Cytosolic replication in epithelial cells fuels intestinal expansion and chronic fecal shedding of Salmonella Typhimurium

SUMMARY

Persistence and intermittent fecal shedding, hallmarks of Salmonella infections, are critical for fecal-oral transmission. In the intestine, Salmonella enterica serovar Typhimurium (STm) actively invades intestinal epithelial cells (IECs) and survives in the Salmonella-containing vacuole (SCV) and the cell cytosol. Cytosolic STm replicates rapidly, expresses invasion factors, and induces extrusion of infected epithelial cells into the intestinal lumen. Here, we engineered STm that self-destructs in the cytosol (STm^CytoKill), but replicates normally in the SCV, to examine the role of cytosolic STm in infection. Intestinal expansion and fecal shedding of STm^CytoKill are impaired in mouse models of infection. We propose a model whereby repeated rounds of invasion, cytosolic replication and release of invasive STm from extruded IECs, fuels the high luminal density required for fecal shedding.

Graphical Abstract

eTOC BLURB

Fecal shedding is crucial for transmission of Salmonella among hosts. Using genetic engineering to selectively deplete cytosolic bacteria from enterocytes, Chong et al. show that intracytosolic replication feeds the expansion of intestinal S. Typhimurium needed for efficient shedding.

INTRODUCTION

The model organism Salmonella enterica serovar Typhimurium (STm) is a common cause of gastroenteritis in humans. Chronically infected asymptomatic livestock animals are a major source of infection, with a subset of supershedders likely responsible for most disease transmission (Gopinath et al., 2012; Lawley et al., 2008). Colonization of the intestinal tract by STm is a complex process that requires multiple virulence factors, including flagella, which are required for motility, and the Type III Secretion System 1 (T3SS1), which mediates invasion of intestinal epithelial cells (IECs) (Hapfelmeier et al., 2005; Stecher et al., 2008; Stecher et al., 2004). Following invasion of IECs, the bacteria either remain in the Salmonella-containing vacuole (SCV) or escape into the cytosol, resulting in two distinct intracellular populations (Knodler et al., 2010; Malik-Kale et al., 2012). Whether these populations have distinct roles in pathogenesis is unknown.

To control cytosolic bacteria, IECs activate inflammasomes and eject infected cells into the intestinal lumen, thus limiting replication and tissue penetration by enteric bacteria such as STm (Crowley et al., 2020; Hefele et al., 2018; Kim et al., 2010; Knodler et al., 2014a; Rauch et al., 2017; Sellin et al., 2014). However, STm may exploit this process as an exit strategy. Significantly, cytosolic STm replicate rapidly and express the two hallmarks of invasive STm, flagella and T3SS1, suggesting that bacteria released from extruded IECs are primed to infect nearby cells (Finn et al., 2017; Knodler et al., 2014b; Knodler et al., 2010; Laughlin et al., 2014; Malik-Kale et al., 2012). Indeed, a recent study using gut enteroids suggests that cycles of T3SS1-driven IEC invasion, intraepithelial replication, and reemergence through infected IEC expulsion is a key mechanism for STm luminal colonization (Geiser et al., 2021).

To address whether intracytosolic replication of STm is involved in expansion of the GI population, we engineered self-destructing STm with an autolytic biosensor that is activated in the cytosol of mammalian cells. These bacteria invade epithelial cells normally, and survive and replicate within the SCV, but rapidly lyse in the cytosol. Using this system in acute and persistent murine infection models, we demonstrate that cytosolic replication within IECs facilitates expansion within the GI tract and is required for chronic fecal shedding.

RESULTS

Autolytic elimination of cytosolic STm restricts bacterial expansion and reinfection in cultured epithelial cells

Fluorescent protein expression under the control of the uhpT promoter (PuhpT), which is induced in the presence of extracellular glucose 6-phosphate (G6P), is an established reporter for intracytosolic STm in cultured cells (Chong et al., 2019; Cooper et al., 2017; Finn et al., 2017; Hausmann et al., 2020; Lau et al., 2019; Roder and Hensel, 2020; Spinnenhirn et al., 2014). Therefore, to develop an autolytic strain that self-destructs in the cytosol of eukaryotic cells, we converted the reporter strain STm^CytoLoc (previously referred to as STm pCHAR) by replacing the gfp gene in the plasmid pCHAR with the λ lysis system, comprising the S, R, Rz and Rz1 genes (Bernhardt et al., 2001; Young, 1992). The resultant plasmid, pCytoKill-SRRz, was transformed into wildtype STm SL1344 (STm^WT) producing STm^SRRz. We first assessed the ability of STm^SRRz to invade and replicate within HeLa cells using a gentamicin protection assay and fluorescence microscopy. At 1.5 h pi, the total numbers of intracellular bacteria and the percentage of infected cells were similar to STm^WT, indicating that STmSRRz has no invasion defect (Figure 1A, 1C). However, during peak intracytosolic replication (4–8 h pi) (Knodler et al., 2014b; Malik-Kale et al., 2012), intracellular STm^SRRz were reduced compared to STm^WT or STm^CytoLoc (Figure 1A left panel). Notably, this phenotype was enhanced by treatment with chloroquine (CHQ), which kills vacuolar bacteria (Knodler et al., 2014b)(Figure 1A right panel, S1A) and at 6 h pi, <10% of STm^SRRz were cytosolic compared to ≅ 75% of STm^WT or STm^CytoLoc (Figure 1B). By 12 h pi, cells containing cytosolic bacteria are dying (Malik-Kale et al., 2012), the cytosolic population of STm^WT and STm^CytoLoc start to decrease and the three STm strains are predominantly vacuolar (<10% cytosolic, Figure 1B). Similar results were observed in C2Bbe1 epithelial cells (Figure S1B, S1C). In macrophages, where STm replicates exclusively in the SCV (Beuzon et al., 2002; Thurston et al., 2016), we detected no replication defect of STm^SRRz (Figure S1D–S1E).

Figure 1. Inhibiting STm replication in the cytosol of epithelial cells limits bacterial expansion and reinfection of host cells.

(A) Total and cytosolic intracellular bacterial CFUs assessed by gentamicin protection assay in the absence (−CHQ) or presence (+CHQ) of chloroquine. HeLa cells were infected with the indicated strains. (B) Percent cytosolic STm relative to total intracellular STm calculated from CFUs in (A). (C-E) Infected HeLa cells were fixed and processed for immunofluorescence labelling (αLAMP1 and αCSA) at the indicated times for quantification by fluorescence microscopy. (C) Quantification of infected cells at 1.5 h pi. (D) Quantification of intracellular STm populations in infected cells at 6 h pi. 100 cells/condition were scored in each experiment. (E) Numbers of STm per infected cell from (D). Each symbol represents an infected cell; data are combined from 3 independent experiments. Values at the top indicate total numbers of cells (LAMP1⁺ or LAMP⁻ STm) or fraction of cells that contain >50 bacteria/cell (LAMP1⁺& LAMP1⁻ STm). (F) Bacterial egress at 6 h pi from HeLa cells infected with the indicated strains. (G) Intracellular CFUs at 7 h pi from primary Hela cell infection on inserts used to infect naïve monolayers (left) or from secondary infection of naïve monolayers incubated under reinfection-permissive conditions for 3 h (right). Data are means ± SD from 3 independent experiments (A-G). Statistics (* P<0.05): unpaired 2-tailed Student’s t-test (C, E left panel), or 2- (A, B) or 1-way (F, G) ANOVA with Tukey’s analysis. See also Figure S1.

