The reagents and antibodies used in the study are shown in Table 1 .

The S. Infantis wild-type strain CAU1508 was isolated from the intestinal contents of diarrhea piglets. S. Infantis with the pFPV-mCherry plasmid has been previously described (12). The SopB mutant strain was derived from the parental S. Infantis wild-type strain CAU1508 and constructed using the λ-Red homologous recombination system.

Caco-2 cells were purchased from Kunming Cell Bank of Chinese Academy of Sciences. The Caco-2 cells were cultured in DMEM/High Glucose media supplemented with 10% FBS and 1% penicillin streptomycin at 37°C in a 5% CO₂ incubator. Cells were seeded in six-well (1×10⁶ cells per well) or 24-well culture plates (1×10⁵ cells per well) and infected when the cell density reached 60% (this is to ensure that bacteria can infect as many cells as possible). Salmonella was grown in LB medium overnight with shaking at 200 rpm and 37°C, then subcultured in 10 ml fresh LB medium (1:40) with shaking under the same conditions for 4 h. Following that, the bacteria were centrifuged at 4,000 g for 15 min at room temperature and resuspended in PBS. The entire infection was according to the experimental procedure of the gentamicin protection assay. The monolayers were infected at an MOI of ~50 for 15 min and washed three times with PBS supplemented with gentamicin (100 μg/ml) to remove extracellular bacteria. Cells were then incubated in fresh growth medium containing gentamicin (100 mg/ml) for 2 h, followed by growth medium supplemented with gentamicin (10 μg/ml) until the infection was complete. After 15 min of Salmonella invasion, the time was 0 hpi (hours post-infection), and the infection lasted for 8 h in total. For groups that required treatment with MK2206 (1 μM) and SC79 (25 μM), both drugs were added 2 h before infection, and their concentrations remained unchanged in the medium until the end of the experiment. For the positive control of apoptosis, cells treated with CCCP (Carbonyl cyanide 3-chlorophenylhydrazone, an apoptosis inducer, 50 μM) for 1 h before other experiments were carried out (immunoblotting or immunofluorescence).

Cell viability was determined using the Cell Counting Kit-8. Cells were seeded in 96-well culture plates. At the end of each treatment, the medium was removed and replaced with 100 μl medium containing 10 μl fresh CCK-8 solution, then incubated at 37°C for 2 h. Following that, absorbance was measured at 450 nm. Experiments were performed six times on each group to ensure the authenticity of the results.

In order to quantify viable intracellular bacteria, monolayers in six-well plates were washed three times with PBS containing gentamicin (100 mg/ml) for 5 min each time before being lysed in 1 ml of 0.3% (v/v) Triton X-100. Serial dilutions were plated on LB agar plates. For quantification of intracellular cytosolic bacteria, cells were co-incubated with media containing chloroquine (700 μM) for 1 h before being solubilized in 1 ml of 1% (v/v) Triton X-100, then plated on LB agar plates.

After Salmonella infection, Caco-2 cells were collected, washed twice with PBS, suspended in 100 μl 1× binding buffer, and stained with the Annexin V-PE/7-AAD Apoptosis Detection Kit. Next, 5 μl each of Annexin V-PE and 7-ADD were added to each sample and incubated in the dark for 15 min. Flow cytometry analysis was performed using a BD FACSVerse™ Flow Cytometer.

Cells were seeded into 24-well plates. After 8 h of Salmonella infection, media was removed and the cells were washed with PBS. Following that, 2 ml of 1× assay buffer supplemented with Calcein-AM (2 μM) and PI (4.5 μM) was added to the cells and incubated at 37°C for 15 min. The samples were then examined with a fluorescence microscope under a 490 ± 10 nm excitation filter.

Total protein of Caco-2 cells was extracted using RIPA buffer (Solarbio, Beijing, China) containing a protease/phosphatase inhibitor cocktail (Cell Signaling Technology, USA) on ice for 30 min. Protein concentration was quantified using the BCA Protein Assay kit (23227, ThermoFisher Scientific). SDS-PAGE was used to separate protein samples, which were then transferred to polyvinylidene fluoride membranes. After incubation with 5% skim milk, the membranes were incubated with primary antibodies. Further details about the primary antibodies are described in the Table 2 . Next, the membranes were incubated with secondary antibodies, then coated with ECL immunoblotting substrate. Images were captured using a Tanon 6200 chemiluminescence imaging workstation.

