Abstract
Vibrio cholerae is a Gram-negative bacterial pathogen that causes the disease cholera, which affects nearly 1 million people each year. In between outbreaks, V. cholerae resides in fresh and salt water environments where it is able to persist through changes in temperature, oxygen, and salinity. One key characteristic that promotes environmental persistence of V. cholerae is the ability to form multicellular communities, called biofilms, that often adhere to biotic and abiotic sources. Biofilm formation in V. cholerae is positively regulated by the dinucleotide second messenger cyclic dimeric guanosine monophosphate (c-di-GMP). While most research on the c-di-GMP regulon has focused on biofilm formation or motility, we hypothesized the c-di-GMP signaling network encompassed a larger set of effector functions than reported. We found that high intracellular c-di-GMP increased catalase activity approximately 4-fold relative to strains with unaltered c-di-GMP. Genetic studies demonstrated that c-di-GMP mediated catalase activity was due to increased expression of the catalase encoding gene katB. Moreover, c-di-GMP mediated regulation of catalase activity and katB expression required the c-di-GMP dependent transcription factors VpsT and VpsR. Lastly, we found that high c-di-GMP increased survival after H2O2 challenge in a katB, vpsR, and vpsT dependent manner. Our results indicate antioxidant production is regulated by c-di-GMP in V. cholerae uncovering a new node in the growing VpsT and VpsR c-di-GMP signaling network.
Importance As a result of infection with V. cholerae, patients become dehydrated leading to death if not properly treated. The marine environment is the natural reservoir for V. cholerae where it can survive alterations in temperature, salinity, and oxygen. The second messenger molecule c-di-GMP is an important signal regulating host and marine environmental persistence because it controls whether V. cholerae will form a biofilm or disperse through flagellar motility. In this work, we demonstrate another function of c-di-GMP in V. cholerae biology: promoting tolerance to the reactive oxygen species H2O2 through differential regulation of catalase expression. Our results suggest a mechanism where c-di-GMP simultaneously controls biofilm formation and antioxidant production, which could promote persistence in human and marine environments.
Introduction
The Gram-negative bacterium Vibrio cholerae is the human pathogen that causes the diarrheal disease cholera. The most common route to infection is consumption of contaminated food or water, after which V. cholerae traverses the stomach and colonizes the small intestines. Cholera patients lose liters of fluid and dissolved ions through toxin-mediated changes to the host intestinal tract, allowing V. cholerae to exit the host, re-enter a water source, and perpetuate its infectious cycle. In addition to the harsh conditions of the human gastrointestinal tract, V. cholerae must adapt to numerous stresses in the marine environment. These environmental stresses include temperature fluctuations, eukaryotic predation, and exposure to chemical insults like reactive oxygen species (ROS) (1).
As an aquatic organism, V. cholerae is exposed to varying concentrations of dissolved oxygen and ROS produced abiotically through photochemical reactions between sunlight and dissolved organic matter in the ocean (2, 3). ROS can also be produced biotically through metabolic processes in aerobic environments by phytoplankton, another potential reservoir of V. cholerae (4). In response to the multiple routes of exposure to ROS, it is not surprising that V. cholerae has multiple ROS defense systems including two paralogues of the oxidative stress responsive transcription factor OxyR, two catalases, and multiple peroxidases (5–7). Another mechanism to increase tolerance to ROS is the production of surface adhered communities encased in an exopolysaccharide matrix also known as biofilms.
