Roles of the Sodium-Translocating NADH:Quinone Oxidoreductase (Na+-NQR) on Vibrio cholerae Metabolism, Motility and Osmotic Stress Resistance

The Na+ translocating NADH:quinone oxidoreductase (Na+-NQR) is a unique respiratory enzyme catalyzing the electron transfer from NADH to quinone coupled with the translocation of sodium ions across the membrane. Typically, Vibrio spp., including Vibrio cholerae, have this enzyme but lack the proton-pumping NADH:ubiquinone oxidoreductase (Complex I). Thus, Na+-NQR should significantly contribute to multiple aspects of V. cholerae physiology; however, no detailed characterization of this aspect has been reported so far. In this study, we broadly investigated the effects of loss of Na+-NQR on V. cholerae physiology by using Phenotype Microarray (Biolog), transcriptome and metabolomics analyses. We found that the V. cholerae ΔnqrA-F mutant showed multiple defects in metabolism detected by Phenotype Microarray. Transcriptome analysis revealed that the V. cholerae ΔnqrA-F mutant up-regulates 31 genes and down-regulates 55 genes in both early and mid-growth phases. The most up-regulated genes included the cadA and cadB genes, encoding a lysine decarboxylase and a lysine/cadaverine antiporter, respectively. Increased CadAB activity was further suggested by the metabolomics analysis. The down-regulated genes include sialic acid catabolism genes. Metabolomic analysis also suggested increased reductive pathway of TCA cycle and decreased purine metabolism in the V. cholerae ΔnqrA-F mutant. Lack of Na+-NQR did not affect any of the Na+ pumping-related phenotypes of V. cholerae suggesting that other secondary Na+ pump(s) can compensate for Na+ pumping activity of Na+-NQR. Overall, our study provides important insights into the contribution of Na+-NQR to V. cholerae physiology.


Introduction
Na + -translocating NADH:quinone oxidoreductases (Na + -NQR) are found in the respiratory chains of a number of marine and pathogenic bacteria (reviewed in [1,2]). In organisms that express Na + -NQR, this enzyme is the gateway for electrons into the respiratory chain. Na + -NQR accepts reducing equivalents from NADH and donates them to the quinone pool [3,4]. The energy from this redox reaction is used to pump sodium ions from the inner to the outer side of the membrane, thereby building a sodium motive force (SMF).
Besides Na + -NQR, two more types of respiratory NADHdehydrogenases have been reported, the proton-pumping NADH dehydrogenases (NDH-1 or Complex I) and NDH-2 [5]. Similar to Na + -NQR, NDH-1 utilizes energy from the redox reaction, but it pumps protons from the inner to the outer side of the membrane to build a proton motive force (PMF). Unlike Na + -NQR and NDH-1, NDH-2 has neither H + nor Na + efflux activities. A few pathogenic bacteria, such as Yersinia pestis and Pseudomonas aeruginosa, have all three types of NADH dehydrogenases, but most bacteria possess one or two of these NADH dehydrogenases [5,6].
Lack of the major NADH dehydrogenase activity usually significantly affect bacterial physiology. Escherichia coli has a single NDH-1 (encoded by nuoA-N) and NDH-2 (encoded by ndh), and it is known that NDH-2 is the major enzyme for aerobic growth, whereas the NDH-1 is essential for anaerobic fumarate and DMSO respiration [7]. Moreover, it is known that the E. coli strain lacking NDH-1 showed growth retardation and increased acetate secretion after the transition to stationary growth phase when grown in mixed amino acids media [8]. It was also shown that the E. coli strain lacking NDH-1 grew poorly when grown in minimal media supplemented with acetate as the sole carbon source [8].
Vibrio cholerae is the causative agent of cholera, a waterborne severe diarrheal disease. Genome sequencing revealed that V. cholerae does not have NDH-1 but has Na + -NQR (encoded by nqrA-F) and NDH-2 (encoded by ndh) [2,6]. Although Na + -NQR is not essential for V. cholerae growth [9], the DnqrA-F mutant strain of V. cholerae shows a marked growth defect when grown in LB [10], similar to the E. coli strain that lacks a functional NDH-1 [8]. In addition, the V. cholerae Dnqr A-F mutant strain shows multiple defects in carbon metabolism [10]. Furthermore, Na + -NQR also affects V. cholerae virulence gene expression possibly by affecting acetyl-CoA metabolism via the TCA cycle [9,10] and survival in vivo [11].
In this study, we aimed to investigate the effects of a DnqrA-F mutation on broad aspects of V. cholerae physiology by using Phenotype Microarray (Biolog), transcriptome, and metabolomics analyses. We found that DnqrA-F mutant strain showed multiple defects in metabolism but did not show any defect in Na + pumping-related phenotypes of V. cholerae.

