A Burkholderia cenocepacia Orphan LuxR Homolog Is Involved in Quorum-Sensing Regulation

Abstract

Burkholderia cenocepacia utilizes quorum sensing to control gene expression, including the expression of genes involved in virulence. In addition to CepR and CciR, a third LuxR homolog, CepR2, was found to regulate gene expression and virulence factor production. All B. cenocepacia strains examined contained this orphan LuxR homolog, which was not associated with an adjacent N-acyl-homoserine lactone synthase gene. Expression of cepR2 was negatively autoregulated and was negatively regulated by CciR in strain K56-2. Microarray analysis and quantitative reverse transcription-PCR determined that CepR2 did not influence expression of cepIR or cciIR. However, in strain K56-2, CepR2 negatively regulated expression of several known quorum-sensing-controlled genes, including genes encoding zinc metalloproteases. CepR2 exerted positive and negative regulation on genes on three chromosomes, including strong negative regulation of a gene cluster located adjacent to cepR2. In strain H111, which lacks the CciIR quorum-sensing system, CepR2 positively regulated pyochelin production by controlling transcription of one of the operons required for the biosynthesis of the siderophore in an N-acyl-homoserine lactone-independent manner. CepR2 activation of a luxI promoter was demonstrated in a heterologous Escherichia coli host, providing further evidence that CepR2 can function in the absence of signaling molecules. This study demonstrates that the orphan LuxR homolog CepR2 contributes to the quorum-sensing regulatory network in two distinct strains of B. cenocepacia.

Members of a group of 17 closely related Burkholderia species termed the Burkholderia cepacia complex (Bcc) have emerged as opportunistic pathogens in people with cystic fibrosis (CF) and chronic granulomatous disease (91-95). Burkholderia cenocepacia and B. multivorans are the most common members of the Bcc involved in CF lung infections (59, 78). Bcc infections are a major concern for CF patients because certain Bcc strains have exhibited high patient-to-patient transmissibility, multidrug resistance, and the potential to cause inflammation and fatal invasive disease (58, 61, 62). These organisms also inhabit diverse ecological niches, and some Bcc strains exhibit bioremediation and biocontrol capabilities (61).

The members of the Bcc utilize N-acyl-homoserine lactone (AHL)-based quorum-sensing (QS) systems for the regulation of diverse physiological processes, including processes involved in virulence. QS is a form of genetic regulation typically mediated by the accumulation and recognition of self-produced signals in local environments. AHL-mediated QS systems are comprised of a luxI homolog, which encodes an AHL synthase that catalyzes the synthesis of AHL signaling molecules, and a luxR homolog, which encodes a transcriptional regulator that mediates gene expression in its active, AHL-bound form (24).

Two QS systems have been described in B. cenocepacia. The CepIR QS system is widely distributed among Bcc strains (28, 52, 57). CepI directs the synthesis of N-octanoyl-homoserine lactone (C₈-HSL) and N-hexanoyl-homoserine lactone (C₆-HSL) (53). CepR positively regulates cepI and negatively controls its own expression in B. cenocepacia (53). Epidemic strains of B. cenocepacia that possess the B. cenocepacia genomic island (cci) also have the CciIR QS system (5). The predominant AHL produced by the AHL synthase, CciI, is C₆-HSL, and minor amounts of C₈-HSL are produced (63). CepR positively regulates the expression of the cciI and cciR genes, which are cotranscribed, and CciR negatively regulates the expression of the cciIR operon. CciR is also involved in the negative regulation of cepI (63).

The B. cenocepacia QS systems form a global gene regulatory system (82). The CepIR QS system in B. cenocepacia is involved in the regulation of swarming motility, mature biofilm development, the production of chitinase, extracellular proteases, the siderophore ornibactin (35, 36, 42, 52, 53, 84, 87), and the nematodicidal protein AidA (43), as well as genes involved in type II secretion, type III secretion, and oxidative stress (86). The CciIR system regulates motility, extracellular proteases, and ornibactin production (63, 86). Both CepIR and CciIR have been shown to contribute to virulence in respiratory infection models (5, 84).

Genome sequence analyses of other bacteria indicate that the number of LuxR homologs present in a genome is often greater than the number of LuxI homologs (23). The “orphaned” LuxR homologs that do not have an associated AHL synthase are predicted to respond to endogenously or exogenously synthesized AHLs. The Pseudomonas aeruginosa QS network contains the orphan LuxR homolog QscR (quorum sensing control repressor) (7). QscR negatively modulates the activity of the AHL-mediated QS systems, LasR and RhlR. It is believed that the mechanism of QscR activity is partially independent of transcriptional control through the formation of QscR-LasR and QscR-RhlR heterodimers or through competition for AHLs or competition for DNA binding sites (7, 48). LasR, RhlR, and QscR have overlapping but distinct regulons (51). QscR requires N-(3-oxodecanoyl)-homoserine lactone to actively bind DNA, but it exhibits a relaxed AHL specificity compared to LasR that may enable it to respond to exogenous AHLs in mixed bacterial populations (49).

Studies investigating B. pseudomallei have used the designations pml, bpm, and bps for components of the QS systems, which are composed of three sets of LuxIR homologs and two additional LuxR homologs (56, 85, 88, 90). Recently, it has been shown that the orphan LuxR homologs BpsR4 and BpsR5 each decrease bpsI3 expression, while BpsR5 activates bpsI1 expression in the absence of AHL (41). B. mallei has two pairs of luxIR homologs, bmaIR1 and bmaIR2, and two orphan luxR homologs, bmaR4 and bmaR5 (89). The BmaIR1 and BmaIR3 QS systems respond to C₈-HSL (17, 18). Three sets of luxIR homologs and two orphan luxR homologs have been identified in B. vietnamiensis (64). However, orphan luxR homologs have not yet been characterized in the Bcc. In this study we report the identification and characterization of an orphan LuxR homolog in B. cenocepacia and its role in gene regulation in two distinct strain backgrounds.