Many bacterial species, including V. cholerae, have increased tolerance to ROS such as hydrogen peroxide (H2O2) when grown in biofilms compared to planktonic counterparts (8–10). Biofilm formation in V. cholerae is regulated by the bacterial second messenger molecule cyclic dimeric guanosine monophosphate (c-di-GMP), which is produced by diguanylate cyclase (DGC) enzymes. C-di-GMP alters bacterial physiology by modulating transcription, translation, and/or protein function (11). In V. cholerae, a common mechanism of c-di-GMP signaling is modulation of gene expression by three c-di-GMP dependent transcription factors: VpsR, VpsT, and FlrA (12–14). c-di-GMP activates the transcription factors VpsR and VpsT, resulting in increased transcription of genes involved in synthesis of the biofilm matrix component Vibrio polysaccharide (VPS) (12, 13, 15). In contrast, c-di-GMP acts as an anti-activator of FlrA, which causes decreased expression of genes necessary for flagellar biosynthesis (14). C-di-GMP can also bind to two riboswitches, Vc1 and Vc2. Binding of c-di-GMP to Vc1 functions as an ON-switch to induce production of the adhesin GbpA, while binding of c-di-GMP to Vc2 functions as an OFF-switch to inhibit production of the transcription factor TfoY (16, 17). As cells receive signals to disperse from the biofilm, phosphodiesterase (PDE) enzymes, which degrade c-di-GMP, become activated. These PDEs then deplete the intracellular concentration of c-di-GMP, promoting a switch from a sessile biofilm lifestyle to a motile one (reviewed in (11)).
While VpsR and VpsT were initially discovered as regulators of biofilm production, other c-di-GMP dependent functions have emerged. For example, c-di-GMP and VpsR transcriptionally regulate genes in the type II secretion operon as well as tfoY, a gene involved in driving dispersive motility and regulating type VI secretion (17–19). VpsT negatively regulates expression of genes involved in flagellar biosynthesis; however, the mechanism is not known (13). Additionally, Ayala et al. demonstrated VpsT negatively regulates the transcription of the stationary phase sigma factor RpoS (20). Recently, we found that c-di-GMP and VpsT induced expression of the DNA repair gene tag to promote survival after alkylation stress (21). These studies demonstrate c-di-GMP regulation extends beyond biofilm formation and motility in V. cholerae.
In this study, we uncovered an additional role for c-di-GMP: positively regulating catalase activity by increasing transcription of the catalase encoding gene katB via a VpsR and VpsT dependent mechanism. We further show that c-di-GMP dependent catalase activity was necessary for survival after exposure to the ROS H2O2. Our results expand the regulatory network of c-di-GMP to include antioxidant production, demonstrating that elevated c-di-GMP enhances the oxidative stress response in V. cholerae.
Materials and Methods
DNA manipulations and growth conditions
V. cholerae C6706 Str2 was used as the wild-type and the low biofilm forming derivative ΔvpsL was used as the Parent strain in the text (22, 23). All vectors were constructed by Gibson Assembly (NEB). Chromosomal deletion strains were constructed using the allele exchange vector pKAS32 digested with KpnI and SacI (NEB, High Fidelity). Luciferase reporters were constructed using the luciferase reporter vector pBBRlux digested with BamHI and SpeI (NEB). Expression vectors for VpsT and VpsR were constructed by removing the ribosome binding site (RBS), green fluorescent protein, and chloramphenicol acetyltransferase from pEVS143 with BamHI and EcoRI digests (NEB) (23). The VpsT purification vector was constructed elsewhere (21). Expression vectors for katB and katG were constructed as follows: pHERD20T was amplified with primers flanking the ampicillin resistance gene using inverse PCR resulting in a linear PCR fragment lacking the ampicillin resistance gene (24). The chloramphenicol resistance gene from pBBRlux was amplified by PCR and the two linear fragments were circularized by Gibson Assembly. Plasmids were introduced into V. cholerae through biparental conjugation using Escherichia coli S17 as the donor strain. V. cholerae harboring the plasmid of interest was selected for using Polymixin B (10 U/mL) with the relevant antibiotic. Antibiotics and reagents were used at the following concentrations unless otherwise stated: Ampicillin (100 μg/mL), Kanamycin (100 μg/mL), chloramphenicol (10 μg/mL), and 100 μM isopropyl β-D-1-thiogalactopyranoside (IPTG). Cultures were grown in Lysogeny Broth (LB, Acumedia) at 35°C, 220 RPM unless otherwise stated.
