138. In Dohoo IR, Martin SW,
Klopfenstein T. 2003. Intervention strategies for Stryhn H (ed), Veterinary Epidemiologic Research.
reduction of E. coli O157:H7 in feedlot steers. AVC, Inc, Charlottetown, PEI, Canada.
Nebraska Beef Cattle Report MP 80-A:2223. 72. Halloran ME. 1998. Concepts of infectious dis-
62. LeJeune JT, Besser TE, Rice DH, Berg JL, ease epidemiology, p 529554. In Rothman KJ,
Stilborn RP, Hancock DD. 2004. Longitudinal Greenland S (ed), Modern Epidemiology, 2nd ed.
study of fecal shedding of Escherichia coli Lippincott-Raven, Philadelphia, PA.
O157:H7 in feedlot cattle: predominance and 73. Beaudeau F, Fourichon C. 1998. Estimating rel-
persistence of specific clonal types despite mas- ative risk of disease from outputs of logistic re-
sive cattle population turnover. Appl Environ gression when the disease is not rare. Prev Vet
Microbiol 70:377384. Med 36:243256.

74. Allen KJ, Rogan D, Finlay BB, Potter AA, Asper 84. McNeilly TN, Mitchell MC, Rosser T, McAteer
DJ. 2011. Vaccination with type III secreted S, Low JC, Smith DGE, Huntley JF, Mahajan A,
proteins leads to decreased shedding in calves Gally DL. 2010. Immunization of cattle with a
after experimental infection with Escherichia coli combination of purified intimin-531, EspA and
O157. Can J Vet Res 75:98105. Tir significantly reduces shedding of Escherichia
75. Garmendia J, Frankel G, Creprin VF. 2005. coli O157:H7 following oral challenge. Vaccine 28:
Enteropathogenic and enterohemorrhagic Esche- 14221428.
richia coli infections: translocation, translocation, 85. Yekta MA, Goddeeris BM, Vanrompay D,
translocation. Infect Immun 73:25732585. Cox E. 2011. Immunization of sheep with a com-
76. Goosney DL, Knoechel DG, Finlay BB. 1999. bination of intimin [gamma]; EspA and EspB
Enteropathogenic E. coli, Salmonella, and Shigella: decreases Escherichia coli O157:H7 shedding. Vet
masters of host cell cytoskeletal exploitation. Immunol Immunopathol 140:4246.
Emerg Infect Dis 5:216223. 86. Vilte DA, Larzabal M, Garbaccio S, Gammella
77. Mahajan A, Currie CG, Mackie S, Tree J, M, Rabinovitz BC, Elizondo AM, Cantet RJC,
McAteer S, McKendrick I, McNeilly TN, Roe A, Delgado F, Meikle V, Cataldi A, Mercado EC.
La Ragione RM, Woodward MJ, Gally DL, 2011. Reduced faecal shedding of Escherichia coli
Smith DGE. 2009. An investigation of the ex- O157:H7 in cattle following systemic vaccination
pression and adhesin function of H7 flagella in the with gamma-intimin C280 and EspB proteins.
interaction of Escherichia coli O157:H7 with bo- Vaccine 29:39623968.
vine intestinal epithelium. Cell Microbiol 11:121 87. Potter AA, Klashinsky S, Li Y, Frey E,
137. Townsend H, Rogan D, Erickson G, Hinkley
78. Erdem AL, Avelino F, Xicohtencatl-Cortes J, S, Klopfenstein T, Moxley RA, Smith DR,
Giron JA. 2007. Host protein binding and adhe- Finlay BB. 2004. Decreased shedding of Esche-
sive properties of H6 and H7 flagella of attaching richia coli O157:H7 by cattle following vaccination
and effacing Escherichia coli. J Bacteriol 189: with type III secreted proteins. Vaccine 22:362
74267435. 369.
79. Bretschneider G, Berberov EM, Moxley RA. 88. Vilte DA, Larzabal M, Mayr UB, Garbaccio S,
2007. Reduced intestinal colonization of adult Gammella M, Rabinovitz BC, Delgado F, Meikle
beef cattle by Escherichia coli O157:H7 tir deletion V, Cantet RJ, Lubitz P, Lubitz W, Cataldi A,
and nalidixix-acid-resistant mutants lacking fla- Mercado EC. 2012. A systemic vaccine based on
gellar expression. Vet Microbiol 125:381386. Escherichia coli O157:H7 bacterial ghosts (BGs)
80. van Diemen PM, Dziva F, bu-Median A, Wallis reduces the excretion of E. coli O157:H7 in calves.
TS, van den BH, Dougan G, Chanter N, Frankel Vet Immunol Immunopathol 146:169176.
G, Stevens MP. 2007. Subunit vaccines based on 89. Rabinovitz BC, Gerhardt E, Tironi Farinati C,
intimin and Efa-1 polypeptides induce humoral Abdala A, Galarza R, Vilte DA, Ibarra C, Cataldi
immunity in cattle but do not protect against A, Mercado EC. 2012. Vaccination of pregnant
intestinal colonisation by enterohaemorrhagic cows with EspA, EspB, gamma-intimin, and Shiga
Escherichia coli O157:H7 or O26:H. Vet Immunol toxin 2 proteins from Escherichia coli O157:H7
Immunopathol 116:4758. induces high levels of specific colostral antibodies
81. Dean-Nystrom EA, Gansheroff LJ, Mills M, that are transferred to newborn calves. J Dairy Sci
Moon HW, OBrien AD. 2002. Vaccination of 95:33183326.
pregnant dams with intimin (O157) protects 90. Thornton AB, Thomson DU, Loneragan GH,
suckling piglets from Escherichia coli O157:H7 Fox JT, Burkhardt DT, Emery DA, Nagaraja
infection. Infect Immun 70:24142418. TG. 2009. Effects of a siderophore receptor and
82. Dziva F, Vlisidou I, Creprin VF, Wallis TS, porin proteins-based vaccination on fecal shed-
Frankel G, Stevens MP. 2007. Vaccination of ding of Escherichia coli O157:H7 in experimentally
calves with EspA, a key colonisation factor of inoculated cattle. J Food Prot 72:866869.
Escherichia coli O157:H7, induces antigen-specific 91. Thomson DU, Loneragan GH, Thornton AB,
humoral responses but does not confer protection Lechtenberg KF, Emery DA, Burkhardt DT,
against intestinal colonisation. Vet Microbiol 123: Nagaraja TG. 2009. Use of a siderophore recep-
254261. tor and porin proteins-based vaccine to control
83. McNeilly TN, Naylor SW, Mahajan A, Mitchell the burden of Escherichia coli O157:H7 in feedlot
MC, McAteer S, Deane D, Smith DGE, Low JC, cattle. Foodborne Pathog Dis 6:871877.
Gally DL, Huntley JF. 2008. Escherichia coli 92. Woerner DR, Ransom JR, Sofos JN, Scanga JA,
O157:H7 colonization in cattle following systemic Smith GC, Belk KE. 2006. Preharvest processes
and mucosal immunization with purified H7 fla- for microbial control in cattle. Food Prot Trends
gellin. Infect Immun 76:25942602. 26:393400.

500 SMITH
93. Standley T, Paterson J, Skinner K, Rainey B, Clowser S. 2009. A two-dose regimen of a vaccine
Roberts A, Geary T, Smith G, White R. 2008. against type III secreted proteins reduced Esche-
The use of an experimental vaccine in gestating richia coli O157:H7 colonization of the terminal
beef cows to reduce the shedding of Escherichia rectum in beef cattle in commercial feedlots.
coli O157:H7 in the newborn calf. The Professional Foodborne Pathog Dis 6:155161.
Animal Scientist 24:4. 104. Peterson RE, Klopfenstein TJ, Moxley RA,
94. Wileman BW, Thomson DU, Olson KC, Jaeger Erickson GE, Hinkley S, Bretschneider G,
JR, Pacheco LA, Bolte J, Burkhardt DT, Emery Berberov EM, Rogan D, Smith DR. 2007. Effect
DA, Straub D. 2011. Escherichia coli O157:H7 of a vaccine product containing type III secreted
shedding in vaccinated beef calves born to cows proteins on the probability of Escherichia coli
vaccinated prepartum with Escherichia coli O157: O157:H7 fecal shedding and mucosal colonization
H7 SRP vaccine. J Food Prot 74:15991604. in feedlot cattle. J Food Prot 70:25682577.
95. Varela NP, Dick P, Wilson J. 2012. Assessing the 105. Smith DR, Moxley RA, Peterson RE, Klopfenstein
existing information on the efficacy of bovine T, Erickson GE, Clowser SL. 2008. A two-dose
vaccination against Escherichia coli O157:H7a regimen of a vaccine against Escherichia coli O157:
systematic review and meta-analysis. Zoonoses H7 type III secreted proteins reduced environ-
Public Health 60:253268. mental transmission of the agent in a large-scale
96. Vogstad AR, Moxley RA, Erickson GE, commercial beef feedlot clinical trial. Foodborne
Klopfenstein TJ, Smith DR. 2014. Stochastic Pathog Dis 5:589598.
simulation model comparing distributions of 106. Smith DR, Moxley RA, Klopfenstein TJ,
STEC O157 faecal shedding prevalence between Erickson GE. 2009. A randomized longitudinal
cattle vaccinated with type III secreted protein trial to test the effect of regional vaccination
vaccines and non-vaccinated cattle. Zoonoses within a cattle feedyard on Escherichia coli O157:
Public Health 61:283289. H7 rectal colonization, fecal shedding, and hide
97. Fox JT, Thomson DU, Drouillard JS, Thornton contamination. Foodborne Pathog Dis 6:885892.
AB, Burkhardt DT, Emery DA, Nagaraja TG. 107. Peterson RE, Klopfenstein TJ, Moxley RA,
2009. Efficacy of Escherichia coli O157:H7 sidero- Erickson GE, Hinkley S, Rogan D, Smith DR.
phore receptor/porin proteins-based vaccine in 2007. Efficacy of dose regimen and observation of
feedlot cattle naturally shedding E. coli O157. herd immunity from a vaccine against Escherichia coli
Foodborne Pathog Dis 6:893899. O157:H7 for feedlot cattle. J Food Prot 70:25612567.
98. Cull CA, Paddock ZD, Nagaraja TG, Bello NM, 108. OConnor AM, Sargeant JM, Gardner IA,
Babcock AH, Renter DG. 2012. Efficacy of a Dickson JS, Torrence ME, Dewey CE, Dohoo
vaccine and a direct-fed microbial against fecal IR, Evans RB, Gray JT, Greiner M, Keefe G,
shedding of Escherichia coli O157:H7 in a ran- Lefebvre SL, Morley PS, Ramirez A, Sischo W,
domized pen-level field trial of commercial Smith DR, Snedeker K, Sofos J, Ward MP, Wills
feedlot cattle. Vaccine 30:62106215. R. 2010. The REFLECT statement: methods and
99. Van Donkersgoed J, Hancock D, Rogan D, processes of creating reporting guidelines for
Potter AA. 2005. Escherichia coli O157:H7 vac- randomized controlled trials for livestock and
cine field trial in 9 feedlots in Alberta and food safety by modifying the CONSORT state-
Saskatchewan. Can Vet J 46:724728. ment. Zoonoses Public Health 57:95104.
100. Moxley RA, Smith DR, Luebbe M, Erickson GE, 109. Vogstad AR, Moxley RA, Erickson GE,
Klopfenstein TJ, Rogan D. 2009. Escherichia coli Klopfenstein TJ, Smith DR. 2013. Assessment of
O157:H7 vaccine dose-effect in feedlot cattle. heterogeneity of efficacy of a three-dose regimen
Foodborne Pathog Dis 6:879884. of a type III secreted protein vaccine for reducing
101. Peterson RE, Klopfenstein TJ, Moxley RA, STEC O157 in feces of feedlot cattle. Foodborne
Erickson GE, Hinkley S, Rogan D, Smith DR. Pathog Dis 10:678683.
2007. Efficacy of dose regimen and observation of 110. Martin SW, Meek AH, Willeberg P. 1987. De-
herd immunity from a vaccine against Escherichia scriptive epidemiology, p 79120. In Martin SW,
coli O157:H7 for feedlot cattle. J Food Prot Meek AH, Willeberg P, Veterinary Epidemiology:
70:25612567. Principles and Methods, 1st ed. Iowa State Uni-
102. Rich AR, Jepson AN, Luebbe M, Klopfenstein versity Press, Ames. IA.
TJ, Smith DR, Moxley RA. 2010. Vaccination to 111. Fine PEM. 1993. Herd immunity: history, theory,
reduce the prevalence of Escherichia coli O157:H7 practice. Epidemiol Rev 15:265302.
in feedlot cattle fed wet distillers grains plus solu- 112. Smith DR, Gray JT, Moxley RA, Younts-Dahl
bles. Nebraska Beef Cattle Report MP93:9495. SM, Blackford MP, Hinkley S, Hungerford LL,
103. Smith DR, Moxley RA, Peterson RE, Klopfenstein Milton CT, Klopfenstein TJ. 2004. A diagnostic
TJ, Erickson GE, Bretschneider G, Berberov EM, strategy to determine the Shiga toxin-producing

Escherichia coli O157 status of pens of feedlot SD, Wheeler TL, Nou X, Koohmaraie M.
cattle. Epidemiol Infect 132:297302. 2009. Longitudinal study of Escherichia coli O157-
113. Irwin KE, Smith DR, Gray JT, Klopfenstein TJ. H7 in a beef cattle feedlot and role of high-level
2002. Behavior of cattle towards devices to detect shedders in hide contamination. Appl Environ
food-safety pathogens in feedlot pens. Bovine Microbiol 7:65156523.
Pract 36:59. 119. Signorini ML, Tarabla HD. 2010. Interventions
114. Arthur TM, Nou X, Bosilevac JM, Wheeler T, to reduce verocytotoxigenic Escherichia coli in
Koohmaraie M. 2011. Survival of Escherichia coli ground beef in Argentina: a simulation study. Prev
O157:H7 on cattle hides. Appl Environ Microbiol Vet Med 94:3642.
77:30023008. 120. Hurd HS, Malladi S. 2012. An outcomes model to
115. Koohmaraie M, Arthur TM, Bosilevac JM, evaluate risks and benefits of Escherichia coli
Guerini M, Shackelford SD, Wheeler TL. 2005. vaccination in beef cattle. Foodborne Pathog Dis
Post-harvest interventions to reduce/eliminate 9:952961.
pathogens in beef. Meat Sci 71:7991. 121. Wood JC, McKendrick IJ, Gettinby G. 2006.
116. Reicks AL, Brashears MM, Adams KD, Brooks A simulation model for the study of the within-
JC, Blanton JR, Miller MF. 2007. Impact of animal infection dynamics of E. coli O157. Prev
transportation of feedlot cattle to the harvest Vet Med 74:180193.
facility on the prevalence of Escherichia coli 122. Vogstad AR. 2012. Modeling the efficacy and
O157:H7, Salmonella, and total aerobic microor- effectiveness of Escherichia coli O157:H7 pre-
ganisms on hides. J Food Prot 70:1721. harvest interventions. MS thesis. University of
117. Lanzas C, Lu Z, Grohn YT. 2011. Mathematical Nebraska-Lincoln, Lincoln, NE.
modeling of the transmission and control of 123. Cross P, Rigby D, Edwards-Jones G. 2012.
foodborne pathogens and antimicrobial resistance Eliciting expert opinion on the effectiveness and
at preharvest. Foodborne Pathog Dis 8:110. practicality of interventions in the farm and rural
118. Arthur TM, Keen J, Bosworth BT, Brichta- environment to reduce human exposure to Esch-
Harhay DM, Kalchayanand N, Shackelford erichia coli O157. Epidemiol Infect 140:643654.

ESCHERICHIA COLI O104:H4

Escherichia coli O104:H4
Pathogenesis: An Enteroaggregative
E. coli/Shiga Toxin-Producing E. coli
Explosive Cocktail of High Virulence
FERNANDO NAVARRO-GARCIA1
26
INTRODUCTION
In May 2011, an outbreak caused by Escherichia coli of serotype O104:H4 spread

throughout Germany (1). The next month, France also reported a cluster of
E. coli O104:H4 infections (2). A total of 46 deaths, 782 cases of hemolytic-
uremic syndrome (HUS), and 3,128 cases of acute gastroenteritis were officially
attributed to this new clone of enterohemorrhagic E. coli (EHEC) (last update
from European Centre for Disease Prevention and Control, 27 July 2011). Most
or all victims (although diagnosed in different countries in Europe) became
infected in Germany or France. The phenotypic and genotypic characterization
of the E. coli O104:H4 indicated that the isolates from the French and German
outbreaks were common to both incidents. Fenugreek seeds imported from
Egypt, from which sprouts were grown, were implicated as a common source.
However, there is still much uncertainty about whether this is truly the com-
mon cause of the infections, as tests on the seeds did not allow the detection of
any E. coli isolate of serotype O104:H4.
This large outbreak was caused by an unusual EHEC strain that is most
similar to an enteroaggregative E. coli (EAEC) of serotype O104:H4. A signifi-
cant difference, however, is the presence of a prophage encoding the Shiga
toxin (Stx), which is characteristic of EHEC strains (35). This combination of
1
Department of Cell Biology, Centro de Investigacin y de Estudios Avanzados del IPN, Mxico DF, Mexico.
505

506 NAVARRO-GARCIA
genomic features, associating characteristics One burning question is what makes this
from both EAEC and EHEC, represents a outbreak EHEC O104:H4 strain so dangerous?
new pathotype (Fig. 1). Because typical EAEC One explanation is that this strain (in short,
strains are isolated primarily from humans, an EAEC with a phage coding for Stx type 2)
the origin of this outbreak may not be zoo- is a better colonizer of the gut (Fig. 2). The
notic. That was recently confirmed by two enhanced adherence of this strain to intestinal
surveys in Germany and France. No evidence epithelial cells might facilitate systemic ab-
of the Stx-producing E. coli O104:H4 outbreak sorption of Stx and could explain the high
strain or EAEC was found in cattle feces frequency of progression to HUS. It is be-
in northern Germany, the hot spot of the lieved that EAEC of serotype O104:H4 is by
2011 HUS outbreak area (6). Similarly, French itself an emerging serovar that has acquired an
cattle were not a reservoir of the highly vir- original set of virulence factors (Fig. 1).
ulent enteroaggregative Stx-producing E. coli EAEC strains of serotype O104:H4 con-
of serotype O104:H4 (7). Recent identifica- tain a large set of virulence-associated genes
tion from sporadic cases of HUS in France of regulated by the AggR transcription factor
EHEC clones similar to the one responsible (Fig. 1). They include the pAA plasmid genes
of the outbreaks (8) suggests that the EHEC encoding the aggregative adherence fimbriae
O104:H4 pathogen has become endemically (AAF), which anchor the bacterium to the
established in Europe and very likely in the intestinal mucosa (the aggregative, so-called
human population. stacked-brick, adherence pattern on intestinal
FIGURE 1 Hybrid characteristics of E. coli O104:H4 outbreak strain (EAEC/STEC). Schematic representation of the
genes harbored by E. coli O104:H4; the main genes from EAEC or STEC are highlighted: stx2 (coding for Stx 2),
pic, sigA, and sepA (coding for the SPATE proteins); Pic, protein involved in intestinal colonization; SigA, a
homolog of Pet, with cytotoxic activity; SepA, a colonization factor of Shigella), set1AB (coding for ShET1, a
holotoxin AB5), iha (coding for Iha, a STEC adhesin that is an IrgA homolog), aggR, aggABCD, aap, aatPABCD
(genes from EAEC plasmids coding for transcription regulator, AAF/I, dispersin, and dispersin transporter,
respectively), lpf1-2 (coding for Lpf of STEC), terZABCDEF (coding for a cluster for Tellurite resistance), CTX-M15
and TEM-1 (antibiotic resistance genes). SigA and SepA are SPATEs detected mainly in Shigella sp.

CHAPTER 26 E. coli O104:H4 Pathogenesis 507
FIGURE 2 Adherence patterns of EAEC, STEC, and E. coli O104:H4 outbreak strain to epithelial cells.
Subconuent epithelial cell cultures are infected with the different bacterial strains. Cells are xed and stained
with Giemsa stain. From Scalesky et al. 1999. Infect Immun 67:3410; Paton et al. 2001. Infect Immun 69:6999;
and Martina Bielaszewska. http://ecdc.europa.eu/en/press/events/Documents/22-231111-Breakthroughs-in-
molecular-epidemiology-Bielaszewska.pdf. doi:10.1128/microbiolspec.EHEC-0008-2013.f2
epithelial cells) and induce inflammation, as biogenesis are located in two regions of the
well as a protein-coat secretion system (Aat) plasmid: region 1, containing the genes that
that secretes the protein dispersin (9). A encode the pilin, chaperonin, and usher, and
switch of the virulence plasmid (pAA) to- region 2, containing the regulator-encoding
gether with the type of AAF could be an ad- gene (designated aggR) (10, 15, 16).
ditional explanation for the higher virulence EAEC strains of serotype O104:H4 also pro-
of this outbreak strain (Fig. 2). Indeed, the duce a variable number of serine protease au-
outbreak EHEC O104:H4 strain is similar to totransporters of Enterobacteriaceae (SPATEs)
EAEC O104:H4 strain 55989, isolated in the implicated in mucosal damage and coloniza-
late 1990s from a patient with persistent dition. The type V secretion pathway enables a
arrhea in the Central African Republic, and family of proteins to reach the surface with a
to EHEC O104:H4 strain HUSEC041 that very limited number of accessory secretion
was associated in 2001 with very few factors because most information necessary to
HUS cases in Germany. Interestingly, EHEC the translocation process is contained within
O104:H4 strain HUSEC041 carries the plas- the secreted protein itself. These proteins,
mid encoding AAF/III (also present in EAEC which can carry out their own transport to the
strain 55989 in a different size). In contrast, outer membrane, are autotransporter pro-
outbreak EHEC O104:H4 isolates acquired a teins. The current isolate diverges from com-
new plasmid, encoding AAF/I, and lost the mon EAEC isolates in the number and nature
AAF/III-encoding plasmid (3, 4). AAFs are of their SPATE proteases (Fig. 1), namely Pic,
encoded in high-molecular-weight plasmids SigA, and SepA (35). Rasko and colleagues
(designated pAA), and their biogenesis em- speculate that the combined activity of these
ploys the chaperone-usher secretion pathway SPATEs, together with other EAEC virulence
(1012). These fimbriae are members of the Dr factors, accounts for the increased uptake of
superfamily of adhesins that includes strains Stx into the systemic circulation, resulting in
of uropathogenic (UPEC) and diarrheagenic the high rates of HUS. The pic gene has a
E. coli with a genetic organization consisting unique characteristic among the autotrans-
of chaperone, usher, and pilin subunits (13, porter proteins since there are two oppositely
14). AAF/I are flexible, bundle-forming fim- oriented genes in tandem within the pic (she)
briae (15), and the genes responsible for their gene, set1B and set1A (17), which encode the

