Substrate preference of 5′-methylthioadenosine/S-adenosylhomocysteine nucleosidase in Burkholderia thailandensis

Authors

  • Qiang Gao,

    1. Key Laboratory of Agricultural and Environmental Microbiology, Wuhan Institute of Virology, Chinese Academy of Sciences, Wuhan, China
    2. Graduate School of the Chinese Academy of Sciences, Beijing, China
    Search for more papers by this author
  • Dasheng Zheng,

    1. Key Laboratory of Agricultural and Environmental Microbiology, Wuhan Institute of Virology, Chinese Academy of Sciences, Wuhan, China
    Search for more papers by this author
  • Zhiming Yuan

    Corresponding author
    • Key Laboratory of Agricultural and Environmental Microbiology, Wuhan Institute of Virology, Chinese Academy of Sciences, Wuhan, China
    Search for more papers by this author

Correspondence: Zhiming Yuan, Wuhan Institute of Virology, Chinese Academy of Sciences, Wuhan 430071, China. Tel.: 86 27 87197242; fax: 86 27 87198120; e-mail: yzm@wh.iov.cn

Abstract

5′-Methylthioadenosine/S-adenosylhomocysteine nucleosidase (MTAN) plays crucial roles in the production of autoinducers and methionine metabolism. Putative genes encoding MTAN and AdoHcyase from Burkholderia thailandensis were cloned and characterized. The Km values of MTAN for 5′-methylthioadenosine (MTA) and S-adenosylhomocysteine (SAH) were 19 and 58 μM, respectively. The catalytic efficiency of MTAN for SAH was only 0.004% of the value for MTA, indicating an almost complete substrate preference of MTAN for MTA. The results of autoinducer-2 assay of B. thailandensis and recombinants indicated that LuxS enzyme activity was lacking in Burkholderia species. Instead, AdoHcyase hydrolysed SAH directly to homocysteine and adenosine in the activated methyl cycle. Meanwhile, the Km value of AdoHcyase for SAH was determined to be 40 μM. Sequence analysis revealed that MTAN had much higher diversity than AdoHcyase, which likely contributes to its substrate preference for MTA. Furthermore, the phylogenetic tree of MTAN sequences revealed that LuxS+ bacteria could be discriminated from LuxS bacteria. These results suggested that the substrate preference of MTAN for MTA and SAH degradation pathway evolved with the bacterial-activated methyl cycle.

Introduction

The activated methyl cycle is a central metabolic pathway responsible for the methylation of cellular components and the recycling of sulphur-containing amino acids. S-Adenosylmethionine (SAM), synthesized from methionine and adenosine triphosphate, is a methyl group donor in a myriad of biological and biochemical events (Lu, 2000). N-acyl-homoserine lactone (AHL) synthase catalyses the transfer of the amino acid moiety of SAM to an acyl acceptor to produce homoserine lactones in the synthesis of AI-1 molecules with 5′-methylthioadenosine (MTA) as a byproduct of this reaction. MTA is also derived from polyamine biosynthesis, where the nucleoside is produced stoichiometrically with spermidine and spermine. 5′-methylthioadenosine nucleosidase/S-adenosylhomocysteine nucleosidase (MTAN) catalyses the physiologically irreversible hydrolysis of MTA to adenine and 5-methylthio-d-ribose (MTR). Adenine is subsequently recycled into the adenine nucleotide pool by the widely distributed adenine phosphoribosyltransferase, while the MTR is subsequently phosphorylated to 5-methylthio-d-ribose-1-phosphate (MTR-1-P) and converted into methionine. Meanwhile, more than 40 metabolic reactions involve the transfer of a methyl group from SAM to various substrates generating S-adenosylhomocysteine (SAH) as a byproduct. SAH is converted into homocysteine (Hcy), adenine and 4, 5-dihydroxy-2, 3-pentanedione (DPD) by the sequential action of the MTAN and LuxS enzymes. The byproduct, DPD, can spontaneously cyclize and interact with borate to form at least two different interconvertible molecules described as a universal signal molecule, AI-2. In every case, an AI-2 synthase, which is highly similar to the Vibrio harveyi LuxS protein, is required for its synthesis. AI-1 and AI-2 are two classes of autoinducers synthesized from SAM.

