|Home | About | Journals | Submit | Contact Us | Français|
Burkholderia pseudomallei, the etiological agent of melioidosis, is a facultative intracellular pathogen. As B. pseudomallei is a gram-negative bacterium, its outer membrane contains lipopolysaccharide (LPS) molecules, which have been shown to have low-level immunological activities in vitro. In this study, the biological activities of B. pseudomallei LPS were compared to those of Burkholderia thailandensis LPS, and it was found that both murine and human macrophages produced levels of tumor necrosis factor alpha, interleukin-6 (IL-6), and IL-10 in response to B. pseudomallei LPS that were lower than those in response to B. thailandensis LPS in vitro. In order to elucidate the molecular mechanisms underlying the low-level immunological activities of B. pseudomallei LPS, its lipid A moiety was characterized using mass spectrometry. The major lipid A species identified in B. pseudomallei consists of a biphosphorylated disaccharide backbone, which is modified with 4-amino-4-deoxy-arabinose (Ara4N) at both phosphates and penta-acylated with fatty acids (FA) C14:0(3-OH), C16:0(3-OH), and either C14:0 or C14:0(2-OH). In contrast, the major lipid A species identified in B. thailandensis was a mixture of tetra- and penta-acylated structures with differing amounts of Ara4N and FA C14:0(3-OH). Lipid A species acylated with FA C14:0(2-OH) were unique to B. pseudomallei and not found in B. thailandensis. Our data thus indicate that B. pseudomallei synthesizes lipid A species with long-chain FA C14:0(2-OH) and Ara4N-modified phosphate groups, allowing it to evade innate immune recognition.
Burkholderia pseudomallei is the etiological agent of melioidosis, a bacterial disease endemic in certain tropical regions, especially in Southeast Asia and northern Australia (9, 11, 12, 13), but with an expanding geographical distribution (10, 23, 43). Infection results in a spectrum of clinical syndromes, ranging from chronic abscesses to acute septicemia (28). Despite the availability of intensive treatment with appropriate antibiotics (57), the fatality rates in countries in which the disease is endemic remain high and recurrence of infection is common (27). In Singapore, the mortality rates average 23.7% (33), although this rate can be as high as 46.5% (30, 34).
Lipopolysaccharide (LPS) is an outer membrane molecule of gram-negative bacteria and is the most common bacterial component that is implicated in initiating sepsis (3). Structurally, LPS is composed of an outer O-antigen-specific polysaccharide and an inner core oligosaccharide that is covalently linked to a lipophilic moiety termed lipid A. Lipid A has been described as being responsible for the endotoxic activity associated with LPS (32, 42). Recognition of LPS by the innate immune system triggers the production of proinflammatory cytokines by host cells, which aids in the clearance of the pathogen (56). However, overstimulation of host cells by LPS can lead to sepsis (29). Sepsis is a major cause of death in patients with melioidosis, which accounts for almost 20% of all community-acquired septicemias in northeastern Thailand (7). The LPS of B. pseudomallei has been implicated in its pathogenesis, as high concentrations of antibodies to LPS are associated with improved survival in severe melioidosis (8, 21). The use of LPSs as subunit vaccines was protective in a murine model of experimental melioidosis (38).
Despite its apparent role in sepsis, the LPS of B. pseudomallei has been shown to have low-level macrophage-activating activity in vitro, which was attributed to a delay in nitric oxide and tumor necrosis factor alpha (TNF-α) production (31, 51, 52), thus enabling the pathogen to evade macrophage killing. As lipid A is the endotoxic center of LPS (32), elucidation of the primary structure of lipid A may shed light on the molecular basis of the low-level immunological activities associated with B. pseudomallei LPS (31, 44, 51). In this report, the ability of LPS from B. pseudomallei to activate macrophages was compared to this ability of LPS from Burkholderia thailandensis, a close relative of B. pseudomallei that rarely causes disease in humans (16, 47). In addition, by using a combination of chemical and mass-spectrometric methods, the structures of lipid A from the two pathogens were compared. Collectively, our results provide insight into the mechanisms of B. pseudomallei virulence.
