Identification of potential CepR regulated genes using a cep box motif-based search of the Burkholderia cenocepacia genome

Background The Burkholderia cenocepacia CepIR quorum sensing system has been shown to positively and negatively regulate genes involved in siderophore production, protease expression, motility, biofilm formation and virulence. In this study, two approaches were used to identify genes regulated by the CepIR quorum sensing system. Transposon mutagenesis was used to create lacZ promoter fusions in a cepI mutant that were screened for differential expression in the presence of N-acylhomoserine lactones. A bioinformatics approach was used to screen the B. cenocepacia J2315 genome for CepR binding site motifs. Results Four positively regulated and two negatively regulated genes were identified by transposon mutagenesis including genes potentially involved in iron transport and virulence. The promoter regions of selected CepR regulated genes and site directed mutagenesis of the cepI promoter were used to predict a consensus cep box sequence for CepR binding. The first-generation consensus sequence for the cep box was used to identify putative cep boxes in the genome sequence. Eight potential CepR regulated genes were chosen and the expression of their promoters analyzed. Six of the eight were shown to be regulated by CepR. A second generation motif was created from the promoters of these six genes in combination with the promoters of cepI, zmpA, and two of the CepR regulated genes identified by transposon mutagenesis. A search of the B. cenocepacia J2315 genome with the new motif identified 55 cep boxes in 65 promoter regions that may be regulated by CepR. Conclusion Using transposon mutagenesis and bioinformatics expression of twelve new genes have been determined to be regulated by the CepIR quorum sensing system. A cep box consensus sequence has been developed based on the predicted cep boxes of ten CepR regulated genes. This consensus cep box has led to the identification of over 50 new genes potentially regulated by the CepIR quorum sensing system.


Background
Burkholderia cenocepacia, belongs to a group of nine related species with common phenotypes, but distinct genotypes collectively named the "Burkholderia cepacia complex" (Bcc) [1,2]. The Bcc are opportunistic pathogens in immunocompromised and cystic fibrosis (CF) patients but have also been isolated from plant rhizopheres as well as urban and suburban soils [1][2][3].
The ability of bacteria to adapt to diverse environments is dependent on the coordinate regulation of factors required to survive and proliferate in each niche. The CepIR quorum sensing system is one regulatory network that contributes to the response of B. cenocepacia to environmental signals (reviewed in [4,5]). Quorum sensing allows bacterial populations to coordinate gene expression in response to population density. CepIR belongs to a group of more than 50 quorum sensing systems that are homologous to the LuxIR system of Vibrio fishceri [6,7]. LuxI homologs are N-acyl homoserine lactone (AHL) synthases that generate AHL signal molecules that are released into the environment. LuxR homologs are transcriptional regulators that complex with AHL and typically bind to a lux-box overlapping the -35 sequence of a promoter to regulate transcription. The lux-box consensus sequence recognized by LuxR homologs typically consists of an inverted repeat with significant consensus among quorum sensing systems [6,[8][9][10]].
The CepIR system was originally identified in B. cenocepacia (formerly B. cepacia) K56-2 [11] and has subsequently been shown to be widely distributed throughout the Bcc [12,13]. CepI directs the synthesis of N-octanoyl homoserine lactone (OHL) and N-hexanoyl homoserine lactone (HHL) and cepR encodes for the transcriptional regulator [11][12][13][14]. CepR has been shown to negatively regulate its own expression, but positively regulate cepI expression at the transcriptional level [14]. The cepIR genes are involved in the regulation of the pvdA gene required for ornibactin biosynthesis [14], the zmpA and zmpB extracellular metalloproteases [15,16], the aidA gene involved in virulence in Caenorhabditis elegans [17][18][19][20], swarming motility and in at least some systems a functional CepIR quorum sensing system is necessary for biofilm formation [21][22][23]. The CepIR system has been shown to contribute to virulence in both plant and animal models. In B. cepacia ATCC 25416 mutations in cepI and cepR attenuated maceration in the onion-rot model [24]. The contribution of CepIR to the severity of B. cenocepacia infections has been demonstrated in two different animal models, a chronic respiratory infection model in rats and an acute intranasal infection model in Cftr (-/-) mice [16]. CepIR have also been shown to be important for virulence in C. elegans [25].
