Salmonella Invasion Is Controlled by Competition among Intestinal Chemical Signals

Salmonella utilizes specific compounds from its local environments as signals to regulate the expression of its invasion genes (6–10, 15, 22, 23). The environment of the murine distal ileum, the preferred site of Salmonella invasion, is rich in the metabolite formic acid (8). Additionally, formic acid was shown to induce Salmonella invasion gene expression, but the mechanism remains undetermined (8). We hypothesized that formic acid induces invasion gene expression by preventing repressors, such as short- and long-chain fatty acids, from quelling the expression of Salmonella invasion genes. To test this, we compared invasion gene expression in Salmonella preincubated with formic acid followed by treatment with intestinal fatty acids known to repress this function. Using a plasmid-borne luxCDABE reporter fusion to the invasion regulator hilA, we measured light production over a time course experiment to quantify hilA expression. We found that in the absence of formic acid preincubation, hilA expression was severely repressed by intestinal fatty acids such as oleic acid and propionic acid but that preincubation with formic acid significantly restored hilA expression (by ~40 and ~6.5 fold, respectively) (Fig. 1A to C). cis-2 hexadecenoic acid (c2-HDA), a potent invasion repressor found in the murine cecum and the colon, drastically reduced hilA expression. However, preincubation with formic acid restored hilA expression significantly (~7.5 fold). To determine whether control of this regulator had effects on SPI-1, we quantified the expression of a downstream and terminal invasion gene, sipC, using a sipC::lacZY reporter fusion. Salmonella strains were preincubated with formic acid prior to addition of fatty acids, and sipC expression was quantified by β-galactosidase assays. As expected, sipC expression was repressed by c2-HDA and oleic acid. In Salmonella that was preincubated with formic acid, however, sipC expression was restored significantly (~5.6 fold and ~5 fold, respectively) (Fig. 1D). These results conclusively show that formic acid preincubation prevents intestinal fatty acids from repressing Salmonella invasion-gene expression.

Salmonella invasion is highly restricted in the murine large intestine (21). Therefore, we reasoned that the formic acid concentrations in the cecum and colon may be lower than that of the ileum and thus inadequate to prevent the fatty acids from repressing invasion genes. To test this, we measured the concentration of formic acid in various regions of the murine intestine. The contents of the ileum, cecum, and colon of C57BL/6 mice (n = 10) were extracted, and an enzymatic assay was performed to assess formic acid concentrations. We found that the ileum had a significantly higher concentration of formic acid than both the cecum (~7 fold) and the colon (~12 fold) (Fig. 1E). Similar results were obtained in a second mouse strain, BALB/c, and with different age groups (age 6 weeks and age 10 weeks), thus showing that the high localized concentration of formic acid is limited to the murine ileum (see Fig. S1 in the supplemental material). These results may explain, at least in part, how Salmonella invasion is targeted to the ileum, the most distal portion of the small intestine, and reduced in the large intestine.

These results imply that as Salmonella moves from the ileum toward environments with lower concentrations of formic acid, the invasion-rescuing effect of formic acid is lost. As the cecum and the colon have lower formic acid and higher fatty acid concentrations (14), the sequential exposure to these chemical cues may be utilized by Salmonella as a signal to switch from induction to repression of invasion. To test this hypothesis, we analyzed invasion gene expression in Salmonella under various conditions, using a hilA::luxCDABE reporter fusion. We found that when Salmonella cultures were preincubated with formic acid followed by addition of fatty acids, formic acid significantly rescued the repression by fatty acids. However, when formic acid and fatty acids were added together or when fatty acids were added prior to formic acid, no substantial rescue from repression was observed (Fig. 1F, Fig. S2). These results thus show that the sequence of exposure to different signals plays a crucial role in their effects on Salmonella invasion gene expression and supports a model of spatial and temporal invasion regulation dictated by the changing chemical environment of the gut.

