Immunogenicity against Human Papillomavirus Type 16 Virus-Like Particles Is Strongly Enhanced by the PhoP^c Phenotype in Salmonella enterica Serovar Typhimurium

Plasmid constructs and bacterial strains used.We have deleted the asd gene from both the PhoP^c mutant, CS022 (29) (a kind gift from John Mekalanos, Boston, Mass.), and the aroA mutant, SL7207 (22) (a kind gift from Irene Corthésy-Theulaz, Lausanne, Switzerland), by P22HTint transduction, yielding Δasd derivatives denoted GL01 (37) and GL04, respectively. Diaminopimelic acid-requiring tetracycline-resistant transductants (42) were purified, and tetracycline-sensitive derivatives were obtained from these transductants by fusaric acid selection (27, 42). Stable expression of HPV16 L1 and PhoQ24 was achieved in these backgrounds with the aspartate β-semialdehyde dehydrogenase balanced-lethal vector-host system (9). To this end, the NcoI-HindIII fragment, encoding HPV16 L1, of plasmid pFS14nsdHPV16-L1 (33) was inserted downstream from the trc promoter into the NcoI and HindIII sites of a medium-copy-number asd- plasmid (pYA3342). In the resulting plasmid, pYA3342HPV16L1, either an XbaI fragment carrying phoPQ24 or an HindIII fragment carrying the phoQ24 open reading frame, including a Shine-Dalgarno (SD) sequence, was inserted. These fragments were generated by PCR performed on the DNA of S. enterica serovar Typhimurium strain CS022 as a template. For phoPQ24, the following primers were used: a 26-mer located 144 nucleotides upstream from the ATG of phoP (45) and containing a XbaI site (underlined), 5′-GGGTCTAGACTGGTCGACGAACTTAA-3′, and a 66-mer containing another XbaI site (underlined) and a t2 terminator (in italics), 5′-GGGTCTAGAAAAAGGCCATCCGTCAGGATGGCCTTCTATGTTAAGTATCCGCAGGCTGGTATCTGA-3′. For the HindIII SDphoQ24 fragment, the primers used were as follows: a 40-mer containing a synthetic SD sequence (in italics), 5′-GGGAAGCTTG AGGAAAAGCTAATGAATAAATTTGCTCGCC-3′, and a 25-mer including a HindIII site (underlined), 5′-GGGAAGCTTGAAATGTTTATTCCTC-3′. The initiation and stop codons are indicated in bold type. The PCR-amplified XbaI phoPQ24 or HindIII SDphoQ24 fragments were cloned in the XbaI or HindIII sites of the pYA3342HPV16L1 plasmid, yielding plasmids pYA3342HPV16L1-PhoPQ24 and pYA3342HPV16L1-SDPhoQ24, respectively. The different plasmid DNAs were introduced into the attenuated S. enterica serovar Typhimurium strains χ⁴⁵⁵⁰ (Δcya Δcrp Δasd [43]), GL04, and GL01 by electroporation as previously described (44). Strains ATCC14028 and CS015 (PhoP⁻ [28]) were a kind gift from John Mekalanos. Table 1 summarizes the different strains and abbreviations used in this study.

