Genetic Similarity of Gonococcal Homologs to Meningococcal Outer Membrane Proteins of Serogroup B Vaccine

Gonorrhea, a sexually transmitted disease, causes substantial global morbidity and economic burden. New prevention and control measures for this disease are urgently needed, as strains resistant to almost all classes of antibiotics available for treatment have emerged. Previous reports demonstrate that cross-protection from gonococcal infections may be conferred by meningococcal serogroup B (MenB) outer membrane vesicle (OMV)-based vaccines. Among 1,525 common proteins shared across the genomes of both N. gonorrhoeae and N. meningitidis, 57 proteins were predicted to be surface expressed (outer membrane proteins [OMPs]) and thus preferred targets for vaccine development. The majority of these OMPs showed high sequence identity between the 2 bacterial species. Our results provide valuable insight into the meningococcal antigens present in the current OMV-containing MenB-4C vaccine that may contribute to cross-protection against gonorrhea and may inform next steps in gonorrhea vaccine development.

Reverse vaccinology, which includes genome and proteome mining, has been utilized to identify potentially surface-exposed antigens that are highly immunogenic and protective for vaccine development (36)(37)(38). The multicomponent MenB-4C vaccine contains the following 3 immunogenic antigens identified by reverse vaccinology: factor H binding protein (FHbp) fused with GNA2091, Neisseria adhesion A (NadA), and Neisseria heparin binding antigen (NhbA) fused with GNA1030 supplemented with the OMV of the New Zealand epidemic strain (NZ98/254) to warrant broader immunogenicity (39). Although the formulation is not identical to that of the New Zealand MenB OMV (MeNZB) vaccine, immunogenic OMV components included in the MenB-4C vaccine may cross-protect against N. gonorrhoeae infections.
In this study, we sequenced and analyzed a diverse genome data set of 970 N. gonorrhoeae isolates collected from Gonorrhea Isolate Surveillance Project (GISP) sites across the United States between 2014 and 2016. We identified common proteins present in both N. gonorrhoeae and N. meningitidis and assessed the diversity of MenB vaccine antigens and OMV proteins to and within the U.S. N. gonorrhoeae strain collection. Bioinformatic tools were applied to predict the subcellular localization of each identified common protein. We found a significant number of predicted OMPs present in both Neisseria species, which could potentially contribute to cross-protection against gonorrhea and may serve as candidate antigens for gonococcal vaccine development.

RESULTS
Identification of common proteins and their subcellular localizations. N. gonorrhoeae and N. meningitidis show the closest genetic relationship within the Neisseria family (1)(2)(3)40). In total, we identified 1,525 proteins that were common among our set of 970 N. gonorrhoeae genomes and 4 NmB references, including MenB-4C vaccine antigens and major abundant proteins previously found in OMVs (see Table S1 in the supplemental material). Using various in silico prediction tools, we predicted the subcellular localization of each common protein found in both Neisseria species. The outer membrane localization was first determined for proteins in NmB and then compared them to those of the respective gonococcal homologs. A protein received its outer membrane-associated assignment only after being confirmed by at least 2 of 3 prediction tools, along with the detection of a signal peptide. A schematic workflow of the OMP identification process is shown in Fig. 1.
Among the major MenB-4C vaccine antigens, only FHbp and NhbA with their corresponding fusion proteins (GNA2091 and GNA1030, respectively) were found in all 4 NmB strains and 970 N. gonorrhoeae isolates (Table 1). A full-length NadA protein was present in the MC58 and CU385/83 NmB strains, but it was truncated in the other two NmB strains (H44/76 and NZ98/254) and absent in all N. gonorrhoeae isolates used in this study. In contrast to the NmB FHbp, the gonococcal FHbp homolog was predicted as not being expressed on the cell surface, based on its subcellular localization and the lack of a signal peptide ( Table 1). The gonococcal NhbA was predicted as being surface exposed. Both fusion proteins (GNA2091 and GNA1030) were predicted as being surface exposed in N. meningitidis; GNA1030 was surface exposed in N. gonorrhoeae, while only 1 tool predicted GNA2091 to be surface exposed in N. gonorrhoeae.
