Expression of Heterologous Antigens in Commensal Neisseria spp.: Preservation of Conformational Epitopes with Vaccine Potential

Abstract

Commensal neisseriae share with Neisseria meningitidis (meningococcus) a tendency towards overproduction of the bacterial outer envelope, leading to the formation and release during growth of outer membrane vesicles (OMVs). OMVs from both meningococci and commensal neisseriae have shown promise as vaccines to protect against meningococcal disease. We report here the successful expression at high levels of heterologous proteins in commensal neisseriae and the display, in its native conformation, of one meningococcal outer membrane protein vaccine candidate, NspA, in OMVs prepared from such a recombinant Neisseria flavescens strain. These NspA-containing OMVs conferred protection against otherwise lethal intraperitoneal challenge of mice with N. meningitidis serogroup B, and sera raised against them mediated opsonophagocytosis of meningococcal strains expressing this antigen. This development promises to facilitate the design of novel vaccines containing membrane protein antigens that are otherwise difficult to present in native conformation that provide cross-protective efficacy in the prevention of meningococcal disease.

Neisseria meningitidis (the meningococcus) causes meningitis and septicemia in healthy children and is widely acknowledged as an important target for vaccine development. Encapsulated meningococcal strains can be divided into 13 serogroups based on chemically and antigenically distinct polysaccharide capsules. Five serogroups (A, B, C, Y, and W135) account for the vast majority of disease, and polysaccharide vaccines are available to prevent infections caused by all of them except serogroup B. Prevention of disease caused by this, the most prevalent meningococcal serogroup in many countries, remains a major challenge, as native serogroup B polysaccharide is not immunogenic in humans (30). The search for alternative vaccine approaches has been intense. The antigens considered have included many single candidates, such as individual outer membrane proteins (OMPs) or lipooligosaccharide subunits, as well as complex preparations such as outer membrane vesicles (OMVs) and extracts of nonpathogenic commensal neisseriae (16).

Meningococcal OMPs present challenging problems as potential vaccine antigens. Multiple hydrophobic domains embedded in the bacterial outer membrane dictate the antigenicity of the native protein, and vaccine preparations of purified, denatured OMPs often lack protective efficacy in humans. Furthermore, many meningococcal OMPs show a high level of sequence variation, particularly involving externally exposed domains and presumably driven by immune selection, so that even conformationally preserved antigens do not elicit cross-protective antibodies against the many potentially infecting strains. The major porin PorA, the serosubtyping antigen, has been extensively researched as a potential vaccine component but suffers from both of these problems.

Preparations of OMVs produced naturally by blebbing meningococci and amenable to pharmaceutical formulation have attracted considerable attention as potential vaccines in which OMPs can be presented in their native conformation. It had been hoped that meningococcal OMVs would elicit cross-protective immunity, as they contain a wide range of different OMPs. Disappointingly, it appears that the protection provided is virtually serosubtype specific, at least in infants, as antibody to PorA dominates the immune response (15, 23, 27). Interest in commensal neisseriae as the basis for a protective vaccine has developed from the appreciation that there is a high level of natural exposure to these organisms, more than to meningococci, during infancy and early childhood (2, 5). These lack PorA and so may be more effective at providing cross-protective immunity than meningococcal OMVs.

In the present study, we report the efficient expression of heterologous antigens, including the meningococcal OMP vaccine candidate NspA, in native form in commensal neisserial OMVs. High-level expression of such strategic components in OMV preparations is anticipated to enhance the capacity of such vaccine preparations to protect against meningococcal disease.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The Neisseria strains N. cinerea NRL32165, N. flava NRL30008, and N. subflava NRL30017 were kindly provided by J Knapp, Centers for Disease Control, Atlanta, Ga.; N. flavescens 2830 was from I. Feavers, National Institute for Biological Standards and Control, Potters Bar, Hertfordshire, United Kingdom, and N. meningitidis NZ98/394 was from D. Martin, Institute of Environmental Science and Research, Porirua, New Zealand. All other strains were from our own collections.

