| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Meningococcal Research Department, Division of Vaccines, Center for Genetic Engineering and Biotechnology, Avenue 31, Cubanacan, Habana 10600, Cuba
2 Department of Molecular Biology, Division of Biotechnology, Finlay Institute, Avenue 27, La Lisa, Habana 11600, Cuba
Correspondence
Gretel Sardiñas
gretel.sardinas{at}cigb.edu.cu
Daniel Yero
dyero{at}finlay.edu.cu
Received July 7, 2008
Accepted October 4, 2008
The significance of Neisseria meningitidis serogroup B membrane proteins as vaccine candidates is continually growing. Here, we studied different aspects of antigen NMB0088, a protein that is abundant in outer-membrane vesicle preparations and is thought to be a surface protein. The gene encoding protein NMB0088 was sequenced in a panel of 34 different meningococcal strains with clinical and epidemiological relevance. After this analysis, four variants of NMB0088 were identified; the variability was confined to three specific segments, designated VR1, VR2 and VR3. Secondary structure predictions, refined with alignment analysis and homology modelling using FadL of Escherichia coli, revealed that almost all the variable regions were located in extracellular loop domains. In addition, the NMB0088 antigen was expressed in E. coli and a procedure for obtaining purified recombinant NMB0088 is described. The humoral immune response elicited in BALB/c mice was measured by ELISA and Western blotting, while the functional activity of these antibodies was determined in a serum bactericidal assay and an animal protection model. After immunization in mice, the recombinant protein was capable of inducing a protective response when it was administered inserted into liposomes. According to our results, the recombinant NMB0088 protein may represent a novel antigen for a vaccine against meningococcal disease. However, results from the variability study should be considered for designing a cross-protective formulation in future studies.
Abbreviations: CREE, Correia repeat-enclosed element; DRVs, dried–reconstituted vesicles; i.n., intranasal; i.p., intraperitoneal; OMVs, outer-membrane vesicles; TM, transmembrane; VCN, vancomycin/colistin/nystatin; VRs, variable regions.
These authors contributed equally to this work. 
The GenBank/EMBL/DDBJ accession numbers for the new nmb0088 gene sequences are EU645484–EU645512.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSReferences |
|---|
Vaccine research against serogroup B meningococcus has mostly focused on cell-surface protein antigens contained in outer-membrane vesicles (OMVs) (Fredriksen et al., 1991; Sierra et al., 1991) or recombinant proteins (Giuliani et al., 2006). The most important vaccines so far evaluated in clinical trials are OMV-based vaccines (Bjune et al., 1991; Sierra et al., 1991). The drawback with the OMV vaccines is that the most immunogenic components are highly variable and, consequently, these vaccines are not able to confer protection against a wide range of heterologous strains (Martin et al., 2000; Tappero et al., 1999). Alternatively, highly conserved N. meningitidis serogroup B membrane antigens have been sought and evaluated as vaccine candidates (Comanducci et al., 2002; Martin et al., 1997; Giuliani et al., 2006). The successful development of a broad-specificity serogroup B vaccine may be attained as a consequence of the genome sequencing of the serogroup B N. meningitidis strain MC58 (Tettelin et al., 2000), which has opened up new possibilities for identifying antigens to be included in candidate vaccines (Pizza et al., 2000). In parallel, techniques for proteomic and immunoproteomic analysis have been applied to characterize the antigen composition of commercial complex vaccines, and particularly some applications have arisen for the characterization of OMVs of pathogenic bacteria (Nally et al., 2005; Wheeler et al., 2007). For the meningococcus, such analytical approaches have recently identified new vaccine candidates (Delgado et al., 2007; Hsu et al., 2008).