B. pseudomallei strain KHW, obtained from the B. pseudomallei collection at the Defense Medical and Environmental Research Institute, DSO National Laboratories (Singapore), was isolated from a national serviceman who died of melioidosis in 1989. B. thailandensis (ATCC 700388) was obtained from the American Type Culture Collection. Both were grown on N-minimal medium (5 mM KCl, 0.5 mM K2SO4, 1 mM KH2PO4, 1.8 μM FeSO4·7H2O, 2 μg/ml thiamine-HCl, 0.1 M Tris-HCl, pH 7.4, 22 mM glucose) (37). The basal N-minimal medium was supplemented with 10 mM MgSO4.
To assess the LPS purification procedures, B. pseudomallei strain K96243 and a K96243 mutant strain (SB04/3518) (kindly provided by Timothy P. Atkins, Defence Science and Technology Laboratory, United Kingdom) were included as controls in this study. The mutant strain has been confirmed to be of a capsular polysaccharide I (CPS I)-negative phenotype (Timothy P. Atkins, personal communication).
LPS was extracted using an LPS extraction kit (Intron Biotechnology, Korea) with some modifications. Briefly, cells were harvested and lysed in lysis buffer (50 mg of cells/ml of lysis buffer) and then subjected to a vigorous vortex to dissolve cell clumps. After the addition of chloroform, the sample was centrifuged for 45 min at 4°C. The upper aqueous layer was collected, and 2 volumes of purification buffer were added to 1 volume of aqueous layer. The mixture was then incubated at −20°C for 2 h, before centrifugation at 20,000 × g for 10 min at 4°C. The resulting pellet was washed twice with 70% ethanol before being lyophilized overnight. Extracted LPS was stored at 4°C. LPS samples were further purified to remove contaminating proteins, nucleic acids, phospholipids, and Toll-like receptor 2 (TLR2) contaminating proteins as described previously (49).
To obtain the whole-cell LPS profile, one full loop of bacterial lawn was harvested from agar plates and 50 μl of solubilization buffer was added to every milligram of bacteria. The bacterial cells were resuspended by use of a vortex to dissolve cell clumps, followed by heating at 99°C for 10 min to kill the bacteria. After the suspension was cooled, an equal volume of solubilization buffer containing 3 mg/ml of proteinase K was added and heated at 60°C for 2 h. Proteinase K was inactivated by boiling for 10 min, and the sample was stored at −20°C until use.
Total fatty acids (FA) in the LPS were liberated and derivatized to FA methyl esters by methanolysis in 2 M methanolic HCl (48) at 90°C for 18 h. An equal volume of 50% saturated NaCl solution was added, and the FA methyl esters were extracted with chloroform. Extracted methyl esters were characterized by gas chromatography-mass spectrometry (GC-MS). Pentadecanoic acid was used as an internal standard. GC-MS analysis was performed on a MAT95XL-T GC-MS spectrometer by using an Rtx-5MS column (30 m by 0.25 mm by 0.25 μm).
Lipid A was isolated from LPS by using modified mild acid hydrolysis (58). Briefly, 1 mg of LPS was dissolved in 500 μl of 1% sodium dodecyl sulfate in 10 mM sodium acetate, pH 4.5, and heated at 100°C for 1 h. The mixture was dried under a vacuum. For the removal of sodium dodecyl sulfate, the sample was resuspended in 100 μl of distilled water and sonicated in an ultrasonic bath for dispersal of the sample. Five hundred microliters of acidified ethanol was added to the suspension and centrifuged at 2,000 × g for 10 min. The pellet was further washed twice with 500 μl of nonacidified ethanol and lyophilized to give fluffy, white, solid lipid A.
Electrospray ionization-MS (ESI-MS) was performed on a Micromass Q-Tof micro-mass spectrometer (Waters Corp., Milford, MA). The capillary voltage and sample cone voltage were maintained at 3.0 kV and 55 V, respectively. The source temperature was 80°C, and the desolvation temperature was set at 250°C. Mass spectra were acquired in the negative-ion mode. Isolated lipid A was resuspended in 2:1 (vol/vol) chloroform-methanol containing 3% of 300 mM piperidine as an ion signal enhancer (45). The sample was directly infused into the mass spectrometer at a flow rate of 10 μl/min.
Negative ESI-tandem MS (MS-MS) analyses were carried out under similar conditions as described above except for collision energies. Most of the major ion peaks, including ions at m/z of 965.60, 900.08, 852.52, and 786.99, were further investigated using MS-MS, either through ESI-MS-MS or high-performance liquid chromatography-ESI-MS-MS, with collision energies ranging from 35 to 80 V. Argon was used as the collision gas.