Catalase assay
Measurement of catalase activity was adapted from (25) with the following changes: Overnight cultures were diluted to a starting OD600 of 0.04 in 5 mL LB in 18 × 150 mm borosilicate test tubes supplemented with necessary antibiotics and IPTG. Cultures were grown at 35°C with shaking at 220 RPM until the OD600 reached approximately 2.0, moved to 15 mL falcon tubes (Corning®), and were pelleted by centrifugation (4,000 × g for 3 minutes). Pellets were resuspended in 100 μL of sterile 1X PBS to create a viscous cell solution that were adjusted to an OD600 of 150 in 100 μL final volume in Pyrex test tubes (13 × 100 mm, borosilicate). 200 μL of catalase reaction buffer (1% Triton-X100, 15% hydrogen peroxide in 1X PBS) was added to the test tubes and the solution was mixed using disposable 10 μL loops (BD Difco). Tubes were incubated at room temperature until gas production subsided (approximately 5-10 minutes). A standard curve was generated by mixing purified bovine catalase (Sigma, 570 U/μL) diluted in 1X PBS with 200 μL of catalase reaction buffer. At 10 minutes of incubation, images of the tubes were taken with an iPad Air (iOS 12.1.4) and the height from the bottom of the tube to the top of the foam were measured in both the standards and samples using the software ImageJ and an internal 1-inch reference mark for each picture. GraphPad Prism was used to generate the standard curve and interpolate the sample catalase activity using linear regression. Data presented is catalase activity (Units) normalized to cell number (OD600).
RNA Isolation and quantitative real-time PCR (qPCR)
Three biological replicate overnight cultures were diluted to a starting OD600 of .04 in 2 mL LB supplemented with ampicillin and IPTG and grown until an OD600 of approximately 1.0 at 35°C and 220 RPM. 1 mL of each replicate was pelleted and RNA was extracted using the TRIzol® reagent following the directions in the manual (Thermo Fischer Scientific). Purified DNA was quantified using a NanoDrop spectrophotometer (Thermo Fischer Scientific). 5 μg of purified RNA was treated with DNAse (Turbo DNAse, Thermo Fischer Scientific). cDNA synthesis was carried out using the GoScript™ Reverse Transcription kit (Promega). cDNA was diluted 1:30 into molecular biology grade water and used as template in qRT-PCR reactions using SYBR Green (Applied Biosystems™) as the method of detection. Reactions consisted of 5 μL 2.5 μM primer 1, 5 μL of 2.5 μM primer 2, 5 μL of diluted cDNA template, and 15 μL of 2X SYBR green (see Table S4 for primer sequences). Each plate had technical duplicates and biological triplicate samples as well as no reverse transcriptase controls to check for genomic DNA contamination. The StepOnePlus Real Time PCR system was used for qRT-PCR with the following thermocycling conditions: 95°C for 20 seconds then 40 cycles of 95°C for 2 seconds and 60°C for 30 seconds. Melting curves were included to ensure PCR products had single amplicons and primer dimers were absent. Data was analyzed by the ΔΔCt method using gyrA as a housekeeping or reference target.
Luciferase reporter assays
A) V. cholerae reporter assays
Overnight cultures of V. cholerae harboring katB transcriptional fusions to luciferase in pBBRlux were diluted 1:100 in 1 mL LB supplemented with ampicillin, chloramphenicol, and IPTG. 200 μL of cell solution was aliquoted into wells of a black 96-well plates (Costar). Plates were incubated at 35°C with shaking at 220 RPM until the OD600 reached approximately 0.25, and luciferase activity was measured using an Envision plate reader (Perkin Elmer). Luciferase activity (RLU) was normalized for cell number by dividing RLU by the OD600 at the time of the reading (Normalized Luminescence). For experiments where H2O2 were added to the cultures, overnight cultures of V. cholerae were diluted as described above except H2O2 was added to cultures at a final concentration of 50 μM when OD600 values reached approximately 0.225, followed by shaking at 35°C for an additional 30 minutes before measuring luciferase and OD600.
B) E. coli DH10b luciferase assays
Overnight cultures of E. coli DH10b harboring vectors to modulate transcription factor, c-di-GMP production, and the luciferase reporter were diluted 1:100 as described above and grown at 35°C, 220 RPM until the OD600 reached 0.45. Luciferase activity was measured and normalized to the OD600 to yield normalized luminescence.
