All V. cholerae strains were derived from the C6706 O1 El Tor pandemic strain and grown in LB liquid media with appropriate antibiotics at 37°C with agitation. All anaerobic strains were grown in LYHBHI liquid media (BHI supplemented to 5 g/l yeast extract, 5 mg/l hemin, 1 mg/ml cellobiose, 1 mg/ml maltose, and 0.5 mg/ml cysteine–HCl) at 37°C in a Coy chamber under anaerobic conditions (5% H₂, 20% CO₂, balance N₂) (Alavi et al., 2020). All anaerobic strains used in animal experiments were grown for 48 h, then 1:100 subcultured into fresh media and grown an additional 48 h prior to gavage. vpsL mutants were constructed using natural transformation based on the published method (Dalia et al., 2014). Approximately a 3-kb vpsL upstream (forward primer: 5′-GTTAAGAGCACCGATTGCAC-3′, reverse primer: 5′-GTCGACGGATCCCCGGAATCTTCATCACTAGACGCTCCTAAC-3′) and downstream (forward primer: 5′-GAAGCAGCTCCAGCCTACAGCGTATTAAGACAGGGCACTTG-3′, reverse primer: 5′-GCATTTTTTACCGTCAGGGTC-3′) sequence was amplified and added to ends of a trimethoprim resistance (Tm^R) cassette amplified from SAD530 using primers ABD123 (5′-ATTCCGGGGATCCGTCGA-3′) and ABD124 (5′-TGTAGGCTGGAGCTGCTTC-3′) by overlapping PCR using Phusion High-Fidelity PCR Master Mix (Thermo Fisher Scientific, Waltham, MA, USA). The first round of PCR consisted of an equimolar vpsL upstream fragment, a vpsL downstream fragment, and Tm^R cassette amplicons with a cycle condition of 98°C for 30 s, followed by 15 cycles (94°C for 10 s, 50°C for 30 s, 72°C for 4 min), and 72°C for 10 min. A vpsL upstream forward primer and a downstream reverse primer were added to the PCR reaction which was then continued using cycle conditions as described above. The final PCR product was added to 0.45 ml of instant ocean sea salt solution (7 mg/ml) with chitin-induced V. cholerae and incubated for 16–24 h at 30°C stationary, allowing natural competence. 0.5 ml LB was then added, and the mixture was incubated for 2–3 h at 30°C with agitation. The final mixture was then plated on an LB agar plate with 10 µg/ml trimethoprim for overnight incubation at 30°C to select for transformants. The final picked mutants were checked by Sanger sequencing.

Specific pathogen-free (SPF) CD-1 suckling animals were purchased from Charles River Laboratories. Upon receipt, suckling mice were fasted, and an equivalent of about 1 mg/g body weight streptomycin was gavaged with 30-gauge plastic tubing. Pups were then placed with a lactating dam for 1 day before being inoculated with communities and V. cholerae. Human gut bacterial strains were prepared by taking a total 300 µl of OD₆₀₀ = 0.4 equivalent culture per animal divided evenly across constituent commensal strains. Appropriate amounts of anaerobic strain culture were pooled and pelleted by centrifugation and resuspended in 25 µl of fresh LYHBHI per animal. Mice then received 25 µl of the community mixture and 25 µl of V. cholerae (1 × 10⁴–1 × 10⁵ CFU) in PBS (Alavi et al., 2020). 4-day-old suckling CD-1 mice were fasted for 1.5 h prior to all gavages. 18 hours after infection, pups were sacrificed, intestinal tissue was homogenized in 5 ml of PBS, and V. cholerae was plated on a streptomycin LB-agar plate for enumeration and fitness assays. For RNA assays, 1 ml of intestinal homogenate from each infected pup was centrifuged at 14,000 rpm, the supernatant was removed, and the pellet was resuspended in 500 µl of TRIzol (Ambion, Foster City, CA, USA) at -80°C.

