Infection-associated gene regulation of L-tartrate metabolism in Salmonella enterica serovar Typhimurium

ABSTRACT Enteric pathogens such as Salmonella enterica serovar Typhimurium experience spatial and temporal changes to the metabolic landscape throughout infection. Host reactive oxygen and nitrogen species non-enzymatically convert monosaccharides to alpha hydroxy acids, including L-tartrate. Salmonella utilizes L-tartrate early during infection to support fumarate respiration, while L-tartrate utilization ceases at later time points due to the increased availability of exogenous electron acceptors such as tetrathionate, nitrate, and oxygen. It remains unknown how Salmonella regulates its gene expression to metabolically adapt to changing nutritional environments. Here, we investigated how the transcriptional regulation for L-tartrate metabolism in Salmonella is influenced by infection-relevant cues. L-tartrate induces the transcription of ttdBAU, genes involved in L-tartrate utilization. L-tartrate metabolism is negatively regulated by two previously uncharacterized transcriptional regulators TtdV (STM3357) and TtdW (STM3358), and both TtdV and TtdW are required for the sensing of L-tartrate. The electron acceptors nitrate, tetrathionate, and oxygen repress ttdBAU transcription via the two-component system ArcAB. Furthermore, the regulation of L-tartrate metabolism is required for optimal fitness in a mouse model of Salmonella-induced colitis. TtdV, TtdW, and ArcAB allow for the integration of two cues, i.e., substrate availability and availability of exogenous electron acceptors, to control L-tartrate metabolism. Our findings provide novel insights into how Salmonella prioritizes the utilization of different electron acceptors for respiration as it experiences transitional nutrient availability throughout infection. IMPORTANCE Bacterial pathogens must adapt their gene expression profiles to cope with diverse environments encountered during infection. This coordinated process is carried out by the integration of cues that the pathogen senses to fine-tune gene expression in a spatiotemporal manner. Many studies have elucidated the regulatory mechanisms of how Salmonella sense metabolites in the gut to activate or repress its virulence program; however, our understanding of how Salmonella coordinates its gene expression to maximize the utilization of carbon and energy sources found in transitional nutrient niches is not well understood. In this study, we discovered how Salmonella integrates two infection-relevant cues, substrate availability and exogenous electron acceptors, to control L-tartrate metabolism. From our experiments, we propose a model for how L-tartrate metabolism is regulated in response to different metabolic cues in addition to characterizing two previously unknown transcriptional regulators. This study expands our understanding of how microbes combine metabolic cues to enhance fitness during infection.

E nteric pathogens coordinate gene expression in a spatial manner to ensure the production of virulence factors in the correct anatomical location within the host.Spatial sensing is achieved by a plethora of cues, many of which are either host-or microbiota-derived metabolites.For example, the Enterohemorrhagic Escherichia coli and Citrobacter rodentium sense microbiota-derived galacturonic acid through ExuR, which in turn activates the expression of the type 3 secretion system (T3SS) genes found on the locus of enterocyte effacement (LEE) pathogenicity island (1).Fucose, another metabolite liberated by the microbiota, is sensed by Enterohemorrhagic E. coli to activate the two-component system FusKR, which in turn controls the expression of virulence and metabolic genes (2).Host-derived oxygen and microbiota-derived formate regulate the virulence program of Shigella flexneri, an intracellular pathogen that invades the colonic epithelium (3).The invasion-associated T3SS (T3SS-1) in Salmonella enterica serovar Typhimurium (S.Tm) is regulated by a large number of metabolic cues, including bile salts, formate, butyrate, acetate, and long-chain fatty acids (4)(5)(6)(7)(8)(9).This complex regulatory network ensures the expression of T3SS-1 in the distal small intestine, the primary site of invasion (10).As such, the location-specific sensing of metabolites is critical for the expression of virulence factors and tissue-specific induction of disease.
Local inflammatory responses induced by invasive enteric pathogens, such as S. Tm, suppresses host-associated microbial communities in the gut while facilitating the expansion of a pathogen population in the gut lumen (11), which in turn enhances transmission success by the fecal-oral route (12).The expansion of the S. Tm pop ulation is driven by changes in the metabolic landscape as inflammation develops (13).In the initial stage of infection, Salmonella utilizes simple sugars, amino acids, and molecular hydrogen (14)(15)(16).Upon tissue invasion, Salmonella induces localized inflammatory responses, and transmigrating phagocytes undergo an oxidative burst to release reactive oxygen species and nitric oxide radicals.These phagocyte-derived compounds react to form nitrate or interact with endogenous molecules such as thiosulfate to form tetrathionate (17,18).Furthermore, a change in epithelial cell metabolism results in the diffusion of molecular oxygen into the gut lumen (19).A lumenal population of Salmonella performs respiration using nitrate, tetrathionate, and oxygen to promote an oxidative central metabolism at later stages of infection (20).Using respiration, Salmonella utilizes a variety of poorly fermentable metabolites, thus bypassing nutritional competition with anaerobic members of the microbiota (20)(21)(22)(23)(24).While several metabolic pathways used by S. Tm for gut colonization have been described, our understanding of how S. Tm responds to temporal changes in metabolite availability is limited.
We have recently shown that at early time points, prior to the overt transmigration of neutrophils, the release of reactive oxygen species and nitric oxide radicals leads to a non-enzymatic decarboxylation of monosaccharides (25).This reaction generates alpha hydroxy carboxylic acids, including the tartrate isomers L-and D-tartrate.Salmonella utilizes L-and D-tartrate in a stereospecific manner (25).After uptake of L-tartrate by the TtdU transporter, the dehydratase TtdAB converts L-tartrate to oxaloacetate.Part of the oxaloacetate pool enters the reductive branch of the tricarboxylic acid (TCA) cycle to support fumarate respiration, while a portion of oxaloacetate is converted to pyruvate via the oxaloacetate decarboxylase sodium pump.Curiously, the utilization of L-tartrate only occurs at earlier time points during infection, and apparently, L-tartrate utilization ceases once overt inflammation has developed (25).Experiencing transient nutrient niches requires an adaptation to the changing environments, and these adaptations are influenced by gene expression.In this study, we investigate the gene regulation for Salmonella L-tartrate metabolism as it relates to infection-relevant cues.