Materials and Methods
Bacterial strains and growth conditions V. cholerae O395N1, V. cholerae O395N1 DnqrA-F [12] and V. cholerae O395N1 DpomAB [13] strains were used in this study. Bacterial strains were kept at 280uC in 20% glycerol stocks. Bacterial strains were grown in Luria-Bertani (LB) medium at 30uC or 37uC. Streptomycin was supplemented at 100 mg/ml. When necessary, 33 mM L-lactate was used.

Phenotype microarray (PM) analyses
Phenotype microarrays were performed using PM3 to PM10 MicroPlate (Biolog) at Biolog's PM Services group. All plates used pyruvate as sole carbon sources.

DNA microarray analyses
Cells of V. cholerae O395N1 and V. cholerae O395N1 DnqrA-F strains, grown in LB (pH 6.5) at 30uC for 4 hours and 8 hours, were treated with RNA Protect Bacteria Reagent (Qiagen). RNA was extracted using the QIAGEN RNeasy Mini Kit (Qiagen). RNA was concentrated using the QIAGEN RNA MiniElute Cleanup kit (Qiagen) and sent to the Center for Genome Research and Biocomputing at Oregon State University, Corvallis, OR, following protocols outlined in the NimbleGen array user's guide, version 5 for cDNA synthesis and microarray analysis using NimbleGen Microarray plates and reagents. The microarray data are accessible at the Gene Expression Omnibus (accession number, GSE56387).

Metabolite extraction
V. cholerae O395N1 and V. cholerae O395N1 DnqrA-F strains were grown in LB (pH 6.5) at 30uC for 2 hours and the supernatant was removed by centrifugation (8000 rpm, 4 uC, 5 min). The pellet was resuspended in 600 mL of LC-MS-grade water in 1 mM HEPES and 1 mM EDTA (pH 7.2). Metabolites were extracted from bacterial cells by using the boiling water method as previously described [14].

LC-MS/MS
Mass spectrometry based metabolomic profiling was performed as previously described [15]. Briefly, liquid chromatography (LC) was performed on a Shimadzu Nexera system and metabolites separated on an Inertsil phenyl-3 stationary phase (GL Sciences, 5 uM, 4.6 6 150 mm). Mass spectrometry was performed on an AB SCIEX Triple TOF 5600 quadrupole-time-of-flight mass spectrometer. MS/MS spectra were gathered on the fly by information dependent acquisition. Most metabolites were identified by mass, isotope distribution, MS/MS fragmentation, and when standards were available, retention time. In the absence of chemical standards, MS/MS spectra were compared to those in the METLIN online database.
To account for analytical and sample preparation variation, samples were normalized to total ion count. Central energy metabolites (metabolites of the tricarboxylic acid cycle, pentose phosphate pathway, and glycolysis and amino acids) were targeted post-data acquisition and in addition, untargeted statistical analysis (Student's t-test p-value plotted against fold-change) revealed large (.10-fold) changes in cyclic AMP and cyclic GMP, and subsequently, related metabolites (purines and pyrimidines) were targeted post-data acquisition.

Growth measurement
Growth measurement assays were performed as previously described [16]. Briefly, overnight grown bacterial cells were inoculated into 200 ml liquid medium in 96-deep-well plates (Whatman) at an initial OD600 of 0.05 and grown at 37 uC for 18 h with vigorous aeration. LBB medium (non-cationic L broth) was supplemented with streptomycin and varying concentrations of ethylene glycol. The initial pH of the media was adjusted with HCl to 6.5 and buffered by the addition of 60 mM BTP. Growth was then measured as OD600 by scanning the plates on a Bio-Rad iMark microplate absorbance reader. All experiments were repeated at least three times in triplicate.