DNA was extracted from an intestinal homogenate and used as template for model strain detection by quantitative PCR. 500 µl of intestinal homogenate was mixed with 210 µl of 20% SDS, 500 µl phenol:chloroform:isoamyl-alcohol (24:24:1, Fisher Scientific), and 500 µl of 0.1-mm zirconia beads (BioSpec, Bartlesville, OK, USA). The mixture was then lysed with a bead beater at 2,400 RPM for 3 min. The qPCR assay to detect model strains was performed using a universal 16S rRNA gene primer set to detect total bacterial load (forward primer: 5′-GTGSTGCAYGGYTGTCGTCA-3′, reverse primer: 5′-ACGTCRTCCMCACCTTCCTC-3′) (Horz et al., 2005). Each reaction was done in triplicates with 12.5 µl of iQ SYBR Green Supermix (BIO-RAD2), 1 µl of 10µM forward and reverse primers, 5.5 µl nuclease-free water, and 5 µl of extracted DNA (200 ng/µl).

Total RNA was extracted using TRIzol according to the manufacturer’s instructions. DNA was degraded using Baseline-ZERO DNase (Lucigen, Middleton, WI, USA), and the samples were then treated with TRIzol again for purification. A sequencing library was prepared using the Ovation Mouse RNA-Seq System with addition of V. cholerae enrichment primers (NuGEN, San Carlos, CA, USA). Final libraries were checked using an Agilent Bioanalyzer, and a 75-bp single-read sequencing was performed on Illumina NextSeq500.

All RNA-Seq-sequenced libraries were trimmed using Trimmomatic v0.39 (Bolger et al., 2014), and reads were aligned to mouse and V. cholerae genomes, using HISAT2 version 2.1.0 (Kim et al., 2015) with reference genomes GCF_000001635.26_GRCm38.p6 and GCF_000006745.1_ASM674v1. HTSeq (Anders et al., 2015) was used to calculate raw counts. After rRNA and tRNA were removed, differential expression analysis was conducted using edgeR (Robinson et al., 2009; Mccarthy et al., 2012). Statistical analysis and visualization of V. cholerae pathway profiles were performed using package clusterProfiler (Yu et al., 2012; Wu et al., 2021) in R with the selected database KEGG, and genes with FDR <0.05 and FC >1 or FC <-1 were used for further analyses.

Intestinal homogenate H₂O₂ levels were measured using the Amplex UltraRed reagent kit (Invitrogen, Carlsbad, CA, USA). 50 µl of intestinal homogenate was collected and mixed with 50 µl of working solution (5 mM Amplex UltraRed reagent, 10 U horseradish peroxidase, HRP, 0.05 M sodium phosphate, pH 7.4) immediately after homogenization. The mixture was incubated at room temperature for 30 min, and absorbance was measured (560 nm). The H₂O₂ level was then determined based on the H₂O₂ standard curve.

50 µl of intestinal tissue homogenates collected from V. cholerae-infected suckling mice colonized with different commensal groups was incubated with 1 mM of hydrogen peroxide for 1 h at room temperature. After incubation, 10 µl of untreated and treated homogenates was collected, serially diluted, and plated on selective agar to determine survival rates.

Intestinal homogenates were prepared as described in previously published methods (Alavi et al., 2020). Briefly, intestines were collected from 6-day-old CD-1 suckling mice fasted for 18 h and homogenized in 2.5 ml sterile water, homogenates were centrifuged, and then aqueous supernatants were collected. The supernatant was treated at 100℃ for 30 min and filter sterilized with a 0.22-µM filter. The sample was then dried with the Savant Integrated SpeedVac System (Fisher Scientific) and resuspended in 0.5 ml sterile water. 2 ml of acetonitrile (Sigma-Aldrich, St. Louis, MO, USA) was added to the sample and incubated at room temperature for 20 min to deproteinize after vortexing (Humbert et al., 2012). The aqueous layer was then filter sterilized, dried, and resuspended in 0.5 ml sterile water. 1.5 ml of OD₆₀₀ = 0.4 pure culture was pelleted and resuspended in purified homogenates then incubated anaerobically for 24 h at 37°C. The culture was heat-treated at 100℃ for 30 min then centrifuged. The aqueous layer was then filter-sterilized with a 0.22-µM filter after being cooled to room temperature and sent for mass spectrometry analysis.

