Neisseria meningitidis Urethritis Outbreak Isolates Express a Novel Factor H Binding Protein Variant That Is a Potential Target of Group B-Directed Meningococcal (MenB) Vaccines

Factor H binding protein (FHbp) is an important Neisseria meningitidis virulence factor that binds a negative regulator of the alternative complement pathway, human factor H (FH). Binding of FH increases meningococcal resistance to complement-mediated killing. FHbp also is reported to prevent interaction of the antimicrobial peptide (AMP) LL-37 with the meningococcal surface and meningococcal killing. FHbp is a target of two licensed group B-directed meningococcal (MenB) vaccines.

ABSTRACT

Factor H binding protein (FHbp) is an important Neisseria meningitidis virulence factor that binds a negative regulator of the alternative complement pathway, human factor H (FH). Binding of FH increases meningococcal resistance to complement-mediated killing. FHbp also is reported to prevent interaction of the antimicrobial peptide (AMP) LL-37 with the meningococcal surface and meningococcal killing. FHbp is a target of two licensed group B-directed meningococcal (MenB) vaccines. We found a new FHbp variant, peptide allele identification no. 896 (ID 896), was highly expressed by an emerging meningococcal pathotype, the nonencapsulated urethritis clade (US_NmUC). This clade has been responsible for outbreaks of urethritis in multiple U.S. cities since 2015, other mucosal infections, and cases of invasive meningococcal disease. FHbp ID 896 is a member of the variant group 1 (subfamily B), bound protective anti-FHbp monoclonal antibodies, bound high levels of human FH, and enhanced the resistance of the clade to complement-mediated killing in low levels of human complement likely present at human mucosal surfaces. Interestingly, expression of FHbp ID 896 resulted in augmented killing of the clade by LL-37. FHbp ID 896 of the clade was recognized by antibodies elicited by FHbp in MenB vaccines.

INTRODUCTION

Neisseria meningitidis clonal complex 11 (cc11) is a hypervirulent meningococcal genetic group that often expresses serogroup C or W capsular polysaccharide; N. meningitidis cc11 is responsible globally for endemic and epidemic invasive meningococcal disease (most often meningitis or septicemia). Since 2015, clusters of N. meningitidis cc11 urethritis have emerged, primarily among heterosexual males residing in several cities in the United States (1–3). To date, over 200 cases have been confirmed, and many others are suspected (4). These infections are caused by a nonencapsulated meningococcal cc11 lineage 11.2/ET-15 clade, termed “US_Nm urethritis clade,” or “US_NmUC” (1), and this emerging meningococcal pathotype is also responsible for cases of mucosal infections (5) and cases of invasive meningococcal disease (4). The clade causing urethritis was also recently reported in the United Kingdom (6).

FHbp, encoded by fHbp, is a 27-kDa meningococcal virulence factor that binds specifically to human factor H (FH) (7–9). The binding of FHbp to FH on the meningococcal surface downregulates human complement cascade activation and bacterial lysis (10). In contrast, the homologue of FHbp in Neisseria gonorrhoeae is not surface expressed or predicted to bind FH (11), and N. gonorrhoeae instead uses porin proteins to recruit FH (12). FHbp also appears to have other functions: for example, providing resistance to killing by the antimicrobial peptide (AMP) LL-37 (13). Furthermore, FHbp is postulated to act as a siderophore binding protein and to recruit enterobactin, which may contribute to iron scavenging on mucosal surfaces (14). Based on amino acid sequence homology and cross-reactivity, FHbp is divided into three distinct variant groups (designated 1, 2 or 3) or two subfamilies (A and B) (15, 16). Variant group 1 corresponds to subfamily B, and variant groups 2 and 3 correspond to subfamily A. FHbp contains highly conserved cross-reactive conformational epitopes across the repertoire of the three naturally occurring FHbp group variants (17). FHbp is the target of two protein-based meningococcal vaccines, MenB-4C (GSK) and MenB-FHbp (Pfizer), developed to protect against invasive serogroup B meningococcal disease. MenB-4C (Bexsero) is a multicomponent vaccine containing recombinant FHbp ID 1, NadA, NHBA, and outer membrane vesicles (18). MenB-FHbp (Trumenba) contains two FHbp antigens: ID 55 is from variant group 1/subfamily B, and the other is from variant group 3/subfamily A (19).

The US_NmUC clade carries a new, unique fHbp gene (DNA allele 1127) and the corresponding protein, FHbp ID 896, belongs to variant 1 group, or subfamily B (1). To assess the potential use of MenB-4C and MenB-FHbp vaccines for control of meningococcal urethritis outbreaks caused by the clade, we investigated the unique fHbp gene and the expressed Fhbp protein. We studied the FH-binding ability of the novel FHbp variant expressed by this clade, the susceptibility of the clade to human complement-mediated killing, and the recognition by MenB vaccine-elicited anti-FHbp antibodies. We also assessed the susceptibility of the clade to the human antimicrobial peptide LL-37.

RESULTS

The fHbp locus of the 11.2 lineage urethritis isolates.

