Attachment organelle ultrastructure correlates with phylogeny, not

gliding motility properties, in Mycoplasma pneumoniae relatives

Jennifer M. Hatchel and Mitchell F. Balish

Department of Microbiology, Miami University, 80 Pearson Hall, Oxford, OH 45056, USA

Correspondence
Mitchell F. Balish
BalishMF@MUOhio.edu

ABSTRACT

The Mycoplasma pneumoniae cluster is a clade of eight described species which all exhibit cellular polarity.
Their polar attachment organelle is a hub of cellular activities including cytadherence and gliding motility, and
its duplication in the species M. pneumoniae is coordinated with cell division and DNA replication. The
attachment organelle houses a detergent-insoluble, electron-dense core whose presence is required for structural
integrity. Although mutant analysis has led to the identification of attachment organelle proteins, the
mechanistic basis for the activities of the attachment organelle remains poorly understood, with gliding motility
attributed alternatively to the core or to the adhesins. In this study we investigated attachment organelle-
associated phenotypes, including gliding motility characteristics and ultrastructural details, in seven species of
the M. pneumoniae cluster under identical conditions, allowing direct comparison. We identified gliding ability
in three species in which it has not previously been reported, Mycoplasma imitans, Mycoplasma pirum and
Mycoplasma testudinis. Across species, ultrastructural features of attachment organelles and their cores do not
correlate with gliding speed, and morphological features of cores are inconsistent with predictions about how
these structures are involved in the gliding process, disfavouring a prominent, direct role for the electron-dense
core in gliding. In addition, we found M. pneumoniae to be an outlier in terms of cell structure with respect to its
close relatives, suggesting that it has acquired a special set of adaptations during its evolution.

Abbreviations: SEM, scanning electron microscopy; TX, Triton X-100

A supplementary figure showing speed distributions for motile cells of each species is available with the online
version of this paper.

INTRODUCTION

The bacterial genus Mycoplasma includes over 100 species. These organisms lack cell walls and in nature are
always associated with vertebrate hosts. Several mycoplasma species exhibit a polarized cell morphology,
exemplified by Mycoplasma pneumoniae, a significant causative agent of assorted respiratory and non-
respiratory human disease conditions (Waites & Talkington, 2004 ). This surface-adherent organism exhibits
the pleomorphy typical of its genus, but is characterized by a prosthecal terminal organelle or attachment
organelle at one pole (Balish, 2006 ). This structure is the site at which adhesins are clustered, enabling the cell
to interact productively with host cells (cytadhere). The attachment organelle is the leading end of the cell
during gliding motility (Balish & Krause, 2006 ) and the location of the gliding motor (Hasselbring & Krause,
2007 ). Gliding appears to be important for spreading from the initial infection site (Jordan et al., 2007 ).
Duplication of the M. pneumoniae attachment organelle at a site adjacent to the existing one accompanies
initiation of DNA replication (Seto et al., 2001 ).
The cytoplasm within the M. pneumoniae attachment organelle lacks DNA, is clear of large particles, and
contains a Triton X-100 (TX)-insoluble structure, the electron-dense core (Biberfeld & Biberfeld, 1970 ), which
appears to be connected to an ill-defined network of cytoskeletal filaments present throughout the cell body
(Meng & Pfister, 1980 ; Göbel et al., 1981 ). Proper assembly of the core is essential for formation of a
functional attachment organelle (Krause & Balish, 2004 ), making the core essential for virulence. The core
consists of several subdomains, including a distal terminal button which is probably in direct contact with the
attachment organelle tip, a perpendicularly striated double rod, and a proximal complex that includes a bowl-
shaped component at the base (Henderson & Jensen, 2006 ; Seybert et al., 2006 ). Although studies of M.
pneumoniae attachment organelle mutants have led to the identification of many attachment organelle proteins
and to the outline of an assembly pathway (Krause & Balish, 2004 ), how each protein contributes to the
structure of the core remains poorly understood.

Mycoplasma gliding is poorly understood compared to other prokaryotic forms of motility. The M. pneumoniae
motility mechanism has been addressed through characterization of mutants with differences in colony-
spreading phenotypes, resulting in the identification of a moderate number of genes of both known and
unknown function (Hasselbring et al., 2006b ). In Mycoplasma mobile, a distant relative whose adherence,
motility and terminal organelle components are unrelated to those of M. pneumoniae (Miyata, 2005 ), gliding is
proposed to involve the cyclical binding and release by the adhesin Gli349 (Uenoyama et al., 2004 ) of
carbohydrate moieties present on a wide variety of host-cell proteins, including those deposited on the surface of
the microscope slide from serum in vitro (Nagai & Miyata, 2006 ). A similar binding-and-release process might
occur in M. pneumoniae, albeit through the use of an unrelated set of adhesins. In support of this hypothesis, the
adhesins P30 (Hasselbring et al., 2005 ) and P1 (Seto et al., 2005a ) have both been implicated in gliding
motility of M. pneumoniae. Alternative hypotheses have focused on the electron-dense core as the major
component of the M. pneumoniae motor, citing differences in fine features of individual M. pneumoniae cores
(Henderson & Jensen, 2006 ) or postulating that the core is capable of twisting within the attachment organelle
(Hegermann et al., 2002 ). Core-centred models require that the morphology of the core would change during
motility, such as by shortening or twisting, and one might further predict that organisms with different motile
properties would exhibit corresponding differences in the morphology of the core. The ability to distinguish
among these models is predicated on an understanding of the relationship between motile properties and
attachment organelle features. One approach to the basis for mycoplasma motility is the characterization of
relatives of M. pneumoniae that have a similar motility mechanism but differing motility characteristics, thereby
constituting a series of ‘naturally occurring mutants’.

