A complete list of all bacterial strains and plasmids utilized in the experiments is provided in Supplementary Tables S1, S2. Salmonella Choleraesuis CVCC2139 was referred to as the wild-type (WT) strain for genetic manipulation to construct the indicated mutants. The Escherichia coli (E. coli) SM10 λ pir strain (19) served as the host for transferring suicide plasmids. All bacterial strains were cultured in Luria Bertani (LB) broth or agar at 37°C containing the appropriate antibiotics: 25 μg/mL chloramphenicol, 100 μg/mL ampicillin, and 50 μg/mL 2,6-diaminopimelic acid. For counterselecting mutant constructs via the sacB gene system, NaCl-free LB agar supplemented with 12.5% (w/v) sucrose was used.

Twenty-two S. Choleraesuis mutants were generated via allelic exchange, employing the suicide plasmid pRE112 as previously detailed (20). The primer sequences designed for gene deletion or complementation in S. Choleraesuis strains are detailed in Supplementary Table S3. To generate the ΔsseJ mutant, upstream and downstream homologous arms were PCR-amplified using primer pairs DsseJ-1F/1R and DsseJ-2F/2R, respectively. These PCR products were joined by overlap PCR and subsequently cloned and inserted into the suicide vector pRE112 (21) through seamless cloning generating plasmid pRE112-ΔsseJ, which carries a deletion of the entire sseJ gene. The pRE112-ΔsseJ plasmid was subsequently introduced into the WT strain via conjugation. This process involved chloramphenicol-mediated positive selection and sacB-mediated sucrose sensitivity screening for the generation of the markerless mutant strain ΔsseJ. Furthermore, to complement sseJ gene in the ΔsseJ, the coding sequence of sseJ were amplified with the primer sseJ-F/R. Then, the PCR product was inserted into the plasmid of pCZb1 (20) via a Seamless Cloning Kit (Sangon Biotech, Shanghai, China), generating plasmid pCZb1-sseJ. Following, the recombinant plasmid was transformed into the mutant strain ΔsseJ to construct the complemented strain named C-ΔsseJ. The same method was applied to constructions of other gene mutants and complemented strains.

The S. Choleraesuis WT and mutant strains were inoculated in 5 mL of LB broth at 37°C with shaking at 180 rpm/min overnight. The following day, cultures of each strain were normalized to an OD₆₀₀ of 0.05 and then cultured in LB broth at 37°C with shaking at 180 rpm/min. The OD₆₀₀ of each culture was measured every 2 h for 12 h.

The swimming motility phenotypes of wild-type (WT) and mutant bacterial strains were evaluated using a previously described protocol (22). Cultures of each strain were grown in LB broth to an optical density (OD₆₀₀) of 0.6–0.8. Bacteria were harvested by centrifugation, washed twice with PBS, and resuspended in the same buffer. Subsequently, 3 μL of the bacterial suspension was applied as droplets to LB agar plates containing 0.25% agar. After incubation at 37°C for 6 h, the diameter of the bacterial migration zone was measured to assess swimming motility.

RAW264.7 macrophages were plated at a density of 5 × 10⁵ cells per well in 24-well plates containing DMEM (Gibco, NY, USA) supplemented with 10% FBS (Tian Hang, Hangzhou, China) and 1% penicillin-streptomycin. S. Choleraesuis WT or mutant strains were added at a multiplicity of infection (MOI) of 100. Following a 30-min incubation in a 5% CO₂ at 37°C incubator to facilitate bacterial adhesion, cell monolayers were washed thrice with PBS to remove non-adherent bacterial cells. The adherent bacteria were then released by lysing the cells with 0.2% Triton X-100, and their numbers were enumerated via serial dilution and colony counting.

For the invasion assay, after the 30-min adhesion step, fresh DMEM was added and cells were incubated for an additional 90 min at 37°C under 5% CO₂. After incubation, the supernatant was discarded, cells were washed twice with PBS and lysed using 0.2% Triton X-100 to enumerate intracellular bacteria.

For intracellular survival analysis, following the invasion step, DMEM supplemented with 10 ng/mL gentamicin was used to eliminate extracellular bacteria. T = 0 h was defined as the initial time point following invasion. At designated time points (T = 0 h and 24 h), serial dilutions of the resulting cell lysates were then plated onto MacConkey agar (Coolaber, Beijing, China) plates for bacterial enumeration and incubated at 37°C for 24 h to count colony-forming units (CFUs).

Female Kunming mice (6 weeks old) were procured from Dashuo Experimental Animal Ltd. (Chengdu, China) and underwent a 1-week acclimation period before experimental procedures. Bacterial colonization and virulence phenotypes were evaluated using methodologies reported in prior studies (20, 22). Ten mice were orally inoculated with PBS or approximately 1 × 10⁸ CFU of S. Choleraesuis WT strain or each mutant strain. Then, spleens and livers were collected from 5 mice at 6 days post-infection. The samples were weighed, ground in PBS, and the bacterial suspensions were serially diluted and spread onto MacConkey agar (Coolaber, Beijing, China) to enumerate viable CFUs, which were expressed as log₁₀ CFU/g. The remaining 5 mice in each group were observed for survival for 1 month after infection.