Measurement of acetate excretion
Acetate excretions from bacterial cells were measured as previously described [10].

Motility assay
Motility of V. cholerae strains was determined on soft agar plates buffered with 100 mM Bis/Tris (pH 6.5) containing 100 mM NaCl, 0.25% agar, 1% tryptone, 0.5% yeast extract and 50 mg ml 21 streptomycin. The endogenous Na + content of the medium was 11 mM as determined by atomic absorption spectroscopy [17]. If indicated, D,L-lactate (33 mM final concentration) was added from a stock solution adjusted to pH 6.5 with KOH. To inoculate the swarming plates, overnight cultures of the DnqrA-F mutant strain, its parent strain or the non-motile V. cholerae O395-N1 DpomAB [13] were diluted in LB medium to a starting OD600 of 0.01. Cells were grown until OD600 of 0.5-0.6 was reached and washed in fresh LB medium. Cell suspensions were adjusted to OD600 of 0.5 with LB, and aliquots of 1.5 ml were spotted on dried soft agar plates. The diameters of swarming diameters were determined after 16 h at 37uC. Mean values and standard error from 16 experiments are presented. P values were calculated using Student's t test.

Results and Discussion
The DnqrA-F mutant strain showed multiple defects in metabolism detected by the phenotype microarray We had previously performed a Phenotype Microarray (Biolog) analysis to address how a DnqrA-F mutation affects V. cholerae metabolism and changes in metabolic profiles were determined for the V. cholerae parent strain and the DnqrA-F mutant strain [10]. We had found that the DnqrA-F mutant has defects in the utilization of several TCA cycle intermediates (fumaric acid, succinic acid and L-malic acid) and many carbon sources that are metabolized into TCA cycle intermediates (L-aspartic acid, L-histidine, L-glutamine and L-glutamic acid). However, because succinic acid was used as the carbon source in the PM3-PM8 plates (nitrogen source, phosphorus source, sulfur sources and nutrient supplements testing plates), the systematic metabolic defects observed in the DnqrA-F mutant and effects of DnqrA-F mutation on these metabolisms were still uncertain. Our previous Phenotype Microarray results indicated that the DnqrA-F mutant utilizes pyruvate as the carbon source similar to the parent strain. Consistent with this, the DnqrA-F mutant grew similar to the parent strain when grown in M9 minimal media supplemented with pyruvate as the sole carbon source (data not shown). Thus, we tested the PM3-PM8 again using pyruvate as a sole carbon source. The results of the phenotype microarray indicated that the V. cholerae O395N1 DnqrA-F mutant strain still showed multiple defects in nitrogen, phosphate and sulfur utilization (Table S1). Using PM 9 and PM 10 plates, we also tested the effects of osmolytes and pH and found that the DnqrA-F mutant is sensitive to sodium chloride, sodium sulfate, ethylene glycol, and urea (Table S1). These data again confirmed that lack of Na + -NQR broadly affects V. cholerae metabolism.