To assess intracellular replication at the single cell level, infected HeLa cells were stained with antibodies to LPS, to reveal STm, and the lysosomal membrane protein LAMP1, which is a marker for SCVs in HeLa cells (Brumell et al., 2002)(Figure S1A). At 6 h pi confocal imaging revealed three populations of infected cells, containing: only LAMP⁺ bacteria (vacuolar); only LAMP1⁻ bacteria (cytosolic); or both LAMP⁺ & LAMP1⁻ bacteria (cytosolic and vacuolar). Intracellular bacteria were counted, as high numbers (here >50/cell) is an indicator of hyper replicating cytosolic bacteria (Knodler et al., 2014b; Knodler et al., 2010). In cells infected with STm^WT or STm^SRRz, >70% of infected cells contained only LAMP1⁺ bacteria and all these cells contained <50 bacteria (Figure 1D, 1E). The major difference between them was that cells containing only LAMP1⁻ (cytosolic) bacteria were not found in cells infected with STm^SRRz whereas ~10% of STm^WT infected cells did and all, bar one, of these had >50 bacteria. Notably, in cells containing a mix of LAMP⁺ & LAMP1⁻ bacteria there were fewer cells containing >50 bacteria compared to those infected with STm^SRRz (6/62, and predominantly LAMP⁺) compared to STm^WT (31/51, predominantly LAMP⁻) (Figure 1E, S1A). Thus, in cells infected with STm^SRRz, the cytosolic (LAMP1⁻) population of bacteria was almost eliminated whereas the vacuolar (LAMP⁺) population appeared unaffected. Altogether, the above data confirm that STm^SRRz has a defect in cytosolic, but not vacuolar, replication.

Since extrusion of IECs containing cytosolic STm is thought to release invasive T3SS1-induced bacteria (Knodler et al., 2010; Laughlin et al., 2014), we next assessed whether STm^SRRz has a defect in egress and spread to other cells. HeLa cells were infected as above but gentamicin was removed 1 h before harvesting the growth medium and solubilizing cells at 6 h pi. Plating to estimate extracellular CFUs revealed that STm^SRRz were reduced by ~2 logs compared to STm^WT or STm^CytoLoc (Figure 1F). To estimate spread of STm following exit from infected HeLa cells we used a two-monolayer system, infection of the secondary monolayers was reduced by ~1 log compared to STm^WT or STm^CytoLoc (Figure 1G right panel) and correlated with a similar reduction in total intracellular numbers of STm^SRRz in the first infected monolayer at 7 h pi (Figure 1G left panel). Thus, cytosolic STm released from epithelial cells are able to spread to other cells.

Systematic analysis of PuhpT-SRRz induction

PuhpT is strongly induced in intracytosolic, but not vacuolar, bacteria (Chatterjee et al., 2006; Hautefort et al., 2008; Runyen-Janecky and Payne, 2002). However, since either expression of SRRz in environments other than eukaryotic cell cytosol or loss of activity would compromise STm^SRRz for in vivo use, we tested PuhpT activity under a variety of conditions. First, we did a dose response assay to G6P in LB liquid culture. The physiological concentration of G6P in IECs is unknown, but in cultured cells it is in the low mM range (Marin-Hernandez et al., 2011). At 1 μM G6P, there was no indication of PuhpT activity by either fluorescent signal (STm^CytoLoc) or OD600 (STm^SRRz) (Figure S2A). GFP fluorescence was slightly elevated at 10 μM G6P and more so at 100 μM and 1 mM (Figure S2A top row) when ~80–90% of STm^CytoLoc were GFP⁺ (Figure S2B). Similarly, STm^SRRz growth was affected slightly at 10 μM G6P and more dramatically at higher concentrations (Figure 2A, S2A). STm^SRRz showed morphological changes within 0.5 h of G6P addition and by 1 h membrane debris and ghosts were prominent (Figure 2E). In LB PuhpT was not induced by changes in osmolarity or low pH, nor was is activated in SPI2-inducing (low phosphate, low magnesium) media (Figure 2A, S2A, Figure S2C–E).

Figure 2. Systematic analysis of PuhpT-SRRz under various conditions.

Growth of the different STm strains in broth culture monitored by OD₆₀₀ (A, C) or CFUs (B, D). GFP (by STm^CytoLoc) and SRRz (by STm^SRRz and STm^CytoKill) expression are controlled by the G6P-responsive promoter, PuhpT. G6P (100 μM) was added at the start or replenished at 2 and 4 h. Total STm (Strep^R) or plasmid-bearing STm (Carb^R) were plated 6 h post inoculation (B, D). Data are means ± SD from triplicate samples from each of 3 independent experiments. (E, F) Transmission electron micrographs of the indicated STm strains at 0, 0.5 and 1 h post induction with 100 μM G6P. Arrows indicate membrane debris; asterisks indicate bacterial ghosts. Scale bar, 1 μm. See also Figure S2.

Growth inhibition of STm^SRRz by G6P was transient, even with G6P replenishment at 2 and 4 h (Figure 2A). Plating on selective media revealed loss of Carb resistance, indicating that plasmid loss occurred (Figure 2B). Therefore, to ensure faithful maintenance of PuhpT-SRRz, we engineered a strain with a balanced-lethal vector (pIVV3) encoding two essential Salmonella genes (asd, murA) to complement lethal chromosomal mutations (ΔasdΔmurA). Comparison of the autolytic strain, ΔasdΔmurA STm pIVV3-CytoKill-SRRz (hereafter STm^CytoKill), with a plasmid control strain ΔasdΔmurA STm pIVV3 (hereafter STm^IVV3), showed that lysis remained G6P inducible and no plasmid loss was detected (Figure 2C, 2D). Lysis of STm^CytoKill was apparent within 0.5 h of induction, and too extensive by 1 h for sampling (Figure 2F).