Analysis of pagN promoter activity.A pagN::lacZ indicator plasmid was first generated. The promoter region of the pagN gene was cloned and identified by cleavage of a PCR fragment containing the S. enterica serovar Typhimurium ATCC 14028 pagN coding sequence (15) and upstream sequences with the enzyme SspI. The resulting 310-bp fragment was inserted into the SmaI site of the transposon Tn 3-based vector pGMK1337 (25), enabling activation of β-galactosidase expression. The plasmid pGMK1338, bearing a 310-bp SspI pagN promoter fragment, and a control plasmid, pGMK1339, bearing an unregulated weak promoter sequence located upstream of the pagN gene, were constructed. To facilitate the analyses, the transposable sequences were integrated into stringently replicating plasmid R751, yielding R751::TnP pagN::folA::lacZ or R751::TnP::folA::lacZ. These plasmids were introduced into the recombinant Salmonella strains by conjugation. The Salmonella exconjugants were selected for resistance to ampicillin (conferred by the conjugative plasmid) and growth on minimal medium. Specific positive agglutination with Salmonella O antiserum group B (Difco) was used to confirm the identity of the exconjugants. β-Galactosidase activity assays were performed as described by Miller (30). Portions (100 μl) of the bacterial cultures were mixed with 900 μl of Z-buffer (0.06 M Na₂HPO₄ · 7H₂O, 0.04 M NaH₂PO₄ · H₂O, 0.01 M KCl, 0.001 M MgSO₄ · 7H₂O, 0.05 M β-mercaptoethanol [pH 7]) containing 0.1% sodium dodecyl sulfate (SDS) and 1% chloroform. A 200-μl volum of o-nitrophenyl-β-d-galactoside (ONPG, 4 mg/ml in Z-buffer [Sigma]) was then added as substrate, and the optical density (OD) was recorded at both 420 and 550 nm. The light-scattering correction factor (CF = OD₄₂₀/OD₅₅₀) was first determined for each Salmonella strain in Luria-Bertani (LB) broth and then the β-galactosidase activity was calculated using Miller's formula: Ui β-Galactosidase = 1,000OD₄₂₀ − (CF × OD₅₅₀)/t × ν × OD₆₀₀, where t is time of the reaction in minutes and v is the volume of the culture used in the assay in milliliters.

PMB resistance assay.PMB resistance was assessed by using a method modified from those of Gunn et al. (14) and Roland et al (40). Briefly, bacterial strains were grown to mid-log phase and diluted to a concentration of approximately 10⁴ CFU/ml in tryptone saline. Cells (100 μl/well) were mixed in a microtiter plate with various concentrations of PMB (Sigma) and incubated at 37°C for 1 h. Then 100 μl of PMB-treated cells was plated on LB agar medium and the relative strain resistances (compared with untreated cells) were determined by CFU counting. The PMB resistance best-fitting sigmoid curves were drawn, and 50% inhibitory concentration were calculated with GraphPad Prism.

Nonspecific acid phosphatase (PhoN) activity assay.PhoN activity was assessed using a method modified from that of Kier et al. (24). Briefly, Salmonella strains were grown overnight, diluted in LB broth at 1:100, and grown to an OD₆₀₀ of 0.6. The bacteria were pelleted and resuspended in 1 ml of 1 M Tris (pH 8.0). The cells were mixed with 200 μl of 0.4% p-nitrophenyl phosphate (Sigma) in 1 M Tris (pH 8) and incubated at 37°C until a yellow color appeared; the reaction was then stopped by adding 200 μl of 1 M K₂HPO₄. The PhoN activity was measured in arbitrary units at OD_420/550 by using the same equation used for the β-galactosidase activity.

Protein analysis.Recombinant Salmonella cells from an exponential-phase culture were lysed in 2.5% SDS. After 30 min at room temperature, the bacteria were boiled for 10 min in a final solution of 3% SDS-2 mM EDTA-50 mM Tris-HCl (pH 7.0)-8% glycerol-1% β-mercaptoethanol-0.1% bromophenol blue. Bacterial lysates were separated on SDS-10% polyacrylamide gels. Expression of L1 in the Salmonella lysates, normalized to the OD₆₀₀ of the cultures, was analyzed by Western blotting as previously described (33), using the anti-HPV16 L1 monoclonal antibody, CAMVIR-1 (Anawa).

Immunization of mice, analysis of the immune response, and recovery of S. enterica serovar Typhimurium.Six-week-old female BALB/c mice were used in all experiments. A 20-μl volume of bacterial inoculum was administered intranasally under anesthesia as described previously (23, 33). For the inoculum, bacteria were grown to mid-log phase and diluted to ca. 10⁷ CFU/20 μl. Sampling of blood, as well as determination of anti-lipopolysaccharide (LPS) and anti-HPV16 VLP titers by enzyme-linked immunosorbent assay were performed as previously described (23, 33). A good correlation was shown between HPV16 VLP-specific antibody titers and HPV16 neutralization (33, 34, 39). Recovery of S. enterica serovar Typhimurium was determined in organs from euthanized mice as previously described (33).

