Persistence, immune response, and antigenic variation of Mycoplasma genitalium in an experimentally infected pig-tailed macaque (Macaca nemestrina)

Abstract

Mycoplasma genitalium is a sexually transmitted pathogen associated with several acute and chronic reproductive tract disease syndromes in men and women. To evaluate the suitability of a pig-tailed macaque model of M. genitalium infection, we inoculated a pilot animal with M. genitalium strain G37 in the uterine cervix and in salpingeal pockets generated by transplanting autologous Fallopian tube tissue subcutaneously. Viable organisms were recovered throughout the 8-week experiment in cervicovaginal specimens and up to 2 weeks postinfection in salpingeal pockets. Humoral and cervicovaginal antibodies reacting to MgpB were induced postinoculation and persisted throughout the infection. The immunodominance of the MgpB adhesin and the accumulation of mgpB sequence diversity previously observed in persistent human infections prompted us to evaluate sequence variation in this animal model. We found that after 8 weeks of infection, sequences within mgpB variable region B were replaced by novel sequences generated by reciprocal recombination with an archived variant sequence located elsewhere on the chromosome. In contrast, mgpB region B of the same inoculum propagated for 8 weeks in vitro remained unchanged. Notably, serum IgG reacted strongly with a recombinant protein spanning MgpB region B of the inoculum, while reactivity to a recombinant protein representing the week 8 variant was reduced, suggesting that antibodies were involved in the clearance of bacteria expressing the original infecting sequence. Together these results suggest that the pig-tailed macaque is a suitable model to study M. genitalium pathogenesis, antibody-mediated selection of antigenic variants in vivo, and immune escape.

Fig 1

(A) Schematic of the mgpB expression site of M. genitalium G37-C. Black boxes (B, EF, and G) indicate the variable regions of mgpB, and intervening white areas represent conserved sequences (82). Arrows connected by broken lines show the location and direction of the primers used for PCR amplification. (B) Alignment of mgpB region B and MgPar8 DNA sequences obtained from the primate inoculum (Inoc), bacteria passaged in broth for 8 weeks (In vitro), and the week 8 primate specimen (Wk8 variant) with the published M. genitalium G37 type sequences. In each set of sequences, the top four lines represent mgpB expression site sequences, while the bottom five lines represent MgPar8 sequences. Nucleotides that are identical to the nucleotides in strain G37 are indicated by dots; nucleotides that differ from the G37 type sequence are shown. Deleted nucleotides are indicated by hyphens. Two MgPar8 sequences were identified in the in vitro-propagated control and are indicated as “In vitro (1)” and “In vitro (2)” (see Results). Nucleotides are numbered relative to the first base pair of G37 mgpB or MgPar8. (C) Alignment of the predicted MgpB region B amino acid sequences showing that the primate week 8 DNA sequence predicts an in-frame protein product with variant amino acids indicated by their single-letter abbreviation. Amino acids are numbered according to the start codon of MgpB.

Fig 2

Immunoblots of primate serum antibody reactivity to M. genitalium whole-cell lysates. Bacterial lysates were electrophoresed through an 8% SDS-polyacrylamide gel, transferred to membranes, and then cut into strips for reaction with primate serum samples collected at weeks 0 to 8 after infection. (A) Serum IgG reactivity. Primate serum samples were diluted 1:5,000, reac-tivity was detected using horseradish peroxidase (HRP)-conjugated goat anti-human IgG secondary antibody diluted 1:10,000, and the membranes were exposed to film for 6 min. (B) Serum IgM reactivity. Primate serum samples were diluted 1:800, reactivity was detected using HRP-conjugated goat anti-human IgM secondary antibody diluted 1:10,000, and then the membranes were exposed to film for 4 min. (C) Serum IgA reactivity. Primate serum samples were diluted 1:200, reactivity was detected using HRP-conjugated goat anti-human IgA secondary antibody diluted 1:8,000, and then the membranes were exposed to film for 13 min. The positions of MgpB (solid black arrows), MgpC (gray dotted arrows), and HMW2 (asterisks) are indicated to the right of the immunoblots. The positions of molecular mass markers (in kDa) are shown to the left of each immunoblot.

