Escherichia coli SCS110 cells (Stratagene, USA) were used for genetic manipulations and cloning. The Salmonella enterica serovar Typhimurium ΔaroC mutant vaccine strain (TML-MD58) was supplied by Microscience Pty Ltd (United Kingdom). This mutant has a deletion in the aroC gene, which encodes chorismate synthase, an enzyme that is crucial for the biosynthesis of tryptophan, tyrosine, phenylalanine, para-aminobenzoic acid and 2,3-dihydroxybenzoate [12]. 2YT media (supplemented, where necessary, with ampicillin and the aromatic amino acids) was used for culture of the recombinant Salmonella.

To construct the Salmonella Gag-expression plasmid, a codon-optimized HIV-1 gag gene, synthesized for us by Geneart (USA), was cloned by standard recombinant DNA protocols [13] into pGEM+GFP, a plasmid designed to express green fluorescent protein that we previously constructed [14]. The gfp gene in pGEM+GFP was replaced with the gag gene and the plasmid designated pGEM+Gag was generated. The expression of Gag was under the control of the E. coli lac (lactose) promoter. Competent ΔaroC Salmonella enterica serovar Typhimurium mutant was transformed with the expression plasmid (pGEM+Gag) and this resulted in the generation of a recombinant Salmonella vaccine clone, designated, aroC+Gag. The parent cloning vector, pGEM^®-T Easy (Promega), was used as a negative control plasmid to generate a vaccine designated aroC+pGEM. The expression of HIV-1 Gag in the recombinant Salmonella bacteria, aroC+Gag, was assessed by SDS-PAGE according to standard protocols. The expression of the Gag protein was further confirmed using the Roche Elecsys^® HIV p24 Ag assay (Roche) according to manufacturer's recommendations.

Stocks of recombinant Salmonella bacterial vaccines were prepared from culture colonies of AroC+Gag (test vaccine) or aroC+pGEM (negative control vaccine). A single colony was inoculated into 200 ml of 2YT liquid media supplemented with ampicillin (100 μg/ml) and aromatic amino acids and grown at 37°C with vigorous aeration. The bacterial cells were harvested when they reached logarithmic phase (OD₆₀₀ of 0.8) by centrifugation and washed once with PBS (pH 7.4). The vaccine was suspended in PBS with 15% glycerol and stored at -80°C prior to use for immunization.

Animal work was approved by the University of Cape Town Animal Ethics Committee. Female H-2^d BALB/c mice (8–10 weeks old) were purchased from South African Vaccine Producers (Johannesburg, South Africa) and were housed (5 per group) at the University of Cape Town Animal Unit. The mice were allowed to adapt to the new environment for at least 10 days before immunization. They were bled on day 0 prior to being inoculated by intragastric gavage with 1 × 10e8 colony forming units (CFUs)/mouse of either the Salmonella test vaccine (aroC+Gag) or the negative control (aroC+pGEM) and again on days 28 and 56 prior to the booster Salmonella vaccine being given. Blood was also taken on day 84 just prior to sacrifice of the mice for harvest of spleens. Serum was collected and used in ELISA assays to measure Gag-specific antibody responses.

Spleens were harvested from the mice at sacrifice on day 84. A pool of 5 spleens per group was made, then passed through a 70-μm cell strainer to obtain a single cell suspension of splenocytes. Splenocytes were suspended in R10 medium (RPMI-1640 with 10% heat inactivated fetal calf serum, 15 mM β-mercaptoethanol, 100 U penicillin and 100 μg streptomycin per ml). Red cells were lysed using erythrocyte lysing solution (0.15 M NH₄Cl, 10 mM KHCO₃, 0.1 mM Na₂EDTA). The splenocytes were suspended at a concentration of 5 × 10e6 cells per ml for use in enzyme-linked immunospot (ELISPOT) assays or at a concentration of 15 × 10e6 cells per ml for stimulation with Gag peptides to quantify Gag-specific cytokines released into the supernatant.