Cytosolic STm promote GI tract colonization in an acute infection model

To test the role of intracytosolic replication in vivo, we used an acute colitis model where cytosolic replication of STm in IECs and sloughing of these cells into the cecal lumen has been described (Crowley et al., 2020; Knodler et al., 2010). To verify PuhpT induction in the GI tract, streptomycin pretreated C57BL/6 mice were orally infected with the fluorescent reporter STm^CytoLoc (Figure 3A, 3B). At 1 d pi, GFP⁺ bacteria were observed only within IECs (cytokeratin 8⁺) that contained >10 bacteria. No GFP⁺ bacteria were detected within the cecal lumen, even though it contained the majority of STm (Figure 3A, auto-fluorescent objects in the lumen, indicated with open arrows, are larger than Alexa Fluor 568 labeled STm). In contrast, at 2 d pi, GFP⁺ STm were predominantly associated with sloughed TUNEL⁺ cytokeratin 8 stained IECs in the lumen (Figure 3B, S3A, S3B). Thus, PuhpT is induced in cytosolic STm within IECs. To test whether PuhpT-dependent expression of SRRz has an impact on intestinal colonization, mice were infected with an equal mixture of STm^WT with either STm^IVV3 or STm^CytoKill. In the intestinal tract (ileum, cecal tissue, cecal contents, feces) there was a marked enrichment of STm^WT over the autolytic strain STm^CytoKill at 2 (Figure 3C) and 4 d pi (Figure 3D), but no discernible difference in spread to the spleen. In contrast, when mice were infected with STm lacking the Salmonella Pathogenicity Island 1 (SPI1)-encoded T3SS1 (ΔSPI1), which do not efficiently invade IECs, but grow readily within the gut lumen and can disseminate to the spleen in C57BL/6 mice (Hapfelmeier et al., 2005; Vazquez-Torres et al., 1999), there was no enrichment of the ΔSPI1 parental strain compared to ΔSPI1^CytoKill in either the intestinal tract or spleen at 2 d pi (Figure S3D). Furthermore, in an ex vivo approach, STm^CytoKill showed no survival defect, compared to STm^IVV3, when incubated in cecal contents from either uninfected or STm^WT infected mice (24 h pi) (Figure S3C). Altogether, these data demonstrate that replication in the cytosol of IECs is required for efficient colonization of the intestinal tract in a murine model of acute colitis.

Figure 3. Cytosolic STm promote GI tract colonization in an acute infection model.

(A, B) Confocal images of cecal tissue from C57BL/6 mice infected (n=2) with STm^CytoLoc (~1×10⁵ CFU/mouse). Cecum sections were immunostained with αCSA to detect all STm (red), αGFP to detect cytosol exposed STm (green) and αCytokeratin-8 to label epithelial cells (CK8; grey). (A) At 1 d pi, luminal STm^CytoLoc are GFP⁻ (left panels); intracellular bacteria are GFP⁺ (cytosolic) or GFP⁻ (vacuolar, right panels). White arrows indicate STm. Open arrows indicate non-specific αGFP labeling of luminal debris. (B) At 2 d pi, luminal STm^CytoLoc are GFP⁺. Scale bar, 20 μm; inset scale bar, 5 μm. See Figure S3A for a larger field image. (C, D) Competitive index in C57BL/6 mice infected orally (~2×10⁸ CFU/mouse) with a 1:1 mixture of STm^WT and STm^IVV3 or STm^CytoKill at 2 d pi (C, n=10/group) and 4 d pi (D; STm^IVV3, n=7; STm^CytoKill, n=9). Limited to no cecal contents due to pronounced inflammation at 4 d pi; CFUs were determined from ceca with contents (D). Geometric means are indicated by the red lines. Each dot represents data from 1 mouse. Statistics (* P<0.05): Wilcoxon rank sum test. See also Figure S3.

Cytosolic STm promote fecal shedding in persistently infected mice

To investigate if intracytosolic replication is required to amplify intestinal loads in persistent infections, genetically resistant 129X1/SvJ mice were orally infected with STm^IVV3or STm^CytoKill and fecal shedding was monitored up to 66 d pi. STm were enumerated by plating on streptomycin (SL1344 is Str^R) and carbenicillin (pIVV3 confers Carb^R) plates, and SRRz activity in STm^CytoKill was confirmed in LB containing G6P (Figure 4 and S4, Table S1 and S2). In two independent experiments, done 14 months apart, mice infected with STm^CytoKill had reduced fecal shedding compared to those infected with STm^IVV3 (Figure 4A–C, S4A, S4B, S4E). In the first experiment, no STm^CytoKill shedding (n=19) could be detected after 20 days whereas 4 out of 19 mice infected with the STm^IVV3 continued shedding for >50 days. In the second experiment, STm^CytoKill shedding was detected in 2 mice after 50 d (n=17) compared to 8 (n=16) for STm^IVV3. In the combined data, fewer mice shed STm^CytoKill vs STm^IVV3 at both early (9 d pi, 22 vs 67%) and late time points (51 d pi, 5 vs 31%) (Figure 4D, S4F). The levels of STm^CytoKill recovered from the spleen, mLN and cecum (tissue and contents) at the end of the experiment were also lower than STm^IVV3 (Figure S4C, S4D). Altogether, these results show that cytosolic replication contributes to intestinal expansion in persistently infected mice and is important for sustained fecal shedding of STm.

Figure 4. Cytosolic STm promote fecal shedding.

129X1/SvJ mice were infected (~1×10⁸ CFU/mouse) with STm^IVV3 or STm^CytoKill. Fecal levels of plasmid-bearing STm (Carb^R). (A, B) Shedding patterns of individual mice in experiment 1 (n=19/group) and experiment 2 (STm^IVV3, n=16; Stm^CytoKill, n=17). (C, D) Shedding levels (C) and percentage of shedders (D) at early (9 d pi) and late (51 d pi) time points from the 2 experiments combined (STm^IVV3, n=35; STm^CytoKill, n=36). ND, not detected. Statistics (*P<0.05): Fisher’s exact test (C) or log-rank test (D). See also Figure S4.

DISCUSSION

IECs are located at the interface of the intestinal lumen and the mucosal immune system where they have an essential role in protection against enteric pathogens such as STm. They limit replication of intracytosolic bacteria by mobilizing inflammasomes to initiate cell death, ultimately ejecting the infected cell, and consequently the pathogen, into the lumen of the intestine (Crowley et al., 2020; Knodler et al., 2014a; Sellin et al., 2014). Yet, extrusion of infected IECs into the lumen is not necessarily a dead end for the intracellular bacteria if they can escape the cell and invade neighboring cells or reseed the lumen for fecal shedding/transmission (Geiser et al., 2021). Here, we have shown that the intracytosolic replicative niche in IECs serves as an amplifier for STm replication in the GI tract, which contributes to bacterial expansion and fuels sustained fecal shedding.

To assess the role of cytosolic replication, we used a conditional lethal approach based on the G6P-activated uhpT promoter (Chatterjee et al., 2006; Hautefort et al., 2008; Runyen-Janecky and Payne, 2002). Our data show that autolysis is induced in cytosolic, but not vacuolar, STm and is dependent on G6P under a variety of in vitro culture conditions. An important question is whether it can also be induced in the lumen of the gut where lysis of STm could significantly impact colonization and persistence. Although G6P is the intracellular form of glucose, release from extruding IECs, particularly in the inflamed gut, could cause transient localized increases in the lumen of the gut. Nevertheless, neither in vivo nor ex vivo control experiments revealed extracellular autolysis and the preponderance of data indicate that it is the intracellular, specifically cytosolic, population of bacteria that are depleted using this system.

