Lactate promotes the biofilm-to-invasive-planktonic transition in Salmonella enterica serovar Typhimurium via the de novo purine pathway

ABSTRACT

Salmonella enterica serovar Typhimurium (S. Typhimurium) infection triggers an inflammatory response that changes the concentration of metabolites in the gut impacting the luminal environment. Some of these environmental adjustments are conducive to S. Typhimurium growth, such as the increased concentrations of nitrate and tetrathionate or the reduced levels of Clostridia-produced butyrate. We recently demonstrated that S. Typhimurium can form biofilms within the host environment and respond to nitrate as a signaling molecule, enabling it to transition between sessile and planktonic states. To investigate whether S. Typhimurium utilizes additional metabolites to regulate its behavior, our study delved into the impact of inflammatory metabolites on biofilm formation. The results revealed that lactate, the most prevalent metabolite in the inflammatory environment, impedes biofilm development by reducing intracellular c-di-GMP levels, suppressing the expression of curli and cellulose, and increasing the expression of flagellar genes. A transcriptomic analysis determined that the expression of the de novo purine pathway increases during high lactate conditions, and a transposon mutagenesis genetic screen identified that PurA and PurG, in particular, play a significant role in the inhibition of curli expression and biofilm formation. Lactate also increases the transcription of the type III secretion system genes involved in tissue invasion. Finally, we show that the pyruvate-modulated two-component system BtsSR is activated in the presence of high lactate, which suggests that lactate-derived pyruvate activates BtsSR system after being exported from the cytosol. All these findings propose that lactate is an important inflammatory metabolite used by S. Typhimurium to transition from a biofilm to a motile state and fine-tune its virulence.

IMPORTANCE

When colonizing the gut, Salmonella enterica serovar Typhimurium (S. Typhimurium) adopts a dynamic lifestyle that alternates between a virulent planktonic state and a multicellular biofilm state. The coexistence of biofilm formers and planktonic S. Typhimurium in the gut suggests the presence of regulatory mechanisms that control planktonic-to-sessile transition. The signals triggering the transition of S. Typhimurium between these two lifestyles are not fully explored. In this work, we demonstrated that in the presence of lactate, the most dominant host-derived metabolite in the inflamed gut, there is a reduction of c-di-GMP in S. Typhimurium, which subsequently inhibits biofilm formation and induces the expression of its invasion machinery, motility genes, and de novo purine metabolic pathway genes. Furthermore, high levels of lactate activate the BtsSR two-component system. Collectively, this work presents new insights toward the comprehension of host metabolism and gut microenvironment roles in the regulation of S. Typhimurium biology during infection.

INTRODUCTION

Salmonella enterica is a common human pathogen that causes diseases of public health concern including inflammatory diarrhea (gastroenteritis) and enteric fever (1). Salmonella enterica serovar Typhimurium (S. Typhimurium), a non-typhoidal Salmonella serovar, causes 94 million cases of gastroenteritis and 155,000 deaths globally each year (2). S. Typhimurium induces gastroenteritis characterized by fever, nausea, and abdominal cramping in immunocompetent patients.

Research suggests that S. Typhimurium virulence factors actively contribute to inflammation, a mechanism that supports Salmonella replication and enables its access to nutrients crucial for colonization, which might otherwise be inaccessible in a healthy gut (3, 4). For instance, it has been reported that the inflammatory response to Salmonella increases nitrate and tetrathionate (5–7), and commensal anaerobic bacteria depletion causes a reduction of butyrate and propionate along with an increase in host-derived lactate (8). Also, during ongoing inflammation, host cells change their catabolism from a β-oxidation to aerobic glycolysis, causing an increase of tricarboxylic acids by-products in the lumen (9). These metabolic changes provide an advantageous niche for replication since S. Typhimurium can use various inflammatory metabolites such as nitrate and tetrathionate for respiration (5, 8), formate as an electron donor (10), or succinate as a carbon source (11).

Like most Enterobacteria, S. Typhimurium exhibits a dynamic lifestyle that alternates between a virulent free-living planktonic state and a non-motile multicellular biofilm state (12). This versatility aids in evading the host immune response and in enduring environmental challenges like disinfectants, antibiotics, or protist predations. With the discoveries of Salmonella-forming biofilms in the intestinal lumen and on gallbladder gallstones (13, 14), it is believed that Salmonella strategically employs both lifestyles to establish a successful infection (15–17). The extracellular matrix of Salmonella biofilm consists of an amyloid-like mesh of curli fibers (mainly composed of the CsgA protein), cellulose, extracellular DNA, O-antigen capsule, and biofilm-associated protein BapA (18). The production of these components is controlled by many factors including the intracellular level of the secondary messenger bis-(3’−5’)-cyclic dimeric guanosine monophosphate (c-di-GMP), which is synthesized and degraded by diguanylate cyclases and phosphodiesterases, respectively (19). An increase in intracellular levels of c-di-GMP leads to the expression of master regulator CsgD, which induces the expression of extracellular matrix components, including proteins CsgA and CsgB (which compose curli fibrils), and cellulose (19–21).

The environmental cues that S. Typhimurium utilizes to control the biofilm-planktonic transition inside the host are largely unknown. We recently discovered that nitrate generated during inflammation increases c-di-GMP accumulation, flagellin-encoding fliC transcription, and reduces csgD and csgB transcription, resulting in a compromised biofilm integrity (6). Lactate is the most abundant metabolite in the inflammatory niche, and it is generated by the host due to a metabolic reprogramming from oxidative phosphorylation to aerobic glycolysis in enterocytes and macrophages (8, 22). In this work, we sought to determine whether S. Typhimurium utilizes lactate to modulate the sessile-to-planktonic transition.