While the porA gene was identified in 966 of 970 (99.5%) N. gonorrhoeae isolates, the DNA sequences contained inactivating mutations in the promoter region (Fig. 2), along with frameshift deletions in the coding sequence resulting in premature internal stops, consistent with previous reports describing porA's pseudogene status in N. gonorrhoeae (25,26).
Common outer membrane proteins shared between NmB and N. gonorrhoeae strains. Among the 1,525 common proteins, 57 were predicted to be OMPs ( Table 2); 28 of these proteins had known biological functions, while the remaining 29 were either putative or hypothetical proteins. Of the 57 predicted OMPs, we identified 25 proteins that had been previously detected in the NZ98/254 OMV components using different laboratory approaches (24,30,32,34), 12 of which were abundantly and consistently expressed across different OMV lots (32) ( Table 2). These 12 proteins include some major immunogenic antigens that have been tested experimentally as potential gonorrhea vaccine candidates, such as PorB, PilQ, OpcA, and Omp85 (BamA)  a Cellular localization and signal peptide predictions were determined using the in silico tools BUSCA, Cello2GO, PSORTdb, and SignalP. OMP, outer membrane protein; ϩ, predicted as OMP by at least 2 tools; -, not predicted as OMP by any of the tools; ϩ/-, predicted as OMP by 1 tool; NA, not applicable due to the absence of an ORF. (9,12,13,41). The remaining 34 OMPs, including the MenB-4C vaccine antigens FHbp and NhbA, had not been previously identified in NZ98/254 OMVs. Two OMV proteins (MtrE and NEIS1428) were predicted as being outer membrane exposed in N. meningitidis; however, a signal peptide was detected only by 1 of the prediction tools. A similar observation was also seen for these 2 OMV proteins in N. gonorrhoeae. Of note, a previous report predicted the meningococcal PilE (NEIS0210) as an OMP (32); however, the prediction tools used in our study indicated extracellular/periplasmic/plasma membrane localization of PilE and absence of a signal peptide in both N. meningitidis and N. gonorrhoeae. In addition, a gene coding for OMP P1 (NEIS0073) contained an internal stop codon in all 970 N. gonorrhoeae isolates. While mafA (NEIS0596) was present in the 4 NmB and 970 N. gonorrhoeae genomes, it was fragmented across multiple contigs within the NZ98/254 genome. Furthermore, Tbp1 (NEIS1690) was only found in 89% of all N. gonorrhoeae strains, which did not meet our threshold of 95% for common protein selection. Therefore, PilE, OMP P1, MafA, and Tbp1 were not included in our list of common proteins and/or OMPs.
Sequence similarity of OMPs found within NmB and N. gonorrhoeae strain collections. Next, we analyzed the genetic similarity of each of the 57 predicted OMPs between the two Neisseria species and within the whole N. gonorrhoeae collection to assess sequence conservation. To this end, we focused our analysis on the NZ98/254 strain, from which the OMV in the MenB-4C vaccine derives. It has been shown that vaccines against Japanese encephalitis virus may induce cross-reactivity to West Nile virus and dengue virus since their common proteins share approximately 80% and 50% amino acid sequence homology, respectively (42,43). Based on this knowledge, we used a minimum sequence similarity of 80% as a threshold for potential crossprotection against N. gonorrhoeae (Fig. 3). The mean sequence similarity between N. meningitidis and N. gonorrhoeae for all 57 OMPs was 91.0% Ϯ 9.8%. At the amino acid level, 50 of 57 (88%) OMPs from the NZ98/254 strain showed 83 to 99% sequence    Values in boldface show mean sequence similarity that is greater than 80%.
Phylogenetic analysis was also conducted to assess the sequence diversity of the OMPs found within the two Neisseria species. Six common OMPs that had previously been identified in the NZ98/254 OMV and showed amino acid sequence similarity above the 80% cutoff to and within the 970 N. gonorrhoeae strains were selected for this purpose. These OMPs were FetA, PilQ, Omp85 (BamA), RmpM, NspA, and MtrE. Our analysis indicated that PilQ, Omp85 (BamA), RmpM, NspA, and MtrE of NZ98/254 formed a distinct cluster from the other 3 NmB strains, whereas FetA of CU385/83 was distinct from the rest of the NmB group (Fig. 4). This pattern of distinction among the NmB strains was also found with other OMV proteins, such as FHbp and its fusion protein (GNA2091), NhbA, AniA, and OpcA (Fig. S1), in contrast to PorB, which formed one cluster. Consistent with the number of unique alleles, RmpM, NspA, and MtrE formed clusters that are less diverse within N. gonorrhoeae than other OMPs, particularly FetA and PorB (Fig. 4 andS1). It is noteworthy that 2 alleles of gonococcal FetA, 1 allele of NhbA, and PorB shared a cluster with the NmB ones, suggesting that these alleles are closely related.