508 NAVARRO-GARCIA
two subunits of the ShET1 toxin (18); this gene strain when expressed from a multicopy plas-
is also present in Shigella flexneri and UPEC. mid (28). Recently, Iha was determined to
Pic is a mucinase (19) that has recently been have a dual-function urovirulence factor for
shown to promote mucous secretion in the gut E. coli clonal group A strains and other path-
(20), and is responsible for the mucoid diar- ogenic E. coli strains. However, it still needs to
rhea that is a classic symptom of Shigella sp. be determined if the siderophore receptor
and EAEC infection. Interestingly, EC55989 activity, the adhesion phenotype, or both are
contains three copies of this gene, two on important for the enhanced in vivo persis-
the chromosome (intact, EC55989_4682 and tence of Iha demonstrated within the urinary
EC55989_3279) and one on the EAEC plasmid tract (29).
(55989p, truncated), all of which are con-
served in the German outbreak strain. In ad-
dition, a fourth pic gene is present in the E. COLI O104:H4 OUTBREAK STRAIN AS
EAEC plasmid of the outbreak strain (but LOCUS OF ENTEROCYTE EFFACEMENT
missing from EC55989) that seems to be in- (LEE)-NEGATIVE STEC
tact. The outbreak strain also encodes SigA, a
SPATE that cleaves the cytoskeletal protein The E. coli O104:H4 outbreak strain has been
spectrin, inducing rounding and exfoliation compared to typical EHEC outbreaks; how-
of enterocytes (21). The third SPATE, SepA, is ever, EHEC only refers to a clinical condition,
associated with increased S. flexneri virulence and the virulence factors of EHEC and STEC
(22), but its function is unknown. can be different. STEC strains isolated from
As mentioned before, EAEC strains of se- humans with specific clinical signs are called
rotype O104:H4, as a hybrid clone, also pro- EHEC (30). In humans, some strains cause
duce Stx2 and an adhesion-siderophore called severe inflammation of a section of the large
Iha; both proteins are found with high prev- intestine accompanied by hemorrhage of the
alence in Stx-producing E. coli (STEC), in- intestinal mucosa and severe diarrhea (hem-
cluding EHEC (Fig. 1). The ability of STEC orrhagic colitis) or HUS, which can lead to
to cause severe disease in humans is mainly kidney failure and even death. Thus, the
associated with the production of Stx; two STEC strains that cause these clinical pictures
distinct groups, Stx1 and Stx2, with similar are designated as EHEC (31). EHEC strains
biological activity but different immunoge- are considered to be a subset of STEC, but
nicity are well known (23). Members of the our knowledge of STEC comes from out-
Stx1 group are antigenically similar, whereas break investigations and studies of E. coli O157
those of the Stx2 group are heterogeneous and (EHEC) infection, which was first identified
comprise several variants or subtypes (24). An as a pathogen in 1982. The non-O157 STEC
interesting finding that highlights the rele- serogroups are not nearly as well understood,
vance of Stx2 is that Stx2 has been epidemi- partly because outbreaks caused by them are
ologically more associated with severe disease rarely identified. As a whole, the non-O157
in humans than Stx1 (25). Moreover, the stx2d serogroup is less likely than E. coli O157 to
activatable subtype has been associated with cause severe illness; however, some non-O157
high virulence and the ability to cause HUS STEC serogroups can cause the most severe
(26). On the other hand, Iha was first de- manifestations of STEC illness (32).
scribed as an adhesin in an EHEC O157:H7 Pathogenic STEC strains require additional
strain and was named IrgA homolog adhesin, virulence factors enabling adherence, for ex-
based on its homology to the IrgA entero- ample, factors that permit colonization of in-
bactin siderophore receptor of Vibrio cholerae testinal epithelial cells (Fig. 2). EHEC strains
(27) and its ability to confer epithelial cell contain an arrangement of virulence genes
adherence capability to a nonadherent K-12 that contribute to their pathogenesis, such as

the LEE pathogenicity island (PAI), the viru- STEC strains have been designated as non-
lence plasmid pO157, and stx1 and/or stx2. O157 strains. Persons with non-O157 STEC
LEE carries all genes necessary for the for- infection usually have less severe illness, and
mation of the attaching and effacing (A/E) non-O157 STEC strains include many sero-
lesion (33). The A/E histopathology is char- groups (400) with varying virulence, some
acterized by effacement of the brush border typically causing only mild diarrhea and
microvilli, intimate bacterial adherence to the others causing HUS and death; non-O157
enterocyte apical plasma membrane, and the isolates predominantly have Stx1. Additional-
accumulation of polymerized actin beneath ly, there are regional variations in prevalence;
the attached bacteria (34). All A/E pathogens however, six serogroups of STEC have been
carry the LEE PAI (35) that encodes gene associated with 70% of the food safety ill-
regulators (36, 37), the adhesin called intimin nesses in the United States (O26, O45, O103,
(38), a type III secretion system (T3SS) (35), O111, O121, and O145) and have been termed
chaperones (39, 40), and several secreted the Big Six. U.S. Department of Agri-
proteins, including the translocated intimin cultures Food Safety and Inspection Service
receptor called Tir (41). Upon contact with considers these six most frequent serogroups
epithelial cells, EHEC injects a variety of adulterants in ground beef, materials intended
effectors into the cells to modulate cellular for ground beef production, and other non-
functions involved in the host defense re- intact beef products. The Food Safety and
sponse, the dynamics of the cytoskeleton, and Inspection Service plans to test for these
the maintenance of tight junctions. A major adulterants using a PCR protocol that initially
target of virulence factors is the cellular sig- targets the detection of Stx genes, stx1 and/or
naling cascade involved in the construction stx2, and the intimin (eae) gene, and then tests
and modulation of the cytoskeleton and for the six O groups. Besides stx genes and
microfilaments (42). While the bacteria re- adherence factors, increasing evidence shows
main mostly extracellular in the lumen of the that differences in virulence between patho-
gut, the T3SS effectors of A/E pathogens ac- genic and nonpathogenic bacterial strains
cess and manipulate the intracellular envi- can be attributed in part to virulence genes
ronment of host cells. The effectors subvert located in PAIs (43). PAIs usually contain
various host cell processes, which enable the blocks of virulence genes and are greater than
bacteria to colonize, multiply, and contribute 10 kb (44). Several PAIs have been identified
to the disease. Thus, two key factors encoded and characterized in STEC. For instance,
by LEE include the adhesin intimin (eae), the chromosomal LEE PAI was identified in
which binds to Tir (tir), and both are essential E. coli O157:H7 strain EDL933 (45). However,
for the intimate attachment of the bacteria some LEE-positive STEC serotypes have
with the cytoplasmic membrane of the host never been associated with disease, and some
cell (33). LEE-negative STEC serotypes can cause HUS
LEE appears to confer enhanced virulence, and outbreaks, indicating that virulence
since LEE-positive STEC strains are much factors other than those in LEE may contrib-
more commonly associated with HUS and ute to pathogenesis of STEC (43). Thus, in
outbreaks than LEE-negative STEC strains addition to the distribution of different PAIs
(43). EHEC has controversial definitions, most (O island [OI]-122, OI-43/48, OI-57, high-
of them indicating that EHEC is a STEC pathogenicity island) and their virulence
harboring the eae gene, or a STEC implicated genes in STEC, the association of the PAIs and
in the illness, or a STEC that has the same individual virulence genes with STEC sero-
clinical, epidemiological, and pathogenic char- pathotypes linked to severe diseases and
acteristics. These definitions are features of outbreaks was evaluated. Most OI-122, OI-43/
the EHEC O157 serogroup, but most of the 48, and OI-57 virulence genes (pagC, sen, nleB,

510 NAVARRO-GARCIA
efa-1, efa-2, terC, ureC, iha, aidA-1, nleG2-3, transcription factor and other virulence factors
nleG6-2, and nleG5-2) were highly prevalent in encoded in EAEC plasmid. These include,
eae-positive STEC, but they were largely ab- among other factors, the pAA plasmid genes
sent in eae-negative STEC, with the exception encoding the AAF, which anchor the bacteri-
of pagC and iha (46). Phylogenetic analysis um to the intestinal mucosa, and the aggrega-
revealed that iha genes from eae-positive tive adherence pattern on intestinal epithelial
STEC had high similarity (99.6%), whereas cells (Fig. 2), also called stacked-brick pattern.
they had lower sequence similarity (91.1 to
93.6%) with iha genes from eae-negative
Stx2
STEC, indicating that iha from eae-positive
and eae-negative STEC may have evolved in- The primary virulence factor in systemic host
dependently or have different origins (46). responses produced by clinical isolates of
Indeed, it has been reported (47) that iha was STEC is Stx2, but some isolates produce both
carried by a 33,014-bp PAI in STEC serotype Stx1 and Stx2, or rarely, only Stx1 (50, 51).
O91:H strains (eae-negative). In addition, iha Genes for Stx are located on a bacteriophage
was found in pO113 plasmid of STEC serotype (a virus of bacteria) that is associated with all
O113:H21 (eae-negative) (48). Furthermore, pathogenic STEC strains (Fig. 3). Numerous
some of these factors can participate in bac- other factors produced by STEC are believed
terial regulation; it has been reported that to act locally in the intestine rather than sys-
lysogeny with Stx2-encoding bacteriophages temically as do the Stxs. Stx includes two
represses T3SS in EHEC. Deletion of Stx2 major immunologically distinct forms (Stx1
phages from EHEC strains increased the level and Stx2), with minor variants of Stx2 (Stx2a
of T3SS whereas lysogeny decreased T3SS. A to h). In contrast to these genotypic differ-
model is proposed in which Stx2-encoding ences, Stxs share many properties, including
bacteriophages regulate T3SS to coordinate molecular structure, enzymatic activity, re-
epithelial cell colonization that is promoted by ceptor specificity, and intracellular trafficking.
Stx and secreted effector proteins (49). All Stxs possess an AB5 structure with an
Interestingly, E. coli O104:H4 represents a enzymatically active A subunit of approxi-
pathogenic STEC paradigm shift, since it lacks mately 32 kDa in noncovalent association with
the classic EHEC markers (the LEE PAI and five identical B subunits, with each B subunit
the pO157) but possesses Stx2a and EAEC being approximately 7.7 kDa in size (52).
virulence factors, including an AAF (Fig. 1 and X-ray crystallographic analyses of Stxs have
2). Thus, the 2011 E. coli O104:H4 outbreak of shown that the pentameric B subunits form a
hemorrhagic diarrhea in Germany is an ex- ring with the carboxy terminus of the A sub-
ample of the explosive cocktail of high viru- unit interdigitated within the central pore.
lence and resistance that can emerge in this The A subunit has N-glycosidase activity, and
species. In other words, E. coli O104:H4 is an the B subunit binds to a membrane glycolipid,
eae-negative STEC (LEE-negative STEC) that globotriaosylceramide (Gb3) (Fig. 4). The as-
does not possess a classical EHEC plasmid sociation of E. coli Stxs with diarrhea-associ-
(pO157). To be pathogenic, a strain must have ated HUS was established in 1985 (33). In
the necessary properties to cause disease in addition, Stx1 and Stx2 do not target exactly
humans. These properties are called virulence the same tissues and organs although both
factors. In the case of E. coli O104:H4 strains, bind Gb3 and are capable of causing diarrhea-
their STEC-EAEC hybrid characteristics might associated HUS (53, 54). Injection of animals
contribute to make this outbreak EHEC O104: with Stx1 or Stx2 results in preferential dam-
H4 strain so dangerous. Thus, EAEC strains of age to organs including kidney and lung. Stx1
serotype O104:H4 contain a large set of viru- appears to target the lung whereas Stx2
lence-associated genes regulated by the AggR prefers the kidney (55).

FIGURE 3 Transmission electron microscopy (TEM) of Stx2 phage (P13374) induction from lysogenic strain
E. coli K-12 strain TPE2364 (C600) infected with phage lysates of E. coli O104:H4 strain CB13374. (A, B) Ultrathin
sections of two bacterial cells (TPE2364) with maturating virion particles within the cytoplasm indicated by
arrows (bars, 500 nm). (C) TEM of CsCl-puried, negatively stained phage (P13374) particles released by strain
TPE2364 (bar, 100 nm). Short tails (arrows) and a hexagonal head are shown. From Beutin et al. 2012. J Virol
86:10444. doi:10.1128/microbiolspec.EHEC-0008-2013.f3
The Stx receptor is a major determi- importance of Gb3 in Stx action is evident
nant and of central importance to diarrhea- from cell culture and animal studies where
associated HUS kidney disease (56). Gb3 is the absence of Gb3 eliminates the response to
synthesized in the Golgi apparatus of select Stx (57). Gb3 may also be referred to as CD77
eukaryotic cells and is transported to the or the Pk blood antigen. Structure/function
plasma membrane where it resides in the studies suggest that each toxin molecule
outer leaflet with its trisaccharide moiety may express 10 to 15 Gb3-binding sites per B-
facing outward and the hydrocarbon ceramide subunit pentamer (58), explaining the high
(C-16 to C-24) moiety noncovalently arranged affinity (dissociation constant [Kd] 109 M)
within the plasma membrane. The binding of toxin binding. All Stxs, with the exception
subunit of Stx specifically recognizes the ter- of one Stx2 variant called Stx2e, bind Gb3;
minal alpha-1,4-di-galactose of the trisaccha- Stx2e shows preferential binding to the gly-
ride. A molecule of Stx contains five binding colipid globotetraosylceramide. Structural dif-
(B) subunits, each capable of binding one or ferences in the toxin receptor also contribute
more molecules of Gb3, resulting in coopera- to toxin susceptibility. Gb3 is heterogeneous,
tive high-affinity binding of Stx to cells. The displaying variability in fatty acid chain length,

512 NAVARRO-GARCIA
FIGURE 4 Schematic representation of E. coli O104:H4 virulence factors and their targets on the mucosal
epithelium. The targets and virulence factors are extrapolated from their known function in other pathogens,
and the action mechanism for ShET1 is hypothetical. doi:10.1128/microbiolspec.EHEC-0008-2013.f4
degree of bond saturation, and hydroxylation. explanation is that more Gb3 is expressed in
Expression of Gb3 isoforms with long-chain, the kidneys of young people. However, this
unsaturated fatty acids was associated with has proven not to be the case as there is more
increased toxin sensitivity (59). It has been Gb3 extractable from kidneys of adults versus
demonstrated that long-chain, unsaturated children (65, 66). An alternative explanation is
Gb3 isoforms were more likely to induce neg- that pediatric kidney expresses more Gb3 on
ative membrane curvature, leading to Gb3 cells that are more involved in biological re-
clustering and formation of tubular invagi- sponses leading to diarrhea-associated HUS.
nations (60). Finally, Gb3 association with It is interesting that only pediatric kidney
lipid rafts was necessary for efficient intra- expressed Gb3 in glomeruli (66). Another
cellular transport of Stx1 B subunits (61). concept, as mentioned above, is that the plas-
Polyunsaturated fatty acid incorporation into ma membrane microenvironment nearby Gb3
cell membranes, known to disrupt lipid rafts, dictates the biological response to Stx. Mem-
protects cells from Stx intoxication (62). Thus, brane Gb3 expression is a critical determinant
much or perhaps most of the Gb3 in a plasma of toxin sensitivity; cells expressing low Gb3
membrane may not be reactive with Stx. This levels are sensitized to toxicity by increased
cryptic Gb3 may be a function of direct in- membrane expression of toxin receptors,
teraction of the Gb3 ceramide moiety with whereas cells selected for loss of Gb3 ex-
other membrane components, including cho- pression are resistant to Stxs (59, 67).
lesterol, other glycolipids, fatty acids, or pro- To be an effective protein synthesis inhib-
teins (63, 64). This possible mechanism helps itor, Stx must reach the cytoplasm to access
explain why a tissue often appears to bind ribosomes (Fig. 4). Stx uses a retrograde trans-
much less Stx compared to the total amount of port pathway to reach the endoplasmic re-
Gb3 that is extractable from the tissue. ticulum (ER), an intracellular compartment
Most diarrhea-associated HUS is related rich in membrane-associated ribosomes and
to young children although individuals of all containing the cellular machinery necessary
ages can develop this syndrome. A simple for protein translocation into the cytoplasm.

Following binding and cross-linking of Gb3, including apoptosis (Fig. 4) of the affected cell
Stx is internalized by clathrin-dependent or (74). The signaling pathways activated in the
clathrin-independent mechanisms (60, 68). ribotoxic stress response include the mitogen-
Following binding to Gb3 on target eukaryotic activated protein kinases (MAPK), such as
cells, the Stx-receptor complex is internalized p38MAPK (74, 75). Remarkably, these activi-
and locates within endosomes. Rather than ties appear to be due to other than inhibition
moving to lysosomes for degradation, the of protein synthesis. (iii) Receptor binding of
complex is transported in a retrograde man- Stx holotoxin or its B subunit alone can initi-
ner through the Golgi apparatus to the ER ate a cytoplasmic signal-transduction cascade
(69). The resulting A1 and A2 fragments are different from the ribotoxic stress response-
linked through a disulfide bond; the latter activated pathway (76). Conserved features
fragment maintains association with the B of Stx-induced apoptosis include routing ac-
pentamer. In the reductive environment of the tive toxin to the ER where prolonged sig-
ER, the disulfide bond is broken, and the naling through the ribotoxic and ER stress
27-kDa A1 fragment translocates across the responses may activate apoptosis; alterations
ER membrane, a process termed retrotrans- in the balanced expression of pro- and anti-
location. By this process, mammalian cells apoptotic Bcl-2 proteins; a rapid activation
may recognize Stx delivered to the ER as of caspase 8, which, in turn, activates both
misfolded proteins and activate the ER stress caspase-dependent and mitochondrial-depen-
response, and the A1 fragment uses the dent apoptotic signaling pathways; and lack
ER-associated protein degradation pathway to of signaling through Fas and tumor necrosis
reach ribosomes. To avoid the proteasome, A1 factor receptors to trigger apoptosis. Com-
fragments contain few (Stx/Stx1) or no (Stx2) plement activation products have been de-
lysine residues (52), which may minimize tected in the serum and plasma of patients
ubiquitination, and A1 fragments associate with HUS, and an in vitro study showed that
with ribosomal proteins, which may inhibit Stx2 not only damages the kidney directly but
proteasomal transport (70). Once the enzy- also indirectly via complement in two ways.
matically active A1 subunit is released into the First, it activates complement, and second,
cytoplasm, it inactivates the eukaryotic ribo- it delays the functions of its control protein
some by removal of a single adenine base factor H on the cell surface, both known to
from 28S rRNA within the large (60S) ribo- damage the kidney (52).
somal subunit (71), acting as a highly specific Thus, Stx can exert different responses in a
N-glycosidase. This is an irreversible process cell type-specific manner. The final result of
that renders the ribosome defective for in- these events can be cell death (apoptosis, ne-
teraction with eukaryotic peptide elongation crosis) or inflammatory responses in cells that
factor for binding of aminoacyl-tRNA and elon- remain viable, and perhaps other intermediate
gation of nascent peptide (72), resulting in responses that can be due to the Stx type, the
inhibition of protein synthesis. host cell type, and/or the bacterial genetic
Stx can exert its effects on eukaryotic cells background. STEC is known to cause hemor-
by one of three known mechanisms: (i) Inac- rhagic enterocolitis and HUS. Stx plays a role
tivation of ribosomes and inhibition of cyto- in the occurrence of blood in the feces and in
plasmic protein synthesis can result in cell HUS by its action on the endothelial cells of
death (73). (ii) Stx-dependent generation of blood vessels in the intestinal submucosa and
the depurinated 28S rRNA in ribosomes ini- in the renal glomeruli (Fig. 4). Epidemiologi-
tiates a signal-transduction response known cally, Stx2 seems to be more important than
as the ribotoxic stress response that leads Stx1 in development of HUS, which was also
to activation of cytokines, chemokines, or seen in the E. coli O104:H4 outbreak strain.
other factors that result in numerous events The action of Stx is not limited to inhibition of

514 NAVARRO-GARCIA
protein synthesis. Stx induces macrophages to STEC. For example, patients may present
express tumor necrosis factor alpha (tumor with acute renal failure in the absence of
necrosis factor-alpha), interleukin (IL)-1 beta, bloody diarrhea (79).
and IL-6 in vitro. These cytokines and lipo- Other events that may influence the path-
polysaccharide (LPS) are reported to increase ogenesis of STEC are the phage origin of
the susceptibility of cells to Stx. A variety of Stx and Stx variability (Fig. 3). Stx phages
cells such as tubular epithelial cells may be display extensive genetic mosaicism; however,
targets for Stx-mediated apoptosis. Apoptosis genes encoding the Stx A and B subunits are
is considered to contribute to the pathogenesis generally located downstream of the anti-
of HUS caused by STEC (77). HUS is pri- terminator Q and the PR promoter. As a
marily due to the production and transloca- consequence of this orientation, toxin genes
tion of Stxs across the intestinal epithelial are late genes optimally expressed upon in-
barrier (52). HUS may manifest after STEC duction of the lytic cycle. STEC may possess
is no longer detectable in the stool. Follow- cryptic lambdoid prophages that serve as
ing toxin-mediated damage to colonic blood sources for recombination events, yielding
vessels, Stxs may enter the bloodstream al- novel toxin-converting phages, and Stx phages
though free toxins have not been detected in expressing new tail assemblies may expand
the circulation of patients with HUS. The risk the host range of toxin-producing organisms
of progression to extraintestinal complications (80). Lysogenic conversion to the toxigenic
is increased in patients infected with STEC phenotype may occur if recipient bacteria
expressing Stx2, either alone or in combina- display phage receptors and possess integra-
tion with Stx1 or Stx2c (78). The kidneys and tion sites within the genome (Fig. 3). Thus, Stx
central nervous system are most frequently phages are responsible for the dissemination
damaged by Stxs. HUS is a constellation of of stx genes in E. coli and other enteric bac-
hematological, renal, and neurological com- teria. Following induction, Stx phages can in-
plications that develops in 10 to 15% of fect other bacteria in vivo and in vitro. Stx
patients with hemorrhagic colitis. Compli- phages may be considered to represent highly
cations include thrombocytopenia and hemo- mobile genetic elements that play an impor-
lytic anemia with schistocytes (fragmented tant role in the expression of Stx, in horizontal
erythrocytes) present in blood smears. The gene transfer, and hence in genome diversifi-
characteristics of acute renal failure, which cation (81). On the other hand, purified Stx
may follow STEC infection, include oliguria alone is capable of producing systemic com-
or anuria, swollen glomerular endothelial cells plications, including HUS, in animal models of
detached from the basement membrane, intra- disease. Stx2a is more potent than Stx1. Epi-
glomerular fibrin deposition, and thrombotic demiologic studies suggest that Stx2 subtypes
microangiopathy. Mesangiolysis or mesangial also differ in potency. Indeed, by examining
hyperplasia has been described in some HUS protein synthesis inhibition using Vero mon-
cases. Renal tubular injury may be present but key kidney cells and inhibition of metabolic
is not a consistent finding late in the course activity using primary human renal proximal
of HUS (79). Approximately 66% of patients tubule epithelial cells, it was found that Stx2a,
with HUS require dialysis. Central nervous Stx2d, and elastase-cleaved Stx2d were at
system involvement may present as lethargy, least 25 times more potent than Stx2b and
irritability, seizures, paresis, and coma. Long- Stx2c; in vivo this potency was also assessed in
term sequelae include renal insufficiency, hy- mice. Stx2b and Stx2c had potencies similar to
pertension, hyperactivity and distractability, that of Stx1, whereas Stx2a, Stx2d, and elas-
and insulin-dependent diabetes mellitus. Mor- tase-cleaved Stx2d were 40 to 400 times more
tality from HUS is 3 to 5%. There is variability potent than Stx1 (82). Furthermore, by using
in signs and symptoms following ingestion of the classification for seropathotypes (A to E),