The genome sequences of 209 clade isolates reported previously (4) were used to define the fHbp locus of the clade. FHbp ID 896, first identified in the clade (1), was encoded by 205 of 209 clade isolates and was identical in the 205 isolates. Among the 4 other clade isolates, two had premature stop codons in fHbp; one isolate each expressed FHbp ID 456 or ID 915, which have 7 amino acid differences and 1 amino acid difference from ID 896, respectively. Querying the >28,000 genomes of N. meningitidis in the PubMLST database with FHbp ID 896 revealed that all isolates with FHbp ID 896 except one (an invasive, nongroupable cc11 2014 isolate, M29650) belong to the clade. Thus, 98% of the clade isolates contained the predicted FHbp ID 896 allele, and the allele is unique for the clade.

FHbp ID 896 contains 255 residues and appears to be a hybrid of ID 1 (94.5% identity) and ID 10 (98% identity), with 14 and 5 residue differences, respectively (Fig. 1). The differences include a unique amino acid at C-terminal residue 247 of the mature protein (Q instead of R for ID 1 and H for ID 10). A total of 21 residues of FHbp were identified to interact with FH by the cocrystal structure of FH and FHbp (20), and an additional 5 residues were identified via altered interaction with FH upon substitution (21, 22). Among these, FHbp ID 896 has D119H, D197Y, and K241E changes, as in ID 10, which previously have been associated with reduced FH binding (23). Interestingly, as an H248L or H248A mutation in the ID 1 FHbp eliminates FH binding (21, 22), the adjacent Q247 residue unique to ID896 might potentially influence FH interaction. Seventeen residues near the C terminus of FHbp ID 1 constituted the binding interface with a vaccine-elicited human antibody, 1A12 (17). All but one of these 17 resides are conserved in ID 896 (Fig. 1A). Alignment with FHbp ID 55, which is the variant group 1/subfamily B FHbp included in MenB-FHbp, revealed additional sequence divergence from ID 896 of the clade (Fig. 1A). Mapping of the functional residues onto the FHbp ID 1 crystal structure shows the interaction interfaces utilized by FH are on the opposite side of those residues interacting with enterobactin and monoclonal antibody (Mab) 1A12 (Fig. 1B). The Q247 residue unique to ID 896 is situated near the center of the FH-interacting surface and atop the C-terminal β-barrel domain of FHbp. A phylogenetic tree of all 734 peptide alleles belonging to FHbp variant 1/subfamily B group presents, in a global context, the relationship of ID 896 to ID 1, 10 and 55 and within this diverse variant group (Fig. 1C).

FIG 1.

(A) Alignment of the protein sequence of FHbp ID 896 with ID 1, 10, and 55. All four variants are assigned to variant group 1/subfamily B. ID 1 is the FHbp present in the MenB-4C vaccine, and ID 55 is the subfamily B variant in the MenB-FHbp vaccine. Residues differing from those of ID 1 are highlighted in gray. The N-terminal 5-amino-acid (aa) insertion of ID 55 is not included in the alignment in order to maintain the amino acid sequence numbering of ID 1. The residues involved in the interaction with FH, as shown by FHbp-FH cocrystal structure (20), are highlighted in yellow on the ID 1 sequence, while five additional residues identified by altered FH binding upon substitution (20, 22, 62) are highlighted in orange. Three FH-interacting residues of ID 896 that are different from those of ID 1 are marked by red asterisks above. The residues that form the vaccine-elicited human antibody 1A12 epitope (17) are indicated by blue asterisks. Residues forming the enterobactin binding site (14) are marked by solid red circles. (B) Residues highlighted in the multialignment (A) are mapped onto the crystal structure of FHbp (PDB ID 3KVD) (66) with the same color scheme (FH binding in yellow, enterobactin binding in red, and MAb 1A12 binding in blue), except the FH-binding residues different between ID 894 and ID 1 (H119, K241, and D197) are colored in magenta with the side chain displayed. The side chain of the unique R247 residue is shown in red. The rendition was generated by EZMol interface (67). (C) Unrooted neighbor-joining phylogenetic tree without distance corrections of 734 alleles belonging to variant 1 group. The locations of ID 1, 10, 55, and 896 within the tree are highlighted with arrows. The branches of nine alleles with greater sequence diversity are truncated in this view. The peptide sequences were retrieved from PubMLST (65).

The fHbp gene of the US_NmUC clade isolates is found in the typical meningococcal genome location adjacent to cbbA, encoding a fructose-1, 6-bisphosphate aldolase, and a gene encoding a hypothetical protein (Fig. 2A). A 181-bp insertion element positioned downstream of the −10 element in isolates belonging to clonal complexes cc1, cc4, and cc5 (24) was not present in the clade. While both monocistronic fHbp and bicistronic cbbA fHbp transcripts have been demonstrated in a cc32 strain, MC58 (25), a study by Sanders et al. (24) suggested that the bicistronic transcriptional coupling is restricted to isolates of cc32 and absent in cc11 isolates. To determine whether cotranscription occurred in the clade isolate CNM3, we performed reverse transcription-PCR (RT-PCR) analysis on total RNA from MC58 and CNM3 across the upstream intergenic region using cDNA generated with a specific reverse primer located in the 3′ end of the fHbp coding region. This analysis resulted in amplified products of the cbbA transcript and across the intergenic region only in the presence of reverse transcriptase (Fig. 2B), confirming that the cc11 urethritis clade isolate expressed bicistronic transcripts. The bicistronic transcript was also detected in another cc11 strain, M11_240294.

