Characterization of Neisseria meningitidis Isolates That Do Not Express the Virulence Factor and Vaccine Antigen Factor H Binding Protein

ABSTRACT Neisseria meningitidis remains a leading cause of bacterial sepsis and meningitis. Complement is a key component of natural immunity against this important human pathogen, which has evolved multiple mechanisms to evade complement-mediated lysis. One approach adopted by the meningococcus is to recruit a human negative regulator of the complement system, factor H (fH), to its surface via a lipoprotein, factor H binding protein (fHbp). Additionally, fHbp is a key antigen in vaccines currently being evaluated in clinical trials. Here we characterize strains of N. meningitidis from several distinct clonal complexes which do not express fHbp; all strains were recovered from patients with disseminated meningococcal disease. We demonstrate that these strains have either a frameshift mutation in the fHbp open reading frame or have entirely lost fHbp and some flanking sequences. No fH binding was detected to other ligands among the fHbp-negative strains. The implications of these findings for meningococcal pathogenesis and prevention are discussed.

Neisseria meningitidis is a Gram-negative bacterium that is a frequent member of the human nasopharyngeal flora, where it causes asymptomatic infection in 10 to 40% of healthy individuals (7,51). Occasionally, the bacterium translocates across the respiratory epithelial barrier, via a transcellular route (45), and establishes disseminated disease by invading into and replicating within the intravascular compartment. From there, the meningococcus can spread to the cerebrospinal fluid, causing meningitis (20,48). The organism remains a leading cause of Gram-negative septic shock and meningitis in developing countries and is responsible for epidemics that can involve hundreds of thousands of children and young adults in Saharan Africa each year (20).
The prognosis of meningococcal disease is directly correlated with levels of circulating lipooligosaccharide (LOS) and bacteremia, which can reach up to 10 9 CFU/ml in individuals with septic shock (5), a condition which still carries a significant case fatality rate and causes substantial long-term disabilities in survivors (44). To attain such high levels within the circulation, the bacterium must avoid killing by the host immune system (39). Complement is essential for defense against meningococcal infection. This is evident from the observation that individuals with deficiency in components of the membrane attack complex (MAC), a pore-forming multiprotein complex that causes bacterial lysis, are highly susceptible to meningococcal sepsis, with over a thousandfold-increased lifetime risk of developing disease (11). Furthermore, polymorphisms or deficits of other complement factors, including C2, C3, and properdin (11), are also associated with increased risk of developing meningococcal disease, while a recent genome-wide association study demonstrated that a region on chromosome 1 harboring the gene encoding factor H (fH), the main negative regulator of the complement system, is linked to susceptibility to meningococcal disease (10).
The meningococcus has evolved multiple mechanisms that promote resistance against complement-mediated lysis. Virtually all invasive isolates recovered from individuals with meningococcal disease express a capsular polysaccharide (17), which is necessary for survival in human serum, while truncation of LOS greatly increases sensitivity to complement (16). More recently it has been shown that the meningococcus recruits fH to its surface (28,39), which downregulates the activity of the alternative complement pathway and increases bacterial survival in the presence of human serum. fH is composed of 20 short consensus repeats (SCRs), each consisting of approximately 60 amino acids, which can engage other complement factors, including C3b, to mediate the regulatory functions of this protein (52). fH is present in the serum and binds to the surface of endothelial cells via polyanions, such as glucosaminoglycans. The meningococcus recruits fH to its surface by expressing factor H binding protein (fHbp) (28), a 27-kDa lipoprotein that consists of two ␤-barrels joined by a short amino acid linker (31,40). While charged carbohydrates on the surface of the vascular endothelium engage fH, charged amino acids in fHbp bind fH at nanomolar affinities at the same site of this complement regulator (40). In addition, it has been shown that fH can also bind to NspA on the surface of some meningococcal strains (24).