MATERIALS AND METHODS

Strains and growth conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. Cultures were routinely grown at 37°C in Miller's Luria broth (Invitrogen, Burlington, Ontario, Canada) or in modified low-salt Luria-Bertani broth (3) for B. cenocepacia H111 and its derivatives. Agrobacterium tumefaciens was grown at 30°C. When appropriate, the following concentrations of antibiotics were used: 100 μg/ml of trimethoprim and 200 μg/ml of tetracycline for B. cenocepacia K56-2 and its derivatives; 100 μg/ml of ampicillin, 50 μg/ml of kanamycin, and 25 μg/ml of chloramphenicol for B. cenocepacia H111 and its derivatives; 1.5 mg/ml of trimethoprim, 15 μg/ml of tetracycline, and 50 μg/ml of kanamycin for Escherichia coli; and 4.5 μg/ml of tetracycline and 50 μg/ml of spectinomycin for A. tumefaciens. For chrome azurol S (CAS) assays and siderophore extraction, cultures were grown in iron-deficient succinate medium with 10 mM l-ornithine dihydrochloride for 40 h (67). MIC broth microdilution assays were conducted using Mueller-Hinton broth.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Descriptiona	Source or reference(s)
A. tumefaciens A136	Ti plasmidless host	C. Fuqua
E. coli strains
DH5α	F− φ80lacZΔM15 Δ(lacZYA-argF) recA1 endA gyrA96 thi-1 hsdR17 supE44 relAl deoR(U169)	Invitrogen
TOP10	F−mcrA Δ(mrr-hsdRMS-mcrBC) φ80lacZΔM15 ΔlacX74 deoR recA1 araD139 Δ(ara-leu)7697 galU galK rpsL(Str) endA1 nupG	Invitrogen
HB101	supE44 hsdS20 (rB mB) recA13 ara-14 proA2 lacY1 galK2 rpsL20 xyl-5 mtl-1	79
MM294	F−endA1 hsdR17 supE44(AS) rfbD1 spoT1 thi-1	66
XL1-Blue	recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac [F′ proAB lacIqZΔM15 Tn10(Tcr)]	Stratagene
Bcc strains
B. cepacia ATTC 25416T	Onion isolate	60
B. multivorans strains
LMG13010T	CF isolate	60
C5393	CF isolate	60
B. cenocepacia strains
J2315	CF isolate, BCESM+cblA+	60
BC7	CF isolate, BCESM+cblA+	60
C5424	CF isolate, BCESM+cblA+	60
C6433	CF isolate, BCESM+	60
C1394	CF isolate BCESM+	60
PC184	CF isolate, BCESM+	60
CEP511	CF isolate, BCESM+	60
J415	CF isolate, BCESM−	60
ATTC 17765	Urinary tract infection isolate, BCESM+	60
PC715j	CF isolate, BCESM−	65
H111	CF isolate, BCESM−	75
H111-R	cepR::Km derivative of H111	35
H111-R2	cepR2::pEX18Gm derivative of H111	This study
H111-R/R2	Kmr Gmr, cepR2::pEX18Gm mutant of H111-R	This study
K56-2	CF isolate, BCESM+cblA+	60
K56-2cepR2a	cepR2::Tp derivative of K56-2, clone a, Tpr	This study
K56-2cepR2b	cepR2::Tp derivative of K56-2, clone b, Tpr	This study
K56-R2	cepR::Tn5-OT182 derivative of K56-2, Tcr	52
K56-dI2	ΔcepI derivative of K56-2	63
K56-2cciI	cciI::Tp derivative of K56-2, Tpr	5
K56-2cciR	cciR::Tp derivative of K56-2, Tpr	63
K56-2cciIR	ΔcciIR derivative of K56-2	63
K56-2cepRcciIR	cepR::Tp, ΔcciIR derivative of K56-2, Tpr	63
B. stabilis strains
LMG14294	CF isolate	60
LMG14086	Respiratory isolate	60
B. vietnamiensis strains
PC259	CF isolate	47
G4	Water treatment facility isolate	71
B. dolosa strains
LMG19468T	CF isolate	J. LiPuma
LMG18943	CF isolate	8
B. ambifaria strains
CEP0996	CF isolate	9
LMG17828	Soil isolate	9
B. athina strains
LMG20980T	Soil isolate	10
LMG20982	Hospital environmental isolate	D. P. Speert
B. pyrrocinia LMG 21822	Soil isolate	10
Plasmids
pEX18Tc	Gene replacement vector, oriT+sacB+ Tcr	32
pUCP26	Broad-host-range vector, Tcr	99
pCR2.1Topo	Cloning vector for PCR products, Apr Kmr	Invitrogen
pRK2013	ColEl Tra (RK2)+, Kmr	21
pRK415	Broad-host-range vector, Tcr	39
p34E-Tp	Source of Tp resistance cassette, Tpr	13
p34S-Tc	Source of Tc resistance cassette, Tcr	12
pCF218	IncP plasmid expressing TraR, Tcr	100
pCF372	pUCD2 with a traI-lacZ fusion, Spr	25
pRM1T6	pCR2.1TOPO containing the 1.6-kb cepR2 fragment, Apr Kmr	This study
pRM613	pUCP26 containing a 1.3-kb SstI/SmaI fragment from pRM1T6 including cepR2, Tcr	This study
pRM516	pRK415 containing the 1.6-kb cepR2 fragment from pRM1T6, Tcr	This study
pBS13	zmpA::luxCDABE transcriptional fusion constructed in pMS402, Kmr Tpr Tcr	B. Subsin
pBS9	zmpB::luxCDABE transcriptional fusion constructed in pMS402, Kmr Tpr Tcr	42
pBBR1MCS	Cmr, broad-host-range cloning vector	44, 45
pBBR-cepR	Cmr, pBBR1MCS containing the cepR gene of B. cenocepacia H111	S. Schmidt
pEX18Gm	Gene replacement vector, oriT+sacB+, Gmr	32
pJBA89	Apr, pUC18Not-luxR-PluxI-RBSII-gfp(ASV)-T0-T1	3
pJBA89luxR−	Apr, pUC18Not-PluxI-RBSII-gfp(ASV)-T0-T1	This study
pJTR2	Cmr, pBBR1MCS containing the cepR2 gene of B. cenocepacia H111	This study
pRK600	Cmr, ColE1 oriV RK2-Mob+ RK2-Tra+	40
pRK2013	Kmr, RK2 derivative, Mob+ Tra+ ColE1	15
pSU11	Gmr, broad-host-range lacZ-based promoter probe vector	L. Eberl
pRN3	Tpr, broad-host-range lacZ-based promoter probe vector, derivative of pSU11	This study

^aBCESM, B. cepacia epidemic strain marker; cblA, cable pilus gene; Ap, ampicillin; Cm, chloramphenicol; Km, kanamycin; Sp, spectinomycin; Tp, trimethoprim; Tc, tetracycline.

DNA manipulations.

DNA manipulations were performed using standard techniques as described by Sambrook et al. (79). Genomic DNA was isolated as described by Ausubel et al. (4) or Walsh et al. (96) or by using a DNeasy tissue kit (Qiagen, Hilden, Germany). Oligonucleotide primers (see Table S1 in the supplemental material) were synthesized by Eurofins MWG Operon (Germany), Invitrogen, or the University of Calgary Core DNA and Protein Services (Calgary, Alberta, Canada). Plasmids were introduced into B. cenocepacia by electroporation (11) or by conjugation employing pRK600 or pRK2013 (21) as the mobilization vector.

Cloning of cepR2 from B. cenocepacia K56-2 and H111.

The cepR2 open reading frame (ORF) was cloned from K56-2 by PCR amplifying a 1.6-kb fragment with the M188F and M188R primers and cloning the fragment into pCR2.1Topo (Invitrogen), resulting in pRM1T6. For K56-2 cepR2 complementation, a 1.3-kb SstI/SmaI fragment from pRM1T6 was cloned into pUCP26 (99), resulting in pRM613, and a 1.6-kb HindIII/XbaI fragment from pRM1T6 was cloned into pRK415 (39), resulting in pRM516. The cepR2 ORF was cloned from H111 by PCR amplification with primers cepR2pBBR-F and cepR2pBBR-R. The amplicon was first inserted into pCR2.1Topo (Invitrogen) and then subcloned into pBBR1MCS using the EcoRV and BamHI restriction sites to create pJTR2.