Protein purification and electrophoretic mobility shift assays (EMSA)
C-terminal HIS-tagged VpsT purification and EMSAs experiments using the FAM labeled katB promoter region from katB2 were carried out as previously described (21). For purification, an overnight culture of E. coli BL21 harboring the pET28b-VpsT expression construct was diluted 1:100 into 250 mL LB supplemented with kanamycin in a 1 L flask. The culture was grown until an OD600 of approximately 0.7 at which point 1 mM IPTG was added and the culture conditions were shifted to 16°C with shaking at 160 RPM for 16 hours to induce protein production. Protein purification was carried out by standard Ni-NTA resin purification protocols (19). For EMSA experiments, varying concentrations of purified HIS-tagged VpsT (0 – 600 nM) were incubated with 2.5 nM FAM-labeled katB probe in VpsT buffer (25 mM Tris-Cl, 150 mM NaCl, 5 mM β-mercapthol, pH – 7.5) at 30°C for 30 minutes. The binding reaction was loaded into pre-run 5% non-denaturing TBE gels and gel electrophoresis was done by applying 90 volts for 90 minutes at 4°C. Fluorescent detection and images of the gels and were taken using a Typhoon FLA 9000 imager and the requisite software (GE Healthcare Life Sciences).
Hydrogen peroxide survival assay
Overnight cultures were diluted to a starting OD600 of 0.04 in 1 mL LB supplemented with necessary antibiotics and IPTG in 1.5 mL microcentrifuge tubes. 140 μL aliquots were added to a 96-well plate (Costar) and grown until an OD600 of 0.3. H2O2 solutions were made from fresh H2O2 stocks in light-impermeable microcentrifuge tubes and sterile 1X PBS. At time 0, 10 μL of H2O2 was added to the cell solution and growth was monitored over time by measuring OD600.
Measurement of intracellular c-di-GMP
Overnight cultures of ΔvpsL harboring pBRP1 (QrgBMut) were diluted 1:100 in 2 mL LB ampicillin in 18 × 150 mm borosilicate test tubes and were grown to an OD600 of 1.0. The cultures were split into two 1 mL aliquots in microcentrifuge tubes, and H2O2 was added to one aliquot at a final concentration of 500 μM. An equal volume of water was added to the other aliquot as the untreated control. Cultures were incubated statically at room temperature for 30 minutes and collected for total protein quantification and nucleotide extraction. Briefly, 100 μL of culture was removed from each tube to quantify total protein, pelleted by centrifugation at full speed (15,000 × g) for 1 minute, resuspended in 100 μL 1X PBS with 10% sodium dodecyl sulfate (SDS), and boiled at 95°C for 10 minutes. Lysed cell solutions were centrifuged at 15,000 × g for 1 minute and the supernatant was removed and placed in new tubes. Total protein was quantified using the DC Protein Assay (Bio-Rad) following the instructions in the manual. Protein standards consisting of bovine serum albumin (provided in DC Protein Assay) were used to generate a standard curve to interpolate unknown sample concentrations. Nucleotide extractions were carried out following the protocol here (26) with the following changes. 900 μL of the remaining culture were pelleted at 15,000 × g for 1 minute in a benchtop microcentrifuge. The supernatants were removed, and the remaining pellets were resuspended in 100 μL nucleotide extraction buffer (40:40:20 methanol/acetonitrile/water with 0.1 N formic acid). The extraction solution was incubated at 20°C for 20 minutes and pelleted for 10 minutes at 15,000 × g. The supernatants were placed into new microcentrifuge tubes, and the solutions were dried overnight using a heated, vacuum centrifuge (SpeedVac Concentrator, Savant). The resulting dried pellets were resuspended in 100 μL HPLC-grade water and subjected to mass spectrometry analysis for quantification of c-di-GMP (27). Data is represented as pmol of c-di-GMP normalized by total cellular protein (mg).
Statistical Analysis
Data are represented as the mean ± SD. Statistical analyses (details in figure legends) were calculated with GraphPad Prism Ver. 6 (GraphPad, San Diego, CA). A p-value of < 0.05 was considered statistically significant.
