LuxS/AI-2 quorum sensing is required by Salmonella Typhimurium to survive in the macrophages

To begin our study, we first investigated the levels of AI-2 produced at different growth phases from Salmonella Typhimurium (STM) WT grown at 37°C in LB medium by using the Vibrio harveyi BB-170 reporter strain (which produces luminescence responses to AI-2). Consistent with previous studies [16], AI-2 dependent light production by V. harveyi BB-170 is maximum in the spent medium of the mid-log phase culture of Salmonella Typhimurium, and it gradually decreases towards the late stationary phase (S1A and S1B Fig). Next, we determined the mRNA expression level in vitro growth in LB medium, and we observed that while Salmonella upregulated the expression of the genes lsrB, lsrK, and lsrR during its mid-log phase to early stationary phase of growth (S1C Fig), it did not alter the expression of luxS during this period in LB medium (S1D Fig). And the deletion of luxS and genes (lsrK, lsrB, lsrR) in Salmonella did not affect the in vitro growth in LB and minimal medium (S1E and S1F Fig). Next, we examined whether AI-2 signaling influences the infection of Salmonella in macrophages. We first assessed bacterial uptake by RAW 264.7 macrophages and observed that, compared to STM WT, the quorum-sensing mutants STM ΔluxS, STM ΔlsrB, and STM ΔlsrK exhibited increased phagocytosis (Fig 1B). In contrast, the STM ΔlsrR mutant and the luxS-complemented STM ΔluxS strain displayed phagocytosis levels comparable to the STM WT (Fig 1B). To determine whether AI-2 signaling is required for intracellular survival rather than uptake, we next evaluated bacterial proliferation within macrophages. Despite their increased phagocytosis, STM ΔluxS, STM ΔlsrB, and STM ΔlsrK showed impaired intracellular survival in RAW 264.7 cells (Fig 1C). By contrast, as expected, the lsrR mutant and the luxS-complemented STM ΔluxS strain behaved similarly to STM WT (Fig 1C). Our confocal microscopy data suggest that, under the tested conditions, STM ΔluxS, STM ΔlsrB, and STM ΔlsrK show reduced intracellular proliferation compared with STM WT (S1G and S1H Fig). Further, we infected RAW 264.7 macrophages with Salmonella strains for an intracellular gene expression study and observed that the luxS and the lsrB, lsrK, and lsrR genes were upregulated 1.5–2-fold from 6h post-infection to 16h compared to 2h post-infection (Fig 1D and 1E). We further corroborated our results by using primary macrophages derived from the peritoneal lavage of C57BL/6J mice for infection. The luxS, lsrB, and lsrK mutants also exhibited an increased uptake by the primary macrophages, but their proliferation was compromised (Fig 1F and 1G). Furthermore, we checked for intracellular AI-2 production by STM WT, STM ∆luxS, and STM ∆lsrB upon their infection of RAW 264.7 cells and observed the AI-2 production by STM WT and STM ∆lsrB till 16h post-infection, but not by STM ∆luxS (Fig 1H). Taken together, these findings imply that Salmonella Typhimurium requires LuxS/AI-2 signaling to survive in the hostile environment of macrophages.

STM ∆luxS acquires AI-2 from STM WT during its in vitro growth

To find out if STM ∆luxS can use AI-2 molecules produced by STM WT, we co-infected RAW 264.7 macrophages with STM WT and STM ∆luxS (1:1 ratio), and observed that STM WT outcompeted STM ∆luxS, hinting that STM ∆luxS may not be able to utilize the AI-2 from STM WT within host macrophages (Fig 2A). The limitation of this experiment is that we cannot conclusively determine whether STM ∆luxS utilizes AI-2 from STM WT in the coinfection study in RAW 264.7 cells. Furthermore, the study does not confirm whether both bacterial strains infect the same cell, as observations reflect a global phenotype. However, it might be possible that STM ∆luxS cannot utilize AI-2 even when infecting the same cell due to a repressed lsr operon. On the contrary, when STM WT and STM ∆luxS were cocultured in LB medium before infection into macrophages, STM WT and STM ∆luxS proliferated similarly in RAW 264.7 cells (Fig 2B). To check if STM ∆luxS can utilize the AI-2 molecule produced by STM WT, we treated STM ∆luxS with the STM WT spent medium and exogenous DPD molecule. Upon infection into RAW 264.7 macrophages, the treated STM ∆luxS proliferated similarly to STM WT, suggesting its ability to use the AI-2 from the spent medium of STM WT or exogenous DPD (Figs 2C, 2D, and S2A–S2D). Next, we used a commercially available inhibitor (Z)-4 bromo-5-(bromomethyl)-3-methylfuran-2(5H) (BF-8) of LuxS/AI-2 signaling [24,25] at different concentrations with Salmonella Typhimurium and observed that while BF-8 did not affect Salmonella Typhimurium growth in vitro, but it attenuated Salmonella survival in RAW 264.7 macrophages, similar to that observed with STM ∆luxS (Figs 2E, 2F, and S2E). We also confirmed that exogenous DPD induces luminescence, whereas the inhibitor BF-8 inhibits luminescence. We observed that upon treatment of Salmonella with synthetic DPD or the inhibitor BF-8, AI-2 activity, measured as luminescence using an autoinducer assay, was increased or decreased, respectively (S2F and S2G Fig). To further validate the effects of synthetic DPD or BF-8, peritoneal macrophages were infected with the respective treated cultures. The data suggested that while exogenous DPD enhanced STM ΔluxS survival, treatment with the AI-2 inhibitor BF-8 significantly attenuated STM WT proliferation (S2H and S2I Fig). These findings imply that the luxS mutant can utilize exogenous AI-2 during its in vitro growth.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. STM ∆luxS acquires AI-2 from STM WT during its in vitro growth.