Proteomics and promoter based approaches have been used to identify genes regulated by the CepIR quorum sensing system. Proteome analysis was used to compare the protein profiles of B. cenocepacia strain H111 and an H111 cepI mutant [19]. Differences in expression were observed for 55 out of 585 proteins and partial N-terminal amino acid sequences were determined for peptide fragments of 11 proteins including AidA, FimA, and SodB. A promoter trap approach was used to identify positively regulated OHL-CepR dependent promoters in B. cepacia ATCC 25416 [17]. A library of ATCC 25416 fragments cloned upstream of a promoterless lacZ gene in a vector that also contained cepR was screened in E. coli in the presence and absence of OHL. Twenty-eight clones with genes upregulated in the presence of OHL were identified. The genes belonged to several functional classes; however, the only overlap in genes identified between the two studies was aidA [17,19]. Mutagenesis with a transposon containing a promoterless lacZ reporter was used to identify seven genes positively regulated by the cepIR quorum sensing system in B. cenocepacia strain K56-2, including cepI and aidA [20].

Identification of genes directly and indirectly regulated by
CepR is a key step to understanding this regulatory system and the regulatory hierarchies that mediate the adaptation B. cenocepacia to the diverse environments it encounters. The above approaches search for genes regulated under defined in vitro conditions and therefore may not identify genes induced only in specific environmental niches including the plant or animal host. Only the study by Aguilar et al. [17] attempted to identify genes that are regulated by the direct interaction of CepR at the promoter.
LuxR homologs have been shown to bind to specific sequences referred to as lux boxes or the boxes for the gene designation of the respective luxR homolog such as tra boxes in the case of recognition sequences for Agrobacterium tumefaciens TraR [26][27][28]. These sequences have dyad symmetries and generally overlap the -35 RNA polymerase binding site. Lewenza et al. demonstrated that CepR was required for the expression of cepI in B. cenocepacia [11,14] and identified a putative lux-box like sequence with imperfect repeats that overlapped the -35 region of the putative cepI promoter [11]. Weingart et al. [20] demonstrated that CepR directly bound to a DNA fragment that contained the cepI promoter using electrophoretic mobility shift assays. They also mapped the transcriptional start site of cepI and using DNAase I footprinting experiments localized the CepR binding site to a region that closely corresponded to the cep box predicted by Lewenza et al [11]. In the present study, we used a functional genomics approach to identify genes in the B. cenocepacia J2315 sequence with a cep box-like sequence in their promoters. We confirmed by site-directed mutagen-esis the cep box sequence located upstream of the cepI gene that is necessary for cepI transcription. Using selected B. cenocepacia CepR regulated genes we predicted a consensus cep box motif sequence and used that motif to search the B. cenocepacia J2315 genome to identify promoters potentially regulated by CepR.

Functional analysis of the CepR binding site
Lewenza et al. identified a potential cep box sequence upstream of cepI [11]. Weingart et al. demonstrated using DNAaseI footprinting of the cepI promoter that CepR protected a region of DNA that corresponded to the predicted cep box [20]. To confirm that the cep box is required for cepI transcription, eleven mutations, each with a 4 bp substitution, were introduced into the region -59 bp to -18 bp from the transcriptional start site of cepI (Fig. 1A). BamHI-XhoI fragments containing the mutations were subcloned into pMS402 directly upstream of the promoterless lux-CDABE operon [29]. The luxCDABE fusions (pCPI302 to pCPI313) were introduced into K56-2 and expression determined by measuring luminescence (Fig. 1B). The K56-2 cepI::luxCDABE fusions with mutations within the 24 bp inverted repeat (pCPI304-310) had luminescence levels below 20% of the wild type K56-2 (pCPI301), whereas promoter fusions containing mutations flanking the inverted repeat (pCPI303, and pCPI311-314) expressed at levels either similar to or higher than wildtype.