An overnight culture of V. cholerae was 1:100 inoculated into fresh LB liquid media containing 100 µg/ml streptomycin and grown to OD₆₀₀ = 0.3. 50 µM of H₂O₂ or an equal volume of PBS was added to the culture and incubated for 2 h at 37°C with shaking. The cultures were collected, and RNA was extracted using TRIzol according to the manufacturer’s instructions. 2 µg of total RNA was then used for reverse transcription PCR using the SuperScript™ IV One-Step RT-PCR System (Invitrogen). qPCR assays were preformed using gene-specific primer sets targeting vpsR (forward primer: 5′-TGGCACACTGCTGCTAAA-3′, reverse primer: 5′-ACATCGACTGCACGAACC-3′), vpsT (forward primer: 5′-ACCCTGATCAAAGGCATGAG-3′, reverse primer: 5′-GTGAGGTCACGACTGAGTTAC-3′), and vpsL (forward primer: 5′-CAGTATGCGAGTGATGGATAATGG-3′, reverse primer: 5′-TCGTGGATCGCCTTTGGT-3′), with recA (forward primer: 5′-ATTGAAGGCGAAATGGGCGATAG-3′, reverse primer: 5′-TACACATACAGTTGGATTGCTTGAGG-3′) as the reference gene. Each reaction was done in triplicate with 12.5 µl of iQ SYBR Green Supermix (BIO-RAD2), 1 µl of 10 µM forward and reverse primers each, 8.5 µl nuclease-free water, and 2 µl of 2-fold-diluted RT-PCR reaction product.

Overnight cultures of E. coli bsh^c and E. coli vector with pZE21 (Alavi et al., 2020) were 1:100 inoculated into fresh LB liquid media containing 50 µg/ml kanamycin and incubated for 24 h at 37℃ with shaking. The cultures were normalized to 1.5 ml of OD₆₀₀ = 0.4 culture, pelleted, and washed with 1.5 ml of PBS. The pellet was resuspended in 1.5 ml of PBS containing 2 mM of taurodeoxycholic acid (TDCA) and transferred to glass tubes then incubated at 37°C stationary for 24 h. After the incubation, supernatants were collected by centrifuge and filter-sterilized with 0.22-µM filters. Collected supernatants were sent for mass spectrometry analysis and tested for V. cholerae growth inhibition.

Supernatants collected as previously described were 1:1 mixed with a 1:1,000 diluted log-phase V. cholerae culture in a 96-well plate. OD₆₀₀ was then taken every 30 min for 20 h with a BioTek Synergy HTX multi-mode reader under 37°C with orbital shaking at 180 cpm (6 mm) using Gen5 2.09.

Bile acids were extracted from in vitro enzymatic assays for LC-MS measurements. An aliquot of 200 µl of the assay was acidified to a final concentration of 2% acetic acid. A liquid–liquid extraction was performed with 300 µl of ethyl acetate. The supernatant was collected (200 µl) and dried down under a stream of nitrogen. The tube content after drying was resuspended in 1 ml methanol and diluted 20× for LC-MS analysis.

Standards for TDCA and deoxycholic acid (DCA) (1 µM) were used to determine retention times and fragmentation patterns for each compound. Chromatographic separations were conducted in a Waters I-Class UPLC System using an ACQUITY UPLC BEH C18 column (2.1 × 100 mm, 1.7 µM; Waters, Milford, MA, USA). Water with 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B) were used as mobile phases. Sample separation was conducted after 1-µl injections with the column kept at 40°C, with a flow rate of 0.4 ml/min under the following gradient: 0 min, 10% B; 1 min, 10% B; 1.5 min, 25% B; 8 min, 45% B, 10 min, 90% B; 16 min, 90% B, 17 min, 45% B, 19 min, 25% B; 21 min, 10% B; 26 min, 10% B. Mass spectrometry measurements were conducted in a SYNAPT G2-Si quadrupole time-of-flight mass spectrometer (Waters). The instrument was operated in negative ion mode under MS/MS conditions with a source capillary voltage of 2.5 kV, source temperature of 150°C, desolvation temperature of 600°C, and desolvation gas at 600 l/h. Leucine enkephalin was used as lock spray standard for mass correction. The following mass transitions (m/z) produced by the indicated collision energies (trap and transfer) were used for signal identification of TDCA and DCA [TDCA, 79.95 and 124.00 (CE 40 V); DCA, 343.26 and 391.28 (CE 25 V)]. The QuanLynx software (Waters) was used for peak integration and peak area calculation and processing. A quality control sample was produced by pooling equal amounts of samples and used to monitor system stability and reproducibility.