Based on differences in the nucleotide and predicted amino acid sequences, fHbps from different strains have been categorized using multiple schemes. These include two subfamilies (A and B) (33) or three variant groups (V1, V2, and V3) (32), with subfamily A corresponding to V2 and V3 and subfamily B to V1 (which is the most abundant). In a manner analogous to using genetic information to type strains by multilocus sequence typing (MLST), fHbp nucleotide and predicted protein sequences have been also assigned allele and peptide numbers (6), respectively, in a publicly available database (www .neisseria.org). For clarity, here we refer to the variant group and specify the allele of fHbp and the peptide subvariants. Of note, fHbps belonging to the same variant group share over 85% amino acid similarity, while there is only 60 to 70% similarity between the three variant groups (1,33). fHbp is also an antigen that elicits serum bactericidal antibody responses in immunized individuals and is a key component of investigational vaccines for the prevention of meningococcal disease, in particular that caused by serogroup B, that are currently being evaluated in clinical trials (12). Immunization of mice with fHbps from variants 2 and 3 generates responses with some degree of immunological cross-reactivity, but these variants do not induce bactericidal antibodies against strains expressing V1 fHbp (32). The vaccine being tested by Pfizer consists of two fHbp protein subvariants (one V1 and one V3) (13), while the Novartis vaccine contains outer membrane vesicle (OMV) and recombinant antigens, including a chimeric protein consisting of a V1 fHbp fused to another protein (14).
During a recent genotypic analysis of potential vaccine antigens, we identified strains from individuals with invasive meningococcal disease in which fHbp either contained a frameshift mutation or was entirely absent; these isolates were likely to be deficient in expression of this important virulence factor and vaccine antigen. Here we present a characterization of the abilities of these strains to express fHbp and their capacities to bind fH.

MATERIALS AND METHODS
Bacterial strains and growth. The bacterial strains used in this work are shown in Table 1. N. meningitidis was grown in the presence of 5% CO 2 at 37°C on brain heart infusion (BHI) agar plates with Levinthal's supplement or on Columbia agar with 5% (vol/vol) horse blood (Oxoid, Basingstoke, United Kingdom). Grouping, typing, subtyping, and MLST were performed using standard methods at the Health Protection Agency Meningococcal Reference Unit (15).
PCR and sequencing. Genomic and plasmid DNAs were isolated as described previously, analyzed by agarose gel electrophoresis, and visualized by staining with 0.1% SYBR green (Invitrogen). For genomic DNA, approximately 20 overnight colonies were suspended in 5 ml of physiological saline by using a sterile swab and adjusted to an absorbance of 0.1 at 650 nm. A 1-ml aliquot of the suspension was then incubated at 60°C for 70 min to ensure killing of the bacteria. The cells were then pelleted at 6,000 ϫ g for 10 min, and DNA was extracted with the DNeasy blood and tissue kit (Qiagen, Crawley, United Kingdom) according to the manufacturer's protocol. DNA was eluted from columns and stored at 4°C. PCRs were performed using the HotStarTaq DNA polymerase kit (Qiagen) or Expand Hi-Fidelity polymerase (Roche). Routine PCR and sequencing protocols for fHbp were performed as previously described (27). Where necessary, primers targeting sequences within flanking genes were used (1869-2F, GAAG AAATCGTCGAAGGCATCAAAC; 1871-Ralt, ATGCCGATACGCAGTCC[ G/C]GTAAAC), and PCR mixtures comprised 2.5 l of 10ϫ PCR buffer, 2.5 l of each primer (5 M stock), 0.5 l deoxynucleoside triphosphate (dNTP) mix (10 mM for each dNTP), 0.125 l HotStarTaq (Qiagen), 14.875 l moleculargrade water, and 2 l of eluted DNA template. Thermocycling conditions comprised an initial step of 96°C for 15 min, followed by 35 cycles of 95°C for 30 s, 63°C for 30 s, and 72°C for 80 s, with a final step of 72°C for 7 min. nspA was amplified with oligonucleotides NspAF (5Ј-GAAGGCGCATCCGGCTTTTAC G-3Ј) and NspAR (5Ј-TCAGAATTTGACGCGCACACCGG-3Ј).