it has been found that the stx(2) variant was TonB-dependent, whereby the protein com-
mainly associated with strains of seropatho- plex TonB/ExbB/ExbD provides the energy
type A, whereas most of the strains of sero- required for active transport (29, 89). Fur-
pathotype C possessed the stx(2-vhb) variant, thermore, iha expression is regulated by
which was frequently associated with stx(2), the protein Fur, a ferric uptake regulator
stx(2-vha), or stx(2c). Levels of stx(2) and (90). In the presence of iron, dimerized Fur
stx(2)-related mRNA were higher in strains binds to iha promoter regions through a
belonging to seropathotype A and in those sequence-specific protein-DNA interaction
strains of seropathotype C that express the stx and represses iha transcription. Under iron-
(2) variant than in the remaining strains of limiting conditions, Fur is unable to interact
seropathotype C. The stx(2-vhb) genes were with the DNA, and as a consequence, iha
the least expressed (83). Another complicating transcription is derepressed (29, 90). Iron is
factor in diarrhea-associated HUS is that an- an important environmental cue, and iron-
tibiotics are not recommended in the earlier limiting conditions induce virulence-related
phases, i.e., before the appearance of bloody genes in a number of pathogens (91, 92). Free
diarrhea because STEC bacteria respond to iron levels are typically low at mucosal sur-
some antibiotics by producing excess Stx (84, faces, due to binding by host lactoferrin, and
85). therefore the induction of iron-scavenging
mechanisms is an important bacterial in vivo
survival strategy (92). Recently, it was found
Iha
that bacterial growth under conditions simu-
The chromosomally encoded adherence- lating colonic, but not ileal, short-chain fatty
conferring protein Iha, a homolog of V. chol- acid (SCFA) concentration increases iha ex-
erae IrgA (28), is one of a range of novel pression in three tested STEC strains, with the
adhesins identified among STEC strains strongest expression detected in LEE-negative
(Fig. 4). Iha was first characterized in E. coli STEC O113:H21 strain 98NK2, as expression
O157:H7, but it is distributed widely among of iha in O157:H7 STEC strain 98NK2 is sub-
LEE-negative and LEE-positive STEC strains ject to Fur-mediated iron repression. How-
and UPEC (28, 86). In a study in which the ever, exogenous iron did not repress iha
difference did not reach statistical signifi- expression in the presence of colonic SCFAs
cance, it was reported that an iha deletion in either 98NK2 or O157:H7 strain EDL933.
mutant of O157:H7 STEC was impaired in Moreover, exposure to the iron chelator 2,2-
adherence to HeLa cells. But there was a dipyridyl caused no further enhancement of
highly significant increase in adherence to iha expression over that induced by colonic
both HeLa and MDBK cells when iha was SCFAs. These findings indicate that SCFAs
expressed from a plasmid in a nonpiliated regulate iha expression in STEC indepen-
recombinant E. coli host (28). The potential dently of iron. Increased expression of iha
importance of this adhesin lies in the high under colonic but not ileal SCFA conditions
prevalence (91%) of iha in STEC belonging to possibly may contribute to preferential colo-
different seropathotypes and the presence nization of the human colon by STEC (93).
of multiple iha copies in some strains (28, 87). Thus, Iha is a potentially important acces-
In the UPEC strain CFT073, Iha functions sory virulence factor for several pathotypes of
as a urovirulence factor by contributing to E. coli, including STEC, UPEC, and avian-
colonization in a mouse urinary tract infec- pathogenic strains (28, 86, 92, 9496), func-
tion model (88). Additionally, Iha from UPEC tioning as an adhesin and, at least for UPEC,
strain UCB34 functions as a catecholate as a catecholate siderophore receptor (29).
siderophore receptor in E. coli K-12. The ca- This latter function is consistent with the
pacity of Iha to transport siderophores is fact that iha is induced under iron-limiting

516 NAVARRO-GARCIA
conditions and is subject to Fur-mediated re- (98). The lpf2 locus contains five genes
pression in iron-replete environments in both (lpfABCDD) but lacks an lpfE homolog, and
UPEC and O157:H7 STEC strains (29, 90) and instead, the lpfD gene is duplicated in O157
in the hypervirulent LEE-negative STEC strains. The in vitro adherence phenotype
strain 98NK2. conferred by this locus is not well understood.
When a nonfimbriated E. coli K-12 strain was
used to express the lpf2 locus, a less adherent
Long Polar Fimbriae (Lpf)
phenotype to Caco-2 intestinal cells was ob-
E. coli O157:H7 possesses two Lpf operons, lpf1 served (98). However, a previous study using
and lpf2, both of which contain genes closely a random transposon mutagenesis showed
related to the Lpf of Salmonella enterica that an insertion in the lpfD2 gene caused
serovar Typhimurium (97). Expression of lpf1 increased bacterial adherence to HeLa cells
and lpf2 is induced during the late exponen- (103). Further, disruption of the gene encoding
tial-phase growth in tissue culture media at the major fimbrial subunit, the lpfA2 gene,
pH 6.5 and 37C or under iron-restricted resulted in a reduction in initial adherence to
conditions (98), and has been found to influ- Caco-2 cells, although adherence to HeLa cells
ence E. coli O157:H7 adherence to cultured was unaffected (98). Finally, expression of
epithelial monolayers (99, 100). The first locus the lpf2 gene in a nonfimbriated E. coli strain
(lpf1) is located in an O157-specific island resulted in the appearance of thin, fibrilla-like
(OI-141) of approximately 5.9 kb, and inserted structures on the bacterial surface that were
in the yhjX-yhjW intergenic region (in relation structurally different from those observed for
to the E. coli K-12 chromosome). The lpf1 a strain expressing the cloned lpf1 genes (98).
operon contains six genes (lpfABCCDE) sim- Homologs of lpf genes have also been iden-
ilar in sequence and gene order to the Sal- tified for non-O157 STEC strains, and their role
monella lpfABCDE genes (101). Expression in adherence has also been explored (Fig. 4). It
of the STEC O157:H7 lpf1 operon in a non- was found that a Tn5phoA mutant of the LEE-
fimbriated E. coli ORN172 strain was reported negative STEC O113:H21 strain exhibited re-
to increase adherence to HeLa and MDCK duced adherence to Chinese hamster ovary-K1
cells, and peritrichous short fimbriae were (CHO-K1) cells, and further analysis mapped
observed (101). It was further demonstrated the mutation to a gene with homology to the
that stx-positive and stx-negative STEC lpfD2 gene (104). Sequencing analysis demon-
O157:H7 mutated in the lpfA1 gene (encoding strated that the STEC O113:H21 strain pos-
the major fimbrial subunit) exhibited a re- sesses an lpf operon (also referred as lpfO113)
duction in adherence to epithelial cells and containing four genes (lpfABCD genes) found
displayed a diffuse adherence pattern (87, at the same chromosomal location as the STEC
101). A recent study showed that STEC O157:H7 lpf2 locus (OI-154) (104). Inactivation
O157:H7 adhered more abundantly to surfaces of the lpfAO113 gene in STEC O113:H21
coated with fibronectin, laminin, and collagen resulted in a significant reduction in micro-
IV and that a reduced binding of the bacteria colony formation on CHO-K1 cells, and when
to these extracellular matrix proteins was the lpfO113 genes were introduced into the
observed for an lpf mutant (lpf1) strain (102). nonfimbriated E. coli K-12 strain ORN103, the
This study demonstrated that Lpf1 and ex- bacteria adhered in a localized pattern, as op-
tracellular matrix proteins interact, and their posed to a diffuse adherence, indicating that
interaction may contribute to STEC O157:H7 the Lpf2 fimbria homologs may promote
colonization of the gastrointestinal tract. The interbacterial interactions (104).
second lpf operon, the lpf2 locus, is approxi- Due to the relatively subtle effects of lpf
mately 6.8 kb and located in OI-154; it is mutations on adherence in vitro, coupled with
inserted in the glmS-pstS intergenic region the divergent findings from in vivo or organ

culture experiments, the precise role of Lpf in (e.g., AAF/I to III), heat-stable enterotoxin,
E. coli O157:H7 adherence remains somewhat transporters, and other secreted proteins
unclear. Therefore, the role of the E. coli (e.g., the serine protease autotransporter Pic,
O157:H7 lpf loci was further tested in an infant which has mucinase activity), as well as mul-
rabbit model, which mimics the diarrhea tiple factors contributing to EAEC-induced
and gut pathology, including the histopatho- inflammation. However, none of these factors
logical A/E lesions, seen in patients with are found in all EAEC isolates, and no single
STEC infection. By performing competition factor has ever been implicated in EAEC vir-
experiments between E. coli O157:H7 and ulence. Thereby, EAEC isolates are genetically
an isogenic lpf1 lpf2 double mutant, it was a heterogeneous group of E. coli strains (112).
found that the mutant was outcompeted in EAEC strains were first associated with per-
the ileum, cecum, and midcolon of rabbits, sistent diarrhea in infants from developing
confirming that Lpf contributes to intestinal countries; since then they have increasingly
colonization (105). Unexpectedly in this study, been linked as a cause of acute and persistent
it was observed that the lpf1 lpf2 double mu- diarrhea in young infants and children in
tant showed an increased adherence to co- developing and industrialized countries, in
lonic epithelial cells in vitro, and transmission individuals infected with human immuno-
electron microscopy revealed curli-like struc- deficiency virus, as a cause of acute diarrhea
tures on the surface of this mutant. Interest- in travelers from industrialized regions, and
ingly, deletion of csgA per se did not appear to with foodborne outbreaks. A major effect of
affect intestinal colonization. Therefore, in EAEC infection is on the malnourished chil-
addition to conclusively demonstrating that dren in developing countries (112). Thus,
Lpf contributes to E. coli O157:H7 intestinal EAEC without Stx causes diarrhea in persons
colonization, the authors indicated that the who reside in developed countries, such as
regulatory mechanisms controlling expression the United Kingdom (113, 114), Switzerland
of Lpf and curli are interconnected (105). (115), and Japan (116), although EAEC is more
commonly associated with acute and persis-
tent (>14 days) diarrhea of infants, children,
E. COLI O104:H4 OUTBREAK human immunodeficiency virus-positive per-
STRAIN AS EAEC sons, and travelers to developing countries
(112, 117). EAEC isolates are a highly hetero-
As mentioned before, the O104:H4 outbreak geneous group of bacteria (118, 119), and each
strain represents a rare combination of EAEC isolate carries a particular subset of EAEC-
characteristics and Stx2 expression; only associated virulence genes; no single virulence
sporadic cases of disease associated with Stx- factor is consistently associated with EAEC
producing EAEC have previously been de- pathogenesis. The addition of stx2 to the
scribed in Germany (106, 107), France (108, EAEC virulence gene repertoire, however, has
109), the United Kingdom (110), the Republic led to a pathogen that has the capacity to
of Georgia (13), and Japan (111). EAEC was cause disease on a large scale with a poten-
first described in 1987, based on the charac- tially deadly outcome due to the possibility of
teristic adherence phenotype with cultured the development of HUS (Fig. 4).
HEp-2 cells (stacked-brick appearance on
epithelial cells) (Fig. 2). This biological test
AggR and AggR-Regulated Genes (AAF/I,
still remains the gold standard of diagnosis,
Dispersin, Dispersin Transporter)
but it does not distinguish between pathogenic
and nonpathogenic strains. Subsequently, a AggR is a member of the AraC/XylS family of
number of virulence factors have been de- bacterial transcriptional activators (10, 120),
scribed for EAEC strains and include adhesins exhibiting the greatest levels of amino acid

518 NAVARRO-GARCIA
identity with the CfaD (68%), Rns (66%), and aggR is located in an open reading frame of
CsvR (62%) regulators of enterotoxigenic 794 bp that encodes a protein with a predicted
E. coli (10). Proteins belonging to this family molecular size of 29.4 kDa. The cloned aggR
are defined by a conserved 99-amino-acid C gene is sufficient to complement a region 1
terminus domain and regulate diverse cellular clone to confer AAF/I expression, while an
functions including metabolism, stress re- aggR mutant is negative for AAF/I expression.
sponse, and the synthesis of virulence factors. Genomic studies of the pAA plasmid
Multiple epidemiologic studies suggest that revealed a small open reading frame just up-
strains expressing AggR are more likely to stream of the aggR gene in most EAEC strains.
cause diarrheal disease than those without This gene was found to encode a 10-kDa se-
it, proposing the term typical EAEC to de- creted protein that was recognized by the sera
scribe strains harboring the AggR regulon of volunteers fed strain 042. A null mutant of
(121, 122). A number of AggR-regulated genes this gene revealed a unique hyperaggregative
have been described previously in archetype phenotype, and scanning electron microscopy
EAEC strain 042. The genes encoding AAF of the bacterial strains showed collapse of the
were the first found to be regulated by AggR AAF onto the bacterial cell surface. The pro-
(10), followed by aap (encoding the dispersin tein product of this gene was therefore termed
surface protein) (123) and the Aat secretion antiaggregation protein (Aap), nicknamed dis-
system, which is required for transport of persin (123). Dispersin is secreted to the sur-
dispersin to the bacterial surface (124). AggR face of EAEC strains and binds noncovalently
also activates expression of the Aai type VI to LPS of the outer membrane. Data suggest
secretion system (T6SS) in strain 042 (125), that the mechanism of dispersins effect may
though the role of Aai in EAEC virulence be mediated predominantly through its ability
remains unknown. to neutralize the strong negative charge of the
Electron microscopy studies demonstrated LPS, so that the AAFs, which carry a strong
different fimbriae in several EAEC strains positive charge, are free to splay out from the
(126128). The best characterized are AAF I, surface and bind distant sites (130). Interest-
II, and III, which are responsible for the ex- ingly, the secretion of dispersin is dependent
pression of aggregative adherence (15, 17, 129). on the presence of an ABC transporter com-
AAFs are encoded in high-molecular-weight plex also encoded on plasmid pAA (124). The
plasmids (designated pAA), and their biogen- genes encoding this transporter (the Aat
esis employs the chaperone-usher secretion complex) correspond to the site of the previ-
pathway (11, 12, 15). These fimbriae are ously cryptic aggregative adherence probe. As
members of the Dr superfamily of adhesins both dispersin and the Aat complex are under
that includes strains of UPEC and diarrhea- the control of AggR, the latter protein is
genic E. coli with a genetic organization con- emerging as the central regulator of virulence
sisting of chaperone, usher, and pilin subunits functions in EAEC.
(13, 14). Expression of AAF genes requires The AggR regulon is not restricted to the
AggR (10); this protein regulates the genes pAA plasmid. It has been found that in addi-
involved in fimbrial biogenesis for both AAF/I tion to the plasmid-encoded genes, AggR reg-
and AAF/II. AAF/I are flexible, bundle- ulates a chromosomal operon inserted on a
forming fimbriae, and the genes responsible PAI at the pheU locus (125). AggR activates the
for their biogenesis are located in two expression of chromosomal genes, including
unlinked regions of the plasmid: region 1, 25 contiguous genes (aaiAY), which are lo-
containing the genes that encode the pilin, calized to a 117-kb PAI inserted at pheU. Many
chaperonin, and usher, and region 2 (sepa- of these genes have homologs in other gram-
rated by 9 kb), containing the regulator- negative bacteria and were recently proposed
encoding gene (designated aggR) (10, 15, 16). to constitute T6SS. AaiC was identified as

a secreted protein that has no apparent with strain 042 mutated in the major flagellar
homologs within GenBank. EAEC strains subunit FliC. IL-8 release from polarized T84
carrying in-frame deletions of aaiB, aaiG, cells was found to require the AggR activator
aaiO, or aaiP synthesized AaiC, but AaiC se- and the AAF fimbriae, and IL-8 release was
cretion was abolished. Cloning of aai genes significantly less when cells were infected
into E. coli HB101 suggested that aaiAP are with mutants in the minor fimbrial subunit
sufficient for AaiC secretion. AafB (134).
Recent studies in the Nataro laboratory
have revealed that AggR positively regulates
SPATEs (Pic, SepA, SigA)
its own expression in a complex fashion.
AggR binds directly and specifically to two SPATEs and other autotransporters use a type
sites flanking the aggR promoter. Additionally, V secretion system for export to the extra-
aggR promoter was found to be positively cellular space (135, 136). The autotransporters
regulated by the DNA-binding protein FIS and contain all the information necessary for pas-
negatively regulated by the global regulator sage through the inner membrane and the
H-NS. EAEC present in the mouse intestine outer membrane. To mediate its own secre-
possessed relatively high levels of fis promoter tion, an autotransporter contains three func-
and aggR promoter activity and a low level of tional domains: an N-terminal signal sequence,
hns promoter when compared with in vitro an extracellular passenger domain, and a C-
experiments. The data provide significant terminal b-barrel domain. The signal sequence
insights into the regulation cascade leading to initiates Sec-dependent transport across the
aggR expression in the mammalian intestine inner membrane and is proteolytically re-
during EAEC infection (131). A recent study moved in the periplasmic space. The C-ter-
showed that there are at least 44 AggR-regu- minal domain forms a b-barrel pore in the
lated genes in the genome of EAEC strain 042. outer membrane, which facilitates the delivery
Twenty-five of the 44 genes were previously of the passenger domain to the extracellular
known to be so regulated, identified as part of space. The passenger domains of some auto-
the Aai T6SS (16 genes), the dispersin secre- transporters remain anchored to the extra-
tion system (5 genes), and the AAF/II fimbrial cellular face of the outer membrane, but
biogenesis system (4 genes). Sixteen of the 44 SPATEs are released from the bacterial cell by
genes are predicted to encode hypothetical proteolytic nicking of a site between the
proteins, and only 5 of these genes showed b-barrel pore and the passenger domain. The
homology to other genes encoding known mature, secreted SPATEs are 104- to 110-kDa
bacterial proteins, suggesting new virulence- toxins that contain a typical N-terminal serine
related functions (132). protease catalytic domain followed by a highly
AggR has also a role in the inflammatory conserved b-helix motif, which is present in
response against EAEC. Jiang et al. (121) nearly all autotransporters (135, 137). Although
reported that IL-8 levels were higher in feces the general process of SPATE secretion is
of patients infected with aggR- or aafA-con- understood, the details of many events in
taining strains compared with those infected SPATE biogenesis (the chaperone function in
with strains negative for these factors. Re- the periplasm, the mechanism of b-barrel in-
cently, it was also shown that EAEC strains sertion into the outer membrane, the translo-
harboring aggR, aggA, and aap were more cation pathway across the outer membrane,
likely to cause IL-8 induction of >4,100 pg/ml the proteolytic release of the mature protein
from nonpolarized HCT-8 intestinal epithelial from the outer membrane, etc.) remain unre-
cells than EAEC negative for those genes solved (135).
(133). In fact, polarized T84 intestinal cells It has been proposed that SPATEs can be
were found to release IL-8 even when infected divided phylogenetically into two distinct

520 NAVARRO-GARCIA
classes, designated 1 and 2 (138). Class 1 mucus, which was accompanied by an in-
SPATEs are cytotoxic in vitro and induce crease in the number of mucus-containing
mucosal damage on intestinal explants (Fig. goblet cells (Fig. 4). This finding is in accord
4). Although the actions of class 1 SPATEs are with one of the hallmarks of EAEC: the for-
not fully understood, several have been shown mation of biofilm, which comprises a mucous
to enter eukaryotic cells and to cleave cyto- layer with immersed bacteria in the intestines
skeletal proteins (21, 139, 140) while the class of patients. Interestingly, an isogenic pic mu-
2 SPATEs induce mucus release, cleave tant (EAEC Dpic) is unable to cause this
mucin, and confer a subtle competitive ad- mucus hypersecretion. Furthermore, purified
vantage in mucosal colonization (20, 141, 142) Pic was also able to induce intestinal mucus
(Fig. 4). A study to determine the prevalence hypersecretion. Thus, Pic mucinase is re-
of SPATEs in EAEC was performed by seek- sponsible for one of the pathophysiologic
ing 10 genes encoding serine protease auto- features of the diarrhea mediated by EAEC
transporter toxins in a collection of clinical (20). It has also been shown that Pic protease
EAEC isolates. Eighty-six percent of EAEC promotes intestinal colonization and growth
strains harbored genes encoding one or more in the presence of mucin, suggesting a novel
class I cytotoxic SPATE proteins (Pet, Sat, metabolic role for the Pic mucinase in EAEC
EspP, or SigA). Two class II noncytotoxic colonization. Interestingly, it has been found
SPATE genes were found among EAEC that Pic targets a broad range of human leu-
strains: pic and sepA, each originally described kocyte adhesion proteins. Substrate specificity
in S. flexneri 2a. Using a multiplex PCR for five is restricted to glycoproteins rich in O-linked
SPATE genes (pet, sat, sigA, pic, and sepA), the glycans, including CD43, CD44, CD45, CD93,
authors found that most of the Shigella sp. CD162 (P-selectin glycoprotein ligand 1), and
isolates also harbored more than one SPATE, the surface-attached chemokine fractalkine,
whereas members of most other E. coli patho- all implicated in leukocyte trafficking, migra-
types rarely harbored a cytotoxic SPATE gene. tion, and inflammation. Additionally, exposure
SPATEs may be relevant to the pathogenesis of human leukocytes to purified Pic results in
of both EAEC and Shigella spp. (143). polymorphonuclear cell activation, but im-
Pic is localized in the EAEC chromosome paired chemotaxis and transmigration; Pic-
(118, 141), and it was found to be identical to a treated T cells undergo programmed cell
protein termed Shmu (Shigella mucinase), death (145).
which is encoded on the Shigellashe PAI (144). Besides Pic, another SPATE from class II
Through functional analysis of Pic, it was found in E. coli O104:H4 is SepA. In 1995,
found that its proteolytic site is involved in Benjelloun-Touimi et al. (22) described a
Pic mucinase activity, serum resistance, and Tsh homolog designated SepA (for Shigella
hemagglutination. Phenotypes identified for extracellular protein), which is the major
Pic suggest that it is involved in the early extracellular protein of S. flexneri (22). In
stages of the pathogenesis and most probably 1994, the first SPATE was described as a
promotes the intestinal colonization (141, 144). temperature-sensitive hemagglutinin (Tsh)
Pic binds mucin, and this binding was blocked in avian-pathogenic E. coli, which causes dis-
in competition assays using monosaccharide seminated infections in birds (146). Investiga-
constituents of the oligosaccharide side chains tion of the proteolytic activity of SepA by using
of mucin. Moreover, Pic mucinolytic activity a wide range of synthetic peptides found that
decreased when sialic acid was removed from SepA hydrolyzed several of these substrates
mucin. Thus, Pic is a mucinase with lectin-like and that the activity was inhibited by the
activity that can be related to its reported serine protease inhibitor phenylmethylulfonyl
hemagglutinin activity (19). Recently, it was fluoride (147). Several SepA-hydrolyzed pep-
shown that Pic induces hypersecretion of tides were described as specific substrates for

cathepsin G, a serine protease produced by altering toxin (136), because it induces con-
polymorphonuclear leukocytes that was pro- traction of the cytoskeleton, loss of actin stress
posed to play a role in inflammation. However, fibers, and release of focal contacts in HEp-2
unlike cathepsin G, SepA degraded neither and HT29/C1 cell monolayers, followed by
fibronectin nor angiotensin I and had no effect complete cell rounding and detachment. In-
on the aggregation of human platelets. The terestingly, Pet cytotoxicity and enterotoxicity
presence of sepA on the virulence plasmid, as depend on Pet serine protease activity (149). It
well as the recognition of SepA by the sera has been shown that Pet enters the eukaryotic
of monkeys infected with Shigella sp., sug- cell and that trafficking through the vesicular
gested that SepA might be involved in Shigella system is required for the induction of cyto-
pathogenicity (22). However, construction and pathic effects. Thus, after clathrin-mediated
phenotypic characterization of a sepA mutant endocytosis, Pet undergoes a retrograde traf-
suggested that SepA is required neither for ficking to the endoplasmic reticulum to be
entry into cultured cells nor for intracellular translocated into the cytosol (150, 151). Finally,
dissemination. Nevertheless, the sepA mutant an intracellular target, a-fodrin (aII spectrin),
demonstrated a reduced ability to induce both has been found for Pet. Pet binds and cleaves
mucosal atrophy and tissue inflammation in epithelial fodrin (between M1198 and V1199)
the rabbit ligated ileal loop model, indicating in vitro and in vivo, causing fodrin redistri-
that SepA may play a role in tissue invasion, bution within the cells, to form intracellular
although this hypothesis remains to be eluci- aggregates as membrane blebs (139).
dated (22).
On the other hand, a SPATE of class I has
Shigella Enterotoxin 1 (Set1)
been reported in O104:H4, SigA. An initial
report showed that the she PAI, which ShET1 toxin is a subunit toxin encoded by
contains the pic (she) gene, also contains a setA and setB, which are thought to form an
gene encoding a second immunoglobulin A oligomeric toxin consisting of a single 20-kDa
protease-like homolog, sigA, lying 3.6 kb SetA protein associated with a pentamer
downstream and in an inverted orientation of five 7-kDa B subunits (SetB) (18). ShET1
with respect to pic (144). Functional analysis appears to induce intestinal secretion via
showed that SigA is a secreted temperature- cyclic AMP and cyclic GMP; however, the
regulated serine protease capable of degrading precise mechanism of action and detailed
casein. Experiments similar to those used with biochemistry remain inconclusive. Unusually,
Pet (another SPATE of EAEC) revealed that the setAB genes are encoded within the pic
SigA is cytopathic for HEp-2 cells, suggesting gene but on the complementary strand and
that it may be a cell-altering toxin with a role thus have the same prevalence characteristics
in the pathogenesis of Shigella infections. In- and disease associations as pic (141). However,
deed, it was found that SigA binds specifically the role of the ShET1 in EAEC pathogenesis
to HEp-2 cells and degrades recombinant has not been studied, even though it can work
human aII spectrin (a-fodrin) in vitro and on adenylyl cyclase (Fig. 4). In the case of
also cleaves intracellular fodrin in situ, causing the S. flexneri 2a strain, culture filtrates cause
its redistribution within cells, suggesting that significant fluid accumulation in rabbit ileal
the cytotoxic and enterotoxic effects mediated loops, when the bacteria are grown in iron-
by SigA are likely associated with the degra- depleted medium. Also, testing in Ussing
dation of epithelial fodrin (21) (Fig. 4). Fur- chambers showed a greater rise in potential
thermore, SigA was at least partly responsible difference and short circuit current with
for the ability of S. flexneri to stimulate fluid such filtrates compared with the medium
accumulation in ligated rabbit ileal loops control. Ultrafiltration and gel exclusion size
(148). In the case of Pet, it is a cytoskeleton- fractionation of M4243 filtrate revealed that

