Effect of temperature on Burkholderia pseudomallei growth, proteomic changes, motility and resistance to stress environments

Burkholderia pseudomallei is a flagellated, gram-negative environmental bacterium that causes melioidosis, a severe infectious disease of humans and animals in tropical areas. We hypothesised that B. pseudomallei may undergo phenotypic adaptation in response to an increase in growth temperature. We analysed the growth curves of B. pseudomallei strain 153 cultured in Luria–Bertani broth at five different temperatures (25 °C–42 °C) and compared the proteomes of bacteria cultured at 37 °C and 42 °C. B. pseudomallei exhibited the highest growth rate at 37 °C with modest reductions at 30 °C, 40 °C and 42 °C but a more marked delay at 25 °C. Proteome analysis revealed 34 differentially expressed protein spots between bacterial cultures at 42 °C versus 37 °C. These were identified as chaperones (7 spots), metabolic enzymes (12 spots), antioxidants (10 spots), motility proteins (2 spots), structural proteins (2 spots) and hypothetical proteins (1 spot). Of the 22 down-regulated proteins at 42 °C, redundancy in motility and antioxidant proteins was observed. qRT-PCR confirmed decreased expression of fliC and katE. Experiments on three B. pseudomallei strains demonstrated that these had the highest motility, greatest resistance to H2O2 and greatest tolerance to salt stress at 37 °C. Our data suggest that temperature affects B. pseudomallei motility and resistance to stress.


Results
Effect of temperature on the growth rate of B. pseudomallei. Temperature  Proteomic alteration in B. pseudomallei under heat stress conditions. We analysed the proteomic changes of B. pseudomallei strain 153 grown at 37 °C and 42 °C by two-dimensional (2D) gel electrophoresis. Approximately 600 protein spots were visualised after Coomassie blue staining, quantitative intensity analysis of which revealed 34 differentially expressed protein spots with ≥ two-fold change between B. pseudomallei cultured at 42 °C versus 37 °C ( Fig. 2 and Table 1). Of these, 12 had increased intensity (2.2 to 24.8 fold) and 22 had decreased intensity (2.3 to 19.0 fold). Two spots were present only at 42 °C, and one spot was absent at this temperature. These proteins were identified using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) and were then categorised based on main biological function using UniProt Knowledgebase (Swiss-Prot and TrEMBL entries) ( Table 1). Proteins with altered expression were identified as chaperones (7 spots), metabolic enzymes (12 spots), antioxidants (10 spots), motility proteins (2 spots), structural proteins (2 spots) and a hypothetical protein (1 spot).
The 22 proteins with decreased expression at 42 °C were metabolic enzymes (9 spots), antioxidants (7 spots), bacterial motility (2 spots), structural proteins (2 spots), a chaperone (1 spot) and a hypothetical protein (1 spot) (Fig. 3A). By contrast, the 12 proteins with increased expression at 42 °C were mostly related to heat shock protein families and chaperones (6 spots), antioxidants (3 spots) and metabolic enzymes (3 spots) (Fig. 3A). Analysis of global protein networks using STRING revealed associations between these proteins produced at different levels ( Fig. 3B), which were consistent with the functional categories derived from the UniProt Knowledgebase database (Table 1).
RNA expression of B. pseudomallei under heat stress conditions. Of the multiple proteins altered by growth at 42 °C, redundancy in motility and antioxidant protein spots were observed among down-regulated proteins. To confirm these proteomic results, qRT-PCR was performed with three clinical isolates of B. pseudomallei from patients with melioidosis, strains 153, K96243 and 1026b for a motility gene (fliC), antioxidant protein-coding genes (katE, katG, tftC and trxA) and a reference gene (16S rDNA) ( Fig. 4A-C). All three bacterial strains demonstrated significant down-regulation of fliC and katE expression at 42 °C compared with 37 °C, while reduction in katG, tftC and trxA expression was variable between the three strains. To further validate the effect of temperature, we analysed the expression of fliC, katE and 16S rDNA genes for B. pseudomallei cultured in LB broth at 25 °C, 30 °C, 37 °C, 40 °C and 42 °C for 18 h. All three strains showed highest fliC expression at 25 °C followed by 30 °C, 37 °C, 40 °C and 42 °C (Fig. 4D). In contrast, these three isolates had highest katE expression at 37 °C and 40 °C and reduced katE expression at other temperatures (Fig. 4E).

