INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Neisseria meningitidis colonizes the nasopharynges and oropharynges of about 10% of healthy individuals. In a small proportion of infected subjects, meningococci can invade the bloodstream and cross the blood-brain barrier, causing septicemia and/or meningitis. Eventually, the bacterium causes sporadic outbreaks and epidemics of invasive meningococcal disease with high mortality and morbidity rates (2, 12). Definition of the factors determining the development of meningococcal disease is, however, difficult, because the available animal models do not adequately reproduce the natural route of infection and human pathology.

The antigenic hypervariability, polysaccharide capsule production, adhesion, and signaling mechanisms of meningococci have all been thoroughly studied and are thought to play an important role in meningococcal carriage and disease (2, 21). Unlike a number of other gram-negative bacterial pathogens, however, no proteinaceous exotoxins have so far been implicated in meningococcal disease. Recently, three iron-regulated frp alleles of N. meningitidis were sequenced (frpA, frpC, and frpC-like). These frp genes encode large secreted proteins of unknown biological activity (17, 19, 20) which possess the characteristic carboxy-proximal RTX (repeat-in-toxin) repetitions of a nonapeptide motif, L-X-G-G-X-G-(D/N)-D-X. Various numbers of such repeats are found in the RTX domains of several cytotoxins involved in the virulence of other gram-negative genera, such as Escherichia, Proteus, Actinobacillus, Morganella, Pasteurella, Bordetella, and Vibrio (1, 9, 22, 23).

The assignment of the Frp proteins to the RTX protein family suggests that they might play a role in meningococcal carriage and/or disease. However, no intact frp gene was found in the sequenced genome of the serogroup A isolate Z2491 (13), which contains only fragments of frp genes scattered around the chromosome. In contrast, two different Frp proteins are expressed and secreted under iron-limited conditions by the serogroup C isolate FAM20 (18-20). These share large portions of identical sequence, but only 13 nonapeptide repeats are found in the 122-kDa FrpA, while 43 repeats are present in the 198-kDa FrpC protein (19, 20). The N-terminal 293 amino acid residues of FrpA and the 407 N-terminal residues of FrpC, however, do not exhibit any sequence homology to each other or to any known protein. This part of the FrpC protein harbors an Arg-Gly-Asp (RGD) sequence, which for a number of other proteins and bacterial virulence factors has been implicated in binding to integrins of mammalian cell membranes (5, 11, 12). A third type, a 141-kDa FrpC-like protein, is encoded in the genome of the serogroup B isolate MC58 (17). It corresponds to a truncated variant of FrpC, missing residues 251 to 377 from the amino-terminal portion and residues 1319 to 1718 from the repeats. The genome of MC58, however, also contains a gene for a longer FrpC protein nearly identical to that of FAM20 (17).

In a limited previous study, production of Frp proteins was detected in five out of eight meningococcal strains tested (19). In this study, we have detected the presence of frp alleles in a set of 65 isolates of N. meningitidis. It is shown for the first time that convalescent-phase sera of a number of patients after invasive meningococcal disease contain high levels of antibodies recognizing the FrpC protein. This suggests that FrpC may be involved in the pathogenesis of meningococcal disease.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, growth conditions and plasmids. Antigenically and phenotypically characterized isolates of N. meningitidis from patients with invasive meningococcal disease and isolates from healthy carriers were from a collection of strains of the National Reference Laboratory for Meningococcal Infections at the National Institute of Public Health in Prague, Czech Republic. The strains, however, were not matching pairs of isolates from patients and their corresponding individual contacts, since such pairs were not available for this study. The isolates were grown on Mueller-Hinton Agar (Bio-Merieux) supplemented with defibrinated sheep blood (chocolate modification) in an atmosphere of 5% CO₂ at 37°C for 24 h. The genomic DNA was isolated from plated pooled bacteria as described elsewhere (7). The Escherichia coli strain XL1-Blue (Stratagene) was used throughout this work for DNA manipulation and was grown at 37°C in Luria-Bertani (LB) medium supplemented with 150 µg of ampicillin/ml for plasmid-containing strains. The BL21(DE3) E. coli strain (Novagen) was used for expression of the FrpC protein and was grown at 30°C in LB medium containing 150 µg of ampicillin/ml. Plasmid pTZ19RMluI is a construct prepared by ligation of PstI-digested pTZ19R (Pharmacia) with a double-stranded adapter, 5'-TACGCGTATGCA, introducing a new MluI site. Plasmid pT7-7BsaAI was constructed by digestion of pT7-7 (16) with BsaAI and religation of the vector with a double-stranded synthetic hexanucleotide 5'-GGATCC. pTYB2 (the intein-mediated purification with an affinity chitin-binding tag protocol [IMPACT] T7 system) was from New England Biolabs.