MTA and SAH need to be removed efficiently, as they are potent feedback inhibitors of polyamine biosynthesis and biological methylation, respectively. MTAN catalyses the irreversible cleavage of the glycosidic bond in MTA and SAH and functions at two steps within bacterial pathways related to polyamine biosynthesis: (1) quorum sensing and methylation and (2) purine and methionine salvage reactions (Cornell & Riscoe, 1998). In mammals and some prokaryotes, MTAN does not exist. Two distinct enzymes catabolize MTA and SAH. MTA is converted to 5-methylthioribose-1-phosphate (MTR-1-P) and adenine by MTA phosphorylase (MTAP) (Sekowska et al., 2004). SAH is hydrolysed by adenosylhomocysteinase (AdoHcyase) to produce Hcy and adenosine. Adenosine is then rapidly converted to inosine (I) and adenosine monophosphate (Baric et al., 2004).

Blocking MTAN activity is expected to cause the accumulation of MTA, resulting in product inhibition of AHL synthase and reduced AI-1 synthesis. In addition, inhibition of MTAN can directly block the formation of S-ribosylhomocysteine (SRH), the precursor of AI-2 (Gutierrez et al., 2009). Therefore, the inhibition of MTAN is expected to inhibit autoinducers synthesis, polyamine biosynthesis and the salvage pathways for adenine and methionine. Structural analogues of MTA and SAH have been synthesized, which display effective antimicrobial activity both in vitro and in vivo (Singh et al., 2005a, b; Gutierrez et al., 2009).

Burkholderia species can be opportunistic or obligate pathogens, causing human, animal and plant diseases. Human Burkholderia infections are usually treated with antibiotics to improve control of the disease and patient survival. Increasing bacterial resistance to antibiotics has become a serious public health problem. Burkholderia thailandensis is a nonfermenting motile gram-negative soil bacterium. It is closely related to Burkholderia pseudomallei and regularly used to model B. pseudomallei because of similarities in the immune response.

In this study, we have cloned and characterized MTAN and AdoHcyase from B. thailandensis. Our study represents the first direct enzyme kinetics study of MTAN and AdoHcyase of B. thailandensis, and we compare the activated methyl cycle of B. thailandensis with other species. Exploitation of potential differences in substrate specificity between the Burkholderia and mammalian enzymes represents a promising opportunity for development of selective antimicrobial agents. These data are of importance for further understanding the role of MTAN and AdoHcyase in activated methyl cycle of Burkholderia species and provide an insight into developing new antibiotic design and curative treatment approaches for melioidosis.

Materials and methods

Bacterial strains and growth conditions

Bacterial strains used in this study are listed in Table 1. Strains of Escherichia coli, B. thailandensis and recombinants were grown in Luria–Bertani (LB) medium with appropriate antibiotics at 37 °C. Antibiotics were added at the following concentrations (per mL): 50 μg kanamycin, 10 μg tetracycline, 100 μg gentamycin for Ecoli and 50 μg tetracycline for B. thailandensis. Vibrio harveyi BB170 was cultured at 30 °C in autoinducer bioassay medium for the AI-2 assay (Taga, 2005).

Table 1. Bacterial strains and plasmids used in this study
StrainRelevant propertiesReference or source
  1. *Kmr, Tcr represent resistance to kanamycin and tetracycline, respectively.

B. thailandensis
E264Wild-type strainATCC
E264Δmtanmtan mutant of E264, Tcr*Unpublished study
E264luxSE264 with the plasmid pRKluxS, TcrThis study
E. coli
BL21(DE3)hsdS gal (λcIts857 ind-l Sam7 nin-5 lacUV5-T7 gene 1)Novagen
S17-1 λ pirTpr Smr recA thi pro hsd RM+::RP4:2-Tcr::Mu Kmr::Tn7, λ pirSimon et al. (1983)
V. harveyi
BB170luxN::tn5KanBassler et al. (1997)
Plasmids
pET28Prokaryotic expression vector, Kmr*Novagen
pET-mtanpET28a with the mtan geneThis study
pET-ahcYpET28a with the ahcY geneThis study
pET-luxSpET28b with the luxS geneThis study
pRK415Broad host range expression vector with lactose-inducible promoter, TcrKeen et al. (1988)
pRKluxSpRK415 with the luxS geneThis study

Plasmid construction and DNA manipulation

Plasmids and primers used in this study are listed in Tables 1 and 2. DNA manipulations and transformations were performed according to Sambrook (Sambrook & Russell, 2001).