Eight species are currently assigned to the M. pneumoniae cluster based on 16S rRNA sequence similarity
(Johansson & Pettersson, 2002 ) and the presence of an attachment organelle that contains a core (Balish &
Krause, 2005 ). Gliding has been described for some of these species, including M. pneumoniae, Mycoplasma
genitalium, Mycoplasma gallisepticum and Mycoplasma amphoriforme, while Mycoplasma alvi has been
reported to be non-motile (Bredt, 1979 ; Hatchel et al., 2006 ). Mean speeds reported in this cluster range from
50 nm s–1 for M. amphoriforme to >300 nm s–1 for M. pneumoniae (Hatchel et al., 2006 ). Homologues of the
M. pneumoniae attachment organelle proteins required for cytadherence and motility are restricted to the M.
pneumoniae cluster (Tham et al., 1994 ; Dhandayuthapani et al., 1999 ; Papazisi et al., 2002 ), making it likely
that a common fundamental mechanism involving this set of proteins underlies attachment organelle-mediated
functions in each of these organisms. For most of these species, detailed analysis of the dimensions of the
electron-dense cores has not been performed. Correlating molecular and structural aspects of attachment
organelle components with cytadherence and motility-related properties in each species will reveal which
components are likely participants in the motility process, directing future research on the mechanistic aspects of
mycoplasma gliding motility.

We have therefore examined each species of the M. pneumoniae cluster with respect to gliding motility and, for
motile species, the morphological features of attachment organelles and cores. Time-lapse
microcinematographic analysis revealed gliding in all species of the M. pneumoniae cluster except M. alvi.
However, scanning electron microscopy (SEM) revealed substantial differences in cell and attachment organelle
dimensions among all species which, while largely associated with phylogenetic relatedness, correlated poorly
with gliding motility characteristics, not supporting a direct role for the core in motility. In addition, SEM
revealed evidence of DNA association with the proximal end of the core in most species.

METHODS

Growth and culture conditions.

The strains used in this study were: M. pneumoniae strain M129; M. genitalium strain G37; M. gallisepticum
strain Rlow; Mycoplasma imitans strain 4229; M. amphoriforme strain A39; Mycoplasma testudinis strain 01008;
Mycoplasma pirum strain 70-159; and M. alvi strain Ilsley. Cells were grown in plastic tissue-culture flasks from
frozen stocks for 2–10 days at 37 °C (30 °C for M. testudinis) in SP-4 broth (Tully et al., 1979 ) to mid-
exponential phase (phenol red indicator was orange). M. alvi was also grown in a modified SP-4 formulation
containing a 1 : 1 mixture of fetal bovine and porcine sera (HyClone) and grown until slightly turbid. Motility
stocks were prepared as previously described (Hatchel et al., 2006 ). For M. testudinis and M. pirum, cells were
passaged three to five times in SP-4 broth to enrich for plastic-attached subpopulations prior to further use and
analysis.

Phylogenetic tree.
The 16S rRNA sequence for each species was entered into BioEdit Sequence Alignment Editor v. 7.0.5.2 (Hall,
1999 ) for trimming and neighbour-joining alignment by CLUSTAL_X v. 1.8 (Thompson et al., 1997 ). The
phylogenetic tree was generated from these data by NJPlot (Perrière & Gouy, 1996 ).

SEM.
SEM was performed as previously described (Hatchel et al., 2006 ) with minor modifications. Briefly, stocks
were grown in 24-well plates containing glass coverslips for 3 h to 1 day before processing. For analysis of TX-
insoluble structures, coverslips were incubated for 30 min at 37 °C in a solution containing 2 % TX in TN buffer
(20 mM Tris/HCl, pH 7.5, 150 mM NaCl). For DNase treatment, cells were grown for 1–4 days on coverslips
before processing. TX extraction was followed by a 30 min incubation with 5 U bovine pancreas DNase I
(Sigma-Aldrich) in 500 µl 1x DNase reaction buffer (10 mM Tris/HCl, pH 7.6, 2.5 mM MgCl2, 0.5 mM CaCl2)
at room temperature. For RNase A treatment, TX-extracted coverslips were incubated 30 min in 500 µl TE
buffer (10 mM Tris/HCl, pH 8.0, 1 mM EDTA) containing 100 µg RNase A (Qiagen). For whole cells,
coverslips were fixed 30 min in fixative A (1.5 % glutaraldehyde/1 % formaldehyde/0.1 M sodium cacodylate,
pH 7.2; M. pneumoniae, M. genitalium, M. gallisepticum, M. imitans and M. pirum) or fixative B (1 %
glutaraldehyde/2 % formaldehyde/0.1 M sodium cacodylate, pH 7.2; M. amphoriforme and M. testudinis). For
visualization of TX-insoluble material, fixative A was used for all species. Dehydration, critical-point drying
and gold coating were performed as previously described (Hatchel et al., 2006 ). Images were viewed and
captured on a Zeiss (Carl Zeiss) Supra 35 FEG-VP scanning electron microscope operating at 5 kV. Dimensions
of whole cells, TX-insoluble structures, and structures following DNase treatment were measured individually
using SPOT software (Diagnostic Instruments).