FIG 2.

(A) Schematic of the N. meningitidis urethritis clade fHbp locus. The cbbA and fHbp promoters responsible for the bicistronic and monocistronic transcripts (dotted arrows) are indicated with bent arrows. The length of the intergenic region is shown between genes, and the PCR products of transcription linkage analysis are marked below: the fHbp coding sequence is indicated by “A” (646 bp), the ccbA-fHbp intergenic region by “B” (231 bp), and the cbbA coding sequence by “C” (667 bp). Primers are shown above the fragment designations. (B) RT-PCR transcription linkage analysis of a cc32 strain, MC58, and cc11 strains, CNM3 and M11_240294. Shown are PCR products for A in lanes 1 to 3, B in lanes 4 to 6, and C in lanes 7 to 9. Lanes 1, 4, and 7 are with reverse transcriptase, lanes 2, 5, and 8 are without reverse transcriptase, and lanes 3, 6, and 9 are the genomic DNA control. (C) Alignment of the fHbp promoter sequence as defined by Biagini et al. (26). Shown are promoter sequences for US_NmUC isolate CNM3 (FHbp ID 896) and, for comparison, promoter clade I sequence for H44/76 (FHbp ID 1). Nucleotides that differ from the promoter clade I sequence are shaded in black. The GTG start codon of fHbp is highlighted in blue, the −10 and −35 promoter elements in yellow, and the FNR binding motif in green.

The fHbp promoter sequence of the clade resembled the promoter clade I sequence as defined by Biagini et al. (26), which exhibited higher levels of FHbp expression. There are five nucleotide differences from the strain H44/76 promoter, and they are scattered within the 184-bp sequence used to define eight major promoter classes (26). The −10 and −35 promoter elements and the previously defined fumarate and nitrate reductase regulatory (FNR) binding motif (25) are conserved in all clade isolates (Fig. 2C). FNR is a transcriptional activator protein required for the switch from an aerobic to anaerobic metabolism (25, 27).

FHbp expression.

Expression of FHbp ID 896 in the US_NmUC clade isolate CNM3 was examined by Western blotting using anti-FHbp ID 1 polyclonal serum (Fig. 3) and by selected reaction monitoring and tandem mass spectrometry (SRM-MS) (26) (Table 1). Four additional US_NmUC clade isolates from three different U.S. cities collected in 2015 and 2016 were also examined analogously to evaluate the general trend of FHbp expression in the clade. All isolates had comparable expression of FHbp when they were grown either on chocolate agar plates or in Muller-Hinton (MH) liquid cultures. The mean expression of FHbp for these 5 clade isolates when grown on chocolate agar was 314 pg/μg total extract (TE) (range, 245 to 347 pg/μg), and that in liquid cultures was 255 pg/μg TE (range, 238 to 294 pg/μg). The numbers of molecules of predicted FHbp protein expressed per cell for the tested clade isolates were 1,752 on chocolate agar and 1,423 in MH broth. These expression results were compatible with moderate to high levels of FHbp detected in other FHbp variant group 1 isolates, which generally express higher levels of FHbp than variant groups 2 and 3 (26). For comparison, the known high-FHbp-expressing strain H44/76 when grown on chocolate agar yielded 722 pg/μg TE and 4,031 molecules of FHbp per cell (26).

FIG 3.

Western blot analyses of FHbp expression in five US_NmUC clade isolates from different U.S. sites. S, total lysate samples prepared from chocolate agar plate cultures; L, total lysate samples prepared from MH broth cultures; C, recombinant ID 1 FHbp protein. The Coomassie-stained gel is shown below for the loading control.

TABLE 1.

FHbp quantification by selected reaction monitoring-mass spectrometry from total extract^a

Strain and culture	fmol/μg TEb	SD	% CV	pg/μg TE	No. of FHbp molecules/cell
CNM3
S	23.92	1.5	6.1	322.46	1,801
L	18.09	2.1	13.1	243.85	1,362
CNM36
S (ND)c
L	17.66	0.9	5.1	238.09	1,330
ATL2
S	25.34	0.6	2.4	341.53	1,907
L	21.83	5.3	24.2	294.26	1,643
ATL98
S	18.14	0.3	1.5	244.48	1,365
L	17.76	2.9	16.4	239.37	1,337
IND414
S	25.72	4.6	17.8	346.71	1,936
L	19.16	3.2	16.6	298.23	1,442

^aS, plate culture; L, broth culture; TE, total extract; CV, coefficient of variation.

^bMean values from three technical replicates.

^cND, not determined.

