BphS, a key transcriptional regulator of bph genes involved in polychlorinated biphenyl/biphenyl degradation in Pseudomonas sp. KKS102.

The bph genes in Pseudomonas sp. KKS102, which are involved in the degradation of polychlorinated biphenyl/biphenyl, are induced in the presence of biphenyl. In this study our goal was to understand the regulatory mechanisms involved in the inducible expression. The bph genes (bphEGF(orf4)A1A2A3BCD(orf1)A4R) constitute an operon, and its expression is strongly dependent on the pE promoter located upstream of the bphE gene. A bphS gene, whose deduced amino acid sequence showed homology with the GntR family transcriptional repressors, was identified at the upstream region of the bphE gene. Disruption of the bphS gene resulted in constitutive expression of bph genes, suggesting that the bphS gene product negatively regulated the pE promoter. The gel retardation and DNase footprinting analyses demonstrated specific binding of BphS to the pE promoter region and identified four BphS binding sites that were located within and immediately downstream of the -10 box of the pE promoter. The four binding sites were functional in repression because their respective elimination resulted in derepression of the pE promoter. The binding of BphS was abolished in the presence of 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid, an intermediate compound in the biphenyl degradation pathway. We concluded that the negative regulator BphS plays a central role in the regulation of bph gene expression through its action at the pE promoter.

In many microorganisms capable of degrading chemical compounds, the transcription of genes involved in the degradation is regulated (8). In most cases, the genes coding for the regulator exist near the structural genes, and their protein products activate the transcription in the presence of their cognate inducer molecule. Repressor-mediated regulation is rare for genes involved in the catabolism of aromatic compounds. However, Mouz et al. (9) reported that the expression of bph genes on transposon Tn4371 was repressed by the product of bphS gene, although the molecular events in the repression and derepression have fully remained to be elucidated.
The bph genes and their organization in KKS102 are highly homologous to the bph genes on transposon Tn4371 (10,11). These two bph gene clusters share 94% identity at the nucleotide level, but DNA sequences in the upstream region of bphE are different from each other (10).
The bph genes in KKS102 are induced when grown in the presence of biphenyl. This induction requires an inducer molecule, 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid (HOPDA), an intermediate metabolite of the biphenyl degradation pathway (12). In this study, we identified a negative regulator of bph genes and revealed its function in the regulation of bph gene expression in KKS102. This report describes for the first time the detailed regulatory mechanism of PCB/biphenyl degradation genes.
Construction of the bphS Disruptant-For the construction of a plasmid for bphS disruption, plasmid pKH1004 carrying a 1.2-kb HincII fragment in the multicloning site of pUC19 was digested by XhoI, and a chloramphenicol resistance gene derived from pHSG399 was inserted into the cleaved site. The resulting plasmid was linearized by BamHI and HindIII digestion and used for electroporation. The gene disruption was confirmed by Southern blot analysis. The Southern blot analysis was performed with an ECL gene detection system (Amersham Pharmacia Biotech) according to the manufacturer's protocol.
Construction of Strains for LacZ Reporter Assay-All of the fusion constructs of the modified upstream region of bphE and lacZ were integrated into the genome of KKS102. For systematic construction of plasmids for integration, plasmid pKLZ-A was constructed. pKLZ-A contains the following DNA fragments: kanamycin resistance gene derived from Tn5 as a marker for integration, a synthetic terminator sequence to prevent read-through of transcription, a multicloning site comprised of EcoRI, SmaI, and BamHI sites, and lacZ gene derived from pMC1403 (13), and these are inserted into randomly selected DNA fragments from the KKS102 genome. The modified promoter-lacZ fusions were constructed by inserting either duplex oligonucleotides or polymerase chain reaction-amplified DNA fragments into the multicloning site of pKLZ-A or of its derivative plasmid. The DNA sequences inserted are presented in Fig. 5.
For integration into the chromosome of KKS strains, resulting plasmids were digested by HindIII within the vector sequence and introduced into KKS102 by electroporation. Integration of promoter-reporter constructs into the genome results in the disruption of a single open reading frame (ORF) that encodes a putative member of the NtrC family regulator. This disruption had no effect on the expression level of bph genes under any conditions tested.
