Molecular Characterization of Putative Virulence Determinants in Burkholderia pseudomallei

The Gram-negative saprophyte Burkholderia pseudomallei is the causative agent of melioidosis, an infectious disease which is endemic in Southeast Asia and northern Australia. This bacterium possesses many virulence factors which are thought to contribute to its survival and pathogenicity. Using a virulent clinical isolate of B. pseudomallei and an attenuated strain of the same B. pseudomallei isolate, 6 genes BPSL2033, BP1026B_I2784, BP1026B_I2780, BURPS1106A_A0094, BURPS1106A_1131, and BURPS1710A_1419 were identified earlier by PCR-based subtractive hybridization. These genes were extensively characterized at the molecular level, together with an additional gene BPSL3147 that had been identified by other investigators. Through a reverse genetic approach, single-gene knockout mutants were successfully constructed by using site-specific insertion mutagenesis and were confirmed by PCR. BPSL2033::Km and BURPS1710A_1419::Km mutants showed reduced rates of survival inside macrophage RAW 264.7 cells and also low levels of virulence in the nematode infection model. BPSL2033::Km demonstrated weak statistical significance (P = 0.049) at 8 hours after infection in macrophage infection study but this was not seen in BURPS1710A_1419::Km. Nevertheless, complemented strains of both genes were able to partially restore the gene defects in both in vitro and in vivo studies, thus suggesting that they individually play a minor role in the virulence of B. pseudomallei.


Introduction
Burkholderia pseudomallei is a Gram-negative motile bacillus, that is, a facultative anaerobe and an environmental saprophyte. It is readily recovered from the soil and surface waters in endemic areas, that is, Southeast Asia and Northern Australia [1]. This bacterium possesses a remarkable capacity to infect humans and animals, causing melioidosis which is an important cause of sepsis in the tropics [2]. It has been considered a potential bioweapon and so virulence factors that correspond to its pathogenesis are being intensively studied at an increasing rate.
The availability of complete genome sequence database of organisms allows researchers to discover many molecules and mechanisms that may be involved in the virulence of B. pseudomallei [3,4]. Several approaches, including subtractive hybridization [5], comparative genomics [6], signaturetagged mutagenesis [7], transposon mutagenesis [8], in vivo expression technology [9], microarray [10], and computational methods [11,12], have accelerated the discovery of virulence factors over the past decades. The availability of the genomic sequences of several B. pseudomallei strains has rapidly added candidate virulence genes to databases.
In general, past studies have focused on genomic differences between species, that is, the virulent B. pseudomallei and a closely related but avirulent family member B. thailandensis [5,6]. PCR-based subtractive hybridization was recently undertaken in our laboratory using a virulent clinical isolate B. pseudomallei (v) and an attenuated strain of the same B. pseudomallei isolate (av) [13]. PCR-based 2 The Scientific World Journal subtractive hybridization successfully demonstrated 6 subtracted DNA fragments that were unique to the virulent strain of B. pseudomallei, whereas these DNA fragments were not seen in its "attenuated" strain [13]. Sequencing of the subtracted DNA fragments revealed 6 unique genes with unknown functions as follows: BPSL2033, BP1026B I2784, BP1026B I2780, BURPS1106A A0094, BURPS1106A 1131, and BURPS1710A 1419.
Besides the 6 putative "virulence" determinants, BPSL3147 is another potential virulence determinant identified by Cuccui et al. [7]. This gene encodes a putative lipoprotein and is reported to be a putative lipoprotein containing 39.16% amino acid that is identical to a VacJ lipoprotein in Shigella flexneri. The Tn10 mutant of S. flexneri YSH6000T VacJ lipoprotein was unable to spread from cell to cell, suggesting VacJ is important for intercellular spread of the organism [14].
The aim of this study was to extensively characterize the 6 putative virulence determinants which were absent in the "attenuated" strain (av) that is believed to have reduced virulence that occurred after several subcultures and longterm storage in the laboratory, together with an additional candidate, BPSL3147, using the same methodology, that is, gene knockout approach, using in vitro and in vivo assays.

