Unusual Genetic Organization of a Functional Type I Protein Secretion System in Neisseria meningitidis
http://www.100md.com
感染与免疫杂志 2005年第9期
Molecular Bacteriology and Immunology Group, Institute of Infection, Immunity and Inflammation, University of Nottingham, Nottingham NG7 2RD, United Kingdom
ABSTRACT
Proteins secreted by Neisseria meningitidis are thought to play important roles in the pathogenesis of meningococcal disease. These proteins include the iron-repressible repeat-in-toxin (RTX) exoprotein FrpC. Related proteins in other pathogens are secreted via a type I secretion system (TOSS), but such a system has not been demonstrated in N. meningitidis. An in silico search of the group B meningococcal genome suggested the presence of a uniquely organized TOSS. Genes encoding homologs of the Escherichia coli HlyB (ATP-binding), HlyD (membrane fusion), and TolC (outer membrane channel) proteins were identified. In contrast to the cistronic organization of the secretion genes in most other rtx operons, the hlyD and tolC genes were adjacent but unlinked to hlyB; neither locus was part of an operon containing genes encoding putative TOSS substrates. Both loci were flanked by genes normally associated with mobile genetic elements. The three genes were shown to be expressed independently. Mutation at either locus resulted in an inability to secrete FrpC and a related protein, here called FrpC2. Successful complementation of these mutations at an ectopic site confirmed the observed phenotypes were caused by loss of function of the putative TOSS genes. We show that genes scattered in the meningococcal genome encode a functional TOSS required for secretion of the meningococcal RTX proteins.
INTRODUCTION
During invasive meningococcal disease large amounts of lipooligosaccharide are released from the bacterial surface leading to massive inflammation and pathology (7). The contribution of secreted proteins to meningococcal pathogenesis, however, is not well understood. Exoproteins play a major role in the pathogenesis of many gram-negative pathogens and we have recently demonstrated that meningococcal secreted proteins (MSPs) play a central role in modulation of host cell genes with a role in inflammation and apoptosis (28). Proteins may be secreted by gram-negative bacteria by one of a number of known pathways, including those listed as types I to V (19). We and others have shown that meningococci secrete a number of proteins of a class called autotransporters, which are secreted via the type V pathway. These include the putative virulence factors immunoglobulin A (IgA) protease (22), the adhesin App (13, 30), AutA (1), and the serine protease AspA (NalP) (35, 36). Meningococci also have genes with homology to fhaB and fhaC of Bordetella pertussis (21). In this pathogen the major virulence factor filamentous hemagglutinin (product of fhaB) is secreted via the two-partner secretion pathway (a variant of the type V pathway) with the help of the fhaC gene product (15). Meningococci are also known to secrete proteins with homology to the repeat-in-toxin (RTX) family of cytotoxins (33, 34). In other gram-negative pathogens such toxins are secreted via a type I secretion system (TOSS) (5), and the recombinant product of one of the meningococcal genes was shown to be secreted via the Escherichia coli hemolysin TOSS (32). The presence of a TOSS in the meningococcus, however, has not been demonstrated.
Type I protein secretion machines are a subset of ABC transporters and comprise three components: an inner membrane ATP-binding protein, an outer membrane channel protein, and a membrane fusion protein. The membrane fusion protein spans the periplasm and contacts both the ATP-binding protein and the outer membrane protein (2, 29). The E. coli hemolysin secretion system is considered to be the prototype TOSS and is composed of the HlyB and HlyD proteins, which are the ATP-binding protein and membrane fusion protein components, respectively, and the outer membrane TolC protein (39). By searching the published meningococcal genome sequences for genes encoding proteins with similarity to E. coli HlyB and HlyD, we identified two noncontiguous sequences encoding proteins with similarity to each of these proteins. Furthermore, the hlyD homolog was adjacent to a homolog of tolC, which is unlinked to the hemolysin secretion genes of E. coli. Here, we show that meningococcal hlyB and hlyD homologs encode components of a functional type I protein secretion machine required for secretion of the meningococcal RTX toxin-like exoproteins. The genetic organization of the meningococcal genes is unlike that of any reported TOSS.
MATERIALS AND METHODS
Bacterial strains and growth conditions. Bacterial strains used in the present study are listed in Table 1. E. coli strains were grown at 37°C in Luria-Bertani broth with agitation or on Luria-Bertani agar supplemented, where appropriate, with ampicillin (50 μg/ml), kanamycin (50 μg/ml), or erythromycin (50 μg/ml). Meningococcal cells were routinely cultured on chocolate agar (Oxoid) at 37°C in 5% CO2. For selection of mutants, meningococcal cells were cultured on Mueller-Hinton agar plates supplemented with Vitox (Oxoid) and, where appropriate, streptomycin and spectinomycin (each at 100 μg/ml), kanamycin (100 μg/ml), or erythromycin (6 μg/ml). Liquid cultures were grown in Mueller-Hinton broth supplemented with Vitox at 37°C with agitation. For preparation of whole-cell extracts, RNA, and concentrated cell-free supernatants, cells were cultured without agitation at 37°C in Dulbecco modified Eagle medium supplemented with 50 μM desferrioxamine (Sigma) or 40 μM FeCl2 (to achieve iron-rich conditions for semiquantitative reverse transcription-PCR [RT-PCR]) in an atmosphere of 5% CO2.
Preparation of secreted proteins and whole-cell lysates. Meningococcal cell cultures grown overnight in 50 ml of Dulbecco modified Eagle medium supplemented with desferrioxamine were centrifuged at 11,000 x g for 20 min. For preparation of concentrated cell-free supernatants, the supernatants were filtered through 0.2-μm-pore-size syringe filters (Sartoris) and concentrated 100-fold by ultrafiltration with a Vivaspin-20 protein concentrator (30,000 molecular weight cutoff; Vivascience) by centrifugation at 2,000 x g until the volume was reduced sufficiently. For whole-cell lysate preparation, cell pellets were washed in 1 ml of ice-cold phosphate-buffered saline (PBS; Oxoid) and resuspended in a further 1 ml of ice-cold PBS. A total of 200 μl of this suspension was transferred to a fresh tube, and cells were harvested by centrifugation at 12,000 x g. After the pellet was suspended in 80 μl of ice-cold water, 20 μl of 5x sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer (0.625 M Tris-HCl [pH 6.8], 5% SDS, 25% glycerol, 12.5% -mercaptoethanol, 0.002% bromophenol blue) was added, and the suspension was mixed and then heated to 95°C for 5 min. The remaining cell suspension was used to prepare RNA (see below).
Protein assays, SDS-PAGE, and immunoblotting. Protein concentration was determined by using Protein Assay Reagent (Bio-Rad) according to the manufacturer's instructions. Preparations of whole-cell lysates or concentrated cell-free supernatants were separated on 10% polyacrylamide gels containing SDS and visualized by using SimplyBlue SafeStain (Invitrogen) according to the manufacturer's recommendations. For immunoblotting, proteins were transferred to nitrocellulose membranes (Schleicher & Schuell) by using a Trans-Blot SD semidry transfer cell (Bio-Rad) according to the manufacturer's recommendations. After being blocked overnight in a solution of PBS containing 5% dried skimmed milk (blocking solution), the membranes were incubated for 1 h at room temperature with monoclonal antibody A4.85 diluted 1:50 in blocking solution. After three 5-min washes in PBS containing 0.1% Tween 20 (Sigma), membranes were incubated in goat anti-mouse immunoglobulin conjugated to horseradish peroxidase (Bio-Rad) diluted 1:2,000 in blocking solution. After being washed as described above, membranes were developed by using either 4-chloro-1-naphthol or with the ECL Western Blotting Analysis System (Amersham).
DNA manipulation. Chromosomal DNA was prepared by using a DNeasy tissue kit (QIAGEN) using the protocol for bacterial cells recommended by the manufacturer. Plasmid DNA was prepared by using a QIAprep spin kit (QIAGEN) according to the manufacturer's recommendations. Restriction enzymes were purchased from Kramel Biotech, New England Biolabs, or Fermentas and used according to the directions of the manufacturer. T4 DNA ligase was purchased from Boehringer Mannheim. Expand Taq DNA polymerase (Roche Diagnostics GmbH) was used in all PCRs. Unless otherwise stated, PCRs contained 100 ng of chromosomal DNA or 1 ng of plasmid DNA. All reactions contained each primer at 200 nM; 200 μM concentrations each of dATP, dCTP, dGTP, and dTTP; 1x Expand polymerase buffer (including MgCl2 to a final concentration of 1.5 mM), and 0.6 U of Expand polymerase in a total volume of 25 μl. After initial incubation at 95°C for 3 min, reactions were subjected to 30 cycles of incubation at 50°C for 1 min, 68°C for between 1 and 6 min, and 94°C for 45 s, with final incubations at 50°C for 1 min and 68°C for 10 min. Products of restriction digestion reactions and PCRs were analyzed by agarose gel electrophoresis.
RNA extraction and RT-PCR. Meningococcal cells from the same overnight cultures used to prepare secreted protein preparations and whole-cell lysates (see above) were pelleted by centrifugation at 2,000 x g and resuspended in 1 ml of TRIzol reagent (Invitrogen). After incubation at room temperature for 30 min, 200 μl of chloroform was added to each suspension, and the mixtures were vortexed for 15 s. Samples were incubated for a further 15 min before centrifugation at 12,000 x g at 4°C. The upper aqueous layer was removed and added to 1.4 ml of RLT buffer (a component of the RNeasy kit; QIAGEN) and 1 ml of ethanol. Mixtures were then processed by using an RNeasy kit according to the manufacturer's recommendations. RNA was quantified by using a NanoDrop spectrophotometer (NanoDrop Technologies). To remove contaminating genomic DNA, RNA samples were treated with RQ1 RNase-free DNase (Promega) according to the manufacturer's instructions.
For RT-PCR, mixtures containing 5 μg of RNA, reverse primer at 0.5 μM, and 1 mM concentrations each of dATP, dCTP, dGTP, and dTTP in a final volume of 10 μl were incubated at 65°C for 5 min and placed on ice for 5 min. To each reaction, 2 μl of 10x reaction buffer (supplied with Superscript II [Invitrogen]), 1 μl of RNAseOUT (Invitrogen), MgCl2 to 50 mM, and dithiotheitol (supplied with Superscript II) to 10 mM were added in a premix, bringing the total volume of each reaction to 19 μl. After incubation at 42°C for 2 min, 1 μl of Superscript II was added to each tube, and reactions were incubated at 42°C for 50 min and 70°C for 15 min and then returned to ice. Prior to PCR amplification, 1 μl of RNase H (supplied with Superscript II) was added to each mixture, and reactions were incubated at 37°C for 20 min. Then, 1 μl of each reaction was used in subsequent PCRs. The primers HlyB-RTF and HlyB-RTR were used to detect the hlyB transcript, the primers HlyD-RTF and HlyD-RTR were used to detect the hlyD transcript, the primers TolC-RTF and TolC-RTR were used to amplify the tolC transcript, and the primers HlyD-RTF and TolC-RTR were used to attempt to detect a transcript spanning hlyD and tolC. For semiquantitative RT-PCR mixtures, the reverse primer gdh-RTR was added, in addition to the gene-specific reverse primer, to cDNA synthesis reactions. The primers gdh-RTF and gdh-RTR were added after five initial cycles of subsequent PCRs, in which only gene-specific primers were present, since when both primer sets were added at the beginning the stronger gdh signal overwhelmed the much weaker hlyB, hlyD, and tolC amplicons. The relative intensities of bands in RT-PCR experiments were determined by densitometry using the ImageJ program (version 1.34i; 1997-2005, Wayne Rasband, National Institutes of Health, Bethesda, Maryland [http://rsb.info.nih.gov/ij/]). All primer sequences are shown in Table 2.
Mutagenesis and complementation. The hlyB gene (NMB1400), together with 830-bp upstream of this open reading frame (ORF), was amplified from chromosomal DNA of strain MC58 by using the primers HlyB-For and HlyB-Rev. This amplicon was cloned into the plasmid pGEM-T Easy (Promega) and subjected to inverse PCR mutagenesis with the primers HlyB-M1 and HlyB-M2. This resulted in the deletion of the first 1,351-bp of the ORF and the introduction of a unique BglII site. This site was used to introduce the omega cassette (encoding resistance to spectinomycin and streptomycin) digested with BamHI from plasmid pHP45 (26). The resulting plasmid was used to mutate the meningococcal strain MC58 by natural transformation and allelic exchange as described previously (13). The deletion in the resulting mutant (MC58hlyB) was confirmed by PCR analysis.
In order to complement the hlyB mutation, hlyB was amplified by using the downstream primer HlyB-ECR and either of the two upstream primers HlyB-ECF1 or HlyB-ECF2 (to amplify the long and short forms of hlyB, respectively), cut with AscI and BamHI, and cloned into the vector pYHS25 cut with BamHI and MluI. The resulting plasmid was used to transform cells of the hlyB mutant strain MC58hlyB. Insertion of the gene at the ectopic site was confirmed by PCR analysis with the primers EV-F and EV-R.
The hlyD homolog (NMB1738) was amplified by using the primers HlyD-For and HlyD-Rev and cloned into pBluescript. The resulting plasmid was subjected to inverse-PCR with the primers HlyD-M1 and HlyD-M2, digested with BclI, and ligated to the kanamycin resistance cassette digested with BamHI from plasmid pJMK30 (37). The resulting plasmid, in which the entire ORF had been replaced with the kanamycin resistance cassette in the opposite orientation to the replaced gene, was used to mutate strain MC58 (Fig. 4B). The deletion in the resulting mutant (MC58hlyD) was confirmed by PCR analysis. In order to complement the hlyD mutation, a region of DNA containing the hlyD and tolC genes was amplified by using the primers HlyD&tolC-ECF and HlyD&tolC-ECR. This fragment was cut with AscI and either BamHI (to clone hlyD alone) or BglII (to clone hlyD and tolC). Both fragments were separately cloned into the vector pYHS25 cut with BamHI and MluI. The resulting plasmids were used to transform cells of the hlyD mutant strain MC58hlyB. Insertion of the gene at the ectopic site was confirmed by PCR analysis by using the primers EV-F and EV-R.
