Salmonella effector SopF regulates PANoptosis of intestinal epithelial cells to aggravate systemic infection

ABSTRACT SopF, a newly discovered effector secreted by Salmonella pathogenicity island-1 type III secretion system (T3SS1), was reported to target phosphoinositide on host cell membrane and aggravate systemic infection, while its functional relevance and underlying mechanisms have yet to be elucidated. PANoptosis (pyroptosis, apoptosis, and necroptosis) of intestinal epithelial cells (IECs) has been characterized as a pivotal host defense to limit the dissemination of foodborne pathogens, whereas the effect of SopF on IECs PANoptosis induced by Salmonella is rather limited. Here, we show that SopF can attenuate intestinal inflammation and suppress IECs expulsion to promote bacterial dissemination in mice infected with Salmonella enterica serovar Typhimurium (S. Typhimurium). We revealed that SopF could activate phosphoinositide-dependent protein kinase-1 (PDK1) to phosphorylate p90 ribosomal S6 kinase (RSK) which down-regulated Caspase-8 activation. Caspase-8 inactivated by SopF resulted in inhibition of pyroptosis and apoptosis, but promotion of necroptosis. The administration of both AR-12 (PDK1 inhibitor) and BI-D1870 (RSK inhibitor) potentially overcame Caspase-8 blockade and subverted PANoptosis challenged by SopF. Collectively, these findings demonstrate that this virulence strategy elicited by SopF aggregates systemic infection via modulating IEC PANoptosis through PDK1-RSK signaling, which throws light on novel functions of bacterial effectors, as well as a mechanism employed by pathogens to counteract host immune defense.


Introduction
Salmonella enterica serovar Typhimurium (S. Typhimurium) is one of the most common foodborne pathogenic bacteria causing diarrhea in a broad range of hosts. The emergence of antibioticresistant S. Typhimurium represents a major public health issue. 1 On the other hand, S. Typhimurium is also a useful model organism for investigating the mechanisms of host-bacterium interactions. 2 It would be potentially significant to provide new paradigms for interactions between bacteria and host immune response, which would provide novel insights for controlling S. Typhimurium infection and other infectious diseases.
As an enteric pathogen, S. Typhimurium invades into the host initially through intestinal epithelial cells (IECs). Salmonella pathogenicity island-1 type III secretion system (T3SS1) and its effectors counteract potent inflammatory responses, enabling S. Typhimurium to enter into IECs. 3 Intriguingly, SopD exerts both pro-and anti-inflammatory response by targeting the Rab8 GTPase. 4 SopB, SopE, and SopE2 elicit membrane ruffling and fuel cooperative invasion through ADP-ribosylation factor 1 (Arf1) and ras-related C3 botulinum toxin substrate 1 (Rac1) dependent actin polymerization. 5 In contrast, SipA and SipC can directly target actin cytoskeleton to drive cell invasion. 6,7 Recent reports demonstrated an additional role of SopE in autophagy flux inside IECs. SopE modulates autophagy flux by interacting with autophagy regulator SP1 when Salmonella is located within Salmonella-containing vacuole (SCV). 8 SopF, a newly discovered T3SS1 effector, is not involved in bacterial invasion into epitheliums but targets host cell membranes via phosphoinositide (PIP) interactions to maintain the integrity of nascent SCV. 9 Moreover, SopF was required for its replication within HeLa cells and full virulence in mice infected with S. Typhimurium. 10,11 How does SopF, a PIP binding effector, make it easy for bacterial dissemination?
IECs per se not only serve as mechanical barriers to protect the mucosa from pathogen invasion but also coordinate a number of innate immune defenses. 12 One possible mechanism linking IECs to bacterial infection is the process of inflammatory programmed cell death (PCD) pathways. PANoptosis is defined as an inflammatory PCD with key features of pyroptosis, apoptosis, and/or necroptosis. PANoptosis of IECs is known to be critically involved in host defense against S. Typhimurium infection, [13][14][15][16] which is often referred to as a double-edged sword from the standpoint of pathogen and host survival. Pyroptosis and apoptosis drive the expulsion of infected IECs to limit S. Typhimurium replication. 16,17 The dead and extruded IECs elicit shrinkage of the neighboring epitheliums to maintain intestinal barrier integrity. 18 In contrast, necroptosis of IECs contributes to intestinal barrier disruption and facilitates S. Typhimurium spread to lamina propria. 19 It remains to be clarified how S. Typhimurium conquers PANoptosis of IECs for survival.
Caspase-8 was originally identified as an extrinsic initiator Caspase of apoptosis. Active Caspase-8 transmits the death signals by cleaving Caspase-3 and Caspase-7, which subsequently stimulate various intracellular molecules to induce apoptotic cell death. 20 Despite its role in promoting apoptosis, Caspase-8 has also been implicated in the onsets of necroptosis and pyroptosis. [21][22][23] Research over the last decade has clearly demonstrated that Caspase-8 had a rather pro-survival function by suppressing necroptosis. 22 Caspase-8 was able to cleave receptor-interacting protein kinase 1 (RIPK1) and RIPK3 and thus prevented the initiation of necroptosis. 24 Moreover, Gasdermin D (GSDMD), the executor of pyroptosis, could be directly cleaved at aspartate 88 (D88) by Caspase-8, which was important in host defense against bacterial infection. 21 Additionally, evidence showed that autophagosomal membranes can serve as a platform for Caspase-8 activation. 25 Overall, Caspase-8 is tightly regulated for its key role in cell death. Interestingly, an intrinsic mechanism, phosphoinositide-dependent protein kinase-1 (PDK1)-p90 ribosomal S6 kinase (RSK) pathway, was revealed to subvert Caspase-8 blockade of necroptosis. 26 Based on the phosphoinositide targeting function of SopF, we hypothesized that SopF might activate the host PDK1-RSK signaling pathway, and thus affects Caspase-8 function.
Herein, by using a streptomycin-pretreated murine model, we revealed that SopF was required for S. Typhimurium to evade host epithelial defense. Mechanistically, we uncovered that SopF activated PDK1-RSK signaling to inhibit Caspase-8 activation in IECs, which resulted in IECs PANoptosis (halted pyroptosis and apoptosis, but promoted necroptosis) and aggravation of systemic infection.