RESULTS

Metabolites found in the inflamed gut reduce S. Typhimurium biofilm development

In order to investigate the role of inflammatory metabolites in the biology of S. Typhimurium, we evaluated the ability of this bacterium to form surface-attached pellicle biofilm mass at different concentrations of gut inflammatory metabolites using a Crystal Violet assay. Sodium nitrate, sodium DL-lactate, sodium succinate, or sodium citrate were tested at increasing concentrations ranging from 12 to 100 mM. Although all of these inflammatory metabolites decreased biofilm formation (Fig. 1A), lactate, succinate, and citrate were more effective in reducing biofilm formation than nitrate, which affected the biofilm mass only at 50 mM and 100 mM concentrations.

Fig 1.

Inflammatory metabolites inhibit S. Typhimurium biofilm formation. (A) S. Typhimurium’s ability to form biofilm pellicles in the presence of increasing concentrations of nitrate, lactate, succinate, or citrate was evaluated using Crystal Violet assays. Cultures were incubated on YESCA medium, 72 h at 28°C on 96-well plates in triplicate. Errors bars indicate standard deviation of the mean of independent determinations. One-way ANOVA was used to evaluate significance vs untreated column. n.s.: not significant; ****P < 0.0001. (B) YESCA plates supplemented with Congo Red and Coomassie blue dyes were used to analyze S. Typhimurium colony morphotypes at indicated concentrations of nitrate or lactate. Plates were incubated at 28°C for 72 h. (C) YESCA plates supplemented with Calcofluor were used to evaluate S. Typhimurium’s ability to synthesize cellulose. After incubation at 28°C for 72 h, plates were exposed to UV light and imaged. As a negative control, a STm mutant in cellulose biosynthesis (ΔbcsE) was also imaged.

Lactate reduces curli and cellulose production via reduced c-di-GMP levels in S. Typhimurium

Lactate emerged as the most significantly elevated compound in metabolic profiling studies conducted during S. Typhimurium infection (8), and it has been reported that lactate plays a significant role in S. Typhimurium expansion within the infected mouse gut (23). Here, we sought to explore whether this metabolite is utilized by S. Typhimurium as a signal to transition between the biofilm and planktonic lifestyles. However, first, to ensure that the decrease in pellicle biofilm observed in Fig. 1 was due to lactate and not from sodium toxicity, we tested potassium L-lactate and saw a similar reduction in biofilm formation (Fig. S1).

In S. Typhimurium, biofilm development can be assessed by monitoring the ability of curli and cellulose to bind Congo Red (CR) and Coomassie blue (CB) dyes (24), resulting in the typical macroscopic red, dry, and rough (rdar) morphotype in solid LB no salt (LBNS) or in yeast extract and casamino acids (YESCA) media. It is known that the lack of curli or cellulose leads to pink or purple/brown colonies, respectively, whereas the lack of both leads to white colonies (21). In order to investigate the role of lactate in S. Typhimurium curli expression, colony morphotypes were monitored at different concentrations of lactate using YESCA agar supplemented with CR and CB. We observed that lactate altered colony biofilm morphology, where the typical wrinkled and purple morphology turned to smooth and white as lactate concentrations increased, suggesting a decrease in curli and cellulose production (Fig. 1B). Loss of biofilm production in the presence of lactate was also assessed by confocal microscopy, where a reduction of extracellular matrix thickness in glass coverslips was observed at 50 mM of lactate (Fig. S2). Cellulose was further evaluated in YESCA plates supplemented with Calcofluor, a fluorescent dye that binds cellulose (20). Our results indicated that lactate inhibited cellulose production as the concentration of lactate increased, reflected by decreasing fluorescence intensity under UV light (Fig. 1C). By analyzing CR and CB plates and calcofluor plates phenotypes together (Fig. 1B and C), we concluded that at 12.5 mM, lactate reduces curli expression while cellulose production remained unchanged. At 50 mM, lactate inhibited both curli and cellulose synthesis as evidenced by the appearance of the saw (smooth and white) morphotype in CR and CB plates and in the lack of fluorescence signals in calcofluor plates. Since similar effects of nitrate were observed on Salmonella biofilms previously (6), we compared the effect of lactate to nitrate in colony biofilms. Lactate started to hinder curli and cellulose at concentrations of 12.5 mM and 25 mM, respectively, whereas 50 mM nitrate did not affect either curli or cellulose. Therefore, we concluded that lactate had a more profound effect than nitrate on biofilm production (Fig. 1B and C).

c-di-GMP is a secondary messenger whose intracellular concentration inversely regulates biofilm formation and motility, playing a critical role in determining the sessile-to-motile lifestyle fate of S. Typhimurium (19, 25). In light of our results, we reasoned that lactate might be altering the intracellular levels of c-di-GMP, leading to changes in biofilm formation and flagellar synthesis in a similar fashion as nitrate (6). Using a reporter plasmid, pFY4950, in which the expression of red fluorescent protein, TurboRFP, is controlled by a c-di-GMP-dependent riboswitch (i.e., cells fluoresce red upon c-di-GMP presence) (6, 26), we monitored the intracellular c-di-GMP accumulation after 4 h of incubation in biofilm-inducing conditions with increasing concentrations of lactate or nitrate (Fig. 2A). We found that c-di-GMP levels started to decrease at 12 mM lactate. In contrast, a significant reduction in c-di-GMP levels was observed only at concentrations of 24 mM or higher (Fig. 2B).

Fig 2.