Motility of B. pseudomallei at different temperatures.
We hypothesised that B. pseudomallei reduces motility and anti-oxidative stress functions to maintain synthesis of vital proteins for growth at high temperatures. The thermal effect on swimming motility was investigated at a range of temperatures for strain 153. Swim plates were incubated at 25 °C, 30 °C, 37 °C, 40 °C and 42 °C for 72 h. The zones of B. pseudomallei increased after incubation at all temperatures; the largest zone diameter was observed at 37 °C, followed by 40 °C, 42 °C and 30 °C, whereas the smallest zone diameter was observed at 25 °C (Fig. 5A,B). We re-examined swimming motility of bacteria at 25 °C, 37 °C and 42 °C for strains 153, K96243 and 1026b and obtained the same results (Fig. 5C). The swimming zone diameters of these strains increased overtime for all temperatures, with the largest diameters observed at 37 °C, followed by 42° and 25 °C. The diameters of all three strains after incubation at 25 °C and 42° were significantly lower than at 37 °C for all time points between 24 and 72 h (P < 0.001 for all strains, at 25 °C versus 37 °C and at 42 °C versus 37 °C).
To determine whether the reduced swimming zone at 42 °C was a result of reduced motility or growth, video-assisted imaging under light microscopy was employed to examine motility of individual bacterial cells grown in LB broth for 18 h. The computational measurement data demonstrated that the average distance of bacterial motility of the three B. pseudomallei strains was significantly lower when incubated at 42 °C compared with 37 °C (Fig. 6A,B). In addition, B. pseudomallei cultured at 25 °C and 37 °C demonstrated forward directional movement, but when cultured at 42 °C showed rotating or circular movements. These data indicate that temperature affects the motility of B. pseudomallei.  Table 1 (Fig. 7A) . The same results were obtained for K96243 and 1026b. These data suggest that temperatures can regulate the expression of genes encoding antioxidant enzymes in B. pseudomallei, and that these enzymes may be involved in the resistance to ROI. Biofilm-forming activity of B. pseudomallei. Biofilm formation has been previously reported to be associated with relapse in patients with melioidosis 24 . We, therefore, examined the effect of temperature on the biofilm-forming activity of B. pseudomallei strains 153, K96243 and 1026b in LB broth at 37 °C and 42 °C. Biofilm activity was significantly higher for all strains when cultured at 37 °C compared with 42 °C (Fig. 7B). The mean biofilm-forming activity varied between individual isolates. When cultured in LB broth at 42 °C, all strains failed to produce biofilm.

Susceptibility of B. pseudomallei to salt.
We previously demonstrated that B. pseudomallei can increase thermal tolerance after exposure to a high concentration of NaCl 13 . In the present study, we further examined the susceptibility of bacteria to a high salt concentration following heat treatment for 6 h. Susceptibility of the three B. pseudomallei strains to salt stress was tested with pre-incubation at 25 °C, 37 °C and 42 °C for 18 h prior to exposure to 450 mM NaCl in LB broth at the same temperatures for 6 h. The average numbers of viable bacteria at all temperatures were lower for bacteria cultured in LB broth with high salt compared with the bacteria cultured in LB without added salt as the control (Fig. 7C). In LB broth with high NaCl concentration, the average number of viable bacteria was highest for bacteria cultured at 37 °C with a bacterial count of 4.17 × 10 7 CFU/ml, which was significantly different from bacteria cultured at 42 °C with a bacterial count of 5.06 × 10 6 CFU/ml (P = 0.047). In LB broth without added salt as the control, the average number of viable bacteria cultured at 37 °C was not significantly different from that of 42 °C (2.17 × 10 8 CFU/ml versus 1.42 × 10 8 CFU/ml, P = 0.102) but was significantly higher than that of 25 °C (2.08 × 10 7 CFU/ml, P = 0.049). These data confirm the effect of temperature on the resistance of B. pseudomallei to salt stress.

