The meningococcal vaccine candidate neisserial surface protein A (NspA) binds to factor H and enhances meningococcal resistance to complement

Abstract

Complement forms an important arm of innate immunity against invasive meningococcal infections. Binding of the alternative complement pathway inhibitor factor H (fH) to fH-binding protein (fHbp) is one mechanism meningococci employ to limit complement activation on the bacterial surface. fHbp is a leading vaccine candidate against group B Neisseria meningitidis. Novel mechanisms that meningococci employ to bind fH could undermine the efficacy of fHbp-based vaccines. We observed that fHbp deletion mutants of some meningococcal strains showed residual fH binding suggesting the presence of a second receptor for fH. Ligand overlay immunoblotting using membrane fractions from one such strain showed that fH bound to a approximately 17 kD protein, identified by MALDI-TOF analysis as Neisserial surface protein A (NspA), a meningococcal vaccine candidate whose function has not been defined. Deleting nspA, in the background of fHbp deletion mutants, abrogated fH binding and mAbs against NspA blocked fH binding, confirming NspA as a fH binding molecule on intact bacteria. NspA expression levels vary among strains and expression correlated with the level of fH binding; over-expressing NspA enhanced fH binding to bacteria. Progressive truncation of the heptose (Hep) I chain of lipooligosaccharide (LOS), or sialylation of lacto-N-neotetraose LOS both increased fH binding to NspA-expressing meningococci, while expression of capsule reduced fH binding to the strains tested. Similar to fHbp, binding of NspA to fH was human-specific and occurred through fH domains 6-7. Consistent with its ability to bind fH, deleting NspA increased C3 deposition and resulted in increased complement-dependent killing. Collectively, these data identify a key complement evasion mechanism with important implications for ongoing efforts to develop meningococcal vaccines that employ fHbp as one of its components.

Figure 1. FH binding to fHbp mutants of select N. meningitidis strains and their isogenic capsule deficient mutants.

A. fH binding to wild-type meningococcal strains H44/76, A2594, BZ198 and Z2087 and their fHbp deletion mutants was examined by flow cytometry. Bacteria were incubated with purified human fH at a concentration of 20 µg/ml and bound fH was detected with polyclonal sheep anti-human fH. Representative controls with the wild-type strains where fH was omitted from the reaction mixture are shown by the broken line. The x-axis represents fluorescence on a log₁₀ scale and the y axis is the number of events. Median fluorescence is indicated to the right of each histogram. B. Capsule expression hinders fH binding to fHbp deletion meningococcal mutants. fHbp deletion mutants of encapsulated strains H44/76, A2594, BZ198 and Z2087 and their isogenic unencapsulated mutants were incubated with fH at a concentration of 20 µg/ml and bacteria-bound fH was detected with sheep anti-human fH. fH binding to the encapsulated (Cap+) strains is shown by the shaded histogram and binding to the isogenic unencapsulated (Cap−) mutant is shown by the solid line. Representative controls with the wild-type strains where fH was omitted from the reaction mixture are shown by the broken line. The x-axis represents fluorescence on a log₁₀ scale and the y axis the number of events. Median fluorescence is indicated to the right of each histogram.

Figure 2. Inverse relationship between fH binding to fHbp mutants and length of HepI glycans extensions.

A. Schematic depicting the LOS HepI glycan extensions of the strains used in this study. B. Binding of fH (10 µg/ml) increases as LOS HepI chain length decreases. fH binding to Cap+ (left panel) and Cap− (right panel) isogenic mutants of A2594 that express either unsialylated LNT LOS (blue graphs), L8 LOS (green graphs) or unsubstituted HepI LOS (red graphs) was measured by flow cytometry. Numbers represent median fluorescence of fH binding from a single representative experiment; color corresponds to the color of the graph. Median fluorescence from three independent experiments was used to perform a Cuzick's nonparametric test for trend across ordered groups. The trend of fH binding increasing as HepI chain length decreases is statistically significant (p-value for trend test p = 0.007). The control (dashed histogram) represents a reaction mixture in which fH was excluded; controls with all mutants yielded similar results and a representative control obtained using the LNT LOS-expressing strain is shown. Axes are as described for Figure 1A.

Figure 3. Sialylation of lacto-N-neotetraose (LNT) LOS enhances binding of fH to fHbp-negative meningococci.

