A variant phenotype of Mycoplasma gallisepticum S6 was isolated from an in vitro antibody-culture system utilizing metabolism-inhibiting antibodies against the 64 kDa lipoprotein (LP64). M. gallisepticum populations grown in medium alone or medium containing normal rabbit serum maintained expression of the parental phenotype. This paper describes the identification of proteins which undergo variable expression. Several of these were integral membrane proteins, with estimated molecular masses of 91, 43, 41, 38, 37, and 18 kDa, which were expressed in the variant phenotype but not in the parental phenotype. Three proteins (LP64, p63 and p47) were expressed in the parental phenotype, but not in the variant phenotype. The data suggest that the interaction of specific immunoglobulins with target epitopes resulted in the selection of a subpopulation of organisms expressing an alternative array of membrane proteins which, lacking the target epitopes, was able to escape the metabolism-inhibiting effects of the specific antibodies.
The continued survival of a bacterium is contingent on the survival of any single cell within a bacterial population. The ability of an organism to vary the phenotypes expressed within a population of cells ensures that some percentage of cells within that population is suitable for the continued survival of that organism. Bacteria possess two basic mechanisms which underlie the control of gene expression: environmental sensing and genetic variation. As a result, phase and antigenic variation provide subpopulations with a means of evading the host immune response, adapting to a variety of environmental changes, and acquiring diversity in their tissue tropism. Due to the lack of a cell wall the mycoplasmal membrane is in direct contact with a wide range of microenvironments in the host, and is therefore critical to the function and survival of these organisms. The ability to alter the expression of membrane surface components may be of considerable importance in understanding pathogenicity.
In vitro antibody-culture systems have been applied to the study of phase and antigenic variation in several mycoplasmas, including Mycoplasma bovis, M. gallisepticum, M. hominis, and M. hyorhinis. Variants expressing an antibody-resistant phenotype were recovered from these systems when propagated in monoclonal antibodies (mAb) or polyclonal serum specific for surface components. Results obtained from the studies on M. bovis, M. gallisepticum, and M. hominis suggest that the resistant phenotypes selected by the antibody-culture systems lacked the epitopes recognized by the specific antibodies, whereas a combination of the phase and antigenic variable expression of the Vlp surface lipoproteins provided the resistance to growth-inhibiting antibodies in the M. hyorhinis variants. Considering the chronic nature of mycoplasmal infections and the existing evidence, it would appear that the constant selection of subpopulations, rather than the adaptation of an entire population to an environmental change, is responsible for the heterogeneous populations of these organisms.
We have examined phenotypic variation in M. gallisepticum S6 using antibodies against LP64  in an in vitro antibody-culture system and have identified a phenotype which expressed several variant proteins. Evidence suggests that this in vitro system relies on an antibody-mediated interaction to enable the selective survival of variant subpopulations. As an alternative to an in vivo model, which is affected by a myriad of host factors, this system provides a rapid and more controlled method for the identification and isolation of antibody-selected phenotypic variants.
Materials and methods
Organisms and culture conditions
M. gallisepticum strains S6 (gift from Bionique Laboratories, Saranac Lake, NY, USA), PG31 (ATCC), and R (gift from Dr. Stanley Kleven, University of Georgia, Athens, GA, USA) were cultured at 37°C in Frey's medium , containing 10% heat-inactivated horse serum and 5% fresh yeast extract. A clonal lineage of parental strain S6, designated S6p1, was established for this study using standard cloning procedures . When noted, subclones were isolated by the same procedure and designated S6 clonal populations. Stock cultures were stored at −70°C.
Preimmune sera from six female New Zealand White rabbits were pooled and used as negative control serum. Polyclonal rabbit serum was generated against whole-cell M. gallisepticum strain PG31 according to the procedure previously described by Forsyth et al. . The preparation and characterization of polyclonal rabbit anti-LP64 serum has been described previously . MAb MyG001  was kindly provided by Dr. David Ley, North Carolina State University, Raleigh, NC, USA. MAb K1  was kindly provided by Dr. Amer Silim, University of Montreal, Quebec, Canada.
Generation and analysis of M. gallisepticum variant populations
M. gallisepticum were propagated in an antibody-culture system based on modifications of the standard metabolism inhibition test . Rabbit serum, ascites fluid (mAb K1), and serum-free hybridoma culture supernatant (mAb MyG001) were diluted in Frey's medium at a final concentration of 1% (v/v). The modified media were passed through 0.45 μm filters prior to their use in experiments. Overnight broth cultures of M. gallisepticum strains were serially diluted in 100 μL Frey's medium or Frey's medium containing normal rabbit serum, anti-LP64 serum, mAb MyG001, or mAb K1. The cultures were incubated at 37°C until a shift in pH (a yellow-orange color) was observed. Subsequent growth of these populations was in Frey's medium in the absence of supplemental antibodies. Organisms from 1 ml of the resulting populations, representing approximately 109 color changing units (CCU), were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western immunoblot analysis.
