|Home | About | Journals | Submit | Contact Us | Français|
Melioidosis is an emerging disease of humans in Southeast Asia and tropical Australia. The bacterium causing this disease, Burkholderia pseudomallei, is also considered a bioterrorism agent, and as yet there is no licensed vaccine for preventing B. pseudomallei infection. In this study, we evaluated selected proteins (LolC, PotF, and OppA) of the ATP-binding cassette systems of B. pseudomallei as candidate vaccine antigens. Nonmembrane regions of the B. pseudomallei proteins were expressed and purified from Escherichia coli and then evaluated as vaccine candidates in an established mouse model of B. pseudomallei infection. When delivered with the monophosphoryl lipid A-trehalose dicorynomycolate adjuvant, the proteins stimulated antigen-specific humoral and cellular immune responses. Immunization with LolC or PotF protein domains afforded significant protection against a subsequent challenge with B. pseudomallei. The most promising vaccine candidate, LolC, provided a greater level of protection when it was administered with immune-stimulating complexes complexed with CpG oligodeoxynucleotide 10103. Immunization with LolC also protected against a subsequent challenge with a heterologous strain of B. pseudomallei, demonstrating the potential utility of this protein as a vaccine antigen for melioidosis.
The bacterium Burkholderia pseudomallei is a gram-negative saprophytic bacillus that is found predominantly in Southeast Asia and tropical Australia and is the causative agent of the emerging infection melioidosis (33). Infection of humans with B. pseudomallei may occur by various routes, including via wounds and existing skin lesions (7), aspiration of contaminated water during near-drowning (10, 28), and inhalation of organisms (30). The clinical manifestations of melioidosis are not uniform, and patients can present with pneumonia, skin abscesses, soft tissue abscesses, or osteomyelitis/septic arthritis. Pneumonic melioidosis is presented in over 50% of acute cases of the disease, and disseminated pneumonia is associated with high mortality (25). Chronic melioidosis usually occurs following acute melioidosis, and relapsing melioidosis can result from the reactivation of a latent B. pseudomallei infection, often due to the withdrawal of antibiotics or the failure to complete prescribed courses of antibiotics.
Melioidosis is typically treated via an antibiotic regimen, and one of the most effective antibiotics is ceftazidime, a β-lactam antibiotic, the use of which has reduced the mortality rate by 50% compared to the previous regimen (1). However, multidrug resistance in B. pseudomallei is a significant problem in the treatment of melioidosis, so alternative countermeasures are needed. Furthermore, there is no human vaccine currently licensed for protection against melioidosis, and effective vaccination is considered likely to require a cell-mediated immune response due to the intracellular nature of B. pseudomallei infection. Subunit vaccines carry no risks associated with persistence, latency, or reactivation and are generally considered safe. Immunization with B. pseudomallei-derived lipopolysaccharide (LPS) and capsular polysaccharide has been shown to result in a delayed time to death in mice challenged with the bacteria (27). To improve the efficacy of the polysaccharide vaccines, it may be necessary to identify additional protein antigens that also afford protection against melioidosis. To our knowledge, only immunization with a DNA vaccine encoding the flagellin protein FliC (8, 9) or sera raised against FliC (5) has been shown to afford any protection against B. pseudomallei infection in mice.
In an effort to identify novel vaccine antigens for B. pseudomallei, we have evaluated proteins of the ATP-binding cassette (ABC) systems of this organism. The ABC systems have roles in bacterial survival, virulence, and pathogenicity, and we have identified ABC system proteins as candidate targets for developing medical countermeasures against bacterial infections (16). The common feature of all ABC systems is the binding of ATP, and highly conserved ABC proteins of the ABC systems are responsible for the generation of energy required for transport. In addition to the ABC proteins, other proteins are required to complete the function of some ABC systems, particularly those involved in import functions (13). In general, the location of ABC systems in bacterial cell membranes implies a likelihood that at least some of the protein components of the systems are exposed to the immune system during infection. Indeed, ABC system proteins that are immunoreactive with convalescent-phase sera have been identified in both gram-positive and gram-negative bacteria (16). Furthermore, an increasing number of ABC system proteins are being considered as candidate vaccine antigens (6, 31, 34, 35).