Identification of CepR regulated genes by transposon mutagenesis
Nine Tn5-OT182 transposon insertion mutants in K56-I2 were identified with differences in β-galactosidase activity on TSB-DC agar with AHL extract and TSB-DC agar without AHL extract. Expression of β-galactosidase activity was increased in the presence of OHL in six mutants and, expression was decreased in three mutants. To locate the Tn5-OT182 insertions in these mutants, the flanking genomic DNA was cloned, sequenced and the sequence obtained was used to search the B. cenocepacia J2315 genome with BlastN to identify the gene containing the insertion (Table 1). A total of 7 distinct genes in 5 regions of the genome were identified. K56-I2-P12, K56-I2-2PB2 and K56-I2-P9 had three distinct insertions within a few hundred base pairs of each other. The P12 transposon inserted into a hemin specific ATPase similar to the phuV gene of Pseudomonas aeruginosa involved in heme iron acquisition [30]. The phuV homolog was predicted to be in an operon with phuR and phuSTUV homologs to of P. aeruginosa. The phuR gene has been shown to be positively regulated by quorum sensing in P. aeruginosa [31]. The insertion in K56-I2-2PB2 transposon was also located in phuV; however, in this case the lacZ fusion was in the opposite orientation to that of the gene. K56-I2-P9 had an insertion in a hypothetical protein which appears to be in an operon and located directly downstream of a pbp1 homolog. K56-2-P1 and P2 were sibling insertions within a predicted acyltransferase that may be involved in lipid metabolism (COG1835). Directly upstream of the acyltransferase is a class D β-lactamase, likely an oxacillin hydrolase. The insertion in K56-I2-P3 was located in a gene, subsequently designated scpB, which belongs to the serine-carboxyl proteinase family [32]. K56-I2-P5 and K56-I2-P10 contained insertions located in aidA, which was also identified in the transposon mutagenesis screen used by Weingart et al. [20]. K56-I2-NB12 contained an insertion in an ORF that has a conserved domain (COG4774) shared with several outer membrane receptors involved in uptake of catechol siderophores, although the other genes flanking this insertion do not appear to be involved in iron acquisition. The insertion in K56-2-2PB2 did not appear to be in a gene. This insertion may result in creation of an artificial promoter-lacZ fusion or influence expression of a regulatory RNA.
To confirm the observations in the plate assay, expression of the unique AHL responsive lacZ fusions was examined over a 24 hr time course in the presence and absence of OHL extract. The growth rates for each mutant were similar to the parent strain K56-I2 ( Fig. 2A), indicating that the insertions did not result in growth defects that might influence lacZ expression. Expression of the Tn5-OT182 fusions in K56-I2-P1 (Fig. 2C) and K56-I2-P10 (Fig. 2D) were similar to that observed for a cepI-lacZ fusion (Fig.  2B). There was little expression in the absence of OHL and expression increased in the presence of OHL. The expression of the K56-I2-P12 fusion was also increased by the presence of OHL in the culture medium but expression started slightly earlier in growth and decreased after 10 hr (Fig. 2E). Three of the insertions appear to be negatively regulated by cepR since their expression was higher in the absence of OHL and decreased markedly when OHL was added to the culture medium (Fig. 2FGH). Positive regulation of β-galactosidase activity was observed for the K56-I2-P3 insertion in the presence of AHL on the plate assay; however, this fusion expressed very poorly in liquid medium (data not shown). When K56-I2-P3 grown on agar plates was analyzed for β-galactosidase activity, expression was significantly higher in cultures from plates supplemented with AHL (data not shown).