522 NAVARRO-GARCIA
the activity was in the fraction of approxi- Stx into the systemic circulation, resulting in
mately 60 kDa. It is thought that ShET1 is the high rates of HUS (Fig. 4).
elaborated in vivo, since it elicits an immune Although more studies are needed to ex-
response and may be important in the patho- plain the increased virulence, E. coli O104:H4
genesis of diarrheal illness due to S. flexneri shows that mixed virulence profiles in enteric
(152). pathogens introduced into susceptible popu-
lations can cause serious outbreaks and have
terrible consequences for infected people.
CONCLUSIONS
ACKNOWLEDGMENTS
The genotypes, phenotypes, and phylogeny of
the outbreak isolates demonstrate that the E. The author was supported by a Conacyt grant
coli O104:H4 outbreak strain is a clone that (128490). I thank Paul S. Ugalde for the ar-
combines virulence potentials of two different tistic work in Fig. 1 and 4 and Lucia Chavez-
pathogens: STEC and EAEC. Thus, EAEC of Dueas for organizing the reference database.
serotype O104:H4 is by itself an emerging I declare no conflicts of interest with regard to
serovariant that has acquired an original set of the manuscript.
virulence factors. All shared virulence profiles
combining typical STEC (stx2, iha, lpfO26,
CITATION
lpfO113) and EAEC (aggA, aggR, set1, pic, aap)
loci and expressed phenotypes that define Navarro-Garcia F. 2014. Escherichia coli O104:H4
STEC and EAEC, including production of pathogenesis: an enteroaggregative E. coli/Shiga
Stx2 and aggregative adherence to epithelial toxin-producing E. coli explosive cocktail of
cells (Fig. 4). Additionally, isolates displayed a high virulence. Microbiol Spectrum 2(4):EHEC-
high antibiotic resistance, specifically those 0008-2013.
related to b-lactamase.
EAEC strains of serotype O104:H4 contain
REFERENCES
a large set of virulence-associated genes
regulated by the AggR transcription factor 1. Frank C, Werber D, Cramer JP, Askar M, Faber
M, an der Heiden M, Bernard H, Fruth A,
(Fig. 1). These include the pAA plasmid genes
Prager R, Spode A, Wadl M, Zoufaly A, Jordan
encoding the AAF, which anchor the bac- S, Kemper MJ, Follin P, Muller L, King LA,
terium to the intestinal mucosa and induce Rosner B, Buchholz U, Stark K, Krause G. 2011.
inflammation (Fig. 2), and a protein-coat se- Epidemic profile of Shiga-toxin-producing Esch-
cretion system (Aat), including its secreted erichia coli O104:H4 outbreak in Germany. N Engl
protein, dispersin. A switch of the virulence J Med 365:17711780.
2. Gault G, Weill FX, Mariani-Kurkdjian P,
plasmid (pAA) together with the type of the Jourdan-da Silva N, King L, Aldabe B, Charron
AAF could be an additional explanation for M, Ong N, Castor C, Mace M, Bingen E, Noel H,
the higher virulence of this outbreak strain. Vaillant V, Bone A, Vendrely B, Delmas Y,
Other adhesion factors (Iha, Lfp) could help Combe C, Bercion R, dAndigne E, Desjardin
M, de Valk H, Rolland P. 2011. Outbreak of
this outbreak strain as a synergistic or initial
haemolytic uraemic syndrome and bloody diar-
adhesion factor to guarantee an efficient ad- rhoea due to Escherichia coli O104:H4, south-west
hesion and perhaps Stx delivery. Additionally, France, June 2011. Euro Surveill 16.
EAEC strains of serotype O104:H4 produce a 3. Bielaszewska M, Mellmann A, Zhang W,
variable number of SPATEs, implicated in Kock R, Fruth A, Bauwens A, Peters G,
mucosal damage and colonization. It allows Karch H. 2011. Characterisation of the Esche-
richia coli strain associated with an outbreak of
speculation that the combined activity of these haemolytic uraemic syndrome in Germany, 2011:
SPATEs, together with other EAEC virulence a microbiological study. Lancet Infect Dis 11:671
factors, accounts for the increased uptake of 676.

4. Mellmann A, Harmsen D, Cummings CA, 13. Servin AL. 2005. Pathogenesis of Afa/Dr diffusely
Zentz EB, Leopold SR, Rico A, Prior K, adhering Escherichia coli. Clin Microbiol Rev 18:
Szczepanowski R, Ji Y, Zhang W, McLaughlin 264292.
SF, Henkhaus JK, Leopold B, Bielaszewska M, 14. Boisen N, Struve C, Scheutz F, Krogfelt
Prager R, Brzoska PM, Moore RL, Guenther S, KA, Nataro JP. 2008. New adhesin of entero-
Rothberg JM, Karch H. 2011. Prospective ge- aggregative Escherichia coli related to the Afa/Dr/
nomic characterization of the German entero- AAF family. Infect Immun 76:32813292.
hemorrhagic Escherichia coli O104:H4 outbreak 15. Nataro JP, Deng Y, Maneval DR, German AL,
by rapid next generation sequencing technology. Martin WC, Levine MM. 1992. Aggregative ad-
PLoS One 6:e22751. herence fimbriae I of enteroaggregative Esche-
5. Rasko DA, Webster DR, Sahl JW, Bashir A, richia coli mediate adherence to HEp-2 cells and
Boisen N, Scheutz F, Paxinos EE, Sebra R, Chin hemagglutination of human erythrocytes. Infect
CS, Iliopoulos D, Klammer A, Peluso P, Lee L, Immun 60:22972304.
Kislyuk AO, Bullard J, Kasarskis A, Wang S, Eid 16. Savarino SJ, Fox P, Deng Y, Nataro JP. 1994.
J, Rank D, Redman JC, Steyert SR, Frimodt- Identification and characterization of a gene
Moller J, Struve C, Petersen AM, Krogfelt KA, cluster mediating enteroaggregative Escherichia
Nataro JP, Schadt EE, Waldor MK. 2011. coli aggregative adherence fimbria I biogenesis.
Origins of the E. coli strain causing an outbreak of J Bacteriol 176:49494957.
hemolytic-uremic syndrome in Germany. N Engl 17. Behrens M, Sheikh J, Nataro JP. 2002. Regula-
J Med 365:709717. tion of the overlapping pic/set locus in Shigella
6. Wieler LH, Semmler T, Eichhorn I, Antao EM, flexneri and enteroaggregative Escherichia coli.
Kinnemann B, Geue L, Karch H, Guenther S, Infect Immun 70:29152925.
Bethe A. 2011. No evidence of the Shiga toxin- 18. Fasano A, Noriega FR, Liao FM, Wang W,
producing E. coli O104:H4 outbreak strain or Levine MM. 1997. Effect of shigella enterotoxin 1
enteroaggregative E. coli (EAEC) found in cattle (ShET1) on rabbit intestine in vitro and in vivo.
faeces in northern Germany, the hotspot of the Gut 40:505511.
2011 HUS outbreak area. Gut Pathog 3:17. 19. Gutierrez-Jimenez J, Arciniega I, Navarro-
7. Auvray F, Dilasser F, Bibbal D, Kerouredan M, Garcia F. 2008. The serine protease motif of Pic
Oswald E, Brugere H. 2012. French cattle is not a mediates a dose-dependent mucolytic activity
reservoir of the highly virulent enteroaggregative after binding to sugar constituents of the mucin
Shiga toxin-producing Escherichia coli of serotype substrate. Microb Pathog 45:115123.
O104:H4. Vet Microbiol 158:443445. 20. Navarro-Garcia F, Gutierrez-Jimenez J, Garcia-
8. Monecke S, Mariani-Kurkdjian P, Bingen E, Tovar C, Castro LA, Salazar-Gonzalez H, Cordova
Weill FX, Baliere C, Slickers P, Ehricht R. 2011. V. 2010. Pic, an autotransporter protein secreted
Presence of enterohemorrhagic Escherichia coli by different pathogens in the Enterobacteriaceae
ST678/O104:H4 in France prior to 2011. Appl family, is a potent mucus secretagogue. Infect
Environ Microbiol 77:87848786. Immun 78:41014109.
9. Huang DB, Mohanty A, DuPont HL, Okhuysen 21. Al-Hasani K, Navarro-Garcia F, Huerta J,
PC, Chiang T. 2006. A review of an emerging Sakellaris H, Adler B. 2009. The immunogenic
enteric pathogen: enteroaggregative Escherichia SigA enterotoxin of Shigella flexneri 2a binds to
coli. J Med Microbiol 55:13031311. HEp-2 cells and induces fodrin redistribution in
10. Nataro JP, Yikang D, Yingkang D, Walker K. intoxicated epithelial cells. PLoS One 4:e8223.
1994. AggR, a transcriptional activator of aggre- 22. Benjelloun-Touimi Z, Sansonetti PJ, Parsot C.
gative adherence fimbria I expression in entero- 1995. SepA, the major extracellular protein of
aggregative Escherichia coli. J Bacteriol 176: Shigella flexneri: autonomous secretion and in-
46914699. volvement in tissue invasion. Mol Microbiol 17:
11. Elias WP Jr, Czeczulin JR, Henderson IR, 123135.
Trabulsi LR, Nataro JP. 1999. Organization of 23. Paton AW, Paton JC. 1998. Detection and char-
biogenesis genes for aggregative adherence fimbria acterization of Shiga toxigenic Escherichia coli by
II defines a virulence gene cluster in enteroaggre- using multiplex PCR assays for stx1, stx2, eaeA,
gative Escherichia coli. J Bacteriol 181:17791785. enterohemorrhagic E. coli hlyA, rfbO111, and
12. Bernier C, Gounon P, Le Bouguenec C. 2002. rfbO157. J Clin Microbiol 36:598602.
Identification of an aggregative adhesion fimbria 24. Scheutz F, Teel LD, Beutin L, Pierard D,
(AAF) type III-encoding operon in enteroaggre- Buvens G, Karch H, Mellmann A, Caprioli A,
gative Escherichia coli as a sensitive probe for Tozzoli R, Morabito S, Strockbine NA, Melton-
detecting the AAF-encoding operon family. Infect Celsa AR, Sanchez M, Persson S, OBrien AD.
Immun 70:43024311. 2005. Multicenter evaluation of a sequence-based

524 NAVARRO-GARCIA
protocol for subtyping Shiga toxins and stan- 36. Elliott SJ, Sperandio V, Giron JA, Shin S,
dardizing Stx nomenclature. J Clin Microbiol Mellies JL, Wainwright L, Hutcheson SW,
50:2951263. McDaniel TK, Kaper JB. 2000. The locus of
25. Friedrich AW, Bielaszewska M, Zhang WL, enterocyte effacement (LEE)-encoded regulator
Pulz M, Kuczius T, Ammon A, Karch H. 2002. controls expression of both LEE- and non-LEE-
Escherichia coli harboring Shiga toxin 2 gene encoded virulence factors in enteropathogenic
variants: frequency and association with clinical and enterohemorrhagic Escherichia coli. Infect
symptoms. J Infect Dis 185:7484. Immun 68:61156126.
26. Bielaszewska M, Friedrich AW, Aldick T, 37. Deng W, Li Y, Hardwidge PR, Frey EA,
Schurk-Bulgrin R, Karch H. 2006. Shiga toxin Pfuetzner RA, Lee S, Gruenheid S, Strynakda
activatable by intestinal mucus in Escherichia coli NC, Puente JL, Finlay BB. 2005. Regulation of
isolated from humans: predictor for a severe type III secretion hierarchy of translocators and
clinical outcome. Clin Infect Dis 43:11601167. effectors in attaching and effacing bacterial
27. Mey AR, Wyckoff EE, Oglesby AG, Rab E, pathogens. Infect Immun 73:21352146.
Taylor RK, Payne SM. 2002. Identification of the 38. Jerse AE, Yu J, Tall BD, Kaper JB. 1990. A ge-
Vibrio cholerae enterobactin receptors VctA and netic locus of enteropathogenic Escherichia coli
IrgA: IrgA is not required for virulence. Infect necessary for the production of attaching and
Immun 70:34193426. effacing lesions on tissue culture cells. Proc Natl
28. Tarr PI, Bilge SS, Vary JC Jr, Jelacic S, Habeeb Acad Sci USA 87:78397843.
RL, Ward TR, Baylor MR, Besser TE. 2000. 39. Abe A, de Grado M, Pfuetzner RA, Sanchez-
Iha: a novel Escherichia coli O157:H7 adherence- Sanmartin C, Devinney R, Puente JL,
conferring molecule encoded on a recently ac- Strynadka NC, Finlay BB. 1999. Enteropatho-
quired chromosomal island of conserved struc- genic Escherichia coli translocated intimin recep-
ture. Infect Immun 68:14001407. tor, Tir, requires a specific chaperone for stable
29. Leveille S, Caza M, Johnson JR, Clabots C, secretion. Mol Microbiol 33:11621175.
Sabri M, Dozois CM. 2006. Iha from an Esche- 40. Elliott SJ, Hutcheson SW, Dubois MS, Mellies
richia coli urinary tract infection outbreak clonal JL, Wainwright LA, Batchelor M, Frankel G,
group A strain is expressed in vivo in the mouse Knutton S, Kaper JB. 1999. Identification of
urinary tract and functions as a catecholate CesT, a chaperone for the type III secretion of
siderophore receptor. Infect Immun 74:3427 Tir in enteropathogenic Escherichia coli. Mol
3436. Microbiol 33:11761189.
30. Pierard D, De Greve H, Haesebrouck F, Mainil 41. Kenny B, DeVinney R, Stein M, Reinscheid DJ,
J. 2012. O157:H7 and O104:H4 Vero/Shiga toxin- Frey EA, Finlay BB. 1997. Enteropathogenic E.
producing Escherichia coli outbreaks: respective coli (EPEC) transfers its receptor for intimate
role of cattle and humans. Vet Res 43:13. adherence into mammalian cells. Cell 91:511520.
31. Gyles CL. 2007. Shiga toxin-producing Esche- 42. Navarro-Garcia F, Serapio-Palacios A, Ugalde-
richia coli: an overview. J Anim Sci 85:E45E62. Silva P, Tapia-Pastrana G, Chavez-Duenas L.
32. Gould LH, Demma L, Jones TF, Hurd S, Vugia 2013. Actin cytoskeleton manipulation by effector
DJ, Smith K, Shiferaw B, Segler S, Palmer A, proteins secreted by diarrheagenic Escherichia
Zansky S, Griffin PM. 2009. Hemolytic uremic coli pathotypes. Biomed Res Int 2013:374395.
syndrome and death in persons with Escherichia 43. Karmali MA. 2003. The medical significance of
coli O157:H7 infection, foodborne diseases active Shiga toxin-producing Escherichia coli infections.
surveillance network sites, 20002006. Clin Infect An overview. Methods Mol Med 73:17.
Dis 49:14801485. 44. Gal-Mor O, Finlay BB. 2006. Pathogenicity
33. Kaper JB, Nataro JP, Mobley HL. 2004. Path- islands: a molecular toolbox for bacterial viru-
ogenic Escherichia coli. Nat Rev Microbiol 2:123 lence. Cell Microbiol 8:17071719.
140. 45. Coombes BK, Gilmour MW, Goodman CD. 2011.
34. Knutton S, Lloyd DR, McNeish AS. 1987. Ad- The evolution of virulence in non-O157 Shiga
hesion of enteropathogenic Escherichia coli to toxin-producing Escherichia coli. Front Microbiol
human intestinal enterocytes and cultured human 2:90.
intestinal mucosa. Infect Immun 55:6977. 46. Ju W, Shen J, Toro M, Zhao S, Meng J.
35. Elliott SJ, Wainwright LA, McDaniel TK, 2013. Distribution of pathogenicity islands OI-122,
Jarvis KG, Deng YK, Lai LC, McNamara BP, OI-43/48, OI-57 and a high-pathogenicity island
Donnenberg MS, Kaper JB. 1998. The complete (in Shiga toxin-producing Escherichia coli. Appl
sequence of the locus of enterocyte effacement Environ Microbiol 79:34063412.
(LEE) from enteropathogenic Escherichia coli 47. Schmidt H, Zhang WL, Hemmrich U, Jelacic S,
E2348/69. Mol Microbiol 28:14. Brunder W, Tarr PI, Dobrindt U, Hacker J,

Karch H. 2001. Identification and characteriza- 59. Schweppe CH, Bielaszewska M, Pohlentz G,
tion of a novel genomic island integrated at sel C Friedrich AW, Buntemeyer H, Schmidt MA,
in locus of enterocyte effacement-negative, Shiga Kim KS, Peter-Katalinic J, Karch H, Muthing
toxin-producing Escherichia coli. Infect Immun J. 2008. Glycosphingolipids in vascular endo-
69:68636873. thelial cells: relationship of heterogeneity in
48. Newton HJ, Sloan J, Bulach DM, Seemann T, Gb3Cer/CD77 receptor expression with differ-
Allison CC, Tauschek M, Robins-Browne RM, ential Shiga toxin 1 cytotoxicity. Glycoconj J
Paton JC, Whittam TS, Paton AW, Hartland 25:291304.
EL. 2009. Shiga toxin-producing Escherichia coli 60. Romer W, Berland L, Chambon V, Gaus K,
strains negative for locus of enterocyte efface- Windschiegl B, Tenza D, Aly MR, Fraisier V,
ment. Emerg Infect Dis 15:372380. Florent JC, Perrais D, Lamaze C, Raposo G,
49. Xu X, McAteer SP, Tree JJ, Shaw DJ, Wolfson Steinem C, Sens P, Bassereau P, Johannes L.
EB, Beatson SA, Roe AJ, Allison LJ, Chase- 2007. Shiga toxin induces tubular membrane
Topping ME, Mahajan A, Tozzoli R, Woolhouse invaginations for its uptake into cells. Nature
ME, Morabito S, Gally DL. 2012. Lysogeny with 450:670675.
Shiga toxin 2-encoding bacteriophages represses 61. Falguieres T, Mallard F, Baron C, Hanau D,
type III secretion in enterohemorrhagic Esche- Lingwood C, Goud B, Salamero J, Johannes L.
richia coli. PLoS Pathog 8:e1002672. 2001. Targeting of Shiga toxin B-subunit to ret-
50. Imamovic L, Jofre J, Schmidt H, Serra-Moreno rograde transport route in association with de-
R, Muniesa M. 2009. Phage-mediated Shiga toxin tergent-resistant membranes. Mol Biol Cell
2 gene transfer in food and water. Appl Environ 12:24532468.
Microbiol 75:17641768. 62. Spilsberg B, Llorente A, Sandvig K. 2007. Poly-
51. Strockbine NA, Jackson MP, Sung LM, Holmes unsaturated fatty acids regulate Shiga toxin
RK, OBrien AD. 1988. Cloning and sequencing of transport. Biochem Biophys Res Commun
the genes for Shiga toxin from Shigella dysenteriae 364:283288.
type 1. J Bacteriol 170:11161122. 63. Mahfoud R, Manis A, Binnington B, Ackerley C,
52. Tesh VL. Induction of apoptosis by Shiga toxins. Lingwood CA. 2010. A major fraction of glyco-
Future Microbiol 5:431453. sphingolipids in model and cellular cholesterol-
53. Tam P, Mahfoud R, Nutikka A, Khine AA, containing membranes is undetectable by their
Binnington B, Paroutis P, Lingwood C. 2008. binding proteins. J Biol Chem 285:3604936059.
Differential intracellular transport and binding of 64. Lingwood CA, Binnington B, Manis A, Branch
verotoxin 1 and verotoxin 2 to globotriaosylcera- DR. 2010. Globotriaosyl ceramide receptor func-
mide-containing lipid assemblies. J Cell Physiol tion where membrane structure and pathology
216:750763. intersect. FEBS Lett 584:18791886.
54. Chark D, Nutikka A, Trusevych N, Kuzmina J, 65. Boyd B, Lingwood C. 1989. Verotoxin receptor
Lingwood C. 2004. Differential carbohydrate glycolipid in human renal tissue. Nephron 51:207
epitope recognition of globotriaosyl ceramide by 210.
verotoxins and a monoclonal antibody. Eur J 66. Lingwood CA. 1994. Verotoxin-binding in human
Biochem 271:405417. renal sections. Nephron 66:2128.
55. Rutjes NW, Binnington BA, Smith CR, Maloney 67. Pudymaitis A, Armstrong G, Lingwood CA.
MD, Lingwood CA. 2002. Differential tissue tar- 1991. Verotoxin-resistant cell clones are deficient
geting and pathogenesis of verotoxins 1 and 2 in in the glycolipid globotriosylceramide: differential
the mouse animal model. Kidney Int 62:832845. basis of phenotype. Arch Biochem Biophys
56. Lingwood CA. 1996. Role of verotoxin receptors 286:448452.
in pathogenesis. Trends Microbiol 4:147153. 68. Sandvig K, Garred O, Prydz K, Kozlov JV,
57. Okuda T, Tokuda N, Numata S, Ito M, Ohta M, Hansen SH, van Deurs B. 1992. Retrograde
Kawamura K, Wiels J, Urano T, Tajima O, transport of endocytosed Shiga toxin to the en-
Furukawa K. 2006. Targeted disruption of Gb3/ doplasmic reticulum. Nature 358:510512.
CD77 synthase gene resulted in the complete 69. Sandvig K, Bergan J, Dyve AB, Skotland T,
deletion of globo-series glycosphingolipids and Torgersen ML. 2010. Endocytosis and retrograde
loss of sensitivity to verotoxins. J Biol Chem transport of Shiga toxin. Toxicon 56:11811185.
281:1023010235. 70. McCluskey AJ, Poon GM, Bolewska-Pedyczak
58. Bast DJ, Banerjee L, Clark C, Read RJ, Brunton E, Srikumar T, Jeram SM, Raught B, Gariepy J.
JL. 1999. The identification of three biologically 2008. The catalytic subunit of Shiga-like toxin 1
relevant globotriaosyl ceramide receptor binding interacts with ribosomal stalk proteins and is in-
sites on the Verotoxin 1 B subunit. Mol Microbiol hibited by their conserved C-terminal domain.
32:953960. J Mol Biol 378:375386.

