OUP user menu

Comparative in vivo and in vitro analyses of putative virulence factors of Burkholderia pseudomallei using lipopolysaccharide, capsule and flagellin mutants

Chanthiwa Wikraiphat, Jaruek Charoensap, Pongsak Utaisincharoen, Surasakdi Wongratanacheewin, Suwimol Taweechaisupapong, Donald E. Woods, Jan G.M. Bolscher, Stitaya Sirisinha
DOI: http://dx.doi.org/10.1111/j.1574-695X.2009.00574.x 253-259 First published online: 1 August 2009


Burkholderia pseudomallei is a gram-negative bacillus that is the causative agent of melioidosis. We evaluated host–pathogen interaction at different levels using three separate B. pseudomallei mutants generated by insertional inactivation. One of these mutants is defective in the production of the polysaccharide side chains associated with lipopolysaccharide; one does not produce the capsular polysaccharide with the structure -3)-2-O-acetyl-6-deoxy-β-d-manno-heptopyranose-(1-; and the third mutant does not produce flagellin. We compared the in vivo virulence in BALB/c mice, the in vitro fate of intracellular survival inside human polymorphonuclear cells (PMNs) and macrophages (Mφs) and the susceptibility to killing by 30% normal human serum, reactive nitrogen and oxygen intermediates and antimicrobial peptides with that of their wild-type counterpart. The lipopolysaccharide and capsule mutants demonstrated a marked reduction in virulence for BALB/c mice, but the flagellin mutant was only slightly less virulent than the parent strain. The results from the BALB/c mice experiments correlated with survival in Mφs. The lipopolysaccharide and capsule mutants were also more susceptible to killing by antimicrobial agents. All bacteria were equally susceptible to killing by PMNs. Altogether, the data suggest that lipopolysaccharide and capsule and, to a much lesser extent, flagella, are most likely associated with the virulence of this bacterium and highlight the importance of intracellular killing by PMNs and Mφs in disease pathogenesis.

  • Burkholderia pseudomallei
  • polymorphonuclear cell
  • macrophage
  • capsule
  • lipopolysaccharide
  • flagella


Burkholderia pseudomallei is a facultative gram-negative bacillus that is the causative agent of a potentially fatal disease known as melioidosis (White, 2003; Wiersinga et al., 2006). Clinical manifestations vary from asymptomatic infection to septic shock. The latter can have a mortality rate as high as 80% despite appropriate antibiotic treatment. The pathogenesis of melioidosis remains to be clarified, but different lines of evidence suggest that innate immunity is important in determining the outcome of infection (Wiersinga et al., 2006). It was shown, for example, that interferon-γ (IFN-γ) played an essential role in resistance in a murine model of infection (Santanirand et al., 1999). Enhancing IFN-γ production before infection via CpG oligodeoxynucleotide administration protected animals against fatal challenge (Wongratanacheewin et al., 2004). Macrophages (Mφs) preactivated with IFN-γ could readily kill B. pseudomallei (Miyagi et al., 1997). More recently, the role of polymorphonuclear cells (PMNs) has attracted the attention of many groups of investigators (Easton et al., 2007; Wiersinga et al., 2007, 2008a, b), but attempts to activate these cells with granulocyte colony-stimulating factor in patients with severe sepsis due to melioidosis have given rise to contradictory and questionable results (Cheng et al., 2007).

Although considerable attention has been paid to identifying and characterizing the virulence factors of B. pseudomallei, no definitive conclusions have been reached. Several putative virulence factors have been implicated in disease due to B. pseudomallei including lipopolysaccharide, the capsular polysaccharide with the structure -3)-2-O-acetyl-6-deoxy-β-d-manno-heptopyranose-(1- (hereafter referred to as the mannoheptan capsule), flagella, the type III secretion system, phospholipase C and several other toxins (Cheng & Currie, 2005; Wiersinga et al., 2006). In the present study, we used three B. pseudomallei mutants that we have developed to compare their virulence in murine infection models, their interaction with human PMNs and Mφs, as well as their susceptibility to killing by antimicrobial agents in order to obtain more informative data regarding virulence and pathogenesis. The results suggest that lipopolysaccharide and the mannoheptan capsule, but not flagella, are relevant to the virulence of this bacterium.

