Mutational Analysis and Biochemical Characterization of the Burkholderia thailandensis DW503 Quorum-Sensing Network

Abstract

Numerous gram-negative bacteria communicate and regulate gene expression through a cell density-responsive mechanism termed quorum sensing (QS), which involves the synthesis and perception of diffusible N-acyl-homoserine lactones (AHL). In this study we genetically and physiologically characterized the Burkholderia thailandensis DW503 QS network. In silico analysis of the B. thailandensis genome revealed the presence of at least three AHL synthases (AHS) and five transcriptional regulators belonging to the LuxIR family of proteins. Mass spectrometry demonstrated that wild-type B. thailandensis synthesizes N-hexanoyl-homoserine lactone (C₆-HSL), N-octanoyl-homoserine lactone (C₈-HSL), and N-decanoyl-homoserine lactone (C₁₀-HSL). Mutation of the btaI1 (luxI) AHS gene prevented accumulation of C₈-HSL in culture supernatants, enhanced beta-hemolysis of sheep erythrocytes, increased lipase production, and altered colony morphology on swarming and twitching motility plates. Disruption of the btaI3 (luxI) AHS prevented biosynthesis of C₆-HSL and increased lipase production and beta-hemolysis, whereas mutagenesis of the btaI2 (luxI) allele eliminated C₁₀-HSL accumulation and reduced lipase production. Complementation of the btaI1 and btaI3 mutants fully restored the synthesis of C₈-HSL and C₆-HSL to parental levels. In contrast, mutagenesis of the btaR1, btaR3, btaR4, and btaR5 (luxR) transcriptional regulators had no effect on AHL accumulation, enhanced lipase production, and resulted in extensive beta-hemolysis on sheep blood agar plates. Furthermore, interruption of the btaI1, btaR1, and btaR3 genes altered colony morphology on twitching and swarming motility plates and induced pigmentation. Additionally, phenotypic microarray analysis indicated that QS in B. thailandensis both positively and negatively affects the metabolism of numerous substrates, including citric acid, formic acid, glucose 6-phosphate, capric acid, γ-hydroxybutyric acid, and d-arabinose. These results demonstrate that mutagenesis of the B. thailandensis QS system affects various cellular processes, including lipase production, swarming and twitching motility, beta-hemolysis of sheep erythrocytes, and carbon metabolism and/or transport.

Burkholderia thailandensis is a gram-negative motile rod that is commonly found in stagnant waters, soils, and rice paddies in the central and northeastern areas of Thailand (5). B. thailandensis is genetically and physiologically similar to Burkholderia pseudomallei, the etiologic agent of melioidosis (5, 22, 45). Biochemically, these two organisms have similar abilities to metabolize various carbon sources, with the exception of l-arabinose, which only B. thailandensis assimilates (45). Most striking are the virulence differences between B. thailandensis and B. pseudomallei. Brett et al. reported that the 50% lethal dose of B. pseudomallei in a hamster model was <10 CFU, whereas the 50% lethal dose of B. thailandensis was >10⁶ CFU (5).

Quorum sensing (QS) is an N-acyl-homoserine lactone (AHL) microbial communication system that is employed by gram-negative bacteria to regulate gene expression and has been shown to both positively and negatively control various cellular processes (2, 4, 6, 8, 9, 11, 14, 15, 27, 28, 30, 33, 34, 37, 44). Recent studies have demonstrated that members of the genus Burkholderia encode functional QS systems that regulate AHL biosynthesis and secretion of extracellular virulence factors and contribute to the in vivo pathogenicity of Burkholderia cepacia (3, 10, 23, 29, 30, 42).

In this study, we both genetically and biochemically characterized the B. thailandensis QS network and determined that at least three functional AHL synthases (AHS) (LuxI) and five transcriptional regulators (LuxR), two of which are orphaned (not flanked by a luxI homologue) for a putative luxI gene (btaR4 and btaR5), are present. The products of the B. thailandensis AHS genes were identified as N-hexanoyl-homoserine lactone (C₆-HSL), N-octanoyl-homoserine lactone (C₈-HSL), and N-decanoyl-homoserine lactone (C₁₀-HSL). By creating merodiploids in each of the B. thailandensis QS alleles, we demonstrate that QS affects lipase production, colony morphology on swarming and twitching motility plates, beta-hemolysis of sheep erythrocytes, and the metabolism and/or transport of numerous carbon sources.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The bacterial strains and cloning vectors used in this study are described in Table 1. B. thailandensis, Agrobacterium tumefaciens A136, and Escherichia coli were cultured in Luria-Bertani (LB) broth or on LB agar at either 30 or 37°C. For screening of recombinant clones, E. coli was grown on LB agar plates containing 25 μg of kanamycin (Sigma, St. Louis, Mo.) per ml and 50 μg of 5-bromo-4-chloro-3-indolyl-β-d-galactoside (X-Gal) (Sigma) per ml by using standard procedures (36).