Table 2. Primers used in this study
Name5′–3′ sequence*
  1. *Restriction endonuclease sites added for cloning purposes are underlined.

mtan-F-EcoRIGGAATTCATGAATCAGCAGACGTCTTTCGACCG
mtan-R-HindIIICCGCTCGAGTCAAGCCACGCGCGCCTTC
achy-F-EcoRIGGAATTCATGAACGCCGCTGTCATCG
achy-R-HindIIICCCAAGCTTTCAGTAACGATAGTGATCCGGC
luxS-F-HindIIICCCAAGCTTATGCCGTTGTTAGATAGCTTC
luxS-R-XhoICCGCTCGAGCTAGATGTGCAGTTCCTGC
luxS-R-XbaIGCTCTAGACTAGATGTGCAGTTCCTGC

To over-express His6-tagged MTAN in E. coli BL21 (DE3), a DNA fragment containing the open reading frame of mtan with the stop codon (BTH_I2631) was amplified by PCR with primers mtan-F-EcoRI and mtan-R-HindIII. The PCR product was digested with EcoRI/HindIII and then inserted into the corresponding site of the pET28a vector to create the plasmid pET-mtan. The primers ahcY-F-EcoRI/ahcY-R-HindIII and luxS-F-HindIII/luxS-R-XhoI were used to amplify ahcY (BTH_I3165) and luxS (ECBD_1033), respectively. Fragments were inserted into the corresponding digested sites of pET28a and pET28b to produce the plasmids pET-ahcY and pET-luxS.

To express the luxS gene in B. thailandensis E264, the luxS gene was amplified with the primers luxS-F-HindIII/luxS-R-XbaI and cloned into the same digested sites of pRK415 to produce pRKluxS.

Phylogenetic analysis of MTAN and AdoHcyase

The cloned ORFs were sequenced, and the corresponding translated protein sequences were submitted to NCBI Conserved Domains Database (CDD) for analysis of the conserved domains. Putative proteins that shared similar functions were selected, and sequences were downloaded from the Kyoto Encyclopaedia of Genes and Genomes (KEGG). All of the sequences were aligned, and phylogenetic trees reconstructed and visualized in mega4 software (Tamura et al., 2007).

Expression and purification of recombinant proteins

The expression of MTAN, AdoHcyase and LuxS in E. coli BL21 were performed according to Sambrook (Sambrook & Russell, 2001). Recombinant proteins were purified from the supernatant with a His-Bind Resin chromatography kit (Novagen). The protein concentration was determined utilizing a BCA-protein assay kit (Pierce).

Expression of LuxS in B. thailandensis E264

Escherichia coli S17-1 cultures containing plasmid pRKluxS and parental B. thailandensis were conjugated using the method described previously (Barrett et al., 2008). A recombinant named E264luxS was screened and confirmed by plasmid identification. E264luxS was grown in LB medium with tetracycline, and LuxS was over-expressed by induction with isopropyl β-d-1-thiogalactopyranoside (IPTG).

In vitro autoinducer-2 synthesis and assay

The assay for synthesizing autoinducer-2 in vitro was based upon the report of Chen-Yu Tsao with some modification (Tsao et al., 2011). The cultured cells of B. thailandensis (wild type and recombinants) were harvested by centrifugation, washed twice with ice-cold PBS (pH 7.4) and finally resuspended in 50 mM Tris-HCl (pH 7.8). Cells were disrupted by ultrasonication, and the supernatant was collected and added to 1 mM SAH in 50 mM Tris-HCl (pH 7.8) and incubated at 37 °C for 12 h. In parallel, purified His6-MTAN and His6-LuxS were added as positive controls. The enzymatic reaction product was extracted twice with chloroform and recovered from the aqueous phase.

Cell-free supernatants (CFS) of B. thailandensis wild type and recombinant were harvested at various growth stages (early log phase, mid to late log phase, stationary phase). Synthesis of autoinducer-2 and CFS were passed through a 0.22 μm filter unit prior to conducting the assay. Autoinducer-2 production was assayed based on the protocol of Taga (2005). CFS of E. coli was tested as the positive control. Data, reported as fold activation, were obtained by dividing the light produced by the reporter cells after the addition of CFS by the light output from the reporter cells when enzymatic reaction buffer alone was added.