How does the cytosolic population of STm in IECs contribute to expansion in the GI tract? Firstly, since cytosolic STm are primed for motility and invasion, expressing both flagella and T3SS1, release via extrusion of infected epithelial cells could lead to multiple rounds of invasion of adjacent IECs while also seeding the lumen (Finn et al., 2017; Geiser et al., 2021; Hausmann et al., 2020; Knodler et al., 2010; Laughlin et al., 2014). Secondly, in the streptomycin treated acute infection model, release of T3SS1-induced STm into the lumen should facilitate inflammation, which is exploited by STm to outcompete the resident microbiota in the lumen of the GI tract (Behnsen et al., 2014; Raffatellu et al., 2009; Stecher et al., 2012; Stecher et al., 2007; Thiennimitr et al., 2011; Winter et al., 2010). Thirdly, we have shown previously that cytosolic STm use the T3SS1 effectors, SopB and SipA, to delay cell death and increase intracytosolic loads of STm (Chong et al., 2019; Finn et al., 2017). Interestingly, Shigella, another enteric pathogen that replicates in the cytosol of IECs and causes inflammatory colitis, uses a T3SS effector homologous to SopB to delay cell death (Ashida et al., 2021). The fact that STm and Shigella use the same mechanism, activation of the pro-survival kinase Akt, to prolong host cell survival indicates that bacterial amplification via cytosolic replication is an essential component of intestinal inflammatory disease.

In the persistent infection model, unlike the streptomycin treated acute model, the native microbiota is intact at the time of infection and inflammation is less prominent, except in supershedder mice, which develop colitis (Gopinath et al., 2013; Lawley et al., 2008; Monack et al., 2004). STm lacking functional T3SS1 have pronounced defects in shedding and transmission although still able to persist at both mucosal and systemic sites (Lam and Monack, 2014; Lawley et al., 2008). Our results suggest that the requirement for T3SS1 in shedding is due to its role in perpetuating the cycle of invasion, cytosolic replication and release of invasive STm. Furthermore, we propose that in supershedders, where there is significant inflammation, this T3SS1-dependent cycle actively fuels massive luminal expansion similar to what occurs in acute infection. Additional studies are needed to test this hypothesis and determine how innate immunity, including specific inflammasomes and autophagy, and the intestinal microbiota influence the outcome.

Host-pathogen interactions are complex, in part because of heterogeneity in pathogen populations in different organs, subpopulations within organs and, as addressed here, within individual host cells. Single cell transcriptome approaches can provide information about the different populations, but it is difficult or impossible to determine the roles of specific populations in pathogenesis. By taking a direct approach, in which the population of interest was specifically eliminated, we discovered what a profound effect a small population can have on the course of infection. In the acute infection model, less than 1% of cecal STm are in the tissue, and an even smaller number are in the cytosol of IECs, yet this dynamic population plays a vital role in establishing and maintaining the intestinal population. This powerful approach could be used to study other populations, such as STm within SCVs in different cell types, assuming that specific inducer-promoter pairs can be identified or developed by genetic engineering.

STAR METHODS

RESOURCE AVAILABILITY

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Olivia Steele-Mortimer (OMortimer@niaid.nih.gov).

Materials Availability

Plasmids and bacterial strains generated in this study are available upon request from the Lead Contact.

Data and Code Availability

This study did not generate datasets/code.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Bacterial Strains and Mutant Construction

Bacterial strains used in this study are listed in the Key Resources Table. The bacteriophage λ Red recombinase system was used to make the mutant strain SL1344 asdA::P_ara c2 FRT/P_murA::P_ara FRT (SL1344 ΔasdAΔmurA) with primers listed in Table S3. The mutant strain SL1344 SPI1::kan/asdA::P_ara c2 FRT/P_murA::P_ara FRT (SL1344 ΔSPI1ΔasdAΔmurA) was made by P22 transduction of the SPI1::kan locus from SL1344 SPI1::kan (Drecktrah et al., 2006) into SL1344 asd::pBAD C2 FRT/PmurA::pBAD FRT (SL1344 ΔasdAΔmurA). To generate asd::TT araC P_ara c2, the chloramphenicol resistance cassette flanked by FRT sites (FRT-cat-FRT) was first amplified from pKD3 using pKD3/4 KpnI 1F and PKD3/4 NdeI 1R, and then cloned into the KpnI and NdeI sites of pCR2.1-Topo, creating pTopo-cat. The rrnB TT araC P_ara c2 fragment was amplified from χ11215 (Kong et al., 2012) using asd insert NdeI 1F and asd insert NotI 1R, and the amplified product cloned into the NdeI and NotI sites of pCR2.1-Topo encoding FRT-cat-FRT, generating the template plasmid pTopo-cat-asd. FRT-cat-FRT rrnB TT araC P_ara c2 was then amplified from pTopo-cat-asd with primers containing flanking homology using asd INS 1F and asd INS 1R, digested with DpnI to remove residual template plasmid and electroporated into SL1344 harboring pKD46. Transformants were recovered in outgrowth media containing 100 μg/mL diaminopimelic acid (DAP, Millipore Sigma), then plated on selection media containing 30 μg/mL chloramphenicol (Millipore Sigma) and 100 μg/mL DAP. Mutants were verified using PCR analysis and dependence on DAP for growth. To construct P_murA::araC P_ara, the araC P_ara fragment was amplified from χ11215 (Kong et al., 2012) using PmurA insert HindIII 1F and PmurA insert SpeI 1R and cloned into the pTopo-cat HindIII and SpeI sites. P_murA::araC P_ara was amplified with primers containing flanking homology using PmurA INS 1F and PmurA INS 1R, digested with DpnI to remove residual template plasmid and electroporated into SL1344 harboring pKD46. Transformants were recovered in outgrowth media containing 0.2% L-arabinose, then plated on selection media containing 30 μg/mL chloramphenicol and 0.2% L-arabinose (Millipore Sigma). Mutants were verified by PCR analysis and dependence on L-arabinose for growth.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Goat polyclonal anti-Salmonella CSA-1	KPL	Cat# 5310–0322; RRID 2890923
Mouse monoclonal anti-LAMP1, Clone H4A3	Developmental Studies Hybridoma Bank	Cat# H4A3; RRID: AB_2296838
Rabbit polyclonal anti-GFP, Alexa Fluor 488-conjugated	Thermo Fisher Scientific	Cat# A-21311; RRID: AB 2217477
Rabbit monoclonal anti-Cytokeratin 8 (EP1628Y), Alexa Fluor 405-conjugated	Abcam	Cat# ab210139; RRID: AB 2890924
Rabbit monoclonal anti-Cytokeratin 8 (EP1628Y)	Abcam	Cat# ab53280; RRID: AB 869901
Bacterial and Virus Strains
STmWT, S. Typhimurium SL1344 wildtype (StrepR)	Steele-Mortimer lab strain collection	Hoiseth and Stocker, 1981
STmΔSPI1, S. Typhimurium SL1344 SPI1::kan (KanR)	Steele-Mortimer lab strain collection	Drecktrah et al., 2006
S. Typhimurium SL1344 ΔasdAΔmurA (StrepR)	This study	N/A
S. Typhimurium SL1344 ΔSPI1ΔasdAΔmurA (StrepRKanR)	This study	N/A
STmCytoLoc, STmWT bearing pCHAR (StrepRCarbR)	This study	N/A
STmSRRz, STmWT bearing pCytoKill-SRRz (StrepRCarbR)	This study	N/A
STmIVV3, SL1344 ΔasdAΔmurA bearing pIVV3 (StrepRCarbR)	This study	N/A
STmCytoKill, SL1344 ΔasdAΔmurA bearing pIVV3-CytoKill-SRRz (StrepRCarbR)	This study	N/A
ΔSPI1IVV3, SL1344 ΔSPI1ΔasdAΔmurA bearing pIVV3 (StrepRKanRCarbR)	This study	N/A
ΔSPI1CytoKill, SL1344 ΔSPI1ΔasdAΔmurA bearing pIVV3-CytoKill-SRRz (StrepRKanRCarbR)	This study	N/A
Biological Samples
Peripheral blood	National Institutes of Health Clinical Center	N/A
Chemicals, Peptides, and Recombinant Proteins
2,6-Diaminopimelic acid	Sigma-Aldrich	Cat# 07036
D-Glucose 6-phosphate sodium salt	Sigma-Aldrich	Cat# G7879
Chloroquine diphosphate salt	Millipore-Sigma	Cat# C6628
Critical Commercial Assays
Dynabeads untouched human monocytes kit	Thermo Fisher Scientific	Cat# 11350D
DeadEnd Fluorometric TUNEL System	Promega	Cat# 3250
Deposited Data
Experimental Models: Cell Lines
HeLa human cervix carcinoma epithelial cell line	ATCC	CCL-2
C2BBe1 human colorectal adenocarcinoma epithelial cell line	ATCC	CRL-2102
Experimental Models: Organisms/Strains
C57BL/6	In-house colony	N/A
C57BL/6	Jackson Laboratory	Cat# 000664
129SvJ/X1	Jackson Laboratory	Cat# 000691
Oligonucleotides
Primers used in this study are listed in Table S3	This study	N/A
Recombinant DNA
pKD3	Datsenko and Wanner, 2000	N/A
pKD46	Datsenko and Wanner, 2000	N/A
pCR2.1-Topo	Thermo Fisher	N/A
pCP20	Cherepanov and Wackernagel, 1995	N/A
pMPMA3DPlac	Ibarra et al., 2010	N/A
pCHAR (aka pPuhpT-gfp)	Finn et al., 2017	N/A
pCytoKill-SRRz	This study	N/A
pIVV3	This study	N/A
pIVV3-CytoKill-SRRz	This study	N/A
Software and Algorithms
Zen 2.3 SP1	Zeiss	N/A
Prism 8.0.1	GraphPad	https://www.graphpad.com/
R	The R foundation	https://www.r-project.org
Other

