Differential thiol-based switches jumpstart Vibrio cholerae pathogenesis

Abstract

Bacterial pathogens utilize gene expression versatility to adapt to environmental changes. Vibrio cholerae, the causative agent of cholera, encounters redox potential changes when it transitions from oxygen-rich aquatic reservoirs to the oxygen-limiting human gastrointestinal tract. We previously showed that the virulence regulator AphB uses thiol-based switches to sense the anoxic host environment and transcriptionally activate the key virulence activator tcpP. Here, by performing a high-throughput transposon sequencing screen in vivo, we identified OhrR as another regulator that enables V. cholerae rapid anoxic adaptation. Like AphB, reduced OhrR binds to and regulates the tcpP promoter. OhrR and AphB displayed differential dynamics in response to redox potential changes: OhrR is reduced more rapidly than AphB. Furthermore, OhrR thiol modification is required for rapid activation of virulence and successful colonization. This reveals a mechanism whereby bacterial pathogens employ posttranslational modifications of multiple transcription factors to sense and adapt to dynamic environmental changes.

Introduction

Pathogenic bacteria must be able to modulate their gene expression in order to adapt to different environmental conditions. They need to survive the stresses imposed by the host immune system, and cope with environmental stresses in reservoirs outside the host. For example, Vibrio cholerae, the causative agent of the diarrheal disease cholera, encounters a myriad of diverse and potentially growth-compromising conditions during its life cycle, which consists of a free-living phase in aquatic environments and infection of the small intestine of humans (Krukonis and DiRita, 2003; Skorupski and Taylor, 1997). In order to colonize and cause disease, V. cholerae uses sophisticated signal transduction networks to activate a set of virulence factors (Matson et al., 2007). The master virulence regulator, ToxT, controls expression of an array of virulence genes, most notably cholera toxin. The expression of toxT requires two transmembrane regulators, ToxR and TcpP; two unlinked regulators, AphA and AphB, in turn regulate TcpP. V. cholerae tightly regulates gene expression timing during infection in response to host stimuli. Virulence genes are induced early by a number of host signals, including bile salts (Yang et al., 2013). Late in infection, virulence genes are repressed and a coordinated “escape response” allows the organism to detach from the intestinal surface in preparation for exit from the host (Larocque et al., 2005; Nielsen et al., 2006; Nielsen et al., 2010b). Virulence gene repression is mediated partially by a combination of RpoS, quorum sensing, and anatomical site controls (Nielsen et al., 2006; Nielsen et al., 2010a; Zhu et al., 2002). V. cholerae also represses a set of genes to evade host defenses during early infection (Hsiao et al., 2006; Liu et al., 2008), and activates them late in infection to facilitate intestinal escape, to prepare for survival during the passage into the aquatic environment, or to become hyperinfectious and ready for transmission to another host (Merrell et al., 2002; Nelson et al., 2009; Schild et al., 2007; Tsou et al., 2008).

How V. cholerae overcomes the stress of changing oxygen tension when it moves from oxygen-rich aquatic reservoirs to the oxygen-limiting human gastrointestinal tract is less well understood. The key virulence activator AphB, a LysR-family protein widely present in prokaryotes, senses oxygen tension. We previously showed (Liu et al., 2011) that under O₂-limiting conditions similar to the gastrointestinal tract, the activity of AphB is enhanced, which leads to the production of virulence factors. This modification is dependent on one key cysteine residue and is reversible between O₂-rich and O₂-limiting conditions, suggesting that V. cholerae uses a thiol-based switch to sense O₂-limiting conditions and activate virulence. In this study, we performed an in vivo high throughput screen and identified a reactive oxygen species (ROS) resistance regulator, OhrR, as an additional anoxic sensor during V. cholerae infection. OhrR belongs to the MarR family of regulators found in both Gram-positive and Gram–negative bacteria (Dubbs and Mongkolsuk, 2012). We found that OhrR works together with AphB to directly regulate the expression of the virulence activator tcpP. More interestingly, we found that when V. cholerae transitions between the host and external environments, AphB and OhrR exhibit different kinetics for conformational changes and thus activity. Therefore, our findings suggest that AphB and OhrR work in coordination to sense changes in oxygen concentration and optimize bacterial fitness during host colonization.