The predicted promoter regions for the three positively regulated genes containing the Tn5-0T182 insertions, phuV, aidA and the acyltransferase, were cloned into pMS402 and expression of the resulting promoter-luxCD-ABE fusions was determined in K56-2, K56-R2 (cepR) and K56-dI2 (cepI) with and without OHL in the medium. The aidA promoter fusion, pAID301, had an expression pattern similar to the cepI promoter with significant activity Functional analysis of the CepR binding site Figure 1 Functional analysis of the CepR binding site. A. Site directed mutagenesis was used to determine the effects of mutations on the luminescence activity of a cepI::luxCDABE fusion. The sequence upstream of the CepI ORF is shown. The ATG start codon is indicated by bold lettering and the predicted -10 hexamer is underlined. A series of 4 bp substitutions used to mutate the promoter region are indicated as bx303-313 and the cep box consensus sequence is enclosed in the rectangle. B. Expression of the cepI::luxCDABE fusions in B. cenocepacia K56-2. Luminescence (CPM) was measured at 22 hours and is represented as CPM/O.D. The numbers on the x axis indicate K56-2 (pCPI303-313) respectively. WT is K56-2 (pCPI301) and the vector control is K56-2 (pMS402).
in K56-dI2 only when OHL was added to the medium ( Fig. 3A and 3B). This expression pattern was similar to the chromosomal Tn5-OT182 lacZ fusion. Expression of the acyltransferase was increased in K56-dI2 in the presence of OHL; however, expression of this fusion in K56-R2 was intermediate between that in K56-dI2 and the parent strain (Fig. 3C). The phuV homolog was predicted to be in an operon with the promoter upstream of a phuR homolog and therefore the phuR promoter was cloned into pMS402. Expression of the phuR promoter was similar to K56-2 until early stationary phase where expression was significantly lower in K56-R2 and K56-dI2 in the absence of OHL (Fig. 3D). Expression of phuR::luxCDABE was slightly enhanced in the presence of OHL in stationary phase. The pattern of expression of the phuR::luxCD-ABE was similar to that of the phuV::lacZ chromosomal fusion (compare Fig. 2E and Fig. 3D). Expression of the scpB promoter was very weak in both the presence and absence of OHL suggesting different growth conditions are required for scpB expression (data not shown).

Construction of the first generation cep box motif and search of the B. cenocepacia genome for match sequences
To identify a consensus cep box motif to search the B. cenocepacia genome for potential CepR regulated genes, promoter regions from cepI, aidA, phuR, the acyltransferase gene identified in K56-I2P2, scpB, and zmpA, which was previously shown to be CepR regulated [16], were analyzed using MEME to identify common motifs. Only positively regulated promoters were analyzed in case there were differences in cep box consensus sequences for positively and negatively regulated promoters. A motif that recognized the defined cep box upstream of the cepI gene was identified using these promoters as the input file ( Table 2). The motif included bp 2-19 of the 24 bp palindrome required for transcription that contained the cep box for the cepI promoter. A single copy of the motif was found in all six of the promoters submitted. The most conserved nucleotides in the 18 bp motif were position 2 (T), 6 (A), 9 (G) and 18 (T). The position specific scoring matrix was then used to search the B. cenocepacia J2315 genome using the MAST program. The search returned 148 hits (numbered consecutively starting from MST001) including the 6 original input sequences (data not shown). The surrounding sequence for each hit was annotated and 49 were located upstream of predicted ORFs. The remaining hits were either within the coding sequence of an ORF or found in non-coding regions.
