|Home | About | Journals | Submit | Contact Us | Français|
A balanced-lethal plasmid expression system that switches from low-copy-number to runaway-like high-copy-number replication (pYA4534) was constructed for the regulated delayed in vivo synthesis of heterologous antigens by vaccine strains. This is an antibiotic resistance-free maintenance system containing the asdA gene (essential for peptidoglycan synthesis) as a selectable marker to complement the lethal chromosomal ΔasdA allele in live recombinant attenuated Salmonella vaccines (RASVs) such as Salmonella enterica serovar Typhimurium strain χ9447. pYA4534 harbors two origins of replication, pSC101 and pUC (low and high copy numbers, respectively). The pUC replication origin is controlled by a genetic switch formed by the operator/promoter of the P22 cro gene (O/Pcro) (PR), which is negatively regulated by an arabinose-inducible P22 c2 gene located on both the plasmid and the chromosome (araC PBAD c2). The absence of arabinose, which is unavailable in vivo, triggers replication to a high-copy-number plasmid state. To validate these vector attributes, the Yersinia pestis virulence antigen LcrV was used to develop a vaccine against plague. An lcrV sequence encoding amino acids 131 to 326 (LcrV196) was optimized for expression in Salmonella, flanked with nucleotide sequences encoding the signal peptide (SS) and the carboxy-terminal domain (CT) of β-lactamase, and cloned into pYA4534 under the control of the Ptrc promoter to generate plasmid pYA4535. Our results indicate that the live Salmonella vaccine strain χ9447 harboring pYA4535 efficiently stimulated a mixed Th1/Th2 immune response that protected mice against lethal challenge with Y. pestis strain CO92 introduced through either the intranasal or subcutaneous route.
Live, attenuated bacteria have been developed to generate safe and immunogenic vaccine strains (50). Attenuated Salmonella enterica has been used as both a homologous vaccine and a delivery system for recombinant heterologous antigens from bacterial, parasitic, viral, and tumor sources (8, 40). The oral administration of Salmonella allows the infection of Peyer's patches via the M cells and colonization of the mesenteric lymph nodes, liver, and spleen, generating a range of humoral and cellular immune responses against Salmonella and the heterologous antigens (8) at local and distal sites such as the mucosa. Because most systems for the expression of heterologous antigenic proteins in Salmonella use plasmids, several approaches have been developed for the antibiotic-free maintenance of plasmid vectors (29, 48). However, a number of factors may affect the immune response to protective antigens, such as the ability of the vaccine strain to invade and colonize the host and the stability of the plasmid expression system. High levels of bacterial protein synthesis specified by multiple-copy plasmids often result in either the rapid loss of the foreign plasmid or a reduction in bacterial growth and the ability to colonize lymphoid tissues due to the demand of the extrametabolic burden. Both of these factors result in a reduction of immunogenicity. The insertion of genes into the bacterial chromosome by homologous recombination can achieve a high degree of stability, but this approach sometimes limits the level of protein synthesis due to the single gene copy and, thus, may lessen the production of a protective immune response with the live vaccine (29).
To overcome some of these problems, we have constructed a balanced-lethal vaccine vector with its copy number regulated by arabinose (pYA4534) that switches to runaway-like high-copy-number replication regulating the delivery and dosage of heterologous antigens. In previous work, the switch from a low to a high copy number of the plasmid was mediated by a temperature change from 30°C to higher than 35°C and is called uncontrolled replication or runaway replication of the plasmid in Escherichia coli (62).
Yersinia pestis is a Gram-negative bacterium that causes plague in humans and is transmitted from rodents to humans by fleas (26, 51). Y. pestis infections present three different clinical forms: bubonic, pneumonic, or septicemic (59). Widespread aerosol dissemination of the bacterium combined with high mortality rates make Y. pestis a deadly pathogen (31). LcrV is a multifunctional protein that forms part of a type III secretion system (T3SS) encoded on Y. pestis 70-kb virulence plasmid pCD1 (16, 52). LcrV along with LcrG helps regulate the expression of Yersinia outer proteins (YOPs) that are injected into the cytosol of the host cell, where they interfere with the cellular signaling involved in phagocytosis and inhibit proinflammatory cytokine production (28, 43, 47). Experimental evidence indicates that antibody responses to LcrV offer protection against plague. Thus, the passive transfer of LcrV monoclonal antibodies (MAbs) or polyclonal-specific serum to LcrV protects animals against bubonic and pneumonic plague (25, 46). Antibodies against LcrV apparently block the translocation of effector YOPs, allowing the phagocytosis of Y. pestis bacilli by macrophages, but the exact mechanism of this protection remains to be determined (19).
In addition to the direct role of LcrV in the formation of the T3SS needle, LcrV has an immunomodulatory function mediated by interleukin-10 (IL-10) induction, which blocks the host protective inflammatory responses and suppresses the proinflammatory cytokines (7). Partial deletions of LcrV and the use of synthetic peptides allowed the identification of two LcrV regions involved in the production of IL-10, which are located from amino acid residues 37 to 57 and from amino acid residues 271 to 285 (34, 49). The induction of IL-10 by LcrV is through the interaction of Toll-like receptor 2 (TLR-2) as well as TLR-6 and cluster of differentiation 14 (CD14) (1, 15, 58). Immunization with full-length LcrV elicited protective immunity (41), but truncated LcrV forms were also able to elicit an immune response that was protective against a lethal challenge with Y. pestis. These variants included rV10 (lacking amino acid residues 271 to 300) (49), the major protective LcrV region (amino acid residues 135 to 275) (25), LcrV196 (amino acid residues 131 to 326) (5), and a small fragment of LcrV (amino acids 135 to 262) (64).
In this work, we describe the construction of pYA4534, a balanced-lethal plasmid expression system containing an arabinose-regulated genetic switch to shift to runaway-like high-copy-number replication in vivo for the regulated delayed dosage of heterologous antigens. The derivative pYA4535, encoding the T2SS β-lactamase N- and C-terminal domains for the export of the lcrV196-encoded fused antigen, was used to transform S. enterica serovar Typhimurium strain χ9447, a new generation of live recombinant attenuated Salmonella vaccines (RASVs) that is phenotypically similar to the wild type at the time of oral vaccination but displays a regulated delayed in vivo attenuation (14), a regulated delayed in vivo synthesis of recombinant antigen (66), and regulated delayed in vivo lysis to release a bolus of protective antigen and confers complete biological containment (35) after host tissue colonization. These RASV strains are able to colonize and persist in the lymphoid tissue without causing disease symptoms, giving an advantageous alternative when carrying heterologous antigens that induce higher protective mucosal and systemic immunity responses. The immune responses of mice immunized orally with this live RASV strain synthesizing an optimized LcrV protein were evaluated for protection against a lethal challenge with virulent Y. pestis CO92 (4). Thus, we offer an alternative for the development of vaccines against clinical forms of plague.
The bacterial strains and plasmids used in this study are listed in Table Table1.1. LB broth (2) supplemented with 0.2% mannose and 0.02% arabinose was used to grow the bacterial strains for colonization and immunization assays. LB broth and LB agar (1.5% agar) or MacConkey agar (Difco) was used for the propagation and plating of bacteria. LB broth, nutrient broth (NB) (Difco) devoid of arabinose and mannose, and minimal salts medium (9) were used to standardize the minimal concentration of arabinose to monitor the switch between the low- and high-copy-number replication of the plasmid. L-broth (0.5% Bacto tryptone, 0.25% NaCl, and 0.25% yeast extract), L-agar (1.5% agar), and Lennox agar (38) were used during RASV strain construction. When required, the medium was supplemented with 0.2% mannose and arabinose (from 0.2% to 0.00002%); for growth of noncomplemented ΔasdA strains and plasmid stability tests, 50 μg/ml diaminopimelic acid (DAP) was added to the growth medium (48).