Results

Tn-seq screens identify OhrR as a redox-dependent colonization factor

We previously showed that the O₂-limiting gastrointestinal tract enhances activity of the virulence activator AphB, which leads to the production of virulence factors (Liu et al., 2011). One of the three cysteine residues in AphB, C235, is critical for this O₂-dependent response, as the non-modifiable AphB^C235S mutant activates tcpP even under aerobic conditions. We thus hypothesized that the aphB^C235S mutant may have a colonization advantage over wildtype if the inoculum is an aerobically grown culture. However, we found that wildtype colonized as well as the aphB^C235S mutant in the infant mouse model (Fig. S1A and S1B), whereas the ΔaphB mutant failed to colonize mice under both conditions, as expected (Fig. S1). These data suggest that there may be additional redox-sensing regulators during infection. To identify such regulators, we performed a transposon insertion site sequencing (Tn-seq) screen in the infant mouse model (Fig. 1A) to look for mutants that have a colonization defect only when they have not adapted to O₂-limiting conditions (aerobic-growth cultures). To avoid issues with bottlenecks, which can lead to a loss of library diversity (Chiang and Mekalanos, 1998), we selected transposon insertions in 296 transcriptional regulators from a defined transposon library (Cameron et al., 2008). We made pools of ≈50 Tn-mutant strains and grew them either aerobically or microaerobically (standing cultures) and then inoculated them into separate infant mice. After a 20-hr incubation, we isolated colonized bacteria. We then extracted bacterial DNA and used Tn-seq (Fu et al., 2013; Kamp et al., 2013) to determine the number of transposon insertions in the input (starting cultures) and output (colonized bacteria) mutant libraries. We compared the output/input ratios of initial inocula consisting of shaking cultures (O₂⁺) and standing cultures (O₂⁻)(Fig. 1B). Most of the mutants that were tested colonized mice either equally well or equally poorly between inocula precultured in O₂⁺ and O₂⁻ conditions (Supplemental data set). For example, key transposon insertions in regulatory genes previously identified as important for colonization, toxT, tcpP, aphA, aphB and luxO (Matson et al., 2007; Zhu et al., 2002), were significantly reduced in both outputs. However, four Tn mutants displayed over a 10-fold decrease in colonization with inoculum from O₂⁺ precultures compared to inoculum from O₂⁻ precultures (Fig. 1B), suggesting that these genes encode regulators that may be involved in V. cholerae adaptation between O₂-rich and O₂-poor environments. We competed these individual Tn mutants with the wildtype strain in the infant mouse model and found that with the exception of the VCA0697 mutant, all other mutants were defective in colonization when they were inoculated as aerobically grown cultures (Fig. S2). One of these Tn mutants disrupted VCA1005, which encodes OhrR, a MarR-family regulator. As OhrR homologs in other bacteria may play a role in sensing redox potential changes and regulating bacterial resistance to organic hydroperoxide (Dubbs and Mongkolsuk, 2012), we selected this gene for further study. We constructed an in-frame deletion mutant of ΔohrR and found that ΔohrR mutants had no growth disadvantage under aerobic conditions (Fig. S3), suggesting that the ΔohrR colonization defect when O₂⁺ precultures were used as inoculum is not due to the oxic growth rate.

Fig. 1. Tn-seq identification of OhrR as a redox-dependent virulence activator.