The standard laboratory strains of Escherichia coli used in this study were S17.1 λ pir, DH5α, and TOP10. E. coli was grown in Luria-Bertani (LB) medium (Oxoid) supplemented with appropriate antibiotics. For both E. coli and neisserial strains, the following antibiotics were used when needed at the indicated concentrations: ampicillin, 100 μg/ml; kanamycin, 75 μg/ml; and nalidixic acid, 20 μg/ml.

Nalidixic acid-resistant (Nal^r) derivatives of the commensal neisseriae were selected by exposure of strains grown to the stationary phase in broth cultures to nalidixic acid at 10 μg/ml, followed by overnight growth on selection medium to isolate spontaneous gyrase mutants.

Neisserial strains were routinely propagated at 37°C with 5% CO₂ on GC agar (Difco) supplemented with 1% Vitox (Oxoid), or at 37°C in Mueller-Hinton (MH) broth (Oxoid) with 1% Vitox or Frantz medium, shaken at 180 rpm. Where necessary, the availability of iron in the growth medium was restricted by the addition of 5 μg of ethylenediamine dihydroxyphenylacetic acid (EDDHA) per ml to broth cultures.

Recombinant DNA methods.

Standard methods were used for DNA preparation, restriction enzyme analysis, cloning, and sequencing (24). Transformations were carried out with a heat shock protocol for chemically competent E. coli TOP10 as described by the manufacturer (Invitrogen).

PCR and sequencing.

Oligonucleotide primers 401US (5′-GCAGTCTCTCGAGCTCAAG-3′) and MIDG200DS (5′-GGCTTTACACTTTATGCTTCCG-3′), which are complementary to the opposing strands at the 5′ and 3′ ends of the multiple cloning site of pMIDG201 (see below), respectively, were synthesized by MWG-Biochem. The cycling conditions for the PCR were as follows: denaturation and enzyme activation for 15 min at 95°C, followed by 30 cycles of 95°C for 30 s, 55°C for 45 s, and 72°C for 1 min. Reactions were performed with Hot Start Taq (Qiagen) in a Perkin-Elmer 2400 thermocycler.

Construction of plasmids.

pMIDG100 (Fig. 1), previously constructed in our laboratory (29), contains the aphA3 gene (encoding kanamycin resistance), transcribed divergently from a reporter domain containing a multiple cloning site, a ribosome-binding site, and a promoterless green fluorescent protein (GFP) gene (gfpmut3), between paired rho-independent transcriptional terminators. pMIDG101 is a derivative of pMIDG100, which contains the strong meningococcal ner promoter upstream of gfpmut3. This was modified to express NspA for the present work as follows. pMIDG101 was digested with BamHI and XbaI to remove the gfp gene and leave the cloned promoter backbone intact. A polylinker cloning domain and two copies of the neisserial uptake sequence, which can act as a transcriptional terminator, the whole flanked by BamHI and XbaI linkers, was ligated to the backbone of pMIDG101 to generate pMIDG201 (Fig. 1). A 600-bp fragment containing nspA and its native ribosomal binding site, the whole flanked by NcoI and PstI linkers, was amplified from N. meningitidis 608B chromosomal DNA by PCR and cloned into the NcoI and PstI sites of pMIDG201 to construct pMIDG201nspA (Fig. 1). The clone was confirmed by sequencing with primers 401US and MIDG200DS.

FIG. 1.

Plasmids and their relationships. Genes aphA3 (kanamycin resistance), gfpmut3 (green fluorescent protein), and nspA are indicated by solid arrows, the ner promoter is indicated by the flag, and relevant restriction sites are labeled by name.

Conjugation.