The complete inventory of ABC systems of B. pseudomallei strain K96243 has recently been compiled (18). During compilation of this inventory, we identified and selected putative ABC system proteins of B. pseudomallei that may constitute candidate vaccine antigens for melioidosis, and here we assessed the potential of these candidates to induce protective immunity against B. pseudomallei in a mouse model of infection.
Escherichia coli TOP10F′ cells (Invitrogen) were used for cloning experiments, and strains of E. coli BL21 (Invitrogen) were used for protein expression and purification studies. E. coli strains were cultured in Luria-Bertani medium or 2× YTG medium at 37°C; when required, the medium was supplemented with ampicillin (100 μg/ml) or ampicillin (100 μg/ml) plus chloramphenicol (34 μg/ml). B. pseudomallei strains were obtained from culture stocks held at the Defense Science and Technology Laboratory. B. pseudomallei strains K96243 and 576 were cultured in Luria-Bertani medium at 37°C. B. pseudomallei genomic DNAs were produced by a lysozyme-phenol-chloroform extraction method (22). All chemicals were obtained from Sigma-Aldrich Co. Ltd. unless otherwise stated. All reagents and kits were obtained from commercial sources, and the manufacturers’ instructions were followed.
The predicted sequences of the selected ABC system proteins were downloaded from the NCBI website (http://www.ncbi.nlm.nih.gov/entrez/). Two Internet-based programs, TMHMM v 2.0 (http://www.cbs.dtu.dk/services/TMHMM-2.0/) and SignalP (http://www.cbs.dtu.dk/services/SignalP/), were used to predict regions of the proteins that encode membrane-spanning domains and signal peptides, respectively.
The nonmembrane, nonsignal peptide regions of the open reading frames (ORFs) encoding LolC, PotF, and OppA (lolC, potF, and oppA, respectively) were amplified from B. pseudomallei K96243 genomic DNA by PCR using oligonucleotide primers listed in Table Table11 and Pfu polymerase (Stratagene). The PCR products were subsequently cloned into the vector pCRT7/NT-TOPO (Invitrogen) for expression of the protein domains with N-terminus-associated His6 tags in E. coli BL21(DE3)(pLysS) cells. The authenticity of the cloned DNA and the polyhistidine tag was confirmed by nucleotide sequencing of these regions. Expression of the PotF and OppA proteins was induced in E. coli cultured at 180 rpm to an optical density at 600 nm of 0.5 using 1 M isopropyl-β-d-thiogalactoside (IPTG). The bacteria were subsequently cultured for a further 3 h at 37°C and 180 rpm, and protein expression was detected in cell culture supernatants by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting using a horseradish peroxidase-labeled mouse anti-His immunoglobulin G (IgG) antibody. The LolC protein was expressed in E. coli BL21 Star(DE3)(pLysS) cells grown to the mid-log phase at 37°C and then induced with 0.1 M IPTG, cooled to 20°C, and cultured overnight. The proteins were purified from cell culture supernatants using immobilized metal affinity chromatography carried out using HisTrap prepacked columns (GE Healthcare). Briefly, E. coli cells were disrupted by sonication three times for 30 s and centrifuged at 15,000 × g for 15 min, and then the resulting cell supernatant was sterilized using 0.45-μm filters. The supernatants were loaded onto the columns and eluted using buffer containing up to 500 nm imidazole, which was subsequently removed by buffer exchange, leaving the protein buffered in phosphate-buffered saline (PBS). The purity of the resulting proteins was examined by SDS-PAGE followed by staining with Coomassie brilliant blue dye, and the protein concentrations were determined using a bicinchoninic acid assay (Pierce Biotechnology). Calculations of the molecular weights of proteins were carried out using the mwt/pI tool at http://www.expasy.ch. Circular dichroism (CD) measurement was performed using an AVIV 62DS spectropolarimeter. Data were recorded at 0.5-nm intervals and 25°C for the LolC domain (0.43 mg/ml) in PBS and PBS baselines, using 0.005-cm-path-length cells. Spectra were analyzed by using the Dichroweb server (21, 36). Comparative sequence and secondary structure predictions for B. pseudomallei and E. coli LolC were performed using ClustalW (http://www.ebi.ac.uk/clustalw) and PHDsec from EMBL (http://www.embl-heidelberg.de/Services/index.html), respectively.