(A) Co-infection assay between STM WT and STM ∆luxS upon infection in RAW 264.7 macrophages at the ratio 1:1. (B) Co-culture grown STM WT and STM ∆luxS followed by infection in RAW 264.7 macrophages (All data in (A) and (B) are represented as mean ± SD from three independent experiments (N = 3, n = 3). (C) Percentage phagocytosis, (D) Fold proliferation of STM in RAW 264.7 macrophages upon 50% spent medium treatment of STM WT and STM ∆luxS (data is from one experiment representative of 3 independent experiments, N = 3,n = 4). (E) Percentage phagocytosis, (F) Fold proliferation of STM in RAW 264.7 macrophages upon treatment of BF-8 inhibitor (data is from one experiment representative of 3 independent experiments, n = 4) (G) AI-2 production by STM WT, STM ∆luxS, and STM ∆lsrB in F-medium (pH 5) (Data is represented as mean ±SD of N = 3, n = 3). STM WT, STM ΔluxS, STM ΔlsrB, STM ΔlsrK, and STM ΔlsrR survival in different ranges of pH [3–8] of PBS (H) pH3 survival, (I) pH4 survival, (J) pH5 survival (data is from one experiment representative of 3 independent experiments, N = 3, n = 4). Unpaired Student’s t-test was used to analyze the data; One-way ANOVA with Dunnett’s post-hoc test was used to analyze the data; p values **** p < 0.0001, *** p < 0.001, ** p < 0.01, * p < 0.05.

https://doi.org/10.1371/journal.ppat.1014244.g002

The loss of a functional AI-2 or its receptor-mediated sensing attenuates survival at acidic pH

The mechanisms underlying bacterial quorum sensing and pH sensing have been thoroughly investigated and shown to be distinct/independent gene regulatory systems. Our AI-2 bioassay in acidic medium shows AI-2 production by Salmonella (Fig 2G). Hence, we hypothesize that LuxS/AI-2 signaling may coordinate with the pH sensing system of Salmonella to regulate survival in an acidic environment. Firstly, when we assessed the expression of mRNAs of luxS, lsrB, lsrK, and lsrR during its growth in F-medium (mimics the acidic vacuolar condition), we observed ~1.5-fold increases in luxS at the mid-log phase to the late-log phase (S3A Fig), and lsrB, lsrK, and lsrR at the mid-log phase (S3B Fig). Moreover, we observed that Salmonella Typhimurium stably produces AI-2 even during its growth in F-medium (at pH 5), which is maximal during the mid to late-log phase. However, this reflects accumulated AI-2 rather than direct per-cell production (Fig 2G). The luminescence production continued in STM ∆lsrB spent medium even beyond 12 h (because there was no accumulation of AI-2 inside the STM ∆lsrB). However, there was no light production from the spent medium of STM ΔluxS (Fig 2G).

To underscore the role of AI-2 in facilitating survival in an acidic environment, we performed an acid survival assay in PBS and LB (a different range of pH) medium and noted that STM WT, STM ΔluxS, STM ΔlsrB, STM ΔlsrK, and STM ΔlsrR survived equally well at pH6 to pH8 in both medium (S3C and S3D Fig). However, at pH3, pH4, and pH5, STM WT survived significantly better compared to STM ΔluxS, STM ΔlsrB, STM ΔlsrK (Figs 2H–2J and S3E–S3G). We conclude that LuxS/AI-2 signaling and communication facilitate Salmonella’s survival strategies in acidic environments under in vitro conditions.

LuxS/AI-2 system tunes phoP/phoQ expression, thereby assisting in pH sensing and adaptation

A well-known two-component system (TCS) of PhoP/PhoQ is essential in mediating the acid tolerance response (ATR) in several Gram-negative bacteria, including Salmonella [26,27]. Thus, we conjectured that LuxS/AI-2 signaling modulates the phoP/phoQ gene expression, which, in turn, aids in the survival of Salmonella under an acidic environment. Thus, we questioned AI-2 signaling is essential to regulate the expression of the downstream gene targets apart from the lsr operon.

We first determined phoP gene expression in STM WT grown in LB (neutral), acidic LB (pH 5), F-medium, and during infection of RAW 264.7 macrophages (S4A–S4D Fig). phoP expression varies across conditions, with altered expression observed in LB (neutral and acidic), F-medium, and RAW macrophages, supporting the context-dependent regulation of phoP under environmental and intracellular conditions. We then compared phoP mRNA levels in STM ΔluxS, STM ΔlsrB, and STM ΔlsrK with STM WT and observed downregulation in LB, acidic LB (pH 5), and F-medium (Figs 3A, S4E, and S4F). Interestingly, the phoP expression was rescued when STM ΔluxS was grown in STM WT spent medium (neutral and acidic LB medium) (Figs 3B and S4G). Additionally, upon infection in macrophages, both phoP and phoQ were underexpressed in the mutants STM ΔluxS, STM ΔlsrB, and STM ΔlsrK compared to STM WT (Fig 3C and 3D). Hence, we conclude that AI-2 sensing and the signaling cascade orchestrate the phoP/phoQ expression in Salmonella Typhimurium.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. LuxS/AI-2 system tunes phoP expression, thereby assisting in pH sensing and adaptation.

(A) phoP gene expression in STM WT, STM ΔluxS, STM ΔlsrB, and STM ΔlsrK in vitro conditions, LB medium (pH5). (B) mRNA expression of phoP in STM ΔluxS upon treatment of STM WT spent medium (LB, acidic pH 5). (C) phoP (D) phoQ expression in STM WT, STM ΔluxS, STM ΔlsrB, and STM ΔlsrK upon infection in RAW 264.7 macrophages. (E) STM WT, STM ΔluxS, STM ΔlsrB, STM ΔlsrK, and STM ΔlsrR acid adaptation and tolerance. (F) Cytosolic pH determination of STM WT, STM ΔluxS, STM ΔlsrB, STM ΔlsrK, and STM ΔlsrR by phuji plasmid in LB medium, in F-medium (pH 5), and upon infection in RAW 264.7 macrophages. (G) Ratio of cadaverine and lysine under acidic conditions (pH 5) (H) cadA, (I) cadB, and (J) cadC, mRNA expression upon infection in RAW 264.7 macrophages. One-way ANOVA with Dunnett’s post-hoc test was used to analyze the data; Two-way ANOVA was used to analyze the grouped data; p values **** p < 0.0001, *** p < 0.001, ** p < 0.01, * p < 0.05. All data are represented as mean ± SD from three independent experiments (N = 3, n ≥ 3).