Lactate reduces c-di-GMP accumulation along with csgA and csgD transcription and stimulates fliC transcription. (A) Confocal microscopy images (x100 oil magnification) of S. Typhimurium cells containing pFY_4950 plasmid under Amcyan (Ex^470nm/Em^490nm) and TurboRFP (Ex^550nm/Em^580nm) filters were taken, and (B) the relative amount of intracellular c-di-GMP at different lactate concentrations were calculated by counting red cells and divided by total green cells. (C) Quantitative PCR evaluating relative gene expression after incubating S. Typhimurium with increasing concentrations of nitrate or lactate for 18 h in YESCA medium at 28°C, relative to 4 h untreated control. (D) csgBAC and csgD promoter activities at different concentrations of lactate or nitrate were assessed by using GFP reporter plasmids. Errors bars indicate standard deviation of the mean of independent determinations. Errors bars indicate standard deviation of the mean of independent determinations. (B and D) Two-way or (C) one-way ANOVA was used to evaluate significance vs. untreated columns (Dunnett’s test; n.s.: not significant, *P < 0.05, ***P < 0.001, ****P < 0.0001).

In enterobacteria, curli-biosynthetic genes are divergently expressed from csgBAC and csgDEFG operons. In the case of Salmonella, there is a regulatory sequence of approximately 700 nucleotides located between the two operons, and their transcription occurs as a result of signal integration of several environmental stimuli. The regulation of CsgD expression is of particular interest in biofilm metabolism since this master regulator controls the expression of additional biofilm-inducing enzymes (27). Transcriptional levels of biofilm master regulator csgD, curli main component csgA, and flagellar subunit fliC were evaluated by quantitative PCR in the presence of nitrate or lactate (Fig. 2C). Although both lactate and nitrate significantly reduced csgA and csgD transcription, fliC transcription was increased only in the presence of lactate. Consistent with the previous results, lactate had a more significant inhibitory effect on the transcription of csgA and csgD than nitrate.

To validate these findings, reporter plasmids containing cloned csgD and csgBAC promoter regions upstream of the green fluorescent protein (GFP) sequence (28) were utilized. As expected, both promoters were less active when lactate or nitrate concentrations increased, which supports the finding that these metabolites inhibit biofilm development (Fig. 2D). Notably, lactate had a greater inhibitory effect on promoter activities than nitrate in agreement with the previous data (Fig. 2C), and although nitrate inhibited csgD and csgBAC transcription at 100 mM, lactate inhibition began at 25 mM (Fig. 2D).

In summary, these findings collectively demonstrate that lactate plays an active role in inducing the transition from the sessile to the motile state in S. Typhimurium. This is achieved by reducing c-di-GMP accumulation, inhibiting csgBAC and csgD transcription, and promoting fliC transcription. Notably, our results indicate that lactate exerts a more significant effect on biofilm inhibition compared with nitrate.

De novo purine synthesis is involved in lactate-modulated biofilm inhibition in S. Typhimurium

In order to identify genes that participate in lactate-associated biofilm regulation, we performed a transposon mutagenesis using T-POP transposon (29), which contains divergent tetracycline-induced promoters to allow the detection of both positive and negative regulatory elements with or without tetracycline addition, respectively. We constructed a reporter S. Typhimurium P_csgBAC-lacZ strain (FJA115) and generated a transposon library in this strain. We screened for the blue-white colonies and used LBNS media as screening media since the sensitivity was more reliable than in YESCA plates (data not shown). Plates were supplemented with tetracycline to select for negative and positive regulatory insertions. Of the 30,000 mutants screened, 24 strains were blue in high-lactate and biofilm-inducing conditions. From those 24 candidates, nine were confirmed as suppressors of the inhibition of csgA transcription in high lactate conditions observed in FJA115 strain (Fig. 3A through C) and sequenced by inverse PCR (Fig. 3A and B). The addition of tetracycline in the β-galactosidase assays helped differentiate phenotypes originating from gene interruptions (cysB^-, rpoN^-, lldR^-, purA^-, recB^-, purG^-, and hns^-) to those resulted from positive regulatory elements (P_lacZ#1 and P_lacZ#2).

Fig 3.

Random T-POP mutagenesis identified mutants that suppress lactate-associated inhibition of csgBAC transcription. (A and B) β-galactosidase assay performed on S. Typhimurium P_csgBAC-lacZ reporter strain (FJA115) showed (A) an inhibition of csgBAC promoter activity after 72 h in lactate-containing LBNS at 28°C and (B) nine random T-POP insertions in FJA115 exhibited reductions of such inhibition. Gain- or loss-of-function in T-POP mutants were evaluated with or without tetracycline, respectively. Errors bars indicate standard deviation of the mean of independent determinations. Unless indicated, significance was evaluated versus FJA115 behavior at 30 mM lactate (A, dotted line). Unpaired t-test: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (C) Evaluation of csgBAC promoter activity in T-POP mutants evidenced as the ability to hydrolyze X-Gal and develop blue colonies.

In P_lacZ#1 and P_lacZ#2 strains, T-POP integration occurred within lacZ promoter, giving support that the genetic screen level of coverage was appropriate. csgA expression in lldR^- and hns^- T-POP mutants under high lactate was expected based on previous reports. LldR is a transcriptional regulator part of lldPRD operon. It is involved in L-lactate metabolism and its transcription is tightly controlled, being induced in the presence of L-lactate (23, 30). The increase in csgA expression in this mutant explains the involvement of lactate catabolism in inhibiting biofilm formation and suggests that lactate is being internalized to accomplish this. H-NS is a histone-like DNA-binding global gene silencer present in many Gram-negative bacteria (31). Since it has been described in S. Typhimurium that H-NS is an activator of csgD transcription (27), an increased expression of csgA - which is positively modulated by CsgD- in a hns^- background, is expected. In addition, the increase in colony mucosity observed in hns::T-POP mutant (Fig. 3C) correlates with the previous finding that hns mutants exhibit more exopolysaccharide production (32–34).

