The subcellular location of a recombinant antigen in recombinant attenuated Salmonella vaccines may influence immunogenicity dependent on exposure of the recombinant antigen to cells involved in systemic immune responses. It has been shown that a recombinant attenuated Salmonella vaccine secreting the recombinant Streptococcus pneumoniae PspA (rPspA) antigen specified by pYA3494 induced protective anti-rPspA-specific immune responses (Kang et al. (2002) Infect. Immun. 70, 1739–1749). A recombinant plasmid pYA3496 specifying a His6-tagged rPspA (His6-rPspA) protein (no apparent signal sequence) caused the rPspA antigen to localize to the cytoplasm of Salmonella. Salmonella vaccines carrying pYA3494 or pYA3496 expressed similar amounts of rPspA. After a single oral immunization in BALB/c mice with 109 colony-forming units (CFU) of the recombinant Salmonella vaccines carrying pYA3494 or pYA3496, IgG antibody responses were stimulated to both rPspA and Salmonella lipopolysaccharide (LPS) antigens. The anti-rPspA IgG titer induced by Salmonella carrying pYA3494 (1.9×107) was 104 times higher than induced by Salmonella carrying pYA3496 (<2.4×103).
Attenuated Salmonella strains have been developed as live vaccines for humans and animals to prevent disease caused by Salmonella infections. Because live attenuated recombinant Salmonella vaccines colonize the gut-associated lymphoid tissues (Peyer's patches) and visceral lymphoid tissues following oral administration, they induce mucosal, humoral and cellular immunities [1,2]. Genetically modified attenuated Salmonella vaccines have also been developed as antigen carriers to deliver heterologous pathogen antigens specified by multicopy plasmids [1,2]. The aspartate β-semialdehyde dehydrogenase (asd) gene has served as the basis for a widely used ‘balanced lethal host–vector system’ for the stable in vivo maintenance of multicopy plasmids, devoid of antibiotic resistance genes, specifying recombinant antigens [3–5]. A series of Asd+ plasmids have been developed to increase their utility for recombinant antigen expression in Salmonella balanced lethal host–vector systems .
Streptococcus pneumoniae is a Gram-positive human pathogen that causes serious health problems, including community-acquired pneumonia, otitis media, meningitis, and bacteremia in persons of all ages . S. pneumoniae is a leading agent of childhood pneumonia worldwide, resulting in about 3 million deaths per year . Although several polysaccharide-based pneumococcal vaccines are available to reduce pneumococcal infection, immunization with pneumococcal polysaccharide vaccines is only moderately effective in reducing the frequency of hospitalization, costs, and mortality caused by pneumococcal pneumonia .
The pneumococcal surface protein A (PspA) expressed in all clinically isolated pneumococcal strains has been considered a pneumococcal subunit vaccine candidate. Immune responses induced by PspA protect mice against virulent S. pneumoniae challenge [9–12]. Among several functional domains of native PspA, the α-helical domain is highly immunogenic and contains the dominant protective epitopes . Mice orally immunized with recombinant attenuated Salmonella typhimurium vaccines expressing a recombinant PspARX1 (rPspA) (pspA gene from S. pneumoniae strain Rx1) elicited PspA-specific immune responses and protected mice against virulent S. pneumoniae challenge [5,14]. In a previous study, it was shown that fusing the α-helical region of PspA (amino acid residues 3–257 of the mature PspARx1 protein) to the β-lactamase signal sequence encoded in pYA3494 resulted in a higher amount of the rPspA being secreted into the periplasm and outside the cells  than achieved by using the PspA signal sequence . This also resulted in induction of a higher anti-rPspA IgG response in mice after oral immunization .
The question addressed in this study is to determine whether rPspA antigen placement in different Salmonella subcellular fractions (cytoplasm vs. periplasm and secretion) influences induction of anti-rPspA-specific immune responses. It has been shown that the rPspA protein specified by pYA3494 in recombinant attenuated S. typhimurium secreted into both the periplasm and culture supernatant without cell lysis . Plasmid pYA3496 carries the same rPspA sequence contained in pYA3494 except that it specifies a His6-tag rather than the β-lactamase signal sequence at the N-terminus of rPspA. With the demonstration of subcellular location of the His6-rPspA specified by pYA3496 in S. typhimurium, we report comparative rPspA immune responses induced by Salmonella vaccines carrying pYA3494 (secretory rPspA) or pYA3496 (cytoplasmic rPspA).