To determine if the putative cep box sequences identified were potentially involved in CepR regulation of downstream genes, eight of the promoter regions identified that were located within 40-250 bp upstream of a predicted ORF were cloned into pMS402 and expression of the resulting luxCDABE fusions was compared in K56-2, K56-dI2 and K56-R2. The three matching motifs with the lowest E-values and five arbitrarily selected motif matches were selected for analysis. When the motifs were located between two putative divergent promoters, one promoter region was chosen for further analysis. The predicted promoters containing putative cep box motifs were located upstream of the following orfs: BCAL0340, a TPR repeat protein (MST005); BCAL0715, a LysR-type transcriptional regulator (MST011); BCAL1354, a conserved hypothetical protein (MST028); BCAL2739, fusA (MST052); BCAL3191, caiA (MST059); BCAM0009, a transcriptional regulator (MST068); BCAM077, hydroxylase (MST072); and BCAM1943, a transcriptional regulator (MST112). The luxCDABE fusions containing the MST005, MST011, MST028, MST059 and MST072 sequences had expression patterns similar to cepI in that expression was higher in K56-2 than in K56-dI2 or K56-R2 and expression was increased in K56-dI2 in the presence of OHL (Fig.  4A,4B,4C,4E and 4G), although expression varied for some fusions depending on the stage of growth. For example, expression of the MST028 fusion peaked at 6 hours and decreased over the remainder of the assay (Fig. 4C). Expression of MST068 was only decreased in K56-R2 in stationary phase although expression was lower in K56-dI2 than in K56-2 and expression in K56-dI2 increased Effect of OHL on β-galactosidase activity in K56-I2 Tn5-OT182 mutants when the medium was supplemented with OHL ( Fig. 4F). MST112, did not appear to be affected by the cepR mutation although expression was lower in K56-dI2 without OHL (Fig. 4H). MST052 did not demonstrate any regulation by CepR in the conditions examined (Fig. 4D).

Construction of the second generation cep box motif and search of the B. cenocepacia genome for potential cep boxes
To improve the specificity of the cep box motif the six promoters with cep box motifs identified by the MAST program with expression patterns similar to that expected for cepIR regulated genes (MST005, MST011, MST028, MST059, MST068 and MST072) were used with the pro-moters for cepI, phuR, aidA and zmpA to generate a second generation cep box consensus motif using MEME ( Table  2). The promoters for scpB, the acyltransferase, MST052 and MST112 did not share the same expression pattern, and therefore were not included. The resulting second generation cep box had the same sequence as the original motif; however the specific score for each position had changed (Fig. 5). The most conserved residues in the second generation motif were in positions 6 (A), 8 (A), 10 (T), 16 (G) and 18 (T).
The new PSSM file was used to search the B. cenocepacia J2315 genome, resulting in 72 sequences matching the motif. Fifty-five of these matches (76%) were potentially Expression of promoter::luxCDABE fusions for OHL responsive genes identified by K56-I2 Tn5-OT182 mutagenesis within a promoter region although it must be noted that the transcriptional start sites of these genes have not been experimentally determined. The genes or operons predicted to be downstream of these matching sequences are listed in Table 3. Both MST designations are included in Table 3 for the six first generation MSTs used in the second generation motif search. Several of the cep boxes identified in the second search had more significant E-values than at least one of the input sequences (data not shown). A cep box was identified upstream of cepR (MST2058), using the second motif. This was the only gene previously determined to be regulated by CepR identified. MST112, which was identified with the first motif, but did not appear to be CepR regulated (Fig. 4H), was not identified with the second motif. Potential cep box sequences were identified on all three chromosomes and the plasmid, suggesting that CepR regulated genes are distributed throughout the genome. Genes downstream of promoters containing cep boxes were classified into seven categories: cell surface or membrane protein genes, genes encoding hypothetical proteins, phage genes, regulatory genes, genes involved in secretion or transport, and genes encoding proteins of unknown function (Table 3). In ten cases the putative cep boxes were located between predicted divergent promoters. In these situations orfs located both downstream and upstream of the cep box are included in Table 3 since it would be possible that cepR regulates genes in both directions. An alignment of the putative cep boxes for each of the MST sequences listed in Table 3 is shown in Fig. 6. The most conserved residues are in position six (A), eight (A), ten (T), sixteen (G) and eighteen (T) which correlates with the motif used in the MEME input file. Further studies are needed to determine if the genes downstream of these predicted promoters and cep box motifs are regulated by CepR.

Discussion
In this study we used a computational genome screen and experimental approaches to identify cepR regulated genes in B. cenocepacia. Transposon mutagenesis was used to The p-value of a site is computed from the the match score of the site with the position specific scoring matrix for the motif. The p-value gives the probability of a random string (generated from the background letter frequencies) having the same match score or higher. (This is referred to as the position p-value by the MAST algorithm.) c The boxed region represents the region determined to be required for cepI expression as determined in Fig. 1. Bold lettering represents the motif predicted by the MEME program. In the case of cepI the motif matches the CepR binding site [20]. d The number of base pairs to the start codons predicted by alignment with homologous genes. e Underlined bases are conserved in at least 4 of 6 sequences. f The promoter regions of 10 genes experimentally determined to be positively regulated by CepR used to search for common motifs with MEME. g Underlined bases are conserved in at least 7 of 10 sequences.