Bacterial Strains, Media, and Culture Conditions.
Bacterial strains and plasmids used are listed in Table 1. The clinical B. pseudomallei strain was isolated from the blood of a patient CMS, at the University Hospital, University of Malaya, Kuala Lumpur, who died from melioidosis as described in previous report [11] and this strain was used throughout the study (henceforth referred to as Bp-CMS). The Escherichia coli strains DH10B, CC118 pir, S17-1 pir, vector pUT-Km, and chloramphenicol acetyltransferase (CAT) cassette were obtained from Prof. Dr. Wang Jin-Town (National Taiwan University, Taiwan). All strains were grown in Luria-Bertani (LB) medium at 37 ∘ C and, when appropriate, antibiotics were used at the following final concentrations: ampicillin 100 g/mL, kanamycin 50 g/mL, chloramphenicol 100 g/mL, and streptomycin 100 g/mL. The mouse leukaemic monocyte macrophage cell line RAW264.7 was obtained from American Type Culture Collection (ATCC, USA). It was cultured and maintained in flasks (Corning, USA) with DMEM (Gifco, USA) supplemented with 10% (v/v) fetal bovine serum (Gifco, USA), 4 mM L-glutamine, and an antibiotic mixture containing 100 U/mL penicillin and 0.1 mg/mL streptomycin at 37 ∘ C in 5% CO 2 . BPSL2033, BP1026B  I2784, BP1026B I2780, BURPS1106A A0094, BURPS1106A 1131, BURPS1710A 1419, and BPSL3147 were disrupted by homologous recombination using a suicide vector pUT-Km [15,16]. Briefly, a partial region of the gene to be inactivated (serving as a significant site of gene exchange through conjugation and recombinant events) was amplified by polymerase chain reaction (PCR) using primer pairs as listed in Table 2.

Construction of Mutants. Genes
The amplicon was subcloned into pGEM-T easy vector, digested with EcoRI, and subcloned into the suicide vector pUT-Km [15]. Selection of transformants by plating onto LB agar containing kanamycin and rapid screening for desired inserts by colony PCR was performed using primers KmF and KmR. The construct was then transformed into E. coli S17-1 pir and introduced into the wild-type B. pseudomallei (Bp-CMS) by conjugation. An insertion mutant was selected using LB agar supplemented with kanamycin and site-specific chromosomal integration was verified by PCR using vectorand corresponding gene-specific primer pairs.

Construction of Complemented Plasmids.
Intact genes BURPS1710A 1419, BPSL2033, BP1026B I2784, BP1026B I2780, BURPS1106A A0094, BURPS1106A 1131, and BPSL3147 and their promoters were amplified by PCR and cloned into a pGEM-T-CAT plasmid. These complemented plasmids were then reintroduced into the corresponding insertion mutants by transformation. The resulting complemented strains were selected using LB agar containing chloramphenicol.

Bacterial Growth
Curve. The growth of B. pseudomallei strains in LB broth was monitored over 8 h by taking 1 mL of culture broth every hour to perform OD600 readings. The wild-type Bp-CMS was used as a positive control for growth in LB.