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Descriptiona	Reference or source
Escherichia coli strains
SM10	Mobilizing strain; RP4 tra genes; Kmr	41
TOP10	Used for cloning and blue-white screening	Invitrogen
Bioreporter strain A136	Agrobacterium tumefaciens strain lacking the Ti plasmid	19
Burkholderia thailandensis strains
DW503	Derived from E264; Δ(amrR-oprA) (Kms Gms Sms); rpsL (Smr)	7
BTRJ1	DW503 derivative; I1::pGSV3; Gmr	This study
BTRJ2	DW503 derivative; I2::pGSV3; Gmr	This study
BTRJ3	DW503 derivative; I3::pGSV3; Gmr	This study
BTRJ4	DW503 derivative; R1::pGSV3; Gmr	This study
BTRJ5	DW503 derivative; R2::pGSV3; Gmr	This study
BTRJ6	DW503 derivative; R3::pGSV3; Gmr	This study
BTRJ7	DW503 derivative; R4::pGSV3; Gmr	This study
BTRJ8	DW503 derivative; R5::pGSV3; Gmr	This study
BTRJ9	BTRJ1 containing plasmid pBHR1-I1	This study
BTRJ10	BTRJ3 containing plasmid pBHR1-I3	This study
BTRJ11	BTRJ6 containing plasmid pBHR1-R3	This study
BTRJ12	BTRJ8 containing plasmid pBHR1-R5	This study
Plasmids
pGSV3	Mobilizable suicide vector; Gmr	13
pCR2.1-TOPO	TA cloning vector; Kmr Apr	Invitrogen
pBHR1	Mobilizable broad-host-range expression vector; Kmr Cmr	MoBiTec
pRUI1	Contains a 369-bp PCR product from the 1026b I1 synthase gene	This study
pRUI2	Contains a 360-bp PCR product from the 1026b I2 synthase gene	This study
pRUI3	Contains a 398-bp PCR product from the 1026b I3 synthase gene	This study
pRUR1	Contains a 397-bp PCR product from the 1026b R1 transcriptional regulator	This study
pRUR2	Contains a 424-bp PCR product from the 1026b R2 transcriptional regulator	This study
pRUR3	Contains a 402-bp PCR product from the 1026b R3 transcriptional regulator	This study
pRUR4	Contains a 391-bp PCR product from the 1026b R4 transcriptional regulator	This study
pRUR5	Contains a 401-bp PCR product from the 1026b R5 transcriptional regulator	This study
pBHR1-btaI1	pBHR1 carrying the btaI1 AHS synthase gene from DW503; Kmr	This study
pBHR1-btaI3	pBHR1 carrying the btaI3 AHS synthase gene from DW503; Kmr	This study
pBHR1-btaR5	pBHR1 carrying the btaR5 gene from DW503; Kmr	This study

^aKm^r, kanamycin resistance; Km^s, kanamycin sensitivity; Gm^r, gentamicin resistance; Gm^s, gentamicin sensitivity; Sm^r, streptomycin resistance; Sm^s, streptomycin sensitivity; Ap^r, ampicillin resistance.

Cloning of the B. thailandensis internal gene amplicons and mutant construction.

Primer design for each B. thailandensis QS allele was performed using the B. pseudomallei K96243 luxIR homologues recovered by blastx analysis using the B. cepacia cepIR and Pseudomonas aeruginosa lasIR and rhlIR genes as query sequences (http://www.sanger.ac.uk/). Genomic DNA for PCR amplification was purified with a MasterPure DNA purification kit (Epicentre Technologies, Madison, Wis.). The PCR cycling parameters used for internal gene amplification, the methods used for mutating each B. thailandensis QS gene, and site-specific integration were confirmed by using the methods of Ulrich and DeShazer (43).

AHL reporter assays and mass spectrometry (MS).

The accumulation of AHLs by wild-type B. thailandensis and each QS mutant was analyzed by using the bioreporter strain A. tumefaciens A136 (19), which responds to exogenously secreted AHLs that are of various sizes and chemical compositions. Thin-layer chromatography (TLC) plates were overlaid with A. tumefaciens A136 (inoculated into LB soft agar containing 50 μg of X-Gal per ml) and incubated for 36 to 48 h at 30°C or until adequate color development was achieved.

AHLs were purified from culture supernatants and separated on TLC plates as described by Shaw et al. (40). Preparative TLC spots visually identified as containing AHLs were scraped and extracted three times with 1 ml of methylene chloride (B&J HPLC grade; VWR Scientific, Bridgeport, N.J.). Residual debris was pelleted by centrifugation at 4,000 rpm (Eppendorf 5415C) for 10 min. Supernatants were pooled and evaporated at 50°C under a gentle stream of nitrogen. Dried samples were reconstituted in 100 μl of 50% acetonitrile (B&J HPLC grade; VWR Scientific) in 0.1% formic acid (Sigma), and aliquots (20 μl) were injected onto a PepMap C₁₈ column (150 mm by 1 mm; 5 μm; 100 Å; LC Packings, San Francisco, Calif.). An ABI 140B syringe pump (Applied Biosystems, Foster City, Calif.) provided a flow rate of 50 μl/min, and a 20-min gradient of 0 to 100% solvent B (0.1% formic acid in 95% acetonitrile) was used to elute the compounds of interest. Solvent A was 0.1% formic acid. The column effluent was directed into an LCQ DECA ion trap mass spectrometer equipped with an API II electrospray interface (Finnigan, San Jose, Calif.). The transfer capillary temperature was 350°C. Full-scan, positive-ion mass spectra were acquired by scanning from m/z 100 to m/z 335 in 1.5 s. For identification, extracts were fragmented by collision-induced dissociation of the corresponding [M+H]⁺ ion by using a relative collision energy setting of 19. The spectra were acquired by scanning from m/z 50 to m/z 335 in 1.5 s. MS-MS spectra of unknown compounds were compared to those of standards acquired under the same conditions in order to confirm the identities of the unknown.