Construction of the pE Promoter-deleted Mutant-For the deletion of the chromosomal pE promoter, plasmid pKH966A was constructed. The plasmid pKH966A carries the following DNA fragments in the cloning site of pHSG399 (see Fig. 8): a bphE upstream region (Ϫ400 to Ϫ1284 where ϩ1 is the start codon for bphE), a DNA fragment derived from pKLZ-A containing the kanamycin resistance gene and a terminator sequence, and a bphE upstream region (Ϫ242 to ϩ3 where ϩ1 is the start codon for bphE). The plasmid pKH966A was digested by HindIII and BamHI in a vector sequence and used for electroporation.
Electroporation-Each of the plasmids was linearized by an appropriate restriction enzyme, extracted with phenol/chloroform, ethanolprecipitated, dissolved in water, and introduced into KKS102 by electroporation. The cells from liquid culture were washed five times with chilled sterile water. The Gene Pulser (Bio-Rad) was used under the following conditions: 0.1-cm cuvette, 200 ohms, 25 microfarads, 1.8 kV, and a pulse time of 4.7-5.0 ms. A 1-ml aliquot of SOC medium (2% tryptone, 0.5% yeast extract, 0.05% NaCl, 10 mM MgCl 2 , and 20 mM glucose, pH 7.0) was added immediately after the electric pulse. The cells were incubated at 30°C for 3 h prior to plating onto 1/3 L broth containing the appropriate antibiotics.
Northern Blot Analysis-The total RNA was isolated by the method described by Hopwood et al. (14). Hybridization and detection were performed by using digoxigenin-labeled DNA with a CSPD system (Roche Molecular Biochemicals) according to the manufacturer's protocol. The 1.0-kb HincII-ApaI fragment, 1.2-kb SphI-SmaI fragment, and 0.9-kb EcoT22I-KpnI fragment were used to generate bphA1, bphC, and bphE probes, respectively (see Fig. 1B).
Assay for BphD Activity-To measure the BphD activity, the cells were washed once in sample buffer (50 mM potassium phosphate buffer (pH 8.0) containing 10% glycerol) and resuspended in 1 ml of the same buffer. After sonication and centrifugation at 15,000 rpm for 10 min, the supernatant (crude extract) was assayed for BphD activity.
For the assay of BphD activity, BphD substrate (HOPDA) was diluted to an A 434 of 0.5 to 1.5 in the sample buffer. HOPDA was prepared as described (12). The crude extract was added to HOPDA solution prewarmed at 30°C. The decrease in A 434 was measured for 3 min, and the portion in which A 434 decreased in proportion to time was used to calculate the BphD activity. We defined 1 unit of BphD activity as the activity necessary to convert 1 nmol of 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid/min. A molar extinction coefficient of 19,800 for 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid was used to calculate the BphD activity (15). The amount of protein in the crude extract was quantified using a protein assay kit (Bio-Rad).
Assay for LacZ Activity-For LacZ activity measurement, crude extract was prepared as described above except that Z buffer (60 mM Na 2 HPO 4 , 40 mM NaH 2 PO 4 , 10 mM KCl, 1 mM MgSO 4 , 50 mM ␤-mercaptoethanol) was used. The crude extract was added to the O-nitrophenyl ␤-D-galactopyranoside solution (4 mg/ml in Z buffer). After 10 min of incubation at 30°C, stop solution (1 M Na 2 CO 3 ) was added to terminate the reaction and A 420 was measured. We defined 1 unit of LacZ activity as the activity necessary to produce 1 nmol of O-nitrophenyl/min. Primer Extension-The total RNA was isolated by the method described by Hopwood et al. (14) from KKS102 cells after 3 h of incubation with biphenyl. The primer extension was performed with oligonucleotides BPHE10-29 (5Ј-CTGGTCGAAACCGTATCTGG-3Ј) (hybridizing to nucleotides 10 -29 of the bphE coding sequence). The primer was annealed to ϳ20 g of the isolated RNA. Primer extension reactions with avian myeloblastosis virus reverse transcriptase (Promega) were performed at 37°C, and the extended product was run beside the DNA sequencing ladder generated by the dideoxy chain termination method using the same primer. Samples were run with a LI-COR model 4000L DNA sequencing system (LI-COR, Lincoln, NE).
