Burkholderia pseudomallei, a pathogenic gram-negative bacterium, causes the severe human disease melioidosis. This organism can survive in eukaryotic host cells by escaping reactive oxygen species via the regulation of stress responsive sigma factors, including RpoS. In B. pseudomallei, RpoS has been reported to play a role in the oxidative stress response through enhanced activity of OxyR and catalase. In this study, the RpoS dependent oxidative stress responsive system was further characterized using comparative proteomic analysis. The proteomic profiles of wild-type B. pseudomallei following exposure to H2O2 and between wild-type and the rpoS mutant strains were analyzed. Using stringent criteria, 13 oxidative responsive proteins, eight of which are regulated by RpoS, were identified with high confidence. It was observed that ScoA, a subunit of the SCOT enzyme not previously shown to be involved directly in the oxidative stress response, is significantly down-regulated after hydrogen peroxide treatment. ScoA and ScoB have been predicted to be organized in a single operon using computational methods: in this study it was confirmed by RT-PCR that these genes are indeed co-transcribed as a single mRNA. The present study is the first to report a role for RpoS in the down-regulation of SCOT expression in response to oxidative stress in B. pseudomallei.
two dimensional electrophoresis
database of prokaryotic operons
- E. coli
electron transport chain
flavin adenine dinucleotide
nicotinamide adenine dinucleotide
peptide mass fingerprinting
reactive oxygen species
quantitative reverse transcription polymerase chain reaction
succinyl-coA:3-ketoacid-coenzyme A transferase
Burkholderia pseudomallei is the causative agent of melioidosis, a deadly infectious disease that is endemic in areas of South-East Asia and northern Australia. It is also a potential bioterror agent . This bacterium can survive in almost all eukaryotic cell types, particularly phagocytic cells [2, 3]. Most eukaryotic cells can produce ROS, such as H2O2, O2− and hydroxyl radical, that promote oxidative damage to intracellular pathogens, leading to their elimination. However, many pathogenic bacteria can escape the effects of these ROS by utilizing enzymatic anti-oxidants, including catalase and superoxide dismutase [4, 5]. For example, in B. pseudomallei the activity of KatG and KatE catalase enzymes is increased after exposure to H2O2. Expression of the KatG enzyme is indirectly controlled by RpoS through its regulation of OxyR, another transcriptional regulator that plays a key role in the response to oxidative stress in bacteria . RpoS also regulates other proteins involved in the response to oxidative stress . However, many of these RpoS-regulated proteins have not been fully characterized. Analysis of the genomic sequence of B. pseudomallei has facilitated the identification of genes encoding putative oxidative responsive proteins and genes involved in pathogenicity . In addition, proteomic reference maps of B. pseudomallei showing protein expression in stationary phase have revealed that RpoS controls many proteins involved in central metabolism, secondary metabolism and stress response [9, 10].
In prokaryotic genomes, some groups of related enzyme-coding genes tend to be co-localized and are often present in the same polycistronic unit, known as an operon. These groups of genes are transcribed together from one regulatory unit , transciption often following the order of the enzyme activities in the corresponding metabolic pathways . In proteomic analysis utilizing 2D-gels, a number of proteins cannot be detected due to low abundance, proteins out-of-range of the molecular weights or pI cutoff, as well as loss of proteins due to degradation in response to stress conditions, however, proteins absent in 2D-gels can be identified to be co-transcribed with other members of the operons that are detected by 2D-gel .
In this study, we took a global approach to genome-wild analysis of proteomic profiles of wide-type B. pseudomallei untreated and exposed to H2O2, and that of the rpoS mutant. We identified target proteins regulated by the RpoS sigma factor and elucidated dynamic regulation under oxidative stress conditions. In addition, we characterized the regulation by RpoS of ScoA, a subunit of SCOT. To the best of our knowledge this is the first report of a role for SCOT in the oxidative stress response. We also demonstrated experimentally that ScoA and ScoB are co-transcribed in an operon and function in SCOT activity. Our finding is the first to elucidate that RpoS down-regulates SCOT expression in response to oxidative stress in B. pseudomallei.
