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 luxCDABE 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 wild-type.

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-luxCDABE 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 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::luxCDABE 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).

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 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).

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 promoters 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 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.

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].

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].

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.

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 microtiter 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].

The Altered Sites^® II in vitro Mutagenesis System (Promega) and the Quick Change^® Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) were used to create mutations spanning the proposed cep box in the cepI promoter (Fig. 2). The template used with the Altered Sites^® II in vitro Mutagenesis System was created by ligating a BamHI-XbaI fragment containing the cepI promoter region from pCPI101 into pALTER-Ex1 (pCPI201). The Altered Sites^® II System was used with mutagenic oligonucleotides CepBx103-106 (Table 5). These oligonucleotides were 5'-phosphorylated using T4 DNA Kinase (Promega) and annealed to single stranded DNA prepared according to the manufacturers instructions from cultures of JM109 F' (pCPI201). The remaining mutagenic oligonucleotides were used with plasmid pCPI101 and the Stratagene Quick Change^® Site-Directed Mutagenesis Kit. Mutagenic oligonucleotides were designed with 4 base pair substitutions that resulted in the introduction of a new restriction enzyme site (Table 5). Mutations were confirmed by restriction enzyme analysis and sequencing. To construct the cepI::luxDCABE fusions, the mutated promoter regions were excised from pCPI101 and pCPI201 by digestion with BamHI-XhoI and ligated into the BamHI-XhoI site of pMS402.

Putative promoters identified in this study were PCR amplified using the primers listed in Table 5 from K56-2 genomic DNA and cloned into the vector PCR2.1^®-TOPO. The promoters were excised from the PCR2.1^®-TOPO clones using BamHI-XhoI and ligated into pMS402 to create plasmids pCPI301, pPHU301, pAYL301, pSCP301 and pAID301, respectively. The eight promoters identified in the first genome search for cep box motifs were cloned using the primers listed in Table 5 for each MST promoter as described above and named pMST005, pMST011, pMST028, pMST052, pMST059, pMST068, pMST072, and pMST112, respectively.

Five ml overnight cultures of K56-2, K56-dI2 and K56-R2 hosting the luxDCABE fusions were grown in TSB supplemented with 100 μg/ml trimethoprim to maintain pMS402. Overnight cultures were diluted with TSB to an A₆₀₀ of 0.05 and aliquots of 150 μl were placed in wells of 96 well clear bottom plates (Costar, Corning Incorporated, Corning, NY). The plates were covered and incubated at 37°C with shaking and the luminescence and absorbance was measured in a Victor²™ multilabel counter at various intervals for 24 hours. Each strain was assayed at least three times in triplicate.

Nucleotide sequence obtained from DNA flanking the transposon insertions was used with BLASTN to determine the location of the insertion in the unpublished genome sequence of B. cenocepacia J2315 [42], a strain of the same lineage as K56-2. Homologues of open reading frames were predicted using BLASTP [43]. Potential promoter elements were identified using BPROM [44]. The cep box consensus sequence was predicted by analyzing the promoter regions of selected positively regulated genes with the motif discovery tool MEME [45]. The MEME program [46] represents motifs as position-dependent letter-probability matrices which describe the probability of each possible letter at each position in the pattern. The output from the MEME program provides a position-specific scoring matrix (PSSM) for the predicted motif. The PSSM for the predicted cep box consensus sequence was used to search the B. cenocepacia J2315 genome with the motif alignment search tool MAST [45,47]. The cep box motifs identified by MAST were also aligned using Multalin [48,49].