Materials and methods

Bacterial cultivation and characterization

The lipopolysaccharide mutant SRM117 (DeShazer et al., 1998), the mannoheptan capsule mutant SR1015 (mutated in the wcbB gene, which encodes for a protein with homology to a glycosyltransferase, Reckseidler et al., 2001) and flagellin mutant MM35 (DeShazer et al., 1997) were generated from wild-type B. pseudomallei strain 1026b by insertional mutagenesis. With regard to the insertional mutagenesis, all the mutant strains contained a tetracycline resistance cassette (Tcr) to indicate their insertional inactivation. All strains were cultured at 37 °C in Luria–Bertani (LB) medium for 16–18 h to obtain bacteria in the late logarithmic phase of growth before harvesting and use in the study. If not indicated otherwise, the medium for culturing the mutants was always supplemented with tetracycline at a concentration of 50 µg mL−1 to prevent possible reversion to the wild type. Under these conditions of growth, the doubling times for all strains were calculated and found to be c. 48 min. The number of viable bacteria, expressed as CFU, was determined by serial dilution and the pour-plating technique performed in duplicate using trypticase soy agar (TSA) in the absence of an antibiotic. The presence or absence of lipopolysaccharide, capsule or flagella in the stock cultures (Table 1) was confirmed as described (DeShazer et al., 1997; Anuntagool et al., 2000; Sirisinha et al., 2000). Biofilm production was also determined (Taweechaisupapong et al., 2005). Lethal dose 50% (LD50) determinations in BALB/c mice for all four strains were performed as described earlier (Wongratanacheewin et al., 2004).

View this table:
Table 1

Phenotypic characteristics of Burkholderia pseudomallei wild type and mutants

Reactive with mAb to
Bacterial strainsLD50 (CFU) (BALB/c, IP)Biofilm formation (OD630 nm)200 kDalipopolysaccharideMotile
Wild type1040.807+++
Capsule mutant>1091.296++
Lipopolysaccharide mutant1082.461++
Flagellin mutant3.16 × 1040.756++
  • +, produces capsule or lipopolysaccharide, motile; −, does not produce capsule or lipopolysaccharide, nonmotile; IP, intraperitoneal challenge.

Mφ and PMN preparations

Mφs and PMNs were isolated from heparinized whole blood of healthy human volunteers by Ficoll–Hypaque density gradient separation (ethical clearance was provided by the Ethical Committee of the Ramathibodi Hospital, Mahidol University, Bangkok, Thailand, clearance number 2549/452). In brief, peripheral blood mononuclear cells were harvested from the fluid at the interface of the separation tube, and monocytes were purified by positive selection of CD14+ cells using a magnetic cell sorter system (MACS@ Miltenyi Biotec, GmbH, Germany). The Mφs were generated from the purified monocytes by culturing the latter in Iscove's modified Dulbecco's medium supplemented with 1%l-glutamine and 10% heat-inactivated human AB serum (GemCell™, West Sacramento, CA) at 37 °C for 5–7 days in a humidified 5% CO2 incubator. The fluid in the lower layer from the Ficoll–Hypaque separation tube was used for PMN preparation (Watson et al., 1992). The PMNs were then purified using 3% Dextran T500 (GE Healthcare, BioSciences, Uppsala, Sweden). Residual contaminating red blood cells were lysed in a hypotonic solution. The PMNs were resuspended in DMEM culture medium supplemented with 10% heat-inactivated fetal bovine serum (HyClone, Logan, UT) and 1%l-glutamine and used immediately.