PCR amplification. The reaction mixtures contained 10 mM Tris-HCl (pH 8.8), 50 mM KCl, 0.1% Triton X-100, MgCl₂ (Table 1), 200 µM (each) deoxynucleoside triphosphate, 1 µM (each) primer (Table 1 and Fig. 1), 0.5 µg of genomic DNA, and deionized water up to the final volume of 100 µl. After 5 min of DNA denaturation at 95°C, PCR was initiated by the addition of 2.5 U of Taq DNA polymerase (Top-Bio, Prague, Czech Republic) for amplification of fragments below 2 kb and 1 U of the LA DNA polymerase mix (Top-Bio) for amplification of longer fragments. Thirty cycles were performed under the conditions specified in Table 1 for each primer pair. Amplification of the conserved 1,779-bp pilAB intergenic region (7) was used as a positive control for the amplification reaction itself and for the quality of template DNA preparations.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the frpA, frpC, and frpC-like alleles of FAM20 and MC58 isolates. The fragments amplified by PCR are indicated by bars with arrowheads pointing in and labeled with the names of the PCR primer pairs used (Table 1). The regions of similarity between the frp alleles are indicated by hatched boxes. Large arrows indicate putative open reading frames.

Southern hybridization analysis. MluI- and HincII-digested genomic DNAs were fractionated on 0.6% agarose gels, transferred to nylon membranes (Hybond N+; Amersham Pharmacia Biotech), and fixed at 80°C for 2 h. The blots were hybridized with digoxigenin-labeled probes (DIG High Prime DNA labeling kit; Roche Molecular Biochemicals) corresponding to the frpA1 and frpA2 regions of frpA (Fig. 1 and Table 1). Hybridization took place at 38°C overnight in DIG Easy Hyb solution (Roche Molecular Biochemicals). The blot was washed twice with low-stringency buffer (2× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate]-0.1% sodium dodecyl sulfate [SDS]) at room temperature for 5 min and twice in high-stringency buffer (0.5× SSC-0.1% SDS) at 65°C for 15 min. The probe was visualized by immunodetection with anti-digoxigenin-alkaline phosphatase-conjugated antibody using the chemiluminescence substrate CSPD (Roche Molecular Biochemicals).

DNA sequencing. Nine of the 1,186-bp and one 2.7-kb PCR products amplified with the orf1-frpC primer pair were cloned and entirely sequenced. The DNA sequences were edited manually, and all ambiguous parts were also resequenced from the complementary strand. Comparison of DNA and deduced protein sequences was performed with the BLAST and MegAlign Lasergene software (DNAStar, Inc., Madison, Wis.).

Cloning of the orf1-frpC locus of N. meningitidis. Based on the nucleotide sequence of strain FAM20, the EcoRI and MluI restriction sites were used to clone the entire orf1-frpC locus from the N. meningitidis isolate 10/96 (C:2a:P1-2,5). The 6- to 10-kb-long MluI-EcoRI fragments of chromosomal DNA were purified on 0.4% agarose gels and ligated with MluI-EcoRI-digested pTZ19RMluI to obtain a partial genomic library. Ten pools of 36 individual clones each were formed and screened by PCR using the orf1-frpC primer pair (Table 1) to identify the recombinants with the cloned orf1-frpC locus. The detection was refined in two subsequent rounds of PCR screening, progressively reducing the pools to six clones each until eight individual genomic clones of the orf1-frpC locus were found. One of them was chosen for further work and was called pTZ19RfrpC locus.

