The bacterial strains and plasmids used in this study are listed in Table 1. The STEC Ee strain (O139, positive for ST-IIe, F18ab) is a virulent field-type strain isolated from pigs in a farm in Wuhan during an outbreak of ED (Liu et al., 2007b). The attenuated S. Choleraesuis vaccine strain C500 and the wild-type, virulent parental strain C78-1 were supplied by the China Institute of Veterinary Drug Control (CIVDC, Beijing, China). C500 was chosen as the parent strain for the generation of genetically modified strains. E. coli and S. Choleraesuis cultures were grown at 37°C in Luria-Bertani (Alborali et al., 2017) broth or on LB agar plates. When required, D L-α, ε-diaminopimelic acid (D L-α, ε-DAP) (Liu et al., 2007a; Gangaiah et al., 2022) was added (50 μg/ml) for the growth of asd- strains (Xu et al., 2006).

The primers used for preparation of the recombinant a virulent vaccine strains are listed in Table 2. S. Choleraesuis C520, which has deletions in both crp and asd, was generated from the S. Choleraesuis C500 vaccine strain as described previously (Xu et al., 2006). Briefly, DNA was introduced into the bacteria by electroporation. A 1048 bp upstream fragment of the crp gene was amplified by PCR as described, with the exception that polymerization was performed at 72°C for 2.5 min from the genomic DNA of S. Choleraesuis C500 strain using two pairs of primers (Accession No: AE008863; N terminal, crp1, pr1 and pr2, crp2, pr3, and pr4). The amplified fragment was cloned into the XbaI and BamHI sites of the pBluescript II SK (+) vector to construct pSK-crpup. Then, the 1743-bp downstream fragment of the crp gene was PCR-amplified using a pair of primers (pr3 and pr4) and cloned into the XhoI and KpnI sites of pSK-crpup to obtain pSKΔcrp, which resulted in a 320-bp deletion, including the crp gene fragment. The 2890-bp fragment, including composed of the upstream and downstream fragments of the crp gene, from an XbaI- and KpnI-digested pSKΔcrp plasmid was ligated to pRE112 plasmid to yield the pREΔcrp suicide plasmid. Transfer of the recombinant suicide plasmid to S. Choleraesuis C500 was accomplished by conjugation using E. coli χ7213 (pREcrp) as the plasmid donor. Strains containing single-crossover plasmid insertions (C500crp: pREΔcrp) were isolated on plates containing chloramphenicol and DAP. Loss of the suicide vector after the second recombination event between homologous regions (i.e., allelic exchange) was selected for by using a sacB-based sucrose sensitivity counter selection system. The presence of the 320-bp crp deletion in S. Choleraesuis C500 was confirmed by sucrose-sensitive growth on media and by PCR using a flanking crp primer set (pr5 + pr6). A pREasd plasmid was constructed by the same method. Then C500Δcrp as recipient bacterium was conjugated with the E. coli χ7213 (pREasd) donor. The presence of the 1,408-bp asd deletion in S. Choleraesuis C520 was confirmed by the inability of the strain to grow on media without DAP and by PCR using a flanking asd primer set (pa5 and pa6).

To construct S. Choleraesuis C522, which is a derivative of C520 that expresses rSF, fragments encoding mature ST-IIe B and fedF were amplified from the genomic DNA of Ee E. coli using two pairs of primers (pB1, pB2, and pF1, pF2) that were designed according to the ST-IIeB gene sequence (GenBank accession no: AY332411) and the fedF gene sequence (GenBank accession no: AFZ26250) (Liu et al., 2007a). These primers contain restriction sites (EcoI, SacI, SacI, and HindIII) to allow the direct cloning of the PCR product into pYA-3493 plasmid. Two rounds of amplification were performed in a total volume of 50 μl containing 200 μM deoxynucleoside triphosphates (dATP, dCTP, dGTP, and dTTP), 1 pmol of each primer, 5 μl of dilution buffer, and 2.5 U of Taq polymerase. Thirty cycles were performed, each consisting of a denaturing step of 1 min at 94°C, an annealing step of 60 s at 55°C, and a 60 s extension step at 72°C. The 1095 bp PCR fragment of ST-IIeB and the fedF fragment were purified and cloned into the EcoRI and HindIII sites of pYA3493, resulting in pYA-SF. In-frame cloning of pYA-SF was confirmed by nucleotide sequencing. pYA-SF (encoding rSF) was electroporated into the C500ΔasdΔcrp strain (named C520), to construct the recombinant S. Choleraesuis C522 (pYA-SF). To prepare a control strain, pYA3493 (vector control) was electroporated into C520, resulting in C521 (pYA3493) being constructed.