Surface exposure of FHbp was further confirmed by flow cytometry using live bacteria of the representative nonencapsulated clade isolate CNM3 and its fHbp (ΔfHbp::aphA3) mutant, CNM3f. The encapsulated strain with high FHbp ID 1 expression, H44/76, was included for comparison. Both CNM3 and CNM3f, as predicted, expressed the PorA P1.5 epitope, while H44/76 expressed PorA P1.7 (Fig. 4A and B). Two anti-FHbp MAbs, JAR 5 (IgG2b, reactive with subfamily B) (28) and JAR 41 (IgG1, reactive with subfamily A and B) (29), were tested at 2 and 10 μg/ml. (Histograms with 2 μg/ml are shown in Fig. 4C and D, respectively.) A third anti-FHbp Mab, JAR 13, specific for FHbp subfamily A (29), was tested at 10 μg/ml as a negative control (Fig. 4E). For H44/76, there were similar median fluorescence intensity (MFI) values between 10 and 2 μg/ml for JAR 5 (6,201 and 7,241, respectively) and for JAR 41 (13,324 and 13,814, respectively). For CNM3, the MFI values for binding of JAR 5 at these two MAb concentrations were 3,978 and 6,003, and for JAR 41, the values were 6,617 and 6,960, respectively. Binding of both of these monoclonal antibodies was eliminated in the CNM3f mutant (MFI of 69 and 65 for JAR 5 and JAR 41 at 10 μg/ml, respectively, compared to an MFI of 69 for the secondary-antibody-only negative control). The negative-control JAR 13 MAb did not bind to either CNM3 (MFI of 82) or H44/76 (MFI of 147). Thus, for both anti-subfamily B JAR 5 and anti-subfamily A/B JAR 41, CNM3 gave within 2-fold binding to the control encapsulated group B H44/76 strain, which is known to be a high expresser of FHbp ID 1. Collectively, these data indicate that CNM3 expresses FHbp at moderate to high levels, comparable to expression in other variant group 1 isolates.

FIG 4.

Binding of anti-PorA and anti-FHbp MAbs as determined by FACS analysis. Anti-PorA P1.5 (A) and P1.7 (B) MAbs confirmed the corresponding PorA expression in CNM3, the CNM3f mutant, and H44/76, respectively. Histograms of the JAR5 MAb (reactive with subfamily B) (C) and JAR 41 MAb (reactive with subfamily A and B) (D) tested at 2 μg/ml and JAR13 MAb (negative control, reactive with subfamily A) at 10 μg/ml (E) are shown. In each panel, H44/76 is shown as a dotted black line, CNM3 as a solid blue line, and the CNM3f mutant as a solid orange line. The secondary-antibody-only samples (gray) served as the negative control.

Detection of human FH binding to the surface of live N. meningitidis.

The ability of a meningococcal strain to bind FH is influenced by the expression level of FHbp and differences in the FHbp protein sequence (23). As noted, the clade FHbp ID 896 shares 94.8 and 98% identity with ID 1 and ID 10, respectively, and has a distinct amino acid at residue 247. Thus, we evaluated the ability of live bacteria of strain CNM3 to interact with human FH by flow cytometry in comparison to a serogroup B strain, H44/76, a high ID 1 expresser (Fig. 5). CNM3 bound significant amounts of human FH at three concentrations. Additional meningococcal surface proteins and moieties that bind FH have been identified, including NspA, whose FH binding is enhanced in unencapsulated meningococci (30, 31). As there was minimal association of FH with the CNM3f fHbp mutant, we examined the levels of NspA expression in CNM3 and CNM3f compared to that in H44/76 (Fig. 5E), using flow cytometry and a mixture of two anti-NspA MAbs. Antibody binding to NspA was slightly higher in both CNM3 and CNM3f than in H44/76 (MFI values of 11,085, 10,313, and 6,885, respectively). Despite the high levels of NspA expression, no significant residual FH binding for the CNM3f mutant was observed.

FIG 5.

Binding of FH to the surface of live meningococcal strains. Three FH concentrations were examined: 50 (A), 10 (B), and 2 μg/ml (C). In each of the histograms, H44/76 is shown as a dotted black line, CNM3 as a solid blue line, and the CNM3f mutant as a solid orange line. The secondary-antibody-only samples are shown in gray. (D) The corresponding units of median fluorescence intensity (MFI) of panels A to C are plotted. (E) Histogram of NspA expression tested with a mixture of two MAbs at 10 μg/ml (14C7 and AL12 at 5 μg/ml each). In the histogram, H44/76 is in black, CNM3 is in blue, the CNM3f mutant is in orange, and the secondary-antibody-only samples are in gray. The MFI values in the absence of FH are 170, 162, and 100 for CNM3, CNM3f, and H44/76, respectively.

FHbp of clade isolates and susceptibility to normal human serum.

As the clade is unencapsulated, sensitivity to killing by normal human serum (NHS) was anticipated. We first examined the relative roles of complement and naturally acquired antibody in killing of the clade by NHS. In 5 to 40% of NHS in which IgG and IgM have been depleted, CNM3 was not killed when treated for 1 h (Fig. 6A). Next we added human complement (with IgG- and IgM-depleted NHS at 30% as the source) to test dilutions of 4 unvaccinated healthy adult sera in which complement activity had been heat inactivated. As shown in Fig. 6B, subject 1 had no bactericidal antibodies, while the other three subjects displayed similar levels of killing at all serum dilutions, and complete killing of CNM3 was observed at the 10% serum dilution.

FIG 6.