Overnight V. cholerae cultures were 1:100 inoculated into fresh LB liquid media with 100 µg/ml streptomycin and 2 ml of dilution aliquoted into 100 × 15-mm sterilized polystyrene petri dishes with 10-mg glass beads (50–100 µm, Polysciences, Warrington, PA, USA), followed by incubation for 24 h at room temperature with slow shaking (10 rpm) (Zhu and Mekalanos, 2003). Planktonic cells were collected by pipetting media, followed by pelleting via centrifugation and resuspension in LB media containing 1 mM of different bile acids (diluted from 100× stock solution in dimethyl sulfoxide, DMSO). Beads were rinsed with 1× PBS to remove residual planktonic cells, and the biofilm-associated bacteria on the beads were then treated with 5 ml LB liquid media containing 1 mM of different bile acids. For both treatments, an identical volume of DMSO as bile acid stock was added to LB liquid media as control. Both planktonic cells and biofilm on the beads were treated for 4 h, and surviving bacteria were then enumerated by serial dilution and plating.

Statistical analysis was performed using GraphPad Prism 9 (GraphPad Software, Inc., La Jolla, CA, USA). Student’s unpaired t-test was used to analyze in vitro and ex vivo CFU and ratio group differences. The Mann–Whitney U-test was used to analyze in vivo CFU level differences between groups. Comparisons where P < 0.05 were considered statistically significant. Results are representative of independent repeated experiments.

Previous studies have shown that specific human gut microbiota compositions can affect V. cholerae colonization using the antibiotic-cleared infant CD-1 suckling mouse model, with dysbiotic gut microbiota being more susceptible to infection compared to prototypical microbial communities characteristic of healthy individuals (Alavi et al., 2020). As the gut microbiota is highly diverse from individual to individual, we used simplified model groups of microbes representative of “healthy” and “unhealthy” or dysbiotic states to examine the effects of microbiota structure on V. cholerae behavior during infection. The simplified infection-resistant model community (SR) consists of the species representing the major taxa found in either the healthy individuals in Bangladesh (Hsiao et al., 2014), Blautia obeum (previously classified as Ruminococcus obeum), Bacteroides vulgatus, and Clostridium scindens. The low-diversity dysbiotic state induced by recurrent diarrhea or malnutrition was represented by Streptococcus salivarius subsp. salivarius (SS). 4-day-old suckling mice were first treated with streptomycin to remove the host microbiome then coinfected with V. cholerae and the community strains after 24 h. The intestine was collected 18 h postinfection and homogenized for CFU enumeration of V. cholerae, with aliquots of homogenates also stored for host and bacterial RNA-Seq analysis. In accordance with previous findings using healthy and dysbiotic human gut microbiota (Alavi et al., 2020), we repeatedly observed that V. cholerae was able to colonize SS-containing animals at around 10⁷ CFU; when coinfected with SR, V. cholerae colonization was around ~4-fold lower (Figure 1A). We also verified the colonization of model microbes by measuring total bacterial load using qPCR assay with universal 16S primers after antibiotic treatment and after gavage; total bacterial 16S copy number was significantly higher after bacterial gavage compared to antibiotic treated mice (Figure 1B). These findings replicate previous studies that these model microbes are capable of colonizing mice intestine (Alavi et al., 2020).

