1539

Journal of Food Protection, Vol. 76, No. 9, 2013, Pages 15391548

doi:10.4315/0362-028X.JFP-12-254
Copyright G, International Association for Food Protection

Impact of Fatty Acid Chain Length of Rosmarinate Esters on Their

Antimicrobial Activity against Staphylococcus carnosus LTH1502
and Escherichia coli K-12 LTH4263
SARISA SURIYARAK,1 CHRISTELLE BAYRASY,2 HERBERT SCHMIDT,3 PIERRE VILLENEUVE,2 AND JOCHEN WEISS1*
1Department

of Food Physics and Meat Science, Garbenstrasse 21/25 and 3Department of Food Microbiology, Institute of Food Science and Biotechnology,
University of Hohenheim, Garbenstrasse 28, 70599 Stuttgart, Germany; and 2Centre de Cooperation Internationale en Recherche Agronomique pour le
Developpement (CIRAD), Unite Mixte de Recherche (UMR), Ingenierie des Agropolyme`res et Technologies Emergentes (IATE), Montpellier, 34060 France
MS 12-254: Received 8 June 2012/Accepted 26 January 2013

ABSTRACT
The effect of the addition of a newly synthesized series of rosmarinic acid (RA) esters (REs) and alcohols with chain lengths
of 1, 4, 8, 10, 12, 16, and 18 carbons (RE1 to 18) on the growth behavior of Staphylococcus carnosus LTH1502 and Escherichia
coli K-12 LTH4263 was investigated. An initial microtiter dilution assay indicated activity of compounds against S. carnosus
LTH1502 only, but not E. coli K-12 LTH4263, and a strong dependence of growth inhibition of S. carnosus LTH1502 on the
esters alkyl acid chain lengths. Esters with chains shorter than 8 and longer than 12 had little effect against S. carnosus
LTH1502, whereas esters with chain lengths between 8 and 12 inhibited growth. To better understand the dependence of the
antimicrobial activity of the REs on alkyl chain lengths, RA, n-methyl rosmarinate (RE1), n-dodecyl rosmarinate (RE12), and noctadecyl rosmarinate (RE18) were used in a time-kill assay against S. carnosus LTH1502. Compounds were added at 0.75 mM
in the lag phase, 5 mM in the exponential phase, and 10 mM in the stationary phase. RA had no effect in the lag and exponential
phases but decreased cell counts during the stationary phase. In contrast, RE1 and RE12 decreased cell numbers in all three
phases, with RE12 reducing counts most rapidly. Addition of RE18 did not affect growth regardless of the growth phase.
Appearance and physiological state of S. carnosus LTH1502 cells treated with 200 mM RA after 48 h and with 10 mM RE12
after 4 h were observed by fluorescence microscopy. Images indicated differences in the way the compounds interacted with and
damaged cells. Results were attributed to the different physicochemical properties of RA and its esters.

Rosmarinic acid (RA) is a naturally phenolic compound

(32) found in many plants that belong to the Lamiaceae
family (31), including rosemary (Rosmarinus officinalis),
sage (Salvia officinalis) (11, 26), lemon balm (Melissa
officinalis), and oregano (Origanum vulgare). RA is a
highly functional compound able to act as an antioxidant (6,
29), a chelator (with Zr4z) (30), an antimicrobial (7, 34),
and an antiinflammatory agent (1, 9). The biological
functionalities of RA have been directly attributed to its
phenolic structure: [(R)-a-[[3-(3,4-dihydroxyphenyl)-1-oxo2E-propenyl]oxy]-3,4-dihydroxy-benzenepropanoic acid], with
RA being a derivative of caffeic acid and 3,4-dihydroxyphenyl
lactic acid. RA exhibits antimicrobial activity only at lower pH
values. There, the compound is fully protonated and able to
penetrate bacterial cells. Moreover, RA is predominately water
soluble and, therefore, difficult to use in lipid-based systems,
such as oil-in-water emulsions. RA is highly volatile, and
activity is lost if RA, or the matrix it is dispersed in, are
subjected to heating.

* Author for correspondence. Tel: z49 711 459 24415; Fax: z49 711 459
24446; E-mail: j.weiss@uni-hohenheim.de.

Recently, a novel class of phenolipids, esters of

phenolic acids and alcohols, have been developed at Centre
de Cooperation Internationale en Recherche Agronomique
pour le Developpement (CIRAD), Montpellier, France (23).
These phenolipids are surface active, i.e., they preferentially
adsorb at interfaces. For a given phenolic group esterified
with different alcohols, the surface activities of the
synthesized esters are governed by the chemical nature of
the hydrophobic group (i.e., its degree of saturation and
chain length) and the type of phenolic acid used (e.g.,
chlorogenic acid or RA). Phenolipids exhibit strong
antioxidant activities in dispersed systems, such as oil-inwater emulsions, in contrast to nonesterified phenolic acids.
For example, esters of RA with alkyl chains lengths varying
between 1 and 18 had high antioxidant activities at medium
alkyl chains lengths, i.e., at chain lengths varying between 8
and 12. This was attributed to the compounds becoming
surface active upon esterification with C8 to C12 alcohols,
increasing their preference for association with oil-in-water
interfaces (22).
Other researchers have reported that antimicrobial
activities of other compounds, notably arginine, are altered
by esterification. Lauric arginate, a conjugate between lauric
acid and arginine, exhibits strong antimicrobial activities

1540

SURIYARAK ET AL.