Substrate induction of the L-tartrate utilization gene cluster
The ttdBAUVW gene cluster in S. Tm encodes for the two subunits of the L-tartrate dehydratase, TtdA and TtdB, the transporter TtdU, and two GntR-type regulators, STM3357 (renamed TtdV) and STM3358 (renamed TtdW) (Fig. 1A).Since genes involved in catabolism are commonly regulated by substrate availability, we initially investiga ted the gene expression of the ttd gene cluster in response to L-tartrate.To this end, we created a chromosomal transcriptional fusion of the ttdA gene and lacZ (encoding β-galactosidase) to capture physiologically relevant transcriptional activities at the L-tartrate gene cluster.The transcriptional fusion did not interfere with L-tartrate utilization (Fig. S1A).We first quantified β-galactosidase activity in response to increasing concentrations of L-tartrate (Fig. 1B).The half-maximal effective concentration (EC 50 ) for L-tartrate was approximately 0.05 mM.Neither D-nor meso-tartrate elicited β-galactosi dase activity, suggesting that transcriptional induction by L-tartrate is stereospecific.Additionally, a transcriptional fusion to ttdU exhibited a similar dose-response curve to L-tartrate, with an EC 50 of approximately 0.15 mM (Fig. S1B).
Next, we analyzed the transcription of other genes in the ttd gene cluster in the presence or absence of L-tartrate.Media, supplemented with 0.1 mM L-tartrate (approximately the EC 50 ) as indicated, was inoculated with the S. Tm wild-type strain.After 3 h of growth under anaerobic conditions, we extracted RNA and quantified mRNA levels by reverse transcription quantitative real-time PCR (RT-qPCR) (Fig. 1C).mRNA levels of the transporter (ttdU) and the two subunits of the L-tartrate dehydratase (ttdA and ttdB) were induced more than 100-fold in response to L-tartrate.In contrast, ttdV and ttdW mRNA levels were unchanged (Fig. 1C), even at higher, likely physiologically irrelevant (25) concentrations of L-tartrate (Fig. S1C).
We considered that a metabolite derived from L-tartrate degradation, rather than L-tartrate per se, might be a cue for the induction of ttdBAU transcription.To test this idea, we generated a mutant lacking L-tartrate dehydratase activity (ttdB mutant) and determined ttdA and ttdU mRNA levels (Fig. 1D and E).L-tartrate stimulated ttdA and ttdU transcription even in the absence of L-tartrate dehydratase activity.Furthermore, induction by L-tartrate was absent in a ttdU mutant (Fig. 1F), consistent with the idea that L-tartrate must enter the cell to influence gene regulation.Collectively, these experi ments suggest that L-tartrate is an inducer of ttdBAU transcription.

TtdV and TtdW are transcriptional repressors for the L-tartrate utilization genes
TtdV and TtdW are members of the GntR family of bacterial transcriptional regulators.These types of transcriptional regulators have a conserved DNA-binding motif at the N-terminus, while the C-terminus serves as a ligand-binding or oligomerization domain (26).Transcriptional regulators in the GntR family can be either transcriptional activa tors or repressors.Transcriptional repressors display high affinity for DNA binding in the absence of ligand; however, ligand presence causes a conformation change that weakens the affinity for DNA binding, allowing transcription to occur (27).To deter mine the function of TtdV and TtdW, we created clean deletion of the ttdV, ttdW, and ttdVW genes.Cultivation in Jordan's Tartrate Agar, a growth medium designed to detect fermentation of tartrate, indicated robust tartrate utilization by the ttdV, ttdW, and ttdVW mutants (Fig. S2A).Transcription of ttdA was significantly elevated in these mutants compared to the wild-type strain, and these strains were non-responsive to L-tartrate (Fig. 2A), suggesting that TtdV and TtdW are negative regulators of ttdBAU gene regulation.A ttdVW double mutant exhibited a phenotype similar to the single mutant.Genetic complementation, either by introducing the ttdV or ttdW coding sequence and its native promoter into the neutral phoN locus or by expressing recombinant histidine-tagged fusion proteins, restored repression of the L-tartrate utilization genes under non-inducing conditions (Fig. 2A; Fig. S2B).Furthermore, we generated mutants in which we replaced the start codon of ttdV (1GTG>GTT) and ttdW (1GTG>TAG).One would predict that with this mutagenesis strategy, any potential cis-regulatory element concealed in the ttdV or ttdW coding sequence should be largely retained.Akin to the single gene deletions, these mutants exhibited increased ttdA transcription in the absence of L-tartrate (Fig. 2B and C).Collectively, these experiments indicate that TtdV and TtdW are repressors of ttdBAU, and that L-tartrate relieves repression by TtdV and TtdW.