The frpC gene was amplified by using primers FrpCM1 and FrpCM2 and cloned into plasmid pGEM-T Easy. The construct was digested with BglII and ligated to the omega cassette digested with BamHI (Fig. 4C). The resulting construct, in which the frpC gene had been disrupted, was used to mutate strain MC58. The mutation in the resulting strain (MC58frpC) was confirmed by PCR analysis.
MS analysis of protein bands. Individual protein bands excised from SDS-PAGE gels stained with SimplyBlue SafeStain (Invitrogen) were destained and prepared as described previously for MS analysis (14). Peptide mixtures eluted from each band were analyzed by MALDI-TOF-MS (Waters/Micromass, Manchester, United Kingdom), and protein identification was deduced from peptide mass fingerprints by searching relevant databases using PeptIdent (http://www.expasy.org/tools/peptident.html). Where peptide identification was not possible by this procedure, samples were subjected to peptide sequence analysis by nano-ESI-Q-TOF-MS/MS (Waters/Micromass), and the sequences obtained were used to search for matches using the BLAST algorithm at http://www.expasy.org/tools/blast/.
Protein and nucleic acid sequence analysis. Public databases of published protein and DNA sequences were searched by using the BLAST programs available at http://www.ncbi.nih.gov/BLAST/ and http://tigrblast.tigr.org/cmr-BLAST/. The genome databases of meningococcal strains Z2491 and FAM18, and the N. lactamica strain were interrogated by using the BLAST servers available at the Sanger Institute (www.sanger.ac.uk). Genome databases of strain MC58 were interrogated by using the BLAST server available at The Institute for Genomic Research (http://www.tigr.org). The genome sequence database of the gonococcal strain FA1090 was interrogated by using the BLAST servers available at www.ncbi.nlm.nih.gov/sutils/genom_table.cgi. Searches for Prosite patterns were made by using the server available at http://us.expasy.org/prosite/. Other DNA and protein sequence analysis was carried out by using the DNAMAN package of programs (Lynnon BioSoft).
RESULTS
The meningococcal genome contains evidence of a type I protein secretion system. The published genomes of group A and group B meningococcal strains Z2491 (25) and MC58 (31), respectively, were searched for genes encoding proteins with similarity to HlyB and HlyD of E. coli (accession number M10133) (8). An ORF encoding a predicted protein of 742 amino acids (aa) that was 70% identical to E. coli HlyB over 678 aa was detected in both genomes (NMA1620 and NMB1400, respectively) (Fig. 1A). A gene with similarity to the E. coli hlyD gene was also found in both genomes (NMA1996 and NMB1738, respectively). In the genome of strain Z2491 this gene contained a frameshift mutation but the equivalent gene in strain MC58 was intact and predicted to encode a protein of 475 aa with 39% identity to E. coli HlyD over 468 aa (Fig. 1B). We also searched both meningococcal genomes with the E. coli TolC protein sequence (accession number NP_755652) and found two ORFs whose predicted protein products had significant similarity: NMB1737 (immediately downstream of hlyD) was the closest match (27.6% identity; P = 2.0 x 10–31) (Fig. 1C); the mtrE gene (NMB1714), which encodes a putative multidrug efflux pump, also had significant similarity (20% identity; P = 4.7 x 10–6). The latter gene was not investigated further. The meningococcal hlyB-, hlyD-, and tolC-like genes are here called hlyB, hlyD, and tolC, respectively. Neither the meningococcal hlyB nor the hlyD/tolC loci were cistronic with any genes with homology to hlyA, including the meningococcal frpC gene, or to a second gene with close similarity to frpC (here called frpC2). The hlyB gene, however, was adjacent to a region of DNA containing frpC (NMB1415), a number of small frpC homologs (NMB1403, NMB1405, NMB1407, and NMB1409), the frpC operon gene frpD (NMB1414), and an additional frpD homolog (NMB1412) (Fig. 2).
The meningococcal hlyB and hlyD/tolC loci are flanked by genes associated with horizontal DNA transfer. The hlyB gene is flanked by the insertion sequences IS1106 and IS1016 (Fig. 2). Downstream of the hlyD/tolC locus there is a pseudogene with homology to a third class of insertion sequences (NMB1736); upstream of this locus are two short ORFs encoding hypothetical proteins of unknown function and a gene homologous to the rstA2 gene of the Vibrio cholerae phage CTXphi (NMB1741; Fig. 2) (38). The G+C content of the hemolysin genes in E. coli is lower than the average value for the E. coli genome (6). To determine whether the composition of DNA encoding the meningococcal hlyB and hlyD/tolC loci differed significantly from the overall composition of the meningococcal genome, we calculated the G+C content of the region bracketed between the IS elements (in the case of hlyB) and between NMB1736 and NMB1741 (including the hlyD and tolC genes). The G+C contents of these regions were 49.2 and 50%, respectively, compared to an overall composition of 51.8% for the entire genome.
The meningococcal hlyB, hlyD, and tolC genes are expressed. To determine whether the meningococcal hlyB, hlyD, and tolC genes are expressed, we analyzed RNA from strain MC58 by RT-PCR using primer pairs specific for each gene in turn (Fig. 3). Transcription of all three genes was detected by using gene-specific primer pairs (lane 2 in each of Fig. 3A to C). Negative control reactions, in which reverse transcriptase was omitted, did not yield equivalent products (lane 6 in Fig. 3A to C), indicating that the amplification products were derived from RNA rather than contaminating DNA.
N. meningitidis possesses a functional type I secretion machine. In order to determine whether the putative meningococcal hlyB and hlyD/tolC loci encode components of a type I secretion machine, each locus was mutated in turn. The hlyB gene (NMB1400) of strain MC58, together with 830 bp of upstream DNA was cloned and subjected to inverse PCR mutagenesis, resulting in the deletion of the first 1,351 bp of the ORF and the introduction of a unique BglII site. The omega cassette, encoding resistance to streptomycin and spectinomycin, was introduced into this site, and the resulting plasmid was used to mutate strain MC58 by natural transformation and allelic exchange (Fig. 4A). The deletion in the resulting mutant (MC58hlyB) was confirmed by PCR analysis (not shown). The mutant was tested for expression of hlyB by RT-PCR and shown not to express the gene, as expected (Fig. 3A, lane 3).
The hlyD homolog (NMB1738) was cloned and subjected to inverse PCR mutagenesis. After introduction of a kanamycin resistance cassette, the resulting plasmid, in which the entire ORF had been deleted, was used to mutate strain MC58 (Fig. 4B). The deletion in the resulting mutant (MC58hlyD) was confirmed by PCR analysis (not shown). The mutant was tested for expression of hlyD by RT-PCR and shown not to express the gene, as expected (Fig. 3B, lane 3).
Concentrated cell-free supernatants of wild-type MC58, MC58hlyB, and MC58hlyD cultures containing secreted proteins were prepared and analyzed by SDS-PAGE. Proteins with apparent molecular masses of 200 and 135 kDa were detectable in the supernatants of wild-type MC58 cells but absent from the supernatants of both mutants (Fig. 5). These data indicate that the products of the meningococcal hlyB and hlyD genes are required for secretion at least two exoproteins.
The endogenous RTX-containing exoproteins are substrates of the meningococcal type I secretion machine. Two RTX-containing proteins have been previously reported to be secreted by the group C meningococcal strain FAM20. These are the closely related iron-regulated proteins FrpA (122 kDa) and FrpC (197 kDa) (32) (Table 3). The genome of strain MC58 contains a number ORFs with similarity to the frpA and frpC genes of strain FAM20 (Table 3); the largest two of which (NMB1415 and NMB0585) encode proteins containing the hemolysin-type calcium-binding region signature sequence (prosite number PDOC00293). NMB1415 corresponds to the frpC gene of strain FAM20, while NMB0585 (here called frpC2) encodes a smaller but closely related protein. The predicted molecular masses of the two proteins (197 and 141 kDa, respectively) are in good agreement with the apparent molecular masses of the proteins present in culture supernatants of the wild-type strain MC58 but absent from culture supernatants of the mutant derivative strains MC58hlyB and MC58hlyD. To determine whether these proteins were the products of the frpC and frpC2 genes, we probed concentrated cell-free culture supernatants of MC58 and MC58hlyB with the monoclonal antibody A4.85. This antibody was raised against meningococcal outer membrane proteins and specifically recognizes the RTX proteins of strain FAM20 (34). Both proteins were detected by the antibody (Fig. 6 and 7), strongly suggesting that they represent the products of frpC and the related frpC2 gene. In addition to the major bands observable in Coomassie-stained gels at 200 and 135 kDa, additional bands with apparent molecular masses of 175 and 85 kDa were detected by the antibody. All of these bands were detected in whole-cell lysates of MC58, as well as the MC58hlyB and MC58hlyD mutant strains (Fig. 7B), indicating that, whereas secretion of these proteins was dependent on the expression of the hlyB and hlyD gene products, their expression was not significantly affected.
To further confirm the identities of the A4.85-reactive bands as FrpC and FrpC2 and to determine which bands corresponded to each gene, we attempted to mutate both genes. The frpC gene was cloned and disrupted by insertion of the omega cassette (Fig. 4C), and the resulting construct was used to mutate strain MC58. The mutation in the resulting strain (MC58frpC) was confirmed by PCR analysis (not shown). Secreted protein preparations from this mutant lacked the 200- and 175-kDa proteins but still contained proteins of 135 and 85 kDa (Fig. 6), indicating that the 200- and 175-kDa proteins were products of the frpC gene. Several attempts to mutate the frpC2 locus were unsuccessful. To confirm the identity of the 135-kDa protein, a preparation of meningococcal secreted proteins was separated by SDS-PAGE, and a close doublet of bands at 135-kDa that was apparent in the wild-type strain but absent from the hlyB mutant strain was observed after staining with Coomassie (not shown). Both bands were excised and subjected to MALDI-TOF analysis. The lower of the two bands was identified as NMB0585 (FrpC2). The upper band produced ambiguous results and was subjected to peptide sequence analysis by nano-ESI-Q-TOF-MS/MS. This procedure identified a mixture of two proteins, one of which was also identified as FrpC2, the other being the previously reported 140-kDa secreted form of the meningococcal autotransporter App (13). The identity of the 85-kDa protein could not be confirmed in this way since it was only detectable by immunoblotting.
Mutations at the hlyB and hlyD/tolC loci can be complemented by functional copies of hlyB and hlyD at an ectopic site. In order to confirm that the observed phenotypes in the mutants MC58hlyB and MC58hlyD were the result of the loss of expression of hlyB and hlyD, respectively, rather than polar effects on neighboring genes or by unknown secondary mutations, we constructed new mutants in which the original mutations were complemented at an ectopic site on the chromosome. We used an ectopic complementation vector in which the gene of interest may be cloned downstream of the porB promoter and upstream of the ermC gene (encoding resistance to erythromycin). These sequences are flanked by the genes NMB102 and NMB103, which are in a head-to-head configuration in both the vector and the meningococcal genome (40). The construct facilitates insertion of genes of interest, under the control of the porB promoter, into the meningococcal genome at a site that is unlikely to effect the expression of any other genes.
A region of hlyB that corresponds to its E. coli counterpart, starting with the methionine at position 33 (hlyBs) and, separately, the entire hlyB ORF encoding an additional 32 aa at the N terminus (hlyBl) were cloned into the complementation vector and reintroduced into the mutant strain MC58hlyB. The resulting complemented mutant strains (MC58hlyB hlyBsEct and MC58hlyB hlyBlEct, respectively) were analyzed by RT-PCR with primers specific for hlyB. The ectopic hlyB genes were expressed in both complemented mutants (Fig. 3A, lanes 4 and 5). When concentrated cell-free supernatants of cultures of the complemented mutant MC58hlyB hlyBlEct were probed in immunoblotting experiments with the monoclonal antibody A4.85, all of the proteins detected in the wild-type supernatants were detected at somewhat higher levels than in similar preparations of wild-type cells (compare Fig. 7A, lanes 3 and 5). These data confirm that the original mutation of the hlyB gene was responsible for the observed secretion defect. They also suggests that, under the conditions in which the cells were grown, the level of hlyB expression in wild-type cells was the rate-limiting condition for secretion of the FrpC and FrpC2 proteins. When culture supernatants of the mutant complemented with the shorter form of hlyB (MC58hlyB hlyBsEct) were probed with the same antibody, only very low levels of the major bands corresponding to FrpC and FrpC2 were observed (Fig. 7A, lane 4).
DNA fragments containing the hlyD gene alone or both hlyD and tolC genes were cloned separately into the ectopic vector and reintroduced into the hlyD mutant MC58hlyD. The expression of both genes was detected by RT-PCR in the cells of both complemented mutants (Fig. 3B and C, lanes 4 and 5). Immunoblotting experiments demonstrated that secretion of FrpC and FrpC2 was restored in both strains (Fig. 7A, lanes 7 and 8), confirming that the hlyD/tolC locus is required for the efficient secretion of these proteins. Levels of secretion by strain MC58hlyD hlyDEct were similar to those of the wild-type strain. In contrast, only a low level of expression was detected in strain MC58hlyD hlyDEct tolCEct.
Detectable amounts of FrpC and FrpC2 were not released by cell lysis. Cell-free concentrated supernatant fractions used in our experiments clearly contained secreted proteins. Although we cannot rule out the possibility that they also contain contaminating proteins derived from intracellular compartments as a result of cell lysis, we were unable to detect FrpC or FrpC2 in preparations prepared from the hlyB and hlyD mutants (Fig. 7A). These proteins were readily detected in whole-cell preparations of the same cells (Fig. 7B), indicating that FrpC and FrpC2 proteins detected in concentrated cell-free supernatants were the substrates of the type I secretion system and were not released nonspecifically from intracellular compartments by cell lysis.