DNA manipulations were carried out as described previously (53). The transformation of E. coli and S. Typhimurium was done by electroporation (Bio-Rad). Transformants containing Asd+ plasmids were selected on LB agar plates without DAP. Plasmid DNA was isolated by using the QIAprep Spin miniprep kit (Qiagen). Restriction enzymes were used as recommended by the manufacturer (New England Biolabs). All primers used in this study (Integrated DNA Technology) were flanked with restriction enzyme recognition sites and are shown in Table Table2.2. All constructs were verified by DNA sequencing (Arizona State University Facilities).
The construction of pYA4534 was performed in several steps that involved the construction of precursor plasmids pYA3972, pYA3971, and pYA3789, described below. The c2 gene from bacteriophage P22 was PCR amplified from pYA3857 (S. Wang and R. Curtiss III, unpublished results) using primer set c2 f NcoI and c2 r StuI KpnI NcoI. The 613-bp PCR product was digested with NcoI and cloned into NcoI-digested pBAD-HA (Invitrogen) to create pYA3972. Plasmid pYA3971 was assembled as follows. An ApaI site was introduced downstream of the araC gene in the pBAD-HA vector by site-directed mutagenesis using primer set f ApaI araC PBAD and r ApaI araC PBAD. The resulting plasmid was digested with NcoI and subcloned downstream of the araC PBAD activator-promoter regulatory system with the 613-bp NcoI fragment released from pYA3972 containing the c2 gene. The resulting plasmid was digested with StuI-KpnI and was cloned downstream of the c2 gene, a 2,415-bp StuI-KpnI fragment containing SD-gtg-murA SD-gtg-asdA (murA and asdA genes harboring modifications in the Shine-Dalgarno [SD] ribosome binding site and a GTG start codon), which was PCR amplified from pYA3650 (12) with primer set murA asd-f StuI and murA asd r SalI KpnI and digested with the same enzymes. Subsequently, the resulting plasmid was digested with SphI and cloned with a 619-bp SphI fragment containing the nucleotide sequence specifying the promoterless RNA II transcript (pUC plasmid replication origin) (67), which was PCR amplified from pUC18 by using primer set pUC STS G f SphI and pUC STS G r KpnI SphI and digested with SphI. The resulting plasmid was digested with KpnI-ApaI, cloned with a 1,344-bp KpnI-ApaI fragment containing the pSC101 replication origin that was PCR amplified from pYA3337 (11) by using primer set pSC101-up-f-KpnI and pSC101-r-ApaI, and digested with the same enzymes. The resulting plasmid was digested with SalI-SphI and cloned upstream of the nucleotide sequence specifying the RNA II transcript, a synthetic DNA containing the operator-promoter of the c2 gene (O/Pc2) (promoter for repressor maintenance [PRM]) from bacteriophage P22 (63), formed by annealing two complementary single-stranded 80-bp oligonucleotides, regulatory f c2 SalI and regulatory r c2 SphI, flanked with sticky SalI and SphI ends, respectively, to generate pYA3971. This plasmid was digested with SalI and SphI to replace the O/Pc2 sequences with the operator-promoter of the cro gene O/Pcro (promoter right [PR]) from bacteriophage P22 (63) and contained a synthetic DNA formed by annealing two complementary single-stranded 88-bp oligonucleotides, f operator cro SalI and r operator cro SphI, flanked with sticky SalI and SphI ends, respectively. The resulting plasmid was digested with KpnI and cloned with a 97-bp KpnI fragment obtained by annealing two complementary single-stranded 97-bp oligonucleotides, f Ptrc-KpnI and r Ptrc-KpnI, flanked with sticky KpnI ends containing the synthetic trc promoter (Ptrc) and NcoI, NheI, and SmaI restriction sites, to generate pYA3789. A 3,096-bp DNA fragment containing an SD sequence and the structural lacZ gene was PCR amplified from pBAD-HA-lacZ (Invitrogen) by using primer set f pBAD lacZ 387 and r pBAD lacZ 3505. The PCR product was digested with SalI and cloned into SalI-digested pYA3789 to form pYA4534.
A 796-bp DNA fragment was PCR amplified from pYA3841 (4) by using primer set f Vag and r Vag, and this contained the nucleotide sequence of optimized Y. pestis lcrV196, encoding amino acid residues 131 to 326 of LcrV, flanked at its amino terminus with the 24 amino acid residues of the β-lactamase sec-dependent signal peptide (32), followed by 12 amino acid residues of the mature β-lactamase protein (60), and flanked at its carboxy terminus with the last 21 amino acid residues of the β-lactamase carboxy-terminal region (blaSS-lcrV196-blaCT, referred to as lcrV196 in this work). Both β-lactamase regions are required for the efficient secretion of the recombinant protein (5, 36) (Fig. (Fig.1A).1A). The PCR product was digested with NheI and was cloned into NheI-digested pYA3789 under the control of Ptrc to create pYA3728. A 796-bp DNA fragment containing the blaSS-lcrV196-blaCT sequence from pYA3728 was released with NheI and subcloned into pYA4534 digested with the same enzyme to yield pYA4535 (Fig. (Fig.1B).1B). The cloned blaSS-lcrV196-blaCT gene of 771 bp encodes a 257-amino-acid polypeptide (LcrV196) with an estimated pI of 5.9 and a molecular mass of 29 kDa.
PCR products were amplified with primers designed to modify the promoter (P), SD, and start codon sequences and were cloned into a suicide vector (Table (Table1).1). Parental S. Typhimurium strains were mated with the E. coli host strain χ7213 harboring the relevant suicide vector on L-agar plates containing 50 μg/ml DAP. Transconjugants were selected by growth on L-agar plates containing antibiotics without DAP. Conjugations with ΔasdA mutant strains were plated onto antibiotic-containing medium with DAP and colicin B to inhibit E. coli growth. Defined deletion mutations with and without insertions were confirmed by PCR and phenotypic verification. The introduction of the defined deletion mutations with or without specific insertions into other Salmonella strains was mediated by P22HTint transduction of suicide vectors integrated into the deletion or deletion-insertion mutations on the chromosome of the Salmonella donor strains (33). Transductants were selected by growth on Lennox agar plates containing either 25 μg/ml chloramphenicol or 12.5 μg/ml tetracycline. The transductants were resuspended in sterile buffered saline with 0.01% gelatin (BSG) (9) and streaked onto L-agar plates containing antibiotics to obtain isolated colonies. Isolated colonies were inoculated into L-broth and grown to 0.8 optical density at 600 nm (OD600) units and plated onto L-agar plates containing 5% sucrose without NaCl or dextrose to select for double-crossover recombinants. Double-crossover recombinants were screened for antibiotic sensitivity, antibiotic-sensitive recombinants were selected, and deletion or deletion-insertion mutations were confirmed by PCR and phenotypic verification. Maps of each deletion-insertion mutation present in χ9373 and its derivative χ9447 were described previously for Δpmi-2426, Δ(gmd-fcl)-26, and ΔrelA198::araC PBADlacI TT (39); ΔPfur81::TT araC PBAD fur, ΔPcrp527::TT araC PBAD crp, ΔaraE25, and ΔaraBAD23 (14); ΔPmurA7::TT araC PBAD murA and ΔendA2311 (35); and Δasd21::TT araC PBAD c2 (24). Evans blue uridine agar plates were used to confirm that transductants were phage free and not P22 lysogens (3). MacConkey agar plates supplemented with 1% maltose were used to confirm the phenotype of the cyclic AMP receptor protein gene (crp) mutants (10). Chrome Azurol S plates were used to confirm the constitutive synthesis of siderophores characteristic of fur mutants (55). The presence of the Δasd21 mutation was confirmed by the inability of the strain to grow on medium without DAP. Lipopolysaccharide (LPS) profiles of Salmonella strains were examined with silver-stained polyacrylamide gels as described previously (27).