DISCUSSION
Increasingly resistant N. gonorrhoeae strains represent a global threat, and no gonococcal vaccine is currently available (13). Two vaccine candidates tested in a full cycle of field trials in humans, a whole-cell vaccine within the Aboriginal population of Inuits in northern Canada (47) and a pilus-based vaccine for high-risk U.S. military personnel stationed in South Korea (48), were unsuccessful due to a lack of efficacy compared to placebo-treated groups. Evidence for potential cross-protection against gonorrhea conferred by MenB OMV vaccines (19)(20)(21) could reduce the disease burden. Our analysis of NmB and N. gonorrhoeae sequences revealed more than 1,500 common proteins shared across the genomes of these 2 Neisseria species. We evaluated antigens of the MenB-4C vaccine, a licensed protein-based vaccine containing OMVs of the New Zealand N. meningitidis epidemic strain (NZ98/254) (32,34), to assess its potential cross-protective effect, particularly on the recent U.S. gonococcal strains.
Major MenB-4C vaccine antigens FHbp, NadA, and PorA are less likely to contribute to cross-protection against N. gonorrhoeae infections. The gonococcal FHbp has previously been shown to not be surface expressed and is unable to bind factor H (49). A recent study demonstrates that antibodies raised against the 3 major MenB-4C antigens did not recognize FHbp in whole-cell lysates of N. gonorrhoeae test strains (22), which confirms that the gonococcal FHbp homolog is not expressed on the cell surface. Furthermore, in accordance with previous reports (2,25,26), NadA was absent in all 970 N. gonorrhoeae isolates analyzed in this study, while porA's promoter appeared to be inactive for DNA transcription. Conversely, NhbA showed moderate amino acid sequence similarity to N. gonorrhoeae strains, as shown by others (2,22), and may therefore contribute to cross-protection to an extent. Similar levels of sequence similarity were found for NhbA of the other 3 NmB strains, including MC58, consistent with the previous finding (2). Interestingly, we found a high level of sequence similarity between NZ98/254 and N. gonorrhoeae for the 2 recombinant fusion proteins GNA2091 and GNA1030. These fusion proteins were shown to induce protective immunity in mice (50) and may therefore exert a cross-protective effect. A recent report showed that antisera raised against recombinant MenB-4C vaccine antigens recognized the gonococcal homolog of NhbA and the 2 fusion proteins (22). However, only NhbA is surface expressed in N. gonorrhoeae, while GNA2091 and GNA1030 are not (22), consistent with previous findings (51,52). Our analysis predicted GNA1030 to be surface exposed in both N. meningitidis and N. gonorrhoeae, suggesting that laboratory experiments are required to confirm the bioinformatic data. Similarity of Gonococcal Proteins to MenB-4C OMVs ® In pathogenic bacteria, OMPs present important virulence determinants through their direct interaction with the host immune system (especially OMPs involved in adhesion and invasion) and have therefore become major targets for vaccine development (53,54). Our sequence and subcellular localization analyses detected 57 common OMPs that were present in both NmB and 970 U.S. N. gonorrhoeae strains used in this study. One of the factors that has hampered gonorrhea vaccine development is the high antigenic variability of surface proteins among gonococcal strains (9,12,41). Our results demonstrate that FHbp, NadA, and PorA in the MenB-4C vaccine would not be effective vaccine antigens against N. gonorrhoeae. However, a number of other immunogenic proteins seem to have a high level of sequence conservation and may therefore serve as potential vaccine targets. In our study, ϳ98% of common OMPs present in both Neisseria species exhibit between 91% and 100% amino acid sequence similarity within 970 N. gonorrhoeae isolates. Approximately 42% of these proteins have been identified in OMV components of the NZ98/254 strain (24,30,32,34). Twelve of these OMPs, including PorB, RmpM, PilQ, OpcA, FetA, Omp85 (BamA), and LbpA, were abundantly and consistently expressed across different OMV lots of the NZ98/254 strain (32), and using bioinformatic analysis, their protein homologs were also identified in publicly available N. gonorrhoeae genomes, including FA1090. Of note, while the gene encoding LbpA is expressed in the majority of N. gonorrhoeae strains, it is a pseudogene in FA1090 (22). The abundances of other OMPs in the NZ98/254 strain and their potential to serve as candidate vaccine antigens warrant further investigation.