Mouse macrophage cells (RAW 264.7 cells) were seeded in 24-well plates at a density of 2.5 × 104 cells/well and stimulated with increasing concentrations of LPS from Escherichia coli strain 055:B5 (Sigma, Singapore), B. pseudomallei strains KHW, K96243, and SB04/3518, and B. thailandensis strain ATCC 700388. The cell culture supernatant was collected at various time points and was analyzed for TNF-α, interleukin-6 (IL-6), and IL-10 release by using the Quantikine enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN). Results given are from 18 h poststimulation.
THP-1 human monocytes were seeded in six-well plates at a density of 1 × 106 cells/well and were prestimulated with 10 ng/ml phorbol 12-myristate 13-acetate to differentiate into macrophage-like cells for 24 h. The differentiated cells were washed with phosphate-buffered saline, and after a recovery phase of 24 h, the cells were stimulated with increasing concentrations of the various LPSs as indicated. The cell culture supernatant was analyzed as described above. Results given are from 24 h poststimulation.
HEK293 cell lines, transfected with either human TLR2 (hTLR2)-CD14 complex or hTLR4-MD2-CD14 complex, were purchased from InvivoGen (San Diego, CA) and cultured in Dulbecco's modified Eagle medium containing 10% fetal calf serum and antibiotics (10 μg/ml of blasticidin and 50 μg/ml of Hydrogold) to maintain the transfected plasmids. Untransfected HEK293 cells not expressing any TLR were used as a negative control. The cells were seeded at a density of 2 × 104 cells/well in 96-well plates without antibiotics and transfected transiently the next day with 20 ng of NF-κB reporter plasmid, pNifty2-luc (InvivoGen, San Diego, CA), and 5 ng of renilla luciferase (plasmid, pRL-TK; Promega, Singapore) by using FuGENE 6 (Roche, Singapore). An empty expression vector of pNifty2-luc, not containing any NF-κB binding sites, was used as a negative control. Twenty-four hours after transfection, cells were stimulated for 7 h with B. pseudomallei strain KHW LPS, B. thailandensis strain ATCC 700388 LPS, ultrapure TLR4 ligand Escherichia coli K-12 LPS (InvivoGen, San Diego, CA), and ultrapure TLR2 ligand lipoteichoic acid from Staphylococcus aureus (InvivoGen, San Diego, CA). The cells were lysed in passive lysis buffer (Promega, Singapore), and luciferase expression was determined using the Dual Luciferase assay kit (Promega, Singapore). Relative light units were obtained by normalizing the firefly luciferase values (pNifty2-luc) to the renilla luciferase values (pRL-TK).
All results were analyzed by Student's t test. Differences with P values of <0.05 were considered significant.
To determine whether antigen-presenting cells could differentiate B. pseudomallei LPS from B. thailandensis LPS, murine macrophage cell line RAW 264.7 was exposed to increasing concentrations of LPS and the supernatant was analyzed for TNF-α, IL-6, and IL-10 (Fig. (Fig.1).1). RAW 264.7 cells produced less TNF-α in response to B. pseudomallei LPS than they did in response to B. thailandensis LPS at concentrations above 1 ng/ml (P < 0.05). LPS from E. coli strain 055:B5, a potent stimulator of macrophage cells, was used as a control. LPSs from both B. thailandensis (P < 0.05) and B. pseudomallei (P < 0.05) were less potent in stimulating the murine macrophages than E. coli LPS. The patterns of IL-6 and IL-10 production were comparable to the pattern of TNF-α production. Overall, LPS from B. pseudomallei was approximately two times less potent in stimulating cytokines release in RAW 264.7 cells than B. thailandensis LPS.
THP-1 human monocytes, which were differentiated into macrophage-like cells, were stimulated with LPS from B. pseudomallei or B. thailandensis to determine if the differences in murine macrophage activation would be similarly observed in human macrophages. Differentiated THP-1 cells produced lower levels of TNF-α, IL-6, and IL-10 in response to B. pseudomallei LPS than they did in response to LPSs from both B. thailandensis (P < 0.05) and E. coli (P < 0.05) (Fig. (Fig.2).2). In contrast to the case for IL-6, E. coli LPS and B. thailandensis LPS induced similar levels of TNF-α and IL-10 release from the human macrophage-like cells.
