RND efflux regulates leuO transcription

The loss of RND efflux in V. cholerae resulted in attenuated CT and TCP production [14]. To identify genes involved in this process, we determined the transcriptome of RND efflux negative V. cholerae strain JB485 by RNA sequencing (RNAseq). Total RNA was isolated and sequenced from JB485 and WT following growth under virulence inducing conditions (AKI conditions). The data from three independent experiments was then analyzed as described in the methods to identify differentially expressed genes. The RNAseq identified 373 genes that were differentially expressed in JB485 (S1 Table). The largest category of differentially expressed genes were those of unknown function (132 genes) followed by metabolism (60 genes), transport and binding (75 genes) and regulatory genes (30 genes) (Table 1). Consistent with previous studies, the expression of many virulence genes was reduced in JB485 (S1 Table). This included genes involved in the production of CT, TCP and multifunctional-autoprocessing repeats-in-toxin (MARTX). Surprisingly, the operons encoding for the VexAB, VexCD, VexGH, VexIJK and VexLM RND efflux systems were upregulated. Additionally, many porin genes were also differentially expressed including the upregulation of VC0972, ompK, ompW and downregulation of ompV and ompS. As reduced OM permeability functions synergistically with active efflux to effect high-level antimicrobial resistance [1], it is possible that the altered expression of the RND systems and porins reflects adaptive responses to modulate the OM barrier properties in response to the loss of RND efflux. Altogether these results suggest that impaired RND efflux activates global responses that facilitate the expression genes involved in environmental adaptation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Categories of differentially expressed genes in JB485.

https://doi.org/10.1371/journal.ppat.1006804.t001

We hypothesized that the link between RND efflux and virulence factor production was being mediated by a transcription factor. There were 30 differentially expressed regulatory genes identified in the RNAseq results (S1 Table); 21 of which were upregulated. Two-component signal transduction regulatory systems (TCS) made up 57% of these identified regulatory genes, which is highly enriched compared to TCSs represented in the V. cholerae genome (~17%). TCS consist of a membrane sensor that relays environmental signals to a response regulator to induce a cellular response. As such, the enrichment of TCSs among the identified regulatory genes suggests that the loss of RND efflux may have resulted in physiological alterations in the periplasmic compartment; perhaps due to the accumulation of substrates that are normally effluxed from the periplasm by the RND systems. Several of the identified TCSs contribute to environmental adaptation and pathogenesis including CpxRA [16–18], CarRS [19, 20], VieSAB [21], and OmpR [22]. We excluded several of the identified genes as virulence regulators in JB485 including cpxRA, vexR, and breR as previous studies suggested that they did not affect CT and TCP production [18, 23, 24]. Among the remaining regulatory genes, the LysR-family transcription factor leuO was one of the most highly upregulated (~6-fold) in JB485. LeuO is a global regulator that has been linked to multiple V. cholerae phenotypes including ToxR regulon expression [11, 12, 25–27]. We therefore further investigated leuO in JB485.

To confirm that leuO expression was increased in JB485 we introduced a leuO-lacZ reporter into WT and JB485 and quantified leuO expression over time during growth under AKI conditions. The results showed increased leuO expression in JB485 relative to WT at each time point (Fig 1A). leuO expression was ~4-fold higher in JB485 than WT at 3.5h and its expression continued to increase through the duration of the experiment, confirming the RNAseq results.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. LeuO represses virulence in RND-deficient V. cholerae.

V. cholerae were cultured under AKI conditions and processed as described in the methods. (A) leuO expression in WT and JB485 bearing pXB266 (leuO-lacZ). (B) Deletion of leuO restores CT and TcpA production (inset) in JB485. (C) TcpP Western blot showing that leuO deletion restores TcpP production in JB485. The data presented (A & B) is the means ± SD from at least three independent experiments. *, P≤0.05.