https://doi.org/10.1371/journal.ppat.1014244.g003

Next, we evaluated the acid tolerance response (ATR) and noted that with a prior adaptation to pH 5, STM WT and STM ΔlsrR were resilient and showed better survival upon exposure to pH 3 compared to STM ΔluxS, STM ΔlsrB, and STM ΔlsrK (Fig 3E). As the STM ΔluxS, STM ΔlsrB, and STM ΔlsrK strains exhibited a diminished ATR, we determined the relative cytoplasmic acidification of all the strains using a pH-sensitive pHuji plasmid. We observed that all Salmonella strains maintained a neutral cytosolic pH in LB medium. However, when grown in acidic F-medium (pH 5), STM WT and STM ΔlsrR could maintain their cytosolic pH within the near-neutral range, while cytosolic pH for STM ΔluxS, STM ΔlsrB, and STM ΔlsrK dropped to an acidic range and a comparable reduction was observed within the hostile macrophage environment (Figs 3F and S4H–S4L). This indicates that these mutants are less able to maintain intracellular pH homeostasis under acidic stress.

The PhoP/PhoQ TCS aids in pH homeostasis via the cadC/BA system (transcriptional regulator, lysine/cadaverine antiporter, and lysine decarboxylase), thereby promoting bacterial survival in mildly acidic environments [28,29]. So, we measured the expression of cadC, cadB, and cadA in STM WT and STM ∆phoP upon infection in RAW 264.7 macrophages and observed diminished mRNA expression of cadC, cadB, and cadA in STM ∆phoP upon infection in RAW 264.7 macrophages, validating the above observation (S5A–S5C Fig). The lysine decarboxylase (CadA) quenches the H⁺ while converting lysine to cadaverine, which is exported to the extracellular milieu by CadB [28]. We performed mass spectrometry analysis to quantify cadaverine and lysine in the culture supernatant. Here, we noted the cadaverine/lysine ratio to be low in STM ΔluxS, STM ΔlsrB, and STM ΔlsrK compared to STM WT under an acidic pH of 5 (Fig 3G). Additionally, in the mutants, the cadA, cadB, and cadC genes were under-expressed at 2h, 6h, and 16h post-infection into RAW 264.7 cells (Fig 3H–3J). From a mechanistic perspective, we conclude that LuxS/AI-2 signaling regulates ATR and maintains the cytosolic pH of Salmonella by regulating the cadC/AB operon via phoP/phoQ.

LuxS/AI-2 signaling regulates phoP/phoQ expression through LsrR interaction with the phoP promoter

We explored whether LsrR, the downstream regulator of the LuxS/AI-2 signaling pathway, could directly interact with the phoP promoter to influence its activity. The AlphaFold3 structure suggested that LsrR could bind to both the DNA strands of the lsr promoter, but to only one of the strands of the phoP promoter [30]. The predicted interactions are mediated by residues Y25, T31, Q32, S33, R43, L44, K45, S47, and R48, all within a distance of <5 Å and a predicted aligned error (PAE) of <14 across multiple replicates (Figs 4A, 4B, S6A, and S6B). These residues are located in the N-terminal domain, spanning positions 25–48, between alpha helices 1, 2, and 3. Notably, Y25, T31, and R43 showed the most significant interactions, with R43 from helix 3 displaying the highest number of contact points [31]. A covariance analysis of the multiple sequence alignment revealed that R43 most frequently co-occurs with Y25, an interacting residue located in helix 1, with the highest probability among all analyzed residues [32]. Given the strong predicted interactions of R43 and Y25 with the lsr and phoP promoters, we generated Y25A, R43A, and Y25A/R43A mutants and used AlphaFold3 to predict structural impacts. The AlphaFold structures indicated that the mutations disrupted promoter interactions, with PAE values exceeding 18 and pLDDT scores for all the DNA-contacting residues dropping below 30. Additionally, we performed NPT ensemble molecular dynamics simulations for the mutant protein-DNA complex, revealing a significant weakening of interactions. We validated our findings in vitro through electrophoretic mobility shift assays (EMSA) using purified wild-type LsrR and its Y25A, R43A, and Y25A/R43A mutants to assess their binding to HPLC-purified lsr and phoP promoters (Figs 4C, 4D, and S6C–S6F). We observed that WT LsrR binds to both the promoter, whereas the double mutant loses binding. Further, R43A mutant LsrR reduced it, and the double mutant lost nearly all binding at 200mM NaCl (S7A and S7B Fig). A similar result was observed for the interaction of LsrR with a single-stranded promoter sequence (S7C–S7F Fig), and a random DNA sequence did not show any interaction (S7G Fig). So, LsrR is predicted, based on in-silico analysis, to preferentially interact with the single-stranded phoP promoter; however, EMSA results show that it binds to both single- and double-stranded forms of the phoP/Q promoter. To assess regulatory effects during infection, we generated a luxS/lsrR double knockout and complemented it with lsrR (WT) and a mutant lsrR gene. Next, we performed infection assays in RAW 264.7 macrophages using the luxS/lsrR double knockout strain complemented with either WT lsrR or mutant lsrR. While the double knockout strain proliferated similarly to STM WT, lsrR complementation reduced proliferation. The Y25A/R43A mutant restored survival to double knockout levels (Fig 4E and 4F). However, we further infected RAW264.7 macrophages with phoP complemented under a non-native promoter in STM ΔluxS and observed rescued survival of STM ΔluxS (S8A and S8B Fig). Together, these data indicate that LsrR binding to the lsr and phoP promoters acts as a negative regulator of Salmonella pathogenicity.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. LuxS/AI-2 signaling regulates phoP/phoQ expression through LsrR interaction with the phoP promoter in Salmonella.