S. Typhimurium effector SopF inhibits intestinal inflammation to aggravate systemic infection
To investigate the potential role of SopF in pathogenesis of S. Typhimurium, we first monitored mortality and body weight change during the course of infection. Streptomycin pretreatment mice were orally infected with STM-WT, STM-ΔsopF, or STM-ΔsopF/psopF. 27 The individuals had a variety moribund state at 4 dpi, and mice gavaged with STM-WT succumbed to infection on the next day. During the same time 70% of the STM-ΔsopF-infected mice were still alive. In addition, mice infected with STM-ΔsopF/psopF succumbed to infection within 6 d (Figure 1(a)). A gradual decrease in body weight in mice was observed as the infection processed, but without showing any significant difference among the three groups (Figure 1(b)). The accelerated mortality of mice highlights the importance of SopF to pathogenesis of S. Typhimurium in early innate immune responses, which in turn greatly affects the outcome of the infection. Consistent with previous research, 10,11 mice infected with S. Typhimurium carrying sopF had about two fold of bacterial burden in their livers and spleens at 48 hpi compared with those infected with STM-ΔsopF, the marked increase in the bacterial burden is observed at 120 hpi ( Figure 1(c-d)). These data confirm that SopF is required for S. Typhimurium dissemination in extra-intestinal organs.
Intestinal inflammatory response in intestinal tract is important for host-pathogen interaction. 4 We next focus on the effects of SopF on S. Typhimurium-induced intestinal inflammation. Importantly, from 12 hpi, we observed that the (i-k) Gene transcription analysis of Il-6, Tnf-α and Il-1β at 48 hpi and 120 hpi, n = 5 mice/group. Gene expression levels are shown relative to Gapdh. Data were compared by one-way ANOVA. Values are expressed as the means ± SD, and statistically significant differences are indicated. *P < .05, **P < .01, ***P < .001, ns: not significant. cecal content of mice infected with STM-ΔsopF were greatly lost, and the ceca became almost transparent and full of bubbles compared to those infected with S. Typhimurium carrying sopF (Supplementary Figure 1(a)). The contraction of the colon length observed at 48 hpi in mice infected with STM-ΔsopF is greater than those in STM-WT and STM-ΔsopF/psopF infected mice ( Figure 1 (e-f)). We next performed Mayer's hematoxylin and eosin (H&E) analysis, no obvious lesions were observed at 12 hpi for all strains (Supplementary Figure 1(b)). We found progressive pathological lesions including swelling and infiltration of inflammatory cells into the lamina propria and submucosa in mice infected with all S. Typhimurium strains. Importantly, the SopF deficiency exacerbated S. Typhimurium-induced histopathological injury of the ceca, including IEC expulsion at 48 hpi and loss of IECs from the mucosal tissue at 120 hpi ( Figure 1(g-h)). 28 To further characterize the intestinal inflammatory responses presented by SopF, we analyzed the level of inflammatory cytokines that are intimately relevant to host defense against intracellular pathogens. 4 As shown in Figure 1(i-k), the transcription levels of Il-6, Tnf-α and Il-1β were most prominently induced in cecum with the lastingness of infection, and significantly increased expression of cytokines was observed in the ceca of mice infected with STM-ΔsopF compared with those in STM-WT and STM-ΔsopF /psopF infected mice at 120 hpi. Collectively, these results suggest that SopF inhibits intestinal inflammation to aggravate systemic infection in response to S. Typhimurium infection.