Time-lapse microcinematographic analysis.

Motility sequences were captured as previously described (Hatchel et al., 2006 ). Briefly, cells were inoculated
into chamber slides containing SP-4+3 % gelatin. After 3 h incubation at 37 °C, cells attached to the slide were
visualized using a Leica DM IRB inverted microscope (Leica Microsystems). The sample was held at 37 °C in a
heated chamber. Twenty-seven consecutive phase-contrast images were taken at appropriate time intervals and
later merged using Adobe Photoshop CS version 8.0. The speed of each cell was computed by measuring the
distance travelled and dividing it by the duration of movement. Cells moving in fewer than ten frames were
scored motile, but were not included in the calculation of mean speed because below this threshold we regarded
the speed measurements as unreliable.
 RESULTS

Gliding motility
Of the eight species found in the M. pneumoniae cluster based on 16S rRNA sequences (Fig. 1 ; Pitcher et al.,
2005 ), only M. alvi did not attach to glass and exhibit gliding motility, as previously reported (Bredt, 1979 ).
M. alvi did not attach under any conditions we tested, and consequently we were unable to investigate it further.
For M. pirum and M. testudinis, only a small minority of the population derived from the initial stocks attached
to glass. In these cases, all further studies were performed on plastic-attached subclones. Efforts to enrich for a
similar population of M. alvi cells were unsuccessful.

Fig. 1. Phylogenetic tree based on 16S rRNA

sequences of mycoplasmas within the M.
pneumoniae cluster. Scale bar, 0.1
substitutions per site.

Gliding characteristics, which were previously determined by our method for M. pneumoniae, M. amphoriforme
and M. gallisepticum, were assayed for the remaining species (Hatchel et al., 2006 ). Mean speeds ranged from
28 nm s–1 for M. pirum to 2970 nm s–1 for M. testudinis (Table 1 ; see also Supplementary Fig. S1, available
with the online version of this paper). M. testudinis was the fastest, though with the lowest percentage of motile
cells. Although cells glided in the direction of the attachment organelle, large-scale cellular movement appeared
generally random for most species. However, 80 % of M. genitalium and M. testudinis cells moved in nearly
circular patterns (not shown). There was no relationship between phylogeny and gliding speed.

Table 1. Gliding motility parameters at 37 °C

Mycoplasma species Mean speed (nm s– Range (nm s– Percentage of cells moving per frame (total
1 1
)±SD ) cells analysed)
M. pneumoniae str. 336±59 197–538 51 (149)
M129*
M. genitalium str. G37 111±22 62–172 72 (126)
M. gallisepticum str. 131±38 50–286 64 (421)
Rlow*
M. imitans str. 4229 107±31 41–194 65 (173)
M. testudinis str. 01008 2970±570 2180–4740 37 (100)
M. amphoriforme str. 49±19 15–133 53 (1034)
A39*
M. pirum str. 70-159 28±8 12–58 45 (289)

*Data from Hatchel et al. (2006).

Attachment organelle and cell morphology

To test whether gliding characteristics correlated with ultrastructural features of the attachment organelle, we
analysed both whole cells and TX-insoluble electron-dense cores in each of the seven adherent, motile species
by SEM. Attachment organelles were divided into substructures for common reference: the distal moiety was
termed the knob, and the proximal portion, the shaft (see Fig. 2h ). Dimensions of individual attachment
organelles and cells (Table 2 ) were measured.

Fig. 2. Scanning electron micrographs of mycoplasma cells grown on

glass coverslips. Three representative cells are shown for each species.
Images are aligned so that the attachment organelle is at the top. (a) M.
gallisepticum; (b) M. imitans; (c) M. amphoriforme; (d) M. testudinis; (e)
M. pirum; (f) M. genitalium; (g) M. pneumoniae; (h) schematic of
mycoplasma cell. Open and black arrows indicate a ridge-like feature at
the base of the attachment organelle shaft in M. gallisepticum and M.
amphoriforme respectively; white arrows indicate striations on the M.
pirum attachment organelle; black arrowheads indicate the curvature of
M. testudinis and M. genitalium attachment organelles. Scale bar,
250 nm.

Table 2. Whole-cell dimensions (nm)±SD, mean of 30

cells

Cell body Attachment organelle Knob

Whole-cell length Length Width Length Width Length Width
ND ND
M. pneumoniae 1500±420 520±90 220±30 290±40 80±10
M. genitalium 590±50 420±50 280±50 170±20 80±10 80±10 80±10
M. gallisepticum 830±120 710±110 330±70 120±20 150±20 100±20 150±20
ND ND
M. imitans 620±50 500±50 350±30 120±10 140±20
M. amphoriforme 770±70 620±70 360±60 150±20 150±20 90±10 140±20
ND ND
M. testudinis 680±100 560±100 280±20 120±10 130±20
M. pirum 720±110 480±90 230±20 250±30 100±10 100±20 100±10

ND, Not determined because knob and shaft were not uniformly distinguishable.