AlphaFold3 structural models visualised using PyMol depicting LsrR (green) interactions with the phoP and lsr promoters (yellow). The N-terminal domain residues, spanning helices 1, 2, and 3, are highlighted in blue, and the key interacting residues of LsrR are represented with a ball and stick model (A, B). Electrophoretic mobility shift assays (EMSA) assessing double-strand DNA binding of wild-type LsrR and its Y25A/R43A mutant to lsr and phoP promoters, respectively. (data is from one experiment representative of 3 independent experiments) (C, D). Intracellular survival assay (ICSA) with mutated Y25A, R43A, and Y25A/R43A lsrR cloned in pQE-60 complemented strains (E) phagocytosis and (F) Fold proliferation (data is from one experiment representative of 3 independent experiments). One-way ANOVA with Dunnet’s post-hoc test was used to analyze the data; p values **** p < 0.0001, *** p < 0.001, ** p < 0.01, * p < 0.05.

https://doi.org/10.1371/journal.ppat.1014244.g004

AI-2 orchestrates the Salmonella Pathogenicity Island-2 gene expression by sensing low pH

PhoP has been shown to control the TCS SsrA/SsrB at the transcription level, which are the master regulators of the SPI-2 genes [33]. The expression of ssrA and ssrB genes was also downregulated in STM ∆phoP upon infection into RAW 264.7 cells (S9A and S9B Fig). Thus, we hypothesized that LuxS/AI-2 signaling might control the SsrB/SsrA system via PhoP/PhoQ. Upon infection into RAW 264.7 macrophages, the mRNA expression of ssrB was downregulated from early stage post-infection to late stage post-infection in macrophages (Fig 5A). On the other hand, ssrA was downregulated at a late stage post-infection (Fig 5B). Furthermore, we observed that expression of ssaV, a Type 3 secretion system (T3SS) needle complex protein encoded by SPI-2, and spiC, an SPI-2-encoded effector molecule, was under-expressed in the absence of luxS, lsrB, and lsrK genes in Salmonella upon infection of RAW 264.7 macrophages (Fig 5C and 5D). Since SPI-2 effectors are known to modulate host immune signaling and promote an anti-inflammatory (M2-like) phenotype favorable for intracellular survival, we hypothesized that quorum-sensing–defective mutants would impair this modulation [34,35]. Notably, this impaired SPI-2 gene expression correlated with altered host cytokine responses (Proinflammatory Tnf-α and anti-inflammatory Il-10) wherein STM WT promoted a more anti-inflammatory (Il-10, M2-like) macrophage profile compared to quorum-sensing–defective mutants (S9C and S9D Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. AI-2 orchestrates the Salmonella pathogenicity island-2 gene expression by sensing low pH.

(A) ssrB (B) ssrA (C) ssaV (D) spiC, mRNA expression in STM WT, STM ΔluxS, STM ΔlsrB, and STM ΔlsrK upon infection in RAW 264.7 macrophages. STM WT, STM ΔluxS, STM ΔlsrB, STM ΔlsrK, STM WT: phoP, STM ΔluxS: phoP, STM ΔlsrB: phoP, STM ΔlsrK: phoP survival in PBS at (E) pH 3, (F) pH 4, (G) pH 5. (H) STM WT, STM ΔluxS, STM ΔlsrB, STM ΔlsrK, STM WT:phoP, STM ΔluxS:phoP, STM ΔlsrB:phoP, STM ΔlsrK: phoP acid adaptation and tolerance. ssrB mRNA expression in phoP/phoQ cloned strain of (I) STM ΔluxS, (J) STM ΔlsrB and (K) STM ΔlsrK. One-way ANOVA with Dunnet’s post-hoc test was used to analyze the data; Two-way ANOVA was used to analyze the grouped data; p values **** p < 0.0001, *** p < 0.001, ** p < 0.01, * p < 0.05. All data are represented as mean ± SD from three independent experiments (N = 3, n = 3).

https://doi.org/10.1371/journal.ppat.1014244.g005

So far, our findings convey that the LuxS/AI-2 signaling regulates phoP/phoQ genes and, thereby, the SPI-2 cluster in Salmonella. However, whether AI-2 directly or the AI-2-PhoP/PhoQ axis controls the SPI-2 genes remains unclear. Thus, we cloned the phoP/phoQ operon under a non-native promoter in the pQE60 vector. Firstly, we confirmed the phoP mRNA overexpression level in STM WT: pQE60-phoP/phoQ and STM ∆luxS: pQE60-phoP/phoQ upon infection in RAW 264.7 (S9E and S9F Fig). Using these complemented Salmonella strains, we observed that STM ΔluxS, STM ΔlsrB, and STM ΔlsrK with a phoP/phoQ cloned under a non-native promoter survived equally at acidic pH3, 4, and 5 as STM WT (Fig 5E–5G). Furthermore, the STM ΔluxS, STM ΔlsrB, and STM ΔlsrK with cloned phoP/phoQ showed enhanced ATR response (Fig 5H). Thus, when phoP/phoQ is not under its native promoter, the regulation by LuxS/AI-2 is not required. Also, ssrB and ssrA expression are significantly higher in phoP/phoQ complemented knockout strains compared to STM ΔluxS, STM ΔlsrB, and STM ΔlsrK strains (Figs 5I–5K and S9G–S9I). Cumulatively, these results suggest that, indeed, the LuxS/AI-2 signalling controls ATR by regulating the phoP/phoQ genes.

The absence of AI-2 mediated signaling compromises the in vivo colonization of Salmonella in mice

