A Live Salmonella enterica Serovar Enteritidis Vaccine Allows Serological Differentiation between Vaccinated and Infected Animals

Bacterial strains. S. enterica serovar Enteritidis phage type 4 strain 76Sa88, isolated from a turkey (8), was used in all experiments. MiniTn5lacZ2 insertion mutants of S. enterica serovar Enteritidis 76Sa88 Rif^r were obtained essentially as previously described (7). S. enterica serovar Enteritidis 76Sa88 contains only one gene that codes for flagellin, fliC.

Construction of guaB and fliC deletion mutants.The guaB and fliC genes of S. enterica serovar Enteritidis 76Sa88 were deleted by homologous recombination of PCR fragments, with the Red recombinase system of bacteriophage Lambda (5). The sequences of the primers used are shown in Table 1.

To construct an S. enterica serovar Enteritidis guaB deletion mutant, overlap PCR (13) was performed with two fragments that were amplified, respectively, with primer sets GuaB3-GuaB4 and GuaB5-GuaB2 and with S. enterica serovar Enteritidis genomic DNA as the template. A nested PCR with primer set GuaB6-GuaB7 was subsequently carried out to amplify a linear fragment in which an 861-bp internal segment of the coding sequence of guaB was replaced with an SmaI site. This PCR fragment was cloned into the SmaI site of pUC18 with the Sureclone ligation kit (Pharmacia Biotech). A chloramphenicol resistance gene with flanking FLP recombinase target (FRT) sites, amplified with plasmid pKD3 DNA as the template and primer set P1-P2 (5), was ligated into the SmaI site of the cloned ΔguaB fragment. The PCR fragment (ΔguaB::catFRT) generated on this clone with primer set GuaB6-GuaB7 was introduced by electroporation (32) into S. enterica serovar Enteritidis, previously transformed with pKD46 (5), after induction of the Red recombinase system with 0.2% arabinose.

For the construction of a ΔfliC deletion mutant and a ΔguaB ΔfliC double mutant of S. enterica serovar Enteritidis, a recombinant fragment harboring the kanamycin resistance gene with flanking FRT sites (kanFRT) and extensions homologous to the initial 50 (1 to 50) and the terminal 50 (1468 to 1518) nucleotides of the fliC coding sequence, was amplified with pKD4 plasmid DNA as the template and primers FliCP1 and FliCP2. The ΔfliC mutants were obtained by electroporation of the resulting PCR fragment into S. enterica serovar Enteritidis(pKD46) and S. enterica serovar Enteritidis ΔguaB::catFRT(pKD46) as previously described (5).

P22 transduction.To avoid additional undesirable mutations, the substitution mutations containing the selectable resistance genes were transduced into a wild-type S. enterica serovar Enteritidis 76Sa88 background with bacteriophage P22 HT int (6). Antibiotic resistance genes were subsequently eliminated with pCP20 (5), and the deletions were confirmed by PCR and phenotypic characterization as described further.

Virulence in and protection of mice.Seven-week-old female BALB/c mice were orally inoculated to evaluate the virulence and efficacy as a vaccine of the ΔguaB and ΔguaB ΔfliC deletion mutants. A ΔfliC mutant was also included to study the effect of inactivation of the fliC gene on the virulence of the wild-type strain. For the virulence assay, mice each received about 10⁸ CFU of the bacterial strains that were cultured overnight in Luria-Bertani (LB) broth (25) and suspended in milk. This dose corresponds to approximately 10⁵ times the 50% lethal dose of the wild-type strain (30). As a positive control, mice were inoculated with wild-type pathogenic S. enterica serovar Enteritidis strain 76Sa88. Noninfected control mice were inoculated with milk. The efficacy of the mutants was determined in vaccinated mice 3 weeks after the initial immunization by oral challenge with about 10⁸ CFU of wild-type S. enterica serovar Enteritidis strain 76Sa88 per mouse. The challenged mice were observed for 21 days for death and clinical signs. The animal experiments were performed by following all relevant national and institutional guidelines.

Invasion and phagocytosis assays.Invasion experiments were performed with the human colon carcinoma cell line T84 and with chicken primary cecal epithelial cells as previously described (43). Phagocytosis assays (29) were performed with the chicken macrophage cell line HD11 (2).