A. Schematic depiction of the Tn-seq screen. B. Volcano plot of Tn-seq results. Total output and input mapped read counts of 296 defined Tn mutants from infant mice infected with aerobic or microaerobic cultures of V. cholerae were normalized against P-value. Each spot represents an identified TnFGL3 mutant. X-axis is the average ratio of output reads/input reads (inocula from O₂⁻-grown cultures) divided by the average ratio of output reads/input reads (from O₂⁺-grown cultures) of each mutant. See Fig. S1–3 for effect of AphB on mouse colonization, ΔohrR growth rate, and Table S1 for primers used for Tn-seq.

OhrR activates AphB-regulated virulence gene

To determine how the ROS resistance regulator OhrR affects colonization, we first examined whether OhrR affects virulence gene expression in vitro. We grew wildtype and Δ ohrR mutant strains containing various virulence gene promoter-luminescence reporter plasmids (Liu et al., 2011) under aerobic conditions. We then shifted the cultures to microaerobic conditions in virulence-inducing AKI medium (Iwanaga et al., 1986). We found that compared to the wildtype strain, the expression of tcpA, encoding a key virulence determinant (Rhine and Taylor, 1994), was significantly reduced in the ΔohrR mutant (Fig. 2A). To elucidate where OhrR acts along the virulence regulatory pathway, transcription levels of various components of the virulence cascade were measured in the ΔohrR mutant background. We found that in the ΔohrR mutant, toxT and tcpP expression were reduced, but the expression of upstream regulatory genes toxR, aphA, and aphB were not (Fig. 2A). These results suggest that OhrR may regulate tcpP expression directly or indirectly. tcpP transcription under the transition from O₂⁺ to O₂⁻ conditions was reduced in the ΔohrR null mutant and increased in an ohrR overexpressing (ohrR^C) strain (Fig. 2B), suggesting that OhrR positively regulates tcpP. Sequence analysis indicated that the tcpP promoter contains a conserved OhrR binding site (Chuchue et al., 2006), (Fig. 2C) and we found that purified recombinant OhrR could bind to tcpP promoter DNA (Fig. 2D). As controls, purified AphB could also bind the same tcpP promoter DNA (Fig. S4). Of note, the activity of OhrR and AphB was dependent on the reducing agent DTT (Fig. 2D and S4), indicating that only reduced OhrR and AphB are functional. Although the family of transcriptional regulators (MarR family) to which OhrR belongs to often serve as repressors, some do serve as both repressors and activators (e.g. SlyA from S. typhimurium)(Stapleton et al., 2002). While we cannot rule out other potential pathways through which OhrR may regulate tcpP, our data suggest that OhrR directly activates tcpP transcription.

Fig. 2. OhrR activation of virulence during oxic-to-anoxic transition.

A. Effects of OhrR on virulence gene transcription. Wildtype and ΔohrR containing promoter-luxCDABE transcriptional fusion reporter plasmids of virulence gene tcpA and other major virulence regulator genes were grown aerobically and then incubated in AKI at 37°C for 1 hr without shaking. Luminescence was then measured and normalized against OD₆₀₀. Expression in ΔohrR was compared with that in wildtype. Results are the means and s.d. of three independent experiments. B. OhrR effect on tcpP expression during O₂⁺→O₂⁻ transition. ΔohrR, wildtype, and P_tac-ohrR (ohrR^C) strains containing a P_tcpP-luxCDABE were grown as above. tcpP expression was calculated by normalizing luminescence against OD₆₀₀. Results are the means and s.d. of three independent experiments. C. Sequence alignment of the OhrR binding sites at the tcpP and ohrA promoters in V. cholerae (Vc), and the ohrA promoter in Xanthomonas campestris (Xc), Bacillus subtilis (Bs), and Agrobacterium tumefaciens (At). Conserved sequences of OhrR binding site are highlighted. D. Gel shift assays using purified OhrR-His₆ and tcpP promoter. 0.1 ng of tcpP promoter DNA in a buffer with or without 1 mM DTT were incubated with different concentrations of OhrR as indicated and the samples were size-fractionated using polyacrylamide gels. See Fig. S4 for AphB gel shift.