The asdA::P_ara c2 and P_murA::P_ara mutations were transduced by P22 either back to SL1344^WT or to transfer between strains followed by selection on chloramphenicol. Chromosomally integrated chloramphenicol resistance cassettes were subsequently removed by FLP1 mediated recombinase using pCP20.

Bacterial Growth Conditions

Bacteria were grown on LB agar supplemented with the appropriate antibiotics and stored at 4°C for up to 1 week. Plasmids were introduced by electroporation and selected with 50 μg/mL carbenicillin (Millipore Sigma). Overnight cultures were grown by inoculating one colony into 2 mL LB-Miller (LB-M) broth with selective antibiotics and incubated at 37°C with shaking at 225 rpm for 16–18 h. For cell infections, strains were sub-cultured under SPI1-inducing conditions to late-log phase: 1:33 dilution into 10 mL LB-M without antibiotics, for 3.5 h at 37°C with shaking at 225 rpm. For animal experiments, strains were grown for 18 hours at 37°C with shaking at 225 rpm to stationary phase in LB–M broth supplemented with antibiotics. Bacteria were pelleted and diluted in sterile pharmaceutical grade saline (SPGS) to the appropriate concentration based on OD₆₀₀. The inoculum was verified by plating ten-fold serial dilutions on LB plates for CFUs.

Mammalian Cell Culture

Human cervical carcinoma cells (HeLa, ATCC CCL-2) were maintained at 37°C in 5% CO₂ in complete medium containing Eagle’s minimal essential medium (MEM, Mediatech) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (Thermo Fisher Scientific), 2 mM L-Glutamine (Thermo Fisher Scientific) and 1 mM sodium pyruvate (Thermo Fisher Scientific). Human colorectal adenocarcinoma epithelial cells (Caco-2 subclone C2BBe1, ATCC CRL-2012) were grown in Dulbecco’s minimal essential medium (DMEM, Mediatech) supplemented with 10% (v/v) heat-inactivated FBS, 4 mM L-glutamine and 10 μg/mL human transferrin (Sigma). Cells were seeded at 4.5 × 10⁴ cells/well (HeLa) or 5 × 10⁴ cells/well (C2Bbe1) in 24-well plates 20 h prior to infection. C2Bbe1 cells were seeded onto collagen coated plates. Cells were passaged as recommended by ATCC and used for experiments within 15 passages of receipt.

Apheresis enriched human peripheral blood monocytes were obtained from peripheral blood provided by the Department for Transfusion Medicine and the National Institutes of Health Clinical Center at the National Institutes of Health (Bethesda, MD). Signed, informed consent was obtained from each donor, acknowledging that the donation would be used for research purposes by intramural investigators throughout the National Institutes of Health. Samples were maintained anonymously. Human monocytes were isolated by negative selection (Dynabeads Untouched human monocytes kit, Thermo Fisher Scientific) and differentiated into macrophages as described previously (Lathrop et al., 2018). Monocytes were plated (day 0) in complete medium containing RPMI 1640 medium (Gibco), 1 mM sodium pyruvate, 1× MEM nonessential amino acids, 10 mM HEPES buffer, 2 mM glutamate, 5% (v/v) heat-inactivated human male AB serum (Millipore Sigma) and 100 ng/ml human recombinant macrophage colony-stimulating factor (M-CSF, PeproTech), and incubated at 37°C in 5% CO₂. On days 3 and 5, half the volume of the cultures was replaced with fresh complete RPMI, 5% human serum, and 200 ng/ml M-CSF. Cells were used for experiments on day 7. Monocytes were plated in 96-well flat bottom plates (Corning) at 4 × 10⁴ cells/well for the gentamicin protection assay or 8-well optically clear plastic chambered coverslips (Ibidi) at 8 × 10⁴ cells/well for immunofluorescence microscopy. Each experiment was conducted with cells prepared at different times using different donors.