J. Food Prot., Vol. 76, No. 9

FIGURE 1. Chemical reaction of rosmarinic acid (RA) and alcohols to generate

rosmarinic acid esters (REs). Molecular
structures and molecular weight of RA and
its esters (RE1, methyl; RE4, butyl; RE8,
octyl; RE10, decyl; RE12, dodecyl; RE16,
hexadecyl; and RE18, octadecyl rosmarinate).

against both gram-positive and gram-negative bacteria,

yeasts, and molds, whereas the individual compounds had
little or no antimicrobial activity (3, 33). We therefore
investigated whether the esterification of RA similarly
altered its antimicrobial activity. We used two selected
bacterial strains, Staphylococcus carnosus LTH1502 and
Escherichia coli K-12 LTH4263. Our hypothesis for this
study was that preference for microbial membranes, and
hence antimicrobial activity, may be increased if RA is
esterified with appropriate alcohols. The antimicrobial
activity of RA and its esters was studied in model microbial
systems. Potential alterations in the appearance and
physiological state of cells were assessed by fluorescence
microscopy.
MATERIALS AND METHODS
Chemicals. RA and dimethyl sulfoxide (DMSO) were
purchased from Sigma Aldrich (Steinheim, Germany). MuellerHinton broth (MHB) and Mueller-Hinton agar (MHA) were
obtained from Oxoid (Basingstoke, Hampshire, England). Standard-I nutrient agar was purchased from Merck (Darmstadt,
Germany). Rosmarinate esters (REs), esterified with fatty acids
with carbon chain lengths of 1, 4, 8 10, 12, 16, and 18 (RE1 to 18,
Fig. 1), were synthesized and purified at CIRAD, Unite Mixte de
Recherche (UMR), Ingenierie des Agropolyme`res et Technologies
Emergentes (IATE) (Montpellier, France). All stock solutions were
prepared with phosphate-buffered saline (PBS) at pH 7.2 to 7.4.
Microorganisms. S. carnosus LTH1502 and E. coli K-12
LTH4263 were grown in MHB and on MHA. Prior to each
experiment, a single colony of culture was activated twice by
incubation in MHB and incubated at 37uC with shaking at 180 rpm
for 18 to 20 h (until the stationary phase was reached).
Antimicrobial activity screening assays. A microtiter
endpoint assay and a macrobroth endpoint assay were initially
used to evaluate the antimicrobial activities of RA and RE1 to 18
against S. carnosus LTH1502 and E. coli K-12 LTH4263.
Inhibitory activities were determined after 48 h of incubation.
Microtiter endpoint assay. A microtiter dilution assay was
used to determine the concentration-dependent inhibition effect of
RA and REs against S. carnosus LTH1502 and E. coli K-12
LTH4263. The wells were loaded with 120 ml of a solution
containing RA or REs ranging in concentrations between 0 and

1.5 mM. Double-strength MHB (120 ml) was added containing the
bacterial inoculum. Final inoculum levels in the wells were
approximately 5 | 105 CFU/ml. Microtiter plates were incubated
at 37uC with shaking at 180 rpm, and optical density (OD) was
measured after 48 h at 630 nm using a Synergy HT multi-mode
microplate reader (Bio-Tek Instruments, Inc., Bad Friedrichshall,
Germany). All measurements were done in triplicate, and samples
were duplicated. The MIC was determined as the concentration at
which the normalized OD (~OD48 h 2 OD0 h) was smaller than
0.05.
Macrobroth endpoint assay. Based on the results of the
microtiter endpoint assay, a concentration below the MIC
determined above was chosen to conduct a macrobroth endpoint
assay with S. carnosus LTH1502 and E. coli K-12 LTH4263. In
that assay, RA and REs were tested at 0.3 mM in PBS containing
5% DMSO to aid in the dissolution of the more lipid-soluble REs
(note that previous studies have shown that DMSO does not inhibit
growth at levels below 10%). The macrobroth assay was
performed by mixing 1 ml of culture in double-strength MHB
(final level: 5 | 105 CFU/ml inoculum) with 1 ml of RA or REs
and incubated at 37uC with shaking at 180 rpm for 48 h. Samples
were withdrawn after 48 h, serially diluted 1:10 in PBS, and plated
on Standard-I nutrient agar for enumeration (Whitley automatic
spiral plater, Don Whitley Scientific Limited, West Yorkshire,
UK). Plates were incubated at 37uC for 24 h, and colonies were
then counted using a colony counter (Synbiosis, Frederick, MD).
All samples were, at a minimum, duplicated.
Microtiter growth-over-time assay. Based on the results of
the above experiments, four compounds (RA, RE1, RE12, and
RE18) were chosen for closer investigation of their activity against
S. carnosus LTH1502, using a microtiter growth-over-time assay.
Wells were loaded with 120 ml of solutions of RA (0 to 1.5 mM),
RE1 (0 to 0.9 mM), RE12 (0 to 0.5 mM), or RE18 (0 to 0.5 mM)
containing DMSO. Twofold-concentrated PBS (120 ml) was added
with the bacterial inoculum. Final inoculation levels were 5 | 105
CFU/ml. Microtiter plates were incubated at 37uC with shaking at
180 rpm, and ODs were measured at 0, 3, 6, 12, and 24 h at 630 nm
using a Synergy HT multi-mode microplate reader (Bio-Tek). All
measurements were done in triplicate, and samples were
duplicated. The MICs were determined.
Macrobroth time-kill assay. To further determine the
functionality of compounds, RA and its esters were added at three
different growth stages to S. carnosus LTH1502 cultures. (i) Lag