TtdV and TtdW bind to the ttdU-ttdV intergenic region
We next sought to uncover the mechanism TtdV and TtdW use to control the transcrip tion of ttdBAU.We hypothesized that TtdV or TtdW could be binding to the region upstream of the ttdU gene to exert control.We generated four biotinylated DNA probes that span the ttdV-ttdU intergenic region as well as part of the ttdV coding sequence (Fig. 3A) and purified C-terminal histidine-tagged TtdV (TtdV-His 6 ) and TtdW (TtdW-His 6 ) to assess protein-DNA interactions by electrophoretic mobility shift assays (EMSAs).Of the four probes tested, we observed significant retardation when probe 4 was incubated with TtdW (Fig. 3B).We also detected more pronounced retardation of DNA probe 4 with increasing concentrations of TtdW (Fig. 3C).Competing labeled probe 4 with unlabeled probe 4 for TtdW binding resulted in the disappearance of the shift (Fig. 4B).TtdV did not bind to any of the probes (Fig. 4A).However, incubating TtdV, TtdW, and labeled probe 4 resulted in a new, slower migrating band, indicating that TtdV, TtdW, and the DNA have formed a complex (Fig. 4B).Overall, the results from these experiments suggest that TtdW binds to a region of DNA closest to the ttdU gene and TtdV requires TtdW to form a complex with the target DNA.

The ArcAB two-component system activates ttdBAU transcription
The ArcAB two-component system is a key regulatory system for fermentation and anaerobic respiration.ArcB, the sensor kinase, senses the redox state of the quinone pools and phosphorylates the response regulator ArcA to control the expression of genes involved in the energy metabolism of the cell (reviewed in reference 28).In E. coli, ArcA was shown to have no effect on ttdA transcription in the absence of oxygen (29).However, it is plausible that ArcAB could influence gene regulation for L-tartrate metabolism since the two-component system coordinates the directionality of the TCA cycle as the cell switches from fermentation to anaerobic respiration.During infection, the central metabolism of Salmonella switches as inflammation-derived electron acceptors become available during infection (reviewed in references 20, 28).We, therefore, hypothesized that the ArcAB two-component system is involved in regulating the L-tartrate utilization genes.We created strains that lack arcA or arcB and monitored ttdA transcription using a β-galactosidase reporter assay (Fig. 5A).The ArcAB system had no discernable effect on ttdA transcription in the absence of L-tartrate; however, induction of ttdA by L-tartrate was absent in both strains that lacked a functional ArcAB two-component system (Fig. 5A).We extended our analysis to quantify ttdVW transcripts in the arcA mutant and observed no differences in expression of ttdV and ttdW, supporting the idea that ArcA is not controlling the expression of ttdVW (Fig. S3A  and B).Lastly, the expression of recombinant ArcA from a plasmid in the arcA mutant restored responsiveness to L-tartrate (Fig. S3C).
Next, we wanted to investigate if phosphorylated ArcA (ArcA~P) would bind to the intergenic region between ttdV and ttdU.To this end, we purified a recombinant His 6 -ArcA fusion protein.We incubated the fusion protein with carbamoyl phosphate, a low-molecular-weight phospho-donor, before adding it to the EMSA binding reactions (Fig. 5B).Of note, all probes incubated with unphosphorylated ArcA exhibited a DNA shift, consistent with previous studies that have shown that ArcA binds non-specifically to DNA (30,31).We found that ArcA~P caused a specific retardation of probe 4. Furthermore, ArcA~P bound to probes 1-3 in a non-specific manner.The addition of an excess of unlabeled probe 4 reversed the observed retardation of probe 4, indicating that the unlabeled probe 4 was competing with labeled probe 4 for ArcA~P binding (Fig. 5B).These experiments suggest that, akin to TtdV and TtdW, ArcA~P specifically binds to the ttdV-ttdU intergenic region.During aerobic growth conditions, the ratio of ubiquinones (Q) to menaquinones (MK)/demethylmenaquinones (DMK) found in quinone pools is mostly comprised of Q and oxidizes ArcB to silence its autophosphorylation activity, leading to ArcA inacti vation.A shift from aerobic to anaerobic growth, however, triggers a change in the composition of the quinone pool in which MK/DMK make up a greater portion of the Q-MK/DMK ratio in the quinone pool.The predominance of MK/DMK activates ArcB kinase activity, leading to the phosphorylation of ArcA (32,33).We hypothesized that artificially changing the Q-MK/DMK ratio by inactivating Q or MK/DMK biosynthesis would lead to altered ArcB/ArcA regulation of the L-tartrate utilization genes.The inactivation of the ubiE gene results in the absence of Q and MK and an increased accumulation of DMK, and menA mutants cannot produce MK or DMK but can synthe size Q (34).These strains were assessed for β-galactosidase activity under inducing and non-inducing conditions.In comparison to the L-tartrate induction of ttdA in the wild-type strain, the accumulation of DMK (ubiE mutant) resulted in the enhanced transcription of ttdA in the presence of L-tartrate (Fig. 5C), consistent with the idea that a greater MK/DMK ratio will lead to activated ArcB sensing and ArcA activation and, thus, an increase in the transcription of ttd genes.In contrast, in a strain that lacks MK or DMK (menA mutant), the inducing effect L-tartrate had on ttdA transcription was drastically reduced (Fig. 5D), supporting the notion that an absence of these fermentation-associ ated quinones leads to silencing of the ArcAB two-component system and, therefore, decreased transcription for the ttd operon.Our data suggest that the activation of the L-tartrate gene cluster is dependent upon sensing of the redox state of the quinone pool.