A previous study described the analysis by two-dimensional gel electrophoresis and mass spectrometry of N. meninigitidis proteins present as the active ingredient of the Cuban vaccine VA-MENGOC-BC®, and the consequent identification of major and minor proteins in the preparation not previously reported as components of this antimeningococcal vaccine (Uli et al., 2006). A set of these antigens that show promise as vaccine candidates is currently under preclinical evaluation (Delgado et al., 2007). One of these antigens is the protein NMB0088 with an apparent molecular mass of approximately 50 kDa. After several proteomic studies, NMB0088 appears to be abundant in N. meningitidis serogroup B OMV preparations obtained from different strains (Vipond et al., 2006; Williams et al., 2007; Uli et al., 2006; Vaughan et al., 2006). In the present study, the nmb0088 gene and the deduced protein sequence were analysed for a broader survey of its sequence conservation across a collection of isolates and in order to predict the topographic localization of regions exposed to the immune system. We also report the purification of a recombinant variant of NMB0088 protein, as well as the characterization of the antibodies elicited in mice against this antigen.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSReferences |
|---|
. Bacteria employed as target strains for the immunological experiments were grown in a humidified atmosphere of 5 % CO2 in air on brain heart infusion (BHI) agar (Oxoid) supplemented with vancomycin/colistin/nystatin (VCN) (Oxoid). For the infant rat protection assay, strain CU385 was grown on BHI agar supplemented with 7 % (v/v) defibrinated goat blood (BHI-blood agar) and containing VCN. The Neisseria lactamica strain 102-CIGB (Sardinas et al., 2006) used for the immunoblotting was obtained from the culture collection of the Center for Genetic Engineering and Biotechnology (Havana, Cuba).
|
Sequence analysis of the nmb0088 gene and the deduced protein. Genomic DNA sequences flanking the nmb0088 gene were obtained from the annotated MC58 genome sequence (accession no. NC_003112). The homologue of nmb0088 and flanking sequences of the N. meningitidis strains Z2491 and FAM18 and a N. lactamica strain (an ST-640 strain identified here as Nl-ST640) were obtained from the Sanger Institute (http://www.sanger.ac.uk/Projects/). The incompletely assembled genome sequence of the N. lactamica vaccine strain Y92-1009 (Vaughan et al., 2006) was used. The Neisseria gonorrhoeae strain FA1090 genome was accessed from the University of Oklahoma website (http://www.genome.ou.edu/gono.html). The recently published sequence from the serogroup C isolate 053442 (ST-4821) (Peng et al., 2008) was also employed. CLUSTAL W version 2.0 (Larkin et al., 2007) was used to perform sequence alignments.
Deduced polypeptide sequences of the open reading frames were analysed with the PSORTb program (www.psort.org/psortb/) to predict their cellular localization and the presence of a signal peptide. BLAST (http://ncbi.nlm.nih.gov/) was used to search the National Center for Biotechnology Information (NCBI) databases to identify previously reported sequences with homology to those that were sequenced. For secondary structure analysis, the prediction server Jpred3 (www.compbio.dundee.ac.uk/
www-jpred/) was applied. An amino acid hydrophilicity plot of NMB0088 was made with the Kyte–Doolittle algorithm available at the ExPASy website (www.expasy.ch/tools/protscale.html).
PCR and nucleotide sequencing.
Genomic DNA from the N. meningitidis strains listed in Table 1
was used as a template for the amplification of an nmb0088 fragment (from amino acid 26 to 464) with the following primers: 0088F1, 5'-CGC AGA TCT CTA CCA CTT CGG CAC AC-3'; and 0088R1, 5'-TTA CTC GAG TTT GTA GGT GTA TTG CAG ACC-3'. Primers were designed for amplification of the segment encoding the mature form of antigen NMB0088, based on the nucleotide sequence reported for this gene in strain MC58. The PCR products were purified and nucleotide sequencing was performed by Macrogen (Seoul, Korea). Nucleotide sequences were edited, aligned and analysed using the program AlignX (vector NTI Suite 7.1; InforMax).
Cloning and expression of the nmb0088 gene. The PCR product corresponding to the NMB0088 sequence from strain CU385 was cloned into the expression vector pM238 using primers 0088F1 and 0088R1 as previously described (Yero et al., 2006). This vector provides six His residues at the C-terminus of the expressed protein, and at the mature N-terminus a stabilizer sequence that consists of the first 47 amino acids of the LpdA (dihydrolipoamide dehydrogenase A) protein from N. meningitidis. The recombinant expression plasmid was transferred to the expression host strain E. coli W3110 and protein expression was induced by growing the bacteria in expression medium for 16 h at 37 °C.