Salmonella imports exogenous formic acid via the formic acid transport channel protein FocA, encoded by focA (24, 25). To determine whether formic acid acts as a cytoplasmic signal to rescue invasion, we tested a focA null mutant of Salmonella carrying the previously described hilA-luxCDABE reporter plasmid. We found that preventing formic acid import through the deletion of focA significantly impaired the ability of formic acid to rescue invasion-gene repression by c2-HDA (Fig. 2A, Fig. S3). Thus, the import of exogenous formic acid is essential to prevent the repression of invasion genes. Salmonella also produces endogenous formic acid by metabolizing pyruvate using the enzyme pyruvate formate lyase, encoded by pflB. Deleting pflB prevents Salmonella from producing endogenous formic acid (8). To determine if endogenous formic acid plays a role in the rescue of Salmonella invasion genes, we constructed a hilA-luxCDABE reporter strain with a pflB null mutation and compared the rescue of invasion gene expression in ΔfocA and ΔpflB mutants. As shown in Fig. 2A, both ΔfocA and ΔpflB null mutants were repressed efficiently by c2-HDA. However, preincubation with formic acid rescued invasion gene expression by ~5 fold in the ΔpflB mutant, while there was no effect in the ΔfocA mutant. Additionally, despite formic acid preincubation, the ΔfocA ΔpflB double mutant strain remained repressed by c2-HDA, showing that only the imported formic acid is sufficient to rescue the expression of invasion genes. These results show that the production of endogenous formic acid by Salmonella is insufficient to rescue invasion genes, requiring instead exogenously supplied formic acid to achieve this effect.

Salmonella metabolizes formic acid by using either the formate dehydrogenase O or N system, encoded by fdoGHI and fdnGHI, respectively (8). Formic acid can also be converted to carbon dioxide and water by fdhF-encoded formate dehydrogenase (8). It is therefore possible that formic acid itself is the chemical essential to rescue invasion-gene expression or that one or more of its metabolic products serves that role. To test this, we prevented formic acid metabolism by constructing strains with null mutations in fdnG, fdoG, and fdhF, eliminating all known mechanisms of formic acid utilization in Salmonella (8). Using the hilA-luxCDABE reporter system, we compared invasion gene repression with fatty acids, with and without formic acid preincubation. We found that formic acid was able to rescue c2-HDA-mediated repression in both the wild-type and the ΔfdnG ΔfdoG ΔfdhF mutant Salmonella (Fig. 2B, Fig. S4), demonstrating that formic acid metabolism is not required for its ability to rescue invasion gene expression. In fact, the inability to metabolize formic acid enhanced hilA expression in the ΔfdnG ΔfdoG ΔfdhF mutant Salmonella strain, as would be expected if formic acid were itself the signal. Together, these results show that formic acid supplied by the environment is imported and utilized by Salmonella to prevent its invasion genes from repression by fatty acid signaling molecules present in the intestine.

HilD is the master transcriptional regulator of Salmonella invasion genes, with most of the environmental signals that control the expression of Salmonella invasion genes doing so through HilD (7, 9, 15). We have previously shown that intestinal fatty acids directly target HilD to control Salmonella virulence by binding to this essential regulator and preventing its ability to interact with its DNA targets (7, 9, 15). We thus hypothesized that formic acid, consisting of a single carboxylate group, may competitively bind to HilD and thereby prevent repressive fatty acids from interacting with this protein. We initially tested this hypothesis by assessing the effect of formic acid and fatty acids on invasion gene expression mediated solely through HilD. We constructed a Salmonella strain with null mutations in hilC and rtsA and placed hilD under a tetracycline-inducible promoter. Upon induction of hilD by tetracycline, hilA expression was induced, and this expression was reduced by fatty acids (Fig. S5). Preincubation with 10 mM formic acid, however, rescued this fatty acid-mediated repression, demonstrating the importance of hilD to this mechanism of regulation. To test the mechanism by which these signals control hilD, we analyzed the binding of fatty acids to the protein HilD with or without preincubation with formic acid using an enzyme-linked immunosorbent assay (ELISA). Fatty acids adherent to polystyrene plates were incubated with His-tagged HilD with or without prior preincubation with formic acid, and binding was detected using an anti-His antibody. As expected, HilD directly bound to c2-HDA (Fig. 3A). However, preincubation of HilD with formic acid significantly reduced its ability to bind to c2-HDA (dissociation constant [K_d], 5.19 μM and 81.98 μM, respectively). We investigated this further by checking the binding of various classes of fatty acids having different carbon lengths and unsaturations. The trans isomer of c2-HDA with the same carbon length, t2-HDA, acted similarly, binding efficiently to HilD but poorly to HilD preincubated with formic acid (K_d, 15.78 μM and 177.1 μM, respectively) (Fig. 3B). A longer cis-2 unsaturated fatty acid, c2-EA with 20 carbons, also bound to HilD but not to the formic acid preincubated protein (K_d, 96.59 μM and 381.1 μM, respectively) (Fig. 3C). Oleic acid, a known intestinal fatty acid of 18 carbons and no cis-2 unsaturation, also showed a similar binding pattern (K_d, 136.3 μM and 1373 μM, respectively) (Fig. 3D). Thus, formic acid preincubation prevented a wide range of fatty acids with different structures from interacting with HilD. This suggests that formic acid may competitively bind to a region of HilD that is required to bind to a chemical moiety common to all fatty acids.