Exoproduct secretion and motility analysis.

Siderophore activity was measured on CAS agar plates by using methods described previously (38). Briefly, wild-type B. thailandensis and each QS mutant were replicated onto CAS agar plates and incubated for 24 to 48 h at 37°C. Iron removal, indicating siderophore secretion, was assayed by measuring the blue-orange halo surrounding each inoculation site. Protease, lipase, and phospholipase C (PLC) production (zone sizes) was determined by using previously described methods (12). Briefly, protease secretion was measured on 3% skim milk agar plates, lipase production was measured on tributyrin plates, and PLC synthesis was measured on egg yolk agar plates. Proteolytic activity (peptide degradation), lipase production (with polyoxyethylene sorbitan Tween 40 as the substrate), and PLC production (lecithin cleavage) were confirmed by measuring zone sizes 24 to 48 h postinoculation. Beta-hemolysis was analyzed on 5% sheep blood agar plates. Positive reactions, as indicated by clearing or lysis of erythrocytes surrounding the bacterial colonies, were determined 48 to 72 h following inoculation. Colony morphologies on twitching and swarming motility plates were examined by using methods described by Köhler et al. and Reimmann et al. (26, 35).

Analysis of substrate utilization by B. thailandensis and each QS mutant.

Carbon utilization was monitored in duplicate by using BiOLOG phenotypic microarrays (BiOLOG, Hayward, Calif.). B. thailandensis and each mutant derivative were cultured on R2A agar plates (BiOLOG) for 24 h at 37°C. Cell densities (as determined from plate scrapings) were adjusted by using IF-0 inoculating fluid (BiOLOG), inoculating (100 μl per well) into PM1 and PM2 assay plates (BiOLOG), and incubating at 37°C for 24 to 48 h or until adequate color development was achieved.

RESULTS

B. thailandensis has multiple luxIR genes.

By using the cepIR, lasIR, and rhlIR genes as query sequences, several luxIR genes were identified in the B. pseudomallei K96243 genome. Incorporating the B. pseudomallei K96243 QS genes for primer design (Table 2), the B. thailandensis QS alleles were PCR amplified (small internal gene amplicons for mutant construction), cloned, sequenced, and used in silico to search the B. thailandensis genome (The Institute for Genomic Research) for putative luxIR homologues. As with B. pseudomallei, in silico analysis of the B. thailandensis genome revealed the presence of at least three luxI and five luxR transcriptional regulators (Table 3). Database (blastx) search results further established that B. thailandensis encodes at least three AHSs and five putative transcriptional regulators, two of which are orphaned (no flanking luxI gene was identified) for cognate AHS alleles (btaR4 and btaR5) belonging to the LuxIR family of proteins (Table 3). The btaIR1 and btaIR2 gene pairs are disrupted by intergenic regions, while the btaIR3 genes are transcribed in the same direction (data not shown). The structural organization, including the surrounding alleles, is identical to that reported for Burkholderia mallei (R. L. Ulrich, D. DeShazer, H. Hines, and J. A. Jeddeloh, submitted for publication). The B. thailandensis genome was also screened in silico for potential LuxS and QscR homologues, none of which were identified, suggesting that B. thailandensis encodes only an AHL-based communication system.

TABLE 2.

Primers used for PCR amplification of internal gene amplicons

Genea	Directionb	Primer sequence	Amplicon size (bp)
btaI1	F	5′-CCGCGACGACGACGGGGAAATC-3′	369
	R	5′-TCGATCCAGCACGCGACGACCAT-3′
btaI2	F	5′-ATAAGCGCCGCGCAACTGGATTCC-3′	360
	R	5′-CAGGATCGCCGTATTGCGGTGAGC-3′
btaI3	F	5′-TCGCGGGCCGATTGAACGAACTGC-3′	398
	R	5′-GAGCGACGCGGCCACCGTGAGCAC-3′
btaR1	F	5′-CGGCTTCGAATATTGCTGCTATGG-3′	397
	R	5′-GAGAAAACGGCTCATCAGCGAGTG-3′
btaR2	F	5′-AGCGACCGGCCCGTGACCTGGAG-3′	424
	R	5′-CGGCCTGTATCTTGTTCGTGGAG-3′
btaR3	F	5′-AGACGTCGTCTCGCTGCACTATCC-3′	402
	R	5′-ACCCACGTGAGGCACATCTGTTCG-3′
btaR4	F	5′-GGCGTTCGACAGATGAAACACGAC-3′	391
	R	5′-GCTCATCTGGCACGACGACCTCTA-3′
btaR5	F	5′-CGCGTGCCGTGGCCGCTGTCCA-3′	401
	R	5′-CCGCGCTCCGGGTCCGCCATCAG-3′

^abtaI1 to btaI3 represent AHSs, while btaR1 to btaR5 encode transcriptional regulators.

^bF, forward; R, reverse.

TABLE 3.