J. Food Prot., Vol. 76, No. 9

ANTIMICROBIAL ACTIVITIES OF ROSMARINATE ESTERS

1541

FIGURE 2. Cell numbers of S. carnosus LTH1502 (A) and E. coli K-12 LTH4263 (B) and its esters obtained by plate counting after
incubation of bacteria in the presence of 0.3 mM RA for 48 h at 37uC. The inoculation levels of S. carnosus LTH1502 and E. coli K-12
LTH4263 were approximately 5 | 105 CFU/ml. Results with different letters (a to d) are significantly different (P , 0.05).
phase. Overnight cultures were inoculated at 5 | 105 CFU/ml
with RA, RE1, RE12, and RE18 at 0.75 mM (1/2 the MIC of RA).
Cells were enumerated at 0, 0.5, 1, 2, 3, 4.5, 6, 9, 12, 24, and 48 h
to generate a growth curve. (ii) Exponential phase (adapted from
Mascio et al., 2007 (28)). Overnight cultures were diluted to 5 |
106 CFU/ml and incubated at 37uC for 2 h before adding RA and
the respective esters RE1, RE12, and RE18 at 5 mM (10| MIC of
RE12). Cells were enumerated at 0, 0.3, 1, 2, 3, and 4 h. (iii)
Stationary phase. A 20- to 22-h activated culture (,108 CFU/ml)
was collected by centrifugation at 3,000 | g for 15 min at 4uC and
was twice washed with PBS. The bacteria were treated with RA,
RE1, RE12, and RE18 at 10 mM (20| MIC of RE12). Cells were
plated for enumeration at 0, 0.5 1, 2, 3, 6, 9, 12, 24, and 48 h.
For enumeration, all samples were serially diluted in PBS and
plated on Standard-I nutrient agar, using a Whitley automatic spiral
plater (Don Whitley Scientific Limited). Plates were incubated at 37uC
for 24 h, and colonies were counted using a colony counter (Synbiosis).
Fluorescence microscopy. Fluorescence microscopy was
used to investigate the appearance and the physiological state of
cells treated with or without RA and its esters. Bacterial cells
imaged were from the stationary phase of the above-described
time-kill assays. Studies have reported that the phenolic group of
RA induces a fluorescence that can be detected if samples are
exciting at 395 nm and emission is recorded at 480 nm. Thus,
direct fluorescence observation of cells without use of a stain was
carried out. Propidium iodide (PI), a fluorescence dye typically
used to monitor changes in cell membrane integrity, was also used
(18); PI is able to give information about a potential cell death. The
excitation and emission ranges of PI are 470 to 490 and 620 to
650 nm, respectively. For staining and imaging, a modified method
from Kim et al. (18) was used: 30 ml of 2 mM PI in DMSO was
mixed with 2 ml of bacterial suspensions treated with RA or RE,
stirred thoroughly, and kept in the dark for 15 min. The samples
were then placed on glass slides and covered with a cover slip.
Treated bacterial cells with or without PI were observed under an

oil immersion objective (100| magnification) with a fluorescence

microscope (Axioplan Universal microscope, Zeiss, Gottingen,
Germany). Images were digitally recorded and adjusted for
brightness and contrast using Image J software (version 1.46,
U.S. National Institutes of Health, Bethesda, MD).
Calculation of HLB values. Hydrophilic-lipophilic balance
(HLB) values of REs were calculated following the equation
described by Davies (13), namely, HLB ~ 7 z g(hydrophile
group numbers) 2 g(lipophile group numbers). HLB group
numbers were taken from Guo et al. (16).
Measurement of f-potential of bacteria. Overnight incubated cultures (S. carnosus LTH1502 and E. coli K-12 LTH4263)
were washed twice with PBS at pH 7.2. Bacterial cells were
centrifuged at 3,000 | g for 15 min at 4uC to collect cell pellets.
Pellets were redispersed in PBS. The procedure was repeated once
more. f-Potential measurements were conducted at a bacterial cell
level of approximately 107 CFU/ml. Suspensions were loaded into
a cuvette of a particle electrophoresis instrument (Zetasizer Nano
ZS, Malvern Instruments, Malvern, UK), and the f-potential was
determined by measuring the direction and velocity of bacteria
movement in the applied electric field (4, 27).
Calculation of means and standard deviations and
statistical analysis. All measurements were repeated at least three
times using, at a minimum, duplicate samples. Means and standard
deviations were calculated from these measurements using Excel
(version 2010, Microsoft, Redmond, WA). Results were subjected
to an analysis of variance (ANOVA). Means were analyzed by a
one-way ANOVA with significance set at P , 0.05.

RESULTS
Microtiter and macrobroth endpoint assays with E.
coli K-12 LTH4263 and S. carnosus LTH1502. The
antimicrobial activity of RA against E. coli K-12 LTH4263

1542

SURIYARAK ET AL.

and S. carnosus LTH1502 was first assessed using a

microtiter endpoint assay. RA inhibited the growth of S.
carnosus LTH1502 but did not affect E. coli K-12
LTH4263. The MIC of RA was 1.5 mM, based on the
standard definition of the MIC in microtiter assays, in which
OD must remain below 0.05 for the duration of the
microtiter test.
We subsequently conducted an endpoint macrobroth
assay with both E. coli K-12 LTH4263 and S. carnosus
LTH1502, whereby the cultures were incubated with 0.3 mM
RA and its esters for 48 h and samples were then withdrawn
for enumeration by plate counting (Fig. 2). To facilitate the
solubilization of the higher alkyl chain length esters (.12),
5% DMSO had to be added to the broth. As demonstrated by
the control, the addition of DMSO did not affect the growth
of bacteria (2). At 0.3 mM, S. carnosus LTH1502 was
affected by the addition of some of the esters (Fig. 2A). For
example, addition of RE4 led to a reduction of cell numbers
from ,105 CFU/ml (the inoculation level) to ,103 CFU/ml.
At 0.3 mM, addition of RE8, RE10, and RE12 reduced cell
numbers below detectable levels. In contrast, exposure of E.
coli K-12 LTH4263 to RA and its esters at 0.3 mM had no
effect (Fig. 2B).
Microtiter growth-over-time assay with S. carnosus
LTH1502. Based on the results obtained above, we
conducted more detailed antimicrobial efficacy studies with
RA and its esters against only S. carnosus LTH1502 since
the compounds were ineffective against E. coli K-12
LTH4263 cells. We carried out a microtiter growth-overtime assay with RA, RE1, RE12, and RE18, in which the
progression of the OD as a function of incubation time was
measured when S. carnosus LTH1502 was exposed to a
variety of concentrations (RA: 0 to 1.5 mM; RE1: 0 to
0.9 mM; RE12: 0 to 0.5 mM; RE18: 0 to 0.5 mM) of the
different compounds (Fig. 3).
The results of the growth-over-time microtiter assay
with RA confirm the previous endpoint assay with RA
against S. carnosus LTH1502. At concentrations of 0.9
and 1.1 mM, the increases in OD over time were less
pronounced than in the control, indicating that growth
occurred at a slower rate (Fig. 3A). With RE1, the OD
remained unchanged at concentrations above 0.7 mM,
indicating suppression of growth (Fig. 3B). At RE1
concentrations of 0.3 and 0.9 mM, increases in OD occurred
at a slower rate compared with the control, indicating again
a reduced growth rate. Results of the assay suggest that
RE12 was the most effective compound tested; concentrations of .0.3 mM were sufficient to completely suppress
growth (Fig. 3C). Even at concentrations as low as 0.1 mM
RE12, increases in OD occurred at a lower rate compared
with the control, indicating that a retardation of growth took
place. Finally, addition of RE18 at concentrations ranging
from 0.0 to 0.5 mM had no effect on growth of S. carnosus
LTH1502; and the ODs increased at a similar rate,
regardless of the concentration of RE18 present (Fig. 3D).
Macrobroth time-kill assay with S. carnosus
LTH1502. To better understand the antimicrobial action