We have previously shown that specific amino acids of HilD are required to bind specific classes of fatty acids (15). Our current results show that formic acid preincubation prevented a range of fatty acids with various chemical structures from binding to HilD (Fig. 3A to D). We therefore tested whether formic acid could prevent the interaction of these repressive signals with HilD even though they utilize different protein binding moieties. We first tested a HilD^Q290A mutant, which can bind to fatty acids harboring a cis-2 unsaturation, such as c2-HDA and c2-EA, but fails to bind ligands such as oleic acid and t2-HDA that lack this structure (15). We found that HilD^Q290A directly bound to c2-HDA but failed to do so in the presence of formic acid (Fig. 3E). In contrast, the HilD^K293A mutant cannot bind fatty acids with a cis-2 unsaturation but does bind those that lack this structure. We found that HilD^K293A directly bound to oleic acid but lost that ability when the protein was preincubated with formic acid (Fig. 3F). These results conclusively show that formic acid competitively inhibits the interaction of HilD with all classes of fatty acids tested, to function as a derepressor.

HilD binds in the hilA promoter region, initiating transcription and thus a cascade of invasion-gene expression, while repressive fatty acid signals prevent this binding (15). We reasoned that competitive binding by formic acid would thus allow the interaction of HilD with its DNA target even in the presence of these repressive signals. To test this, we performed an electrophoretic mobility shift assay (EMSA), assessing the effects of repressive fatty acids on the migration of a hilA DNA fragment in the presence of HilD with or without preincubation with formic acid. HilD bound to the hilA DNA, impeding its migration, while the fatty acids c2-HDA, t2-HDA, and oleic acid disrupted the HilD-hilA complex (Fig. 4A). However, preincubation of HilD with formic acid prevented this disruption, restoring the HilD-hilA complex. Additionally, formic acid prevented the disruption of the HilD^Q290A-hilA complex by c2-HDA and that of the HilD^K293A-hilA complex by t2-HDA (Fig. 4B and C). These results show that formic acid prevents all classes of fatty acids from disrupting the ability of HilD to bind to its target DNA. Together, they support a model by which Salmonella can utilize a microbial metabolite commonly found in the intestine to prevent a wide range of repressive signals from interacting with its principal invasion regulator.

Growth of Salmonella in the presence of repressive fatty acids leads to a drastic reduction in the HilD half-life, which halts invasion gene expression (15, 22). As formic acid prevents fatty acids from binding to HilD, we reasoned that preincubation with formic acid may increase the stability of HilD in the presence of these fatty acids, thus sustaining expression of invasion genes. To test this hypothesis, we compared the stability of HilD in growing bacterial cultures, in the presence of fatty acids, with or without preincubation with formic acid. We used a Salmonella strain expressing hilD under the control of a tetracycline-inducible promoter and with a 3×FLAG tag that allows the quantification of HilD using Western blots. Salmonella cultures were grown with tetracycline to induce hilD expression, and equal numbers of bacteria were then preincubated with or without formic acid before the addition of repressive fatty acids. New protein production was stopped by using a cocktail of transcription- and translation-inhibiting antibiotics, and the stability of HilD was analyzed at subsequent time points. As shown in Fig. 5, treatment with c2-HDA led to rapid degradation of HilD, reducing its half-life from ~227 min to a mere ~9 min (Fig. 5B). Preincubation with formic acid, however, significantly increased the stability of HilD and extended its half-life from ~9 to ~139 min. Similar results were obtained with the stereoisomer t2-HDA, where formic acid preincubation increased the half-life of HilD significantly (from ~28 min to ~127 min) in the presence of t2-HDA (Fig. 5C). These results thus demonstrate a precise molecular mechanism by which Salmonella may exploit the ileal constituent formic acid to shield its primary invasion regulator HilD from interacting with repressive fatty acids, thus increasing its stability, and sustaining its DNA-binding ability in the presence of these compounds, thereby sustaining invasion gene expression.