Survival and replication of B. pseudomallei wild type and mutants in PMNs and Mφs

PMNs (2 × 106 cells) and Mφs (1 × 105 cells) were infected with the bacteria as described previously (Kespichayawattana et al., 2000), using a multiplicity of infection of 2 : 1 and 1 : 1, respectively. Briefly, after 2 h of infection, extracellular bacteria were killed by adding kanamycin to the cell culture to a final concentration of 250 µg mL−1. Two hours later, the antibiotic concentration was changed to 20 µg mL−1. In addition to the kanamycin, for experiments using the mutants, tetracycline at a concentration of 50 µg mL−1 was also present. The number of viable intracellular bacteria at different time intervals was determined by lysing the infected cells with 0.1% Triton X-100, and the number of bacteria in the lysate was quantified as described above. At the end of all experiments, the viability of the infected cells was determined by trypan blue staining, and experiments with viability <90% were discarded. Data were analyzed using a paired-samples t-test. P-values <0.05 were considered statistically significant.

Host-cell free killing assays

The sensitivity of the three mutants and the wild type to killing by 30% fresh pooled normal human serum (NHS) and of human neutrophil peptide-2 (HNP-2) was determined as described, respectively, by DeShazer (1998) and West (2005). In brief, 106 CFU mL−1 of late log-phase bacteria were incubated in LB broth containing 30% fresh or heat-inactivated pooled human serum (obtained from donors living in a nonendemic area of infection), 10 µg mL−1 of HNP-2 or in phosphate-buffered saline for 2 h before residual bacteria were determined. In addition to HNP-2, the killing activities of other antimicrobial peptides, namely, cathelicidin (LL-37), histatins (histatin5 and its variant, dhvar5) and lactoferrin peptides (LFcin 17-30 and LFampin 268-284), were also investigated. These peptides were synthesized as described (Den Hertog et al., 2004; van der Kraan et al., 2004). Bacteria (1 × 107 CFU mL−1) in 1 mM potassium phosphate buffer (PPB), pH 7.0, were added to equal volumes of the antimicrobial peptide solutions. The final concentration of all the peptides was 100 µM. A bacterial suspension in PPB without a peptide served as a control. After incubation at 37 °C for 1 h, the number of residual bacteria was determined and the percentage of killing was calculated using the formula [1−(CFU sample/CFU control)] × 100%. Differences in susceptibility to each antimicrobial peptide of each mutant compared with that of the wild type were analyzed using Student's t-test. Statistical significance was considered when the P-value was <0.05.

Susceptibility to killing by reactive nitrogen (RNI) and oxygen (ROI) intermediates

The sensitivity of the bacteria to RNI killing was assessed by culturing them in different concentrations of nitric oxide (NO)-generating compounds, S-nitrosoglutathione (GSNO) or 3-morpholinosydnonimine hydrochloride (SIN-1) (Chakravortty et al., 2002). After 4 h of incubation, residual bacteria were determined by serial dilution and plating on TSA. The concentrations that killed 50% of the bacteria were then calculated and expressed as lethal concentration 50%. H2O2 was used to test their sensitivity to ROI. For this test, the bacteria were similarly cultured in LB broth that contained different concentrations of H2O2, and the number of residual bacteria was determined after 2 h of incubation.


Virulence of B. pseudomallei wild type and mutants for BALB/c mice

The LD50 values of the wild type and mutants in BALB/c mice are shown in Table 1. The virulence of lipopolysaccharide and capsule mutants was markedly reduced compared with that of the wild-type counterpart or the flagellin mutant. Judging from these values, the capsule appeared to be the most important virulence factor in this animal model, as its LD50 value decreased by >5 log10 compared with that of the wild type, and this was followed by that of the lipopolysaccharide mutant.