J. Food Prot., Vol. 76, No. 9

of RA and its esters, a time-kill assay was conducted with

RA, RE1, RE12, and RE18 against S. carnosus LTH1502.
The objective was to elucidate the effect of addition of the
compounds with respect to the developmental stage of the
microorganisms. Thus, compounds were added to S.
carnosus LTH1502 cultures during the lag, exponential,
and stationary phases. Cell numbers after addition of
compounds were determined by plate counting (Fig. 4).
In the lag phase (Fig. 4A), cultures had an initial
inoculum level of 5 | 105 CFU/ml. In the absence of any
compounds (control), cultures grew to ,109 CFU/ml within
a period of 48 h. When 0.75 mM RA, RE1, RE12, and
RE18 were added to the system, progressions of cell
numbers were noticeably altered depending on the type of
compound added. Only a slight retardation in the growth
was observed upon addition of RA, whereas addition of
RE1 resulted in cell numbers remaining unchanged for a
period of 12 h and decreasing below detectable levels
thereafter. Addition of RE12 rapidly decreased cell numbers
below detectable levels in as little as 4.5 h. Finally, addition
of RE18 had no effect on growth of S. carnosus LTH1502.
In the exponential phase, cell numbers had increased to
5 | 106 CFU/ml, at which point the compounds were
added at 5 mM (Fig. 4B). Typically, cells are more resistant
in the exponential phase, and thus a substantially higher
concentration had to be used to observe an effect. Note that
the concentrations shown in the figures were selected such
that an optimal distinction of inhibitory activities of the
different compounds was possible (data for concentration
selection experiments are not shown). Controls grew to
,109 CFU/ml within a period of 4 h. A slight retardation in
growth was observed upon addition of 5 mM RA. In
contrast, addition of 5 mM RE1 led to a reduction of cell
numbers to ,104 CFU/ml over the course of 4 h. RE12, the
most effective compound, reduced cell numbers below
detectable levels within 2 h, whereas again, RE18 had no
impact on the growth behavior.
Finally, in the stationary phase, with cell numbers at 5
| 108 CFU/ml, 10 mM RA and its esters were added
(Fig. 4C). Cellular levels in controls remained stationary for
the observation period but decreased more or less rapidly
upon addition of compounds. With 10 mM RA, S. carnosus
LTH1502 decreased to ,104 CFU/ml over a period of 48 h.
In contrast, with RE1 and RE12, cell numbers decreased
below detectable levels within 48 and 9 h, respectively.
RE18 had little effect on cell numbers, with levels
remaining at ,107 CFU/ml after 48 h.
The results of the time-kill assay confirm that RE12
was the most effective compound tested against S. carnosus
LTH1502, regardless of growth stage. Note that we
conducted an additional experiment (data not shown), in
which we applied a very low concentration of RE12 (0.5 mM)
during the exponential phase. Even then, cell numbers
decreased by ,3 log within 30 min after incubation.
Fluorescence microscopy images of RA- and REtreated S. carnosus LTH1502. We observed in the timekill assay above that addition of RE12 to bacterial cultures
led to the formation of bacterial clusters or bacterial

J. Food Prot., Vol. 76, No. 9

ANTIMICROBIAL ACTIVITIES OF ROSMARINATE ESTERS

1543

FIGURE 3. Effect of (A) rosmarinic acid (RA), (B) methyl rosmarinate (RE1), (C) dodecyl rosmarinate (RE12), and (D) octadecyl
rosmarinate (RE18) on the growth behavior of S. carnosus LTH1502 over 24 h of incubation. Growth was assessed by measuring the
optical density (OD) of cultures incubated at 37uC in the presence of different concentrations of RA and its esters over the course of 24 h.
Shown is the normalized OD (OD 2 OD0), where OD0 is OD at time zero.

flocs, which was notably different than when other

compounds had been added. We thus investigated changes
to the appearance and physiological state of bacterial cells
treated with no compounds (controls), with RA, and with
RE12, using fluorescence microscopy (Fig. 5). Fluorescence microscopy was used since the phenol group of RA is
known to fluoresce when excited at 395 nm. Moreover, PI, a
red fluorescent dye that can be excited at 488 nm and that is
often used as an indicator of cell viability due to its ability to
bind to DNA as an indicator for dead bacteria cells, was
used to obtain information about the state of treated cells.
We selected cells in the stationary growth stage and

incubated them for 48 h with 200 mM RA and for 4 h

with 10 mM RE12.
Cells treated with 200 mM RA displayed no change in
appearance (image not shown). Nevertheless, cells were
dead, which was confirmed both by enumeration and by the
strong red fluorescence signal obtained after the PI stain had
been added. In contrast, upon addition of RE12, a dramatic
change occurred in appearance and physiological state,
including the formation of cellular clusters (Fig. 5). The
strong blue fluorescence signal in Figure 5 (left), which
originates from the RA ester RE12 itself, suggests that RE12
was present both in membranes as well as in the cellular

1544

SURIYARAK ET AL.