V. cholerae OhrR belongs to the two-cysteine OhrR family, and the conserved cysteine residues are essential for its redox sensing mechanism (Panmanee et al., 2006). We mutagenized each of the four cysteine residues in OhrR to serine and examined the activity of the mutants. We found that OhrR^C23S and OhrR^C128S could activate tcpP under aerobic conditions (Fig. 3A), suggesting that C23 and C128 residues are critical for OhrR activity. Under similar conditions, AphB^C235S is constitutively active (Fig. 3B), consistent with our previous report (Liu et al., 2011). Taken together, the above results suggest that both AphB and OhrR activate virulence gene expression using a thiol-based switch mechanism. In addition, we also examined the effects of constitutive OhrR^C128S mutant on colonization. We found that similar to the aphB^C235S mutant, the ohrR^C128S mutant colonized infant mice as well as wildtype, even when O₂⁺ precultures were used as an inoculum (Fig. S5A and S5B). This again implies that wildtype V. cholerae is optimized for adapting to the oxic-to-anoxic transition.

Fig. 3. The effects of cysteine residues of OhrR (A) and AphB (B) on tcpP expression.

ΔohrR mutants containing P_tac-ohrR (wildtype and cysteine mutant derivatives) plasmids, and ΔaphB mutants containing P_BAD-aphB (wildtype and cysteine mutant derivatives) plasmids were grown aerobically in LB containing 0.1 mM IPTG or 0.1% arabinose at 37°C until early-log phase. RNA was extracted and qRT-PCR was performed to determine tcpP transcription levels. recA was used as an internal control for each sample. Fold changes of tcpP expression in different strains were shown as means and s.d. of three independent experiments. *: Student t-test P < 0.01. See Fig. S5A&B for ohrR^C128S colonization data.

OhrR is more sensitive to redox potential changes

Next, we asked why V. cholerae uses two distinct regulators, AphB and OhrR, to regulate the same gene, tcpP. As both OhrR and AphB reversibly modify their activity through cysteine residues in response to redox potential changes (Dubbs and Mongkolsuk, 2012; Liu et al., 2011), we hypothesized that these proteins might react to different levels of redox potential so that gene expression can be “fine–tuned” to better adapt to multiple environments. To test this hypothesis, we first examined the difference in conformational change of purified recombinant proteins AphB and OhrR in response to reducing agents using an ANS (1-anilinonaphthalene-8-sulfonic acid) binding assay (Matulis et al., 1999). ANS is a hydrophobic fluorescent dye that can bind to exposed hydrophobic patches on proteins and has been used to probe intermediates during protein folding (Ali et al., 1999). We found that ANS-OhrR fluorescence was more sensitive to DTT than ANS-AphB fluorescence (Fig. 4A). Although the ANS assay does not directly measure reduction kinetics, the above results suggest that OhrR is more sensitive to reduction than AphB.

Fig. 4. Dynamics of OhrR and AphB reduction during during oxic-to-anoxic transition.

A. AphB and OhrR conformational changes in response to DTT by using ANS binding assays (Matulis et al., 1999). The function of DTT concentration on AphB or OhrR structural changes was calculated using the following formula: Δ482nm (F_R) = (R_certain λ_482nm-Control λ_482nm)/(R_max λ_482nm-Control λ_482nm)×100, where F_R indicates the function of reagent (DTT), R_certain indicates the certain concentration of DTT used in experiment, control indicates without DTT, R_max indicates 1000 μM DTT. Results are means and s.d. of three independent experiments. B & C. Thiol trapping experiments. Early-log aerobically-grown cultures of ΔaphBΔohrR mutants carrying P_tac-ohrR-FLAG and P_BAD-aphB plasmids were shifted to the anaerobic chamber. At the time points indicated, trichloroacetic acid was added. The precipitated proteins were reacted with mPEG-Maleimide and detected by SDS-PAGE and Western blots. Representative images from three independent experiments are shown (B and C) and the percentage of reduced proteins at different time points was quantified using ImageQuant software (Life Sciences) (D). *: P < 0.01. See Fig. S6 for AphB reduction rate in ΔohrR mutants.

