INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacillus anthracis, a spore-forming bacterium, is the etiologic agent of anthrax. The bacterium secretes a lethal toxin composed of two proteins, the protective antigen (PA, 83 kDa) and the lethal factor (LF, 85 kDa) (4, 16). The pag and lef genes, carried by the virulence plasmid pXO1 (185 kbp), encode PA and LF, respectively (5, 20, 31). PA is the binding moiety of the toxin, and LF is the intracellular enzyme that damages the cells. A mode of action of the lethal toxin has been proposed from in vitro and in vivo experiments (6, 17). PA binds to a ubiquitous receptor on the surface of mammalian cells and is cleaved by furin-like proteases (9, 14, 26). This processing results in the release of a 20-kDa amino-terminal fragment, and the cell-associated 63-kDa fragment interacts with LF. The PA63-LF complex is then internalized by receptor-mediated endocytosis, and at acidic pHs, LF is translocated into the cytosol. LF displays zinc metalloprotease activity specific for mitogen-activated protein kinase kinases 1 and 2 (8, 15, 30). The amino-terminal part of LF (LF254) binds to PA63 (25). Fusion proteins consisting of LF254 and heterologous antigens have been shown to be successfully delivered to cells via PA (1, 2, 3).

The B. anthracis Sterne strain, which is attenuated, is currently used as the live veterinary vaccine against anthrax. The immunity induced by this live vaccine is associated with stimulation of a humoral response to the toxin components. Studies performed with toxin-deficient B. anthracis strains (24) have shown that the antibody response specific for LF is dependent on the presence of PA. The level of LF-specific antibodies is much higher in animals immunized with bacteria producing both PA and LF than in animals receiving bacteria producing LF only. The molecular mechanisms underlying this adjuvant effect of PA have been further investigated. Strains carrying site-directed mutations in the functional domains of toxin genes have been constructed (6). Analysis of their immunogenic properties in mice clearly indicated that potentiation of the LF-specific antibody response requires only the binding of LF to PA. Neither the biological activity of LF nor binding of the PA63-LF complex to the cell receptor is involved in this phenomenon.

The successful in vivo delivery of heterologous antigens by B. anthracis has been reported. Strains producing listeriolysin O, a hemolysin from Listeria monocytogenes, or the Ib component of iota toxin from Clostridium perfringens, which is under the control of the pag promoter, elicit specific immune responses and protection in mouse models (27, 28).

We investigated the ability of PA to enhance the humoral response against a heterologous antigen fused to LF254 when delivered in vivo by B. anthracis. We used the C fragment of tetanus toxin (ToxC). Tetanus toxin (150 kDa) is composed of a light chain (50 kDa) carrying the catalytic site and a heavy chain (100 kDa) containing domains involved in cell binding and translocation of the toxin into the cytoplasm (13, 21). ToxC (50 kDa) is part of the heavy chain and corresponds to the cell-binding part of tetanus toxin (12, 22). ToxC has been produced in various live vectors including Salmonella enterica serovar Typhi, Vibrio cholerae, and Lactoccocus lactis (7, 11, 32). Analysis of the antibody response against ToxC secreted by the recombinant bacteria suggested that this protein is only weakly immunogenic. Recently, ToxC has been successfully anchored to the surface of B. anthracis vegetative cells (19). However, no antibody response against ToxC was observed after a single injection of bacilli, even in the presence of adjuvant, and several injections were required for both antibody response and protection. Protection against tetanus is known to be antibody mediated (10). This fact and the weak immunogenicity of ToxC in heterologous systems make this antigen particularly suitable as a model for analysis of the ability of PA to exert its adjuvant effect on foreign antigens. In this study, recombinant B. anthracis strains that produced LF254-ToxC protein in a PA-producing or PA-deficient background were constructed. The capacity of the recombinant strains to stimulate a humoral response and protective immunity against tetanus toxin were analyzed.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial strains and media. Escherichia coli and B. anthracis strains were cultured in Luria-Bertani and brain heart infusion (BHI) (Difco, Detroit, Mich.) media, respectively. Ampicillin (100 µg/ml), spectinomycin (60 µg/ml), and erythromycin (5 µg/ml and 180 µg/ml for B. anthracis and E. coli, respectively) were added as appropriate. The B. anthracis Sterne (7702) strain and the RPL strains have been described previously (6).

Construction of the hybrid gene lef-toxC. A MluI restriction site was created after the codon encoding leucine 254 in the lef gene in the recombinant plasmid pUC1835, using the Quickchange site-directed mutagenesis kit (Stratagene, La Jolla, Calif.), giving rise to pLE254 (1, 5, 23). The DNA fragment encoding ToxC was amplified by PCR using the primers toxc1 (5'-CGACGCGTTCAACACCAATTCCATTTTCTTATTC-3'), containing an MluI restriction site (underlined), and toxc2 (5'-CATGCCATGGTCATGAACATATCAATCTGTTTAATC-3'), containing an NcoI restriction site. Total plasmid DNA from Clostridium tetani CN655 (kindly provided by M. R. Popoff) was used as the template. pLE254 and the 1,420-bp PCR fragment were cleaved with MluI and NcoI and ligated, and the resulting plasmid, pTET2541, was used to transform E. coli. Finally, the SacI-SphI fragment of pTET2541 was ligated into the gram-positive bacterium suicide vector pAT113 (29), giving rise to pBF254.