Regulation of L-tartrate metabolism is required for optimal fitness in the murine gut
During infection, L-tartrate is generated as a byproduct of reactive oxygen and reactive nitrogen metabolism (25).Since other electron acceptors, such as oxygen, nitrate, and tetrathionate, become available during infection at later time points (17,19,35), we sought to clarify the link between these exogenous electron acceptors and L-tartrate utilization.Consistent with previous observations, anaerobic L-tartrate utilization ceases to provide a fitness advantage in the presence of alternative electron acceptors (Fig. 6A) (25).To determine whether these changes in L-tartrate utilization were driven by gene regulation, we cultured the ttdA transcriptional reporter strain in the presence of L-tartrate and determined the impact of oxygen, nitrate, and tetrathionate on ttdA transcription (Fig. 6B).Consistent with our observation that electron acceptors interfere with the fitness advantage conferred by L-tartrate utilization, ttdA transcription was significantly repressed by all three exogenous electron acceptors.
We next asked if the ability to regulate L-tartrate metabolism contributes to Salmonella fitness during infection.Groups of CBA mice were intragastrically inoculated with equal amounts of the S. Tm wild-type strain and a ttdV mutant, or the S. Tm wild-type strain and a ttdW mutant.After 4 days, colonic and cecal content was collected and plated on selective media (Fig. 6C through F).Fitness was assessed by calculating the ratio of the wild-type strain and the mutant in the contents, corrected by the corre sponding ratio of the two strains in the inoculum (competitive index).Compared to the wild-type strain, the ttdV and ttdW mutants were less fit in the colon and cecum content, indicating that TtdV-and TtdW-mediated gene regulation is required for optimal fitness of S. Tm.

DISCUSSION
S. Tm has fine-tuned metabolism and virulence to cope with the transient niches it experiences during infection.Here, we investigated how L-tartrate, a metabolite generated as a byproduct of inflammatory ROS and RNS metabolism through the oxidative decarboxylation of monosaccharides, is sensed by S. Tm.Our in vitro studies enable us to formulate a specific model of how the L-tartrate utilization gene cluster is regulated (Fig. 7).Both TtdV and TtdW are negative regulators (Fig. 2), while the activated response regulator ArcA mediates positive regulation and is required for optimal transcription (Fig. 5).We propose that both TtdV and TtdW interact and that this complex represses transcription of the ttdBAU operon (Fig. 3 and 4).In the presence of L-tartrate, this repression is relieved (Fig. 1).Akin to TtdV and TtdW, ArcA~P binds the ttdV-ttdU intergenic region (Fig. 5), and it is plausible that these two binding sites overlap.Our data are consistent with a scenario in which TtdV and TtdW prevent ArcA~P from binding to the ttdV-ttdU intergenic region and prevent transcriptional activation by ArcA~P.For substrate-induced activation, TtdV and TtdW are removed, possibly by L-tartrate binding to the complex, before ArcA~P can activate the transcription of ttdBAU.In the presence of respiratory electron acceptors, such as oxygen, unphosphorylated ArcA is unable to bind the ttdV-ttdU intergenic region, and transcription is minimal regardless of L-tartrate availability.One limitation of our study is that we were unable to conclusively demonstrate L-tartrate binding to purified TtdV and TtdW proteins.However, it is likely that L-tartrate, and not a downstream metabolite, is the actual cue to relieve TtdV-and TtdW-mediated repression, since a mutant unable to degrade L-tartrate is fully responsive to L-tartrate-mediated induction.Our study focused on how the L-tartrate utilization genes are regulated and the role TtdV and TtdW have in this regulatory mechanism; however, it is possible that additional layers of regulation, for example, autoregulation of ttdV and ttdW, may exist.Our model (Fig. 7) illustrates how Salmonella integrates two infection-relevant cues, i.e., the availability of exogenous electron acceptors and the availability of the sub strate, L-tartrate.We had previously shown that tartrate isomers, including L-tartrate, are generated as byproducts of inflammatory ROS and RNS metabolism and that the utilization of L-tartrate confers a fitness advantage early during infection (25).During that time, S. Tm utilizes L-tartrate early during infection in a branched TCA cycle, with a portion of L-tartrate being ultimately converted to fumarate to support fumarate respiration.At later time points, electron acceptors released by inflammatory macro phages and neutrophils, such as nitrate and tetrathionate, enable a population of S. Tm to respire and switch from a branched TCA cycle to a complete, oxidative TCA cycle (25).At this point, the fitness advantage conferred by L-tartrate utilization fades, likely due to the repression of L-tartrate utilization genes by increased availability of electron acceptors (25) (Fig. 6).Only in the absence of energetically more favorable electron acceptors, and when the substrate is present at appreciable concentrations, genes involved in L-tartrate utilization are optimally transcribed.These data could explain why L-tartrate is primarily converted to fumarate for fumarate reduction, and L-tartrate conversion to oxaloacetate is not used as an anaplerotic reaction.In the presence of exogenous electron acceptors and more efficient electron transport chains, fumarate reduction becomes superfluous and L-tartrate degradation via the ttdBAU gene cluster ceases due to gene repression.
The regulation of gene expression is critical for Salmonella to establish infection.Many of the cues involved in virulence gene regulation are metabolites as well as host-derived antimicrobials.Metabolites regulate the expression of T3SS machinery to minimize any metabolic cost of expression while enabling tissue-specific invasion and tissue replication (36,37).Our data suggest that TtdV and TtdW, two regulatory proteins required for repression of genes involved in L-tartrate utilization, are required for optimal fitness during gut colonization.Regardless of whether the cell exhibits a branched or complete TCA cycle, the central metabolism fulfills key physiological functions, i.e., redox balance, interconversion of metabolic building blocks, and energy generation.As such, the TCA cycle, as well as major pathways connecting to it, are subject to complex control mechanisms (38).TtdAB converts L-tartrate directly to the TCA cycle intermediate oxaloacetate, thus likely requiring a tight regulation of TtdAB activity.In the absence of TtdV-and TtdW-mediated repression, uncontrolled uptake of L-tartrate and immediate conversion to oxaloacetate could perturb metabolic flux in the TCA cycle, thus leading to slower growth and decreased gut colonization (Fig. 6).
L-tartrate metabolism has been studied in detail in commensal E. coli (29,39,40).The L-tartrate utilization gene cluster in E. coli encodes a single LysR-type transcriptional regulator TtdR, whereas the presence of a second transcriptional regulator in S. Tm deviates from the genetic structure of the operon found in E. coli.The primary amino acid sequences of TtdR (E.coli), TtdV (S.Tm), and TtdW (S.Tm) exhibit no apparent similarities, suggesting that gene regulation of L-tartrate utilization differs between E. coli and Salmonella.Furthermore, ArcAB regulates L-tartrate utilization in S. Tm (Fig. 5), but not commensal E. coli (29).These observations raise the possibility that the L-tartrate gene regulation in Salmonella could be an adaptation to the pathogenic lifecycle of Salmonella.