Purification of the recombinant fusion protein. After growth in expression medium, E. coli cells expressing the recombinant NMB0088 (rNMB0088) protein were collected by centrifugation for 15 min at 4000 g. The recombinant polypeptide was then purified by the washed pellet procedure. Briefly, the pellet was suspended by gentle stirring in chilled rupture buffer (10 mM Tris-Cl, 1 mM EDTA, pH 7.5), sonicated (five cycles, 30 s pulses at 60 s intervals on ice) and the insoluble material was collected by centrifugation for 10 min at 7000 g at 4 °C. The inclusion bodies contained in the isolated pellet were washed successively with rupture buffer containing 10 mM MgCl2, 0.8 M NaCl and 0.1 % Nonidet P-40 and recentrifuged. After solubilization in loading buffer [6 M urea in carbonate–bicarbonate buffer solution (0.1 M sodium carbonate and 0.1 M sodium hydrogen carbonate, pH 10.0)], the recombinant protein was purified by affinity chromatography with a Chelating Sepharose Fast Flow column (Amersham Pharmacia Biotech) previously loaded with Cu2+ according to the manufacturer's instructions. The column was washed with loading buffer containing 10 mM imidazole and contaminants bound to the column were eluted. Finally, the recombinant protein was eluted with 50 mM imidazole in loading buffer. The eluates were analysed by SDS-PAGE, the fractions containing a single protein band were pooled and dialysed against phosphate buffer (pH 8.0), and the protein concentrations were determined. The expression of recombinant protein, after each purification step, was analysed by gel electrophoresis and the proportions were determined by densitometry analysis using software Image J 1.37h (National Institutes of Health, USA).
Insertion of the recombinant protein into liposomes. Liposomes were obtained using the dried–reconstituted vesicles (DRVs) method (Kirby & Gregoriadis, 1984). Briefly, a mixture of 18 mg dipalmitoylphosphatidylcholine and 9 mg cholesterol in chloroform was dried by rotatory evaporation at 30 °C for 1 h. Dried phospholipids were suspended in 5 ml PBS (pH 8.0) by vigorous agitation. The suspension was homogenized in an ultrasonic water bath (five bursts of 2 min), resulting in a suspension of multilamellar vesicles. The suspension of empty liposomes was mixed with 6 mg rNMB0088 and freeze-dried three times in order to achieve the insertion of rNMB0088 into the vesicles. DRVs were centrifuged at 100 000 g for 2 h at 4 °C, and the pellet was suspended in 1 ml PBS.
Immunization of animals. To evaluate the immunogenicity of the recombinant protein, female BALB/c mice (CENPALAB; Havana, Cuba) 8–10 weeks of age were immunized by the intraperitoneal (i.p.) or the intranasal (i.n.) routes. The animals used in the study were housed and used strictly in accordance with the Institutional Guidelines for Care and Use of Laboratory Animals. Three groups of 10 mice each received three i.p. doses at 2-week intervals (100 µl per dose) containing 10 µg rNMB0088 in Freund's or aluminium hydroxide adjuvants or inserted in liposomes (rNMB0088-DRVs). An additional group received 50 µl (25 µl per nostril) rNMB0088-DRVs, containing 10 µg of the recombinant protein, by the i.n. route. As positive control, a mouse group was immunized i.p. following the same schedule with 10 µg per dose of meningococcal OMVs (with aluminium adjuvant). The sera were collected by centrifugation from blood extracted on days 0 (preimmune) and 45 and samples were stored at –20 °C.
SDS-PAGE and Western blotting. Cell lysates of N. meningitidis strains were separated by SDS-PAGE, transferred onto a nitrocellulose membrane (Hybond ECL; Amersham Pharmacia Biotech) and analyses were performed as described previously (Wedege & Froholm, 1986), with the following modifications. Membrane strips containing transferred proteins were blocked with 5 % (w/v) skim milk powder (Oxoid) in PBS for 3 h at room temperature. Pooled sera from immunized and control mice were diluted 1 : 100 in PBS containing 0.05 % (v/v) Tween 20 (PBS-T) and 1 % (w/v) skim milk, and incubated with the membranes overnight at 4 °C with continuous shaking. The membranes were washed three times with PBS-T and incubated for 1 h at room temperature with goat anti-mouse IgG peroxidase conjugate (Sigma-Aldrich) diluted 1 : 10 000 in PBS-T. The membranes were washed again and the antigen–antibody reactions were visualized using a substrate solution containing 3-amino-9-ethylcarbazole (Sigma-Aldrich) as the chromogen.