Discussion
Melioidosis is a public health burden in tropical countries and an emerging disease that is estimated to cause 89,000 deaths globally each year 2 . Patients with melioidosis present with abscess and protracted fever that requires prolonged treatment. Our study focused on the effect of temperatures that are encountered by B. pseudomallei in the environment and the human host during infection. Our data demonstrated that B. pseudomallei could grow and survive in a wide range of temperatures. Thermal changes were found to trigger several responses in B. pseudomallei at the proteomic level. The proteins produced by B. pseudomallei due to thermic change included metabolic enzymes, chaperones, cell motility proteins, antioxidants, and structural proteins. These proteins were particularly produced by B. pseudomallei at 37 °C, a temperature which was found to support bacterial growth, motility, biofilm formation and resistance to oxidative stress. The rapid growth of B. pseudomallei at 37 °C suggests that B. pseudomallei may proficiently replicate upon entry into hosts. The ability of the bacteria to grow at 38 °C-40 °C suggests that B. pseudomallei are capable of replicating in a host with a high fever. The range of temperatures used here of 25 °C-42 °C are similar to environmental conditions in tropical regions that support B. pseudomallei in soil and water. Indeed, it has been found rDNA was used as a reference for the calculation of relative expression levels of other genes. The normalised expression levels were calculated by using 2 −ΔΔCt method 43 . Data represent the mean, and error bars represent the standard deviation. *P < 0.05; **P < 0.01, ***P < 0.001, for 37 °C versus other temperatures.
that B. pseudomallei can remain viable with high bacterial numbers of more than 10,000 CFU per gram of soil in northeast Thailand 3,4,25 . Although the growth of B. pseudomallei is reduced at lower temperatures such as 25 °C, stationary phase is finally achieved and cell viability remains high. However, our study demonstrated that B. pseudomallei grew at a high rate at 42 °C but viability slowly declined after stationary phase if the high temperature was maintained. Our results suggested that the dynamics of temperature changes during the day and night or in different seasons may affect the growth of B. pseudomallei, but that the bacterium can tolerate fluctuation in environmental temperature in tropical climates. Different growth temperatures had a broad impact on protein expression by B. pseudomallei. Proteomic analysis identified up-and down-regulated proteins in response to temperature changes, suggesting that temperature is an important environmental signal for cellular metabolism aimed at maintaining the growth and biological activity of the bacterium. Many studies have shown that alteration of bacterial cellular proteins is a necessary mechanism for cell survival under heat stress 15,16,26 . During early exposure to high temperatures, reactive oxygen species (ROS) are generated, and there is an accumulation of misfolded membrane proteins. Our result demonstrated that at 42 °C, B. pseudomallei had reduced levels of outer membrane proteins (OmpA) and chitin-binding proteins (CBPs). OmpA is a structural and immunogenic protein 27,28 and CBPs of many bacterial pathogens contain surface-exposed domains that support infection of non-chitinous mammalian hosts 29 . The increased levels of many chaperones at high temperatures are consistently observed in other bacteria 17,19 . Heat shock proteins (HSPs) are well known to participate in protein folding, refolding and removal of non-functional damaged proteins 16,18 . Escherichia coli respond to high temperatures by activating a heat shock response. The stability of mRNAs of heat shock genes has been shown to be maintained by the cold shock protein C (CspC) 15 , but this mechanism is still unknown for B. pseudomallei.
An adaptive mechanism utilised by bacteria to survive at elevated temperatures is the reduction of metabolic activity, which is crucial for energy conservation 30 . Our findings confirmed that expression of several metabolic proteins including acetoacetyl-CoA reductase, D-beta-hydroxybutyrate dehydrogenase, ATP synthase, elongation factor Tu, non-ribosomally encoded peptide/polyketide synthase, phosphoglycerate kinase, inorganic pyrophosphatase, rhodanese-related sulfotransferase and carbamate kinase were decreased. We previously showed that carbamate kinase encoded by arcC is a component of the arginine deiminase system, which facilitates acid tolerance in B. pseudomallei 12 . The decreased level of carbamate kinase under heat stress would potentially lead to defective cell survival processes. However, some cellular metabolic proteins were upregulated, including carboxypeptidase protein, pyruvate aminotransferase and phosphoserine aminotransferase.
