α-2,3-sialyltransferase expression level impacts the kinetics of lipooligosaccharide sialylation, complement resistance, and the ability of Neisseria gonorrhoeae to colonize the murine genital tract

Abstract

Neisseria meningitidis and Neisseria gonorrhoeae modify the terminal lacto-N-neotetraose moiety of their lipooligosaccharide (LOS) with sialic acid. N. gonorrhoeae LOS sialylation blocks killing by complement, which is mediated at least in part by enhanced binding of the complement inhibitor factor H (FH). The role of LOS sialylation in resistance of N. meningitidis to serum killing is less well defined. Sialylation in each species is catalyzed by the enzyme LOS α-2,3-sialyltransferase (Lst). Previous studies have shown increased Lst activity in N. gonorrhoeae compared to N. meningitidis due to an ~5-fold increase in lst transcription. Using isogenic N. gonorrhoeae strains engineered to express gonococcal lst from either the N. gonorrhoeae or N. meningitidis lst promoter, we show that decreased expression of lst (driven by the N. meningitidis promoter) reduced LOS sialylation as determined by less incorporation of tritium-labeled cytidine monophospho-N-acetylneuraminic acid (CMP-NANA; the donor molecule for sialic acid). Diminished LOS sialylation resulted in reduced rates of FH binding and increased pathway activation compared to N. gonorrhoeae promoter-driven lst expression. The N. meningitidis lst promoter generated sufficient Lst to sialylate N. gonorrhoeae LOS in vivo, and the level of sialylation after 24 h in the mouse genital tract was sufficient to mediate resistance to human serum ex vivo. Despite demonstrable LOS sialylation in vivo, gonococci harboring the N. meningitidis lst promoter were outcompeted by those with the N. gonorrhoeae lst promoter during coinfection of the vaginal tract of estradiol-treated mice. These data highlight the importance of high lst expression levels for gonococcal pathogenesis.

Importance: Neisseria gonorrhoeae has become resistant to nearly every therapeutic antibiotic used and is listed as an "urgent threat" by the Centers for Disease Control and Prevention. Novel therapies are needed to combat drug-resistant N. gonorrhoeae. Gonococci express an α-2,3-sialyltransferase (Lst) that can scavenge sialic acid from the host and use it to modify lipooligosaccharide (LOS). Sialylation of gonococcal LOS converts serum-sensitive strains to serum resistance, decreases antibody binding, and combats killing by neutrophils and antimicrobial peptides. Mutant N. gonorrhoeae that lack Lst (cannot sialylate LOS) are attenuated in a mouse model. Lst expression levels differ among N. gonorrhoeae strains, and N. gonorrhoeae typically expresses more Lst than Neisseria meningitidis. Here we examined the significance of differential lst expression levels and determined that the level of LOS sialylation is critical to the ability of N. gonorrhoeae to combat the immune system and survive in an animal model. LOS sialylation may be an ideal target for novel therapies.

FIG 1

Incorporation of CMP-NANA onto LNT-LOS by Lst and the relationship of sialylation kinetics to binding of FH and MAb 3F11. (A) Incorporation of CMP-NANA by N. gonorrhoeae F62ΔD expressing lst either from its native promoter (Ng P_lst-Ng) or from the N. meningitidis 126E lst promoter (Ng P_lst-Nm) and N. meningitidis NMB SiaA (Nm NMB SiaA). Strains were grown in the presence of 2 µg/ml CMP-NANA (50:1 [wt/wt] unlabeled CMP-NANA/labeled CMP sialic acid [sialic-6-³H]), and incorporation was measured over time. Negative controls included N. gonorrhoeae Δlst lacking the sialyltransferase enzyme, Lst (Ng Δlst), and N. gonorrhoeae P_lst-Ng grown in the absence of CMP-NANA [Ng P_lst-Ng (No CMP-NANA)]. Values are the means and standard deviations of the results of triplicate measurements from three independent experiments. (B) Kinetics of FH binding in relation to LOS sialylation. Flow cytometry was used to measure binding of human FH (10 µg/ml) and MAb 3F11 (recognizes the unsialylated LNT-LOS epitope) to N. gonorrhoeae F62ΔD over time following the addition of CMP-NANA (2 µg/ml) to log-phase cultures. Bound human FH was detected with MAb 90X, and sialylation of the LNT LOS epitope was detected by loss of reactivity with MAb 3F11. The x axis represents time in minutes, and the y axis represents the median fluorescence. The analysis was conducted three times, and results from a representative experiment are shown.

