Abstract
Salmonella enterica is a frequent cause of foodborne diseases, which is attributed to its adaptability. Even within a single host, expressing a gene can be beneficial in certain infection stages but neutral or even detrimental in others as previously shown for flagellins. Mutants deficient for the conserved glycerol-3-phosphate and phosphate antiporter glpT have been shown to be positively selected in nature, clinical, and laboratory settings. This suggests that different selective pressures select for the presence or absence of GlpT in a context dependent fashion, a phenomenon known as antagonistic pleiotropy. Using mutant libraries and reporters, we investigated the fitness of glpT-deficient mutants during murine orogastric infection. While glpT-deficient mutants thrive during initial growth in the gut lumen, where GlpT’s capacity to import phosphate is disadvantageous, they are counter-selected by macrophages. The dichotomy showcases the need to study the spatial and temporal heterogeneity of enteric pathogens’ fitness across distinct lifestyles and niches. Insights into the differential adaptation during infection may reveal opportunities for therapeutic interventions.
Similar content being viewed by others
Introduction
Enteric bacterial pathogens must continuously adapt to environmental challenges, including changes in nutrient abundance and challenges by the immune system, such as oxidative stress, to cause disease1. In the past decade, much progress has been made in discovering critical virulence factors that allow enteric bacterial pathogens to overcome specific obstacles (summarized in recent reviews2,3,4). Since the host’s fitness depends on reducing damage by the pathogen, specific virulence factors face opposing selective pressures, where gene expression is advantageous in some contexts but detrimental in others. One example is the O-antigen, made up of glycans linked to the outer membrane of bacterial pathogens, which protects against membrane stress but is also recognized by intestinal antibodies5,6,7,8. Knowledge of this tradeoff has been leveraged to design vaccines that select for the loss of the O-antigen, which attenuated the pathogen9. Similarly, the expression of type-3-secretion systems (T3SS) is crucial to causing inflammation and invading host cells but reduces the pathogen’s replication rate and increases susceptibility to membrane stress, which results in avirulent cells being selected10,11,12,13. Another example is motility conferred by flagellins, needed for enteric pathogens to reach, adhere to, and invade the epithelial cells that form the intestinal barrier14,15. Flagellin expression represents an Achilles heel for many enteric pathogens as the host’s Toll-like receptor 5 selectively recognizes them and initiates an immune response16. As a result, the advantage of motility is thought to outweigh the risk of detection during early growth in the lumen. Once the epithelial barrier has been breached, facultative intracellular pathogens like Salmonella enterica serovar Typhimurium (S. Tm) can reside and grow inside host cells, where the disadvantage of flagella is thought to govern the selective pressure17,18. Understanding how a gene can increase or decrease a pathogen’s fitness in a context-dependent fashion is important for developing effective strategies to control infectious diseases and understanding the evolution of pathogens with and without interventions such as antibiotic use or vaccination. For example, understanding that antagonistic pleiotropy affects the evolution of S. Tm’s genes encoding surface-exposed structures helped explain why conventional vaccination attempts have been mostly unsuccessful (recently reviewed in refs. 19,20,21,22) On a positive note, the understanding allowed us to create a new form of vaccination termed evolutionary trap vaccine or EvoVax, which works by exploiting that escape variants are selected for by the immunized host immune system are attenuated9,23.
So far, most examples of antagonistic pleiotropy have been attributed to the risk of detection by the host’s immune system. Given that nutrient, mineral, and electron donor or acceptor availability differ considerably between the luminal and intracellular lifestyles, we hypothesized that some metabolic genes should similarly have fitness effects that are highly lifestyle-dependent and face opposing selective pressures. Evidence for antagonistic pleiotropy in metabolic genes adds to our understanding of the evolution of bacterial pathogens, with likely implications for preventative and therapeutic interventions.
The fitness conferred by a particular gene is determined by the context of the selective pressures, meaning that inactivating mutants are selected for in some niches, while the functional gene is selected for in others. The concept of such types of genes conferring opposing fitness effects is called antagonistic pleiotropy24. Antagonistic pleiotropy has been proposed as a driver of niche specialization25,26. A signature of antagonistic pleiotropy is the enrichment of loss-of-function mutations such as premature stop codons, compared to genes that are important for fitness in all niches27. A recent screen of over 100,000 S. Tm genomes found that the flagellin methylase gene fliB, which contributes to adhesion and host cell invasion, has four times as many premature stop codons as would be expected under neutral selection27,28. These mutations have likely been selected in natural environments rather than inside laboratories27,28. The increased frequency of premature stop codons suggests that strains proficient for the genes are favored in certain conditions, whereas certain niches select for inactivating mutations27. In addition to fliB, the screen identified a metabolic gene with an eightfold increased frequency of such premature stop codons: glpT27. The underlying selective pressures have remained unclear. One complementary approach to studying the fitness effects of particular mutations is to study the within-host selection of spontaneous pathogen mutants29. Recently, our lab studied within-host evolution using an S. Tm long-term murine infection model13. Strikingly, glpT mutations were found in 70% of isolates, including single nucleotide polymorphisms (SNPs) and deletions (“dels”) at distinct locations (illustrated in Fig. 1a)13. This included cases where wildtype S. Tm prevented the re-establishment of a complex gut microbiota and a case where the microbiota was re-established by co-housing13. Moreover, the gene is not encoded in a genomic region that displays an unusually high frequency of mutations. Together, this indicates that these mutants were positively selected within the host (Fig. S1a). Our findings made in mouse models and the evidence in wild isolates support the existence of niches where the inability to express glpT confers a selective advantage, while other niches select for the wild-type (WT) glpT gene and that many S. enterica strains of different origins face opposing selective pressures acting for/against functional glpT. However, the nature of these selective niches has remained unclear.
a Single nucleotide polymorphisms and deletions found in the glpT gene of SL1344 after murine infection described in Gül et al.70 PLoS Biology 2023 (upper half) and enriched premature stop codons found in sequences of different S. Tm isolates identified in Cherry, GBE, 2020 (lower half). Created in BioRender. Santamaria de Souza (2025) https://BioRender.com/f23o207b The glpT mutant has a fitness advantage on the 1st day that gradually decreases over time. The data points show the aggregated competitive infection data of glpT vs. wild-type to highlight the data spread and the consistency of the change in the normalized ratio over time. The inoculum (“ino”) has a normalized ratio of 1 since the mutant vs wild-type ratio is normalized to their ratio in the inoculum. The orange lines connect the data obtained from individual mice, whereas the black lines represent the median. The statistical significance was tested using a Kruskal–Wallis test. n = 40 mice. Source data are provided as a Source Data file.
glpT encodes for an antiporter of Sn-Glycerol-3-phosphate (G3P) and inorganic phosphate (Pi), across a concentration-gradient27,30,31,32. So far, there is very little information about the role of phosphate availability during enteric infection. Homologs of GlpT are found across all phyla, and highly conserved among Enterobacteriaceae, including several pathogens of global importance of the genera Salmonella, Escherichia, Citrobacter, Shigella, and Klebsiella (Fig. S1b)33. While GlpT has mostly been described for the import of G3P and export of Pi, we show that it can also import Pi. G3P is an essential building block of phospholipids and is thus found in every membrane. In addition, G3P can serve as an important carbon and or phosphate source for S. Tm and other mammalian pathogens, making the selection of glpT mutations even more surprising31,34,35.
Here, we systematically assess the fitness defects of glpT mutants during distinct stages of orogastric mouse infection. We show that glpT-deficient mutants outcompete WT S. Tm during luminal growth when phosphate availability is high. In contrast, we demonstrate that the intracellular lifestyle in macrophages selects for a functional glpT gene. The lifestyle-dependent selection highlights the contextual fitness effect of glpT. Our findings emphasize the importance of antagonistic pleiotropy resulting from context-dependent metabolic trade-offs for pathogen evolution.
Results
The fitness of the glpT mutant changes during infection
To analyze the role of glpT in S. Tm gut-luminal growth, we compared the fitness of the glpT mutant to the isogenic WT strain. To this end, we infected streptomycin-pretreated mice with a 1:1 mix of a glpT knock-out (KO) strain and the WT S. Tm SL1344 for 4 days (see Table 1; 107−108 Colony Forming Units or CFUs, by orogastric injection). We determined the abundance of both strains in the feces of the mice by plating them on selective media and calculated the ratio between the strains, normalized to the ratio in the inoculum. By day 1 of infection, the glpT mutant outcompeted the WT, yielding a normalized ratio of ~10 (Fig. 1b). Notably, the WT outcompeted from day 2 and the mutant/WT ratio progressively declined from around 10 to around 1 by day 4 p.i. (Fig. 1b). The data implies that the glpT KO strain has an advantage during the initial stages of gut colonization, whereas another selective pressure favors WT S. Tm during days 2–4. This provided us with the opportunity to investigate the cause of the change in the fitness effect of glpT.