The tolC gene is expressed independently of hlyD. The meningococcal hlyD and tolC genes are adjacent, in the same orientation, and separated by 269 bp. In order to determine whether the two genes were components of an operon, RNA from cells of the wild-type strain, the hlyD mutant MC58hlyD, and from this mutant ectopically complemented with either hlyD alone (MC58hlyD hlyDEct) or hlyD and tolC (MC58hlyD hlyDEct tolCEct) was subjected to RT-PCR with either a tolC-specific primer pair or a primer pair spanning the hlyD and tolC genes. Expression of tolC was detected in all four strains (Fig. 3C), but we were unable to detect a transcript spanning both genes in any strain (Fig. 3D). In strain MC58hlyD the hlyD gene had been completely replaced with a kanamycin resistance cassette in the opposite orientation with respect to the tolC gene; the gene upstream of the kanamycin cassette insertion is also transcribed in the opposite orientation with respect to tolC (Fig. 2). It is highly unlikely, therefore, that tolC could be transcribed from a promoter either within or upstream of the insertion site in strain MC58hlyD. These data indicate that tolC and hlyD are independently expressed.
Expression of the meningococcal hlyB, hlyD, and tolC genes is not iron repressible. The meningococcal RTX toxins are expressed at higher levels under conditions of iron limitation (34). To test whether the genes encoding their secretion apparatus were also iron repressible, RNA was prepared from meningococcal cells grown under iron-rich and iron-limiting conditions and analyzed by semiquantitative RT-PCR to detect relative expression levels of the hlyB, hlyD, and tolC genes. The gdh gene, encoding glucose-6-phosphate dehydrogenase, was used as an internal control since this gene would not be expected to be regulated by iron levels. None of the three genes was significantly upregulated under conditions of low iron availability; indeed, transcription of hlyB was actually reduced by a moderate but significant amount under iron-limiting conditions (Fig. 8).
RTX toxin-related genes are conserved in group A, group B, and group C strains of N. meningitidis but are not present in the genomes of N. gonorrhoeae or N. lactamica. In order to further examine the distribution of frpC-like genes and related secretion genes in the genus Neisseria, we searched the genomes of the group C meningococcal strain FAM18 and N. lactamica (both sequenced at the Sanger Institute), and of the N. gonorrhoeae strain FA1090 (sequenced at The University of Oklahoma). Homologs of frpC were detected in strain FAM18, including one gene encoding a 238-kDa protein containing the repeats characteristic of RTX toxins (Table 3 and Fig. 9). Homologs of frpC could not be detected in the N. gonorrhoeae or N. lactamica genome sequences. Genes with strong (>99%) identity to the hlyB, hlyD, and tolC genes of strain MC58 were detected in the genome of strain FAM18 (NMC1342, NMC1658, and NMC1657, respectively, in the preliminary annotation of this genome); no sequences with significant homology to these genes were detected in the N. lactamica or N. gonorrhoeae genome sequences.
To determine whether the meningococcal genomes also contained genes with similarity to the toxin activator genes of typical RTX operons, we searched the meningococcal genomes with the E. coli HlyC protein sequence (accession number M10133). The genome of strain MC58 contained two genes (NMB1210 and NMB1763), which encoded predicted proteins with 40% (P = 5.7 x 10–23) and 34% (P = 7.7 x 10–19) identity, respectively, to the E. coli sequence. Corresponding genes were not found, however, in the genomes of strains Z2491 or FAM18.
DISCUSSION
The presence of a TOSS for the export of meningococcal RTX-containing exoproteins was predicted over a decade ago (32). The availability of the genome sequence of the group B meningococcal strain MC58 (25) has made it possible to identify likely components of such a system. Here, we have bioinformatically identified hlyB, hlyD, and tolC-like genes and demonstrated that they were required for RTX protein secretion.
Most RTX proteins are found as part of an operon taking the form rtxCABD (18), in which the first gene encodes an activator protein required for the fatty acid acylation of the rtxA toxin gene product. The rtxB and rtxD genes encode the ATP-binding protein and membrane fusion protein components, respectively, of the secretion system. In E. coli the tolC gene, which encodes the outer membrane component of both the hemolysin exporter and a distinct TOSS required for colicin V secretion, is genetically unlinked to the hlyCADB operon (39). In some other RTX operons, the genes encoding each of the three secretion system components are linked: the tolC-like cyaE gene of B. pertussis, which is required for secretion of the bifunctional adenylate cyclase/hemolysin, is downstream of the cyaABD operon (11). Similarly, a tolC-like gene is located downstream of the mbxCABD RTX operon of Moraxella bovis (3). The arrangement of genes encoding the meningococcal RTX proteins and their secretion apparatus is unique. In all other type I secretion systems, to our knowledge, the genes encoding the ATP-binding protein and membrane fusion protein components are contiguous and, in most systems, components of operons containing the genes encoding their secretion substrates. In contrast, the meningococcal hlyB and hlyD genes were not linked to each other. The hlyB gene was, however, adjacent to a region containing frpC and a number of smaller related genes (Fig. 2). This region also contained the frpD gene, which is cistronic with frpC and encodes an outer membrane protein with a high affinity for FrpC (27), as well as an additional frpD-like gene. The juxtaposition of hlyD and tolC is also unusual, although these genes were shown to be expressed independently. The observation that both loci are flanked by genes associated with mobile genetic elements is intriguing: in strain MC58 hlyB is flanked by transposases of the IS1106 and IS1016 families (Fig. 2), while the hlyD/tolC locus is flanked by a pseudogene with homology to a third class of transposases (NMB1736 in strain MC58) and a gene homologous to the rstA2 gene of the V. cholerae phage CTXphi (NMB1741; Fig. 2): a gene required for phage replication (38). This genetic arrangement is conserved in the group A and group C strains Z2491 and FAM18, respectively, although in the preliminary sequence of strain FAM18 the ORFs corresponding to IS1016 (downstream of hlyB) and the rstA2 homolog (downstream of tolC) both contain frameshifts. These observations suggest that the meningococcal genes may have been acquired by horizontal transfer, although the DNA composition (G+C content) of these genes is not significantly different from that of the meningococcal genome. Previously, the region containing the hlyD pseudogene in strain Z2491 and its neighboring tolC-like gene has been identified as one of eight species-specific DNA islands present in N. meningitidis but not in N. gonorrhoeae (16). The authors of that study suggested that these islands may have arisen by import of DNA from other species and recombination via homologous flanking DNA. Evidence of horizontal gene transfer of RTX operons of a number of species belonging to the Pasteurellaceae, as well as some other gram-negative bacteria, has led to the suggestion that RTX toxins may have originated within the this family (10).
The frpC gene is well conserved in different strains of meningococci, whereas the closely related frpA gene is much less well conserved (23). In strain MC58 a second gene, NMB0585, which is almost identical to frpC but which lacks a number of internal fragments, is also present in the genome. Here we named this gene frpC2 to indicate its close similarity to frpC and to distinguish it from a number of other, smaller, ORFs with similarity to frpC in the MC58 genome. These latter ORFs lack the characteristic repeats of classical RTX proteins and do not contain regions homologous to the C-terminal region, which in other systems is required for secretion via TOSSs (4) (Fig. 9 and Table 3). Similar ORFs are also present in the genome of strain Z2491, but this genome does not contain any ORFs of a size equivalent to frpC/frpC2 or frpA. It is also of note that in this strain the hlyD homolog (NMA1976) contained a 61-bp insertion containing a stop codon 1,311-bp after the ATG start codon, followed by a direct duplication of the 73-bp preceding the insertion. This would be predicted to result in the expression of a truncated protein lacking a key signature motif found in members of the HlyD family of proteins (Prosite PS00543) (20). As a result of the frameshift, this signature motif is present in the ORF immediately downstream of NMA1976. The absence of both an intact hlyD gene and complete frpC-related genes in this strain suggests that the lineage has lost the capacity to produce and secrete RTX exoproteins. Sequence analysis of the 3' end of the hlyD gene in 13 strains of N. meningitidis chosen to represent the major clonal lineages of this pathogen, and including one group A isolate, revealed that the insertion was absent from all of these strains (16). It would be of interest to know whether other group A isolates contain intact RTX exoproteins, as well as an intact type I secretion system, since this may have important implications for the pathogenesis of isolates belonging to this serogroup. It is likely that the genes encoding the meningococcal RTX proteins and their cognate secretion systems were acquired by the species N. meningitidis after the divergence of this species and the related pathogen N. gonorrhoeae, since there are no genes with close homology to any of these dispersed genes in the genomes of the latter species or that of N. lactamica.
The first 32 aa encoded by the meningococcal hlyB gene (NMB1400) are predicted to encode an N-terminal region not present in HlyB of E. coli. Furthermore, the first methionine residue of the E. coli sequence aligns with an internal methionine residue at position 33 in the meningococcal sequence (Fig. 1). This led us to speculate that the second ATG codon of the meningococcal ORF was the true start of the gene. When we expressed the shorter form of the gene at an ectopic site in the hlyB mutant strain, however, we were only able to detect very low levels of protein secretion. In contrast, when the complete ORF was used to complement the mutation, the levels of substrate protein secretion were even higher than the wild-type levels. Since both ectopic genes were transcribed, this observation suggests that the entire meningococcal ORF, including the 32-aa meningococcus-specific N terminus, are required for the efficient functioning of the meningococcal HlyB protein, although we cannot rule out the alternative hypothesis that the shorter form of the gene was not efficiently translated in the complemented mutant.
Complementation of the hlyD mutation with an ectopic copy of hlyD resulted in secretion of FrpC and FrpC2 at levels similar to those of wild-type cells. When we designed our complementation strategy, we postulated that expression of tolC may be disrupted in the hlyD mutant. For this reason, we also made a complementation construct that contained both hlyD and tolC. The results of RT-PCR analysis subsequently demonstrated that tolC expression was unaffected by the mutation in the hlyD gene. These data were consistent with the observation that complementation with hlyD alone resulted in efficient phenotypic complementation of the hlyD mutation. When both hlyD and tolC were introduced into the ectopic site, the resulting levels of FrpC/FrpC2 secretion were much lower than wild-type levels. It is likely that the altered stoichiometry in this complemented mutant, in which two functional copies of tolC were present, resulted in interference with the normal interactions between the different components of the secretion system.
The neisserial TolC homologue is shorter by 56 C-terminal amino acid residues than its E. coli counterpart (Fig. 1C), and this may reflect a functional difference between the two proteins. The crystal structure of the E. coli TolC protein has been solved to a high resolution (2.1 ) (17). TolC assembles into a homotrimer, in which each monomer contributes 4 of 12 beta strands that together form a channel through the outer membrane in a beta barrel conformation. Each monomer also contributes a number of alpha-helical regions that form a channel continuous with the beta barrel and projecting up to 100 into the periplasm. In order to crystallize the E. coli TolC protein, it was necessary to proteolytically remove the carboxy-terminal 43 residues (17). Moreover, it was reported that this truncation did not affect the function of TolC (17). TolC functions that were tested included hemolysin secretion, colicin sensitivity, and drug efflux (V. Koronakis, unpublished data). The region removed by proteolysis prior to crystallization is largely coincident with the C-terminal part of E. coli TolC that does not have a counterpart in N. meningitidis. Thus, the structure of this region is unknown. More importantly, it is known to be dispensable for at least some functions of the E. coli molecule.
Low iron levels have been known for many years to be used by bacteria as a signal for the upregulation of a number of virulence factor genes (41). The meningococcal RTX proteins were originally identified as being regulated by iron (34), and it might be expected, therefore, that the genes encoding the secretion machinery for the products of these genes might be coordinately regulated in response to the same signal. In an extensive microarray-based study of the iron-regulated genes of strain MC58, Grifantini et al. showed that frpC, the frpC-like genes NMB0365 and NMB1405, frpD, and the frpD-like genes NMB0364, NMB0584, and NMB1412 were all iron repressible (12). For some reason frpC2 was not identified in this screen, although the adjacent frpD homolog was shown to be iron regulated. The genes encoding the meningococcal hlyB, hlyD, or tolC were also not among the iron-repressible genes reported (12). Our results using semiquantitative RT-PCR confirm these findings and show that the meningococcal TOSS genes are not iron repressible. The significance of this is not known, but it is possible that since each secretion channel can process many exoprotein molecules it may be of less importance in regulating the genes encoding the channel components than those encoding the secreted proteins themselves.
Multiple bands were detected by the monoclonal antibody A4.85 in secreted protein preparations and whole-cell extracts. The secreted protein bands at 200 and 175 kDa were clearly the products of the frpC gene, since they were not observed in the frpC mutant strain MC58frpC. The band observed at 135-kDa was confirmed to be a product of the frpC2 gene by mass spectroscopy. It is likely that the 85-kDa band is also a product of this gene. The meningococcal RTX-containing exoproteins have recently been shown to belong to a new class of RTX proteins that are capable of calcium-dependent autoproteolytic cleavage and cross-linking. FrpC was shown to cleave the peptide bond between Asp414 and Pro415; the resulting amino-terminal fragment is capable of covalent linkage via its carboxy-terminal aspartate residue to the -amino group of an internal lysine residue of another FrpC molecule (24). It is likely that the 175-kDa FrpC-related band is the carboxy-terminal fragment of FrpC reported previously (24). The smaller of the two proteins detected in secreted protein preparations of MC58frpC (as well as wild-type cells) could be a similar cleavage product of FrpC2, since the cleavage site identified previously within FrpC is conserved in the smaller protein. Additional bands detected in whole-cell lysates might represent additional products of FrpC or FrpC2 resulting from autocleavage and/or cross-linking events, degradation products resulting from the activity of other intracellular proteases, or the products of the smaller frpC-related ORFs.