526 NAVARRO-GARCIA
71. Endo Y, Tsurugi K, Yutsudo T, Takeda Y, seropathotypes A and C. Microbiology 154:176

Ogasawara T, Igarashi K. 1988. Site of action of a 186.
Vero toxin (VT2) from Escherichia coli O157:H7 84. Zhang X, McDaniel AD, Wolf LE, Keusch GT,
and of Shiga toxin on eukaryotic ribosomes. Waldor MK, Acheson DW. 2000. Quinolone
RNA N-glycosidase activity of the toxins. Eur J antibiotics induce Shiga toxin-encoding bac-
Biochem 171:450. teriophages, toxin production, and death in mice.
72. Obrig TG, Moran TP, Brown JE. 1987. The J Infect Dis 181:664670.
mode of action of Shiga toxin on peptide elon- 85. Wong CS, Jelacic S, Habeeb RL, Watkins SL,
gation of eukaryotic protein synthesis. Biochem J Tarr PI. 2000. The risk of the hemolytic-uremic
244:287294. syndrome after antibiotic treatment of Escherichia
73. Obrig TG, Del Vecchio PJ, Brown JE, Moran coli O157:H7 infections. N Engl J Med 342:1930
TP, Rowland BM, Judge TK, Rothman SW. 1936.
1988. Direct cytotoxic action of Shiga toxin on 86. Johnson JR, Russo TA, Tarr PI, Carlino U,
human vascular endothelial cells. Infect Immun Bilge SS, Vary JC Jr, Stell AL. 2000. Molecular
56:23732378. epidemiological and phylogenetic associations of
74. Jandhyala DM, Ahluwalia A, Obrig T, Thorpe two novel putative virulence genes, iha and iroN
CM. 2008. ZAK: a MAP3Kinase that transduces (E. coli), among Escherichia coli isolates from
Shiga toxin- and ricin-induced proinflammatory patients with urosepsis. Infect Immun 68:3040
cytokine expression. Cell Microbiol 10:14681477. 3047.
75. Iordanov MS, Pribnow D, Magun JL, Dinh TH, 87. Toma C, Martinez Espinosa E, Song T,
Pearson JA, Chen SL, Magun BE. 1997. Miliwebsky E, Chinen I, Iyoda S, Iwanaga M,
Ribotoxic stress response: activation of the stress- Rivas M. 2004. Distribution of putative ad-
activated protein kinase JNK1 by inhibitors of the hesins in different seropathotypes of Shiga toxin-
peptidyl transferase reaction and by sequence- producing Escherichia coli. J Clin Microbiol 42:
specific RNA damage to the alpha-sarcin/ricin 49374946.
loop in the 28S rRNA. Mol Cell Biol 17:33733381. 88. Johnson JR, Jelacic S, Schoening LM, Clabots
76. Walchli S, Aasheim HC, Skanland SS, Spilsberg C, Shaikh N, Mobley HL, Tarr PI. 2005. The
B, Torgersen ML, Rosendal KR, Sandvig K. IrgA homologue adhesin Iha is an Escherichia coli
2009. Characterization of clathrin and Syk inter- virulence factor in murine urinary tract infection.
action upon Shiga toxin binding. Cell Signal Infect Immun 73:965971.
21:11611168. 89. Postle K, Kadner RJ. 2003. Touch and go: tying
77. Nakao H, Takeda T. 2000. Escherichia coli Shiga TonB to transport. Mol Microbiol 49:869882.
toxin. J Nat Toxins 9:299313. 90. Rashid RA, Tarr PI, Moseley SL. 2006. Ex-
78. Orth D, Grif K, Khan AB, Naim A, Dierich MP, pression of the Escherichia coli IrgA homolog
Wurzner R. 2007. The Shiga toxin genotype adhesin is regulated by the ferric uptake regula-
rather than the amount of Shiga toxin or the tion protein. Microb Pathog 41:207217.
cytotoxicity of Shiga toxin in vitro correlates with 91. Litwin CM, Calderwood SB. 1993. Role of iron in
the appearance of the hemolytic uremic syn- regulation of virulence genes. Clin Microbiol Rev
drome. Diagn Microbiol Infect Dis 59:235242. 6:137149.
79. Scheiring J, Andreoli SP, Zimmerhackl LB. 92. Touati D. 2000. Iron and oxidative stress in
2008. Treatment and outcome of Shiga-toxin- bacteria. Arch Biochem Biophys 373:16.
associated hemolytic uremic syndrome (HUS). 93. Herold S, Paton JC, Srimanote P, Paton AW.
Pediatr Nephrol 23:17491760. 2009. Differential effects of short-chain fatty
80. Allison HE. 2007. Stx-phages: drivers and acids and iron on expression of iha in Shiga-
mediators of the evolution of STEC and STEC- toxigenic Escherichia coli. Microbiology 155:3554
like pathogens. Future Microbiol 2:165174. 3563.
81. Herold S, Karch H, Schmidt H. 2004. Shiga 94. Ewers C, Li G, Wilking H, Kiessling S, Alt K,
toxin-encoding bacteriophagesgenomes in mo- Antao EM, Laturnus C, Diehl I, Glodde S,
tion. Int J Med Microbiol 294:115121. Homeier T, Bohnke U, Steinruck H, Philipp HC,
82. Fuller CA, Pellino CA, Flagler MJ, Strasser JE, aWieler LH. 2007. Avian pathogenic, uropatho-
Weiss AA. Shiga toxin subtypes display dramatic genic, and newborn meningitis-causing Esche-
differences in potency. Infect Immun 79:1329 richia coli: how closely related are they? Int J Med
1337. Microbiol 297:163176.
83. de Sablet T, Bertin Y, Vareille M, Girardeau JP, 95. Ons E, Bleyen N, Tuntufye HN, Vandemaele F,
Garrivier A, Gobert AP, Martin C. 2008. Dif- Goddeeris BM. 2007. High prevalence iron re-
ferential expression of stx2 variants in Shiga ceptor genes of avian pathogenic Escherichia coli.
toxin-producing Escherichia coli belonging to Avian Pathol 36:411414.

96. Rodriguez-Siek KE, Giddings CW, Doetkott C, 107. Mellmann A, Lu S, Karch H, Xu JG, Harmsen
Johnson TJ, Fakhr MK, Nolan LK. 2005. D, Schmidt MA, Bielaszewska M. 2008. Re-
Comparison of Escherichia coli isolates implicated cycling of Shiga toxin 2 genes in sorbitol-
in human urinary tract infection and avian fermenting enterohemorrhagic Escherichia coli
colibacillosis. Microbiology 151:20972110. O157:NM. Appl Environ Microbiol 74:6772.
97. Baumler AJ, Heffron F. 1995. Identification and 108. Boudailliez B, Berquin P, Mariani-Kurkdjian P,
sequence analysis of lpfABCDE, a putative fim- Ilef D, Cuvelier B, Capek I, Tribout B, Bingen E,
brial operon of Salmonella typhimurium. J Piussan C. 1997. Possible person-to-person
Bacteriol 177:20872097. transmission of Escherichia coli O111associated
98. Torres AG, Kanack KJ, Tutt CB, Popov V, hemolytic uremic syndrome. Pediatr Nephrol 11:
Kaper JB. 2004. Characterization of the second 3639.
long polar (LP) fimbriae of Escherichia coli O157: 109. Morabito S, Karch H, Mariani-Kurkdjian P,
H7 and distribution of LP fimbriae in other Schmidt H, Minelli F, Bingen E, Caprioli A.
pathogenic E. coli strains. FEMS Microbiol Lett 1998. Enteroaggregative, Shiga toxin-producing
238:333344. Escherichia coli O111:H2 associated with an out-
99. Jordan DM, Cornick N, Torres AG, Dean- break of hemolytic-uremic syndrome. J Clin
Nystrom EA, Kaper JB, Moon HW. 2004. Long Microbiol 36:840842.
polar fimbriae contribute to colonization by 110. Willshaw GA, Scotland SM, Smith HR, Rowe B.
Escherichia coli O157:H7 in vivo. Infect Immun 1992. Properties of Vero cytotoxin-producing
72:61686171. Escherichia coli of human origin of O serogroups
100. Torres AG, Milflores-Flores L, Garcia-Gallegos other than O157. J Infect Dis 166:797802.
JG, Patel SD, Best A, La Ragione RM, Martinez- 111. Iyoda S, Terajima J, Wada A, Izumiya H,
Laguna Y, Woodward MJ. 2007. Environmental Tamura K, Watanabe H. 2000. Molecular epi-
regulation and colonization attributes of the long demiology of enterohemorrhagic Escherichia coli.
polar fimbriae (LPF) of Escherichia coli O157:H7. Nihon Saikingaku Zasshi 55:2936.
Int J Med Microbiol 297:177185. 112. Estrada-Garcia T, Navarro-Garcia F. 2012.
101. Torres AG, Giron JA, Perna NT, Burland V, Enteroaggregative Escherichia coli pathotype: a
Blattner FR, Avelino-Flores F, Kaper JB. genetically heterogeneous emerging foodborne
2002. Identification and characterization of enteropathogen. FEMS Immunol Med Microbiol
lpfABCC'DE, a fimbrial operon of enterohemor- 66:281298.
rhagic Escherichia coli O157:H7. Infect Immun 113. Smith HR, Cheasty T, Rowe B. 1997.
70:54165427. Enteroaggregative Escherichia coli and outbreaks
102. Farfan MJ, Cantero L, Vidal R, Botkin DJ, of gastroenteritis in UK. Lancet 350:814815.
Torres AG. 2011. Long polar fimbriae of entero- 114. Tompkins DS, Hudson MJ, Smith HR, Eglin RP,
hemorrhagic Escherichia coli O157:H7 bind to Wheeler JG, Brett MM, Owen RJ, Brazier JS,
extracellular matrix proteins. Infect Immun 79: Cumberland P, King V, Cook PE. 1999. A study
37443750. of infectious intestinal disease in England: mi-
103. Torres AG, Kaper JB. 2003. Multiple elements crobiological findings in cases and controls.
controlling adherence of enterohemorrhagic Commun Dis Public Health 2:108113.
Escherichia coli O157:H7 to HeLa cells. Infect 115. Pabst WL, Altwegg M, Kind C, Mirjanic S,
Immun 71:49854995. Hardegger D, Nadal D. 2003. Prevalence of
104. Doughty S, Sloan J, Bennett-Wood V, enteroaggregative Escherichia coli among children
Robertson M, Robins-Browne RM, Hartland with and without diarrhea in Switzerland. J Clin
EL. 2002. Identification of a novel fimbrial gene Microbiol 41:22892293.
cluster related to long polar fimbriae in locus of 116. Itoh Y, Nagano I, Kunishima M, Ezaki T.
enterocyte effacement-negative strains of entero- 1997. Laboratory investigation of enteroaggre-
hemorrhagic Escherichia coli. Infect Immun gative Escherichia coli O untypeable:H10 associ-
70:67616769. ated with a massive outbreak of gastrointestinal
105. Lloyd SJ, Ritchie JM, Torres AG. 2012. Fim- illness. J Clin Microbiol 35:25462550.
briation and curliation in Escherichia coli O157: 117. Harrington SM, Dudley EG, Nataro JP. 2006.
H7: a paradigm of intestinal and environmental Pathogenesis of enteroaggregative Escherichia coli
colonization. Gut Microbes 3:272276. infection. FEMS Microbiol Lett 254:1218.
106. Bockemuhl J, Aleksic S, Karch H. 1992. Sero- 118. Czeczulin JR, Whittam TS, Henderson IR,
logical and biochemical properties of Shiga-like Navarro-Garcia F, Nataro JP. 1999. Phylogenetic
toxin (verocytotoxin)-producing strains of Esch- analysis of enteroaggregative and diffusely ad-
erichia coli, other than O-group 157, from patients herent Escherichia coli. Infect Immun 67:2692
in Germany. Zentralbl Bakteriol 276:189195. 2699.

528 NAVARRO-GARCIA
119. Okeke IN, Wallace-Gadsden F, Simons HR, adherence in enteroaggregative Escherichia coli.
Matthews N, Labar AS, Hwang J, Wain J. Infect Immun 65:41354145.
2010. Multi-locus sequence typing of entero- 130. Velarde JJ, Varney KM, Inman KG, Farfan M,
aggregative Escherichia coli isolates from Nigerian Dudley E, Fletcher J, Weber DJ, Nataro JP.
children uncovers multiple lineages. PLoS One 5: 2007. Solution structure of the novel dispersin
e14093. protein of enteroaggregative Escherichia coli. Mol
120. Gallegos MT, Michan C, Ramos JL. 1993. The Microbiol 66:11231135.
XylS/AraC family of regulators. Nucleic Acids Res 131. Rossiter AE, Browning DF, Leyton DL,
21:807810. Johnson MD, Godfrey RE, Wardius CA,
121. Jiang ZD, Greenberg D, Nataro JP, Steffen R, Desvaux M, Cunningham AF, Ruiz-Perez F,
DuPont HL. 2002. Rate of occurrence and path- Nataro JP, Busby SJ, and Henderson IR.
ogenic effect of enteroaggregative Escherichia coli 2011. Transcription of the plasmid-encoded toxin
virulence factors in international travelers. J Clin gene from enteroaggregative Escherichia coli is
Microbiol 40:41854190. regulated by a novel co-activation mechanism
122. Huang DB, Mohamed JA, Nataro JP, DuPont involving CRP and Fis. Mol Microbiol 81:179191.
HL, Jiang ZD, Okhuysen PC. 2007. Virulence 132. Morin N, Santiago AE, Ernst RK, Guillot SJ,
characteristics and the molecular epidemiology Nataro JP. 2013. Characterization of the AggR
of enteroaggregative Escherichia coli isolates regulon in enteroaggregative Escherichia coli.
from travellers to developing countries. J Med Infect Immun 81:122132.
Microbiol 56:13861392. 133. Huang DB, DuPont HL, Jiang ZD, Carlin L,
123. Sheikh J, Czeczulin JR, Harrington S, Hicks S, Okhuysen PC. 2004. Interleukin-8 response
Henderson IR, Le Bouguenec C, Gounon P, in an intestinal HCT-8 cell line infected with
Phillips A, Nataro JP. 2002. A novel dispersin enteroaggregative and enterotoxigenic Escherich-
protein in enteroaggregative Escherichia coli. ia coli. Clin Diagn Lab Immunol 11:548551.
J Clin Invest 110:13291337. 134. Harrington SM, Strauman MC, Abe CM,
124. Nishi J, Sheikh J, Mizuguchi K, Luisi B, Burland Nataro JP. 2005. Aggregative adherence fimbriae
V, Boutin A, Rose DJ, Blattner FR, Nataro JP. contribute to the inflammatory response of epi-
2003. The export of coat protein from entero- thelial cells infected with enteroaggregative
aggregative Escherichia coli by a specific ATP- Escherichia coli. Cell Microbiol 7:15651578.
binding cassette transporter system. J Biol Chem 135. Henderson IR, Navarro-Garcia F, Desvaux M,
278:4568045689. Fernandez RC, AlaAldeen D. 2004. Type V
125. Dudley EG, Thomson NR, Parkhill J, Morin NP, protein secretion pathway: the autotransporter
Nataro JP. 2006. Proteomic and microarray story. Microbiol Mol Biol Rev 68:692744.
characterization of the AggR regulon identifies a 136. Navarro-Garcia F, Elias WP. 2011. Autotrans-
pheU pathogenicity island in enteroaggregative porters and virulence of enteroaggregative E. coli.
Escherichia coli. Mol Microbiol 61:12671282. Gut Microbes 2:1324.
126. Knutton S, Shaw RK, Bhan MK, Smith HR, 137. Dautin N, Bernstein HD. 2007. Protein secretion
McConnell MM, Cheasty T, Williams PH, in gram-negative bacteria via the autotransporter
Baldwin TJ. 1992. Ability of enteroaggregative pathway. Annu Rev Microbiol 61:89112.
Escherichia coli strains to adhere in vitro to 138. Dutta S, Lalitha PV, Ware LA, Barbosa A,
human intestinal mucosa. Infect Immun 60:2083 Moch JK, Vassell MA, Fileta BB, Kitov S,
2091. Kolodny N, Heppner DG, Haynes JD, Lanar
127. Suzart S, Guth BE, Pedroso MZ, Okafor UM, DE. 2002. Purification, characterization, and im-
Gomes TA. 2001. Diversity of surface structures munogenicity of the refolded ectodomain of the
and virulence genetic markers among entero- Plasmodium falciparum apical membrane antigen
aggregative Escherichia coli (EAEC) strains with 1 expressed in Escherichia coli. Infect Immun
and without the EAEC DNA probe sequence. 70:31013110.
FEMS Microbiol Lett 201:163168. 139. Canizalez-Roman A, Navarro-Garcia F. 2003.
128. Vial PA, Robins-Browne R, Lior H, Prado V, Fodrin CaM-binding domain cleavage by Pet
Kaper JB, Nataro JP, Maneval D, Elsayed A, from enteroaggregative Escherichia coli leads to
Levine MM. 1988. Characterization of entero- actin cytoskeletal disruption. Mol Microbiol
adherent-aggregative Escherichia coli, a putative 48:947958.
agent of diarrheal disease. J Infect Dis 158:7079. 140. Navarro-Garcia F, Canizalez-Roman A, Luna J,
129. Czeczulin JR, Balepur S, Hicks S, Phillips A, Sears C, Nataro JP. 2001. Plasmid-encoded toxin
Hall R, Kothary MH, Navarro-Garcia F, Nataro of enteroaggregative Escherichia coli is internal-
JP. 1997. Aggregative adherence fimbria II, a ized by epithelial cells. Infect Immun 69:1053
second fimbrial antigen mediating aggregative 1060.

141. Henderson IR, Czeczulin J, Eslava C, Noriega 147. Benjelloun-Touimi Z, Si Tahar M, Montecucco
F, Nataro JP. 1999. Characterization of pic, a C, Sansonetti PJ, Parsot C. 1998. SepA, the
secreted protease of Shigella flexneri and entero- 110 kDa protein secreted by Shigella flexneri:
aggregative Escherichia coli. Infect Immun 67: two-domain structure and proteolytic activity.
55875596. Microbiology 144(Pt 7):18151822.
142. Harrington SM, Sheikh J, Henderson IR, Ruiz- 148. Al-Hasani K, Henderson IR, Sakellaris H,
Perez F, Cohen PS, Nataro JP. 2009. The Pic Rajakumar K, Grant T, Nataro JP, Robins-
protease of enteroaggregative Escherichia coli Browne R, Adler B. 2000. The sigA gene which is
promotes intestinal colonization and growth in borne on the she pathogenicity island of Shigella
the presence of mucin. Infect Immun 77:2465 flexneri 2a encodes an exported cytopathic pro-
2473. tease involved in intestinal fluid accumulation.
143. Boisen N, Ruiz-Perez F, Scheutz F, Krogfelt Infect Immun 68:24572463.
KA, Nataro JP. 2009. Short report: high preva- 149. Navarro-Garcia F, Sears C, Eslava C, Cravioto
lence of serine protease autotransporter cytotoxins A, Nataro JP. 1999. Cytoskeletal effects induced
among strains of enteroaggregative Escherichia by pet, the serine protease enterotoxin of entero-
coli. Am J Trop Med Hyg 80:294301. aggregative Escherichia coli. Infect Immun 67:
144. Rajakumar K, Sasakawa C, Adler B. 1997. Use of 21842192.
a novel approach, termed island probing, identi- 150. Navarro-Garcia F, Canizalez-Roman A,
fies the Shigella flexneri she pathogenicity island Burlingame KE, Teter K, Vidal JE. 2007. Pet, a
which encodes a homolog of the immunoglobulin non-AB toxin, is transported and translocated
A protease-like family of proteins. Infect Immun into epithelial cells by a retrograde trafficking
65:46064614. pathway. Infect Immun 75:21012109.
145. Ruiz-Perez F, Wahid R, Faherty CS, 151. Navarro-Garcia F, Canizalez-Roman A, Vidal
Kolappaswamy K, Rodriguez L, Santiago A, JE, Salazar MI. 2007. Intoxication of epithelial
Murphy E, Cross A, Sztein MB, Nataro JP. 2011. cells by plasmid-encoded toxin requires clathrin-
Serine protease autotransporters from Shigella mediated endocytosis. Microbiology 153:2828
flexneri and pathogenic Escherichia coli target a 2838.
broad range of leukocyte glycoproteins. Proc Natl 152. Fasano A, Noriega FR, Maneval DR Jr,
Acad Sci USA 108:1288112886. Chanasongcram S, Russell R, Guandalini S,
146. Provence DL, Curtiss R 3rd. 1994. Isolation and Levine MM. 1995. Shigella enterotoxin 1: an en-
characterization of a gene involved in hemagglu- terotoxin of Shigella flexneri 2a active in rabbit
tination by an avian pathogenic Escherichia coli small intestine in vivo and in vitro. J Clin Invest
strain. Infect Immun 62:13691380. 95:28532861.

THE WAY FORWARD

The Way Forward
VANESSA SPERANDIO1
27
RESEARCH INTO THE MOLECULAR MECHANISMS OF EHEC VIRULENCE
We have observed a vertical leap in our understanding of EHECs virulence.