Surprisingly, in two of the mutants, the transposon disrupted purA and purG genes, which are involved in purine biosynthesis. The role of purine metabolism on biofilm formation has not been described in S. Typhimurium. Bacteria uptake purines from the environment or synthesize them de novo. In Salmonella, the de novo purine pathway is controlled by the PurR repressor and is comprised of 10 reaction steps that begin in the synthesis of 5-phosphoribosil-1-pyrophosphate (PRPP) from ribose-5-phosphate (R5P) and ends in the conversion of inosine 5-monophosphate (IMP), precursor of the ulterior adenosine monophosphate (AMP), and guanosine monophosphate (GMP) synthesis (Fig. 4A). In turn, AMP and GMP give rise to adenosine triphosphate (ATP) and guanosine triphosphate (GTP), which are further metabolized to signaling molecules like c-di-GMP and cyclic-AMP (35, 36). In human cells, the de novo purine biosynthetic enzymes are held together in a metabolon called the purinosome (37). Recently, it has been described in E. coli that pur enzymes also interact physically, suggesting the existence of a purinosome-like structure in bacteria (38). According to our results, PurA and PurG are involved in the lactate-mediated inhibition of csgA transcription (Fig. 3B and C), suggesting an important role of the de novo purine pathway in Salmonella biofilm metabolism. To further explore the involvement of this pathway in biofilm formation, we grew pur deletion mutants at different concentrations of lactate and concluded that disruption of the de novo purine pathway favors the rdar morphotype in high lactate, in contrast to that occurring in the S. Typhimurium parental strain (Fig. 4B). Exogenous supplementation of IMP further showed the importance of this pathway in the Salmonella rdar morphotype in the presence of lactate, as pur mutants utilize the IMP present in the environment and inhibit the rdar colony type in high lactate, as parental strain does (Fig. 4C). Of note, purA mutant exhibited a distinct colony morphology regardless of the presence of lactate or IMP consistent with a curli⁺/cellulose^- phenotype (red and smooth morphotype, Fig. 4D). We hypothesize that such purA^- phenotype might be related to the need for AMP or cyclic-AMP to synthesize cellulose. To answer this question, we added exogenous AMP to CR and CB plates and found that the typical wrinkled morphology re-appeared in purA^- strain regardless of the presence of lactate, confirming the importance of AMP to synthesize cellulose (Fig. 4E).

Fig 4.

S. Typhimurium (STM) de novo purine pathway is involved in lactate-mediated biofilm inhibition. (A) De novo purine pathway biosynthesis in Salmonella. R5P, Ribose-5-P; PRPP, 5-phosphoribosyl-1-pyrophosphate; IMP, inosine monophosphate; AMP, adenosine monophosphate; GMP, guanosine monophosphate; Gln, glutamine. (B) S. Typhimurium wild-type and pur mutants colony morphotypes in LBNS agar supplemented with Congo Red and Coomassie blue at different concentrations of lactate. (C) Colony morphotypes at different concentrations of lactate and exogenous IMP in LBNS agar supplemented with Congo Red and Coomassie blue. (D) S. Typhimurium wild-type, curli (csgBA^-), cellulose (bcsE^-), and purA^- colony morphotypes in LBNS agar supplemented with Congo Red. Cellulose⁺/curli^- and cellulose^-/curli⁺ genotypes are characterized by pink-rough and white/smooth phenotypes, respectively. (E) S. Typhimurium wild-type and purA^- colony morphotypes at different concentrations of lactate and exogenous AMP in LBNS agar supplemented with Congo Red and Coomassie Blue.

In conclusion, these results show that S. Typhimurium utilizes the de novo purine pathway to modulate its metabolism toward a non-biofilm state when lactate is present in the environment.

A transcriptomic approach identified further lactate-related biofilm pathways

To investigate the influence of lactate on S. Typhimurium global gene expression, RNA-Seq analysis was carried out on the wild-type strain with or without 30 mM lactate after 18 h of biofilm-inducing growth condition. The differential expression analysis identified 40 upregulated and 49 downregulated genes in lactate-treated compared with untreated samples (fold-change ≥2 and a P-value ≥ 0.05; Fig. 5A).

Fig 5.

Global transcriptomic analysis of S. Typhimurium in high lactate conditions. (A) Volcano plot representing gene transcripts found in lactate vs untreated S. Typhimurium RNA-seq. Dotted lines indicate fold-change ≥2 and P-value ≤ 0.05 cutoff criteria (−1≥log₂FC≥−1 and -log₁₀ P-value ≥ 1.3). yjiY(btsT) transcription ratio is shown in the upper right side of the panel. (B) Abbreviated heat map depicting fold-change values of independent determinations. Gene functional classification is indicated on the left. T3SS: Type-3 Secretion System. (C) Quantitative PCR evaluating btsT transcription after incubating S. Typhimurium with or without lactate for 18 h in LBNS at 28C. Errors bars indicate the standard deviation of the mean of independent determinations (unpaired t-test, ***P < 0.001). (D) Schematic representation of TCS BtsSR regulation mechanism that represses btsT and stimulates csgBAC transcription in E. coli. (E) Quantitative PCR evaluating csgA and btsT transcription after incubating S. Typhimurium wild-type or btsR mutant with or without lactate for 18 h in biofilm-inducing conditions. Errors bars indicate the standard deviation of the mean of independent determinations (two-way ANOVA Tukey’s test. n.s.: not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). (F) S. Typhimurium wild-type, btsR, and btsT mutants’ colony morphotypes in LBNS agar supplemented with Congo Red and Coomassie blue at different concentrations of lactate.