J. Food Prot., Vol. 76, No. 9

interior. Moreover, the red color upon addition of the PI

stain (Fig. 5, right) suggests complete cell death, which was
also confirmed by enumeration (e.g., cell numbers were
below detectable levels).
Physical properties of RA and its esters. We
hypothesized that the results shown above are due to
differing physical properties of RA and its esters (see
introductory paragraph), i.e., their affinity for bilayer
membranes. Thus, HLB values were calculated by the
Davies rule from the knowledge of the molecular structures
of RA, RE1, RE12, and RE18 (Fig. 1) and were 30.5, 14.1,
7.9, and 5.0, respectively (Table 1). The calculations
suggest that RA has little surface activity, with the
compound preferring to be in the aqueous phase. In
contrast, attachment of a single alkyl chain to RA induces
a substantial increase in affinity for interfaces, as indicated
by a HLB of 14.1. With an HLB value of 7.9, RE12 is still
water soluble but has a strong preference for association
with amphiphilic compounds of similar HLB values,
namely, phospholipids of which bacterial membranes are
composed. Finally RE18 with an HLB of 5.0 has
characteristics of an oil-soluble surfactant, able to form
reverse micelles in an oil phase.
f-Potential of bacterial cells. Finally, we measured
the surface charge of bacteria in order to develop and
confirm a mechanistic model (see below). The f-potential of
S. carnosus LTH1502 was 222 mV, whereas that of E. coli
K-12 LTH4263 was 218 mV.
DISCUSSION
General considerations. Phenolipids synthesized
of RA or chlorogenic acid and a variety of alcohols have
been mainly used as antioxidants to prevent lipid oxidation
in emulsified systems (22, 23). They have been considered
safe for use in foods or pharmaceuticals since the ester bond
between the phenolic moiety and the lipid carbon chain is
readily hydrolyzed under physiological conditions (6).
Phenolipids composed of RA and alcohols with alkyl chain
lengths of ,18 exhibited more or less pronounced
antimicrobial activities depending on the length of the alkyl
chain. In initial studies (results not shown) the activity of
esters of chlorogenic acid was also assessed, but compounds
did not inhibit the growth of our test organisms. Thus one
may conclude that both the chemical nature of the
hydrophilic moiety (e.g., rosmarinic or chlorogenic acid)
and the hydrophobic moiety impact antimicrobial activity of
phenolipids.

r
FIGURE 4. Assessment of the effect of addition of rosmarinic
acid and its esters at different stages of growth of S. carnosus
LTH1502. Application concentrations were adjusted to account for
the different sensitivities of cells during the different growth stages:
(A) lag phase, 0.75mM; (B) exponential phase, 5 mM; (C)
stationary phase, 10 mM. Cell numbers were determined by plate
counting at 0, 0.5, 1, 2, 3, 4.5, 6, 9, 12, 24, and 48 h.

J. Food Prot., Vol. 76, No. 9

ANTIMICROBIAL ACTIVITIES OF ROSMARINATE ESTERS

1545

FIGURE 5. Fluorescence microscopy images of S. carnosus LTH1502 cells incubated with dodecyl rosmarinate (RE12) at 10 mM for 4 h.
Left, fluorescence images observed under 395- to 440-nm filter without PI stain; right, images observed with PI stain.

RA has been shown to be an effective antimicrobial,

inhibiting growth of both aerobic and anaerobic bacteria
depending on system pH (12, 14, 42, 44). In contrast,
chlorogenic acid has not been reported to exhibit any
significant antimicrobial activity against foodborne bacteria.
Wen and coauthors, for example, reported that chlorogenic
acid was ineffective against Listeria monocytogenes even at
levels of .1.0% (wt/vol) (45). Because the alcohols used
for esterification in our study are not inhibitory in
themselves at the levels used, we suggest that at least one
of the moieties needs to have an antimicrobial function in
order for phenolipids to exhibit antimicrobial activities.
Specificity of activity against test organisms. RA and
its esters were effective only against S. carnosus LTH1502
but not E. coli K-12 LTH4263; this indicates that their
activities are somehow related to the structure of the cell
membrane since the principal difference between the two
types of organisms is the presence or absence of a
lipopolysaccharide (LPS) layer. Membrane-active antimicrobials are often more effective against gram-positive
bacteria than gram-negative bacteria, since the LPS layer
may prevent insertion of the compounds into their
phospholipid bilayers (39). In these cases, addition of
chelators that disrupt the integrity of the LPS layer by
binding the LPS-layer stabilizing metal ions may lead to the
induction of activity against gram-negative bacteria. One
may thus speculate that an activity enhancement may be
similarly feasible if chelators are added to phenolipids.

More studies that also include a larger number of microbial

strains will be needed to confirm this hypothesis.
Impact of alkyl chain length. The activity of RA
esters against S. carnosus LTH1502 showed a nonlinear cutoff effect depending on alkyl chain length. The results of all
assays show that antimicrobial activity peaks if the alkyl
chain length is 12 and that activity decreases at both shorter
and longer alkyl chain lengths. Such a cut-off effect in a
surfactant-based system has not only been observed with
respect to antimicrobial activity (10, 21, 37, 46) but also
with respect to a number of other properties (17), such as the
thermodynamic stability of self-assembled structures formed
of such compounds (41) and the antioxidant activity in
systems in which interfaces are present (22, 36). This may
be attributed to the overall free energy gain that can be
realized when such compounds are removed from the
continuous phase and are inserted into an interface or a
membrane (5, 17, 25).
Particularly at alkyl chain lengths of 10 to 12, a
pronounced partitioning into membrane structures occurs
(35). Investigations with other compounds of carbon chain
lengths of 10 to 12, such as daptomycin (38), monolaurin
(24), capric acid (37), lauric acid (46), dodecyl gallate (21),
and dodecyl dimethylamine oxide (C12NO) (20) have
shown that an optimal length of the alkyl group maximizes
the driving force for an insertion into lipid bilayer structures.
For example, the insertion of surfactants such as n-dodecyl
octaethylene glycol monoether and dodecyl dimethylamine

TABLE 1. Physical characteristics of rosmarinic acid and its esters a

Compound

HLB

Behavior of compound

RA (rosmarinic acid)
RE1 (methyl rosmarinate)
RE12 (dodecyl rosmarinate)
RE18 (octadecyl rosmarinate)

30.5
14.1
7.9
5

Water-soluble compound
Amphiphilic compound with a preference for stabilizing oil-in-water dispersions, forms micelles
Amphiphilic compound, both oil and water soluble, with preference for bilayer structures, forms micelles
Amphiphilic compound, oil-soluble, with a preference for water-in-oil dispersions, forms reverse micelles

Hydrophilic-lipophilic balance (HLB) numbers were calculated using the Davis rule (16). Behavior of compounds was visually assessed
by dispersing them in phosphate buffer at pH 7.2. Formation of micelles was assessed by dynamic light scattering studies.