https://doi.org/10.1371/journal.ppat.1006804.g001

Previous studies showed differential contributions of the six V. cholerae RND transporters to antimicrobial resistance and virulence factor production [28, 29]. VexB, VexD, VexH, and VexK pumps functioned in in vitro antimicrobial resistance while all six pumps contributed to CT and TCP production. Given this, we tested if individual RND pumps differentially affected leuO expression by quantifying leuO-lacZ expression in a panel of RND mutant strains following growth under AKI conditions. The results showed that cells lacking the vexB, vexD or vexBD RND pumps did not affect leuO expression (S1 Fig). The expression of leuO-lacZ significantly increased in the ΔvexBDH mutant. Deletion of four RND pumps (ΔvexBDHK) increased leuO expression to a level similar to JB485 (ΔvexBDFHKM). This indicated that vexBDHK were linked to the increased leuO transcription. The expression of leuO was unaffected in strains DT422 (ΔvexBFHKM; vexD+) and JB464 (ΔvexDFHKM, vexB+), suggesting that the presence of vexB or vexD alone could suppress the RND efflux-dependent induction of leuO transcription. The fact that vexB or vexD suppressed leuO expression is consistent with previous studies showing redundancy among the RND efflux systems [14, 28, 29].

Mutation of leuO restores virulence factor production in an RND negative mutant

Increased leuO expression in the RND negative mutant JB485 suggested that LeuO repressed CT and TCP production. To test this, we compared CT and TcpA production in WT, JB485, ΔleuO and JB485ΔleuO following growth under AKI conditions. Consistent with previous results [14], CT and TcpA production were attenuated in JB485 relative to WT (Fig 1B). Deletion of leuO in WT did not significantly affect CT or TcpA production. By contrast, leuO deletion in JB485 restored CT and TcpA production to WT levels (Fig 1B), confirming that increased leuO expression was responsible for reduced CT and TcpA production in JB485.

Transcription of tcpPH was reported to be repressed in JB485 [14]. We therefore tested if leuO deletion in JB485 affected TcpP production under AKI conditions. As shown in Fig 1C, TcpP production in WT peaked at 4h and then declined to very low levels following overnight growth. TcpP production in JB485 was greatly reduced relative to WT, and still declined over time. Deletion of leuO in JB485 elevated TcpP production, with TcpP peaking at 6h before declining. In contrast to WT, TcpP was still relatively abundant in overnight JB485ΔleuO cultures (Fig 1C). Although TcpP production appeared to be higher in JB485ΔleuO at 6h and 24h, CT and TcpA production in JB485ΔleuO was like WT (Fig 1B); this is likely due to TcpP indirectly regulating CT and TCP production. Collectively these results suggested that the loss of RND efflux increased leuO expression and that LeuO was responsible for reduced TcpP, CT and TcpA production.

LeuO represses the ToxR regulon in the absence of RND efflux

We hypothesized that if leuO was repressing specific genes in the ToxR regulon, then ectopic expression of the repressed genes would restore CT production. We tested this by overexpressing aphA, aphB, tcpPH and toxT in JB485 and quantifying CT production following growth under AKI conditions. Both toxT and tcpPH restored CT production (S2A Fig). As toxT is downstream of tcpPH in the ToxR regulon, this is consistent with previous studies linking tcpPH repression to attenuated CT/TCP production in JB485. AphA and AphB function upstream of tcpPH and bind to the tcpPH promoter to activate its expression [30]. The expression of aphA restored CT production in an arabinose-concentration dependent manner (S2B Fig). The expression of aphB partially restored CT production in JB485, but in contrast to aphA, CT production was independent of the arabinose concentration (S2B Fig). This suggested that decreased tcpPH expression may be due to LeuO repressing aphA transcription. We further tested this by using qRT-PCR to quantify aphA, aphB, leuO, and tcpP expression in WT and JB485 during growth under AKI conditions. The results showed ~6-fold increase in leuO expression and a ~2-fold decrease in tcpP expression in JB485 relative to WT (Fig 2A), while aphA expression was reduced by ~1.5-fold. By contrast, the expression of aphB and toxR was not significantly changed. Consistent with the qRT-PCR results we also observed that ToxR and AphB protein production was not different between WT and JB485 (Fig 2C and S3 Fig). These results suggested that leuO repressed aphA expression in JB485.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. LeuO represses aphA in JB485.