Cells were fixed with 4% paraformaldehyde and permeabilized with 1% (v/v) Triton X-100. Following that, the samples were blocked with 2% bovine serum albumin at room temperature. The samples were then incubated with anti-LAMP1 (1:100); anti-Cleaved-caspase-3 (1:100); anti-Cytochrome c (1:50); anti-Phospho-Akt (Thr308) (1:800); anti-Tom-20 (1:200) at 4°C overnight. The Mito-Tracker Red CMXRos was used to label mitochondria. Next, the samples were incubated with secondary antibodies at room temperature. Either DAPI or Hoechst 33342 was used to stain the DNA. Images were taken using a Nikon A1 confocal laser scanning microscope.

Mitochondria were labeled with Tom-20 and imaged using confocal microscopy. The Mitochondrial Network Analysis (MiNA) toolset, which consists of a relatively simple pair of macros using existing ImageJ plug-ins, was used to analyze the mitochondrial networks.

All animal work was performed in accordance with the Guidelines for Laboratory Animal Use and Care of the Chinese Center for Disease Control and Prevention and the Rules for Medical Laboratory Animals (1998) of the Chinese Ministry of Health. A total of 36 six-week-old male C57BL/6 mice were obtained from Charles River Laboratory Animal Technology Co., Ltd (Beijing, China). Mice were provided food and water ad libitum throughout the entire experiment. All mice were administered a single dose of streptomycin (15 mg per mouse) via gastric gavage before being infected via gavage 24 h later (2 × 10⁶ Salmonella in 200 μl of PBS). Control mice were orally fed an equal volume of PBS. The mice were then euthanized, and their ileum tissues were harvested 3 days after infection. Mice feces were homogenized in 1 ml of PBS, and serial dilutions were plated on LB agar plates to quantify bacterial burdens.

The severity of diarrhea was evaluated using the fecal score and the dry/wet weight of fecal pellets. The fecal scoring criteria were as follows: 1 (normal stool); 2 (slightly wet, soft stool and formed stool); 3 (wet and unformed stool with mucus); 4 (watery stool). In order to determine the fecal dry/wet weight ratio, mice were separately placed in a clean cage, without food or water. Next, 0.5 g of feces was collected and weighed. Following that, the feces were placed in a 60°C oven for 24 h until the weight change was less than 1% before being weighed. The dry/wet weight ratio was then calculated.

In order to evaluate ileal pathology, the mid-segments of the ileum were excised, rinsed with saline, then fixed with 4% paraformaldehyde for 48 h. Paraffin-embedded tissue samples were sectioned (3 μm) and stained with hematoxylin and eosin. A neutral resin was used for sealing. Tissue sections were observed and imaged using an Olympus CX23 microscope equipped with an imaging system.

For gene expression analysis, total RNA was extracted from ileal tissues using the Trizol reagent (Invitrogen, Carlsbad, CA, USA). RNA transcription was performed using the PrimeScriptTM RT Reagent Kit according to the manufacturer’s instructions (RR047A, TaKaRa, Japan). Quantitative real-time RT-PCR was performed using the SYBR Green PCR Master Mix (LS2062, Promega, USA). The cycle threshold (CT) values of target genes were normalized to the CT value of hypoxanthine gene hypoxanthine phosphoribosyl-transferase. The results were presented as fold-change using the 2^−ΔΔCT method. Primer sequences for PCR are listed in Table 3 .

All statistical analysis was performed using GraphPad Prism 7 with a one-way ANOVA or a t-test with Bonferroni correction. Data were presented as means ± SEM. P < 0.05 was considered statistically significant.

The curves showed that the bacterial load began to rapidly increase at 4 hpi and peaked at 8 hpi ( Figure 1A ). We discovered two bacterial subpopulations in the cell by adding chloroquine (used to selectively kill vacuolar bacteria): one in the SCV and the other free in the cytoplasm. The bacterial load was mainly contributed by cytosolic bacteria ( Figure 1A ). Next, we used mCherry-S. Infantis to infect the cells and LAMP-1 to label SCV for confirmation. Confocal microscopy images revealed that cytosolic S. Infantis was the absolute dominant subpopulation ( Figure 1B ).

The cell survival rate significantly decreased at 6 hpi and remained almost unchanged after 8 hpi ( Figure 1C ). Therefore, we focused on the 0–8 hpi period, which encompasses the peak of cytosolic replication and the low phase of cell viability. The results showed that Caspase-9, Cleaved-caspase-3, and Cleaved-PARP were not activated during S. Infantis infection ( Figure 1D ). Morphological analysis of apoptosis revealed that S. Infantis infection did not result in apoptotic characteristics of the nucleus ( Figures 1B, E ). These findings indicated that cytosolic S. Infantis hyper-replicated in Caco-2 cells without inducing apoptosis.