A single colony of χ9447(pYA4534) was inoculated into 1 ml of 1× minimal salts medium (9) supplemented with 0.2% (vol/vol) glycerol, 0.2% glucose, and 2% Casamino Acids. The next day, the culture grown overnight was diluted at a 1:50 ratio to inoculate 3 ml of minimal medium without glucose and supplemented with arabinose from 0.00002% to 0.2% in independent tubes and grown at 37°C until the cultures reached an OD600 of 0.8.
The plasmid stability of strain χ9447(pYA4534) or χ9447(pYA4535) was determined with LB broth with 0.02% arabinose and 0.2% mannose under selective and nonselective conditions (presence of DAP) (48). A culture from each vaccine strain grown overnight was diluted 1:1,000 into prewarmed LB broth supplemented as described above and grown standing for 24 h at 37°C. The next day, the culture of each vaccine strain was diluted and grown as indicated above, and this process was repeated for five consecutive days (approximately 50 generations). The proportions of cells retaining the Asd+ plasmids were determined for each culture, and the cultured cells from each day were diluted, spread onto LB agar plates (0.2% arabinose, 0.2% mannose, and 50 μg/ml X-gal [5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside]) with 50 μg/ml DAP, and grown overnight. Afterwards, 100 colonies from each culture for each day were picked and patched onto LB agar plates (0.2% mannose, 0.2% arabinose, and 50 μg/ml X-gal) with or without 50 μg/ml DAP. The next day, the percentage of clones retaining the plasmids from each culture was determined by counting the colonies grown on these LB agar plates with and without DAP. At the same time, 10 clones from each day's plating were chosen and grown in LB broth (0.2% mannose) for plasmid extraction and restriction pattern analysis. Afterwards, one clone from each culture was chosen and grown in LB broth (0.2% mannose, from 0.00002% to 0.2% arabinose, and without arabinose) until the cultures reached an OD600 of 0.8 for agarose gel analysis to evaluate the amount of supercoiled plasmid DNA and for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblot analysis to confirm the synthesis of the correct size and the amount of LcrV196 by each clone.
A frozen glycerol stock of strain χ9447(pYA4534) or χ9447(pYA4535) was streaked onto LB agar plates (0.2% arabinose, 0.2% mannose, and 50 μg/ml X-gal) and incubated at 37°C. A single colony from each plate was inoculated in NB (75 ml in a 250-ml flask) with 0.02% arabinose and grown at 37°C to 0.8 OD600 units. Each culture (25 ml) was used to inoculate prewarmed NB (225 ml in a 500-ml flask) with or without 0.02% arabinose at a dilution of 1:10. The cultures were then grown at 37°C with shaking, and each hour, an aliquot of 8 ml was taken from each culture until 12 h. Afterwards, aliquots of 8 ml were taken every 24 h (5 ml for plasmid extraction, 1 ml to read the OD600 and to determine the CFU, 1 ml for the β-galactosidase [β-gal] activity assay, and 1 ml for the protein profile analysis).
To detect the level of β-gal activity on plates, LB agar was supplemented with 50 μg/ml X-gal. The β-gal assays were performed with permeabilized cells using the substrate o-nitrophenyl-β-d-galactopyranoside (Sigma). The β-gal activity was measured as previously described by Miller (45). Average values (± standard deviations) for activity units were calculated from three independent assays.
The expression of the recombinant 6×His-tagged full-length LcrV (6×His-LcrV) protein encoded by plasmid pHT-V was described previously (21), and the protein was purified by nickel-nitrilotriacetic acid agarose chromatography under native conditions. Eluted fractions containing purified 6×His-LcrV were selected after SDS-PAGE, pooled, dialyzed against phosphate-buffered saline (PBS), and then concentrated. The protein concentration was determined by the Bradford assay (3a) using bovine serum albumin as a standard. Endotoxins were removed by using Detoxi-Gel endotoxin-removing gel (Pierce, Rockford, IL). The amount of endotoxin contamination in the recombinant protein was measured quantitatively with the Limulus amebocyte lysate assay (Cambrex Bio Science Walkersville, Inc.), which was used according to the manufacturer's instructions. The amount of endotoxin found was <0.01 endotoxin units (EU) per μg of recombinant protein. The purified 6×His-LcrV protein was used for enzyme-linked immunosorbent assays (ELISAs) and enzyme-linked immunospot (ELISPOT) assays. Salmonella outer membrane proteins (SOMPs) were obtained using a sonication and Triton X-100 extraction procedure described previously (44).
Cell fractions were prepared from 1 ml of culture by using the PeriPreps periplasting kit (Epicentre) according to the manufacturer's instructions. For the isolation of proteins released into the culture supernatants, 1 ml of each sample was filtered (0.22-μm pore size; Corning) and precipitated with 10% trichloroacetic acid (TCA), and the pellet obtained by centrifugation was resuspended in 100 μl of lithium dodecyl sulfate (LDS) sample buffer (Invitrogen). For the total extract, 10 μl of 1 ml of culture was used for gel loading, whereas the load samples for the other fractions were at a final dilution factor of 1:10. Thus, 10 μl was mixed with 10 μl of LDS sample buffer, boiled for 5 to 10 min, analyzed by SDS-PAGE (NuPAGE 4 to 12% Bis-Tris; Invitrogen), and immunoblotted. The recombinant protein was identified with rabbit anti-LcrV serum (1:15,000) (61), followed by alkaline phosphatase-conjugated anti-rabbit IgG (Sigma). Mouse anti-σ70 monoclonal antibody (1:1,000) was used as a loading marker and for cell integrity (Neoclone). All experiments were performed three times.
Female BALB/c mice, 6 to 7 weeks old, were purchased from Charles River Laboratories. The Arizona State University Animal Care and Use Committee approved all animal procedures. Mice were acclimated for 7 days before the experiments were started. Mice were deprived of food and water for 6 h before oral immunization. The RASV χ9447 strain harboring pYA4534 or pYA4535 was grown statically for 18 h in 5 ml LB broth with 0.2% mannose and 0.02% arabinose at 37°C. These cultures were used to inoculate 100 ml of LB broth with 0.2% mannose and 0.02% arabinose and grown at 37°C with shaking (180 rpm) to 0.8 OD600 units. Cells were pelleted by centrifugation at room temperature (4,000 × g for 15 min), and the pellet was resuspended in 1 ml of BSG. Dilutions of the vaccine strains were plated onto LB agar plates (0.2% arabinose, 0.2% mannose, and 50 μg/ml X-gal) to determine bacterial titers. One group of 40 mice (for vaccination with RASV χ9447 harboring pYA4534), another group of 40 mice (for vaccination with RASV χ9447 harboring pYA4535), and a third group of 24 mice (for BSG alone) were orally inoculated with 20 μl of the respective RASV suspended in BSG containing 2 × 109 CFU or BSG on days 0 and 9. Water and food were returned to the mice 30 min after immunization. Blood samples were obtained by submandibular vein puncture 2 days before vaccination and at days 22 and 32 after the first vaccination. Next, the blood was incubated at 37°C for 60 min. Afterwards, the blood was centrifuged at 4,000 × g for 5 min. The serum was removed and stored at −70°C. Vaginal secretion specimens were collected into 50 μl of BSG 2 days before immunization and at days 22 and 32. Samples were stored at −20°C.
