Mycoplasma pulmonisVsa proteins and polysaccharide modulate adherence to pulmonary epithelial cells


  • Jeffrey R. Bolland,

    1. Department of Microbiology, The University of Alabama at Birmingham, Birmingham, AL, USA
    Current affiliation:
    1. U.S. Department of Agriculture, Agricultural Research Service, Crop Diseases, Pests, and Genetics Unit, Parlier, CA, USA
    Search for more papers by this author
  • Kevin Dybvig

    Corresponding author
    1. Department of Genetics, The University of Alabama at Birmingham, Birmingham, AL, USA
    • Department of Microbiology, The University of Alabama at Birmingham, Birmingham, AL, USA
    Search for more papers by this author

Correspondence: Kevin Dybvig, Department of Genetics, University of Alabama at Birmingham, KAUL 720, Birmingham, AL 35294-0024, USA. Tel.: +1 205 934 9141; fax: +1 205 975 4418; e-mail:


The Mycoplasma pulmonisVsa proteins are a family of size- and phase-variable lipoproteins that shield the mycoplasmas from complement and modulate attachment to abiotic surfaces. Mycoplasmas producing a long Vsa protein hemadsorb poorly and yet are proficient at colonizing rats and mice. The effect of the length of the Vsa protein on the attachment of mycoplasmas to epithelial cells has not been previously explored. We find that independent of Vsa isotype, mycoplasmas producing a long Vsa protein with many tandem repeats adhere poorly to murine MLE-12 cells compared with mycoplasmas producing a short Vsa. We also find that mutants lacking the EPS-I polysaccharide of M. pulmonis exhibited decreased adherence to MLE-12 cells, even though it has been shown previously that such mutants have an enhanced ability to form a biofilm.


The mycoplasmas are prokaryotic pathogens of humans and other animals, distinguished by the lack of a cell wall, diminutive size, and a limited genome. They are parasitic obligate scavengers of numerous host factors required for growth. Colonization occurs predominantly at the mucosal surfaces of the genital and respiratory tracts and is a prerequisite for infection (Hu et al., 1976; Cassell et al., 1985; Razin et al., 1998; Simmons & Dybvig, 2009). Mycoplasma pulmonis is the causative agent of murine respiratory mycoplasmosis (MRM), which is among the most serious of naturally acquired diseases of rodent colonies. Exposing the upper respiratory tract of mice to M. pulmonis reveals a classic model of chronic mycoplasmal pneumonic disease, and numerous studies have utilized this model system to better elucidate the host–pathogen interactions in chronic respiratory disease caused by various species of Mycoplasma including the human pathogen M. pneumoniae (Cartner et al., 1998). The capability of M. pulmonis to attach to the pulmonary epithelium is one of the critical initial steps in the colonization of the host (Cassell et al., 1985).

The size- and phase-variable Vsa (variable surface antigen) lipoproteins influence virulence and the ability of the mycoplasma to adhere to inert surfaces and hemadsorb (Simmons & Dybvig, 2003; Simmons et al., 2007). In the mycoplasma strains used in this study, there are a suite of seven unique phase-variable Vsa isotypes; VsaA, C, E, F, G, H, and I. Isotype switching occurs when a silent vsa gene is combined with the vsa expression site by means of a site-specific DNA inversion (Shen et al., 2000). Size variation is a result of slipped-strand DNA mispairing during replication of the tandem repeat regions of the vsa gene. Mycoplasmas producing the long form of the Vsa protein, containing about 40–60 tandem repeats, attach to glass and plastic surfaces poorly, while mycoplasmas producing a short Vsa with 0–5 repeats exhibit significantly greater attachment (Simmons & Dybvig, 2003). It is thought that the innate immune response of the host exerts selection pressure for size variants. For example, exposure to complement can select for mycoplasmas producing a long Vsa protein (Simmons et al., 2004). Both long Vsa- and short Vsa-producing mycoplasmas are readily isolated from infected rats and mice (Gumulak-Smith et al., 2001; Denison et al., 2005). The possible role of the Vsa proteins in modulation of cytoadherence to epithelial cells has not previously been examined.