Construction of the K56-2cepR2, H111-R2, and H111-R/R2 mutants.

To construct the K56-2 cepR2::dhfRII mutant, a SacII digest of pRM1T6 removed 33 bp of cepR2, and the trimethoprim resistance cassette from a SmaI digest of p34E-Tp (13) was inserted into the cepR2 ORF. The disrupted cepR2 fragment was subcloned into pEX18Tc with HindIII and XbaI, resulting in pRM1X6. Two independently isolated K56-2cepR2 mutants were characterized and designated K56-2cepR2a and K56-2cepR2b (the mutant primarily used in this study). An H111 cepR2 mutant (H111-R2) was constructed as follows. A 384-bp internal fragment of cepR2 was amplified with primers cepR2-R and cepR2-F, digested with HindIII, ligated into the gene replacement vector pEX18Gm (32), and transformed into E. coli XL1-Blue. The final construct was transferred to H111, and one mutant was designated H111-R2. An H111 cepR cepR2 double mutant, designated H111-R/R2, was constructed in an analogous manner using the H111 cepR mutant (H111-R) as the recipient strain. Allelic exchange in K56-2 and H111 cepR2 mutants was confirmed by both PCR and Southern hybridization.

Construction of plasmid pJBA89luxR⁻.

Plasmid pJBA89luxR⁻ was constructed by partial digestion of pJBA89 with restriction enzymes EcoRI and HindIII, which deleted a 703-bp fragment containing luxR from plasmid pJBA89.

Analysis of cepR2 in the Bcc.

The presence of cepR2 in 17 strains representing nine Bcc species was determined by Southern hybridization using normal stringency conditions and a hybridization temperature of 65°C. The presence of cepR2 in 12 B. cenocepacia strains was determined by PCR using primers M188scF and M188scR.

Sequence analysis.

The nucleotide sequence of the B. cenocepacia J2315 cepR2 locus was obtained from the EMBL database (http://www.ebi.ac.uk/) (33). Sequence analysis was performed using Artemis (77), DNAMAN (Lynnon Biosoft, Vandreuil, Quebec, Canada), gapped BLASTX (2), and BLASTP (2), as well as the Conserved Domain Database at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi).

RNA manipulations.

B. cenocepacia was subcultured to obtain an initial optical density at 600 nm (OD₆₀₀) of 0.02 and grown for the time indicated below for each experiment without selection. There were no differences in growth between K56-2cepR2a or K56-2cepR2b and K56-2. Total RNA was isolated using a RiboPure bacterial RNA isolation kit (Ambion, Streetsville, Ontario, Canada). DNase treatment was performed, and samples were confirmed by PCR using Platinum Taq polymerase (Invitrogen) to be free of DNA prior to cDNA synthesis.

RT-PCR and qRT-PCR.

Reverse transcription PCR (RT-PCR) was performed using a Titan one-tube RT-PCR kit (Roche, Mississauga, Ontario, Canada) or an iScript Select cDNA synthesis kit (Bio-Rad). Oligonucleotide primers (see Table S1 in the supplemental material) were designed with Primer Express software (version 2; Applied Biosystems, Foster City, CA) or Primer3 (76). The sigma factor gene sigA (BCAM0918) and the NADH dehydrogenase gene ndh (BCAM0166) were used as reference standards as described previously (64, 86). The expression of these genes was not significantly altered, as determined by microarray analysis (data not shown). RT-PCR and agarose gel electrophoresis were used to semiquantify gene expression. For quantitative RT-PCR (qRT-PCR), quantification and melting curve analyses were performed with an iCycler and iQ SYBR green Supermix (Bio-Rad) according to manufacturer's instructions. qRT-PCRs were performed in triplicate, and the data shown below represent data from at least two independent experiments. Relative expression values for each gene were calculated using the ΔΔC_t equation (55).

Microarray sample preparation.

Three independent RNA samples of B. cenocepacia K56-2 and K56-2cepR2b grown for 8 h were used in microarray experiments. Gene expression profiles were generated using custom B. cenocepacia microarrays (Agilent, Santa Clara, CA) (33, 50). cDNA samples were fluorescently labeled with Cy3 (K56-2) or Cy5 (K56-2cepR2b), mixed, hybridized to the array according to the Agilent 60-mer oligonucleotide microarray processing protocol (version 2.1), and scanned. Dye-swap experiments were conducted to eliminate genes which showed a dye bias. Labeling, hybridization, and scanning were performed by the Mahenthiralingam Laboratory, Cardiff University, Wales.

Microarray data analysis.

Microarray data analysis was performed using GeneSpring GX 7.3.1 software (Agilent). Initial data were preprocessed by employing the enhanced Agilent FE import method, and then per-spot and per-chip normalizations were performed for all arrays. Consistency across the arrays of the signal intensities of spike-in control genes (added prior to cDNA synthesis) and prelabeled control genes (added prior to hybridization) was confirmed for data quality evaluation prior to further analyses (data not shown). At least 87% of all probes on the three arrays were found to be present (data not shown). After filtering on flags (present in at least two of three arrays), genes were selected on the basis of changes, for which a 1.5-fold cutoff (in at least two of three arrays) was used for comparisons of K56-2 and K56-2cepR2b. Genes which showed a dye bias or inconsistent changes were also excluded from the analysis. Subsequent to microarray analysis, certain genes were noted because they appeared to be in transcriptional units associated with differentially regulated genes.

Transcriptional fusions to luxCDABE.

Transcriptional fusions of zmpA::luxCDABE (pBS13) and zmpB::luxCDABE (pBS9) (42) were constructed in pMS402 (16). The 642-bp zmpA promoter region was amplified using primers PFLX and PRLX and cloned into the XhoI-BamHI site upstream of luxCDABE in pMS402. A SacI fragment from p34S-Tc (12) containing the tetracycline resistance cassette was ligated into this construct digested with XhoI to generate pBS13. Luminescence assays were carried out as previously described (64). The level of promoter activity is expressed below as the ratio of luminescence to turbidity (cps/OD₆₀₀).

Transcriptional fusions to lacZ.

Transcriptional fusions of the two promoters driving expression of the pyochelin regulator (BCAM2231 [pchR] to BCAM2221 and the biosynthesis operon BCAM2232 [pchD] to BCAM2235) with lacZ were constructed in H111 as follows. The two promoter regions were amplified using primers pyo2231P-R and pyo2231P-F and primers pyo2232P-R and pyo2232P-F. The two amplicons were digested with XhoI and HindIII and inserted into the promoter-probe vector pSU11 or pRN3. β-Galactosidase activity was determined as described previously (72).

Phenotypic assays.