Statistics.Comparisons of the different data were made by one-way analysis of variance and a Bonferoni post-test using GraphPad Prism.

Extrachromosomal expression of phoPQ24 activates the pag gene phoN in both χ4550 and GL04.The mutant phoQ gene, phoQ24, was inserted into plasmid pYA3342-HPV16L1, already carrying the HPV16 L1 gene and the asd gene but no antibiotic resistance marker. The two plasmids constructed encode either the complete phoPQm operon, including the promoter-operator, or only the phoQ24 coding sequence, including an SD sequence. Both constructs were introduced into the attenuated S. enterica serovar Typhimurium strains, χ4550 (Δcya Δcrp Δasd [43]) and GL04 (ΔaroA Δasd [Table 1]). In the resulting S. enterica serovar Typhimurium recombinants, plasmid-located phoQ24 is expected to be dominant over the chromosomal phoQ and thus to induce the PhoP^c phenotype. The phenotypes of strains bearing plasmid-located phoQ24 and those in which the full phoPQ24 operon was present were anticipated to be different. In the former, the global regulatory protein PhoP will be mostly in the activated state due to unregulated phosphorylation by PhoQ24, whereas in the latter, overphosphorylation of PhoP is expected to be accompanied by higher expression of the phoPQ24 operon as a result of a positive autoregulatory loop (45). However, in both strains, the expression of the pag and prg genes is expected to be deregulated in comparison to that in strains harboring the wild-type phoQ.

One of the aspects of the complex PhoP^c phenotype is its high phosphatase activity, which helps the identification of this strain on solid media as dark blue colonies when 5-bromo-4-chloro-3-indolylphosphate (BCIP) is added to the medium (29). Indeed, both strains χ4550 and GL04 carrying the plasmid containing phoPQ24 generated dark blue colonies on BCIP-containing plates. To our surprise, however, we could not distinguish χ4550 and GLO4 recombinants, harboring the plasmid containing phoQ24 from those bearing pYA3342-HPV16L1 as controls, since they all generated light blue colonies on BCIP-containing plates. Determination of the non-acid phosphatase PhoN activities (24) in all these strains further confirmed these observations (Table 2). PhoN activities were similar in the three strains harboring a wild-type phoPQ operon (33U, 31U, and 27U in ATCC14028, χ4550 L1, and GL04 L1, respectively) and plasmid-located phoQ24 (22U and 25U in χ4550 L1-SDQm and GL04 L1-SDQm, respectively). In contrast, PhoN activity was 5- and 8-fold higher in χ4550L1-PQm and GL04L1-PQm (P < 0.05 for 146U and P < 0.001 for 214U), while even higher activation of PhoN (17-fold [p < 0.001]) was observed in the PhoP^c strain compared to the wild-type strain. These data show that extrachromosomal expression of phoPQ24, but not phoQ24 activates phoN.

Activation of the pagN promoter in different phoQ24 backgrounds.An indirect functional analysis of phoQ24 was performed by introduction of a plasmid (PpagN-lacZ) in which lacZ is expressed under the control of the phoPQ-regulated pagN promoter (15). A PhoP^c strain bearing the PpagN-folA::lacZ plasmid exhibited 20-fold-higher β-galactosidase activity than did the same strain bearing a plasmid constitutively expressing low levels of LacZ (control-lacZ) (Fig. 1); β-galactosidase activities in PhoP⁻ recombinants invariably remained low (data not shown). When the same plasmid was introduced into χ4550L1-SDQ24 or χ4550L1-PQ24, β-galactosidase activities were 12- and 13-fold higher, respectively, than in the control strains that expressed lacZ from a phoPQ-independent promoter. Most importantly, a significant increase (five- to sevenfold [fig. 1]) in the β-galactosidase activity could be clearly attributed to phoQ24 expression. Interestingly, and in contrast to what we had observed in the regulation of phoN, activation of pagN was similar in χ4550 L1-SDQ24 and χ4550 L1-PQ24, suggesting that different threshold levels of the activated PhoP are necessary to turn on the expression of different pag genes.