Accurate assessment of protein expression may require optimal protein detection methods. It is worth noting that most OMV vaccines were produced by extraction of vesicles using detergents to remove endotoxin from lipopolysaccharides. This extraction method could lower the content of immunogenic proteins released in OMVs compared to detergent-free methods (55)(56)(57). Therefore, the protein composition isolated from OMVs may vary depending on the preparation process. Interestingly, the vaccine antigen FHbp was found exclusively in detergent-free OMVs (30,56), which may explain its absence in the detergent-extracted OMVs (24,30,32,34) and underscores the importance of choosing proper laboratory methods in these investigations. In regard to cross-protection, one recent finding shows that a MenB vaccine made of native OMVs elicited gonococcal bactericidal antibodies in a mouse model (57).
Quantitative proteomic analyses have been demonstrated to facilitate the identification of surface-expressed gonococcal antigens for vaccine development, such as ubiquitous proteins localized to the cell envelopes and membrane vesicles, including BamA, LptD, TamA, MetQ, and NGO2054 (58,59). In these studies, assessment of protein subcellular localization was performed using a number of prediction tools, SoSuiGramN, PSORTb, CELLO, and SignalP, of which the last 3 tools and BUSCA were used in our study. While a combination of proteomic and bioinformatic approaches was utilized to identify surface-expressed proteins in the previous studies (58,59), our data are strictly based on the bioinformatic analysis of bacterial genomes. Among common OMPs identified in our study, PorB, PilQ, Omp85 (BamA), OpcA, LbpA, NspA, MtrE, MetQ, TbpB, and AniA have been experimentally tested as potential gonorrhea vaccine antigens due to their immunogenic and antigonococcal properties in vitro and/or in a mouse model (9,12,13,41). Previous findings describe the significance of MetQ (60) and AniA (44)(45)(46) as targets for antigonococcal vaccines and therapeutic intervention. Of those previously tested as gonorrhea vaccine candidates, PilQ, Omp85 (BamA), NspA, MtrE, MetQ, and LbpA have been identified within OMVs of the NmB strain NZ98/254 (24,30,32,34) and had Ն93% amino acid sequence similarity to N. gonorrhoeae, suggesting their potential contribution to mechanisms of cross-protection. On the other hand, PorB and OpcA only showed Յ70% amino acid sequence similarity to the respective N. gonorrhoeae homologs, which may reduce their cross-protective role against gonorrhea. These 4 OMPs (PorB, PilQ, OpcA, and LbpA) were abundant in the total protein amount quantified across different NZ98/254 OMV lots (32). Interestingly, copper-containing nitrite reductase AniA, the N. gonorrhoeae pivotal anaerobic respiration factor, was shown to be expressed only in urethritis isolates from men who have sex with men (61), providing the rationale for not being detected in the NZ98/254 OMV. Among gonococcal vaccine targets that were not part of our analysis, a monoclonal antibody against the conserved lipooligosaccharide (LOS) 2C7 epitope has been shown to hold promise in preclinical testing (13,62). Similarly, Liu and colleagues demonstrated that intravaginal immunization of N. gonorrhoeae-challenged mice with a mixture of gonococcal OMVs and microencapsulated interleukin-12 resulted in accelerated clearance of infection (63).
Taken together, our bioinformatics approach to examine the genetic similarities between NmB (particularly NZ98/254) and N. gonorrhoeae strains contributes to a better understanding of the OMPs present within the MenB-4C vaccine that may be involved in cross-protection and offers valuable insight into antigens of interest to guide next steps for gonorrhea vaccine development. Recognition of gonococcal OMV proteins by MeNZB-like OMV-induced antibodies provides an explanation for the previously observed decrease in gonococcal cases following MeNZB vaccination (22). Investigations of immunogenicity and types of immune responses are required to further investigate the contribution of host defense mechanisms induced by specific OMPs within the MenB-4C OMV against N. gonorrhoeae infections.