The growth curves of the strains in LB were determined, and carbohydrate fermentation or utilization assays were processed according to the manufacturer’s protocol. MH medium, sugar fermentation tube purchased from Tianhe Microbial Reagent Co., Ltd (Hangzhou, China). The group O serovar and H antigen were identified by slide agglutination with antisera supplied by the CIVDC (Beijing, China). The rSF fragment expression in the cytoplasm and culture supernatant of C522 was monitored by 12% SDS-PAGE, and immunoblot analyses were performed with the anti-rSF rabbit polyclonal antibody as previously described (Liu, 2008). Specific band intensities were further analyzed by densitometry using the Quant Studio™ 5 real-time PCR instrument (MA, USA).

Duroc × Landrace × Yorkshire (DLY) Hybrid Pigs (F18R+, 28–30 days of age) were obtained from a breeding farm in Shandong Province, China to the experimental animal house of Huazhong Agricultural University. All pigs were confirmed either to be without antibody against both O antigen of S. Choleraesuis and rSF by their individual ELISA or to be culture-negative for both S. Choleraesuis and STEC by the enrichment of rectal swabs. In the process of the study, all animal experiments were approved by the Animal Ethics Committee (AEC) of the Huazhong Agricultural University, ethics number HZAUSW-2006-0005 and were performed according to the ARRIVE guidelines. When immunizing pigs (30 days of age), 2.0 × 10⁹ CFU immunization doses of vaccine strain was administered orally in accordance with the vaccine instructions for the use of the C500 strain. A total of 128 pigs were randomly assigned to four groups, as follows: group 1 (pigs 1–32) received C522; group 2 (pigs 33–64) received C521; group 3 (pigs 65–96) received C500; group 4 (pigs 97–128) was non-infected controls, respectively, via the oral route (Narita and Ishii, 2004). Twelve pigs per group, group B, group C and group D were used for the evaluation of C521 to evaluate whether or not having ability to elicit a mucosal immune response via the oral route in swine. In addition, twelve pigs in group A were monitored as well as other three groups. Clinical manifestations and rectal temperature changes were observed daily over the first 5 days and on day 14 and 21 post-vaccination. And fecal consistency score was evaluated on a continuous scale (0–4/4) as described by Fairbrother et al. (2017). Fecal consistency scores of 2, 3, and 4 were considered as indicative of mild, moderate and severe diarrhea. Pigs with rectal temperature ≥41°C are fever (Yu et al., 2012). Additionally, rectal swabs and blood samples from each pig were collected. Autopsies were conducted as presently as possible for simultaneous death or on day 0, day 1, day 14, and day 21 post-vaccination. Three grams of spleen, mesenteric lymph nodes and Peyer’s patches were immediately prepared in triplicates for cytokine determination. To determine the bacteria counts and antibody titers in the infected organs, 1 g of tissue samples from the spleen, mesenteric lymph nodes and Peyer’s patches were homogenized in 10 ml phosphate-buffered saline (PBS). Enumeration of the bacteria strains in these organs or tissues was performed by plating a dilution series of lysates on MacConkey agar after overnight incubation at 37°C. Also, 2 ml PBS were used to wash the intestinal mucosa of the part of terminal ileum and then the wash fluid was collected. Antibody titers were determined in supernatants of homogenized organs and wash fluid. The other organ samples were fixed in 10% (w/v) buffered formalin and then subjected to histopathological examination.