M. gallisepticum (Fig. 2a ), M. imitans (Fig. 2b ), M. amphoriforme (Fig. 2c ), and M. testudinis (Fig. 2d )

exhibited short, wide attachment organelles, with mean lengths from 120 to 150 nm, and widths from 130 to
150 nm (Table 2 ). The knob was distinct from the shaft in M. gallisepticum and M. amphoriforme, sometimes
in M. testudinis, but not in M. imitans. The M. pirum attachment organelle, which also featured a prominent
knob, was long, at 250 nm, and moderately wide, at 100 nm (Table 2 ), and uniquely exhibited three or four
parallel striations approximately perpendicular to the long axis (Fig. 2e ). M. genitalium (Fig. 2f ) and M.
pneumoniae (Fig. 2g ) had narrow shafts, only 80 nm in width. The M. genitalium structure, at 170 nm, was
substantially shorter than that of M. pneumoniae, which at 290 nm was the longest of the cluster, consistent with
previously noted differences (Lind et al., 1984 ). Knobs were indistinct in M. pneumoniae attachment organelles
(Fig. 2g ). The two species that exhibited circular gliding paths had curved attachment organelles. For M.
testudinis, 25 % of attachment organelles were bent (Fig. 2d ), whereas 100 % of M. genitalium attachment
organelles were curved an average of 2 °±1 ° from the long axis of the cell (Fig. 2f ), and the knob, prominent in
this species, was off-centre in the direction of this curvature.

Measurements of cell bodies revealed a positive correlation between cell width and attachment organelle width
(Table 2 ). M. pneumoniae and M. pirum cells were distinctly narrow in comparison to the others (Fig. 2e, g ;
Table 2 ). In contrast, cell length appeared unrelated to other features (Table 2 ). M. pneumoniae cells were
unique in the prominence of the trailing filament found at the pole opposite the attachment organelle (Fig. 2g ).
This structure, whose length varied greatly, drove the mean length of the M. pneumoniae cell to 1500 nm
(Table 2 ). M. gallisepticum and M. amphoriforme cell bodies often had a ridge-like feature in the vicinity of
the base of the attachment organelle shaft (Fig. 2a, c ).

Cell division
We observed SEM fields of cells from each of the seven
adherent species of this cluster (Fig. 3 ) to characterize
cells that appeared to be in the process of dividing. Either
the appearance of two attachment organelles at the same
pole on a single cell or two attachment organelles pointing
in different directions was taken to indicate that cell
division was in progress. All species except M. imitans
(Fig. 3b ) were observed to have cells exhibiting one or
both of these phenotypes. Dividing M. pirum cells were
only found to have attachment organelles at opposite
poles (Fig. 3e ), and in M. gallisepticum, adjacently
paired attachment organelles were not observed (Fig. 3a
), suggesting that in these species the new attachment
organelle might not form immediately adjacent to the old
one as in M. pneumoniae.
Fig. 3. Representative scanning electron micrographs of mycoplasma cells grown on glass coverslips. Each image is a representative
field of cells. (a) M. gallisepticum; (b) M. imitans; (c) M. amphoriforme; (d) M. testudinis; (e) M. pirum; (f) M. genitalium; (g) M.
pneumoniae. Arrows indicate cells that have two attachment organelles pointing at opposite poles; arrowheads indicate cells with two
attachment organelles not yet at opposite poles on a single cell. Scale bar, 1 µm.

TX-insoluble structures
Substructures of the electron-dense core of the attachment organelle, namely the terminal button, the rod, the
base and a tuft of fibrous material (Fig. 4h ; Hatchel et al., 2006 ), have been described. Although we
previously had difficulty resolving structures in M. gallisepticum (Hatchel et al., 2006 ), the use of a fresh
preparation of TX resulted in consistent observation of core-like structures in all seven species (Fig. 4 ).

Fig. 4. Scanning electron micrographs of

mycoplasma electron-dense cores. Under ‘TX
only’, cells were grown on glass coverslips and
extracted with 2 % TX for 30 min at 37 °C (see
Methods). Under ‘TX+DNase I’, cells were
grown on glass coverslips, extracted with TX as
above, and then treated with DNase I for 30 min
at room temperature (see Methods). Two
representative cores are shown for each species
with and without DNase treatment, with terminal
buttons aligned at the tops of images. (a) M.
gallisepticum; (b) M. imitans; (c) M.
amphoriforme; (d) M. testudinis; (e) M. pirum;
(f) M. genitalium; (g) M. pneumoniae; (h)
schematic of the mycoplasma electron-dense core
substructures; (i) M. gallisepticum treated with
TX and then RNase A. Arrows point to notches
in M. gallisepticum and M. imitans bases. Scale
bar, 250 nm.