Alavi et al. determined that one contributor of microbiota-dependent variation in infection resistance was a variable level of commensal-encoded bile salt hydrolase (BSH) that de-conjugates the attached amino acid of the bile acid taurocholic acid (TCA), leading to differences in induction of virulence gene expression; B. obeum (an SR organism) was a key BSH producer, while S. salivarius lacked BSH activity (Alavi et al., 2020). To examine other potential contributors to SR/SS infection differences, we examined V. cholerae and host gene expression during infection of suckling mice. We extracted RNA from intestinal homogenates of animals colonized with both communities and infected with V. cholerae and used RNA-Seq to examine community-dependent V. cholerae gene expression patterns. Using pathway analysis with the Kyoto Encyclopedia of Genes and Genomes (KEGG) (Kanehisa and Goto, 2000; Kanehisa, 2019; Kanehisa et al., 2021), we found that V. cholerae coinfected with SS significantly upregulated biofilm formation and phosphotransferase system pathways relative to infection of SR mice; in SR mice, 16 pathways were increased in expression relative to SS, predominantly metabolic pathways including carbon metabolism, amino acid metabolism, and oxidative phosphorylation (Figure 2A). Biofilm is a population of microbes aggregated and attached to a surface in a matrix of extracellular polysaccharide (EPS), nucleic acid, and proteins. The production and manipulation of Vibrio extracellular polysaccharide (VPS) are driven by gene products encoded by two gene clusters vps-I (vpsU, vpsA-K) and vps-II (vpsL-Q) (Yildiz and Schoolnik, 1999; Fong et al., 2010). We observed that the expression of most VPS genes is upregulated 2–4-fold during infection of SS-colonized animals compared to SR-colonized animals. Among VPS genes, vpsL displayed the strongest difference in expression between microbiota groups, with 4.5-fold higher in expression in SS vs. SR. VpsL is known to be essential for forming a stable biofilm structure (Fong et al., 2010). Other than the exopolysaccharide matrix production genes, another vps-coregulated protein, Bap1 (biofilm associated protein), also displayed a 2-fold increase in expression in infection of SS animals (Figure 2B). Bap1 is an extracellular matrix protein that modulates biofilm structure and pellicle formation and mediates cell-to-surface adhesion (Fong and Yildiz, 2008; Yan et al., 2016). Biofilms protect microbes, including V. cholerae, from environmental stress including extreme environments and host defense mechanism (Roilides et al., 2015; Yin et al., 2019), and increases infectivity (Zhu and Mekalanos, 2003; Tamayo et al., 2010). Our findings suggest that infection of SS-like permissive, dysbiotic microbiota could induce V. cholerae to form a more robust biofilm structure either directly or indirectly, leading to stronger defenses against environmental stress.

Given that the SS triggered the expression of the genes associated with biofilm formation in V. cholerae, and the critical role of biofilms in defense against environmental stresses, we next examined whether the host environment during infection was modulated by the microbiota through transcriptomic measures of host responses to V. cholerae infection in the presence of SR and SS model microbes. We found that mice coinfected with SR and V. cholerae exhibited a significantly higher expression of NOS1, NOS2, and DUOX2 in the total intestine (Figures 3A–C). NOS2 and DUOX2 are known to be involved in the production of compounds such as nitric oxide (NO) and extracellular hydrogen peroxide (H₂O₂), respectively (Geiszt et al., 2003; El Hassani et al., 2005). In response to numerous gastrointestinal infections, host enzymes are able to generate increased local reactive nitrogen stress (RNS) and reactive oxygen stress (ROS), which collectively form a key mechanism to inhibit pathogens (Paiva and Bozza, 2014). Nitric oxide synthases (NOS) catalyze the production of nitric oxide (NO), which includes three different isoforms, NOS1–3. The NO synthesized by NOS1 and NOS3 mainly acts as a neurotransmitter in the brain and peripheral nervous system and maintains a healthy cardiovascular system (Forstermann and Munzel, 2006; Knott and Bossy-Wetzel, 2009). NOS2 is involved in immune response and NO synthesis. ROS synthesis involves the expression of dual oxidase (DUOX) that catalyzes hydrogen peroxide production (Sommer and Backhed, 2015). Prior studies have demonstrated an increase in both ROS and RNS during V. cholerae infection; Ellis et al. showed that cholera patients show higher nitric oxide synthase 2 (NOS2) protein abundant during the acute phase (Ellis et al., 2015), while an examination of cholera patients’ mucosal gene expression by Bourque et al. demonstrated that dual oxidase 2 (DUOX2) and NOS2 are higher in expression during the acute phase (Bourque et al., 2018).