Enzymatic assay of recombinant MTAN and AdoHcyase

MTA/SAH hydrolase activity of MTAN was determined by the enzyme-linked spectrophotometric assay (Dunn et al., 1994). A 0.28 units of xanthine oxidase were mixed with 200 μL of 1 mM p-iodonitrotetrazolium violet (INT) dissolved in 100 mM HEPES buffer (pH 7.5). For enzyme kinetics, the concentration of MTA and SAH substrates was varied from 0.625 to 100 μM. Enzyme reactions were initiated with the addition of 10 nM–4 μM MTAN (250 μL final volume). The A470 of the reaction mixture was recorded every 10 min at 25 °C for 1 h using a 96-well plate reader (BioTek). The molar extinction coefficient (ε470 = 15 400 M−1 cm−1) was used to measure the conversion of INT to its formazan salt. Enzyme reactions containing similar concentrations of adenine served as the positive control. Reactions lacking the addition of MTAN were used as the negative control.

SAH hydrolase activity of AdoHcyase was determined according to a method reported previously (Lozada-Ramirez et al., 2008). Enzyme reactions were conducted at 37 °C in 50 mM phosphate buffer (pH 7.6) that contained SAH (0.625–200 μM) and 100 μM 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB). Reactions (final volume 250 μL) were initiated by the addition of 1 μM purified AdoHcyase. AdoHcyase activity was measured every 10 min for 1 h by monitoring the conversion of DTNB to 5-thio-2-nitrobenzoic acid (TNB) at 412 nm (ε412 = 13 700 M−1 cm−1) (Ellman et al., 1961). Enzyme reactions containing cysteine or lacking AdoHcyase were used as positive and negative controls, respectively.

Reaction conditions were optimized by measuring the effect of pH (4.5–9.5) and temperature (4, 20, 37, 45, 55 °C) on recombinant enzyme activity. For the calculation of enzyme activity in the Synergy HT plate spectrophotometer, the path length was determined to be 0.8 cm. Enzyme assays were performed in triplicate at each experimental condition. The results represent the mean (±SD) from three independent experiments.

Data analyses

Data for determining the kinetic constants were fitted to the Michaelis-Menten equation using graphpad prism v.5.0 (GraphPad Software). The Student's t-test was applied for statistical analysis with spss software (version 13) between the groups.

Results

Sequence analysis and phylogenetic tree construction

We identified a 789-bp fragment corresponding to a putative mtan gene and a 1422-bp fragment corresponding to a putative ahcY gene based on the complete B. thailandensis E264 genome sequence (Kim et al., 2005). MTAN matched to a multi-domain model belonging to PRK06698 (bifunctional 5′-MTA/SAH nucleosidase/phosphatase), and AdoHcyase matched to a specific hit on cd00401 (SAH hydrolase) from CDD.

A consensus tree was generated from the bootstrap replicates using mega4 with default parameters. Branches with bootstrap scores of < 50 were collapsed in the tree (Fig. 1). MTAN and AdoHcyase were conserved among different Burkholderia species with 95% homology. The phylogenetic tree of MTAN was subdivided into three subgroups: (1) LuxS+ bacteria (Vibrio cholerae, E. coli, Staphylococcus aureus and Streptococcus pneumoniae), (2) LuxS bacteria (B. thailandensis, B. pseudomallei, Burkholderia mallei, Mycobacterium tuberculosis, Brucella melitensis, Ralstonia pickettii and Pseudogulbenkiania_NH8B) and 3) eukaryotic organisms including plants (Arabidopsis thailana) and protozoa (Tetrahymena thermophila) (Fig. 1a). AdoHcyase was subdivided into three subgroups: (1) LuxS bacteria (Burkholderia spp., pseudogulbenkiania_NH8B, M. tuberculosis, R. pickettii and B. melitensis), (2) plants (Arabidopsis thaliana) and protozoa (T. thermophila and Paramecium tetraurelia) and (3) animals (Danio rerio, Homo sapiens and Branchiostoma floridae) (Fig. 1b).

Figure 1.