Because AI-2 is a universal signaling molecule, it is an excellent choice for mediating interactions between cells in the mammalian gut, where hundreds of bacterial species live and interact. Thus, we aimed to understand the role of LuxS/AI-2 signalling in regulating Salmonella pathogenesis in in vivo mouse models. We infected C57BL/6J mice by orally gavaging with STM WT and knockout strains at a CFU of 10⁷ per mouse (Fig 6A). We observed that STM ΔluxS, STM ΔlsrB, and STM ΔlsrK showed a significantly lower organ burden in the intestine, mesenteric lymph node (MLN), spleen, and liver, and lower bacteraemia upon oral gavage (Fig 6B–6F). Oral gavage mimics the physiological route of Salmonella infection into its host, and Salmonella is required to be able to breach the intestinal barrier successfully. The lower organ colonization in STM ΔluxS, STM ΔlsrB, and STM ΔlsrK can be explained by their inability to cross the gut epithelial barrier. Next, we bypassed the gut epithelial barrier by infecting C57BL/6 mice intraperitoneally and found that STM ΔluxS, STM ΔlsrB, and STM ΔlsrK still exhibited reduced colonization in the spleen and liver and less dissemination in blood than STM WT (S10A–S10D Fig). Furthermore, upon infection of mice by oral gavage at a CFU of 10⁸ per mouse, we noted that the mice infected with STM WT and STM ΔlsrR succumbed to death as early as the 6th day of post-infection, while STM ΔluxS, STM ΔlsrB, and STM ΔlsrK infected mice survived longer (Fig 6G–6H) with less weight reduction (S10E Fig). Further, we observed that the BF-8 inhibitor at 4 mg/kg reduced the STM WT colonization to different sites of infection in C57BL/6 mice (Fig 6I–6N) and improved the survival of mice with delayed weight reduction compared to untreated mice (Figs 6O, 6P, and S10F). Liver tissue histopathology results also suggest that treatment with BF-8 reduced the disease score (S9G Fig). Finally, we confirmed our observation by using phoP complemented under the non-native promoter STM ∆luxS. We infected C57BL/6J mice by orally gavaging strains at a CFU of 10⁷ per mouse. Post 5^th day of infection, we determined the organ burden and observed that STM ∆luxS: pQE60-phoP/phoQ (phoP complemented under non-native promoter) colonized significantly higher in the intestine, MLN, Liver, Spleen, and more disseminated in blood compared to STM ∆luxS (S11A–S11E Fig). Furthermore, upon infection of mice by oral gavage at a CFU of 10⁸ per mouse, we noted that the mice infected with STM ∆luxS: pQE60-phoP/phoQ succumbed to death earlier compared to STM ΔluxS, infected (S11F and S11G Fig). Conclusively, our results show that LuxS/AI-2 signaling is critical for the in vivo pathogenesis and virulence of Salmonella Typhimurium.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. The absence of AI-2 mediated signaling compromises the in vivo colonization of Salmonella in mice.

(A) The experimental protocol for organ burden in C57BL/6J mice by orally gavaging 10⁷ CFU per mouse [36]. The organ burden post 5 days of oral gavage in (B) Intestine, (C)Mesenteric lymph node (MLN), (D) Spleen, (E) Liver, and dissemination in (F) Blood. (G). The experimental protocol for C57BL/6J mice survival by orally gavaging 10⁸ CFU of STM per mouse. (H) Percentage survival of Mice upon infection with STM WT, STM ΔluxS, STM ΔlsrB, STM ΔlsrK, STM ΔlsrR, and STM ΔluxS:luxS. (I) The experimental protocol for organ burden in C57BL/6J mice by orally gavaging 10⁷ CFU per mouse upon treatment of BF-8 inhibitor (2mg/kg and 4mg/kg through the Intraperitoneal route). The organ burden post 5 days of oral gavage in (J) Intestine, (K) Mesenteric lymph node (MLN), (L) Spleen, (M) Liver, and blood dissemination in (N) Blood. (O) The experimental protocol of mice survival in C57BL/6J mice by orally gavaging 10⁸ CFU per mouse upon treatment of BF-8 inhibitor (2mg/kg and 4mg/kg through the Intraperitoneal route). (P) Percentage survival of Mice upon infection with STM WT, STM WT upon treatment, and STM ΔluxS. Non-parametric One-way ANOVA (Kruskal-Wallis) with Dunn’s post hoc test was used to analyze organ burden in mice. (data is from one experiment representative of 2 independent experiments, N = 2, n = 5).

https://doi.org/10.1371/journal.ppat.1014244.g006

Ethics statement

All experiments comply with the rules set forth by the Indian Institute of Science, Bangalore’s Institutional Animal Ethics Committee (IAEC). The approved protocol numbers are CAF/Ethics/852/2021 and CAF/Ethics/116/2025. The Committee for Control and Supervision of Experiments on Animals (CPCSEA), a statutory committee established under Chapter 4, Section 15(1) of the Prevention of Cruelty to Animals Act 1960, and National Animal Care provided guidelines that were meticulously adhered to during all animal experiments, all of which were approved by the Institutional Animal Ethics Committee. (Registration No. 435 48/1999/CPCSEA). The Institutional Animal Ethics Committee approved every animal experiment, and the National Animal Care Guidelines were meticulously followed.

Bacterial strains and growth conditions

The wild-type Salmonella enterica serovar Typhimurium strain 14028S (STM WT) used in all experiments in this study was a kind gift from Professor Michael Hensel of the Max Von Pettenkofer-Institute for Hygiene und Medizinische Mikrobiologie in Germany. The bacterial strains were revived on LB agar with or without antibiotics. The LB broth culture of wild-type, knockout, and complemented strains was cultured at 37˚C (170 rpm) in an orbital shaker incubator (S1 Table). Antibiotics Kanamycin (50 μg/ml), Chloramphenicol (25 μg/ml), and Ampicillin (50 μg/ml) were used whenever required. Vibrio campbellii ATTC BAA-1117 strain was used for autoinducer (AI-2) bioassay. This strain was grown in ATCC Medium: 2746 Autoinducer Bioassay (AB) Medium [47].

A growth kinetics study was conducted in LB and M9 minimal medium [48,49] (S2 Table). Briefly, STM WT and mutants were inoculated in fresh LB broth (1:100 ratio from overnight primary culture). The cultures were incubated at 37 ˚C (170 rpm) in an orbital shaker incubator. The OD at 600nm was measured at regular intervals until 24 hours post-inoculation.

Bacterial gene knockout and strain generation

Salmonella enterica serovar Typhimurium strain 14028S and isogenic mutants were used for all assays. luxS, lsrB, lsrK, and lsrR gene knockout strains were made by λ-Red recombination system as described previously in one gene step chromosomal gene inactivation method demonstrated by Datsenko and Wanner [50]. Briefly, STM WT strains were transformed with pKD46 plasmid expressing λ-red recombinase system under Arabinose inducible plasmid. The pKD3 and pKD4 plasmid were used as the template to amplify the chloramphenicol and kanamycin resistance cassette with knockout primers (S3 Table). The amplified reaction products were purified using chloroform isopropanol precipitation. The purified PCR product was transformed into pKD46- containing STM WT expressing λ-Red recombinase system. The transformed cells were plated on LB agar with chloramphenicol and kanamycin accordingly. The knockout strains were confirmed by PCR using primers (S3 Table) corresponding to the regions ~100 bp upstream and downstream of the relevant genes.