Plasmids, bacterial strains, and generation of mutants
The bacterial strains and plasmids used in this study are found in Table S1, and all primers are listed in Table S2.The upstream and downstream regions of ttdV, ttdW, ubiE, and menA were amplified by PCR from the IR715 genome using Q5 High-Fidelity Master Mix (New England Biolabs).PCR products were purified using the QIAEX II Gel Extraction Kit (Qiagen).Using Gibson Assembly (New England Biolabs), the upstream and downstream regions of ttdV, ttdW, and ttdVW were inserted into SphI-digested pGP706, and the upstream and downstream regions of ubiE and menA were inserted into SphI-digested pRDH10, giving rise to pSC1, pSC2, pSC3, pMW31, and pMW35, respectively.E. coli DH5α λpir served as the cloning strain for all suicide plasmids.The DNA sequence of upstream and downstream regions cloned into the suicide plasmid pGP706 was verified by Sanger sequencing.Suicide plasmids were introduced into S. Tm via conjugation.E. coli S17-1 λpir served as the donor strain for conjugation.Exconjugants (merodiploid) were selected on LB agar plates supplemented with Nal and Kan.To perform counterse lection, a merodiploid colony was grown overnight in LB broth at 37°C with aeration, and a diluted sample of this culture was plated on sucrose plates (5% sucrose, 15 g/L agar, and 8 g/L nutrient broth base).Clean deletions were verified via PCR using Quick-Load Taq Master Mix (New England Biolabs) or Q5 High-Fidelity Master Mix.This cloning strategy was used to make bacterial strains SC1 (ΔttdV), SC2 (ΔttdW), SC3 (ΔttdVW), MW201 (ΔmenA), and MW208 (ΔubiE).
To create complementation strains VR3, VR29, and VR28, we first created pPhoKm.We designed pPhoKm by inserting PCR amplified upstream and downstream DNA regions of the neutral phoN locus that would remove the existing SphI cut site from the plasmid and introduce a new SphI cut site in between the two phoN flanking regions and inserted these modified flanking regions into SphI-digested pGP706 using Gibson Assembly.We then PCR-amplified the native promoter and gene(s) of interest and inserted the fragment into SphI-digested pPhoKm.We used the same strategy of making clean deletions to make clean insertions of the promoter and gene of interest in the phoN region.

β-Galactosidase assay
Media prepared for all β-galactosidase assays that included L-tartrate were supplemen ted with a final concentration of 0.1 mM L-tartrate, unless indicated otherwise.Potassium tetrathionate (Sigma) or sodium nitrate (Sigma) was added at a final concentration of 4 mM.Three milliliters of media was preincubated in the anaerobic chamber (5% hydrogen, 5% CO 2 , and 90% nitrogen) 1 day prior to inoculation.Overnight cultures grown aerobically in LB broth and Cm were subcultured in the prereduced media at a 1:20 dilution and incubated in the anaerobic chamber at 37°C for 3 h.The subcultures were taken out of the anaerobic chamber and placed on ice for 20 min.The protocol described by Miller (42) was used to perform the assays.Briefly, after the optical density at a wavelength of 600 nm (OD 600 ) was measured, 100 or 500 µL of the subculture was combined with 900 or 500 µL sterile PBS supplemented with β-mercaptoethanol (Thermo Fisher), respectively.Bacterial lysis was carried out by adding 40 µL chloroform and 20 µL 0.1% SDS, then vortexing for 10 s, followed by 5-min incubation at room temperature.To begin the reaction, 200 µL of 4 mg/mL ortho-nitrophenyl-β-galactoside (Thermo Scientific) was added, and after allowing the reaction sufficient time to develop a yellow color, the reaction was stopped by adding 500 µL of 1 M Na 2 CO 3 and a final mix (vortex).The absorbance at a wavelength of 420 and 550 nm was measured, and the Miller units were calculated as follows: Miller units = 1,000 × ([A 420 -(1.75 × A 550 )]/(t × v × OD 600 )), where t is the time in minutes and v is the volume in milliliters.

Quantification of mRNA in S. Tm grown in vitro
Nine milliliters of freshly prepared media was placed in the anaerobic chamber 1 day prior to inoculation.Overnight cultures grown with aeration in LB broth were subcul tured in the prereduced media at a 1:10 dilution then incubated at 37°C for 3 h.At the end of the incubation, the cultures were removed from the anaerobic chamber and placed in a 4°C centrifuge, and the suspension was centrifuged for 5 min at 3,220 × g.After discarding the media, total RNA was extracted using the Aurum Total RNA Mini Kit (Bio-Rad), treated with DNase I (Thermo Fisher) twice then stored at −80°C.