The LD₅₀ of the ΔsseLΔsopD₂ΔsteA was determined as previously described (23). 10-fold serial dilutions of the CFU of the ΔsseLΔsopD₂ΔsteA strain were orally inoculated into groups of Kunming mice (n = 5/per dose). Animals were monitored daily for 30 days post-infection to assess survival rates. The median lethal dose (LD₅₀) was determined using the Reed-Muench method. To ensure humane endpoints, mice exhibiting severe distress—characterized by labored breathing, tremors, unresponsiveness to tactile stimuli, or inability to access food/water—were humanely euthanized via CO₂ inhalation. Deceased animals were immediately subjected to sterilization, sealed in biohazard bags, and transferred to the Sichuan Agricultural University Laboratory Animal Center for compliant biosafety disposal.

Female Kunming mice (6–8 weeks old) were randomly divided into three groups (n = 20/group). The experimental group received an oral gavage of 1 × 10⁹ CFU ΔsseLΔsopD₂ΔsteA in 200 μL PBS, while the control group received an equal volume of PBS alone. A booster immunization was administered on day 14 using the same protocol. Serum samples were collected via retro-orbital bleeding on days 7 and 21 from six randomly selected mice per group. On day 42, all mice were orally challenged with 10-fold LD₅₀ of S. Choleraesuis CVCC2139. Five mice per group were euthanized on day 6 post-challenge, and samples from the spleen and liver were collected for measurement of bacterial loads. Survival of the remaining mice was monitored and recorded daily for 30 days.

Antibody titers against inactivated S. Choleraesuis antigens were quantified using an indirect ELISA protocol as previously described (24). In brief, 100 μL of 10⁹ CFU/ml of the heat-killed S. Choleraesuis antigens was added to wells of a 96-well ELISA plate coated with antigen with overnight incubation at 4°C. The next day, the plates were washed three times with PBST followed by blocking with 5% BSA (BD, San Diego, CA) in PBS at 37°C for 2 h. Following antigen coating and blocking, the plate was washed three times with PBST. Serum samples, diluted 1:200 in blocking buffer (5% BSA in PBS), were added to each well (100 μL/well) and incubated at 37°C for 1 h in a humidified chamber. The plate was then washed five times with PBST to remove unbound antibodies. Subsequently, 100 μL of HRP-conjugated goat anti-mouse IgG (Abclonal, Wuhan, China), diluted 1:5,000 in antibody diluent, was added to each well and incubated at 37°C for 1 h. After five additional PBST washes, the plate was ready for substrate development. 100 μL of TMB substrate solution (Macgene, Shanghai, China) was added, and the plates were incubated in the dark at 25°C for 10 min. After adding 50 μL of 2 M H₂SO₄ to stop the reaction, absorbance was measured at 450 nm using a Bio-Rad microplate reader (Bio-Rad, California, USA).

All animal procedures were conducted in strict adherence to the Guide for the Care and Use of Laboratory Animals published by China's Ministry of Science and Technology. The study protocol was approved by the Animal Ethics Committee of Sichuan Agricultural University and the Sichuan Laboratory Animal Management Committee (permit number: SYXK2019-187), ensuring compliance with national and institutional welfare guidelines.

Data are presented as the mean ± standard deviation (SD) and analyzed using one-way analysis of variance (ANOVA) followed by Tukey's post-hoc multiple-comparison test with GraphPad Prism software (GraphPad Software, California, USA). Statistical significance was defined as P < 0.05. All in vitro experiments were independently repeated three times to ensure reproducibility.

Twenty-one effector genes (sseJ, sseG, slrP, sseF, gtgE, gogB, sspH, gtgA, sifA, sifB, sseK, steA, steC, sseL, sopD₂, spiC, sseI, pipB, pipB₂, sopD, steD) were screened and each of them was deleted from the WT S. Choleraesuis strain CVCC2139. The ssaV mutant strain (ΔssaV) was also constructed as a positive control as the SsaV is a structural component forming the inner ring of the SPI-2 T3SS injectosome that is involved in effector protein translocation (25). Then, the mutant strains were compared to the WT strain in terms of the ability of bacteria to adhere to, invade and survive within RAW264.7 macrophages. Deletion of the SPI-2 T3SS effector genes neither affected bacterial adhesion to nor changed bacterial invasion into macrophages (Figures 1A, B). Nevertheless, bacterial replication in the mutant strain ΔssaV, ΔsseF, ΔsseJ, ΔsifB, ΔsseK, ΔsifA, ΔsopD₂, ΔsteC, or ΔsteD was significantly decreased compared to that in the WT strain, while loss of either of the other 13 genes had on adverse effects (Figure 1C). Gene complementation in trans in the mutant strains fully or partially restored the WT phenotypes (Figure 1C). This finding indicated that effector genes including sseF, sseJ, sifB, sseK, sifA, sopD₂, steC, and steD promoted bacterial survival in macrophages.