Serum bactericidal activity (SBA) against the clade isolate CNM3 tested with human complement. (A) Survival of CNM3 incubated for 60 min at 37°C with a 5 to 40% commercial IgG- and IgM-depleted normal human serum pool (PelFreez). This NHS pool was used at 30% as the complement source in panels B, C, and D. Percentages of survival are normalized to CFU/ml of bacteria incubated for 60 min with heat-inactivated complement controls. Data are the mean ± range from duplicate determinations. (B) SBA of sera from four unvaccinated healthy adults. Dilutions of heat-inactivated sera were tested with 30% complement. (C) SBA using serum raised against recombinant FHbp ID 1 (rFHbp [inverted red triangles]) and the aluminum hydroxide-only control (△). Serum from immunized mice obtained 2 weeks post-dose 3 in a previous study (64) were examined. Controls are also shown for two serum pools, each from 2 mice immunized in the same study with aluminum hydroxide without an antigen. (D) SBA of serum from MenB-FHbp vaccination. Tested sera (blue circles) were from two serum pools (5 mice per pool) obtained 2 weeks post-dose 3. The negative-control serum (△) was from mice immunized in the same study with aluminum hydroxide without an antigen. Sera tested in panels B, C, and D were assayed in parallel, and the results are expressed as the mean percentage of survival ± range. (E) Survival of the clade isolate CNM3 (■) and its fHbp mutant, CNM3f (○), in 6% (dotted line) or 10% (solid line) pooled normal human sera (NHS) over the course of 90 min. The percentage of survival was calculated by dividing the viable CFU at the respected time point by the counts at T = 0 (100%). Shown are representative data done in duplicate. (F) The wild type, CNM3 (■), and its fHbp mutant, CNM3f (○), were incubated with various concentrations of NHS for 60 min. The viable CFU counts at T = 0 were set as 100% for normalization. Data are the means and standard deviations from a representative experiment performed in triplicate. Asterisks indicate P < 0.05 by Student's t test.

Given the high sequence homology of FHbp ID 896 to ID 1 and the fact that ID 896 is recognized by monoclonal antibodies that bind to key epitopes that contribute to anti-FHbp protective activity, it was anticipated that the clade would be killed by antibodies raised against subfamily B FHbp variants in meningococcal vaccines (FHbp ID 1 in MenB-4C and ID 55 in MenB-FHbp). Thus, survival of CNM3 was examined with murine immune sera raised against recombinant FHbp ID 1 (Fig. 6C) and murine sera obtained post-MenB-FHbp vaccination (Fig. 6D) using the same complement source as in Fig. 6A and B. Two independent serum pools for each antigen were examined. Both types of sera killed CNM3 proficiently, with the serum against FHbp ID 1 displaying a significantly higher bactericidal titer than the MenB-FHbp-vaccinated sera.

At mucosal surfaces, complement component levels may be at 10% or less of serum (32). Thus, we further compared CNM3 and its fHbp mutant, CNM3f, to confirm the contribution of FHbp to survival at low levels of NHS. As shown in Fig. 6E, CNM3 survived in 6% commercial pooled NHS over 90 min, but the CNM3f mutant was rapidly killed, and its survival was reduced 10-fold over the time course of the assay. At 10% NHS concentrations, both CNM3 and the CNM3f mutant were efficiently killed. There was no killing with 10% NHS in which complement activity had been heat inactivated. A titration experiment with a 30-min incubation showed that the CNM3f mutant was more sensitive to NHS than its parental strain across 6 to 10% serum concentrations (Fig. 6F). Of note, two other clade isolates, M27423 and M38873, which caused invasive meningococcal disease (4), were also tested. Both isolates were completely killed by 10% NHS, similar to the finding observed for CNM3, which was recovered from a case of urethritis.

A deletion of fHbp enhanced survival in the presence of antimicrobial peptide LL-37.

FHbp is reported to provide protection against host antimicrobial peptide LL-37 in encapsulated serogroup B strains expressing FHbp ID 1 or ID 10 (13). We examined whether FHbp ID 896, which resembles a hybrid of these two variants, affected LL-37 sensitivity of CNM3. The CNM3f mutant and the parental CNM3 isolates were treated with or without 20 μM LL-37, and viable CFU counts were determined after 30, 60, and 90 min of incubation (Fig. 7A). Surprisingly, the CNM3f mutant displayed enhanced survival compared to CNM3. A 30-min incubation, which showed no change in viability, was then selected to conduct a dose-dependent killing assay. Again, the CNM3f mutant was more resistant to LL-37 at 2.5 to 10 μM concentrations (Fig. 7B).

FIG 7.

(A) Survival of the wild-type parental strain, CNM3 (■), and its fHbp mutant, CNM3f (○), in the absence (solid line) or presence of the antimicrobial peptide LL-37 (20 μM [dotted line]) over the course of 90 min. The viable CFU counts were normalized to the counts at T = 0, which were set as 100%. (B) Dose-dependent killing of strains CNM3 (■) and CNM3f (○) by LL-37 with 30 min of incubation. The survival CFU counts were normalized to the counts of samples without peptide after 30 min of incubation. The data shown are the averages and standard deviations from a representative of multiple independent assays performed. Asterisks indicate P < 0.05 by Student’s t test.