Rod length was conserved across species over a narrow range from 130 to 170 nm; inclusion of the terminal
button raised the values to 180–240 nm (Table 3 ). However, rod width varied substantially among species,
from 30 to 80 nm (Table 3 ), and the variation correlated well with phylogenetic relatedness. The length of the
M. imitans rod (Table 3 ) is underestimated on account of being variably obscured by the unusual base (see
below). The slight difference between the current results and those previously reported for rod width in M.
amphoriforme might be due to the use of a different fixative preparation (Hatchel et al., 2006 ). Terminal button
length and width varied somewhat across species, and although the variation appeared unrelated to gliding
speed, width appeared to have a phylogenetic component, with M. pneumoniae and M. genitalium exhibiting the
narrowest terminal buttons (Table 3 ). The M. pneumoniae terminal button was frequently difficult to
distinguish from the rod (Fig. 4g ), as previously described (Hatchel et al., 2006 ). The M. genitalium rod
exhibited a pronounced lateral curvature, and the terminal button was consistently offset in the direction of this
curvature (Fig. 4f ). Curvature of the electron-dense core was also visible in 25 % of M. testudinis cells,
though in this species the curvature was more abrupt and located close to the terminal button (Fig. 4d ).
Although we could not image the non-adherent M. alvi, it is known to have an electron-dense core (Gourlay et
al., 1977 ).
 Table 3. Electron-dense core dimensions (nm)±SD, mean of 30 cells

Terminal button Rod Base

Length Width Length Width Terminal button+rod length Length Width
ND ND
M. pneumoniae 60±10 50±10 240±20 70±10 70±10
M. genitalium 50±10 60±20 160±20 50±10 220±20 110±20 120±20
M. gallisepticum 90±10 90±20 140±30 30±0 230±30 120±40 140±40
M. imitans 60±20 60±20 130±60 40±10 180±70 240±60 160±50
M. amphoriforme 70±10 90±20 160±20 60±10 220±20 90±20 140±30
M. testudinis 80±10 80±10 170±20 80±10 240±20 150±40 120±20
M. pirum 90±10 90±10 130±10 80±10 210±10 100±20 110±10

ND , Not determined because rod and terminal button were difficult to distinguish.

Except in M. pneumoniae, the base of the electron-dense core was obscured by irregular material. Treatment of
adherent TX-insoluble fractions with DNase I resulted in highly consistent loss of this material, exposing the
base of the core (Fig. 4 ). Incubation in buffer alone had no effect, nor did treatment with RNase A (Fig. 4i ).
Pronase caused loss of all visible structures (data not shown). In M. genitalium (Fig. 4f ) and M. amphoriforme
(Fig. 4c ) the DNase-sensitive material was predominantly filamentous, with some clumpier regions; in M.
gallisepticum (Fig. 4a ), M. imitans (Fig. 4b ), and M. testudinis (Fig. 4d ) this mass was mostly nodular, with
occasional filaments; and in M. pirum (Fig. 4e ) the material had a smooth appearance with some short
filaments extending from it. These filaments are consistent with the appearance of loops of DNA that have been
locally denuded of protein (Cunha et al., 2001 ). Taken together, these data are consistent with the identification
of the core-associated material as DNA, rendered condensed by treatment with detergent and salt (Cunha et al.,
2001 ), although whether the interaction between the DNA and the core is physiologically relevant is unclear.
Treatment with DNase had no significant effects on measurements of other core substructures, except for
occasional lengthening of the rod due to its being less obscured.

The dimensions of the exposed bases were highly variable across species (Table 3 ; Fig. 4 ). The base of M.
pneumoniae was by far the smallest in both length and width. At the other extreme was M. imitans, which had a
very large and highly irregularly shaped base (Fig. 4b ). The DNase-treated bases of M. gallisepticum and M.
imitans consistently had a notch-like feature of irregular orientation (Fig. 4a, b ), suggesting that the removal of
DNA was equally complete in each species. No relationship between base dimensions and gliding speed or
phylogeny was apparent across species. In fact, although attachment organelle morphology generally correlated
well with phylogeny, no relationship was detected between gliding speed and any element of attachment
organelle morphology.

DISCUSSION

Gliding motility and ultrastructure

In the present study, all species of the M. pneumoniae cluster except M. alvi exhibited gliding motility, raising
the total number of motile mycoplasma species to nine with the addition of M. imitans, M. testudinis and M.
pirum. These seven organisms exhibited a wide range of speeds and proportions of cells moving at a given time.
M. testudinis populations had a low frequency of motile cells, but interestingly, the motile cells glided as rapidly
as M. mobile, whose gliding speed was previously unchallenged among mycoplasmas (Miyata, 2005 ). Its
ultrastructure and phylogenetic position suggest that M. testudinis has a mechanistically similar motor to that of
its close relatives, but that it and M. mobile have evolved, perhaps under similar unidentified pressures, to glide
rapidly. Alternatively, despite its close relatedness to M. pneumoniae and its similar appearance to its relative,
M. testudinis might have acquired a different mechanism for motility, perhaps through horizontal gene transfer.
However, the absence of a relationship between the motor mechanism and gliding speed is observed in
Mycoplasma pulmonis, which appears to have the same gliding machinery as M. mobile (Seto et al., 2005b ),
despite its mean speed being much slower (Bredt & Radestock, 1977 ).