Mating procedure. The B. anthracis mutant strains were constructed by heterogramic mating, as previously described (23, 29). The wild-type lef gene was replaced with the fusion gene in two steps as previously described (6). First, pBF254, which is resistant to erythromycin, was introduced into the lef-inactivated strain, RPL, which carries a spectinomycin resistance cassette in the lef gene (6). We selected clones in which pBF254 had been integrated into pXO1 by a single crossover event on agar containing both erythromycin and spectinomycin. The resulting merodiploids were grown in 5 ml of fresh BHI medium inoculated with 5 µl of a culture with an optical density at 600 nm of 0.8. After seven subcultures, 20% of the bacteria were sensitive to the marker antibiotics and harbored the hybrid gene under the control of the lef promoter (Fig. 1A). One of the resulting clones, RPL254, was used to construct its PA-deficient counterpart, RPA254. The suicide plasmid pBAFH114, harboring a deleted pag gene inactivated by a spectinomycin resistance cassette, was used for the mating (27). The RPA254 strain was obtained by allelic exchange between the wild-type pag gene on pXO1 from RPL254 and the deletion-containing pag gene of pBAFH114. The two recombinant strains were checked by PCR, and the junction between lef and toxC was sequenced (Sequenase; Amersham, Cleveland, Ohio).

View larger version (43K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the hybrid gene and Western blot analysis of the recombinant strains. (A) Schematic representation of the hybrid gene, lef-toxC, at the lef locus on pXO1. plef corresponds to the promoter of the lef gene. The MluI restriction site adds two codons encoding Thr and Arg at the junction of the fusion protein. The open reading frame is indicated by arrows. (B) In vitro production of LF254-ToxC by recombinant B. anthracis strains. The recombinant strains, RPL254 and RPA254, were grown in BHI medium in the presence of bicarbonate to stimulate toxin production. Supernatants from 500 µl of culture were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting with anti-LF and anti-PA monoclonal antibodies and with rabbit sera specific for ToxC.

Detection of B. anthracis toxin components. Recombinant B. anthracis was grown in BHI medium supplemented with 0.4% (wt/vol) bicarbonate at 37°C in tightly closed flasks to an optical density at 600 nm of 0.7 (27). The cultures were centrifuged (13,000 × g, 5 min), and proteins were precipitated from the supernatants with 10% trichloroacetic acid and subjected to electrophoresis in a 12% polyacrylamide gel. The proteins were either stained with Coomassie blue or transferred to a nitrocellulose membrane (Hybond-C; Amersham). For immunoblot analysis, monoclonal antibodies specific for PA and LF and polyclonal serum specific for ToxC were used at a dilution of 1/10,000. A secondary antibody conjugated to peroxidase was added at a dilution of 1/20,000. Proteins were detected by adding luminol (ECL kit; Amersham).

In vivo experiments. Seven-week-old female OF/1 outbred mice (Iffa Credo, l'Arbresle, France) were immunized by subcutaneous injection with 3 × 10⁸ to 5 × 10⁸ spores per animal (in a volume of 500 µl). After 35 days, serum samples were obtained from the retro-orbital plexus as previously described (24). If necessary, a booster of 3 × 10⁸ spores per mouse was given at 35 days and sera were obtained on day 50. The titer of antibody specific for the tetanus toxin was measured by enzyme-linked immunosorbent assay (ELISA) using tetanus anatoxin (a formaldehyde-inactivated protein) at 25 µg/ml to coat the microplates. ELISA for PA-, LF- and extractable antigen 1 (EA1)-specific-antibody titers were performed as previously described (24). An arbitrary A₄₉₂ value of 0.5 was used to calculate the endpoint titers. However, for PA, LF, EA1, and ToxC, the A₄₉₂ value obtained with sera of untreated animals diluted 1/100 never exceeded 0.02. For protection experiments, mice were challenged intramuscularly, 35 days after the first or 15 days after the second injection, with 4 ng of purified tetanus toxin per animal (30 50% lethal doses [LD₅₀], in a volume of 250 µl). Mortality and morbidity (paralysis) were monitored for 10 days after challenge.