Results
C-di-GMP Positively Regulates Catalase Activity
We have shown that c-di-GMP regulates genes involved in DNA repair and that this regulation increased tolerance to the methylating agent methyl methanesulfonate (MMS) (21). Thus, we hypothesized c-di-GMP had a role in mitigating other forms of cellular stress besides DNA methylation damage. We chose to test if c-di-GMP increased H2O2 tolerance because it is a common ROS produced by aerobic microorganisms as a byproduct of cellular respiration, and H2O2 is found in high concentrations in marine environments (reviewed in (28)). To test the effects of c-di-GMP on H2O2 tolerance in V. cholerae, we expressed the Vibrio harveyi DGC QrgB in a strain unable to form mature biofilms (ΔvpsL, designated as the parent strain for this text). We used this strain because expressing QrgB in WT V. cholerae induces biofilm production, which would introduce a confounding variable in our experiments (23). As a control, we expressed an inactive allele of QrgB (QrgBMut) which is unable to synthesize c-di-GMP. Immediately after the addition of H2O2, both cultures began to produce gas bubbles, which is a phenomenon indicative of catalase activity as H2O2 is degraded into water and gaseous oxygen. Interestingly, the culture with higher intracellular c-di-GMP exhibited a higher amount of gas production despite the cultures having similar numbers of bacteria, suggesting c-di-GMP increased catalase activity. Quantification of catalase activity revealed an approximate 5-fold increase when comparing strains expressing QrgB to QrgBMut (Figure 1), indicating c-di-GMP increased catalase activity. Importantly, we have previously demonstrated that the concentration of c-di-GMP generated by QrgB overexpression is similar to that seen naturally in the low-cell-density quorum sensing state, demonstrating that these results are physiologically relevant (17).
c-di-GMP can directly modulate protein activity or change gene expression through allosteric interactions with c-di-GMP-dependent transcription factors or riboswitches (13–15). The c-di-GMP dependent transcription factors VpsT and VpsR induce transcription of genes involved in biofilm formation, protein secretion, and DNA repair in high c-di-GMP conditions (12, 13, 17, 18, 21). Therefore, we hypothesized that these transcription factors control c-di-GMP regulated catalase activity. To test this hypothesis, we repeated the catalase assay using ΔvpsT, ΔvpsR, and the ΔvpsTΔvpsR V. cholerae mutants and observed a loss of c-di-GMP mediated induction of catalase activity, suggesting the increased catalase activity was part of the VpsR/VpsT/c-di-GMP regulatory network (Figure 1).
V. cholerae encodes two catalases: katG (VC1560), a bifunctional enzyme with both catalase and peroxidase functions, and katB (VC1585), which only exhibits catalase activity (7, 29). Mutants of these enzymes render V. cholerae more susceptible to H2O2 treatment; however, regulation of either katB or katG by c-di-GMP has not been described (29). We therefore measured c-di-GMP induction of catalase activity in the mutants ΔkatB, ΔkatG, or ΔkatBΔkatG. We observed that the ΔkatB background did not display c-di-GMP regulated catalase activity but did possess basal level catalase activity similar to the parent strain expressing QrgBMut (Figure 1). Expression of katB from an arabinose-inducible, multicopy plasmid in the ΔkatB background complemented catalase activity regardless of the levels of c-di-GMP (Figure S1). Strains lacking katG were still able to induce catalase activity by c-di-GMP; however, the level of catalase activity induced by c-di-GMP was approximately 20% lower than that of the parent strain (Figure 1). Additionally, catalase activity in the ΔkatG strain expressing QrgBMut was approximately 2-fold lower than the parent strain in the same condition (Figure 1). Strains lacking both katG and katB did not have measurable catalase activity under these conditions, which is expected because these are the only annotated genes encoding catalase activity in the El Tor V. cholerae reference genome N16961 (7) (Figure 1).
Characterization of the katB Promoter and regulation of katB expression
Our results suggest katB, but not katG, is positively regulated by c-di-GMP at the transcriptional level. We addressed this hypothesis by measuring katB mRNA from cultures inducing QrgB and QrgBMut using qRT-PCR. We found that katB expression increased approximately 12-fold in strains expressing QrgB relative to QrgBMut (Figure 2).