Detection of immune response by ELISA. Antibodies against the protein on the surface of the bacteria were detected by whole-cell ELISA as described previously (Rosenqvist et al., 1990). Strains were grown overnight at 37 °C on BHI agar and then subcultured onto a fresh plate under the same conditions for a further 4 h. The bacteria were harvested and suspended in PBS with 0.02 % sodium azide. The OD620 of the suspension was adjusted to 0.1. One hundred microlitres of this suspension was added to individual wells in polystyrene plates and left to evaporate overnight at 37 °C. The rest of the experiment followed a conventional ELISA protocol. Titres were expressed as the reciprocal of the dilution showing a twofold increase in absorbance over that obtained with a negative control.
Bactericidal antibody assay.
The bactericidal assay was carried out using homologous strain CU385 in 96-well plates, as described previously (Maslanka et al., 1997), with few modifications. Baby rabbit serum was used as a source of exogenous complement, and this serum was adsorbed on ice with inactivated meningococcal cells immediately before use. Briefly, bacteria (target strain) were adjusted to OD620 1.0 in PBS and the suspension was incubated for 30 min at 56 °C. A bacterial pellet, corresponding to 2.5 ml heat-inactivated bacterial suspension, was added to 1.0 ml rabbit serum and the mixture was incubated for 50 min on ice with agitation. Finally, bacteria were removed by centrifugation in a refrigerated centrifuge. The pre-adsorbed baby rabbit serum was used in the bactericidal assays. The bactericidal titre of the serum was defined as the reciprocal of the highest serum dilution giving 
50 % killing compared to the mean from complement-independent control wells for each assay plate.
Infant rat protection assay. The ability of the sera from immunized mice to confer protection against N. meningitidis bacteraemia in infant rats was performed as previously described (Yero et al., 2005). In brief, 5–6-day-old infant rats (outbred Wistar; CENPALAB) were randomly distributed in groups of six animals per pool of sera. Rat pups were injected i.p. with 100 µl serum diluted 1 : 10, 1 h before the i.p. bacterial challenge with approximately 107 c.f.u. per pup in a final volume of 150 µl. For the challenge, bacteria were suspended in sterile PBS containing iron dextran. A pool of preimmune sera was used as negative control and a pool of hyperimmune sera from OMV-immunized mice was used as positive control for protection. Development of bacteraemia was assessed by culturing diluted blood samples taken 4 h after challenge on BHI-blood agar containing VCN. The results of blood cultures were transformed to logarithmic values to calculate the geometric mean value of the c.f.u. in each group of six animals. Statistical significance was evaluated using a one-way analysis of variance (ANOVA) followed by a Dunnett's post test. GraphPad Prism version 4.0 (GraphPad Software) was used and a P value less than 0.05 was considered statistically significant.
|
RESULTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSReferences |
|---|
). CREEs have been divided into four families (
, ', β and β'), distinguished by a 50 bp internal deletion and five point mutations (Buisine et al., 2002). In N. meningitidis strain MC58, the downstream region of nmb0088 contains a 155 bp -family CREE with 26 bp terminal inverted repeats. In strains MC58 and FAM18, the upstream regions of this gene contain a member of the '-family CREE that is 106 bp long and has terminal inverted repeats of 26 and 27 bp. These upstream CREEs carry putative –35 and –10 promoter elements that could act as transcription regulators for the nmb0088 gene. In strains Z2491 and 053442, and in N. lactamica Nl-ST640, the upstream CREEs are 101 bp long due to a five nucleotide internal deletion (Fig. 1). In N. lactamica Y92-1009, this upstream repetitive element is absent, and different promoter elements could act on the expression of the nmb0088 gene.
|
. We detected stretches with secondary structure typical of transmembrane (TM) domains, three -helical motifs at the N-terminus of the protein and several regions with β-sheet prediction (reliability of accuracy >7) along the rest of the amino acidic sequence.
|
). The crystal structure of FadL from E. coli has been reported indicating this protein forms a 14-stranded β-barrel that is occluded by a central hatch domain (van den Berg et al., 2004). The N-terminal region of the protein contains three short helices, which plug the barrel. One of the helices is capped by the conserved sequence NPA, a signature sequence in aquaporins (Fu et al., 2000) that is also present in NMB0088 (amino acids 56–58).