MATERIAL AND METHODS
Bacterial strains and hydrogen peroxide treatment
The clinical isolate B. pseudomallei strain 844  and the rpoS mutant strain  were grown in Luria Bertani medium with and without tetracycline (i.e., under normal conditions) at 37°C with shaking at 220 rpm as previously described [10, 13, 14]. For H2O2 treated conditions, cultures were grown for 12 hrs (A600 of about 1.0) corresponding to early stationary phase, after which 0.5 mM H2O2 was added to the culture medium every 10 mins for 1 hr before harvesting. Cultures treated with H2O2 or untreated controls were harvested for further use.
Protein extraction and two dimensional electrophoresis analysis
Proteins were extracted using 500 µL lysis buffer: 8 M urea, 4% (w/v) 3-[3-(cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), 2 mM tributyl phosphine, 1% (v/v) Immobilized pH gradient (IPG) buffer pH 4–7 (Amersham Biosciences, Uppsala, Sweden) and 1% v/v protease inhibitor cocktail set II (CalBiochem, La Jolla, CA, USA). Cell lysis was performed on ice using sonication, followed by the removal of cell debris by centrifugation at 13,000 g for 30 mins at 20°C. Lysates were stored at −80°C prior to use. Protein concentrations were determined with an RC DC protein assay kit (Bio-Rad, Hercules, CA, USA) and used for 2DE as previously described . Protein spots were visualized by Coomassie Brilliant Blue G-250 staining and gels were scanned with an ImageMaster Scanner (Amersham Biosciences). Image analysis was performed using PDQuest software version 7.1 (Bio-Rad). Images from three independent cultures were compared. The relative intensity of each protein spot was determined by normalizing to the sum of the intensity of all protein in the gel. Proteins with changes in expression with intensities representing at least a threefold difference for both up- and down-regulated and P < 0.05 were considered in this analysis.
Tryptic in-gel digestion and peptide mass fingerprinting
Protein spots were excised from 2D gels of the wild-type strain and transferred into 1.5 mL microcentrifuge tubes. In-gel digestion was carried out in 96-well microtiter plates using an Ettan Spot Handling workstation (Amersham Biosciences). Digestion was performed as previously described  with minor modifications. Peptides produced after digestion by trypsin (Promega, Madison, WI, USA), were collected by extraction with 30 µL of 20 mM ammonium bicarbonate and 30 µL of 50% acetonitrile with 5% formic acid. PMF was performed using a reflex IV MALDI-TOF mass spectrometer (Bruker Daltonik, Bremen, Germany) as previously described . The mass spectrometer was operated in the positive reflectron mode and covered the mass range between 800 and 3500 Da. The resulting PMF of each protein spot was visualized using FlexAnalysis software version 2.2 (Bruker Daltonik).
Protein identification by database searching
Identification of proteins from the tryptic fragments was performed by comparison of their PMFs against the subgroup of “other proteobacteria” in the NCBInr protein database using the MASCOT search engine (www.matrixscience.com). This subgroup was selected because there is no specific Burkholderia spp. database. Protein identification was based on the assumption that peptides are monoisotopic, with fixed modifications at cysteine residues and variable oxidation at methionine residues; one missed tryptic cleavage was allowed. Only identified proteins that matched B. pseudomallei, B. mallei or B. thailandensis were included in this analysis. Because the MALDI-TOF approach imposed limitations in that it was not able to precisely identify multiple proteins present in one spot of the 2D-gel , an attempt was made to maximize confidence in the identifications of proteins at each of these spots by using two stringent criteria for selecting the most likely proteins. Firstly, proteins in spots within 10% differential pIs and MWs between the experimental and the expected values in the database were considered. Secondly, when more than one protein spot linked to one particular peptide, only the best-matched spot with the highest scores from the peptide match using MASCOT analysis was selected. For consistency, in this experiment the same criteria were also utilized to re-analyze the RpoS-regulated data of Osiriphun et al. .