Groups of six or nine female 5- to 6-week-old BALB/c mice (Charles River Laboratories) were immunized with the purified recombinant LolC, OppA, or PotF protein given with adjuvant. The proteins were prepared for immunization at a concentration of 100 μg/ml in PBS mixed 1:1 with a monophosphoryl lipid A-trehalose dicorynomycolate (MPL+TDM) adjuvant (Sigma-Aldrich) or 1:1 with Emulsigen adjuvant (MVP Laboratories) or mixed with 12.5 μg of CpG oligodeoxynucleotide (ODN) 10103 (Coley Pharmaceuticals), either in complex with 12.5 μg immune-stimulating complex (ISCOM) AbISCO 100 (Isconova AB), in complex with Emulsigen, or alone. The proteins were delivered by intraperitoneal (i.p.) injection of 100 μl of each protein in MPL+TDM adjuvant on days 0, 14, and 28 or by subcutaneous (s.c.) injection with Emulsigen, CpG, or Emulsigen-CpG on days 0, 14 and 28. Proteins with the adjuvant ISCOM-CpG complex were delivered s.c. on days 0, 28, and 49. Alternatively, groups of six female 5- to 6-week-old BALB/c mice were immunized with an auxotrophic mutant of B. pseudomallei referred to as 2D2 (2, 19). On days 0 and 14, 1 × 106 CFU of 2D2 in sterile PBS was delivered by i.p. injection in 200-μl doses. Groups of age-matched mice were left untreated (naïve) or given adjuvant only as controls.
Mice immunized with the LolC, PotF, or OppA protein or appropriate controls were challenged with B. pseudomallei K96243 on day 70. B. pseudomallei K96243 was routinely cultured for 16 to 18 h and prepared at concentrations of 4 × 105 and 7 × 105 CFU/ml. Alternatively, B. pseudomallei 576 was cultured, subsequently stored frozen, and then thawed prior to use at 6 × 106 CFU/ml. Approximately 50% of the bacteria remained viable after freezing and thawing. Mice were challenged via the i.p. route with a 100-μl inoculum. The median lethal doses of B. pseudomallei K96243 and B. pseudomallei 576 were previously shown to be approximately 103 CFU (unpublished data) and 80 CFU (2), respectively. Subsequently, the animals were closely observed for up to 56 days after challenge to determine their protected status. Humane endpoints were strictly observed so that any animal which displayed a collection of clinical signs that indicated a lethal infection (piloerection, hunching, mobility problems) was culled.
Approximately 100 μl of blood was removed from the tail veins of immunized mice 7 days prior to challenge with B. pseudomallei. The blood was allowed to clot at 4°C for 24 h and then centrifuged at 13,000 rpm, and the serum was removed and stored at −20°C. Subsequently, the LolC-specific, OppA-specific, and PotF-specific IgG1 or IgG2a responses in the serum samples were determined by an enzyme-linked immunosorbent assay (ELISA). Briefly, microtiter plates were coated with 5 μg/ml of the appropriate recombinant protein in PBS for 16 to 18 h. Nonspecific binding was blocked using 2% (wt/vol) bovine serum albumin in PBS (BSA-PBS) for 1 h at 37°C. The plates were washed three times using 0.05% (vol/vol) Tween 20 in PBS, and twofold dilutions of test serum samples in BSA-PBS were added in duplicate and incubated for 2 h at 37°C. Following washing, horseradish peroxidase-labeled goat anti-mouse IgG1 or IgG2a (Oxford Biotechnology) diluted in BSA-PBS was added and incubated for 2 h at 37°C. The substrate 2,2′-azino-di-3-ethylbenzthiazoline sulfonate (ABTS) was added, and the optical density at 414 nm was measured 20 min after addition of the substrate. Antibody concentrations (in ng/ml) were calculated from standard curves generated with IgG1 or IgG2a antibody preparations whose concentrations were known using the Ascent software, version 2.4.2 (Thermo Electron Corporation).