We next were interested in the promoter architecture of the region upstream of katB and hypothesized specific regions in the katB promoter were necessary for c-di-GMP mediated transcriptional control. We tested this hypothesis by measuring luciferase activity from a series of katB transcriptional reporters that have been truncated at the 5’ end in the presence of QrgB or QrgBMut expression (Figure 3A). In the full-length promoter construct katB1, QrgB induced katB expression 5-fold compared to strains over-producing QrgBMut (Figure 3A). Inclusion of 212 bp upstream of katB was sufficient to maintain c-di-GMP induction (Figure 3A, katB2 and katB3). However, if the promoter was truncated to include 172 bp upstream of katB (katB4), c-di-GMP dependent induction of katB was abrogated. This result suggested the necessary cis-acting sequences for c-di-GMP mediated activation of katB expression are found between −172 and −212 bp upstream of the katB relative to the ATG start codon (Figure 3A). Deleting 30 bps from katB4 resulted in expression levels similar to that of the promoter-less vector control regardless of c-di-GMP concentrations (katB5), suggesting components necessary for the basal expression of katB are between −172 and −142 relative to the ATG start codon (Figure 3A). We also measured the effect of c-di-GMP on a katG-luciferase transcriptional fusion and did not observe significant c-di-GMP dependent changes in expression, in agreement with our initial hypothesis (Figure 3B).
VpsT and VpsR are necessary for the c-di-GMP-dependent induction of catalase activity, suggesting that these transcription factors activate transcription of katB at high intracellular concentrations of c-di-GMP (Figure 1). Thus, we hypothesized that c-di-GMP mediated induction of katB would be lost in strains lacking vpsT, vpsR, or both vpsT and vpsR. We tested this hypothesis by measuring katB2 reporter activity under different c-di-GMP conditions in the parent, single, and double knockout strains. We found that katB2 expression increased 5-fold in the parent background. katB expression increased 2.5 to 3-fold in the ΔvpsT and ΔvpsR backgrounds, but the differences in expression between QrgBmut and QrgB were not statistically significant (Figure 4A). It was only in the double mutant ΔvpsTΔvpsR background that c-di-GMP mediated katB expression was completely lost (Figure 4A). These data suggest that both VpsR and VpsT are needed for full induction by c-di-GMP.
katB expression is up-regulated in response to hydrogen peroxide in V. cholerae through the transcriptional activator OxyR (6, 29). To determine if the same region necessary for c-di-GMP mediated regulation of katB was necessary for H2O2 induction of katB, we measured the ability of H2O2 to induce katB4, the promoter region that no longer responded to c-di-GMP (Figure 3B). We found that 50 μM H2O2 induced katB4 expression approximately 6-fold in the parent background (Figure 4B). Next, to determine if VpsT and VpsR contributed to the H2O2 inducible response of the katB4 promoter, we repeated the assay in a ΔvpsTΔvpsR background. In this background, H2O2 increased katB4 expression to the same extent as the parent, however the basal and induced expression level was 1.5-fold lower compared to the parent (Figure 4B).
In Mycobacterium smegmatis, H2O2 can act as a first messenger to promote c-di-GMP synthesis (30). Whether H2O2 acts as a first messenger to modulate c-di-GMP in V. cholerae has not been demonstrated. To test the hypothesis that H2O2 acts as a first messenger, we measured intracellular c-di-GMP in control and H2O2 treated cultures and found no differences in intracellular c-di-GMP (Figure S2). Together, these results indicate that katB transcription is induced by c-di-GMP and the oxidative stress response through distinct regulatory mechanisms and demonstrate that H2O2 does not alter global levels of intracellular c-di-GMP.