The family of RTX-containing proteins include pore-forming cytotoxins, metalloproteases, and lipases; they contribute to the virulence of a number of important pathogens (18). Evidence for a role for the meningococcal RTX-containing exoproteins in pathogenesis is thus far lacking. A mutant strain of N. meningitidis lacking functional frpC and frpC2 genes was not attenuated in an infant rat model of meningococcal septicemia (9). Similarly, a mutant of strain MC58, in which the meningococcal-specific DNA island containing the hlyD and tolC genes was deleted, was able to multiply in the infant rat bloodstream at levels similar to those of the wild-type parent strain (16). Furthermore, no target cells or receptors have thus far been identified for the meningococcal RTX proteins. It can also be concluded that the meningococcal RTX proteins are not always required to cause meningococcal disease: strain Z2491, which has lesions in both RTX protein genes and in the hlyD secretion gene, was isolated from a case of meningitis (assuming the genetic lesions in this strain did not accumulate during laboratory passage). However, RTX protein genes have been shown to be present in a wide range of group B and group C meningococcal lineages, and high levels of serum antibodies have been detected in patients suffering from invasive meningococcal disease (23). The hlyD and tolC genes have also been shown to be universally present in a range of isolates chosen to represent the major clonal lineages of N. meningitidis (16). These observations, especially in the light of the known role of related toxins in other bacteria, suggest that the meningococcal RTX proteins are likely to play an important role in meningococcal carriage and/or pathogenesis. Since these genes are not found in the closely related pathogen N. gonorrhoeae, which characteristically causes localized inflammation of the genitourinary tract but rarely causes disseminated septicemia or meningitis, and are also not found in the closely related commensal species N. lactamica, it is tempting to speculate that secretion of the meningococcal RTX proteins contributes in some way to the ability of this pathogen to cause disseminated disease. It remains an important goal to identify target cells and receptors for the meningococcal RTX proteins and to determine what, if any, roles they play in meningococcal pathogenesis.
ACKNOWLEDGMENTS
We thank Fred Sparling for supplying monoclonal antibody A4.85 and Chris Tang for providing the ectopic complementation plasmid. Mass spectrometry was carried out by Kevin Bailey and Matthew Carlile at the University of Nottingham Proteomics facility.
Present address: Department of Chemistry, Faculty of Science and Art, University of Dicle, 21280 Diyarbakir, Turkey.
REFERENCES
1. Ait-Tahar, K., K. G. Wooldridge, D. P. Turner, M. Atta, I. Todd, and D. A. Ala'Aldeen. 2000. Auto-transporter A protein of Neisseria meningitidis: a potent CD4+ T-cell and B-cell stimulating antigen detected by expression cloning. Mol. Microbiol. 37:1094-1105.
2. Akama, H., T. Matsuura, S. Kashiwagi, H. Yoneyama, S.-i. Narita, T. Tsukihara, A. Nakagawa, and T. Nakae. 2004. Crystal structure of the membrane fusion protein, MexA, of the multidrug transporter in Pseudomonas aeruginosa. J. Biol. Chem. 279:25939-25942.
3. Angelos, J. A., J. F. Hess, and L. W. George. 2003. An RTX operon in hemolytic Moraxella bovis is absent from nonhemolytic strains. Vet. Microbiol. 92:363-377.
4. Benabdelhak, H., S. Kiontke, C. Horn, R. Ernst, M. A. Blight, I. B. Holland, and L. Schmitt. 2003. A specific interaction between the NBD of the ABC-transporter HlyB and a C-terminal fragment of its transport substrate hemolysin A. J. Mol. Biol. 327:1169-1179.
5. Binet, R., S. Letoffe, J. M. Ghigo, P. Delepelaire, and C. Wandersman. 1997. Protein secretion by gram-negative bacterial ABC exporters: a review. Gene 192:7-11.
6. Boyd, E. F., and D. L. Hartl. 1998. Chromosomal regions specific to pathogenic isolates of Escherichia coli have a phylogenetically clustered distribution. J. Bacteriol. 180:1159-1165.
7. Brandtzaeg, P., A. Bjerre, R. Ovstebo, B. Brusletto, G. B. Joo, and P. Kierulf. 2001. Neisseria meningitidis lipopolysaccharides in human pathology. J. Endotoxin Res. 7:401-420.
8. Felmlee, T., S. Pellett, E. Y. Lee, and R. A. Welch. 1985. Escherichia coli hemolysin is released extracellularly without cleavage of a signal peptide. J. Bacteriol. 163:88-93.
9. Forman, S., I. Linhartova, R. Osicka, X. Nassif, P. Sebo, and V. Pelicic. 2003. Neisseria meningitidis RTX proteins are not required for virulence in infant rats. Infect. Immun. 71:2253-2257.
10. Frey, J., and P. Kuhnert. 2002. RTX toxins in Pasteurellaceae. Int. J. Med. Microbiol. 292:149-158.
11. Glaser, P., H. Sakamoto, J. Bellalou, A. Ullmann, and A. Danchin. 1988. Secretion of cyclolysin, the calmodulin-sensitive adenylate cyclase-haemolysin bifunctional protein of Bordetella pertussis. EMBO J. 7:3997-4004.
12. Grifantini, R., S. Sebastian, E. Frigimelica, M. Draghi, E. Bartolini, A. Muzzi, R. Rappuoli, G. Grandi, and C. A. Genco. 2003. Identification of iron-activated and -repressed Fur-dependent genes by transcriptome analysis of Neisseria meningitidis group B. Proc. Natl. Acad. Sci. USA 100:9542-9547.
13. Hadi, H. A., K. G. Wooldridge, K. Robinson, and D. A. Ala'Aldeen. 2001. Identification and characterization of App: an immunogenic autotransporter protein of Neisseria meningitidis. Mol. Microbiol. 41:611-623.
14. Hopkinson, A., R. S. McIntosh, R. Layfield, J. Keyte, P. J. Tighe, and H. S. Dua. 2005. Optimised two-dimensional electrophoresis procedures for the protein characterisation of structural tissues. Proteomics doi:10.1002/pmic.200401073.
15. Jacob-Dubuisson, F., C. Buisine, N. Mielcarek, E. Clement, F. D. Menozzi, and C. Locht. 1996. Amino-terminal maturation of the Bordetella pertussis filamentous haemagglutinin. Mol. Microbiol. 19:65-78.
16. Klee, S. R., X. Nassif, B. Kusecek, P. Merker, J.-L. Beretti, M. Achtman, and C. R. Tinsley. 2000. Molecular and biological analysis of eight genetic islands that distinguish Neisseria meningitidis from the closely related pathogen Neisseria gonorrhoeae. Infect. Immun. 68:2082-2095.
17. Koronakis, V., A. Sharff, E. Koronakis, B. Luisi, and C. Hughes. 2000. Crystal structure of the bacterial membrane protein TolC central to multidrug efflux and protein export. Nature 405:914-919.
18. Lally, E. T., R. B. Hill, I. R. Kieba, and J. Korostoff. 1999. The interaction between RTX toxins and target cells. Trends Microbiol. 7:356-361.
19. Lee, V. T., and O. Schneewind. 2001. Protein secretion and the pathogenesis of bacterial infections. Genes Dev. 15:1725-1752.
20. Letoffe, S., P. Delepelaire, and C. Wandersman. 1996. Protein secretion in gram-negative bacteria: assembly of the three components of ABC protein-mediated exporters is ordered and promoted by substrate binding. EMBO J. 15:5804-5811.
21. Locht, C., M. C. Geoffroy, and G. Renauld. 1992. Common accessory genes for the Bordetella pertussis filamentous hemagglutinin and fimbriae share sequence similarities with the papC and papD gene families. EMBO J. 11:3175-3183.
22. Lomholt, H., K. Poulsen, and M. Kilian. 1995. Comparative characterization of the iga gene encoding IgA1 protease in Neisseria meningitidis, Neisseria gonorrhoeae and Haemophilus influenzae. Mol. Microbiol. 15:495-506.
23. Osicka, R., J. Kalmusova, P. Krizova, and P. Sebo. 2001. Neisseria meningitidis RTX protein FrpC induces high levels of serum antibodies during invasive disease: polymorphism of frpC alleles and purification of recombinant FrpC. Infect. Immun. 69:5509-5519.
24. Osicka, R., K. Prochazkova, M. Sulc, I. Linhartova, V. Havlicek, P. Sebo, S. Forman, X. Nassif, V. Pelicic, J. Kalmusova, and P. Krizova. 2004. A novel "clip-and-link" activity of repeat in toxin (RTX) proteins from gram-negative pathogens. J. Biol. Chem. 279:24944-24956.
25. Parkhill, J., M. Achtman, K. D. James, S. D. Bentley, C. Churcher, S. R. Klee, G. Morelli, D. Basham, D. Brown, T. Chillingworth, R. M. Davies, P. Davis, K. Devlin, T. Feltwell, N. Hamlin, S. Holroyd, K. Jagels, S. Leather, S. Moule, K. Mungall, M. A. Quail, M. A. Rajandream, K. M. Rutherford, M. Simmonds, J. Skelton, S. Whitehead, B. G. Spratt, and B. G. Barrell. 2000. Complete DNA sequence of a serogroup A strain of Neisseria meningitidis Z2491. Nature 404:502-506.
26. Prentki, P., and M. Krisch. 1984. In vitro insertional mutagenesis with a selectable DNA fragment. Gene 29:303-3113.
27. Prochazkova, K., R. Osicka, I. Linhartova, P. Halada, M. Sulc, and P. Sebo. 2005. The Neisseria meningitidis outer membrane lipoprotein FrpD binds the RTX protein FrpC. J. Biol. Chem. 280:3251-3258.
28. Robinson, K., M. Taraktsoglou, K. S. J. Rowe, K. G. Wooldridge, and D. A. A. Ala'Aldeen. 2004. Secreted proteins from Neisseria meningitidis mediate differential human gene expression and immune activation. Cell Microbiol. 6:927-938.
29. Schlor, S., A. Schmidt, E. Maier, R. Benz, W. Goebel, and I. Gentschev. 1997. In vivo and in vitro studies on interactions between the components of the hemolysin (HlyA) secretion machinery of Escherichia coli. Mol. Gen. Genet. 256:306-319.
30. Serruto, D., J. Adu-Bobie, M. Scarselli, D. Veggi, M. Pizza, R. Rappuoli, and B. Arico. 2003. Neisseria meningitidis App, a new adhesin with autocatalytic serine protease activity. Mol. Microbiol. 48:323-334.
31. Tettelin, H., N. J. Saunders, J. Heidelberg, A. C. Jeffries, K. E. Nelson, J. A. Eisen, K. A. Ketchum, D. W. Hood, J. F. Peden, R. J. Dodson, W. C. Nelson, M. L. Gwinn, R. DeBoy, J. D. Peterson, E. K. Hickey, D. H. Haft, S. L. Salzberg, O. White, R. D. Fleischmann, B. A. Dougherty, T. Mason, A. Ciecko, D. S. Parksey, E. Blair, H. Cittone, E. B. Clark, M. D. Cotton, T. R. Utterback, H. Khouri, H. Qin, J. Vamathevan, J. Gill, V. Scarlato, V. Masignani, M. Pizza, G. Grandi, L. Sun, H. O. Smith, C. M. Fraser, E. R. Moxon, R. Rappuoli, and J. C. Venter. 2000. Complete genome sequence of Neisseria meningitidis serogroup B strain MC58. Science 287:1809-1815.
32. Thompson, S. A., and P. F. Sparling. 1993. The RTX cytotoxin-related FrpA protein of Neisseria meningitidis is secreted extracellularly by meningococci and by HlyBD+ Escherichia coli. Infect. Immun. 61:2906-2911.
33. Thompson, S. A., L. L. Wang, and P. F. Sparling. 1993. Cloning and nucleotide sequence of frpC, a second gene from Neisseria meningitidis encoding a protein similar to RTX cytotoxins. Mol. Microbiol. 9:85-96.
34. Thompson, S. A., L. L. Wang, A. West, and P. F. Sparling. 1993. Neisseria meningitidis produces iron-regulated proteins related to the RTX family of exoproteins. J. Bacteriol. 175:811-818.
35. Turner, D. P., K. G. Wooldridge, and D. A. Ala'Aldeen. 2002. Autotransported serine protease A of Neisseria meningitidis: an immunogenic, surface-exposed outer membrane and secreted protein. Infect. Immun. 70:4447-4461.
36. van Ulsen, P., L. van Alphen, J. ten Hove, F. Fransen, P. van der Ley, and J. Tommassen. 2003. A neisserial autotransporter NalP modulating the processing of other autotransporters. Mol. Microbiol. 50:1017-1030.
37. van Vliet, A. H. M., K. G. Wooldridge, and J. M. Ketley. 1998. Iron-responsive gene regulation in a Campylobacter jejuni fur mutant. J. Bacteriol. 180:5291-5298.
38. Waldor, M. K., E. J. Rubin, G. D. Pearson, H. Kimsey, and J. J. Mekalanos. 1997. Regulation, replication, and integration functions of the Vibrio cholerae CTXphi are encoded by region RS2. Mol. Microbiol. 24:917-926.
39. Wandersman, C., and P. Delepelaire. 1990. TolC, an Escherichia coli outer membrane protein required for hemolysin secretion. Proc. Natl. Acad. Sci. USA 87:4776-4780.
40. Winzer, K., Y. H. Sun, A. Green, M. Delory, D. Blackley, K. R. Hardie, T. J. Baldwin, and C. M. Tang. 2002. Role of Neisseria meningitidis luxS in cell-to-cell signaling and bacteremic infection. Infect. Immun. 70:2245-2248.