Briefly, cells of the human colon carcinoma cell line T84 were seeded in 96-well cell culture plates at a density of 10⁵ cells/well in cell culture medium (Dulbecco modified Eagle medium plus 10% fetal calf serum and 2% l-glutamine without antibiotics) and grown overnight. The bacteria were grown overnight at 37°C in 5 ml of LB broth (25) on a shaker platform, subcultured 1:100 in fresh LB broth (5 ml), and grown to late log phase at 37°C without shaking. After 5 h of incubation, the bacterial suspensions were centrifuged and resuspended in Dulbecco modified Eagle medium with 10% fetal calf serum. The number of CFU per milliliter was determined by plating on brilliant green (BG) agar at 37°C. The suspensions were kept at 4°C until they were used in the assay, diluted to a density corresponding to a multiplicity of infection of 10, and added to the cultured cells. To ensure close contact between the bacteria and the cells, the plates were centrifuged for 10 min at 1,500 rpm and incubated for 1 h at 37°C under 5% CO₂. The cells were subsequently rinsed three times with Hanks’ balanced salt solution (Life Technologies, Paisley, Scotland). Cell culture medium with gentamicin (50 μg/ml) was added, and the plates were incubated for 1 h at 37°C under 5% CO₂. Thereafter, the cells were rinsed three times with phosphate-buffered saline (PBS) and lysed with 1% Triton X-100 (Sigma, St. Louis, MO) in distilled water. From this lysate, a 10-fold dilution series was plated on BG agar plates. Invasion and phagocytosis assays with chicken primary cells and chicken macrophage cells (HD11) were performed essentially as described above, with some modifications (29, 43). All assays were performed in duplicate, with three repeats of each experiment. Percentages of invasion and phagocytosis by defined S. enterica serovar Enteritidis mutants, relative to invasion and phagocytosis by the S. enterica serovar Enteritidis wild-type strain (100%), were calculated. Analysis of variance was performed with SPSS 12.0 software.

Safety assessment with chickens.The safety of the ΔguaB and ΔguaB ΔfliC isogenic mutant strains was evaluated with chickens in two experiments by inoculation of 1-day-old chicks (SPAFAS). In the first experiment, the safety of the S. enterica serovar Enteritidis ΔguaB strain was determined with four groups of 10 1-day-old chicks. Chicks in groups 1 and 2 were inoculated by the intratracheal route and by oral gavage, respectively. Birds in control groups 3 and 4 were inoculated with PBS, respectively, by the intratracheal and oral gavage routes. Mortality was recorded for 38 days.

In the second experiment, the safety of the S. enterica serovar Enteritidis ΔguaB ΔfliC mutant was also evaluated with four groups of 1-day-old chicks. Similar to the first experiment, 10 chicks in groups 1 and 2 were inoculated by the intratracheal route and by oral gavage, respectively. Ten birds in group 3 were inoculated by the intratracheal route with the Poulvac ST aroA S. enterica serovar Typhimurium vaccine (Fort Dodge Animal Health). Five birds in group 4 were inoculated with PBS by the intratracheal route. The body weight of each bird from all four groups was measured at 21 days postinoculation. Body weight was compared among groups in an analysis-of-variance model with body weight as the dependent variable and treatment included as an independent variable. The estimated body weight and its 95% confidence interval were constructed. Group comparisons were made with Tukey's test for multiple comparisons. The frequency distribution of the continuous outcome (body weight) was assessed by PROC UNIVARIATE. All analysis of data was performed with the SAS system (SAS Institute Inc.). The level of significance was set at P < 0.05.

Efficacy assessment with chickens.One-day-old specific-pathogen-free white Leghorn chicks were randomly divided into five groups. Birds in groups 1 and 1A were vaccinated with S. enterica serovar Enteritidis ΔguaB by coarse spray on the first day. At 2 weeks of age, these birds received a second vaccine dose by drinking water (group 1) or by oral gavage (group 1A). Birds in groups 2 and 2A were vaccinated with S. enterica serovar Enteritidis ΔguaB ΔfliC by coarse spray on the first day. At 2 weeks of age, these birds received a second vaccine dose by drinking water (group 2) or by oral gavage (group 2A). Birds in group 3 were administered PBS by spray on the first day.

At 6 weeks of age, the birds from all groups were challenged by oral gavage with virulent S. enterica serovar Enteritidis PT4 (nalidixic acid resistant) and observed for clinical signs and death.