The fitness effect of glpT depends on substrate availability
GlpT is a G3P and Pi antiporter with transport activities that depend on G3P and Pi concentrations36. Because of that, we wondered if the competition between the glpT mutant and the WT depends on substrate availability36,37. To test that, we wanted to modulate the substrate availability to S. Tm in the intestinal lumen in three independent ways. First, we procured mouse chow with a reduced Pi content of 0.1% as one way to reduce the phosphate availability to S. Tm in the intestinal lumen. We then compared the normalized ratio between the glpT mutant and the WT in mice fed the regular 0.8% Pi chow to the normalized ratio between the glpT mutant and the WT in mice fed the 0.1% Pi chow36,37. The diet with a reduced Pi content significantly reduced the advantage of the glpT KO strain on day 1 p.i., to a normalized ratio of around 3 (Fig. 2a). This suggested that the fitness effect of glpT depends on the transport intensity and/or direction, which depend on substrate availability. As a second way to influence substrate availability, we supplemented 1% G3P in the drinking water of the mice (Fig. 2b). Similar to the reduction of Pi availability (black data points, Fig. 2a), this should lead to more G3P import and Pi export, which should increase the fitness of the glpT-proficient WT. As expected, the G3P supplementation too significantly lowered the normalized ratio between the glpT knockout strain and the isogenic WT (Fig. 2b). As a third way to validate that the fitness effect of glpT depends on substrate availability, we knocked out all three known Pi transporters encoded by S. Tm (Δ3Pi; pitA, pstS, yjbB)38,39,40,41. The resulting strain (Δ3Pi; pitA, pstS, yjbB) likely has less access to phosphate than the WT. While YjbB has been shown to be able to export Pi when overexpressed, we did not know its transport activity at normal expression levels and in the absence of pitA and pstS42,43. Thus, we decided to take the conservative approach of removing it too, which should, if anything, reduce the effect of potential GlpT Pi transport, if YjbB acts as a Pi exporter in the tested conditions41. Next, we tested the relative fitness of the glpT KO in the Δ3Pi background compared to the parental strain (Δ3Pi). Similarly to the low Pi diet, the low availability resulting from this genetic background (Δ3Pi), strongly decreased the ratio between the glpT KO and its isogenic parent (WT background) to around 1 (Δ3Pi background) (Fig. 2b). The decrease in the normalized ratio between the glpT KO and its parental strain, when access to Pi was reduced with the modified diet, or the lack of Pi transporters, establishes a link between the fitness effect of glpT and phosphate availability. The normalized ratio between the ∆glpT pitA::cat, ∆glpT pstS::aphT, and ∆glpT yjbB::aphT double knock-out strains can be found in Fig. S2a. We additionally tested the competitive fitness of the individual KOs (pitA, pstS, and yjbB, visualized in Fig. 2c) and the triple KO (Δ3Pi) in competition against the WT on day 1 p.i. All of them led to a nonsignificant ratio, close to 1, even when all three transporters were knocked out in the same strain (Fig. 2d). The neutral competition suggested that the luminal phosphate availability for S. Tm is surprisingly high.
a The fitness advantage conferred by the glpT mutant is reduced when mice are fed chow with an eightfold lower phosphate concentration (0.1%). Statistical significance was tested with a two-sided unpaired t-test. Lines represent the median. n = 12 mice, combined from four independent experiments, indicated by different shapes. Please note that four data points for the regular 0.8% Pi chow control are also shown in panel 2b, as the same mice served as the control for both experiments for the sake of 3 R; these data points are shown as open circles. b The fitness advantage conferred by the glpT mutant is reduced when mice are fed 1% G3P via their drinking water. The statistical significance was tested with a two-sided unpaired t-test. Lines represent the median. n = 7 mice. The different shapes indicate they originate from two independent experiments. c The fitness advantage conferred by the glpT mutant is strongly reduced in the background of a strain deficient for the phosphate transporters pitA, pstS, and yjbB (Δ3Pi). The Δ3Pi strain likely has less access to Pi, which should lower the intracellular Pi concentration. The statistical significance was tested with a two-sided unpaired t-test. Lines represent the median. n = 7 mice for the wild-type background and n = 8 mice for the Δ3Pi background combined from three independent experiments. d Visualization of the three Pi transporters encoded by S. Tm. Adapted from Bruna et al., Adv. in Exp. Med and Bio., 2022 for illustrative purposes. Created in BioRender. Santamaria de Souza (2025) https://BioRender.com/x62f158 (e) Competitive indices of phosphate transport mutants show that the Pi concentration is high enough for mutants not to have a competitive disadvantage on day 1 post-infection. Significance was tested by comparing the median to the theoretical median of 1 using a one-sample Wilcoxon test (two-sided). Lines mark the median. n = 5 mice combined from two independent experiments. Source data are provided as a Source Data file.
We, therefore, decided to measure the total Pi concentration in the cecum lumen of our mice. The median concentration was around 20 mM in the cecal content of untreated C57BL/6 mice and germ-free mice (Fig. S2b). We also quantified fumarate (detectable in lysed cells but not cecal content) and acetate (detectable in both) by targeted liquid chromatography/mass spectrometry to ensure that there was no contribution by the lysis of microbiota cells44,45 (Fig. S2c). The range of phosphate concentration is in line with a recent study (40–50 mM Pi) that obtained mouse chow from the same provider46. The intestinal concentration of Pi in humans is reported to reach 12 mM47,48. Such Pi concentrations should result in a high Pi availability, given that in vitro studies on the phosphate starvation response in E. coli and S. Tm use defined media with 0.5–1 mM Pi as the non-limiting control37,49,50.
GlpT can import Pi in vitro
While traditionally known for importing G3P and exporting Pi, the GlpT transporter is bidirectional and can additionally import Pi by the non-equimolar exchange of Pi in E. coli36,51. Thus, Pi import should be the main transport direction when the extracellular Pi concentrations are considerably higher than G3P (visualized in Fig. 3a, scenarios 2 and 3)31,36. Because GlpT could directly contribute to Pi import, we decided to test if the presence of glpT impacts intracellular Pi availability in vitro. We constructed a reporter for PpstS that is highly sensitive to intracellular Pi limitation37,43,52. All primers and plasmids used in this study can be found in Tables 2 and 3. We quantified the PpstS-nanoluc reporter activity in strains with differing capacities of importing Pi: the WT, a glpT KO, a pitA yjbB double KO, and a pitA yjbB glpT triple knock-out strain. We did not knock out pstS, as it influences PpstS activity, and chose the nanoluc reporter gene since it is highly compatible for use in vivo, allowing us to use the same reporter in subsequent animal experiments (see Fig. S3b)37,41,43,53. As expected, the PpstS promoter was active in all strains when no Pi was added to the medium (Fig. 3b). With the addition of Pi to the media, the PpstS activity decreased with distinct cutoffs for the different strains. The additional glpT mutation in the pitA, yjbB deficient background changed the cutoff Pi concentration fourfold (40 µM vs 10 µM for pitA, yjbB, Fig. 3b). These results suggest that GlpT can act as a Pi importer in our S. Tm strain. GlpT’s capacity for Pi-import was also reflected in the OD600nm at 24 h after dilution, which was lower for the triple KO (ΔpitA yjbB::aphT ΔglpT) than the double KO (ΔpitA yjbB::aphT) strain when no Pi was added to the medium (Fig. 3c). Thus, GlpT’s transport activity depends on environmental parameters, such as the G3P and Pi availability and GlpT might contribute to Pi import during luminal growth, a milieu where we found phosphate availability to be high (Fig. 2d and S2a, b). The reduction in the fitness advantage for the glpT KO strain when Pi availability is reduced (Fig. 2a, b) indicates that GlpT’s additional Pi-import capability is harmful during day 1 p.i. luminal growth.
a The three known transport modes of GlpT, depending on the availability of Glycerol-3-phosphate (G3P) and Pi. GlpT’s G3P and Pi antiport is bi-directional, in a concentration-dependent manner. In the first hypothetical scenario, there is more G3P than Pi in the periplasm and the main transport direction in G3P import and Pi export, by equimolar G3P:Pi antiport. In the second hypothetical scenario, there is more Pi than G3P in the periplasm, and the main transport direction is Pi import and G3P export, by equimolar Pi:G3P antiport. GlpT can additionally import Pi by non-equimolar exchange of 2Pi:1Pi in E. coli, which is shown in the third scenario. The different numbers of representative Pi and G3P molecules aim at representing different periplasmic/cytoplasmic concentrations. Created in BioRender. Santamaria de Souza (2025) https://BioRender.com/d47n170 (b) PpstS promoter activity as a readout for intracellular phosphate levels reveals that GlpT can increase intracellular phosphate levels. Cells were harvested after 24 h of incubation in a modified M9 medium. The statistical significance was calculated with an ordinary Two-Way ANOVA with Šidák multiple tests correction to account for the effect of the genotype and concentration. The lines represent the median. n = 3 biological replicates. c The final OD600nm after 24 h of incubation in a modified M9 medium shows that GlpT contributes to growth when phosphate availability and or uptake is limited. The lines represent the median. Statistical significance was tested using an ordinary one-way ANOVA with Dunnet correction. n = 3 biological replicates. Source data are provided as a Source Data file.
G3P metabolism does not explain the d1 fitness advantage
To investigate if the luminal fitness advantage of the glpT knockout strain on day 1 p.i. is related to G3P metabolism, we constructed a library of isogenic knockout strains lacking genes encoding enzymes that have been implicated in glycerol metabolism (visualized in Fig. 4a)31,54. The WT and each isogenic mutant encode distinct fitness-neutral genetic barcodes, which allowed us to quantify the abundance of each strain in the pool by qPCR (visualized in Fig. 4b)55,56. UgpB is the substrate binding component of the UgpB-AEC2 ABC transporter for G3P57. Since the ugp operon is induced during phosphate limitation37,43, we expected to find a neutral normalized ratio between the ugpB KO and the WT strain (≈1, Fig. 4c), further indicating a high Pi availability in the gut lumen. GlpF and GlpK enable the glycerol uptake and phosphorylation, respectively58. The neutral ratio for the glpFK knock-out strain suggested that uptake via the GlpFK route does not influence the fitness in the gut lumen, potentially because GlpT could export the G3P58. The neutral ratio (≈1) for the two G3P dehydrogenases glpABC and glpD also suggested that there is little G3P conversion (Fig. 4c). We also analyzed mgsA, encoding for the methylglyoxal synthase, which catalyzes the formation of methylglyoxal, a strong electrophile with toxic effects59,60. The production of methylglyoxal is thought to be important during phosphate limitation as it leads to the release of the phosphate group from DHAP, during high glycolytic flux, and oxidative stress61,62. Since MgsA is allosterically inhibited by Pi binding (at 0.3–0.5 mM Pi in E. coli), the high phosphate availability in the intestinal lumen probably inhibits MgsA (Fig. 4c)59,60,63. The lack of methylglyoxal production was supported by the competition with the mutants deficient in enzymes involved in methylglyoxal detoxification (gldA, gloA, gloB, gloC, yqhD, fghA, yeaE), which were not outcompeted by the WT on day 1 p.i. (Fig. 4d)60,64,65.
a Visualization of the glycerol metabolism in S. Tm. Mutants with a significant normalized ratio are highlighted in orange and genes or reactions affected by phosphate are highlighted in blue. Abbreviations: DHA(P) for Dihydroxyacetone (phosphate), AKR for Aldo-keto reductase, and HPA for Hydroxypropionic acid. The visualization was adapted from Tran et al.122 and Subedi et al.60 for illustrative purposes. Created in BioRender. Santamaria de Souza (2025) https://BioRender.com/v23o781 (b) Visualization of the setup of the screen to quantify the effect of the different players in the glycerol metabolism and methylglyoxal pathway on day 1 post-infection of streptomycin-pretreated C57BL/6 mice. Created in BioRender. Santamaria de Souza (2025) https://BioRender.com/o63l097 (c) The normalized ratios for the mutants in the upper part of the glycerol metabolism pathway show that only the glpT mutation leads to a significant competitive advantage. The statistical significance was tested using the Wilcoxon signed rank test (two-sided) against a hypothetical mean of 1 for a neutral competition. Lines indicate the median. n = 6 mice, except for ugpB::aphT n = 5 mice. d The normalized ratios for the mutants defective in different parts of the methylglyoxal detoxification pathway show that the absence of detoxification enzymes does not lead to a fitness disadvantage. The statistical significance was tested using the Wilcoxon signed rank test against a hypothetical mean of 1 for a neutral competition. Lines represent the median. n = 4 mice. Source data are provided as a Source Data file.