The resulting increase in NO and H₂O₂ production leads to high local oxidative stress, which forms a key element of the initial host response toward pathogen infection (Kim and Lee, 2014; Król and Kepinska, 2021). These antimicrobial effects are mediated by several pathways, including conversion of NO to peroxynitrite (OONO^-), and form OH· and nitrogen dioxide (·NO₂) (Nathan and Ding, 2010) to cause oxidative damage on phagocytosed pathogens, and pathogen in phagosome, and H₂O₂ or through a catalytic reaction to form hydroxyl radicals (OH·) which leads to pathogen DNA breakage, base oxidation and deamination, and lipid peroxidation (Paiva and Bozza, 2014). This increase in ROS production has been observed in cholera patients, as measured by NO metabolite levels in serum and urine (Janoff et al., 1997). Cholera patients’ duodenum samples at the acute phase showed more abundant NOS2 and DUOX2 transcript levels compared to the convalescent phase (Ellis et al., 2015; Bourque et al., 2018). Other than higher ROS/RNS enzyme expression, host catalase (CAT) gene expression is also induced by infection in the presence of SS (Figure 3D). Since catalase is an antioxidant enzyme that promotes the conversion and detoxification of ROS and H₂O₂ to prevent oxidative damage to host cells, this further supports an increase in ROS stress during infection of individuals with SR-like microbiota compared to individuals with an SS-like commensal microbe. We then measured H₂O₂ production in the homogenates from each model microbiome and found that colonization with the SR community induced significantly higher production of H₂O₂ (Figure 3E). Given that a significant difference in overall H₂O₂ was observed from total organ samples, local differences in levels of H₂O₂ at epithelial microenvironments as a function of microbiome and infection may be even higher (Bhattacharyya et al., 2014; Taylor and Colgan, 2017; Burgueno et al., 2019; Imlay, 2019).

Previous studies have shown that biofilm formation enhances V. cholerae resistance toward oxidative stresses driven by early infection (Wang et al., 2018), and our data suggest that SS can induce higher biofilm formation gene expression in V. cholerae. To test whether microbiome-dependent differences in biofilm-associated gene expression lead to an increase in oxidative stress resistance, we used an ex vivo ROS survival assay. We coinfected suckling animals with wild-type and biofilm defective vpsL mutants (Figure 4A) and our two model commensal collections. The mutant showed no significant differences in colonization between the two model microbiota (Figure 4B). We then isolated V. cholerae and associated extracellular structures by homogenizing infected small and large intestinal tissues and then treated V. cholerae-containing homogenates with H₂O₂. Survival rates of V. cholerae were then determined by plating. We found that wild-type V. cholerae exhibited >90% H₂O₂ survival after infection of the SS-co-colonized small intestine, while V. cholerae that colonized SR-bearing animals exhibited only ~55% survival. When the vpsL mutant was co-gavaged with the model microbes, this SS-dependent advantage was ablated (Figure 4C), suggesting that biofilm formation was necessary for this difference in oxidative stress resistance. We observed the same pattern from the large intestine where the wild-type V. cholerae from SS animals exhibited higher levels of ex vivo ROS resistance compared to V. cholerae from SR animals, and vpsL mutants in either microbiota context (Figure 4D), although the rate of survival in both colonization contacts was significantly reduced compared to the small intestine, which may indicate that the regulatory effects of commensal microbes are concentrated at the preferred site of colonization of V. cholerae, the distal small intestine. To determine whether biofilm biogenesis was induced by hydrogen peroxide, we treated V. cholerae culture with H₂O₂ in vitro and examined vpsR, vpsT, and vpsL expression. Quantitative RT-PCR analysis showed no significant differences in expression level between H₂O₂-treated and untreated cultures (Figure 4E), suggesting that other mechanisms, potentially microbial in nature, are responsible for observed in vivo microbiome-specific differences in V. cholerae biofilm production. Taken together, these findings suggest that i) microbiota-associated biofilm expression leads to different levels of biofilm-dependent ROS resistance in V. cholerae and ii) that this effect is much more substantial in the preferred site of colonization in the distal small intestine. In combination with higher levels of oxidative stress during infection of SR-like communities, this suggests that modulation of biofilm or biofilm formation potentially contributes to the higher V. cholerae colonization level observed during infection of SS-bearing individuals.