Bacterial polysaccharides are often virulence factors that can contribute to immune modulation, immune evasion, biofilm formation, and cellular adherence (Comstock & Kasper, 2006). In Pseudomonas aeruginosa, polysaccharides have a positive role in both biofilm formation and cellular attachment (Byrd et al., 2009). Streptococcus pneumoniae modulates the adherence to the epithelia of the upper respiratory tract through regulation of the production of its capsular polysaccharide. Reduced production of capsular polysaccharide results in a transparent colony morphology and an enhanced ability to adhere to respiratory epithelium. Cells fully producing the polysaccharide have opaque colony morphology and adhere poorly (Kim & Weiser, 1998; Magee & Yother, 2001). Daubenspeck et al. (2009) studied the EPS-I polysaccharide of M. pulmonis. EPS-I contains equimolar amounts of glucose and galactose residues, with galactose being the terminal sugar. When compared with wild-type mycoplasmas producing a Vsa protein of equivalent length and isotype, mutants that have no detectable EPS-I have an enhanced ability to form a biofilm on abiotic surfaces. Genetic complementation of the mutants restored the wild-type phenotype.

This study investigates the attachment of M. pulmonis to murine pulmonary epithelial cells. Mycoplasma pulmonis that produced a long Vsa protein was found to attach to epithelial cells less robustly than did mycoplasmas producing a short Vsa. Thus, the length of the Vsa protein has a similar effect on the adherence of the mycoplasmas to epithelial cells as it does on the ability of the mycoplasma to form a biofilm. These results are in contrast to the effect of the EPS-I polysaccharide, which has a negative effect on the ability of the mycoplasma to form a biofilm on abiotic surfaces, but a positive effect on cytoadherence.

Materials and methods

Mycoplasma strains

Mycoplasma pulmonis was cultured in mycoplasma broth (MB) and assayed for CFU on mycoplasma agar (MA; Simmons & Dybvig, 2003). Cells from 15-mL cultures were harvested by centrifugation, washed three times with 1 mL of fresh MB, and suspended in 1 mL of freezing medium (MB 80%, glycerol 20%). Cells were sonicated at 50% power with a 50% duty cycle on a Branson Sonifier 450 for 30 s to gently disrupt cell aggregates. Aliquots were frozen at −80 °C. A frozen aliquot of each strain was thawed and assayed for CFU to determine the titer of the stocks.

The strains of M. pulmonis utilized in this study are presented in Table 1 and have been described elsewhere (Simmons & Dybvig, 2003; Simmons et al., 2004; Daubenspeck et al., 2009). Strains CTG38 and CTG-R5 produce a VsaG protein with 35 tandem repeats and five tandem repeats, respectively. CT182-R40 and CT182-R3 are isogenic Vsa size variants producing a VsaA protein with 40 and three repeats, respectively. The strains VsaI-R40 and VsaI-R4 are isogenic Vsa size variants producing a VsaI containing 40 and four repeats, respectively. The strain CT-H.8 produces VsaH, which lacks a tandem repeat region (Simmons et al., 2004). The production of the EPS-I polysaccharide by the strains of mycoplasma used in this study was verified by gas chromatography as described (Daubenspeck et al., 2009). The CTG1701 and CTG1291 strains have transposon disruptions in the genes MYPU_7410 and MYPU_7420, respectively, and hence lack the EPS-I polysaccharide. Strain CTG1701-C is the complemented CTG1701 with restored EPS-I production. These mutants and the complemented mutant are described elsewhere (Daubenspeck et al., 2009).

Table 1. Summary of strains used in this study
StrainVsa typeVsa lengthEPS-I productionHemadsorption
Mean scoreSEMedianMode
CTG-R5GShort (R5)Yes0000
CTG38GLong (R35)Yes0000
CTG1291GLong (R40)No0000
CTG1701GLong (R40)No0000
CTG1701-CGLong (R40)Yes0000
CT182-R3AShort (R3)Yes3.40.344
CT182-R40ALong (R40)Yes2.10.122
CTI-R4IShort (R4)Yes3.50.244
CTI-R40ILong (R40)Yes2.40.122

Cell culture

Mouse lung epithelial cells 12 (MLE-12) is a pulmonary epithelial cell line that can form a monolayer bronchiolar and alveolar epithelium, originally isolated from pulmonary tumors in mice transgenic for a gene expressing the simian virus 40 large tumor antigen (Wikenheiser et al., 1993). Stocks of MLE-12 cells were grown to confluence in D-MEM/F-12 medium (Invitrogen) containing 2.5 mM l-glutamine, 15 mM HEPES, 0.5 mM sodium pyruvate, 1200 mg L−1 sodium bicarbonate, and 2% fetal bovine serum in a humidified atmosphere of 5% CO2/95% air at 37 °C.