Aside from the curvature of the rod in relation to large-scale near-circular movement in M. genitalium and M.
testudinis, no measured dimension of the attachment organelle appeared to correlate with gliding characteristics.
Curved paths were previously indicated for M. genitalium (Pich et al., 2006 ). The curvature of these structures
is quite different in the two species, with the rod of the M. genitalium core exhibiting a shallow curvature
throughout (Fig. 4f ), and that of the M. testudinis core, when curved, being sharply bent at a single location
(Fig. 4d ). Importantly, this curvature is distinct from the bend observed in M. pneumoniae cores in cells that are
not attached to the substrate (Henderson & Jensen, 2006 ). We do not observe this bend under our conditions, as
confirmed by high-angle tilt SEM (not shown), perhaps because when attached to the substrate the attachment
organelle is under tension.

In no case did we observe cores that were twisted, as predicted by one model (Hegermann et al., 2002 ).
Although the M. pneumoniae core is clearly suggested to have a broad and a narrow side (Regula et al., 2001 ;
Hegermann et al., 2002 ; Henderson & Jensen, 2006 ), there was little variation in rod width within any
species. The inchworm model, an alternative model for core function during gliding (Henderson & Jensen, 2006
), invokes repeated contraction and expansion of the core along its length, which might have been borne out by
substantial differences in rod length of cells within a single species. However, except for M. imitans, whose rod
was difficult to measure because of obscuration by the large, irregular base, distributions of the length of the
terminal button plus the rod were not dramatic. It is possible that the degree of contraction is too small to be
measured by our techniques, but also possible that contraction does not occur. Also, observations of gliding
cells did not suggest any quantized movement, though the step size could be smaller, or the steps faster, than our
ability to detect. The consistency in the dimensions of cores within each species leads us to propose that the
electron-dense core, though an important determinant in attachment organelle biogenesis, is not especially
dynamic, and not directly involved in motility.

Cell division
Whether or not the electron-dense core has direct involvement in gliding motility, it might well play a direct role
in some other process. Its roles in normal attachment organelle formation and adhesin localization in M.
pneumoniae are well documented (Krause & Balish, 2004 ), though other core-lacking mycoplasmas such as M.
mobile (Shimizu & Miyata, 2002 ) form head-like structures that appear to function analogously to attachment
organelles. The special relationship between attachment organelle formation and DNA replication in M.
pneumoniae might highlight a role for the electron-dense core in coordination between the two events. An
increase in the amount of cellular DNA is accompanied by appearance of a second attachment organelle
adjacent to the first one in M. pneumoniae, and the distance between the two organelles increases with
continually increasing DNA levels (Seto et al., 2001 ). Furthermore, time-lapse microcinematography of
dividing M. pneumoniae cells reveals that a new attachment organelle generally appears adjacent to the old one,
and the new one anchors the cell in place while the old one glides away from it (Hasselbring et al., 2006a ),
resulting in the two structures being present at opposite poles. Finally, in M. gallisepticum, newly synthesized
DNA is specifically enriched in a biochemical fraction itself enriched for attachment organelles (Maniloff &
Quinlan, 1974 ). Although it could be artefactual, the physical interaction observed in this study between the
cellular DNA and the electron-dense core is intriguing and will be a subject of future studies.

Evolution
Clearly, the electron-dense core and the attachment organelle, as well as gliding motility, were present in the last
common ancestor of the M. pneumoniae cluster; these features were presumably conferred by acquisition of the
set of cytadherence-associated genes present in M. pneumoniae and its relatives from an unidentified source.
The considerable lengths of the M. pneumoniae and M. pirum structures appear to have arisen independently,
since the M. genitalium attachment organelle is not much longer than that of M. amphoriforme. Curvature in M.
genitalium and M. testudinis also appears to have developed independently, as discussed above. Thus, the
progenitor of these species most likely harboured a short, wide attachment organelle like that of M.
gallisepticum and M. amphoriforme. As the knob is markedly reduced only in M. pneumoniae, it was probably
prominent in the common ancestor. In addition, the trailing filament appears to be unique to M. pneumoniae and
therefore unlikely to have been present in the parent species. Finally, the presence of faster gliding only in M.
pneumoniae and especially M. testudinis suggests that slower motility was probably ancestral.

M. pneumoniae appears to be the most derived member of its cluster with respect to attachment organelle and
other ultrastructural phenotypes. Its speed is not close to that of any of its relatives, and the reduction in the size
of the distal elements of the structure is likely an adaptation to an unknown pressure. Numerous proteins
involved in gliding are apparently unique to the subclade containing M. pneumoniae and M. genitalium
(Hasselbring et al., 2006b ), and it is conceivable that some of these proteins are involved in processes special to
one or both of these organisms. Thus, extrapolation of M. pneumoniae attachment organelle-related data to other
species ought to be applied cautiously. In contrast, M. amphoriforme might be the most primitive, with its slow
speed and average dimensions; as such, it might constitute a more general model for attachment organelle
function in species of the M. pneumoniae cluster.

ACKNOWLEDGEMENTS

This work was supported by funds from Miami University. SEM was performed at the Miami University
Electron Microscopy Facility. We thank the following for strains: L. Duffy and K. Waites, University of
Alabama-Birmingham (M. genitalium); S. Kleven, University of Georgia (M. imitans); M. Brown, University of
Florida (M. testudinis); M. Davidson, Mollicutes Culture Collection, Purdue University (M. alvi, M. pirum).
Thanks also to R. Balish and E. Bridge for suggestions about the manuscript.