To assess the protective effects of RASV against Y. pestis in immunized mice, groups of eight immunized or control mice were challenged with Y. pestis strain CO92 (biovar Orientalis) (4) at 28 days after the second immunization (day 35). For subcutaneous (s.c.) challenge, a low dose (LD) of 4.49 × 102 CFU or a high dose (HD) of 5.63 × 103 CFU of Y. pestis grown at 28°C in heart infusion broth (HIB; Difco) supplemented with 0.2% xylose was inoculated subcutaneously for each group of mice. For intranasal (i.n.) challenge, mice were anesthetized with a mixture of 0.5% ketamine and 0.01% xylazine administered intramuscularly, and 25 μl of PBS containing an LD of 4.1 × 103 CFU or an HD of 4.4 × 104 CFU of Y. pestis grown in HIB at 37°C supplemented with 0.2% xylose and 2 mM CaCl2 was inoculated intranasally for each group of mice. The mice were observed daily, and mortality was recorded until 14 days after the challenge. The surviving animals were euthanized at the end of the experiment, and blood samples were obtained by intracardiac puncture. The blood samples were left overnight at 4°C and then centrifuged at 4,000 × g for 5 min. The sera were removed, and half of each sample was incubated for 3 days on tryptose blood agar (TBA) plates at 28°C to verify that they were free of bacteria. The serum samples were used for serological analysis.
IgG, IgG2a, IgG1, and IgA antibody titers against LcrV196 and SOMP from mice were determined by ELISA. Briefly, Nunc Immunoplate Maxisorb F96 plates (Nalge Nunc, Rochester, NY) were coated with purified LcrV (500 ng/well) or SOMP (500 ng/well) suspended in 0.05 M carbonate-bicarbonate buffer (pH 9.6). The coated plates were incubated overnight at 4°C. The plates were washed three times with PBS containing 0.05% Tween 20. Free binding sites were blocked with PBS-0.1% Tween 20 containing 2% bovine serum albumin (BSA). Vaginal washes and sera obtained from the same experimental group were pooled. The serum pools were serially diluted by 2-fold dilutions from an initial dilution of 1:200 in PBS. The vaginal secretion pools were serially diluted by 2-fold dilutions from an initial dilution of 1:10 in PBS. Aliquots of 100 μl of the serially diluted samples were added to duplicated wells and incubated overnight at 4°C. Plates were washed and treated with horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG, IgG1, IgG2a, or IgA (1:4,000 dilution; Southern Biotechnology Inc., Birmingham, AL). Wells were developed with o-phenylenediamine dihydrochloride (OPD) at 0.4 mg/ml in phosphate-citrate buffer with H2O2 (Sigma) at 200 μl/well. Color development was stopped by the addition of 50 μl 3 M H2SO4 per 200 μl of reaction solution. Absorbance was recorded at 492 nm by using an automated ELISA plate reader (Labsystems Multiskan MCC/340). End-point titers were expressed as the reciprocal log10 of the last sample dilution given an absorbance of 0.1 OD units above negative controls after 1 h of incubation.
An ELISPOT assay was performed to enumerate the gamma interferon (IFN-γ), IL-4, and IL-10 cytokine-secreting cells (CSC) in the spleens of immunized and naïve mice in order to determine the potential cellular immune response to immunization. This was performed with ELISPOT kits (mouse IFN-γ, IL-4, and IL-10 ELISPOT sets; eBioscience) according to the manufacturer's instructions. Briefly, three mice from each group were euthanized 3 weeks after the second vaccination. Single-cell suspensions were prepared by mechanical dissociation, and splenocytes from the mice belonging to the same group were pooled. The 96-well polyvinylidene difluoride (PVDF) membrane ELISPOT plates (MultiScreenHTS-IP filter plate; Millipore, Bedford, MA) were coated with either anti-mouse IFN-γ, IL-4, or IL-10 and incubated at 4°C overnight. The plates were washed and blocked with RPMI 1640 medium plus 10% fetal bovine serum (Invitrogen Life Technologies) at room temperature for 1 h. The blocking solution was removed, and spleen cell suspensions at various densities ranging from 2.5 × 105 to 1 × 106 spleen cells in 10% fetal bovine serum (FBS)-RPMI 1640 medium were added to each well and restimulated with recombinant LcrV (1 μg/well) or medium. The spleen cells from all groups of mice were incubated in the presence of 10 U/ml of mouse IL-2 (Roche) for 24 h (for IFN-γ-secreting cells) or 48 h (for IL-4- or IL-10-secreting cells) at 37°C in a humidified (5% CO2-in-air) incubator. After the incubation period, the cells were removed by washing, and the properly diluted biotinylated anti-IFN-γ, -IL-10, or -IL-4 detection antibodies were added to each well and incubated at room temperature for 2 h. The antibody solution was then decanted, and the plates were washed four times. Avidin-horseradish peroxidase was added and incubated at room temperature for 45 min. The plates were washed, and freshly prepared AEC (3-amino-9-ethylcarbazole) substrate solution (BD Bioscience) was used to develop the spots at room temperature for 10 to 60 min. The substrate reaction was stopped by washing the wells once with distilled water. The spots were counted by using an automated ELISPOT plate reader (CTL Analyzers; Cellular Technology Ltd., Cleveland, OH).
To measure the secretion of IFN-γ, IL-10, and IL-4 by spleen cells, ELISAs were performed (IFN-γ, IL-10, and IL-4 ELISA set kits; eBioscience). Briefly, 21 days after the second immunization, pools of 5 × 106 spleen cells from three mice belonging to the same group were suspended in 1 ml of 10% FBS-RPMI 1640 medium in the presence of 10 U/ml of mouse IL-2 (Roche), restimulated with recombinant LcrV (10 μg) or medium, and then incubated for 48 h at 37°C in an humidified (5% CO2 in air) incubator. Afterwards, the cell culture supernatants were collected, and assays were performed according to the kit instructions.
Single-cell suspensions of spleens obtained as described above for ELISPOT assays were incubated with an anti-CD16/CD32 MAb for 10 min to block Fc receptors and then stained for 1 h with fluorescein isothiocyanate (FITC) anti-mouse CD4 (L3T4) or phycoerythrin (PE) anti-mouse CD8 (Ly-2), and the isotype controls used were FITC-rat IgG2a(κ) and PE-rat IgG2a(κ) (all antibodies were purchased from eBioscience). Cells were washed twice with fluorescence-activated cell sorter (FACS) buffer (PBS containing 0.1% sodium azide and 1% fetal calf serum [FCS]), fixed in fixation buffer (eBioscience) diluted 1:4 with PBS, and then analyzed by using a Cytomics FC 500 flow cytometer (Beckman Coulter Inc.). Data analysis was performed with CXP software.
Differences in antibody responses and cytokine secretion levels measured by ELISA and ELISPOT assay between groups were determined by one-way analysis of variance (ANOVA), followed by Tukey's multiple-comparison test. Differences with P values of <0.05 were considered significant at the 95% confidence interval. Statistical analysis was performed by using GraphPad Prism software (GraphPad Software, San Diego, CA).