Bacterial Replication
Assays. Intracellular bacterial survival and replication was assayed in the mouse macrophagelike cell line RAW 264.7. Cells were seeded at a density of 1 × 10 5 cells/well into 24-well culture plates (Corning, USA) and incubated overnight at 37 ∘ C in 5% CO 2 . An overnight culture of B. pseudomallei strain was diluted 1 : 100 and grown at 37 ∘ C for 3 h with shaking to reach mid-log phase. Cell monolayers were then washed twice with PBS and incubated in fresh DMEM without antibiotic for 1 h prior to infection with bacteria. The bacterial suspension was added at a MOI of 100 : 1 and the coculture was immediately centrifuged at room temperature at 170 ×g for 5 min to bring the bacteria in direct contact with the host cells. After 1 h, the cells were washed twice with PBS and incubated in fresh DMEM containing 300 g/mL of tetracycline to suppress the growth of residual extracellular bacteria. Tetracycline was used instead of gentamicin, as Bp-CMS is resistant to gentamicin. At 2, 4, and 8 h after infection, infected monolayers were washed twice with PBS and lysed with 0.1% (v/v) Triton X-100 for 15 min. Serial dilutions of the released bacteria (expressed as colony forming units CFU) were plated on tryptic soy agar (TSA) plates to enumerate bacterial loads by direct colony counts. The number of internalized bacteria obtained at 2 h after infection represented the initial entry of bacteria, whereas at 4 and 8 h after infection represented intracellular bacterial replication. Bacterial survival was normalized to counts obtained at 2 h after infection and the relative survival rate presented as a percentage.
2.6. Virulence Testing with C. elegans. The wild-type C. elegans N2 was obtained from Carolina Biological Supply Company (USA). The nematode was propagated on nematode Killing assays were performed as previously described by Tan et al. [17] with minor modifications. All nematodes were agesynchronized by a bleaching procedure prior to the killing assay [18]. All B. pseudomallei derived strains (wild-type Bp-CMS, 7 insertion mutants, and 7 complemented strains) and E. coli OP50-1 were grown overnight at 37 ∘ C and 40 L of each culture was spread on NGM plates containing 50 g/mL 5-fluorodeoxyuridine (Merck, USA) to inhibit the eggs of C. elegans from hatching. Plates were incubated at 37 ∘ C 4 The Scientific World Journal  to gentle touch by the worm pick. E. coli OP50-1 was used as a negative control. The resulting data was analyzed with GraphPad Prism 5 software and plotted using the Kaplan-Meier survival plot.

Insertion Mutant Construction and Growth.
Positive chromosomal integration mutants were successfully constructed for the 7 candidate genes, on the first or second attempt, and a representative of construct is shown in Figure 1. All 7 mutant strains demonstrated similar growth rates to the parental strain in liquid media over 8 h.  (Figures 2(a) to 2(c)).