Nucleotide Sequence Determination-The nucleotide sequence was determined by the dideoxy chain termination method with the Applied Biosystems model 310 DNA sequencing system (Applied Biosystems, Foster City, CA).
In the competition assay, 40 ng of cold target DNA or 1 g of unrelated DNA (sermon sperm DNA) was added. After incubation for 10 min at room temperature, the mixtures were separated by electrophoresis at 60 V constant voltage in 5% polyacrylamide gels buffered with 1ϫ Tris borate-EDTA buffer.
For preparation of CFE, E. coli harboring plasmid pKH701, which carries the bphS gene under the control of the lac promoter, was cultivated in L broth. When the turbidity reached a value of 0.5, isopropylthio-␤-D-galactoside was added at a final concentration of 1 mM. After 5 h of additional incubation, cells were collected, washed once with 50 mM potassium phosphate buffer (pH 7.4), resuspended in the same buffer, and disrupted by sonication. After centrifugation of the sonicated cell suspension, the supernatant was used as CFE. The protein concentration in the CFE was determined by using the Bio-Rad protein assay kit. Bovine serum albumin was used as a standard.
DNase I Footprinting-A 183-bp fragment used for gel retardation with both ends labeled with 32 P was used after the following procedures were performed. The fragment was digested by either BamHI or XhoI. In either digestion, 177-and 6-bp fragments were generated. The longer fragments, digested by BamHI and XhoI had 32 P in the coding and noncoding strand, respectively. Two rounds of ethanol precipitation removed the shorter fragment. The single end-labeled fragments were incubated with CFE prepared from E. coli cells expressing BphS protein in the same buffer conditions as were present for the gel retardation assay. The total volume of the reaction mixture was 40 l, and 0 -20 g of protein in CFE was used. After 10 min of incubation at room temperature, DNase I solution (diluted in 10 mM MgCl 2 , 5 mM CaCl 2 ) was added and incubated at 30°C for 2 min. The reaction was stopped by the addition of 40 l of phenol followed by vortexing and addition of 100 l of stop solution. Protected bands were identified by comparison with the migration of the same fragment treated for A ϩ G sequencing reactions by the method of Maxam and Gilbert (16).

Nucleotide Sequence of the Upstream Region of bphE-
The upstream region of bphE, the first gene of the bph gene cluster, was sequenced for about 3 kb. Two ORFs were found in this region in an orientation opposite that of the bph genes cluster (Fig. 1B). The start codon of the ORF proximal to bphE was 511 bp distant from that of bphE. Their nucleotide and deduced amino acid sequences are shown in Fig. 2. The deduced amino acid sequence of the ORF proximal to bphE showed homology to transposases of several transposons. These are IS1405 from Ralstonia solanacearum (DDBJ accession number AF167984), IS1384 from Pseudomonas putida (DDBJ accession number AF052751), and ISPSMC from Pseudomonas syringae (DDBJ accession number AB023075). Identities between this ORF and their transposases are 74, 64, and 62%, respectively. Typical terminal 4-bp direct repeats and 16-bp inverted repeats flank this ORF. Because these are common features of an insertion sequence, we designated this region and the transposase ORF as ISBPH and tnpBPH, respectively. Southern blot analysis using a 1374-bp SphI fragment (Fig. 1B) as a probe revealed that this insertion sequence existed as a single copy in the genome of KKS102 (data not shown).
The deduced amino acid sequence of another ORF distal from bphE showed homology to transcriptional repressors of the GntR family (17), BphS of transposon Tn4371 (9) and AphS of Comamonas testosteroni (18). This ORF had 74.4 and 37.0% identities to these repressors, respectively. The sequence also showed 40.2% identity to ORF0 from Pseudomonas pseudoalcaligenes KF707, which also belongs to the GntR family but exceptionally works as a positive regulator (19). Because the product of this ORF functioned as a regulator of the bph genes (see below), we designated it as bphS.