The glpT mutant is outcompeted by day 4 of infection
Since the ratio between the glpT KO and the WT declines around 10-fold between days 2–4 p.i. (Fig. 1d), we considered that the luminal phosphate availability might decrease in the presence of inflammation. From around day 2 p.i. inflammation causes a dramatic shift in the chemical composition of the gut lumen, which substantially influences the fitness of enteric bacteria. Therefore, we examined the ratio of a glpT mutation in an ΔinvG ΔssaV background, a strain deficient for S. Tm’s T3SS and incapable of eliciting pronounced gut inflammation66,67. In this strain background, the glpT mutant also had a fitness advantage by day 1 p.i. (ratio ≈ 10; Fig. 5a, shown in blue). Remarkably, the ratio of the glpT mutant did not decrease in the ΔinvG ΔssaV background on subsequent days (Fig. 5a, shown in blue). However, when we quantified the Pi concentration in mice infected with WT S. Tm, we saw no difference between day 1 and day 4 p.i. (Fig. S3a). We additionally measured the activity of our PpstS-nanoluc reporter (as a readout for intracellular Pi) and found no difference between day 1 and 4 p.i. in WT or ΔinvG ΔssaV infected mice (Fig. S3b). Despite the supplementation of phosphate (100 mM K2HPO4) in the drinking water of the mice, we observed no difference in their luminal concentration or the ratio of the glpT mutant (Fig. S3c, d). This suggested that there was no significant decrease in luminal Pi availability.
a The normalized ratio only decreases when the strains are proficient for the two type three secretion systems and, therefore, virulent. The lines connect the means. n = 9 mice in three individual experiments. b Visualization of the setup and effect on S. Tm’s capacity to cause inflammation and or invade the lamina propria. Representative microscopy images (10×) of cecum cryosection demonstrate the absence of inflammation on the top left and strong inflammation in all other setups. Illustration created in BioRender. Santamaria de Souza (2025) https://BioRender.com/x00c911 (c) Triple infection experiment, with a wild-type strain that leads to high inflammation levels in both groups. The normalized ratio only decreases when the strain is proficient for the type three secretion systems, enabling access to the lamina propria and intracellular replication. The lines connect the means. n = 7 mice in three individual experiments. Source data are provided as a Source Data file.
We additionally wanted to exclude that the transcription of glpT changed over time. To do so, we created a PglpT-gfp transcriptional reporter and quantified the gfp+ fraction of cells stained with an anti-O5 LPS antibody to detect all S. Tm cells using a gating strategy shown in Fig. S4a. We ensured the plasmid was maintained over the infection without any significant fitness cost (Fig. S4b–d) and observed no significant difference in PglpT activity between day 1 and day 4 p.i. (Fig. S4c). These observations suggested that the murine gut luminal niche selects for glpT mutants and that other niches within the mouse may select against glpT mutants by days 2–4 p.i. (Fig. 1b). The nature of such niches remained unclear.
A T3SS-accessible niche selects against the glpT mutant
In addition to the inflammation state of the cecum, the expression of functional T3SS enables S. Tm to take up its intracellular lifestyle and form gut tissue reservoirs68,69. Thus, the declining fitness of the glpT mutant detected from day 2 and onwards might result from the selection of the WT over the glpT knockout when the pathogen resides in intracellular compartments within host tissues. To discriminate between the effect of inflammation on the luminal subpopulation and a potential effect on the intracellular subpopulation, we set up an experiment in which the same two sets of strains competed against one another (glpT::aphT vs. WT and ΔinvG ΔssaV glpT::aphT vs. ΔinvG ΔssaV). However, this time, we diluted the mix of the two competing strains in a 100-fold excess of untagged WT S. Tm. The excess of WT S. Tm leads to pronounced gut inflammation in both setups (visualized in Fig. 5b)70. As anticipated, in the competition between glpT mutant and the isogenic WT strain proficient for gut tissue invasion, the glpT mutant outcompeted the WT on day 1 p.i., while the ratio declined to around 1 by day 4 p.i. (Fig. 5c, shown in orange). In contrast, in the competitive infections between ΔinvG ΔssaV glpT::aphT vs. ΔinvG ΔssaV strains in the inflamed gut that cannot invade the gut tissue, the glpT mutant outcompeted throughout the infection (ratio≈7, Fig. 5c, shown in turquoise). This indicated that the selection of the WT is not due to inflammation-associated changes in the cecal lumen milieu. Instead, it seems that a niche mainly accessible to cells with functioning T3SS is selected for the WT.
The glpT mutant has a replication deficit inside macrophages
In the lamina propria, many S. Tm cells are lodged within macrophages and S. Tm can replicate in tissue-resident macrophages, a Pi-scarce environment52,71. We considered that WT S. Tm could benefit from glpT expression in this environment. To test that, we infected Raw264.7 macrophages either with the WT or the glpT mutant and quantified S. Tm’s replication rates (by dividing the CFUs quantified for 24 h p.i. by the CFUs for 2 h p.i.). We found that the glpT mutant has a replication deficit (ΔglpT empty vector) that can be rescued by plasmid complementation (ΔglpT Pnat-glpT, Fig. 6a). To test if the activation state of macrophages matters, we also quantified the replication rate in Raw264.7 that were primed with IFN-γ prior to infection. The priming with IFN-γ slightly increased the importance of glpT for intra-macrophage replication (Fig. 6b). In both cases, the complementation plasmid was stably maintained without antibiotic selection (Fig. 6c). Since host cells can limit the amount of phosphate available to intracellular pathogens when they are infected, we hypothesized that the effect of the glpT mutation on intra-macrophage replication should not depend on the presence of the three Pi transporters pitA, pstS, and yjbB since Pi availability is already limited in the niche72,73,74. To test that, we compared the replication rate of the glpT mutant in the WT background to the ratio in the Δ3Pi (pitA, pstS, and yjbB deficient) background. We found that the glpT mutation significantly attenuated intracellular replication in both cases (Fig. 6d). We also hypothesized that with the lower Pi availability, the entire glycerol pathway should be active inside macrophages. Since the methylglyoxal pathway was recently shown to contribute to the resistance against oxidative stress generated by macrophages, we decided to investigate if glpT contributes to the methylglyoxal pathway inside macrophages62. To this end, we constructed a transcriptional reporter for PgloC, which regulates the expression of a type II glyoxalase involved in the last step of the methylglyoxal pathway (see Fig. 4a)75. The higher PgloC promoter activity for the WT inside Raw264.7 macrophages suggests that glpT contributes to the methylglyoxal pathway and can, consequently, contribute to the resistance against the oxidative stress challenge faced by S. Tm with an intra-macrophage lifestyle (Fig. 6e)62. To discriminate between a methylglyoxal-dependent effect, such as a contribution to resistance against oxidative stress, and a methylglyoxal-independent effect, such as import of G3P as a carbon source, we compared the replication rate of the glpT deficient strain, to the replication rate of strains deficient for only mgsA, gloA, and gloB or in combination with the glpT knock out. MgsA is responsible for the formation of methylglyoxal, while GloA and GloB are involved in its detoxification (see Fig. 4a)59,60,64,65,76,77. mgsA, gloA, and gloB deficient strains should all be less protected against oxidative stress than WT S. Tm62. The replication rate was similar for all strains (around 1, Fig. S5). This indicated that our replication assays do not have sufficient resolution to be able to disentangle GlpT’s contribution to replication in Raw264.7 macrophages in methylglyoxal-dependent and/or -independent ways, even when we combined datasets from three separate experiments, to increase the statistical power (Figure S5)64,76,77.
a Replication rate inside infected Raw264.7 macrophage cells as CFUs at 24 h post-infection divided by the CFUs at 2 h post-infection after a gentamicin protection assay. The statistical significance was tested by an ordinary one-way ANOVA with Dunnet correction. The lines indicate the median. n = 3 Raw264.7 biological replicates. b Replication rate inside infected Raw264.7 macrophage cells. The Raw264.7 macrophages were primed with 2 ng/mL IFNγ for 24 h prior to infection. The lines indicate the median. The statistical significance was tested by an ordinary one-way ANOVA with Dunnet correction. n = 3 biological replicates (except for the empty vector condition where n = 2). c Plasmid retention for the experiments shown in panels a and b was quantified for the glpT mutant carrying the Pnat-glpT complementation plasmid by replica plating from LB Agar containing Streptomycin onto LB Agar with Ampicillin and then Streptomycin. The percentage of cells resistant to Ampicillin and Streptomycin is shown. The lines indicate the mean and the whiskers and the smallest and largest values, respectively. n = 3 agar plates. d The replication rates in Raw264.7 macrophage cells show that glpT is important for replication regardless of whether the strain is proficient for the canonical Pi transporters. The statistical significance was tested by an ordinary one-way ANOVA with Šidák correction. The lines indicate the median. n = 3 biological replicates. e The PgloC promoter activity suggests that the methylglyoxal pathway is less active in the glpT mutant. The lines indicate the median. The statistical significance was tested with a two-sided unpaired t-test. n = 4 biological replicates. f Normalized ratio between the glpT mutant and the wild-type in mice intraperitoneally injected with PBS (open symbols), anti-CSF1R (green), or an isogenic antibody control (black). The different shapes (circles and squares) indicate the origin of two independent experiments and the lines indicate the median. In the absence of macrophages, the glpT mutant continues to outcompete the wild-type on day 4 post-infection. The statistical significance was tested using an ordinary one-way ANOVA with Dunnet correction. n = 6 or 4 (isogenic antibody) mice. Source data are provided as a Source Data file.