An additional 20 pigs in each group were selected for protection studies. After 3 weeks of inoculation, 10 pigs were challenged with 2 × 10¹⁰ CFU of C78-1 and another 10 pigs with 2.5 × 10¹¹ CFU of Ee strain. Monitoring and sample collection from challenged pigs were performed as described above in the post-vaccination and post-challenge periods. Necropsies were performed after simultaneous death on day 21 post-infection. Organ and tissue samples were prepared as described above. Rectal swabs and organs were examined to determine whether the orally administered live vaccine candidates or challenge strain were shed in the feces or located in the organs. Isolation and identification of the vaccine candidates and challenge strains were performed according to previously described methods (Xu et al., 2006). Blood and other samples were collected and stored at −80°C until use.

Total RNA was extracted from the spleens of the vaccinated pigs using ISOGEN (Invitrogen). Reverse transcriptase-PCR (RT-PCR) was monitored using a one-step RNA PCR kit (Takara) with three pairs of primers according to the operating instructions. Primers are listed below: IFN-Υ (5′-GTTTTTCTGGCTCTTACTGC-3′; 5′-CTTCCGCTT TCTTAGGTTAG-3′) (Chaoprasid and Dersch, 2021); TNF-α (5-ACTGCACTTCGAGGTTATCGG-3′, 5′-GGCGACGGGCTTATCTGA-3′) (Meissonnier et al., 2008); interleukin (IL)-4 (5′-GTCTGCTTACTGGCATGTACCA-3′; 5′-GCTCCATGCACGAGTTCTTTCT-3′) (Duvigneau et al., 2005); GAPDH (5′-AACGACCCCTTCATTGAC-3′; 5′-TCCACGACATACTCAGCAC-3′).

Quantitative reverse transcriptase-PCR (qRT-PCR) was performed on the 7900HT Sequence Detection System (Applied Biosystems) using SYBR Green (Meissonnier et al., 2008). Each sample was analyzed in triplicate. Data were analyzed by a comparative CT method (Applied Biosystems). Transcript levels were calculated by normalizing to the levels of GAPDH mRNA. In addition, cytokines (IFN-γ, IL 4, TNF-α) in the spleen were analyzed in relation to the expression of antibodies (IgG, IgA) in the mucus, MLN and serum utilizing JMP software.¹

Salmonella Somatic or rSF ELISA was used to monitor antibodies in serum samples, lymphoid-associated tissue and intestinal mucosal from each piglet. For the determination of anti-rSF titers, 100 ng of purified rSF dissolved in 100 μl 0.1 M carbonate buffer (pH 9.6) was coated in each well of polystyrene 96-well flat-bottomed microtiter plates (Kangjia Ltd., China). For determining the anti-Salmonella Somatic antibody level, S. Choleraesuis C500 cells were diluted in PBS at 3 × 10¹¹ CFU/ml and grown overnight. The cells were then harvested by centrifugation, inactivated for 10 min at 80°C and stored at −80°C. Plates coated with 100 μl of this suspension at 100-fold dilution in carbonate buffer were incubated at 37°C for 1 h, followed by overnight incubation at 4°C. Then they were blocked with a blocking buffer (PBS, 0.1% tween 20, and 5% skimmed milk). Samples of serum, lymphoid associated tissue and intestinal mucosa were diluted at the optimal dilution factor and added to each well and incubated at 37°C for 30 min. After three washes, plates were treated with biotinylated goat anti-pig IgG (Southern Biotechnology Inc., Birmingham, AL, USA) and lymphoid associated tissue homogenates or IgA at 37°C for 30 min, followed by five washes. The substrate solution TMB and H₂O₂ (50 μl) were then added to each well, and the plates were incubated at room temperature in the dark for approximately 10 min. The catalytic reactions were stopped with 50 μl 1% SDS. The optical densities were read at 630 nm using an ELISA reader (Liu, 2008).

The wild-type S. Choleraesuis strain C78-1 and the STEC field strain Ee were cultured in tryptose soy agar medium for 24 h at 37°C. The cultures were diluted of 1:100 in the tryptose soy broth and was incubated for 6 h with gentle shaking. All the pigs were fasted for 24 h prior to infection. The pigs were then intramuscularly injected with azaperone (stresnil, 4 mg/kg body weight), and 10 ml of a 10% (w/v) sodium bicarbonate solution was imported via gastric intubations followed by administration of bacterial emulsion. The Ee strain infection model was applied according to the method of B. T. Bosworth (Cornick et al., 1999). All of the pigs in this model were fed commercial rations containing 21.0% crude protein ad libitum, and until day 21 after the challenge, feed consumption, diarrhea, disorientation, abnormal behavior, and histopathology were all regularly assessed. For light microscopy, lungs and intestines were fixed in 10% formaldehyde solution embedded in paraffin and sections stained with Hematoxylin and eosin. Microscopically histological changes of these tissues caused by C78-1 and the Ee were the representative.