In our study, qRT-PCR analysis confirmed the proteomic results in the modification of transcription levels of fliC (encoding flagellin) and katE expression by temperature changes. Experiments on three B. pseudomallei strains using swim agar validated the proteomic results that motility and the presence of flagella were affected by temperature change, with the highest function observed at 37 °C. Temperature-dependent regulation of flagella expression was previously reported in B. thailandensis, wherein bacterial growth at 28 °C enhanced cell motility and flagella expression through a mechanism involving the regulation of the fliC gene at the mRNA stability level 31 . B. thailandensis is a non-virulent environmental species of Burkholderia which is closely related to B. pseudomallei. We observed similar transcriptomic results in B. pseudomallei that flicC expression was highly expressed at 25 °C. However, our results were based on three B. pseudomallei strains that showed larger swimming zones at 37 °C compared with those at 25 °C and 30 °C. The discrepancy between transcriptomic results and swimming zone at 25 °C and 30 °C may be associated with bacterial growth rate or a complex process of a translation inhibition of FliC, an issue that requires further investigation. Our data from live cell imaging and electron microscopy also confirmed that the large zone diameter at 37 °C was associated with individual cell movement and flagella production. However, swimming motility was reduced when B. pseudomallei were cultured at 42 °C, which was consistent with proteomic and mRNA transcription of the fliC gene. The proteomic analysis suggested that response to increased temperature may enable bacterial cells to produce more important stress response proteins, namely chaperones and metabolic enzymes and antioxidants. Temperature has been reported to regulate flagellar motility genes in other bacteria such as Pseudomonas syringae 32 and Listeria monocytogenes 33 .
We also found that B. pseudomallei underwent a number of phenotypic adaptations under different growth temperatures. Altered flagella expression could be linked to the virulence and modulation of the immune response during infection. French et al. demonstrated that bacterial flagella support rapid intracellular motility and efficient cell-to-cell spread, thus forming multinucleate giant cells (MNGC), leading to cell death 34 . A notable virulence feature of B. pseudomallei is the formation of MNGC, which occurs in many cell types including phagocytic and non-phagocytic cells 35,36 . Flagellin (FliC) of B. pseudomallei is an important innate immune ligand for Toll-like receptor (TLR) 5. It activates NF-κB and leads to the production of inflammatory cytokines 37 . Our data highlight the importance of flagellin expression, which may be transiently regulated by the alteration of host temperature.
All three B. pseudomallei strains were efficiently resistant to ROI, with the highest resistance observed when cultured at 37 °C. This observation on oxidant agar was consistent with the results of proteomic and transcriptional analyses showing high expression of katE at 37 °C and 42 °C compared with those at 42 °C and other temperatures. katE (BPSS2214) and katG (BPSL2865) encode for catalase and peroxidase enzymes, respectively, which catalyse H 2 O 2 degradation. These enzymes are important for the survival of facultative aerobic organisms exposed to oxidative stress. katG and katE expression in B. pseudomallei has also been reported to be regulated by rpoS and oxyR 38 , whereas katE expression is regulated by rpoN2 39 in response to several stress conditions, such as nutritional starvation and oxidative stress. Further studies are required to investigate the effect of temperature on the expression and roles of these regulators. Our data demonstrated that biofilm activity was significantly higher for all three strains when cultured at 37 °C compared with 42 °C. Consistent with our study, the effect of temperature on biofilm formation has recently been reported in B. pseudomallei, in which the level of biofilm formation by B. pseudomallei strain 1026b in LB medium decreased ∼6-fold at 40 °C and 42 °C compared with 37 °C 40 . The role of temperature in the regulation of biofilm formation has been linked to cyclic di-GMP signaling in B. pseudomallei 40 . Our data suggest that biofilm formation is associated with B. pseudomallei cells that are more metabolically active during growth at 37 °C. The decrease in biofilm seen at 42 °C was possibly related to the down-regulation of the flagella expression and other metabolic enzymes.