S. Typhimurium effector SopF restricts the dislodging of IECs to promote bacterial dissemination
IECs negotiate essential roles in separating the microbial community of the lumen from the sterile systemic milieu. Expulsion of infected IECs restricts the pathogen's intraepithelial proliferation, and serves as a general defense mechanism against enteric pathogen infection. 16,29 To determine the interaction between S. Typhimurium and IEC mediated by SopF, we employed fluorescence microscopy to visualize the integrity of the IEC barrier and the distribution of S. Typhimurium. In accordance with the histopathological examination, staining for the epithelial marker EpCAM was clear and intact at 12 hpi (Supplementary Figure 1(c)). Accordingly, we observed significant loss of total IEC numbers and more epithelial gaps per 10x field of view in mucosal tissue of mice infected with STM-ΔsopF at 120 hpi (Figure 2(a-d)). We further visualized the ultrastructure of ceca with transmission electron microscope (TEM) (Figure 2(b)). Bacteria were seen in the intestinal lumen at the early stage of infection (12 hpi) (Figure 2(b-c)). To assess the possible correlation between IEC expulsion and the onset of bacterial dissemination, we extend these studies by monitoring the time course of the infection. At 48 hpi, anomalies of the brush border structures, such as disordered, fewer, and shorter microvilli, were observed in mice ceca infected with STM-ΔsopF. In contrast, mice infected with S. Typhimurium carrying sopF appeared minor injury in IECs at 48 hpi but significantly higher bacteria load in the submucosa at 120 hpi ( Figure 2  Together, these results suggest that SopF restricts IEC expulsion to promote bacterial dissemination.