Operon prediction and microarray data prediction
The common names (under “synonym” tag in the database) of eight oxidative responsive genes regulated by RpoS were used as queries to search against groups of genes computationally predicted by DOOR (http://csbl1.bmb.uga.edu/OperonDB/) as operons . Furthermore, genes predicted to be in the same operon were checked to verify that they were co-transcribed according to the time-series microarray data (Gene Expression Omnibus, GEO, Series GSE5495, a genome-wide mapping of growth regulated genes in B. pseudomallei ). The expression data of all genes in all-time series were used to find pair-wise correlations (using Pearson's method). The correlations of experimentally verified operons  (namely BPSL1216 and BPSL1217, BPSS0182 and BPSS0183, BPSS1954 and BPSS1955 and BPSL1468 and BPSL1469) were used as positive controls. Random correlations were used as negative controls. These values were then used as criteria for determining the cut-offs for correlation of expression for genes that were likely to be present in operons.
RNA isolation and operon identification by quantitative reverse transcription polymerase chain reaction
Ribonucleic acid extraction was performed using TriZol reagent (Invitrogen, Carlsbad, CA, USA). One milliliter of 12 hr culture (OD600 of 1.5 approximately) of untreated and H2O2-treated wild type cells and the RpoS mutant were harvested and collected by centrifugation. RNA extraction was carried out following the manufacturer's instructions. RNA samples were treated with RQ1 RNase-free DNase (Promega) and tested for DNA contamination. Measurement of RNA concentration was performed using a Nanodrop 2000 (Thermo Fisher Scientific, Houston, TX, USA).
To determine if, as predicted, ScoA and ScoB are co-transcribed in an operon, RNA was incubated at 70°C for 5 mins with 20 pmol/µL of specific reverse primers as described below. Reverse transcription mixes and the steps of reverse transcription followed the manufacturer's instructions (Promega). The specific primers used in this study, which were designed to anneal pair-wise between two genes, scoA and scoB (BPSL1954-1955 forward primer 5′-AGTTCGACGGCAAGCATTAC-3′ and BPSL1954-1955 reversed primer 5′-ATGAGATCCGCGTCGACTTC-3′), have an expected product size of approximately 500 bp. PCR products were analyzed on 1.5% agarose gels stained with ethidium bromide.
Degree of scoA–scoB operon expression in differential stress conditions and regulon
To investigate whether scoA–scoB are regulated by the RpoS sigma factor in response to H2O2 stress, RNA from H2O2-treated and untreated wild-type and the RpoS mutant were extracted as describe above. The semi-quantitative band intensity of RT-PCR products in each condition were visualized using 1.5% agarose gels stained with ethidium bromide and the intensities compared using Gene Tools software (Syngene, Cambridge, UK). 16sRNA was used as an internal control for determining comparative degrees of gene expression.
Using the statistical package R version 2.13.0 (2011-04-13), Fisher's exact test was used to calculate P values for the overlap between RpoS-regulated proteins and the proteins responsive to H2O2.