(A & B) The indicated V. cholerae strains were cultured under AKI conditions for 3.5h when total RNA was isolated and used for qRT-PCR to quantify aphA, aphB, tcpP and toxR expression. P-values are relative to a hypothetical ratio of 1.0 and were determined using the Student’s t-test. * P<0.05. (C) ToxR Western blot. The indicated strains were grown under AKI conditions for 6h when aliquots were collected, normalized by OD₆₀₀, and subjected to Western blotting with anti-ToxR antibody. The extra non-specific bands serve as loading controls. (D & E) Overexpression of leuO represses aphA-lacZ, but not rxtB-lacZ expression in E. coli. E. coli containing pBAD33::leuO or pBAD33 and a (D) aphA-lacZ or (E) rtxB-lacZ reporter plasmid was cultured in LB broth to mid-log phase (5h) when β-galactosidase activity was quantified. The data are the mean ±SD of 3 independent experiments. * P≤0.05 relative to the no arabinose control.

https://doi.org/10.1371/journal.ppat.1006804.g002

To further confirm that LeuO repressed aphA transcription in JB485, we compared aphA expression in JB485ΔleuO to WT. We hypothesized that if leuO was responsible for reduced aphA expression, then aphA expression should increase in JB485ΔleuO. The results showed an approximately 3-fold increase in aphA expression in JB485ΔleuO relative to WT, but no change in aphB expression, confirming our hypothesis (Fig 2B). Increased aphA transcription in JB485ΔleuO potentially explains the elevated TcpP protein observed in JB485ΔleuO (Fig 1C). These results provide additional evidence to suggest that LeuO-dependent repression of aphA decreased tcpPH expression and CT and TcpA production in JB485.

To address if LeuO was acting directly at the aphA promoter, we introduced pBAD33-leuO into E. coli bearing an aphA-lacZ reporter. Cultures were grown in the presence of arabinose to induce leuO expression before quantifying aphA-lacZ expression. The results showed an arabinose dose-dependent decrease in aphA-lacZ expression in the pBAD33-leuO culture and no effect on aphA expression in the pBAD33 control (Fig 2D). As an additional control we examined the effect of pBAD33-leuO on rtxB, which was also differentially regulated in JB485 (S1 Table). Overexpression of leuO did not affect rtxB expression in E. coli, indicating that LeuO did not directly regulate rtxB and exhibited specificity for the aphA promoter (Fig 2E). Gel-shift assays were also performed and showed that LeuO can directly bind to the aphA promoter (S4 Fig). Based on these results we concluded that LeuO directly repressed aphA transcription.

ToxR positively regulates leuO in response to impaired RND efflux

ToxR positively regulates leuO transcription in response to extracellular cues [11, 12]. We therefore tested if ToxR activated leuO transcription in cells lacking RND-mediated efflux. We cultured WT, ΔtoxRS, JB485, and JB485ΔtoxRS bearing pXB266 (leuO-lacZ) under AKI conditions and quantified leuO expression at 3h and 5h. The results showed increased leuO expression in JB485 by ~3.5-fold at 3h and 4.5-fold at 5h relative to WT (Fig 3A). By contrast, leuO expression was greatly diminished in JB485ΔtoxRS, indicating that toxRS was required for leuO expression in the absence of RND efflux. To confirm that ToxR was responsible for leuO expression in JB485, we complemented toxRS mutants during growth under AKI conditions by introducing pBAD33::toxRS (pXB302) into the ΔtoxRS and JB485ΔtoxRS mutants bearing pXB266 (leuO-lacZ). Strains grown in the presence of arabinose increased leuO expression relative to the no arabinose control (Fig 3B). Interestingly, the magnitude of leuO expression in JB485ΔtoxRS was ~30% greater than the efflux positive ΔtoxRS mutant at the same arabinose concentration. While the significance of this is unclear, we speculate that it reflects differences in ToxR activation in the RND negative mutant vs WT background. These results support the conclusion that ToxR positively regulates leuO and that ToxR is responsible for increased leuO transcription in JB485. These results also suggest that RND mediated efflux may influence ToxR activation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. LeuO expression in JB485 is dependent upon the ToxR PPD.