We previously showed that a high NaCl concentration affects the survival and adaptive mechanisms of B. pseudomallei. In response to exposure to high NaCl concentration, B. pseudomallei isolates show an increased thermal tolerance, oxidative resistance, and plaque-forming efficiency 13 . However, the result of the present study demonstrated that B. pseudomallei had the greatest tolerance to salt stress following pre-incubation at 37 °C in comparison with that at 25 °C and 42 °C. Increased temperature of 42 °C was negatively associated with reduced bacterial viability under the same salt stress conditions. Resistance to salt at 37 °C may provide a major benefit for B. pseudomallei stability in environmental niches where the bacteria encounter a warm temperature but high salt concentrations.
Our study suggested that the human body temperature of 37 °C facilitates maximum growth rate, motility and anti-oxidative stress functions in B. pseudomallei. These mechanisms may be a prerequisite for acute infection and dissemination. In conditions with high temperatures, B. pseudomallei expresses a number of stress-response proteins but decreases its motility by reducing flagella expression. This may ultimately lead to latent infection if the patients do not recover from melioidosis.

Materials and Methods
Ethics statement. The study was approved by the Ethics Committee of the Faculty of Tropical Medicine, Mahidol University (approval number: MUTM2018-009-01). All bacterial isolates obtained from humans were anonymous.
B. pseudomallei strains. Three B. pseudomallei isolates from patients with melioidosis in northeast Thailand were used: strains 153, 1026b and K96243. All isolates grew with a type I colony morphology on Ashdown agar after incubation for 4 days at 37 °C in air 41 . Unless otherwise stated, a single colony of B. pseudomallei was cultured on LB, Lennox agar or broth and incubated at 37 °C in air for 18 h before being used in each experiment.
Bacterial growth curve analysis. Bacterial culture was performed in a BSL-3 laboratory. Several colonies of B. pseudomallei were suspended in sterile PBS and the optical density (OD) at 600 nm adjusted to obtain a bacterial concentration of approximately 1 × 10 8 colony-forming units (CFU)/ml. Five microliters of bacterial suspension was added to 5 ml of LB broth to make a final concentration of 1 × 10 5 CFU/ml. Cultures were then incubated in duplicate at five different temperatures under static conditions: 25 °C, 30 °C, 37 °C, 40 °C and 42 °C and sampled at time intervals (0-, 6-, 12-, 24-, 36-, 48-, 60-and 72 h), aliquots from which were serially diluted in sterile PBS. Ten microliters of each dilution were inoculated on Columbia agar in triplicate, followed by incubation at 37 °C in air for 16 h before a colony count was performed.
Protein extraction. A single colony of B. pseudomallei on Ashdown agar was picked and inoculated into 100 ml of LB broth, which was incubated at 37 °C or 42 °C under static conditions for 18 h. Two independent experiments were performed. Bacteria were harvested by centrifugation at 4,500 × g for 30 min followed by washing with PBS. The bacterial pellet was resuspended in 1 ml of a cold lysis buffer (5 mM EDTA and 1 mM phenylmethylsulfonyl fluoride) 12 and then sonicated on ice at 22% amplitude for 3 min. The cell lysate was centrifuged at 14,000 × g at 4 °C for 3 min, and the supernatant was then collected and filtered using a 0.2-μm filter. Protein samples were stored at −80 °C until use.

2D gel electrophoresis and quantitative spot intensity analysis.
After cleaning the protein samples, protein concentration was measured using a 2D quantification kit (GE Healthcare Bio-Sciences). Protein complexes (700 µg) derived from two independent experiments for each condition were separated by isoelectric focusing on Immobiline DryStrip (linear pH gradient of 4-7, 18 cm) followed by 12% polyacrylamide gel electrophoresis, as previously described 12 . Separated protein spots were visualised with colloidal Coomassie blue G-250 stain and then captured using an Image Scanner II and LabScan software version 5.0 (GE Healthcare Bio-Sciences). Image Master 2D Platinum software 7.0 (GE Healthcare Bio-Sciences) was used for matching and analysis of the protein spots in 2D gels. The fold change in the intensity of matched protein spots between 37 °C and 42 °C was calculated. The differential spots representing more than a two-fold change were then subjected to protein identification by mass spectrometry.