To assess the LPS purification method and to ensure that the observed differences between the endotoxicities of B. pseudomallei LPS and B. thailandensis LPS were not confounded by the presence of capsular polysaccharide, RAW 264.7 and THP-1 cells were stimulated with increasing amounts of LPS extracted from B. pseudomallei strain K96243 and its CPS I-negative mutant, SB04/3518. No significant difference between the abilities of the two LPSs to induce TNF-α (Fig. (Fig.3),3), IL-6, and IL-10 (data not shown) production from THP-1 and RAW 264.7 cells was observed, indicating that the amount of LPS in the wild-type strain of K96243 was similar to that of the CPS I-negative strain, SB04/3518, prepared using the purification procedures. B. thailandensis LPS was similarly found to be a more potent LPS than K96243 and SB04/3518 LPSs (P < 0.05). Collectively, the results indicate that both murine and human macrophages respond differently to B. pseudomallei LPS relative to B. thailandensis LPS.
In order to determine the receptor specificity of B. pseudomallei LPS, HEK293 cells expressing either hTLR2-CD14 or hTLR4-MD2-CD14 were stimulated with LPS and the level of NF-κB activation was determined using a NF-κB-dependent luciferase reporter (Fig. (Fig.4).4). TLR4-dependent activation of NF-κB was identified in response to B. pseudomallei LPS (P = 0.001). LPS from B. thailandensis similarly elicited a robust TLR4-dependent activation of NF-κB (P = 2.8 × 10−4). The level of NF-κB induced by B. pseudomallei LPS was significantly lower than that induced by B. thailandensis LPS (P = 0.007). TLR2-dependent activation of NF-κB was not observed when the TLR2-transfected cells were stimulated with LPS from either B. pseudomallei or B. thailandensis. These results indicate that both B. pseudomallei and B. thailandensis LPSs specifically activate the TLR4 complex but not TLR2.
To investigate if differences in lipid A structure may contribute to the differential recognition by RAW 264.7 cells, the LPSs from B. pseudomallei and B. thailandensis were hydrolyzed and their lipid A moieties were subjected to a combination of structural analyses. FA analysis of B. pseudomallei LPS revealed the presence of tetradecanoic acid (C14:0), 2-hydroxytetradecanoic [C14:0(2-OH)], 3-hydroxytetradecanoic acid [C14:0(3-OH)], hexadecanoic acid (C16:0), and 3-hydroxyhexadecanoic acid [C16:0(3-OH)] in its lipid A (data not shown). The LPS of B. thailandensis contained all the above FA with the exception of 2-hydroxytetradecanoic acid [C14:0(2-OH)].
To further characterize the differences detected by FA compositional analysis and to investigate the heterogeneity of purified lipid A from B. pseudomallei and B. thailandensis, lipid A was analyzed by ESI-QTOF MS. Isolated lipid A was successfully detected by using piperidine as a signal enhancer. The negative-ion ESI-QTOF mass spectra of purified lipid A from B. pseudomallei revealed a heterogeneous mixture with major doubly charged ions at m/z of 973.59 and 965.60 (Fig. (Fig.5A),5A), indicative of penta-acylated species that were most likely diphosphorylated. Minor peaks of doubly charged ions at m/z of 908.07 and 900.08 were detected as well. The mass differences of 131 units between doubly charged ions at m/z of 965.60 and 900.08 and between ions at m/z of 973.59 and 908.07 indicate a likely difference of a 4-amino-4-deoxy-arabinose (Ara4N) residue for the corresponding lipid A pairs (5, 35, 46). The doubly charged ion at an m/z of 973.59 possessed one more hydroxyl (—OH) group than the corresponding lipid A with a doubly charged ion at an m/z of 965.60. Based on the accurate masses of the doubly charged species, the ions at m/z of 973.59, 965.60, 908.07, and 900.08 were assigned to lipid A species with molecular formulae of C96H182N4O31P2, C96H182N4O30P2, C91H173N3O28P2, and C91H173N3O27P2, respectively.
