A prokaryotic expression cassette was developed in which the gfp gene was fused in-frame with an E coli β-galactosidase α-fragment sequence (N-terminus) (Figure 1). The gfp gene was amplified and cloned into pGEM-Teasy plasmid vector. The β-galactosidase α-fragment with DNA sequence (5'-ATG ACC ATG ATT ACG CCA AGC TAT TTA GGT GAC ACT ATA GAA TAC TCA AGC TAT GCA TCC AAC GCG TTG GGA GCT CTC CCA TAT GGT CGA CCT GCA GGC GGC CGC GAA TTC ACT AGT GAT-3') had 24 amino acids (MTMITPSYLG DTIEYSSYAS NALGALPYGR PAGGREFTSD) and the peptide was 4.2 kDa in size. A small linker (L) sequence with 15 codons (5-TAT GGC GCC AAA GAC TCC GGC TCC GCC GGT TCC GCC GGC TCA GCT-3) was incorporated between the β-galactosidase α-fragment and gfp. The linker peptide had 15 amino acids (YGAKDSGSAG SAGSA) and a molecular weight of 1.266 kDa. The gfp gene had 237 amino acids (SKGEELFTGV VPILVELDGD VNGHKFSVSG EGEGDATYGK LTLKFICTTG KLPVPWPTLV TTFSYGVQCF SRYPDHMKRH DFFKSAMPEG YVQERTISFK DDGNYKTRAE VKFEGDTLVN RIELKGIDFK EDGNILGHKL EYNYNSHNVY ITADKQKNGI KANFKIRHNI EDGSVQLADH YQQNTPIGDG PVLLPDNHYL STQSALSKDP NEKRDHMVLL EFVTAAGITH GMDELYK) and a molecular weight of 26.6 kDa. The GFP contains a Balb/C mouse CD8+ T cell epitope, HYLSTQSAL. The whole β-galactosidase-GFP fusion protein was 32.1 kDa. A preferred translation stop codon (

TAA

TAA

Very high level constitutive expression of GFP antigen by the recombinant Salmonella enterica serovar Typhimurium, AroC+GFP, was demonstrated (Figure 2). Colonies and cultures of the bacterial vaccine, AroC+GFP, fluoresced brightly green under UV light. SDS-PAGE analysis showed that GFP antigen was the most highly expressed antigen by the Salmonella vaccine vector (Figure 2A). The GFP protein band was visible on the Coomassie-stained gel. Western blotting further confirmed that GFP antigen was expressed at very high levels by the vaccine vector (Figure 2B). There was no expression of GFP by the negative control vaccine, AroC+pGEM.

The induction of GFP-specific CD8+ T cells in the spleen was evaluated on Day 84 after oral vaccination of mice with a dose of 10e8 colony-forming units either with AroC+GFP or the control (AroC+GEM) on Days 0, 28 and 56. Both IFN-γ and IL-4 producing GFP-specific CD8+ T cells were evaluated after sacrifice of mice. A high magnitude of GFP-specific CD8+ T cells was detected when the GFP peptide was included in the IFN-γ and IL-4 ELISPOT assays (Tables 1 and 2). The number of cells secreting IFN-γ after stimulation with a GFP CD8 peptide were significantly higher in AroC+GFP than in the negative control vaccine, AroC+pGEM (p < 0.05) (Table 1). There was no significant difference in response between the two groups when the cells were stimulated with media or full-length GFP (p > 0.05). Response to the LPS stimulation differed between the two groups (p < 0.05). Analysis of the responses within the AroC+GFP group showed that the number of cells producing IFN-γ were significantly higher when stimulated with GFP CD8 peptide than when not stimulated (p < 0.05) (Table 1). In the negative vaccine group, AroC+pGEM, there was no difference in response in GFP peptide-stimulated cells and unstimulated cells (p > 0.05) (Table 1).

The number of cells from AroC+GFP vaccine group producing IL-4 were also significantly higher than in AroC+pGEM after stimulation with GFP CD8 peptide (p < 0.05) (Table 2). However no significant difference in response between the two groups were observed when the cells were stimulated with media, full-length GFP or Salmonella LPS (p > 0.05). Within the AroC+GFP group, the number of cells producing IL-4 were significantly higher when stimulated with GFP CD8 peptide than when unstimulated (p < 0.05) (Table 2). No difference in IL-4 responses was observed within the AroC+pGEM group between GFP peptide-stimulated and unstimulated cells (p > 0.05) (Table 2).