S. Typhimurium effector SopF regulates the PANoptosis of intestinal epithelial cells in host defense against S. Typhimurium infection
Although epithelium is primarily considered as a mechanical barrier, studies have indicated that the cell fate of IECs allows them to counteract pathogens. 19,30 We investigated the role of SopF in cell death of IECs in response to S. Typhimurium infection. A reduced release of lactate dehydrogenase (LDH) was found in the human colon carcinoma Caco-2 cells and human normal colonic epithelial NCM460 cells infected with S. Typhimurium carrying sopF, revealing that SopF suppressed the cell death of IECs (Figure 3(a-b)). Recent evidence has shown an interplay between pathogens and hosts with respect to PANoptosis. 31 To assess the effects of SopF on PANoptosis of IECs, we isolated IECs from the ceca of S. Typhimurium-infected mice. Western blot analysis showed a higher protein level of GSDMD-NT fragments, which can form membrane pores to induce pyroptosis, in STM-ΔsopF-infected mice. Epithelial-derived GSDME that is cleaved by Caspase-3 also participates in the pathogenesis of intestinal inflammation. 32 Accordingly, a robust GSDME activation was found in STM-ΔsopFinfected mice (Figure 3(c-d)). The aforementioned results suggest that SopF inhibits pyroptosis of IECs. In addition to pyroptosis, we find an increased cleavage of apoptotic executor caspase, Caspase-3, in mice infected with STM-ΔsopF ( Figure 3(e-f)). Pyroptotic and apoptotic IECs apicaled-out to maintain epithelial barrier integrity, while SopF prevented the elimination of infected IECs. 15,33 Necroptosis of IECs also leads to epithelial barrier dysfunction that promotes S. Typhimurium dissemination. 19 Next, we examined whether SopF was also involved in the necroptosis of IECs. As expected, robust phosphorylation of mixed lineage kinase domain-like protein (MLKL) was observed in mice infected with S. Typhimurium carrying sopF at 12 hpi, suggesting that SopF promotes necroptosis of IECs, which enables the dissemination of S. Typhimurium to lamina propria or even extra-intestinal organs ( Figure 3(g-h)). Similar results were shown in Caco-2 cells (Figure 3(i-n)) and NCM460 cells (Supplementary Figure 2(a-f)). Collectively, these data indicate that SopF modulates PANoptosis by inhibiting apoptosis and pyroptosis but facilitating the necroptosis of IECs, which may relate to the S. Typhimurium systemic infection.

S. Typhimurium effector SopF inhibits the activation of Caspase-8, a molecular switch for PANoptosis
Caspase-8, activated by cell surface death receptor ligation and oligomerization, is a molecular switch for apoptosis, necroptosis, and pyroptosis. 34 Besides, Caspase-8 orchestrate epithelial PCD and prevents barrier dysfunction in response to S. Typhimurium infection. 19 To understand the potential impact of SopF on Caspase-8, we tested the isolated IECs from the ceca of S. Typhimuriuminfected mice. Increased cleavage of Caspase-8 was observed in mice infected with Salmonella lacking sopF compared with those infected with WT or complemented mutant strains at 12 and 48 hpi. (Figure 4(a-b)). Correspondingly, the deletion of sopF resulted in an increased cleavage of Caspase-8 at 24 hpi in Caco-2 cells and NCM460 cells (Figure 4(c-f)). Overall, these data indicate that SopF abrogates the activation of Caspase-8 to modulate IECs PANoptosis.