Macrophages tip the balance between the glpT mutant and the WT
Finally, we tested if intramacrophage replication could impact the ratio between the glpT mutant and the WT in fecal samples. Since 99.999% of the luminal population is killed by neutrophils on day 2 p.i., S. Tm that have replicated inside macrophages could contribute to the re-colonization of the intestinal lumen78,79,80. Therefore, the higher intra-macrophage replication rate of the WT compared to the glpT mutant might shift the balance towards a neutral ratio (≈1) in fecal samples by the last days of infection. To test that, we injected mice with an anti-CSF1R blocking antibody that selectively reduces the precursors of macrophages in the lamina propria of the mice81,82,83. As a control, we intraperitoneally injected mice with PBS or an isogenic antibody control. As expected, the ratio between the glpT mutant and the WT was neutral for the mice injected with PBS or with the isogenic antibody control (Fig. 6f). Strikingly, in the mice treated with anti-CSF1R, the glpT mutant consistently outcompeted the WT (Fig. 6f). This provided additional evidence supporting that the macrophages select S. Tm cells expressing glpT.
Discussion
Previous studies have shown that classical virulence factors face opposing selective pressures, which have been leveraged as a therapeutic strategy9. In this study, we investigated whether the distinct niches encountered by the enteric pathogen S. Tm lead to opposing fitness effects and selective pressures on metabolic genes. We characterized glpT as a candidate metabolic gene that faces opposing selective pressures. We found that it has a lifestyle-dependent fitness effect, where glpT is harmful during luminal growth, characterized by a high phosphate availability. However, the same gene is important for intracellular replication in Raw264.7 macrophages, and the wild type is selected in the presence of macrophages in vivo.
The availability of sugars, lipids, and metals has been shown to impact the fitness of S. Tm and other enteric pathogens during luminal growth1,84. However, it was unclear what the impact of phosphate availability on S. Tm’s luminal growth would be. Since the Pi concentration impacts GlpT’s transport and glycerol utilization, we considered phosphate availability during luminal growth. We found that there is enough Pi in the cecal lumen (~20 mM, Fig. S2b) for S. Tm to stay fit (neutral ratio between KO and WT) even when we deleted its canonical Pi transporters (pitA, pstS, and yjbB, Fig. 2b–d and S2b). PitA and PstS are believed to be more efficient at importing Pi than GlpT, during in vitro growth85,86,87. Interestingly, the effect of glpT on luminal growth depends on phosphate availability, as the selection of the glpT KO is diminished when mice are fed a low-phosphate diet or when the strain is deficient for its canonical Pi transporters (Fig. 2a, b), suggesting that glpT is only harmful when Pi availability is high. While nutrient limitation has been tied to protection against enteric pathogens, the negative effect of high phosphate availability demonstrates that an excess can also be harmful88. While GlpT has mostly been described for the import of G3P in exchange for Pi, the reverse transport direction was discussed several decades ago36. Additionally, GlpT has been shown to be able to import by non-equimolar Pi exchange in E. coli (visualized in Fig. 3a)36,51. Our PpstS reporter data indicates that in vitro, GlpT can also import Pi in S. Tm and, thus, directly contribute to the intracellular Pi concentration (Fig. 3). The high luminal phosphate availability and capacity for GlpT to import Pi is supported by the neutral fitness effect of knocking out the other genes involved in glycerol metabolism on day 1 p.i. (Fig. 4). GlpT’s Pi import may reduce WT S. Tm growth in the intestinal lumen and explain the fitness advantage of the glpT KO during luminal growth.
The luminal disadvantage for glpT expression could have multiple non-exclusive reasons. If Pi import by GlpT is coupled to G3P export, the reduction in the important metabolite G3P could reduce fitness. In addition, the Pi import in a niche with such a high phosphate availability could also pose a risk for phosphate intoxication. Although phosphate intoxication has not previously been investigated in the context of luminal colonization, there is precedent for S. Tm’s phosphate intoxication during in vitro growth41. The mechanism for the Pi intoxication proposed by the authors is that phosphate chelates magnesium ions, which prevents magnesium-dependent processes, including translation89. Magnesium chelation seems a severe risk for S. Tm as it evolved a process that inhibits its own F1F0 ATPase to free Mg2+ 90. Moreover, increased intracellular Pi concentrations directly reduce the Gibb’s free energy for ATP hydrolysis under physiological concentrations and, thus, the concentration of ATP in the cell91.
It is currently unclear if expressing glpT in the lumen is only disadvantageous or if it prepares the WT for other niches. Importing Pi via GlpT could be beneficial for two reasons. Firstly, to prevent the accumulation of G3P, which, like other sugar phosphates, can be growth-inhibiting92. Secondly, it could be a means to accumulate Pi since PitA does not import much Pi, and the Pst complex is repressed by intracellular Pi37,43. The 70% reduced mortality rate in mice infected with S. Tm mutants that cannot store Pi in the form of polyphosphate suggests that phosphate storage greatly affects its virulence93. Since Pi levels are limited at several sites of infection, such as the blood and intracellular niches, many pathogens might benefit from storing phosphate while they grow in a high phosphate environment, such as the intestinal lumen. Access to phosphate at Pi-limited infection sites seems to affect virulence as sepsis patients with higher serum phosphate levels have a higher mortality94,95. Given its potential for therapeutic intervention, it would be interesting to explore the links between phosphate storage during luminal growth and virulence in future experimental work71.
In clinical contexts, glpT mutations are also selected by treatment with Fosfomycin and related antibiotics that have been proposed as a treatment for infections with multidrug-resistant Enterobacteriaceae30,96,97. Thus, it might be important to prescribe Fosfomycin in combination with another antibiotic to prevent the selection of glpT mutations that could lead to a dangerous short-time expansion of enteric pathogens in the gut lumen, especially in immunosuppressed patients who might not be able to mount an immune response that could limit glpT mutants.
In the second part, we examined the striking change in the selection pressure acting on the glpT KO on days 2–4 p.i. when the WT outcompetes the glpT KO. We found that this selection of the WT depends on S. Tm’s two T3SS, which enable the invasion and replication inside host cells98,99. The experiments with Raw264.7 cells showed that glpT is important for replication inside macrophages. When we depleted macrophage populations in vivo, the glpT KO outcompeted the luminal population on day 4 p.i. The correlation between the presence of glpT and PgloC activity suggested that the entire pathway is active inside the phosphate-limited macrophages. This suggests that S. Tm imports G3P in this niche and that the methylglyoxal pathway is not inhibited by Pi. The importance of GlpT inside macrophages could be linked to its contribution to the methylglyoxal pathway. Detoxification of methylglyoxal has recently been shown to be crucial for S. Tm’s survival of the oxidative stress generated by macrophages62. Additionally, G3P metabolism could be an important carbon source for intracellular replication, given the niche’s nutrient and mineral scarcity100. G3P metabolism entails less phosphorylation than the utilization of glucose, as it enters the glycolytic pathways at a midway point. Such a benefit for G3P metabolism was recently shown for another pathogen that can replicate inside phagocytic host cells, Listeria monocytogenes, in J774.7 macrophages101. Furthermore, G3P metabolism has been found important for S. Tm colonizing the spleen of mice, a site where the majority of S. Tm reside in red pulp macrophages35,102.
An interesting side observation was the heightened replication rate of the Δ3Pi strain in Raw264.7 macrophages (Fig. 6d). The finding was in line with a prior publication, which found that a phoB deficient S. Tm strain showed a higher replication rate in J774 macrophages103. A similar conclusion can be drawn for replication in HeLa cells, where phosphate availability was inversely correlated with S. Tm’s replication rate74. The authors knocked out the eukaryotic PIT1 gene, which should increase phosphate availability to S. Tm, and saw a reduced replication rate for S. Tm74. Accordingly, overexpression of PIT1 increased S. Tm’s replication rate74. In contrast, a different study found that a phoB- and pstSCAB-deficient S. Tm strain was attenuated in replication in HeLa cells and Raw264.7 macrophage cells52. The conflicting results show that the effect of phosphate availability and lack thereof at intracellular sites on S. Tm fitness is unresolved and should be addressed in future studies. Since the luminal S. Tm population is decimated by neutrophils that kill 99.999% of luminal cells on day 2 p.i., we have suspected that re-seeding events from intracellular sites contribute to the re-plenishing of the luminal S. Tm population after the luminal population has been decimated by neutrophils or antibiotic treatment78,104. The contribution of intramacrophage replication to the recolonization of the lumen would be in line with a study that used an engineered (STmCytoKill) S. Tm strain to show that cytosolic replication in enterocytes significantly contributed to luminal expansion and fecal shedding79. Here, we saw that the S. Tm cells expressing glpT outcompeted glpT deficient cells in the intestinal lumen when macrophages were present as an intracellular niche in the host tissue. The change in ratio in the intestinal lumen by 4 p.i. suggests that the higher replication rate of glpT-expressing cells inside macrophages promotes re-colonization of the intestinal lumen. Overall, this selection for WT glpT could explain why this gene is evolutionarily conserved.
The evolution dynamic of glpT is particularly interesting given that enteric pathogens infecting via the fecal-oral route reach the gut lumen before establishing an intracellular lifestyle. The number of cells and replication rate are higher in the lumen, where glpT mutants are selected. glpT mutations could, therefore, represent an example of antagonistic pleiotropy, where a mutation is selected during an (early) context but later results in a disadvantage. A famous example of antagonistic pleiotropy is E. coli’s resistance to phage T4 infection, which reduces its fitness in the absence of the phage105,106. Alternatively, the selection of glpT mutants observed in our study may result from reduced levels of colonization resistance, which are characteristic of the mouse models that we have used107. This should enhance the chances of selecting for mutants with enhanced fitness in such low-colonization-resistance niches. Transient alleviation of colonization resistance by food, antibiotics or other stressors may alter the gut-luminal niche to select for mutations in particular genes including in natural Salmonella populations, explaining why the loss of function of glpT and several dozen other genes are selected often enough to be picked up in when analyzing 100,000 genomes27. Positive selection of glpT-deficient mutants is only possible when the glpT gene is expressed by the parental strains in the respective niches.