The cytometric bead array (CBA) assay and flow cytometry analysis were used to quantify the IFN-γ and IL-4 simultaneously produced by splenocytes after stimulation with the GFP H-2K^d binding peptide (HYLSTQSAL). Cells from AroC+GFP produced higher levels of IFN-γ when stimulated with the GFP CD8 peptide than when unstimulated (p < 0.05) (Figure 3). The IFN-γ cytokine levels were also higher in the test group (AroC+GFP) than in the negative vaccine group (AroC+pGEM) (p < 0.05) (Figure 3). There was no difference in the amount of IFN-γ produced between stimulated and unstimulated cells within the negative control vaccine (p > 0.05) (Figure 3). As with IFN-γ, the same trends were observed with IL-4 (Figure 4). Cells from AroC+GFP produced higher levels of IL-4 when stimulated with GFP peptide than when unstimulated (p < 0.05). The level of the IL-4 produced by the stimulated cells was 10.4-fold above background. GFP peptide-stimulated cells from AroC+GFP also produced significantly higher levels of IL-4 than cells from the negative control vaccine, AroC+pGEM (p < 0.05). The results from CBA further confirmed the ELISPOT results which showed that a significant number of cells produced both IFN-γ and IL-4 after stimulation with the GFP CD8 peptide. Further analysis of the both ELISPOT and CBA assay results showed that GFP-specific IFN-γ was produced at a rate of 2.56 pg/cell as opposed to GFP-specific IL-4 which was produced at 1.63 pg/cell.

Competent Escherichia coli SCS110 cells (Stratagene, USA) were used in cloning and genetic manipulations. An auxotroph, ΔaroC Salmonella enterica serovar Typhimurium mutant vaccine strain (TML-MD58) (Microscience Pty Ltd, UK) was used as an attenuated vaccine for the expression of GFP. The strain has a deletion in the aroC gene, which encodes chorismate synthase, an enzyme necessary for the biosynthesis of aromatic compounds, tryptophan, tyrosine, phenylalanine, para-aminobenzoic acid and 2,3-dihydroxybenzoate 37. The bacteria were grown in 2YT media supplemented, where necessary, with ampicillin and aromatic amino acids (tryptophan, tyrosine, phenylalanine, para-aminobenzoic acid and 2,3-dihydroxybenzoate) as previously described 3738.

Unless stated otherwise, DNA manipulations were performed using standard recombinant DNA methods 38. A recombinant plasmid, designated pGEM+GFP, was constructed. The gfp gene was amplified using GFP2 (forward), 5'-ATG GCG CCA AAG ACT CCG GCT CCG-3' and GR (reverse), 5'- AAG CTT ATT TGT ATA GTT CAT CCA TGC-3') synthetic oligonucleotides as primers. The primers were rationally designed so that GR could have a preferred gram-negative bacterial stop codon, 5'-

TAA

TAA

TAA

The polymerase chain reaction for amplification of gfp was conducted in a 50 μl volume with 4.5 units AmpliTaq Gold™ DNA polymerase (Applied Biosystems), 1× PCR buffer, 1.5 μM of each primers (GFP2 and GR), 0.2 mM deoxynucleotide triphosphates, 1.5 mM magnesium chloride and 10 ng of PEHAOGFP plasmid (provided by Dr W Bourn, University of Cape Town). The PCR cycling conditions were as follows: 1 cycle of 95°C for 5 min, 5 cycles of 95°C for 45 s, 55°C for 30 s, 72°C for 2 min, 25 cycles of 95°C for 45 s, 64°C for 30 s, 72°C for 2 min, and a final extension of 72°C for 7 min. Analysis of the gfp amplicon aliquot (5 μl) was done by agarose gel electrophoresis. An aliquot (1 μl) of remaining amplicon was ligated into a linearized pGEM-Teasy (Promega, USA) according to manufacturer's recommendations. The ligation reaction was used in the genetic transformation of competent E. coli SCS110 cells using the heat-shock method. The recombinant SCS110 clones harbouring the recombinant plasmid (pGEM+GFP) were screened for presence of gfp fragment and its orientation by blue-white screening procedure and UV-illumination. The white and fluorescing (candidate) clones were cultured using standard protocols. To investigate the presence of the recombinant gfp gene in plasmids, restriction mapping was performed initially with EcoR1 followed by double digestion with NarI and HindIII. The gfp gene in the candidate pGEM+GFP plasmid was sequenced.

To investigate the expression of recombinant GFP, pGEM+GFP and pGEM (negative control) plasmids were used in the genetic transformation of competent aroC Salmonella enterica serovar Typhimurium mutant by a standard heat-shock method 33. The agar plates were incubated overnight and fluorescence of colonies viewed under UV light the following morning. Single colonies were cultured in 100 ml 2 YT liquid broth with ampicillin (100 μg/ml). To determine the expression of GFP by the recombinant Salmonella, total bacterial protein was extracted from each culture, separated by a standard 12% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and visualized in the gel by Coomassie blue staining. A standard Western blotting was performed to identify and to confirm the specificity and integrity of the GFP antigen band seen on the SDS-PAGE after Coomassie blue staining. A mixture of two anti-GFP mouse monoclonal antibodies (Clones 7.1 and 13.1) (Roche Diagnostics) was used as a primary antibody (diluted at 1.1000). Goat-anti-mouse immunoglobulins conjugated to horseradish peroxidase (Biorad), diluted at 1.1000 were used as secondary antibody. The immunoblot was visualized by enhanced chemiluminescence (Roche Diagnostics) and autoradiography according to manufacturer's recommendations