DISCUSSION

Recent multicity outbreaks in the United States of N. meningitidis urethritis and other mucosal infections were caused by a distinct nonencapsulated meningococcal cc11 clade (4, 5). The present study characterizes FHbp ID 896, which is expressed by 98% of the clade isolates. In contrast, FHbp found in urethritis isolates from sporadic cases demonstrates considerable sequence diversity as well as fHbp mutations. Taha et al. described a frameshift allele in each of six urethritis strains studied (33). Another report with 38 meningococcal urethritis isolates found considerable fHbp diversity, with all three major variant groups (or two subfamilies) represented (34). Similarly, diverse fHbp alleles were present in 33 cc11 urogenital isolates curated from the PubMLST database, with 17 isolates harboring nonfunctional alleles (1). Thus, FHbp ID 896 encoded by the clade is distinct from historic meningococcal urethritis isolates (35).

The control of fHbp transcription is complex (25, 26, 36) and has been shown to vary among genetic lineages (24). In contrast to previous report (24), bicistronic and monocistronic fHbp transcripts were detected in the cc11 clade isolates (Fig. 2B). Oriente et al. reported that the fHbp monocistronic transcript becomes the major transcript as a result of activation by FNR (25) under oxygen-limiting conditions, which is commonly found during urethritis (37–39). Since the clade’s fHbp promoter elements and the FNR binding motif are conserved, the clade may benefit from the upregulation of fHbp under anaerobic conditions in dealing with a host mucosal immune response. The importance of transcriptional coupling with the upstream gene cbbA, encoding a fructose bisphosphate aldolase, remained to be determined. CbbA, while typically a cytoplasmic glycolytic enzyme, has also been found localized to the outer membrane of N. meningitidis, is surface accessible, and is required for optimal adhesion to human cells (40).

The surface-exposed residues of FHbp make contact with FH and are targets of antibodies that elicit complement-mediated bacteriolysis. We detected moderate to high levels of FHbp expression in clade isolates, and FHbp ID 896 of the clade was recognized by the JAR 5 and JAR 41 MAbs (Fig. 4). The epitope of JAR 5 is composed of FHbp residues 84 to 91 and 115 to 123 (41). With the exception of one change, D for H at residue 119 (Fig. 1), the others are conserved in FHbp ID 896. JAR 5, in concert with a second anti-FHbp MAb such JAR 41, elicits human complement-mediated bactericidal activity and together confers passive protection against meningococcal bacteremia in human FH transgenic rats (29). Thus, FHbp ID 896 is recognized by protective anti-FHbp MAbs. Furthermore, polyclonal antibodies in sera from mice immunized with recombinant FHbp ID 1 (the FHbp antigen in MenB-4C) or MenB-FHbp elicited human complement-mediated bacteriolysis.

The ability of a meningococcal strain to bind FH is influenced by expression levels of FHbp and differences in protein sequences (23). Our data showed that FHbp ID 896 expressed by the unencapsulated CNM3 binds as much FH as encapsulated strain H44/76, a high expresser of ID 1 (Fig. 5D) and a strong FH binder (23). FH is typically present in serum at concentrations of ∼200 to 400 μg/ml. Complement components also are present at human mucosal surfaces (42), albeit at lower concentrations (∼10% of complement lytic activity in serum) (43, 44), can be produced through local mucosal synthesis, and are increased during inflammation. Furthermore, even low concentrations of FH are sufficient to downregulate complement activation and enhance the resistance of meningococci to complement-mediated bactericidal activity (45). Of note, the CNM3f mutant has no detectable binding of FH, as examined by fluorescence-activated cell sorter (FACS), despite expressing relatively high levels of NspA and not having a capsule. The interaction between FH and NspA on meningococci is also heavily influenced by lipooligosaccharide (LOS), with LOS sialylation increasing FH-NspA interactions (30). Meningococcal strains with capsules containing sialic acid (e.g., the serogroup B control strain H44/76) can either sialylate their LOS endogenously or scavenge host CMP-sialic acid to sialylate their LOS (46). The interactions of complement and antibody on the surface of the unencapsulated CNM3 and CNM3f clade isolates and H44/76 were investigated using bacteria grown with exogenous CMP-sialic acid to promote LOS sialyation of all three strains and permit comparison of the respective results to each other (Fig. 4 and 6). Given the relatively high NspA expression, why there is no detectable binding of FH to the fHbp null mutant is unclear, and further studies are needed. A lack of FH binding despite NspA expression had also been previously observed for an fHbp mutant of H44/76 (47).

Encapsulated meningococcal strains expressing FHbp ID 1 and ID 10 showed increased sensitivity to LL-37 when FHbp was deleted (13). However, the unencapsulated CNM3 behaved differently. Many cellular factors contribute to LL-37 resistance in both N. meningitidis and N. gonorrhoeae, including encapsulation, LOS, and the Mtr efflux pump, as well as hypothetical proteins with unknown functions (48–52). It is unclear which of these potentially confounding factors led to the opposite phenotypes of an fHbp mutation in different strain backgrounds. The fact that FHbp might sensitize the clade to the action of AMPs yet provide protection against serum bactericidal activities suggests FHbp may play a subtle and differential role during the various stages of meningococcal infection and colonization.