Context-dependent selection for mutants has also been observed for other enterobacteriaceal genes. The expression of the virulence regulator hilD of S. Tm is one prominent example12,13. Here, costs of expression in the gut lumen result in the selection of hilD-inactivating mutations when the pathogen grows over extended periods of time at high densities within the lumen of a microbiota-depleted inflamed gut. Similarly, gut luminal growth of E. coli K12 in germ-free mice selected for mutants in flagella, ompB and the maltose regulator gene malT108. Another study found that gut-luminal growth can select for the co-existence of E. coli strains, which can utilize galactitol, and gat mutants, which cannot109. Bioinformatics analysis of carbohydrate utilization genes in Enterobacteriaceae provided evidence that such genes might be under diverse context-dependent selection pressures, resulting in substantial differences in the control of carbohydrate utilization gene expression between different Enterobacteriaceae110. In conclusion, our data indicates that glpT mutations are selected in the gut luminal niche, resulting in a disadvantage in the phagocytic and potentially other phosphate- and nutrient-limited niches. The counterselection follows much slower kinetics, as the ratio reaches 10 within the 1st day of infection but takes the following 3 days to decline to the same extent. Yet, the data suggest that the capacity for intramacrophage replication has a crucial effect on the luminal population and, by extension, fecal shedding and infection of new hosts.
In conclusion, the fitness gains and losses of glpT show that opposing selective pressures can indeed act on metabolic genes. The example highlights how the heterogeneity between the luminal and intracellular niches results in lifestyle-dependent selection, even for a single gene within the same host. Similar dynamics likely exist in other facultative intracellular pathogens. The results reveal a need for studying the dynamic heterogeneity in pathogen populations at a higher resolution. In addition, such metabolic tradeoffs reveal an opportunity for therapeutic intervention, as past studies have shown that environmental changes during infection impose severe bottlenecks on pathogens. Consequently, the heterogeneity in selective pressures, need for adaptation and tradeoffs represents a unique opportunity to identify therapeutic targets, especially for the growing number of bacterial pathogens that cannot efficiently be treated with existing antibiotics.
Methods
Ethics statement
All animal experiments conducted in this study were reviewed and approved by Tierversuchskommission, Kantonales Veterinäramt Zürich, under the license numbers ZH158/2019, ZH108/2022, and ZH109/2022 and complied with the cantonal and Swiss legislation. We hold publication rights for all figures displayed in this study.
Mouse infections
Male and female mice aged between 8 and 12 weeks were randomly assigned to experimental groups. All mice are bred and housed in individually ventilated cages in SPF and E. coli-free conditions at the EPIC facility at ETH Zürich. Germ-free mice were bred in flexible film isolators at the EPIC isolator facility. The SPF mice are offspring of C57BL/6 originally acquired from the Jackson laboratory. The mice were fed a standard chow diet acquired from KLIBA NAFAG, Switzerland, ad libitum. For the low Pi diet experiment, the food was changed to low phosphate chow (KLIBA NAFAG #2169) 24 h prior to infection. The mice had free access to water and were kept on a 12 h light/dark cycle at 21 ± 1 °C. SPF mice were orally pre-treated with 25 mg of Streptomycin 1 day before infection, according to our established streptomycin colitis model107. The S. Tm strains used for infection were grown in LB with 3 M NaCl overnight with appropriate antibiotics. The cultures were then diluted 1:10 and subcultured in LB with 3 M NaCl for 4 h without antibiotics. The bacteria were washed in DPBS before being diluted 1:2 in DPBS and used as the inoculum (~5 × 107 CFUs in 100 µL). To quantify CFUs in cecal tissue samples, we conducted gentamicin protection assays. For this assay, we cut the cecum open longitudinally, washed it in DPBS, and incubated it in 400 μg/ml gentamicin in DPBS (Sigma–Aldrich) at room temperature for 1 h. CFU quantification of in vivo samples was performed by plating diluted samples on MacConkey agar with appropriate antibiotics and counting. For competitive infections with two different S. Tm strains, the two strains were quantified by differential plating, where the two strains were quantified after plating them on solid media containing antibiotics, which only the strain of interest was resistant to. The normalized ratio is defined as the number of CFUs for the mutant divided by the number of CFUs by the isogenic parental strain, normalized to the ratio present in the inoculum. C57BL/6 mice were sacrificed by day 4 p.i. by carbon dioxide asphyxiation followed by cervical dislocation, in accordance with our license.
Bacterial strains, plasmids, and primers used in this study
The bacterial strains used in this study are indicated in Table 1 and derived from S. Tm SL1344, which was passaged through a mouse and designated SB30012,111. Bacterial strains were cultivated in 1.5% lysogeny broth (LB) at 37 °C with shaking. LB was supplemented with ampicillin (100 µg ml−1), kanamycin (50 µg ml−1), streptomycin (50 µg ml−1), or chloramphenicol (35 µg ml−1), when necessary. All oligos and plasmids used in this study are indicated in Tables 2 and 3.
Genetic barcoding of strains using WITS and WISH tags
WITS and WISH were introduced in the strains of interest by p22 transduction55,56,112. The WISH tags used in this study are a beta version of the ones published in ref. 56 that were compatible with qPCR but were only later made compatible with quantification by next-generation sequencing. All the WITS and WISH tags are integrated between malX and -Y genes along with a kanamycin, chloramphenicol, or ampicillin resistance cassette.
List of fitness-neutral genetic tags
WITS01, WITS02, WITS11, WITS13, WITS17, WITS19, WITS21.
WISH028, WISH115, WISH182, WISH194, WISH205, WISH206, WISH207, WISH208, WISH209, WISH210, WISH212, WISH213, WISH214.
The sequences of WISH tags can be found in the supplementary table provided by Daniel et al., Nature Microbiology, 202445.
Gene disruption
Novel gene knockouts were created using the λ/red one-step protocol, where genes were disrupted with a cat or aphT gene, conferring chloramphenicol or kanamycin resistance, respectively113. The primers for the knockout were designed by taking the 40 bp flanking the gene(s) of interest, followed by 20 bp for the amplification of the plasmids pKD3 (encodes the cat gene) and pKD4 (encodes the aphT gene). Primers were used to amplify the pKD3 and pKD4 plasmids and create overhangs that lead to the integration of the antibiotic resistance gene inside the gene(s) to be replaced. The PCRs were done using high-fidelity Phusion polymerase (Thermo Fisher Scientific F530L). The length of the PCR was verified on a gel and the DNA was cleaned up by ethanol precipitation. Briefly, the PCR product was pooled and precipitated in 0.1 vol of 3 M sodium acetate and 2.5 vol ethanol. The DNA was allowed to precipitate at −20 °C for 10 min before being spun down for 5 min at 4 °C and full speed. The DNA was washed twice in EtOH before being resuspended in ddH2O. S. Tm cells carrying the pKD46 plasmid were grown in SOB medium and induced for the expression of phage-derived genes encoded on pKD46 using 10 mM arabinose (Sigma–Aldrich) and then made electrocompetent. After electroporation, cells were rescued in SOC, incubated at 37 °C for 1 h, and then plated on LB with kanamycin or chloramphenicol. p-22 lysates were prepared and used to transduce the mutation into the ancestral strain and other strains of interest114. p-22 transduction was also used to introduce mutations available in the McClelland collection to strains of interest115. Where indicated, the cat or aphT resistance cassettes were removed by FLP recombination113.
Promoter fusion constructs
Promoter fusion constructs were created by amplifying 500 bp upstream of the gene of interest and its first three bases. The promoter region was fused to the reporter gene (gfpmut2 or a codon-optimized nanoluc) by Gibson Assembly116. Gibson Assembly Master Mix (New England Biolabs, E2611L) was used according to the manufacturer’s instructions. Primers were designed with NEBuilder.
Quantitative PCR (qPCR)
One hundred microliter of the homogenized samples was diluted in 2 mL of LB with Streptomycin and incubated for 4 h at 37 °C with shaking. The pellet of the enrichment culture was used for genomic DNA extraction with the QIAamp DNA Mini Kit (QIAGEN). The primers were mixed to a final concentration of 1 µM. Duplicates of the DNA samples (2 µL) were amplified using the FastStart Universal SYBR Green Master Mix (Roche, Cat# 4385610). The qPCR was carried out using a QuantStudio 7 Flex instrument (Applied Biosystems) with the following protocol: initial denaturation at 95 °C for 14 min followed by 40 cycles of 94 °C for 15 s, 61 °C for 30 s, and 72 °C for 20 s. The qPCR primers were designed to have equal amplification efficiencies55,56. The DNA extracted from the inoculum was used for a standard curve, which was used to calculate the CT values and the relative frequency of every barcode in the pooled genomic DNA.
Defined media for Pi titration and the PpstS-nanoluc reporter assay
Bacteria were grown in a minimal medium: 0.5 g/L NaCl, 1 g/L NH4Cl, 100 mM CaCl2, 2 mM MgSO4 supplemented with 250 mg/L Histidine-HCl, 4 mM fumarate and 40 mM pyruvate. For the overnight growth, 1 mM Pi in the form of K2HPO4 was added to the medium. In the morning, cells were diluted 1:100 in the same medium. After 4 h, cells were washed and resuspended in the base medium free of added Pi with a titration of K2HPO4 at the indicated concentration. The time of resuspension in the medium with the addition of the indicated Pi concentration is considered time point 0 h. The cultures were incubated at 37 °C in aerobic conditions and with shaking. After 24 h, the OD600–nm was measured, the cells were pelleted and frozen until the luciferase assay was conducted. The reporter activity was normalized to the OD600 nm.
Luciferase assay
A nanoluciferase gene was codon-optimized and synthesized by IDT technologies. In vivo, samples were centrifuged at 30 × g for 1 min to remove debris. All samples were centrifuged at 21,000 × g for 5 min. The cells in the pellet were lysed by exposing them to 3 cycles of cold shock in liquid nitrogen followed by 5 min at 98 °C. The lysed cells were resuspended in the Promega NanoGlo luciferase buffer and processed according to the vendor’s protocol.
Antibody-mediated depletion of macrophages and neutrophils
Anti-CSF1R (LuBio Science #BE0213) antibodies were used for macrophage depletion. The antibodies were diluted in DPBS to a concentration of 5 g/L for the first dose and 1.5 g/L for injections on days 0, 1, 2, and 3 p.i. 200 µl of antibody solution was administered with every injection using a 1 mL syringe and a 26 G needle. To control for the effect of the injection, the cohort of control mice was injected with 200 µl of DPBS or 150 µl of isogenic antibody control (IgG2a raised in rats, LuBio Science #BE0089).