VpsT Induces katB Expression in a Heterologous Host and Binds to the katB Promoter In Vitro
To test whether VpsT or VpsR directly regulates katB, we used Escherichia coli as a heterologous host to measure katB expression at high versus low concentrations of c-di-GMP in the presence of either VpsR or VpsT. We reasoned the genetic dissimilarity between V. cholerae and E. coli would allow us to isolate the effects of VpsR and VpsT directly on the katB promoter without the effect of these transcription factors regulating each other’s expression (23, 31, 32). Expression of QrgB with an empty vector increased katB expression in E. coli approximately 2.5-fold when compared to expression of QrgBMut through an unknown mechanism (Figure 5). Similarly, co-expression of VpsR along with QrgB resulted in approximately the same fold change (2-fold) as the empty vector, indicating VpsR is not sufficient to induce katB expression when expressed alone in E. coli along with increased c-di-GMP (Figure 5). Expression of VpsT and QrgB, however, resulted in transcription from the katB promoter increasing 25-fold (Figure 5), suggesting that VpsT directly regulates katB transcription in response to c-di-GMP.
As a transcription factor, VpsT binds to DNA in the presence of c-di-GMP to modulate gene expression (13, 20, 21, 32). Since VpsT was necessary for c-di-GMP mediated katB expression and was able to induce katB expression when expressed in a heterologous host, we hypothesized VpsT directly interacted with its promoter in a c-di-GMP-dependent manner. Thus, we purified C-terminal HIS-tagged VpsT and measured its ability to bind to the katB2 promoter in vitro. VpsT only partially shifted the katB2 probe in the absence of c-di-GMP at the highest concentration tested (600 nM) (Figure 6, lane 5). However, with the addition of 50 μM c-di-GMP, VpsT was able to decrease the intensity of the unshifted band at 150 nM and completely shift the probe at 300 nM (Figure 6, lanes 10-12). Addition of a 100-fold molar excess unlabeled 20bp oligonucleotide, composed of the VpsT binding site found in the vpsL promoter, was able to outcompete VpsT binding to the labeled katB2 probe in the presence or absence of c-di-GMP (Figure 6, lanes 7, 14) (32). When the unlabeled competitor had transversion mutations introduced into the palindromic region, it was no longer able to abrogate the VpsT-katB2 band migration (Figure 6, lanes 6, 13) (32). Together, the in vivo and in vitro data suggest VpsT directly interacts with the katB promoter to stimulate expression under high c-di-GMP conditions.
C-di-GMP mediated HOOH survival is dependent on catalase
Since c-di-GMP increased katB expression and catalase activity, we hypothesized that high intracellular c-di-GMP would provide a survival advantage during H2O2 stress. We tested this hypothesis by measuring survival after H2O2 treatment in V. cholerae backgrounds (ΔvpsL) unable to make mature biofilms to specifically test if the transcription regulation of katB by c-di-GMP was responsible for any observed protection as opposed to matrix production or formation of multicellular biofilms (8, 10, 33, 34). V. cholerae expressing QrgBMut or QrgB were challenged with H2O2 and the gross culture viability (OD600) was monitored for three hours. We found that in the parent strain, expression of QrgB, but not QrgBMut, led to significant protection from H2O2 stress (Figure 7A). To test this if this protection is dependent on c-di-GMP mediated catalase activity, we measured H2O2 survival in the ΔkatB, ΔkatG, and ΔkatBΔkatG mutants. Consistent with our catalase activity results, deletion of katB in either the ΔkatB and ΔkatBΔkatG mutants displayed no c-di-GMP mediated H2O2 survival (Figure 7E and G) while survival of the ΔkatG mutant during H2O2 treatment resembled the parent strain (Figure 7F).
As vpsT and vpsR were necessary for c-di-GMP dependent induction of katB expression and catalase activity (Figure 1 and 4), we hypothesized deletion mutations of these transcription factors would decouple c-di-GMP signaling from survival during H2O2 treatment. Indeed, strains lacking vpsT, vpsR, or both vpsT and vpsR lost c-di-GMP-mediated survival during H2O2 treatment (Figure 7B-D). Taken together, these data suggest c-di-GMP increases katB expression and KatB catalase activity through the c-di-GMP dependent transcription factors vpsT and vpsR resulting in increased survival during H2O2 treatment (Figure 8).