1546

SURIYARAK ET AL.

J. Food Prot., Vol. 76, No. 9

FIGURE 6. Proposed mechanism of interaction of rosmarinic acid and RE12 with gram-positive bacterial cells.

oxide in small unilamellar liposomes has been studied.

There, an increase in diameter and a decrease in liposomal
curvature, together with a decrease in the unfavorable
interactions of the compounds with the water phase, have
been attributed to a free energy gain upon solubilization
(20). For daptomycin (a 10-carbon chain compound),
substantial morphological alterations of the membranes of
Staphylococcus aureus and Enterococcus faecalis have been
observed after treating the cells with the compound (43).
Similarly, microcopy images in this study showed that
addition of the 12-carbon ester of RA induced changes in
appearance and physiological state in S. carnosus
LTH1502. In contrast, no such modifications were
observed when treating cells with only RA. We therefore
conclude that the esterification alters the mode of action of
the antimicrobial, with the compound not simply disrupting
the cellular mechanism anymore but now disrupting the
integrity of the membranes, and that this disruption is
particularly pronounced if an ester with an alkyl chain
length of 12 is used.
Impact of growth stage of S. carnosus LTH1502.
Organisms are more or less susceptible to the superimposition of stresses, depending on the growth stage they are in

(12). Generally, the growth of bacteria can be associated

with one of three phases: the lag, exponential, or stationary
phase. In the lag phase, cells adapt to their new environment
and do not proliferate at a maximal rate. Cells in the
exponential phase are metabolically highly active, rapidly
increase in number, and consume available nutrients. In the
stationary phase, growth of cells is reduced or stops since
nutrients have been largely used up and secondary
metabolites generated during the growth phase may inhibit
further growth. Due to the superimposed stress in this stage,
cells are relatively resistant.
In each stage, a change in membrane composition
occurs, which is one reason for the different susceptibilities
of cells depending on type, age of cell, and characteristic of
the active compounds (19). This is particularly true for
antimicrobials that are membrane active. For instance,
chitosan and their derivatives, highly effective antimicrobials that interact electrostatically with bacterial membranes,
have been shown to be more or less effective against S.
aureus CCRC 12657, depending on the growth stage (8).
This is in contrast to the activity of some antibiotic
compounds such as beta-lactams (40) or quinolones (15),
which have been shown to act preferentially on dividing
cells rather than nondividing cells; this phenomenon is not

J. Food Prot., Vol. 76, No. 9

ANTIMICROBIAL ACTIVITIES OF ROSMARINATE ESTERS

related to membranes but rather to the synthesis of DNA and

RNA (15).
Effect of pH and charge of compounds. Our
antimicrobial activity studies were performed at pH 7.2 to
7.4. Since the pKa of the phenolic acid group is
approximately 2 to 3, the RA molecules were negatively
charged. In addition, the alcohol groups connected to the
base phenol contribute (albeit to a lower degree) to the
overall negative charge of the molecules. With the
esterification, the negative charge somewhat decreases since
the acid group is converted into an ester bond disallowing
deprotonation. Bacterial species tested were also negatively
charged; S. carnosus LTH1502 had a f-potential of 222 mV
and E. coli K-12 LTH4263 had a f-potential of 218 mV.
Thus, one may expect that both RA and its esters experience
an electrostatic repulsion force that prevents insertion and
penetration of the compounds into the membranes.
Proposed mechanism of action. The proposed
mechanism of action of compounds is shown in Figure 6.
When we used a surfactant-based approach to determine the
physicochemical properties of surfactant, namely the HLB
number and the solubilization behavior (Table 1), it was
apparent that the behavior of the RA is governed by the
alkyl chain length. RA and its esters with alkyl chain lengths
of 1 and 4 are predominantly hydrophilic and do not need a
solubilizer to be dispersed in water (i.e., RA, RE1, and RE4
are completely soluble at .10, .1, and .1 mM,
respectively). RA esters having chain lengths of 8 to 12
are partially soluble in water and require only small amounts
of solubilizer. Compounds that fall into those two categories
can therefore readily interact with bacterial cells. In contrast,
RA esters with alkyl chain lengths of 16 and 18 are oil
rather than water soluble, and high concentrations of an
effective solubilizer such as DMSO are needed to disperse
them in water. Their potential interaction with bacterial cells
may therefore be reduced. In addition to the solubilization
behavior of compounds, their affinity for lipid phases
changes. At chain lengths of 8 to 12, a maximal affinity for
lipid phases can be observed since HLB values match. Both
at shorter and longer alkyl chain lengths, this affinity
decreases, explaining the observed cut-off effect. In
addition, since compounds are all negatively charged at the
pH tested, the electrostatic repulsion preventing interaction
of compounds with bacterial cells will have to be
compensated by the enthalpy gain that is realized when
the compounds are inserted in the membranes. This
enthalpy gain is largest if alkyl chain lengths vary between
8 and 12.
In conclusion, our results demonstrate that chain length
of fatty acids in RA esters influences their antimicrobial
activities. Our study provides a new approach to increase
activity of naturally occurring antimicrobials by changing
their physical properties. Selection of appropriate alcohols
for esterification is of critical importance. This study
confirms the previously reported magic ability of lauryl
alcohol (C12) to induce affinity for bacterial membranes.
Whereas our results have not shown activity of long chain

1547

RA esters (C14 and above) in aqueous microbiological

model systems, these compounds may yet exhibit activity in
complex food systems in which fats and oils are present.
Moreover, carrier systems such as surfactant micelles could
help overcome the solubility limitations encountered in the
longer alkyl chain esters.
ACKNOWLEDGMENTS
This project was supported by funds from the University of
Hohenheim Experiment Station, Stuttgart, Germany, and by the Strategic
Fellowship for Frontier Research Networks from the Commission on
Higher Education, Thailand. We greatly appreciate the support of Dr. A.
Heller of the Department of General Botany (210a) at the University of
Hohenheim for her assistance in generating the microscopic images. We
thank Dr. M. Laguerre and Dr. J. Lecompte at CIRAD UMR IATE,
Montpellier, France, for their valuable suggestions and insights.