The negative-ion ESI-QTOF mass spectra of purified lipid A from B. thailandensis revealed a complex pattern as well (Fig. (Fig.5B).5B). In addition to the doubly charged ions at m/z of 965.60 and 900.08 detected in B. pseudomallei, the mass spectra of B. thailandensis showed major doubly charged ions at m/z of 852.52 and 786.99 and a minor doubly charged ion at an m/z of 834.55 (not labeled), which was indicative of a heterogeneous mixture of penta- and tetra-acylated species. The lipid A species represented by the doubly charged ion at an m/z of 973.59 was unique to B. pseudomallei and was not detected in B. thailandensis. The mass difference of 226 units between doubly charged ions at m/z of 965.60 and 852.52 and between ions at m/z of 900.08 and 786.99 could likely be attributed to a difference of one FA C14:0(3-OH) for the corresponding lipid A pairs. The doubly charged ions at m/z of 834.55 and 786.99 represent the sequential loss of Ara4N (change in m/z, 131) from the ion at an m/z of 900.08. Accordingly, the ions at m/z of 852.52, 834.55, and 786.99 were assigned to lipid A, with molecular formulae of C82H156N4O28P2, C86H164N2O24P2 and C77H147N3O25P2, respectively, based on the accurate masses of the doubly charged species.
The tandem mass spectra of lipid A fractions with doubly charged parental ions at m/z of 786.99, 852.52, 900.08, and 965.60 are shown in Fig. Fig.6.6. The MS-MS fragmentation spectrum of the abundant lipid A species (m/z of 965.60), which was present in both B. pseudomallei and B. thailandensis, showed a doubly charged daughter ion at an m/z of 900.07 (change in m/z, 131), corresponding to the loss of Ara4N (fragment B) from the doubly charged parental ion at an m/z of 965.60 (Fig. (Fig.6A).6A). The subsequent loss of ions with m/z of 131 and 80 indicated that the Ara4N residue was attached through a phosphate group (35). Losses of phosphorylated Ara4N (fragment C) were commonly observed among parental ions at m/z of 965.60, 900.08, 852.52, and 786.99 (Fig. 6A to D). The fragmented ion at an m/z of 210.21 (Fig. (Fig.6A)6A) refers to a loss of one water molecule (H2O) from phosphorylated Ara4N (fragment A). The doubly charged daughter ions at m/z of 852.48 and 739.5 (not labeled) arose from the sequential loss of FA C14:0(3-OH) from the doubly charged parental ion at an m/z of 965.60 (Fig. (Fig.6A).6A). Similar results were observed in the sequential MS-MS pattern of doubly charged parental ion from m/z of 900.08 to 786.99 and subsequently to 673.93 (not labeled) (Fig. (Fig.6B).6B). This indicates that FA C14:0(3-OH) is the major acyl chain attached to the disaccharide backbone. The presence of fragmented ion at an m/z of 243.31 further confirmed the presence of FA C14:0(3-OH) in the lipid A structure (Fig. (Fig.6A6A).
Using the information above, we hypothesized the structure for the major lipid A fraction at an m/z of 973.59, which was present only in B. pseudomallei (Fig. (Fig.7).7). This peak was consistent with a lipid A structure of a biphosphorylated disaccharide backbone modified with Ara4N at both phosphate groups and penta-acylated with a combination of FA C14:0(2-OH), FA C14:0(3-OH), and FA C16:0(3-OH). The substitution of FA in the lipid A backbone was not determined in this study and is not represented in any particular order in our hypothetical structure, as further nuclear magnetic resonance studies will be needed to confirm the positions of these FA. In contrast, the major lipid A species from B. thailandensis with a doubly charged ion at an m/z of 965.60 is likely to have the same structure as the ion at an m/z of 973.59 but with the substitution of C14:0 for FA C14:0(2-OH). This is consistent with the FA compositional analysis, which revealed the presence of C14:0(2-OH) only in the lipid A of B. pseudomallei.
Likewise, the lipid A with an m/z of 900.08, which was present in both B. pseudomallei (minor species) and B. thailandensis (major species), was assigned to be a penta-acylated lipid A with only one Ara4N group. The other lipid A species present in only B. thailandensis were assigned as follows: lipid A with an m/z of 852.52 is a tetra-acylated lipid A containing two Ara4N groups; lipid A with an m/z of 786.99 is a tetra-acylated lipid A containing one Ara4N group; and lipid A with an m/z of 834.55 is a penta-acylated lipid A containing no Ara4N group.
In addition to FA C14:0(3-OH), FA C14:0, and FA C16:0(3-OH), depicted in our hypothesized structure, fatty acyls C16:0 and C18:0 were also identified in the MS-MS spectra of the lipid A species (not labeled for clarity), signifying the existence of other structural combinations with different fatty acyl numbers.