Concordant to the colony morphology under CR and CB staining, global transcriptomic analysis showed decreased expression of curli and cellulose genes and increased expression of motility- and invasion-associated genes (Fig. 5B). These changes again highlighted the relevant role of lactate in modulating S. Typhimurium sessile-planktonic lifestyles. Also, the de novo purine genes were overexpressed in high lactate conditions, suggesting, as mentioned before, their involvement in the lactate-dependent inhibition of biofilm.

yjiY was the most highly expressed transcript in the high lactate condition (Fig. 5A), and this was validated by qPCR (Fig. 5C). In S. Typhimurium, YjiY (also known as BtsT) is a membrane-associated peptide transporter whose expression is triggered during the onset of the stationary phase and positively controlled by the two-component system BtsSR upon activation by pyruvate (Fig. 5D) (39). Previous studies observed that BtsT plays a role in biofilm formation and cell adhesion in S. Typhimurium (39, 40) and that BtsR negatively regulates csgD and csgBAC in E. coli (41) (Fig. 5D). Based on these previous results, we hypothesized that the BtsSR-BtsT system might act on Salmonella biofilm formation under high lactate conditions. Thus, we explored the role of BtsSR in modulating csgA and btsT expression at the early stages of biofilm formation by qPCR using deletion mutants. We found that in the absence of lactate, the response regulator BtsR acts as and activator of csgA expression, since csgA transcription was reduced in btsR mutant. The presence of lactate further reduced csgA transcription in btsR^- background, pointing out the presence of BtsR-independent mechanisms inhibiting csgA transcription in high lactate (Fig. 5E). btsR^- and btsT^- colony morphotypes with or without the addition of lactate were also investigated. Their behavior in CR and CB plates was undistinguished from wild-type strain: Strains developed rdar morphotypes in the absence of lactate that disappeared when lactate concentration increased (Fig. 5F). Thus, our experiments conclude that albeit BtsR exerts some degree of negative regulation on csgA transcription in untreated condition, the BtsSR-BtsT system does not play a relevant role in regulating Salmonella biofilm formation. Regarding the increase in BtsT expression observed in high lactate, we propose that this is the result of activation of BtsSR system occurring in such environment.

In conclusion, we presented here that lactate stimulates the transcription of invasion and motility genes while inhibiting curli and cellulose genes in S. Typhimurium. Also, we showed evidence that lactate activates the two-component system BtsSR and that the response regulator BtsR exerts some degree of activation on csgA transcription.

DISCUSSION

During Salmonella-induced inflammation, immune cells go through a metabolic reprogramming from oxidative phosphorylation to aerobic glycolysis that leads to substantial changes in metabolic by-product concentrations (42). In activated macrophages, the increase in succinate triggers Salmonella antimicrobial resistance and virulence factor expression (43). Additionally, during inflammation, phagocytes within the intestinal lumen secrete antimicrobial compounds such as reactive oxygen and nitrogen species. Although these compounds typically eliminate bacteria through direct contact, they can also form metabolites that act as final electron acceptors, thereby promoting Salmonella anaerobic respiration and growth (5, 7). Similar metabolic reprogramming has also been observed in enteric epithelial cells of Salmonella-infected mice causing a 70-fold increase in luminal lactate concentration, which provided Salmonella a metabolic fitness advantage (8). All of these findings indicate that Salmonella has evolved to adapt its biology to survive and colonize within an inflammatory environment.

There has been a long-standing debate about whether Salmonella can integrate host environmental signals and establish a sessile (biofilm) state within the host. Recent evidence from multiple investigations supports that Salmonella is indeed capable of forming biofilm structures in the gallbladder and gastrointestinal tract (13, 14). Furthermore, nitrate can serve as an environmental cue to trigger biofilm dispersal in the gut lumen (6). Given that lactate is the most abundant inflammatory metabolite found in the infected gut (8), we sought to study whether lactate has an active role in modulating Salmonella colony morphologies that relate to biofilm formation.

Using colony biofilms and targeted transcriptional analysis, we showed that a high lactate environment intercedes with S. Typhimurium physiology and biofilm formation. The pathogen adapts by reducing c-di-GMP accumulation, restraining curli and cellulose formation by inhibiting csgA and csgD transcription and increasing fliC transcription (Fig. 1 and 2). We also compared the effect of nitrate, another compound highly produced in inflammation, on Salmonella biofilm production (6) and found that lactate had a more dramatic effect in inhibiting Salmonella biofilm formation. Our transcriptomic study further revealed that lactate stimulates flagella and type three secretion system gene expression, providing additional evidence to support the idea that this metabolite promotes the Salmonella biofilm-to-planktonic transition. Previous studies that demonstrated an inhibitory effect of lactic acid bacteria on Salmonella biofilm formation (44, 45) were not able to uncover the mechanism behind this inhibitory activity. These studies suggested that leveraging the anti-biofilm and antimicrobial activity of lactic acid bacteria could be beneficial for inhibiting Salmonella in food or during infections (46). The fact that our studies revealed that lactate drives the expression of Salmonella virulence genes provides a new understanding of the interactions between lactic acid bacteria and Salmonella. This insight could lead to innovative strategies for controlling Salmonella biofilms and reducing its virulence, potentially improving food safety and therapeutic approaches for Salmonella infections. Further research is needed to explore the precise mechanisms and develop practical applications based on these findings.