Discussion
In this work, we sought to determine if c-di-GMP provides resistance to ROS in V. cholerae. Using a plasmid-based system to modulate intracellular c-di-GMP, we observed increased gas production after the addition of H2O2, suggesting c-di-GMP positively influenced catalase activity. We determined that katB was responsible for c-di-GMP regulated catalase activity and found that katB transcription was increased 5-fold in the parent, but this induction was lost in strains deficient for the c-di-GMP dependent transcription factors VpsT and VpsR. Measuring katB transcription in a heterologous host revealed that VpsT was sufficient to induce katB expression under high c-di-GMP conditions, and in vitro DNA binding assays demonstrated VpsT specifically bound to the katB promoter in a c-di-GMP dependent fashion. Lastly, we showed that c-di-GMP mediated survival after H2O2 treatment was dependent on vpsT, vpsR, and katB.
In agreement with our findings, other groups have observed an increase in katB expression under certain biofilm inducing conditions which modulate intracellular c-di-GMP, such as incubation with norspermidine (35). Additionally, transcriptomic data from experiments where c-di-GMP or VpsT was artificially induced in V. cholerae suggested katB expression was a positively regulated by c-di-GMP (13, 36). Interestingly, c-di-GMP and VpsT were shown to down-regulate production of the stationary phase sigma factor RpoS, in turn decreasing survival to various stressors including H2O2, which is in contrast to our results (Figure 6A) (20). These experiments were done with the El Tor biotype strain C7258, which is part of the serogroup Ogawa while the strain used in the current study (C6076 str2) is part of the serogroup Inaba. While there are no studies describing differences in transcriptional regulation between the two serogroups, we hypothesize there may be serogroup dependent differences in c-di-GMP signaling, which would explain the contrasting results.
Unlike the ΔvpsTΔvpsR double mutant, c-di-GMP was able to induce the express of the katB reporter in the single vpsT and vpsR deletion background, albeit the difference was not statistically significant. However, only production of VpsT and c-di-GMP, but not VpsR and c-di-GMP, increased expression of the katB promoter in E. coli (Figure 4). C-di-GMP mediated catalase activity and survival in H2O2 were also lost in both the ΔvpsT and ΔvpsR backgrounds. Thus, our evidence suggests that phenotypes are primarily controlled by VpsT (Figure 8). VpsR is required for c-di-GMP mediated induction of vpsT in V. cholerae, and thus we conclude it has an indirect effect on katB transcription (31) (Figure 8).
A predicted VpsT binding site was previously reported approximately 150 bps upstream of the ATG start codon (37). Our results indicate that even when the VpsT binding site was present in the katB4 transcriptional fusion, induction by c-di-GMP was lost. We note that the predicted VpsT sequence was found in the template strand while predicted and validated VpsT binding sites in the vpsL and vpsA promoters were found in the coding strand (32, 37). Whether this difference in strand binding alters VpsT transcriptional regulation warrants further investigation.
Although our work focuses on V. cholerae, the association between c-di-GMP and protection against ROS has been shown in other bacteria. In Listeria monocytogenes, deletion of genes encoding PDE domains have elevated biofilm production and H2O2 tolerance (9). However, it is not known whether H2O2 tolerance was caused by increased EPS production, antioxidant production, or both. In Mycobacterium smegmatis, a relative of the human pathogen M. tuberculosis, H2O2 stimulates production of intracellular c-di-GMP, which inactivates the transcriptional repressor HpoR. As a result, ROS defense genes are up-regulated and increase H2O2 tolerance (30). This differs from our work in two facets: first, H2O2 does not act as a first messenger that stimulates c-di-GMP activity in V. cholerae (Figure S2) and second, VpsT and c-di-GMP regulate a wide array of genes whereas HpoR-c-di-GMP has been shown to only regulate expression of the hpoR operon. Despite the differences in regulation, the connection between c-di-GMP and ROS tolerance is evident in bacteria from diverse phylogenetic backgrounds.
Acknowledgments
This material was supported by NIH grants GM109259, GM110444, and AI130554 awarded to CMW. We would like to thank the College of Natural Science and the Department of Microbiology and Molecular Genetics at Michigan State University for the University Enrichment Fellowship and the Rudolph Hugh and Bertina Wentwoth Fellowship awarded to NLF. We would also like to thank Geoffrey B. Severin and Brian Y. Hsueh for critically reading and providing comments on the manuscript.