FHbp is a principal target of two meningococcal serogroup B-directed vaccines, MenB-4C (ID 1) and MenB-FHbp (ID 55) (18). Considering the high sequence homology of FHbp ID 896 to FHbp ID 1, the clade isolates were, as expected, killed by antibodies against FHbp ID 1. Although the sequence divergence between FHbp ID 896 and ID 55 is greater (Fig. 1), antibodies elicited by MenB-FHbp also showed bactericidal activities against the clade. Whether MenB vaccines might prevent urethritis caused by the clade may depend on the generation of mucosal immune responses. Both MenB vaccines were licensed based on eliciting serum bactericidal activity, which correlates with protection against invasive meningococcal disease. In contrast to meningococcal polysaccharide-protein conjugate vaccines, meningococcal protein antigen vaccines appear to have limited (53) or no (54) efficacy in decreasing meningococcal carriage. However, other data suggest that MenB vaccines containing outer membrane vesicles (OMVs) might be effective in decreasing gonococcal infections in human populations (55, 56). Our data indicate that FHbp expressed by the meningococcal urethritis clade is readily expressed on the meningococcal surface, enhances resistance to complement-mediated killing in low levels of human complement, likely is present at mucosal surfaces, and is recognized by cross-reactive protective anti-FHbp antibodies. FHbp is a potential vaccine target for control of urethritis caused by the U.S. meningococcal clade.

MATERIALS AND METHODS

Bacterial isolates and growth conditions.

CNM3 and CNM36 are from the collection of US_NmUC isolates from the sexual health clinic at Columbus Public Health, Columbus, OH (3). In addition, one US_NmUC isolate from Indianapolis, IN (IND414), and two from Atlanta, GA (ATL2 and ATL98), were examined in this study. Meningococcal strains were grown with 5% CO₂ at 37°C on gonococcal base agar (GC agar; BD) plates or GC broth as previously described (57).

Whole-genome sequence analysis.

Whole-genome sequence (WGS) data from 209 US_NmUC isolates (4) were extracted. Draft genome assemblies of urogenital sourced N. meningitidis and closed reference genomes of N. meningitidis, N. gonorrhoeae, and Neisseria lactamica were downloaded from PubMLST (http://pubmlst.org/neisseria/) (58). The MegAlign program in the DNASTAR Lasergene 16 suite was used for multialignments of short DNA and protein sequences by ClustalW. The 734 FHbp peptide allele sequences of the variant 1 group were retrieved from PubMLST and aligned using Clustal Omega, which uses seeded guide trees and HMM profile-profile techniques to generate alignments at the EMBL-EBI search and sequence analysis tools (59) and rendered by iTOL (60).

FHbp mutation.

An fHbp::aphA3 mutational construct was created by overlapping PCR. The 5′ fragment was generated with primers fHbp-5F (GACGGCAAAACCCCTTCTTCTTAT) and fHbp-5RA3 (ttcctcctagttagtcacccTGATTTTGCCTGCCTGACCTTC), while the 3′ fragment was made with primers fHbp-3FA3 (cctggagggaataatgacccGCCATACCGCCCTAGGACACG) and fHbp-3R (CGCCCGCCCGACCTGATA). (The lowercase sequence is complementary to those of aphA3 primers for annealing during overlapping PCR. The uppercase sequences are of fHbp locus.) The aphA3 cassette was PCR amplified with primers aphA3-SmF (GGGTGACTAACTAGGAGGAA) and aphA3-SmR (GGGTCATTATTCCCTCCAGG). The template of the first overlapping PCR was the mixture of the 5′ fragment and the aphA3 cassette and amplified with primers fHbp-5F and aphA3-SmR. The resulting PCR product was then mixed with the 3′ fragment as the template for the second PCR with primers fHbp-5F and fHbp-3R. The purified final PCR product was verified by sequencing and used to transform strain CNM3 with kanamycin selection (80 μg/ml) on brain heart infusion (BHI) agar plates supplemented with 2.5% fetal bovine serum.

Transcription linkage analysis.

Total RNAs were purified from mid-log-phase supplemented GC broth cultures. Fhbp-specific reverse primer fhbp-R (GGCGATTTCAAATGTTCGATTT) was used to generate cDNA in the presence or absence of reverse transcriptase and the reaction mixtures used as the template for RT-PCR transcription linkage analysis. PCR amplifications were performed for the fHbp coding sequence using primer pair fhbp-F (CTATTCTGCGTATGACTAGGAG) and fhbp-R (646 bp), the intergenic region using primers IGF (AAACCTGTTTCGTTGGAAAAA) and IGR (TTCACAGGTTTACTCCTAGTCAT) (231 bp), and the cbbA coding sequence using primers cbbA-F (GACGGCAAAACCCCTTCTTCTTAT) and cbbA-R (TGATTTTGCCTGCCTGACCTTC) (667 bp).

Western blot.

Bacteria grown in tryptic soy broth to early exponential phase (L) and overnight on the chocolate agar plate (S) were resuspended in phosphate-buffered saline (PBS) at the same optical density at 600 nm (OD₆₀₀). Bacterial lysates were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis using 4% to 12% NuPAGE polyacrylamide gels and 2-(N-morpholino)ethanesulfonic acid (MES) gel electrophoresis buffer. The proteins were transferred to polyvinyl difluoride membranes using a semidry electrophoretic transfer cell. The membranes were blocked using PBST (PBS plus Tween 20) containing 2% nonfat dry milk. The membranes were washed and incubated with the anti-FHbp polyclonal serum at 1:1,000 dilution. The polyclonal serum showed no nonspecific reactivity when probing the lysate of an fHbp mutant (13). The membranes were washed in PBST and incubated with a 1:10,000 dilution of a rabbit anti-mouse IgG–horseradish peroxidase conjugate, washed, and developed with a chemiluminescent substrate.