Surprisingly, the RNA-seq analysis also revealed that high lactate stimulated Salmonella to overexpress the BtsSR-controlled peptide transporter YjiY (BtsT), which was the most upregulated gene compared with unstimulated control. In a previous work carried out in E. coli, it was shown that the response regulator BtsR exhibits an inverse regulatory function by inhibiting csgBAC and stimulating btsT expression (41). Here, we explored BtsR function using a S. Typhimurium btsR mutant and found that BtsR also activates btsT transcription in high lactate, but, on the contrary, BtsR only activates csgBAC transcription in untreated conditions (Fig. 5E). However, in spite of what was found in a previous report (40), we could not associate the BtsSR-BtsT system with regulation of Salmonella biofilm formation. We propose that the BtsT overexpression observed in this work might be occurring due to a stimulation of histidine kinase BtsS by excreted lactate-derived pyruvate, based on the previous finding that BtsS does not sense lactate per se (47) and because it has been shown that the BtsSR system is activated by sensing extracellular pyruvate excreted from an overflow metabolism (48, 49). This context suggests that lactate catabolism plays a key role in regulating Salmonella biofilm formation. In agreement, we also showed here an increased transcription of lldD and lldR in high lactate (Fig. 5A); genes that belong to the lactate-regulated lldDPR operon involved in lactate utilization (23, 30).

Finally, we used a transposon genetic screen and loss-of-function mutants to demonstrate that S. Typhimurium utilizes the de novo purine pathway to modulate its lifestyle toward a non-biofilm state in high lactate environment (Fig. 3 and 4). In human cells, there is strong evidence showing that pur gene products interact physically through the formation of a supramolecular metabolon-like structure, the purinosome, to synthesize purine and purine derivatives (37). It has been recently proposed that a similar mechanism is present in bacteria as well (38). We envision a scenario in which lactate exerts changes in Salmonella metabolism that ends in an exacerbation of the putative purinosome activity or in the stabilization of its structure. Although it has been shown that disrupting the de novo purine genes attenuates virulence in Brucella abortus, Staphylococcus aureus, and Francisella tularensis (35, 50, 51) and curli synthesis in E. coli (52, 53), there is no previous report about the involvement of this pathway in S. Typhimurium pathogenesis or biofilm formation. Thus, this work provides novel insights into the Salmonella biofilm regulation, suggesting that this bacterium utilizes purine metabolism to induce a transition from sessile to a planktonic state.

In conclusion, the work presented here contributes to a better comprehension of Salmonella behavior within the intestinal tract. We showed solid evidence that this bacterium benefits from the inflammatory environment to transition from biofilm to planktonic state to readily infect other niches.

MATERIALS AND METHODS

Strain and plasmid construction

Strains used in this work are derivatives of Salmonella sp. strain IR715, a fully virulent, spontaneous nalidixic acid-resistant derivative of strain ATCC 14028 and are listed in Table 1. Plasmids and oligonucleotides used are provided in Tables 2 and 3, respectively. For pFA65 construction, csgBAC promoter region was amplified by PCR using primers PcsgB-Fw and PcsgB-Rv and cloned in SmaI-digested pFUSE plasmid. Suitable orientation was checked with primers PcsgB-Fw and lacZ-int-Rv. Phusion Taq (Invitrogen) was used for PCRs, and cycling parameters were 95°C for 5 minutes; 32 cycles of 95°C for 30 seconds, 55°C for 15 seconds, 72°C for 1 minute, last extension of 10 minutes at 72°C.

TABLE 1.

Strains used in this work

Strain	Relevant genotype	Reference or source
Salmonella enterica subsp. enterica serovar Typhimurium strain 14028
Wild-type (STM)	Wild-type isolate ATCC 14028, Nalr	(54)
FJA115	PcsgB::(PcsgB-lacZYA), Cmr	S17-1 pFA65 X STM conjugation. This work
FJA117	cysB::T-POP, Tetr, Cmr	T-POP insertion in FJA115 background. This work
FJA118	rpoN::T-POP, Tetr, Cmr	T-POP insertion in FJA115 background. This work
FJA119	lldR::T-POP, Tetr, Cmr	T-POP insertion in FJA115 background. This work
FJA120	purA::T-POP, Tetr, Cmr	T-POP insertion in FJA115 background. This work
FJA121	recB::T-POP, Tetr, Cmr	T-POP insertion in FJA115 background. This work
FJA122	purG::T-POP, Tetr, Cmr	T-POP insertion in FJA115 background. This work
FJA123	hns::T-POP, Tetr, Cmr	T-POP insertion in FJA115 background. This work
FJA124	PlacZ::T-POP-lacZYA, Tetr, Cmr	T-POP insertion in FJA115 background. This work
FJA125	PlacZ::T-POP-lacZYA, Tetr, Cmr	T-POP insertion in FJA115 background. This work
FJA126	purA::Km, Kmr	BEI Resources
FJA127	purD::Km, Kmr	BEI Resources
FJA128	purF::Km, Kmr	BEI Resources
FJA129	purG::Km, Kmr	BEI Resources
FJA130	purK::Km, Kmr	BEI Resources
FJA131	btsT::Km, Kmr	BEI Resources
FJA132	btsR::Km, Kmr	BEI Resources
CT10	csgBA::Km, Kmr	(6)
CT259	bcsE::Km, Kmr	(6)
Escherichia coli
S17-1 (λpir)	recA, thi, pro, RP4-2-Tc::Mu-Km::Tn7 λpir Tpr Smr rK- mK+	(55)
DH5α (λpir)	φ80dlacZ ΔM15 Δ(lacZYA- argF)U169 recA1 hsdR17 deoR thi-l supE44 gyrA96 relA1/λpir	(56)
TH3923	pJS28 (Carbr P22-9+)/F′114ts Lac+zzf-20::Tn10[tetA::MudP](Tcs) zzf-3823::Tn10dTc[del-25](T-POP)/leuA414 hsdSB Fels2−	Hughes

TABLE 2.