RT-PCR and qPCR
The concentration of the DNase-treated RNA was determined by measuring the A 260 / A 280 ratio (NanoDrop One, Thermo Scientific).A 25-µL reaction was prepared for each sample using the following TaqMan reverse transcription reagents: 2.5 µL 10× RT-PCR buffer, 5.5 µL 25 mM MgCl 2 , 5 µL dNTP mix, 2.5 mM each, 1.3 µL of 50 µM random hexamers, 0.5 µL RNase inhibitor, and 0.6 µL reverse transcriptase (RT) and 9.6 µL of template RNA.To account for DNA contamination, control reactions lacking RT were prepared for each sample by mixing RT-PCR buffer, template RNA, and water.Samples were incubated in a thermal cycler at 25°C for 10 min, 48°C for 30 min, and 95°C for 5 min.
To perform real-time PCR (qPCR), we prepared a 25-µL reaction for each sample: 12.5 µL SYBR Green Mix, 0.625 µL of forward and reverse primer, 7.25 µL water, and 4 µL cDNA.Primer sequences are listed in Table S2.Two technical replicates for each biological replicate were assayed using the SYBR green qPCR supplied protocol from QuantStudio 6 Real-Time System.QuantStudio qPCR software was used for data analysis, determining baselines through the baseline threshold algorithm.Subsequently, Microsoft Excel and Prism 10 were used to further analyze the data, employing the comparative threshold cycle method.

Protein expression and purification
The strains and plasmids used to generate and maintain the recombinant proteins are described in Tables S1 and S2, respectively.To generate pVR17 and pVR18, we PCR-ampli fied the ttdV and ttdW genes, leaving out the stop codon, from genomic material of S. Tm IR715 as the template and using Q5 High-Fidelity Master Mix.The PCR products were purified using the same DNA purification kit for generating mutant strains and Gibson assembled with XhoI-digested pET21(+).The Gibson Assembly reaction was transformed into E. coli DH5α λpir, and the transformed bacteria were plated on LB agar plates supplemented with Carb.The plasmids were confirmed via PCR and subsequently cultured in 100 mL LB broth with 100 µL Carb overnight to extract the recombinant vectors using the Qiagen Plasmid Midi Kit.pVR17 and pVR18 were used to transform E. coli BL21 (DE3) cells, and the transformants were plated on LB agar with Carb.The presence of the recombinant plasmids was confirmed by PCR.
To express the recombinant proteins from BL21 cells, we inoculated 10 mL LB broth supplemented with 1 M glucose for a final concentration of 20 mM and 10 µL Carb with one colony of the BL21 strains harboring the recombinant plasmids pVR17 and pVR18.The starter culture was grown overnight at 37°C with aeration.The following day, 5 mL of the starter cultures was aliquoted into 100 mL LB broth with 100 µL Carb.The subculture was grown at 37°C with vigorous shaking for about 1 h or until the OD 600 range was between 0.5 and 0.7.The subculture was then supplemented with 100 µL of a 1 M ITPG stock for a final concentration of 1 mM to induce the protein expression of recombinant TtdV and TtdW.Protein induction lasted for 4-5 h at 37°C shaking at 250 rpm.The subculture was split into 50-mL tubes and centrifuged at 4°C at 3,220 × g for 20 min.The supernatant was discarded, and the cell pellets were stored at −20°C until ready for further extraction.The cell pellets were thawed on ice for 15 min before adding 5 mL of ice-cold lysis buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, and 10 mM imidazole), and 5 mg of lysozyme for a final concentration of 1 mg/mL was added to the lysate and incubated on ice for 30 min.To lyse the cells, the 5-mL cell suspension was kept on ice and sonicated with a microtip with 10 s bursts and 10 s rest periods six times.The cleared lysate was centrifuged at 10,000 × g for 20 min at 4°C.The supernatant was decanted into a new 15 mL tube (soluble fraction), and the pellet was resuspended with 5 mL lysis buffer (insoluble fraction).The whole-cell lysates pre-and post-IPTG induction and cellular fractions were loaded on a 10% polyacrylamide gel and Coomassie-stained to assess the expression and solubility of recombinant TtdV-His 6 or TtdW-His 6 .Recombinant TtdV-His 6 or TtdW-His 6 were both found in the soluble fraction when expressed from pET21(+) vector.To isolate recombinant protein from the soluble fraction, we followed the vendor-supplied protocol of the His-Pur Ni-NTA Purification Kit (3 mL, Thermo Fisher).The flowthrough, wash, and elution fractions were analyzed on a 10% polyacrylamide gel that was Coomassie-stained.We used the Pierce BCA Protein Assay Kit (Thermo Fisher) to measure the protein concentration from all three elution fractions.The third elution fraction had the least amount of protein contamination and was stored at 4°C and used for subsequent EMSA reactions.
To generate pVR55, we restriction-digested pET14b with BamHI, resolved the digested vector on a 0.7% agarose gel, and purified the DNA using the QIAEX II Gel Extraction Kit.We PCR-amplified the arcA gene, leaving out the start codon, from the S. Tm IR715 genome.After agarose gel electrophoresis, DNA was extracted, and Gibson Assembly was performed with BamHI-digested pET14b as described above.Next, we transformed E. coli DH5α λpir with 2 µL of the Gibson Assembly and plated the trans formants on LB agar supplemented with Carb.We verified the insertion of arcA into pET14b by PCR, then grew a single colony in 100 m LB broth supplemented with 100 µL Carb overnight to prepare for plasmid extraction.We used the Qiagen Plasmid Midi Kit to extract the plasmid.We used E. coli BL21 pLysE strain for the expression of arcA from pET14b due to potential toxic side effects that recombinant ArcA would have on endogenous ArcAB sensing.We plated the electroporated cells onto LB agar with Carb.Colonies were purified and confirmed to harbor the plasmid via PCR.
To purify His 6 -ArcA, we grew a 6 mL starter culture with LB broth, 6 µL Carb, 3 µL Cm, and 120 µL 1 M glucose (final concentration 20 mM) overnight at 37°C with aeration.We then subcultured 5 mL of the overnight starter culture into 250 mL LB broth containing 250 µL Carb.The subculture was shaken at 250 rpm at 37°C until the culture reached an optical density (OD 600 ) of 0.4-0.6.The subculture was induced with 250 µL of 1 M ITPG (final concentration 1 mM) and kept shaking for 5 h.One-milliliter aliquots were taken before and after induction to serve as whole-cell lysate controls.The subculture was split into two 250-mL centrifuge bottles and then centrifuged at 3,220 × g for 20 min at 4°C.The supernatant was removed, and pellets were stored at −20°C until the next day.The pellets were thawed on ice for 15 min before resuspending with 7 mL of ice-cold lysis buffer, collecting the pellets into one 15-mL tube and then adding 7 mg of lysozyme.The pellets were kept on ice for 30 min then sonicated with a microtip six times with 10 s bursts and 10 s rests.The cleared lysate was centrifuged at 10,000 × g for 20 min at 4°C; the supernatant (soluble fraction) was decanted into a new 15-mL tube, and the pellet (insoluble fraction) was resuspended in 7 mL lysis buffer.The whole-cell lysates from the uninduced and induced culture and the fractions were analyzed via SDS-PAGE and Coomassie-stained to verify His 6 -ArcA expression, and it was found in the soluble fraction.We used the same protocol and kits to purify His 6 -ArcA and measure each fraction's concentration.