Protease activity was determined using skim milk agar (83). Thin-layer chromatography AHL bioassays were performed using A. tumefaciens A136(pCF218)(pCF372) as a reporter strain (81). Biofilm formation was quantified by crystal violet staining (64). Pellicle formation assays were performed as described previously (6). Swarming motility was determined using semisolid agar (0.5%) motility assays (54). For MIC assays, overnight cultures were normalized to a concentration of 1 × 10⁵ CFU/ml and incubated in medium containing twofold serial dilutions of heavy metals, antibiotics, or detergent. Growth was assessed after incubation at 37°C for 24 h. Ceftazidime resistance was determined with Etest strips (AB Biodisk, Solna, Sweden). Resistance to heat shock at 42°C was determined as described previously (14). Siderophore activity present in the culture supernatants was measured by CAS assays (80). For strain H111, CAS agar was prepared without MM9 buffer and contained Luria-Bertani medium instead of Casamino Acids. Pyochelin and ornibactin were extracted from 50-ml culture supernatants. Cell-free culture supernatants were acidified (pH 2.0), and pyochelin was extracted with dichloromethane. Extracts were dried, and the residue was resuspended in 20 μl methanol. The aqueous phase containing ornibactin was concentrated on a rotary evaporator to a final volume of 1 ml.

Activation of the AHL biosensor pJBA89 and its derivative pJBA89luxR⁻.

Overnight cultures of E. coli MT102 harboring plasmid pJBA89 or pJBA89luxR⁻ (in the presence or absence of plasmid pJTR2) were subcultured in 50 ml Luria-Bertani medium containing appropriate antibiotics to an OD₆₀₀ of 0.1. At an OD₆₀₀ of 1.5, cells were distributed in 200-μl aliquots into wells of a microtiter plate. The wells contained no HSL or C₁₄-HSL, 3-oxo-C₁₂-HSL, C₁₂-HSL, 3-oxo-C₁₀-HSL, C₁₀-HSL, 3-oxo-C₈-HSL, C₈-HSL, 3-oxo-C₆-HSL, C₆-HSL, or C₄-HSL at a final concentrations of 1,500, 750, 375, 187.5, 93.8, 46.9, and 23.4 nM. After incubation of the plates at 30°C for 6 h, the green fluorescence of the monitor strains was measured using a microtiter reader (Synergy HT; Bio-Tek, Bad Friedrichshall, Germany) with an excitation wavelength of 485 nm and detection of emission at 528 nm.

Microarray data accession number.

The entire microarray data set has been deposited in the ArrayExpress database (http://www.ebi.ac.uk/arrayexpress) under accession number E-MEXP-1935.

RESULTS

Identification of CepR2 as a LuxR homolog unique to B. cenocepacia.

In silico identification of LuxR homologs was carried out by performing a BLAST search (2) of the B. cenocepacia J2315 genome sequence (33) with CepR (accession no. AAD12726) (52). Multiple ORFs with predicted helix-turn-helix DNA binding motifs typical of LuxR response regulators were identified throughout the genome. Only one of these ORFs, BCAM0188, also contained an AHL binding domain. BCAM0188 is located on chromosome 2 and does not have a proximal luxI homolog. BCAM0188 contains all seven residues that are conserved among LuxR transcriptional regulators (26).

BCAM0188 encodes a 237-amino-acid product with a predicted molecular mass of 26,046 Da. The BCAM0188 amino acid sequence is 40% identical to the sequences of the orphan LuxR homologs B. pseudomallei BpmR5 (88) and B. mallei BmaR5 (89). Using the nomenclature used for B. pseudomallei and B. mallei, the product encoded by BCAM0188 has been designated CepR2. CepR2 is 38% identical to Ralstonia solanacearum SolR (accession no. AAC4597) (22), 36% identical to B. vietnamiensis BviR (accession no. AAD12726) (57) and B. cenocepacia CepR (accession no. AAD12726) (52), and 21% identical to B. cenocepcacia CciR (33).

To determine the distribution of cepR2 in the Bcc, Southern hybridization was performed with 17 strains representing nine Bcc species (10, 60) (Fig. 1). The cepR2 probe hybridized to the three B. cenocepacia strains but not to any other Bcc species. To determine if cepR2 is common in different B. cenocepacia strains, a cepR2 fragment was PCR amplified from 11 B. cenocepacia strains in an experimental Bcc strain panel (60), as well as B. cenocepacia strains H111 and PC715j (Fig. 1). All of these B. cenocepacia strains contained cepR2. The distribution data, together with the amino acid identities, indicate that the cepR2 product is a distinct B. cenocepacia LuxR homolog.

FIG. 1.

Distribution of cepR2 in the Bcc. (A) Detection of cepR2 in Bcc strains by Southern hybridization. Genomic DNA was digested with BamHI and hybridized with a cepR2 probe labeled with [³²P]dCTP. Lane 1, B. cepacia ATTC 25416^T; lane 2, B. multivorans LMG13010^T; lane 3, B. multivorans C5393; lane 4, B. cenocepacia K56-2; lane 5, B. cenocepacia J2315; lane 6, B. cenocepacia Pc715j; lane 7, B. stabilis LMG14294; lane 8, B. stabilis LMG14086; lane 9, B. vietnamiensis PC259; lane 10, B. vietnamiensis G4; lane 11, B. dolosa LMG18943; lane 12, B. dolosa LMG19468^T; lane 13, B. ambifaria CEP0996; lane 14, B. ambifaria LMG17828; lane 15, B. athina LMG20980; lane 16, B. athina LMG20982; lane 17, B. pyrrocinia LMG21822; lane 18, cepR2 probe fragment. (B) Detection of cepR2 in B. cenocepacia strains by PCR. Products were amplified with the M188scF and M188scR primers and electrophoresed on a 1% agarose gel. Lane 1, 1-kb Plus ladder (Invitrogen); lane 2, C5424 (ET12 lineage); lane 3, BC7 (ET12 lineage); lane 4, K56-2 (ET12 lineage); lane 5, C6433; lane 6, C1394; lane 7, PC184 (Midwest lineage); lane 8, CEP511; lane 9, J415; lane 10, ATCC 17765; lane 11, Pc715J; lane 12, J2315; lane 13, H111; lane 14, no-template control.

Regulatory relationship between cepR2, cepIR, and cciIR.

A K56-2cepR2 mutant was constructed by insertional inactivation. Preliminary RT-PCR expression kinetic studies determined that the optimal time for cepR2 expression in the mutant was after 8 h of growth (data not shown). The expression of cepR2 was quantified by qRT-PCR analysis of cultures of K56-2, K56-R2 (cepR), K56-dI2 (cepI), K56-2cciR, K56-2cciI, and K56-2cepR2b. The expression levels of cepR2 were 170.3- ± 50.6-fold greater in K56-2cepR2b than in K56-2 (Table 2). K56-2cciR showed a 3.5- ± 2.0-fold increase in cepR2 expression compared to K56-2, while the cepR2 expression levels were unchanged in the other K56-2 mutants (data not shown). Together, these results indicate that cepR2 is maximally expressed during the mid- to late log phase of growth, that CepR2 is involved in negative autoregulation, and that CciR negatively regulates cepR2 expression. Transcriptional analysis was also performed to determine if a mutation in cepR2 affected the expression of the cepR and cciIR genes. The expression of cepR and cciIR was similar in K56-2 and K56-2cepR2b as determined by RT-PCR and qRT-PCR (data not shown). These data indicate that CepR2 is not involved in the transcriptional regulation of cepIR or cciIR.