REFERENCES
1. Al-Sereiti, M. R., K. M. Abu-Amer, and P. Sen. 1999. Pharmacology
of rosemary (Rosmarinus officinalis Linn.) and its therapeutic
potentials. Indian J. Exp. Biol. 37:124130.
2. Ansel, H. C., W. P. Norred, and I. L. Roth. 1969. Antimicrobial
activity of dimethyl sulfoxide against Escherichia coli, Pseudomonas
aeruginosa, and Bacillus megaterium. J. Pharm. Sci. 58:836839.
3. Bakal, G., and A. Diaz. 2005. The lowdown on lauric arginate: food
antimicrobial hammers away at plasma membrane, disrupting a
pathogens metabolic process. Food Qual. Saf. February/March.
Available at: http://www.foodquality.com/details/article/878217/
The_Lowdown_on_Lauric_Arginate.html. Accessed 5 June 2012.
4. Balaev, A. E., K. N. Dvoretski, and V. A. Doubrovski. 2002.
Refractive index of Escherichia coli cells. Proc. SPIE 4707:253
260.
5. Balgavy, P., D. Uhrkova, J. Gallova, K. Lohner, and G. Degovics.
1993. Interaction of tertiary amine anesthetics with phosphatidylcholine bilayers, p. 184185. In P. Laggner and O. Glatter (ed.), Progress
in colloid and polymer science, vol. 93. Trends in colloid and
interface science VII. Springer, Berlin.
6. Bors, W., M. Christa, K. Stettmaier, L. Yinrong, and F. L. Yeap.
2004. Antioxidant mechanisms of polyphenolic caffeic acid oligomers, constituents of Salvia officinalis. Biol. Res. 37:301311.
7. Chakraborty, D., S. M. Mandal, J. Chakraborty, P. K. Bhattacharyaa,
A. Bandyopadhyay, A. Mitra, and K. Gupta. 2007. Antimicrobial
activity of leaf extract of Basilicum polystachyon (L) Moench. Indian
J. Exp. Biol. 45:744748.
8. Chen, Y.-L., and C.-C. Chou. 2005. Factors affecting the susceptibility of Staphylococcus aureus CCRC 12657 to water soluble lactose
chitosan derivative. Food Microbiol. 22:2935.
9. Cheung, S., and J. Tai. 2007. Anti-proliferative and antioxidant
properties of rosemary Rosmarinus officinalis. Oncol. Rep. 17:1525
1531.
10. Chu-Kung, A. F., R. Nguyen, K. N. Bozzelli, and M. Tirrell. 2010.
Chain length dependence of antimicrobial peptide-fatty acid conjugate activity. J. Colloid Interface Sci. 345:160167.
11. Cuvelier, M.-E., H. Richard, and C. Berset. 1996. Antioxidative
activity and phenolic composition of pilot-plant and commercial
extracts of sage and rosemary. J. Am. Oil Chem. Soc. 73:645652.
12. Davidson, P. M., J. N. Sofos, and A. L. Branen. 2005. Antimicrobials
in food. CRC Press, Boca Raton, FL.
13. Davies, J. T. 1957. A quantitative kinetic theory of emulsion type. I.
Physical chemistry of the emulsion agent, p. 426438. In Proceedings
of the 2nd International Congress on Surface Activity. Butterworths,
London.
14. Gaysinsky, S., and J. Weiss. 2007. Aromatic and spice plants: uses in
food safety. Stewart Postharvest Solutions 4:916.
15. Gradelski, E., B. Kolek, D. Bonner, and J. Fung-Tomc. 2002.
Bactericidal mechanism of gatifloxacin compared with other
quinolones. Chemotherapy 49:185188.

1548

SURIYARAK ET AL.