PDK1-RSK signal is required for Caspase-8 blockade during SopF-mediated S. Typhimurium infection
Caspase-8 may function as a gatekeeper to regulate PANoptosis. Previous work by Yang et al. has shown that the 3-phosphoinositide-dependent protein kinase 1 (PDK1)-ribosomal S6 kinase (RSK) signal is an intrinsic mechanism for diminishing the Caspase-8 blockade. 26 Of note, SopF binds to the phosphoinositide of the host cell membrane to facilitate the stability of nascent Salmonellacontaining vacuole (SCV). 9 Thus, we hypothesized that PDK1-RSK signaling was involved in the SopF-mediated blockade of Caspase-8 activation.
Western blot analysis revealed that the protein levels of PDK1 and phosphorylated RSK were potentially higher in mice infected with S. Typhimurium carrying sopF than those in STM-ΔsopF-infected groups ( Figure 5(a-b)), as well as in Caco-2 cells (Supplementary Figure 3(a, b)) and NCM460 cells (Supplementary Figure 3(c, d)), suggesting that SopF could activate PDK1-RSK pathway to abolish Caspase-8 activation. His or His-SopF, expressed in 293 T cells, was immunoprecipitated and assessed for its ability to interact with endogenous PDK1. Results showed that PDK1 was undetected in His-SopF immunoprecipitated samples ( Figure 5(c)). We then tested whether SopF constructed to establish the infection model. The PDK1 was also undetected in SopF:His immunoprecipitated samples both at early and late stages of infection, suggesting that SopF could not directly interact with PDK1 ( Figure 5(d)). AR-12 and BI-D1870 are ATP antagonists for PDK1 and RSK, respectively. 35,36 The PDK1 inhibitor AR-12 potentially eliminated the increased phosphorylation of RSK but subverted the decreased activation of Caspase-8 presented by SopF at 24 hpi. Moreover, the RSK inhibitor BI-D1870 had the same effect on Caspase-8 in Caco-2 cells infected with S. Typhimurium ( Figure 5(e-g)).
The results indicate that SopF inhibits the cleavage of Caspase-8 through PDK1-RSK signaling pathway. Given the fact that Caspase-8 played a central role in regulating PANoptosis, we further pretreated Caco-2 cells with either AR-12 or BI-D1870. Western blot analysis showed that the protein levels of GSDMD-NT, GSDME-NT, as well as, cleaved Caspase-3 were increased, and the phosphorylation of MLKL was decreased after administration of AR-12 and BI-D1870. No statistical differences were observed among the three infection groups after the treatment of inhibitors. The data indicate that both AR-12 and BI-D1870 restore the halted pyroptosis and apoptosis and the prompted necroptosis ( Figure 5(f-h)).
SopF was reported to decrease LC3 decoration by targeting ATP6V0C in the V-ATPase for ADP-ribosylation at the initial stage of xenophagy. 9 Autophagosomal membranes have been reported to provide a platform for activation of Caspase-8 25 . To our surprise, the administration of Bafilomycin A1, a V-ATPase inhibitor, was unable to restore Caspase-8 blockade presented by SopF, indicating that LC3-II accumulation was dispensable for SopFmediated Caspase-8 inactivation (Supplementary Figure 4(a-d)). Thus, our data demonstrated that PDK1-RSK signaling that was activated by SopF was the upstream of Caspase-8.