Gag-specific CD4+ T cells secreting IFN-γ or IL-4 in the spleen were enumerated using IFN-γ and IL-4 ELISPOT assay kits (BD Pharmingen) as previously described [15]. Briefly, splenocytes were cultured in triplicate wells at 0.5 × 10e6 cells/well in a final volume of 200 μl with either R10 medium only (background responses) or R10 medium containing the Gag CD4 peptide (NPPIPVGDIYKRWIILGLNK, an MHC class II-restricted, CD4 binding peptide) at 4 μg/ml. After incubation (37°C, 5% CO₂) for 24 hours (IFN-γ ELISPOT assay) or 48 hours (IL-4 ELISPOT assay), the cells secreting IFN-γ or IL-4 were detected using Nova Red substrate (Vector Labs) according to the manufacturer's instructions. Spots were counted using CTL Analyzer and ImmunoSpot Version 3.2 software (Cellular Technology, USA). The mean number of spots ± SD in triplicate wells was calculated and expressed as spot-forming units (SFU) per 10e6 splenocytes. Differences in immune responses between vaccine groups were analyzed by a two-sample student's t-test. A p < 0.05 was considered statistically significant. The differences in response between stimulated cells and unstimulated cells within the same vaccine group were also analyzed by a student's t-test and p < 0.05 was considered statistically significant.

Splenocytes (1.5 × 10e6) were cultured (37°C, 5% CO₂) in a final volume of 200 μl R10 medium only (background responses) or R10 medium containing the Gag CD4 peptide (as used in the ELISPOT assays) at 4 μg/ml [15]. Culture supernatants were harvested at 48 hours and the content of Gag-specific Th1 cytokines (IFN-γ and TNF-α) and Th2 cytokines (IL-4 and IL-5) were quantified using a mouse Th1/Th2 Cytokine Bead Array (CBA) assay (BD Pharmingen) according to manufacturer's instructions. Results were expressed as pg cytokine per 10e6 splenocytes. Differences in immune responses between vaccine groups were analyzed by a two-sample student's t-test and p < 0.05 was considered statistically significant. The differences in response between stimulated cells and unstimulated cells within the same group were also analyzed by a student's t-test and p < 0.05 was considered statistically significant.

Antibodies (total IgG) to HIV Gag were detected in mouse serum (pooled from 5 mice per group) collected on day 0, 28, 56 and 84 and Gag-specific IgG1 and IgG2a detected at day 84 using HIV-1 Gag (P55) enzyme linked immunosorbent assay (ELISA). Briefly, flat-bottom, 96-well MaxiSorp ELISA plates (AEC-Amersham) were coated with 50 μg of HIV-1 P55 (Quality Biological Inc. USA) in 50 μl 0.1 M bicarbonate buffer (pH 9.5) overnight at 4°C. Plates were washed with PBS-Tween (PBS containing 0.3% Tween 20, Merck) and blocked with blocking solution (PBS containing 0.3% Tween 20, 1% goat serum, 3% milk powder) overnight at 4°C. Mouse serum samples were diluted (1/100), and 100 μl added to duplicate wells and the reaction incubated overnight at 4°C. After washing with PBS-Tween containing 1% goat serum and 1% milk powder, total Gag (P55)-specific IgG, was detected by the addition of 100 μl/well of biotinylated goat anti-mouse IgG (Southern Biotechnology, Birmingham, AL). After incubation at 37°C for 2 hours, plates were washed with PBS-Tween and streptavidin-alkaline phosphatase (50 μl/well) was added and the plates incubated a further 1 hour at 37°C. To detect HIV-1 Gag (P55)-specific IgG1 and IgG2a isotypes, alkaline phosphatase-conjugated antibodies specific for IgG1 or IgG2a) (Serotec) were used. Alkaline phosphatase was reacted with 50 μl/well para-nitrophenyl phosphate (PNP) diluted to a final concentration of 1 mg/ml. The optical density (OD) was read at 405 nm and means of duplicates was calculated. For total IgG, the ratios of vaccinated to prebleed were calculated and represented graphically. The OD readings were represented graphically for IgG1 and IgG2a. Differences in immune responses between vaccine groups were analyzed by a two-sample student's t-test and a p value < 0.05 was considered statistically significant.