Identification of oxidative stress responsive proteins in Burkholderia pseudomallei
A B. pseudomallei protein reference map was previously created by the present authors ; this detected 282 protein spots on the 2D-gels. In this study, the accuracy of protein identification was further improved by using more stringent criteria than previously employed (see Materials and Methods Section for more details). Firstly, only protein spots with a ≤ 10% differential percentage between experimental and theoretical values of pI and MW were selected for further analysis. Secondly, when several protein spots matched the MASCOT database, the one with the highest score was chosen. Thus, 282 reference proteins were the basis for a new B. pseudomallei protein reference map containing 102 high quality proteins (see Supplement Material 1). Using this updated reference map, oxidative responsive proteins were identified by comparing proteomic profiles of B. pseudomallei untreated and treated with H2O2, which was used as a source of oxidative stress (Fig. 1). Thirteen proteins with ≥ threefold differential degrees of expression are here reported as oxidative stress responsive proteins in B. pseudomallei (Table 1). Two of the thirteen proteins were up regulated when treated with H2O2 as compared to untreated cultures; they include pyridoxal phosphate biosynthetic protein (PdxJ; YP_109018) and a protein belonging to the universal stress protein family (YP_110056). Four of these thirteen proteins have previously been shown to be differentially expressed under oxidative stress in other bacterial species, namely non-ribosomally encoded peptide/polyketide synthase protein (YP_111196) [19, 20], cysteine synthase B (CysM; YP_109102) [21, 22], glutathione s-transferase (YP_108349) [23-25] and universal stress-related protein (YP_110056) [26, 27]. The other nine proteins, namely succinyl-CoA:3-ketoacid-coenzyme A transferase subunit A (ScoA; YP_108553), oxidoreductase (YP_109344), 3-methyl-2-oxobutanoate hydroxyl-methyltransferase (PanB; YP_109419), pyridoxal phosphate biosynthetic protein (PdxJ; YP_109018), flagella biosynthesis/type III secretory pathway ATPase (FliI; YP_106855), flagellum-specific ATP synthase, Burkholderia lethal factor 1 (BLF1; YP_108169; a former hypothetical protein BPSL1549)  and some proteins of unknown function (hypothetical proteins BPSS1486 [YP_111493], BPSS0683 [YP_110696] and BPSS2129 [YP_112128]), have not previously been studied for a direct role in responding to oxidative stress in bacteria. Some of the proteins with unknown function may be involved in protection from environmental stresses. Oxidoreductase is similar to antioxidant oxidoreductase EC. 18.104.22.168 of Ralstonia solanacearum, the function of which is to convert H2O2 to water; 3-methyl-2-oxobutanoate hydroxymethyltransferase (PanB; YP_109419) catalyzes the first committed step in the biosynthesis of pantothenate (vitamin B5) and its derivatives have activities that protect cells during oxidative stress [29-31]; and pyridoxine 5′-phosphate synthase (PdxJ; YP_109018), is involved in the synthesis of pyridoxine (vitamin B6) by a 1-deoxy-D-xylulose 5-phosphate-dependent pathway . Pyridoxine and its derivatives are capable of inhibiting ROS production and preventing lipid peroxidation [33, 34].
|Proteins||Spot no. ||Accession no.||Locus tag.||H2O2 ratio||RpoS promoter group ||Reference|
|Succinyl-CoA:3-ketoacid-coenzyme A transferase subunit A (ScoA)||62||YP_108553||BPSL1955||0.3||2||This study|
|Non-ribosomally encoded peptide/polyketide synthase||73||YP_111196||BPSS1183||0.2||1||[19, 20]|
|Amino acid metabolism|
|Cysteine synthase B (CysM)||80||YP_109102||BPSL2507||0.1||1|||
|3-methyl-2-oxobutanoate hydroxymethyltransferase (PanB)||100||YP_109419||BPSL2824||0.2||This study|
|Pyridoxal phosphate biosynthetic protein (PdxJ)||102||YP_109018||BPSL2426||3.2||This study|
|Cell motility, intracellular trafficking and secretion|
|Flagella biosynthesis/type III secretory pathway ATPase (FliI)||188||YP_106855||BPSL0227||0.2||This study|
|Burkholderia lethal factor 1 (BLF1) (hypothetical protein BPSL1549)||261||YP_108169||BPSL1549||0.1||1|||
|Universal stress protein family||208||YP_110056||BPSS0032||3.0||2||[26, 27]|
|Hypothetical protein BPSS1486||277||YP_111493||BPSS1486||0.3||3||This study|
|Hypothetical protein BPSS0683||271||YP_110696||BPSS0683||0.1||1||This study|
|Hypothetical protein BPSS2129||282||YP_112128||BPSS2129||0.3||This study|
RpoS regulates hydrogen peroxide stress response proteins in Burkholderia pseudomallei
Earlier studies of RpoS regulated proteins in B. pseudomallei have shown that the RpoS sigma factor plays a key role in stress response systems . In addition, it has been demonstrated that OxyR and DpsA, proteins involved in oxidative stress response, are regulatory target of RpoS . Here, a global approach was utilized to explore other B. pseudomallei proteins that respond to oxidative stress and are regulated by RpoS. The high confidence data of 13 H2O2 stress response proteins was combined with 22 RpoS regulated proteins (Fig. 2) obtained using highly stringent criteria as described in the Materials and Methods section. It was found that 8/13 proteins responsive to oxidative stress are regulated by RpoS (P = 0.0009, Fisher's exact test) , as shown in Table 1. These proteins are: succinyl-CoA:3-ketoacid-coenzyme A transferase subunit A (ScoA; YP_108553), non-ribosomally encoded peptide/polyketide synthase (YP_111196), cysteine synthase B (CysM; YP_109102), oxidoreductase (YP_109344), universal stress-related protein (YP_110056), Burkholderia lethal factor 1 (BLF1; YP_108169; former hypothetical protein BPSL1549) and two hypothetical proteins BPSS1486 (YP_111493), BPSS0683 (YP_110696).