Plasmids used in this work

Plasmid	Description	Reference
pFY4950	pMMB67EH containing the Bc3 c-di-GMP biosensor. Gmr	(26) Yildiz UCSC
pFUSE	lacZYA oriT traJ traK R6Kγori, CmR	(57)
pDW6	pBR322-derived with promotorless gfp	(58)
pCT125	pDW6::PcsgBAC-gfp	(28)
pCT126	pDW6::PcsgD-gfp	(28)
pNK972	pBR333-derived with IS10 transposase	(59)
pFA65	pFUSE::PcsgBAC-lacZ	This work

TABLE 3.

Primers used in this work

Primer	Sequence	Reference
T-POP-int-Fw	CTAAGTCATCGCGATGGAGC	(60)
T-POP-int-Rv	ATTGGCCCATTACTGTTTGC	(60)
PcsgB-Fw	GAGACGTGGCATTAACCTGG	This work
PcsgB-Rv	GCTGTCACCCTGGACCTGG	This work
lacZ-int-Rv	TGCATCTGCCAGTTTGAGG	This work
16SRNA-F-qRT1-AM	GGAAACGGTGGCTAATAC	This work
16SRNA-R-qRT1-AM	CCTCACCAACAAGCTAATC	This work
csgD-qPCR-F-1-AM	ACGCTACTGAAGACCAGGAAC	This work
csgD-qPCR-R-1-AM	GCATTCGCCACGCAGAATA	This work
csgA-qPCR-F-1-AM	AGCATTCGCAGCAATCGTAG	This work
csgA-qPCR-R-1-AM	AATGCTCAACGTGGAATCCG	This work
RT-FliC-NM-For	GTAACGCTAACGACGGTATC	This work
RT-FliC-NM-Rev	ATTTCAGCCTGGATGGAGTC	This work
RT_yjiY_Fw	CGCCTCTCTTGGCGAGATGATC	This work
RT_yjiY_Rv	CGGCACCGTGGAGCAGACG	This work

Media and growth conditions

Cultures were grown at 37°C in Luria-Bertani (LB, Fisher Scientific) media. For biofilm-inducing conditions, LBNS (10 g/L tryptone, 5 g/L yeast extract) or YESCA (1 g/L yeast extract, 10 g/L casamino acids) were used, and cultures were incubated at 28°C with no agitation. When necessary, sodium DL-lactate (Thermo Scientific) was added to the media. Antibiotics and other reagents used: Nalidixic acid (Nal), 50 mg/L; kanamycin (Km), 100 mg/L; tetracycline (Tc), 20 mg/L; carbenicillin (Carb), 100 mg/L; and 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal), 40 mg/L.

For curli morphotypes development LBNS agar or YESCA agar were autoclaved and Congo Red (50 mg/L final concentration) and Coomassie blue (10 mg/L final concentration) were added. For cellulose morphotype development, filtered calcofluor white (10 µg/mL final concentration) was added to YESCA Agar plates after autoclaving. Overnight LB cultures were spotted in the plates and imaged under UV light after 72 h incubation at 28°C.

Crystal violet staining

Crystal violet (CV) staining to assess biofilm formation was performed as previously described with modifications (61). Briefly, strains grown overnight in LB at 37°C were diluted back 1:100 in LBNS with or without the addition of lactate. Two hundred microliters of cultures were placed in a flat-bottom 96-well polystyrene plate (Fisher Scientific, USA) and incubated at 28°C. After 72 h, cultures were gently discarded after growth determination at absorbance 600 nm (BMG POLARstar Omega plate reader), and wells were washed 3 times with 250 µL PBS. Afterward, wells were treated with 200 µL CV Solution (0.05% Crystal Violet, 5% methanol, and 5% isopropanol in PBS) for 1 h, washed 5 times with 250 µL diH₂O and air-dried completely at room temperature for an extra hour. CV was resuspended in situ with 200 µL 33% acetic acid and measured at absorbance 595 nm (BMG POLARstar Omega plate reader).

Indirect c-di-GMP quantification by confocal microscopy

To measure the levels c-di-GMP in S. Typhimurium, pMMB-Gm-Bc3-5AAV plasmid carrying a c-di-GMP biosensor (pFY4950, kindly provided by Dr. Fitnat Yildiz, UC Santa Cruz) was used (26). After transforming pFY4950 by electroporation, overnight culture strains (LB, 37°C) were diluted back 1:10 in YESCA medium with or without lactate and grown for 4 h at 28°C with no agitation to induce biofilm condition. Cells were concentrated 10 times in PBS and immobilized for imaging by placing 25 µL culture on a 100 µL 1% agarose LE pad prepared as previously described (62). Cells were visualized on a Leica SP5 microscope with a TCS confocal system using a sequential scan at 100× oil magnification. Images were taken from a minimum of 10 fields for each condition. Amcyan- and TurboRFP-positive cells were enumerated using ImageJ. The percentage of c-di-GMP-positive cells was determined as the number of TurboRFP-positive cells divided by the number of Amcyan-positive cells for each individual field. The final mean percentage for each condition was determined by averaging the percentages of each individual field.

Fluorescence assays

csgBAC and csgD promoter activities were evaluated in vivo by measuring GFP expression using pCT125 and pCT126 plasmids. To this aim, S. Typhimurium harboring pCT125, pCT136, or promotorless GFP plasmid (pDW6) were grown in YESCA media on 96-well plates for 72 h at 28°C. Bacteria growth was estimated by measuring turbidity at OD₆₀₀, and GFP fluorescence was measured at OD₄₉₀^Em / OD₅₂₀^Ex (BMG POLARstar Omega plate reader). Arbitrary Units = (Fluorescence / OD₆₀₀)_sample – (Fluorescence / OD₆₀₀)_pDW6.