At 7 days postchallenge, all surviving birds were necropsied. Tissue samples of approximately 1 g each of the spleen, kidney, and liver from each bird were obtained aseptically. These were pooled in sterile swirl pack bags containing 10 ml BG tetrathionate broth (BGTB; prepared with BG broth and tetrathionate broth base obtained from Difco Laboratories, Becton, Dickinson and Company, Sparks, MD), macerated for 30 s in a Stomacher blender, and incubated at 42°C for 24 h. Also, 10-mm lengths of the duodenum (bottom of the duodenal loop below the pancreas), jejunum (region of yolk sac diverticulum), and ileum (anterior to the ileocecal junction) were aseptically collected from each bird, flushed internally and externally with sterile PBS, pooled in sterile swirl pack bags containing 10 ml BGTB medium, macerated for 30 s in a Stomacher blender, and incubated at 42°C. After 24 h of incubation, a loopful of culture from each swirl pack bag was streaked onto BG and xylose-lysine-tergitol 4 agar plates (26) supplemented with 100 μg/ml nalidixic acid. The swirl pack bags were incubated further for another 24 h. If Salmonella isolation was negative after 48 h, 1 ml culture from each bag was transferred to a tube containing 10 ml BGTB medium. The tube was incubated for 24 h, and the culture was streaked onto BG and xylose-lysine-tergitol 4 agar plates containing 100 μg/ml nalidixic acid. When S. enterica serovar Enteritidis PT4 grew on either of the plates, the pool was considered positive.

For assessing Salmonella colonization of the cecum, approximately 1 g of cecal content collected from each bird was placed in a swirl pack bag containing 100 ml of sterile PBS and mixed thoroughly. A 0.1-ml sample of the suspension was plated in duplicate on BG agar plates supplemented with 100 μg/ml nalidixic acid. When S. enterica serovar Enteritidis PT4 grew on either of the plates, the pool was considered positive. An agglutination test with group D Salmonella antiserum was performed with at least one colony from all positive plates to confirm the presence of a group D Salmonella.

Data analysis was performed with the SAS system (SAS Institute Inc.). For statistical analysis, the experimental unit was the individual bird. The outcome was Salmonella isolation, and the study tested the null hypothesis that there would be no difference in Salmonella isolation among groups. Salmonella isolation (percentage or number of organisms recovered) was compared between each group and its control group by chi-square analysis. When expected cell sizes were too small, comparisons were made by Fisher's exact test. Group comparisons were made by the bootstrap method (PROC MULTTEST). The relative risk and its 95% confidence interval were constructed for each comparison.

A guaB mutant is attenuated in mice and induces protection.In a set of miniTn5lacZ2 insertion mutants of S. enterica serovar Enteritidis 76Sa88 Rif^r, several auxotrophic mutants were identified. One of these mutants was characterized as a guaB mutant, since it grows on minimal A medium (25) complemented with 0.3 mM guanine, guanosine, xanthine, or xanthosine but not on medium complemented with hypoxanthine or inosine. The guaB gene (47) encodes the enzyme IMP dehydrogenase (EC 1.1.1.205), which converts IMP to XMP, the penultimate step in the biosynthetic pathway of GMP. The insertion of the miniTn5lacZ2 transposon into the guaB gene of the mutant was confirmed by PCR (data not shown).

Three mice at 7 weeks of age were orally inoculated with the guaB insertion mutant by administering 3.5 × 10⁸ CFU per mouse. All three mice remained healthy during the 21-day observation period and were then challenged by oral administration of 2.7 × 10⁸ CFU of wild-type strain S. enterica serovar Enteritidis 76Sa88 per mouse. All three mice survived the challenge. As a positive control, wild-type strain S. enterica serovar Enteritidis 76Sa88 (4.4 × 10⁸ CFU) was orally given to three BALB/c mice of the same age that were not previously immunized. All three mice died within 6 days. These results demonstrate that the guaB mutant is highly attenuated in mice and that oral immunization of mice with this mutant protects them against infection by wild-type S. enterica serovar Enteritidis.