Prior to sequence analysis, PCR products were purified using ExoSAP-IT (USB Corporation) or the Qiagen PCR cleanup kit followed by cloning into pGEMTeasy (Promega). Sequencing reactions were performed using the BigDye v3.1 kit (Applied Biosystems). Sequencing reactions comprised 1.75 l 5ϫ sequencing buffer, 0.5 l BigDye master mix, 0.66 l primer (5 mM stock), 6.09 l molecular-grade water, and 1 l purified PCR product. Sequence analyses were performed on a 3130xl sequence analyzer (Applied Biosystems). Contig assembly and manual adjustment of bases were performed using Sequencher v4.8 (Gene Codes Corporation).
ELISAs for detection of fHbp. Bacteria were cultured overnight on solid medium, resuspended in phosphate-buffered saline (PBS), and then fixed in the presence of 3% paraformaldehyde for 1 h. Cells were resuspended to a final absorbance of 0.2 at 650 nm in bicarbonate buffer (15 mM sodium carbonate, 35 mM sodium bicarbonate; pH 9.6), and 50 l was added to each well of an Immulon 2 HB enzyme-linked immunosorbent assay (ELISA) plate (Thermo Labsystems, Franklin, MA) followed by overnight incubation at 4°C. After washing, plates were blocked with 100 l per well of 3% bovine serum albumin in PBS at room temperature for 1 h.
After washing, sera raised in rabbits against fHbp V1, -2, or -3 (kindly donated by Novartis Vaccine Research) were added to wells in column 1 of the ELISA plates at a starting dilution of 1:200 in serum diluent (SD) buffer (5% newborn bovine serum and 2% dried milk in PBS containing 0.1% Tween 20 [PBST]), and doubling dilutions were made across the plates, with only SD buffer added to the final column. Following overnight incubation at 4°C, plates were washed, and then anti-rabbit IgG-alkaline phosphatase conjugate polyclonal antibody (pAb; Sigma-Aldrich, Dorset, United Kingdom) was applied at a final dilution of 1:1,000 in SD buffer for 2.5 h at room temperature. Following a final wash step, p-nitrophenol phosphatase (Sigma-Aldrich) was added to wells at a concentration of 1 mg/ml in diethanolamine buffer (1 M diethanolamine, 0.5 mM MgCl 2 ⅐ 6H 2 O; pH 9.8) followed by a 2-h incubation at room temperature. Enzyme activity was stopped by adding 25 l of 3 M NaOH to each well, and the optical density (OD) was read at 405 nm on a Versamax microplate reader (Molecular Devices, Sunnyvale, CA). Polyacrylamide gels were submerged in Coomassie blue staining buffer (0.2% Coomassie blue R-250, 40% ethanol, 10% glacial acetic acid) for at least 10 min at room temperature. Gels were destained with buffer I (40% methanol, 10% glacial acetic acid) for at least 10 min, followed by buffer II (10% glacial acetic acid, 4% glycerol) overnight.

SDS-PAGE and Western and far Western analyses. SDS-PAGE was per-
Proteins separated by SDS-PAGE were transferred to nitrocellulose membranes (Hybond; Amersham Biosciences, Buckinghamshire, United Kingdom) and equilibrated in transfer buffer (48 mM Tris, 39 mM glycine; pH 8.3) prior to use. Transfers were performed using standard protocols in a Mini Trans-Blot cell (Bio-Rad), filled with ice-cold transfer buffer (48 mM Tris, 39 mM glycine; pH 8.3), and run at 300 mA for 1 h. Membranes were then blocked in PBST with 5% skimmed milk either at room temperature for 1 h or at 4°C overnight.
For detection of fHbp, after blocking, membranes were rinsed with PBS and incubated with murine anti-V1, -V2, or -V3 fHbp primary antibodies (1:10,000; kindly provided by Novartis) diluted in PBST with 1% skimmed milk (PBSTM). Membranes were washed in PBST and then incubated in the secondary antibody conjugated to horseradish peroxidase (HRP) diluted in PBSTM. The duration of incubation and the dilution of antibodies were optimized for each antibody. After a further four washes with PBST, membranes were incubated for 1 min in enhanced chemiluminescence Western blotting detection reagent (Amersham Biosciences) at room temperature.