Alignment of NMB0088 sequences and variable region definition
Proteins homologous to NMB0088 have been found in other meningococcal genomes with identities ranging from 93 % to 99 %, and in N. lactamica strains Nl-ST640 and Y92-1009 with 93 % and 89 % identity, respectively. However, in the gonococcus, the homologue to nmb0088 appears to be a pseudogene. A first alignment analysis using these reported sequences showed discrete regions of high variability in the deduced amino acid sequence of NMB0088. To confirm the existence of variable regions (VRs) within the NMB0088 sequence and in order to define the topology model of this protein, we refined the alignment analysing orthologous sequences in a panel of meningococcal strains belonging to different serogroups, serotypes and sero-subtypes.
Comparison of the deduced amino acid sequences from 34 meningococcal strains and those from the two N. lactamica strains revealed large areas of highly conserved sequence and several well-defined regions of variation (Table 1 and Fig. 2). The alignment showed that NMB0088 variability is confined to three specific segments, designated VR1 (193–216), VR2 (268–287) and VR3 (339–345). The two first regions were the most variable among strains at both the nucleotide and amino acid level. The number of amino acids in the three regions is different between strains tested, a fact that contributes to the variability of the VRs. Comparison of the per cent similarities in NMB0088 sequences among N. meningitidis isolates identified four variants with amino acid sequences of 
96 % similarity (Table 1). The two N. lactamica sequences added to the alignment represented two new variants. In the N. meningitidis strains tested, the DNA regions coding for the VR1 in NMB0088 variants 2, 3 and 4 are distinguished by a 9 bp internal deletion. No correlation was found between the serological classification of strains and their NMB0088 variant, except for the two serogroup A strains that were grouped into variant 3. However, it was interesting that 70 % of the strains belonging to variant 2 were isolated from healthy carriers (Table 1).
Membrane topology model for protein NMB0088
In an attempt to construct a topology model for this membrane protein, we compared sequences of different NMB0088 proteins, and hypervariable regions (putative exposed regions) were predicted from the variability profile of the alignment (Fig. 2). The pattern of alternating β-sheets and VRs is similar to that of other neisserial porins and TonB-dependent proteins, where it has been demonstrated that variable stretches are located at the surface, while conserved regions form membrane-spanning β-sheets (van der Ley et al., 1991; Derrick et al., 1999). We assumed that VRs are present on the surface, and used the location of hydrophilic maxima in the hydropathy profile (Kyte–Doolittle) as an additional criterion for surface exposure.
The membrane-spanning segments in the proposed model for NMB0088 were predicted based on a number of features concerning TM β-strands and their organization in well-known β-barrel proteins (Schulz, 2000). The number of TM β-strands in barrels has been shown to range from 6 to 22 (most frequently 12) residues. TM β-strands show an inside–outside dyad repeat motif of alternating residues facing the lipid bilayer and the inside of the barrel. Outside (lipid bilayer facing) residues are typically hydrophobic whilst inside (facing inside of barrel) residues are of intermediate polarity. TM β-strands are also often flanked by a layer of aromatic residues, believed to be involved in maintaining the protein's stability within the membrane (Yau et al., 1998). Subsequently, we examined the NMB0088 alignment for stretches of 12 amino acid residues, which should be amphipathic when present in a β-configuration, flanked by aromatic amino acids and located in regions predicted as β-sheets by Jpred and by homology with E. coli FadL.
Finally, the topographic map was built starting at the C-terminal region with the last TM segment directed towards the periplasm. In our model, protein NMB0088 is thought to span the membrane 14 times, thereby exposing seven hydrophilic loops beyond the outer membrane. Based on the topology model proposed here, the VRs were located in the third, fourth and fifth extracellular loops, respectively (Fig. 2). Regardless of whether or not the first 11 residues in the mature protein were included in the alignment, the N-terminal portion of protein NMB0088 appears to form an -helical plug similarly to the homologous region in E. coli FadL.
Immunogenicity of the recombinant protein and functional activities of the elicited antibodies
As described in detail in Methods, the complete nmb0088 DNA sequence of CU385, minus its signal sequence, was amplified and inserted into an E. coli expression vector. The recombinant protein formed insoluble cytoplasmic inclusion bodies that were harvested, purified and concentrated. To study the immunogenicity of rNMB0088 using different formulations, BALB/c mice were immunized (i.p.) with different adjuvants. Taking into account that NMB0088 is probably an integral β-barrel protein and to attempt to present the protein in its native conformation for immunization experiments, rNMB0088 was also incorporated into liposomes and administered by the i.p. and i.n. routes. Sera collected after immunization with the recombinant protein were analysed by whole-cell ELISA and Western blotting experiments.