A prokaryotic expression plasmid that contained a wild-type HIV-1 Subtype C gag gene was initially constructed. As Gag expression was poor, probably because of the presence of bacterial rare codons (data not shown), we redesigned the recombinant vaccine. For this the wild-type HIV-1 gag gene was replaced with a synthesised gene that reflected codon usage in Salmonella and the expression plasmid, pGEM+Gag, containing the codon-optimized gag was successfully constructed (Figure 1A). The plasmid contained the HIV-1 gag gene fused to the β-galactosidase α-fragment and the expression was constitutively driven by the E. coli lac promoter and other lac operon transcription and translation domains found in pGEM-Teasy plasmid. The recombinant plasmid-carrying Salmonella expressed very high levels of HIV Gag (Figure 1B). The Gag protein band was visible in Coomassie blue-stained SDS-PAGE gels (Figure 1B, Lane 2). The high-level of Gag protein expressed by the Salmonella was further confirmed by Roche Elecsys^® HIV p24 Ag Capture assay (Figure 1C). The expression of codon-optimized Gag was higher than previously found with wild-type Gag (data not shown).

CD4+ Th1/Th2 T cells induced in the spleens of mice in response to oral vaccination with the recombinant codon-optimized HIV-1 Gag-expressing Salmonella vaccine vector, aroC+Gag, were evaluated using ELISPOT assays. Mice vaccinated with aroC+Gag developed HIV-1 Gag-specific CD4+ Th1 (IFN-γ) and Th2 (IL-4) cells in the spleen. The frequency of Gag-specific IFN-γ spot-forming units from aroC+Gag was above the background (p < 0.05) (Figure 2A). Cells from the aroC+Gag group responded to Gag CD4 peptide stimulation. The number of IL-4-producing cells from aroC+Gag was also significantly higher than the number of cells producing IL-4 from the negative vaccine control (p < 0.05) (Figure 2B). It was, therefore, evident that both Gag-specific IFN-γ and IL-4 cytokines were elicited by the recombinant Salmonella vaccine vector.

A CBA assay was used to quantify Gag-specific IFN-γ, TNF-α, IL-4 and IL-5 cytokines secreted by the splenocytes upon antigenic stimulation. The amounts of CD4+ Th1 (INF-γ and TNF-α) and Th2 (IL-4 and IL-5) cytokines secreted by the splenocytes into the supernatant were quantified. The levels of Gag-specific Th1 and Th2 cytokines were all significantly elevated above the background (p < 0.05) (Figure 3) in mice vaccinated with aroC+Gag. The levels of Th1 cytokines, IFN-γ and TNF-α were 7.5-fold and 29.1-fold above the background, respectively (Figure 3A and 3B). The Th2 cytokine levels, IL-4 and IL-5 were 26.2- and 89.3-fold above the background, respectively (Figure 3C and 3D). The CBA results, therefore, confirmed that vaccination of mice orally with the Salmonella vaccine vector induced systemic Gag-specific Th1 and Th2 cytokine responses.

HIV-1 Gag-specific humoral immune responses in mice vaccinated with the HIV Gag-expressing Salmonella vaccine were evaluated on days 28, 56 and 84. Anti-Gag total IgG and IgG subtypes IgG2a and IgG1 in the serum of vaccinated mice were determined. A very low serum HIV-1 Gag specific IgG response (1.52-fold OD₄₀₅ reading above prebleed) was detected on day 28 in serum (1/100 dilution) in mice vaccinated with aroC+Gag and this was not significantly above the control (p > 0.05) (Figure 4A). On day 56, the antibody response was boosted significantly (5-fold OD₄₀₅ reading above prebleed) (p < 0.05). The Gag-specific antibody response was further boosted by day 84 (22-fold OD₄₀₅ reading above prebleed) (p < 0.05) (Figure 4A). HIV-1 Gag-specific IgG responses were confirmed using the New LAV Blot I HIV-1 Western blotting kit (Biorad). The serum from the aroC+Gag-vaccinated mice reacted specifically with Gag bands (P55, P40, P24/25 and P17/18) on the blot (results not shown). Gag-specific serum IgG subtypes measured on day 84 indicated the presence of significant (p < 0.01) levels of IgG1 and IgG2a (p < 0.05) (Figure 4B). The Gag-specific Ig2a response was slightly greater than the IgG1 response. In summary, both Th2-related IgG1 and Th1-related IgG2a antibody responses were induced in the mice vaccinated with the HIV Gag-expressing Salmonella vaccine.