For detection of fH binding, following blocking in PBSTM membranes were rinsed in PBS and then incubated with purified fH (5 g/ml; Sigma Aldrich) in PBSTM for 2 h. After washing four times in PBST, membranes were incubated with goat polyclonal anti-human fH antibody (Calbiochem, EMD Biosciences), diluted 1:2,000 in PBSTM for 1 h. After four further washes in PBST, membranes were incubated in murine anti-goat HRP-conjugated IgG (Sigma-Aldrich) diluted 1:10,000 in PBSTM for 45 min. Membranes were exposed to Hyperfilm (Amersham Biosciences).

RESULTS
The fHbp polymorphisms ⌬T366 and ⌬A650 are associated with loss of fHbp expression. Without the signal sequence, the open reading frame (ORF) of V1 fHbp typically ranges from 765 to 789 bp in length and encodes a protein consisting of 255 to 263 amino acids. Initially we identified a strain of N. meningitidis with a polymorphism (⌬T366) in fHbp corresponding to alleles 765 bp in length that could affect protein expression through loss of a single base (T) at nucleotide (nt) 366 in the ORF (Fig. 1A). This change is expected to result in a frameshift mutation followed by 12 missense mutations prior to a TAG stop codon at nt 400, resulting in a truncated protein of 133 amino acids (Fig. 1B). Initially, four clonal complex 11 (cc11), sequence type 11 (ST-11) isolates with truncated V1 fHbp alleles were identified among a total of 22 cc11 isolates from all English and Welsh invasive disease isolates from 2007 to 2008 (total isolates, 613) (26). To determine whether this was an isolated recent phenomenon, fHbp was characterized among a further 18 English/Welsh cc11 isolates (C:P1.5,2, ST-11) from between 1998 and 2001 (which coincided with the peak in cc11-associated serogroup C disease in England and Wales [15]), among which a further three isolates harboring related truncated fHbp alleles were identified. A further two English/Welsh ST-11 isolates (M05 0240072 and M06 241270) containing truncated fHbp alleles were found from among eight English/Welsh ST-11 strains from 2005 to 2006. The uncharacterized cc11 isolate 0030/01 from the Czech Republic harbored the sole truncated fHbp allele among 28 Czech isolates (representing every seventh isolate between 2001 and 2006) (33), although its clonal complex is unknown. Therefore, a total of 10 strains were found to harbor fHbp alleles with the ⌬T366 polymorphism ( Table 1). Nine of these strains were from patients in the United Kingdom, of a total of 48 cc11 United Kingdom isolates analyzed (frequency of the ⌬T366 polymorphism in cc11 isolates, 18.8%) and comprised six serogroup C and three serogroup B strains, all belonging to ST-11 (cc11), except for one serogroup C strain which was an ST-7664 (cc11) isolate.
Sequence alignment showed that eight of the strains share the same fHbp sequence, which is identical to allele 82 (peptide 78) in the databases of www.neisseira.org when the frameshift mutation is corrected. The fHbp alleles from M05 240072 and 0030/01 differ from each other by just a single base (altering one amino acid, A22T) and from allele 82 by 5 or 6 nucleo-tides, with alterations of 2 (G30S and M35T) or 3 (A22T, G30S, and M35T) amino acids, respectively ( Fig. 1A and B). Neither the 0030/01 nor the M05 240072 fHbp subvariants had corresponding, nontruncated counterparts in the fHbp or other sequence databases or among other isolates we have tested to date. However, the ST-11 isolate M01 240074 (fHbp allelic subvariant 82) expresses an fHbp that is identical to eight of the strains when the frameshift mutation is taken into consid- eration and was therefore selected as a positive control for all isolates with the ⌬T366 polymorphism in subsequent assays, while MC58⌬fhbp was used as a negative control.
To determine whether these strains express any detectable fHbp, we performed Western blot and whole-cell ELISA analyses with polyclonal sera raised against recombinant V1 fHbp (protein subvariant 1) (Fig. 2). Both methods demonstrated that, as with the negative control (MC58⌬fHbp), none of the isolates harboring fHbp with the ⌬T366 polymorphism expressed detectable fHbp, with no truncated version of the protein detected by Western blot analysis. In contrast, fHbp was clearly expressed by the control strain M01 240074, which contains the intact subvariant 82 allele, as demonstrated by both ELISA and Western blot analyses. Therefore, the ⌬T366 polymorphism is associated with loss of fHbp expression in these isolates.