To examine the reduction rate of AphB and OhrR in bacterial cells, we shifted V. cholerae from O₂-rich to O₂-poor environments and monitored the production of free thiol groups in AphB and OhrR by labeling cell lysates with Poly (ethylene glycol) methyl ether maleimide (mPEG- Maleimide). mPEG-Maleimide is able to react with free thiols of cysteine residues and pegylated proteins have higher molecular weights (Tetsch et al., 2011). We found that under aerobic conditions, both OhrR and AphB were oxidized (Fig. 4B and 4C, 0 min). Approximately 40% of the OhrR protein was reduced within 10 minutes of being shifted to the anaerobic condition, whereas most of the AphB protein remained oxidized (Fig. 4B–D). The AphB protein started to get reduced at 30 min, and by 60 min, the majority of both proteins were reduced (Fig. 4B–D). These results suggest that OhrR protein is reduced more rapidly than AphB protein during the transition from aerobic to anaerobic conditions.

OhrR jumpstarts virulence gene expression in vitro and in vivo

We hypothesized that the sensitivity of OhrR to reducing conditions may allow V. cholerae to activate virulence genes during the transition from O₂-rich aquatic environments to the O₂-poor environment of the intestine more rapidly than would be possible with the less sensitive AphB alone. To test this, we first examined tcpP expression in different V. cholerae strains during the switch from oxic to anoxic environments. We found that compared to that of wildtype, tcpP expression was significantly lower in the ΔohrR mutant 20 minutes after the switch but caught up after 60 minutes (Fig. 5A, blue bars). We also verified that tcpP expression did not change with alterations in oxygen level in the aphB^C235S and ohrR^C128S mutants (Fig. 5A, green and orange bars), consistent with a previous report (Liu et al., 2011) and data shown in Fig. 3A. In addition, tcpP expression was abolished in both ΔaphB and ΔaphBΔohrR mutants (Fig. 5A, gray and white bars) as AphB is the key activator for tcpP. We then compared induction of the downstream virulence gene tcpA in the wildtype and ΔohrR strains upon shifting from oxic to anoxic environments. Fig. 5B shows that in the presence of OhrR (wildtype, solid line), tcpA was quickly induced, whereas in the ΔohrR mutant (dashed line), tcpA reached its maximal expression level at a slower rate. However, if the wildtype and ΔohrR mutant cultures were inoculated directly into anoxic conditions without first growing aerobically, no difference in tcpA expression was observed (Fig. 5C). These data suggest that the delayed tcpA activation in the ΔohrR mutant results from the lack of rapid OhrR activation of tcpP transcription during aerobic-to-anaerobic shifts.

Fig. 5. The effect of sensitive reduction of OhrR on virulence in vitro.

A. OhrR effects on tcpP during the transition from oxic to anoxic environments. Early-log aerobically grown cultures of wildtype, ΔohrR, aphB^C235S, ohrR^C128S, ΔaphB, and ΔaphBΔohrR mutants were reinoculated into AKI medium and incubated at 37°C in an anaerobic chamber. Samples were collected at the time points indicated for RNA purification and real-time qPCR examination of tcpP. The induction of tcpP was normalized to wildtype at 0 min. Results are the means and s.d. of three independent experiments. *: P < 0.01 (compared to wildtype at the same time point). B. OhrR effects on downstream virulence gene tcpA. Early-log aerobically grown cultures of wildtype and ΔohrR containing P_tcpA-luxCDABE plasmids were reinoculated into AKI medium and incubated at 37°C in an anaerobic chamber. Samples were taken out of the anaerobic chamber at different time points and luminescence was measured. C. Anaerobically grown ΔohrR mutants do not affect tcpA expression. Overnight cultures of wildtype and ΔohrR mutants containing a P_tcpA-luxCDABE plasmid were refreshed in LB medium and were incubated at 37°C anaerobically to OD₆₀₀ ≈ 0.5. The cultures were spun down and resuspended with fresh AKI media and incubated at 37°C anaerobically. Luminescence and OD₆₀₀ were then measured. The tcpA expression was calculated by normalizing luminescence against OD₆₀₀. Results are the means and s.d. of three independent experiments.