Protein homology searches for the B. thailandensis QS genes

Gene	% Identity	% Similarity	Homologya	Protein accession no.b
btaI1	96	97	B. pseudomallei AHS	AF501236_1
	96	97	B. mallei AHS	AF525414_1
	80	88	B. cepacia AHS CepI	AF330012_1
btaI2	48	65	B. cepacia AHS BviI	AAD12727_1
	48	65	Burkholderia ambifaria AHS BafI	AAK27305_1
	48	65	B. cepacia AHS CepI	AF330012_1
btaI3	40	54	B. mallei AHS	AF525414_1
	40	54	B. pseudomallei AHS	AF501236_1
	38	51	Ralstonia solanacearum AHS SolI	CAD17074_1
btaR1	80	88	B. vietnamiensis AHL receptor	AF333007_1
	80	89	Burkholderia stabilis AHL receptor	AF330021_1
	80	89	B. cepacia AHL receptor	AF330020_1
btaR2	43	64	R. solanacearum transcriptional activator protein	CAD17929_1
	40	58	R. solanacearum transcriptional activator protein SolR	AF330023_1
	42	59	B. cepacia AHL receptor CepR	AF330023_1
btaR3	45	62	Burkholderia fungorum DNA-binding protein	ZP_00030469.1
	36	57	Bradyrhizobium japonicum transcriptional regulatory protein	NP_768520.1
	35	56	Mesorhizobium loti unknown protein	BAB52313.1
btaR4	35	50	Pseudomonas aeruginosa PAO1 transcriptional regulator	AE004616_1
	33	49	Listonella anguillarum transcriptional regulator VanR	AAC45213.1
	35	50	R. solanacearum transcriptional activator protein SolR	CAD17075.1
btaR5	51	66	R. solanacearum transcriptional activator protein SolR	CAD17929.1
	41	56	B. cepacia transcriptional activator BviR	AF296284_2
	39	57	R. solanacearum transcriptional activator protein SolR	CAD17075.1

^aLuxI or LuxR proteins similar to that of B. thailandensis.

^bGenBank protein accession number.

Characterization and detection of the AHLs produced by B. thailandensis.

In order to characterize the AHLs synthesized by B. thailandensis, the A. tumefaciens A136 bioreporter strain was used, which allowed detection of multiple AHLs with variable side chain lengths and chemical modifications (C₄-HSL) (40). To confirm that strains BTRJ1 (btaI1), BTRJ2 (btaI2), and BTRJ3 (btaI3) were deficient in AHL synthesis, TLC overlays with A. tumefaciens A136 were examined. As shown in Fig. 1, lanes 2 to 4, BTRJ1 (btaI1), BTRJ2 (btaI2), and BTRJ3 (btaI3) failed to produce detectable levels of C₈-HSL, C₁₀-HSL, and C₆-HSL, respectively. However, it should be noted that AHL detection in this study was dependent on the sensitivity of the bioreporter strain, and it is possible that B. thailandensis produces additional AHLs at concentrations below the limit of detection with A. tumefaciens A136. To determine if any of the five LuxR homologues encoded by B. thailandensis affected AHL biosynthesis, TLC overlays were performed with AHL extracts from each transcriptional regulator mutant. Although the A. tumefaciens A136 bioassay is only semiquantitative, it appeared that the remaining B. thailandensis LuxR QS mutants (BTRJ4 to BTRJ8) accumulated similar levels of C₆-HSL, C₈-HSL, or C₁₀-HSL, suggesting that these LuxR proteins have no effect on AHL production (data not shown). To determine if AHL synthesis could be complemented, wild-type copies of btaI1 and btaI3 were heterologously expressed in the corresponding luxI mutants. Both BTRJ1 (btaI1) and BTRJ3 (btaI3) were fully complemented, and AHL production was restored to parental levels (Fig. 1, lanes 5 and 6). Attempts to clone and maintain the btaI2 gene in E. coli were unsuccessful.

FIG. 1.

Synthesis of the AHL molecules produced by B. thailandensis and each luxI mutant. AHL extracts were spotted (10 μl) onto C₁₈ reversed-phase TLC plates and visualized by using an A. tumefaciens A136 overlay. Lane 1, wild-type B. thailandensis; lane 2, BTRJ1 (btaI1); lane 3, BTRJ2 (btaI2); lane 4, BTRJ3 (btaI3); lane 5, BTRJ9 (pBHR1-btaI1); and lane 6, BTRJ10 (pBHR1-btaI3).

To determine the chemical structure of each AHL produced by B. thailandensis, MS of the TLC-purified molecules was performed. Figures 2D to F show that B. thailandensis synthesized C₈-HSL, C₆-HSL, and C₁₀-HSL. The mass spectra for the C₆-HSL extract and the standard compound had ions at m/z 200.0 and m/z 200.2, respectively, in addition to a product ion at m/z 102.0, which is characteristic of the lactone ring (Fig. 2A and D). The spectra for the C₈-HSL extract and the standard had molecular ions at m/z 228.3 and m/z 227.9, respectively, and also contained the AHL signature lactone ion at m/z 102.0 (Fig. 2B and E). Like that of C₆-HSL and C₈-HSL, fragmentation of the C₁₀-HSL standard and extract produced ion products at m/z 102 (lactone ring) and m/z 155 representing the acyl side chain of C₁₀-HSL (Fig. 2C and F).

FIG. 2.