Adherence assay

MLE-12 cells were grown to confluence in 12-well tissue culture plates (Corning). The cells were counted with a hemocytometer (Hausser Scientific) after trypsinizing the monolayer. Mycoplasma strains were thawed at room temperature and dispensed into each well containing MLE-12 at a multiplicity of infection (MOI) of 1 : 1 in D-MEM/F-12 medium. Plates were incubated in a humidified atmosphere of 5% CO2/95% air at 37 °C for 2.5 h. The wells were washed three times in MB that lacked supplemental serum. The wells were treated with a 0.05% trypsin/0.53 mM EDTA solution (Mediatech) for about 10 min, until the MLE-12 monolayer detached and the cells went into suspension. The suspension was then sonicated to disrupt aggregates and assayed for mycoplasmal CFU. Control experiments demonstrated that the treatment with the typsin/EDTA solution had no discernible effect on mycoplasmal CFU.

Hemadsorption assay

The mycoplasmas were grown on MA in a humidified atmosphere at 37 °C for 5–7 days as previously described (Simmons & Dybvig, 2003). MA plates with 30–120 colonies were overlaid with 3 mL of 0.5% sheep red blood cells (sRBC) in phosphate-buffered saline (PBS) and incubated for 30 min at 37 °C without agitation. The sRBC suspension was drawn off, and the plates were washed three times with 3 mL of PBS while rocking gently. The colonies were observed with a Leica dissecting microscope and scored for the level of hemadsorption. A colony was assigned a score of 0 when few or no sRBC were attached, a score of 1 when up to 25% of the surface area was covered, a score of 2 when between 25% and 50% of the surface area was covered, a score of 3 when between 50% and 75% of the surface area was covered, and a score of 4 when > 75% of the colony surface area was covered. The mean, median, and mode hemadsorption scores were determined after pooling the data from four experiments.

Statistical analysis

Statistical analysis was performed with the jmp version 8 software package (SAS Institute Inc., Cary, NC). Data were analyzed by analysis of variance followed by the Tukey post hoc test for a pairwise comparison of the means of epithelial attachment between strains, as well as hemadsorption. The CFU data were log transformed prior to analysis. All data reported as statistically significant have a P-value of < 0.001.


Cellular attachment is dependent upon Vsa length but not isotype

An evaluation was undertaken to determine whether the length or isotype of the Vsa proteins influenced attachment to MLE-12 cells. The reduced attachment of mycoplasmas producing a long VsaG protein, strain CTG38, as compared to mycoplasmas producing a short VsaG protein, strain CTG-R5, was statistically significant (Fig. 1). The same pattern existed among strains producing the other Vsa isotypes. Strains producing short Vsa proteins attached to MLE-12 cells in statistically significant higher numbers than did those strains that produced long proteins. These findings were true for strains producing a short or long VsaA protein, a short or long VsaI protein, and VsaH. There is no long form of the VsaH because this protein lacks tandem repeats (Simmons et al., 1996, 2004). There were no statistical differences observed between the Vsa isotypes examined. The only significant differences were between strains producing short and long Vsa proteins.

Figure 1.

Influence of Vsa isotype and tandem repeats on Mycoplasma pulmonis cytoadhesion. The short VsaH strain CT-H.8, the long and short VsaG strains CTG38 and CTG-R5, the long and short VsaA strains CT182-R40 and CT182-R3, and the long and short VsaI strains CTI-R40 and CTI-R4 were assayed. Data are presented as the CFU of short Vsa-producing (black box) and long Vsa-producing (gray box) mycoplasmas recovered after incubation with MLE-12 cell monolayer. Data represent three experiments performed in triplicate, n = 18 for CTG38 and CTG-R5, and n = 9 for all other strains. Means with the same number of asterisks were not significantly different from each other, but there were significant differences between means with different number of asterisks (P < 0.001).