Akt regulates cell survival and suppresses apoptosis via phosphorylation at Thr308 and Ser473 sites (27–29). Several studies have found that Akt is constantly phosphorylated after S. Typhimurium invades epithelial cells (18, 20). In this study, we hypothesized that continuous Akt phosphorylation contributes to the inhibition of apoptosis. As seen in Figure 2A , phosphorylation of Akt occurred in a discontinuous manner during S. Infantis infection. The first phase occurred at 0.5 hpi and rapidly decreased to near the background level. The second phase was observed at 3–4 hpi, with the expression of p-Akt higher than in the first phase. The levels of Cleaved-caspase-3 and Cleaved-PARP significantly increased after Akt phosphorylation was inhibited by MK2206, an Akt inhibitor that inhibits Akt phosphorylation at Thr 308 and Ser 473 ( Figures 2B, C ).

Next, we explored which parts of the bacteria induced Akt phosphorylation. At 0.5 hpi, Akt phosphorylation was observed in all infected cells ( Figure 2D ). Images at 4 hpi revealed high levels of p-Akt in cells containing cytosolic hyper-replicating S. Infantis (bacteria number >20), which were not observed at 8 hpi ( Figure 2E ). In infected cells without cytosolic bacteria, phosphorylation of p-Akt was not detected at 4 and 8 hpi ( Figure 2F ). These results demonstrated that S. Infantis-induced Akt phosphorylation occurred in two distinct phases. The first phase is widely induced after the invasion and rapidly depleted within 30 min, whereas the second phase is only induced by cytosolic bacteria at 3–4 hpi. Both phases of Akt phosphorylation inhibited apoptosis of the infected cells.

Notably, the SPI1 effector SopB, which contributes to invasion and SCV maturation, has 4-phosphatase activity, which can induce Akt activation during S. Typhimurium infection (18, 20). As expected, the p-Akt level was almost completely diminished in cells infected with the ΔSopB mutant ( Figures 3A, C ). Interestingly, LY294002 (Ly, a pan Akt inhibitor) completely inhibited Akt phosphorylation, but there was a certain expression of p-Akt after Wortmannin (Wor, a PI3K/Akt inhibitor) treatment ( Figure 3B ). Furthermore, infection with the SopB mutant induced apoptosis, as demonstrated by nuclear chromatin condensation and increase of Cleaved-caspase-3 and Cleaved-PARP levels ( Figures 3C, D ). The SC79 (an Akt phosphorylation activator) was used to activate Akt ( Figure 3E ), with findings confirming that SopB-mediated Akt intermittent phosphorylation inhibited apoptosis of infected cells ( Figures 3F, G ). Importantly, wild-type (WT) S. Infantis had enough time for intracellular replication, resulting in increased bacterial load by inhibiting apoptosis ( Figure 3H ).

The mitochondrion is the primary control organelle responsible for endogenous apoptosis. In normal cells, individual mitochondria connect to form tubules and shape dynamic networks through continuous division and fusion (30). Therefore, we evaluated the morphology of the infected cells’ mitochondrial network. At 4 hpi, the mitochondria of WT S. Infantis–infected cells still maintained an abundant network structure, but infection with the SopB mutant disrupted the mitochondrial network ( Figure 4A ). The mitochondrial cavity also appeared to be expanding, as evidenced by the ring-shaped structure (yellow arrows) ( Figure 4A ). Morphological analysis of the mitochondrial network revealed that infection with the WT strain had no discernible effect on the dynamics of the mitochondrial network ( Figure 4B ).

The key events of mitochondria-mediated apoptosis are the opening of mitochondrial permeability transition pore (MPTP) and the release of cytochrome c (31, 32). In the WT group, cytochrome c and mitochondria remained co-localized, while mk2206 treatment resulted in cytochrome c translocation from the mitochondria to the cytoplasm ( Figures 4C, D ). Infection with the SopB mutant resulted in massive cytochrome c release into cytoplasm, which was reversed by the addition of SC79, partially restoring the mitochondrial network ( Figures 4C, D ). Interestingly, co-localization of cytochrome c and mitochondria was also restored by the addition of CSA (a MPTP blocker) ( Figure 4C ). The mitochondrial membrane permeability may be affected by p-Akt induced by SopB, and the Bcl-2 family regulates the permeability of the mitochondrial outer membrane by inhibiting Bax translocation from the cytosol to the mitochondria (33, 34). We extracted mitochondrial protein and detected the distribution of Bcl-2 and Bax. As shown in Figure 4D , WT S. Infantis infection significantly increased the distribution of Bcl-2 in the mitochondria, while the addition of mk2206 decreased Bcl-2 and increased the distribution of Bax in the mitochondria. Infection with the SopB mutant also resulted in the decrease of Bcl-2 and the increase of Bax in mitochondria, which could be reversed by adding SC79. In addition, WT S. Infantis infection could phosphorylate Bad and Caspase-9 ( Figure 4E ), which enhanced the inhibition of apoptosis. In summary, S. Infantis-mediated Akt phosphorylation by SopB maintained mitochondrial dynamic network homeostasis, hence suppressing the apoptosis in infected cells.