Structural analysis of the AHLs biosynthesized by B. thailandensis. (A) Synthetic C₆-HSL; (B) synthetic C₈-HSL; (C) synthetic C₁₀-HSL; (D to F) AHLs (C₆-HSL, C₈-HSL, and C₁₀-HSL, respectively) extracted from an overnight culture of B. thailandensis.

Mutagenesis of the B. thailandensis QS system results in altered lipase and beta-hemolysis phenotypes.

Semiquantitative plate assays for detection of beta-hemolysis and protease, siderophore, lipase, and PLC production were performed with each B. thailandensis QS mutant. Disruption of the B. thailandensis QS system had no effect on protease secretion (as measured on 3% skim milk agar plates) or PLC activity (as assayed on egg yolk agar plates) and had a marginal effect on siderophore production (measured on CAS agar plates) (data not shown).

Unlike the effects on protease, PLC, and siderophore synthesis, mutagenesis of the B. thailandensis QS genes both positively and negatively affected lipase production. Strains BTRJ2 (btaI2) and BTRJ5 (btaR2) exhibited reductions in the zone radii on lipase plates, while BTRJ1 (btaI1), BTRJ3 (btaI3), BTRJ4 (btaR1), BTRJ6 (btaR3), BTRJ7 (btaR4), and BTRJ8 (btaR5) exhibited zone sizes that were larger than those for wild-type B. thailandensis (data not shown). Table 4 summarizes the phenotypes characterized in this investigation. Addition of exogenous AHL (C₈-HSL for BTRJ1, C₁₀-HSL for BTRJ2, and C₆-HSL for BTRJ3) at a concentration of 200 nM to lipase plates marginally complemented BTRJ2 (btaI2) and BTRJ3 (btaI3), but not BTRJ1 (btaI1) (data not shown). However, trans complementation with the parental btaI1 gene in BTRJ1 (btaI1) restored the defective lipase phenotype (data not shown). Cloning the B. thailandensis luxR homologues in the broad-host-range expression vector pBHR1 was problematic, and only the btaR3 and btaR5 genes were successfully cloned and maintained in E. coli. As a result, complementation of the remaining mutants (BTRJ4 [btaR1], BTRJ5 [btaR2], and BTRJ7 [btaR4]) was not pursued. The elevated lipolytic phenotypes observed for BTRJ6 (btaR3) and BTRJ8 (btaR5) were reduced to parental levels by expression of the functional allele in trans (data not shown).

TABLE 4.

Phenotype characterization of the B. thailandensis DW503 QS mutants

Mutant	Protease	PLC	Lipase	Swarming motility	Twitching motility	Beta-hemolysis	C6-HSL	C8-HSL	C10-HSL	Carbon sourcesb
										Formic acid	Glucose 6-phosphate	Capric acid	γ-Hydroxybutyric acid	d-Arabinose	Citric acid
BTRJ1 (btaI1)	NAa	NA	−	v	v	−	NA	+	NA	NA	NA	−	−	NA	NA
BTRJ2 (btaI2)	NA	NA	+	NA	NA	NA	NA	NA	+	NA	NA	−	−	NA	NA
BTRJ3 (btaI3)	NA	NA	−	NA	NA	−	+	NA	NA	NA	NA	−	−	−	NA
BTRJ4 (btaR1)	NA	NA	−	NA	v	−	NA	NA	NA	NA	NA	−	−	−	−
BTRJ5 (btaR2)	NA	NA	+	NA	NA	−	NA	NA	NA	NA	NA	−	−	−	NA
BTRJ6 (btaR3)	NA	NA	−	NA	v	−	NA	NA	NA	+	NA	−	−	NA	−
BTRJ7 (btaR4)	NA	NA	−	NA	NA	NA	NA	NA	NA	NA	NA	−	−	NA	NA
BTRJ8 (btaR5)	NA	NA	−	NA	NA	−	NA	NA	NA	NA	+	−	−	−	NA

^aNA, not affected; −, negatively regulated; +, positively regulated; v, different from wild-type B. thailandensis.

^bFor carbon sources, negatively regulated indicates a gain of function and positively regulated indicates a loss of function.

A recent study revealed that B. pseudomallei produces a heat-stable extracellular toxin, determined to be a rhamnolipid, that is hemolytic and cytotoxic to both phagocytic (HL60) and nonphagocytic (HeLa) cell lines (24). Lipid extracts from B. thailandensis and each QS mutant indicated that these strains synthesize a rhamnolipid that is structurally analogous to the B. pseudomallei hemolytic rhamnolipid (data not shown). To determine if QS in B. thailandensis was involved in the beta-hemolysis of sheep erythrocytes, each QS mutant was screened on blood agar plates. BTRJ1 (btaI1), BTRJ3 (btaI3), BTRJ4 (btaR1), BTRJ5 (btaR2), BTRJ6 (btaR3), and BTRJ8 (btaR5) displayed hyper-beta-hemolytic phenotypes (Fig. 3, panel 1). Particularly striking were the levels of beta-hemolysis observed for strains BTRJ6 (btaR3) and BTRJ8 (btaR5⁻) on sheep blood agar plates (Fig. 3, panel 1). In contrast, BTRJ2 (btaI2) and BTRJ7 (btaR4) produced zones of beta-hemolysis equivalent to those of wild-type B. thailandensis (data not shown). Interestingly, trans complementation of BTRJ1 (btaI1), BTRJ6 (btaR3), and BTRJ8 (btaR5) did not reduce beta-hemolysis to parental levels. Also, exogenous addition of C₈-HSL or C₆-HSL (200 nM) to blood agar plates inoculated with BTRJ1 (btaI1) and BTRJ3 (btaI3), respectively, failed to reduce bacterial hemolysis to parental levels.