41. Wooldridge, K. G., and P. H. Williams. 1993. Iron uptake mechanisms of pathogenic bacteria. FEMS Microbiol. Rev. 12:325-348.(Karl G. Wooldridge, Murat)
ABSTRACT
Proteins secreted by Neisseria meningitidis are thought to play important roles in the pathogenesis of meningococcal disease. These proteins include the iron-repressible repeat-in-toxin (RTX) exoprotein FrpC. Related proteins in other pathogens are secreted via a type I secretion system (TOSS), but such a system has not been demonstrated in N. meningitidis. An in silico search of the group B meningococcal genome suggested the presence of a uniquely organized TOSS. Genes encoding homologs of the Escherichia coli HlyB (ATP-binding), HlyD (membrane fusion), and TolC (outer membrane channel) proteins were identified. In contrast to the cistronic organization of the secretion genes in most other rtx operons, the hlyD and tolC genes were adjacent but unlinked to hlyB; neither locus was part of an operon containing genes encoding putative TOSS substrates. Both loci were flanked by genes normally associated with mobile genetic elements. The three genes were shown to be expressed independently. Mutation at either locus resulted in an inability to secrete FrpC and a related protein, here called FrpC2. Successful complementation of these mutations at an ectopic site confirmed the observed phenotypes were caused by loss of function of the putative TOSS genes. We show that genes scattered in the meningococcal genome encode a functional TOSS required for secretion of the meningococcal RTX proteins.
INTRODUCTION
During invasive meningococcal disease large amounts of lipooligosaccharide are released from the bacterial surface leading to massive inflammation and pathology (7). The contribution of secreted proteins to meningococcal pathogenesis, however, is not well understood. Exoproteins play a major role in the pathogenesis of many gram-negative pathogens and we have recently demonstrated that meningococcal secreted proteins (MSPs) play a central role in modulation of host cell genes with a role in inflammation and apoptosis (28). Proteins may be secreted by gram-negative bacteria by one of a number of known pathways, including those listed as types I to V (19). We and others have shown that meningococci secrete a number of proteins of a class called autotransporters, which are secreted via the type V pathway. These include the putative virulence factors immunoglobulin A (IgA) protease (22), the adhesin App (13, 30), AutA (1), and the serine protease AspA (NalP) (35, 36). Meningococci also have genes with homology to fhaB and fhaC of Bordetella pertussis (21). In this pathogen the major virulence factor filamentous hemagglutinin (product of fhaB) is secreted via the two-partner secretion pathway (a variant of the type V pathway) with the help of the fhaC gene product (15). Meningococci are also known to secrete proteins with homology to the repeat-in-toxin (RTX) family of cytotoxins (33, 34). In other gram-negative pathogens such toxins are secreted via a type I secretion system (TOSS) (5), and the recombinant product of one of the meningococcal genes was shown to be secreted via the Escherichia coli hemolysin TOSS (32). The presence of a TOSS in the meningococcus, however, has not been demonstrated.
Type I protein secretion machines are a subset of ABC transporters and comprise three components: an inner membrane ATP-binding protein, an outer membrane channel protein, and a membrane fusion protein. The membrane fusion protein spans the periplasm and contacts both the ATP-binding protein and the outer membrane protein (2, 29). The E. coli hemolysin secretion system is considered to be the prototype TOSS and is composed of the HlyB and HlyD proteins, which are the ATP-binding protein and membrane fusion protein components, respectively, and the outer membrane TolC protein (39). By searching the published meningococcal genome sequences for genes encoding proteins with similarity to E. coli HlyB and HlyD, we identified two noncontiguous sequences encoding proteins with similarity to each of these proteins. Furthermore, the hlyD homolog was adjacent to a homolog of tolC, which is unlinked to the hemolysin secretion genes of E. coli. Here, we show that meningococcal hlyB and hlyD homologs encode components of a functional type I protein secretion machine required for secretion of the meningococcal RTX toxin-like exoproteins. The genetic organization of the meningococcal genes is unlike that of any reported TOSS.
MATERIALS AND METHODS
Bacterial strains and growth conditions. Bacterial strains used in the present study are listed in Table 1. E. coli strains were grown at 37°C in Luria-Bertani broth with agitation or on Luria-Bertani agar supplemented, where appropriate, with ampicillin (50 μg/ml), kanamycin (50 μg/ml), or erythromycin (50 μg/ml). Meningococcal cells were routinely cultured on chocolate agar (Oxoid) at 37°C in 5% CO2. For selection of mutants, meningococcal cells were cultured on Mueller-Hinton agar plates supplemented with Vitox (Oxoid) and, where appropriate, streptomycin and spectinomycin (each at 100 μg/ml), kanamycin (100 μg/ml), or erythromycin (6 μg/ml). Liquid cultures were grown in Mueller-Hinton broth supplemented with Vitox at 37°C with agitation. For preparation of whole-cell extracts, RNA, and concentrated cell-free supernatants, cells were cultured without agitation at 37°C in Dulbecco modified Eagle medium supplemented with 50 μM desferrioxamine (Sigma) or 40 μM FeCl2 (to achieve iron-rich conditions for semiquantitative reverse transcription-PCR [RT-PCR]) in an atmosphere of 5% CO2.
Preparation of secreted proteins and whole-cell lysates. Meningococcal cell cultures grown overnight in 50 ml of Dulbecco modified Eagle medium supplemented with desferrioxamine were centrifuged at 11,000 x g for 20 min. For preparation of concentrated cell-free supernatants, the supernatants were filtered through 0.2-μm-pore-size syringe filters (Sartoris) and concentrated 100-fold by ultrafiltration with a Vivaspin-20 protein concentrator (30,000 molecular weight cutoff; Vivascience) by centrifugation at 2,000 x g until the volume was reduced sufficiently. For whole-cell lysate preparation, cell pellets were washed in 1 ml of ice-cold phosphate-buffered saline (PBS; Oxoid) and resuspended in a further 1 ml of ice-cold PBS. A total of 200 μl of this suspension was transferred to a fresh tube, and cells were harvested by centrifugation at 12,000 x g. After the pellet was suspended in 80 μl of ice-cold water, 20 μl of 5x sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer (0.625 M Tris-HCl [pH 6.8], 5% SDS, 25% glycerol, 12.5% -mercaptoethanol, 0.002% bromophenol blue) was added, and the suspension was mixed and then heated to 95°C for 5 min. The remaining cell suspension was used to prepare RNA (see below).
Protein assays, SDS-PAGE, and immunoblotting. Protein concentration was determined by using Protein Assay Reagent (Bio-Rad) according to the manufacturer's instructions. Preparations of whole-cell lysates or concentrated cell-free supernatants were separated on 10% polyacrylamide gels containing SDS and visualized by using SimplyBlue SafeStain (Invitrogen) according to the manufacturer's recommendations. For immunoblotting, proteins were transferred to nitrocellulose membranes (Schleicher & Schuell) by using a Trans-Blot SD semidry transfer cell (Bio-Rad) according to the manufacturer's recommendations. After being blocked overnight in a solution of PBS containing 5% dried skimmed milk (blocking solution), the membranes were incubated for 1 h at room temperature with monoclonal antibody A4.85 diluted 1:50 in blocking solution. After three 5-min washes in PBS containing 0.1% Tween 20 (Sigma), membranes were incubated in goat anti-mouse immunoglobulin conjugated to horseradish peroxidase (Bio-Rad) diluted 1:2,000 in blocking solution. After being washed as described above, membranes were developed by using either 4-chloro-1-naphthol or with the ECL Western Blotting Analysis System (Amersham).
DNA manipulation. Chromosomal DNA was prepared by using a DNeasy tissue kit (QIAGEN) using the protocol for bacterial cells recommended by the manufacturer. Plasmid DNA was prepared by using a QIAprep spin kit (QIAGEN) according to the manufacturer's recommendations. Restriction enzymes were purchased from Kramel Biotech, New England Biolabs, or Fermentas and used according to the directions of the manufacturer. T4 DNA ligase was purchased from Boehringer Mannheim. Expand Taq DNA polymerase (Roche Diagnostics GmbH) was used in all PCRs. Unless otherwise stated, PCRs contained 100 ng of chromosomal DNA or 1 ng of plasmid DNA. All reactions contained each primer at 200 nM; 200 μM concentrations each of dATP, dCTP, dGTP, and dTTP; 1x Expand polymerase buffer (including MgCl2 to a final concentration of 1.5 mM), and 0.6 U of Expand polymerase in a total volume of 25 μl. After initial incubation at 95°C for 3 min, reactions were subjected to 30 cycles of incubation at 50°C for 1 min, 68°C for between 1 and 6 min, and 94°C for 45 s, with final incubations at 50°C for 1 min and 68°C for 10 min. Products of restriction digestion reactions and PCRs were analyzed by agarose gel electrophoresis.
RNA extraction and RT-PCR. Meningococcal cells from the same overnight cultures used to prepare secreted protein preparations and whole-cell lysates (see above) were pelleted by centrifugation at 2,000 x g and resuspended in 1 ml of TRIzol reagent (Invitrogen). After incubation at room temperature for 30 min, 200 μl of chloroform was added to each suspension, and the mixtures were vortexed for 15 s. Samples were incubated for a further 15 min before centrifugation at 12,000 x g at 4°C. The upper aqueous layer was removed and added to 1.4 ml of RLT buffer (a component of the RNeasy kit; QIAGEN) and 1 ml of ethanol. Mixtures were then processed by using an RNeasy kit according to the manufacturer's recommendations. RNA was quantified by using a NanoDrop spectrophotometer (NanoDrop Technologies). To remove contaminating genomic DNA, RNA samples were treated with RQ1 RNase-free DNase (Promega) according to the manufacturer's instructions.
For RT-PCR, mixtures containing 5 μg of RNA, reverse primer at 0.5 μM, and 1 mM concentrations each of dATP, dCTP, dGTP, and dTTP in a final volume of 10 μl were incubated at 65°C for 5 min and placed on ice for 5 min. To each reaction, 2 μl of 10x reaction buffer (supplied with Superscript II [Invitrogen]), 1 μl of RNAseOUT (Invitrogen), MgCl2 to 50 mM, and dithiotheitol (supplied with Superscript II) to 10 mM were added in a premix, bringing the total volume of each reaction to 19 μl. After incubation at 42°C for 2 min, 1 μl of Superscript II was added to each tube, and reactions were incubated at 42°C for 50 min and 70°C for 15 min and then returned to ice. Prior to PCR amplification, 1 μl of RNase H (supplied with Superscript II) was added to each mixture, and reactions were incubated at 37°C for 20 min. Then, 1 μl of each reaction was used in subsequent PCRs. The primers HlyB-RTF and HlyB-RTR were used to detect the hlyB transcript, the primers HlyD-RTF and HlyD-RTR were used to detect the hlyD transcript, the primers TolC-RTF and TolC-RTR were used to amplify the tolC transcript, and the primers HlyD-RTF and TolC-RTR were used to attempt to detect a transcript spanning hlyD and tolC. For semiquantitative RT-PCR mixtures, the reverse primer gdh-RTR was added, in addition to the gene-specific reverse primer, to cDNA synthesis reactions. The primers gdh-RTF and gdh-RTR were added after five initial cycles of subsequent PCRs, in which only gene-specific primers were present, since when both primer sets were added at the beginning the stronger gdh signal overwhelmed the much weaker hlyB, hlyD, and tolC amplicons. The relative intensities of bands in RT-PCR experiments were determined by densitometry using the ImageJ program (version 1.34i; 1997-2005, Wayne Rasband, National Institutes of Health, Bethesda, Maryland [http://rsb.info.nih.gov/ij/]). All primer sequences are shown in Table 2.
Mutagenesis and complementation. The hlyB gene (NMB1400), together with 830-bp upstream of this open reading frame (ORF), was amplified from chromosomal DNA of strain MC58 by using the primers HlyB-For and HlyB-Rev. This amplicon was cloned into the plasmid pGEM-T Easy (Promega) and subjected to inverse PCR mutagenesis with the primers HlyB-M1 and HlyB-M2. This resulted in the deletion of the first 1,351-bp of the ORF and the introduction of a unique BglII site. This site was used to introduce the omega cassette (encoding resistance to spectinomycin and streptomycin) digested with BamHI from plasmid pHP45 (26). The resulting plasmid was used to mutate the meningococcal strain MC58 by natural transformation and allelic exchange as described previously (13). The deletion in the resulting mutant (MC58hlyB) was confirmed by PCR analysis.
In order to complement the hlyB mutation, hlyB was amplified by using the downstream primer HlyB-ECR and either of the two upstream primers HlyB-ECF1 or HlyB-ECF2 (to amplify the long and short forms of hlyB, respectively), cut with AscI and BamHI, and cloned into the vector pYHS25 cut with BamHI and MluI. The resulting plasmid was used to transform cells of the hlyB mutant strain MC58hlyB. Insertion of the gene at the ectopic site was confirmed by PCR analysis with the primers EV-F and EV-R.
The hlyD homolog (NMB1738) was amplified by using the primers HlyD-For and HlyD-Rev and cloned into pBluescript. The resulting plasmid was subjected to inverse-PCR with the primers HlyD-M1 and HlyD-M2, digested with BclI, and ligated to the kanamycin resistance cassette digested with BamHI from plasmid pJMK30 (37). The resulting plasmid, in which the entire ORF had been replaced with the kanamycin resistance cassette in the opposite orientation to the replaced gene, was used to mutate strain MC58 (Fig. 4B). The deletion in the resulting mutant (MC58hlyD) was confirmed by PCR analysis. In order to complement the hlyD mutation, a region of DNA containing the hlyD and tolC genes was amplified by using the primers HlyD&tolC-ECF and HlyD&tolC-ECR. This fragment was cut with AscI and either BamHI (to clone hlyD alone) or BglII (to clone hlyD and tolC). Both fragments were separately cloned into the vector pYHS25 cut with BamHI and MluI. The resulting plasmids were used to transform cells of the hlyD mutant strain MC58hlyB. Insertion of the gene at the ectopic site was confirmed by PCR analysis by using the primers EV-F and EV-R.