Oxidative stress responsive proteins co-expressed from operons
Whether the eight oxidative responsive proteins under RpoS regulation identified in this study are also part of operons was further investigated in order to extrapolate the functional characterization of other genes in the same operons.
To do this, the computational prediction of operons in the DOOR operon database  was first exploited. The database predicts operons in prokaryotic genomes based on genome organization, promoter positions, distance between genes and conservation across closely related species [35, 36]. Three (scoA, cysM and BPSS1486) of the eight genes were predicted to be part of different operons, namely scoA and scoB in operon ID:120978, cysM with two hypothetical proteins (BPSL2507A and BPSL2508) in operon ID:121097 and BPSS1486 and two hypothetical proteins (BPSS1484 and BPSS1485) in operon ID:121554. Because genes in the same operon are co-transcribed into the same mRNA, their degrees of co-expression were checked using the available microarray data of time-series B. pseudomallei . Because the dataset from the microarray comprises the expression profiles of 5289 genes, there were 13,984,116 pair-wise correlations in total. The correlation coefficients of all genes pairs were plotted in a histogram to show the correlation distribution among these data (data not shown). The average correlation was 0.004 ( ± 0.35 SD), suggesting no correlation between unrelated genes; however, the maximal correlation was 0.996. Note that genes in operons are expected to be co-transcribed in any given condition such that the correlations of co-transcription are close to 1. However, the average correlation of four verified operons (as described in the Materials and Methods Section) was 0.676, which gives a P-value of 0.027. From this cutoff, it can be concluded that BPSL1954 and BPSL1955 (scoB and scoA), which have a correlation of 0.917, are indeed predicted to be organized into an operon; whereas the other two predicted operons (BPSL2507, BPSL2507A, BPSL2508 and BPSS1484, BPSS1485, BPSS1486) have correlations of −0.035 and 0.278, respectively, and are thus not likely to be operons.