FIG. 3.

Mutations in the B. thailandensis QS system result in hyper-beta-hemolysis of sheep erythrocytes and altered colony morphologies on twitching and swarming motility plates. Experiments were performed in triplicate, as described in Materials and Methods. Panel 1 shows beta-hemolytic activity, while panels 2 and 3 show the results of twitching and swarming motility assays, respectively. DW503 represents wild-type B. thailandensis, btaI corresponded to luxI mutants, and btaR was disrupted luxR homologues.

Disruption of the B. thailandensis QS system results in altered colony morphologies on swarming and twitching motility plates.

Several studies have demonstrated that QS is involved in bacterial motility and plays an integral role in biofilm maturation (1, 21, 35). To determine if QS influenced motility in wild-type B. thailandensis and each QS mutant, bacterial strains were analyzed on a medium that promotes swarming and twitching motility. Compared to wild-type B. thailandensis, the QS mutants BTRJ1 (btaI1), BTRJ4 (btaR1), and BTRJ6 (btaR3) displayed altered growth phenotypes on twitching mofility plates (Fig. 3, panel 2). Wild-type B. thailandensis colonies had a mucoidal, symmetrical morphology without visible pigmentation (Fig. 3, panel 2). BTRJ1 (btaI1) and BTRJ4 (btaR1) exhibited a wrinkled phenotype in which cells proliferated from the center of the inoculation site and grew on the surface of the underlying colony (Fig. 3, panel 2). Interestingly, BTRJ6 (btaR3) produced a faint orange pigment and displayed extensive wrinkling without the glistening appearance observed for wild-type B. thailandensis (Fig. 3, panel 2). Furthermore, BTRJ6 (btaR3) did not exhibit the symmetrical concentric colony morphology of the parental strain. Heterologous expression of the wild-type genes (btaI1 and btaR3) in BTRJ1 (btaI1) and BTRJ6 (btaR3) and addition of C₈-HSL (200 nM) to the twitching motility plates failed to restore the defective phenotype (data not shown). Likewise, swarming motility was affected by mutagenesis of the B. thailandensis QS system. On swarming motility plates, wild-type B. thailandensis grew in an irregular and spreading fashion at 24 h, and the colony margins were undulating or wavy and had a glistening appearance (Fig. 3, panel 3). Interestingly, most of the QS mutants tested in this study on swarming motility plates resembled (in terms of size and shape of growth) wild-type B. thailandensis, except for BTRJ1 (btaI1), which exhibited a hyperswarming phenotype (Fig. 3, panel 3). Complementation of this luxI mutant by heterologous expression of the btaI1 gene in BTRJ1 or exogenous addition of C₈-HSL (200 nM) to the swarming motility plate was negligible.

Mutagenesis of the B. thailandensis QS system affects carbohydrate metabolism.

Recent studies have linked QS to numerous physiological processes, including transcription, translation, amino acid biosynthesis, chemotaxis, and, interestingly, carbohydrate utilization (37, 44). To determine if carbon metabolism is linked to QS, wild-type B. thailandensis and each QS mutant were assayed by using the BiOLOG phenotypic microarray kit. Mutation of the btaR1 (BTRJ4) and btaR3 (BTRJ6) genes negatively affected (i.e., gain of function) citric acid catabolism (Table 4). The substrates that were positively affected (with a loss of function) included formic acid (BTRJ6 [btaR3]) and glucose 6-phosphate (BTRJ8 [btaR5]) (Table 4). Individual mutagenesis of each of the B. thailandensis luxIR alleles positively influenced capric acid and γ-hydroxybutyric acid metabolism (Table 4). Furthermore, mutations in the btaI3 (BTRJ3), btaR1 (BTRJ4), btaR2 (BTRJ5), and btaR5 (BTRJ8) QS alleles resulted in a gain-of-function phenotype (negatively regulated in B. thailandensis) for d-arabinose utilization (Table 4). It should be noted that disruption of the B. thailandensis QS system imposed no apparent bacterial growth phenotype (growth curve for optical density at 600 nm) when organisms were cultured in complex medium (LB medium).

DISCUSSION

QS is a complex mode of intra- and interspecies communication utilized by both gram-positive and gram-negative bacteria. Signaling mechanisms of this type, which function in response to population density, have been shown to regulate the transcription of various genes associated with human and plant infections (20). In turn, this provides microorganisms with an ecological niche that allows subversion of host defense mechanisms.