Edited by: C. Citti

REFERENCES
Balish, M. F. (2006). Subcellular structures of mycoplasmas. Front Biosci 11, 2017–2027.[CrossRef][Medline]
Balish, M. F. & Krause, D. C. (2005). Mycoplasma attachment organelle and cell division. In Mycoplasmas: Molecular Biology,
Pathogenicity, and Strategies for Control, pp. 189–237. Edited by A. Blanchard & G. Browning. Norfolk, UK: Horizon Bioscience.
Balish, M. F. & Krause, D. C. (2006). Mycoplasmas: a distinct cytoskeleton for wall-less bacteria. J Mol Microbiol Biotechnol 11,
244–255.[CrossRef][Medline]
Biberfeld, G. & Biberfeld, P. (1970). Ultrastructural features of Mycoplasma pneumoniae. J Bacteriol 102, 855–861.
[Abstract/Free Full Text]
Bredt, W. (1979). Motility. In The Mycoplasmas, vol. 1, pp. 141–155. Edited by M. F. Barile & S. Razin. New York: Academic
Press.
Bredt, W. & Radestock, U. (1977). Gliding motility of Mycoplasma pulmonis. J Bacteriol 130, 937–938.[Abstract/Free Full Text]
Cunha, S., Odijk, T., Süleymanoglu, E. & Woldringh, C. L. (2001). Isolation of the Escherichia coli nucleoid. Biochimie 83, 149–
154.[Medline]
Dhandayuthapani, S., Rasmussen, W. G. & Baseman, J. B. (1999). Disruption of gene mg218 of Mycoplasma genitalium through
homologous recombination leads to an adherence-deficient phenotype. Proc Natl Acad Sci U S A 96, 5227–5232.
[Abstract/Free Full Text]
Göbel, U., Speth, V. & Bredt, W. (1981). Filamentous structures in adherent Mycoplasma pneumoniae cells treated with nonionic
detergents. J Cell Biol 91, 537–543.[Abstract/Free Full Text]
Gourlay, R. N., Wyld, S. G. & Leach, R. H. (1977). Mycoplasma alvi, a new species from bovine intestinal and urogenital tracts. Int
J Syst Bacteriol 27, 86–96.[Abstract/Free Full Text]
Hall, T. A. (1999). BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT.
Nucleic Acids Symp Ser 41, 95–98.
Hasselbring, B. M. & Krause, D. C. (2007). Cytoskeletal protein P41 is required to anchor the terminal organelle of the wall-less
prokaryote Mycoplasma pneumoniae. Mol Microbiol 63, 44–53.[CrossRef][Medline]
Hasselbring, B. M., Jordan, J. L. & Krause, D. C. (2005). Mutant analysis reveals a specific requirement for protein P30 in
Mycoplasma pneumoniae gliding motility. J Bacteriol 187, 6281–6289.[Abstract/Free Full Text]
Hasselbring, B. M., Jordan, J. L., Krause, R. W. & Krause, D. C. (2006a). Terminal organelle development in the cell wall-less
bacterium Mycoplasma pneumoniae. Proc Natl Acad Sci U S A 103, 16478–16483.[Abstract/Free Full Text]
Hasselbring, B. M., Page, C. A., Sheppard, E. S. & Krause, D. C. (2006b). Transposon mutagenesis identifies genes associated
with Mycoplasma pneumoniae gliding motility. J Bacteriol 188, 6335–6345.[Abstract/Free Full Text]
Hatchel, J. M., Balish, R. S., Duley, M. L. & Balish, M. F. (2006). Ultrastructure and gliding motility of Mycoplasma
amphoriforme, a possible human respiratory pathogen. Microbiology 152, 2181–2189.[Abstract/Free Full Text]
Hegermann, J., Herrmann, R. & Mayer, F. (2002). Cytoskeletal elements in the bacterium Mycoplasma pneumoniae.
Naturwissenschaften 89, 453–458.[CrossRef][Medline]
Henderson, G. P. & Jensen, G. J. (2006). Three-dimensional structure of Mycoplasma pneumoniae's attachment organelle and a
model for its role in gliding motility. Mol Microbiol 60, 376–385.[CrossRef][Medline]
Johansson, K. E. & Pettersson, B. (2002). Taxonomy of Mollicutes. In Molecular Biology and Pathogenicity of Mycoplasmas, pp.
1–29. Edited by S. Razin & R. Herrmann. New York: Kluwer Academic/Plenum Publishers.
Jordan, J. L., Chang, H. Y., Balish, M. F., Holt, L. S., Bose, S. R., Hasselbring, B. M., Waldo, R. H., III, Krunkosky, T. M. &
Krause, D. C. (2007). Protein P200 is dispensable for Mycoplasma pneumoniae hemadsorption but not gliding motility or
colonization of differentiated bronchial epithelium. Infect Immun 75, 518–522.[Abstract/Free Full Text]
Krause, D. C. & Balish, M. F. (2004). Cellular engineering in a minimal microbe: structure and assembly of the terminal organelle of
Mycoplasma pneumoniae. Mol Microbiol 51, 917–924.[CrossRef][Medline]
Lind, K., Lindhardt, B. Ø., Schütten, H. J., Blom, J. & Christiansen, C. (1984). Serological cross-reactions between Mycoplasma
genitalium and Mycoplasma pneumoniae. J Clin Microbiol 20, 1036–1043.[Abstract/Free Full Text]