TABLE 2.

Selected genes or operons regulated by CepR2

Gene	Functiona	Transcriptional unitb	Change (fold) for K56-2cepR2b vs K56-2
			Microarrayc	qRT-PCR or RT-PCRd
BCAL1234	Putative heat shock Hsp20-related protein	BCAL1234-BCAL1233	−2.2 ± 0.5	−17.9 ± 21
BCAL1233	Putative heat shock protein		−2.0 ± 0.5
bclA	Lectin BclA	BCAM0186	2.9 ± 0.8	2.2 ± 0.1
cepR2	Orphan LuxR homolog CepR2		21.1 ± 3.4	170 ± 50
BCAM0189	Putative AraC-family regulatory protein		188 ± 46	26.7 ± 6.6
BCAM0190	Putative aminotransferase, class III		11.6 ± 0.9
BCAM0191	Putative nonribosomal peptide synthetase	BCAM0191-BCAM0190	9.5 ± 3.0	—e
BCAM0192	Conserved hypothetical protein	BCAM0192-BCAM0196	113 ± 27	—e
BCAM0193	Conserved hypothetical protein		171 ± 74
BCAM0194	Hypothetical protein		151 ± 48
BCAM0195	Putative nonribosomal peptide synthetase		58.3 ± 62.1
BCAM0196	Conserved hypothetical protein		80.5 ± 31.1
BCAM0199	Putative outer membrane efflux protein	BCAM0199-BCAM0202	2.5 ± 0.6	—e
BCAM0200	Putative secretion protein, HlyD family		8.4 ± 5.6
BCAM0201	Putative transporter		2.3 ± 0.8
BCAM0202	Hypothetical protein		1.8 ± 0.2
BCAM1290	RpiR family transcriptional regulator		−1.7 ± 0.2	−2.6 ± 2.0
BCAM1420	Putative multidrug efflux protein		−2.3 ± 1.2	−8.1 ± 0.3
BCAM2307	Zinc metalloprotease ZmpB		−1.7 ± 0.7	2.6 ± 1.2f
BCAS0293	Nematocidal protein AidA		1.7 ± 0.1	5.8 ± 2.9f
BCAS0409	Zinc metalloprotease ZmpA		1.6 ± 0.3	1.2 ± 0.1f

^aFunction derived from B. cenocepacia J2315 (33).

^bTranscriptional units were experimentally determined by RT-PCR.

^cUnless indicated otherwise, change (mean ± standard deviation) in 8-h cultures of K56-2cepR2b compared to 8-h cultures of K56-2 as determined by microarray analysis with at least two biological replicates.

^dChange (mean ± standard deviation) in 8-h cultures of K56-2cepR2b compared to 8-h cultures of K56-2 as determined by qRT-PCR.

^eExpression was measured by RT-PCR (see Fig. 3).

^fExpression was measured in cultures grown for 18 h.

Phenotypic characterization of K56-2cepR2.

A variety of phenotypic tests were carried out with K56-2cepR2b to determine if CepR2 is involved in the regulation of known B. cenocepacia QS-regulated phenotypes. The only phenotypic change identified in the K56-2cepR2b mutant was that it had slightly greater protease activity than K56-2 on skim milk agar; however, the protease activity of the mutant was not restored to parental levels upon addition of cepR2 in trans (Fig. 2A). A second independently constructed cepR2 mutant (K56-2cepR2a) also produced significantly more protease than K56-2 (P < 0.001, analysis of variance, Student-Newman-Keuls test), suggesting that the hyper-protease-production phenotype of K56-2cepR2b was not due to a secondary mutation in K56-2. No differences were observed between K56-2cepR2b and K56-2 in any of the following phenotypic assays: AHL production profiles, swarming motility, biofilm formation on polystyrene pegs, pellicle formation at the air-liquid interface, resistance to a selection of heavy metals, antibiotics, and detergent, or sensitivity to heat shock at 42°C (data not shown).

FIG. 2.

Comparison of protease production and expression in K56-2 and K56-2cepR2b. (A) Protease production on skim milk agar plates as indicated by a zone of clearing around the culture growth. (B and C) Expression of zmpA::luxCDABE (B) and zmpB::luxCDABE (C) transcriptional fusions monitored in K56-2 (▪) and K56-2cepR2b (□) during growth in Luria broth. The results are expressed in cps/OD₆₀₀ unit. *, P < 0.05, unpaired t test, Welch corrected.

Identification of the CepR2 regulon.

Transcriptional profiling using a custom B. cenocepacia microarray was employed in order to determine whether CepR2 acted as a functional regulator of genes other than cepR2. Genome-wide expression profiles revealed differential expression of 191 genes (64 genes with increased expression and 127 genes with decreased expression) in K56-2cepR2b compared with K56-2 (see Tables S2 and S3 in the supplemental material). The greatest number of differentially regulated genes were located on chromosome 2 (91 genes); 76 and 24 genes were located on chromosomes 1 and 3, respectively. None of the 98 genes present on the plasmid were CepR2 regulated under the conditions examined (data not shown). The microarray data confirmed the lack of regulation of the cepIR or cciIR genes by CepR2 (data not shown). CepR2 differentially regulated proteins involved in virulence, chemotaxis, and motility and positively regulated proteins involved in heat shock and signal transduction.

Eight members of a cluster of 14 genes located adjacent to cepR2 exhibited a >10-fold increase in expression in K56-2cepR2b compared to K56-2 (Table 2), indicating that there was negative regulation by CepR2. Overall, this gene cluster was typically regulated to a greater extent than genes which were not adjacent to cepR2 (Table 2). BCAM0189, which is predicted to encode an AraC-family regulatory protein, showed the highest level of change (188-fold) among the genes represented on the microarray. Consistent with the qRT-PCR data, cepR2 expression as measured by microarray analysis was significantly increased in K56-2cepR2b (Table 2). RT-PCR was performed for a subset of these negatively regulated genes in order to experimentally verify the presence of putative transcriptional units and to validate the microarray data. Three multigene transcriptional units were identified, encompassing BCAM0191 and BCAM0190, BCAM0192 to BCAM0196, and BCAM0199 to BCAM0202 (Fig. 3A and B). The lectin-encoding gene bclA (46) is located upstream of BCAM0185 and BCAM0184 encoding two putative lectins. RT-PCR revealed that BCAM0185 and BCAM0184 were transcribed as a unit. However, no transcript was detected for a region encompassing BCAM0186 and BCAM0185, indicating that these genes are not cotranscribed (Fig. 3B), which is consistent with the fact that there was no difference in the transcription of either BCAM0185 or BCAM0184 between K56-2cepR2b and K56-2 as determined by the microarray analysis. Semiquantitative RT-PCR confirmed the microarray data for bclA, the putative AraC-family transcriptional regulator encoded by BCAM0189, the two operons putatively involved in the synthesis of nonribosomal peptides (BCAM0191 and BCAM0190 and BCAM0192 to BCAM0196), and the putative efflux pump (BCAM0199 to BCAM0202) (Fig. 4C). qRT-PCR confirmed that there was increased expression of bclA and BCAM0189 in K56-2cepR2b compared to K56-2 (Table 2).