Transcriptome analyses
We next performed DNA microarray analyses to investigate how the lack of Na + -NQR broadly affects gene expression pattern in V. cholerae. We had previously found that the DnqrA-F mutant increased virulence gene expressions, including toxT, ctxB and tcpA, but only at the early phase of the logarithmic growth [18]. This finding suggested that the lack of Na + -NQR affected V. cholerae gene expressions differently at different phases of growth. Because we aimed to understand the overall changes of gene expression pattern in the DnqrA-F mutant strain, we performed microarray analyses using RNA prepared from bacterial cultures both at the early and late phases of the logarithmic growth. At the early logarithmic growth, 612 genes were up-regulated and 660 genes were down-regulated in the DnqrA-F mutant compared to the parent strain, whereas at the late logarithmic growth phase, 119 genes were up-regulated and 264 genes were down-regulated in the DnqrA-F mutant (data not shown). Consistent with our previous  data, virulence genes, including toxT, tcpA-F and ctxB were upregulated in the DnqrA-F mutant at the early logarithmic growth phase but not at the late logarithmic growth phase (data not shown). Among these changes, 31 genes were commonly upregulated and 55 genes were commonly down-regulated in both growth phases (Table 1 and 2).
The cadBA genes showed the highest increases in gene expression levels in the DnqrA-F mutant in both growth phases ( Table 1). The cadA gene encodes a lysine decarboxylase and cadB encodes a lysine/cadaverine antiporter. The expression of cadBA is regulated by a ToxR-type transcriptional regulator, CadC [19], and a LysR type transcriptional regulator, AphB [20]. However, expression of the other AphB regulated genes, such as tcpP and nhaB [20], were not affected by the DnqrA-F mutation. Thus, CadC might be responsible for the induced expression of cadAB in response to lack of Na + -NQR. CadC is known to induce cadAB expression in response to acidic pH [19] and we previously found that the V. cholerae DnqrA-F mutant showed increased acetate production and caused acidification of the external medium [10]. Thus, it is likely that cadAB expression was induced by the acidic pH that results when Na + -NQR function is impaired.
Several transporter genes including tripartite ATP-independent periplasmic (TRAP) transporters, siaPQM and dctMQP, were down-regulated in the DnqrA-F mutant strain compared to the parent strain (Table 2). It was recently reported that SiaPQM is a Na + -dependent sialic acid-specific TRAP transporter [21,22]. Consistent with this, multiple genes encoding enzymes in the sialic acid utilization pathway, which convert sialic acid to fructose 6phosphate, and neuraminidase, which convert host cell surface polysialogangliosides to GM1 monoganglioside and release sialic acid [23], were also down-regulated in the DnqrA-F mutant ( Table 2). Given that Na + -NQR and sialic acid catabolic pathways are essential for V. cholerae colonization in the small intestine of mice [11,24], such decreased expression of genes in the sialic acid utilization pathway might explain why the DnqrA-F mutant showed defects in colonization in the small intestine of mice. The dctMQP genes were recently shown to encode a C4-dicarboxylate-specific TRAP transporter and to be partly responsible for V. cholerae C4dicarboxylates, succinate, malate and fumarate, utilization [21]. Thus, it might be possible that the decreased utilization of succinate, malate and fumarate by the V. cholerae DnqrA-F mutant [10] was simply caused by the decreased uptakes of these C4dicarboxylates. Further studies to better investigate the intriguing links between gene regulation events in response to loss of Na + -NQR and changes in V. cholerae carbon utilization are required to better understand these observations.
We also found that two methyl-accepting chemotaxis proteins (MCPs), encoded by VCA0176 and VCA0864, were downregulated in the V. cholerae DnqrA-F mutant compared to the parent strain (Table 2). Interestingly, we found that a VCA0864 mutant strain of V. cholerae C6706 showed a hypermotile phenotype in minimal media supplemented with N-acetylglucosamine chemotaxis plate (unpublished data). Together with the fact that the sialic acid degradation pathway and transport system were downregulated in the V. cholerae DnqrA-F mutant, these data suggest that the DnqrA-F mutant might be sialic acids starved.
The Class II, III and IV flagellar genes were systematically down regulated in the DnqrA-F mutant compared to the parent strain (Table 2). One might speculate that the decreased amount of regulatory and structural components of the polar flagellum might result in diminished motility of the DnqrA-F mutant strain. Indeed, the diameter of its swarming rings on soft agar plates (pH 6.5) reached only 80% of the diameter observed with the parent strain, but was clearly motile when compared with the non-motile DpomAB V. cholerae strain lacking essential components of the flagellar stator.
When lactate was added, both parent and mutant strain exhibited diminished swarming, and the improved motility of the parent strain was no longer apparent (p = 1, Fig.1). This indicates that depending on the external medium, the mutant strain suffers from a limitation in energy supply for flagellar rotation, resulting in diminished motility.