The specificity of antisera produced against purified rNMB0088 was tested by Western blotting and results showed that this protein was expressed by a panel of meningococcal strains belonging to different serogroups and different NMB0088 variants, and also by a N. lactamica strain (Fig. 3). An easily identifiable band was recognized in each strip corresponding to a meningococcal protein with an apparent molecular mass of approximately 50 kDa. These antisera produced against purified rNMB0088 in all the immunization variants recognized epitopes contained in the sample of homologous strain CU385 in a whole-cell ELISA (Fig. 4a). Similar results were obtained in the immunoassay when we evaluated pooled sera from mice immunized with rNMB0088 in aluminium hydroxide using as target cells strains B16B6 (titre 3690), Z4181 (titre 4200), M982 (titre 5890), NZ124 (titre 1750) and Z1127 (titre 2100). These results indicate that NMB0088 was expressed by the N. meningitidis strains tested and levels of expression appear to be nearly the same.
|
|
). Only sera from mice immunized with protein emulsified in Freund's adjuvant showed no detectable bactericidal titres. The other antisera promoted bactericidal activity against meningococci, with titres ranging from 1 : 16 to 1 : 64. In addition, sera from mice immunized with the recombinant protein were bactericidal against two heterologous strains (B16B6 and M982) with the same NMB0088 variant but with different serological classification and sequence type (Table 2). These sera did not kill two other heterologous strains with different sequence variants (Table 2).
|
). Animals passively immunized with pooled sera from mice injected with formulation containing rNMB0088-DRVs by the i.n. or i.p. route had a significant reduction in the level of bacteraemia compared to the control group. No significant differences were detected between the other formulations. |
DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSReferences |
|---|
The high variability observed in the NMB0088 VRs suggests that possible insertions, deletions and/or recombination events occurred. The nmb0088 gene is located in a chromosome region that may be genetically unstable, due to the presence of repeat sequences flanking the gene. The N. meningitidis genome contains many hundreds of repetitive sequence elements ranging from simple sequence repeats to duplications of gene clusters. It was suggested that these repetitive arrays may encourage sequence variation in neighbouring genes by increasing the frequency of recombination with exogenous DNA (Bentley et al., 2007). Among the most abundant repeat types are the CREEs, which are often located upstream of genes (Liu et al., 2002; De Gregorio et al., 2003a), have been shown to affect gene expression (De Gregorio et al., 2003b), and may be transposable or mobilizable (Buisine et al., 2002). The downstream CREE reported in the present study could also be involved in the expression of NMB0088 since a family of long CREEs has been considered to be a transcriptional terminator in the dcw cluster of N. gonorrhoeae CH811 (Francis et al., 2000). The CREE upstream of the nmb0088 gene has been previously reported for the MC58 strain (De Gregorio et al., 2003a) and our sequence analysis (Fig. 1) showed that this element is not identical in all reported sequences, and is absent in one N. lactamica strain. However, the fact that N. lactamica Y92-1009 lacks the 5' CREE does not impair the expression of protein NMB0088 as was previously demonstrated in proteomic studies by Vaughan et al. (2006), indicating that other promoter elements are present on the upstream region of nmb0088. The precise localization of the promoter elements in the nmb0088 gene needs to be demonstrated and further transcription studies should be conducted to clarify the regulation of transcription in this gene.
The function of NMB0088 in the meningococcus is still unclear. This antigen has homology with a family of bacterial membrane proteins that includes FadL from E. coli. The FadL family includes several distantly related proteins, all probably outer-membrane proteins, from the human pathogens Haemophilus influenzae, Pseudomonas sp. and Moraxella catarrhalis. Many researchers have focused on these proteins as vaccine candidates. In M. catarrhalis, the protein homologue to FadL is known as the outer-membrane protein OMP-E, and many studies have indicated that OMP-E should be considered a potential vaccine antigen (Murphy et al., 2001). Furthermore, NMB0088 is a homologue of OmpP1 from H. influenzae, which has been evaluated as a vaccine candidate against experimental otitis media due to nontypable strains (Bolduc et al., 2000). Passive immunization with H. influenzae OmpP1 induced protection against bacteraemia in the infant rat model (Munson & Hunt, 1989) as was also demonstrated for protein NMB0088 in the present study. OmpP1 has four immunogenic surface-exposed loops which show sequence variability among different strains (Bolduc et al., 2000) and contain positively selected codons (Mes & van Putten, 2007). Thus development of a polyvalent vaccine reflecting the variability of OmpP1 would be necessary to construct an efficacious vaccine against H. influenzae, and a similar approach could be carried out with protein NMB0088.