Phylogenetic tree of MTAN and AdoHcyase. (a) Phylogenetic tree of MTAN. MTAN was subdivided into LuxS+ bacteria, LuxS bacteria and another group including a plant and protozoa. The LuxS+ and LuxS subgroups could be separated according to the phylogenetic relationship. Plants and protozoa were more distant from the bacteria. (b) Phylogenetic tree of AdoHcyase. AdoHcyase was subdivided into three subgroups: LuxS bacteria, plant and protozoa, and animals.

Cloning, expression and purification of MTAN, AdoHcyase and LuxS

MTAN and AdoHcyase of B. thailandensis and LuxS of E. coli were expressed in the recombinant E. coli BL21 strains after induction with IPTG. Sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) showed purified MTAN, AdoHcyase and LuxS each produced a single protein band with estimated molecular weights of 34, 58 and 26 kD, respectively (Fig. 2a).

Figure 2.

Enzyme dynamics of MTAN and AdoHcyase. (a) SDS-PAGE of purified MTAN, AdoHcyase and LuxS. Lane 1 standard protein marker, lane 2 purified Burkholderia thailandensis MTAN, lane 3 purified Escherichia coli LuxS, lane 4 purified B. thailandensis AdoHcyase. (b) MTAN catalytic activity for MTA. (c) MTAN catalytic activity for SAH. (d) AdoHcyase catalytic activity for SAH. Data represent mean values ± SD for three replicates.

Enzyme kinetics of MTAN and AdoHcyase from B. thailandensis

Studies were performed to determine the optimal pH and temperature for MTAN activity. The results indicate that the enzyme has similar activity across a broad pH (pH 5–8) and temperature (25–37 °C) range (data not shown). Burkholderia thailandensis MTAN hydrolyses both MTA and SAH, although hydrolytic activity towards SAH is weak. Hydrolysis of MTA required only 10 nM MTAN in the assay (Fig. 2b), while 4 μM MTAN was required to measure SAH hydrolysis (Fig. 2c). All data were fitted utilizing the Michaelis–Menten equation and summarized in Table 3. A kcat/Km ratio of B. thailandensis MTAN was calculated as 9.5 × 104 M−1 s−1 for MTA; and 3.6 for SAH, only 0.004% of the value for MTA. Burkholderia thailandensis AdoHcyase also has hydrolytic activity against SAH (Fig. 2d). The AdoHcyase concentration for SAH was 1 μM, the catalytic efficiency of AdoHcyase for SAH was about 188-fold higher than that of MTAN towards SAH.

Table 3. Enzyme kinetic constants of MTAN and AdoHcyase
ProteinsSubstratesKm (μM)Vmax (μM min−1)kcat (s−1)kcat/Km (mol−1 L s−1)
MTANMTA191.081.89.5 × 104
SAH580.052.1 × 10−43.6
AdoHcyaseSAH401.632.7 × 10−26.8 × 102

Autoinducer-2 assay of B. thailandensis

CFS collected from cultures at various time points were tested to analyse the amount of AI-2. However, no growth of the reporter strain, V. harveyi, could be detected at the tested CFS concentration of 1%. Some secondary metabolites produced by B. thailandensis might inhibit the growth of V. harveyi (Duerkop et al., 2009). In vitro derived, synthetic AI-2 molecules induced 120-fold more light than V. harveyi BB170, which was comparable to the 372-fold increase in light produced by E. coli. Burkholderia thailandensis E264 and E264Δmtan were negative in the assay. Burkholderia thailandensis E264 with supplementation of LuxS in vivo (E264luxS) and in vitro (E264 + LuxS) gave 260-fold and 162-fold increases in light produced, respectively. Burkholderia thailandensis E264Δmtan with supplementation of LuxS was still negative (Fig. 3).

Figure 3.

AI-2 assay of Burkholderia thailandensis E264 and E264 derivatives. E264 is the B. thailandensis wild-type, and E264Δmtan is the mtan mutant of B. thailandensis. E264Δmtan LuxS is E264Δmtan supplemented with purified LuxS. E264 + LuxS was E264 supplemented with purified LuxS in vitro. E264luxS is utilized for LuxS overexpression in B. thailandensis; Escherichia coli is the positive control in the assay. The Y-axis was the bioluminescence induced fold change of the sample with respect to the negative control. Data represent mean values ± SD from at least six replicates. Three samples (E264Δmtan + LuxS, E264luxS, E. coli) showed significantly higher levels of AI-2 production than the negative control (P < 0.0001).