Degrees of expression of scoA and scoB in response to oxidative stress
Of the three operons involved in the oxidative stress response and under regulation of RpoS, only one operon, namely scoA–scoB, was confirmed; we therefore focused on this operon (BPSL1954 and BPSL1955: succinyl-CoA:3-ketoacid-coenzyme A transferase subunit B [scoB] and succinyl-CoA:3-ketoacid-coenzyme A transferase subunit A [scoA], respectively). To the best of our knowledge, the role of these proteins in the oxidative stress response has not been studied in any bacteria. In our comparative proteomic profiles of B. pseudomallei untreated and treated with H2O2, only the ScoA protein was detected to exhibit reduced expression during H2O2 stress (threefold difference); in contrast the ScoB protein did not pass this stringent criterion, perhaps because ScoB is out-of-range of the pI cutoff. However, both DOOR operon prediction database and microarray data analysis indicated that scoA and scoB are co-transcribed as an operon. In order to illustrate the co-transcription of scoA–scoB genes, RT-PCR was utilized with primers designed to cover the two genes in the operon (Fig. 3a). This showed that scoA and scoB mRNA are present on the same cDNA molecule, which indicates that they are co-transcribed to the same mRNA, thus confirming that scoA and scoB are organized as an operon, as shown in Figure 3b (WT lane). According to a previous study by Osiriphun et al. , RpoS promoters are predicted to be located upstream of the scoA–scoB operon, suggesting that this operon is directly regulated by RpoS during stationary phase when grown under normal conditions (see Materials and Methods Section). To confirm this experimentally, semi-quantitative RT-PCR was performed for the scoA and scoB operon in wild-type and RpoS mutant strains, grown in normal and oxidative stress conditions as shown in Figure 3b. It was found that the RT-PCR product of the scoA–scoB operon in the RpoS mutant decreased about 2.4-fold as compared to that of the wild-type. This indicates that the scoA–scoB operon is indeed down-regulated in the absence of RpoS.
In addition, we had previously seen that the ScoA protein is expressed less in cells exposed to H2O2 than in normal growth conditions. To confirm this finding at the transcriptional level, semi-quantitative RT-PCR was performed in the wild type and RpoS mutant treated with H2O2. It was observed that the degrees of transcription of the operon under oxidative stress decreased 4.2- and 20.8-fold for the wild type and RpoS mutant treated with H2O2, respectively, as compared to wild type cells in normal conditions (Fig. 3b). Note that the RNA levels of 16sRNA were used as internal controls and are equal in each condition (Fig. 3c).
Oxidative stress response mechanisms are found throughout all aerobic organisms . ROS can be generated internally by any organisms as part of energy production, or produced by other organisms such as competing bacteria or phagocytic cells to defend against pathogenic bacteria . To enhance their survival in eukaryotic host cells, pathogenic bacteria utilize the regulators OxyR and SoxRS, as well as the alternative sigma factor RpoS, to regulate target genes in response to oxidative stress conditions . The proteobacterium B. pseudomallei, an intracellular pathogen, can survive and reproduce in several eukaryotic cell types, including phagocytic cells [3, 13, 40]. Generally, eukaryotic cells produce ROS as a defense, to induce oxidative stress and damage in infecting bacteria . However, many bacteria can survive during oxidative stress conditions, exemplifying the important role of their oxidative stress response systems. Previous studies have reported that RpoS is one of the key regulators of the oxidative stress response in B. pseudomallei .
Using high-throughput proteomics, we performed a global analysis to explore other oxidative responsive proteins regulated by RpoS in B. pseudomallei. This analysis utilized genome-wide oxidative stress from H2O2 response proteins together with the publicly available RpoS database . We identified eight RpoS regulated oxidative responsive proteins, including ScoA, a non-ribosomally encoded peptide/polyketide synthase, CysM, oxidoreductase, universal stress-related protein, BLF1, and two hypothetical proteins: BPSS1486 (YP_111493) and BPSS0683 (YP_110696). In contrast, the universal stress-related protein from E. coli is not regulated by RpoS . This indicates that the regulatory system for stress response in B. pseudomallei and E. coli are not identical. Our results also indicate that 5/13 oxidative responsive proteins are not directly regulated by RpoS, namely PanB (YP_109419), PdxJ (YP_109018), glutathione s-transferase (YP_108349), FliI (YP_106855) and the hypothetical protein BPSS2129 (YP_112128) . This is not surprising: other regulators such as OxyR and SoxRS, which are also observed in the other proteobacteria [41-43], may participate in the regulation of these five proteins, as is true for the E. coli universal stress-related protein .