Our results show that Salmonella imports exogenous formic acid to sustain invasion-gene expression in the presence of repressive fatty acids. As the murine ileum is rich in formic acid, we hypothesized that Salmonella may utilize the formic acid present as a bona fide intestinal signal to derepress invasion in the ileum. To test this, we compared invasion gene expression in the murine ileum of a Salmonella wild-type strain with that of a focA mutant that is unable to import formic acid. To quantify gene expression, we used Salmonella strains with a green fluorescent protein (GFP) reporter fused to the promoter of the terminal invasion gene sicA (PsicA→GFP). Expression of invasion genes was prevented in cultures prior to animal inoculation, thus allowing us to assess their induction within the murine intestine. C57BL/6 mice were orally inoculated with equal numbers of bacteria, and after 90 min mice were euthanized, and bacteria were recovered from the ileum and analyzed by flow cytometry to evaluate sicA induction (Fig. 6A). We found that the proportion of the wild-type Salmonella population recovered from the ileum that demonstrated detectable sicA expression was ~44%, compared to ~0.5% in the inoculated culture (Fig. 6B), demonstrating clear induction of expression within the murine intestine. The ΔfocA mutant, however, showed a significantly lower capacity to induce gene expression in the ileum, with sicA induced in only ~11% of its population (~4-fold less than the wild type) (Fig. 6C). In fact, invasion gene expression in the ΔfocA mutant was not significantly different from that of a ΔhilD null mutant or of a control strain lacking the PsicA→GFP fusion, demonstrating complete repression due to the lack of focA. To ensure that these effects were indeed due solely to the loss of focA, we complemented the mutant using a plasmid constitutively expressing focA, finding that it restored sicA expression to the level of the wild type (~45%). These results show that Salmonella imports the formic acid present in the murine distal ileum and utilizes it as a signal to induce invasion genes in that region.

Short- and long-chain fatty acids that repress Salmonella invasion exist in high concentrations within the large intestine. We thus hypothesized that the lower concentration of formic acid present in this organ may not be sufficient to counter these repressive effects, thus providing a plausible mechanism of invasion induction. To test this hypothesis, we preincubated Salmonella with a lower concentration of formic acid (1 mM) prior to adding relatively higher concentration of fatty acids (40 μM c2-HDA) and assessed sipC gene expression. We found that formic acid preincubation was unable to rescue sipC repression caused by c2-HDA under these conditions (Fig. S6). As expected, both the focA mutant and the complemented strain remained repressed. These results show that the import of formic acid can prevent repression only when fatty acid concentrations are relatively lower in the surrounding environment.

All experiments were performed using Salmonella enterica subsp. enterica serovar Typhimurium 14028s or E. coli BL21 DE3 unless otherwise noted. All strains, plasmids and primers used in the study are listed in Table 1 and Table 2 respectively. Fatty acids c2-EA (93772-82-8), oleic acid (90260), and t2-HDA (11132) were purchased from Cayman Chemicals. c2-HDA was purchased from both Cayman Chemicals (11133) and Larodan (10-1609). Propionic acid was purchased from Sigma (W292400). All antibiotics were used at their standard concentrations (chloramphenicol, 25 μg/mL; kanamycin, 50 μg/mL; ampicillin, 100 μg/mL) unless otherwise mentioned.

Salmonella strains with deletions in various genes were constructed as previously described (16, 22). Using plasmid pKD4, the DNA region encoding resistance against kanamycin was amplified by PCR with a 40-bp homology extension flanking focA. The PCR fragment was electroporated into a strain expressing λ Red recombinase (36). Loss of focA was confirmed using PCR and sequencing. Mutations were unmarked by the pCP20 recombinase protocol (37). P22 transduction was used to transfer constructs among strains (38).

Salmonella strains carrying luxCDABE reporter plasmids were grown overnight in 5 mL LB broth with appropriate antibiotics. These were subcultured (1:100) in M9 minimal medium with 0.2% glucose and antibiotics for 16 h. Equal numbers of bacteria (optical density at 600 nm [OD₆₀₀] of 0.02) were inoculated into LB broth containing 100 mM MOPS (morpholinepropanesulfonic acid), pH 6.7, and antibiotics with 10 mM formic acid or left untreated. These were incubated for 3 h at 37°C in an incubator without aeration. Dimethyl sulfoxide (DMSO) or different concentrations of fatty acids were added to the reaction, and replicate aliquots were introduced into the BioTek Synergy H1 microplate reader. Luminescence and OD₆₀₀ were recorded every half-hour for 24 h. Data were analyzed in GraphPad Prism. For every treatment, the area under the curve (AUC) was calculated. DMSO was normalized to 100%, and other reactions were calculated accordingly.