Repressive Function of the bphS Gene Product in KKS102-The bphS gene of KKS102 was disrupted as indicated in Fig.  3A (for details see "Experimental Procedures"), and we analyzed the effects of disruption on the expression of the bph genes. The strains were grown in liquid culture with or without biphenyl, and their BphD activities were measured at 1, 3, 5, and 7 h after the addition of biphenyl. In the wild type KKS102, BphD activity was kept at a low level in the absence of biphenyl, whereas it gradually increased in the presence of biphenyl and reached a level of activity 5-fold that of the original level after 5 h. In the bphS disruptant (KKS⌬S), a high level of BphD activity was detected even in the absence of biphenyl (Fig. 3B).
We also measured the bph mRNA level in the bphS disruptant by Northern blot analysis. Strains were incubated with or without biphenyl before mRNA preparation. mRNAs were blotted onto nitrocellulose membranes and were hybridized with bphA1-, bphC-, or bphE-specific probe. Signals on the membranes were quantitatively analyzed, and the amounts of bph mRNAs that were detected with the above probes were compared with those of uninduced wild type strain (Fig. 3C). In the wild type strain KKS102, the bph mRNA level was high in the presence of biphenyl but not in its absence. In contrast, a high level of bph mRNA was detected in the bphS disruptant even in the absence of biphenyl.
To rule out the possible polar effect of the disruption of bphS with a chloramphenicol resistance gene on the expression of bph genes, a fusion construct of the bphE upstream region and lacZ as a reporter gene (Fig. 3D) was integrated into a randomly selected site on the KKS102 genome (see "Experimental Procedures"), and the effect of bphS disruption was analyzed. In the bphSϩ strain (KLZ12), LacZ activity was about 3 times higher in the presence of biphenyl than in its absence. In contrast, we detected a high level of LacZ activity in the bphS disruptant (KLZ12⌬S) in the absence of biphenyl, and this level was even higher than that in the presence of biphenyl. These results clearly demonstrate that the bphS gene product negatively regulates bph gene expression in KKS102.
Characterization of a Promoter Located Upstream of bphE-Because the bphE gene, the first gene in the bph gene cluster, is induced by biphenyl, we searched for a promoter in the upstream region of bphE. To identify the transcription start site in vivo, primer extension analysis was performed. When the primer BPHE10-29, which hybridized to the nucleotides 10 -29 of the BphE-coding sequence, was used, a single band corresponding to a T residue located at 317 nucleotides upstream from the bphE start codon was detected (Fig. 4). The transcription start site was preceded by a conserved E. coli Ϫ10 box (TATAAT). The Ϫ35 box was not highly consistent with that of E. coli (GTGTTT versus TTGACA of E. coli). The location of these elements was in good agreement with the results from the LacZ reporter assay (see below). No other transcriptional start site was identified.
Hereafter we refer to the DNA region that includes the Ϫ10 and Ϫ35 boxes (Ϫ326 to Ϫ354 where ϩ1 is the translation start point for bphE) as the pE core promoter and the DNA region that includes the pE core promoter and the other elements involved in the transcriptional regulation as the pE promoter.