The Kaplan-Meier method (GraphPad Prism software) was applied to obtain the survival fractions following s.c. or i.n. Y. pestis challenges of orally immunized mice. By using the Mantel-Haenszel log-rank test, the P values for statistical differences between survival rates of the plague challenges in the Salmonella-vaccinated groups or oral BSG controls were determined at the 95% confidence interval.
The genes, promoters, and plasmid replication origins used to generate pYA4534 were flanked with restriction enzyme sites and assembled as cassettes for the development of a versatile vector that can be further improved depending on the characteristics of the antigens under evaluation. Plasmid pYA4534 contains two plasmid replication origins from low-copy-number plasmid pSC101 and high-copy-number plasmid pUC. To tightly control the high copy number of the plasmid, the nucleotide sequence specifying the promoterless RNA II transcript (pUC replication origin) was cloned downstream of the cro operator-promoter O/Pcro (PR), which is regulated by the P22 C2 repressor protein. The synthesis of the C2 repressor protein, encoded by c2 gene copies on both the plasmid and chromosome of the RASV strain, was under the control of the araC PBAD activator-promoter regulatory system, inducible by l-arabinose. Thus, in the presence of 0.02% arabinose, C2 is synthesized and binds to the cro operator sites (OR1, OR2, and OR3), blocking the binding of RNA polymerase to the PR promoter upstream of the RNA II transcript, leading to pSC101-dependent replication. In the absence of arabinose, the PR promoter is accessible to RNA polymerase and produces the RNA II transcript, resulting in pUC-dependent replication. Thus, in vivo, the RASV χ9447 strain transformed with pYA4534 is in an arabinose-free environment, and during its multiplication in host tissues, the synthesis of the C2 protein ceases, and its concentration gradually decreases during each cell division until its depletion, leading to a high-copy-number plasmid state along with a high level of expression of the genes encoding the heterologous antigens.
Plasmid pYA4534 also contains the SD-gtg-asdA and SD-gtg-murA genes, whose expression is under the control of the araC PBAD regulatory system, which is inducible by arabinose. The asdA gene complements the lethal chromosomal ΔasdA mutant allele of the RASV χ9447 strain. The asdA and murA genes each contain a GTG start codon, which replaces the ATG codon to reduce their translation efficiency but increases their stability in the cell (35). The asdA selectable marker encodes an enzyme involved in lysine biosynthesis necessary for the synthesis of diaminopimelic acid (DAP), which is a basic component of the peptidoglycan layer of the bacterial cell wall and allows the maintenance of this antibiotic resistance-free plasmid. The murA gene encodes a key enzyme, UDP-N-acetylglucosamine enolpyruvyl transferase, which is involved in bacterial cell wall peptidoglycan synthesis (6). An SD-lacZ promoterless gene was also included in pYA4534, which allows the monitoring of the switch between low- and high-copy-number replication of the plasmid during bacterial growth. For the expression of the genes encoding the heterologous antigens, synthetic Ptrc was introduced into pYA4534. During the growth of RASV χ9447(pYA4535) in the presence of arabinose, the Ptrc repressor LacI is synthesized (due to the presence of the ΔrelA198::araC PBADlacI TT insertion-deletion mutation), and this blocks the synthesis of the Ptrc-controlled synthesis of the LcrV196 protective antigen. In vivo, where arabinose is absent, LacI synthesis ceases and is diluted at each cell division, leading to the derepressed synthesis of the LcrV196 antigen to stimulate the induction of immunity.
To enhance the immunogenicity of the RASV strains synthesizing and delivering the recombinant heterologous antigens, RASV strain χ9447 was constructed. Strain χ9447, derived from strain χ9373 (24), harbors deletion-insertion mutations to exhibit a regulated delayed attenuation in vivo. This strain also exhibits regulated delayed in vivo synthesis and secretion of the protective LcrV196 antigen. It is therefore phenotypically wild type at the time of immunization and becomes attenuated and commences to synthesize protective antigen only after colonization of host tissues. To achieve this goal, we replaced the promoter sequences of genes involved in metabolic functions or virulence with the tightly arabinose-regulated araC PBAD activator-promoter cassette (14, 35) and repressed the synthesis of protective antigen by the arabinose-regulated synthesis of LacI to prevent the expression of the lcrV196-carrying sequence on pYA4535 until after lymphoid tissue colonization. Strain χ9447 harbors the deletion-insertion mutations ΔasdA21::TT araC PBAD c2 and ΔPmur7::TT araC PBAD murA. The lethal asdA deletion is rescued by the addition of DAP to the growth medium and is needed for the antibiotic resistance-free maintenance of the runaway-like replication plasmid pYA4534, and the c2 and murA promoters were replaced with the araC PBAD activator-promoter regulatory system such that C2 and the MurA enzyme cease to be synthesized in vivo due to the lack of free arabinose. Benefits of other deletion-insertion mutations present in χ9447, such as Δpmi-2426, Δ(gmd-fcl)-26, ΔPfur81::TTaraC, PBAD fur, ΔPcrp527::TT araC, PBAD crp, ΔaraE25, ΔaraBAD23, ΔrelA198::araC, and PBAD lacI TT ΔendA2311, were described in detail previously (13, 14, 24, 35, 39, 57, 66).
The minimal concentration of arabinose required to maintain a low-copy-number plasmid state was analyzed with several culture media, such as minimal medium, NB, and LB broth. The addition of 0.02% arabinose to the different growth media was the minimal concentration of arabinose required to maintain a low-copy-number state of the plasmid and repress pUC-dependent replication (Fig. (Fig.2A).2A). However, in the presence of 0.00002% arabinose in minimal medium or NB, there was the maximum amount of supercoiled plasmid DNA as observed without arabinose. In contrast, in LB broth, the presence of 0.002% arabinose permitted substantial plasmid replication, while 0.0002% and 0.00002% arabinose allowed maximum pUC-dependent replication of supercoiled plasmid (Fig. (Fig.2A).2A). This is due to the presence of low-level amounts of arabinose in the yeast extract used to prepare the medium (66).
The expression profile of χ9447 expressing lcrV196 from pYA4535 was analyzed in LB broth with and without 0.02% arabinose until the culture reached an OD600 of 0.8. One protein band with an approximate molecular mass of 29 kDa in total cell extracts and in cell fractions reacted specifically with the anti-LcrV serum as detected by Western blotting with an anti-LcrV polyclonal antibody (Fig. (Fig.2B).2B). The synthesis of the LcrV196 protein and its secretion profile were proportional to the plasmid copy number and inversely proportional to the concentration of arabinose in the culture medium. The amount of LcrV196 in the total extract without arabinose was less than that in the cytoplasmic fraction due to the ratio of each fraction loaded onto the gel (for total extract, 1:100; for all other fractions, 1:10). σ70, a cytoplasmic protein, was used as a loading marker and to evaluate cell integrity. The discrepancy in the levels of σ70 in periplasmic and supernatant fractions may be due to the fact that there was slight cell lysis in the culture in the absence of arabinose compared to that in the presence of arabinose. Plasmids pYA4534 and pYA4535 were stably maintained for 50 or more generations in the χ9447 host grown under nonselective conditions, such as in the presence 50 μg/ml DAP (Fig. (Fig.3A).3A). Similar amounts of supercoiled plasmid were observed for χ9447(pYA4534) and χ9447(pYA4535) when both strains were grown in LB broth supplemented with 0.2% mannose, 0.00002% to 0.2% arabinose, and without arabinose. We provide the results only for χ9447(pYA4535) (Fig. (Fig.3B).3B). The synthesis of LcrV196 by χ9447(pYA4535) was shown to be stably maintained after 50 or more generations when the strain was grown in LB broth as described in Materials and Methods (Fig. (Fig.3C3C).