A. fH (10 µg/ml) binding to fHbp deletion mutants of unencapsulated (Cap−) derivatives of strains A2594 (upper panel) and Z2087 (lower panel) grown either with (sia+) or without (sia−) CMP-NANA (5 µg/ml) added to the growth media was measured by flow cytometry using anti-fH mAb 90×. Binding to sialylated (sia+) strains is shown by the shaded graphs and binding to strains without LOS sialic acid (sia−) by the solid line. Controls (no fH added) are shown by the dashed line. Numbers represent the median fluorescence of the corresponding histogram. B. fH binding to fHbp− LNT-bearing meningococci is dose dependent and saturatable. Z2087 mynB fHbp (Cap−, fHbp−, LNT LOS) was grown in media containing increasing concentrations of CMP-NANA ranging from 0 to 10 µg/ml. Bacteria were incubated with fH (20 µg/ml) and fH bound to bacteria was detected by flow cytometry using mAb 90×. The average median fluorescence from 3 independent experiments is plotted. Error bars represent standard deviations. The increases in fH binding were statistically significant (p-value<0.05) for all CMP-NANA concentrations prior to saturation. Controls and axes are as described in Figure 3A.

Figure 4. Identification of NspA as a meningococcal ligand for fH.

A. Membrane extracts from strains A2594 and H44/76 and their fHbp deletion mutants (fHbp−) were separated on a 4–12% Bis-Tris gel, transferred to a PVDF membrane by western blotting and probed with pure human fH (1 µg/ml). Bound fH was detected with sheep polyclonal anti-human fH. The locations of PorA, PorB and fHbp are indicated. A ∼17 kD molecule that also bound fH is indicated by the arrow. This protein was identified by MALDI-TOF analysis of the co-migrating band on a Coomassie blue stained gel (indicated by the asterisk) as NspA (see text). B. PorA and PorB are not ligands for fH on intact meningococci. porA and porB3 were deleted from the background of A2594 Cap− (right panel) and A2594 Cap− (left panel), respectively, and fH (10 µg/ml) binding was measured by flow cytometry using anti-fH mAb 90× as the detecting Ab. The shaded graph represents fH binding to the porin deletion mutants and the solid line represents fH binding to the parent strain. Controls and axes are as described in Figure 1A.

Figure 5. Deleting NspA decreases binding of fH to N. meningitidis.

A. Expression of fHbp and NspA in BZ198 and A2594 derivatives as determined by western blotting of whole cell lysates followed by detection with polyclonal anti-fHbp (variant 1,2 and 3) or anti-NspA mAb Me-7 as indicated. After transfer, proteins migrating above ∼50 kD were stained with Coomassie blue and served as a loading control, proteins migrating between ∼20 kD and 40 kD were probed to detect fHbp and proteins migrating below 20 kD were probed to detect NspA. NspA migrates with an apparent molecular mass of approximately 15 kD when 4–12% Bis-Tris gels are used with MES running buffer. Also, of note, NspA is a heat-modifiable protein and the second larger anti-NspA-reactive band seen in some lanes is the result if incomplete heat denaturation. B. Strains BZ198 Cap+/LNT sia+, BZ198 Cap+/L8 LOS and A2594 Cap−/L8 LOS, and their fHbp, nspA or fHbp nspA double mutants were examined for their ability to bind to fH (20 µg/ml) by flow cytometry. The boxed numbers accompanying each histogram represents the median fluorescence of fH binding of the entire bacterial population. Controls (shown by the broken lines) represent fluorescence where fH was omitted from the reaction mixture; all strains yielded similar background binding and, for simplicity, only tracings obtained with the parent strains have been shown. Axes are as described for Figure 1A.

Figure 6. Recombinant NspA expressed in E. coli vesicles binds to fH.

A. Binding of fH to microvesicles prepared from an E. coli strain expressing recombinant NspA (squares) or to vesicles prepared from the same E. coli strain transformed with the plasmid without the nspA gene (circles) was measured by ELISA. Binding of fH to vesicles harboring recombinant NspA was blocked by the anti-NspA mAb 14C7 (triangles). Each data point represents the arithmetic mean of the OD_405nm reading from three independent experiments and error bars represent the standard deviation. B. Binding of fH to A2594 Cap− fHbp− LNT LOS sia− was measured in the presence of anti-NspA mAb Me-7 (purple shaded histogram), anti-NspA mAb 14C7 (pink shaded histogram) or anti-PorA mAb P1.9 (green shaded histogram); all mAbs were used at a concentration of 30 µg/ml and fH binding was detected using sheep anti-human fH. The control (histogram depicted by a dashed line) represents bacteria incubated with mAb Me-7, followed by addition of anti-fH and anti-sheep IgG FITC. A control where fH binding was measured in the absence of any added mAb is shown by the grey shaded histogram in the left panel. Surface binding of each mAb (shading as described above) to A2594 Cap− fHbp− LNT LOS sia− is also shown. The control (dashed line) represents bacteria incubated with anti-mouse IgG FITC. Median fluorescence is indicated to the right of each histogram.

Figure 7. fH binding increases with increasing NspA expression.