MATERIALS AND METHODS
Bacterial strain collection. This study included 970 N. gonorrhoeae isolates from 27 sites participating in GISP, a national sentinel surveillance system (https://www.cdc.gov/std/gisp/default.htm), obtained between 2014 and 2016. These isolates were selected to represent the geographic, temporal, and genetic diversity among the domestic N. gonorrhoeae population from an original set of 1,518 isolates previously sequenced for surveillance and for antimicrobial susceptibility studies. The NmB reference strain (MC58) and 3 NmB epidemic strains from New Zealand (NZ98/254), Norway (H44/76), and Cuba (CU385/83) were included as meningococcal reference strains. NZ98/254 belongs to clonal complex 41/44 (CC41/44), while the other 3 NmB strains belong to CC32. The 3 NmB epidemic strains were used in the OMV-containing MenB vaccines; the licensed MenB-4C vaccine contains an OMV from NZ98/254. DNA sequencing of N. gonorrhoeae isolates. Genomic DNA extraction was performed using Qiagen DNeasy kit, MagNA Pure 24 system (Roche), or Promega Wizard DNA extraction kit (Promega). Sequencing libraries were generated using the Nextera XT library preparation kit (Illumina) or NEBNext Ultra DNA library prep kit for Illumina (New England BioLabs), following the manufacturers' instructions or PulseNet protocols (https://www.cdc.gov/pulsenet/). High-quality libraries were (paired-end) sequenced on an Illumina MiSeq instrument using V2 reagents at Antibiotic Resistance Laboratory Network laboratories (in Maryland, Washington State, and Texas) and the Hawaii Department of Health or on an Illumina HiSeq instrument using HiSeq 2500 V2 reagents at the CDC Biotechnology Core Facility Branch.
Genome analysis and selection of OMPs. Illumina sequence reads for 970 N. gonorrhoeae isolates were trimmed using Cutadapt 1.8 (64) and assembled using SPAdes 3.7.0 (65). Low-coverage contigs relative to the overall genomic coverage were discarded. The assemblies for 4 NmB strains were included in the analysis. All genomes were annotated using the Neisseria allele collection from PubMLST (66) to identify and characterize the genomic features found in the assembly. The annotation files were compared to find genes that were common to all 4 NmB strains and found at a frequency of 95% or greater (n Ն 922) within N. gonorrhoeae genomes. The threshold of 95% represents the "soft-core genes," which includes genes that may otherwise have not been considered due to sequencing, assembly, or annotation errors. These common gene sequences were obtained and translated using custom Python (https://www.python.org/) scripts. Any gene sequence without a complete ORF or with premature internal stops was discarded. The resulting collection of protein sequences was used to generate a consensus sequence for each locus for N. meningitidis and N. gonorrhoeae separately. The consensus protein sequences for NmB and N. gonorrhoeae were provided as input to 3 subcellular localization Web servers, Cello2Go (67), BUSCA (68), and PSORTb (69). The molecular function of the proteins was predicted using InterProScan (70) software. Two of these tools, BUSCA and PSORTb, provided signal peptide predictions, while Cello2Go did not. As a result, a third signal peptide prediction software, SignalP (71), was included. The results of these tools were combined and used to identify OMPs of interest.
Analysis of sequence similarity. Unique protein sequences were obtained for the selected OMPs from all 970 N. gonorrhoeae and 4 NmB genomes. Protein similarity values were calculated using BLAST (72) for the following comparisons: all 4 NmB strains against all N. gonorrhoeae genomes, within all N. gonorrhoeae genomes, the NZ98/254 NmB strain against all N. gonorrhoeae genomes, and the other three NmB strains against all N. gonorrhoeae genomes.
Phylogeny analysis. Protein-based phylogenetic trees were created for selected OMPs. Alignments were created for each OMP using Clustal Omega 1.2 (73), and these alignments were used as input for RAxML 8.2.9 (74) to generate each phylogeny. All phylogenies were visualized and annotated using the Interactive Tree of Life (75) platform.
Data availability. All raw read data have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA317462. The accession numbers of the 970 gonococcal genomes used in this study are listed in Table S3 in the supplemental material.