Construction of expression vectors. To construct a vector for high-level expression, the open reading frame of frpC was subcloned into the expression vector pT7-7BsaAI under the control of the transcription and translation initiation signals of gene 10 from bacteriophage T7 (16). First, the 5' end of frpC was amplified with the primer pair frpC5' (Table 1). The 466-bp PCR product was cloned as an NdeI-EcoRI fragment into pT7-7BsaAI, yielding the pT7-7NdeIEcoRI construct. Next, the 3' end of frpC was amplified by PCR using the second pair of primers, frpC3' (Table 1), and the 257-bp PCR fragment was inserted as an EcoRI-BamHI fragment into pT7-7NdeIEcoRI, yielding pT7-7NdeIBamHI. The absence of undesired mutations in the subcloned PCR products was verified by DNA sequencing and comparison to the published sequence of frpC from the FAM20 strain (19). Next, the entire frpC gene was reconstituted by insertion of a 5,220-bp BsaAI fragment, encoding the remaining part of frpC, between the two BsaAI sites within the subcloned 5' and 3' ends of frpC on pT7-7NdeIBamHI to yield the plasmid pT7-7frpC. The integrity of the subcloned frpC gene was confirmed by restriction analysis and partial DNA sequencing. Expression of the 198-kDa FrpC protein was assessed by comparative SDS-polyacrylamide gel electrophoresis of extracts from induced cultures of strains carrying pT7-7frpC and mock-transformed E. coli.

Expression and purification of FrpC. For a typical purification of recombinant FrpC, 500 ml of LB medium supplemented with 150 µg of ampicillin/ml was inoculated 1:100 with an overnight culture of E. coli BL21(DE3) harboring the expression vector pTYB2frpC. The culture was grown with shaking at 30°C to an optical density at 600 nm of 0.6, and isopropyl--D-thiogalactopyranoside (1 mM final concentration) was added to induce the synthesis of the FrpC-intein-CBD fusion protein. After an additional 3 h of growth, the cells were collected by centrifugation, washed, and resuspended in cold 50 mM Tris-HCl-100 mM NaCl-1 mM EDTA (pH 7.4). The cells were then disrupted by sonication at 4°C, and the extract was cleared at 20,000 × g for 30 min and used for purification of the soluble form of FrpC.

Serum samples. Human sera were obtained from 19 individuals who suffered from invasive meningococcal disease (septicemia and/or meningitis). Serum samples were drawn from the patients immediately after their admission to the hospital (serum 1), 2 weeks later (serum 2), and 4 to 5 weeks later (serum 3). Furthermore, serum samples were obtained from 11 healthy individuals who were in close contact with patients during the first stages of invasive disease. While isolation of meningococci from blood or fluids of most these patients was successful, isolation of meningococci from nasopharynx swabs of all but one of the respective contact individuals failed. All contacts were also free of any clinical symptoms of the meningococcal disease. The serum samples were assayed for the presence of bactericidal antibodies against meningococci by standard methods. In corresponding cerebrospinal fluid or blood samples, the ethiological agent of the disease, N. meningitidis, was identified by cultivation, latex agglutination, and PCR methods. A control group of human sera was obtained from 18 healthy adult volunteers who had no contact with invasive meningococcal disease.

ELISA of human sera. Wells of the PolySorp enzyme-linked immunosorbent assay (ELISA) plates (Nunc, Roskilde, Denmark) were coated overnight at 4°C with 100 µl of the purified FrpC protein solution at 1 µg/ml in 0.05 M sodium carbonate buffer (pH 9.6). After being washed with phosphate-buffered saline (PBS), pH 7.4, the plates were blocked for 2 h at 37°C with PBS containing 1% bovine serum albumin (Sigma, Steinheim, Germany) and 0.05% Tween-20 in PBS (PBST) and washed three times with PBS. The human sera were initially diluted 1:100 and then diluted threefold for seven dilutions (from 1:100 to 1:72,900) in 1% bovine serum albumin-PBST. Duplicate 100-µl samples of the diluted sera were incubated in antigen-coated wells at 37°C for 2 h. After the plates were washed with PBST, 100 µl of either horseradish peroxidase-conjugated swine anti-human immunoglobulin G (IgG) (SwAHu IgG-Px; diluted 1:5,000) or horseradish peroxidase-conjugated swine anti-human IgA (SwAHu IgA-Px; diluted 1:1,000) (SEVAC, Prague, Czech Republic) per well was added in 10% normal swine serum-PBS and incubated for 45 and 60 min, respectively. After the plates were rinsed with PBST, o-phenylenediamine peroxidase substrate was added, and the plates were incubated at room temperature in the dark for 10 (SwAHu IgG-Px) or 40 (SwAHu IgA-Px) min. The reaction was stopped by the addition of 50 µl of 2 M H₂SO₄, and absorbance at 492 nm was measured. The cutoff value for determination of both anti-FrpC IgG and anti-FrpC IgA antibody titers was determined as the mean plus 2 standard deviations from the test results of negative sera from healthy volunteers at 1:100 dilution. The titers of FrpC-specific antibodies were then defined as the reciprocal of the last dilution yielding an A₄₉₂ greater than the cutoff value.