Responses to immunization with LolC, PotF, or OppA were analyzed by ELISA measurement of gamma interferon (IFN-γ) in restimulated spleen and lymph node cell culture supernatants. Briefly, three mice from groups of nine animals were humanely culled, and spleens and lymph nodes were removed on day 63. Single-cell suspensions of the spleens and lymph nodes were prepared by passage through 100-μm cell strainers (Becton Dickinson), and erythrocytes were removed from the suspensions using lysis buffer (Sigma). The remaining cells were washed and resuspended in RPMI 1640 medium (Life Technologies) supplemented with 10% (vol/vol) fetal calf serum, 10 mM l-glutamine, 200 U/ml penicillin, 200 μg/ml streptomycin, and 50 μM 2-mercaptoethanol (R10 medium). The cells were plated in U-bottom 96-well plates at a density of 2.5 × 106 cells/ml. Subsequently, triplicate samples were stimulated with appropriate concentrations of LolC, PotF, or OppA and incubated at 37°C in 5% (vol/vol) CO2, and supernatants were collected for ELISA on day 3. Alternatively, mice vaccinated with B. pseudomallei strain 2D2 were humanely culled on day 35, and spleens were removed and prepared for stimulation as described above. Briefly, cultures were stimulated with 5 × 104 CFU of gamma-irradiated B. pseudomallei 576 or different concentrations of LolC, PotF, or OppA. Cultures were incubated at 37°C in 5% CO2, and supernatants were collected for ELISA on day 3.
Rat anti-murine IFN-γ antibody pairs, AN18, and biotinylated R46A2 (Mabtech) were used in ELISAs according to the manufacturer's instructions. The ELISAs were developed with 1 μg/ml strepavidin-conjugated peroxidase and SureBlue substrate (KPL Laboratories), and the development was stopped by addition of 2 M H2SO4. Each sample from a triplicate restimulation group was tested in duplicate wells, and the mean value for duplicate wells was read with a Dynex MRXII plate reader at a wavelength of 450 nm. The subsequent analysis of data was performed using Revelation software, version 4.25.
The genomic DNA from strains of B. pseudomallei, B. mallei, and B. thailandensis (Table (Table2)2) was screened by PCR for the presence of lolC using oligonucleotide primers described in Table Table11 and Pfu polymerase.
Log rank tests were used to assess the significance of protection at the 95% confidence level and were performed using GraphPad Prism, version 3.02 for Windows software (GraphPad Software).
Three putative B. pseudomallei proteins were selected for evaluation as candidate vaccine antigens for melioidosis: LolC (BPSL2277), PotF (BPSS0467), and OppA (BPSS2141). This selection was based upon the original B. pseudomallei K96423 genome annotation (19), homology to proteins in other pathogens, and reports of known virulence and immunogenicity properties of the proteins in other bacteria. Briefly, LolC is a membrane protein associated with the LolCDE ABC system, which is involved in lipoprotein sorting between the inner and outer membranes of gram-negative bacteria (38). In E. coli, disruption of lolC is lethal and prevents the release of lipoproteins from the inner membrane (26). PotF is a periplasmic binding protein of the PotFGHI system involved in putrescine import in E. coli (31), a system that has homologues required for attachment to plant cells and for full virulence in the phytopathogen Agrobacterium tumefaciens (24). Finally, OppA is an oligopeptide-binding protein of the Opp system in E. coli (14). It is also a virulence factor involved in intracellular survival in Listeria monocytogenes (4) and has recently been shown to be a protective antigen of Yersinia pestis (34).