All data analysis was performed using the T Test or ANOVA in the SPSS 27 software² for comparison of the differences in specific antibody levels between different groups. P-value less than 0.05 (typically < 0.05) was statistically significant.

To improve the S. Choleraesuis C500 vaccine, we introduced deletion mutations in both asd and crp genes. The resulting strain, C520, lacked the ability to synthesize DAP and was unable to grow on media without DAP, as expected. Its ability was restored when transfected with the control vector pYA3493, resulting in the creation of C521; or with the rSF expression vector pYA-SF, resulting in the creation of C522 (Figure 1D). The mean generation in Luria broth of the recombinant S. Choleraesuis strains C522 (31.2 min) and C521 (28.1 min) were similar to that of the parent attenuated C500 vaccine (27.9 min). We further compared their fermentation patterns on different carbohydrates. As expected, S. Choleraesuis C500, but not C521 and C522, was able to use maltose, glucose, mannose and xylose. Moreover, C521, C522 and the parent C500 strain were confirmed to share the same O and H antigenic type, 6, 7: C: 1, 5.

To confirm that S. Choleraesuis C522 expressed rSF we performed immunoblotting and Coomassie blue staining of SDS-polyacrylamide gels. C522 expressed the rSF of a molecular weight of 37 kDa, which is consistent with the calculated size of rSF (Figure 1A). Analysis of Coomassie blue-stained SDS-polyacrylamide gels showed that the amount of the rSF protein accounted for up to approximately 2.1% of the total C522 (pYA-SF) protein (Figures 1B, C). Approximately 69.8% of the rSF was detected in the cell lysates and 30.2% in the culture supernatants. To examine the stability of plasmids pYA3493 and pYA-SF in C522 and C521 in vitro, cells were cultured with daily passage of 1:1,000 dilution for five consecutive days in LB broth containing DAP. The last day, the amounts of the 37-kDa rSF that were expressed were similar to those from the first day, suggesting that the expression of rSF is stable from rearrangements.

To evaluate the colonization abilities of the engineered strains, we vaccinated swine. All C500-inoculated piglets had diarrhea from day 1 to day 3 post-vaccination, and 20% of C500-inoculated piglets had fever after 14 days post vaccination; however, piglets inoculated with C520, C522, or C521 did not show side effects such as diarrhea, fever, depressed spirit, or abnormal behavior during the corresponding period (Figure 2A). The vaccine strains were detected in rectal swabs collected from immunized piglets from day 1 to day 10 after oral immunizations, but only the C500 strain was recovered from shedding fecal samples from day 1 to day 3 post-vaccination. These findings suggest that the vaccine strains have the ability to colonize the intestine, which is vital to the mucosal immunity and also suggests that C521 and C522 are safer than their parent strain C500.

To verify that the vaccine strains can settle in lymphoid tissue and to assess the residence time, we monitored the quantities of vaccine strain in the Peyer’s patches, mesenteric lymph nodes and spleen at different times after inoculation. Between 0 and 1.4 × 10³ CFU were detected from day 0 to day 21 post-inoculation (Figure 2B). These results confirm that the vaccine strains can colonize both the intestinal mucosa and lymphoid tissues.

To further evaluate the cellular immune response to the C521 vaccine strain, we assessed the expression of inflammatory cytokines in the spleen of vaccinated piglets and naive piglets. On days 1, 14, and 21 after inoculation, the expression of IFN-γ, IL4, and TNF-α were elevated as assessed by qRT-PCR in the vaccinated piglets in contrast to piglets without inoculation of C521, on day 0. From day 1 to day 21 post-vaccination, IFN-γ expression level declined from day 1 to day 21 post-vaccination, whereas for TNF-α expression level increased. As for the expression of IL4, it sustains a certain level at three timepoints, day 1, 14, and 21 (Figure 2C). In addition, the correlation analysis showed that the Pearson correlation between the expression of cytokines (IFN-γ, IL 4, TNF-α) in the spleen and the expression of antibodies (IgG, IgA) in the mucus (Supplementary Table 1), MLN (Supplementary Table 2) and serum (Supplementary Table 3) was between 0.4 and 0.8, indicating better correlation. These results indicate that inoculation with strain C521 may elicit cellular immunity in swine.