The frpC gene was amplified by using primers FrpCM1 and FrpCM2 and cloned into plasmid pGEM-T Easy. The construct was digested with BglII and ligated to the omega cassette digested with BamHI (Fig. 4C). The resulting construct, in which the frpC gene had been disrupted, was used to mutate strain MC58. The mutation in the resulting strain (MC58frpC) was confirmed by PCR analysis.
MS analysis of protein bands. Individual protein bands excised from SDS-PAGE gels stained with SimplyBlue SafeStain (Invitrogen) were destained and prepared as described previously for MS analysis (14). Peptide mixtures eluted from each band were analyzed by MALDI-TOF-MS (Waters/Micromass, Manchester, United Kingdom), and protein identification was deduced from peptide mass fingerprints by searching relevant databases using PeptIdent (http://www.expasy.org/tools/peptident.html). Where peptide identification was not possible by this procedure, samples were subjected to peptide sequence analysis by nano-ESI-Q-TOF-MS/MS (Waters/Micromass), and the sequences obtained were used to search for matches using the BLAST algorithm at http://www.expasy.org/tools/blast/.
Protein and nucleic acid sequence analysis. Public databases of published protein and DNA sequences were searched by using the BLAST programs available at http://www.ncbi.nih.gov/BLAST/ and http://tigrblast.tigr.org/cmr-BLAST/. The genome databases of meningococcal strains Z2491 and FAM18, and the N. lactamica strain were interrogated by using the BLAST servers available at the Sanger Institute (www.sanger.ac.uk). Genome databases of strain MC58 were interrogated by using the BLAST server available at The Institute for Genomic Research (http://www.tigr.org). The genome sequence database of the gonococcal strain FA1090 was interrogated by using the BLAST servers available at www.ncbi.nlm.nih.gov/sutils/genom_table.cgi. Searches for Prosite patterns were made by using the server available at http://us.expasy.org/prosite/. Other DNA and protein sequence analysis was carried out by using the DNAMAN package of programs (Lynnon BioSoft).
RESULTS
The meningococcal genome contains evidence of a type I protein secretion system. The published genomes of group A and group B meningococcal strains Z2491 (25) and MC58 (31), respectively, were searched for genes encoding proteins with similarity to HlyB and HlyD of E. coli (accession number M10133) (8). An ORF encoding a predicted protein of 742 amino acids (aa) that was 70% identical to E. coli HlyB over 678 aa was detected in both genomes (NMA1620 and NMB1400, respectively) (Fig. 1A). A gene with similarity to the E. coli hlyD gene was also found in both genomes (NMA1996 and NMB1738, respectively). In the genome of strain Z2491 this gene contained a frameshift mutation but the equivalent gene in strain MC58 was intact and predicted to encode a protein of 475 aa with 39% identity to E. coli HlyD over 468 aa (Fig. 1B). We also searched both meningococcal genomes with the E. coli TolC protein sequence (accession number NP_755652) and found two ORFs whose predicted protein products had significant similarity: NMB1737 (immediately downstream of hlyD) was the closest match (27.6% identity; P = 2.0 x 10–31) (Fig. 1C); the mtrE gene (NMB1714), which encodes a putative multidrug efflux pump, also had significant similarity (20% identity; P = 4.7 x 10–6). The latter gene was not investigated further. The meningococcal hlyB-, hlyD-, and tolC-like genes are here called hlyB, hlyD, and tolC, respectively. Neither the meningococcal hlyB nor the hlyD/tolC loci were cistronic with any genes with homology to hlyA, including the meningococcal frpC gene, or to a second gene with close similarity to frpC (here called frpC2). The hlyB gene, however, was adjacent to a region of DNA containing frpC (NMB1415), a number of small frpC homologs (NMB1403, NMB1405, NMB1407, and NMB1409), the frpC operon gene frpD (NMB1414), and an additional frpD homolog (NMB1412) (Fig. 2).
The meningococcal hlyB and hlyD/tolC loci are flanked by genes associated with horizontal DNA transfer. The hlyB gene is flanked by the insertion sequences IS1106 and IS1016 (Fig. 2). Downstream of the hlyD/tolC locus there is a pseudogene with homology to a third class of insertion sequences (NMB1736); upstream of this locus are two short ORFs encoding hypothetical proteins of unknown function and a gene homologous to the rstA2 gene of the Vibrio cholerae phage CTXphi (NMB1741; Fig. 2) (38). The G+C content of the hemolysin genes in E. coli is lower than the average value for the E. coli genome (6). To determine whether the composition of DNA encoding the meningococcal hlyB and hlyD/tolC loci differed significantly from the overall composition of the meningococcal genome, we calculated the G+C content of the region bracketed between the IS elements (in the case of hlyB) and between NMB1736 and NMB1741 (including the hlyD and tolC genes). The G+C contents of these regions were 49.2 and 50%, respectively, compared to an overall composition of 51.8% for the entire genome.
The meningococcal hlyB, hlyD, and tolC genes are expressed. To determine whether the meningococcal hlyB, hlyD, and tolC genes are expressed, we analyzed RNA from strain MC58 by RT-PCR using primer pairs specific for each gene in turn (Fig. 3). Transcription of all three genes was detected by using gene-specific primer pairs (lane 2 in each of Fig. 3A to C). Negative control reactions, in which reverse transcriptase was omitted, did not yield equivalent products (lane 6 in Fig. 3A to C), indicating that the amplification products were derived from RNA rather than contaminating DNA.
N. meningitidis possesses a functional type I secretion machine. In order to determine whether the putative meningococcal hlyB and hlyD/tolC loci encode components of a type I secretion machine, each locus was mutated in turn. The hlyB gene (NMB1400) of strain MC58, together with 830 bp of upstream DNA was cloned and subjected to inverse PCR mutagenesis, resulting in the deletion of the first 1,351 bp of the ORF and the introduction of a unique BglII site. The omega cassette, encoding resistance to streptomycin and spectinomycin, was introduced into this site, and the resulting plasmid was used to mutate strain MC58 by natural transformation and allelic exchange (Fig. 4A). The deletion in the resulting mutant (MC58hlyB) was confirmed by PCR analysis (not shown). The mutant was tested for expression of hlyB by RT-PCR and shown not to express the gene, as expected (Fig. 3A, lane 3).
The hlyD homolog (NMB1738) was cloned and subjected to inverse PCR mutagenesis. After introduction of a kanamycin resistance cassette, the resulting plasmid, in which the entire ORF had been deleted, was used to mutate strain MC58 (Fig. 4B). The deletion in the resulting mutant (MC58hlyD) was confirmed by PCR analysis (not shown). The mutant was tested for expression of hlyD by RT-PCR and shown not to express the gene, as expected (Fig. 3B, lane 3).
Concentrated cell-free supernatants of wild-type MC58, MC58hlyB, and MC58hlyD cultures containing secreted proteins were prepared and analyzed by SDS-PAGE. Proteins with apparent molecular masses of 200 and 135 kDa were detectable in the supernatants of wild-type MC58 cells but absent from the supernatants of both mutants (Fig. 5). These data indicate that the products of the meningococcal hlyB and hlyD genes are required for secretion at least two exoproteins.
The endogenous RTX-containing exoproteins are substrates of the meningococcal type I secretion machine. Two RTX-containing proteins have been previously reported to be secreted by the group C meningococcal strain FAM20. These are the closely related iron-regulated proteins FrpA (122 kDa) and FrpC (197 kDa) (32) (Table 3). The genome of strain MC58 contains a number ORFs with similarity to the frpA and frpC genes of strain FAM20 (Table 3); the largest two of which (NMB1415 and NMB0585) encode proteins containing the hemolysin-type calcium-binding region signature sequence (prosite number PDOC00293). NMB1415 corresponds to the frpC gene of strain FAM20, while NMB0585 (here called frpC2) encodes a smaller but closely related protein. The predicted molecular masses of the two proteins (197 and 141 kDa, respectively) are in good agreement with the apparent molecular masses of the proteins present in culture supernatants of the wild-type strain MC58 but absent from culture supernatants of the mutant derivative strains MC58hlyB and MC58hlyD. To determine whether these proteins were the products of the frpC and frpC2 genes, we probed concentrated cell-free culture supernatants of MC58 and MC58hlyB with the monoclonal antibody A4.85. This antibody was raised against meningococcal outer membrane proteins and specifically recognizes the RTX proteins of strain FAM20 (34). Both proteins were detected by the antibody (Fig. 6 and 7), strongly suggesting that they represent the products of frpC and the related frpC2 gene. In addition to the major bands observable in Coomassie-stained gels at 200 and 135 kDa, additional bands with apparent molecular masses of 175 and 85 kDa were detected by the antibody. All of these bands were detected in whole-cell lysates of MC58, as well as the MC58hlyB and MC58hlyD mutant strains (Fig. 7B), indicating that, whereas secretion of these proteins was dependent on the expression of the hlyB and hlyD gene products, their expression was not significantly affected.
To further confirm the identities of the A4.85-reactive bands as FrpC and FrpC2 and to determine which bands corresponded to each gene, we attempted to mutate both genes. The frpC gene was cloned and disrupted by insertion of the omega cassette (Fig. 4C), and the resulting construct was used to mutate strain MC58. The mutation in the resulting strain (MC58frpC) was confirmed by PCR analysis (not shown). Secreted protein preparations from this mutant lacked the 200- and 175-kDa proteins but still contained proteins of 135 and 85 kDa (Fig. 6), indicating that the 200- and 175-kDa proteins were products of the frpC gene. Several attempts to mutate the frpC2 locus were unsuccessful. To confirm the identity of the 135-kDa protein, a preparation of meningococcal secreted proteins was separated by SDS-PAGE, and a close doublet of bands at 135-kDa that was apparent in the wild-type strain but absent from the hlyB mutant strain was observed after staining with Coomassie (not shown). Both bands were excised and subjected to MALDI-TOF analysis. The lower of the two bands was identified as NMB0585 (FrpC2). The upper band produced ambiguous results and was subjected to peptide sequence analysis by nano-ESI-Q-TOF-MS/MS. This procedure identified a mixture of two proteins, one of which was also identified as FrpC2, the other being the previously reported 140-kDa secreted form of the meningococcal autotransporter App (13). The identity of the 85-kDa protein could not be confirmed in this way since it was only detectable by immunoblotting.
Mutations at the hlyB and hlyD/tolC loci can be complemented by functional copies of hlyB and hlyD at an ectopic site. In order to confirm that the observed phenotypes in the mutants MC58hlyB and MC58hlyD were the result of the loss of expression of hlyB and hlyD, respectively, rather than polar effects on neighboring genes or by unknown secondary mutations, we constructed new mutants in which the original mutations were complemented at an ectopic site on the chromosome. We used an ectopic complementation vector in which the gene of interest may be cloned downstream of the porB promoter and upstream of the ermC gene (encoding resistance to erythromycin). These sequences are flanked by the genes NMB102 and NMB103, which are in a head-to-head configuration in both the vector and the meningococcal genome (40). The construct facilitates insertion of genes of interest, under the control of the porB promoter, into the meningococcal genome at a site that is unlikely to effect the expression of any other genes.
A region of hlyB that corresponds to its E. coli counterpart, starting with the methionine at position 33 (hlyBs) and, separately, the entire hlyB ORF encoding an additional 32 aa at the N terminus (hlyBl) were cloned into the complementation vector and reintroduced into the mutant strain MC58hlyB. The resulting complemented mutant strains (MC58hlyB hlyBsEct and MC58hlyB hlyBlEct, respectively) were analyzed by RT-PCR with primers specific for hlyB. The ectopic hlyB genes were expressed in both complemented mutants (Fig. 3A, lanes 4 and 5). When concentrated cell-free supernatants of cultures of the complemented mutant MC58hlyB hlyBlEct were probed in immunoblotting experiments with the monoclonal antibody A4.85, all of the proteins detected in the wild-type supernatants were detected at somewhat higher levels than in similar preparations of wild-type cells (compare Fig. 7A, lanes 3 and 5). These data confirm that the original mutation of the hlyB gene was responsible for the observed secretion defect. They also suggests that, under the conditions in which the cells were grown, the level of hlyB expression in wild-type cells was the rate-limiting condition for secretion of the FrpC and FrpC2 proteins. When culture supernatants of the mutant complemented with the shorter form of hlyB (MC58hlyB hlyBsEct) were probed with the same antibody, only very low levels of the major bands corresponding to FrpC and FrpC2 were observed (Fig. 7A, lane 4).
DNA fragments containing the hlyD gene alone or both hlyD and tolC genes were cloned separately into the ectopic vector and reintroduced into the hlyD mutant MC58hlyD. The expression of both genes was detected by RT-PCR in the cells of both complemented mutants (Fig. 3B and C, lanes 4 and 5). Immunoblotting experiments demonstrated that secretion of FrpC and FrpC2 was restored in both strains (Fig. 7A, lanes 7 and 8), confirming that the hlyD/tolC locus is required for the efficient secretion of these proteins. Levels of secretion by strain MC58hlyD hlyDEct were similar to those of the wild-type strain. In contrast, only a low level of expression was detected in strain MC58hlyD hlyDEct tolCEct.
Detectable amounts of FrpC and FrpC2 were not released by cell lysis. Cell-free concentrated supernatant fractions used in our experiments clearly contained secreted proteins. Although we cannot rule out the possibility that they also contain contaminating proteins derived from intracellular compartments as a result of cell lysis, we were unable to detect FrpC or FrpC2 in preparations prepared from the hlyB and hlyD mutants (Fig. 7A). These proteins were readily detected in whole-cell preparations of the same cells (Fig. 7B), indicating that FrpC and FrpC2 proteins detected in concentrated cell-free supernatants were the substrates of the type I secretion system and were not released nonspecifically from intracellular compartments by cell lysis.