Experimental Animals

Animal experiments in this study adhered to the Guide for Care and Use of Laboratory Animals, 8^th Edition (National Research Council). The animal protocols were reviewed and approved the Rocky Mountain Laboratories Animal Care and Use Committee (Animal Protocol numbers 2016-035-E and 2019-037-E). Male and female 6–8-week-old C57BL/6J (C57BL/6) mice (Jackson Laboratories or in-house colony) and 7–9-week-old 129X1/SvJ mice (Jackson Laboratory) were used in these studies. Mice were housed up to 5 animals per cage under specific pathogen-free conditions in filter-top cages that were changed bi-weekly. Except when specified, food and water were provided ad libitum. Infected mice were monitored daily for signs of clinical illness.

METHOD DETAILS

Plasmid Construction

Plasmids are listed in the Key Resources Table. Primer and synthetic gene fragment (gBlock) sequences are listed in Table S3. Primers and gBlocks were purchased from Integrated DNA Technologies (IDT). Standard cloning techniques were used to generate the plasmids used in this study.

pCytoKill-SRRz – The PuhpT-SRRz-TT gBlock, composed of the S. Typhimurium uphT promoter (−158 to −1 bp from start codon), the lysis cassette (genes S, R, Rz, Rz1) from λ phage (NC_001416.1) and a transcriptional terminator (synTT, Bba_B0015, parts.igem.org), was subcloned into the NotI and XhoI sites of pMPMA3ΔPlac.

pIVV3 - A second synTT was amplified using NotI 1F and SacI 1R; the amplicon was cloned into the NotI and SacI sites of pMPMA3ΔPlac TT. The SL1344 murA gene region including a 667 bp 5’ region, the murA ORF and 182 bp 3’ of the translational stop codon was amplified using murA comp BamHI 1F and murA comp NotI 1R; the amplicon was cloned into the BamHI and NotI sites. The SL1344 chromosomal asd gene region including a 449 bp 5’ region, the asd ORF and 144 bp 3’ of the translational stop codon amplified using asd comp PsiI 1F and asd comp SacI 1R; the amplicon was cloned into the PsiI and SacI sites. The murA and asd regions were sequence verified.

pIVV3-CytoKill-SRRz - The uphT promoter (−158 to −1 bp from start codon) and the lysis cassette (genes S, R, Rz, Rz1) from lambda phage (PuhpT-SRRz) was subcloned into the PstI and XhoI sites of pIVV3.

Infection of Mammalian Cells

Bacteria were grown under SPI1-inducing conditions (as described in Bacterial Growth Conditions), pelleted (8,000 x g, 2 min) at room-temperature, re-suspended in an equal volume of Hanks’ Balanced Salt Solution without Ca²⁺/Mg²⁺ (HBSS, Mediatech), and diluted in either complete MEM (for HeLa cell infections), complete DMEM (for C2Bbe1 cell infections) or RPMI (for macrophage infections). Mammalian cells were infected following established protocols (Finn et al., 2017; Lathrop et al., 2018). For all infections, media were prewarmed to 37°C, and cells were incubated at 37°C in 5% CO₂. HeLa and C2Bbe1 monolayers were infected at a MOI of ~50. Human macrophages were infected at a MOI of ~15. At 10 min pi, extracellular bacteria were removed by washing with HBSS with Ca²⁺/Mg²⁺ and cells were incubated in antibiotic-free complete medium until 0.5 h pi. At 0.5 h pi, medium was exchanged for complete medium containing L-Histidine (500 μg/mL) and gentamicin (50 μg/mL) and incubated for another 1 h. For infected HeLa cells, medium was exchanged for complete medium containing L-Histidine (500 μg/mL) and gentamicin (10 μg/mL) at 1.5 h pi; for infected macrophages, medium was exchanged at 45 min pi.

Gentamicin Protection and Bacterial Exit Assays

Cells were infected as described above. For the gentamicin protection assay, culture medium was removed from wells, infected cells were lysed in 1 mL of 0.2% (w/v) sodium deoxycholate (DOC; Millipore Sigma) in PBS at the indicated time points, and serial dilution plated on LB agar for CFUs. For the chloroquine (CHQ) resistance assay, infected monolayers were treated with 400 µM (Millipore Sigma) chloroquine 1 h prior to lysis in 0.2% DOC and dilution plating. For the exit assay, gentamicin-containing medium was removed from wells, replaced with complete medium without antibiotics, and incubated for an additional 1 h at 37°C in 5% CO₂. Supernatants containing exited bacteria were collected, and serial dilutions plated on LB agar for CFUs.

Reinfection Assays

The first HeLa cell monolayers were seeded at 1 × 10⁴ cells/insert in 100 μL on inverted polyester membrane tissue culture-treated 24-well inserts with 0.4 μm pore size (Corning) 20 h prior to infection. After 1 h at 37°C in 5% CO₂, the inserts were flipped and incubated in wells with complete MEM. For infection, the inserts were inverted and incubated for 10 min with 50 μL of medium containing SPI1-induced bacteria at a MOI of ~50. The inserts were flipped into wells for washes and media changes after removal of the inoculating medium as described above. At 5 h pi, the inserts were washed once in antibiotic-free complete MEM, then moved into wells with fresh HeLa cell monolayers in complete MEM containing L-Histidine (500 μg/mL). The secondary HeLa cell monolayers were seeded at 4.5 × 10⁴ cells/well in 24-well plates 24 h prior to infection. The infected inserts were incubated with the secondary monolayers for 2 h. The inserts were then removed, rinsed twice with PBS, treated with 100 μL 0.2% (w/v) DOC in PBS to lyse infected cells and serial dilutions plated.

The medium in the wells were replaced with complete medium containing L-Histidine (500 μg/mL) and gentamicin (50 μg/mL), and incubated for another 1 h. The medium was then removed from the wells, treated with 1 mL of 0.2% (w/v) DOC in PBS to lyse infected cells and serial dilutions plated.

Plate reader growth and GFP reporter assays

For glucose-6-phosphate titration, overnight cultures of STm were diluted 1:30 in fresh LB-Miller without or with supplemented D-glucose-6-phosphate sodium salt (G6P, Millipore Sigma) to the final concentrations indicated (1 μM, 10 μM, 100 μM or 1 mM). For the G6P replenishment assay, 100 μM G6P was added to the 96-well plate at the indicated timepoints (0, 2, and 4 h). At the end of the experiment (6 h), total (Strep^R) or plasmid-bearing (Carb^R) CFUs from the cultures were determined by dilution plating. For plate reader assays under SPI2 inducing conditions, overnight cultures of STm were diluted 1:30 in low phosphate, low magnesium-containing medium (Coombes et al., 2004). To assess growth under varying osmolarity, overnight cultures of STm were diluted 1:30 in fresh LB containing 0.1 M, 0.2 M or 0.4 M NaCl. To assess growth at different pH, overnight cultures were diluted in LB-M to an OD₆₀₀ of 0.1 and grown at 37°C with shaking at 225 rpm until they reached an OD₆₀₀ of 0.3. Samples were then centrifuged to pellet the bacteria, resuspended in LB-M at pH 7, pH 5.8 or pH 3 and incubated at 37°C with shaking at 225 rpm for 15 minutes. Post incubation, samples were pelleted, resuspended in LB-M. Subcultures were aliquoted in triplicate (200 μL/well) into 96-well plates, and grown with shaking at 37°C for the indicated duration. OD₆₀₀ and/or GFP fluorescence (Ex 478 nm, Em 515 nm) were measured every 10 min in a Tecan Infinite 200 Pro plate reader.