FIG. 3.

Genomic organization, transcriptional units, and RT-PCR of genes adjacent to cepR2. (A) Scale diagram of ORFs represented by BCAM0186 to BCAM0202, with experimentally determined transcriptional units indicated by the bars below the arrows. Bar 1, BCAM0190 and BCAM0191; bar 2, BCAM0192 to BCAM0196; bar 3, BCAM0199 to BCAM0202. (B) Agarose gel electrophoresis indicating the absence or presence of PCR products generated from templates. Lanes a, negative control; lanes b, genomic DNA; lanes c, cDNA; lanes M, markers. The amplified regions are indicated above the gels. Primer pairs were designed to span intergenic regions to determine transcriptional units. (C) RT-PCR of BCAM0189, BCAM0191, BCAM0192, BCAM0199, and sigA performed with cDNA generated from 8-h cultures of K56-2 and K56-2cepR2b. (D) RT-PCR of BCAM0189, BCAM0191, BCAM0192, BCAM0199, and sigA performed with cDNA generated from 8-h cultures of K56-2 and K56-2cepR2b with the vector control (pUCP26) or cepR2 in trans (pRM613).

FIG. 4.

CAS activity of the B. cenocepacia parent H111, the cepR mutant H111-R, the cepR2 mutant H111-R2, and the cepR cepR2 double mutant H111-R/R2, as well as the complemented mutant strains. Strains were spotted on CAS plates and incubated for 2 days. wt, wild type.

Microarray analysis revealed that CepR2 positively regulates many genes, including genes encoding putative heat shock proteins (BCAL1234 and BCAL1233), a putative transcriptional regulator (BCAM1290), and a component of a putative multidrug efflux pump (BCAM1420). Subsequent analysis by qRT-PCR verified that there was reduced expression of BCAL1234, BCAM1290, and BCAM1420 (−17.9-, −2.6-, and −8.1-fold changes, respectively) in K56-2cepR2b compared to K56-2 (Table 2). Although the levels of the differences obtained in the microarray and qRT-PCR experiments were different, the two methods of analysis produced similar results for the selected genes examined.

Microarray analysis indicated that CepR2 influenced expression of the zinc metalloprotease genes, zmpA and zmpB (Table 2). The protease activity measured on skim milk agar was greater in K56-2cepR2b than in K56-2 (Fig. 2A), which suggested that at least one of these protease genes was negatively regulated. Temporal expression of these genes was monitored throughout growth using transcriptional fusions. Significantly increased expression of both zmpA::luxCDABE and zmpB::luxCDABE in K56-2cepR2b compared to K56-2 was observed between 18 and 24 h and between 14 and 18 h, respectively (P < 0.05, unpaired t test, Welch corrected) (Fig. 2B), indicating that both of these genes are negatively regulated by CepR2 and likely account for the differences in protease activity observed between the mutant and parent strains. The difference between the microarray and transcriptional fusion data for zmpB is likely due to the fact that zmpB expression is very low at 8 h (Fig. 2B), which corresponds to the culture time used for the microarray analysis. Using the transcriptional fusions, negative regulation was consistently observed at later stages of growth. To confirm these findings, qRT-PCR was performed using 18-h cultures, and the data were consistent with the transcriptional fusion data (Table 2).

B. cenocepacia virulence against the nematode Caenorhabditis elegans is positively influenced by AidA protein production (34). The increase in aidA expression in K56-2cepR2b compared to K56-2 observed by microarray analysis was independently confirmed by qRT-PCR (Table 2).

Effect of cepR2 in trans on expression of CepR2-regulated genes.

An RT-PCR gene expression analysis was performed with K56-2cepR2b and K56-2 carrying either pUCP26 or pRM613 (pUCP26 containing cepR2) to determine if it was possible to restore CepR2-regulated gene expression to parental levels. Introduction of pRM613 into K56-2cepR2b slightly increased expression of BCAM0189, BCAM0191, and BCAM0199 compared to the results obtained with the vector control rather than reducing expression (Fig. 3D). Surprisingly, K56-2(pRM613) had markedly increased levels of expression of these three genes plus BCAM0192 compared to K56-2(pUCP26) (Fig. 3D). In addition, qRT-PCR was employed for analysis of one of these negatively regulated genes (BCAM0189) and two positively regulated genes (BCAL1234 and BCAM1420). Introduction of cepR2 in trans into K56-2 led to increased expression of BCAM0189 and decreased expression of BCAL1234 and BCAM1420 (data not shown). The phenotype was also altered in the presence of the vector alone. The negative regulation exerted by CepR2 on cepR2 and BCAM0189 in the independently constructed K56-2cepR2a mutant was confirmed by qRT-PCR since these genes were found to be upregulated 13- ± 4-fold and 112- ± 13-fold, respectively, in the mutant compared to K56-2.

CepR2 is involved in the regulation of pyochelin production in B. cenocepacia H111.

Although several CepR2-regulated genes were identified in strain K56-2, the only phenotypic difference observed between the cepR2 mutant and the parent strain was a difference in protease activity. Attempts to complement either gene regulation or protease expression were unsuccessful. Since CciIR negatively regulates cepR2 expression, a defined cepR2 mutant (H111-R2) and a cepR cepR2 double mutant (H111-R/R2) of strain H111, which lacks the cciIR QS genes, were constructed to facilitate the analysis of cepR2.

H111-R2 was indistinguishable from H111 with respect to biofilm formation, swarming motility, and proteolytic activity. There was also no difference in virulence in a C. elegans infection model (data not shown). Unlike the results for K56-2, however, significantly decreased production of siderophores was observed in H111-R2 compared to H111 (Fig. 4). No significant siderophore activity was observed in H111-R2, the cepR mutant H111-R, and the double mutant H111-R/R2. In addition, the presence of cepR2 on plasmid pJTR2 greatly stimulated siderophore production to a level that was well above the level of production in the parent strain. When cepR was provided in trans on plasmid pBBR-cepR, the mutant phenotype of H111-R, but not the mutant phenotype of H111-R2 or H111-R/R2, was rescued. These data demonstrate that the production of at least one siderophore is CepR2 regulated and suggest that there is a regulatory cascade in which CepR is required for CepR2 expression in H111.