In-gel tryptic digestion and protein identification by MALDI-TOF MS. Candidate proteins were excised from the gel using the Ettan Spot Handling Workstation (GE Healthcare Bio-Sciences) followed by digestion with trypsin, as previously described 12  were classified based on their main biological functions using UniProt Knowledgebase (Swiss-Prot and TrEMBL entries). The differentially expressed proteins isolated at 42 °C compared with at 37 °C were further analysed for protein interaction network using a STRING tool software (version 8.3) (http://string.embl.de/) 42 . Predicted protein-protein associations were queried through experimentally derived physical protein interactions from the research literature and databases of curated biological pathway knowledge. RNA extraction. B. pseudomallei strains K96243, 153 and 1026b were inoculated into 5 ml LB broth and incubated at 25 °C, 30 °C, 37 °C, 40 °C and 42 °C for 18 h under static conditions, after which bacteria were harvested by centrifugation at 12,000 rpm for 2 min. The pellet was washed with 1 ml of PBS, and RNA was extracted using RNeasy kit (Qiagen, Germany). The integrity of the purified RNA was examined by agarose gel electrophoresis, and the concentration was measured using a Nanodrop spectrophotometer (Thermo Scientific, USA). Contaminant genomic DNA was removed using DNase I treatment according to the manufacturer's instructions (Thermo Scientific, USA), and the presence of residual DNA was checked by PCR using primers for the 16S rDNA gene (Table S1).
Quantitative reverse transcriptase PCR (qRT-PCR). Two-step qRT-PCR was used to quantitatively measure gene expression using the iScript TM Reverse Transcription Supermix for RT-qPCR (Bio-Rad, USA) and a CFX96 Touch TM Real-Time PCR Detection System with Bio-Rad CFX Manager software version 3.0 (Bio-Rad). All primer pairs were designed using NCBI PrimerBlast (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). The primer sequences are shown in Table S1. The amplification was performed for the fliC, katE, katG, tftC, trxA and 16 S rDNA genes in duplicate in 10 µl total volume containing 5 µl of 2 iTaq Universal SYBR Green (Bio-Rad), 1 µl of cDNA, 0.4 µl of each primer (10 µmol/l), and 3.2 µl of distilled water. The cycle conditions were as follows: 1 cycle of 95 °C for 30 s followed by 40 cycles of 95 °C for 10 s and 60 °C for 30 s. After amplification, melting curve analysis was conducted by increasing the annealing temperature by 0.1 °C per step from 65 °C to 95 °C. The 16 S gene was used as the reference for calculating the relative expression levels of other genes. The normalised expression levels were calculated by using the 2 −ΔΔCt method, where ΔΔCt = (Ct target gene − Ct 16S rDNA gene ) at other temperatures − (Ct target gene − Ct 16S rDNA gene ) at 37 °C 43 .
Swimming motility assay. Two single colonies of B. pseudomallei were cultured in 5 ml of LB broth at a range of temperatures (25 °C, 30 °C, 37 °C, 40 °C and 42 °C) under static conditions for 18 h. The OD at 600 nm of the bacterial suspension was adjusted in sterile PBS to a concentration of 1 × 10 8 CFU/ml and point inoculated into the centre of swimming agar plate (1% tryptone, 0.5% NaCl and 0.3% agar) using a sterile toothpick, as described previously 41 . The swim plates were further incubated at the same temperatures. The swimming zone diameters were measured in three different positions including the widest, medium and shortest diameters of the circular turbid zone formed by the bacterial cells migrating away from the point of inoculation at 0-, 6-, 12-, 24-, 36-, 48-, 60-and 72 h time intervals and the average zone diameters are presented.
Live cell imaging of cell motility in B. pseudomallei. Live cell imaging was used to monitor the motility of B. pseudomallei using the hanging drop method. B. pseudomallei were cultured on Ashdown agar at 37 °C overnight. Colonies of each isolate were harvested and suspended in 10 ml of LB broth, and their OD at 600 nm was adjusted to a concentration of 1 × 10 10 CFU/ml. Equal volumes of the bacterial suspension (2.5 ml) were added into two tubes containing LB broth to make a final volume of 5 ml and incubated separately at 37 °C and 42 °C for 6 h under static conditions to obtain log-phase cultures. Thereafter, 30 µl of the bacterial suspension was dropped on a glass slide with a depression, over which a cover slip was placed. Motility was observed using a light microscope (Leica DM750, Wetzlar, Germany) at 1000 × magnification. A video of the motility was recorded and the motility of 20 individual cells tracked for 20 s using ImageJ program (http://rsb.info.nih.gov/ij/) with the manual tracking plug-in.