For complement strain, colony PCR amplified the luxS gene with gene-specific primers (containing BamH1 and HindIII restriction site sequence) (S3 Table). The amplified PCR product was purified by chloroform-isopropanol precipitation and cloned into the empty pQE60 vector using BamH1 and HindIII sites. The double-digested insert and vector were subjected to ligation by T4DNA ligase in ligation buffer (NEB) overnight at 16 °C. The respective ligated vector was transformed into the STM ∆luxS strains to generate the complemented strain. Complement strain was initially confirmed by colony PCR with cloning primers and internal primers of the genes, and finally by restriction digestion of recombinant plasmid isolated from complement strains to study insert release.

Cell culture and maintenance

Murine macrophage cell line RAW 264.7 was maintained in Dulbecco’s Modified Eagle Medium (DMEM, Lonza), supplemented with 10% FBS (Gibco) and 1% penicillin-streptomycin (Sigma-Aldrich) in a humidified incubator with 5% CO₂ at 37 °C. The cells were seeded into the respective cell culture plate for each experiment to conduct intracellular survival assays and intracellular gene expression experiments.

Peritoneal macrophage isolation

The primary macrophage isolation was performed as previously described [36]. Briefly, C57BL/6J mice (5–6 weeks old) were injected with Brewer’s thio-glycolate medium (8% w/v, HIMEDIA). At four days of treatment, 5 ml of cold PBS was injected into the peritoneal cavity, followed by aspiration of the fluid from the cavity and dispensed into a 15 ml centrifuge tube. The centrifugation was done at 300g for 10 min, and the resuspended cell pellet [in Roswell Park Memorial Institute 1640 (RPMI 1640)] was supplemented with 10% FBS and an antibiotic cocktail (penicillin-streptomycin). Cells were counted by hemocytometer and seeded onto 24 well cell culture plates.

Chemical synthesis of acetonide-protected (S)-4,5-dihydroxy-2,3-petanedione (DPD)

Acetonide-protected DPD 6 was synthesized from commercially available L-serine 1 in five steps (for details, see the supplemetary Information). These steps include stereoretentive displacement of the amino group of L-serine through diazotisation and esterification to form methyl (S)-2,3-dihydroxypropanoate 2 [51] followed by protection of 1,2-diol as acetonide (3 [52]. Transformation of methyl ester of 3 to N,N-dimethyl amide 4 [53] paved the way for its conversion to isopropenyl ketone 5 by reacting with isopropenyl Grignard reagent. Ozonolytic cleavage of the alkene in 5 resulted in acetonide-protected (S)-DPD 6 [54].

AI-2 bioassay

The culture of Salmonella Typhimurium WT and mutants was obtained at 0h, 1h, 3h, 6h, 9h, and 12h, and centrifuged at 6000rpm for 10 min. After filtration using a 0.22 μm filter, the supernatant was stored at -20 °C until required. The AI-2 was quantified using the sensor strain Vibrio campbellii ATTC BAA-1117 (strain designation V. harveyi BB-170, which only responds to AI-2). The AI-2 bioassay was performed as previously described [1,55] with minor modifications. A single colony of V. harveyi BB-170 was incubated in 5 ml of Autoinducer Bioassay medium (ATCC Medium: 2746 AB Medium) at 30°C and 200 rpm. A 1:1000 dilution was obtained by transferring a 10 μl sensor strain culture to 10 mL of fresh AB medium. Then, 20 μl of Salmonella cell-free culture supernatant was added to 180 μl diluted culture of the sensor strain. The mixture was added to a white, flat-bottomed 96-well microtiter plate (Corning, USA), followed by shaking in a rotary shaker at 200 rpm, at 30 °C. Luminescence was measured at 3 h in a fluorescent microtiter plate reader. Autoinducer signal calculated as – (luminescence produced by Vibrio in the presence of Salmonella Typhimurium free supernatant - luminescence in control medium)/luminescence produced by Vibrio with sterile medium. To validate the activity of exogenous DPD molecule and BF-8 inhibitor, an autoinducer bioassay was performed in the presence of DPD and BF-8 inhibitor (Cat# 53796, Sigma). For the intracellular AI-2 assay, RAW 264.7 macrophages were infected, and cell lysates were collected at different time points.

Gentamicin protection assay

RAW 264.7 macrophages and peritoneal macrophages isolated from C57BL/6J mice were seeded in tissue culture plates for infection. Cells were infected with STM WT, STM ∆luxS, STM ∆lsrB, STM ∆lsrK, STM ∆lsrR, and STM ∆luxS:luxS (stationary phase culture growing overnight in LB broth). The Multiplicity of Infection (MOI) of 10 was used for the intracellular survival assay. MOI of 20 was used for the Salmonella Typhimurium gene expression study (qRT-PCR). Following the infection of STM strains into the cell line, the plate was spun for 5 min at 600 rpm using a Rota-Superspin R–NV swing bucket centrifuge. The plate was then incubated for 25 min at 37 °C with 5% CO₂ in a humidified incubator. The infected cells were washed twice with 1X PBS and treated with a 100 μg/mL concentration of gentamycin. After 1 h, the medium were changed with a reduced concentration of gentamycin (25 μg/mL) and incubated until the desired time point. The time points of 2h, 6h, and 16h were taken for the qRT-PCR samples and 2h and 16h for the intracellular survival assay.

Also, for studying the effect of exogenous DPD in the recovery of the phenotype of mutant luxS, were grown in vitro with supplementation of 25 μM, 50 μM,100 μM, and 200 μM DPD molecules and 10%, 20%, 30%, and 50% spent medium of STM WT collected at 3h in LB growing. For inhibition, the inhibitor (Z)-4- bromo-5-(bromomethylene)-3-methylfuran-2(5H), known as BF-8 treatment, was added to STM WT in LB medium. DPD or spent medium supplemented and BF-8 treated culture was used for gentamycin protection assay.

Intracellular survival assay and phagocytosis assay

The cells were lysed with 0.1% triton-X 100 in PBS at specific time points post-infection. For the intracellular survival assay, 2h and 16h post-infection samples were collected and plated, and the corresponding CFU at 2h and 16h was determined. Fold proliferation was determined using the formula [CFU at 16 h]/[CFU at 2 h]. For phagocytosis, the bacterial number was determined in the inoculum and 2h time point post-infection. Percent phagocytosis was by using the formula- (CFU at 2 h)/(CFU of pre-inoculum)]×100.