Intracellular survival and replication of B. pseudomallei in human PMNs and Mφs

The fates of B. pseudomallei wild type and the three mutants after being internalized by human PMNs and Mφs were followed for c. 16–18 h when the experiments were terminated and host cells were still viable. Inside the PMNs, both the wild type and the mutants displayed similar survival profiles (Fig. 1). The absolute numbers of intracellular bacteria, expressed as CFU of viable count, gradually declined during these time intervals, and at the end of the experiments there was a reduction of between 1 and 2 log10 of viable intracellular bacterial counts for all strains compared with their initial 3-h time points. However, unlike the results with the PMNs, the Mφs could not suppress the growth of the B. pseudomallei wild type and at the end of the experiment the absolute numbers of intracellular bacteria increased by at least 1 log10. Its doubling time in the Mφs was about 1 h 50 min, a value similar to that obtained with other wild-type strains that we have tested (Charoensap et al., 2009). The profile of the intracellular lipopolysaccharide mutant, on the other hand, was markedly different from that of the wild type, and by the end of the experiment the residual numbers of the lipopolysaccharide mutant were significantly lower than those of the wild type by >2 log10 (P=0.036). Hence, the oligosaccharide component of the lipopolysaccharide appeared to be one of the components of B. pseudomallei that conferred resistance against intracellular killing by human Mφs.

Figure 1

Intracellular survival kinetics of Burkholderia pseudomallei wild type and mutants inside human PMNs (a) and Mφs (b). The results are expressed as mean and SE of the mean of absolute number of intracellular bacteria (CFU), each from five different donors. Survival data of the wild type and the three mutants inside Mφs were also expressed as percent internalization (c).

It is apparent from Fig. 1b that the numbers of the capsule mutant at the 16-h time point were markedly different from those of the wild type. The difference was more obvious when these end-point values were transformed into the percentage of their respective initial values at the 3-h time point (P<0.01). When calculated, the doubling time of the capsule mutant was considerably prolonged, i.e., >5 h. The survival profile of flagellin mutant, on the other hand, was almost the same as that of the wild type; however, it appeared to replicate slightly slower, having its doubling time slightly prolonged from 1 h 50 min found with the wild type to 2 h 16 min. Altogether, the survival data of B. pseudomallei in the Mφs were in agreement with the results obtained from the LD50 values in animals in that both the lipopolysaccharide and the capsule appeared to be more related to virulence than the flagella.

Susceptibility of B. pseudomallei wild type and mutants to killing by NHS

The contribution of lipopolysaccharide, capsule and flagella to resistance to killing by NHS was analyzed by simultaneously testing all four strains in a cell-free system in the presence or absence of 30% fresh NHS. The data summarized in Table 2 showed that the capsule and flagellin mutants exhibited similar degrees of resistance to killing as that of the wild type. The level of resistance to killing of the lipopolysaccharide mutant, on the other hand, decreased significantly (P=0.006) compared with other strains. At the end of the 2-h exposure, essentially all lipopolysaccharide mutant organisms were killed. The bactericidal activity of NHS for B. pseudomallei appeared to be associated with heat-labile component(s), as its activity could be totally destroyed by heating at 56 °C for 30 min (Table 2).

View this table:
Table 2

Lethal concentration 50% (LC50) values and susceptibility levels (with ± SD) of Burkholderia pseudomallei wild type and mutants to killing by serum and different antimicrobial agents

Residual bacteria (CFU)% Killed (n=6)LC50
Histatins (100 µM)Lactoferrins (100 µM)Cathelicidin (100 µM)RNIs (mM)ROI (µM) (n=4)
Bacteria30% NHS30% heat-inactivated NHSPBSHistatin 5Dhvar5LFcin 17-30LFampin 268-284LL-37SIN-1GSNOH2O2
Wild type1.14 × 1061.23 × 1061.58 × 10658.0 ± 6.939.3 ± 9.225.2 ± 7.920.7 ± 18.999.9 ± 0.100.652.7536 ± 6
Capsule mutant1.11 × 1061.18 × 1061.45 × 10678.1 ± 6.174.7 ± 5.971.1 ± 9.480.5 ± 7.599.5 ± 0.060.552.7273 ± 14
Lipopolysaccharide mutant411.22 × 1061.53 × 10658.1 ± 9.780.0 ± 9.560.2 ± 18.157.8 ± 19.999.9 ± ± 3
Flagellin mutant1.45 × 1061.19 × 1061.81 × 10640.7 ± 21.851.9 ± 11.138.9 ± 13.816.8 ± 19.399.7 ± 1.021.853.6555 ± 13
  • P<0.01 compared with strain 1026b.

  • PBS, phosphate-buffered saline.