An additional United Kingdom clinical isolate (M08 240219) was found to possess an fHbp gene with a different frameshift mutation, ⌬A650, which is predicted to truncate the protein at amino acid 238, preceded by a series of 20 amino acids with no homology to fHbp (Fig. 3A). M08 240219 is a United Kingdom serogroup B clinical isolate belonging to ST-162 (cc162). This single cc162 isolate was identified among a total of 10 cc162 isolates from all invasive disease isolates in the epidemiological year 2007 to 2008 (frequency of the ⌬A650 polymorphism in cc162 strains, 10%); no other strains from this clonal complex were examined. Aside from the frameshift, the fHbp allele in this strain is classified as V2, allele/peptide 21, which is expected to be expressed as a full-length protein by strains M08 240039, M08 240804, and M08 240374, all of which are clinical serogroup B, ST-162 isolates. Of note, Western blot analysis showed that fHbp was detected in whole-cell lysates of M08 240219 by using polyclonal sera raised against recombinant V2 fHbp (Fig. 3B), although the protein was of a lower molecular mass than that present in the control strain, M08 240039; this is consistent with M08 0240219 expressing a truncated version of fHbp. In contrast, no protein was detected on the surface of this strain by whole-cell ELISA (Fig. 3C), unlike results with the control strains M08 240039 and M08 240374.
Identification of strains lacking an fHbp locus. As well as strains with frameshift mutations, we identified seven serogroup B clinical isolates from which the fHbp gene could not be amplified by PCR using primers targeting sequences flanking the fHbp ORF or highly conserved regions within the gene     5). This indicates a probable horizontal transfer event between the two species. Interestingly the 1,200-bp region contained an ORF of between 739 and 899 bp in the opposite orientation to fHbp. The ORF, annotated as a putative opacity protein in the N. lactamica strain 020-06 genome, contains a tract of 5 to 25 GCGTTCCT repeats. Addition or subtraction of repeat units in the ORF alters the reading frame; two of the seven cc286related isolates harbor an in-frame ORF. A tBLASTn search of the translated ORF against the nucleotide collection database revealed up to 36% amino acid identity (99% query coverage) against the opacity protein (Opa) of various neisserial species.
ELISA and Western blotting results demonstrated that the strains harboring a deleted fHbp locus did not express fHbp when polyclonal antibodies raised against recombinant V1, V2, or V3 fHbp were used (Fig. 6) Analysis of fH binding to strains not expressing fHbp. fH bound to the surface of the meningococcus retains its activity as a cofactor for fI-mediated cleavage of C3b and enhances survival of N. meningitidis in the presence of the human complement system (38). Therefore, we examined whether the strains with frameshift mutations in fHbp or lacking the entire fHbp gene bound fH. Far Western analysis was used to detect binding of full-length fH, as this method can differentiate between interactions with truncated fHbp and other potential targets, such as the 17-kDa protein NspA (24). There was no evidence of fH binding to the strains with the ⌬T366 mutation, even though binding was detected to the control strains expressing full-length fHbps (Fig. 7A). Similarly, no fH binding was detected to M08 240219, which has the ⌬A650 polymorphism (Fig. 7B); of note, even though the control strains M08 240039 and M08 240374 expressed fHbp based on Western analysis and ELISA, this protein did not bind fH in the far Western analysis (Fig. 7B). However, in lysates of M08 240374, far Western blotting detected a faint band of around 17 kDa, consistent with NspA; nspA was detected by PCR in all strains not expressing fHbp (data not shown). There was no fH binding detected in strains with the fHbp locus deleted (Fig. 7C).

DISCUSSION
The ability of certain bacteria to avoid elimination by the immune system is critical to their success as pathogens. Here we report the characterization of serogroup B and C meningococcal strains with polymorphisms predicted to result in truncations or a complete lack of fHbp. These isolates retain their capacity to cause disease despite their failure to express a functional fHbp for recruiting fH, the key negative regulator of the alternative complement pathway.