A. Comparison of NspA expression in Cap− derivatives of strains A2594, Z2087, H44/76, C2120, W171, Y2220 and BZ198 by western blotting of whole cell lysates followed by detection with anti-NspA mAb Me-7. After transfer, proteins migrating above 50 kD were stained with Coomassie blue and served as a loading control. B. Comparison of fHbp expression in strains A2594, Z2087, H44/76, BZ198 and Y2220 by western blotting of whole cell lysates followed by detection with polyclonal anti-variant 1, 2 and 3 fHbp. After transfer, proteins migrating above 40 kD were stained with Coomassie blue and served as a loading control. C. Overexpression of NspA. The porA promoter was used to increase NspA expression in the backgrounds of Y2220 Cap+/LNT sia+ and Y2220 Cap+/L8 LOS. Bacterial lysates were subject to western blotting and probed with mAb Me-7. Similar loading of parent and mutant strains was confirmed by Coomassie staining as described in A. D. Overexpression of NspA enhances fH binding to Y2220 Cap+/LNT sia+ and Y2220 Cap+/L8 LOS. fH binding to the parent strain, expressing wildtype levels of NspA, is shown by the solid line and binding to the NspA overexpressing isogenic mutant is shown by the grey shaded histogram. A representative control (no added fH) is shown by the dashed line. Axes are as described in Figure 1A.

Figure 8. Binding of fH/Fc fusion proteins to N. meningitidis fHbp and NspA mutants of A2594.

Binding of fH/Fc fusion constructs to A2594 Cap− L8 LOS and to its fHbp−, NspA− and fHbp− NspA− double mutants was assessed by flow cytometry using anti-mouse IgG FITC to disclosure the bound constructs. Strains that expressed either fHbp, NspA or both bound only to fH/Fc fusion constructs that contained SCR 6 and SCR 7 (SCR 1–7/Fc and SCR 6–10/Fc) of fH. In all graphs, the x-axis represents fluorescence on a log₁₀ scale and the y-axis represents the number of events. No fusion protein is present in the control tube. One representative experiment of at least three independent experiments is shown.

Figure 9. Meningococcal NspA binds selectively to human fH.

A. Unencapsulated N. meningitidis A2594 expressing L8 LOS and its fHbp−, NspA−, and fHbp−NspA− isogenic mutants were incubated with 10% (v/v) heat-inactivated human, chimpanzee and rhesus sera and Western blots were performed using polyclonal goat anti- human fH. B. Human, chimpanzee and rhesus serum controls at a dilution of 1/200 (v/v) were performed to ascertain that the goat polyclonal anti-human fH Abs recognized all primate fH tested.

Figure 10. NspA expression enhances resistance of meningococci to complement-dependent killing and limits C3 deposition on bacteria.

A. Strains BZ198 Cap+/L8 LOS and A2594 Cap−/L8 LOS and their isogenic mutant derivatives that lacked fHbp, NspA, or both fHbp and NspA were tested for their ability to resist killing by normal human serum in a serum bactericidal assay. The y-axis represents percent survival. Error bars indicate standard deviation calculated from 3 independent experiments. In all cases the decreased survival observed in strains that lack NspA was statistically significant (P<0.02 by a t-test) compared to the parent strain expressing NspA and fHbp B. C3 deposition on strains BZ198 Cap+/L8 LOS and A2594 Cap−/L8 LOS and their isogenic mutants that lacked either fHbp or NspA expression. The BZ198 Cap+ mutants were incubated with 40% NHS while the A2594 Cap− mutants were incubated with 20% NHS. C3 deposition on bacteria was detected by flow cytometry. Axes are as described for Figure 1A. Data with the double fHbp−/NspA− mutant was similar to the NspA− mutant and has been omitted for ease of visualization.

Figure 11. Complementation of fHbp NspA double mutants with NspA restores fH binding and serum resistance.

A. Meningococcal strain A2594 Cap−/L8 LOS (“parent”), its fHbp nspA double mutant (fHbp− NspA−) and its fHbp nspA double mutant complemented with NspA (fHbp− NspA−/NspA_A2594 comp) were examined for their ability to bind to MAb 14C7 (anti-NspA) and to fH (20 µg/ml) by flow cytometry. The boxed numbers accompanying each histogram represents the median fluorescence of 14C7 or fH binding to the entire bacterial population. Controls (shown by the broken lines) represent fluorescence where either mAb 14C7 or fH or was omitted from the reaction mixture. Axes are as described for Figure 1A. B. Strain A2594 Cap−/L8 LOS (fHbp+ NspA+), its fHbp nspA double mutant (fHbp− NspA−) and its fHbp nspA double mutant complemented with NspA (fHbp− NspA−/NspA_A2594 comp) were tested for their ability to resist killing by NHS at concentrations of 3% (left graph) or 1.5% (right graph) in a serum bactericidal assay. The y-axis represents percent survival. Error bars indicate standard deviation calculated from 3 independent experiments. With 1.5% NHS the decreased survival observed in strains that lack NspA was statistically significant (P<0.02 by a t-test) compared to the parent strain expressing both NspA and fHbp.