Confocal microscopy

To study the Salmonella proliferation and their intracellular vacuolar status, all STM strains (STM WT, STM ∆luxS, STM ∆lsrB, STM ∆lsrK, and STM ∆lsrR) were transformed with pFPV 25.1-mCherry-Amp^R plasmid, and used to infect RAW264.7 cells. Briefly, 1 × 10⁵ RAW 264.7 cells were seeded on the coverslip and infected with different strains of bacteria at MOI 25. After appropriate post-infection durations, cells were washed twice with PBS and fixed with 3.5% paraformaldehyde (PFA), and incubated with a specific antibody (LAMP1 monoclonal antibody-clone 1D4B; sc-19992, Santa Cruz Biotechnology, USA) in a blocking buffer containing 2% BSA and 0.01% saponin for 2 hours at RT or overnight at 4°C. The cells were washed twice with PBS and incubated with the appropriate secondary antibody conjugated to a fluorochrome (Alexa Fluor 488 or DyLight 488, dilution 1:200) for 1 hour at RT. The coverslips were then mounted onto a clean glass slide and imaged under a confocal microscope (Zeiss LSM-710) using a 63X objective.

In vitro Cell culture competition assay

RAW 264.7 macrophages were infected at a 1:1 ratio with STM WT and STM ∆luxS. At 16 h, post-infection cells were lysed, and the sample was collected as described above. The competition was also done with the coculture of STM WT and STM ∆luxS. This involved the co-cultivation of STM WT and STM ∆luxS together, followed by infection in RAW 264.7 macrophages. The competitive index was calculated for a WT strain and ∆luxS by dividing the ratio between CFU (STM ∆luxS) and CFU (STM WT) by the ratio of both strains in the inoculum.

RNA isolation and RT-qPCR

For gene expression study in Luria Bertani (LB) medium and F-medium (acidic media that mimics the cell vacuole environment) [49,56]. An overnight-grown culture in LB medium was subjected to subculture (at 1:100) in LB or F-medium. At 3h, 6h,9h, and 12h, bacterial cell pellets were resuspended in TRIzol (from TaKaRa, RNA isoPlus- 9109) and kept at -80˚C. Total RNA was isolated by chloroform extraction followed by isopropanol precipitation. To evaluate the quality, the amount of RNA was quantified in nanodrop (ThermoFischer) and examined on a 2% agarose gel.

To produce cDNA, 3 μg of RNA sample was treated with DNase I (TaKaRa) at 37˚C for 1 h followed by heating for 10 min at 65˚C. PCR using 16S primers showed no amplification in DNase-treated samples, confirming the absence of genomic DNA contamination and validating RNA-specific quantification. DNA-free RNA samples were used to make cDNA using the PrimeScript RT reagent Kit provided by TaKaRa (Cat# RR037A).

For gene expression in Salmonella Typhimurium upon infection in RAW 264.7 macrophages, infection was performed at MOI of 20. At 2h, 6h, and 16h post-infection, the cells were lysed using TRIzol and kept at -80˚C. Total RNA isolation and cDNA synthesis were carried out using the manufacturer’s protocol. The list of expression primers is provided in S3 Table. Relative gene expression was determined using the ΔΔCt method, with the 3 h time point (LB culture) and 2 h post-infection (macrophage assays) used as the respective reference conditions for normalization. For comparative analyses, expression levels were calculated relative to STM WT, which served as the reference strain.

Acid survival assay

Overnight cultures of STM WT, STM ∆luxS, STM ∆lsrB, STM ∆lsrK, and STM ∆lsrR were adjusted to OD₆₀₀ - 0.3. For survival in 1XPBS, pH was adjusted to 3,4,5,6,7, or 8 using concentrated HCl. Strains of Salmonella Typhimurium (10 ⁷ bacteria/ml) were exposed to different pH as mentioned in 1X PBS and incubated for 2h. At 2h post-treatment, CFU were enumerated by plating on LB agar medium. The same protocol was followed for survival in LB medium with similar pH ranges. For the strain of STM harboring plasmids pQE60 complemented with phoP/phoQ, a survival assay was done in pH of 3, 4, and 5 in 1X PBS.

Acid tolerance assay

A standard acid tolerance assay was performed as previously described with slight modification [26]. Briefly, acid tolerance assay (ATR) was conducted with strains grown overnight at 37°C in LB broth containing the appropriate antibiotic. The culture corresponding to 0.3 OD₆₀₀ of the overnight culture (unadapted) was centrifuged and the pellet was resuspended into 2 ml of LB broth (pH 5) and incubated at 37°C with shaking for 2 h (known as adapted culture). The acid challenge of unadapted and adapted cultures involved readjusting the pH to 3. CFU was determined for adapted and unadapted cultures after exposure to pH 3 (2 h of pH 3 treatment). ATR assay was also done with the strain of Salmonella containing plasmid pQE60 complemented with phoP/phoQ gene.

Intracellular pH measurement by pHuji plasmid

The plasmid pBAD:pHuji (Plasmid #61555) was used to transform STM WT, STM ∆luxS, STM ∆lsrB, STM ∆lsrK, and STM ∆lsrR. pHuji is a pH-sensitive red fluorescent protein demonstrating a more than 20-fold fluorescent intensity change from pH 5.5 to 7.5 [57]. Overnight-grown bacterial strains were subjected to a pH range of 3.0 to 8.0 in phosphate buffer (PBS) in 40 μM of sodium benzoate to ensure equilibration to the desired pH. The resulting fluorescence intensity ratio, obtained from flow cytometric analysis, was plotted as a function of pH and fitted to get a standard curve. The standard curve was used to interpolate ratios measured within the intracellular pH of bacteria in LB medium, F-medium, and RAW 264.7 macrophages.

