Gardnerella Vaginolysin Potentiates Glycan Molecular Mimicry by Neisseria gonorrhoeae

Abstract

Bacterial vaginosis (BV) is a dysbiotic condition of the vaginal microbiome associated with higher risk of infection by Neisseria gonorrhoeae-the cause of gonorrhea. Here we test if one known facet of BV-the presence of bacterial cytolysins-leads to mobilization of intracellular contents that enhance gonococcal virulence. We cloned and expressed recombinant vaginolysin (VLY), a cytolysin produced by the BV-associated bacterium Gardnerella, verifying that it liberates contents of cervical epithelial (HeLa) cells, while vector control preparations did not. We tested if VLY mediates a well-known gonococcal virulence mechanism-the molecular mimicry of host glycans. To evade host immunity, N. gonorrhoeae caps its lipooligosaccharide (LOS) with α2-3-linked sialic acid. For this, gonococci must scavenge a metabolite made inside host cells. Flow cytometry-based lectin-binding assays showed that gonococci exposed to vaginolysin-liberated contents of HeLa cells displayed greater sialic acid capping of their LOS. This higher level of bacterial sialylation was accompanied by increased binding of the complement regulatory protein factor H, and greater resistance to complement attack. Together these results suggest that cytolytic activities present during BV may enhance the ability of N. gonorrhoeae to capture intracellular metabolites and evade host immunity via glycan molecular mimicry.

Keywords: Gardnerella; bacterial vaginosis; complement; cytolysin; factor H; gonorrhea; sialic acid; vaginolysin.

Figure 1.

Exposure of human cells to Gardnerella VLY leads to a release of intracellular contents from RBCs and cervical epithelial cells. A, Coomassie blue SDS-PAGE gel showing recombinant purified VLY and a parallel control preparation using cells containing the empty expression vector. B, Cytolytic activity of VLY against human RBCs determined by heme release (detected as absorbance at 545 nm). The EC₅₀ was calculated using a 4-parameter nonlinear regression model. The data are a combination of 3 independent experiments, with 3 biological replicates per experiment. C, Cytolytic activity of VLY against HeLa cells determined by LDH release. The EC₅₀ was calculated using a 4-parameter nonlinear regression model. Data are a combination of 2 independent experiments, with 3 to 4 biological replicates per experiment. B and C, error bars represent standard deviation. Abbreviations: Abs, absorbance; EC₅₀, 50% maximum effective concentration; LDH, lactate dehydrogenase; RBC, red blood cell; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; VLY, vaginolysin.

Figure 2.

MAL-I binds to sialylated gonococcal LOS. A, Schematic of Neu5Ac addition to the terminal lactosamine of gonococcal LOS. MAL-I binding site is indicated. B, MAL-I binding to Neisseria gonorrhoeae F62 wild type following bacterial incubation with increasing concentrations of CMP-Neu5Ac, presented as MFI. Data are from a single experiment, with 3 biological replicates per CMP-Neu5Ac concentration. Error bars represent standard deviation. Abbreviations: CMP-Neu5Ac, cytidine-5′-monophospho-N-acetyl neuraminic acid; Gal, galactose; GlcNAc, N-acetylglucosamine; LDmanHep, L-glycero-D-manno-heptose; Kdo, 2-keto-3-deoxy-D-mannooctanoic acid; LOS, lipooligosaccharide; Lst, LOS-sialyltransferase; MAL-I, Maackia amurensis lectin I; MFI, median fluorescence intensity.

Figure 3.

Treatment of cervical epithelial cells with Gardnerella VLY potentiates Lst-dependent Neisseria gonorrhoeae sialylation. N. gonorrhoeae F62 WT and Δlst were incubated with the supernatants of HeLa cells that had been exposed to either 1 μg/mL VLY or a parallel dilution of the vector control preparation. A, Schematic of the method for Figure 3 experiments, and those shown in Figure 4C and 4D. B–D, MAL-I binding to N. gonorrhoeae F62 WT and Δlst after incubation with HeLa cell supernatants. B and C, Representative histograms of MAL-I binding to (B) F62 WT or (C) Δlst. D, MAL-I binding presented as MFI. Data are a combination of 2 independent experiments, with 3–4 biological replicates per experiment. Error bars represent standard deviation. Statistical significance was evaluated using ordinary 1-way ANOVA. *P < .05, ****P < .0001. Abbreviations: CMP, cytidine-5′-monophosphate; LOS, lipooligosaccharide; Lst, LOS-sialyltransferase; MAL-I, Maackia amurensis lectin I; MFI, median fluorescence intensity; PerCP-Cy5-5-A, flow cytometer channel detecting red light emissions; ns, not significant; VLY, vaginolysin; WT, wild type.