Metabolomic analyses
To further understand how lack of Na + -NQR affects V. cholerae metabolism, we performed a mass spectrometry based metabolomic profiling. We found that 12 metabolites were significantly decreased and 3 metabolites were increased in the V. cholerae DnqrA-F mutant compared to the parent strain (Table 3). We had previously suggested that TCA cycle activity is decreased in the V. cholerae DnqrA-F mutant [10]. Consistent with this, intracellular isocitrate levels were decreased in the DnqrA-F mutant (Table 3). However, we unexpectedly found that intracellular malate levels were significantly increased in the DnqrA-F mutant (Table 3). We also detected slightly increased levels of succinate in the DnqrA-F mutant (data not shown). Taken together, this may suggest that the reductive pathway of the TCA cycle is more active in the V. cholerae DnqrA-F mutant (Fig. 2).
We found that intracellular lysine levels were decreased, while cadaverine levels were increased in the V. cholerae DnqrA-F mutant compared to the parent strain (Table 3). Since CadA catalyzes the conversion of lysine to cadaverine, these changes could be caused by the increased cadA expression that we detected in the transcriptome analyses.
The acetate switch is broken in the V. cholerae DnqrA-F mutant We had previously shown that the V. cholerae DnqrA-F mutant showed increased acetate production compared to the parent strain [10]. When grown in tryptone-based rich media, E. coli first produces acetate by using the PTA-ACK pathway but at a certain point in its growth phase, it shifts to utilize the excreted acetate by using the AMP forming acetyl-CoA synthetase (AMP-ACS). This transition from acetate production to acetate utilization is called the acetate switch [25]. To further examine acetate production in the V. cholerae DnqrA-F mutant, we measured a time course of acetate excretion. Between 6 hr and 8 hr of growth in LB media, the V. cholerae parent strain started to reduce external acetate production, indicating that V. cholerae also has an acetate switch (Fig. 2). Interestingly, the V. cholerae DnqrA-F mutant did not show this phenotype and kept excreting acetate until the end of growth (Fig. 3). This is consistent with the Phenotype Microarray data that showed that the V. cholerae DnqrA-F mutant had a defect in acetate utilization (Table S1). The AMP-ACS protein is known to be regulated by protein acetylation with the acetylated form of AMP-ACS being inactive and these defects in acetate utilization in the DnqrA-F mutant suggested that AMP-ACS might be acetylated.
The DnqrA-F mutation does not affect osmotic resistance in V. cholerae The results of the phenotype microarrays showed that the V. cholerae DnqrA-F mutant strain was sensitive to osmotic stress including NaCl and ethylene glycol (Table S1). Thus, we performed growth assays in LBB-based medium (noncationic L broth), containing increasing concentrations of NaCl and ethylene glycol. Consistent with the phenotype microarrays, the V. cholerae DnqrA-F mutant showed growth defect compared with the parent strain but the growth defect was independent to the increasing concentrations of NaCl (Quinn et. al. submitted) or ethylene glycol (data not shown). To further determine whether the effect of Na + -NQR on growth was primarily related to loss of quinone reduction or more general Na + homeostasis, L-lactate was added to the growth media, which is expected to allow the L-lactate dehydrogenase to replenish the quinone pool directly. Interestingly, addition of L-lactate restored the growth of the V. cholerae DnqrA-F mutant to the parent level even in the presence of high concentration of NaCl (Quinn et.al. submitted) or ethylene glycol (data not shown). These data indicated that lack of Na + -NQR does not directly affect osmotic stress resistance in V. cholerae. Our transcriptome analyses detected the increased expression of a major Na + /H + antiporter gene, nhaA in the V. cholerae DnqrA-F mutant ( Table 1), suggesting that NhaA might complement the Na + pumping activity of Na + -NQR.

Conclusions
Lack of Na + -NQR broadly affects V. cholerae physiology but mainly affects its central metabolism but not Na + pumping-related phenotypes.   cholerae O395N1 and V. cholerae O395N1 DnqrA-F strains were inoculated into LB (pH 6.5) at 30uC. Acetic acids levels in the medium were measured using the acetic acid enzymatic assay kit (R-Biopharm). doi:10.1371/journal.pone.0097083.g003