In order to determine growth rates, the viability of cells, and the time during growth when the shift from low to high copy numbers of the plasmid occurs, one colony each of Salmonella strains χ9447(pYA4534) (vector control) and χ9447(pYA4535) (expressing lcrV196) was inoculated into and grown in NB with 0.02% arabinose as described in Materials and Methods. An aliquot from each culture was taken every hour until 12 h and then at 24 h and every 24 h afterwards. The growth rates of strains χ9447(pYA4535) and χ9447(pYA4534) in NB with or without 0.02% arabinose during the first 12 h of growth were similar (Fig. (Fig.4A).4A). The OD600 of the cultures increased over time until 12 h of growth, but additional growth was not observed between 12 and 24 h, and after 24 h, the OD600 began to decrease moderately in either strain until the end of the experiment (216 h, at day 9) (data not shown). The OD600 between 24 and 216 h was moderately lower without arabinose than with 0.02% arabinose in cultures of both strains χ9447(pYA4535) and χ9447(pYA4534) (data not shown). Viable cell numbers were determined by counting the CFU on LB agar plates (0.2% arabinose, 0.2% mannose, and 50 μg/ml X-gal). The numbers of CFU under both conditions of growth, with and without arabinose, were similar from those for the initial adaptation period of 3 h at the beginning of the culture until 24 h (Fig. (Fig.4B).4B). The number of bacterial generations of growth that occurred from the start of the culture to 24 h of growth was approximately 2 to 2.5. After 24 h of growth, the cell viability of the cultures with arabinose began to decline drastically compared with the cell viability of the cultures without arabinose, which gradually decreased over the course of the experiment (data not shown). The OD600 values of the cultures with and without arabinose were similar and consistent with the cell viability during the first 24 h of growth. After this time, the OD600 and cell viability of the cultures with arabinose were not proportional (data not shown). The faster decrease in cell viability observed after 24 h of culture with arabinose might be the result of a low-copy-number plasmid state in the bacterial cell that maintained a low expression level of the asdA and murA genes carried by plasmids pYA4534 and pYA4535, whose products are necessary for bacterial cell wall synthesis.
The shift from low to high copy numbers of the plasmid in the cultures without arabinose occurred during the first bacterial generation from 0 to 5 h, as determined by β-gal activity assays (Fig. (Fig.4C)4C) and by visualization of the increase in amounts of plasmid DNA on agarose gels from 4 to 6 h (data not shown), which has been established as the second generation of growth. The ratio of high-copy-number to low-copy-number cells reached its maximum at 24 h of growth, during which the cultures produced 2 to 3 bacterial cell divisions.
χ9447(pYA4534) (vector control) and χ9447(pYA4535) synthesizing LcrV196 were grown in LB broth supplemented with 0.02% arabinose and 0.2% mannose until an OD600 of 0.8 was reached, and colonization was evaluated with BALB/c mice inoculated orally with 2 × 109 CFU at day 0 and euthanized at days 2, 4, 6, and 8. We observed colonization in several tissues, including Peyer's patches, mesenteric lymphoid nodes, and spleen by χ9447 harboring pYA4534. However, colonization by χ9447(pYA4535) synthesizing LcrV196 was observed only in Peyer's patches and mesenteric lymphoid nodes (data not shown).
To investigate the immunogenicity of the LcrV196 protein delivered by the RASV strain, we orally immunized groups of BALB/c mice on days 0 and 9 with S. Typhimurium χ9447 harboring either the control vector (pYA4534) or pYA4535 for the synthesis of LcrV196. Serum IgG responses to LcrV196 and SOMP from immunized mice were measured by ELISA. Systemic IgG responses to LcrV196 were observed 2 weeks postimmunization and showed an incremental increase at 3 weeks. The LcrV196-vaccinated group had significantly higher anti-LcrV196 antibody titers after 2 and 3 weeks postimmunization (P < 0.001) (Fig. (Fig.5A).5A). Low levels of anti-LcrV196 IgG were detected in sera obtained from mice immunized with vector control strain χ9447(pYA4534) or in preimmune serum. Similar results were observed previously (41). These results indicate that LcrV196 delivered by χ9447(pYA4535) induces systemic antibody IgG titers against LcrV196. The anti-SOMP IgG responses significantly increased in mice vaccinated with χ9447(pYA4534) and χ9447(pYA4535) 2 and 3 weeks postimmunization compared to preimmune serum (P < 0.001) (Fig. (Fig.5B).5B). However, the anti-SOMP IgG responses elicited by χ9447(pYA4534) at 2 and 3 weeks postimmunization were higher than the titers observed with χ9447(pYA4535).
The immune response induced by χ9447(pYA4535) synthesizing LcrV196 was further examined by measuring the levels of IgG isotype subclasses IgG1 and IgG2a in preimmune serum and at 2 and 3 weeks postimmunization by ELISA. In mice, IgG1 is associated with a Th2-like response, while a Th1 response is associated with the induction of IgG2a (23). The IgG1 and IgG2a titers to LcrV196 were increased at 2 and 3 weeks postimmunization (Fig. (Fig.5C)5C) and were significantly higher than those of χ9447(pYA4534)-vaccinated mice (P < 0.001). At 2 and 3 weeks postimmunization, the IgG2a/IgG1 values were 0.89 and 1.79, respectively, indicating an initial IgG1 response that was slightly stronger than the IgG2a response, which switched to a slight IgG2a bias by week 3 and postchallenge (see below). These data indicate that the RASV synthesizing LcrV196 induced both Th2-related IgG1 and Th1-related IgG2a antibody responses. Anti-LcrV196 IgG1 and IgG2a were barely detected in sera obtained from mice vaccinated with control strain χ9447(pYA4534) or in preimmune serum.
IgA antibodies are the principal mucosal antibody class produced by local plasma cells and are the first line of immune defense against infection (30). Moderate secretory IgA titers against LcrV196 were detected in vaginal fluids from mice immunized with χ9447(pYA4535) synthesizing LcrV196 on week 2 postimmunization and showed a significant increase at week 3 postimmunization compared to the control strain or preimmune serum (P < 0.05) (Fig. (Fig.5D).5D). These results indicate that LcrV196 delivered by strain χ9447(pYA4535) promoted secretory antibody LcrV-specific titers against LcrV196.
The serum immune response to LcrV196 was examined by measuring the levels of total IgG isotype and subclasses IgG1 and IgG2a in surviving mice after s.c. or i.n. challenge. The total IgG titer was higher in mice post-s.c. challenged mice than in orally immunized mice (P < 0.05) (Fig. (Fig.5A).5A). In contrast, the total IgG titers were similar between the post-i.n. challenged mice and the orally immunized mice (Fig. (Fig.5A).5A). The total IgG titers from post-s.c. challenged mice were significantly elevated compared with those of post-i.n. challenged mice (P < 0.05) (Fig. (Fig.5A).5A). The IgG2a-to-IgG1 ratios for mice after s.c. and i.n. challenge were similar. The slight IgG2a bias observed for immunized mice 3 weeks after vaccination was maintained with the surviving mice challenged by the s.c. and i.n. routes (P < 0.001) (Fig. (Fig.5C).5C). These results indicated that a mixed Th1/Th2 response was developed over time in the mice surviving lethal Y. pestis challenge. The total IgG-to-SOMP ratio was not significantly different between mice immunized with χ9447(pYA4534) at 3 weeks after vaccination and the post-s.c. or post-i.n. challenged mice (Fig. (Fig.5B).5B). In addition, the anti-SOMP IgG levels between mice that survived the challenge by the i.n. route with high and low doses were similar. In contrast, slightly lower IgG-to-SOMP ratios were detected for mice that survived the challenge with HD by the s.c. route (P < 0.05) than for mice that survived the challenge with an LD by the s.c. route (Fig. (Fig.5C5C).