Plasmids were transformed into E. coli S17-1 λ pir and transferred into the commensal neisserial Nal^r strains by conjugation. The E. coli S17-1 donor strain containing the plasmid was grown overnight in LB broth containing kanamycin, and the culture was then diluted 1:100 in MH broth with no antibiotics and incubated without shaking for 1.5 h at 37°C in 5% CO₂. Overnight plates of the recipient Nal^r neisserial strains were harvested into MH broth to 2 × 10⁶ CFU/ml. This subculture was incubated without shaking for 2 to 4 h at 37°C with 5% CO₂. The neisserial (300 μl) and E. coli cell suspensions (50 μl) were mixed on a 0.45-μm membrane filter (Millipore) placed on an GC agar plate and incubated overnight at 37°C in 5% CO₂. Thereafter, the bacteria were removed into MH broth, and aliquots were plated on GC plates containing nalidixic acid and kanamycin to select neisserial transformants.

Monoclonal antibodies.

A bank of NspA-specific murine monoclonal antibodies (MAbs) was generated in Quebec. The MAbs used in this study were Me-7 (immunoglobulin G2a [IgG2a]) (4), which binds to the exposed loop 3 between amino acids 109 and 124; Me-23 (IgG1), which binds to a transmembrane domain between amino acids 140 and 174 in the native protein and is not accessible on intact cells; and Me-14 (IgG2a), which binds NspA only on intact cells, inferentially to a nonlinear (conformational) epitope, perhaps incorporating more than one surface-exposed loop (3).

Western blot analysis.

Bacterial colonies were harvested into phosphate-buffered saline (PBS) from overnight GC plate cultures and washed, and the pellet was resuspended in reducing buffer (24). Equal amounts of protein were subjected to sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) on polyacrylamide gels and electrotransferred on to nitrocellulose membranes (24). The primary antibody was anti-GFP antibody (Clontech) or an anti-NspA MAb. The secondary antibody was anti-mouse IgG plus IgM conjugated to horseradish peroxidase (Jackson ImmunoResearch Laboratories). The signal was detected after incubation in peroxidase solution [1 M KH₂PO₄, 1 M (NH₄)₂SO₄, 0.15% (vol/vol) peroxidase, 200 μl of O-dianisidine, 0.5% (vol/vol) methanol] or with the ECL Plus reagents according to the manufacturer's instructions (Amersham Biosciences).

Detection of GFP expression by flow cytometry.

Formaldehyde-fixed bacterial samples were evaluated by flow cytometry. The fluorescent signal was quantified with a FACSCalibur flow cytometer running Cellquest software (Becton Dickinson). All subsequent analysis was performed with WinMidi software (http://www.bio.umass.edu/mcbfacs/flowcat.html#contents).

Detection of surface labeling by flow cytometry.

A flow cytometric assay was carried out to study the attachment of NspA-specific antibodies at the intact bacterial surface as described by Rioux et al. (21). For this assay, bacteria were grown to early exponential phase in MH broth, the optical density at 490 nm was adjusted to 0.5, and bacteria were harvested by centrifugation for 2 min at 14,000 × g.

Immunogold electron microscopy.

The attachment of NspA-specific antibodies at the surface of whole neisserial cells was carried out as described by Rioux et al. (21).

Preparation of NOMVs.

Native outer membrane vesicles (NOMVs) were extracted from overnight cultures in Frantz medium containing 5 μg of EDDHA per ml as described by Saunders et al. (25). This method does not involve treatment with detergent, which may interfere with the correct folding of NspA.

Protein determination.

Protein concentrations were determined with the Pierce bicinchoninic acid protein assay.

Immunogold labeling of NOMVs.

NOMVs were dried onto a carbon-coated copper grid, which was then placed in a solution of anti-NspA MAb and incubated at room temperature for 2 h. The grid was washed twice in blocking buffer (PBS containing 2% bovine serum albumin) and placed in a 1:20 dilution of 12-nm colloidal gold particles conjugated to AffiniPure anti-mouse IgG (Jackson ImmunoResearch laboratories) in 0.02% (vol/vol) Tween 20, 0.1% (wt/vol) bovine serum albumin, 5% (vol/vol) newborn calf serum in PBS. Following 1 h of incubation at room temperature, the grids were washed twice in blocking buffer and once in distilled H₂O and stained with 1% (wt/vol) potassium phosphotungstate for 5 to 10 s. Samples were examined by electron microscopy.