EMSA using TtdV and TtdW
We prepared the biotinylated (labeled) and unbiotinylated (unlabeled) DNA probes by PCR amplification using S. Tm IR715 genomic DNA as the template and Q5 High-Fidelity Master Mix using the following thermocycling profile: denature at 98°C for 1 min, 35 cycles of melting at 98°C for 10 s, annealing at 50°C-72°C for 30 s, and extension at 72°C for 20 s with a final extension of 72°C for 1 min and an infinite hold at 4°C.The PCR products were loaded onto a 1% agarose gel, excised, and purified using the QIAEX II Gel Extraction Kit.We used the LightShift Chemiluminescent EMSA Kit (Thermo Scientific) and protocol for all EMSA reactions.For the competition reactions, we incubated unlabeled probe 4 with TtdW-His 6 for 5 min at room temperature before adding labeled probe 4.After a 20-min incubation at room temperature, we loaded the EMSA reactions onto TBE gels (0.5× TBE buffer, 4.5%-6% polyacrylamide, 0.5% glyc erol, 0.06% ammonium persulfate, and 0.2% tetramethylethylenediamine).TBE buffer contained 5.4 g/L Tris base, 2.8 g/L boric acid, and 10 mM EDTA.To transfer the binding reactions to the nylon membrane (Biodyne B Nylon Membrane, Thermo Scientific), we semi-dry-transferred the reactions at 15 V for 45 min and immediately crosslinked the transferred DNA to the membrane using a commercial UV light crosslinking oven (UV Stratalinker, Stratagene) with a 45-60 s exposure using the auto-crosslink function.

In vitro ArcA phosphorylation reaction and DNA binding assays using ArcA and ArcA~P
To phosphorylate ArcA before adding the modified protein to the EMSA reactions, we incubated purified His 6 -ArcA at a final concentration of 80 µg/mL, with carbamoyl phosphate (final concentration of 50 mM) and phosphorylation buffer (50 mM Tris-Cl, pH 7.0, 50 mM KCl, 20 mM MgCl 2 , 20 mM dithiothreitol, and 10 µg/mL bovine serum albumin, all final concentrations) for a 60-µL reaction.After 1 h incubation at room temperature, the phosphorylation reaction was stopped with 200 mM EDTA pH 8.0.The phosphorylated ArcA was then added to the appropriate EMSA reactions.
The same probes were used to assess ArcA/ArcA~P binding, and the same kit and protocol were used to perform all EMSA reactions, with minor modifications.We included 10 mM EDTA to the binding reactions.To assess competing DNA binding between labeled probe 4 and unlabeled probe 4, we incubated ArcA~P with unlabeled probe 4 for 5 min at room temperature before adding labeled probe 4.The EMSA reactions were loaded onto a 4.5% 0.5× TBE gel, and the same transfer and crosslinking process was followed.

Anaerobic growth of S. Tm in M9 media supplemented with casamino acids
Media that were preincubated in the anaerobic chamber for 1 day were inoculated with equal ratios of AJB715 and LS25 at a final concentration of 1 × 10 3 CFU/mL of each strain.Final concentrations of 20 mM L-tartrate, 20 mM glycerol, and 0.2% casamino acids, as well as 40 mM sodium nitrate or 40 mM potassium tetrathionate, were added as indicated.The strains were grown in the anaerobic chamber for 18 h at 37°C, and recovery of each strain was determined by spread-plating serial dilutions on selective LB agar plates.

CBA infection model
All mouse experiments were performed in accordance with the Institutional Animal Care and Use Committees at UT Southwestern Medical Center and UC Davis.Groups of CBA mice (aged 8-10 weeks old) were intragastrically infected with 5 × 10 8 CFU of each strain for competitive infection experiments.After 4 days, colonic and cecal luminal content was collected for bacterial enumeration.Serial dilutions of intestinal content were plated on selective media.