The tolC gene is expressed independently of hlyD. The meningococcal hlyD and tolC genes are adjacent, in the same orientation, and separated by 269 bp. In order to determine whether the two genes were components of an operon, RNA from cells of the wild-type strain, the hlyD mutant MC58hlyD, and from this mutant ectopically complemented with either hlyD alone (MC58hlyD hlyDEct) or hlyD and tolC (MC58hlyD hlyDEct tolCEct) was subjected to RT-PCR with either a tolC-specific primer pair or a primer pair spanning the hlyD and tolC genes. Expression of tolC was detected in all four strains (Fig. 3C), but we were unable to detect a transcript spanning both genes in any strain (Fig. 3D). In strain MC58hlyD the hlyD gene had been completely replaced with a kanamycin resistance cassette in the opposite orientation with respect to the tolC gene; the gene upstream of the kanamycin cassette insertion is also transcribed in the opposite orientation with respect to tolC (Fig. 2). It is highly unlikely, therefore, that tolC could be transcribed from a promoter either within or upstream of the insertion site in strain MC58hlyD. These data indicate that tolC and hlyD are independently expressed.
Expression of the meningococcal hlyB, hlyD, and tolC genes is not iron repressible. The meningococcal RTX toxins are expressed at higher levels under conditions of iron limitation (34). To test whether the genes encoding their secretion apparatus were also iron repressible, RNA was prepared from meningococcal cells grown under iron-rich and iron-limiting conditions and analyzed by semiquantitative RT-PCR to detect relative expression levels of the hlyB, hlyD, and tolC genes. The gdh gene, encoding glucose-6-phosphate dehydrogenase, was used as an internal control since this gene would not be expected to be regulated by iron levels. None of the three genes was significantly upregulated under conditions of low iron availability; indeed, transcription of hlyB was actually reduced by a moderate but significant amount under iron-limiting conditions (Fig. 8).
RTX toxin-related genes are conserved in group A, group B, and group C strains of N. meningitidis but are not present in the genomes of N. gonorrhoeae or N. lactamica. In order to further examine the distribution of frpC-like genes and related secretion genes in the genus Neisseria, we searched the genomes of the group C meningococcal strain FAM18 and N. lactamica (both sequenced at the Sanger Institute), and of the N. gonorrhoeae strain FA1090 (sequenced at The University of Oklahoma). Homologs of frpC were detected in strain FAM18, including one gene encoding a 238-kDa protein containing the repeats characteristic of RTX toxins (Table 3 and Fig. 9). Homologs of frpC could not be detected in the N. gonorrhoeae or N. lactamica genome sequences. Genes with strong (>99%) identity to the hlyB, hlyD, and tolC genes of strain MC58 were detected in the genome of strain FAM18 (NMC1342, NMC1658, and NMC1657, respectively, in the preliminary annotation of this genome); no sequences with significant homology to these genes were detected in the N. lactamica or N. gonorrhoeae genome sequences.
To determine whether the meningococcal genomes also contained genes with similarity to the toxin activator genes of typical RTX operons, we searched the meningococcal genomes with the E. coli HlyC protein sequence (accession number M10133). The genome of strain MC58 contained two genes (NMB1210 and NMB1763), which encoded predicted proteins with 40% (P = 5.7 x 10–23) and 34% (P = 7.7 x 10–19) identity, respectively, to the E. coli sequence. Corresponding genes were not found, however, in the genomes of strains Z2491 or FAM18.
DISCUSSION
The presence of a TOSS for the export of meningococcal RTX-containing exoproteins was predicted over a decade ago (32). The availability of the genome sequence of the group B meningococcal strain MC58 (25) has made it possible to identify likely components of such a system. Here, we have bioinformatically identified hlyB, hlyD, and tolC-like genes and demonstrated that they were required for RTX protein secretion.
Most RTX proteins are found as part of an operon taking the form rtxCABD (18), in which the first gene encodes an activator protein required for the fatty acid acylation of the rtxA toxin gene product. The rtxB and rtxD genes encode the ATP-binding protein and membrane fusion protein components, respectively, of the secretion system. In E. coli the tolC gene, which encodes the outer membrane component of both the hemolysin exporter and a distinct TOSS required for colicin V secretion, is genetically unlinked to the hlyCADB operon (39). In some other RTX operons, the genes encoding each of the three secretion system components are linked: the tolC-like cyaE gene of B. pertussis, which is required for secretion of the bifunctional adenylate cyclase/hemolysin, is downstream of the cyaABD operon (11). Similarly, a tolC-like gene is located downstream of the mbxCABD RTX operon of Moraxella bovis (3). The arrangement of genes encoding the meningococcal RTX proteins and their secretion apparatus is unique. In all other type I secretion systems, to our knowledge, the genes encoding the ATP-binding protein and membrane fusion protein components are contiguous and, in most systems, components of operons containing the genes encoding their secretion substrates. In contrast, the meningococcal hlyB and hlyD genes were not linked to each other. The hlyB gene was, however, adjacent to a region containing frpC and a number of smaller related genes (Fig. 2). This region also contained the frpD gene, which is cistronic with frpC and encodes an outer membrane protein with a high affinity for FrpC (27), as well as an additional frpD-like gene. The juxtaposition of hlyD and tolC is also unusual, although these genes were shown to be expressed independently. The observation that both loci are flanked by genes associated with mobile genetic elements is intriguing: in strain MC58 hlyB is flanked by transposases of the IS1106 and IS1016 families (Fig. 2), while the hlyD/tolC locus is flanked by a pseudogene with homology to a third class of transposases (NMB1736 in strain MC58) and a gene homologous to the rstA2 gene of the V. cholerae phage CTXphi (NMB1741; Fig. 2): a gene required for phage replication (38). This genetic arrangement is conserved in the group A and group C strains Z2491 and FAM18, respectively, although in the preliminary sequence of strain FAM18 the ORFs corresponding to IS1016 (downstream of hlyB) and the rstA2 homolog (downstream of tolC) both contain frameshifts. These observations suggest that the meningococcal genes may have been acquired by horizontal transfer, although the DNA composition (G+C content) of these genes is not significantly different from that of the meningococcal genome. Previously, the region containing the hlyD pseudogene in strain Z2491 and its neighboring tolC-like gene has been identified as one of eight species-specific DNA islands present in N. meningitidis but not in N. gonorrhoeae (16). The authors of that study suggested that these islands may have arisen by import of DNA from other species and recombination via homologous flanking DNA. Evidence of horizontal gene transfer of RTX operons of a number of species belonging to the Pasteurellaceae, as well as some other gram-negative bacteria, has led to the suggestion that RTX toxins may have originated within the this family (10).
The frpC gene is well conserved in different strains of meningococci, whereas the closely related frpA gene is much less well conserved (23). In strain MC58 a second gene, NMB0585, which is almost identical to frpC but which lacks a number of internal fragments, is also present in the genome. Here we named this gene frpC2 to indicate its close similarity to frpC and to distinguish it from a number of other, smaller, ORFs with similarity to frpC in the MC58 genome. These latter ORFs lack the characteristic repeats of classical RTX proteins and do not contain regions homologous to the C-terminal region, which in other systems is required for secretion via TOSSs (4) (Fig. 9 and Table 3). Similar ORFs are also present in the genome of strain Z2491, but this genome does not contain any ORFs of a size equivalent to frpC/frpC2 or frpA. It is also of note that in this strain the hlyD homolog (NMA1976) contained a 61-bp insertion containing a stop codon 1,311-bp after the ATG start codon, followed by a direct duplication of the 73-bp preceding the insertion. This would be predicted to result in the expression of a truncated protein lacking a key signature motif found in members of the HlyD family of proteins (Prosite PS00543) (20). As a result of the frameshift, this signature motif is present in the ORF immediately downstream of NMA1976. The absence of both an intact hlyD gene and complete frpC-related genes in this strain suggests that the lineage has lost the capacity to produce and secrete RTX exoproteins. Sequence analysis of the 3' end of the hlyD gene in 13 strains of N. meningitidis chosen to represent the major clonal lineages of this pathogen, and including one group A isolate, revealed that the insertion was absent from all of these strains (16). It would be of interest to know whether other group A isolates contain intact RTX exoproteins, as well as an intact type I secretion system, since this may have important implications for the pathogenesis of isolates belonging to this serogroup. It is likely that the genes encoding the meningococcal RTX proteins and their cognate secretion systems were acquired by the species N. meningitidis after the divergence of this species and the related pathogen N. gonorrhoeae, since there are no genes with close homology to any of these dispersed genes in the genomes of the latter species or that of N. lactamica.
The first 32 aa encoded by the meningococcal hlyB gene (NMB1400) are predicted to encode an N-terminal region not present in HlyB of E. coli. Furthermore, the first methionine residue of the E. coli sequence aligns with an internal methionine residue at position 33 in the meningococcal sequence (Fig. 1). This led us to speculate that the second ATG codon of the meningococcal ORF was the true start of the gene. When we expressed the shorter form of the gene at an ectopic site in the hlyB mutant strain, however, we were only able to detect very low levels of protein secretion. In contrast, when the complete ORF was used to complement the mutation, the levels of substrate protein secretion were even higher than the wild-type levels. Since both ectopic genes were transcribed, this observation suggests that the entire meningococcal ORF, including the 32-aa meningococcus-specific N terminus, are required for the efficient functioning of the meningococcal HlyB protein, although we cannot rule out the alternative hypothesis that the shorter form of the gene was not efficiently translated in the complemented mutant.
Complementation of the hlyD mutation with an ectopic copy of hlyD resulted in secretion of FrpC and FrpC2 at levels similar to those of wild-type cells. When we designed our complementation strategy, we postulated that expression of tolC may be disrupted in the hlyD mutant. For this reason, we also made a complementation construct that contained both hlyD and tolC. The results of RT-PCR analysis subsequently demonstrated that tolC expression was unaffected by the mutation in the hlyD gene. These data were consistent with the observation that complementation with hlyD alone resulted in efficient phenotypic complementation of the hlyD mutation. When both hlyD and tolC were introduced into the ectopic site, the resulting levels of FrpC/FrpC2 secretion were much lower than wild-type levels. It is likely that the altered stoichiometry in this complemented mutant, in which two functional copies of tolC were present, resulted in interference with the normal interactions between the different components of the secretion system.
The neisserial TolC homologue is shorter by 56 C-terminal amino acid residues than its E. coli counterpart (Fig. 1C), and this may reflect a functional difference between the two proteins. The crystal structure of the E. coli TolC protein has been solved to a high resolution (2.1 ) (17). TolC assembles into a homotrimer, in which each monomer contributes 4 of 12 beta strands that together form a channel through the outer membrane in a beta barrel conformation. Each monomer also contributes a number of alpha-helical regions that form a channel continuous with the beta barrel and projecting up to 100 into the periplasm. In order to crystallize the E. coli TolC protein, it was necessary to proteolytically remove the carboxy-terminal 43 residues (17). Moreover, it was reported that this truncation did not affect the function of TolC (17). TolC functions that were tested included hemolysin secretion, colicin sensitivity, and drug efflux (V. Koronakis, unpublished data). The region removed by proteolysis prior to crystallization is largely coincident with the C-terminal part of E. coli TolC that does not have a counterpart in N. meningitidis. Thus, the structure of this region is unknown. More importantly, it is known to be dispensable for at least some functions of the E. coli molecule.
Low iron levels have been known for many years to be used by bacteria as a signal for the upregulation of a number of virulence factor genes (41). The meningococcal RTX proteins were originally identified as being regulated by iron (34), and it might be expected, therefore, that the genes encoding the secretion machinery for the products of these genes might be coordinately regulated in response to the same signal. In an extensive microarray-based study of the iron-regulated genes of strain MC58, Grifantini et al. showed that frpC, the frpC-like genes NMB0365 and NMB1405, frpD, and the frpD-like genes NMB0364, NMB0584, and NMB1412 were all iron repressible (12). For some reason frpC2 was not identified in this screen, although the adjacent frpD homolog was shown to be iron regulated. The genes encoding the meningococcal hlyB, hlyD, or tolC were also not among the iron-repressible genes reported (12). Our results using semiquantitative RT-PCR confirm these findings and show that the meningococcal TOSS genes are not iron repressible. The significance of this is not known, but it is possible that since each secretion channel can process many exoprotein molecules it may be of less importance in regulating the genes encoding the channel components than those encoding the secreted proteins themselves.
Multiple bands were detected by the monoclonal antibody A4.85 in secreted protein preparations and whole-cell extracts. The secreted protein bands at 200 and 175 kDa were clearly the products of the frpC gene, since they were not observed in the frpC mutant strain MC58frpC. The band observed at 135-kDa was confirmed to be a product of the frpC2 gene by mass spectroscopy. It is likely that the 85-kDa band is also a product of this gene. The meningococcal RTX-containing exoproteins have recently been shown to belong to a new class of RTX proteins that are capable of calcium-dependent autoproteolytic cleavage and cross-linking. FrpC was shown to cleave the peptide bond between Asp414 and Pro415; the resulting amino-terminal fragment is capable of covalent linkage via its carboxy-terminal aspartate residue to the -amino group of an internal lysine residue of another FrpC molecule (24). It is likely that the 175-kDa FrpC-related band is the carboxy-terminal fragment of FrpC reported previously (24). The smaller of the two proteins detected in secreted protein preparations of MC58frpC (as well as wild-type cells) could be a similar cleavage product of FrpC2, since the cleavage site identified previously within FrpC is conserved in the smaller protein. Additional bands detected in whole-cell lysates might represent additional products of FrpC or FrpC2 resulting from autocleavage and/or cross-linking events, degradation products resulting from the activity of other intracellular proteases, or the products of the smaller frpC-related ORFs.