N. meningitidis is a highly diverse bacterium that expresses 1 of 12 capsular serogroups, each with a distinct chemical composition. Examination of the complete genome sequences of the meningococcus reveals that it possesses multiple mecha-nisms that enable phenotypic variation. These include homopolymeric tracts that mediate phase variation of surface antigens, including lipopolysaccharide, repeat sequences that can undergo recombination (such as DNA uptake sequences and Correia elements), and transposable elements (42). This variation enables the bacterium to circumvent killing by the immune system, and it is likely that different lineages of the pathogen have evolved distinct mechanisms to survive in the hostile environment of the host.
While a recent epidemiological analysis detected strains with a single point mutation that leads to a potential truncation of a V1 fHbp (18), we have described isolates from three distinct clonal complexes of the meningococcus, cc11, cc162, and those centered on ST-286, in which fHbp contains a frameshift mutation or has been replaced (along with flanking regions) by sequences shared with the commensal species N. lactamica. As all these isolates were obtained from patients with disseminated meningococcal disease, this demonstrates that fHbp is not essential for pathogenesis of these strains, even if recruitment of fH contributes to meningococcal survival in serum and human blood (38,43). We could not detect fH binding to any of these strains except M08 240374 by far Western analysis using purified fH and available polyclonal antibodies. This method has been used to successfully detect fH binding to meningococci either via fHbp or other potential ligands (24,38) and makes it possible to exclude nonspecific associations (by performing blotting assays in the absence of fH or the primary antibody).
In the absence of fHbp, N. meningitidis could recruit fH by alternative receptors. It was shown recently that certain meningococcal strains can bind fH to their surface via NspA, which has been evaluated as a vaccine antigen (30). NspA binds fH, especially in the absence of capsule and LOS sialylation, and can promote bacterial survival in dilute human serum (i.e., 1.5 to 5%) in strains expressing low levels of fHbp (24). We did detect fH binding to a protein of the same molecular mass as NspA in M08 240374 by far Western analysis. It is possible that the other strains utilize distinct complement regulators to promote their survival during bloodstream infection. For instance, it has been reported that C4BP, the negative regulator of the classical pathway, is also bound by N. meningitidis; we did not examine the strains for binding to C4BP, as this regulator is not recruited by the meningococcus in any appreciable amounts at physiologically relevant osmolarities (21). However, it is possible that the strains compensate for the lack of fHbp by recruiting other complement regulators.
cc11 is a hyperinvasive lineage associated with relatively high levels of mortality and morbidity, has caused outbreaks of serogroup C and W135 disease affecting individuals in developed and developing countries, and is not commonly recovered from healthy individuals in carriage studies (29). Of note, cc11 strains now account for approximately 18% of serogroup B disease in North America (19). A recent survey showed that cc11 strains from across the world express a diverse range of fHbps belonging to any of the three variant groups (1). However, the subset of cc11 strains identified here have the same frameshift mutation and are predicted to express V1 fHbps that are either identical to each other or differ by 3 amino acids at the most when the frameshift is taken into account. Therefore, it is likely that these alleles originated following a single We recently characterized three serogroup C ST-11 strains from Spain that produce high levels of capsule due to the presence of an insertion sequence, IS1301, in the capsule biosynthesis locus (cps), which promotes resistance against complement-mediated lysis (47). A diverse collection of other meningococcal isolates demonstrated that this genetic change is also found in a significant proportion of cc269 and other isolates. However, in cc269 this polymorphism is almost invariably accompanied by other changes in the cps that counter the effect of IS1301 on capsule biosynthesis (23). Therefore, the cc11 strains may have particular mech-  , and M1239 (control for V3 fHbp) and seven N. meningitidis cc286-related strains lacking the fHbp locus were separated by SDS-PAGE. Proteins were transferred to membranes, and fHbp was detected using a 1:10,000 dilution of polyclonal sera raised against fHbp variant 3. Similar results were obtained using polyclonal sera raised against variant 1 and variant 2 fHbp (data not shown). (B and C) The fHbp-deficient cc286-related isolates, positive-controls MC58 (fHbp variant 1), M08 240039 (fHbp variant 2), and M02 240629 (fHbp variant 3), and