Since different bile acids have been associated with inhibition of V. cholerae growth in vitro and regulation of biofilm production (Hay and Zhu, 2015), we examined how host bile might affect microbiota- and biofilm-dependent V. cholerae fitness during infection. Bile is a complex mixture secreted by the gallbladder into the proximal intestine in order to aid with the emulsification and absorption of dietary fats. A dominant component of bile are bile acids and their salt forms, the primary forms of which are synthesized in the liver from cholesterol, often in amino-acid conjugated forms. The cycle of bile secretion and subsequent reabsorption in the small intestine and the detergent effects of bile mean that commensals and pathogens that target this body site have evolved methods to detoxify bile components and also use them as signaling cues to time body site-specific gene expression programs. One key element of microbial bile response is the expression of bile salt hydrolase (BSH) enzymes, which catalyze the deconjugation of amino acids from primary bile, leading to a reduction in detergent activity and subsequent toxicity (Chand et al., 2017). Subsequent microbial action can also dehydroxylate the sterol backbone of bile acids, leading to a variety of different forms of bile acids that circulate within the intestine of animals housing resident microbes (De Smet et al., 1995). Taurocholic acid (TCA), an abundant amino-acid-conjugated bile species in mouse and human gut, is able to strongly activate the V. cholerae virulence regulatory cascade by promoting the formation of disulfide bonds in the upstream regulator TcpP (Yang et al., 2013). In contrast, the de-conjugated form, cholic acid (CA), and the deconjugated and dehydroxylated form deoxycholic acid (DCA) cannot strongly induce virulence. A previous study has shown that SR microbes and SS can help drive differential colonization resistance by modulating the level of TCA conjugation and virulence gene expression (Alavi et al., 2020); however, the interaction of V. cholerae with other TCA-derived bile acid metabolites has not been well described. This is especially important as DCA generated by BSH-dependent deconjugation of TCA is reabsorbed and reconjugated to taurine by the host, being secreted as taurodeoxycholic acid (TDCA) (Figure 5A). We therefore used mass spectrometry to analyze the bile salt composition in the intestinal homogenates, before and after incubation with B. obeum and S. salivarius. We found that TDCA signals in both control and SS are significantly higher than when incubated with B. obeum (Figure 5B). DCA levels in B. obeum-treated intestinal homogenates were several logs higher compared to Streptococcus-treated and control homogenates (Figure 5C). The accumulation of DCA suggests that B. obeum is able to de-conjugate TDCA to DCA in addition to de-conjugating TCA to CA. To examine whether the B. obeum BSH enzyme is able to catalyze this reaction, we examined TDCA and DCA levels in intestinal homogenates incubated with E. coli expressing B. obeum BSH (bsh^c ) compared to isogenic vector strains lacking any endogenous BSH activity. We observed the same low TDCA and high DCA signal patterns compared to vector control (Figure 5). To confirm the conversion from TDCA to DCA via B. obeum BSH, pure solutions of TDCA were added to bsh^c cell pellets in PBS and incubated for 24 h. The supernatant was collected, filter sterilized, and used for mass spectrometry analysis. Consistent with our previous findings, the level of TDCA remaining was significantly higher in the vector control compared with bsh^c , while DCA levels were ~600-fold higher in bsh^c compared to vector control (Figures 6A, B). Previous studies have shown that DCA is inhibitory to V. cholerae growth (Hung et al., 2006). To confirm this, TDCA treated with bsh^C or vector control was added to fresh V. cholerae culture and OD₆₀₀ was taken every 30 min for 20 h. We observed that V. cholerae growth was inhibited by bsh^c -processed supernatant compared to the vector strain(Figure 6C). Importantly, the presence of B. obeum bsh activity is able to provide inhibitory activity in the presence of Streptococcus, supporting the idea that the presence of infection-resistant microbes are able to restore colonization resistance to otherwise susceptible microbiota configurations (Alavi et al., 2020). Taken together, these results indicate that B. obeum is capable of converting TDCA to DCA via BSH expression, leading to V. cholerae growth inhibition during colonization.

Previous studies have indicated that crude bile is able to induce biofilm formation, and that biofilm-associated cells are more resistant to bile toxicity (Hung et al., 2006). To test whether biofilm will contribute to V. cholerae resistance against different microbially modified bile acids, we first generated biofilm V. cholerae populations in media with beads as a surface for microbial attachment and subsequent biofilm development. We then exposed planktonic and biofilm-formed populations with products of microbial TCA metabolism: TCA, DCA, and TDCA. We found that TCA and TDCA do not significantly affect V. cholerae planktonic cell fitness in comparison with no bile acid (Figure 7A), but that after 4 h of treatment with 1 mM of DCA, planktonic V. cholerae exhibited a significantly lower survival. In contrast, biofilm-associated V. cholerae were insensitive to all bile treatments (Figure 7B). These repeatable data suggest that the formation of biofilm provides V. cholerae resistance toward certain bile acids present in the intestinal tract.