Nucleotide sequence accession numbers. The GenBank accession numbers of the frpA, frpC, and frpC-like alleles of the FAM20 and MC58 isolates are L06302, L06299, and NMB0585, respectively.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

In order to examine whether Frp proteins might be involved in the virulence of N. meningitidis, we analyzed the presence of the frp alleles in a set of 65 meningococcal isolates of all serogroups. This comprised 38 invasive isolates recovered from blood or cerebrospinal fluids of patients with invasive meningococcal disease and 27 carrier isolates recovered from the nasopharynges of healthy individuals.

The frpA allele is not present in most meningococcal isolates. The complete genome sequencing of the Z2491 and MC58 isolates revealed that in parallel with intact frp alleles, numerous fragments of the frp genes can also be scattered around the meningococcal chromosome (13, 17). To limit their detection, specific PCR amplification of entire frpA and frpC alleles (Fig. 1 and Table 1) had to be employed instead of Southern DNA hybridization. As shown in Fig. 2A, in control PCR the 1,779-bp conserved pilAB intergenic region was unambiguously amplified from the DNA of all isolates. In contrast, with primers matching the ends of the frpA open reading frame, the amplification of an frpA allele of the expected size of 3,303 bp was obtained with only 1 of the 65 isolates (Tables 2 and 3). Moreover, 1- to 8-kb-longer frpA fragments were amplified from 22 of the isolates (34%), showing that large DNA insertions occurred within their frpA alleles (Fig. 2B and Tables 2 and 3). However, no frpA-specific PCR product was obtained for the rest of the isolates (65%), suggesting that the frpA allele was poorly conserved. To check whether the entire frpA allele was absent in most isolates or whether a diversity of its 5'- and/or 3'-terminal sequences prevented PCR, we performed selective amplification of two different portions of frpA. With the frpA1 primer pair (Fig. 1 and Table 1), an expected 366-bp fragment of the 5'-proximal part of the frpA allele was amplified from only 7 out of 38 invasive isolates (Fig. 2C and Table 2) and 12 out of 27 carrier isolates (Table 3). With the frpA2 primer pair (Fig. 1), a 322-bp PCR product corresponding to DNA downstream of the frpA1 target was amplified for 16 out of 38 invasive isolates (Fig. 2D and Table 2). Moreover, a FAM20-derived frpA-specific probe, encompassing both frpA1 and frpA2 primer sites, failed to hybridize with genomic DNA of the PCR-negative isolates (not shown). Hence, an frpA allele was absent in most tested meningococcal isolates.

View larger version (53K): [in this window] [in a new window]   FIG. 2.   PCR detection of frpA alleles. The positions of the primer pairs used are listed in Table 1 and indicated in Fig. 1. (A) Amplification products of the pilAB locus specific for Neisseria spp. (7), which served as a positive control for template DNA quality and functionality of the amplification reaction. (B) Products of amplification of the entire frpA alleles. (C and D) frpA fragments amplified by primer pairs frpA1 (C) and frpA2 (D). The codes of the N. meningitidis isolates from which DNA was extracted are indicated above the lanes. The DNA of isolate FAM20 was used as a positive control.  DNA digested with PstI (A and B) and a 100-bp DNA ladder (C and D) were used as size standards in the first lane.

frpC-like alleles of variable size are present in all invasive and most carrier isolates. In contrast to frpA, the frpC alleles could be amplified from all 38 invasive meningococcal isolates and from 24 out of 27 carrier isolates when a pair of primers specific for the full-length frpC allele of FAM20 was used (Fig. 1 and Table 1). However, as illustrated in Fig. 3A and summarized in Tables 2 and 3, the sizes of the frpC-specific products varied considerably. Therefore, conservation of two different frpC portions was investigated.

View larger version (47K): [in this window] [in a new window]   FIG. 3.   PCR detection of the frpC loci. The positions of the primer pairs used are listed in Table 1 and indicated in Fig. 1. (A) Amplification products of the entire frpC alleles. (B) Amplification products of the repetitive portions obtained with the primer pair frpC-rtx. (C) Products of amplification of the nonrepetitive portions obtained with the frpC-non-rtx primer pair. (D) PCR products obtained with the orf1-frpC primer pair. The codes of the N. meningitidis isolates are indicated above the lanes. DNA of strain FAM20 was used as a positive control. Fragments of  DNA digested with PstI were used as size standards loaded in lane 1.