To prepare vaccine stocks, a single colony of recombinant Salmonella was inoculated into 200 ml of 2 YT liquid media supplemented with ampicillin (100 ug/ml), and aromatic amino acids (1×) and grown at 37°C with strong aeration. The bacterial cells were harvested in the logarithmic phase (OD₆₀₀ = 0.8–1.0) by centrifugation at 3000 rpm for 5 mins, washed once with equal volume of phosphate buffered saline (PBS, pH 7.4) and suspended in PBS with 15% glycerol. The cultures were stored in aliquots at -80°C until vaccination. The bacterial count in the vaccine stocks was determined by plating of serial dilutions. The vaccines were designated AroC+GFP (a recombinant aroC Salmonella enterica serovar Typhimurium mutant expressing GFP) and AroC+pGEM (a recombinant control aroC Salmonella enterica serovar Typhimurium harbouring an empty plasmid, pGEM-Teasy).

All animal procedures were approved by the University of Cape Town Animal Ethics Committee. Female H-2^d BALB/c mice (8–10 weeks old; and five per group) were purchased from South Africa Vaccine Producers Pty Ltd (Johannesburg, South Africa), housed at the University of Cape Town Animal Unit and allowed to adapt for a minimum of 10 days before vaccinations. Groups of female BALB/c mice were inoculated by intragastric gavage with 10e8 colony forming units (CFUs)/mouse of either Salmonella vaccine (AroC+GFP) or negative control (AroC+pGEM) on Days 0, 28 and 56. Mice were sacrifice on Day 84 and spleens were pooled for each group. The spleens were meshed using a rubber stopper and metal grid (Sigma) placed in a petri dish to generate a single cell suspension in RPMI 1640 medium (Invitrogen, USA). The cell suspension was transferred to a 50 ml conical centrifuge tube. The volume was made up to 50 ml with RPMI 1640 medium. The cell suspension was centrifuged at 1500 rpm for 5 minutes to pellet the cells. The pellet was re-suspended in 50 ml of RPMI 1640 medium and centrifuged as before. The pellet was then washed twice with 50 ml of RPMI 1640 medium. The cells were re-suspended in 50 ml R10 medium (RPMI 1640, 10% heat-inactivated fetal calf serum, a mixture of pernicillin and streptomycin (Invitrogen, USA), and 15 mM 2-mercaptoethanol (Sigma, USA)). A single cell suspension of splenocytes was prepared and red cells were lysed using erythrocyte lysing buffer (0.15 M NH₄Cl, 10 mM KHCO₃, 0.1 mM Na₂EDTA) for 1 min at room temperature. To count the cells and determine viability, 1/10 dilution of the suspension was made in Trypan Blue and Neubauer counting chamber used. Cell concentration in suspension was calculated and adjusted to an appropriate concentration. For use in ELISPOT assay, the splenocytes were adjusted to a concentration of 5 × 10e6 cells per ml and 100 ul of this stock was added to a single well which contained 100 ul of the stimulant. For use in CBA assay, the splenocytes were adjusted to a concentration of 15 × 10e6 cells per ml and 100 ul of this stock was added to a single well which contained 100 ul of the stimulant.

The IFN-γ and IL-4 ELISPOT kits (BD Pharmingen) were used according to manufacturer's recommendations. Splenocytes were plated in triplicate at 0.5 × 10e6 cells/well in a final volume of 200 μl of R10 medium (RPMI-1640 with 10% heat-inactivated fetal calf serum, 15 mM β-mercaptoethanol, 100 U penicillin per ml, and 100 μg streptomycin) either alone or with stimulants at 4 μg/ml. The stimulants used assays were media (no peptide), GFP H-2K^d binding peptide (HYLSTQSAL), full-length GFP, Salmonella lipopolysaccharide (LPS) (at a final concentration of 0.5 μg/ml). After incubation for 24 hrs (IFN-γ ELISPOT assay) or 48 hr (IL-4 ELISPOT assay), the plates were processed to detect IFN-γ- or IL-4-spot-forming units (SFUs) using Nova Red substrate (Vector Laboratories, UK) according to the kit instructions. Spots were counted using a CTL Analyzer (Cellular Technology, OH, USA) and ImmunoSpot Version 3.2 software (Cellular Technology OH, USA). The mean number of spots ± SD in triplicate wells was calculated and expressed as SFUs/10e6 splenocytes. Differences in immune responses between vaccine groups were analyzed by the two-sample t-test.

Splenocytes at a concentration of 1.5 × 10e6 per 200 ul R10 culture medium (RPMI-1640 with 10% heat inactivated fetal calf serum, 100 U penicillin per ml, and 100 μg streptomycin) were cultured alone or with the individual stimulants as in the ELISPOT assay. CD8+ Tc1 (IFN-γ) and Tc2 (IL-4) cytokines secreted by the splenocytes were quantified using a mouse Th1/Th2 cytokine cytometric bead array (CBA) assay (BD Biosciences kit) and flow cytometry analysis according to manufacturer's instructions. Results were expressed as pg cytokine per 1 × 10e6 splenocytes. Differences in immune responses between vaccine groups were analyzed by the two-sample t-test.