To further assess the immunogenicity of the vaccine strains, we evaluated the formation of antibodies against S. Choleraesuis in the sera, gut mucosa and mesenteric lymph nodes (MLN) of piglets that were orally vaccinated with C522, C521, and C500. ELISA was performed using whole cell S. Choleraesuis as antigen. C522, C521, and C500 vaccine elicited IgA and IgG antibody responses in the sera (Figures 3A, B), gut mucus (Figures 3C, D) and MLN (Figures 3E, F) on days 14 and 21 that were 7–9-fold higher than on day 0 (P < 0.01). Additionally, the antibody response to C522 was similar to that elicited by vaccination with C500 and C521. These results suggest C521 and C522 confer a vigorous systemic immune response, which verifies that the asd and crp deletions in these strains had a minimal influence on the immune response to Salmonella itself.

We also assessed the production of antibodies against rSF by C522, C520, and C521, which is the only vaccine strain that expresses rSF. Our results demonstrate that C522-vaccinated pigs produced IgA and IgG antibody responses in the sera (Figures 4A, B), gut mucus (Figures 4C, D), and MLN (Figures 4E, F) at days 14 and 21.

As a test of the efficacy of the vaccine strains, we assessed the ability of oral administration of the recombinant C521 vaccine, C522 vaccine, and the parental C500 vaccine to induce immunity in swine against oral challenge with an absolute lethal dose of the wild type, virulent parent strain, C78-1 (2 × 10¹⁰ CFU) or the STEC Ee strain (2.5 × 10¹¹ CFU) (Zhao, 2009). Animals were fed a high-protein diet (comparable to commonly used commercial weaning rations). Complete protection against C78-1 was afforded by C521, C522, and C500, whereas there were no survivors in a group of ten naïve piglets in the PBS group (Table 3). Furthermore, almost no Ee strain was detected in rectal swabs from pigs orally vaccinated with C522 from day 3 until day 21 post-challenge, whereas large quantities were detected in the blank control group (Figure 5A). Only the C522 strain provided protection against STEC Ee. The other groups of animals showed clinical signs that were consistent with ED. Three pigs in the PBS group died of ED after seven days, and two were in lateral recumbence for 2 days and were euthanized. These results indicate that oral immunization with C521 or C522 (pYA-SF) can provide complete protection from S. Choleraesuis infection, and that C522 can additionally provide protection against STEC Ee.

Continual fevers typically occur in pigs undergoing necrotic enteritis. To further determine the in vivo efficiency of the recombinant strains, we monitored the temperatures during the entire animal experiment. Vaccinated pigs had marginally increased in body temperature 1 day after inoculation, followed by temperatures that immediately return to basal levels (Figure 5B), which suggests that continued necrotic enteritis was not present.

We also directly examined tissues for pathologic signs of necrotic enteritis in 21 days after vaccination. Naive pigs (PBS) inoculated with C78-1 and Ee showed severe pathological changes in the lungs and intestines (Figures 6A1–8). All of these pigs had extensive systemic lesions including focal necrosis in the liver and lymph nodes, fibrous necrotic enteritis in the ileum and caecum and mild interstitial pneumonia. In contrast, no significant pathological lesions were observed for the vaccinated groups upon challenge with C78-1 and Ee (Figures 6B1–6). These findings confirm the efficacy of C522 in protection swine against both S. Choleraesuis and STEC Ee.

To verify the protective response in C522-vaccinated pigs, the IgG and IgA responses in the MLN were assessed on day 21. C522 induced high antibody (IgG and IgA) response against both rSF and S. Choleraesuis Somatic (Figure 5C). Similar results were observed in the sera and mucosa (Figure 4). These results demonstrate that rSF fusion protein expressed by the strain C522 confers systemic rSF-specific immunity.