Materials and methods
Bacterial strains, plasmids, and growth conditions
Escherichia coliχ6212 (F−λ−Φ80 Δ(lacZYA-argF) endA1 recA1 hsdR17 deoR thi-1 glnV44 gyrA96 relA1ΔasdA4)  was used as the host strain for construction of Asd+ vectors. A recombinant attenuated S. typhimuriumχ8501 (hisGΔcrp-28ΔasdA16) constructed in a previous study  was used for the delivery of rPspA antigens. S. typhimuriumχ8599 (hisGΔasdA16 atrB13::MudJ) was used for rPspA protein localization analysis . Plasmids pYA3342, pYA3493, pYA3494 and pYA3496 were described elsewhere : pYA3342 and pYA3493, Asd+ vectors; pYA3494, specifying rPspA fused to the β-lactamase signal sequence; pYA3496, specifying His6-rPspA. E. coli and S. typhimurium cultures were grown at 37°C in Lennox broth  or Luria–Bertani (LB) broth, or on LB agar . Diaminopimelic acid (DAP) was added (50 µg ml−1) for growth of Asd− strains . Phosphate-buffered saline containing 0.01% gelatin (BSG) was used for the resuspension of concentrated vaccines.
General DNA procedures
DNA manipulations were carried out as described in the procedures of Sambrook et al. . Transformation of E. coli and Salmonella was done by electroporation (Bio-Rad, Hercules, CA, USA). Transformants containing Asd+ plasmids were selected on L agar plates without DAP. Only clones containing the recombinant plasmids were able to grow under these conditions.
Protein samples were separated by discontinuous sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and then transferred electrophoretically to nitrocellulose membranes. The membranes were incubated with mouse monoclonal antibodies specific for PspA (Xi126) , OmpC , or β-galactosidase (Sigma), and then with a peroxidase-conjugated goat anti-mouse immunoglobulin G (Bio-Rad). Immunoreactive bands were detected by the addition of 4-chloro-1-naphthol (Sigma) in the presence of H2O2.
Immunization of mice
Three groups of five inbred 7 weeks old female BALB/c mice were used for the immunization experiment. Food and water were withdrawn 4 h prior to immunization and resupplied 30 min after immunization. The recombinant S. typhimuriumχ8501 (pYA3493), χ8501 (pYA3494), and χ8501 (pYA3496) vaccines grown in LB broth to an OD600 of 0.8 were concentrated, suspended in BSG and orally administered (109 colony-forming units (CFU) in 20 µl) to BALB/c mice. Blood was obtained by retro-orbital puncture with heparinized capillary tubes at biweekly intervals. Following centrifugation at 4000×g for 5 min, the serum was removed from the whole blood and stored at −20°C.
Enzyme-linked immunosorbent assay (ELISA)
ELISA was used to assay antibody titers to rPspA and S. typhimurium lipopolysaccharide (LPS). The ELISA procedure used in this study is the same as described by Kang et al. . The purified His6-rPspA antigen was prepared as described in a previous study . Polystyrene 96-well flat-bottom microtiter plates were coated with S. typhimurium LPS (100 ng per well; Sigma) or purified His6-PspA (100 ng per well). Sera obtained from the same experimental group (five mice per group) were pooled and diluted serially. A 100 µl volume of diluted samples was added to individual wells in duplicate and incubated for 2 h at 37°C.
Subcellular localization of rPspA specified by pYA3496 in Salmonella
Plasmid pYA3496 constructed in a previous study  encodes a His6-tagged α-helical region of PspA from amino acid residue 3 to 257 of the mature PspARx1 protein (588 amino acids) (Fig. 1). An apparent signal sequence is not observed at the N-terminus of His6-rPspA. To observe His6-rPspA expression and localization in Salmonella, plasmid pYA3496 was introduced into S. typhimuriumχ8599. β-Galactosidase production from the atrB13::MudJ allele in χ8599 was used as a cytoplasmic protein marker and as an indicator of membrane leaking in the examination of subcellular fractions.