Likely membrane-associated regions in the sequences of selected ABC system homologues, coupled with known structures or functions of the proteins in other bacteria, were used to reach the following conclusions. The LolC protein is predicted to contain multiple transmembrane domains, as well as a large periplasmic domain, residues 44 to 266, which is similar to that predicted for the E. coli LolC protein (20). The PotF protein is predicted to be secreted via the Sec machinery with an N-terminal signal peptide, and similarly, the OppA protein is predicted to have a single N-terminal transmembrane domain that could act to anchor the periplasmic protein to the inner membrane.
Subsequently, the predicted nonmembrane, nonsignal peptide-encoding regions of LolC (residues 47 to 243), PotF (residues 24 to 364), and OppA (residues 44 to 554) were amplified from B. pseudomallei K96243 genomic DNA by PCR and cloned into the pCRT7/NT-TOPO vector (Invitrogen Ltd.) for expression as proteins fused to N-terminal His6 tags. Induction of the lac promoter encoded by this vector with IPTG was performed to express the protein domains in E. coli BL21-based cells. Recombinant proteins with predicted molecular masses of approximately 25 kDa (LolC), 38 kDa (PotF), and 57 kDa (OppA) were detected by SDS-PAGE and Western blotting. Subsequently, the LolC, PotF, and OppA protein domains were purified (Fig. (Fig.1)1) from culture using a nickel-Sepharose column, eluted with imidazole, and buffer exchanged into PBS. Protein concentrations were determined by the bicinchoninic acid assay. In Western blots these proteins reacted with an antibody directed against the polyhistidine tag.
The structural integrity of LolC was also assessed by CD. Analysis of the spectrum for the secondary structure content indicated that this protein, representing most of the predicted periplasmic domain of LolC, contained approximately 45% ± 3% α-helix, 14% ± 3% β-strand, 19% ± 3% β-turn, and 24% ± 5% other secondary structure components. A recent secondary structure prediction of the E. coli LolC protein suggests that residues 42 to 265 form a domain in the periplasm flanked by transmembrane helices (20). Alignment of this predicted periplasmic domain with the equivalent region of B. pseudomallei LolC (residues 46 to 266) using ClustalW indicated 30% sequence identity and 69% sequence similarity and very similar patterns of predicted helices, strands, and coils using PhDsec (data not shown). Taken together, these data support the observation that the recombinant domain of B. pseudomallei LolC produced in these studies adopts a folded conformation and predict similar secondary structure features in the related E. coli LolC periplasmic domain.
To determine whether the recombinant LolC, PotF, or OppA protein domains are recognized by B. pseudomallei-immune T cells, groups of BALB/c mice were immunized as described above. On day 35 postvaccination, spleens were removed and stimulated with 5 × 104 CFU of gamma-irradiated B. pseudomallei 576 or different concentrations of LolC, PotF, or OppA. Spleen cells immunized with B. pseudomallei 2D2 recognized all three proteins, generating appreciable levels of IFN-γ in culture supernatants by day 3 (Fig. (Fig.22).
To evaluate the immunogenicity of the recombinant truncated B. pseudomallei LolC, PotF, and OppA protein domains, groups of BALB/c mice were inoculated three times at 2-week intervals with 10 μg of the individual proteins mixed with the MPL+TDM adjuvant by i.p. injection. Sera from each group of mice sampled on day 63 were pooled and analyzed by ELISA for protein-specific IgG1 and IgG2a responses. In all cases the IgG1 or IgG2a responses to a purified His-tagged Francisella tularensis protein (L7 protein) were more than 10-fold lower than the responses to the B. pseudomallei proteins. This indicates that the majority of the antibody response was directed to the B. pseudomallei proteins rather than to the polyhistidine tag. Immunization with the LolC, PotF, and OppA protein domains generated measurable LolC-, PotF-, and OppA-specific IgG responses, respectively, with a bias towards IgG2a (Table (Table3),3), indicative of a TH1-type immune response in mice. In a separate experiment, groups of BALB/c mice were immunized as described above, and spleens and lymph nodes were removed on day 63 and then pooled for T-cell restimulation assays, as measured by IFN-γ ELISAs of cell culture supernatants. These assays showed that LolC, PotF, and OppA were able to stimulate protein-specific T-cell responses in cells isolated from the spleens of immunized mice, although LolC was the only antigen able to generate a measurable response in cells isolated from the lymph nodes (Table (Table33).