Epithelial attachment is modulated by polysaccharide

The mutants that lack EPS-I that are available all produce a long VsaG protein. The EPS-I mutants CTG1291 and CTG1701 exhibited statistically significant reduced attachment to MLE-12 cells as compared to all strains of mycoplasma that produced the polysaccharide (Fig. 2). The complemented mutant that had restored EPS-I production, strain CTG1701-C, attached as well as did the wild-type long VsaG-producing strain CTG38. The reduced attachment of the mutants is attributed to the loss of the EPS-I polysaccharide, because the VsaG proteins produced by the mutant, wild-type, and complemented strains are indistinguishable by Western blot (Daubenspeck et al., 2009). As above, mycoplasmas producing a short VsaG exhibited statistically significant more adherence than did the strains that produced a long VsaG.

Figure 2.

Impact of EPS-I on Mycoplasma pulmonis cytoadhesion. The wild-type strains CTG38 and CTG-R5, the EPS-I mutants CTG1291 and CTG1701, and the CTG1701-C-complemented mutant were assayed. Data are presented as the CFU of adherent mycoplasmas recovered after incubation with MLE-12 cell monolayer. Data represent six experiments performed in triplicate, n = 12 for CTG1701-C and n = 18 for all other strains. Means with the same number of asterisks were not significantly different from each other, but there were significant differences between means with different number of asterisks (P < 0.001).

Differential hemadsorption among Vsa isotypes and length

Because the EPS-I mutants have an enhanced ability to form a biofilm on glass and plastic surfaces (Daubenspeck et al., 2009), the poor cytoadherence of the mutants was unexpected. Hence, hemadsorption was used as another approach to assess the adherence properties of the strains under study. Mycoplasma pulmonis colonies were scored according to the percent sRBC adsorbed (HA score). The results are shown in Table 1. The median HA scores of all strains expressing the VsaG isotype were 0, regardless of Vsa length. The median HA scores of CT182-R3, producing a short VsaA protein, and CT182-R40, producing a long VsaA, were 4 and 2, respectively. Similarly, the median HA scores of CTI-R4, short VsaI, and CTI-R40, long VsaI, were 4 and 2, respectively. The CT-H.8 median HA score was 4. Excluding the VsaG-producing strains, the median score of all the short Vsa-producing mycoplasmas was significantly greater than all the long Vsa-producing mycoplasmas (P = 0.0008). The HA assays for the CTG-R5 (n = 466), CTG38 (n = 509), CTG1291 (n = 472), CTG1701 (n = 603), and CTG1701-C (n = 524) were performed separately from the assays for CT182-R3 (n = 411), CT182-R40 (n = 641), CTI-R4 (n = 276), CTI-R40 (n = 437), and CT-H.8 (n = 385). Because the VsaG-producing strains did not hemadsorb regardless of whether EPS-I was produced, no conclusion could be reached as to a possible role of EPS-I in hemadsorption.


The Vsa proteins serve as a model for understanding the roles of highly repetitive proteins in mycoplasmal pathogenesis. The MBA proteins of ureaplasmas appear to be particularly analogous to the Vsa proteins because of the size variation that results from gain or loss of tandem repeats and the phase variation that results from DNA inversion (Zimmerman et al., 2011). Proteins that undergo extensive size variation because of gain or loss of tandem repeats has been described for other pathogenic species of Mycoplasma including M. agalactiae, M. arthritidis, M. bovis, M. hominis, and M. hyorhinis (Yogev et al., 1991; Lysnyansky et al., 1996; Zhang & Wise, 1996; Ladefoged, 2000; Tu et al., 2005; Nouvel et al., 2009), but the pathogenic significance and effect on cytoadherence of protein size variation in these species have for the most part not been studied.