Absorption of sera.

N. flavescens 2830 was grown overnight on BHI agar supplemented with 1% (vol/vol) horse serum. Bacteria were harvested into PBS, the optical density at 600 nm was adjusted to 10, and 400 μl of this suspension was harvested by centrifugation for 10 min at 14,000 × g. The bacteria were then resuspended in 300-μl aliquots of pooled murine sera raised against N. flavescens NOMVs containing NspA and incubated at room temperature for 90 min, with brief agitation every 20 min. This was repeated three times (i.e., four rounds of absorption) with a fresh 400-μl aliquot of N. flavescens 2830 from the suspension in PBS for each round.

Protection by passive immunization.

N. flavescens NOMV vaccines containing the adjuvant aluminum hydroxide (Alhydrogel; Superfos Biosector; final concentration, 4 mg ml⁻¹) were prepared as described by Oliver et al. (18). NIH mice (6 to 8 weeks old) (Harlan) were immunized with 10 μg of protein in 0.2 ml by subcutaneous injection on days 1, 21, and 28. Polyclonal sera were prepared from blood taken at a terminal bleed on day 35. A 1:5 dilution of murine polyclonal antiserum (50 μl) was administered by intraperitoneal injection to groups of 6- to 8-week-old NIH mice 2 h before challenge and 2 h and 3 to 5 h postchallenge. The growth medium used to prepare the challenge inoculum was used as the diluent. Mice were then challenged by intraperitoneal injection of N. meningitidis strain 608B (B:2a:P1.2:L3) as described by Oliver et al. (18). In brief, bacteria grown in iron-restricted Frantz medium were mixed with an equal volume of a solution of human transferrin (40 mg/ml; Sigma) in PBS. Mice received challenge doses, and 24 h later, a second intraperitoneal injection of 0.2 ml of saline containing human transferrin (50 mg/ml) was administered. Signs of ill health were monitored for up to 70 h postinfection, and mice were humanely killed before the preset severity limit of the experiment was exceeded.

Opsonophagocytosis assay.

N. meningitidis strains were subcultured into 10 ml of Frantz medium, centrifuged at 5,000 × g for 15 min, and resuspended in 1 ml of Sørnes buffer (Dulbecco's PBS with 5 mM glucose, 0.9 mM CaCl₂, 0.5 mM MgSO₄ and 5 mg of bovine serum albumin per ml) (8) containing 100 μg of the fluorochrome 2′,7′-bis-(2-carboxyethyl)-5-(and 6-)-carboxyfluorescein-acetoxymethyl ester (Molecular Probes) per ml. Incubation was continued for 1 h before bacteria were centrifuged at 4,000 × g for 15 min and resuspended in 10 ml of Sørnes buffer. Bacteria were killed by the addition of sodium azide and phenylmethylsulfonyl fluoride to 0.2% (wt/vol) and 100 μM, respectively, and the suspension was held for 48 h at 37°C. After lack of viability was determined, killed bacteria were washed, resuspended in Sørnes buffer, and kept at 4°C until used.