Since B. cenocepacia produces at least two siderophores, ornibactin and pyochelin, the contribution of these two compounds to the overall siderophore activity seen on CAS plates was determined. Supernatants of cultures grown in minimal medium were extracted with dichloromethane, and the organic and aqueous fractions were spotted on CAS plates (Fig. 5). Analyses of the organic fractions revealed that H111 produced only minute amounts of pyochelin under these conditions. Likewise, extracts of H111-R, H111-R2, or H111-R/R2 cultures showed no activity, and complementation of the mutant strains with cepR did not stimulate pyochelin production. Complementation with cepR2 resulted in strong overproduction of pyochelin in all three mutants, indicating that CepR2 is a major regulator of pyochelin biosynthesis (Fig. 5A). Analysis of the aqueous fractions showed that CepR2 has no influence on ornibactin production. However, we observed that inactivation of cepR stimulated ornibactin synthesis and that complementation of the mutants with cepR repressed ornibactin production (Fig. 5B). These results are in full agreement with previous work demonstrating that the CepIR system is a negative regulator of ornibactin biosynthesis in B. cenocepacia K56-2 (53). In this context it is also important to note that K56-2 is unable to produce pyochelin as a result of a point mutation in the pyochelin synthetase gene pchF (BCAM2228) (33).

FIG. 5.

Production of pyochelin (A) and ornibactin (B) in the B. cenocepacia parent H111, the cepR mutant H111-R, the cepR2 mutant H111-R2, and the cepR cepR2 double mutant H111-R/R2, as well as the complemented mutant strains. Supernatants of cultures grown in minimal medium were extracted with dichloromethane, and the organic and aqueous fractions were analyzed separately on CAS plates. Extracts of supernatants of B. cenocepacia H111 and K56-2 were included as controls (wt).

CepR2 positively regulates pyochelin biosynthesis.

To analyze the role of CepR2 in the regulation of pyochelin biosynthesis, the promotorless lacZ gene was fused to the promoter regions of pchR (encoding a regulator of pyochelin production) and pchD (the first gene in the pyochelin biosynthesis gene cluster). In Luria-Bertani medium the pchD promoter was inactive (data not shown), while the pchR-lacZ fusion showed high activity in the early growth phase in both H111 and H111-R (Fig. 6). Consistent with our data on pyochelin production, the pchR-lacZ fusion did not show any activity in H111-R2. In succinate minimal medium both promoters were active in H111 (see Fig. S1 in the supplemental material). As it was in Luria-Bertani medium, expression of the pchR-lacZ fusion was dramatically decreased in the cepR2 mutant, while the activity of the pchD-lacZ fusion was slightly reduced in the cepR cepR2 mutant. These data support a model in which CepR negatively regulates expression of ornibactin biosynthesis genes and positively regulates expression of cepR2, which in turn is required for full expression of pyochelin biosynthesis genes.

FIG. 6.

CepR2 controls activity of the pchR promoter. The activities of the pchR promoter driving expression of one of the pyochelin biosynthesis operons were determined using Luria-Bertani medium. Growth (open symbols) and β-galactosidase activity (filled symbols) were monitored throughout the growth curve. The transcriptional fusions were measured in the parent B. cenocepacia H111 (squares), in the cepR mutant H111-R (triangles), and in the cepR2 mutant H111-R2 (circles). The values are means ± standard deviations (n = 3).

CepR2 activates transcription of target promoters in the absence of AHL signal molecules.

Pyochelin production by the cepR mutant and the cepR cepR2 double mutant was restored by complementation with cepR2 (Fig. 5A). As neither of these mutants produces AHLs, we hypothesized that CepR2 may not require AHL signal molecules for its activity. To test this hypothesis, plasmid pJTR2 was transferred into E. coli MT102 harboring the green fluorescent protein (GFP)-based AHL sensor plasmid pJBA89 (3). This sensor is based on components of the LuxIR QS system of Vibrio fischeri and thus is most sensitive to 3-oxo-C₆-HSL, but it is also responsive (with decreased sensitivity) to various related AHL molecules (Fig. 7A). In the presence of cepR2 (pJTR2), the GFP expression of the sensor plasmid was strongly induced, and addition of AHL signal molecules did not further increase fluorescence (Fig. 7C). To rule out the possibility that the luxR gene on pJBA89 influenced the results, luxR was deleted, generating plasmid pJBA89luxR⁻. As expected, this sensor plasmid could not be activated by external addition of AHLs (Fig. 7B). However, when plasmid pJTR2 (containing cepR2) was provided in trans, GFP expression was strongly stimulated independent of the presence or absence of AHLs (Fig. 7D). Taken together, these results strongly suggest that CepR2 acts as a transcriptional activator in an AHL-independent manner.

FIG. 7.

CepR2 activates a PluxI-gfp transcriptional fusion in an AHL-independent manner. The GFP expression of the AHL biosensor E. coli MT102(pJBA89) (A and C) and its luxR deletion derivative E. coli MT102(pJBA89luxR⁻) (B and D) was measured in the absence (A and B) and in the presence (C and D) of plasmid pJTR2 (cepR2⁺). Measurements were also obtained in the absence or presence of various amounts of different AHL signal molecules. RFU, relative fluorescence units.

As the sensor plasmid contains a lux box operator sequence that is similar but not identical to the cep box sequence, the effect of CepR2 on transcription of a cepI-lacZ fusion in an E. coli background was determined. The activity of this promoter fusion was very low (<1 Miller unit) independent of whether cepR2 was present in trans (data not shown). Addition of 200 nM C₈-HSL to the medium had no effect on expression, indicating that CepR2 does not recognize the cep box sequence upstream of cepI but may bind a somewhat different operator sequence that is more similar to a lux box sequence.

We also measured the activity of the pchR promoter in E. coli in the presence or absence of cepR2. Approximately threefold C₈-HSL-independent induction of β-galactosidase activity was observed in the presence of cepR2, although the promoter activity was very low (1.3 ± 0.6 and 5.2 ± 1.9 Miller units in the absence and presence of cepR2, respectively). Given the strong effect of CepR2 on transcription of pchR in H111, it has to be assumed that either additional host factors are required for full activation of the promoter or CepR2 controls expression of a downstream regulator, which stimulates pchR transcription. The fact that inspection of the promoter region did not reveal any obvious lux or cep box-like sequences may favor the latter possibility.

DISCUSSION

QS mediated by the CepIR and CciIR systems in B. cenocepacia regulates virulence gene expression and contributes to pathogenesis in infection models. In this study we show that the B. cenocepacia QS system includes a third component, an orphan LuxR homolog, CepR2, that regulates expression of numerous genes. The cepR2 gene was identified in all B. cenocepacia strains examined, suggesting that, unlike cciR, cepR2 is not associated with epidemic strains. Analysis of sequenced genomes at www.burkholderia.com and www.broad.mit.edu revealed that cepR2 is located on an 11-kb genomic segment that is present in B. cenocepacia strains HI2424, AU 1054, PC184, and MC0-3 but not in B. lata 383, B. multivorans strain ATCC 17616, B. vietnamiensis strain G4, or B. xenovorans strain LB400, which confirms the Southern and PCR data.