Identification of the DNA Region Required for Inducible Expression of the pE Promoter-We performed a deletion analysis of the bphE upstream region to define the sequence necessary for inducible promoter activity. A series of fusion constructs of the partially deleted 5Ј region of the bphE and lacZ gene were integrated into the genome of KKS102 (see "Experimental Procedures" for details), and LacZ activity was measured in the presence or absence of biphenyl. By integrating the lacZ fusion constructs into the genome, the effect of difference in copy number of the reporter gene could be excluded from the measurement of the promoter activity. The results are summarized in Fig. 5A. First, pE is the sole promoter that resides within 1555 bp upstream of the bphE start codon. The same level of LacZ activity was observed in strain KLZ10 (containing up to 1555 bp from the bphE translation start site) and in strain KLZ12 (containing up to 387 bp from the bphE translation start site), indicating that there is no promoter activity be- 1. A, the PCB/biphenyl degradation pathway in Pseudomonas sp. KKS102. Biphenyl is converted to dihydrodiol compound by BphA activity (1). The dihydrodiol compound is converted to 2,3-dihydroxybiphenyl (1) by BphB activity. 2,3-Dihydroxybiphenyl is converted to HOPDA by meta cleavage activity exerted by BphC (4). BphD is a hydrolase and converts HOPDA to two molecules, benzoic acid and 2-hydroxypenta-2,4-dienoate (4). Further, 2-hydroxypenta-2,4-dienoate is converted to acetyl-CoA and pyruvate by 2-hydroxypenta-2,4-dienoate hydratase (BphE), 4-hydroxy-2-oxovalerate aldolase (BphF), and acetaldehyde dehydrogenase (BphG) (5). B, the bph gene cluster in KKS102. The two ORFs found in this study are also shown. Lines under the gene cluster represent the DNA fragments used to generate probes for Northern and Southern blot analyses. tween nucleotides Ϫ388 and Ϫ1555. The deletion of the pE core promoter resulted in a markedly low level of LacZ activity (compare strains KLZ12 and KLZ9). Deletion of the sequence just upstream of the Ϫ35 box resulted in a decrease in LacZ activity (see strains KLZ14 and KLZ8). This decrease may be due to deletion of the UP element, the third DNA element of the prokaryotic core promoter (20). Second, the DNA region from Ϫ387 to Ϫ243 (where ϩ1 is the bphE translation start site) is sufficient and necessary for promoter activity and the inducible expression because strain KLZ23 (which contains Ϫ387 to Ϫ243) showed enhanced LacZ activity that was much higher in the presence of biphenyl.
Specific Binding of BphS to the pE Promoter-The results described above suggested that the bphS gene product binds within the DNA region spanning from nucleotides Ϫ387 to Ϫ243. To determine whether the BphS binds to this DNA region, we conducted gel retardation experiments. The DNA region from Ϫ387 to Ϫ243 of the pE promoter was excised as an EcoRI-HindIII fragment from pYO12R and was end-labeled with 32 P (Fig. 6A). The BphS protein was expressed in E. coli, and the CFE was used for the gel retardation assay. The retarded bands were observed only in the reactions containing BphS protein (Fig. 6B). There were two shifted bands at higher protein concentrations, suggesting that BphS binds to multiple sites within the fragment. When the nonradioactive DNA fragment was added to the binding reaction in excess of the 32 Plabeled fragment, the retarded band was greatly reduced (Fig.  6B, lane 6). The addition of an unrelated DNA fragment did not affect the BphS binding (Fig. 6B, lane 7). The retarded band was not observed when CFE of E. coli harboring vector pHSG399 was used (Fig. 6B, lane 8). These results clearly indicate that BphS specifically binds to the DNA region spanning Ϫ387 to Ϫ243. Inhibition of Binding of BphS to the pE Promoter by HOPDA-In our previous work, we demonstrated that HOPDA, the intermediate metabolite of the biphenyl degradation pathway, is the inducer molecule of bph genes in KKS102 (12). Here we performed a gel retardation assay in the presence of varying concentrations of HOPDA (0 -0.5 mM). The amount of protein of CFE was fixed at 3 g. Under this condition only one retarded band was observed (Fig. 6B, lane 3). As shown in Fig. 6C, the intensity of the retarded band was reduced in a concentration-dependent manner. In the presence of HOPDA at 0.5 mM, the retarded band disappeared almost completely (Fig. 6C, lane 6). In contrast, a saturated concentration of biphenyl (ϳ0.1 mM) or 0.5 mM 2,3-dihydroxybiphenyl did not inhibit the binding. These results indicate that the BphS protein loses its ability to bind to the pE promoter in the presence of the inducer molecule HOPDA. This inhibition of binding of BphS to the pE promoter by HOPDA is consistent with the in vivo function of HOPDA as an inducer.