To demonstrate the importance of this OhrR-dependent “jumpstart” of virulence in vivo, we used recombinase-based in vivo expression technology (RIVET)(Lee et al., 1999) to detect the patterns of tcpP and downstream tcpA induction in V. cholerae strains with OhrR (wildtype) or without OhrR (ΔohrR) in the infant mouse model. We infected infant mice with aerobically grown cultures of wildtype or ΔohrR harboring lacZ::res1-tet-res1 and tcpP::tnpR or tcpA::tnpR (Lee et al., 1999). Induction of resolvase TnpR controlled by tcpP or tcpA promoter results in permanent removal of the tetracycline-resistant gene cassette in cells, rendering them tetracycline-sensitive (Tet^S). Therefore, tetracycline sensitivity can be used as a measure of tcpP or tcpA expression in vivo. At different time points of infection, the bacterial cells were recovered from mice and examined for resolution of the tetracycline-resistant cassette. In wildtype, we found that both tcpP and tcpA expression was rapidly activated early (greater percentage of Tet^S cells) in infection and reached maximal expression level at 10-hr post-infection (Fig. 6A and 6B, solid lines), consistent with a previous report (Lee et al., 1999). In contrast, the expression of tcpP and tcpA in ΔohrR was not fully induced until late in infection (Fig. 6A and 6B, dashed lines). To assess if the delayed induction of tcpP and tcpA in ΔohrR has an impact on V. cholerae pathogenesis, we examined colonization ability of ΔohrR using inocula precultured in anoxic (O₂⁻) or oxic (O₂⁺) conditions. We found that the ΔohrR mutant colonized as well as wildtype when O₂⁻ cultures were used as inocula, whereas when O₂⁺ cultures were used as inocula, ΔohrR mutants were defective in colonization (Fig. 6C and S5C). As expected, ΔaphBΔohrR double mutants were defective in colonization in both conditions (Fig. 6C and S5C). These results suggest that OhrR plays an important role in sensing host anoxic environments and facilitates AphB initiating virulence gene expression to colonize the host small intestine.

Fig. 6. The effect of sensitive reduction of OhrR on virulence in vivo.

tcpP (A) and downstream tcpA (B) expression kinetics of wildtype and ΔohrR in vivo. Wildtype and ΔohrR harboring lacZ::res1-tet-res1 and tcpP::tnpR or tcpA::tnpR (Lee et al., 1999) were grown aerobically at 37°C in LB before inoculation into infant mice. At the time points indicated, bacteria were recovered from mice and examined for resolving of tetracycline-resistance cassettes. For each data point, over 300 colonies were examined. The percentage of Tet sensitive (Tet^S) CFU was then calculated. *: P<0.01. C. ΔohrR and ΔaphBΔohrR colonization. Wildtype and mutants were grown microaerobically (O₂⁻) or aerobically (O₂⁺) to mid-log. Colonized bacteria were enumerated and competitive index (CI) was calculated by normalizing the output ratio of mutant/WT with the input ratio of mutant/WT. Horizontal lines represent average CI. *: P < 0.01. D. Working model. AphB and OhrR respond to redox potential changes differently. This results in differential regulation of tcpP so that promptly reduced OhrR “jumpstarts” virulence during the transition from aquatic environments into the host. See Fig. S5C for additional data on effect of OhrR on colonization.