Neutralization assays. Sera from animals immunized with the RPL254 and RPA254 strains were pooled separately and heated for 30 min at 56°C. For each mouse, 1 ng of tetanus toxin (eight LD₅₀) was incubated for 2 h at room temperature with diluted serum (final dilutions from 1:10 to 1:40). The mixture was then injected intramuscularly into naive mice (four animals per dilution) (28). Neutralizing activity was scored as the highest dilution at which all animals survived. Surviving animals were killed 10 days after challenge.

Statistics. Statistical analyses were performed using Student's unpaired t test for antibody titers and ² for protection assessment.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Characterization of recombinant B. anthracis strains harboring the chimeric gene, lef-toxC. The hybrid gene, lef-toxC, was introduced by homologous recombination into the lef gene on pXO1 of the Sterne strain, as described in Materials and Methods (Fig. 1A). The resulting recombinant strain, RPL254, was used as a recipient for the construction of its PA-deficient counterpart. The wild-type pag gene was replaced by allelic exchange with an inactivated copy carrying a spectinomycin resistance cassette (6) to produce the RPA254 strain. The hybrid gene, lef-toxC, is under the control of the in vivo inducible lef gene promoter. The recombinant B. anthracis strains were then grown under conditions known to induce the synthesis of PA and LF in vitro (27). The secreted proteins were analyzed by immunoblotting using antibodies specific for PA, LF, and ToxC (Fig. 1B). As expected, PA was found exclusively in the supernatants of the parental Sterne strain and of RPL254. LF-specific antibodies detected an 85-kDa band in the supernatant of the Sterne strain, corresponding to the full-length LF molecule. In contrast, the recombinant strains RPL254 and RPA254 secreted a 78-kDa protein which reacted with antibodies against LF and ToxC. The size of this protein was as expected for the hybrid LF254-ToxC molecule. The two recombinant strains also appeared to produce similar amounts of LF254-ToxC.

Antibody response to tetanus toxin induced by B. anthracis strains. Mice were injected subcutaneously with one or two doses of RPL254 or RPA254 spores, and serum-specific antibodies were analyzed by ELISA (Fig. 2). The in vivo development of the RPL254 and RPA254 strains was monitored by the antibody response raised against EA1, one of the two S-layer proteins which are exclusively found at the surface of vegetative forms of B. anthracis (18). Titers of antibody against EA1 thus reflect the germination of spores and the development of bacilli within mice. The humoral responses obtained with the two strains were similar (1/12,800) and close to those routinely obtained following immunization with B. anthracis strains (6). The in vivo development of the spores was therefore similar for the two LF254-ToxC producing strains. The antibody response (1/10,600) against PA in animals injected with RPL254 was similar to that usually found after the injection of PA-producing strains. An antibody response to the tetanus toxin was detected after the first injection for both recombinant strains (Fig. 2). However, the antibody titers elicited by the PA-producing strain were 8 times higher than those for the PA-deficient strain (1/1,700 versus 1/200; P < 0.05). Booster injections of RPL254 and RPA254 stimulated a secondary immune response, as observed by the increase in antibody titers (1/5,000 versus 1/1,000, respectively). These data clearly indicate that in vivo, LF254-ToxC is efficiently produced and presented to the immune system. Moreover, it is established that LF254 is the region involved in the binding of LF to PA63 (1, 2, 3, 25). This property allows the formation of the PA63-LF254-ToxC complex. It is likely that the higher stability of the complex in vivo accounts for the immunostimulatory effect of PA (6). The data also provide evidence that PA is able to potentiate a humoral response against a heterologous antigen secreted by B. anthracis. This effect was clearly demonstrated by the detection of an antibody response to ToxC after single-spore immunization, in contrast to results reported for other delivery systems (7, 19, 32). The adjuvant property of PA is thus of particular interest for the delivery in vivo of proteins that are only weakly immunogenic.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Serum titers of antibody against tetanus toxin in mice immunized with recombinant B. anthracis. Animals were immunized once or twice with RPL254 or RPA254 spores. Tetanus toxin-specific antibodies were determined by ELISA. Circles represent individual titers, and the black bar represents the arithmetic mean of the group. The strains used for immunization are indicated at the top.

Protective immunity against tetanus toxin induced by recombinant strains. The ability of recombinant B. anthracis to protect mice against a challenge with the tetanus toxin (30 LD₅₀) was assessed (Fig. 3). All control animals died within 2 days. A single immunization was sufficient to protect 75 to 90% of mice injected with the RPL254 strain whereas, under the same conditions, only 11 to 20% of animals injected with the RPA254 strain were protected (Fig. 3A). After the booster, the level of protection was 100 and 80% for mice immunized with the RPL254 and RPA254 strains, respectively (Fig. 3B).

View larger version (17K): [in this window] [in a new window]   FIG. 3.   Efficacy of protection against lethal tetanus toxin challenge. Mice were immunized by a single injection (A) or two injections (B) of RPA254 () or RPL254 () spores. Untreated animals were used as controls (×). Animals were challenged with 30 LD50 of tetanus toxin, and mortality was monitored over 10 days.