DNase I Footprinting Analysis-To obtain further information on the binding site of the BphS protein, DNase I footprinting analysis was carried out. A 183-bp pE promoter fragment containing the region from Ϫ387 to Ϫ243 was analyzed with both coding and noncoding strands (Fig. 7). The results of the DNase I footprinting are summarized in Fig. 7B. We identified four BphS binding sites and named them (beginning with the furthest upstream) BS I (binding site I), BS II, BS III, and BS IV. When DNA labeled in the coding strand was incubated with a relatively low amount of CFE containing BphS, two regions ranging from nucleotides Ϫ333 to Ϫ319 (BS I) and Ϫ315 to Ϫ299 (BS II) were protected from DNase I digestion (Fig. 7A,  lane 3). When DNA labeled in the noncoding strand was used under the same protein and DNA concentrations, DNA regions ranging from Ϫ329 to Ϫ315 (BS I) and Ϫ312 to Ϫ295 (BS II) were protected (Fig. 7A, lane 8). At higher protein concentrations, additional DNA regions from Ϫ296 to Ϫ281 (BS III) and Ϫ279 to Ϫ263 (BS IV) (Fig. 7A, lane 5, analyzing coding strand) and from Ϫ291 to Ϫ278 (BS III) and Ϫ274 to Ϫ260 (BS IV) (Fig.  7A, lane 10, analyzing noncoding strand) were protected. In Fig. 7, the DNA regions protected at low and high protein concentrations are indicated by thick and thin lines, respectively. No protection was observed when CFE of E. coli harboring vector plasmid was used (data not shown). These results demonstrate that the BphS protein binds to four sites near the pE core promoter and has greater affinity to the two upstream binding sites (BS I and BS II) than to the two downstream sites (BS III and BS IV). That a part of the Ϫ10 box hexamer of the To investigate the function of these two sets of BphS binding sites in the repression of the pE promoter in vivo, a series of promoter constructs was integrated into the genome of KKS102 as described under "Identification of the DNA Region Required for Inducible Expression of the pE Promoter" and assayed for promoter activity (Fig. 5B). The strain KLZ23, which had a construct with all four binding sites as well as the pE core promoter, showed low LacZ activity when grown in the absence of biphenyl. In contrast, the strain KLZ22, the integrated construct of which had BS I and BS II but lacked BS III and BS IV, showed high LacZ activity even in the absence of biphenyl. The LacZ activity in the absence of biphenyl was increased ϳ3-fold,  KKS102 and the bphS disruptant were incubated in 100 ml of 1/3 L broth. When the turbidity at OD 660 reached a value of 0.5, the culture was divided into halves and further incubated in the presence or absence of biphenyl. After 3 h of additional incubation, cells were harvested, and the total RNAs were prepared. RNA samples were slot-blotted onto the nitrocellulose membrane and probed with bphA1, bphC, and bphE. The results of Northern blot analysis were quantitatively analyzed. The relative amounts of the mRNA are shown. D, LacZ activity in the strain harboring the chromosomally integrated bphE promoter-lacZ transcriptional fusion. The bphE upstream region up to position Ϫ387 (ϩ1 represents the bphE start codon) was fused with the lacZ gene and integrated into the genome of KKS102 (strain KLZ12). KLZ12 and its bphS disruptant derivative, KLZ12⌬S, were incubated in 1/3 L broth, and when turbidity at OD 660 reached a value of 0.3, the culture was divided into halves and further incubated in the presence or absence of biphenyl. After 6 h, the cells were harvested for LacZ activity measurement.
although it did not exceed that in the presence of biphenyl, by deletion of BS III and BS IV (compare strains KLZ23 and KLZ22 in the absence of biphenyl), demonstrating the in vivo function of BS III and BS IV. In strain KLZ21, where the integrated construct had no BphS binding site, LacZ activity in the absence of biphenyl was further elevated to the level observed in the strain KLZ15, whose integrated construct lacked BSI and BS II but had BS III and BS IV. This result indicates that the presence of BS I and BS II is essential for repression of the pE promoter.
In the strains KLZ12⌬S, KLZ21, and KLZ15, the presence of biphenyl resulted in lower lacZ activities than its absence. This might have been due to the cytotoxicity of biphenyl (21) and/or the catabolite-repressive effect on the pE promoter as a result of biphenyl assimilation. In conclusion, BS I and BS II were found to play an essential role in repression in vivo, and another set of two binding sites was shown to be functional, although its role was auxiliary.