Susceptibility of B. pseudomallei wild type and mutants to killing by antimicrobial peptides

The susceptibility of B. pseudomallei wild type and the three mutants to killing by different antimicrobial peptides was investigated. While both the wild type and all three mutants were found to be resistant to killing by 10 µg mL−1 of HNP-2 (data not shown), all strains were highly susceptible to killing by 100 µM of cathelicidin LL-37 (Table 2). On the other hand, when compared with their wild-type counterpart, only the lipopolysaccharide and the capsule mutants showed enhanced susceptibility to killing by histatins and lactoferrin peptides (Table 2), whereas the flagellin mutant appeared to be indistinguishable from the wild type with regard to this susceptibility.

Susceptibility of B. pseudomallei wild type and mutants to killing by RNIs and ROI

Although these intermediates are known to be generated intracellularly by some phagocytic cells, in order to avoid misinterpretation of the data, the experiments were performed in a cell-free system. The results presented in Table 2 showed that, compared with the wild type, the lipopolysaccharide and, to a lesser extent, the flagellin mutant, appeared to have enhanced resistance to killing by the NO-generating compounds. With the ROI testing using H2O2, only the capsule and, to a lesser extent, flagellin mutants appeared to have enhanced resistance against killing. On the other hand, there was a slight, if any, increase in susceptibility for the lipopolysaccharide mutant compared with the wild type (Table 2). These observations altogether suggested that the lipopolysaccharide component may be one of the inherent properties associated with the resistance and susceptibility of B. pseudomallei to killing by RNIs and ROI and, as a consequence, may influence the outcome of the organism's intracellular fate in the phagocytes. The capsule and, to a lesser extent, the flagellin mutants appeared to be more resistant to ROI killing than the wild type. The latter also had a considerable delay in stimulating respiratory burst in human PMNs compared with the wild type and the other two mutants, judging from a chemiluminescence assay using 200 µM luminol (data not shown).


In the present study, we attempted to provide additional evidence to further identify and verify the possible role of three putative virulence factors, namely, lipopolysaccharide, capsule and flagella, in disease caused by B. pseudomallei. This was done by comparative analyses of the three respective mutants with their wild-type counterpart in inducing disease in susceptible BALB/c mice, their intracellular fates in human PMNs and Mφs and their susceptibility to killing by serum, antimicrobial peptides, ROI and RNIs. The results from our studies strongly suggest that the lipopolysaccharide and capsule, and to a much lesser extent flagella, are associated with the virulence of this bacterium. Moreover, because both the lipopolysaccharide and the capsule mutants demonstrated noticeably higher biofilm production than either the flagellin mutant or the wild type (Table 1), it appears that biofilms do not play a significant role in the disease process. This conclusion is consistent with our previous report using a B. pseudomallei wild type possessing different biofilm contents and a biofilm mutant (Taweechaisupapong et al., 2005).