(A) Expression of leuO in JB485 requires toxRS. WT, ΔtoxRS or JB485 bearing pXB266 (leuO-lacZ) were cultured under AKI conditions for 3h or 5h when β-galactosidase activity was quantified. (B) Ectopic toxRS expression complements for leuO expression. The ΔtoxRS and JB485ΔtoxRS mutants bearing pXB266 and pBAD33::toxRS were cultured under AKI conditions without or with 0.1% arabinose for 5h when β-galactosidase was quantified. (C) ToxR^ΔPPD production. Strains were grown under AKI conditions for 5h when aliquots were collected, normalized by OD₆₀₀, and used for Western blotting with anti-ToxR antibody. A toxRS mutant carrying pBAD18::toxR^ΔPPDS grown under AKI conditions in the presence of 0%, 0.01% and 0.1% arabinose is included as a marker for ToxR^ΔPPD protein. (D & E) Differential requirements of the ToxR PPD for ompU and leuO expression. The indicated strains carrying an ompU-lacZ or leuO-lacZ reporter were cultured under AKI conditions for 5h when β-galactosidase activity was quantified. *, P≤0.05 relative to the parental strain.

https://doi.org/10.1371/journal.ppat.1006804.g003

The ToxR PPD is thought to function as an environmental sensor and is required for the activation of leuO transcription in response to extracellular cues [11, 12]. Since leuO expression in JB485 was dependent on ToxR (Fig 3A), we tested if the ToxR PPD was required for leuO expression in the RND null mutant. Therefore, we generated chromosomal ToxRΔPPD mutants in WT and JB485 by truncating ToxR at amino acid 188 to remove the C-terminal PPD as previously reported (hereafter referred to as toxR^ΔPPD) [31]. We then introduced pXB266 (leuO-lacZ) and pAL144 (ompU-lacZ) into WT, ΔtoxRS, toxR^ΔPPD, JB485, JB485ΔtoxRS and JB485toxR^ΔPPD, grew the resulting strains under AKI conditions, and quantified leuO and ompU expression.

We first validated that the toxR^ΔPPD allele was functional by examining ompU expression. Expression of ompU is dependent on ToxR, but independent of the ToxR PPD in rich media [31]. Consistent with this, toxRS deletion ablated ompU expression in WT and JB485 while PPD deletion had no effect on ompU expression (Fig 3D). In fact, ompU expression was elevated in the toxR^ΔPPD mutants relative to their WT parental strains; as previously observed [31]. These studies also revealed elevated ompU expression in JB485 relative to WT. Since ToxR protein abundance is correlated with increased ompU expression in nutrient limiting conditions [32], we assessed if the loss of RND-mediated efflux affected ToxR protein abundance. ToxR Western blots revealed similar levels of ToxR production in WT and JB485 (Fig 2C) and their isogenic ΔPPD mutants (Fig 3C). Similarly, deletion of the PPD in WT did not affect aphA, toxT, ctxA, and tcpA expression (S5 Fig), as previously reported [31]. Together these results confirmed that the toxR^ΔPPD allele was functional in WT and JB485 and indicated that the PPD was dispensable for ompU expression and ToxR regulon expression.

We next quantified leuO expression in the above strains. The results showed increased leuO transcription in JB485 relative to WT and impaired leuO expression in the isogenic ΔtoxRS mutants of both strains (Fig 3E); confirming that toxR positively regulated leuO expression in JB485. In contrast, deletion of the ToxR PPD in WT resulted in a ~5-fold reduction in leuO expression, indicating that the PPD differentially affects ToxR activity at the ompU and leuO promoters. Deletion of the PPD in JB485 did not abolish leuO expression, but rather reduced leuO expression to WT levels (Fig 3E). This suggests that ToxR lacking its PPD (i.e. ToxR^ΔPPD) maintains the capacity to activate leuO expression in JB485, but not in WT. From these results, we concluded that the ToxR PPD is important for leuO transcription including increased leuO expression in cells lacking RND efflux. However, since leuO was still expressed in JB485toxR^ΔPPD, albeit at a reduced level, other factors also influence leuO expression in the absence of RND efflux.