The roles of many RpoS regulated proteins that respond to H2O2-induced oxidative stress have previously been studied. However, the connection between oxidative stress and ScoA, a subunit of SCOT (EC.22.214.171.124) protein has not been investigated. Our study is the first to report that the SCOT enzyme is directly regulated by RpoS and responds to oxidative stress. The data from proteomic profiles identified only ScoA, a subunit of the SCOT enzyme, as being regulated by RpoS. ScoA was down-regulated by H2O2 treatment whereas we did not detect ScoB. Because it has a pI of approximately four, which is close to the cutoff for the 2D gel system used in this study (pI 4–7), ScoB may have been masked by other proteins or out-of-range of the pI cutoff. Interestingly, the ScoA protein in B. pseudomallei has an amino acid sequence similar to the Helicobacter pylori enzyme . H. pyroli SCOT enzyme is dimeric, containing two subunits, subunit A and B, that are co-transcribed . In addition, operon prediction using the DOOR database also suggested that, in B. pseudomallei, scoA and the adjacent scoB gene are organized as an operon, which is consistent with microarray expression profiles of B. pseudomallei. Using RT-PCR, we further demonstrated experimentally that scoA and scoB are co-transcribed as an operon. In addition, we also observed that the amount of mRNA of this operon was significantly less in the RpoS mutant strain than in the wild-type strain, confirming that the operon is regulated by the RpoS sigma factor.
How does the SCOT enzyme function to respond to external ROS? When exogenous energy sources have been depleted, such as in stationary phase, bacteria change to use energy stores. PHB is a storage molecule found in many prokaryote cells, including B. multivorans, Rhodospirillum rubrum and Sphaerotilus natans . D-(−)-β-hydroxybutyrate from PHB is oxidized to acetoacetate, after which SCOT functions to transfer CoA from succinyl-CoA to acetoacetate producing acetoacetyl-CoA . Subsequently, two molecules of acetyl-CoA are produced; these can enter glycolysis or the TCA cycle to generate NADH and FADH2, which can be utilized by the ETC. Many studies have reported that the ETC is a source of intracellular ROS in prokaryotic and eukaryotic cells (Fig. 4a) [47-52].
According to our semi-qRT-PCR findings in wild type and RpoS mutant cells exposed to H2O2, the amount of mRNA in the scoA–scoB operon is down-regulated in the presence of ROS (in the form of added H2O2), as compared to normal conditions. The smaller amount of SCOT enzyme may result in both reduction in acetoacetyl-CoA production and reduction of NADH and FADH2, ultimately resulting in less endogeneous ROS (Fig. 4b). This could be a component of the mechanism employed by the bacterium to reduce intracellular ROS, allowing for enhanced tolerance to external oxidants.
Interestingly, although RpoS is a positive regulator, the mRNA levels of the scoA–scoB genes, which are regulated by RpoS, are down regulated during H2O2 stress both in wild type and RpoS mutant strains. We therefore looked for additional factors involved in regulation of the scoA–scoB operon under oxidative stress. Using a bioinformatics prediction of the promoter region, we identified binding sites for other transcription regulators, including OxyR  and SoxRS  (data not shown). OxyR reportedly has a positive regulator role in cells exposed to oxidative stress , whereas SoxRS negatively regulates expression in response to ROS stress in E. coli . We therefore suggest that the SoxRS regulator may play a role in repression of scoA–scoB expression in B. pseudomallei.
This study provides the first detailed functional analysis of how the SCOT enzyme and its regulation are important for the response of B. pseudomallei to exogenous oxidative stress. We anticipate that analysis of the role of SoxRS together with RpoS in regulating this operon will add further insights into how this deadly pathogenic bacterium survives in conditions of oxidative stress.
The research project was supported by the Thailand Research Fund. We thank the Faculty of Science of Mahidol University for providing a new researcher grant and a top-up grant to V. Charoensawan. P. Wongtrakoongate was supported by the Junior Science Talent Project. We are grateful to Dr. Sittiruk Roytrakul from the Bioservice Unit, National Science and Technology Development Agency, Pathumthani, Thailand for mass spectrometric analysis and to Dr. Laran T. Jensen, Department of Biochemistry, Faculty of Science, Mahidol University for proofreading and correcting the English version of this article.
We declare that our financial supports do not involve any conflicts of interest and that the grants received were only for academic study.