The 5'-terminal sequences of frpC are highly conserved among isolates. It was important to assess how conserved are the upstream and 5'-terminal sequences of frpC, which contain the transcriptional as well as the translational initiation signals determining the level of expression of the gene. Therefore, an orf1-frpC primer pair was used to amplify a 1,186-bp fragment from the 5' end of the orf1-frpC locus (Fig. 1 and Table 1). As shown in Fig. 3D, a PCR product of the expected length was obtained for all 38 invasive isolates tested (Table 2) and for 24 out of the 27 noninvasive isolates (Table 3).

View larger version (31K): [in this window] [in a new window]   FIG. 4.   Locations (boxed) of the sequence differences in the amino-terminal 118-residue portions of FrpC from nine clinical isolates and FAM20. The sequences were aligned using MegAlign software. The residue positions are indicated above the sequence blocks. Only sequence portions containing substitutions are shown.

Cloning of the orf1-frpC locus and expression and purification of recombinant FrpC. It was of interest to produce a recombinant and highly purified FrpC antigen, which would allow determination of the levels of FrpC-specific antibodies in convalescent-phase sera. Therefore, the orf1-frpC locus was cloned from an N. meningitidis 10/96 (C:2a:P1-2,5) strain, which is representative of the most common type of invasive isolate in the Czech Republic.

View larger version (86K): [in this window] [in a new window]   FIG. 5.   Purification of FrpC from E. coli BL21(DE3)/pTYB2frpC using a combination of affinity and ion-exchange chromatographies. Lanes: 1, crude extract from uninduced cells; 2, crude extract from cells induced for production of the FrpC-intein-CBD fusion protein; 3, clarified crude extract from induced cells; 4, chitin column flowthrough; 5, chitin column wash; 6, fraction of eluted 198-kDa FrpC after intein-mediated excision of intein-CBD; 7, FrpC after ion-exchange chromatography on a DEAE-Sepharose column; 8, high-molecular-mass standards. The samples were analyzed on a 7.5% polyacrylamide gel and stained with Coomassie blue.

FrpC-related proteins induce specific serum antibodies during invasive meningococcal disease. Positive serological evidence would indicate that FrpC-related proteins are expressed during natural infections. Therefore, a specific ELISA protocol was developed, using the purified recombinant FrpC as a coating antigen. In addition to standard negative controls, a mock extract of E. coli BL21, processed by the same chromatographic procedures as FrpC, was also used for coating. This allowed us to control for the reaction of sera with any residual E. coli components within the FrpC preparations.

View larger version (33K): [in this window] [in a new window]   FIG. 6.   Titration of sera of patients after invasive meningococcal disease. Results for six patients are given. The sera were taken at the time of admission of the patient to the hospital () and 2 () and 4 to 5 () weeks later. Detection of FrpC-specific IgG (A) and IgA (B) antibodies is shown. The dashed line indicates the cutoff value calculated as the mean plus 2 standard deviations of the test results of negative sera from healthy volunteers diluted 1:100. The three serum numbers for each patient are given in the upper right corner of each panel.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We show here that the sera of a majority of patients who suffered from invasive meningococcal disease contained antibodies specifically recognizing the FrpC RTX protein of N. meningitidis. Indeed, in some of these patients, the levels of FrpC-specific antibodies were already rather high 2 weeks after diagnosis of the disease. This demonstrates that an FrpC-related antigen was expressed by most of the invasive meningococcal clones and was immunogenic during a natural infection. It also raises the possibility that FrpC-related proteins contribute to meningococcal pathology.

Interestingly, 7 out of the 11 healthy contacts of diseased patients also had increased levels of FrpC-specific antibodies. Moreover, in the group of 18 healthy individuals with no reported contact with meningococcal disease, three sera were also positive well above the cutoff values. The circulation of meningococci in healthy populations and the probability of infection are, indeed, very high, with about 10% of individuals being colonized at a given time (3). Therefore, a plausible interpretation of the results is that FrpC may also be produced during the unrecognized-carrier state, when meningococci are just colonizing the nasopharynx. Experiments are under way to determine whether antibodies against FrpC are bactericidal.