In the past edition of this book, the locus of enterocyte effacement (LEE) and its
encoded type 3 secretion system (T3SS) had been recently discovered (1, 2).
However, few effectors were known at the time, with Tir (3) and intimin (4)
dominating research on the molecular mechanisms involved in the formation of
attaching and effacing (AE) lesions. Structural insights into the T3SS came later,
with the description of the EscF needle (5) and the EspA filament (6) forming
the unique translocon of the EHEC and EPEC T3SSs. The number of effectors
quickly expanded from the six LEE-encoded effectors, to the first hints that
effectors encoded outside of the LEE existed (7), to the large expansion of
their repertoire (8). Next-generation sequencing of many EHEC genomes also
highlighted the fact that different strains of EHEC and enteropathogenic E. coli
(EPEC) carry different combinations of these effectors (9). Vigorous research
was initially devoted to understanding the mechanism through which EHEC
engaged the actin cytoskeleton to form AE lesions. These studies involved Tir
and intimin interactions, but also extensive studies on the EspFu/TccP effectors
(3, 1016). More recently, studies of non-AE-related effectors and their role in
1
Department of Microbiology and Department of Biochemistry, The University of Texas Southwestern Medical Center,
Dallas, TX 75390.
doi:10.1128/9781555818791.ch27
533

534 SPERANDIO
more discrete actin rearrangements, as well as lipopolysaccharide has profound immune

in modulation of the host immune response, modulation properties that we dont com-
have taken the front seat (1724). Looking pletely understand (28). There is also not
forward, we need to understand how different enough information on the central nervous
combinations of T3SS effectors affect the vir- system action of Shiga toxin, where research
ulence potential of EHEC strains. We are also is still in its infancy although this is a key
starting to study the hierarchy of secretion of aspect of EHEC-mediated disease (29). We
these effectors (25, 26) and how they work in still have to deal with the mystery of why
concert. Knowledge of which effectors are some patients develop Stx-mediated hemo-
acting within a mammalian cell at any given lytic uremic syndrome (HUS) and others
time, and how their functions amplify or an- dont, let alone explaining the age distribu-
tagonize their phenotypes, is the next frontier tion usually associated with HUS. Could this
in understanding the role of these proteins in be dependent on the hosts Gb3 affinity to
the bacterial/host interplay. Another still un- Shiga toxin? Is age distribution dependent on
resolved issue is how the T3SS is regulated to changes in the gastrointestinal (GI) microbi-
shift from secreting the translocon proteins ota? It is documented that usually children
(EspA, EspB, and EspD) to secreting bona fide under 5 years of age are more susceptible to
effectors within epithelial cells. There is also HUS. As it turns out, ones GI microbiota does
the question of how the EspA filament is not become stable until the age of 5 (30, 31).
disassembled during the infection process to The influence of microbiota, and potentially
allow the close contact between the bacteria probiotics, on Shiga toxin-mediated disease
and the host. Finally, the big question that has again been at center stage. Fukuda et al.
remains open is, how does EHEC cause diar- (32) reported that strains of Bifidobacterium
rhea in the human intestine, and which are (largely employed as probiotics) with en-
the main players in this disease process? hanced ability to produce acetate can inhibit
An ongoing puzzle that remains largely Shiga toxin translocation from the intesti-
unanswered is how EHEC initially adheres to nal lumen to the blood, having a protective
the host enterocytes. Is this adherence medi- effect against EHEC infections. Hence, these
ated by the T3SS per se, and if so, what is the studies beg the question, do differences in
receptor? Would this implicate the EspA fila- the microbiota composition change the hosts
ment and maybe EspB and EspD? Conversely, susceptibility to EHEC infection? Which role
could this adherence be mediated by other does dysbiosis play in susceptibility to GI
fimbrial adhesins? EHEC does encode for sev- pathogens?
eral adhesins in its genome; certain adhesins
such as the long polar fimbriae (Lpf) have
been implicated in tissue tropism, and other EHEC/MICROBIOTA/HOST RELATIONSHIPS
structures have been shown to contribute to
cattle carriage of this pathogen (27). How- EHEC colonizes the human colon, where it
ever, a clear picture of the contribution of encounters trillions of bacterial species that
these adhesive structures to EHEC intestinal comprise the microbiota. EHEC learned to
adherence and their role in intestinal tropism exploit signal and nutritional cues provided by
is still missing. this microbiota to time regulation of its viru-
Although we still have many mysteries to lence traits. EHEC senses the signal AI-3
solve in the intestinal phase of EHEC infec- produced by the microbiota, the levels of the
tion, we are also searching for many answers carbon sources glucose and fucose, and the
regarding Shiga toxin. The mechanism of nitrogen source ethanolamine to fine-tune
action of the A subunit is known, but inde- expression of the LEE genes (3336). Cues
pendently, Shiga toxin in combination with from the microbiota also seem to affect the

CHAPTER 27 The Way Forward 535
production of Shiga toxin (37) and, as men- signals and the microbiota AI-3 converge at
tioned above (32), the hosts susceptibility to the biochemical level to the receptor QseC,
this toxin. Given the profound association which controls a complex virulence program
that humans have with their GI microbiota within the EHEC cell (33, 4244). However,
(38) and the interplay between the microbial there are many other human hormones pres-
GI flora and enteric GI pathogens, several ent in the GI tract that can be exploited by
questions will populate future research. Is the EHEC as excellent cues towards successful
microbiota protective against pathogens, or host colonization. Consequently, one has to
can it be exploited by pathogens to gauge ask, which other human hormones can EHEC
their niche within the intestine? Does EHEC respond to? How will responses to different
adapt itself metabolically to subvert competi- human hormones be integrated? Do medi-
tion for nutrients with the microbiota? Do cations that alter human signaling processes
differences in microbiota affect host suscep- affect susceptibility to EHEC disease?
tibility to EHEC infection? Indeed, there is
a report suggesting that microbiota trans-
plants from susceptible mice to Citrobacter TREATMENT
rodentium (a murine surrogate model for
EHEC infections) to resistant mice, and vice The main worry surrounding EHEC infec-
versa, changed host susceptibility to infection, tions is the development of HUS, which is
suggesting the very exciting possibility that caused by Shiga toxin (45). Because Shiga
manipulations of the GI microbiota composi- toxins are encoded by lysogenic phages and
tion could have a protective effect against their expression is coupled to the phage
EHEC infections (39). Moreover, recent re- entering its lytic cycle (46, 47), the use of
search suggested that virulence traits such as antibiotics to treat EHEC infections is con-
the T3SS are actually competition tools em- troversial, due to their potential induction
ployed by pathogens to compete for coloniza- of toxin expression. In fact, several studies
tion niches within the GI tract and that they demonstrated an increased risk for the de-
are only required in the presence of the com- velopment of HUS following antibiotic treat-
peting microbial flora (40). The Kamada study ment (48, 49). However, some antibiotics do
goes as far as suggesting an Achilles heel to not seem to increase Shiga toxin release or
GI infections, in which C. rodentium promotes increase the risk of HUS development (50).
overgrowth of Gammaproteobacteria during Nevertheless, treatment of EHEC infections
infection, which ultimately compete with with antibiotics is controversial, and the usual
sugar sources, eliminating the pathogen from recommendation is to perform supportive
the GI tract and functioning as a natural treatment by giving isotonic saline to patients
probiotics treatment. to avoid dehydration (51).
Besides sensing cues from the microbiota Translation of basic knowledge of EHEC
to colonize the human GI tract, EHEC also pathogenesis allowed the investigation of new
adapted to sense acyl-homoserine lactone strategies to treat EHEC infections. Many
(AHL) signals produced by the rumen mic- potential new therapeutics have been devel-
robiota to modulate gene expression toward oped, the majority of them designed to pre-
successful colonization of its main reservoir, vent toxin binding to its Gb3 receptor or to
ruminants (41). limit Gb3 receptor generation. There are also
In addition to reading chemical cues therapies that inhibit toxin trafficking, pro-
provided by the GI microbiota, EHEC also cessing, or activity (reviewed in chapter 22
senses chemical signals from the host, specif- of this volume). Other therapeutics prevent
ically the stress hormones epinephrine and Shiga toxin and LEE expression, or kill
norepinephrine. The sensing of these host EHEC without promoting Shiga toxin release

536 SPERANDIO
(reviewed in chapter 22). However, although Nonetheless, investigations of STEC in animal

many new therapies are being developed reservoirs, in the farm environment, and in
and are currently under different stages of foods all have great potential to improve hu-
testing (ranging from efficacy in vitro, to ani- man health. Several of the remaining big
mal models, to clinical trials), no unified new issues in this area of research include under-
therapeutic to treat EHEC infections is cur- standing the following:
rently used in clinics. Consequently, the full
the role of Shiga toxin in the bovine gas-
development of novel therapies to treat EHEC
trointestinal tract and in the environment
infections and either prevent or ameliorate
the carriage of EHEC by cattle and the
HUS is still one of the biggest challenges in
impact of the competing microbiota, im-
the field.
mune response, and molecular mecha-
nisms of microbial attachment to mucosa,
hide, water troughs, etc.
RESERVOIRS
the survival of EHEC in the farm envi-
ronment and farm management practices
Although many recent outbreaks have been
that impact environmental reservoirs
associated with produce, the main reservoir
(water, feed, birds, etc.)
for EHEC is ruminants. Much effort has been
seasonal prevalence (summer for E. coli
put forth to control infection of meat products
O157:H7, but spring and fall for the other
by EHEC within abattoirs. Also, many vac-
six EHEC)
cines have been developed to control EHEC
effective preharvest and processing inter-
colonization of cattle (reviewed in chapters 26
ventions
to 31). Complete eradication of EHEC from its
environmental reservoir is quite a challenge!
However, it is foreseeable that controlling its
presence in its environmental reservoir will EPIDEMIOLOGY
lead to fewer outbreaks.
STEC is part of the normal intestinal mi- EHEC outbreaks occur throughout the globe.
crobiological flora of ruminants. Its extent and The serotype most prevalent in the U.S. is
diversity are staggering, with more than 400 usually O157:H7; however, there is a steep
STEC serotypes identified as transiently col- increase in outbreaks caused by non-O157
onizing cattle. The 1994 ruling by the U.S. serotypes (reviewed in chapters 11, 16, and 23
Department of Agriculture-Food Safety and of this volume). Additionally, the rapid evo-
Inspection Service of E. coli O157:H7 as an lution of E. coli is responsible for the quick rise
adulterant in raw ground beef, and the sub- of more virulent or different types of STEC.
sequent 2011 designations of O26, O45, O103, The O157 spinach outbreak strain had two
O111, O121, and O145 as adulterants in certain copies of Shiga toxin, leading to a higher per-
fresh beef products, are more than challenging centage of HUS cases than usual in this out-
for the industry and will have economic con- break. The German O104:H4 STEC strains are
sequences. The adulterant status of these more closely related to enteraggregative E. coli
STEC strains does not seem fully scientifically (EAEC) than to EHEC and evolved from
based, since serotype does not equal pathogen EAEC acquiring a Stx prophage (chapters 32
and other more prevalent and more virulent and 33). Epidemiological surveys generally
bacteria, such as Salmonella, have not been include serotyping, detection of Shiga toxin
designated as adulterants in fresh product. To by various methods (enzyme-linked immuno-
meet this challenge, improved accurate, rapid sorbent assays, PCR, etc.), multiplex PCRs
methodologies to test fresh products (meat including defined sets of virulence factors, and
and vegetables) for EHEC will be required. restriction fragment length polymorphisms

(RFLPs). The speed and accessibility of next- 4. Jerse AE, Yu J, Tall BD, Kaper JB. 1990. A ge-
generation sequencing are revolutionizing netic locus of enteropathogenic Escherichia coli
necessary for the production of attaching and
and increasing the speed with which out- effacing lesions on tissue culture cells. Proc Natl
break strains are identified. A striking exam- Acad Sci USA 87:78397843.
ple is the German outbreak where, through 5. Sekiya K, Ohishi M, Ogino T, Tamano K,
international collaboration among sequenc- Sasakawa C, Abe A. 2001. Supermolecular
ing and epidemiology laboratories and real- structure of the enteropathogenic Escherichia coli
type III secretion system and its direct interaction
time public genomic data release, the outbreak
with the EspA-sheath-like structure. Proc Natl
strain was quickly identified. Although next- Acad Sci USA 98:1163811643.
generation sequencing-based epidemiology is 6. Knutton S, Rosenshine I, Pallen MJ, Nisan I,
powerful, several challenges exist to its world- Neves BC, Bain C, Wolff C, Dougan G, Frankel
wide implementation on EHEC epidemiolog- G. 1998. A novel EspA-associated surface organ-
elle of enteropathogenic Escherichia coli involved
ical surveys, especially in developing nations,
in protein translocation into epithelial cells.
where the cost of this technology is still a EMBO J 17:21662176.
significant hurdle. Moreover, a real bottleneck 7. Deng W, Puente JL, Gruenheid S, Li Y, Vallance
is the bioinformatics power needed to process BA, Vazquez A, Barba J, Ibarra JA, O'Donnell
these large genomic datasets. P, Metalnikov P, Ashman K, Lee S, Goode D,
Pawson T, Finlay BB. 2004. Dissecting virulence:
systematic and functional analyses of a pathoge-
nicity island. Proc Natl Acad Sci USA 101:3597
CONCLUDING REMARKS 3602.
8. Tobe T, Beatson SA, Taniguchi H, Abe H, Bailey
As a community, EHEC researchers, public CM, Fivian A, Younis R, Matthews S, Marches
O, Frankel G, Hayashi T, Pallen MJ. 2006. An
health officials, medical doctors, veterinarians, extensive repertoire of type III secretion effectors
and food safety experts are living in exciting in Escherichia coli O157 and the role of lambdoid
and revolutionary times. Our understanding phages in their dissemination. Proc Natl Acad Sci
of EHEC virulence and evolution has dra- USA 103:1494114946.
matically increased, leading to designs of new 9. Hazen TH, Sahl JW, Fraser CM, Donnenberg
MS, Scheutz F, Rasko DA. 2013. Refining the
potential therapies, vaccines, and detection
pathovar paradigm via phylogenomics of the
methods. Globalization and the speed of in- attaching and effacing Escherichia coli. Proc Natl
formation sharing through the press and so- Acad Sci USA 110:1281012815.
cial networks have led to faster identification 10. de Grado M, Abe A, Gauthier A, Steele-
and control of outbreaks. Mortimer O, DeVinney R, Finlay BB. 1999.
Identification of the intimin-binding domain of
Tir of enteropathogenic Escherichia coli. Cell
Microbiol 1:717.
REFERENCES
11. DeVinney R, Stein M, Reinscheid D, Abe A,
1. McDaniel TK, Jarvis KG, Donnenberg MS, Ruschkowski S, Finlay BB. 1999. Enterohemor-
Kaper JB. 1995. A genetic locus of enterocyte rhagic Escherichia coli O157:H7 produces Tir,
effacement conserved among diverse enterobac- which is translocated to the host cell membrane
terial pathogens. Proc Natl Acad Sci USA 92:1664 but is not tyrosine phosphorylated. Infect Immun
1668. 67:23892398.
2. Jarvis KG, Giron JA, Jerse AE, McDaniel TK, 12. Luo Y, Frey EA, Pfuetzner RA, Creagh AL,
Donnenberg MS, Kaper JB. 1995. Enteropatho- Knoechel DG, Haynes CA, Finlay BB, Strynadka
genic Escherichia coli contains a putative type III NC. 2000. Crystal structure of enteropathogenic
secretion system necessary for the export of Escherichia coli intimin-receptor complex. Nature
proteins involved in attaching and effacing lesion 405:10731077.
formation. Proc Natl Acad Sci USA 92:79968000. 13. Hartland EL, Batchelor M, Delahay RM, Hale
3. Kenny B, DeVinney R, Stein M, Reinscheid DJ, C, Matthews S, Dougan G, Knutton S,
Frey EA, Finlay BB. 1997. Enteropathogenic Connerton I, Frankel G. 1999. Binding of intimin
E. coli (EPEC) transfers its receptor for intimate from enteropathogenic Escherichia coli to Tir and
adherence into mammalian cells. Cell 91:511520. to host cells. Mol Microbiol 32:151158.

538 SPERANDIO
14. Garmendia J, Phillips AD, Carlier MF, Chong Y, 25. Berger CN, Crepin VF, Baruch K, Mousnier A,
Schuller S, Marches O, Dahan S, Oswald E, Rosenshine I, Frankel G. 2012. EspZ of entero-
Shaw RK, Knutton S, Frankel G. 2004. TccP is pathogenic and enterohemorrhagic Escherichia
an enterohaemorrhagic Escherichia coli O157:H7 coli regulates type III secretion system protein
type III effector protein that couples Tir to the translocation. MBio 3:e00317-12. doi:10.1128/mBio.
actin-cytoskeleton. Cell Microbiol 6:11671183. 00317-12.
15. Liu H, Magoun L, Luperchio S, Schauer DB, 26. Mills E, Baruch K, Aviv G, Nitzan M, Rosenshine
Leong JM. 1999. The Tir-binding region of I. 2013. Dynamics of the type III secretion system
enterohaemorrhagic Escherichia coli intimin is activity of enteropathogenic Escherichia coli. MBio
sufficient to trigger actin condensation after bac- 4:e00303-13. doi:10.1128/mBio.00303-13.
terial-induced host cell signalling. Mol Microbiol 27. Farfan MJ, Torres AG. 2012. Molecular mecha-
34:6781. nisms that mediate colonization of Shiga toxin-
16. Campellone KG, Robbins D, Leong JM. 2004. producing Escherichia coli strains. Infect Immun
EspFU is a translocated EHEC effector that 80:903913.
interacts with Tir and N-WASP and promotes 28. Obrig TG, Karpman D. 2012. Shiga toxin patho-
Nck-independent actin assembly. Dev Cell 7:217 genesis: kidney complications and renal failure.
228. Curr Top Microbiol Immunol 357:105136.
17. Alto NM, Scott JD. 2004. The role of A-Kinase 29. Obata F, Tohyama K, Bonev AD, Kolling GL,
anchoring proteins in cAMP-mediated signal Keepers TR, Gross LK, Nelson MT, Sato S,
transduction pathways. Cell Biochem Biophys Obrig TG. 2008. Shiga toxin 2 affects the central
40:201208. nervous system through receptor globotriaosyl-
18. Orchard RC, Kittisopikul M, Altschuler SJ, Wu ceramide localized to neurons. J Infect Dis 198:
LF, Suel GM, Alto NM. 2012. Identification of 13981406.
F-actin as the dynamic hub in a microbial- 30. Murray CS, Tannock GW, Simon MA, Harmsen
induced GTPase polarity circuit. Cell 148:803815. HJ, Welling GW, Custovic A, Woodcock A.
19. Wong AR, Clements A, Raymond B, Crepin VF, 2005. Fecal microbiota in sensitized wheezy and
Frankel G. 2012. The interplay between the non-sensitized non-wheezy children: a nested
Escherichia coli Rho guanine nucleotide exchange case-control study. Clin Exp Allergy 35:741745.
factor effectors and the mammalian RhoGEF 31. Tannock GW. 2005. New perceptions of the
inhibitor EspH. MBio 3:e00250-11. doi:10.1128/ gut microbiota: implications for future research.
mBio.00250-11. Gastroenterol Clin North Am 34:361382, vii.
20. Hemrajani C, Berger CN, Robinson KS, 32. Fukuda S, Toh H, Hase K, Oshima K, Nakanishi
Marches O, Mousnier A, Frankel G. 2010. NleH Y, Yoshimura K, Tobe T, Clarke JM, Topping
effectors interact with Bax inhibitor-1 to block DL, Suzuki T, Taylor TD, Itoh K, Kikuchi J,
apoptosis during enteropathogenic Escherichia Morita H, Hattori M, Ohno H. 2011. Bifido-
coli infection. Proc Natl Acad Sci USA 107:3129 bacteria can protect from enteropathogenic in-
3134. fection through production of acetate. Nature
21. Pearson JS, Riedmaier P, Marches O, Frankel 469:543547.
G, Hartland EL. 2011. A type III effector protease 33. Sperandio V, Torres AG, Jarvis B, Nataro JP,
NleC from enteropathogenic Escherichia coli Kaper JB. 2003. Bacteria-host communication:
targets NF-kappaB for degradation. Mol Microbiol the language of hormones. Proc Natl Acad Sci
80:219230. USA 100:89518956.
22. Hartland EL, Leong JM. 2013. Enteropathogenic 34. Pacheco AR, Curtis MM, Ritchie JM, Munera
and enterohemorrhagic E. coli: ecology, patho- D, Waldor MK, Moreira CG, Sperandio V. 2012.
genesis, and evolution. Front Cell Infect Microbiol Fucose sensing regulates bacterial intestinal col-
3:15. onization. Nature 492:113117.
23. Gao X, Wang X, Pham TH, Feuerbacher LA, 35. Njoroge JW, Nguyen Y, Curtis MM, Moreira
Lubos ML, Huang M, Olsen R, Mushegian A, CG, Sperandio V. 2012. Virulence meets metab-
Slawson C, Hardwidge PR. 2013. NleB, a bacte- olism: Cra and KdpE gene regulation in entero-
rial effector with glycosyltransferase activity, hemorrhagic Escherichia coli. MBio 3:e00280-12.
targets GAPDH function to inhibit NF-kappaB doi:10.1128/mBio.00280-12.
activation. Cell Host Microbe 13:8799. 36. Kendall MM, Gruber CC, Parker CT, Sperandio
24. Pham TH, Gao X, Tsai K, Olsen R, Wan F, V. 2012. Ethanolamine controls expression of
Hardwidge PR. 2012. Functional differences and genes encoding components involved in inter-
interactions between the Escherichia coli type III kingdom signaling and virulence in enterohemor-
secretion system effectors NleH1 and NleH2. rhagic Escherichia coli O157:H7. MBio 3:e00050-12.
Infect Immun 80:21332140. doi:10.1128/mBio.00050-12.

37. de Sablet T, Chassard C, Bernalier-Donadille A, signaling cascade in enterohemorrhagic E. coli

Vareille M, Gobert AP, Martin C. 2009. Human (EHEC). PLoS Pathog 5:e1000553. doi:10.1371/
microbiota-secreted factors inhibit shiga toxin journal.ppat.1000553.
synthesis by enterohemorrhagic Escherichia coli 45. Karmali MA, Petric M, Lim C, Fleming PC,
O157:H7. Infect Immun 77:783790. Arbus GS, Lior H. 1985. The association between
38. Gordon JI, Klaenhammer TR. 2011. A rendez- idiopathic hemolytic uremic syndrome and in-
vous with our microbes. Proc Natl Acad Sci USA fection by verotoxin-producing Escherichia coli. J
108(Suppl 1):45134515. Infect Dis 151:775782.
39. Willing BP, Vacharaksa A, Croxen M, 46. Neely MN, Friedman DI. 1998. Functional and
Thanachayanont T, Finlay BB. 2011. Altering genetic analysis of regulatory regions of coliphage
host resistance to infections through microbial H-19B: location of shiga-like toxin and lysis genes
transplantation. PLoS One 6:e26988. doi:10.1371/ suggest a role for phage functions in toxin release.
journal.pone.0026988 Mol Microbiol 28:12551267.
40. Kamada N, Kim YG, Sham HP, Vallance BA, 47. Wagner PL, Livny J, Neely MN, Acheson DW,
Puente JL, Martens EC, Nunez G. 2012. Regu- Friedman DI, Waldor MK. 2002. Bacteriophage
lated virulence controls the ability of a pathogen control of Shiga toxin 1 production and release by
to compete with the gut microbiota. Science Escherichia coli. Mol Microbiol 44:957970.
336:13251329. 48. Slutsker L, Ries AA, Maloney K, Wells JG,
41. Hughes DT, Terekhova DA, Liou L, Hovde CJ, Greene KD, Griffin PM. 1998. A nationwide
Sahl JW, Patankar AV, Gonzalez JE, Edrington case-control study of Escherichia coli O157:H7
TS, Rasko DA, Sperandio V. 2010. Chemical infection in the United States. J Infect Dis 177:
sensing in mammalian host-bacterial commensal 962966.
associations. Proc Natl Acad Sci USA 107:98319836. 49. Wong CS, Jelacic S, Habeeb RL, Watkins SL,
42. Clarke MB, Hughes DT, Zhu C, Boedeker EC, Tarr PI. 2000. The risk of the hemolytic-uremic
Sperandio V. 2006. The QseC sensor kinase: a syndrome after antibiotic treatment of Escherichia
bacterial adrenergic receptor. Proc Natl Acad Sci coli O157:H7 infections. N Engl J Med 342:1930
USA 103:1042010425. 1936.
43. Rasko DA, Moreira CG, Li DR, Reading NC, 50. Panos GZ, Betsi GI, Falagas ME. 2006. System-
Ritchie JM, Waldor MK, Williams N, Taussig R, atic review: are antibiotics detrimental or benefi-
Wei S, Roth M, Hughes DT, Huntley JF, Fina cial for the treatment of patients with Escherichia
MW, Falck JR, Sperandio V. 2008. Targeting coli O157:H7 infection? Aliment Pharmacol Ther
QseC signaling and virulence for antibiotic de- 24:731742.
velopment. Science 321:10781080. 51. Holtz LR, Neill MA, Tarr PI. 2009. Acute bloody
44. Hughes DT, Clarke MB, Yamamoto K, Rasko diarrhea: a medical emergency for patients of all
DA, Sperandio V. 2009. The QseC adrenergic ages. Gastroenterology 136:18871898.