Discussion

In most bacteria, MTANs show dual-substrate specificity for MTA and SAH, while plant MTANs (Lupinus luteus and A. thaliana) exhibit a distinct preference for MTA with little or no activity towards SAH (Guranowski et al., 1981; Siu et al., 2008). The comparison of available A. thaliana MTAN1 and MTAN2 structures (PDB ID: 2H8G/3BSF) suggests that a combination of differences in the 5′-alkylthio binding region and reduced conformational flexibility in the A. thaliana active site likely contribute to its reduced efficiency in binding substrate analogues with longer 5′-substituents (Siu et al., 2011). Burkholderia thailandensis MTAN showed comparable activity for MTA when compared to other homologous proteins. The catalytic efficiency for MTA was 26-fold less than that of the A. thaliana MTAN1 isoform and 6-fold less than that of the A. thaliana MTAN2 isoform (Siu et al., 2008). However, at high enzyme concentration (4 μM) a small amount of SAH hydrolysis could be measured. MTAN exhibited extremely low activity towards SAH, with only 0.004% catalytic efficiency when compared to MTA, indicating the preference of MTAN for MTA. This result was similar to the A. thaliana MTAN's substrate preference. Homology modelling analysis suggested that Phe148/Phe120 of B. thailandensis MTAN probably hinders the divergence in specificity of the enzyme and is responsible for the low activity of MTAN for SAH (data not shown). Burkholderia thailandensis AdoHcyase activity and catalytic efficiency was lower than the other reports (Lozada-Ramirez et al., 2008). The probable reason was the recombinant protein did not represent the native protein from B. thailandensis, and specific reasons will require further investigations in the future.

The activated methyl cycle in B. thailandensis were confirmed based upon the qualitative AI-2 assay. Although the assay showed that B. thailandensis wild-type (E264) and mutant knockout strains (E264Δmtan and E264Δmtan + LuxS) were negative for AI-2 production, the complementation of LuxS activity in B. thailandensis E264 (E264luxS and E264 + LuxS) yielded strong positive results in the AI-2 assay. This demonstrated that the activated methyl cycle of Burkholderia spp. was in agreement with KEGG (http://www.genome.jp/kegg-bin/show_pathway?bte00270). Burkholderia thailandensis MTAN (EC 3.2.2.9) could hydrolyse SAH to SRH, while hydrolysis of SRH is blocked due to the absence of LuxS (EC 4.4.1.21). Meanwhile, B. thailandensis AdoHcyase (EC 3.3.1.1) hydrolyses SAH to release Hcy and Adenosine.

According to the different activated methyl cycle in bacteria, bacteria could be divided into two groups: LuxS+ and LuxS. Homology analysis of the MTAN sequences and phylogenetic results showed that sequence diversity of MTAN was distinguishable between LuxS and LuxS+ bacteria. We propose that sequence diversity could contribute to the decreased activity of MTANs for SAH. In the methionine metabolism, MTAN and LuxS complete one pathway (LuxS+ bacteria), whereas MTAP and AdoHcyase complete another (mostly mammals, some bacteria). MTAN and AdoHcyase complete the third pathway (LuxS bacteria, protozoa, plants). These results indicated that the substrate preference of MTAN for MTA and SAH degradation pathway evolved with the activated methyl cycle. Escherichia coli and V. cholerae (LuxS+) MTANs showed comparable activity towards substrate MTA and SAH, while B. thailandensis (LuxS) and A. thaliana MTANs showed obviously substrate preference for MTA. We considered that MTANs of LuxS bacteria probably have different degrees of substrate preference for MTA as is reported for the plant A. thaliana (Siu et al., 2008).

MTAN and AdoHcyase play very crucial functions by regulating the autoinducers produced and transmethylation activity in bacterial-activated methyl cycle. Inhibition of MTAN would prevent autoinducer production and block quorum sensing, generating novel antimicrobials against pathogenic bacteria (Kamath et al., 2006; Cornell et al., 2009). Insight into the B. thailandensis MTAN three-dimensional structure will give clearer explanation for the substrate preference of MTAN and facilitate drug design against melioidosis.

Acknowledgements

This project was supported by a grant (24GFCX-YJ-2306) from the Chinese Academy of Sciences and National S&T Major Project (2012ZX10004403 and 2009ZX09301-14).

Ancillary