Salmonella strains carrying lacZY fusions were preincubated with 1 or 10 mM formic acid for 3 h and grown in 1 mL LB broth with 100 mM MOPS, pH 6.7, antibiotics, and different concentrations of fatty acids or an equal volume of DMSO, without aeration for 16 to 18 h. The β-galactosidase assay was performed using the method of Miller, as previously described in detail (39). Miller units were calculated using the following equation: Miller units = (OD₄₂₀ × 1,000)/(OD₆₀₀ × t × v), where t is the time (minutes) the reaction was allowed to continue before adding the stop solution, v is the volume of culture (mL) assayed, and OD₄₂₀ and OD₆₀₀ are the densities of the cultures at 420 and 600 nm, respectively. All reactions were performed in quadruplicates. Data were analyzed in Excel and plotted in GraphPad Prism.

Mice were dissected, and distal ileum (~6 cm above the ileocecal junction), whole cecum, and colon (~6 cm from the end of cecum) were collected. The contents were extracted in 1 mL of PBS, vortexed vigorously, and centrifuged at 5000 × g for 10 min to precipitate debris. Supernatants were collected for measuring formic acid concentration. These were measured using an enzymatic bioanalysis detection kit, following the manufacturer’s protocol (R-Biopharm). This method uses the enzymatic conversion of formic acid and NAD⁺ to bicarbonate and NADH through the action of the enzyme formate dehydrogenase. The NADH concentration is then measured at 340 nm to calculate the formic acid concentration. A standard curve was plotted before the experiment to correlate concentrations and optical densities.

Assays were performed as previously described in detail in Chowdhury et al. (15, 16). Briefly, 96-well polystyrene plates coated with various concentrations of fatty acids in coating buffer (0.05 M sodium carbonate-bicarbonate, pH 9.3) were incubated overnight at 4°C. The wells were washed three times with PBS, followed by the addition of 200 μL blocking buffer (1% Ficoll 400 in PBS). After 2 h of blocking at room temperature, the wells were washed three times with PBS, and different concentrations of proteins, preincubated with 500 μM formic acid, in blocking buffer were added and incubated at room temperature for 2 h. The wells were washed, incubated with a 1:1,000 dilution of anti-His tag antibody, and incubated at room temperature for an hour. After washing, the wells were incubated with a 1:5,000 dilution of anti-mouse secondary antibody for an hour. Finally, the wells were washed three times with PBS-T (PBS plus 0.1% Tween 20), and 100 μL OPD (o-phenylenediamine) substrate was added. After color development, the reaction was ended with 1 N sulfuric acid and read spectrophotometrically at 490 nm. Data were analyzed in GraphPad Prism. Nonlinear regression was plotted, and one-site total binding was performed to calculate K_d values.

Assays were performed as previously described (15, 16). HilD, HilD^Q290A, or HilD^K293A (200 μM) was preincubated with either 500 μM formic acid or an equal volume of PBS for 1 h. hilA promoter DNA (–286 to +31) (50 nM) was added to the reaction in binding buffer (100 mM Tris, pH 7.5, 10 mM EDTA, 1 M KCl, 1 mM dithiothreitol [DTT], 50% vol/vol glycerol). Various concentrations of the fatty acids (c2-HDA, 20 μM; t2-HDA, 40 μM; or oleic acid, 50 μM) or control solvents were added, and the reaction was incubated at room temperature for 1 h. Dye solution (10 mM Tris, 1 mM EDTA, 50% vol/vol glycerol, 0.001% wt/vol bromophenol blue) was added to a final concentration of 1×, and samples were separated on 7% acrylamide gels. DNA was stained using SYBR Safe and visualized in a Bio-Rad GelDoc imaging system.