Construction of guaB and fliC deletion mutants.A guaB deletion mutant was constructed to avoid reversion of the miniTn5lacZ2 insertion mutant and to remove the kanamycin resistance gene of miniTn5lacZ2. After electroporation of a linear fragment (containing a ΔguaB::catFRT mutation) in S. enterica serovar Enteritidis, chloramphenicol-resistant S. enterica serovar Enteritidis ΔguaB::catFRT mutants that require supplementation of minimal A medium (25) with 0.3 mM guanine were selected. The replacement of guaB with the chloramphenicol resistance cassette was further confirmed by PCR with primer combinations GuaB6-GuaB7, GuaB6-P2, GuaB7-P1, and P1-P2.

To delete the flagellin gene fliC, an internal 1,416-bp segment (bp 51 to 1467) of the fliC coding sequence (AY649742.1) was replaced with the kanamycin resistance gene in S. enterica serovar Enteritidis(pKD46) and S. enterica serovar Enteritidis 76Sa88 ΔguaB::catFRT(pKD46). By homologous recombination, S. enterica serovar Enteritidis ΔfliC::kanFRT and the double mutant S. enterica serovar Enteritidis ΔguaB::catFRT ΔfliC::kanFRT were generated. Isogenic strains were obtained after P22 transduction of the antibiotic resistance-encoding substitution mutations to a wild-type background. The antibiotic resistance genes were subsequently excised with plasmid pCP20 (5). The colonies were tested for carbenicillin, chloramphenicol, and/or kanamycin sensitivity to ascertain the loss of the plasmid pCP20 and the elimination of the antibiotic resistance genes. Both the S. enterica serovar Enteritidis ΔguaB and S. enterica serovar Enteritidis ΔguaB ΔfliC mutants required guanine (0.3 mM) for growth on minimal A medium. The S. enterica serovar Enteritidis ΔguaB ΔfliC and S. enterica serovar Enteritidis ΔfliC mutants were nonmotile on LB medium containing 0.4% agar. The deletion mutants were confirmed by PCR with primer combinations GuaB6-GuaB7 for guaB and FliC1-FliC2 for fliC. Sequencing (33) of the resulting fragments with primers GuaB10 and FliC3, respectively, confirmed the presence of the P1-FRT-P2 scar sequence.

The ΔguaB ΔfliC double mutant is safe and confers full protection on mice.To study the effect of the deletion of the fliC gene on the immunogenicity of the S. enterica serovar Enteritidis ΔguaB vaccine strain, virulence and protection tests with both mutants were carried out with 7-week-old female BALB/c mice. Two independent experiments were performed (Table 2). Mice infected with the S. enterica serovar Enteritidis ΔguaB mutant showed typical Salmonella disease symptoms (reduced activity, untidy coat, and curved back), and 1 out of 10 died in the first experiment, while no disease symptoms were observed in the second experiment. The ΔguaB ΔfliC double mutant did not induce disease symptoms in both experiments. All mice inoculated with the wild-type S. enterica serovar Enteritidis strain died within 9 days after infection, while the noninfected control mice remained healthy during the observation period of 21 days.

One mouse immunized with the ΔguaB mutant died after challenge with the wild-type S. enterica serovar Enteritidis 76Sa88 strain. All other immunized mice survived the challenge without observable disease symptoms. All nonimmunized control mice died after challenge. These data show that both mutants are attenuated and confer protection against a challenge with the corresponding wild-type parent strain.

Additional mutations, due to unwanted recombination events caused by the Red recombinase, can influence the results. Therefore, the experiment was repeated with isogenic strains that were constructed by P22 transduction.

A ΔfliC mutant was also included to study the effect of inactivation of the fliC gene on the virulence of the prototrophic wild-type strain. This ΔfliC mutant remained as virulent as the wild-type strain under these conditions. Data obtained from the ΔguaB and ΔguaB ΔfliC transductants (Table 3) confirmed the observations made in the first experiments. The ΔguaB ΔfliC double mutant is more attenuated in BALB/c mice and confers better protection against challenges with high doses of the wild-type strain than the ΔguaB mutant.

Immunization with ΔguaB and ΔguaB ΔfliC mutants generates high anti-lipopolysaccharide (LPS) IgG titers.Fifty-four days following initial oral immunization of BALB/c mice with approximately 10⁸ CFU per mouse of, respectively, the S. enterica serovar Enteritidis ΔguaB and S. enterica serovar Enteritidis ΔguaB ΔfliC mutants, blood samples were collected from the tail arteries of five mice. Anti-LPS IgG titers were determined by means of an enzyme-linked immunosorbent assay with the use of 0.5 μg of S. enterica serovar Enteritidis LPS (Sigma) per well for coating. Comparison between sera of mice immunized with S. enterica serovar Enteritidis ΔguaB and mice immunized with S. enterica serovar Enteritidis ΔguaB ΔfliC showed that, in both cases, comparable anti-LPS serum IgG responses were elicited. A second and third boost did not enhance the anti-LPS serum IgG levels (data not shown).