16. Guo, X., Z. Rong, and X. Ying. 2006. Calculation of hydrophilelipophile balance for polyethoxylated surfactants by group contribution method. J. Colloid Interface Sci. 298:441450.
17. Karlovska, J., K. Lohner, G. Degovics, I. Lacko, F. Devnsky, and P.
Balgavy. 2004. Effects of non-ionic surfactants N-alkyl-N, Ndimethylamine-N-oxides on the structure of a phospholipid bilayer:
small-angle X-ray diffraction study. Chem. Phys. Lipids 129:3141.
18. Kim, Y.-M., S. Farrah, and R. H. Baney. 2007. Membrane damage of
bacteria by silanols treatment. Electron. J. Biotechnol. 10:252259.
19. Kong, M., X. G. Chen, K. Xing, and H. J. Park. 2010. Antimicrobial
properties of chitosan and mode of action: a state of the art review.
Int. J. Food Microbiol. 144:5163.
20. Kragh-Hansen, U., M. Le Maire, and J. V. Maller. 1998. The
mechanism of detergent solubilization of liposomes and proteincontaining membranes. Biophys. J. 75:29322946.
21. Kubo, I., K. Fujita, K. Nihei, and N. Masuoka. 2003. Non-antibiotic
antibacterial activity of dodecyl gallate. Bioorg. Med. Chem. 11:573
580.
22. Laguerre, M., L. J. Lopez Giraldo, J. Lecomte, M.-C. FigueroaEspinoza, B. Barea, J. Weiss, E. A. Decker, and P. Villeneuve. 2010.
Relationship between hydrophobicity and antioxidant ability of
phenolipids in emulsions: a parabolic effect of the chain length
of rosmarinate esters. J. Agric. Food Chem. 58:28692876.
23. Lecomte, J., L. J. Lopez Giraldo, M. Laguerre, B. Barea, and P.
Villeneuve. 2010. Synthesis, characterization and free radical
scavenging properties of rosmarinic acid fatty esters. J. Am. Oil
Chem. Soc. 87:615620.
24. Lieberman, S., M. G. Enig, and H. G. Preuss. 2006. A review of
monolaurin and lauric acid: natural virucidal and bactericidal agents.
Altern. Complement. Ther. 12:310314.
25. Lohner, K. 1991. Effects of small organic molecules on phospholipid
phase transitions. Chem. Phys. Lipids 57:341362.
26. Lu, Y., and L. Y. Foo. 1999. Rosmarinic acid derivatives from Salvia
officinalis. Phytochemistry 51:9194.
27. Marquis, R. E. 1973. Immersion refractometry of isolated bacterial
cell walls. J. Bacteriol. 116:12731279.
28. Mascio, C. T., J. D. Alder, and J. A. Silverman. 2007. Bactericidal
action of daptomycin against stationary-phase and nondividing
Staphylococcus aureus cells. Antimicrob. Agents Chemother. 51:
42554260.
zcan, M. M., and D. Arslan. 2011. Antioxidant effect of essential
29. O
oils of rosemary, clove and cinnamon on hazelnut and poppy oils.
Food Chem. 129:171174.
zturk, M., M. E. Duru, B. Ince, M. Harmandar, and G. Topcu. 2010.
30. O
A new rapid spectrophotometric method to determine the rosmarinic
acid level in plant extracts. Food Chem. 123:13521356.
31. Petersen, M. S. 1991. Characterization of rosmarinic acid synthase
from cell cultures of Coleus blumei. Phytochemistry 30:28772881.
32. Razzaque, A., and B. E. Ellis. 1977. Rosmarinic acid production in
Coleus cell cultures. Planta 137:287291.

J. Food Prot., Vol. 76, No. 9

33. Rodriguez, E., J. Seguer, X. Rocabayera, and A. Manresa. 2004.

Cellular effects of monohydrochloride of L-arginine, N-lauroyl
ethylester (LAE) on exposure to Salmonella typhimurium and
Staphylococcus aureus. J. Appl. Microbiol. 96:903912.
34. Romano, C. S., K. Abadi, V. Repetto, A. A. Vojnov, and S. Moreno.
2009. Synergistic antioxidant and antibacterial activity of rosemary
plus butylated derivatives. Food Chem. 115:456461.
35. Rowe, E. S., F. Zhang, T. W. Leung, J. S. Parr, and P. T. Guy. 1998.
Thermodynamics of membrane partitioning for a series of n-alcohols
determined by titration calorimetry: role of hydrophobic effects.
Biochemistry 37:24302440.
36. Sasaki, K., J. Alamed, J. Weiss, P. Villeneuve, L. J. Lopez Giraldo, J.
Lecomte, M.-C. Figueroa-Espinoza, and E. A. Decker. 2010.
Relationship between the physical properties of chlorogenic acid
esters and their ability to inhibit lipid oxidation in oil-in-water
emulsions. Food Chem. 118:830835.
37. Skrivanova, E., M. Marounek, V. Benda, and P. Brezina. 2006.
Susceptibility of Escherichia coli, Salmonella sp. and Clostridium
perfringens to organic acids and monolaurin. Vet. Med. 51:8188.
38. Straus, S. K., and R. E. W. Hancock. 2006. Mode of action of the new
antibiotic for gram-positive pathogens daptomycin: comparison with
cationic antimicrobial peptides and lipopeptides. Biochim. Biophys.
Acta Biomembr. 1758:12151223.
39. Taylor, T. T., B. D. Bruce, J. Weiss, and P. M. Davidson. 2007.
Listeria monocytogenes and Escherichia coli O157:H7 inhibition in
vitro by liposome-encapsulated nisin and ethylene diamine tetraacetic
acid. J. Food Saf. 28:183197.
40. Tenover, F. C. 2006. Mechanisms of antimicrobial resistance in
bacteria. Am. J. Med. 119:S3S10.
41. Tristram-Nagle, S., R. N. A. H. Lewis, J. W. Blickenstaff, M.
DiPrima, B. F. Marques, R. N. McElhaney, J. F. Nagle, and J. W.
Schneider. 2005. Thermodynamic and structural characterization of
amino acid-linked dialkyl lipids. Chem. Phys. Lipids 134:2939.
42. Vegara, S., L. Funes, N. Mart, D. Saura, V. Micol, and M. Valero.
2011. Bactericidal activities against pathogenic bacteria by selected
constituents of plant extracts in carrot broth. Food Chem. 128:872
877.
43. Wale, L. J., A. P. Shelton, and D. Greenwood. 1989. Scanning
electron microscopy of Staphylococcus aureus and Enterococcus
faecalis exposed to daptomycin. J. Med. Microbiol. 30:4549.
44. Weckesser, S., K. Engel, B. Simon-Haarhaus, A. Wittmer, K. Pelz,
and C. M. Schempp. 2007. Screening of plant extracts for
antimicrobial activity against bacteria and yeasts with dermatological
relevance. Phytomedicine 14:508516.
45. Wen, A., P. Delaquis, K. Stanich, and P. Toivonen. 2003. Antilisterial
activity of selected phenolic acids. Food Microbiol. 20:305311.
46. Yang, D., D. Pornpattananangkul, T. Nakatsuji, M. Chan, D. Carson,
C.-M. Huang, and L. Zhang. 2009. The antimicrobial activity of
liposomal lauric acids against Propionibacterium acnes. Biomaterials
30:60356040.

Reproduced with permission of the copyright owner. Further reproduction prohibited without
permission.