Recent studies have demonstrated that Burkholderia species encode functional QS networks that direct the biosynthesis of numerous AHL molecules, regulate virulence factor production, and contribute to animal pathogenicity (3, 10, 23, 25, 29-31, 42). In silico analysis of the B. thailandensis genome revealed the presence of at least three luxI and five luxR homologues. Using MS analysis after TLC separation, this study demonstrated that B. thailandensis produces the signaling molecules C₈-HSL (BtaI1), C₆-HSL (BtaI3), and C₁₀-HSL (BtaI2). Similar findings have been reported for B. cepacia and Burkholderia vietnamiensis (10, 29). Although genetically related, especially with regard to the similarity of encoded (at the DNA level) luxIR homologues, B. pseudomallei synthesizes a wide range of additional AHLs, including N-(3-hydroxyoctanoyl)-homoserine lactone, N-(3-hydroxydecanoyl)-homoserine lactone, and N-(3-oxotetradecanoyl)-homoserine lactone, which are not detected in B. thailandensis. Furthermore, disruption of the B. pseudomallei luxI genes (three total) has a negligible effect on AHL accumulation (Ulrich et al., submitted for publication). Mutagenesis of the btaI1, btaI2, and btaI3 alleles prevented the accumulation of C₈-HSL, C₁₀-HSL, and C₆-HSL in culture supernatants, respectively (Fig. 1, lanes 2, 3, and 4). However, it is possible that additional signaling molecules not recognized by A. tumefaciens A136 are synthesized. In numerous QS networks, luxI expression has been shown to be regulated by LuxR proteins (16-18, 29, 39). Although the results are preliminary, disruption of the B. thailandensis luxR genes appeared not to have an effect on AHL accumulation (data not shown).

In B. cepacia the secretion of exoproducts (lipases, proteases, and siderophores), which enhance microbial virulence, is regulated at the level of transcription via QS (29, 30). DeShazer et al. demonstrated that B. pseudomallei produces several potential extracellular virulence factors (12). Because of the genetic and biochemical relatedness between B. thailandensis and B. pseudomallei, each QS mutant and wild-type B. thailandensis were tested for protease, PLC, lipase, and siderophore production. Using semiquantitative plate assays, it was found that QS in B. thailandensis both positively and negatively regulates lipase production (Table 4). Similar findings have been reported for B. cepacia K52-2 (29, 30). Disruption of the cepR allele increased ornibactin secretion, reduced lipase activity, and resulted in a protease-defective phenotype (29). Deletion of the cepI gene also resulted in hyperproduction of ornibactin and eliminated protease activity (29). Interestingly, with the multiple B. thailandensis QS mutants tested, only mutagenesis of the btaIR2 genes resulted in a reduction in lipase production (Table 4). These results suggest that the btaIR2 gene products function as positive regulators of lipase synthesis. Unlike the results obtained with mutations in the B. cepacia cepIR genes, which eliminated protease activity (30), analysis of protease and PLC biosynthesis in each B. thailandensis QS mutant suggested that these exoproducts are not regulated in a cell density-dependent manner via QS.

Both swarming and twitching motility are modes of bacterial translocation that allow microorganisms to avoid inhibitory compounds, promote nutrient uptake, and establish cell-to-cell contact with plant and animal tissues. Several investigations have shown that QS is involved in the regulation of bacterial motility (25, 26, 35). In B. cepacia, swarming motility is affected by QS, and mutagenesis of the cepIR genes results in a pronounced reduction in swarming motility (25). Disruption of the B. thailandensis QS system revealed that both twitching and swarming motility were affected (Fig. 3, panels 2 and 3). On twitching motility plates, BTRJ1 (btaI1), BTRJ4 (btaR1), and BTRJ6 (btaR3) produced a wrinkled phenotype that was not apparent in wild-type B. thailandensis (Fig. 3, panel 2). In addition, the appearance of the colony in the wrinkled domain was rough instead of the characteristic smooth and glistening appearance observed for B. thailandensis (Fig. 3, panel 2). Interestingly, BTRJ6 (btaR3) also produced an uncharacterized pale orange pigment not found in wild-type B. thailandensis. It should be noted that extensive incubation (5 to 8 days) of B. thailandensis at 30 or 37°C resulted in altered colony morphologies similar to the wrinkled appearance observed for BTRJ6 (btaR3). However, prolonged incubation of B. thailandensis (in our laboratory) has never induced pigment biosynthesis. Subtle variations (colony size and shape) in swarming motility were also observed for each B. thailandensis QS mutant tested (data not shown). The most prominent difference on motility plates was observed for BTRJ1 (btaI1), which exhibited a hyperswarming phenotype. The mechanisms for these defects in twitching and swarming motility following mutagenesis of the B. thailandensis QS system are unclear and remain to be determined. Analysis of pilus (twitching) and flagellum biosynthesis (by electron microscopy) in BTRJ1 (btaI1), BTRJ4 (btaR1), and BTRJ6 (btaR3) revealed no differences compared to wild-type B. thailandensis (data not shown). These findings suggest that disruption of the B. thailandensis QS system has no effect on pilus and flagellum biosynthesis, indicating that an additional factor(s) associated with QS contributes to the observed twitching and swarming motility defects.