HL60 cells (European Collection of Cell Cultures) were differentiated by incubation for 5 days in RPMI medium, 20% (vol/vol) bovine serum, and 0.8% (vol/vol) dimethylformamide. For an opsonophagocytosis assay, 20 μl of a suspension of bacteria at 6.25 × 10⁸/ml was added to 20 μl of 1:10 diluted serum samples in the wells of a microtiter plate (all dilutions with Sørnes buffer). The plate was then incubated for 30 min at 37°C with shaking, followed by the addition to each well of 10 μl of Pel-Freez baby rabbit Complement (Mast Diagnostics) and further incubation for 15 min; 40 μl of a suspension of differentiated cells (2.5 × 10⁷/ml in RPMI medium) was added to the serum-bacteria mixture, and the incubation was continued for a further 30 min. The reaction was stopped by adding 80 μl of ice-cold Sørnes buffer, and the contents of each well was transferred to tubes containing 400 μl of ice-cold Dulbecco's PBS with 0.02% (wt/vol) EDTA. Samples were kept on ice until analyzed by flow cytometry; 10⁴ cells from each sample were evaluated, with a horizontal fluorescence gate set to include ≤10% of the population of antibody-negative, complement-only controls. The fluorescence index (FI) was calculated by taking the mean fluorescence of a sample cell population (X-mean) and multiplying by the proportion of cells that are present in the fluorescence gate (percent gated). The FI ratio was calculated by taking the mean FI of each serum and dividing by the mean FI of the complement-only controls. Thus, background (complement-only) levels of phagocytosis have a value of 1.

Serum bactericidal assays were performed according to standardized methodology (Centers for Disease Control, Report of the 2nd International Workshop on Meningococcal Immunology and Serology, 1992) as described by Oliver et al. (18).

RESULTS

Conjugation-positive strains.

A collection of commensal neisseriae were screened for their ability to take up plasmids pMIDG100 and pMIDG101 (Fig. 1) with a conjugation technique previously shown to work in N. meningitidis serogroups B and C (29). pMIDG100 (promoterless gfp gene) and pMIDG101 (gfp under the control of the strong meningococcal ner promoter) were successfully conjugated into four of four N. cinerea strains, two of four N. flavescens strains, one of one N. flava strain, one of one N. subflava strain, and two of three N. sicca strains. Attempts to transfer pMIDG100 into a total of 53 N. lactamica and two N. polysaccharea strains proved unsuccessful. The conjugation-positive strains selected for further work were N. cinerea NRL32165, N. flavescens 2830, N. subflava NRL30017, and N. flava NRL30008. Transformants were initially screened by colony PCR with primers designed against pMIDG100 to detect the presence of the plasmid. The plasmid was isolated from conjugation-positive strains to ensure that it had not inserted into the chromosome, and confirmatory restriction digestions were carried out to verify that it was intact.

Detection of GFP expression.

GFP expression was first detected by Western blotting (Fig. 2A), followed by flow cytometric analysis (Fig. 2B). No GFP expression was detected in the wild-type neisserial strains or in strains harboring pMIDG100. In contrast, both methods clearly demonstrated that transformed commensals expressed GFP under the control of the ner promoter.

FIG. 2.

A) Western immunoblot of whole cells, demonstrating GFP expression in commensal neisseriae. Recombinant GFP (5 μg) was used as a positive control. GFP was detected in the strains harboring pMIDG101, while all wild-type strains and those harboring pMIDG100 showed no reactivity. Gel 1 lanes: 1, rGFP; 2 to 4, N. cinerea NRL32165, wild type and carrying pMIDG100 or pMIDG101, respectively; 5 to 7, N. subflava NRL30017, wild type and carrying pMIDG100 or pMIDG101, respectively; 8 to 10, N. flava NRL30008, wild type and carrying pMIDG100 or pMIDG101, respectively. Gel 2 lanes: 1, recombinant GFP; 2 to 4, N. flavescens 2830 wild type and carrying pMIDG100 or pMIDG101, respectively. B) Flow cytometric analysis, demonstrating expression of GFP in commensal neisseriae. Shaded profiles represent no expression of GFP in strains containing pMIDG100 (promoterless GFP); open profiles represent expression of GFP in strains containing pMIDG101 (GFP under the control of the ner promoter).

nspA expression in commensal neisseriae.