Discussion
IECs are the gatekeepers for S. Typhimurium invasion. In addition to acting as a mechanical barrier, IECs also participate in host immune defense. Intestinal pathogens like S. Typhimurium have evolved strategies to counteract this host defense. SopF is a newly discovered effector secreted by Salmonella T3SS1. It was reported that SopF could promote bacterial dissemination in mice. 9 In accordance with previous work, our work validated that mice were susceptible to S. Typhimurium carrying sopF infection, as observed by increased mortality and bacterial loads in systemic organs. SopF exhibited an anti-inflammatory role in the intestinal tract, as visualized by shorter ceca, severe pathological lesions, and higher levels of inflammatory cytokines in mice infected with STM-ΔsopF. Taken together, these results demonstrated that SopF ameliorated gut inflammation and aggravated system infection.
Notably, sopF deletion mutant only partially reduced the bacterial loads in both liver and spleen, indicating that SopF was necessary but not sufficient for bacterial dissemination. Many additional effectors have been demonstrated to contribute significantly to S. Typhimurium systemic infection. [37][38][39] Our previous work showed that another S. Typhimurium virulence determinate, SpvC, attenuated intestinal inflammation to aggravate systemic infection like the effect of SopF in this study. However, the mechanisms were vastly different. 40 SpvC, mainly secreted by T3SS2, was found to interrupt the anti-bacterial function of macrophages. Considering that T3SS1 effectors are generally associated with bacterial entry into non-phagocytes, we next focus on the lesions of IECs affected by SopF. Previous publications highlighted the role of IECs expulsion in restricting bacterial colonization. 15,16 However, this process may help S. Typhimurium complete its infectious cycle and also contribute to boosted recolonization of the enteroid lumen. 41,42 These observations may correlate with other noted mechanisms for luminal population restriction, such as commensal microbiota competition. 43 We observed reduced IEC expulsion but increased bacterial migration into the submucosa via morphological observation in mice ceca infected with S. Typhimurium carrying sopF. These results indicate that SopF-mediated IEC expulsion plays a critical role during S. Typhimurium infection. Inflammatory programmed cell death in intestinal epithelium cells involves in multiple signaling pathways that are activated during the whole stage of S. Typhimurium infection, and the regulation of these signaling pathways has been a useful strategy for S. Typhimurium to escape from the host immune system. 44 It is generally accepted that the PANoptosis (pyroptosis, apoptosis, and necroptosis) of IECs plays a critical role during the course of S.
Typhimurium infection. 15,19,[45][46][47] S. Typhimurium T3SS1 effectors, SipB and SipD, were reported to activate Caspase-3 by impeding the translocation of NF-κB subunit p65 into the nucleus and thereby induce apoptosis. 48 The T3SS2 effectors, SseK1 and Ssek3, could directly target the death domain of TRADD and inhibit necroptotic cell death in Salmonella-infected macrophages. 49,50 Our findings demonstrate that SopF inhibits pyroptosis and apoptosis but promotes necroptosis of IECs in vivo. This notion was supported by data in epithelial cell lines Caco-2 and NCM460 cells. Collectively, our results indicate that SopF manipulates PANoptosis of IECs.
Caspase-8 not only promotes apoptosis but also inhibits necroptosis by suppressing the function of RIPK to activate MLKL, an executor of necroptosis. 24 Recent investigations have discovered additional functions of Caspase-8 for the regulation of inflammation in various ways depending on its catalytic activity and scaffolding role. 51 Taken together, Caspase-8 is defined as the molecular switch for PANoptosis. 24,34 Our data suggest that SopF inhibits Caspase-8 activation to regulate the PANoptosis of IECs.
To ensure the best bacterial growth inside the cell, Salmonella had developed a series of strategies to escape the host immune system. SopB can decrease gut inflammation via activating Akt signaling to inhibit PANoptosis of host cells. [52][53][54] Furthermore, Akt activation was blocked by the presence of PDK1 inhibitor, suggesting that PDK1-Akt signaling pathway was responsible for Salmonella induced cell death. 53 Previous studies have shown that SopF can bind PIP to the host cell membrane, which may lead to the activation of PDK1. 9,55 We found that the protein level of PDK1 was higher in mice infected with STM-WT, as well as increased phosphorylation of RSK, which inhibited Caspase-8 activation. 26 Immunoprecipitation evidence showed that PDK1 was not an interaction partner of SopF, providing additional evidence that SopF targeted PIP to activate PDK1. Accordingly, SopF-mediated Caspase-8 inactivation was restored after the administration of PDK1 and RSK inhibitors. PDK1 phosphorylates Ser221 in the N-terminal kinase domain (NTKD) of RSK, while extracellular signal regulated kinase (ERK) permits Ser380 site phosphorylation in the C-terminal kinase domain (CTKD). 36 As our experiments focus on the role of RSK Ser221, we shall not rule out the possibility of an ERK pathway in mediating RSK phosphorylation. In addition to RSK, AKT and protein kinase C (PKC) are also well-established activation molecules downstream of PDK1. 56 PKC phosphorylated NLRC4 is critical for inflammasome activation in response to S. Typhimurium infection. 57 This may imply that SopF inhibits PDK1 activation, leading to blockade of NLRC4 inflammasome and thus inhibiting pyroptosis of IECs.
Xenophagy, a member of noncanonical autophagy, could serve as host defense to capture intracellular pathogens and subject them to lysosomal clearance. 10 Bafilomycin A1 inhibits V-ATPase acidification independently of its effect on autolysosome fusion, implying that this drug could disrupt both canonical autophagy and xenophagy. 58 In canonical autophagy, the inhibition of autolysosome fusion results in LC3-II accumulation, whereas, the disruption of V-ATPase in xenophagy leads to the inactivation of LC3. LC3-II accumulation in Bafilomycin A1-treated groups indicates that other S. Typhimurium effectors could induce canonical autophagy, which, to our knowledge, might mask the contribution of SopF-blockedxenophagy on Caspase-8. 59 Therefore, a sitedirected mutant strain STM-ΔsopF/psopF Y224D, which abolishes its xenophagy-inhibition activity, is needed to rule out its polar effect on other effectors. On the other hand, studies have revealed that autophagy plays a multiple role in Caspase-8 activation, the interaction between Caspase-8 and autophagy needs further study. 60,61 In summary, we identified a novel function of SopF in regulating IECs PANoptosis to aggravate dissemination of S. Typhimurium, as a model in Figure 6. In this paradigm, SopF, a PIP binding effector, blocked Caspase-8 activation through PDK1-RSK signaling. Owing to its importance in PCD, Caspase-8 sites as a crossroad to regulate PANoptosis. The suppression of Caspase-8 decreased pyroptosis and apoptosis but increased necroptosis. Blocking potential therapeutic targets of either PDK1 or RSK subverted Caspase-8 blockade and the subsequent IECs PANoptosis. SopFinhibited pyroptosis and apoptosis impeded the elimination of infected IECs, and SopF-provoked necroptosis potentiated S. Typhimurium to internalization thus contributing to bacterial dissemination. These findings reveal a novel mechanism whereby SopF manipulates cell fate decisions of IECs, which may provide attractive therapeutic strategies for the control of S. Typhimurium infection and other corresponding diseases.