The 21 S. Choleraesuis mutant strains were subjected to detection of growth curves and swimming. Compared to the WT strain, the growth rates of the ΔgtgA, ΔsteA, ΔspiC, ΔsopD₂, ΔslrP, ΔpipB₂, ΔsopD, and ΔsteD strains were significantly reduced in LB broth at 37°C (Figure 2A). In contrast, the remaining 14 mutants exhibited similar growth rates to the WT strain (Figures 2B, C). This finding suggested that the effector genes gtgA, steA, spiC, sopD₂, slrP, pipB₂, sopD, and steD promotes the in vitro growth of S. Choleraesuis. Also, swimming of the ΔsseJ, ΔslrP, ΔsifB, and ΔsopD₂ mutants were significantly enhanced compared to that of the WT strain, whereas the ΔpipB₂ and ΔsspH strains exhibited reduced motility (Figure 2D). Complementation of sseJ, slrP, sifB, sopD₂, pipB₂and sspH in corresponding mutants restored WT swimming phenotype. Therefore, the effector genes sseJ, slrP, sifB, and sopD₂ restrain the swimming ability of S. Choleraesuis, whereas sspH and pipB₂ positively influence motility.

To determine roles of the SPI-2 T3SS effector in virulence of S. Choleraesuis, mice were orally administered with 1 × 10⁸ CFU of the WT strain or each of the mutant strains. The survival of mice was monitored for 1 month. The WT strain led to 40% of survival post-infection, while the five strains including ΔspiC, ΔssaV, ΔsseL, ΔsopD₂, and ΔsteA caused no death (Table 1). Also, ΔgtgA, ΔgtgE, ΔsseF, ΔsifA, and ΔsspH resulted in increased survival (80%), by contrast, all mice succumbed to the ΔsteD infection. The remaining strains caused a 40% or 60% survival (Table 1). Thus, SPI-2 T3SS effector genes sseL, sopD₂, and spiC, and the gene ssaV contributed remarkably to S. Choleraesuis virulence in mice, functioning as virulence genes. To further detect the roles of the five virulence genes in bacterial colonization, mice were inoculated with 10⁸ CFU of the WT or mutant strains, then bacterial loads in liver and spleen were measured 6 days post-infection. No bacteria were recovered from the ΔspiC and ΔssaV groups. However, the bacterial loads of ΔsseL, ΔsopD₂, and ΔsteA in liver and spleen tissues were comparable to those of the wild-type strain (Figures 3A, B). This finding suggests that the spiC and ssaV contribute to the colonization of S. Choleraesuis in the liver and spleen of mice, while sseL, sopD₂, and steA are not involved in bacterial colonization.

Although the two genes spiC and ssaV play a significant role in bacterial virulence, their mutant strains lost colonization ability in mice, hinting their poor immunogenicity. To develop a suitable live attenuated strain, the other three virulence genes sseL, sopD₂, and steA were deleted from the WT strain simultaneously, generating a triple mutant ΔsseLΔsopD₂ΔsteA. This strain colonized of the spleen and liver at a significantly lower level than the WT strain post oral infection (Figure 4). Also, the LD₅₀ of the triple mutant was >1.65 × 10¹⁰ CFU, demonstrating at least a 150-fold reduction in virulence compared with that of the wild-type strain with the LD₅₀ of 1.08 × 10⁸ CFU (Table 2).

To evaluate the vaccine potential of the attenuated strain ΔsseLΔsopD₂ΔsteA, mice were orally administered with 10⁹ CFU of the vaccine strain twice with an interval of 14 days and then were challenged orally with a lethal dose of the S. Choleraesuis WT strain 28 days post-second immunization. Immunization with the ΔsseLΔsopD₂ΔsteA induced significantly higher serum IgG responses to whole bacterial antigens than with the PBS group on Day 7 and 21 post-immunization (Figure 5A). Following the challenge, the bacterial loads in the spleen and liver of the ΔsseLΔsopD₂ΔsteA group were significantly lower than those in the PBS group (Figures 5B, C). Furthermore, all the mice in the PBS control group died, whereas 40% of the mice in the immunized group survived (Figure 5D). Thus, immunization with the ΔsseLΔsopD₂ΔsteA strain induced a robust antibody response, significantly reduced the tissue loads of the challenge strain, and provided 40% protection efficacy against lethal S. Choleraesuis infection.