Flow Cytometry

Plate-reader growth assays with or without supplemented G6P were set up as described above. Samples were harvested and prepared for flow cytometry at 0, 2 and 4 h as previously described (Cooper et al., 2017). STm^CytoLoc were stained with 10 μM Syto41 (ThermoFisher) and processed on a BD LSR Fortessa flow cytometer (BD Bioscience). Data were analyzed with FlowJo software (Tree Star). Samples were gated on Syto41⁺ events and the % of GFP⁺ events measured.

Transmission Electron Microscopy

Strains were sub-cultured under SPI1-inducing conditions to late-log phase as described in Bacterial Growth Conditions, then induced with 100 μM G6P and sampled (0.5 mL) at the indicated times. Samples were centrifuged at 3000 x g for 5 min at room temperature. The cell pellets were inactivated and fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer. All subsequent TEM processing steps were carried out in a PELCO BioWave Pro laboratory microwave (TedPella Inc) at 250 Watts. The bacterial pellets were rinsed with buffer and post fixed with 1% osmium tetroxide reduced with 0.8% potassium ferrocyanide in sodium cacodylate buffer. After water washes, the pellets were treated with 1% tannic acid and stained with uranyl acetate replacement (UAR) stain. The pellets were dehydrated in gradient ethanol series and infiltrated with 1:1 ethanol:Spurr’s resin. Subsequently, the cells were infiltrated with 100% resin and polymerized overnight in a 65°C oven. The bacterial pellets were sectioned using a Leica UC6 ultramicrotome (Leica Microsystems) and observed using a Hitachi 7500 TEM at 80 kV (Hitachi High-Technologies). Images were captured with a Hamamatsu Orca digital camera (Advanced Microscopy Techniques).

Immunofluorescence Staining

HeLa cells were seeded on acid-washed glass coverslips at 4.5 × 10⁴ cells/well in 24-well plates 20 h prior to infection. Monocytes were seeded at 8 × 10⁴ cells/well in 8-well optically clear plastic chambered coverslips (Ibidi). Cells were infected as described above. At the indicated times post infection, infected cells were fixed in 2.5% PFA in PBS for 10 min at 37°C. Fixed cells were washed three times with PBS, permeabilized and blocked for 45 min in 0.1% (w/v) saponin, 10% (v/v) normal donkey serum in PBS (PBS-SS). Cells were then incubated for 1 h sequentially in primary and secondary antibodies with three PBS washes between antibody incubations. Primary and secondary antibodies were diluted in PBS-SS. Stained cells were washed three times with PBS, incubated with 4′,6-Diamidino-2-phenylindole (DAPI) (300 nM in water) for 5 min, washed twice with water and then either mounted onto glass slides (coverslips) or overlaid (Ibidi chambers) with a Mowiol solution supplemented with 2.5% (w/v) DABCO. Primary antibodies used were goat anti-Salmonella CSA1 (KPL, 1:300) and mouse anti-LAMP1 (Developmental Studies Hybridoma Bank, clone H4A3,1:500). Alexa Fluor conjugated secondary antibodies (1:400) were purchased from Thermo Fisher Scientific.

Immunohistochemistry

Streptomycin pretreated C57BL/6J mice were orally inoculated with 1 × 10⁵ CFU of STm^CytoLoc and euthanized as described above (2 mice/time point). Immediately after euthanasia, mice were transcardially perfused with 5–10 mL of SPGS with heparin sodium (100 U/mL, Sagent Pharmaceuticals), then with 5–10 mL of 4% (w/v) paraformaldehyde (PFA). Ceca were harvested, fixed in 10% neutral buffered formalin for a minimum of 24 h (Leica Biosystems), paraffin embedded and sectioned (5 μm thick). Immunostaining of cecum sections was performed as previously described (Bauler et al., 2017). Prior to labelling, the Alexa Fluor 488-conjugated rabbit anti-GFP antibody (Thermo Fisher Scientific, 1:400) was pre-adsorbed with PFA-fixed goat anti-CSA labelled STm^WT for 1 h at room temperature to reduce cross-reactivity with the CSA antibody. Sections were blocked (2% donkey serum, 1% bovine serum albumin, 0.1% Triton X-100, 0.05% Tween-20 in PBS) for 0.5 h, and stained first with the anti-GFP antibody overnight at 4°C, followed by the goat anti-Salmonella CSA (1:400) and Alexa Fluor 405-conjugated rabbit anti-cytokeratin 8 (Abcam, 1:200) for 2 h at room temperature. Alexa Fluor 568–conjugated donkey anti-goat secondary antibody was used (1:400; Thermo Fisher Scientific). Samples were mounted with ProLong Gold (Thermo Fisher Scientific).

TUNEL staining for apoptotic cells in formalin fixed paraffin embedded (FFPE) cecal tissue sections was performed using the DeadEnd Fluorometric TUNEL System (Promega) according to the manufacturer’s instructions with minor changes. All steps were performed in a humidity chamber. Briefly, the slides were deparaffinized and immunofluorescence stained for Cytokeratin 8 and Salmonella. Alexa Fluor 568–conjugated donkey anti-rabbit and Alexa Fluor 647–conjugated donkey anti-goat secondary antibodies (1:400; Thermo Fisher Scientific) were used. The slides were washed in PBS, incubated with equilibration buffer for 10 minutes at room temperature, then incubated with recombinant Terminal Deoxynucleotidyl Transferase (rTdT) incubation buffer for 1 h at 37°C. The reaction was stopped by submerging slides in 2X SSC buffer for 15 minutes. Positive control slides were prepared according to the manufacturer’s instructions by treating FFPE slides of cecum from naïve mice with 10 U/ml DNAse I (Sigma-Aldrich) for 10 minutes at room temperature prior to TUNEL staining. For negative controls, FFPE slides of the cecum from streptomycin treated STm^WT infected mice (36 h pi) were processed like the sample slides, but omitting the primary antibodies for the immunostaining and exchanging the rTdT enzyme in the rTDT incubation buffer for dH₂O. Slides were mounted using ProLong Glass antifade mountant containing NucBlue nuclear stain (Thermo Fisher Scientific) and allowed to cure. TUNEL stain image acquisition parameters were set according to the signal in positive control slides.