Antigen-specific T-cell cytokines were evaluated by ELISPOT assays and ELISA. To compare stimulations of the production of the proinflammatory Th1 cytokine IFN-γ and the anti-inflammatory Th2 cytokines IL-4 and IL-10 (54), cultured lymphocytes from spleens isolated from χ9447(pYA4535)-immunized BALB/c mice and control mice at 3 weeks after the last immunization were assessed by cytokine ELISPOT assay. Lymphocytes were restimulated for 24 h (for IFN-γ) or 48 h (for IL-4 and IL-10) with 1 μg/well of recombinant LcrV or medium. Antigen-specific Th1 and Th2 cells were induced in the spleen of mice orally vaccinated with recombinant LcrV196 delivered by the χ9447(pYA4535) vaccine. The number of LcrV196-specific IFN-γ spot-forming units (SFU) from splenocytes of mice immunized with strain χ9447(pYA4535) was significantly higher than the IFN-γ SFU from splenocytes from mice vaccinated with vector control strain χ9447(pYA4534) (P < 0.001) (Fig. (Fig.6A).6A). The number of LcrV196-specific IL-10-producing cells from mice immunized with strain χ9447(pYA4535) was significantly higher than the number of cells producing IL-10 from mice vaccinated with χ9447(pYA4534) (P < 0.05) (Fig. (Fig.6B).6B). The number of LcrV196-specific IL-4 SFU from splenocytes of mice immunized with strain χ9447(pYA4535) was significantly higher than the IL-4 SFU from splenocytes of mice vaccinated with χ9447(pYA4534) (P < 0.05) (Fig. (Fig.6C).6C). These results indicate that antigen-specific IFN-γ, IL-4, and IL-10 cytokine-forming T cells were elicited by χ9447(pYA4535) and reveal that a mixed Th1/Th2 immune response was elicited in immunized mice.
An ELISA was used to quantify LcrV196-specific IFN-γ, IL-4, and IL-10 cytokines secreted by the splenocytes upon antigen stimulation. At 3 weeks after the last immunization, 5 × 106 cultured lymphocytes from spleens isolated from χ9447(pYA4535)-immunized BALB/c mice and control mice were restimulated with 10 μg recombinant LcrV or medium for 48 h. Afterwards, the cytokines secreted by the splenocytes into the supernatant were quantified. The levels of LcrV196-specific Th1 cytokine IFN-γ and Th2 cytokine IL-10 were significantly higher in mice vaccinated with χ9447(pYA4535) (P < 0.01) than in mice immunized with vector control strain χ9447(pYA4534) (Fig. 6D and E). The levels of the Th2 IL-4 cytokine were within the detection limits (<5 pg/ml) in the supernatants of cultured lymphocytes from all groups of mice (data not shown). The low IL-4 levels detected in the supernatant could be the result of the uptake of IL-4 by the cells, as was previously described (20). These results confirmed that the oral vaccination of mice with a Salmonella strain synthesizing LcrV196 from pYA4535 induced LcrV196-specific systemic proinflammatory IFN-γ and anti-inflammatory cytokine IL-10, consistent with our results from the ELISPOT analysis.
The effect of a RASV synthesizing LcrV196 was evaluated for the stimulation of T-cell responses. CD4+ and CD8+ T cells are central components of the acquired immune system and may contribute to successful vaccination (56). At 28 days after the last immunization, spleen cells were isolated from Salmonella-immunized BALB/c mice and control mice, and the percentage of T cells (CD4+ and CD8+) from these spleen cells was determined by FACS analysis. A significant increase (P < 0.0001) in the percentages of both CD4+ and CD8+ T cells was observed for both groups of mice immunized with χ9447(pYA4535) (23.7% CD4+ and 8.2% CD8+ T cells) and control χ9447(pYA4534) (23.1% CD4+ and 8.7% CD8+ T cells) compared to the percentages from naïve mice (17.1% CD4+ and 7.1% CD8+ T cells). However, Salmonella delivering LcrV196 did not promote a significant increase in percentages of these T cells compared to the percentages from control mice immunized with χ9447(pYA4534). This result suggested that T-cell immune responses are induced against Salmonella, which is an intracellular bacterium that is taken up by antigen-presenting cells, which lead to major histocompatibility complex (MHC) presentation of the bacterial antigens (42). However, with these experiments, we cannot evaluate an overlapping T-cell immune response induced against both LcrV and Salmonella antigens, which might both provide protection against Yersinia infection.
To test the protective efficacy of the χ9447 strain expressing lcrV196 from pYA4535, BALB/c mice were orally immunized with χ9447(pYA4535), χ9447(pYA4534), or BSG on days 0 and 9 as described above. At 28 days after the second immunization, mice were challenged s.c. with 4.49 × 102 CFU (LD) or 5.63 × 103 CFU (HD) of virulent Y. pestis CO92. All nonimmunized mice (BSG control) succumbed to infection by day 7 postchallenge (Fig. (Fig.7).7). RASV strain χ9447(pYA4535) synthesizing LcrV196 provided the greatest efficacy, with 87.5% survival postchallenge with LD and HD of Y. pestis CO92. Mice vaccinated with the χ9447(pYA4534) vector control and challenged with an HD of virulent Y. pestis had a median survival time of 4 days postchallenge, and all died by day 11 postchallenge. The mortality rate was lower for mice vaccinated with the χ9447(pYA4534) vector control and challenged with an LD of virulent Y. pestis. Although these results confirm previous results (4), future experiments will be needed to distinguish between Salmonella-induced innate immunity and the induction of cross-protective immunity. In this regard, no protection was afforded by vaccination with BSG after an LD challenge.
Since strain χ9447(pYA4535) induced protective immunity against bubonic plague, we tested if it would protect against the more lethal pneumonic plague. BALB/c mice were orally immunized as described above and challenged 28 days after the second immunization i.n. with 4.1 × 103 CFU (LD) or 4.4 × 104 CFU (HD) of Y. pestis CO92 (Fig. (Fig.8).8). χ9447(pYA4535) showed great efficacy, with 75% and 50% survival post-i.n. challenge with HD and LD, respectively. These results correlate with the slightly higher LcrV antigen-specific IgG, IgG1, and IgG2a responses observed for immunized mice challenged with an HD of Y. pestis than those for mice challenged with an LD (Fig. (Fig.5A),5A), which could account for the protection obtained. None of the mice from the BSG or vector control groups were protected, and all died by day 5 postchallenge. These results show that our RASV strain, χ9447(pYA4535), synthesizing LcrV196 was protective against bubonic and pneumonic plague, consistent with the immune responses observed.
In this study we designed and developed a versatile vaccine vector that combined several strategies to make the synthesis of antigens proportional to the gene dosage based on a balanced-lethal plasmid expression system with conditional runaway-like replication.