Index
A1 and A2 fragments, 513 research on, 536

AAF fimbriae, 518 for STEC, 211230, 264, 278, 458463
aaiC gene, 30 Animal waste
Aat protein-coat secretion system, 507 organism survival in, 427
AB5 structure, 510 VTEC in, 463465
Abattoir Antelopes, STEC in, 212
postharvest interventions in, 437456 Antiaggregation protein, 518
preharvest interventions in, 421436 Antibiotics
Abl protein, 107 EHEC infections and, 403404
Acetic acid, for beef processing, 442, 449450 for HUS, controversy over, 303306
N-Acetyl-D-glucosamine, 113 for isolating organisms, 272
Acid tolerance, 186, 382 for meat processing, 441445, 467468
EHEC, 176177 for STEC infections, controversy over, 342343
STEC, 366 for typing, 274
Acid washes, for beef processing, 442445, 449451 Antibodies
Actin, activation of, 106107, 115 EHEC, 384
Acuity pyramid, 302 lipopolysaccharide, 479481
Acute respiratory distress syndrome, in HUS, 309 Anti-inflammatory action, of short-chain fatty acids,
Acyl-homoserine-lactone autoinducers, 182, 404
405406, 412413 Antisilencers, 180
Adenosine A2a receptor protein, 77 Anuria, in HUS, 303, 306
adfO gene, 22 AP transcription factors, 323
Adherence, to epithelium, 56 Apoptosis, 85, 116, 383, 385, 388, 390, 513
Adhesins, 131155 Aquaporin 4, 85
AdiA protein, in acid tolerance, 176177 arcA gene, 21
adk gene, 21 Arg protein, 107
Aerobic plate counts, 445446 Arginine-dependent system, for acid tolerance,
aggR gene and AggR protein, 30, 506, 517519 176177
Aggregative adherence fimbriae, 8, 506507 argW gene, 2526
AHLs (acyl-homoserine-lactone autoinducers), 182, aroE gene, 21
405406, 412413 Arp2/3 complex, 107, 115116
AI-2 and AI-3 quorum sensing molecules, 177178 Arrhythmogenic media, 272
AI autoinducers, 404406 aspC gene, 21
Alpacas, STEC in, 212 Asymptomatic carriers, 299
Ammonium hydroxide, for beef processing, 442 of EHEC, 205
Amyloid component B, for STEC, 344, 349350 ruminants as, 214
Amyloid protein, 80 of STEC, 221
Anemia, hemolytic, 67 Attaching and effacing (A/E) lesions, 56, 97130,
Animal models 132134, 179182, 254, 326327, 412413, 509
for antibiotic use, 343 Autoinducers, 182, 412413
for EHEC, 157174 Autotransporters, 142144, 517
for STEC, 220221
Animal reservoirs Baboons, as disease models, 166167
classification of, 2021 Bacterial adrenergic sensor kinases, 100
for E. coli O157:H7, 437441, 458463 Bacteriophage lysotyping, 274
541
Index.proof.3d 541 Manila Typesetting Company 04/06/15 16:23

542 INDEX
Bacteriophages, 2527, 109 Carvacrol, for beef processing, 450451

for beef processing, 442, 446, 451 Caspase-3 activation assay, 85, 90
STEC in, 249 Cat(s), STEC in, 212, 220, 429
Bacteroides thetaiotaomicron, 408, 412 Catabolite repressor/activator protein, 178179, 184,
Band 0.03, 386 408
Barriers, to colonization, 382 Catecholamines, 382
Basal ganglia lesions, in HUS, 308309 Cathelicidin, 382
Bax inhibitor-1, 116 Cattle, see also Beef
BBG Escherichia coli, 2627 as disease models, 220221
Bcl-2 proteins, 513 EHEC in, 56, 203205, 412413
Beef, see also Cattle E. coli O157:H7 in, 421456
ground, outbreaks in, 78, 56, 256, 301, 372, STEC in, 212214, 245, 264, 362367, 426427,
440441 460462
hides and carcasses of, see Hides and carcasses vaccines for, 423425, 487501
outbreaks due to, 203205 CDC42, 115
processing of CD40L, 386
postharvest interventions for, 437456 Cell binding, of Stxs, 4143
preharvest interventions for, 421436 C-EnterNet, 362363
STEC in, 438441, 466468 Centers for Disease Control and Prevention,
Behavioral factors, STEC susceptibility and, 367 247248, 362
bfpA gene, 56 Central nervous system, Stx effects on, 8191, 161,
Big five genes 390
for EHEC testing, 269 Ces proteins, 102
for STEC testing, 268 Cetylpyridinium chloride, for beef processing, 445,
Big seven genes, for EHEC testing, 269 447
Big seven group, for STEC testing, 268 CfaD protein, 518
Big Six adulterants, 509 CG Escherichia coli, 2627
Big six genes, for produce testing, 238 Chaperone-usher secretion pathway, 8, 507
Biological reservoir, 425 Cheese, STEC contamination of, 466
Biotyping, 274 Chemical dehairing, 446447
Birds, STEC in, 212, 218, 429 Chemoattractants, 77
Bison, STEC in, 212, 216 Chemokine(s), 7779, 383384, 388390
Bovine animals, see also Cattle Chemokine ligand 1, 79
STEC in, 212214 Chickens
Brain, Stx effects on, 8191, 390 as disease models, 166167, 220221
Brefeldin A, for STEC, 351 STEC in, 212, 218219
Butyrate, 179 Chitosan acetate, for beef processing, 451
anti-inflammatory role of, 404, 411 Chlorine compounds, for beef processing, 442, 445,
in EHEC colonization, 382 447, 450
Chlorofoam, for beef processing, 447
C-9, for STEC, 344, 347348 Chloroquine, for STEC, 345, 351
Caenorhabditis elegans, 428 Cholesterol, in lipid rafts, 4142
Cah protein, 144 Cif protein, 109, 116
Calcium imaging, 90 Citric acid, for beef processing, 442, 450
Calcium-binding antigen 43, 144 Citrobacter rodentium animal models, 157, 164166,
Canines 324326
as disease models, 166167, 220221 Classification, of Escherichia coli, 56, 1718, 2021
STEC in, 212, 220, 429 Clathrin, in cell binding, 42
Capillary defects, in central nervous system lesions, Cleaning, of cattle hides and carcasses, 441452
83, 85 clpX gene, 21
Carbohydrate metabolism, disruption of, 404, Coagulation cascade, 388
408411 Coliforms, in produce, 232233
Carbohydrate recognition, 178179, 187 Colitis, hemorrhagic, see Hemorrhagic colitis
Carcasses, see Hides and carcasses Collagen, 138
Cardiovascular disorders, in HUS, 309 Colonization
Carriage, see also Animal reservoirs EHEC, 382384, 412413
in asymptomatic humans, 205, 214, 299 STEC, 438

INDEX 543
Colony hybridization, 272 Dogs

Colony immunoblot test, for Stxs, 266267 as disease models, 166167, 220221
Coma, in HUS, 308309 STEC in, 212, 220, 429
Commission for European Normalization, Donkeys, STEC in, 216217
268269 Dr superfamily, 8, 507
Complement, activation of, 390391, 513 Drinking water, contamination of, in farms, 213
Complement receptor, 114 DsrA protein, 186
Composting, of animal waste, 464 Ducks, STEC in, 212, 218
Contamination Dysbiosis, EHEC infections and, 403417
with E. coli O157:H7, 422423, 425429
interventions for, 441453 eae gene, 6, 23, 30, 46, 56, 104, 132, 237, 238, 277278
of milk, 213 EAEC (enteroaggregative Escherichia coli), 17
prevalence of, 440441 genomics of, 6365
sources of, 438440 outbreaks, 505529
with STEC, 211230 pathogenesis of, 505529
vaccination preventing, 493 EC55989, 508
of water, 213, 218, 232, 464465 ECP (Escherichia coli common pilus) protein,
COPII coat protein, 184 140141, 178, 183
Costs, of STEC infections, 364365 Eculizumab
Cra protein, 178179, 184, 408 for HUS, 310
CRISPR sequences, 276 for STEC, 345, 352
Critical Control Point Program, 421, 452453 EDL933 isolate, of Escherichia coli, 57, 6162,
Crl protein, 140 99100, 102, 113, 183, 185
Cross-contamination, 440, 494 Eeyarestatin, for STEC, 345, 351
Cross-immunity, 384 efa gene and Efa protein, 29, 183
CsgA protein (curli), 66, 139140 efa1 gene and Efa1 protein, 100, 117118
CsvR protein, 518 Effector molecules, injection of, 184185
Cultural factors, STEC susceptibility and, 367 Eha proteins, 142143, 183, 240
Culture, 201, 271273 EHEC (enterohemorrhagic Escherichia coli), 17
Cultured substances, for beef processing, 443 adhesins of, 131155
Curli, 66, 139140, 183 animal models for, 157174
cyaA gene, 21 colonization by, 382384
Cysteine methyltransferase, 328 definitions of, 300
Cytokeratins, 116 with EAEC characteristics, 506
Cytoskeleton, modification of, 115116 emerging strains of, 269
Cytotoxicity assays, for Stxs, 265266 genomics of, 5571
host response to, 381402
Dairy products, contamination of, 465466 inflammatory response in, 321339
Daisy, for STEC, 344, 348 locus of enterocyte effacement of, 97130
Decontamination, of cattle, 422 microbiota interactions with, 403417
Deer, STEC in, 212, 215216, 429 nomenclature of, 56
Defensins, 382 outbreaks of, 19, 199209
Dehairing, chemical, 446447 pathogenesis of, 381402
Department of Agriculture, Microbiological Data phylogeny, 22
Program, 233234 research topics for, 533539
Diabetes mellitus, in HUS, 309 testing for, 237
Dialysis, for HUS, 306 treatment of, 341358
Diarrhea vaccines for, 477485
in HUS, 310311 virulence gene regulation by, 175195
pathogenesis of, 116117 EHEC factor for adherence 1, 117118
Diffusely adherent Escherichia coli, 17 ehxA gene and EhxA protein, 186, 238
Diploscapter, 428 EibG protein, 145
Direct screening, for, STEC, 271272 EivF protein, 182
Dispersin, 507, 518 Elands, STEC in, 212
DksA protein, 181 Electrophysiologic studies, 90
DNA analysis, for Stxs, 267270 ELF protein, 141
dnaG gene, 21 Elk, STEC in, 216

544 INDEX
Endothelial cells, 83, 85, 90 genomics of, 5571

Enrichment, for STEC testing, 235241, 270271 genotyping of, 430
Enteroaggregative Escherichia coli (EAEC), see history of, 48, 5557
EAEC in Japan, 199209
Enterohemolysin, 20, 186, 239240, 273 locus of enterocyte effacement, 104117
Enterohemorrhagic Escherichia coli (EHEC), see outbreaks of, 1820, 56, 231244, 437
EHEC pathogenic effects of, 6566, 384385
Enteroinvasive Escherichia coli, 17 in produce, 231244
Enteropathogenic Escherichia coli (EPEC), 17, 20, 56 reservoirs for, 57, 458463
Enterotoxigenic Escherichia coli (ETEC), 17, 234 risk factors for, 361380
Enterotoxins, 234, 521522 typing of, 5760
Environment vaccines for, 423425, 477485, 487501
E. coli O157:H7 survival in, 427429 Esp protein family, 99118, 134, 143, 181, 184185,
STEC in, 264 333, 408409, 413, 423424, 491
water contamination in, 213, 218, 232 ETEC (enterotoxigenic Escherichia coli), 17, 234
Enzyme immunoassays, 301302 Ethanolamine, 179, 186, 187
Enzyme-linked immunosorbent assays, for Stxs, EtrA protein, 182
266267 European Centre for Disease Prevention and
EPEC (enteropathogenic Escherichia coli), 17 Control, 246247, 254255, 363
nomenclature of, 56 European Food Safety Authority, 249, 255
pathologic effects of, 20 European Food Safety on Biological Hazards
Epidemiology, see also Outbreaks (BIOHAZ) panel, 1920
E. coli O104:H4, 8967 European Surveillance System, 1920
E. coli O157:H7, 67, 5571, 421422, 437 European Union, surveillance in, 245259
research on, 536537 European Union Reference Laboratories, 251253
STEC, 299300, 361380 European Union Summary Report on Trends and
in animals, 211230 Sources of Zoonoses, Zoonotic Agents and
detection methods for, 263295 Food-Borne Outbreaks, 249250
Epinephrine/norepinephrine signaling, 187, 406 Evolution
Epithelial cells, 99, 322 of E. coli O157:H7, 65
Equine animals, STEC in, 212, 216217 of STEC, 371
Erk protein, 107, 323 Exo2, for STEC, 345, 351
esc gene family, 101102 External quality assessment, 250253
Escherichia coli, as indicator, for produce quality,
232233 F9 fimbriae, 141
Escherichia coli common pilus protein, 140141, 178, F9 protein, 183
183 Fabrys disease, 41
Escherichia coli immunoglobulin-binding protein, facD gene, 21
145 Factor H, 391
Escherichia coli laminin-binding fimbriae, 141 Farms, STEC contamination in, 211230
Escherichia coli O26, 248249 Fatty acids, short-chain, 404, 410411, 515516
Escherichia coli O103:H25, 462 FCg receptor, 114
Escherichia coli O104:H4, 253, 255257, 373 FDA Bacteriological Analytical Manual, 239
history of, 4, 810 Fecal coliforms, in produce, 232233
pathogenesis of, 505529 Fecal pat prevalence, 422423
Escherichia coli O113:H21, 441 Feces
Escherichia coli O127:H6, 99, 104106 interventions for reducing contamination by,
Escherichia coli O157:H7 441453
animal models for, 157174 organism survival in, 463465
animal reservoirs for, 211230 STEC in, 211230, 421429, 438440
in cattle, 437456 Fenugreek sprouts, 205, 373, 505
cattle preharvest interventions for, 421436 Feral pigs, STEC in, 212, 217218, 429, 462463
clinical perspectives of, 299310; see also Ferrets
Hemolytic-uremic syndrome; as disease models, 166167
Hemorrhagic colitis STEC in, 212
epidemiology of, 67, 5571, 421422, 437 Fiber, dietary, microbiota interactions with, 411
evolution of, 301 Fibrin deposition, 78, 80

INDEX 545
Fibronectin, 138 Genomic islands, 2829

Filaments, in Type III secretion system, 102103 Genotyping, of Escherichia coli O157, 430
fimA gene, 142 Glial cells, 90
Fimbriae, 136145, 183, 518 Globotriaosylceramide, 38, 40, 75, 90, 158, 411
Fimbrial locus, for vaccine development, 425 Gluconeogenesis, 178179, 186
Fish, STEC in, 212, 219 Glucose metabolism, disruption of, 404, 408411
Flagella, 144145, 177178, 186 Glucose-repressed system, for acid tolerance,
Flagellin, 411, 413 176177
in inflammatory response, 323324 Glutamate decarboxylase acid resistance system,
vaccine based on, 425, 491 412
flg genes, 411 Glutamate-dependent system, for acid tolerance,
flh genes, 177178 176177
Fli proteins, 177178 Glutamic acid, as active site, 39, 40
flicC gene, 144145, 276 Glycosphingolipid (G3), 4143, 45
Flies, STEC in, 220, 428429 Glycosyltranferases, 330331
Fluid management, for HUS, 303 gnd gene, 276
Fluorescein actin stain, 6, 9899 Goats, STEC in, 212, 214215, 462
Fodrin, 521 Goblet cells, 520
Follicle-associated epithelium, 105, 136137, 139, Golgi matrix protein, 116
382 GrlRA regulators, 181
Food and Waterborne Diseases and Zoonoses, Ground beef, outbreaks of, 78, 56, 256, 301, 372,
246248, 250, 254255 440441
Food safety grpE gene, 21
cattle preharvest interventions for, 421436 Guanine nucleotide exchange factors, 115
fresh produce testing for, 231244 gyrB gene, 21
risk analysis in, 361380
surveillance for, 245259 H antigen, for typing, 275276
FoodNet, 248, 250251, 362, 364365 ha gene and Iha protein, 145, 183, 382, 508, 510,
Fresh produce, see Produce 515516
Fruit, see Produce Hamburger outbreaks, 78, 56, 256, 301, 372,
Fucose, 179, 409410 440441
fumC gene, 21 Hazard Analysis and Critical Control Points,
Fur protein, 515516 452453
Furin inhibitors, for STEC, 345, 351352 HCP protein, 141142, 324
FusK protein, 179 Health care costs, of STEC infections, 364365
Fyn protein, 107 Heat-labile enterotoxin, 234
Heat-stable enterotoxin, 234
Gad proteins, in acid tolerance, 176177 Hematologic complications, in HUS, 308
Galactose catabolism, 187 Hemodialysis, for HUS, 306
1,4-Galactosyltransferase, 90 Hemolysis, in HUS, 308
Gammaproteobacteria, 410 Hemolytic anemia, in HUS, 67
Gastrointestinal disorders, in HUS, 309 Hemolytic-uremic syndrome
Gb3 polymers, 344, 348, 511512 acuity pyramid of, 302
Gb3 receptor, 385386 animal models for, 157174
Geese, STEC in, 212, 218 clinical perspectives of, 299319
Gel electrophoresis, 267 definitions of, 300
Gene(s), 2123; see also specific genes diagnosis of, 300302
Stx A subunit, 40 history of, 67, 300301
Stxs, 5 outbreaks of, 2325, 2930, 45, 203
Gene array analysis, of gene response, 7980 pathogenesis of, 4546, 7595, 381402, 505529
Gene probe hybridization, 267268 prognosis for, 302, 306
Genome sequence-based high-resolution research on, 533534
genotyping, 5960 risk assessment for, 278282
Genome sequences, Q genes, 2728 surveillance for, 246249, 254, 256
Genomic(s) toxins of, 185186
EHEC, 5571 Hemorrhagic colitis
E. coli O157:H7, 5571 animal models for, 157174

546 INDEX
Hemorrhagic colitis (continued) Infectious Disease Surveillance Center (Japan),

history of, 78 200201
outbreaks of, 18 Infectious Diseases Control Law (Japan), 200
pathogenesis of, 508 Inflammatory bowel syndrome, 404
risk assessment for, 278282 Inflammatory response, in EHEC infections, 321339
Hemorrhagic Escherichia coli pilus, 141142 Inhibitory proteins, 113
Herd immunity, 493 Innate immunity, modulation of, 113118
Hfq protein, 181 Insects, STEC in, 220
Hides and carcasses Insulin receptor substrate p53 (IRSp53), 107108,
chemical dehairing of, 446447 134
cleaning of, 441452 Insulin receptor tyrosine kinase substrate (IRTKS),
contamination of, 422423, 428, 438453 107108, 134
interventions for reducing contamination in, Integrative elements, 328
441453, 467468 Integrins, 106, 134136
Histidine sensors, 406 Interferon(s), in inflammatory response, 325
Histopathology, 83, 85, 159160 Interferon-gamma-inducible protein-10 (IP-10),
History, 49 7980
of E. coli O104:H4, 410 Interleukin(s), 322327, 383384, 388389
of E. coli O157:H7, 5557 International Escherichia and Klebsiella Centre, 275
of hemolytic-uremic syndrome, 67 International Organization for Standardization,
H-NS protein, 133, 140, 176, 180 technical specification, 249, 252253
Holotoxin, 75, 513 Intimin, 6, 46, 104106, 133136, 237, 459, 509
Horizontal gene transfer, 65 types of, 2829
Horses, STEC in, 212, 216217, 429 vaccine based on, 491
Host(s) IrgA protein, 508
response to EHEC, 381402 Iron transport, vaccines blocking, 424
of STEC, 211230, 429 IRSp53 (insulin receptor substrate p53), 107108,
Host factors 134
in animal studies, 167168 IRTKS (insulin receptor tyrosine kinase substrate),
in STEC susceptibility, 366367 107108, 134
Houseflies, STEC in, 220, 428429 IS printing system, 201
Hp90 protein, 104, 134 Isolation, of STEC, 271273
HUS, see Hemolytic-uremic syndrome
HuSAP, for STEC, 344, 349350 Jack in the Box Escherichia coli O157:H7 outbreak,
Hybridoma cell lines, for Stxs, 266267 421422
Hydrogen peroxide, for dehairing, 446 Japan, EHEC outbreaks in, 199209
Hyperkalemia, in HUS, 306307 JNK, 323
Hypertension, in HUS, 307308
Hyperuricemia, in HUS, 307 Kdp proteins, 178179, 182, 184, 407408
Hypoalbuminemia, in HUS, 307 Keratinocyte-derived chemokine, 79
Hypobromous acid, for beef processing, Kidney failure, see also Hemolytic-uremic
443, 447 syndrome
Hypocalcemia, in HUS, 303, 306307 in HUS, 67, 309310
Hypochlorous acid, for beef processing, 443 pathogenesis of, 388390
Hypocomplementemia, 390391 Knife trimming, of contaminated spots, for beef
Hypokalemia, in HUS, 306307 processing, 448449
Hyponatremia, in HUS, 306307
Laboratories
icd genes, 21 reference, 235, 250253
IkB kinase, 323 for surveillance, 246253
Imatinib, for STEC, 345, 352 Lactic acid, for beef processing, 443445, 449450,
Immunity, innate, modulation of, 113118 451
Immunoblot test, 266267, 270, 272 Lactoferrin, for beef processing, 443
Immunodetection assays, 90 Lairage, contamination during, 439
Immunological assays, for Stxs, 266267 Laminin, 138
Immunomagnetic separation assays, 239, 272 Lauramide arginine ethyl ester preparation, for beef
Incubation time, 271 processing, 443