Assays were performed as previously described (22). Salmonella strains carrying a chromosomal tetracycline-inducible hilD with a C-terminal 3×FLAG tag and a hilA::lacZY fusion (for invasion gene expression assessment) were grown overnight in LB and then diluted 1:100 into LB buffered with 100 mM MOPS (pH 6.7) with 2.5 μg/mL tetracycline. This concentration of tetracycline induced hilD to wild-type levels. After 3 h of growth, cultures were treated with 10 mM formic acid or left untreated and incubated at room temperature for 3 h without aeration, followed by fatty-acid treatment. After 1 h of incubation at 37°C with aeration, samples were collected to measure hilA::lacZY expression in a β-galactosidase assay. Individual cultures were then equilibrated to an OD₆₀₀ of 1. New protein production was halted by adding an antibiotic combination (100 μg/mL rifampin, 200 μg/mL streptomycin, and 50 μg/mL spectinomycin). These cultures were incubated at 37°C with aeration, and 200-μL aliquots were collected every 30 min for 4 h. Samples were lysed in 5× Laemmli buffer, boiled, and immunoblotted with an anti-FLAG antibody to monitor HilD levels.

hilD alleles were cloned into pET15b(+), expressed, and purified as described in detail previously (15). Briefly, the pET15b-6XHis-hilD or pET15b-6XHis-hilD^Q290A or pET15b-6XHis-hilD^K293A plasmid was transformed into E. coli BL21(DE3) for expression. Transformants were induced with 5 mM IPTG (isopropyl β-d-1-thiogalactopyranoside) for 4 h. Cultures were centrifuged, lysed, treated with TALON beads, and applied to protein purification columns. Bound protein was eluted and confirmed by SDS-PAGE. Following confirmation, eluted samples were dialyzed overnight. The final protein concentration and quality after dialysis were assessed by Bradford estimation and SDS-PAGE.

Studies involving vertebrate animals were approved by the Cornell University Institutional Animal Care and Use Committee (protocol 2012-0074). Euthanasia was performed using carbon dioxide inhalation in accordance with the American Veterinary Medical Association Guidelines for Euthanasia of Animals. The Cornell University Animal Care and Use program and associated animal facilities are operated in accordance with the U.S. Department of Agriculture Animal Welfare Act (1966), Regulation (C.F.R., 2009), and policies, the Health Research Extension Act (1985), the Public Health Service Policy on Humane Care and Use of Laboratory Animals (PHS, 2002), the Guide for the Care and Use of Laboratory Animals (NRC, 2011), the Guide for the Care and Use of Agricultural Animals in Research and Teaching (2010), the New York State Health Law (Article 5, Section 504), and other applicable federal, state, and local laws, regulations, policies, and guidelines.

Salmonella strains, either wild type or with different null mutations, carrying a constitutive blue fluorescent protein (BFP) only (phoN::BFP, as negative control) or with a GFP reporter fusion to the promoter of invasion gene sicA (PsicA→GFP), were grown overnight in LB broth with 1 mM nonanoic acid (to stop the expression of invasion genes). On the day of the experiment, they were subcultured 1:100 in LB broth with 1 mM nonanoic acid and grown to an OD₆₀₀ of 1. Equal volumes of cultures were precipitated, washed twice, and resuspended in 1 mL PBS. Female C57BL/6 mice (n = 5), 6 to 8 weeks of age, were gavaged with 100 μL of 3% sodium bicarbonate prior to bacterial inoculation (4 × 10⁸ CFU in 100 μL PBS). As a control, 100 μL of the inoculum was stored on ice. Feed was removed 2 h before inoculation and supplied again after infection. After 90 min of infection, mice were euthanized, and the distal ileum (~6-cm length above the cecum) was extracted. All contents were transferred to 1 mL chilled PBS. Samples were filtered through a 5-μm filter to eliminate debris and centrifuged to precipitate bacteria. Inoculum was precipitated similarly, and all bacterial pellets were fixed in 1 mL of 4% PFA (paraformaldehyde) in PBS, on ice and shielded from direct light, for 30 min. Following fixation, the fixative was aspirated, and the pellet was resuspended in 1 mL PBS. These samples were analyzed by flow cytometry (Thermo Fisher Attune NxT analyzer), gating on the BFP-expressing population (to determine total Salmonella counts), and calculating the proportion of Salmonella expressing invasion genes (expression of both BFP and GFP). Data were plotted in GraphPad Prism. Illustrations were created in BioRender.

All experiments were repeated at least twice. Statistical significances were calculated in GraphPad Prism using the Mann-Whitney test. Differences were considered significant when the P value was <0.05, <0.01, or <0.001, indicated by *, **, and ***, respectively.

All relevant data have been provided in the manuscript and the supplemental information files.