Expression of promoter::luxCDABE fusionsidentified in the first cep box motif screen identify OHL responsive genes in an approach similar to that described by Weingart et al [20]. Since we had previously determined that genes involved in production of the siderophore ornibactin were cepIR regulated [14], we performed our screen in low iron medium in an attempt to identify other iron regulated genes that were responsive to OHL. We also had previously determined that cepR could both positively and negatively regulate gene expression, and therefore, the transposon library was screened for insertion mutants in which β-galactosidase activity was either turned on or off in the presence of exogenous AHLs. Four unique positively regulated and three negatively regulated lacZ fusions were identified. We identified two genes potentially involved in iron transport, a putative outer membrane receptor (BCAM1187) and phuV, a hemin ATP binding protein (BCAM2630). Interestingly, expression of the outer membrane receptor gene was negatively influenced by OHL, whereas phuV expression was positively influenced.
In a screen of approximately 25,000 transposon mutants we only identified six loci with AHL responsive genes. The screening assay was dependent on the visual identification of colonies that were either blue or white in the presence or absence of AHL on medium with X-gal. Although we were able to detect as little as two-fold differences in expression with this assay, we would not detect gene fusions expressed in both the presence and the absence of AHL since we did not attempt to identify mutants with varying shades of blue. For example, although CepR positively regulates zmpA, the CepIR system is not required for its expression since zmpA is expressed at low levels in the absence of AHL and in cepI or cepR mutants [16]. The lacZ fusions in the positively regulated genes identified with transposon insertions were only expressed at significant levels in the presence of OHL. The three negatively regulated fusions had very low expression in the presence of OHL (Fig. 2). It was surprising that we did not identify cepI since CepR tightly regulates cepI expression [14] and cepI Graphical representation of the cep box consensus sequence Figure 5 Graphical representation of the cep box consensus sequence. Nucleotide sequence logos are derived from the sequences in Table 2. The relative sizes of the letters indicate their likelihood of occurring at a particular position. The upper logo is based on the six sequences used in the first generation consensus search and the lower logo is based on the ten sequences used in the second generation consensus search.   appeared to be a hotspot for transposon insertions in the study by Weingart et al. [20]. The aidA gene which is tightly regulated by cepIR was identified in both transposon screens, as well as the proteomics and promoter trap approaches [17,19].
Lewenza et al [11] identified a putative CepR binding site in the cepI promoter. During the course of this current study it was reported that CepR directly interacted with a cep box that overlapped this region and directly bound to a cep box within the aidA promoter [20]. We demonstrated using site directed mutagenesis of the cep box region that a 24 bp sequence that contained the cep box was required for cepI expression. All cepI::luxCDABE promoter fusions with mutations in the 24 bp cep box had levels of expression less than or equal to 20% of K56-2 (pCPI301). Similar mutations constructed flanking the cep box had either no effect or in one case increased transcription.
The use of bioinformatics to identify CepIR regulated genes has several advantages that are complementary to the experimental methods used to search for CepIR regulated genes. Procedures such as transposon mutagenesis, promoter libraries, microarray analysis or proteomics are dependent on the transcription and expression levels of the genes and on the conditions used in the study. Furthermore, the genes and proteins identified by these approaches may be regulated directly or indirectly by CepR. The use of a motif in a genome-wide search for CepIR regulated genes may identify niche specific genes that may only be expressed in certain conditions. Identification of a cep box motif may also be used to predict whether CepIR genes are directly regulated by interaction with CepR at the promoter or indirectly by CepR interaction with a promoter for an intermediate regulatory gene. In fact, 14 of the 55 putative cep boxes identified were in the predicted promoter regions for regulatory genes. We are currently characterizing some of these regulatory genes to confirm that they are cepR regulated and to determine their regulatory properties.