The observation that the lipopolysaccharide represented one of the virulence components of B. pseudomallei is not really unexpected (Table 1 and Fig. 1) because it was suggested previously by our own group (DeShazer et al., 1997; Arjcharoen et al., 2007). In one of these reports, we showed that compared with the wild type, this same lipopolysaccharide mutant was more susceptible to killing during the early phase of infection by a mouse Mφ cell line RAW 264.7 and the data obtained suggested that this may be related to its ability to activate IFN-β and iNOS, both of which are known to upregulate NO production that killed this intracellular bacterium (Arjcharoen et al., 2007). It could be argued, however, that in that report a mouse cell line was used and so the conclusion must be interpreted with caution. Therefore, in the present study, we similarly tested host–bacterial interactions using freshly isolated human phagocytic cells infected with B. pseudomallei wild type and three different mutants. With the human system, we found again that the lipopolysaccharide mutant was more susceptible to killing by the Mφs than its wild-type counterpart during the entire period of observation (Fig. 1). We would like to suggest that the decreased virulence of the lipopolysaccharide mutant in our in vivo mouse model (Table 1) is a combined result of their intracellular fates in Mφs, together with enhanced susceptibility to killing by serum, antimicrobial agents and, to a lesser extent, by ROI (Table 2). The data presented also allow us to conclude that the capsule is another important virulence factor for this bacterium. Its ability to confer some degree of resistance against intracellular killing by human Mφs in particular sheds additional light on its possible mechanism of pathogenicity. The association of the capsule with virulence had been reported previously in another animal model (Reckseidler et al., 2001), but it was then suggested that this is probably attributable to the fact that the capsule is known to interfere with complement activation and subsequently phagocytosis of B. pseudomallei (Reckseidler et al., 2005). In addition to these possible mechanisms, our data in Fig. 1, showing that the capsule mutant survived rather poorly in the Mφs when compared with the wild type, provide another possible mechanism of resistance against killing. It should be mentioned that the slight increase in the number of viable capsule-mutant organisms in the lysate at the later time points (Fig. 1) was unlikely to be caused by a premature release of bacteria from dying infected cells into the supernatant fluids because the viability of these infected cells was still higher than 90% when the experiment was terminated. The ability of the capsule to confer enhanced resistance against intracellular killing may represent a novel mechanism of capsule function, and this phenomenon has never been reported before for B. pseudomallei. This observation is, however, in accordance with the results reported recently for Neisseria meningitidis (Spinosa et al., 2007). It was demonstrated further in that model that the presence of the capsule correlated with its enhanced resistance to killing by defensins, cathelicidins, protegerins and polymyxin B, a finding that is also similar to our data on the susceptibility of Burkholderia to killing by the antimicrobial peptides used in our study (Table 2). Moreover, the enhanced resistance of the capsule mutant to ROI killing as noted in the table may contribute to its prolonged survival in the Mφs. The importance of Mφs in host defense against B. pseudomallei infection was convincingly demonstrated with the data showing that Mφ depletion by dichloromethylene biphosphonate-containing liposomes rendered mice more susceptible to B. pseudomallei infection (Breitbach et al., 2006). Consistent with the animal study reported earlier (Miyagi et al., 1997), we have recently demonstrated in the human system that although B. pseudomallei was able to survive intracellular killing inside the Mφs, the killing capacity could be readily enhanced after activation by IFN-γ (Charoensap et al., 2009). On the other hand, the role of PMNs should not be overlooked as it was shown earlier that they were needed in defense against B. pseudomallei infection in a mouse model (Easton et al., 2007; Wiersinga et al., 2008a, b). Our results (Fig. 1), demonstrating the efficacy of PMNs in killing B. pseudomallei in vitro, supported its potential role in host defense against B. pseudomallei infection.

Lastly, it should be mentioned that the role of flagella is yet to be established. For example, although we recently reported that flagella were needed for the invasion of B. pseudomallei into nonphagocytic cells (Chuaygud et al., 2008), it is inconsistent with the data in the present study using primary human phagocytic cell cultures. It is therefore possible to attribute the difference to the types of cells used in the two studies. Our in vivo data using flagellin mutant bacteria in BALB/c mice (Table 1) were also different from what was reported earlier by Chua (2003). The discrepancy between our results and those of the previous results may be explained by a difference in the route of bacterial challenge. In their study, they clearly demonstrated that flagella were a virulence determinant as their flagellin-defective mutant was avirulent for BALB/c mice when deposited intranasally. When challenged by the intraperitoneal route, however, as shown in the current studies, flagella are not required for virulence. In summary, while we showed here, using in vitro and in vivo experiments, that lipopolysaccharide and the capsular polysaccharide with the structure -3)-2-O-acetyl-6-deoxy-β-d-manno-heptopyranose-(1- are definitely important virulence factors for B. pseudomallei, the role of flagellin remains unsettled and needs to be investigated further.


This work was supported by grants from the National Science and Technology Development Agency (Thailand), the Royal Golden Jubilee PhD Program of the Thailand Research Fund, the Institutional Strengthening Program, Faculty of Science, and the RA Scholarship, Faculty of Graduate Studies, Mahidol University.


  • Editor: Richard Marconi


View Abstract