RNA isolation and quantitative RT-PCR

Total RNA was extracted using the hot phenol method previously described (63) with modifications. 1 mL cultures incubated statically at 28°C were spun down and resuspended in 450 µL AE Buffer (sodium acetate 50 mM, EDTA 10 mM, pH 5.2). Cell lysis was carried out by adding 40 µL SDS 20% and 450 µL acid phenol (pH 4.5) to the samples following an incubation at 65°C for 10 minutes. Samples were incubated 5 minutes on ice, centrifugated for 10 minutes at 8,000 g 4°C, and upper phases were transferred to new tubes. Then, 0.1 vol of sodium acetate pH 5.5 and 1 vol isopropanol were added, and samples were mixed by inversion and incubated at −80°C for 1 h. RNA was precipitated by centrifugation (13,000 g, 30 minutes, 4°C), pellets were washed with 500 µL ethanol 70%, dried completely, and resuspended in water (25 µL). RNA samples were treated with DNAse (Thermo Scientific), repurified with rounds of acid phenol and chloroform extractions, and precipitated again. RNA quality and quantity were assessed using a nanodrop (Thermo Scientific, USA).

cDNA was synthesized using Random Primers and Reverse Transcriptase according to the manufacturer’s protocol (Applied Biosystems). Quantitative real-time PCR was carried out using a standard SYBR green method (Applied Biosystems) using a real-time PCR system (Step One Plus, Applied Biosystem). Cycling parameters: 95°C, 15 seconds; 95°C 3 seconds; 60°C 30 seconds. Melting curve parameters: 95°C, 15 seconds; 60°C, 1 minute; 95°C, 15 seconds. 16S rRNA primers were used as internal control, and relative quantification was analyzed using the ΔΔCt method (64).

Transposon mutagenesis and inverse PCR

For transposon mutagenesis, a lacZ reporter strain (FJA115) was constructed by introducing pFA65 into S. Typhimurium wild-type by conjugation. The ability of this strain to modulate β-galactosidase expression in the presence or absence of lactate was verified in LBNS and YESCA plates containing X-Gal. P22 phage lysate containing a Tn10-derivative transposon T-POP (29) was obtained as previously described (65) using E. coli TH3923 as a donor strain. Then, a FJA115 overnight culture containing plasmid pNK972 was diluted back to 10⁸ CFU/mL in fresh LB, and P22 TH3923 lysate was added in a final concentration of 10⁷ PFU/mL. After 30 minutes of incubation at 37°C, 100 µL of the preparation was plated in LBNS agar containing X-Gal, Tetracycline (to induce the expression of T-POP adjacent genes) and 30 mM lactate. Plates were incubated at 28°C for 72 h.

Inverse PCR was utilized to identify T-POP insertions (66, 67). Briefly, after standard phenol-chloroform DNA extraction was performed on candidate clones, genomic DNA was digested with BssHII, and fragments were circularized using a DNA ligase. Circularized DNA region containing T-POP insertion was identified by carrying out a PCR using primers T-POP-int-Fw and T-POP-int-Rv and by sequencing the PCR product. The PCR cycling parameters were 95°C for 5 minutes; 32 cycles of: 95°C for 30 seconds, 53°C for 15 seconds, 72°C for 2.5 minutes; last extension of 10 minutes at 72°C.

β-galactosidase measurements

β-galactosidase assays were carried out as described previously (68). Briefly, after measuring growth (OD₆₀₀), cells were resuspended in 1 mL Z-Buffer (0.06M Na₂HPO₄, 0.04M NaH₂PO₄, 0.01M KCl, 0.001M MgSO₄, and β-mercaptoethanol 0.35% vol/vol) and permeabilized with SDS and chloroform for 10 minutes. Two hundred microliters of o-nitrophenyl-beta-D-galactopyranosidase (ONPG, 4 mg/mL) were added, and reactions were stopped adding 500 µL Na₂CO₃ 1M. Colored product was quantified at OD₄₂₀. Miller Units were calculated as previously described (68).

Library preparation, RNA sequencing, and data analysis

Isolated RNA sample quality was assessed by High Sensitivity RNA Tapestation (Agilent Technologies Inc., California, USA) and quantified by Qubit 2.0 RNA HS assay (ThermoFisher, Massachusetts, USA). Ribosomal RNA depletion was performed with QIAseq FastSelect 5 s/16 s/23 s kit (Qiagen, Hilden, Germany) per manufacturer’s instructions. All library construction was prepared according to the NEBNext UltraTM II Directional RNA Library Prep Kit for Illumina (New England BioLabs Inc., Massachusetts, USA). Final library quantity was assessed by Qubit 2.0 (ThermoFisher, Massachusetts, USA), and quality was assessed by TapeStation HSD1000 ScreenTape (Agilent Technologies Inc., California, USA). Average final library size was about 400 bp with an insert size of 250 bp. Illumina 8-nt dual indices were used. Equimolar pooling of libraries was performed based on QC values and sequenced on an Illumina NovaSeq X plus (Illumina, California, USA) with a read length configuration of 150 PE for 20 M PE reads per sample (10M in each direction).

Raw reads were subjected to quality control using FastQC. Trimming of low-quality bases was performed using Sickle. Reads were aligned to the Salmonella enterica serovar Typhimurium LT2 genome (NC_003197) using EDGE-pro, which then calculated the RPKM (reads per kilobase of transcript per million reads mapped) values. The raw counts were then normalized using the DESeq2 package in R. Differential expression analysis was performed using the DESeq2 package in R with a false discovery rate (FDR) of 0.05, and a P value of less than 0.05 was used to determine significance. Data visualization was created using ggplot2, RColorBrewer, and EnhancedVolcano.

Raw files and processed data are available online in the Gene Expression Omnibus repository (Accession number: GSE255285).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/iai.00266-24.

Supplemental figures. iai.00266-24-s0001.pdf.

Figures S1 and S2.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.