A previous comparative proteomic study of OMVs produced from N. meningitidis and N. lactamica strain Y92-1009 revealed substantial differences between the antigen repertoires of the two species (Vaughan et al., 2006). However, as most of the N. lactamica proteome components have close meningococcal orthologues, like NMB0088, it has been proposed that an OMV vaccine prepared from this bacterium may offer sufficient cross-protection to prevent meningococcal disease (Gorringe et al., 2005). In our work, sera from mice immunized with rNMB0088 recognized a protein in Western blotting of a N. lactamica strain (Fig. 3), indicating that differences in the sequence do not impair immune recognition. Previous studies have shown that OMVs from N. lactamica are immunogenic when given by the i.n. route and they are able to induce bactericidal antibodies against N. meningitidis (Sardinas et al., 2006). The results of the i.n. vaccination with rNMB0088 in the present work corroborate the utility of this route for developing a protective response and encourage studying the mucosal immune response after immunization with Neisseria antigens.
In the present study, results from the sequence alignment also impacted on the topological reconstruction of protein NMB0088. It appears to share common folding with other Gram-negative integral outer-membrane proteins, i.e. they form a TM β-barrel consisting of antiparallel amphipathic β-strands. The identification of β-barrel regions here was based on the X-ray crystallography of the protein FadL of E. coli (van den Berg et al., 2004). Despite the low level of amino acid sequence congruence, the fit between FadL and NMB0088 was sufficient to pinpoint the approximate localization of TM regions and surface-exposed loops. More divergent structural templates and targets, such as OmpF and PhoE of E. coli, have successfully been used for the reconstruction of 3D models of porins in Neisseria (Derrick et al., 1999). Recently, a homology model for the vaccine candidate OmpP1 of H. influenzae has been proposed based on the FadL structure, demonstrating that the VRs are located on surface-exposed loops (Mes & van Putten, 2007).
More results from the present study deserve further comments. Recombinant NMB0088 protein was easily purified by affinity chromatography with Chelating Sepharose Fast Flow, obtaining a preparation with 90 % purity. However, this protein was obtained as inclusion bodies and denaturing and renaturing steps were required during its purification. The reconstitution of recombinant bacterial membrane proteins into their native conformations after purification has been a major problem in their use as effective vaccines. Liposomes have been shown to be an attractive approach as they provide a native-like environment for membrane proteins, particularly with a β-barrel structure (Buchanan, 1999). Several studies with recombinant meningococcal antigens including porins PorA (Christodoulides et al., 1998; Niebla et al., 2001) and PorB (Wright et al., 2002) have clearly demonstrated that the production of bactericidal antibodies was dependent on refolding of the protein to induce native conformation. In this study, immunization with formulations containing rNMB0088 induced a humoral immune response, characterized by the presence of specific IgG in sera (as detected by whole-cell ELISA and Western blotting) and by the induction of bactericidal antibodies. Nevertheless, the only pooled sera that conferred passive protection in the infant rat model were from the group immunized with the recombinant protein inserted into liposomes, suggesting that some additional protective epitopes of the NMB0088 antigen are better represented in a conformation of the protein resembling that in the outer membrane. However, based on this study and the methods of protein expression and preparation, the antigen would not be a sufficiently good candidate for inclusion in future vaccines. Further studies should be conducted to increase the titre of bactericidal antibodies after immunization with NMB0088, e.g. to change the dose of the antigen, the adjuvant and/or the refolding procedure during purification.
In the fight against human pathogens, a number of the most promising new vaccine candidates have been identified as a result of applying proteomic-based techniques to a large number of antigens, enabling more rapid development of these candidates at a reasonable cost compared with traditional strategies (Mora et al., 2003). The protein presented here is an example of how high-throughput strategies contribute to protein identification, production and evaluation as a vaccine candidate. NMB0088 should be considered in future studies as an attractive antigen for a subunit recombinant vaccine against the meningococcus. However, the level of diversity of the exposed loops in NMB0088 and the result of the bactericidal activity against different strains suggest that vaccines based on this antigen should be carefully designed in order to provide broad coverage in N. meningitidis.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSReferences |
|---|
|
References |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