Selected reaction monitoring-mass spectrometry analysis.

Samples obtained from bacterial growth at 37°C performed under both liquid tryptic soy broth (OD of 0.25) and chocolate agar plate (OD of 0.8) conditions were examined. Bacterial lysates and protein digestions were prepared as described previously (26). For each strain, digestions of independent broth and plate cultures were analyzed in triplicate measurements. Three peptide fragments (TYGNGDSLNTGK, FDFIR, and LTYTIDFAAK) were monitored.

Flow cytometric detection of binding of mouse polyclonal or monoclonal antibodies and human FH to the surface of live N. meningitidis bacteria.

The method was performed as described previously (61), except the bacteria were grown to early log phase in liquid culture in Frantz medium containing 4 mM dl-lactate and 2 mM CMP-sialic acid (Sigma-Aldrich) instead of MH broth containing lactate and CMP-sialic acid. Aliquots of ∼10⁷ cells/ml were incubated for 1 h at room temperature with murine anti-P1.5 and anti-P1.7 MAbs, each tested at a dilution of 1:1,000. Anti-FHbp MAbs JAR 5 (reactive with subfamily B) and JAR 41 (reactive with subfamilies A and B) were each tested at 2 and 10 μg/ml, and JAR 13 (negative control, reactive with subfamily A) was tested at 10 μg/ml. Anti-NspA was a mixture of two MAbs tested at 10 μg/ml (14C7 and AL12, each at 5 μg/ml). After the bacteria were washed, the cells were incubated for 1 h at room temperature with goat anti-mouse IgG (H+L) conjugated with Alexa Fluor 488 (Jackson ImmunoResearch), diluted 1:500. The bacteria were washed twice with buffer and fixed with 0.5% (vol/vol) formaldehyde in PBS, and binding of the MAb was detected by flow cytometry. For detection of binding of human FH to the bacterial surface, the bacterial cells were incubated for 1 h at room temperature with various amounts of purified human FH (Complement Technology) as previously described (62). Sheep anti-human FH (AbCam) and donkey anti-sheep IgG–Alexa Fluor 488 (Life Technology), both diluted 1:500 and incubated for 45 min, were used for secondary detection. Group B strain H44/76 [15 (PorB3), P1.7, 16, ST-32] from an epidemic in Norway was tested in parallel with the clade strain as a positive bacterial control.

Serum bactericidal assays with fixed concentration of human complement.

The serum bactericidal assay (SBA) protocol was as described previously (63). Commercial IgG- and IgM-depleted normal human serum pool (PelFreez) was used as the complement source at 30% and assayed with serum from healthy unvaccinated adults, murine serum raised against recombinant FHbp ID 1 protein, and murine serum collected after MenB-FHbp vaccination, in all of which the complement had been heat inactivated at 56°C for 30 min. The sera from mice immunized with recombinant FHbp ID 1 or aluminum hydroxide alone were from a published study (64). For the MenB-FHbp mouse study: 6- to 8-week-old female CD1 mice (Charles River) were immunized with 1/5 of a human dose of MenB-FHbp (Trumenba) vaccine. Mice were given three doses each, 2 weeks apart, and blood samples were collected 2 weeks post-dose 3. The two pools were the combination of 5 mice per pool (10 mice total). Control mice (n = 10) were immunized with three doses of aluminum hydroxide (600 μg per dose) without antigen. Two serum pools were tested for SBA.

Serum survival assay.

The resistance of N. meningitidis to serum-mediated killing was tested by incubating ∼10⁴ CFU in RPMI with the desired dilution of commercial pooled normal human serum (Atlanta Biologicals) at 37°C with 5% CO₂. Mid-log-phase supplemented GC broth cultures were collected and adjusted in RPMI medium to an OD₅₅₀ of 0.15, followed by a further 5,000-fold dilution. Aliquots of diluted cells were added into 96-well plates containing an equal volume of serum dilution at 2× the concentration to be tested to initiate the incubation. At the designated time point, aliquots were removed and plated onto GC plates to determine the number of viable bacterial CFU. The percentage of survival was calculated by normalizing to the CFU counts at time zero (T = 0). Assays were conducted in duplicate for each condition and repeated at least three times.

Antimicrobial peptide LL-37 killing assay.

Overnight plate cultures were harvested into 0.2× GC broth diluted 5-fold with water (48) and adjusted to an OD₅₅₀ of 0.3. The standardized suspensions were diluted 100-fold and then 50-fold to have approximately 10⁵ CFU/ml. Each assay was started by the addition of 90 μl of cells into a well containing 10 μl of LL-37 to reach the desired final concentrations, 1.25 to 10 μg/ml. The microtiter plate was incubated at 37°C and 5% CO₂. Two 20-μl aliquots of the sample were removed after 30 min, and the number of viable CFU was determined by plating onto GC agar plates. The time course experiment was performed once to determine the appropriate incubation time, and the dose-dependent killing experiments in duplicate wells were subsequently performed three times. Unpaired and two-sided Student's t tests were used to determine the statistical significance of survival of the mutant with respect to that of the wild-type strain, with P < 0.05 considered significant.