Published data from Martin et al. (10) suggested that nspA would not be found in the conjugation-positive strains of the commensal Neisseria species studied. We confirmed this by Southern hybridization under high-stringency conditions with labeled nspA from N. meningitidis 608B used to probe chromosomal DNA isolated from different Neisseria strains. nspA was found to be present in N. lactamica and N. polysaccharea but absent from all N. cinerea, N. flavescens, N. subflava, N. flava, and N. sicca strains tested (results not shown).

pMIDG201nspA was conjugated into the five Nal^r conjugation-positive commensal species, and whole-cell proteins extracted from the wild-type and transconjugant strains were examined by Western blot analysis with MAbs Me-7 and Me-23 (Fig. 3). As expected, there was no expression of NspA from the wild-type strains, in contrast to substantial expression in strains harboring pMIDG201nspA. The blots were also probed with MAb Me-14, which binds to a conformational epitope; this failed to bind or bound very weakly (results not shown).

FIG. 3.

NspA expression in commensal neisseriae detected by Western blotting with anti-NspA MAbs. The 18.4-kDa NspA band was detected in strains harboring pMIDG201nspA. No wild-type strain showed reactivity. Lanes: 1 and 2, N. flava wild-type and with pMIDG201nspA, respectively; 3 and 4, N. subflava wild-type and with pMIDG201nspA, respectively; 5 and 6, N. cinerea wild-type and with pMIDG201nspA, respectively; 7 and 8, N. flavescens wild-type and with pMIDG201nspA, respectively.

Detection of NspA expression by flow cytometry.

A flow cytometric assay was used to study the attachment of NspA-specific antibodies at the intact bacterial surface. Whole cells were probed with fluorescein-conjugated secondary antibody only, as a negative control; Me-23, its cognate epitope inaccessible on intact cells, also therefore a negative control; Me-7, binding one exposed loop; and Me-14, binding a conformational epitope. Antibodies did not bind to the wild-type strains, but all anti-NspA MAbs except Me-23 clearly attached to intact cells of all strains expressing NspA (Fig. 4).

FIG. 4.

Flow cytometric analysis, demonstrating binding of anti-NspA MAbs to N. flavescens 2830. Shaded profiles represent binding to wild-type N. flavescens 2830, and open profiles represent binding to N. flavescens 2830/pMIDG201nspA.

The results of the surface-labeling experiments, summarized in Table 1, indicated that NspA was exposed on the surface of the transconjugant commensal neisseriae. The anti-NspA MAbs efficiently recognized their corresponding surface-exposed epitopes on all strains containing pMIDG201nspA while failing to bind to the wild-type strains. The N. flavescens and N. cinerea transconjugants expressing NspA had high binding indices, comparable to that seen to native NspA from N. meningitidis 608B.

TABLE 1.

Surface accessibility for three anti-NspA MAbs and a panel of Neisseria strains^a

Strain	Binding index
	Me-23	Me-7	Me-14
N. flavescens/pMIDG201nspA	1	73	71
N. cinerea/pMIDG201nspA	1	53	59
N. subflava/pMIDG201nspA	1	15	13
N. flava/pMIDG201nspA	1	42	42
N. meningitidis 608B	1	66	15

^aAll wild-type strains had a binding index of 1 for all three MAbs. The binding index = median fluorescence of samples probed with anti-NspA MAb + fluorescein isothiocyanate-conjugated secondary antibody, divided by the median fluorescence obtained with secondary antibody alone.

Immunogold labeling.

Binding of anti-NspA MAbs to the surface of whole neisserial cells was confirmed by immunogold electron microscopy. The wild-type strains remained unlabeled when incubated with the anti-NspA MAbs followed by a gold-conjugated secondary antibody (results not shown). Cells expressing NspA reacted with Me-7 and Me-14 gold conjugates. Uniform labeling of all cells was seen in the N. flavescens (Fig. 5A) and N. cinerea (results not shown) transconjugants. Only sparse binding of anti-NspA antibody was seen in the case of N. meningitidis 608B (results not shown). This may be explained by the facts that, in contrast to the commensals, this strain is encapsulated and the capsule may mask the antigen, and that nspA is expressed from a single chromosomal copy of the gene rather than a plasmid.