Discussion

Bacterial pathogens have the ability to reprogram diverse aspects of their physiology to adapt to different niches during infection and other environmental transitions. For example, during the transition from oxic aquatic environments to anoxic host small intestines, V. cholerae must cope with numerous changes in the intestinal environment, such as temperature, osmolarity, pH, nutrient availability, the presence of antimicrobial agents, and oxygen concentration. In this study, we have identified a mechanism based on the coordination of two thiol-based switches that allows V. cholerae to rapidly adapt to the anoxic host environment and activate virulence (Fig. 6D). Both OhrR and AphB are involved in regulating the virulence gene tcpP. In the oxygen-rich aquatic environment, both OhrR and AphB are oxidized and thus not active. In the oxygen-limiting environment of the intestinal tract, OhrR is reduced more rapidly than AphB, thus helping jumpstart virulence gene expression.

It is intriguing that the key V. cholerae virulence activator TcpP is regulated by a number of factors. We show that AphB plays a key role in activating tcpP, and OhrR is important when V. cholerae enters the oxygen-limiting host environment from oxygen-rich aquatic reservoirs. In addition, the transcription of tcpP is AphA- (Skorupski and Taylor, 1999) and CRP-cAMP-dependent (Kovacikova and Skorupski, 2001). In fact, the involvement of multiple transcription factors in regulating a single promoter is a common strategy in bacterial transcriptional regulation (Gama-Castro et al., 2008). Since different transcription factors monitor different environmental conditions or metabolic states (Ishihama, 2010), their combined action at a promoter provides a mechanism for integrating multiple signals to control gene expression. In future studies, it will be of interest to explore and better understand these transcriptional regulators of tcpP and how V. cholerae precisely coordinates virulence expression in response to constantly changing host environments. In addition, TcpP activity is also regulated post-transcriptionally. It has been shown that a set of bile salts enhance TcpP activity by affecting its disulfide bond formation (Yang et al., 2013). TcpP protein is also subjected to proteolysis regulation (Matson and DiRita, 2005), and TcpP-ToxR interaction is regulated by oxygen concentration (Fan et al., 2014). These multiple levels of control suggest that virulence gene expression in V. cholerae is tightly regulated. We considered the possibility that OhrR not only facilitates tcpP transcriptional activation, but may also help to reduce AphB upon shifting from oxic to anoxic environments. To test this, we examined the reduction rate of AphB in ΔohrR strains. We found that AphB was reduced at a similar rate in ΔohrR as in wildtype (Fig. S6), suggesting that at least under the conditions tested, OhrR is not involved in AphB reduction.

We have previously shown that V. cholerae uses a thiol-based switch in AphB to sense redox potential change during the transition into the O₂-limiting small intestine and responds by activating virulence genes (Liu et al., 2011). In bacteria, thiol-based regulatory switches play central roles in cellular stress responses, particularly oxidative stresses, due to the reduction potential of protein sulfhydryls (Antelmann and Helmann, 2011; Paget and Buttner, 2003). Examples include OxyR in Escherichia coli (Choi et al., 2001), OhrR in Bacillus subtilis (Fuangthong and Helmann, 2002), and MgrA, SarZ, AgrA in Staphylococcus aureus (Chen et al., 2006; Chen et al., 2009; Sun et al., 2012), many of which are regulators for ROS resistance and also involved in virulence (Chen et al., 2011). Our results indicate that V. cholerae uses two regulators to monitor redox potential changes and control virulence. Rather than being redundant, the combined action of these regulators enable V. cholerae to optimize its response to changes in the redox potential of the environment by using differential redox kinetics. As AphB and OhrR belong to LysR and MarR families respectively, both of which are widespread in prokaryotes, we speculate that other bacterial pathogens may use a similar mechanism to modulate gene expression in order to rapidly adapt to different environmental niches.