Total inorganic orthophosphate quantification
Samples were collected in water instead of DPBS, and phosphate levels were quantified using a commercially available kit (Sigma–Aldrich MAK030) according to the manufacturer’s instructions.
Bacterial flow cytometric analysis
Fecal and cecal content samples were harvested in 1 mL PBS and homogenized for 2.5 min at 25 Hz. The samples were then centrifuged for 1 min at 500 × g and RT to remove debris. The supernatant was transferred to a 1.5 mL Eppendorf tube, which was centrifuged for 5 min at 1500 × g and 4 °C. The pellet was resuspended in 300 µl of DPBS and transferred to a 96 well plate. The plate was centrifuged at 1500 × g and 4 °C for 5 min. The pellet was resuspended in 200 µl of 2% PFA in DPBS for fixation and incubated for 20 min at RT. Then the cells were washed in 200 µl of 1% BSA in DPBS. Next, the cells were resuspended in 50 µl of 1% BSA in DPBS containing human anti-O12 antibody (hSTA5, 1:200). The samples were incubated with the primary antibody for 30 min at RT and in the dark. Then 150 µl of 1% BSA in DPBS were added and the samples were washed with 200 µl of 1% BSA in DPBS. The pellet of bacterial cells was resuspended in 50 µl of 1% BSA in DPBS containing anti-human IgG Fc AF647 (1:200) and incubated for 30 min at 4 °C and in the dark. 150 µl of DPBS were added and the samples were washed with 200 µl of DPBS. The cells were resuspended in 200 µl of DPBS and analyzed using a CytoflexS cytometer (Beckmann Coulter).
Targeted liquid chromatography mass spectrometry
Murine cecal contents were sampled for the quantification of short-chain fatty acids. After harvesting, the samples were immediately derivatized with 3- Nitrophenylhydrazine and N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride in 70% isopropanol in LCMS-grade H2O. Short-chain fatty acids analysis by reversed-phase LC-MS was carried out on an Ultimate 3000 ULPC instrument (Thermo Scientific, Waltham, MA, USA) hyphenated to a QExative plus mass spectrometer (Thermo Scientific, Waltham, MA, USA) as described in117 with slight modifications. For LC separation we used a Kinetex XB C18 column (particle size 1.7 µm; 50 × 2.1 mm, Phenomenex, Torrance, CA, USA) with 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile as mobile phases. The sample injection volume was 2 µl. We ran a linear gradient at a flow rate of 500 µl min−1 with solvent B changing as follows: 0 min: 2%; 3 min: 95%; 5 min: 95%; 5.3 min:2%. Subsequently, the column was equilibrated for 2 min at initial condition. Mass analysis was performed in the negative FTMS mode at a mass resolution of 70,000 (at m/z 200) by applying heated electrospray ionization. Source parameters were set as follows: spray voltage: 2.7 kV; S-lens RF level: 50; sheath gas: 50; aux gas: 20; sweep gas: 0; heater: 350 °C. The metabolites were quantified using stable isotope-labeled as internal standards [13C4]-fumaric acid (99%), and [13C4, 15N]-L-aspartic acid (99%), purchased from Cambridge Isotope Laboratories.
Histopathology of in vivo samples
Murine cecal tissue sections were carefully dissected, embedded in Optimal Cutting Temperature embedding medium (O.C.T., Tissue Tek), and flash-frozen in liquid nitrogen. Samples were stored at −80 °C until they were processed. Samples were cut in 5 μm sections using a CryoStar NX50 Cryostat (Fisher Scientific) and air-dried on glass slides. The samples were stained with hematoxylin and eosin, and images were acquired on an AxioVision light microscope.
Phylogenetic analysis
An enzyme similarity analysis of the SL1344 GlpT (UniProtKB #A0A0H3NJJ9) amino acid sequence was conducted with EFI, where the UniProt database (version 2024–01) served as the reference 118,119. The search retrieved 792 similar proteins within the Enterobacteriaceae family. All subsequent analyses were performed using MEGA7120. The Sequence identity matrix was calculated with the Poisson model and visualized as a matrix of the within and between genus identity. All sequences are accessible on the public UniProt database.
Statistical analysis
All statistical analyses were conducted using GraphPad Prism 9 or 10 for Windows. Where applicable, a paired or unpaired t-test, Wilcoxon test, One-way or Two-way ANOVA test were used to assess statistical significance. The test used for each graph is indicated in the figure legends.
Biological materials availability
Those interested in our strains, plasmids, and/or tags are encouraged to contact Prof. Wolf-Dietrich Hardt (wolf-dietrich.hardt@micro.biol.ethz.ch).
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
All data needed to evaluate the conclusions of this study are presented in the Article and Supplementary Information. Additional raw data, such as flow cytometry raw data, microscopy metadata, and mass spectrometry raw data relevant to this paper, will be made openly accessible by the ETH Research Collection in agreement with ETH Zürich’s open-access policy. Source data are provided with this paper.
References
Wotzka, S. Y., Nguyen, B. D. & Hardt, W. D. Salmonella typhimurium diarrhea reveals basic principles of enteropathogen infection and disease-promoted DNA exchange. Cell Host Microbe 21, 443–454 (2017).
Gan, J., Giogha, C. & Hartland, E. L. Molecular mechanisms employed by enteric bacterial pathogens to antagonise host innate immunity. Curr. Opin. Microbiol. 59, 58–64 (2021).
Herzog, M. K.-M. et al. Mouse models for bacterial enteropathogen infections: insights into the role of colonization resistance. Gut Microbes 15, 2172667 (2023).
Spiga, L. & Winter, S. E. Using enteric pathogens to probe the gut microbiota. Trends Microbiol. 27, 243–253 (2019).
Collins, L. V., Attridge, S. & Hackett, J. Mutations at rfc or PMI attenuate Salmonella typhimurium virulence for mice. Infect. Immun. 59, 1079–1085 (1991).
Vito, R., Dexian, Z., Christopher, T. & Piddock, L. J. V. The O-antigen epitope governs susceptibility to colistin in Salmonella enterica. mBio 11, e02831 (2020).
Qingke, K. et al. Effect of deletion of genes involved in lipopolysaccharide core and O-antigen synthesis on virulence and immunogenicity of Salmonella enterica Serovar Typhimurium. Infect. Immun. 79, 4227–4239 (2011).
Moor, K. et al. High-avidity IgA protects the intestine by enchaining growing bacteria. Nature 544, 498–502 (2017).
Diard, M. et al. A rationally designed oral vaccine induces immunoglobulin A in the murine gut that directs the evolution of attenuated Salmonella variants. Nat. Microbiol. 6, 830–841 (2021).
Sturm, A. et al. The cost of virulence: retarded growth of Salmonella typhimurium cells expressing type III secretion system 1. PLoS Pathog. 7, e1002143 (2011).
Sobota, M. et al. The expression of virulence genes increases membrane permeability and sensitivity to envelope stress in Salmonella typhimurium. PLoS Biol. 20, e3001608 (2022).
Diard, M. M. et al. Stabilization of cooperative virulence by the expression of an avirulent phenotype. Nature 494, 353–356 (2013).
Gül, E. et al. The microbiota conditions a gut milieu that selects for wild-type Salmonella typhimurium virulence. PLoS Biol. 21, e3002253 (2023).
Chaban, B., Hughes, H. V. & Beeby, M. The flagellum in bacterial pathogens: for motility and a whole lot more. Semin. Cell Dev. Biol. 46, 91–103 (2015).
Duan, Q., Zhou, M., Zhu, L. & Zhu, G. Flagella and bacterial pathogenicity. J. Basic Microbiol. 53, 1–8 (2013).
Clasen, S. J. et al. Silent recognition of flagellins from human gut commensal bacteria by Toll-like receptor 5. Sci. Immunol. 8, eabq7001 (2023).
Cummings, L. A., Wilkerson, W. D., Bergsbaken, T. & Cookson, B. T. In vivo, fliC expression by Salmonella enterica Serovar Typhimurium is heterogeneous, regulated by ClpX, and anatomically restricted. Mol. Microbiol. 61, 795–809 (2006).
Imke, S. et al. Regulation of flagellum biosynthesis in response to cell envelope stress in Salmonella enterica Serovar Typhimurium. mBio 9, https://doi.org/10.1128/mbio.00736-17 (2018).
Garima, B., Mostafa, G., Sears, K. T., Galen, J. E. & Tennant. S. M. Genetic engineering of Salmonella spp. for novel vaccine strategies and therapeutics. EcoSal Plus 0, eesp-0004-2023 (2024).
Takaya, A., Yamamoto, T. & Tokoyoda, K. Humoral immunity vs. Salmonella. Front. Immunol. 10, 3155 (2020).
Baliban, S. M., Lu, Y.-J. & Malley, R. Overview of the nontyphoidal and paratyphoidal Salmonella vaccine pipeline: current status and future prospects. Clin. Infect. Dis. 71, S151–S154 (2020).
Sears, K. T., Galen, J. E. & Tennant, S. M. Advances in the development of Salmonella‐based vaccine strategies for protection against Salmonellosis in humans. J. Appl. Microbiol. 131, 2640–2658 (2021).
Lentsch, V. et al. “EvoVax” – A rationally designed inactivated Salmonella typhimurium vaccine induces strong and long-lasting immune responses in pigs. Vaccine 41, 5545–5552 (2023).
Chen, P. & Zhang, J. Antagonistic pleiotropy conceals molecular adaptations in changing environments. Nat. Ecol. Evol. 4, 461–469 (2020).
Cooper, V. S. & Lenski, R. E. The population genetics of ecological specialization in evolving Escherichia coli populations. Nature 407, 736–739 (2000).
MacLean, R. C., Bell, G. & Rainey, P. B. The evolution of a pleiotropic fitness tradeoff in Pseudomonas fluorescens. Proc. Natl. Acad. Sci. USA 101, 8072–8077 (2004).
Cherry, J. L. Selection-driven gene inactivation in Salmonella. Genome Biol. Evol. 12, 18–34 (2020).
Horstmann, J. A. et al. Methylation of Salmonella typhimurium flagella promotes bacterial adhesion and host cell invasion. Nat. Commun. 11, 2013 (2020).