negative-control MC58⌬fHbp were fixed in the presence of 3% paraformaldehyde and added to wells in an ELISA plate; fHbp was detected using doubling dilutions of anti-fHbp variant 1 (B) or 2 (C) polyclonal sera using a starting dilution of 1:200 (shown as 200). OD values were adjusted by subtracting the readings from wells to which no serum had been added. VOL. 18, 2011 N. MENINGITIDIS ISOLATES THAT DO NOT EXPRESS fHbp anisms for subverting complement-mediated killing, allowing fHbp to be dispensable. cc162 strains account for a small but consistent proportion (5 to 10%) of serogroup B cases in developed countries, such as those in Europe and the United States (19), and are not frequently found in carriage studies (M. Maiden, personal communication). Although little is known about the features of cc162, it is interesting that several strains lack a functional two-partner secretion (TPS) system, through a frameshift mutation in the gene encoding the cognate transporter TpsB (49), which is responsible for transporting TpsA across the other membrane. The TPS system contributes to the adhesion of the bacterium to epithelial cells through unknown mechanisms. We identified a single strain from this clonal complex, M08 240219, with a frameshift mutation that truncates the protein prior to the site corresponding to a critical amino acid in V1 proteins, Glu 304 , which is a Thr in V2 and V3 fHbps (40). It is also apparent that the truncated form of the protein does not reach the cell surface, thus abrogating further a potential role in recruiting fH. As we found only one strain with this change, it is possible that the mutation represents an isolated occurrence and/or occurred during its isolation or passage while being sent to the Meningococcal Reference Unit. However, of note, the fHbp expressed by the control strains (i.e., peptide 21) does not bind fH by far Western analysis, although the basis of this is being investigated.
Initially a single isolate with an absent fHbp locus (M07 240677; ST-1867) was identified from all disease isolates (n ϭ 613) in the epidemiological year 2007 to 2008, which all underwent MLST and fHbp characterization. To determine whether this was unique, isolates belonging to ST-1867 and related STs (centered around ST-286 [ Fig. 4A]) were sought to determine whether they contained fHbp. Among the 14 ST-286-related isolates identified from our collection spanning 35 years, 7 had the deleted allele, representing STs 286, 1867, 3309, and 4019; disease caused by ST-286-related strains is infrequent. Nonetheless, we found that 50% of isolates lacked fHbp, with the locus replaced by sequences which are also found in N. lactamica strains and are likely to have been acquired by horizontal transfer between this species and N. meningitidis. Interestingly, this sequence includes an open reading frame that is predicted to encode a protein related to the opacity proteins. Of the opacity proteins, the meningococcal Opas are phase-variable proteins that have been shown to be the target of the activated complement factors C4b and C3b (25), although there is no evidence that they recruit regulatory proteins. It is not clear what is the function of the additional opacity protein in the ST-1867 strains, although it appears to be a phase-variable antigen, due to the presence of a repeat sequence in its ORF. Of note, the opacity protein Opc binds vitronectin (50), which is a negative regulator of the complement system (4). Therefore, it is possible that expression of Opa can compensate for the lack of fHbp in these strains. fHbp is an important antigen for two investigational vaccines against serogroup B N. meningitidis currently undergoing advanced clinical trials. The occurrence of disease strains lacking fHbp is reminiscent of strains lacking other vaccine candidates, such as PorA and FetA (9,34), and has implications for vaccines containing these antigens. However, cc162 and cc286 strains are not frequent causes of meningococcal disease in the United Kingdom, while cc11 has been mostly associated with serogroup C disease, which is preventable by the MenC conjugate vaccine. Altogether, the fHbp-nonexpressing strains account for 3/539 (0.6%) of Men B disease isolates and 6/613 (1%) of all isolates in the United Kingdom based on the last year for which we have complete MLST data. Just under 7% of MenB isolates in the United States between 2000 and 2005 belong to corresponding cc's and so may also include fHbpdeficient lineages (19). Therefore, strains lacking fHbp would have relatively little impact on the efficacy of the vaccines being evaluated. However, the occurrence of fHbp-deficient strains capable of causing invasive meningococcal disease does illustrate that active surveillance of disease and carriage strains is vital throughout and following vaccine implementation, in order to detect the potential emergence of strains that have evolved under selective pressure of any vaccine to no longer express this antigen.