SDS-PAGE and Western immunoblot analysis
Mycoplasma proteins were separated by SDS-PAGE according to the method of Laemmli , electrophoretically transferred to nitrocellulose membrane according to the method of Towbin et al.  and immunostained according to the procedure described by Forsyth et al.  using anti-LP64 serum, anti-M. gallisepticum PG31 serum, or mAb MyG001 followed by peroxidase-conjugated secondary antibody (goat anti-rabbit IgG or -mouse IgG). Membranes were developed with 4-chloro-1-naphthol chromogenic substrate. Immunoblots were scanned and the figures were prepared using computer software. Each figure was prepared from a single immunoblot. Lanes were arranged using computer software to aid in the presentation of the data. Correct alignment of the lanes was determined using the 64 kDa band and an invariant protein, as indicated in each figure by an asterisk. Protein molecular mass estimations were made using computer software.
Triton X-114 phase partitioning of mycoplasma proteins
Mycoplasma proteins were phase partitioned according to the method of Bordier . Briefly, organisms harvested from 25 ml of overnight broth culture were solubilized, insoluble materials were removed by centrifugation, and the soluble material was subjected to three rounds of phase partitioning. Proteins partitioning in the aqueous and detergent phases were subjected to SDS-PAGE and Western immunoblot analysis.
Metabolic labeling of mycoplasmas
Stock cultures of parental and variant M. gallisepticum S6 were diluted 1:30 in Frey's medium containing 1 mCi [9,10-3H]palmitic acid (specific activity, 50 Ci mmol−1) and incubated for 24 h at 37°C. Organisms were harvested, washed, and subjected to Triton X-114 phase partitioning and SDS-PAGE. Gels containing radiolabeled proteins were stained with Coomassie blue, dried under vacuum, and exposed to radiographic film for 3 weeks at −70°C.
Effect of normal rabbit serum, anti-LP64 serum, mAb K1, and mAb MyG001 on the phenotype of M. gallisepticum S6
When cultures containing approximately 109 CCU ml−1 were serially diluted in medium alone or medium containing anti-LP64 serum, mAb K1, mAb MyG001, or normal rabbit serum, the number of viable organisms did not vary, however, the media containing anti-LP64 serum, mAb K1, or mAb MyG001 exhibited an inhibitory effect on the rate of metabolism.
Immunoblot analysis using anti-M. gallisepticum PG31 serum demonstrated that populations grown in the presence of anti-LP64 serum, mAb K1, or mAb MyG001 possessed a variant pattern of recognition when compared to that of parental populations grown in medium alone or medium containing normal rabbit serum (Fig. 1a). These variant populations expressed several proteins, approximately 91, 43, 41, 38, 37, and 18 kDa, not recognized in the parental population (Table 1). Subsequent immunoblot analysis revealed that anti-LP64 serum did not react with these variant proteins (Fig. 1b). Immunoblot analysis of Triton X-114 phase partitioned proteins from the parental and variant populations revealed that the 91, 43, 41, 38, 37, and 18 kDa proteins partitioned within the detergent phase, indicating that they are integral membrane proteins (Fig. 2). The 91, 38, 37, and 18 kDa proteins metabolically labeled with [3H]palmitic acid (data not shown).
|Parental phenotype||Variant phenotype|
|medium alone||NRSa||anti-LP64 serumb||mAb K1c||mAb MyG001d|
|Parental phenotype||Variant phenotype|
|medium alone||NRSa||anti-LP64 serumb||mAb K1c||mAb MyG001d|
aMedium containing 1% normal rabbit serum (NRS).
bMedium containing 1% rabbit anti-LP64 serum.
cMedium containing 1% mAb K1.
dMedium containing 1% mAb MyG001.
Western blot analysis using mAb MyG001 revealed the existence of two additional variations between the parental and variant populations (Table 1). Proteins with estimated molecular masses of 47 kDa and 63 kDa were expressed in parental populations but not in variant populations (data not shown). Triton X-114 phase partitioning revealed that the 47 kDa protein partitioned into the detergent phase, while the 63 kDa protein partitioned into the aqueous phase (data not shown).