Mass spectrometry for the determination of lysine and cadaverine

Overnight grown STM WT, STM ∆luxS, STM ∆lsrK, STM ∆lsrB, and STM ∆lsrR were sub-cultured in acidic pH 5 LB medium and incubated at 37°C until mid-log phase. Thereafter, bacteria were spun down and the spent medium were collected. Spent medium were kept with an equal volume of acetone at -20 °C for overnight. Next, samples were collected after centrifugation at 13k rpm for 10 min. Samples were analysed in Orbitrap fusion connected to UHPLC Vanquish (Thermo Scientific). Water and acetonitrile with 0.1% formic acid were used as a mobile phase. The flow rate was kept at 0.3 ml/min linear gradient was started with 5% to 95% using Hypersil Gold C18 column (2.1 × 100 mm & particle size 1.9μ).

Plasmid construction, and site-directed mutagenesis

The vector pET28a(+) was used to construct plasmids plsrR for the overexpression of LsrR as a C-terminal 6xHis-tagged fusion protein. The DNA fragment containing the lsrR coding region was amplified by polymerase chain reaction (PCR) from the chromosome of Salmonella Typhimurium 14028s using primers (F-lsrR and R-lsrR containing restriction site NcoI and XhoI, respectively). The PCR products were digested with NcoI and XhoI, and ligated with NcoI and XhoI digested pET28a(+). The inserts were checked by DNA sequencing.

For Y25A and R43A site-directed mutagenesis in LsrR protein, PCR for SDM was set up with template DNA from pETa(+)- lsrR and amplified with Q5 polymerase (NEB). The PCR products were digested with DpnI and transformed into E. coli DH5-alpha. Y25A and R43A double mutants were generated using pET28a(+) lsrR R43A plasmid as a template for SDM using primers specific to the second set of mutations. All plasmids were confirmed by DNA sequencing.

Overexpression and purification of LsrR

LsrR protein was purified by using the protocol as described previously [58]. pET28a(+)-lsrR and its mutants containing C-terminal 6xHis tag were transformed into E. coli Rosetta (DE3). The transformed colonies were inoculated into 5 ml LB containing kanamycin (Kan) and chloramphenicol (Cm) and grown overnight. Inoculum (1%) was added to 1.2 L LB containing Kan and Cm, grown at 37 °C to an OD₆₀₀ of 0.4 under shaking, and thereafter supplemented with 0.1 mM IPTG, and grown at 16°C overnight. Cells were harvested by centrifugation at 4°C, resuspended in 10 ml buffer A [20 mM Tris–HCl pH 8, M NaCl, 10% glycerol (v/v), 1.5 mM β-mercaptoethanol, 1 mM PMSF] lysed by sonication, and centrifuged at 4°C in a pre-cooled centrifuge at 10000 rpm for 10 min. The supernatant was loaded onto a 1 ml Ni-NTA column equilibrated with buffer A, washed with 20 ml buffer A, and eluted with a gradient of imidazole (20–500 mM) in the same buffer. The fractions were analysed on 12% SDS-PAGE. Fractions enriched for LsrR were pooled, loaded onto a Superdex75 gel filtration column, and eluted in buffer B [20 mM Tris–HCl pH 8, 1M NaCl, 10% glycerol (v/v) and 1.5 mM β-mercaptoethanol]. The purity of LsrR was checked on 12% SDS-PAGE (S7H Fig). Fractions with apparent homogeneity were pooled, concentrated using a 10 kDa cut-off Centricon (Millipore), and estimated by Bradford’s method using bovine serum albumin (BSA) as standard. The proteins were dialyzed against buffer A containing 50% glycerol (v/v) and stored at –20˚ C.

Gel shift assay

HPLC-purified 90 bp single-strand DNA oligos of the lsrR and phoP operon promoters (Sigma) were used for a gel shift assay. Binding reactions involved incubating promoter DNA with varying concentrations of LsrR or its mutated form. The incubation buffer contained 50 mM Tris-Cl (pH 7.5), 150 mM NaCl, 3 mM magnesium acetate, 0.1 mM EDTA, and 0.1 mM DTT. After 15 min at room temperature, the mixture was combined with gel loading buffer (60% 0.25 × TBE, 40% glycerol, 0.2% bromophenol blue) and run on a 6% native polyacrylamide gel. The staining was done for DNA with SYBR Gold Nucleic Acid Gel Stain (Invitrogen) and analyzed as per the manufacturer’s instructions.

In vivo animal experiment

Oral gavaging of 10⁷ CFU of STM WT, STM ΔluxS, STM ΔlsrB, STM ΔlsrK, STM ΔlsrR, and STM ΔluxS:luxS was used to infect 5- to 6-week-old C57BL/6J mice. Five days post-infection, the intestine (Peyer’s patches), MLN (mesenteric lymph node), spleen, liver, and blood were aseptically extracted (in a Biosafety level 2 cabinet) to examine the colonization in the organs. The CFU were enumerated on differential and selective Salmonella-Shigella (SS) agar.

Mice survival assay

Male C57BL/6, aged 5–6 weeks, were obtained from the Central Animal Facility, IISc. The mice were given 10⁸ CFU of each strain (overnight grown cultures) orally to compare the survival and weight alterations of mice post-infection. Following infection, the mice were observed every day to determine their survival and weight, and the results were expressed as a percentage of survival (Kaplan-meier curve) and weight reduction.

All experiments comply with the rules set forth by the Indian Institute of Science, Bangalore’s IAEC. The approved protocol number is CAF/Ethics/852/2021. The Institutional Animal Ethics Committee approved every animal experiment, and the National Animal Care Guidelines were scrupulously followed.

Statistical analysis

As mentioned in the figure legends, each experiment has been independently repeated two to five times. GraphPad Prism 8.4.3 was utilized for all statistical analyses. Normality of the data was assessed using the Shapiro–Wilk test. In addition, Q–Q plots were visually inspected to evaluate deviations from normality, particularly given the small sample size. Parametric tests were applied only when data did not significantly deviate from normality (p > 0.05). As the figure legends state, the statistical analyses included an unpaired, two-tailed Student’s t-test, One-way ANOVA with Dunnett’s post hoc test, and Two-way ANOVA with Tukey’s post hoc test. A non-parametric one-way ANOVA (Kruskal-Wallis) test with Dunn’s post-hoc test was performed for the animal experiment. P-values less than 0.05 were regarded as significant. The analysis is presented as mean ±SD with information on group sizes and p values mentioned in the respective figure legends.