To determine if the recombinant truncated B. pseudomallei LolC, PotF, or OppA protein domains may constitute protective antigens, groups of BALB/c mice were immunized as described above and were subsequently challenged on day 63 via i.p. injection of approximately 4 × 104 CFU of B. pseudomallei K96243. The challenged mice were observed for 6 weeks, and humane endpoints were observed strictly. Naïve mice and control mice given only the MPL+TDM adjuvant all succumbed to the infection by day 19 and day 10 postchallenge, respectively (Fig. (Fig.3).3). In comparison, the recombinant LolC or PotF protein afforded significant protection against the B. pseudomallei challenge. However, the recombinant OppA protein did not offer significant protection compared to either naïve or adjuvant-only controls. Overall, LolC afforded the greatest level of protection in this mouse model of melioidosis, protecting five of six mice from succumbing to the B. pseudomallei infection for 6 weeks.
In order to attempt to improve the protection afforded by the recombinant LolC protein against B. pseudomallei, a range of adjuvants were compared for delivery of the protein. The adjuvants were selected based upon the hypothesis that they may be appropriate for stimulation of cell-mediated immune responses that are considered important in B. pseudomallei infection. At 2-week intervals, groups of BALB/c mice were given three s.c. doses of 10 μg of LolC protein in either the adjuvant MPL+TDM, Emulsigen, CpG ODN 10103, Emulsigen in complex with CpG ODN 10103, or ISCOMs in complex with CpG ODN 10103. On day 63, the immunized mice were challenged via the i.p. route with approximately 7 × 104 CFU B. pseudomallei K96423 and monitored for 13 days postchallenge as described above. Following challenge, naïve control mice all succumbed to the infection by day 5. There was no significant protection of mice immunized with LolC in the adjuvant MPL+TDM, Emulsigen, CpG ODN 10103, or Emulsigen plus CpG ODN 10103 (Fig. (Fig.4).4). However, there was significant protection of mice immunized with ISCOMs plus CpG ODN 10103 compared to either naïve (P = 0.019) or adjuvant-treated control mice (P = 0.024). In an additional study (data not shown), we have demonstrated the equivalence of the levels of protection against B. pseudomallei infection afforded by LolC with the adjuvant ISCOM-CpG complex and the protection afforded by the live attenuated vaccine candidate 2D2 (2). In order to better understand the protective immune response to the recombinant LolC protein with the adjuvant ISCOM-CpG complex, serum removed from the immunized mice on day 63 was assayed for LolC-specific IgG responses. As before, LolC with the adjuvant MPL+TDM resulted in a strong IgG2-biased LolC-specific IgG response, and, similarly, LolC with the ISCOM-CpG complex adjuvant also stimulated an IgG21-biased response, indicative of a TH2-type immune response in mice (data not shown).