Transmission electron microscopy. B. pseudomallei strain 153 was cultured in LB broth at 25 °C, 37 °C and 42 °C for 6 h. A preparation of the bacterial cells was examined using transmission electron microscopy (TEM) for the presence of flagella, as previously described 44 . The bacterial cells were observed with a Hitachi Electron Microscope HT-7700 (Hitachi, Japan). The presence of bacterial flagella was recorded for 50 bacteria in each temperature.

Determination of susceptibility of B. pseudomallei to reactive oxygen intermediates (ROI).
The susceptibility of B. pseudomallei to ROI at different temperatures was determined on oxidant agar plates, as previously described 7 . Briefly, B. pseudomallei were inoculated in 10 ml of LB broth and incubated at different temperatures (25 °C, 30 °C, 37 °C, 40 °C and 42 °C) for 16 h. The bacterial suspension from each condition was diluted with sterile PBS to make 10-fold dilutions. Ten microliters of a serially diluted inoculum were inoculated onto LB agar containing 0, 78 and 156 μM H 2 O 2 in triplicate. The colonies were counted after incubation at 37 °C in air for 24 h. The number of colonies on the plates containing H 2 O 2 was normalised with those from plates without H 2 O 2 by presenting as the percentage of bacterial survival.
Assessment of biofilm-forming activity of B. pseudomallei. The biofilm-forming activity of B. pseudomallei was assessed as described previously 41 . Briefly, the bacteria were inoculated into 10 ml of LB broth and incubated at 37 °C with shaking for 18 h. The OD at 600 nm of the bacterial suspension was adjusted to 1.0 with fresh LB broth and 200 µl of the suspension was added into a 96-well flat-bottomed plastic tissue culture plate (FALCON, USA), with eight wells/strain. Wells with only LB broth were included as negative controls. The plates were incubated at 37 °C or 42 °C to allow adhesion for 3 h. The supernatant was gently aspirated and replaced with 200 µl of fresh LB broth, and the plate was continually incubated at the same temperature for an additional 21 h. The supernatant was replaced with a fresh LB broth and incubated as above for an additional 24 h. Thereafter, the wells were washed with PBS, fixed with methanol for 15 min, and air-dried at room temperature. The plates were stained with 1% crystal violet and solubilised with 250 µl of 33% (v/v) glacial acetic acid per well before measuring the OD at 630 nm using a TECAN microplate reader (TECAN, Switzerland).
Salt stress assay. The survival of B. pseudomallei strains K96243, 1026b and 153 under salt stress conditions were determined by observing the number of viable bacteria after exposure to 450 mM NaCl, as previously described 13 . Briefly, B. pseudomallei were incubated in 5 ml of LB broth at 25 °C, 37 °C and 42 °C for 18 h. The bacterial cells were harvested, washed and resuspended in PBS. The OD at 600 nm of the bacterial suspension was adjusted to 0.15 to obtain a bacterial concentration of 1 × 10 8 CFU/ml. The bacteria were then inoculated at a dilution of 1:10 into 10 ml of LB broth containing 450 mM NaCl or LB broth without added salt as the control. The bacteria were further incubated at the same temperatures for 6 h with shaking. A 10-fold dilution of the bacteria was prepared and plated onto LB agar. After incubation at 37 °C for 24 h, colonies were counted. The mean numbers of colonies of the three B. pseudomallei strains were presented as the CFU.
Statistical analysis. Statistical analyses were performed using Stata, version 12 (StataCorp LP, College Station, TX, USA). One-way ANOVA or Student t-test was used to test differences in quantitative data among different groups. All data in this study were presented as mean ± standard deviation. Differences were considered statistically significant at a P value < 0.05. Data availability. All data generated or analysed during this study are included in this published article.