FIG 2

FH binding to N. gonorrhoeae correlates with lst expression level. Binding of FH to N. gonorrhoeae F62ΔD expressing the F62 Lst amino acid sequence driven by either the F62 lst promoter (P_lst-Ng) or the N. meningitidis 126E lst promoter (P_lst-Nm) (A and B) or binding of FH to strain F62ΔD expressing either the N. gonorrhoeae Lst or the N. meningitidis Lst amino acid sequence (C) was measured by FACS analysis using human FH and affinity-purified polyclonal goat anti-human FH. Binding was measured over time (0 to 30 min as indicated) following the addition of CMP-NANA (0.5 µg/ml) to bacteria. Control reactions included (i) bacteria incubated with high concentrations of CMP-NANA (25 mg/ml) for 30 min, (ii) bacteria incubated without CMP-NANA, and (iii) bacteria that were not incubated with FH. An lst mutant (Δlst) also served as a negative control (data not shown). (A) Histograms from a representative experiment comparing P_lst-Ng (grey shaded) and P_lst-Nm (bold black line) are shown. The y axis represents the number of events (counts), and the x axis represents fluorescence on a log₁₀ scale. FH was omitted in the negative control (thin black line). Numbers alongside the histograms (e.g., the shaded box corresponds to the shaded histograms) are the median fluorescence values. Results showing binding of FH to both strains following incubation with 25 µg/ml CMP-NANA for 30 min are shown (lower left histograms). (B) Median fluorescence with standard errors of the means (SEM) of the results from 6 independent experiments comparing FH binding to P_lst-Ng (grey bars) and P_lst-Nm (black bars) are shown on the y axis. Each bar represents the mean of the median fluorescence values. *, P < 0.05; **, P < 0.01; ***, P < 0.0001 (Bonferroni’s posttest 2-way analysis of variance [ANOVA]). (C) Median fluorescence values with SEM from 5 independent experiments comparing levels of FH binding to strain F62ΔD expressing either the N. gonorrhoeae Lst (Lst_Ng) or N. meningitidis Lst (Lst_Nm) amino acid sequence are shown on the y axis. Each bar represents the mean of the median fluorescence values. ns, not significant (P > 0.05; Bonferroni’s posttest 2-way ANOVA).

FIG 3

Complement deposition (C3 and factor B) on strain F62ΔD expressing high (P_lst-Ng) and low (P_lst-Nm) levels of lst. Data represent levels of binding of total C3 and factor B to strain F62ΔD expressing lst using either the F62 lst promoter (P_lst-Ng) or the N. meningitidis 126E lst promoter (P_lst-Nm) measured by ELISA. Gonococcal strains grown in the presence or absence (“+” or “-”) of CMP-NANA (1 µg/ml) were incubated with 33.3% PNHS for 10 min, washed, and dispensed onto microtiter wells. Bound C3 and fB were detected using goat anti-human C3 (Complement Technology, Inc.) and goat anti-human factor B (Quidel), respectively. Binding to strain Δlst is shown as a negative control. The mean OD₄₀₅ values with SEM from two independent experiments performed in duplicate are plotted on the y axis. **, P < 0.01; ****, P < 0.0001 (1-way ANOVA using Tukey’s multiple comparison test).

FIG 4

Serum resistance as a function of Lst expression level. (A) In vitro sialylation and serum resistance of gonococci expressing low and high levels of lst. Data represent survival of isogenic F62ΔD expressing gonococcal Lst from either the gonococcal F62 lst promoter (P_lst-Ng) or the meningococcal 126E lst promoter (P_lst-Nm) in 33.3% PNHS following growth in the absence (0 µg/ml) or presence (0.1 or 1.0 µg/ml) of CMP-NANA. The percentages of survival at 30 min relative to time zero are plotted on the y axis. Each bar represents the mean and SEM of the results of 2 independent observations. **, P < 0.01; ***, P < 0.001 (Tukey’s posttest 1-way ANOVA). (B) In vivo sialylation and serum resistance of gonococci expressing low and high levels of lst. Bacteria harvested from the genital tract of mice infected with N. gonorrhoeae expressing lst_Ng using either the gonococcal lst promoter (P_lst-Ng) or the meningococcal lst promoter (P_lst-Nm) were tested for their ability to resist killing by 30% PNHS ex vivo. Samples from 3 infected mice in each group were pooled and tested on 2 consecutive days, 24 and 48 h postinfection. Neuraminidase removes sialic acid from LOS and was used to determine the in vivo role of LOS sialylation (23) in the serum resistance of bacteria harvested from infected mice. Bacteria harvested from mouse vaginal samples were treated with either neuraminidase (+) or buffer (−) prior to bactericidal testing. Mice infected with an lst mutant (Δlst) that lacks the enzyme required for LOS sialylation were used as a control. The percentages of survival relative to time zero are plotted on the y axis. Each bar represents the mean and SEM of 2 independent observations. ns, not significant; ****, P < 0.0001 (1-way ANOVA using Tukey’s multiple comparison test).

FIG 5

Attenuation of N. gonorrhoeae expressing Lst from the meningococcal lst promoter in the mouse vaginal colonization model. Data representing recovery of strain F62ΔD expressing lst from the N. meningitidis promoter (P_lst-Nm) relative to recovery of strain F62ΔD expressing lst from its native gonococcal promoter (P_lst-Ng) from estradiol-treated BALB/c mice (wild type or human FH transgenic) coinfected with both strains are expressed as competitive indices (CI); each point represents the CI calculated from an individual mouse. Data representing recovery of an lst mutant (Δlst) relative to that of strain F62ΔD expressing lst from its native gonococcal promoter (P_lst-Ng) are also shown. A CI of <1.0 indicates a decrease in the ratio of mutant to wild type with respect to that of the inoculum. The ratios of strain P_lst-Nm to strain P_lst-Ng and of strain Δlst to strain P_lst-Ng in the inocula were 1 and 0.85, respectively. Horizontal bars represent the geometric mean CI values; open symbols represent mice from whom at least 50 wild-type (P_lst-Ng) CFU/ml, but no mutant (P_lst-Nm or Δlst) CFU, were recovered. The limit of detection (40 CFU/ml) was used as the number of mutant CFU recovered in these cases.