Furthermore, we used a new live attenuated Salmonella enterica serovar Typhimurium χ9447 strain, which is phenotypically wild type at the time of infection and gradually manifests an attenuated phenotype during colonization as a result of the decrease in the production of Fur, Crp, and MurA, whose genes are also under the regulation of araC PBAD activator-promoter cassettes.
We used the multifunctional protein LcrV of Y. pestis to evaluate the runaway-like replication vaccine vector because it is one of the major antigens known to afford protection against plague. RASV strain χ9447(pYA4534) (empty vector control) or χ9447(pYA4535) synthesizing Y. pestis LcrV196 stably maintained the plasmids over 50 generations or more when cultured onto LB agar plates (0.2% arabinose, 0.2% mannose, and 50 μg/ml X-gal) with or without DAP in vitro. Also, growth was not affected by the synthesis of LcrV196, as similar growth rates consistent with the change in CFU were observed for both strains χ9447(pYA4534) and χ9447(pYA4535).
We could isolate χ9447(pYA4535) from Peyer's patches and mesenteric lymph nodes, but not the spleen, during the first 10 days after the oral inoculation of mice. Mice immunized orally on days 0 and 9 with live χ9447 delivering the LcrV196 antigen encoded on runaway-like replication vector pYA4535 elicited a mixed Th1/Th2 antibody response that was protective against lethal Y. pestis CO92 challenge in both bubonic and pneumonic plague models.
Previous work using recombinant Salmonella required 5 immunizations by the intragastric route with 109 CFU on days 0, 14, 28, 42, and 56 of S. enterica serovar Typhimurium SL3261 aroA expressing full-length lcrV from pTR-LcrV, an antibiotic resistance-based vector (pBR322 derivative) to obtain limited success, affording 30% protection after s.c. Y. pestis strain GB challenge (22). The SL3261(pTR-LcrV) immunization elicited predominantly an IgG2a (Th1) immune response. However, the protected mice did not show significantly higher LcrV-specific IgG1 or IgG2a responses than the nonprotected mice.
Despite the fact that Liu et al. (41) observed that RASV χ8501 (hisG Δcrp-28 ΔasdA16) delivering full-length flanked BlaSS-LcrV encoded on pYA3495 (AsdA+ pBR derivative) to mice immunized by the i.n. route three times (at 2-week intervals between each boost) elicited comparable levels of LcrV-specific IgG1 (Th2) and IgG2a (Th1) and offered 60% protection after intraperitoneal Y. pestis strain Yokohama-R challenge, they observed only 20% protection for mice immunized by a combination of oral priming with two subsequent i.n. boosts. However, Branger et al. (4) showed that RASV χ8501 delivering a truncated-flanked BlaSS-LcrV196-BlaCT encoded on pYA3841 (AsdA+ pBR derived) to mice immunized orally on days 0 and 9 elicited an LcrV-specific IgG (Th1/Th2) response that afforded strong protection against Y. pestis CO92 challenge by the s.c. route. Moreover, Branger et al. (5) observed that χ8501(pYA3841)-immunized mice were also protected against Yersinia enterocolitica infection. In addition, the RASV χ8501 strain itself harboring the vector control offered protection against Yersinia pseudotuberculosis infection. These results clearly demonstrated that the ratios of LcrV-specific IgG1 (Th2) and IgG2a (Th1) immune responses developed in the Salmonella-immunized mice appeared to be different depending on the plasmid expression system, immunization route, immunization regimen, length of LcrV, and attenuated Salmonella strain.
Several studies demonstrated the essential role of the humoral immune response in protection against plague. Interestingly, in the absence of protective antibodies or optimal antibody-mediated protection, Th1 cytokines such as TNF-α and IFN-γ also play a role in protection against plague (37). IFN-γ and TNF-α promote the phagocytosis of bacilli, which activates the production of reactive oxygen and nitrogen to kill internalized bacteria. Thus, the ideal plague vaccine should achieve protection by both humoral and cellular immune responses (37).
The level of LcrV196-specific IgG in the serum of all mice immunized with RASV χ9447(pYA4535) is consistent with the protective immune response against Y. pestis challenge, and secretory IgA should also have a role in protection against pneumonic plague. The mixed IgG2a/IgG1 humoral immune response elicited by χ9447(pYA4535) in mice immunized orally offered protection against both bubonic and pneumonic plague, and this was supported by the secretion of the LcrV-specific Th1 IFN-γ and Th2 IL-10 and IL-4 cytokines, which correlated with the increased level of CD4+ and CD8+ T cells detected in immunized mice compared with unvaccinated mice. These results indicate that the T-cell immune response against Salmonella itself afforded partial protection of mice inoculated with the χ9447(pYA4534) vector control and that χ9447(pYA4535) elicited an LcrV humoral response along with specific LcrV cell-mediated immunity in immunized mice, resulting in even stronger protection against Y. pestis. In contrast, in parallel studies with the Y. pestis PsaA antigen and using the same schedule of immunization and challenge with LD and HD Y. pestis, mice orally immunized with χ9558 (24) expressing optimized psaA from pYA3342 (AsdA+ pBR derivative) exhibited limited protective immunity against Y. pestis strain CO92 by i.n. challenge, and all mice succumbed to both LD and HD s.c. challenges (A. Torres-Escobar, M. D. Juárez-Rodríguez, C. B. Branger, and R. Curtiss III, submitted for publication).
Our results showed that χ9447 harboring the balanced-lethal plasmid expression system with conditional runaway replication offers a fine-tuned synthesis of LcrV196 regulated by arabinose, which is absent in host tissues. Our Salmonella vector-host was fully capable of increasing the synthesis of LcrV196 in vivo to induce a robust immune response with 75% protection against an i.n. lethal challenge with 4.4 × 104 CFU of Y. pestis CO92 and even higher protection (87.5%) against a lethal s.c. challenge compared to previous work using strains with the AsdA+ pYA3841 vector system (4, 5). This demonstrates that this system has all the features required for the delivery of antigens that can elicit protective responses to pathogens displaying those antigens. The LcrV protein is known to act as an immunosuppressive agent, inducing IL-10 in host immune cells via interactions with TLR-2/TLR-6/CD14 and suppressing proinflammatory cytokines such as IL-12p40, IFN-γ, and TNF-α, which are required to limit the pathogenesis of plague. In this study we used a fragment of LcrV fused with the β-lactamase signal and carboxy-terminal sequences for its secretion by the RASV strain. This LcrV196 antigen was able to induce a robust immune response and the specific production of IL-4 and IL-10 as well as an increased production of IFN-γ by T cells in the LcrV restimulation assays. The high level of synthesis of this antigen in χ9447(pYA4535) and its efficient secretion could be necessary for the induction of this potent immune response that includes the development of T-cell immunity and antibody responses effective against bubonic and pneumonic plague. Several other groups reported a similar or slightly better protection against a similar or higher Y. pestis challenge dose by both subcutaneous and intranasal routes using the purified LcrV protein for immunization or a DNA vaccine specifying the expression of the LcrV antigen (49, 65). However, our system developed in this work confers greater protection than other live attenuated vaccine strains. Future modification of the properties of LcrV delivery by the χ9447 host could offer better protection and at lower cost than immunization with the purified LcrV protein. Moreover, this host-plasmid system can be modified and improved to delay the delivery of antigen to achieve a higher-level colonization of host tissues, thus stimulating a more potent immune response.
This research was supported by National Institutes of Health grant AI057885.
We thank Clara Espitia (UNAM, México) for her critical reading of the manuscript.
Editor: J. B. Bliska
Published ahead of print on 22 March 2010.