Malate induces leuO transcription via ToxR

The above data suggested that ToxR was activated upon loss of RND efflux. From this we speculated that cell metabolites accumulated in in the RND negative strain JB485 and were responsible for the changes in ToxR activity and increased leuO expression. The RNAseq results revealed differential expression of genes involved in malate metabolism with fumC (fumarate hydratase) being downregulated (S1 Table) and mdh (malate dehydrogenase) being upregulated by ~1.6-fold. As gene expression can be regulated by product feedback, this suggested that malate may accumulate in JB485 to effect ToxR activation. To test this, we quantified malate in JB58 and JB485 cells and culture supernatants following growth under AKI conditions. The results showed differential malate accumulation between WT and JB485. Cell lysates of strain JB485 contained elevated levels of malate relative to WT whereas WT culture supernatants contained more malate than JB485 supernatants (Fig 4A). These results supported the hypothesis that the absence of RND-mediated efflux resulted in intracellular malate accumulation which could be influencing ToxR activity in JB485.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. ToxR activates leuO transcription in response to malate.

(A) Malate quantification in WT and RND null mutant JB485 culture supernatants and cells following growth under AKI conditions for 3.5h. Malate was quantified as described in the methods. The reported results are the means ± SD of six biological replicates with each replicate consisting of two technical replicates. * P≤0.05. (B) The indicated V. cholerae strains bearing pXB266 (leuO-lacZ) were cultured under AKI conditions for 4h when malate was added to 0, 1 or 2 mM. The cultures were then incubated with shaking for an additional hour when β-galactosidase production was quantified. The data are the mean ±SD of 3 independent experiments. * P≤0.05 relative to the no malate control.

https://doi.org/10.1371/journal.ppat.1006804.g004

We next asked if the addition of exogenous malate stimulated ToxR-dependent expression of leuO. To test this, we quantified leuO-lacZ expression in WT, toxR^ΔPPD and ΔtoxRS following growth under AKI conditions for 1hr in the presence of varying concentrations of malate. We found that malate addition did not affect leuO expression in the toxR^ΔPPD and ΔtoxRS mutants, but increased leuO expression in a concentration-dependent manner in WT (Fig 4B), suggesting that malate can influence the activation state of ToxR by a process that was dependent upon the ToxR PPD. The effects of exogenous malate on leuO expression in WT were somewhat reduced relative to what was observed in JB485. The reasons for this are unclear, but could reflect physiological differences in WT that limit periplasmic malate accumulation in WT including active efflux by the RND systems and the production of the cation-selective porin OmpU which could restrict malate diffusion across the outer membrane [33]. When this assay was repeated with fumarate we did not observe changes in leuO expression (S6 Fig), suggesting that the effects of malate on ToxR were not the due to nonspecific effects of C4 dicarboxylic acids on the cell.

RND efflux is essential for colonization

The RND null mutant JB485 was highly attenuated in the infant mouse model and was not recovered from the small intestine following challenge [14]. Since the loss of RND-mediated efflux was pleiotropic, with cells exhibiting antimicrobial hypersensitivity and attenuated virulence factor production, the relative contribution of these two phenotypes to colonization was unclear. As the studies here revealed that leuO deletion restored in vitro virulence factor production in JB485, we tested whether leuO deletion also restored the ability of JB485 to colonize infant mice. To test this, we performed colonization competition assays between JB485ΔleuO and WT. The results showed that neither JB485ΔleuO nor JB485 could be recovered from the intestine of challenged mice, further confirming that the RND transporters were essential for colonization. We considered the possibility that deletion of leuO may have dysregulated virulence factor production in vivo and thus led to a colonization defect. To address this, we performed a competition assay between JB485 and JB485toxR^ΔPPD; ToxR^ΔPPD mutants have been shown to activate virulence factors normally (S5 Fig and [31]). However, we were again unable to recover JB485 or JB485toxR^ΔPPD from the challenged mice. Taken together, these results suggest that the function of the RND transporters in antimicrobial resistance, and/or their contribution to unknown physiological functions, is likely epistatic to virulence factor production for colonization.

Bacterial strains and growth conditions

The bacterial strains used in this study are listed in Table 2. E. coli strain EC100λpir was used for cloning experiments. E. coli strain SM10λpir was used for conjugating plasmids into V. cholerae. Bacterial strains were routinely grown in Lysogeny Broth (LB) broth or on LB agar at 37°C. AKI growth conditions were used to induce virulence gene expression in V. cholerae as previously described [41]. Antibiotics were used at the following concentrations: carbenicillin (Cb), 100 μg/ml; kanamycin (Km), 50 μg/ml; and streptomycin (Sm), 100 μg/ml.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Strains and plasmids used in this study.