Fluorescence Microscopy

Images were captured with a Zeiss LSM 710 confocal laser-scanning microscope equipped with a Plan Apochromat 63X/1.4 N.A. objective. The large tiled overview images (Figure S3A and S3B) were acquired using a Plan Apochromat 20X/0.8 N.A. objective. Z-slices were processed into maximum intensity projections and tiled overview images were stitched together with Zen v.2.3 SP1, and assembled using Adobe Photoshop CC. Composite and single channel (Figure 3A, 3B, S1A, S3A, S3B) or composite only (Figure S1E, S3B controls) images are shown.

Mouse Infections

For oral infections, water and food were withdrawn 3–4 h before each gavage. C57BL/6J mice were used for the acute colitis model studies - 20 mg of streptomycin dissolved in 100 μL sterile water was administered by oral gavage (o.g.). Twenty-four hours after streptomycin treatment, the mice were inoculated with STm by o.g. For immunohistochemistry, 2 mice/group were infected with ~ 10⁵ CFU in 100 μL of SPGS and euthanized at 1 or 2 d pi. Ceca were harvested, fixed and processed for immunohistochemistry. A lower inoculating dose was used for the imaging study to limit intestinal tract damage from inflammation induced by STm infection. For the competitive index experiments, ~2 × 10⁸ CFU in 100 μL of SPGS of a 1:1 mixture of the indicated STm strains by o.g. Competitive indices were calculated by dividing the CFUs of plasmid-bearing STm recovered by the CFUs of STm^WT recovered or plasmid-bearing ΔSPI1 recovered by the CFUs of STm^ΔSPI1 recovered, which was then divided by the same ratio from the inoculum.

For the persistent infection studies, 129X1/SvJ mice were infected by o.g. with ~1 × 10⁸ CFU in 100 μL of SPGS. Individual mice were distinguished by ear tags. Fecal pellets were collected and plated 1 day prior to infection to ensure the animals were not colonized by streptomycin or carbenicillin resistant microflora. Twenty mice were infected with each strain in each experiment (Figure 4A, 4B, S4A and S4B). In experiment 1 (Figure 4A, S4A), 1 animal from the control group (STm^IVV3) was excluded due to the presence of carbenicillin resistant fecal colonies at −1 d pi. One STm^CytoKill infected mouse was excluded because it expired from unrelated causes before the end of the experiment. In experiment 2 (Figure 4B, S4B), 4 STm^IVV3 and 2 STm^CytoKill infected mice expired from unrelated causes before the end of the experiment, and colonies recovered from 1 STm^CytoKill infected mouse did not respond to G6P induction. Data from these animals were excluded.

For shedding studies, one fecal pellet/mouse was collected into pre-weighed tubes containing 3–4 sterile 2 mm zirconia beads (BioSpec Products) and 0.5 mL SPGS, weighed, homogenized with a Bead Mill 24 (Fisher Scientific) at 4.85 m/s for 20 seconds, and ten-fold serial dilutions plated for CFUs on LB agar plates containing the appropriate antibiotic. Mice that shed at >10⁴ CFU/gram of feces for 2 consecutive collection time points were singly housed for the duration of the experiment. At the indicated time points, mice were euthanized by isoflurane overdose and cervical dislocation. For CFU determinations, tissues (blood, spleen, liver, mesenteric lymph nodes, ileum, cecum with contents, and fecal pellets) were collected in pre-weighed tubes containing 3–4 sterile 2 mm zirconia beads and 0.5 mL SPGS, weighed, and homogenized before plating ten-fold serial dilutions. For tissue invasion assays (Figure 3C, 3D, S3D, S4C, S4D, S4H), cecal contents were removed and cecal tissue washed once in SPGS, incubated in SPGS containing 100 μg/mL gentamicin for 1 h at room temperature and washed twice with SPGS. We were unable to collect cecal contents for bacterial loads at 4 d pi due to pronounced cecal inflammation. Instead, bacterial loads were determined from the cecum containing cecal contents (Figure 3D). The gentamicin treated tissues and the luminal cecal contents were collected into pre-weighed tubes containing 3–4 sterile 2 mm zirconia beads and 0.5 mL SPGS, weighed, homogenized, and dilutions plated. All homogenates were plated on LB agar with 100 μg/mL streptomycin and LB agar with 50 μg/mL carbenicillin. After enumeration, representative colonies from the feces and tissues of each mouse were replica plated onto either streptomycin or carbenicillin containing LB agar for verification. Colonies on carbenicillin replica plates from STm^CytoKill infected animals were assessed for G6P-induced lysis in broth cultures by plate reader as described above.

Ex vivo cecal content growth

Cecal contents were harvested from uninfected C57BL/6 forty-eight h after streptomycin treatment, uninfected 129X1/SvJ, or 129X1/SvJ infected for 24 h by o.g. with ~10⁸ CFU STm^WT (n=10/group). Cecal contents were harvested from 10–15 C57BL/6 mice that were streptomycin pretreated then infected for 24 h with ~10⁸ CFU STm^WT (o.g.). For each condition, the contents were divided into two portions, placed in sterile 50 mL conical tubes (Corning), weighed and inoculated at ~1 × 10⁷ CFU/g with either STm^IVV3 or STm^CytoKill. Inocula were prepared from stationary phase STm cultures and diluted in SPGS (~1 × 10⁸ CFU/mL). After inoculation, the tubes were capped, the contents vortexed for 1 min, and incubated at 37°C. Duplicate samples from each pool were removed at 0, 3, 6, and 9 h into pre-weighed tubes containing 3–4 sterile 2 mm zirconia beads and 0.5 mL SPGS, weighed, homogenized, and dilutions plated on LB agar with 100 μg/mL streptomycin and LB agar with 50 μg/mL carbenicillin.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical analyses were performed using Prism (GraphPad) or R (Figures 4A and S4A) (R Core Team, 2019). Statistical parameters are reported in the figures and corresponding legends. P values of <0.05 were considered significant. Differences are not significant unless noted.

Supplementary Material

Table S1

Table S1 (Related to Fig 4 and S4). Growth phenotypes of colonies recovered from STm^CytoKill infected mice.

Table S2

Table S2 (Related to Fig 4 and S4). Growth phenotypes of colonies recovered from STm^IVV3.

Inclusion and Diversity.

We worked to ensure sex balance in the selection of non-human subjects. One or more of the authors of this paper self-identifies as an underrepresented ethnic minority in science. One or more of the authors of this paper self-identifies as a member of the LGBTQ+ community. One or more of the authors of this paper self-identifies as living with a disability. The author list of this paper includes contributors from the location where the research was conducted who participated in the data collection, design, analysis, and/or interpretation of the work.

HIGHLIGHTS.

• A lethal biosensor restricts S. Typhimurium intracytosolic growth in epithelial cells

• Intracytosolic replication facilitates the expansion of S. Typhimurium loads in the gut

• Sustained fecal shedding of S. Typhimurium is fueled by intracytosolic replication