In order for a subunit vaccine against melioidosis to have broad utility, the selected antigen should be able to protect against different strains of the organism. To this end, PCR was used to demonstrate the presence of the LolC-encoding gene in DNAs from 10 different strains of B. pseudomallei, including strains which express a different form of LPS O antigen than strain K96243. Using this approach, the LolC-encoding gene was also found to be present in the closely related pathogen B. mallei. To test the cross-protective efficacy of LolC, a challenge experiment using a different mouse model for melioidosis was performed. Groups of BALB/c mice were immunized on days 0, 14, and 28 with 10 μg of recombinant LolC protein with the MPL+TDM adjuvant via the i.p. route or on days 0, 28, and 49 with 10 μg of recombinant LolC protein with the adjuvant ISCOM-CpG via the s.c. route. On day 63 (MPL+TDM group) or day 84 (ISCOM-CpG group), mice were challenged with approximately 6.6 × 105 CFU of B. pseudomallei strain 576 administered via the i.p. route. Naïve mice succumbed to the infection by day 19. Similarly, MPL+TDM- and ISCOM-CpG-treated control mice succumbed by days 16 and 24, respectively (Fig. (Fig.5).5). In comparison, immunization with LolC administered with either adjuvant provided significant protection against the B. pseudomallei 576 challenge compared to naïve or adjuvant-only controls.
The aim of this study was to determine whether components of ABC systems may be exploited as vaccine antigens against melioidosis. ABC system proteins were selected as potential vaccine candidates because members of this family of proteins have been shown generally to be immunogenic and to play roles in bacterial virulence (16). In addition, some components of ABC systems have been shown to be vaccine antigens (6, 31, 34, 35). Homologues of the LolC, PotF, and OppA proteins identified in B. pseudomallei and soluble domains were produced in E. coli, with membrane-spanning regions removed in order to aid expression and purification for the study. Additionally, it was considered likely that the protein regions predicted to be present in the periplasm would be more likely to be involved in immunogenicity than the regions present within the inner membrane of the bacteria are.
Hamsters, diabetic rats, and outbred and inbred strains of mice have previously been used for small-animal models of melioidosis (37). We have previously used BALB/c mice inoculated by the i.p. route to evaluate live attenuated mutants of B. pseudomallei as vaccines (2, 11, 17). In addition we have shown that it is possible to demonstrate protection of immunized mice against an intranasal challenge with B. pseudomallei (11). Therefore, in the longer term we could use this model to test the ability of ABC transporter proteins to protect against pneumonic melioidosis. The initial study was designed to determine whether immunization of BALB/c mice with truncated LolC, PotF, or OppA afforded any protection against B. pseudomallei K96243 delivered by the i.p. route. In previous work evaluating subunit vaccines against B. pseudomallei, the MPL+TDM adjuvant was found to be effective (27) and therefore was considered appropriate for the initial evaluation of the ABC system proteins. MPL+TDM comprises nontoxic, highly refined, monophosphoryl lipid A from Salmonella enterica serovar Minnesota and synthetic trehalose dicorynomycolate, an analogue of trehalose dimycolate, from the cord factor of Mycobacterium tuberculosis. This adjuvant, an oil-in-water emulsion, applies a depot effect to the antigen, resulting in retention at the immunization site for an extended period of time, as well as stimulating the immune system due to lipid A stimulation of proinflammatory cytokines (23). In a previous study using MPL+TDM as an adjuvant for proteins, an IgG1 response was found to predominate in immunized mice (15). However, in this study the predominant IgG isotype response to the three different proteins tested was IgG2a, reflecting a TH1-type immune response. The importance of this type of immune response has been confirmed by the requirement for IFN-γ in controlling B. pseudomallei infection (17). Such TH1-type immunity results in the activation of macrophages, the secretion of IFN-γ and other cytokines, and the subsequent stimulation of the cell-mediated immune response likely required for protection against B. pseudomallei infection.
In this study, both the LolC and PotF proteins offered significant protection to immunized mice against approximately 40 minimal lethal doses (MLDs) of B. pseudomallei. However, immunization with LolC was more effective at both stimulating systemic cell-mediated immunity and affording protection against infection. When a range of adjuvants were tested for improving the protection afforded by LolC, ISCOMs complexed with CpG ODN 10103 were found to offer a level of protection against a larger challenge (approximately 70 MLDs) of B. pseudomallei greater than the level of protection observed with LolC with the MPL+TDM adjuvant. Although no significant protection against 70 MLDs of B. pseudomallei was observed in mice given LolC in complex with the MPL+TDM adjuvant in this experiment, other unpublished data obtained in our laboratories have confirmed this protection afforded by this combination of protein and adjuvant against B. pseudomallei. Furthermore, we have demonstrated the equivalence in protective efficacy of the combinations compared to the live attenuated mutant of B. pseudomallei, 2D2.