• Open in new tab
• Download powerpoint

FIG. 1.

Extrachromosomal expression of PhoQ24 is able to activate the pagN promoter. β-Galactosidase activity was measured in the different conjugative strains, as indicated in Materials and Methods. Mean β-galactosidase activities in three independent experiment are shown for each conjugative strain. Black and white bars indicate strains containing plasmids encoding β-galactosidase driven by the pagN promoter and the control promoter, respectively. SEM, standard error of the mean.

Expression of phoQ24 or phoPQ24 in S. enterica serovar Typhimurium χ4550 and GL04 confers resistance to PMB.We next examined whether extrachromosomal expression of phoQ24 or phoPQ24 could influence the expression of another endogenous PhoPQ-regulated gene. The pmrAB locus, a pag gene responsible for the PMB resistance (14), was chosen for this assay. The 50% inhibitory concentrations of PMB in different S. enterica serovar Typhimurium strains were determined (Table 3). As expected, the PhoP^c strain exhibited a significant increase of PMB resistance (sevenfold) compared to the wild-type ATCC 14028 (P < 0.001) whereas the PhoP⁻ strain was more sensitive to PMB than was the wild-type strain (P < 0.001), which is consistent with the inactivation of the PhoPQ regulatory system in this strain (11, 31). Interestingly, a significant increase in PMB resistance was conferred on χ4550 and GL04 harboring either the L1-SDQ24 or L1-PQ24 plasmid (7.5- and 17.5-fold, respectively, for χ4550 and 4.4- and 6.2-fold, respectively, for GL04 [P < 0.001]). In addition, similar to phoN, a higher activation of pmrAB was induced when phoPQ24 was expressed in both strains χ4550 and GL04 (P < 0.01).

Extrachromosomal expression of phoPQ24, but not phoQ24, altered the ability of χ4550 and GL04 to invade and persist in the mouse.Combinations of Δcya Δcrp mutations (χ4550) with mutations in the phoPQ locus have not yet been evaluated in animal models, while a double PhoP^c ΔaroA mutant was shown to be 100-fold more attenuated than the single ΔaroA mutant (32). It was therefore important to determine whether expression of PhoQ24 in a strain that already harbored attenuating mutations would not cut short the infectious pathway in the mouse. This might result in poor antigen presentation due to cumulative effect of attenuating mutations on mouse virulence. For this purpose, we intranasally immunized groups of mice with the different recombinant Salmonella strains. The mice were euthanized 2 weeks after immunization, and recombinant Salmonella organisms were recovered from representative sites of bacterial invasion and persistence, i.e., lungs, cervical lymph nodes, Peyer's patches, and spleen (33). While the two strains harboring the phoPQ24-containing plasmid showed a dramatic decrease in survival and/or replication in all organs examined (Fig. 2), only a slight decrease was observed in the lungs when these strains were harboring the phoQ24-containing plasmid, suggesting that phoPQ24, but not phoQ24, had deleterious effects on invasiveness and persistence of the Δcya Δcrp and ΔaroA Salmonella strains in vivo.

• Open in new tab
• Download powerpoint

FIG. 2.

Recovery of S. typhimurium χ4550 (A) and GL04 (B) 2 weeks after intranasal vaccination. Groups of three to five 6-week-old BALB/c mice were immunized with ca. 10⁷ CFU of χ4550 or GL04 recombinant strains. The different organs were prepared and plated on agar as previously described (33). Data are expressed as the geometric mean (log₁₀) CFU per organ. The level of detection (20 CFU/organ, or log₁₀ 1.3) is indicated by a horizontal dashed line. Error bars indicate the standard error of the means.