Microarray analysis demonstrated that CepR2 regulates genes on each chromosome and that CepR2 repressed more genes than it activated. Included in the set of repressed genes were genes located in the cepR2 genomic region and cepR2 itself. Self-regulation of LuxR homologs is common in Burkholderia species and other species (53, 63, 70, 73). CepR and CciR both feedback inhibit their own expression, and CciR negatively regulates cepR and cepR2 expression (53, 63; this study). CepR2 is not transcriptionally regulated by CepR in strain K56-2. However, in strain H111, cepR expression was required for CepR2-mediated positive control of siderophore activity in strain H111. This indicates that the B. cenocepacia QS hierarchy is different in strains K56-2 and H111, which may in part result from the absence of cciIR in the latter strain, in addition to possibly other unidentified components. CepR2 also negatively regulated zmpA and zmpB protease gene expression in K56-2; however, there was no detectable difference in protease production in the H111 cepR2 mutant. Whether this was due to the influence of CciIR or other proteases contributing to the activity observed on skim milk plates was not examined.

Our studies determined that the CepR2 regulon includes some genes previously determined to be regulated by the CepIR and CciIR QS systems. For example, zmpA, zmpB, and aidA are negatively regulated by CepR2. CciR negatively regulates aidA expression (R. J. Malott and P. A. Sokol, unpublished data), whereas CepR positively regulates aidA (1, 34, 74, 98). CepR2-mediated negative regulation may be a mechanism to balance the timing and level of gene expression. Early expression of CepR2 followed by negative autoregulation appears to limit the influence of CepR2, thus allowing CepR and CciR to exert their influence in late growth phase.

A number of genes were selected for further analysis because they represent novel QS-regulated genes in B. cenocepacia and they may contribute to the virulence of this organism. Lectins play a role in adhesion, biofilm formation, and host recognition (37). Of three recently identified lecB-like genes in B. cenocepacia (46), CepR2 regulates only the BclA gene. The members of the AraC family of regulators generally act as positive transcriptional regulators and have been shown to influence pathogenesis in a number of species, including Pseudomonas, Salmonella, Yersinia, and Vibrio species (27). The largest differential effect on expression observed in the K56-2cepR2b mutant was the effect of an araC regulatory gene, BCAM0189, which is transcribed divergently from cepR2. Sequence analysis of the intergeneric region identified a 20-bp imperfect palindrome with a high level of similarity to cep box sequences (data not shown). It is possible that the derepression of BCAM0189 by mutation of cepR2 contributed to differences in expression of differentially regulated genes adjacent to cepR2 identified in the array experiments.

CepR2 was shown to regulate the expression of pchR in strain H111, which is an AraC regulator required for pyochelin synthesis and transport (29, 30, 68). In P. aeruginosa, PchR induces expression of pchD and other promoters involved in pyochelin transport. In H111, the pchR-lacZ fusion was more strongly induced than the pchD-lacZ fusion in the parent compared to the cepR2 mutant. It is possible that the induction of the pchD promoter by CepR2 is mediated by PchR. In P. aeruginosa, pyochelin is required as an effector for pchR binding to target promoters (69). This is likely also the case in B. cenocepacia since expression of the pyochelin genes was not detectable in K56-2, possibly due to the mutation in the pyochelin biosynthesis pathway. This would also account for the low levels of expression of the pyochelin promoters in E. coli.

Expression of cepR2 in trans led to unexpected effects on target gene transcription. In K56-2, expression of cepR2 in trans led to increased transcription of several genes which exhibited increased expression in K56-2cepR2b compared to K56-2. Similar results were obtained whether cepR2 was expressed from a high-copy-number vector (pUCP26) or a low-copy-number vector (pRK415). Increased protease activity detected on skim milk agar was also observed in K56-2 containing cepR2 carried on several different expression vectors (data not shown). Similar results were obtained for strain K56-2 containing pJTR2 (cepR2) (data not shown). Expression of cepR2 in trans in cepR or cepR2 mutants of strain H111 also led to overproduction of pyochelin at levels greater than those seen in the parent strain. A possible explanation for these apparent discrepancies is that the expression vectors supply artificially high levels of constitutively active CepR2, whereas the usual level of the cepR2 transcript is barely detectable in K56-2 and H111.

Some, but not all, orphan LuxR homologs require AHLs in order to function. P. aeruginosa QscR and Sinorhizobium meliloti ExpR, both of which are orphan regulators, require AHL binding in order to properly function (31, 51). QscR requires N-(3-oxo-decanoyl)-homoserine lactone to actively bind DNA, but it may also utilize other AHLs (49). In Xanthomonas oryzae pv. oryzae, a variety of exogenous AHLs are not sufficient to solubilize the orphan LuxR homolog, OryR; however, solubilization occurs in the presence of a rice signal molecule (19). Furthermore, OryR regulates target genes in the absence or presence of the rice signal molecule (20). Protein solubilization and AHL binding are important factors when workers consider how LuxR regulators exert their effects. The positive regulatory effect of CepR2 on luxI promoter activity was demonstrated using a heterologous E. coli host both in the absence and in the presence of AHLs. CepR2 was able to restore pyochelin production in the H111 cepR mutant, which does not synthesize any AHLs. These data indicate that CepR2 does not require AHL for solubility or activity. It is possible that CepR2 may be constitutively active and may not be modulated by endogenously or exogenously produced AHLs. In this scenario, control of CepR2 expression levels could be the most important factor in determining the influence of CepR2. Previous studies have identified LuxR homologs which are active in the absence of AHLs. In Pantoea stewartii, the LuxR homolog EsaR binds target DNA in the absence of AHL in order to repress transcription. Subsequent AHL binding by EsaR leads to derepression (97). Recent work has also shown that B. pseudomallei orphan LuxR homologs can exert regulatory effects in the absence of AHL. BpsR4 and BpsR5 decrease bpsI3 expression, while BpsR5 activates bpsI1 expression in the absence of AHL (41).

Reversible AHL binding may occur in LuxR homologs which do not require an AHL for proper folding (18). Preliminary binding studies using an E. coli strain expressing CepR2 suggested that CepR2 is indeed capable of binding various AHL molecules (K. Riedel and L. Eberl, unpublished), and thus the possibility that CepR2 competes with CepR and CciR for signal molecules cannot be ruled out. Nor can we entirely exclude the possibility that CepR2 is in fact activated by an unknown molecule and thus is particularly important in a specific niche.

It is known that CepR requires C₈-HSL in order to fold and be active (98). AHL sequestration by CepR2 would limit the activity of CepR and possibly CciR so that CepR2 regulation of certain target genes would be achieved in an indirect manner. Alternatively, CepR2 could directly bind CepR- and CciR-controlled promoters or form inactivating heterodimers with CepR or CciR, as has been postulated for QscR inactivation of LasR and RhlR (48). Work is under way to test these possibilities in order to continue elucidating the contribution of CepR2 to the B. cenocepacia QS network.