Reduced invasiveness and phagocytosis of S. enterica serovar Enteritidis mutants.The isogenic ΔguaB and ΔfliC single mutants and the ΔguaB ΔfliC double mutant were less invasive than the wild-type S. enterica serovar Enteritidis strain in the human colon carcinoma cell line T84 and in isolated primary chicken cecal epithelial cells (Table 4). Internalization of these mutants in the chicken macrophage cell line HD11 was also reduced compared to that of wild-type strain (Table 4).

The ΔguaB ΔfliC mutant is safe in 1-day-old chicks.Ten 1-day-old chicks were inoculated intratracheally with the S. enterica serovar Enteritidis ΔguaB mutant strain (group 1) at 1.3 × 10⁸ CFU per chick. One bird died during inoculation. Two birds died at 2 days postinoculation, and three birds died, respectively, at days 3, 5, and 13 postinoculation. Out of the 10 birds inoculated with the same dose of this strain by oral gavage (group 2), 1 chick died because of inoculation trauma during oral gavage and 1 bird died at day 5 postinoculation. No birds in the PBS-treated groups (groups 3 and 4) died. These results indicate that the S. enterica serovar Enteritidis ΔguaB mutant strain is not safe in chickens when administered at 1.3 × 10⁸ CFU per bird at 1 day of age by the intratracheal or oral gavage route.

One-day-old chicks that were inoculated with the S. enterica serovar Enteritidis ΔguaB ΔfliC mutant strain at 2.5 × 10⁷ CFU per chick by the intratracheal or oral gavage route survived the 21-day observation period, and no deaths were recorded. Also, chicks inoculated with Poulvac ST (a licensed commercial vaccine), used as an intratracheal procedural control, and chicks inoculated with PBS survived the inoculation. Twenty-one days after inoculation, the average weight and standard deviation of the birds in groups 2, 3, and 4 were, respectively, 0.224 ± 0.018 kg, 0.200 ± 0.019 kg, and 0.205 ± 0.009 kg. The body weights of the chicks inoculated with the S. enterica serovar Enteritidis ΔguaB ΔfliC mutant are not statistically significantly different from the body weights of the chicks vaccinated with the Poulvac ST vaccine or the body weights of the chicks inoculated with PBS. The average body weight (0.181 ± 0.030 kg) of the birds in group 1, which received the vaccine strain by intratracheal administration, was statistically significantly lower (P = 0.0009) than the average body weight of birds that were immunized with the same strain by oral gavage. It can be concluded that S. enterica serovar Enteritidis ΔguaB ΔfliC is safe in 1-day-old chicks after inoculation with 2.5 × 10⁷ CFU per bird by the intratracheal or oral gavage route.

Immunization with the ΔguaB ΔfliC mutant protects chickens against infection of the internal organs by S. enterica serovar Enteritidis.Immunization of chicks with S. enterica serovar Enteritidis ΔguaB and S. enterica serovar Enteritidis ΔguaB ΔfliC and oral challenge with a virulent nalidixic acid-resistant S. enterica serovar Enteritidis PT4 strain were performed as described in Materials and Methods and in Table 5.

The vaccine strains were administered at 1 day of age by coarse spray and at 2 weeks of age by drinking water or oral gavage. The vaccines were deemed safe because no birds died and no adverse effects were observed in the postvaccination period, confirming the results of the previous experiment.

Four weeks post the second vaccination, the birds were challenged with 8.5 × 10⁷ CFU of a virulent nalidixic acid-resistant S. enterica serovar Enteritidis PT4 strain. One week after the challenge, samples of organs (kidney, liver, and spleen), intestines (duodenum, jejunum, and ileum), and cecal contents were tested for the presence of the S. enterica serovar Enteritidis challenge strain. The results summarized in Table 5 show that the vaccine candidates S. enterica serovar Enteritidis ΔguaB and S. enterica serovar Enteritidis ΔguaB ΔfliC protect the organs against infection but do not protect the intestines and ceca against colonization by pathogenic S. enterica serovar Enteritidis.