Didelot, X., Walker, A. S., Peto, T. E., Crook, D. W. & Wilson, D. J. Within-host evolution of bacterial pathogens. Nat. Rev. Microbiol. 14, 150–162 (2016).
Hengge, R., Larson, T. J. & Boos, W. sn-Glycerol-3-phosphate transport in Salmonella typhimurium. J. Bacteriol. 155, 186–195 (1983).
Lin, E. C. C. Glycerol dissimilation and its regulation in bacteria. Ann. Rev. Microbiol. 30, 535–578 (1976).
Hayashi, S., Koch, J. P. & Lin, E. C. Active transport of l-α-glycerophosphate in Escherichia coli. J. Biol. Chem. 239, 3098–3105 (1964).
Lemieux, M. J., Huang, Y. & Wang, D.-N. Glycerol-3-phosphate transporter of Escherichia coli: structure, function and regulation. Res. Microbiol. 155, 623–629 (2004).
Xavier, K. B., Kossmann, M., Santos, H. & Boos, W. Kinetic analysis by in vivo 31P nuclear magnetic resonance of internal Pi during the uptake of sn-glycerol-3-phosphate by the pho regulon-dependent Ugp system and the glp regulon-dependent GlpT system. J. Bacteriol. 177, 699–704 (1995).
Steeb, B. et al. Parallel exploitation of diverse host nutrients enhances Salmonella virulence. PLoS Pathog. 9, e1003301 (2013).
Maloney, P. C., Ambudkar, S. V., Anatharam, V., Sonna, L. A. & Varadhachary, A. Anion-exchange mechanisms in bacteria. Microbiol. Rev. 54, 1–17 (1990).
Bruna, R. E., Kendra, C. G. & Pontes, M. H. Phosphorus starvation response and PhoB-independent utilization of organic phosphate sources by Salmonella enterica. Microbiol. Spectr. 0, e02260–23 (2023).
Konings, W. N. & Rosenberg, H. Phosphate transport in membrane vesicles from Escherichia coli. Biochim. Biophys. Acta 508, 370–378 (1978).
Rosenberg, H., Gerdes, R. G. & Chegwidden, K. Two systems for the uptake of phosphate in Escherichia coli. J. Bacteriol. 131, 505–511 (1977).
Van Veen, H. W., Abee, T., Kortstee, G. J. J., Konings, W. N. & Zehnder, A. J. B. Translocation of metal phosphate via the phosphate inorganic transport system of Escherichia coli. Biochemistry 33, 1766–1770 (1994).
Bruna, R. E., Kendra, C. G., Groisman, E. A. & Pontes, M. H. Limitation of phosphate assimilation maintains cytoplasmic magnesium homeostasis. Proc. Natl. Acad. Sci. USA 118, e2021370118 (2021).
Motomura, K. et al. Overproduction of YjbB reduces the level of polyphosphate in Escherichia coli: a hypothetical role of YjbB in phosphate export and polyphosphate accumulation. FEMS Microbiol. Lett. 320, 25–32 (2011).
Bruna, R. E., Kendra, C. G. & Pontes, M. H. An intracellular phosphorus-starvation signal activates the PhoB/PhoR two-component system in Salmonella enterica. mBio 0, e01642–24 (2024).
Spiga, L. et al. An oxidative central metabolism enables Salmonella to utilize microbiota-derived succinate. Cell Host Microbe 22, 291–301.e6 (2017).
Nguyen, B. D. et al. Import of aspartate and malate by DcuABC drives H2/fumarate respiration to promote initial salmonella gut-lumen colonization in mice. Cell Host Microbe 27, 922–936.e6 (2020).
Knöpfel, T. et al. Paracellular transport of phosphate along the intestine. Am. J. Physiol. Gastrointest. Liver Physiol. 317, G233–G241 (2019).
McClure, S. T., Chang, A. R., Selvin, E., Rebholz, C. M. & Appel, L. J. Dietary sources of phosphorus among adults in the United States: results from NHANES 2001–2014. Nutrients 9, 95 (2017).
Davis, G. R. et al. Absorption of phosphate in the jejunum of patients with chronic renal failure before and after correction of vitamin D deficiency. Gastroenterology 85, 908–916 (1983).
Hwan Baek, J. & Yup Lee, S. Novel gene members in the Pho regulon of Escherichia coli. FEMS Microbiol. Lett. 264, 104–109 (2006).
Rong, G. & Stock, A. M. Quantitative kinetic analyses of shutting off a two-component system. mBio 8, e00412 (2017).
Hall, J. A. & Maloney, P. C. Altered oxyanion selectivity in mutants of UhpT, the Pi -linked sugar phosphate carrier of Escherichia coli*. J. Biol. Chem. 280, 3376–3381 (2005).
Röder, J., Felgner, P. & Hensel, M. Single-cell analyses reveal phosphate availability as critical factor for nutrition of Salmonella enterica within mammalian host cells. Cell Microbiol. 23, e13374 (2021).
Gardner, S. G., Johns, K. D., Tanner, R. & McCleary, W. R. The PhoU protein from Escherichia coli interacts with PhoR, PstB, and metals to form a phosphate-signaling complex at the membrane. J. Bacteriol. 196, 1741–1752 (2014).
Bumann, D. & Schothorst, J. Intracellular Salmonella metabolism. Cell Microbiol. 19, e12766 (2017).
Grant, A. J. et al. Modelling within-host spatiotemporal dynamics of invasive bacterial disease. PLoS Biol. 6, e74 (2008).
Daniel, B. B. J. et al. Assessing microbiome population dynamics using wild-type isogenic standardized hybrid (WISH)-tags. Nat. Microbiol. 9, 1103–1116 (2024).
Wuttge, S. et al. Determinants of substrate specificity and biochemical properties of the sn-glycerol-3-phosphate ATP binding cassette transporter (UgpB–AEC2) of Escherichia coli. Mol. Microbiol. 86, 908–920 (2012).
Sweet, G. et al. Glycerol facilitator of Escherichia coli: cloning of glpF and identification of the glpF product. J. Bacteriol. 172, 424 LP–430 (1990).
Hopper, D. J. & Cooper, R. A. The purification and properties of Escherichia coli methylglyoxal synthase. Biochem. J. 128, 321–329 (1972).
Subedi, K. P., Kim, I., Kim, J., Min, B. & Park, C. Role of GldA in dihydroxyacetone and methylglyoxal metabolism of Escherichia coli K12. FEMS Microbiol. Lett. 279, 180–K187 (2008).
Lowry, O. H., Carter, J., Ward, J. B. & Glaser, L. The effect of carbon and nitrogen sources on the level of metabolic intermediates in Escherichia coli. J. Biol. Chem. 246, 6511–6521 (1971).
Kant, S., Liu, L. & Vazquez-Torres, A. The methylglyoxal pathway is a sink for glutathione in Salmonella experiencing oxidative stress. PLoS Pathog. 19, e1011441 (2023).
Hopper, D. J. & Cooper, R. A. The regulation of Escherichia coli methylglyoxal synthase; a new control site in glycolysis? FEBS Lett. 13, 213–216 (1971).
Ozyamak, E. et al. The critical role of S-lactoylglutathione formation during methylglyoxal detoxification in Escherichia coli. Mol. Microbiol. 78, 1577–1590 (2010).
Lee, C., Kim, I. & Park, C. Glyoxal detoxification in Escherichia coli K-12 by NADPH dependent aldo-keto reductases. J. Microbiol. 51, 527–530 (2013).
Hapfelmeier, S. et al. The salmonella pathogenicity island (SPI)-2 and SPI-1 type III secretion systems allow Salmonella Serovar typhimurium to trigger colitis via MyD88-dependent and MyD88-independent mechanisms. J. Immunol. 174, 1675–1685 (2005).
Stecher, B. et al. Salmonella enterica Serovar Typhimurium exploits inflammation to compete with the intestinal microbiota. PLoS Biol. 5, e244 (2007).
Li, W. et al. Strategies adopted by Salmonella to survive in host: a review. Arch. Microbiol. 205, 362 (2023).
Fattinger, S. A., Sellin, M. E. & Hardt, W.-D. Salmonella effector driven invasion of the gut epithelium: breaking in and setting the house on fire. Curr. Opin. Microbiol. 64, 9–18 (2021).
Gül, E. et al. Salmonella T3SS-2 virulence enhances gut-luminal colonization by enabling chemotaxis-dependent exploitation of intestinal inflammation. Cell Rep. 43, 113925 (2024).
Zhang, H. et al. YaeB, expressed in response to the acidic pH in macrophages, promotes intracellular replication and virulence of Salmonella typhimurium. Int. J. Mol. Sci. 20, 4339 (2019).
Koumakis, E. et al. Novel function of PiT1/SLC20A1 in LPS-related inflammation and wound healing. Sci. Rep. 9, 1–15 (2019).
Pang, Y. et al. Bladder epithelial cell phosphate transporter inhibition protects mice against uropathogenic Escherichia coli infection. Cell Rep. 39, 110698 (2022).
Yang, W. et al. Phosphate (Pi) transporter PIT1 induces Pi starvation in Salmonella-containing vacuole in HeLa cells. Int. J. Mol. Sci. 24, 17216 (2023).
Stamp, A. L. et al. Structural and functional characterization of Salmonella enterica serovar Typhimurium YcbL: an unusual Type II glyoxalase. Protein Sci. 19, 1897–1905 (2010).
MacLean, M. J., Ness, L. S., Ferguson, G. P. & Booth, I. R. The role of glyoxalase I in the detoxification of methylglyoxal and in the activation of the KefB K+ efflux system in Escherichia coli. Mol. Microbiol. 27, 563–571 (1998).
Clugston, S. L. et al. Isolation and sequencing of a gene coding for glyoxalase I activity from Salmonella typhimurium and comparison with other glyoxalase I sequences. Gene 186, 103–111 (1997).
Maier, L. et al. Granulocytes impose a tight bottleneck upon the gut luminal pathogen population during Salmonella typhimurium colitis. PLoS Pathog. 10, e1004557 (2014).
Chong, A. et al. Cytosolic replication in epithelial cells fuels intestinal expansion and chronic fecal shedding of Salmonella typhimurium. Cell Host Microbe 29, 1177–1185.e6 (2021).
Bakkeren, E. et al. Salmonella persisters promote the spread of antibiotic resistance plasmids in the gut. Nature 573, 276–280 (2019).