Methods

In vivo Tn-seq screen and deep sequencing data analysis

296 transposon mutants with a transposon inserted in genes encoding transcriptional regulators were selected from the defined V. cholerae Tn library (Cameron et al., 2008) and grown individually in LB with appropriate antibiotics at 30°C overnight. Cultures of fifty or fewer mutants per group were then mixed in equal volume and used to inoculate fresh LB medium with appropriate antibiotics. The cultures were then incubated at 37°C aerobically (shaking) or microaerobically (standing) until early-log. Approximately 10⁷ bacterial cells were then inoculated into 5-day-old infant CD-1 mice (at least 3 mice/group). All input shaking or standing cultures were pooled and DNA was extracted. After 20-hr colonization, infant mice were sacrificed and homogenized small intestines were filtered through a 40 μm membrane. The filtrates were centrifuged, and bacterial pellets were resuspended into 20 ml LB medium with appropriate antibiotics and were grown to saturation for DNA extraction (output library). The transposon junctions were amplified from sheared gDNA samples and subjected to massive parallel sequencing using Illumina MiSeq as describe previously (Kamp et al., 2013). Modified primers and adaptor DNA oligos used in this study are listed in Table S1. All read mapping and primary data analysis were performed using the Galaxy server (https://usegalaxy.org). The C-tail sequences generated in the C-tailing procedure were removed using the “clip adapter sequences script” with the 3′-adapter set to 26C and minimum read length set to 26. The normalized reads were aligned to the V. cholerae strain N16961 reference chromosome I and II genomes, accession number NC_002505.1 and NC_002506.1 respectively, using Bowtie script with default settings. The output bowtie files were then submitted for further analysis using custom script. The hit numbers of all individual insertion sites in genomes were tabulated in Excel file, with its gene name and gene locus to which that position maps. The total number of reads in each Tn-seq group was normalized to 5,000,000. For each infant mouse colonization experiment, at least three replicates of input and output libraries from at least three individual animals were used for data analysis. Student t test was performed to assess the significance where P<0.01 was considered significant.

Measuring protein conformation changes in response to reductants

AphB and OhrR proteins were purified from aerobically grown cultures. 100 μM 1-anilinonaphthalene-8-sulfonic acid (ANS) (Invitrogen) was added to 6 μM proteins in a buffer containing 30 mM NaCl (pH 3.4). The samples were then incubated with different amounts of DTT and equilibrated at 25°C for 60 mins in the dark. The ANS fluorescence was then measured using a BioTek Synergy H1 spectrophotometer using 390 nm excitation wavelength and the emission spectra were scanned between 420 nm and 600 nm. The intensities at the emission spectra peak (482 nm) were recorded and plotted as a function of DTT concentration.

Free thiol-trapping assays

V. cholerae ΔaphBΔohrR mutants containing P_tac-ohrR-FLAG and P_BAD-aphB plasmids were grown in LB medium containing 0.5 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG), 0.1% arabinose, and appropriate antibiotics at 37°C aerobically until OD₆₀₀≈0.1. Bacterial cells were pelleted and resuspended in fresh LB medium in the absence of IPTG and arabinose. The cultures were then incubated at 37°C in an anaerobic chamber. At the time points indicated, the samples were retrieved and proteins were fixed and precipitated with cold trichloroacetic acid (TCA, 10% final concentration) in the anaerobic chamber. The pellets were dissolved in freshly prepared mPEG-Maleimide buffer [15 mM Poly (ethylene glycol) methyl ether maleimide (mPEG-Maleimide), 1 M Tris-HCl, pH 8.0, 0.1% (w/v) SDS] and incubated at 37 C for 1 hr. The samples then were mixed with SDS-PAGE loading buffer without DTT and proteins were separated on a non-reducing SDS-PAGE. Western immunoblots were then performed to detect OhrR with anti-FLAG and AphB with anti-AphB antibodies.