Subcellular fractionation was preformed by the same procedures as described by Kang at al. . The S. typhimuriumχ8599 (pYA3496) culture grown in LB broth to an OD600 of 0.8 was centrifuged and the supernatant fluid saved. The extracellular proteins in the culture supernatant fraction (750 µl) were precipitated with 10% trichloroacetic acid (1 h, 4°C). A portion of the cell pellet was resuspended in 800 µl of 100 mM Tris–HCl buffer (pH 8.6) containing 500 mM sucrose and 0.5 mM ethylenediamine tetraacetic acid (EDTA) and subjected to the lysozyme-osmotic shock treatment to prepare the periplasmic fraction . Cell lysates prepared by cell disruption with a French press were centrifuged at 132,000×g at 4°C for 1 h to separate the soluble fraction and the pelleted cell envelopes. The soluble fraction contained the cytoplasmic proteins. To isolate the outer membrane fraction, total envelope pellets were suspended in 4 ml of 20 mM Tris-HCI (pH 8.6) containing 1% Sarkosyl to remove the inner membrane as a soluble form from the total envelope. The outer membrane fraction was obtained as a pellet after centrifugation. An equal volume (30 µl) of each fraction except supernatant fluids was separated by SDS–PAGE for Western blot analysis.
Although the deduced size of His6-rPspA was approximately 30 kDa, it was detected as an approximately 35 kDa protein (Fig. 2), which is a similar migration of rPspA seen in a previous study . A large amount of the His6-rPspA resided in the cytoplasmic fraction, not in the periplasm of the Salmonella host. The rPspA was detected in the concentrated culture supernatant fluid (21-fold). Little or no rPspA was detected in the outer membrane fraction. In the immunoblot analyses of subcellular fractions with anti-OmpC and anti-β-galactosidase monoclonal antibodies, the OmpC protein was detected in the outer membrane fraction, but the β-galactosidase was detected in both the cytoplasm and concentrated supernatant fluids. The growth rate of χ8599 (pYA3496) in LB broth at 37°C was less than that of χ8599 (pYA3494). The β-galactosidase activities in the log-phase culture supernatant of χ8599 (pYA3496) cultures increased proportional to the increase of cell density, while the culture supernatants of χ8599 (pYA3494) cultures maintained a constant very low level of β-galactosidase activity (data not shown). These results suggest that the His6-rPspA detected in concentrated culture supernatant fluid was due to non-specific membrane leaking or cell lysis instead of an active secretory mechanism. This is in accord with the observed lower growth rate. χ8599 harboring pYA3342 (vector alone) was used as the control and grew with a growth rate similar to that of χ8599 (pYA3494).
S. typhimurium χ8501 vaccine strains expressing rPspA antigen
pYA3496 encoding His6-rPspA were electroporated into the Δcrp-28ΔasdA16 strain χ8501. The S. typhimuriumχ8501 (Δcrp-28ΔasdA16) vaccine strain containing pYA3496 also expressed the His6-rPspA protein at an approximate molecular mass of 35 kDa. In the immunoblot analyses, the amount of rPspA protein expressed in χ8501 (pYA3496) was at a similar level as expressed in χ8501 (pYA3494) (Fig. 3). To examine the stability of plasmid pYA3496 under unpressured culture conditions in vitro, χ8501 cells containing pYA3496 were cultured with daily passage of 1:1000 dilutions for five consecutive days in LB broth containing DAP. All χ8501 clones examined (100 clones per day) kept the Asd+ plasmid pYA3496, indicating that pYA3496 was very stable in the χ8501 vaccine strain.
Immune responses induced by recombinant S. typhimurium-rPspA vaccines
The recombinant S. typhimuriumχ8501 (pYA3494) and χ8501 (pYA3496) vaccines (1.2×109 and 1.8×109 CFU, respectively, in 20 µl BSG) were administered to BALB/c mice. The recombinant S. typhimuriumχ8501 (pYA3493) vaccine (2.1×109 CFU in 20 µl of BSG) was used as a vector control. All immunized mice survived without any signs of disease in the immunized mice during the 6 weeks experimental period. The antibody responses to Salmonella LPS and to the foreign antigen rPspA were measured from the sera of the immunized mice.
The serum IgG responses to LPS and rPspA are presented in Fig. 4. χ8501 (pYA3493), χ8501 (pYA3494) and χ8501 (pYA3496) vaccines elicited anti-LPS IgG responses, suggesting that all Salmonella vaccines colonized lymphoid effector tissues in mice. Vaccines χ8501 (pYA3494) and χ8501 (pYA3496) elicited anti-rPspA IgG responses. Surprisingly, χ8501 (pYA3494) secreting rPspA antigen induced significantly higher levels of IgG than those induced by χ8501 (pYA3496) expressing His6-rPspA antigen. End point titer analysis with the sera at the 6 weeks time point showed that the IgG titer induced by χ8501 (pYA3494) (1.9×107) was 104-fold higher than induced by χ8501 (pYA3496) (<2.4×103), indicating that the rPspA antigen expressed in χ8501 (pYA3494) might more efficiently associate with or be presented to the systemic effector immune systems than that expressed in χ8501 (pYA3496). Anti-rPspA IgG was not detected in sera obtained from mice immunized with χ8501 (pYA3493), the vector control vaccine.