Flow cytometry results showed that the proportion of 7-ADD⁺/Annexin-V PE⁺ cells significantly increased during 6–8 hpi, and the cells entered later stage of apoptosis without the early phase ( Figure 5A ). This indicated that there was a change in the membrane permeability of infected cells. In order to validate this conjecture, we performed double staining with Calcein-AM and PI during infection with S. Infantis. Images revealed that a subset of cells in the S. Infantis infection group had been damaged, as evidenced by PI-positive staining ( Figure 5B ). The scanning electron microscope images revealed that plasmalemma was destroyed and bacteria had been drilled out along the pores ( Figure 5C ), suggesting the occurrence of pyroptosis.

Next, the protein markers of pyroptosis, caspase-1 (p10), and GSDMD-N were examined. Infection with WT S. Infantis significantly activated pyroptosis ( Figure 5E ). Surprisingly, the ΔSopB strain induced caspase-1 (p10) and GSDMD-N activation 2 h earlier than WT S. Infantis ( Figure 5F ). However, the SopB mutant induced lower levels of caspase-1 (p10) and GSDMD-N compared to the WT strain, indicating a weaker degree of pyroptosis induced by the SopB mutant ( Figure 5G ). A recent study reported that p-Akt suppressed inflammasome activation in Salmonella-infected macrophages (35). Pyroptosis can be regulated by S. Infantis through p-Akt. MK2206 significantly reduced the levels of caspase-1 (p10) and GSDMD-N induced by WT S. Infantis, while SC79 treatment caused the WT strain to display a similar regulation as the ΔSopB strain ( Figures 5H, I ). Furthermore, SC79 treatment resulted in a significant decrease in caspase-1 (p10) and GSDMD-N levels during infection with the ΔSopB strain ( Figure 5J ), indicating that Akt phosphorylation both delayed pyroptosis and aggravated the severity of pyroptosis of infected Caco-2 cells. Intracellular bacterial load detection revealed that the number of bacteria in cells infected with the WT strain was much higher than in cells infected with the ΔSopB strain ( Figure 5D ). This may explain the two phenotypes of S. Infantis causing varying degrees of pyroptosis: the WT strain has a greater bacterial load and stimulates the inflammasome more strongly.

In order to verify our findings in vivo, we infected the C57BL/6 mouse model with Salmonella. The severity of diarrhea was determined by fecal score and the dry/wet weight of fecal pellets. The results revealed that WT S. Infantis caused more severe diarrhea than the ΔSopB strain ( Figures 6A, B ). In addition, the fecal bacterial load in the WT group was also significantly higher than the ΔSopB group ( Figure 6C ). Since Salmonella infection can cause severe ileal injury, the pathological changes in the ileum were examined. Infection with WT S. Infantis resulted in more severe ileal damage than infection with the ΔSopB strain ( Figures 6D, E ). Consistent with the in vitro results, p-Akt (Ser473 and Thr308) was found to be highly expressed in the WT group ( Figure 6F ). As shown in Figure 6G , infection with WT S. Infantis significantly increased the mRNA level of inflammatory factors, while the mRNA level of inflammatory factors induced by the ΔSopB strain was lower compared to WT S. Infantis. In combination with the detection of caspase-1 level ( Figure 6H ), we discovered that the infection with WT strain led to more severe intestinal inflammatory injury. Furthermore, immunoblotting and the TUNEL fluorescence assay showed that the ΔSopB strain caused more severe apoptosis of intestinal cells than the WT stain ( Figures 6I, J ). Intestinal cells infected with the ΔSopB strain may shed rapidly from the epithelium through apoptosis and be eliminated from the body, reducing the gut Salmonella load and inflammatory response.

In conclusion, S. Infantis delayed the death of infected Caco-2 cells through intermittent activation of Akt mediated by SopB, allowing intracellular cytosolic bacteria sufficient time for replication and resulting in more severe intestinal inflammation.