ISCOMs are cage-like structures assembled from cholesterol, phospholipids, and saponins that can selectively target antigen to phagocytic cells and enhance the induction of a cell-mediated immune response (32). Synthetically produced CpG ODNs act as immune response modifiers through activation of the innate immune response, since CpG DNA is able to up-regulate costimulatory molecules and increase cytokine secretion, antigen processing, and peptide-major histocompatibility complex stability through interaction with Toll-like receptor 9 (12). In this study, the use of ISCOMs together with CpG ODN as an adjuvant for LolC resulted in an LolC-specific IgG2a bias similar to that observed when MPL+TDM was used, indicating that a TH1-type immune response was stimulated.
The ability to protect against heterologous strains is an important feature of vaccines designed to prevent bacterial infection. Therefore, the LolC protein with either the MPL+TDM adjuvant or the adjuvant ISCOM-CpG complex was also evaluated for protection against B. pseudomallei 576, a strain of B. pseudomallei with an LPS O-antigen type different than that of K96243. This experiment showed that LolC delivered with either adjuvant was able to protect immunized mice. The lolC ORF encoding LolC was shown by PCR screening to be present in a range of B. pseudomallei strains, indicating the potential ability of the protein to protect against melioidosis caused by different strains of B. pseudomallei. Since lolC was also present in B. mallei, it is possible that the LolC protein may protect against glanders, the disease caused by B. mallei. The protective efficacy of immunization with LolC against glanders should be tested.
The mechanism by which LolC protects mice against melioidosis is not yet understood. The components of ABC transporters that may be expected to interact with the host immune system in gram-negative bacteria are the associated outer membrane proteins. However, LolC is an inner membrane protein in B. pseudomallei, and the truncated form used to immunize in this study is predicted to be located in the periplasm. Nevertheless, it is possible that LolC may be able to interact with the immune system since there are many examples of antibodies in convalescent-phase sera which are directed to cytoplasmic proteins of bacteria (16). This may reflect the surface location of at least part of the protein population or that bacteria with disrupted cell walls or bacteria which are taken up by antigen-presenting cells display cytoplasmic or periplasmic protein to the immune system. Since LolC stimulates strong humoral immune responses in immunized mice, it is possible that LolC-specific antibody could directly block the Lol lipoprotein transport system that is known to be essential in E. coli (26). Alternatively, LolC-specific antibody may have an opsonizing effect, or cell-mediated immunity to LolC may play a role in protection. Further work is required to determine the role of LolC in the virulence of B. pseudomallei and to elucidate how immunization with LolC provides protection against melioidosis.
The work described here identified a novel protein antigen that offers protection against infection with heterologous strains of B. pseudomallei. To our knowledge, this is the first example of a soluble recombinant B. pseudomallei protein domain, the periplasmic domain of LolC, which is protective against this organism. The LolC protein is able to stimulate strong antibody and cell-mediated immune responses when it is combined with either the MPL+TDM adjuvant or an ISCOM-CpG ODN complex adjuvant. Since it is considered possible that B. pseudomallei could be used as a biological weapon (29), future work to assess this antigen will include an evaluation of the protection afforded against an aerosol challenge with B. pseudomallei, as well as the cross-protective efficacy provided against B. mallei.
We gratefully acknowledge the provision of Burkholderia strains by T. Pitt, Central Public Health Laboratory, Colindale, London, United Kingdom. We also acknowledge Bonnie Wallace at Birkbeck College, London, United Kingdom, for useful discussions concerning the CD data and Peter Sharratt, PNAC Facility, University of Cambridge, for quantitative amino acid analysis.
Editor: D. L. Burns
Published ahead of print on 21 May 2007.