Since the first development of the asd gene-based balance lethal host–vector system comprising a bacterial host with an asd gene deletion mutation and an Asd+ vector in 1988 [3,4], the utilities of this system have been demonstrated in many researches to deliver heterologous pathogen's recombinant antigens by live attenuated recombinant Salmonella vaccines [5,14]. Asd+ vectors used in this system have been serially developed by modifying genetic elements of the plasmid to increase multicopy plasmid stability and to optimize expression of recombinant antigens in Salmonella vaccines . A preferable system would have a recombinant antigen secreted from the cytoplasm of Salmonella vaccines [21,22]. Recently, we have developed an Asd+ plasmid pYA3493 facilitating β-lactamase signal sequence fusion to the N-terminus of recombinant antigen for sending out expressed recombinant antigen from the cytoplasm .
In our previous study, a DNA fragment encoding a highly antigenic α-helical region of PspA was subcloned in the multicopy Asd+ pYA3493 vector to create pYA3494. Approximately 50% of the recombinant PspA (rPspA) expressed in Salmonella strains harboring pYA3494 was detected in the combined supernatant and periplasmic fraction of broth-grown recombinant Salmonella. S. typhimuriumΔcrpχ8501 vaccine carrying pYA3494 stimulated IgG immune responses to the rPspA antigen . Although Salmonella-secreting rPspA vaccine stimulated protective rPspA immune responses, it remained to be solved how important the rPspA antigen cellular location (secreting vs. cytoplasmic) was in stimulating systemic immune responses. To answer this question, a comparison is needed of rPspA-specific immune responses induced by a secreted versus a cytoplasmic form of rPspA in Salmonella vaccines.
BALB/c mice orally immunized with a single dose (~109 CFU) of S. typhimuriumχ8501 expressing either rPspA or His6-rPspA induced immune responses to both Salmonella LPS and rPspA (Fig. 4). Detection of anti-LPS IgG in sera from Salmonella-rPspA vaccine immunized mice indicates that both χ8501 (pYA3494) and χ8501 (pYA3496) vaccines likely reached appropriate lymphoid tissues to stimulate a systemic immune response. There was, however, a significant difference in rPspA-specific IgG immune responses stimulated by Salmonella vaccines carrying pYA3496 or pYA3494. In the end point titer analyses at the 6 weeks time point, the anti-rPspA IgG titer induced by Salmonella carrying pYA3494 (1.9×107) was much higher (104-fold) than induced by Salmonella carrying pYA3496 (<2.4×103), suggesting that rPspA antigen placed in the periplasm and extracellular fluid is much more immunogenic than that mostly retained in the cytoplasm in Salmonella vaccines. Antigens located in the periplasm and/or secreted may contribute to the increased immune responses by facilitating adequate exposure and folding structures of rPspA antigen-presenting cells for efficient processing. Alternatively, although anti-LPS titers suggests normal colonization of χ8501 (pYA3496), we cannot exclude the possibility that Salmonella expressing the antigen intracellularly grew slowly in vivo due to toxicity of the intracellular antigen as seen in vitro. It should be noted that the vaccine strains were concentrated prior to oral immunization thus discarding any rPspA or His6-rPspA in the culture supernatant fluids.
In conclusion, this study demonstrates that a Salmonella vaccine secreting the rPspA antigen enhances systemic immune responses to the recombinant antigen compared to the antigen retained in the cytoplasm. Although there are differences at the N-terminus of each rPspA, it may not affect very much in the immune responses due to their short sequences. His6-tag is poorly immunogenic (QIAexpress Handbook, Qiagen) and does not interfere with the structure of the tagged proteins [23,24]. We have also elicited high-titer antibodies to His6-rPspA in parenterally immunized rabbits. To our knowledge, this is the first report comparing the immune responses to the same Salmonella vaccine delivered recombinant antigens residing in different subcellular locations. The results of this study provide a direction to enhance the immunogenicity of recombinant antigens in live recombinant attenuated Salmonella vaccines.
This research project was supported by a grant from the U.S. Public Health Service through the National Institutes of Health (DE06669).