INDEX 547
LED209 small molecule, for STEC, 343346 research on, 533534

Ler protein, 133, 180182, 410 for typing, 274278
Lettuce Molecular risk assessment, 265
coliforms in, 233 Molecular syringe, 179180
STEC in, 235 Monitoring, for public health, 245259
Leucine-responsive protein, 410 Monkeys, as disease models, 167
Leukocytosis, 388389 Monoclonal antibodies, for STEC, 345, 350351
lifA gene, 29, 100, 117118 Monocyte(s), in inflammatory response, 385387
Lineage-specific polymorphism typing assay, Monocyte chemotactic protein 1, 7778, 324
5960 Monosialotetrahexosylganglioside, 41
Lipid rafts, 41, 512 Most probable number method, 233
Lipocalin-2, 80 Motility, 177178
Lipopolysaccharide, 7781, 383387, 478482 mtlD gene, 21
Livestock, see Cattle; Goats; Sheep Muc1 protein, 144
Llamas, STEC in, 212 Mucins, 333, 408409, 520
Locus of enterocyte effacement, 6, 20, 2223, 2829, Multilocus sequence typing, 2122, 234, 274275,
4546, 97130, 132145, 179184, 265, 366, 509 277
Long-polar fimbriae, 136139, 516517 Multilocus variable number tandem repeat analysis,
Loop-mediated isothermal amplification, 267 5859, 277
lpf genes and Lpf proteins, 136139, 183, 425, Multiple hurdle treatment, for beef processing,
516517 451452
lrp gene and Lrp protein, 382, 410 Multiple-locus variable-number tandem-repeat
Lux proteins, 405406 analysis, 201
lysP gene, 21 Multiplex PCR assays, 268
for ETEC, 234
Macrophage inflammatory proteins, 45, 7778, 324, for produce, 238
386 for STEC, 267
Macropinocytosis, 42, 382 Murine interleukin-8 mimic, 79
Maintenance host population, 429 Murine species, see Mice
Manganese, for STEC, 345, 351 mutS gene, 21
Mannose-binding lectin pathway, 391 MyD88, 326, 383384
Map protein, 109113, 115117, 184
Mat protein, 140141 Na(+)/H(+) exchanger regulatory factor 2
mdh gene, 21 (NHERF2), 116117
Meat, from cattle, see Beef National Epidemiological Surveillance of Infectious
Meat Inspection Act of 1906, 421 Diseases (Japan), 200
Mesangiolysis, 514 National Health Surveillance System (Argentina),
Metalloproteinases, 329330 363
Methyltransferase, 113 National reference laboratories, 250251
Mice Nationally Notifiable Diseases Surveillance System,
as disease models, 164166, 220221, 324326 362
STEC in, 212, 219 Nck protein, 107
Microarray analysis, 7980, 276 NEDD 8 protein, 109
Microbiological Data Program, 233235 Nematodes
Microbiology Laboratory Guidebook, 441442 as disease models, 166167
Microbiota, EHEC interactions with, 403417, E. coli O157:H7 in, 428
534535 Neural Wiskott-Aldrich syndrome protein, 107109,
Milk, STEC contamination of, 213, 218, 465466 115
mirA gene, 25 Neurologic disorders, in HUS, 308309
Mitochondrial-associated protein (Map protein), Neutrophils
109113, 115117, 184 in inflammatory response, 322, 324, 333, 384385,
Mitogen-activated protein kinases, 322323, 383, 387
513 migration of, 7879
Mitomycin C, for enrichment, 270 NHERF2 [(Na)/H(+) exchanger regulatory factor
MlrA protein, 140 2], 116117
MMA-tet, for STEC, 344, 349 Nisin, for beef processing, 443444
Molecular methods Nitrobenzyl-thioinosine, for STEC, 345, 351

548 INDEX
nle genes and Nle proteins, 29, 112, 116, 184185, of hemolytic-uremic syndrome, 4546, 7595,
277279, 282, 328332 381382, 505529
NM strain, typing of, 276 of hemorrhagic colitis, 508
NOD proteins, 333 Pathogenicity islands, 365366, 509
Nomenclature, of EHEC, 56 Pathotypes, 1723, 2931, 264, 274278
Nonselective culture media, 272 pch genes and Pch proteins, 181, 182, 410
Novobiocin, for enrichment, 270 PCR testing
NRIR sequences, 185 for ground beef, 441
Nuclear factor-kappa B, 113, 322323, 326328, for produce, 236241
331333, 383 for Stxs, 266270, 276, 278
Nucleolin, 106, 135136 Penn State University, Escherichia coli reference
Nucleotide polymorphism-derived genotyping, laboratory in, 235
2122 PerC protein, 182
Nucleotide sequencing, 277278 Periharvest interventions, for Escherichia coli
Nutrient deprivation, 181 O157:H7, 437456
Nutrient signaling, 407411 Peripheral nervous system, Stx effects on, 90
N-WASP (neural Wiskott-Aldrich syndrome Peritoneal dialysis, for HUS, 306
protein), 107109, 115 Peroxyacetic acid preparations, for beef processing,
444, 450451
O antigen, 275276, 478 Pet protein, 521
Octamer-based genome scanning typing assay, 59 Pets, STEC in, 220, 429
OI-122 pathognicity island, 29 Petting zoos, STEC in, 214
O-island regulators, 182 Phage lysotyping, 274, 277
Oligoanuria, in HUS, 303, 306 Phagocytosis, inhibition of, 114
Open reading frames, 282 phe genes, 100
orf gene family, 101 Phenotyping, 275276
Organic acids, for beef processing, 449451 Phosphoric acid, for beef processing, 447, 450
Organic produce, vs. conventional produce, 233 Phylogenetic analysis-derived genotyping, 2122
OSP vaccine, 478482 Pic protein, 507508, 517, 519521
OspZ protein, 328 Pigeons, STEC in, 212, 218
Oubain, for STEC, 345, 352 Pigs
Outbreaks, 1820 as disease models, 160161
EAEC, 459, 505529 STEC in, 212, 217, 429, 462463
EHEC, 158160 Pilin, 507
E. coli O157:H7, 1820, 5657, 231244, 437 Plasmapheresis, for HUS, 310
hemolytic-uremic syndrome, 2325, 2930, 45 Plasmid profiles, for typing, 274
hemorrhagic colitis, 18 Platelets, in inflammatory response, 386390
produce-related, 231232 Polyclonal antibodies, for STEC, 344345, 350351
sequencing studies of, 6265 Porcine edema disease, 217, 278, 462463
STEC, 361380 Postharvest interventions, 437456, 467
surveillance and, 245259 Poultry, STEC in, 212, 218219, 429
Outer membrane protein A, 145, 183 PpdD protein, 141142
Ozone, for beef processing, 444445, 450 ppGpp protein, 181
Predicted protein D, 141142
p38, 323 Preharvest interventions, 467
pagC gene, 29 for E. coli O157:H7, 421436
Pancreatitis, in HUS, 309 vaccination, 423425, 487501
Paralysis, 81, 83, 85 prfC gene, 25
Pasteurization, steam, for beef processing, Primates, as disease models, 220221
448449 Probiotics, 344, 349, 411
patC gene, 25 Produce, see also specific items, e.g., Spinach
Pathogen Reduction and Hazard Analysis, 421 EHEC in, 203205
Pathogen-associated molecular patterns, 383384 ETEC in, 234
Pathogenesis organism attachment to, 6566
of E. coli O104:H4, 505529 Shiga-toxin producing organisms in, 231244
of E. coli O157:H7, 6566, 384385 STEC in, 231244
of EHEC, 381402 Prophages, 109, 185186, 514

INDEX 549
Protozoa, Escherichia coli O157:H7 in, 428 Rotifers, Escherichia coli O157:H7 in, 428
Public health Rpo proteins, 140, 176
cattle vaccination and, 487501 Ruminants, see also Cattle; Goats; Sheep
genome sequencing and, 66 EHEC in, 412413
Japanese outbreak, 199209 STEC in, 212, 214215, 249250
microbiologic surveillance for, 245259
taxonomy and, 2931 Saa protein, 143144, 237238, 240
Pulsed-field gel electrophoresis, 21, 58, 201, Sab protein, 143144
276277, 364, 439 Safety, food, see Food safety
PulseNet, 201, 248, 364 Sakai isolate, of Escherichia coli, 57, 6162, 185
purD gene, 21 Sakai, Japan, EHEC outbreaks in, 199209, 372
Pure Food and Drug Act, 421 Salads, ready-to-eat, 231244
Pyocin, for STEC, 343, 344 Salicylidene acylhydrazides, 102
Sanitizers, for beef processing, 441445
Q antiterminator protein, 2728 scbB gene, 25
Qse proteins, 133, 167168, 177178, 182, 184, 186, SdiA protein, 177, 182, 405406, 413
187, 406407 Seasonal variation, in cattle contamination,
Quorum sensing, 182, 382, 404411 422423, 424, 426430, 439
Seizures, 83
Rab protein, 116 sel genes, 100
Rabbits Selective culture media, 272273
as disease models, 161164, 220221 sen gene, 29
STEC in, 212, 219220 sep gene family and Sep proteins, 101102, 109, 181,
Rac proteins, 115, 333 507508, 519521
Raccoons, STEC in, 212, 220 Septrin, 521
RANTES, 77, 385, 389 Serine protease autotransporters of
Rats Enterobacteriaceae (SPATEs), 8, 507508,
as disease models, 167, 220221 519521
STEC in, 212, 219 Serologic tests, for Stxs, 266267
Ready-to-eat produce, see Produce Serotypes, 1718, 2931
recA gene and RecA protein, 21, 185186 Serotyping, 274
Recto-anal junction, 213215, 412 serU gene, 25
Reference laboratories, 235, 250253 Serum bactericidal assay, 480
Regulations, for STEC monitoring, 255257 Sfp (sorbitol-fermenting fimbriae protein), 141
Rel proteins, 181, 322323 Shedding, of bacteria, 213214, 424, 440, 488489
Research, topics for, 533539 Sheep, STEC in, 212, 214215, 429, 438, 462
Reservoirs, see also Animal reservoirs Shellac coating, for beef processing, 446
for E. coli O157:H7, 57, 422429 Shellfish, STEC in, 212, 219
research on, 536 ShET1 toxin, 508, 521522
for STEC, 264 SHIBAM medium, 240
Restriction fragment length polymorphism, 267, 277 Shiga toxin-producing Escherichia coli (STEC), 17;
Retinal hemorrhage, in HUS, 309 see also Stx(s)
Retro-2, for STEC, 345, 351 SHP-1 protein, 332
Retrograde trafficking, 43, 512513 Siderophore receptor and porin cattle vaccine,
rfbE gene, 276 423425, 491492
Rho proteins, 115 SigA protein, 507508, 519521
Ribotoxic stress response, 44, 513 Single nucleotide polymorphism typing, 60, 276
Risk assessment, 254255 Slaughter, of cattle
for cattle contamination, 425 postharvest interventions for, 437456
for STEC, 235241, 265, 278282 preharvest interventions of, 421436
Risk factors, for STEC, 361380 Slide agglutination tests, 275
RNaseE, 181 Small molecules, for STEC, 343, 344346
Rns protein, 518 SNX9 protein, 115
Rodents Sodium decyl sulfate, for beef processing, 451
as disease models, 220221 Sodium hydroxide, for beef processing, 447
STEC in, 212, 219 Sodium sulfide preparation, for dehairing, 446
ror gene family, 101102 Sodium-D-glucose transporter, 117

550 INDEX
Solid agar plates, 272 cell binding of, 4143

SopE protein, 115, 333 detection of, 265271
Sorbitol-fermenting Escherichia coli, 237, 301 discovery of, 2, 37
Sorbitol-fermenting fimbriae protein, 141 diseases associated with, 38, 4546
Sorbitol-MacConkey agar, 301302 food containing, 234235
SPATEs (serine protease autotransporters of genetic analysis of, 4041
Enterobacteriaceae), 8, 507508, 519521 history of, 49
Spinach overview of, 3738
coliforms in, 233 pathogenic effects of, 7595, 381402
outbreaks related to, 5657, 6566, 235, 236, 366, pathophysiology of, 508
372 in prophages, 185186
STEC in, 235 retrograde trafficking of, 43
Spinal cord, Stx effects on, 8191 structure of, 3941
SpoT protein, 181 subtypes of, 265
Sprouts, outbreaks from, 89, 256, 372373, 505 typing of, 3839
SRP proteins, 423425, 491492 vaccines based on, 481
ssrA gene, 25 Stx1
Stacked-brick adherence pattern, 506507, 517 bacteriophages associated with, 2527
STARFISH, for STEC, 344, 348 discovery of, 37
Starlings, STEC in, 212, 218, 429 nomenclature of, 3738
stcE gene and StcE protein, 22, 183, 184, 385 overview of, 3738
Steam treatment, for beef processing, 448449 pathologic effects of, 18
STEC (Shiga toxin-producing Escherichia coli), see properties of, 2325
also specific organisms, e.g., Escherichia coli variants of, 2425
O157:H7 Stx1a
animal reservoirs for, 211230 action of, 4445
classification of, 19 genetic analysis of, 4041
clinical perspectives of, 299319; see also nomenclature of, 3738
Hemolytic-uremic syndrome; Hemorrhagic structure of, 3941
colitis versus Stx2a, 4445
definitions of, 300 Stx2, 24
detection of, 263295, 441442 nomenclature of, 3738
disease caused by, 2325 pathogenic effects of, 18, 510515
genomic islands of, 2829 properties of, 2325
isolation of, 271273 variants of, 2425
nonhuman sources of, 263295 Stx2a, 2324
outbreaks, 19, 361380, 465 action of, 4445
pathologic effects of, 18 disease associated with, 45
phylogeny, 2122 nomenclature of, 3738
in produce, 231244 versus Stx1a, 4445
risk analysis for, 235241, 278282 subtypes of, 39
risk factors for, 361380 stx2b gene, 25
serotypes of, 2931, 362 stx2c gene, 24
subtypes of, 264265 Stx2d, 2324, 39
surveillance for, 245259 Stx2e, 25
testing for, 235241 Stxb, 24
treatment of, 341358 Stxc, 24
types of, 264, 274278 subAB gene, 236238, 240
Stress response, ribotoxic, 44 Subendothelial matrix proteins, 387
Stress, STEC susceptibility and, 366367 Subtilase cytotoxin, 237238
Stringent response, 181 Subtypes
Stroke, in HUS, 308 STEC, 274278
Stromal cell-derived pathway, 389390 toxins, 2325
Structures, of Stxs, 3841 Sugar, for beef processing, 443
Stx(s), 3753; see also specific subtypes SUPER TWIG, for STEC, 344, 348
actions of, 4344 Super-shedding animals, 213214, 367, 424, 460, 489
animal models for, 157174 Surveillance

INDEX 551
EHEC infections, 200202 TVP, for STEC, 344, 349

for public health, 245259 Type I fimbriae, 142
STEC infections, 362364 Type III secretion system, 2829, 101102, 109113,
Swine, see Pigs 132, 178184, 490491
SYNSORB, for STEC, 344, 348 cattle vaccines based in, 423425
proinflammatory activity of, 332333
TAB proteins, 113, 328 Type V secretion system, 507
TaqMan minor groove primer, 268 Type VI secretion system, 518
Taxonomy, 1721, 2931 Typing, of Escherichia coli O157:H7, 5760
Tccp protein, 107108, 184
Tellurite media, 272273 Ubiquitin, 109
Temperature-sensitive hemagglutinin, 520 UhpB protein, 409
ter genes, 273 uidA gene, 21
Terminal deoxynucleotidyltransferase-mediated Urtoxazumab, for STEC, 345, 350
dUTP-biotin nick and labeling (TUNEL), 85, 90
Terminology, 1718 Vaccines
Tetrahymena, 428 cattle, 423425, 487501
Thrombocytopenia, 67, 386 for humans, 477485
Thrombosis intimin, 136
pathogenesis of, 387388 Vacuuming, with steam, for beef processing,
renal, 7779 448449
Tir protein, 2829, 133136, 184, 332, 509 Vascular cell adhesion molecule-1, 79
in intimin binding, 105109, 113 Vegetables, see Produce
vaccine based on, 491 Verocytotoxigenic Escherichia coli reverse passive
TIRAP protein, 384 latex agglutination assay, for Stxs, 266267
Tissue factor, 387, 390 Verocytotoxin-producing Escherichia coli, see
Tissue tropism, 135136, 139 VTEC
Toll-like receptors, 323, 326, 383384, 386387 Verotoxins, see also Stx(s)
torS/T gene, 25 terminology of, 4
Total microbial populations, in produce, 232 Veterinary reference laboratories, 251253
ToxB protein, 117118 Virulence factors, 366
Toxins, 2325; see also specific toxins of EHEC, 97130, 381402
TRAF2 (tumor necrosis factor receptor-associated microbiota interactions and, 407411
factor 2), 113, 328 for typing, 56
TRAM protein, 384 Virulence gene regulation, 175195
Transfusions, for HUS, 308 Volume expansion, for HUS, 303
Translocons, 101104 Volunteer studies, of EHEC, 159
Transmission von Willebrand factor, 387
of EHEC, 204205 VTEC (verocytotoxin-producing Escherichia coli),
of STEC, 264 17; see also EAEC; STEC
Treatment outbreaks of, 20
of EHEC infections, 341358 pathologic effects of, 18
of HUS, 299319 public health approach to, 457476
research on, 533534 reservoirs for, 457476
of STEC infections, 341358 vtx1 gene, 25
TRIF protein, 384 vtx2 gene, 25
Trisodium phosphate, for beef processing, 447 vtx2a gene, 2425
Tropism, extension of, 105 vtx2c gene, 27
Tubulin, 116
Tumor necrosis factor, 324325, 330, 390 Washed blood agar medium, 239240
Tumor necrosis factor alpha, 85, 90, 383 Washing methods, for cattle hide, 441, 445446.448
Tumor necrosis factor receptor-associated factor 2 Water
(TRAF2), 113 contamination of, 232, 464465
TUNEL (terminal deoxynucleotidyltransferase- in environment, 213, 218
mediated dUTP-biotin nick and labeling), 85, as reservoir, 427428
90 Water buffalo, STEC in, 212, 215
Turkeys, STEC in, 212 Water fowl, STEC in, 212, 218

552 INDEX
Whole-genome mapping, 59 Ycb proteins, 183

Whole-genome sequence typing, 60 yciD gene, 25
Wildlife, STEC in, 429 Yeast-2-hybrid analysis, 101
Wiskott-Aldrich syndrome protein, 107109, 115 yecE gene, 25
wrbA gene, 2526 yehV gene, 2525
WxxxE proteins, 115, 333 yjbM gene, 25
wyz gene, 276 ynfH gene, 25
wzx gene, 276
Z4322 gene, 29, 117118
Yaks, STEC in, 212 Zinc, for STEC, 344, 346

About the Editors
Vanessa Sperandio is a Professor in the

Departments of Microbiology and Biochem-
istry at UT Southwestern. Dr. Sperandios
research investigates chemical, stress, and
nutritional signaling at the interface among
the mammalian host, beneficial microbi-
ota, and invading pathogens. Her labora-
tory research focuses on how bacterial cells
sense several mammalian hormones leading
to rewiring and reprogramming of bacterial
transcription towards host and niche adap-
tation. Dr. Sperandio is chair-elect of Divi-
sion D of the ASM and was elected a fellow
of the American Academy of Microbiology in 2013. She is the recipient of
the 2015 ASM Eli-Lilly and Company-Elanco research award and a winner of
the GSK Discovery Fast Track challenge program 2014. She was a Pew Latin
American Fellow in Biological Sciences, an Ellison Medical Foundation New
Scholar in Global Infectious Diseases, a Burroughs Wellcome Fund Inves-
tigator in Pathogenesis of Infectious Diseases, and a Kavli Frontiers of Sci-
ence Fellow from the National Academy of Science since 2007. She currently
serves on the editorial boards of mBio, Infection and Immunity, Journal of
Bacteriology, and Gut Pathogens.
Carolyn J. Hovde is a University Distinguished

Professor who has served as the Idaho NIH INBRE
Director since 2006. Dr. Hovdes laboratory stud-
ies E. coli O157:H7 with a primary focus on under-
standing the relationship between this human
pathogen and its silent reservoir, healthy cattle.
Dr. Hovde has authored more than 100 scientific
publications and holds one patent. She has been the
recipient of numerous honors and awards includ-
ing election as a Fellow to the American Associa-
tion for the Advancement of Science and winner of
the ASM Carski Foundation Distinguished Under-
graduate Teaching Award. She currently serves as
the President of the National Association of IDeA Principal Investigators.
553

E Coli 2015

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

E Coli 2015

Uploaded by

Copyright:

Available Formats

Enterohemorrhagic

Ecoli TItle pages.indd 1 3/26/15 3:52 PM

Ecoli TItle pages.indd 2 3/26/15 3:52 PM

Library of Congress Cataloging-in-Publication Data

Printed in the United States of America

INCIDENCE, EPIDEMIOLOGY, AND ECOLOGY

11 Animal Reservoirs of Shiga Toxin-Producing Escherichia coli211

DIAGNOSIS, DETECTION, AND STRAIN CHARACTERIZATION

CLINICAL, PATHOLOGICAL, AND PATHOPHYSIOLOGICAL ASPECTS

HOST DETERMINANTS OF DISEASE AND HOST RESPONSE

PREVENTION AND CONTROL STRATEGIES

22 Peri- and Postharvest Factors in the Control of Shiga Toxin-Producing

ESCHERICHIA COLI O104:H4

THE WAY FORWARD

This book is an exceptional compilation of our current worldwide understand-

Part01.proof.3d 1 Manila Typesetting Company 03/31/15 12:06

JAMES B. KAPER1 and ALISON D. OBRIEN2

Chap01.proof.3d 3 Manila Typesetting Company 03/31/15 10:44

Our understanding of what constitutes a acquisition, through horizontal genetic ex-

Chap01.proof.3d 4 Manila Typesetting Company 03/31/15 10:44

Chap01.proof.3d 5 Manila Typesetting Company 03/31/15 10:44

Chap01.proof.3d 6 Manila Typesetting Company 03/31/15 10:44

Chap01.proof.3d 7 Manila Typesetting Company 03/31/15 10:44

Chap01.proof.3d 8 Manila Typesetting Company 03/31/15 10:44

Chap01.proof.3d 9 Manila Typesetting Company 03/31/15 10:44

Chap01.proof.3d 10 Manila Typesetting Company 03/31/15 10:44

REFERENCES 11. Dubos RJ, Geiger JW. 1946. Preparation and

Chap01.proof.3d 11 Manila Typesetting Company 03/31/15 10:44

Chap01.proof.3d 12 Manila Typesetting Company 03/31/15 10:44

Chap01.proof.3d 13 Manila Typesetting Company 03/31/15 10:44

Part02.proof.3d 15 Manila Typesetting Company 03/31/15 12:07

The term enteropathogenic Escherichia coli was originally used to refer to

Chap02.proof.3d 17 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 18 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 19 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 20 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 21 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 22 Manila Typesetting Company 03/31/15 10:28

lambdoid stx-encoding phages showed that

TOXIN AND DISEASE

The Shiga toxin family is divided into two

Chap02.proof.3d 23 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 24 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 25 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 26 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 27 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 28 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 29 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 30 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 31 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 32 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 33 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 34 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 35 Manila Typesetting Company 03/31/15 10:28

Chap02.proof.3d 36 Manila Typesetting Company 03/31/15 10:28

Chap03.proof.3d 37 Manila Typesetting Company 03/30/15 16:57

same mode of action as Stx/Stx1 but is im- TYPING AND NOMENCLATURE

TABLE 1 Prototype toxins and strains that produce those toxins

Stx 3818T Yes 133

Chap03.proof.3d 38 Manila Typesetting Company 03/30/15 16:57

Chap03.proof.3d 39 Manila Typesetting Company 03/30/15 16:57

Chap03.proof.3d 40 Manila Typesetting Company 03/30/15 16:57

Chap03.proof.3d 41 Manila Typesetting Company 03/30/15 16:57