Figure 4.

Neisseria gonorrhoeae caps LOS galactose when exposed to VLY-liberated HeLa contents. ECA recognizes the unsialylated terminal epitope of gonococcal LOS, and therefore binding of ECA acts as a negative readout of gonococcal sialylation. A, Diagram of N. gonorrhoeae unsialylated LOS, with ECA-binding epitope indicated. B, ECA binding to N. gonorrhoeae F62 WT following bacterial incubation with increasing concentrations of CMP-Neu5Ac, presented as MFI. Data are from a single experiment, with 3 biological replicates per concentration. Error bars represent standard deviation. C and D, ECA binding after N. gonorrhoeae F62 WT was incubated with the supernatants of HeLa cells that had been exposed to either 1 μg/mL VLY or a parallel dilution of the pET28a vector-only control. C, Representative histogram of ECA binding to F62 WT. D, ECA binding presented as MFI. Data are a combination of 2 independent experiments, with 3–4 biological replicates per experiment. Error bars represent standard deviation. Significance determined by Mann-Whitney U-test. ***P < .001. Abbreviations: CMP-Neu5Ac, cytidine-5′-monophospho-N-acetyl neuraminic acid; ECA, Erythrina cristagalli lectin; Gal, galactose; LOS, lipooligosaccharide; Lst, LOS-sialyltransferase; MFI, median fluorescence intensity; PerCP-Cy5-5-A, flow cytometer channel detecting red light emissions; VLY, vaginolysin; WT, wild type.

Figure 5.

Exposure of Neisseria gonorrhoeae to VLY-liberated HeLa contents increases sialylation-dependent recruitment of human FH. A, Schematic of chimeric FH protein S2534 (Fc3/FH*). S2534 consists of human IgG3 Fc receptor (Fc3) fused to the 3 C-terminal domains 18–20 of human FH. A point mutation in FH domain 19 (D to G at position 1119), which was introduced to abrogate binding to host cells, does not interfere with binding of the C-terminus of FH to sialylated N. gonorrhoeae. B, Binding of Fc3/FH* to N. gonorrhoeae F62 WT and Δlst following incubation with either 0 μM or 1 μM CMP-Neu5Ac. Data are presented as the percentage of gated events positive for FITC. C and D, Fc3/FH* binding after N. gonorrhoeae F62 WT and Δlst were incubated with the supernatants of HeLa cells that had been exposed to either 1 μg/mL VLY or a parallel dilution of the pET28a vector-only control. C, Representative dot-plot of Fc3/FH* binding to N. gonorrhoeae WT and Δlst after exposure to VLY-extracted HeLa contents. D, Fc3/FH* binding presented as the percentage of gated events positive for FITC. B and D, Data are a combination of 2 independent experiments, with 2–3 (B) or 4–5 (D) biological replicates per experiment. Error bars represent standard deviation. Significance determined by 1-way ANOVA. ****P < .0001. Abbreviations: CMP-Neu5Ac, cytidine-5′-monophospho-N-acetyl neuraminic acid; FH, Factor H; FITC, fluorescein isothiocyanate; lst, lipooligosaccharide-sialyltransferase; ns, not significant; SSC, side scatter; VLY, vaginolysin; WT, wild type.

Figure 6.

Exposure of Neisseria gonorrhoeae to VLY-liberated HeLa contents increases sialyaltion-dependent survival in human serum. N. gonorrhoeae F62 WT and Δlst were incubated with the supernatants of HeLa cells that had been exposed to either 1 μg/mL VLY or a parallel dilution of the vector-control preparation. A and B, Survival of N. gonorrhoeae F62 WT (A) and Δlst (B) after 40-minute incubation in 4% normal human serum. Percent survival determined by colony enumeration at 0 and 40 minutes. Error bars represent standard deviation. Significance determined by Mann-Whitney U-test. ***P < .001. Abbreviations: lst, lipooligosaccharide-sialyltransferase; ns, not significant; VLY, vaginolysin; WT, wild type.

Figure 7.

Proposed model of Gardnerella VLY potentiating Neisseria gonorrhoeae virulence. Gardnerella secretes VLY, which forms pores in cervical epithelial cells. Pore formation leads to a release of intracellular contents, including CMP-Neu5Ac. N. gonorrhoeae scavenges released CMP-Neu5Ac to sialylate its LOS, which enhances its ability to evade host immunity. Abbreviations: CMP-Neu5Ac, cytidine-5′-monophospho-N-acetyl neuraminic acid; LOS, lipooligosaccharide; Lst, LOS-sialyltransferase; VLY, vaginolysin.