MacDonald, K. P. A. et al. An antibody against the colony-stimulating factor 1 receptor depletes the resident subset of monocytes and tissue- and tumor-associated macrophages but does not inhibit inflammation. Blood 116, 3955–3963 (2010).
Hausmann, A. et al. Intercrypt sentinel macrophages tune antibacterial NF-κB responses in gut epithelial cells via TNF. J. Exp. Med. 218, e20210862 (2021).
Arnold, I. C. et al. CD11c+ monocyte/macrophages promote chronic Helicobacter hepaticus-induced intestinal inflammation through the production of IL-23. Mucosal Immunol. 9, 352–363 (2016).
Ducarmon, Q. R. et al. Gut microbiota and colonization resistance against bacterial enteric infection. Microbiol. Mol. Biol. Rev. 83, e00007-19 (2019).
Ambudkar, S. V., Larson, T. J. & Maloney, P. C. Reconstitution of sugar-phosphate transport systems of Escherichia coli. J. Biol. Chem. 261, 9083–9086 (1986).
Willsky, G. R. & Malamy, M. H. Characterization of two genetically separable inorganic phosphate transport systems in Escherichia coli. J. Bacteriol. 144, 356–365 (1980).
Harris, R. M., Webb, D. C., Howitt, S. M. & Cox, G. B. Characterization of PitA and PitB from Escherichia coli. J. Bacteriol. 183, 5008–5014 (2001).
Stecher, B. B. & Hardt, W.-D. D. Mechanisms controlling pathogen colonization of the gut. Curr. Opin. Microbiol. 14, 82–91 (2011).
Pontes, M. H., Sevostyanova, A. & Groisman, E. A. When too much ATP is bad for protein synthesis. J. Mol. Biol. 427, 2586–2594 (2015).
Lee, E. J., Pontes, M. H. & Groisman, E. A. A bacterial virulence protein promotes pathogenicity by inhibiting the bacterium’s own F1Fo ATP synthase. Cell 154, 146–156 (2013).
Thauer, R. K., Jungermann, K. & Decker, K. Energy conservation in chemotrophic anaerobic bacteria. Bacteriol. Rev. 41, 100–180 (1977).
Kadner, R. J., Murphy, G. P. & Stephens, C. M. Two mechanisms for growth inhibition by elevated transport of sugar phosphates in Escherichia coli. J. Gen. Microbiol. 138, 2007–2014 (1992).
McMeechan, A. et al. Inactivation of ppk differentially affects virulence and disrupts ATP homeostasis in Salmonella enterica serovars Typhimurium and Gallinarum. Res. Microbiol. 158, 79–85 (2007).
Wang, H. et al. Hyperphosphatemia rather than hypophosphatemia indicates a poor prognosis in patients with sepsis. Clin. Biochem. 91, 9–15 (2021).
Nan, W., Huang, Q., Wan, J. & Peng, Z. Association of serum phosphate and changes in serum phosphate with 28-day mortality in septic shock from MIMIC-IV database. Sci. Rep. 13, 21869 (2023).
Livermore, D. M. et al. What remains against carbapenem-resistant Enterobacteriaceae? Evaluation of chloramphenicol, ciprofloxacin, colistin, fosfomycin, minocycline, nitrofurantoin, temocillin and tigecycline. Int. J. Antimicrob. Agents 37, 415–419 (2011).
Falagas, M. E., Kastoris, A. C., Kapaskelis, A. M. & Karageorgopoulos, D. E. Fosfomycin for the treatment of multidrug-resistant, including extended-spectrum β-lactamase producing, Enterobacteriaceae infections: a systematic review. Lancet Infect. Dis. 10, 43–50 (2010).
Monack, D. M., Raupach, B., Hromockyj, A. E. & Falkow, S. Salmonella typhimurium invasion induces apoptosis in infected macrophages. Proc. Natl. Acad. Sci. USA 93, 9833–9838 (1996).
Chen, L. M., Kaniga, K. & Galán, J. E. Salmonella spp. are cytotoxic for cultured macrophages. Mol. Microbiol. 21, 1101–1115 (1996).
Pokorzynski, N. D. & Groisman, E. A. How bacterial pathogens coordinate appetite with virulence. Microbiol. Mol. Biol. Rev. 87, e00198–22 (2023).
Eylert, E. et al. Carbon metabolism of Listeria monocytogenes growing inside macrophages. Mol. Microbiol. 69, 1008–1017 (2008).
Burton, N. A. et al. Disparate impact of oxidative host defenses determines the fate of salmonella during systemic infection in mice. Cell Host Microbe 15, 72–83 (2014).
Choi, S. et al. The Salmonella virulence protein MgtC promotes phosphate uptake inside macrophages. Nat. Commun. 10, 1–14 (2019).
Diard, M. et al. Antibiotic treatment selects for cooperative virulence of Salmonella typhimurium. Curr. Biol. 24, 2000–2005 (2014).
Lenski, R. E. & Levin, B. R. Constraints on the coevolution of bacteria and virulent phage: a model, some experiments, and predictions for natural communities. Am. Nat. 125, 585–602 (1985).
Lenski, R. E. Experimental studies of pleiotropy and epistasis in Escherichia coli. I. Variation in competitive fitness among mutants resistant to virus t4. Evolution 42, 425–432 (1988).
Barthel, M. et al. Pretreatment of mice with streptomycin provides a Salmonella enterica serovar Typhimurium colitis model that allows analysis of both pathogen and host. Infect. Immun. 71, 2839–2858 (2003).
De Paepe, M. et al. Trade-off between bile resistance and nutritional competence drives Escherichia coli diversification in the mouse gut. PLoS Genet. 7, e1002107 (2011).
Sousa, A. et al. Recurrent reverse evolution maintains polymorphism after strong bottlenecks in commensal gut bacteria. Mol. Biol. Evol. 34, 2879–2892 (2017).
Cherrak, Y. et al. Non-canonical start codons confer context-dependent advantages in carbohydrate utilization for commensal E. coli in the murine gut. Nat. Microbiol. 9, 2696–2709 (2024).
Hoiseth, S. K. & Stocker, B. A. D. Aromatic-dependent Salmonella typhimurium are non-virulent and effective as live vaccines. Nature 291, 238–239 (1981).
Schubert, C. et al. Monosaccharides drive salmonella gut colonization in a context-dependent manner. bioRxiv 2024.08.06.606610 https://doi.org/10.1101/2024.08.06.606610 (2024).
Datsenko, K. A. & Wanner, B. L. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97, 6640–6645 (2000).
Sternberg, N. L. & Maurer, R. B. T.-M. in E. [2] Bacteriophage-mediated generalized transduction in Escherichia coli and Salmonella typhimurium. in Bacterial Genetic Systems, Vol. 204 18–43 (Academic Press, 1991).
Porwollik, S. et al. Defined single-gene and multi-gene deletion mutant collections in Salmonella enterica sv Typhimurium. PLoS ONE 9, e99820 (2014).
Gibson, D. G. et al. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat. Methods 6, 343–345 (2009).
Liebisch, G. et al. Quantification of fecal short chain fatty acids by liquid chromatography tandem mass spectrometry—investigation of pre-analytic stability. Biomolecules 9, 121 (2019).
Oberg, N., Zallot, R. & Gerlt, J. A. EFI-EST, EFI-GNT, and EFI-CGFP: enzyme function initiative (EFI) web resource for genomic enzymology tools. J. Mol. Biol. 435, 168018 (2023).
Zallot, R., Oberg, N. & Gerlt, J. A. The EFI Web resource for genomic enzymology tools: leveraging protein, genome, and metagenome databases to discover novel enzymes and metabolic pathways. Biochemistry 58, 4169–4182 (2019).
Kumar, S., Stecher, G. & Tamura, K. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874 (2016).
Diard, M. et al. Inflammation boosts bacteriophage transfer between Salmonella spp. Science 355, 1211–1215 (2017).
Tran, K. T., Maeda, T., Sanchez-Torres, V. & Wood, T. K. Beneficial knockouts in Escherichia coli for producing hydrogen from glycerol. Appl. Microbiol. Biotechnol. 99, 2573–2581 (2015).
Acknowledgements
We want to acknowledge the highly appreciated work of the staff at the ETH animal facility (EPIC and RCHCI, especially Manuela Graf, Katharina Holzinger, Dennis Mollenhauer, Sven Nowok, and Dominik Bacovcin). We are grateful to Patrick Andreassen, Joshua Newson, and Luca Maurer for the introduction to macrophage infections, intraperitoneal injections, and the microtome, respectively. We also want to thank Dirk Bumann, Randall Platt, and members of the Hardt, Sunagawa, and Slack labs for helpful comments and discussions. Furthermore, we appreciate Andrew Abi Younes and Patrick Andreassen’s input and feedback on this manuscript. NSdS was funded by a Boehringer Ingelheim Fonds PhD fellowship. The Swiss National Science Foundation funded this work via grant 310030_192567 to WDH and via the NCCR Microbiomes grant to JAV and WDH. C.S is supported by the German Research Foundation (SCHU 3606/1-1). Y.C. is supported by an EMBO long-term fellowship (ALTF-234-2020) and a flexibility grant from the SNF/NCCR Microbiome (51NF40_180575). Figures 1a, 2d, 3a, 4a, b, and 5b were created by the first author (NSdS) for this study using BioRender.com. The figures are subject to a CC-BY license, where we hold publication rights granted by BioRender.com.
Author information
Authors and Affiliations
Contributions
NSdS, BDN, CS, and WDH conceived the project and designed the experiments. NS carried out the mouse experiments and analyzed the data. NSdS, MV, TBA, EB, CS, and BDN cloned the strains, plasmids, and tags. NSdS and SK carried out flow cytometry experiments. MB performed histopathology. PC, PK, and JAV contributed the LC-MS measurement and analysis. WDH provided the resources for the study. NS wrote the initial draft of the manuscript. NSdS, YC, EB, and WDH edited the draft. All authors read, commented on, and approved this manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Santamaria de Souza, N., Cherrak, Y., Andersen, T.B. et al. Context-dependent change in the fitness effect of (in)organic phosphate antiporter glpT during Salmonella Typhimurium infection. Nat Commun 16, 1912 (2025). https://doi.org/10.1038/s41467-025-56851-5
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-025-56851-5