Fig 3

Reactivity of primate serum antibodies to whole-cell lysates of wild-type M. genitalium G37-C (WT) and to mgpB (Δ191) and mgpC (Δ192) deletion mutants. (A) Serum IgG reactivity. Primate serum samples were diluted 1:5,000, reactivity was detected using HRP-conjugated goat anti-human IgG secondary antibody diluted 1:10,000 and then exposed to film for 5 min. (B) Serum IgM reactivity. Primate serum samples were diluted 1:800, and reactivity was detected using HRP-conjugated goat anti-human IgM secondary antibody diluted 1:10,000 and then exposed to film for 4 min. (C) Serum IgA reactivity. Due to the limited volume of week 8 serum samples available, the reactivity of week 2 IgA is shown. Primate serum samples were diluted 1:200, and reactivity was detected using HRP-conjugated goat anti-human IgA secondary antibody diluted 1:8,000 and then exposed to film for 20 min. The positions of MgpB (solid black arrows), MgpC (gray dotted arrows), and HMW2 (asterisks) are indicated. The positions of molecular mass markers (in kDa) are shown to the left of each immunoblot.

Fig 4

Immunoblot analysis of primate genital tract antibody reactivity to whole-cell lysates of wild-type M. genitalium. Cervical exudates collected prior to inoculation (week 0 [Wk 0]) or after 8 weeks of infection (Wk 8) were diluted 1:50 and reacted with whole-cell lysates of M. genitalium G37-C. To identify the positions of the 140-kDa MgpB and 110-kDa MgpC proteins, bacterial lysates were reacted with antibodies produced in rabbits (Rb), combined, and diluted 1:10,000. HRP-conjugated secondary antibody (anti-human IgG or anti-rabbit IgG) was diluted 1:10,000. After reaction with chemiluminescent reagents, membranes were exposed to film for 5 min or 1 min, respectively. The positions of MgpB (solid black arrows) and MgpC (gray dotted arrow) are indicated.

Fig 5

Analysis of reactivity of primate serum to recombinant wild-type and variant MgpB region B peptides. (A) Amino acid sequences of the wild-type (rMgpB:B^G37) and week 8 variant (rMgpB:B^Wk8) MgpB region B His-tagged recombinant peptides. Plasmid vector-associated amino acids, including the His tag, are shown in gray type. Amino acids that differ between the wild-type and week 8 peptides are indicated by gray background, and deleted amino acids are represented by a dash. The asterisk indicates the TGA codon that encodes stop in E. coli. Predicted linear B-cell epitopes are shown in bold type with wavy underlines; conformational B-cell epitopes for mycobacterial G37-C region B are overlined. His tag and vector-encoded amino acids were omitted from B-cell epitope predictions. (B) (Left) Coomassie blue stain of purified recombinant peptides. (Right) Reactivity of rabbit anti-MgpB region B (α-rMgpB:B) antibodies to purified recombinant peptides. Rabbit serum samples from animals immunized with a recombinant MgpB region B peptide consisting of amino acids 185 to 352 (Iverson-Cabral and Totten, unpublished data) were used at 1:1,000,000 and detected with HRP-conjugated goat anti-rabbit IgG secondary antibody diluted 1:1,000. (C) ELISA reactivity of primate serum to recombinant peptides over the course of infection. Reactivity of serum to rMgpB:B^G37 (black circles) increased over time, while reactivity to rMgpB:B^Wk8 (gray circles) remained unchanged from the high background reactivity observed at weeks −2 and 0 (prior to infection). Primate serum samples were diluted 1:50 and detected with HRP-conjugated goat anti-human IgG secondary antibodies diluted 1:25,000. Data shown are averages of triplicate wells from a typical experiment repeated at least three times, and the error bars show standard errors. A no-antigen control was included for each plate, and this value was subtracted from the data shown. (D) ELISA data presented as fold increase compared to preinoculation values (average of week −2 and week 0). The graph shows the averages of five experiments, each with triplicate wells. Dark bars, wild-type peptide; light bars, week 8 variant peptide. Error bars indicate standard errors. The fold increase for reactivity to the rMgpB:B^G37-C peptide was significantly different by Student's t test for weeks 1 through 6 (P < 0.001) and for week 8 (P = 0.0035) compared to preinoculation values.