Discussion
There was no difference in the growth rates of all 7 mutants compared to that of the wild type when assayed in rich media (data not shown), ruling out the possibility that these genes are not affecting the growth but involved in other aspects such as pathogenicity. The infection assay on phagocytic cells showed that, without the presence of gene BPSL2033 (putative transport-related membrane protein), the mutants demonstrated reduced ability to replicate and survive over 8 h. This result was supported by a plasmid-encoded complemented strain, which demonstrated partial restoration of 7 intracellular survival when compared to wild-type Bp-CMS. Similar outcomes were seen with BURPS1710A 1419 (putative lipoprotein) and BPSL3147 (lipoprotein). At 8 h post infection, loss of gene BPSL2033 showed a 5-fold reduction in survival time in RAW264.7 cells while both BURPS1710A 1419 and BPSL3147 resulted in a 3-fold reduction in survival, indicating that these genes may be involved in intracellular survival. There was a weak but still significant difference ( = 0.049) exhibited by BPSL2033::Km compared to Bp-CMS. Thus, BPSL2033 may act in concert with other genes and play an essential role in virulence. Subtractive hybridization, as reported in our earlier study, demonstrated that Bp-CMS contained 6 DNA sequences that were not found in the attenuated strain, indicating that maximal virulence probably requires multiple genes acting together in concert [13]. This hypothesis is further supported by the present study in which B. pseudomallei virulence in the phagocytic cell line model was not critically dependent on any single putative gene tested.
Several studies have suggested that a double mutant and not a single mutant of B. pseudomallei contributed significantly to the growth inside murine macrophage [19,20]. Future experiments involving the use of double mutants (i.e., BPSL2033 and BURPS1710A 1419) may prove this possibility. In the context of BPSL3147, there may have been other unidentified gene(s) acting together for full virulence in B. pseudomallei infections. It is unclear why these genes BPSL2033, BURPS1710A 1419, and BPSL3147 with their corresponding complemented plasmids only restored intracellular survival and replication to approximately 50% of the wildtype level. One possible explanation for this may be due to the gene being present on multiple copies of the plasmid.
C. elegans has been used as a simple surrogate host for modeling bacterial diseases [21]. It has been shown that on a low nutrient nematode growth medium (NGM) B. pseudomallei killed C. elegans strain N2 within 3 days and this type of killing is referred to as "slow killing" [22][23][24]. In our study, wild-type strain Bp-CMS killed 74% of the nematode population at 72 h time point in NGM. Our results are in agreement with published reports that differences in killing efficiency of C. elegans occur among wild-type strains of B. pseudomallei [23,25,26]. For instance, the percentage of killing of worms by various B. pseudomallei strains, that is, ATCC23343, EY4, number #40, and KHW, was approximately 50, 60, 75, and 90%, respectively, at 72 h time points [23]. Virulence of B. pseudomallei for the nematode is likely to be variable due to the different genetic determinants in the strains.
In the present study, all the constructed mutants were less effective in killing C. elegans under slow-killing conditions; that is, twice as many worms survived when fed on mutant strains compared to worms fed on wild-type Bp-CMS after 72 h of coculture. However, there was no significant difference of C. elegans killing between the mutants and wild type. Two complemented strains of BPSL2033::Km and BURPS1710A 1419::Km achieved at least partial restoration of virulence at 72 h coculture, thus suggesting that both genes are most probably involved in bacterial virulence.
The low level attenuation of virulence in the mutants, compared to the wild type, may possibly be due to the following.
(i) A single gene was insufficient to mediate full killing in the animal model. At least 2 genes are probably required to act together for achieving virulence in B. pseudomallei. A double mutant ΔrelAΔspoT was reported to exhibit significant and severe attenuation in larva of the wax moth Galleria mellonella and C57BL/6 in mice models, which was not seen with the single mutant [20].
(ii) C. elegans possesses mechanisms to avoid or move away from pathogenic bacteria like B. pseudomallei. It has a simple nervous system that consists of 302 neurons that facilitate the identification of molecules, neurons, and circuits involved in their behavior [27]. It uses chemotaxis to find food on the plate and is able to discriminate food both physically based on size and chemically based on taste and olfaction [28]. Worms can modify their olfactory preference via the neurotransmitter serotonin to avoid odours from pathogens and this learning occurs with exposure as short as 4 h [29]. In our study, C. elegans was cultivated on E. coli OP50-1 that best supports growth, so the worms had already experienced good food, which might have increased their exploratory bahaviour when switching to very bad food such as B. pseudomallei, especially in leaving bahaviours. Nonetheless, it is impossible to raise worms on B. pseudomallei alone because of the virulence of the organism.
BPSL2033 is a 428-amino-acid protein with a molecular mass of 46 kDa as a transport-related membrane protein; BURPS1710A 1419 is a 74-amino-acid protein with a calculated molecular mass of 8 kDa which is a putative lipoprotein. Further bioinformatics analysis suggests that amino acids 23-324 of BPSL2033 encode a domain belonging to major facilitator superfamily (MFS). MFS transporters are ubiquitous and found in all classes of organisms and in several pathogens such as Francisella tularensis [30] and Legionella pneumophila [31]. Chatfiled and colleagues [32] have shown that MFS protein plays a role in virulence by promoting bacterial ironsiderophore import. The Phyre 2 model [33] predicts that BPSL2033 forms a major facilitator superfamily fold and shares very low sequence identity (16%) to glycerol-3-phosphate transporter protein from E. coli with known three-dimensional structure (PDB template: 1pw4A). The results imply that BPSL2033 might exhibit new structural and/or functional characteristics and further X-ray crystallography analysis may provide valuable information. At present, we postulate that BPSL2033 transports nutrients (probably glycerol-3phosphate) that are essential for the replication of B. pseudomallei. Database searches with the NCBI blastp tool identified BURPS1710A 1419, as a putative lipoprotein within genus level, is diverged from 37 to 82% with no conserved domain identified, as well as a lack of three-dimensional structural 8 The Scientific World Journal information, possibly suggesting that BURPS1710A 1419 gene product has new or different functional characteristics.
In conclusion, our results suggest that BPSL2033 and BURPS1710A 1419 individually are likely to contribute to a minor role in virulence and provide a basis for further characterization of their role in pathogenesis. We hypothesize that the combination effect of both genes can provide a clear virulence role in B. pseudomallei.