Role of the pE Promoter in the Expression of Entire bph Genes-The results presented above demonstrate the role of BphS in the regulation of the pE promoter. Under "Repressive Function of the bphS Gene Product in KKS102," we have demonstrated that disruption of the bphS gene resulted in constitutive expression of the bphA1, bphC, and bphE genes (Fig. 3). This suggests that BphS plays a central role in the regulation of entire bph genes and raises the following questions. Does the pE promoter drive the transcription of the entire bph gene cluster? Are there any other BphS-regulated promoters in the bph gene cluster? To address these questions, the pE promoter region from Ϫ242 to Ϫ400 in the strain KKS102 was replaced with a kanamycin resistance gene and a transcription terminator as depicted in Fig. 8. Since the absence of the pE promoter could hinder the accumulation of the inducer molecules and could lead to persistent repression of any promoters in the bph genes cluster by BphS, the effect of the deletion of the pE promoter was also tested in a strain with a bphS-disrupted background. The resulting KKS102 and KKS⌬S derivatives were designated as KKS⌬pE and KKS⌬pE⌬S, respectively. We then investigated the induction of BphD activity (Table I). The bphD gene is located relatively downstream in the bph genes cluster, and therefore BphD activity serves as an indicator of the presence of any intervening promoters. The BphD activities in KKS⌬pE were very low irrespective of the presence or absence of biphenyl, even lower than that of the uninduced wild type strain. In addition, the bph mRNA level detected by the bphA1-, bphA4-, bphC-, or bphDspecific probe was low, and no induction by biphenyl was observed. We thus concluded that pE is the primary promoter for the transcription of most bph genes.

The bphS Gene Product Is a Negative Regulator of bph
Genes-In the bphS gene disruptant, expression of BphD and LacZ reporter activity and amounts of bph gene transcripts were elevated even in the absence of biphenyl to the level of those in the induced wild type strain, indicating that the bphS gene product plays an essential role in repression of the bph genes. This repression results from the direct action of the bphS gene product at the pE promoter because the BphS protein specifically bound to the pE promoter and deletion of BphS binding sites led to constitutive production of LacZ activity. The repressor function of BphS is consistent with the fact that BphS belongs to the GntR family of transcriptional repressors.
pE Promoter, an Essential Promoter for Transcription of bph Genes-The LacZ reporter assay of the upstream region of bphE revealed that only one promoter, which was designated the pE promoter, exists in the region. Elimination of the pE promoter resulted in weak and constitutive production of BphD activity as well as of bphA1, bphA4, bphC, and bphD transcripts, demonstrating that the pE promoter plays an essential role in the expression of bph genes and that bph genes (at least from bphE to bphA4) constitute an operon. Although some other parts of the nucleotide sequence may exert promoter activity in the long (12-kb) bph gene cluster, it seems that such promoter activities are trivial compared with the promoter activity of the pE promoter. For example, we detected promoter activity upstream of the bphA1 gene, but fusion with the lacZ gene showed that the activity was significantly lower (20 times lower) than that observed for the pE promoter (data not shown). In conclusion, the pE promoter is the primary promoter driving transcription of the bph operon.
BphS Binding to the pE Promoter-In vivo and in vitro experiments showed specific binding of BphS protein to the pE promoter, indicating that the BphS protein plays an essential role in regulation of the bph operon because the pE promoter is the primary promoter involved in expression of the bph operon.
To identify additional binding sites for the BphS protein that might be involved in repression, we performed a gel retardation assay. We used various DNA fragments derived from the bphE upstream region as well as the bphE-coding region (from Ϫ2262 to ϩ396 where ϩ1 represents the bphE translation start codon), but we did not identify any additional BphS binding sites (data not shown).