When searching the genome using the first generation cep box motif we identified some sequences that were not identified with the refined motif used in the second screen of the genome (data not shown). It is possible that these genes are regulated by CepR but have less conserved cep box sequences. Of the eight promoter-lux fusions constructed from sequences identified in the first generation search, six were determined to have cepR regulated expression. There was no difference between the expression of the pMST112 in K56-R2 and K56-2; however, luminescence was increased in K56-dI2 in medium with OHL. The MST112 motif was not detected in the second cep box motif, suggesting that this BCAM1943 may not be cepR Alignment of the putative cep box sequences Figure 6 Alignment of the putative cep box sequences. The MST sequences listed are described in Table 3. Bases conserved in at least 70% of the sequences are shown in red and indicated by an upper case letter in the consensus sequence at the bottom of the alignment, and those conserved in at least 50% of the sequences are shown in blue and indicted by a lower case letter in the consensus sequence. Other bases are indicated in black.
regulated. Mutations in cepI or cepR did not influence the expression of pMST052 (BCAL2739). Although this promoter region was excluded from the group used to generate the second motif, this potential cep box was also detected in the second search (MST2035). It is possible that BCAL2739 is, in fact, CepR regulated in different medium or growth conditions.
Interestingly, the MEME program identified a cep box motif farther upstream of the aidA cepR binding site identified by Weingart et al. [20]. It is possible that there is more than one CepR binding site upstream of aidA. The additional site might contribute to its tight regulation by CepR and dependency on OHL for expression, features that may have resulted in aidA being detected in all of the approaches to date to identify CepR regulated genes.
We identified a cep box in the cepR promoter region that contains all of the most conserved bases. We have previously shown that cepR negatively regulates itself [14]. This is the first confirmed negatively regulated gene identified in the motif search.
It is difficult to compare the extent of overlap between the genes identified using the bioinformatics approach to those identified by Aguilar et al. [17] and Weingart et al.  [20] since the same annotation of the J2315 sequence was not used, although Aguilar et al. identified in addition to aidA, a lysR regulator and a putative short chain dehydrogenase which may be the same ones we identified. Concurrent with this study, we employed a random promoter library approach to identify promoter::lux fusion clones that were differentially expressed in the presence or absence of OHL in K56-dI2 [33]. Of the 86 promoter clones identified, surprisingly only 4 genes overlapped between the two approaches, BCAM0009, BCAM0010, cepI and zmpA. A putative cep box was identified in the promoter regions of 30/89 OHL responsive genes from the promoter library, but generally with only 50-60% identity to the cep box consensus identified in this study. Therefore, these would not have been identified with the stringency employed in the search. It is surprising that more genes that were identified using the cep box motif were not found in the promoter library, although the promoter library also lacks other known CepR regulated genes indicating that it is not complete. Some of the genes with cep boxes may not be expressed in the conditions used to screen the library.
Strains of B. cenocepacia, including K56-2, that contain the cenocepacia island (cci) have a second set of quorum sensing genes [34]. CciI is an AHL synthase that produces predominantly HHL and small amounts of OHL. CciR is the transcriptional regulator. CciIR are co-transcribed and cepR is required for cciIR expression [35]. Little is currently known about the regulatory targets of cciIR, although the zinc metalloproteases zmpA and zmpB are regulated by cciIR, and CciR negatively regulates cepI [35,36]. There is no apparent cep box upstream of cciIR; however, there is one located within the coding sequence 13 bp downstream of the predicted start codon. This putative cep box TTGCTGAAGTTGTTCGGT lacks the conserved A in position 6 present in all the sequences in Table 3 but contains the other conserved bases. It is currently unknown whether CciR binds to a similar site as CepR, but we have determined that some cepR regulated genes are not regulated by cciIR (data not shown). It is possible that some of the cep boxes we have identified might be CciR binding sites. Further studies are in progress to explore the regulatory relationships between these two quorum sensing systems in B. cenocepacia.