Expression of phoQ24 in both χ4550 and GL04 L1-expressing strains confers immunogenicity against HPV16 L1 VLPs.The aim of this study was to test whether the expression of the phoQ24 gene in S. enterica serovar Typhimurium backgrounds that were previously shown to be poorly immunogenic would improve the immunogenicity of the HPV16VLP antigens. Although successfully used to induce protective immunities against a number of other heterologous antigens, cya crp, and aroA mutants yielded only barely detectable antibody responses to the HPV16 L1 protein (5; J. Benyacoub, unpublished results). Strains χ4550 and GL04 harbor a chromosomal deletion that eliminates the asd gene. This gene is essential for the growth of bacteria on unsupplemented media, but its defect can be complemented in trans by using a plasmid. This creates a balanced lethal state that can be exploited to ensure stable plasmid inheritance. Since we had previously encountered serious problems in the maintenance of L1-encoding plasmids in χ4989 and other isogenic strains (5), we decided to use this system to evaluate the role of phoQ24 in the immunogenicity of HPV16 VLPs delivered by live Salmonella strains in the mouse. The levels of expression of HPV16L1 were examined in all the strains to determine whether addition of phoQ24 to the L1-encoding plasmid would alter L1 expression. This did not turn out to be the case, as illustrated in Fig. 3. To test the immunogenicity of the VPLs produced by S. enterica serovar Typhimurium recombinants, mice were intranasally vaccinated twice with the L1-expressing χ4550 and GL04 recombinants and specific antibody responses were measured in serum 4 weeks after the second immunization. As expected, χ4550 L1 and GL04 L1 induced low or barely detectable HPV16 VLP antibodies. In contrast, expression of phoQ24 in these backgrounds induced a significant increase of HPV16 VLP immunoglobulin G (IgG) titers in both χ4550 L1-SDQ24 and GL04 L1-SDQ24 (P < 0.05 and P < 0.001, respectively [Fig. 4 ]). In contrast to this, and in agreement with the observed penetration and survival of Salmonella recombinants in vivo, bacteria expressing phoPQ24 induced only very low antibody titers against HPV16 VLP, the heterologous antigen, and against bacterial LPS. This supported our hypothesis that these recombinants were overattenuated (Fig. 4). There was no statistical difference in the anti-LPS responses induced by the other four strains (Fig. 4B), suggesting similar infection of the mice. We confirmed by a sandwich ELISA (14) that in agreement with the Western blot analysis, no difference in VLP assembly occurred whether PhoQ24 was expressed or not (7.0 ± 1.3 and 5.1 ± 0.4 μg of VLP/10¹¹ CFU for χ4550 L1 and χ4550 L1-SDQ24, respectively; 5.6 ± 0.3 and 7.3 ± 1.5 μg of VLP/10¹¹ CFU for GL04 L1 and GL04 L1-SDQ24, respectively). In addition, we did not find any statistical difference in the anti-HPV16 VLP IgG titers induced by GL01, a PhoP^c Δasd strain (37), harboring either the L1 gene or the L1-phoQ24-carrying plasmid (anti-HPV16 VLP IgG log₁₀ titers of 2.42 ± 0.34 and 2.84 ± 0.22 [mean ± standard error of the mean] respectively). Taken together, our data demonstrate that a relatively low expression level of PhoQ24 can be combined with attenuations in Δaro, as well as in Δcya and Δcrp, although to a lesser extent, to induce immunogenicity against the HPV16 VLP antigen.

• Open in new tab
• Download powerpoint

FIG. 3.

HPV16 L1 expression in the χ4550 (A) and GL04 (B) recombinant strains. Scanning of the L1 protein bands obtained after Western blotting (four independent experiments) of the bacterial lysates from the different recombinant strains was performed by using NIH Image software. The results are shown as the means of the pixel densities of the L1 proteins bands normalized to the content in bacteria (OD₆₀₀ of the culture) and are expressed as fold increase in comparison to the strains harboring the plasmid-encoded L1.

• Open in new tab
• Download powerpoint

FIG. 4.

Anti-HPV16 VLP (A) and anti-LPS (B) IgG in the serum of immunized mice. BALB/c mice were intranasally immunized twice with ca. 10⁷ CFU of the different recombinant strains, and serum samples were obtained 4 weeks later. Data are expressed as the geometric means (log₁₀) of the reciprocal dilutions of specific IgG from individual mice. Error bars indicate the standard errors of the means.