Sera of four patients diagnosed with a characteristic meningococcal disease were seronegative for FrpC. Individual immune responses against a given antigen, however, are known to differ significantly within an outbred human population, depending also on the individual course of infection and applied treatment. Therefore, it could not be concluded whether the seronegative patients were infected by strains which did not produce an FrpC-like protein. The frpC alleles could be amplified from all of the invasive isolates. It was not systematically determined whether these alleles of various sizes were functional genes or had disrupted reading frames due to deletions and/or insertions. Twelve matching pairs of patient sera and the corresponding isolates, however, could be analyzed in this study. While only one of the 12 patients was seronegative for FrpC, a range of different sizes was observed for the frpC alleles detected in the group of corresponding meningococcal isolates (cf. Table 5). This suggests that these size-variable frpC alleles in fact encode a family of immunogenic FrpC-like proteins of various molecular masses. This conclusion tends to be supported also by Western blot analysis of iron-limited cultures of a set of meningococcal strains, where FrpC-like proteins of various molecular masses were detected (data not shown).

The variability of frpC alleles was at least partly due to differences in the sizes and/or numbers of the highly redundant RTX repeat blocks. The observed size differences of amplified repetitive domains suggest that these might consist of integral multiples of a 600-bp DNA block. Such a block is, indeed, present once in the frpA alle of FAM20, twice in the frpC-like allele of MC58, and four times in the full-length frpC gene of both the FAM20 and MC58 strains (17, 19, 20). It is worth noting that similar variation in the size of the RTX domain was recently also observed for a related RTX protein, ApxIVA from Actinobacillus (14). For frpC, however, not all size variation could be attributed to the repetitive sequences. Deletions as well as insertions were also observed in the nonrepetitive portion of the gene encoding the first 872 residues of FrpC. This further documents the genomic fluidity and high recombination frequency of meningococci (4, 6, 8, 13, 17). Moreover, there was no correlation between the detected size of the frpC allele and the serological characteristics of the isolates. This was also to be expected, since the serotype and subtype characteristics were previously found to reflect only poorly, if at all, the clonal lineages in meningococci (4, 6, 8).

Interestingly, a rather high sequence conservation of the 5' end of the frpC gene was found in the portion encoding the 118 amino-terminal residues of FrpC from nine local isolates. Five of the sequences were fully conserved in respect to that of the FAM20 strain, while the remaining four sequences had a limited set of amino acid substitutions at conserved positions. The sequence encoding the RGD tripeptide, shown to be involved in binding of other proteins to cellular integrins, was, however, intact in all of the frpC alleles examined. These results suggest that the N-terminal portion of the FrpC protein may be rather conserved and subject to selective pressure for conservation of function.

Characterization of the biological activities of the better-conserved FrpC proteins is now intensely pursued. Except for the repeat domain, which is highly homologous in all RTX proteins, the FrpC protein exhibits significant sequence similarity only to the newly identified RTX protein ApxIVA from Actinobacillus pleuropneumoniae (15). The area of similarity is located in the central 300 amino acids of ApxIVA and between residues 300 and 600 of the prototype FrpC protein of FAM20, with 42% identical and 50% identical plus similar amino acid residues, respectively. The biological function of ApxIVA also remains unknown. It appears to exhibit a weak hemolytic activity suggestive of a channel-forming capacity (15). However, no hemolytic activity of FrpC could be detected in our experiments (data not shown).

In contrast to frpC, the entire frpA allele was conserved in only 1 out of 65 isolates tested. It is unlikely that the additional 22 frpA alleles detected, which contain large DNA insertions, represent functional frpA genes. Some of these isolates apparently harbored only a part of the frpA-specific DNA, which could be amplified by only one primer pair (frpA2) and not by another matching primer pair immediately upstream (frpA1). This suggests that fragments of frpA were present in the chromosomes of this portion of the isolates tested. Moreover, the combination of Southern blot and PCR analyses showed that an frpA allele was absent in most of the strains tested. Indeed, a complete sequence of frpA was also not found in the two meningococcal strains whose genomes have been sequenced. Collectively these data suggest that a complete frpA gene may not be present in many strains.

Finally, when the observed size variation of frpC alleles is considered, it appears inappropriate to continue the differentiation of the meningococcal RTX proteins into FrpA and FrpC-like merely on the basis of size. Since large portions of identical sequence are shared by FrpA and FrpC, we suggest considering FrpA an insertion variant of FrpC and calling all the RTX proteins of meningococci FrpC. These could then be further classified according to their sizes.