Mycoplasmas producing a long Vsa protein have been shown in previous studies to be shielded from the lytic effects of complement and to have a markedly reduced ability to form a biofilm on glass and plastic surfaces (Simmons & Dybvig, 2003; Simmons et al., 2004, 2007). We extend these findings in this study to include cytoadherence. Independent of the Vsa isotype, mycoplasmas producing a long Vsa protein adhered less efficiently to epithelial cells than did mycoplasmas producing a short Vsa protein. We also found that the EPS-I polysaccharide is required for full cytoadherence. The mutants that lack the polysaccharide exhibited extremely poor adherence to MLE-12 cells, despite the fact that these mutants adhere robustly to abiotic surfaces and efficiently colonize mice in vivo (Daubenspeck et al., 2009). Cytoadherence in this system may be complex. Although it is possible that EPS-I, which is composed of glucose and galactose, serves as an adhesin, the addition of the pertinent monosaccharides galactose and glucose or the disaccharide lactose to binding assays failed to inhibit the adherence of M. pulmonis to erythrocytes (Minion et al., 1984). Nevertheless, EPS-I may be required for an initial interaction that brings the mycoplasma into close contact with host cells. The adhesins that are responsible for cytoadherence may be partially buried within the Vsa shield when the Vsa proteins are long. These adhesins would be exposed leading to efficient adherence when the Vsa proteins are short. This suggested process for adherence of the mycoplasma to host cells is reminiscent of a model proposed by Minion et al. (1984). Because of the high frequency of Vsa size variation that occurs during growth of the mycoplasma (Simmons et al., 2007), varied subpopulations of mycoplasmas would be present throughout the infection with some cells more adherent than others. The presence of the adherent population would be needed for colonization. The nonadherent population might better resist interactions with macrophages and other cells of the host immune system. Perhaps, the adherent population initially binds to the epithelium and forms a base onto which the less adherent population can interact to protect a colony of bacteria from host defenses.

Efforts to identify EPS-I mutants that produce a short Vsa protein have been unsuccessful. Thus, it cannot be ascertained whether EPS-I is required for efficient adherence when Vsa is short. No mutants that lack Vsa protein have been identified in our robust transposon library, suggesting that these proteins are essential (Dybvig et al., 2010). Past studies have concluded that M. pulmonis cells producing a short Vsa are sensitive to lysis by complement, leading to the hypothesis that the Vsa proteins form a protective shield (Simmons et al., 2004). Cultures inoculated with mycoplasmas that produce a short Vsa protein have a longer lag phase than cultures inoculated with cells producing a long Vsa but have a rapid growth rate in exponential phase and reach a high titer (Dybvig et al., 1989), suggesting an initial toxicity that is labile and from which the long Vsa can protect. EPS-I may also have a role in protection, rendering EPS-I mutants with a short Vsa protein difficult to isolate.

Because it was intriguing that EPS-I promoted cytoadherence, but inhibited biofilm formation, hemadsorption assays were utilized as an additional approach to examine interactions among the mycoplasma and host cells. Hemadsorption has often been used as an indicator for adherence to pulmonary epithelia in multiple species of mycoplasma (Hasselbring et al., 2005). The utter lack of hemadsorption that was observed for mycoplasmas that produce the VsaG isotype was totally unexpected and perplexing. In all prior studies, variation in Vsa length but not isotype resulted in phenotypic differences. For example, the Vsa isotype has no known association with any tissue tropism (Gumulak-Smith et al., 2001; Denison et al., 2005). The EPS-I mutants are currently available only in the VsaG background, thus nothing can be said about the role of EPS-I in hemadsorption. However, the remaining Vsa isotypes A, I, and H all exhibit hemadsorption profiles in concurrence with previously published data, with short Vsa-producing strains exhibiting significantly greater hemadsorption than long Vsa-producing strains (Simmons & Dybvig, 2003).

Bacterial pathogens generally produce multiple adhesins, and colonization is a complex process. The adhesins and receptors involved in the colonization of the murine host by M. pulmonis are unknown as are the precise roles of the Vsa proteins and the EPS-I polysaccharide. Cells producing a long Vsa protein or lacking EPS-I may retain the ability to colonize animals because although cytoadherence is reduced, it is not eliminated. Mycoplasmal structures resembling the towers of biofilms that develop on glass or plastic surfaces have been observed ex vivo and in vivo on the mouse trachea (Simmons & Dybvig, 2009). Thus, although mutants lacking EPS-I cytoadhere poorly, their enhanced ability to form a biofilm may be a contributing factor to their ability to efficiently colonize animals.


We thank Ping Lao, Portia Caldwell, and Warren Simmons for technical assistance. This work was supported by NIH grants AI63909 and AI64848.