FIG. 5.

A) Transmission electron micrographs of whole N. flavescens/pMIDG201nspA. The strain was probed with the anti-NspA MAbs Me-23, Me-7, and Me-14. B) Constituents of N. flavescens 2830 NOMVs; SDS-12% PAGE gel, silver stained. Lanes: 1, wild-type N. flavescens 2830; 2, N. flavescens 2830/pMIDG201nspA; 3, purified recombinant NspA.

NOMVs.

NOMVs were isolated from N. flavescens 2830 because it showed the highest level of NspA expression. The NOMV preparations were separated on a 12% Bis-Tris NuPAGE gel, and constituent components were visualized by silver staining (Fig. 5B). NspA is clearly present in the N. flavescens NOMVs containing pMIDG201nspA and the positive control purified recombinant NspA, prepared as described by Martin et al. (10), while absent in the wild-type strain. Expression of NspA on the surface of NOMVs was confirmed by immunogold electron microscopy (results not shown).

Role of anti-NspA antibodies in passive protection.

To determine the possible contribution that anti-NspA antibodies make to the protection that NOMVs afford against experimental meningococcal infection, sera raised in mice against N. flavescens NspA-enhanced NOMVs were used in a passive-protection experiment. NOMVs from N. flavescens and other commensal neisseriae themselves provide effective protection (10, 17). Pooled sera were subjected to four rounds of absorption to remove background N. flavescens 2830 NOMV antibodies, confirmed by Western blotting (Fig. 6). The identity of the prominent band seen on probing with absorbed serum was confirmed as NspA with MAb Me-7 (Fig. 6C). This absorbed serum retained full protective efficacy against experimental infection with 10⁷ CFU of N. meningitidis in the mouse model. Five mice (each) were passively immunized by intraperitoneal injection of (i) pooled serum collected from mice previously vaccinated with N. flavescens NspA-enhanced NOMVs and (ii) such serum from which background anti-N. flavescens 2830 NOMV antibody was removed by absorption. An additional five (control) mice were given pooled normal mouse serum. At 72 h after experimental infection with 10⁷ CFU of N. meningitidis 608B, three of the mice given unabsorbed serum from previously vaccinated animals survived; four of the mice given absorbed serum, retaining anti-NspA activity, survived; but none of the mice given normal mouse serum survived.

FIG. 6.

Western blot of proteins extracted from NOMVs. Lanes: 1, wild-type N. flavescens; 2, N. flavescens bearing pMIDG201nspA. Panels show results obtained with serum following vaccination with NOMVs from N. flavescens bearing pMIDG201nspA: A) unabsorbed; B) absorbed with wild-type N. flavescens; C) probed with anti-NspA MAb Me-7. The arrow indicates NspA.

Opsonophagocytic activity of anti-NspA antibodies.

Sera from mice vaccinated with meningococcal OMVs have previously been shown to opsonize whole meningococci or OMV-coated beads (9). To investigate the capacity of NspA to elicit antibody with opsonophagocytic activity, sera from mice vaccinated with N. flavescens wild-type OMVs and N. flavescens OMVs expressing NspA were assessed with the wild-type meningococcal strain 608B (expressing NspA) and its nspA knockout mutant. Opsonophagocytic activity was also assessed against an unrelated control meningococcal strain, NZ98/254 (lineage III) (Fig. 7). A marked increase in phagocytosis was observed in the presence of the anti-NspA serum against strains NZ98/254 and 608B, with no activity observed for strain 608B nspA. Antibodies raised against killed whole cells of wild-type E. coli and E. coli expressing NspA were also used and showed a similar NspA-dependent response.

FIG. 7.

Opsonophagocytic activity of mouse sera. Activity in murine sera raised against N. flavescens vesicles, killed wild-type E. coli, vesicles from N. flavescens expressing NspA, and killed E. coli expressing NspA. Serogroup B target strains (x axis) were wild-type 608B and its nspA mutant and NZ98/254.