Following analysis of clones isolated from a population of parental S6p1 grown in medium containing anti-LP64 serum, clone S6c1 was selected as a representative of 12 identical variant clones. Variant clone S6c1 was subsequently maintained in broth culture in the absence of anti-LP64 serum for a total of 15 passages. Western blot analysis of 10 subclones isolated from the resulting population was performed using anti-M. gallisepticum PG31 serum (Fig. 3). Fifty percent (5 out of 10) of these subclones exhibited the variant immunoblot pattern characteristic of the low passage variant population. The remaining clones exhibited a reversion to a parental-like immunoblot pattern.
Effect of anti-LP64 serum on the phenotypes of M. gallisepticum strains PG31 and R
Western blot analysis using anti-M. gallisepticum PG31 serum demonstrated that parental populations of strain PG31 expressed proteins which were 2–3 kDa larger in molecular mass than the variant proteins recognized in strain S6 (Fig. 4, lanes 4 and 5). These strain PG31 proteins were expressed at a reduced level following growth in the presence of anti-LP64 serum (Fig. 4, lane 5). Western blot analysis using anti-M. gallisepticum PG31 serum demonstrated that populations of strain R grown in the presence of anti-LP64 serum expressed the proteins in the 64 kDa band at a reduced level (Fig. 4, lane 7). Western blot analysis using mAb MyG001 revealed that strains PG31 and R expressed the 47 kDa protein observed in parental populations of strain S6, and that populations grown in media containing anti-LP64 serum no longer expressed this protein (data not shown).
This paper describes the complement-independent, antibody-mediated selection of a variant phenotype of M. gallisepticum S6 which is resistant to the metabolism-inhibiting effects of antibodies which recognize LP64. The examination of several strains of M. gallisepticum revealed an intraspecies strain variability in the phenotypes expressed in response to these metabolism-inhibiting antibodies. The variant phenotype identified in strain S6 coordinately expresses proteins with estimated molecular masses of 91, 43, 41, 38, 37, and 18 kDa which are inversely expressed with proteins LP64, p63, and p47. With the exception of the 63 kDa protein, these proteins are integral membrane proteins, and p91, LP64, p38, p37, and p18 are lipoproteins. The reaction of mAb MyG001 with p47, p63, and LP64 indicates that these proteins possess homologous epitopes. The observation that the 63 kDa hydrophilic protein undergoes variable expression is the first report of a variable, non-integral membrane protein in M. gallisepticum. Analysis of the population which resulted from the continued growth of variant clone S6c1 in medium alone demonstrated that the variant phenotype persisted in the absence of specific antibodies and that expression of these proteins was phase variable. These results suggested that the antibodies do not act as effector molecules required for the induced expression of the variant proteins, rather that the observed phenotypic variations are the result of a reversible genetic mechanism allowing for the selection of subpopulations which evade the inhibiting effects of specific antibodies. Failure of the polyclonal anti-LP64 serum or mAb MyG001 to react with the variant proteins suggested that the epitopes recognized by the metabolism-inhibiting antibodies are absent from proteins expressed in the escape variant phenotype and that the variant proteins are not immunologically related to LP64.
The examination of M. gallisepticum strains S6, PG31, and R revealed an intraspecies strain variability in the phenotypes expressed in response to the anti-LP64 metabolism-inhibiting antibodies. Polyclonal serum against whole-cell M. gallisepticum PG31 demonstrated that the variant proteins identified in strain S6 are immunologically related to proteins expressed in strain PG31. Immunoblots using anti-M. gallisepticum PG31 serum demonstrated a variability in the phenotypes expressed by populations of strains S6, PG31, and R grown in the presence and absence of rabbit anti-LP64.
This report, as well as others, demonstrates that proteins in the 64–67 kDa region of M. gallisepticum are susceptible to phase variable expression. Markham et al.  have previously reported that a 67 kDa lipoprotein, designated pMGA, undergoes variable expression in vitro when grown in media containing pMGA-specific antibodies, resulting in the expression of 82 and 45 kDa proteins. In addition, isolates obtained from experimentally infected chickens demonstrated variable expression of a second 67 kDa lipoprotein, designated p67 . Bencina et al.  have reported that mAbs K1 and MyG001 do not react with the pMGA molecule, therefore, it remains to be determined whether pMGA- or p67-specific antibodies exhibit an additional effect on M. gallisepticum phenotype similar to that observed in this study. It may be hypothesized that the phase variability of proteins in this region plays a role in the survivability and pathogenicity of this organism. Identification and characterization of the mechanisms controlling this phenotypic variation will lend to an understanding of pathogenicity, providing important insights into future vaccine development and aid in our general understanding of the functions of this organism.
This research was supported in part by USDA Grant 94-37208-1069 and USDA Agricultural Experiment Station Grant CONSOO640.