Maniloff, J. & Quinlan, D. C. (1974). Partial purification of a membrane-associated deoxyribonucleic acid complex from
Mycoplasma gallisepticum. J Bacteriol 120, 495–501.[Abstract/Free Full Text]
Meng, K. E. & Pfister, R. M. (1980). Intracellular structures of Mycoplasma pneumoniae revealed after membrane removal. J
Bacteriol 144, 390–399.[Abstract/Free Full Text]
Miyata, M. (2005). Gliding motility of mycoplasmas: the mechanism cannot be explained by current biology. In Mycoplasmas:
Molecular Biology, Pathogenicity, and Strategies for Control, pp. 137–163. Edited by A. Blanchard & G. Browning. Norfolk, UK:
Horizon Bioscience.
Nagai, R. & Miyata, M. (2006). Gliding motility of Mycoplasma mobile can occur by repeated binding to N-acetylneuraminyllactose
(sialyllactose) fixed on solid surfaces. J Bacteriol 188, 6469–6475.[Abstract/Free Full Text]
Papazisi, L., Frasca, S., Jr, Gladd, M., Liao, X., Yogev, D. & Geary, S. J. (2002). GapA and CrmA coexpression is essential for
Mycoplasma gallisepticum cytadherence and virulence. Infect Immun 70, 6839–6845.[Abstract/Free Full Text]
Perrière, G. & Gouy, M. (1996). WWW-Query: an on-line retrieval system for biological sequence banks. Biochimie 78, 364–369.
[Medline]
Pich, O. Q., Burgos, R., Ferrer-Navarro, M., Querol, E. & Piñol, J. (2006). Mycoplasma genitalium mg200 and mg386 genes are
involved in gliding motility but not in cytadherence. Mol Microbiol 60, 1509–1519.[CrossRef][Medline]
Pitcher, D. G., Windsor, D., Windsor, H., Bradbury, J. M., Yavari, C., Jensen, J. S., Ling, C. & Webster, D. (2005).
Mycoplasma amphoriforme sp. nov., isolated from a patient with chronic bronchopneumonia. Int J Syst Evol Microbiol 55, 2589–
2594.[Abstract/Free Full Text]
Regula, J. T., Boguth, G., Görg, A., Hegermann, J., Mayer, F., Frank, R. & Herrmann, R. (2001). Defining the mycoplasma
'cytoskeleton': the protein composition of the Triton X-100 insoluble fraction of the bacterium Mycoplasma pneumoniae determined
by 2-D gel electrophoresis and mass spectrometry. Microbiology 147, 1045–1057.[Abstract/Free Full Text]
Seto, S., Layh-Schmitt, G., Kenri, T. & Miyata, M. (2001). Visualization of the attachment organelle and cytadherence proteins of
Mycoplasma pneumoniae by immunofluorescence microscopy. J Bacteriol 183, 1621–1630.[Abstract/Free Full Text]
Seto, S., Kenri, T., Tomiyama, T. & Miyata, M. (2005a). Involvement of P1 adhesin in gliding motility of Mycoplasma
pneumoniae as revealed by the inhibitory effects of antibody under optimized gliding conditions. J Bacteriol 187, 1875–1877.
[Abstract/Free Full Text]
Seto, S., Uenoyama, A. & Miyata, M. (2005b). Identification of a 521-kilodalton protein (Gli521) involved in force generation or
force transmission for Mycoplasma mobile gliding. J Bacteriol 187, 3502–3510.[Abstract/Free Full Text]
Seybert, A., Herrmann, R. & Frangakis, A. S. (2006). Structural analysis of Mycoplasma pneumoniae by cryo-electron
tomography. J Struct Biol 156, 342–354.[CrossRef][Medline]
Shimizu, T. & Miyata, M. (2002). Electron microscopic studies of three gliding mycoplasmas, Mycoplasma mobile, M. pneumoniae,
and M. gallisepticum, by using the freeze-substitution technique. Curr Microbiol 44, 431–434.[CrossRef][Medline]
Tham, T. N., Ferris, S., Bahraoui, E., Canarelli, S., Montagnier, L. & Blanchard, A. (1994). Molecular characterization of the P1-
like adhesin gene from Mycoplasma pirum. J Bacteriol 176, 781–788.[Abstract/Free Full Text]
Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F. & Higgins, D. G. (1997). The CLUSTAL_X Windows interface: flexible
strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 25, 4876–4882.[Abstract/Free Full Text]
Tully, J. G., Rose, D. L., Whitcomb, R. F. & Wenzel, R. P. (1979). Enhanced isolation of Mycoplasma pneumoniae from throat
washings with a newly-modified culture medium. J Infect Dis 139, 478–482.[Medline]
Uenoyama, A., Kusumoto, A. & Miyata, M. (2004). Identification of a 349-kilodalton protein (Gli349) responsible for cytadherence
and glass binding during gliding of Mycoplasma mobile. J Bacteriol 186, 1537–1545.[Abstract/Free Full Text]
Waites, K. B. & Talkington, D. F. (2004). Mycoplasma pneumoniae and its role as a human pathogen. Clin Microbiol Rev 17, 697–
728.[Abstract/Free Full Text]
Received 28 August 2007; revised 5 October 2007; accepted 11 October 2007.