B. pseudomallei is a facultative intracellular pathogen that stimulates macrophages weakly, enabling the pathogen to evade macrophage killing (1, 31, 51). It has been postulated that unique features of B. pseudomallei LPS may contribute to the low-level immunological activities and thus facilitate the intracellular survival of the pathogen (51). However, the structure of B. pseudomallei LPS, specifically its lipid A, has not been determined until now. Structural information for B. pseudomallei lipid A, in comparison with that of the closely related but rarely pathogenic B. thailandensis, may be an essential component in the overall understanding of the molecular mechanisms underlying pathogenesis.
In this study, we observed that the LPS of B. pseudomallei stimulated both human and murine macrophages to produce levels of TNF-α, IL-6, and IL-10 lower than those in response to the LPS of B. thailandensis in vitro. In order to dissect potential molecular features contributing to this difference, the lipid A structure of B. pseudomallei LPS was characterized and compared to that of B. thailandensis LPS. The major lipid A species in B. pseudomallei consists of a biphosphorylated disaccharide backbone, which was modified with Ara4N at both phosphate groups and penta-acylated with FA C14:0(3-OH), FA C16:0(3-OH), and either FA C14:0 or FA C14:0(2-OH). On the contrary, the major lipid A species identified in B. thailandensis was a heterogeneous mixture of penta- and tetra-acylated structures varying in Ara4N substitutions and acylation of C14:0(3-OH). In addition, the substitution of FA C14:0(2-OH) into the lipid A backbone was unique to B. pseudomallei and was not found in B. thailandensis.
Modification of LPS is a strategy used by various gram-negative bacteria to evade antibacterial mechanisms initiated by the host innate immune system (36). The endotoxic activity of lipid A varies strongly with its primary structure, namely, the FA, polar head groups, and carbohydrate components that constitute it (44). The hexa-acylated lipid A of Escherichia coli, Neisseria meningitidis, and Vibrio cholerae with side chains of 12 to 14 carbons represents the most biologically active form of the molecule (6, 61). Any deviation from this structure, such as a difference in the number or length of fatty acyl chains, will reduce the magnitude of the signal (2, 6, 40, 41). The lipid A structures of closely related species Burkholderia cepacia and Burkholderia mallei have recently been characterized (5, 46). Similar to our data for B. thailandensis, B. cepacia and B. mallei contain a mixture of tetra- and penta-acylated lipid A structures with substitutions of C14:0, C14:0(3-OH), and C16:0(3-OH) but not C14:0(2-OH), which was unique only to B. pseudomallei. Although B. cepacia and B. mallei express only tetra- and penta-acylated lipid A species, the position of these FA relative to the lipid A backbone was described to result in LPS molecules with biological properties similar to those of the potent hexa-acylated species (5, 60). Hence, the observed lowered immunological activities of B. pseudomallei relative to those of E. coli and B. thailandensis may be due to the presence of FA C14:0(2-OH), the order of fatty acyl chains to the lipid A backbone, and the relative concentrations of individual lipid A species. The differences in the lipid A forms of B. pseudomallei and B. mallei are surprising, as the latter, although causing more equine diseases, shares similarities in virulence mechanisms and clinical features with B. pseudomallei (5, 9). Hence, it is tempting to speculate that B. mallei may require an external environmental stimulus to alter the high-level biological activity of its lipid A, as observed with Salmonella enterica serovar Typhimurium, to evade host recognition (18). On the contrary, the lipid A of B. pseudomallei contains the structural features required for host evasion and does not depend on an external stimulus. Further work will be needed to test this hypothesis.
Pathogenic bacteria also evade the immune system through capping phosphate groups at the terminal ends of lipid A with Ara4N and phosphoethanolamine. These modifications confer resistance to the bactericidal effects of endogenously produced host cationic antimicrobial peptides (CAMPs) (14, 17, 50). In Salmonella serovar Typhimurium, the terminal phosphate groups of lipid A are modified with Ara4N residues when grown under magnesium-deficient conditions (18, 59). This decreases the overall negative charge on the pathogen's cell surface and thus lowers the affinity for CAMPs and cationic antibiotics (25). In vitro, B. pseudomallei has been shown to be resistant to the cationic peptides protamine sulfate and purified human defensin HNP-1 (24). In support of this, our structural analysis revealed that both phosphate groups in the major lipid A species identified in B. pseudomallei are capped with Ara4N residues. Modification with Ara4N may increase the resistance of B. pseudomallei to CAMPs and allow the bacterium to survive and replicate within host cells.