The family of RTX-containing proteins include pore-forming cytotoxins, metalloproteases, and lipases; they contribute to the virulence of a number of important pathogens (18). Evidence for a role for the meningococcal RTX-containing exoproteins in pathogenesis is thus far lacking. A mutant strain of N. meningitidis lacking functional frpC and frpC2 genes was not attenuated in an infant rat model of meningococcal septicemia (9). Similarly, a mutant of strain MC58, in which the meningococcal-specific DNA island containing the hlyD and tolC genes was deleted, was able to multiply in the infant rat bloodstream at levels similar to those of the wild-type parent strain (16). Furthermore, no target cells or receptors have thus far been identified for the meningococcal RTX proteins. It can also be concluded that the meningococcal RTX proteins are not always required to cause meningococcal disease: strain Z2491, which has lesions in both RTX protein genes and in the hlyD secretion gene, was isolated from a case of meningitis (assuming the genetic lesions in this strain did not accumulate during laboratory passage). However, RTX protein genes have been shown to be present in a wide range of group B and group C meningococcal lineages, and high levels of serum antibodies have been detected in patients suffering from invasive meningococcal disease (23). The hlyD and tolC genes have also been shown to be universally present in a range of isolates chosen to represent the major clonal lineages of N. meningitidis (16). These observations, especially in the light of the known role of related toxins in other bacteria, suggest that the meningococcal RTX proteins are likely to play an important role in meningococcal carriage and/or pathogenesis. Since these genes are not found in the closely related pathogen N. gonorrhoeae, which characteristically causes localized inflammation of the genitourinary tract but rarely causes disseminated septicemia or meningitis, and are also not found in the closely related commensal species N. lactamica, it is tempting to speculate that secretion of the meningococcal RTX proteins contributes in some way to the ability of this pathogen to cause disseminated disease. It remains an important goal to identify target cells and receptors for the meningococcal RTX proteins and to determine what, if any, roles they play in meningococcal pathogenesis.
ACKNOWLEDGMENTS
We thank Fred Sparling for supplying monoclonal antibody A4.85 and Chris Tang for providing the ectopic complementation plasmid. Mass spectrometry was carried out by Kevin Bailey and Matthew Carlile at the University of Nottingham Proteomics facility.
Present address: Department of Chemistry, Faculty of Science and Art, University of Dicle, 21280 Diyarbakir, Turkey.
REFERENCES
1. Ait-Tahar, K., K. G. Wooldridge, D. P. Turner, M. Atta, I. Todd, and D. A. Ala'Aldeen. 2000. Auto-transporter A protein of Neisseria meningitidis: a potent CD4+ T-cell and B-cell stimulating antigen detected by expression cloning. Mol. Microbiol. 37:1094-1105.
2. Akama, H., T. Matsuura, S. Kashiwagi, H. Yoneyama, S.-i. Narita, T. Tsukihara, A. Nakagawa, and T. Nakae. 2004. Crystal structure of the membrane fusion protein, MexA, of the multidrug transporter in Pseudomonas aeruginosa. J. Biol. Chem. 279:25939-25942.
3. Angelos, J. A., J. F. Hess, and L. W. George. 2003. An RTX operon in hemolytic Moraxella bovis is absent from nonhemolytic strains. Vet. Microbiol. 92:363-377.
4. Benabdelhak, H., S. Kiontke, C. Horn, R. Ernst, M. A. Blight, I. B. Holland, and L. Schmitt. 2003. A specific interaction between the NBD of the ABC-transporter HlyB and a C-terminal fragment of its transport substrate hemolysin A. J. Mol. Biol. 327:1169-1179.
5. Binet, R., S. Letoffe, J. M. Ghigo, P. Delepelaire, and C. Wandersman. 1997. Protein secretion by gram-negative bacterial ABC exporters: a review. Gene 192:7-11.
6. Boyd, E. F., and D. L. Hartl. 1998. Chromosomal regions specific to pathogenic isolates of Escherichia coli have a phylogenetically clustered distribution. J. Bacteriol. 180:1159-1165.
7. Brandtzaeg, P., A. Bjerre, R. Ovstebo, B. Brusletto, G. B. Joo, and P. Kierulf. 2001. Neisseria meningitidis lipopolysaccharides in human pathology. J. Endotoxin Res. 7:401-420.
8. Felmlee, T., S. Pellett, E. Y. Lee, and R. A. Welch. 1985. Escherichia coli hemolysin is released extracellularly without cleavage of a signal peptide. J. Bacteriol. 163:88-93.
9. Forman, S., I. Linhartova, R. Osicka, X. Nassif, P. Sebo, and V. Pelicic. 2003. Neisseria meningitidis RTX proteins are not required for virulence in infant rats. Infect. Immun. 71:2253-2257.
10. Frey, J., and P. Kuhnert. 2002. RTX toxins in Pasteurellaceae. Int. J. Med. Microbiol. 292:149-158.
11. Glaser, P., H. Sakamoto, J. Bellalou, A. Ullmann, and A. Danchin. 1988. Secretion of cyclolysin, the calmodulin-sensitive adenylate cyclase-haemolysin bifunctional protein of Bordetella pertussis. EMBO J. 7:3997-4004.
12. Grifantini, R., S. Sebastian, E. Frigimelica, M. Draghi, E. Bartolini, A. Muzzi, R. Rappuoli, G. Grandi, and C. A. Genco. 2003. Identification of iron-activated and -repressed Fur-dependent genes by transcriptome analysis of Neisseria meningitidis group B. Proc. Natl. Acad. Sci. USA 100:9542-9547.
13. Hadi, H. A., K. G. Wooldridge, K. Robinson, and D. A. Ala'Aldeen. 2001. Identification and characterization of App: an immunogenic autotransporter protein of Neisseria meningitidis. Mol. Microbiol. 41:611-623.
14. Hopkinson, A., R. S. McIntosh, R. Layfield, J. Keyte, P. J. Tighe, and H. S. Dua. 2005. Optimised two-dimensional electrophoresis procedures for the protein characterisation of structural tissues. Proteomics doi:10.1002/pmic.200401073.
15. Jacob-Dubuisson, F., C. Buisine, N. Mielcarek, E. Clement, F. D. Menozzi, and C. Locht. 1996. Amino-terminal maturation of the Bordetella pertussis filamentous haemagglutinin. Mol. Microbiol. 19:65-78.
16. Klee, S. R., X. Nassif, B. Kusecek, P. Merker, J.-L. Beretti, M. Achtman, and C. R. Tinsley. 2000. Molecular and biological analysis of eight genetic islands that distinguish Neisseria meningitidis from the closely related pathogen Neisseria gonorrhoeae. Infect. Immun. 68:2082-2095.
17. Koronakis, V., A. Sharff, E. Koronakis, B. Luisi, and C. Hughes. 2000. Crystal structure of the bacterial membrane protein TolC central to multidrug efflux and protein export. Nature 405:914-919.
18. Lally, E. T., R. B. Hill, I. R. Kieba, and J. Korostoff. 1999. The interaction between RTX toxins and target cells. Trends Microbiol. 7:356-361.
19. Lee, V. T., and O. Schneewind. 2001. Protein secretion and the pathogenesis of bacterial infections. Genes Dev. 15:1725-1752.
20. Letoffe, S., P. Delepelaire, and C. Wandersman. 1996. Protein secretion in gram-negative bacteria: assembly of the three components of ABC protein-mediated exporters is ordered and promoted by substrate binding. EMBO J. 15:5804-5811.
21. Locht, C., M. C. Geoffroy, and G. Renauld. 1992. Common accessory genes for the Bordetella pertussis filamentous hemagglutinin and fimbriae share sequence similarities with the papC and papD gene families. EMBO J. 11:3175-3183.
22. Lomholt, H., K. Poulsen, and M. Kilian. 1995. Comparative characterization of the iga gene encoding IgA1 protease in Neisseria meningitidis, Neisseria gonorrhoeae and Haemophilus influenzae. Mol. Microbiol. 15:495-506.
23. Osicka, R., J. Kalmusova, P. Krizova, and P. Sebo. 2001. Neisseria meningitidis RTX protein FrpC induces high levels of serum antibodies during invasive disease: polymorphism of frpC alleles and purification of recombinant FrpC. Infect. Immun. 69:5509-5519.
24. Osicka, R., K. Prochazkova, M. Sulc, I. Linhartova, V. Havlicek, P. Sebo, S. Forman, X. Nassif, V. Pelicic, J. Kalmusova, and P. Krizova. 2004. A novel "clip-and-link" activity of repeat in toxin (RTX) proteins from gram-negative pathogens. J. Biol. Chem. 279:24944-24956.
25. Parkhill, J., M. Achtman, K. D. James, S. D. Bentley, C. Churcher, S. R. Klee, G. Morelli, D. Basham, D. Brown, T. Chillingworth, R. M. Davies, P. Davis, K. Devlin, T. Feltwell, N. Hamlin, S. Holroyd, K. Jagels, S. Leather, S. Moule, K. Mungall, M. A. Quail, M. A. Rajandream, K. M. Rutherford, M. Simmonds, J. Skelton, S. Whitehead, B. G. Spratt, and B. G. Barrell. 2000. Complete DNA sequence of a serogroup A strain of Neisseria meningitidis Z2491. Nature 404:502-506.
26. Prentki, P., and M. Krisch. 1984. In vitro insertional mutagenesis with a selectable DNA fragment. Gene 29:303-3113.
27. Prochazkova, K., R. Osicka, I. Linhartova, P. Halada, M. Sulc, and P. Sebo. 2005. The Neisseria meningitidis outer membrane lipoprotein FrpD binds the RTX protein FrpC. J. Biol. Chem. 280:3251-3258.
28. Robinson, K., M. Taraktsoglou, K. S. J. Rowe, K. G. Wooldridge, and D. A. A. Ala'Aldeen. 2004. Secreted proteins from Neisseria meningitidis mediate differential human gene expression and immune activation. Cell Microbiol. 6:927-938.
29. Schlor, S., A. Schmidt, E. Maier, R. Benz, W. Goebel, and I. Gentschev. 1997. In vivo and in vitro studies on interactions between the components of the hemolysin (HlyA) secretion machinery of Escherichia coli. Mol. Gen. Genet. 256:306-319.
30. Serruto, D., J. Adu-Bobie, M. Scarselli, D. Veggi, M. Pizza, R. Rappuoli, and B. Arico. 2003. Neisseria meningitidis App, a new adhesin with autocatalytic serine protease activity. Mol. Microbiol. 48:323-334.
31. Tettelin, H., N. J. Saunders, J. Heidelberg, A. C. Jeffries, K. E. Nelson, J. A. Eisen, K. A. Ketchum, D. W. Hood, J. F. Peden, R. J. Dodson, W. C. Nelson, M. L. Gwinn, R. DeBoy, J. D. Peterson, E. K. Hickey, D. H. Haft, S. L. Salzberg, O. White, R. D. Fleischmann, B. A. Dougherty, T. Mason, A. Ciecko, D. S. Parksey, E. Blair, H. Cittone, E. B. Clark, M. D. Cotton, T. R. Utterback, H. Khouri, H. Qin, J. Vamathevan, J. Gill, V. Scarlato, V. Masignani, M. Pizza, G. Grandi, L. Sun, H. O. Smith, C. M. Fraser, E. R. Moxon, R. Rappuoli, and J. C. Venter. 2000. Complete genome sequence of Neisseria meningitidis serogroup B strain MC58. Science 287:1809-1815.
32. Thompson, S. A., and P. F. Sparling. 1993. The RTX cytotoxin-related FrpA protein of Neisseria meningitidis is secreted extracellularly by meningococci and by HlyBD+ Escherichia coli. Infect. Immun. 61:2906-2911.
33. Thompson, S. A., L. L. Wang, and P. F. Sparling. 1993. Cloning and nucleotide sequence of frpC, a second gene from Neisseria meningitidis encoding a protein similar to RTX cytotoxins. Mol. Microbiol. 9:85-96.
34. Thompson, S. A., L. L. Wang, A. West, and P. F. Sparling. 1993. Neisseria meningitidis produces iron-regulated proteins related to the RTX family of exoproteins. J. Bacteriol. 175:811-818.
35. Turner, D. P., K. G. Wooldridge, and D. A. Ala'Aldeen. 2002. Autotransported serine protease A of Neisseria meningitidis: an immunogenic, surface-exposed outer membrane and secreted protein. Infect. Immun. 70:4447-4461.
36. van Ulsen, P., L. van Alphen, J. ten Hove, F. Fransen, P. van der Ley, and J. Tommassen. 2003. A neisserial autotransporter NalP modulating the processing of other autotransporters. Mol. Microbiol. 50:1017-1030.
37. van Vliet, A. H. M., K. G. Wooldridge, and J. M. Ketley. 1998. Iron-responsive gene regulation in a Campylobacter jejuni fur mutant. J. Bacteriol. 180:5291-5298.
38. Waldor, M. K., E. J. Rubin, G. D. Pearson, H. Kimsey, and J. J. Mekalanos. 1997. Regulation, replication, and integration functions of the Vibrio cholerae CTXphi are encoded by region RS2. Mol. Microbiol. 24:917-926.
39. Wandersman, C., and P. Delepelaire. 1990. TolC, an Escherichia coli outer membrane protein required for hemolysin secretion. Proc. Natl. Acad. Sci. USA 87:4776-4780.
40. Winzer, K., Y. H. Sun, A. Green, M. Delory, D. Blackley, K. R. Hardie, T. J. Baldwin, and C. M. Tang. 2002. Role of Neisseria meningitidis luxS in cell-to-cell signaling and bacteremic infection. Infect. Immun. 70:2245-2248.
41. Wooldridge, K. G., and P. H. Williams. 1993. Iron uptake mechanisms of pathogenic bacteria. FEMS Microbiol. Rev. 12:325-348.(Karl G. Wooldridge, Murat)