Conclusion
We have identified several new CepR regulated genes using transposon mutagenesis and lux promoter fusions.
We have also used a cep box consensus sequence to identify several genes or operons potentially regulated by CepR. To confirm that these genes are regulated by cepR or possibly cciR, experimental approaches such as transcriptional fusions, microarrays, or demonstration of direct binding of CepR to their promoter regions will be required. These studies reveal a significant number of genes that may be further studied to increase our understanding of the CepR regulon.

Reagents, bacterial strains and culture conditions
Unless otherwise stated all molecular biology reagents were purchased from Invitrogen Life Technologies (Burlington, Ontario) and all chemicals purchased from Sigma Chemical Co. (St. Louis, Mo.). The strains and plasmids used in this study are listed in Table 4. For genetic manipulations, B. cenocepacia and Escherichia coli strains were grown at 37°C in Luria-Bertani (LB) broth (Invitrogen) or on 1.5% LB agar plates. Concentration of antibiotics in selective medium for E. coli were 100 μg/ml ampicillin, 1.5 mg/ml trimethoprim and 15 μg/ml tetracycline, and for B. cenocepacia were 200 μg/ml tetracycline and 100 μg/ ml trimethoprim. For transcription assays, B. cenocepacia strains were grown in tryptic soy broth (TSB, Difco, Detroit, Mich.) or TSBD-C [37].

AHL extraction and OHL purification
AHLs were extracted from culture supernatants of K56-2 as previously described [14]. The extract from 50 ml culture fluid was resuspended in 1 ml distilled water and 20 μl aliquots of this stock solution were spread onto agar plates to screen for mutants in which lacZ expression was altered in the presence of AHL. This quantity of AHL extract was found to restore wild-type protease activity to B. cenocepacia K56-I2 as indicated by the zones of clearing observed on skim milk plates. OHL was purified from culture supernatants of B. cenocepacia K56-I2 (pSLS225), a strain that carries the cepI gene in trans as previously described [38].

Molecular biology and sequence analysis
DNA manipulations were performed generally as described by Sambrook et al. [39]. T4 DNA ligase was purchased from Promega Corporation (Madison, WI) and New England Biolabs Inc. (Beverly, MA). Custom oligonucleotides were synthesized by Invitrogen Life Technologies. DNA sequencing was performed at the Univeristy of Calgary Core DNA Services (Calgary, Canada) using an ABI1371A DNA sequencer or at Macrogen Inc. (Seoul, Korea) on an ABI3730 XL automatic DNA sequencer.

Transposon mutagenesis
Mutagenesis of B. cenocepacia K56-I2 (Tp r ) with Tn5-OT182 was performed as described by Lewenza et al. [11]. Tn5-OT182 is a self-cloning transposon with a promoterless lacZ gene that is transcribed from the promoter of a host gene when it is fused in the direction of transcription [40]. Transposon insertion mutants were picked using a robot (Norgren Systems, Palo Alto, CA) into Becton Dickenson microtest flat bottom polystyrene 96 well micro-titer plates containing 200 μl medium per well and grown overnight at 37°C with shaking at 200 rpm. Cultures were stamped onto TSBD-C (200 μg/ml tetracycline, 100 μg/ml trimethoprim and 40 μg/ml X-gal) agar with and without the addition of AHL extract and grown for 48 hours at 37°C. β-galactosidase expression was visually monitored at 24 and 48 hours for differences in blue color. Approximately 25,000 tetracycline and trimethoprim resistant transposon insertion mutants from five independent mutagenesis experiments were screened. Positively regulated insertion mutants appeared blue in the presence of AHL and X-gal and white in the absence of AHL. The reverse is true in the case of negatively regulated genes. Nine mutants exhibiting reproducible differences in AHL dependent β-galactosidase expression were chosen for further characterization. The DNA flanking the Tn5-OT182 insertions was self-cloned from XhoI or EcoRI digests of genomic DNA and sequenced using oligonucleotides OT182-LT and OT182-RT [41].