S. typhimurium infection in vivo and ethics statement
Female C57BL/6 mice (6-8 weeks) were bred and maintained at the experimental animal center of Soochow University. All animal experiments were approved by the Ethics Committee of Soochow University and conducted in accordance with the Guidelines for the Care and Use of Research Animals established by Soochow University. The streptomycin-pretreated mouse model was established as described previously. 40 Mice that fasted for 4 h were administered intragastrically with streptomycin (100 μl of 200 mg/ml solution in sterile water; Sigma, USA). After 24 h, mice were randomly divided into four groups, including control, STM-WT, STM-ΔsopF, and STM-ΔsopF/  psopF infection groups. 100 µL of either 5 × 10 7 colony-forming unit (CFU) S. Typhimurium strains or sterile PBS were used for oral gavage to mice that were fasted for 4 h before infection. Mice were sacrificed using CO 2 asphyxiation at indicated time points. For histopathology, tissue samples were harvested and fixed in 4% paraformaldehyde, processed according to standard procedures for dehydration, paraffin embedding, section cutting, and deparaffinization. The sections were stained with hematoxylin-eosin (Baso, Zhuhai, China) and observed under a light microscope (Olympus, Japan). Total inflammatory scores were assessed based on the following parameters according to previous literature: 28 neutrophil infiltration (0, none; 1, slight increase; 2, marked increase), fibrin deposition, submucosal neutrophil margination, submucosal edema, epithelial necrosis, epithelial ulceration (0, absent; 1, present). The percentage of pathological lesions was counted on a total scale of 0-20.
For bacterial burden measurement, spleen, and liver were harvested and then immersed in 100 µg/ ml of amikacin for 1 h. Tissues were homogenized in PBS containing 0.3% Triton (Sigma, USA) for 30 min. CFU values were quantified by plating lysates with appropriate dilutions onto Salmonella-Shigella agar (Hangwei, China), followed by incubation overnight.
Sample processing for transmission electron microscopy (TEM) was carried out in the School of Biology and Basic Medical Science, Medical College of Soochow University. Ceca samples were fixed in ice-cold 2.5% glutaraldehyde for at least 4 h. Samples were washed twice using 0.1 M phosphate buffer for 15 min at room temperature. Subsequently, the samples were post-fixed in 1% OsO4 for 1 h, dehydrated through an acetone series and embedded in epoxy resin. Ultra-thin sections were stained and observed using a 120 kV Transmission electron microscope (HT7700, Hitachi, Japan).