FIG 1
FIG 1 Substrate induction of the L-tartrate utilization gene cluster.(A) Schematic representation of the ttdBAU-ttdVW gene cluster in S. Tm LT2.The gene locus number and (putative) function of the gene product are listed.STM3357 and STM3358 were renamed ttdV and ttdW consistent with their proposed function of regulating L-tartrate metabolism in S. Tm. (B) The transcriptional ttdA-lacZ fusion strain (ttdA::pVR7) was cultured in media containing M9 salts and casamino acids (M9C) supplemented with various concentrations of L-, D-, or meso-tartrate and grown anaerobically.After 5 h, β-galactosidase activity was assessed.(C) The S. Tm wild-type strain was grown anaerobically in M9C media supplemented without or with 0.1 mM L-tartrate for 3 h.Total RNA was extracted, and mRNA levels were quantified by RT-qPCR; mRNA levels were normalized to the housekeeping gene gmk.(D and E) mRNA levels of ttdA (D) and ttdU (U) were determined in the wild-type strain and the ttdB mutant, cultured in the absence (black bars) and presence (gray bars) of L-tartrate.(F) mRNA levels of ttdA were determined in the wild-type strain and the ttdU mutant.Bacteria were cultured in the absence (black bars) and presence (gray bars) of L-tartrate.Bars represent the geometric mean with geometric standard deviation.Each dot represents one biological replicate.WT; wild-type.*** P < 0.001; ns, not significant (ANOVA).

FIG 2
FIG 2 TtdV and TtdW are transcriptional repressors for the L-tartrate utilization genes.(A) The indicated S. Tm ttdA-lacZ transcriptional fusion strains were grown anaerobically without (black bars) or with (0.1 mM; gray bars) L-tartrate.After 3 h, β-galactosidase activity was determined.The pertinent genotype is listed.For genetic complementation, genes of interest and their native promoters were inserted into the neutral phoN gene.(B and C) The start codon of the ttdV gene (B) and the ttdW gene (C) was altered as indicated.Bacteria carrying a ttdA-lacZ transcriptional fusion were cultured in the absence (black bars) and presence (gray bars) of L-tartrate, and β-galactosidase activity was quantified as described above.Bars represent the geometric mean with geometric standard deviation.Each dot represents one biological replicate.* P < 0.05; *** P < 0.001; ns, not significant (ANOVA).

FIG 3 FIG 4
FIG 3 TtdW bind to the intergenic region between ttdU and ttdV.(A) Schematic representation of the DNA probes used to probe for interaction with TtdW.(B and C) Electrophoretic mobility shift assays contained 20 fmol of DNA probes and were resolved on a 6% 0.5× TBE polyacrylamide gel.TtdW-His 6 was added at a concentration of 400 nM (B) or at increasing concentrations (C).Gel images are representative of two independent experiments.The approximate location of free, unbound DNA and a TtdW-DNA complex is indicated on the left side of the gel.* indicates a non-specific band that is present in all binding reactions.

FIG 5
FIG 5 The ArcAB two-component system activates ttdBAU transcription.(A) ttdA-lacZ transcriptional reporter strains were anaerobically cultured in M9C media in the absence or presence of 0.1 mM L-tartrate.After growth for 3 h, β-galactosidase activity was determined.The relevant genotype of each strain is shown below the graph.(B) EMS assays contained 50 fmol of DNA probes and were resolved on a 4.5% 0.5× Tris-borate-EDTA polyacrylamide gel.ArcA-His 6 and phosphorylated ArcA-His 6 were added at a concentration of 1,500 nM, as indicated.A 50-fold excess amount of competitor unlabeled probe 4 to the EMS reaction mixture, as indicated.The approximate location of free DNA and DNA-protein complexes is indicated on the left side of the gel image.(C and D) The ttdA-lacZ transcriptional reporter fusion strains, including a ubiE mutant (C) and a menA mutant (D), were anaerobically cultured in M9C media in the absence or presence of 0.1 mM L-tartrate.After growth for 3 h, β-galactosidase activity was determined.Bars represent the geometric mean with geometric standard deviation.Each dot represents one biological replicate.* P < 0.05; *** P < 0.001; ns, not significant (ANOVA).

FIG 6
FIG 6 Regulation of L-tartrate metabolism is required for optimal fitness in the murine gut.(A) Minimal M9 media containing 20 mM L-tartrate, 20 mM glycerol, and 0.2% casamino acids were supplemented with 40 mM tetrathionate or 40 mM nitrate as indicated.An equal mixture of the S. Tm WT strain and a ΔttdBAU mutant strain was added, and the suspension was incubated in an anaerobic chamber for 18 h.The fitness of the strains in each condition was measured by determining the competitive index.(B) ttdA-lacZ transcriptional reporter strains were anaerobically cultured in M9C media in the absence or presence of 0.1 mM L-tartrate and 4 mM tetrathionate or 4 mM nitrate and grown anaerobically or aerobically, as indicated.β-Galactosidase activity was determined after 3 h of growth.(C-F) Groups of CBA mice were intragastrically inoculated with a 1:1 ratio of the S. Tm wild-type strain (black bars) and a ttdV mutant (gray bars) (C and D) or the wild-type strain (black bars) and a ttdW mutant (white bars) (E and F).Samples of the cecal (C and E) and colon (D and F) content were collected 4 days after infection, and the fitness of each strain in the colon and cecal content was measured by determining the competitive index.Bars represent the geometric mean with geometric SD.Each dot represents one biological replicate.* P < 0.05; ** P < 0.01; *** P < 0.001.

FIG 7
FIG 7 Proposed model for the regulation of the L-tartrate utilization gene cluster ttdBAU ttdVW in S. Tm.In the absence of exogenous electron acceptors and L-tartrate, TtdV and TtdW repress the transcription of ttdBAU (top left panel).Increased availability of L-tartrate, in the absence of exogenous electron acceptors, leads to de-repression by TtdV and TtdW; phosphorylated ArcA binds in the ttdU-ttdV intergenic region and enhances transcription of ttdBAU (top right panel).In the presence of respiratory electron acceptors, the ArcAB system is inactive and unable to enhance the transcription of ttdBAU, regardless of whether L-tartrate is absent (bottom left panel) or present (bottom right panel).