The serum samples and meningococcal strains described above were also used to test for serum bactericidal activity. None was detected.

DISCUSSION

In this study, both a cytosolic eukaryotic protein (GFP) and a surface-exposed meningococcal OMP (NspA) were successfully expressed in a range of commensal neisseriae. The surface localization of NspA in intact bacterial cells was established with well-characterized MAbs, and the protein was demonstrated in apparently native conformation at the surface of recombinant commensal neisserial strains.

NspA has been shown to be highly conserved across diverse meningococcal strains and has been studied intensively as a potential vaccine antigen (4, 10, 11, 12, 13, 14, 19). Sera from mice vaccinated with purified recombinant meningococcal NspA prepared in E. coli are bactericidal to a range of different serogroups and have protected animals against experimental intraperitoneal challenge with a potentially lethal infecting dose of serogroup B meningococci (10). Purified recombinant NspA also elicits antimeningococcal bactericidal responses in rabbits and monkeys (12). Despite these promising results, purified recombinant NspA prepared in this way has so far failed to induce significant levels of bactericidal antibody in humans.

Hou et al. (6), using two protective anti-NspA MAbs which recognize conformational epitopes, recently demonstrated that optimally immunogenic vaccines with purified recombinant NspA will require formulations that permit proper folding of the protein. Recombinant E. coli expression has also been found to be unsuitable for expression of other meningococcal OMP vaccine components (20), leading this group to develop systems for the genetic upregulation of certain OMPs in the meningococcus and then isolation of the modified OMVs.

We sought bactericidal activity in sera from mice vaccinated with NspA-enhanced NOMVs but found no significant difference in serum bactericidal activity between these and control samples. However, these sera clearly protected against meningococcal infection and this protection most likely derives from opsonophagocytic activity (Fig. 7). NOMVs isolated from the commensal neisseriae are themselves very effective in protecting against experimental meningococcal disease, which makes demonstration of any additional protection provided by NspA in NspA-enriched NOMVs difficult. Nonetheless, the absorption experiment demonstrates that substantial protection appears to derive from NspA-specific antibody. (We acknowledge the formal possibility that protection may derive from antibody to one or more of the unidentified, low-abundance N. flavescens proteins remaining in the absorbed serum [Fig. 6b]; unlikely as it must be that one of these is a protective antigen, their identification will be pursued.)

Meningococcal OMV vaccines have been used more widely than any other formulation in attempts to control serogroup B disease. Two such vaccines, used in outbreaks in Cuba and Norway, demonstrated protective efficacy in clinical settings in adults and older children (1, 26), and a similar vaccine is to be used to combat the current highly clonal epidemic of serogroup B disease in New Zealand (22). However, the vaccines used in Cuba and Norway do not appear to be as effective in infants (the population at greatest risk) as they are in older individuals (15), and the dominance of PorA in these preparations and their corresponding lack of cross-protective efficacy means that their main value is likely to be restricted to the kind of outbreak settings in which they have so far been deployed.

Tondella et al. (28) calculated that, based on PorA serosubtypes currently causing disease in the United States, some 20 PorA types would need to be included to give 80% coverage against these strains. In addition, a rapid change in the prevalent serosubtypes has been observed in the United Kingdom (7), which suggests that such vaccines will have to be reformulated regularly. OMVs prepared from commensal neisseriae lack PorA, and the immune response that they elicit, which is cross-protective against different meningococcal strains (18), appears not to be directed towards any other single immunodominant protein (18). They possess the desirable quality of being safely prepared in bulk (obviating the need for the high-level containment facilities required for vaccine manufacture with large volumes of N. meningitidis), and formulations enhanced by the strategic incorporation of antigens such as NspA, which elicit unequivocal protective antibody when presented in this way, hold promise as a potential cross-protective vaccine to prevent disease caused by N. meningitidis.