S. Typhimurium infection in vitro
Cells were seeded in 12-well plates (5 × 10 5 cells/ well) and co-cultured with S. Typhimurium at a multiplicity of infection (MOI) of 100:1. At 1 hpi, cells were washed with PBS three times and incubated further for 2 h with a fresh medium containing amikacin (100 µg/ml, MilliporeSigma, Burlington, MA, USA) to kill the extracellular bacteria. Afterward, cells were washed again, and a fresh medium containing amikacin (10 µg/ml, MilliporeSigma) was added to limit extracellular replication of bacteria. For inhibition of PDK1 activation, cells were pretreated with AR-12 (1 μM; S1106, Selleck, USA) 1 h before infection. For inhibition of RSK phosphorylation, cells were pretreated with BI-D1870 (10 μM; HY-10510, MCE) 1 h before infection. Cells were pretreated with Bafilomycin A1 (100 nM; ab120497, Abcam) 2 h before infection to interfere with autophagy. Proteins were extracted using a radioimmunoprecipitation assay (RIPA, Beyotime Biotechnology) buffer containing protease inhibitors and phosphatase inhibitors (Beyotime Biotechnology).

Lactate dehydrogenase (LDH) release assay
Cells were seeded in 12-well plates (4 × 10 5 cells/ well). The infection model was established as described previously. Cell culture supernatant was collected for detection according to manufacturer's protocol (Beyotime Biotechnology). The absorbance was read at 490 nm with Infinite® F50 Absorbance Microplate Reader (Tecan, Switzerland).

Immunoprecipitation
Caco-2 cells were seeded in 10-cm dishes and cocultured with STM-ΔsopF-pBAD or STM-ΔsopF/ psopF: His at an MOI of 100 for 2 hpi and 24 hpi. Following bacterial infections, cells were washed three times with cold PBS. Protein extracts were incubated with 20 μl protein A/G Plus-Agarose (Santa Cruz Biotechnology, Dallas, TX, USA) for 30 min at 4°C with rotation. The suspension was then incubated with 7 μl His-Tag Monoclonal antibody (66005-1-Ig, proteintech) for 1 h at 4°C with rotation. About 45 μl protein A/G Plus-Agarose was added and incubated at 4°C with rotation overnight. Beads were then washed five times with 1 ml PBS, and bound proteins were eluted by the addition of 20 μl SDS loading buffer. Boil the samples for 5 min and analyze 10 µl aliquots by western blot.

IECs isolation
IECs of cecum were isolated as described previously. 62 In brief, the ceca from the mice were opened longitudinally and washed with PBS to remove intestinal contents. Subsequently, the intestinal segments were incubated in a solution containing 5 mM EDTA, 1 mM DTT, and 5% FBS at 37°C for 15 min. After three times, the medium was collected and centrifuged (1,200 rpm, 5 min) to pellet the cells for subsequent protein extraction.

Quantitative PCR
Total RNA from mouse ceca were isolated using TRIzol reagent (Beyotime Biotechnology, Shanghai, China) according to the manufacturer's protocol and was reverse-transcribed with All-inone RT MasterMix kit (Applied Biological Materials, Richmond, BC, Canada). qPCR was performed using EvaGreen MasterMix-Low ROX (Applied Biological Materials) in a ViiA7 realtime PCR instrument (Applied Biosystems, Carlsbad, CA, USA). Specific primer sequences are listed in Table 2. All expression levels were normalized to Gapdh expression. Values were expressed as fold induction in comparison to untreated control mice.

Statistics
Data are presented as mean ± SD. Comparisons among multiple groups were performed with oneway ANOVA. Survival curves were analyzed with log-rank (Mantel-Cox) test. Values of P < .05 were considered statistically significant.