https://doi.org/10.1371/journal.ppat.1006804.t002

Plasmid and mutant construction

The plasmids and oligonucleotides used in this study are listed in Table 2 and S3, respectively. Chromosomal DNA from strain N16961 was used as the template for cloning experiments. Transcriptional reporter plasmids were constructed as described below. The promoter regions of ctxAB, tcpA and rtxB were generated by PCR using the PctxAB-F/PctxAB-R, PtcpA-F/PtcpA-R, or PrtxB-F/PrtxB-R oligonucleotide primers. The resulting amplicons were digested with XhoI and XbaI restriction endonucleases and ligated into similarly digested pTL61T to generate the plasmids pXB193, pXB194 and pVA195, respectively. pXB209 was constructed similarly using the aphB-F/aphB-R primers with the resulting PCR amplicon being digested with EcoRI and XbaI restriction endonucleases before being ligated into similarly digested pBAD18Km. Deletion of leuO and the toxR PPD domain was generated using the plasmid pWM91ΔleuO or pWM91::ΔtoxRppd as previously described [11, 12]. Repair of lacZ in JB485 was accomplished by using PCR to amplify the lacZ locus from N16961 strain JB3 using the Vc.lacZ.F.BamHI and Vc.lacZ.R.SacI PCR primers. The resulting 5.1 kb amplicon, containing the lacZ locus and ~1 kb of flanking DNA sequence, was digested with BamHI and SacI restriction endonucleases before being ligated into similarly digested pWM91 to generate pWM91-lacZ. The resulting plasmid was then used to repair the lacZ allele by allelic exchange as previously described [11, 12].

Quantitative real time PCR

V. cholerae strains were grown under AKI conditions in AKI broth. Total RNA was then isolated from the cultures at 3h using Trizol per the manufacturer’s directions. cDNA was generated from the total RNA using the Superscript III RT (Invitrogen) and used as a template for qRT-PCR with the SYBR Green Dye Kit (Fisher Scientific). The expression level of specific genes was quantified by amplifying 25 ng cDNA with 0.3 μM primers using the SYBR green PCR mix on a StepOnePlus real-time PCR System (Applied Biosystems). The gene encoding for DNA gyrase (gryA) was used as the internal control. Gene expression changes were calculated using the 2^−ΔΔCT method and the presented results are the means ± standard error of the means (SEM) from three biological replicates, with each biological replicate generated from three technical replicates.

RNA sequencing

Total RNA from V. cholerae strains JB58 and JB485 grown under AKI conditions for 3h were isolated using TRIzol per the manufacturer’s directions (Invitrogen) and further purified using an RNeasy kit with in column DNase treatment (Qiagen). The resulting RNA samples were assessed using a Qubit 2.0 fluorimeter (Thermo Scientific) and Agilent Tapestation 2200 (Agilent Technologies). Sequencing libraries were generated using the Illumina TruSeq RNA Access library prep kit (Illumina). Cluster generation and 75-bp single-read single-indexed sequencing was performed on Illumina NextSeq 500 (Illumina). The resulting raw reads were trimmed to remove adaptor/primer sequences. CLC Genomics Workbench version 10.1 (Qiagen) was then used to map the reads from three independent experiments to the N16961 genome [42]. The identification of differentially expressed genes was accomplished using the CLC Genomics Workbench RNA-Seq Analysis tool. Genes showing a 2-fold or greater difference in expression and a P-value and False Discovery Rate P-value of less than or equal to 0.05 were identified as differentially expressed genes. The RNA sequencing data was deposited in the National Center for Biotechnology Information Sequence Read Archive under accession number SRP109296.

Transcriptional reporter assays

Expression of promoters that were transcriptionally fused to lacZ were assayed following growth under AKI conditions. Samples from the cultures were taken at the indicated time points and processed in triplicate to assay for β-galactosidase activity as previously described [43]. The reporter experiments were independently performed at least three times. Statistical analyses were determined using the Student’s t-test. The concentrations of malate and fumarate used in the reporter assays was based on the physiological concentrations reported for E. coli (i.e. 1.7 mM for malate and 0.12 mM for fumarate) [44].