In a study that compared the cellular FA profiles of B. pseudomallei and B. thailandensis, the two pathogens were described as sharing the same FA profile, with the exception of FA C14:0(2-OH), which was detected only in B. pseudomallei (22). When comparing the lipid A structures of the two species, we similarly observed the presence of FA C14:0(2-OH) only in the lipid A of B. pseudomallei. FA C14:0(2-OH) appears to be unique to B. pseudomallei, as lipid A species of other closely related species, such as B. cepacia and B. mallei, do not synthesize this FA (5, 46). Based on these findings, the major doubly charged ion at an m/z of 973.59 in B. pseudomallei, which differed by the presence of one —OH group from the corresponding lipid A at an m/z of 965.60 present in both B. pseudomallei and B. thailandensis, was attributed to the substitution of FA C14:0 for FA C14:0(2-OH). In Salmonella serovar Typhimurium, hydroxylation of FA C14:0 to FA C14:0(2-OH) occurs in response to stimuli from host microenvironments, such as the reduced level of magnesium in the phagosome. This modifies the lipid A and has been described to confer resistance to CAMPs (18), permitting a prolonged survival of the bacteria inside the host cell (15). Our data thus indicate that the FA C14:0(2-OH) in B. pseudomallei may enable the bacterium to subvert host cellular responses and survive within the host cell.
LPS is recognized by the innate immune system through interactions with the TLR4 complex present on immune cells (61). However, some LPSs have been shown to activate immune cells via TLR2 instead (4, 20). In this study, we observed that the LPSs from both B. pseudomallei and B. thailandensis specifically activated hTLR4. West and coworkers similarly characterized B. pseudomallei LPS as a TLR4 ligand (53). These findings are contrary to the results of Wiersinga and coworkers, who described B. pseudomallei LPS as acting as a TLR2 agonist (55). The discrepancy observed could be due to the presence of contaminants in LPS, which activate TLR2. A recent study has demonstrated that the ability of some LPSs to activate TLR2 is attributed to the presence of lipoproteins in the preparations (26). B. pseudomallei LPS is typically extracted using the modified hot aqueous-phenol extraction method (39, 55). However, in the case of B. pseudomallei, the LPS is closely associated with the proteins and partitions into the phenol phase together with the proteins (39). This isolation may result in an LPS preparation which is contaminated with proteins (49). Extensive purification of the LPS fraction is therefore required to remove the contaminating proteins, especially for biological assays in which the presence of contaminants may confound results (19). In accordance with these studies, we also observed that B. pseudomallei LPS extracted using the traditional method potently activates both TLR2 and TLR4. Thus, extensive purification of LPS is required to remove contaminating proteins and to eliminate recognition by TLR2.
Multiple factors probably contribute to the pathogenesis of melioidosis, with the LPS of the bacterium being one of the important factors (1, 51, 54). Based on the structural data determined in this study, it appears that the lipid A of B. pseudomallei, with its different fatty acyl chains, induces weak immunological activities and thus evades early host defenses. The presence of Ara4N-modified phosphate groups and C14:0(2-OH) in the lipid A may confer resistance to the effects of CAMPs and allow the pathogen to survive intracellularly. In contrast, the more potent LPS synthesized by B. thailandensis may activate the innate immune system more strongly. Consequently, B. thailandensis becomes more susceptible to the bactericidal effects of host innate immune responses, resulting in efficient clearance of the pathogen. The significance of these modifications has to be determined in vivo in order to fully comprehend the role of LPS in the pathogenesis of B. pseudomallei. Further studies to determine the complete structure of lipid A and the relevance of the minor lipid A species in B. pseudomallei to establish a specific structure-function relationship are ongoing.
This study was supported by a grant from Defense Science & Technology Agency, Singapore. M.R.W. is supported in part by the Singapore National Research Foundation under CRP award no. 2007-04, the Academic Research Fund (grant R-183-000-160-112), the Biomedical Research Council of Singapore (grant R-183-000-211-305), and the National Medical Research Council (grant R-183-000-224-213).
We thank Timothy P. Atkins from Defence Science and Technology Laboratory, United Kingdom, for kindly providing us with the CPS I-negative strain of K96243 (SB04/3518) and our colleagues Angeline Lim and Brendon Hanson from DSO National Laboratories for their advice in the preparation of plasmids.
Published ahead of print on 19 August 2009.