A recent study characterized a cytotoxic and hemolytic exolipid synthesized by B. pseudomallei (24). Elevated levels of cytotoxicity and hemolysis were detected in supernatant extracts of B. pseudomallei that had reached stationary phase (24). Considering the relatedness of B. thailandensis and B. pseudomallei strains, we tested wild-type B. thailandensis and each QS mutant for hemolytic activity on sheep blood agar plates. Disruption of the B. thailandensis QS system, particularly the luxR homologues, resulted in hyper-beta-hemolysis of erythrocytes, suggesting that QS regulates a factor(s) that contributes to beta-hemolysis (Fig. 3, panel 1). The mechanism (protein or lipid) for this hyper-beta-hemolysis is not fully understood. However, using TLC and MS, we purified and characterized a monounsaturated rhamnolipid from B. thailandensis that is chemically analogous to the B. pseudomallei hemolytic rhamnolipid (data not shown). In addition, using semiquantitative plate assays (egg yolk agar), we found that QS has no effect on PLC production which could also contribute to hemolysis (Table 4). It is conceivable the hyper-beta-hemolytic phenotypes observed in this study following disruption of the B. thailandensis QS system were the result of increased rhamnolipid production. Attempts to quantify the rhamnolipid in each B. thailandensis QS mutant were unsuccessful, and additional studies will be necessary to determine the mechanisms of this beta-hemolysis.

Preliminary analysis using a semiquantitative approach (BiOLOG phenotypic array plates) suggested that QS either directly or indirectly plays a role in the central metabolism of B. thailandensis (Table 4). Interestingly, mutagenesis of the btaR5 gene (BTRJ8) resulted in a positive phenotype, that is, a substrate that was positively regulated by QS in B. thailandensis for the metabolism of glucose 6-phosphate (Table 4). A similar finding, at the transcriptional level, was recently reported by Schuster et al., who found that the P. aeruginosa QS system regulates the NADPH-dependent glucose-6-phosphate dehydrogenase, an essential enzyme needed for glucose metabolism (37). Another interesting phenotype affected by QS in B. thailandensis is d-arabinose utilization (Table 4). Although l-arabinose is structurally different, l-arabinose utilization is one of the assays used to distinguish B. thailandensis, which assimilates l-arabinose, from B. pseudomallei, which cannot use l-arabinose as a primary carbon source (45). Disruption of the btaI3 (BTRJ3), btaR1 (BTRJ4), btaR2 (BTRJ5), and btaR5 (BTRJ8) genes resulted in a positive phenotype for d-arabinose utilization; that is, disruption conferred the ability to use this substrate as a sole carbon source (Table 4). However, neither BTRJ4 (btaR1), BTRJ5 (btaR2), nor BTRJ8 (btaR5) grew in M9 minimal broth (36) with d-arabinose as the primary substrate (data not shown). These results were likely concentration (d-arabinose) dependent and will require further analysis. Also, considering the results of Schuster et al. (37), it is possible that QS regulates the biosynthesis of an additional factor(s) needed to support the growth of B. thailandensis in d-arabinose minimal medium. However, these findings suggest that QS affects carbon metabolism in B. thailandensis. The mechanisms for these differences in carbon utilization remain to be determined, but it is plausible that B. thailandensis employs QS in the environment to prevent excess energy expenditure for carbon acquisition and metabolism when adequate substrates are present in situ.

Complementation of the numerous defective phenotypes observed in this study was unsuccessful. Similar results have been reported for B. cepacia, in which lipase activity was not restored to parental levels by heterologous expression of the cepR and cepIR genes in trans (29). The disruption cassettes used in this investigation were not designed to create in-frame mutations within the target gene. It is possible that the complementation difficulties are the result of polar effects on downstream alleles (i.e., transcriptional regulators) that are required for normal swarming and twitching motility in addition to lipase production. Furthermore, the relative copy number of the gene being transcribed and translated could result in complementation problems, especially when the hypothesis described by Schuster et al. (37), which stated that LuxR levels are essential for activation of QS-controlled genes, is applied to B. thailandensis. Another possibility is that the B. thailandensis QS system forms a hierarchy; that is, there is cross- regulation of the luxI and luxR genes, and overexpression of these genes disrupts this regulatory cascade. However, full restoration of C₈-HSL and C₆-HSL accumulation in BTRJ1 (btaI1) and BTRJ3 (btaI3) was achieved by trans expression of the btaI1 and btaI3 genes, respectively. The latter finding suggests that the relative concentration of the signaling molecule is essential for effective complementation of defects in swarming motility (BTRJ1 [btaI1]) and lipase production.

Extensive genetic analysis will be required to determine the functional relationship between QS and the environmental survival of B. thailandensis. As in B. cepacia, mutations in the B. thailandensis QS system resulted in both positive and negative regulation of numerous phenotypes (Table 4). With regard to the multiple B. thailandensis luxR mutants assayed for phenotypic variations in this investigation and the findings of Medina et al. (32), who demonstrated that the rhlR transcriptional regulator is both an activator when it is associated with the signaling molecule C₄-HSL and an antiactivator when it is not bound to C₄-HSL, we hypothesize that the B. thailandensis LuxR transcriptional regulators have dual functions for modulation of twitching and swarming motility, lipase production, and substrate utilization (Table 4). Similar to data obtained for P. aeruginosa, B. cepacia, and B. thailandensis, disruption of the solIR QS system in Ralstonia solanacearum resulted in a 1.7-fold increase in polygalacturonase synthesis (18). Negative gene regulatory mechanisms of this type may provide bacteria a way to reduce energy expenditure for protein synthesis in environments that contain adequate substrates and cofactors. With the multiple phenotypes altered by mutagenesis of the B. thailandensis QS system, it is apparent that extensive molecular analysis, likely by expression profiling studies with whole-genome DNA microarrays, will be required to unravel the complex regulatory circuits encoded by the QS network.