Characterization of CT, TcpA, AphB, ToxR and TcpP production

CT production in V. cholerae strains grown under AKI conditions was determined by GM1 enzyme-linked immunosorbent CT assays as previously described using purified CT (Sigma) as a standard [28]. The production of TcpA, ToxR, AphB and TcpP was determined respectively by Western immunoblotting as previously described using the polyclonal rabbit antisera against TcpA, ToxR, AphB, and TcpP [12]. Immunoreactive proteins on the Western blots were visualized using the SuperSignal West Pico chemiluminescent detection kit (Thermo Scientific).

Electrophoretic mobility shift assay (EMSA)

LeuO was purified from middle logarithmic phase LB broth cultures of E. coli ER2566 bearing pVA175 (pMAL-c2::leuO) following induction with isopropyl β-D-1-thiogalactopyranoside (0.3 mM) for 2h as previously described [11, 25]. The cells were then harvested by centrifugation, resuspended in column buffer (20 mM Tris-HCl, 200 mM NaCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride), and lysed using a M-11P Microfluidizer according to the manufacturer’s instructions (Microfluidics). The lysates were cleared by centrifugation at 15,000 x g for 20 min and the clarified lysates were diluted 1:6 with column buffer and run over an amylose resin chromatography column (New England Biolabs). Proteins bound to the amylose resin were then eluted with elution buffer (20 mM Tris-HCl, 200 mM NaCl, 1 mM EDTA, 10 mM maltose), before LeuO was liberated from the MBP using factor Xa protease. Protein purity was assessed by SDS–PAGE with Coomassie brilliant blue R-250 staining. Protein concentration was determined using the Bradford Assay kit according to the manufacturer’s instructions (Thermo Scientific).

The gel-shift assays were performed as previously described [25]. Briefly, biotinylated probes were generated by PCR using primers that were purchased from the manufacturer (IDT) with a 5’ biotin label. The biotinylated probes were gel purified before being used in the gel shift assays. The DNA binding reactions were performed in a final volume 10 μl of binding buffer (10 mM Tris-HCl (pH 7.4), 5 mM NaCl, 50 mM KCl, 1 mM EDTA, 50 μg/ml BSA, 1.5 nM biotinylated probe and 10 μg/ml sheared salmon sperm DNA). The binding reactions were incubated at 30°C for 30 minutes before being subjected to electrophoresis on non-denaturing 5% polyacrylamide TBE gels. The resolved gels were electroblotted to positively charged nylon membrane before the membranes were subjected to UV crosslinking at 120,000 microjoules using a Stratalinker 1800 Crosslinker (Strategene). The biotinylated probes were then detected using the Pierce Chemiluminescent Nucleic Acid Detection Module (Thermo Scientific) and visualized using a FluorChem E imaging system (Protein Simple).

Malate determination

JB58 and JB485 were cultured under AKI conditions for 3.5h when the culture optical density at 600 nm was recorded and a one mL culture aliquot was collected. The cells were separated from the supernatant by centrifugation and the supernatant and cell pellet retained. Cell lysates were generated by resuspending the cell pellet in one mL of ddH₂O before subjecting the cells to three freeze-thaw cycles followed by sonication. Aliquots of the culture supernatant and cell lysates were then assayed for malate using the Malate Assay Kit (Sigma) according to the manufacturer’s directions and normalized by optical density. The reported results are the means ± standard deviation from six biological replicates, with each biological replicate generated from two technical replicates.

Infant mouse competition assay

Cultures of XBV255 (JB485ΔleuOΔlacZ) and WT strain JB3 (N16961 lacZ positive) or DK296 (RND null, lacZ positive) and XBV468 (JB485 ΔPPD ΔlacZ) were combined at a 1:1 ratio before being delivered perorally to infant mice as previously described [14]. The following day the small intestine from the stomach to the cecum was excised and homogenized in phosphate buffered saline. Serial dilutions of the homogenates were then plated onto LB agar containing X-gal and streptomycin to enumerate the bacterial strains based on colony color. A competitive index was then calculated as the ratio of the test strains in the input divided by the ratio of test strains in the output.

Ethics statement

Animal protocols were approved by the University of Pittsburgh Institutional Animal Care and Use Committee (Protocol number 15015310) and met the standards for humane animal care and use as set by the Animal Welfare Act and the NIH Guide for the Care and Use of Laboratory Animals.