Consensus Operator Sequence of BphS Protein-DNase foot-printing analysis identified four binding sites for BphS just downstream of the pE promoter. The binding sites were named, beginning with the most upstream site, BS I, BS II, BS III, and BS IV (Fig. 7B). The gel retardation assay revealed that the affinity of BphS to these sites is greater in the order of BS II Ͼ BS I ϾBS IV ϾBS III. 2 We also tested an inverted repeat sequence found in the vicinity of the promoter of the bph genes on Tn4371 (9) (Fig. 9) and found that BphS of KKS102 bound to this sequence at affinity greater than to BS II. 2 The operator sequences for GntR family members have been suggested to contain perfect or imperfect inverted repeat sequences (22). BS I contains an imperfect inverted repeat sequence and BS II contains a perfect inverted repeat sequence, while BS III and BS IV have no distinct inverted repeat sequence. Alignment of the three stronger binding sites with distinct inverted repeat sequences (BS I, BS II, and the inverted repeat sequence found in Tn4371) identified the consensus sequence of AN 12 T (Fig. 9) in an AT-rich symmetric sequence, which was reminiscent of the binding motif (TN 11 A) for LysR-type transcriptional regulators (23). Conservation of A and T nucleotides separated by 12 base pairs seems important based on the fact that, in several DNA binding proteins that use a helix-turn-helix motif for binding, two recognition helices in the dimers are separated about one turn of DNA helix long (24). In the DNA sequences of BS III and BS IV, the AN 12 T motifs were found, although these binding sites were less symmetric, which might be reflected in the comparably weak binding affinity for BphS (Fig. 7B). Further investigations will be needed to clarify the importance of the AN 12 T motif of BphS binding sites. The AT content around the binding site was 78.4% (in the 74 nucleotides from Ϫ333 to Ϫ260), and this value was very high for the bph genes in which the average GC content was 62%. The abundance of AT base pairs may have helped to form a structure favored by BphS binding.
Repression Mechanism Mediated by Four BphS Binding Sites and BphS Protein-The data obtained from the promoter-lacZ fusion indicates that the proximal two binding sites (BS I and BS II), as well as the distal sites (BS III and BS IV), are functional in vivo for repression and that BS I and BS II play a primary role in repression. How, then, do these four binding sites and the BphS protein mediate transcriptional repression? Protection from DNase I cleavage of part of the Ϫ10 box by BphS indicates that the BphS protein bound at BS I masks the Ϫ10 box of the pE promoter and prevents the access of RNA polymerase as has been described in many other cases, as for example, the binding of the phage cI repressor to the operators O R1 and O R2 (25) and also the binding of LacI to the O 1 operator of the lac promoter (26). The role of BphS proteins bound at BS II may be to enhance repression through stabilization of BphS binding to BS I through protein-protein interaction. BphS bound to BS I may also stabilize binding of BphS to BS II, and thus the binding of BphS repressors to these sites could be cooperative. In support of this possibility, only two species of retarded band were observed in the gel retardation assay; one might represent binding to BS I and BS II and the other might represent additional binding to BS III and BS IV, suggesting preferable binding of BphS protein to each set of operator sites. In addition, the finding of simultaneous protection of BS I and BS II and of BS III and BS IV in the DNase footprinting analysis was consistent with the notion that the BphS protein has a cooperative binding property. In our recent study, purified His 6 -tagged BphS protein was shown to bind to a DNA fragment containing both BS I and BS II with 10 times 2 Y. Ohtsubo, M. Delawary, K. Kimbara, M. Takagi, A. Ohta, and Y. Nagata, unpublished data. more efficiency than to a DNA fragment containing only BS II (as mentioned under "Consensus Operator Sequence of BphS Protein," the affinity for the BphS protein is stronger than that of BS I), supporting the cooperative binding of BphS to BS I and BS II. 2 In a recent review on bacterial transcriptional regulation, possible cooperative interaction of GntR dimer pairs was suggested because the candidate binding sites for GntR occurred in pairs in several cases (27). A GntR family protein, AphS from C. testosteroni TA441, binds to two sites in the promoter region, although the implications of the presence of these two binding sites are not known (18).
It has been reported that multiple binding sites for other types of negative regulators are necessary for efficient repres-sion. For example two GalR dimers bound to two operators have been shown to interact with each other and repress transcription of the gal operon (28). And LacI binds to three operators and cooperates in repression (29).
Derepression of BphS-mediated Repression by HOPDA-We demonstrated in vitro that BphS protein binds specifically to the pE promoter and that the binding affinity decreases in the presence of HOPDA. This feature of the BphS protein enables high levels of the expression of bph genes only when the intermediate of the biphenyl degradation pathway is present. This finding is well consistent with the previous finding that  HOPDA is the inducer of bph genes in vivo (12). BphS Protein, a Key Regulator of bph Gene Transcription-The question of the mechanisms by which gene expression occurs only under particular circumstances is of great interest. A model depicting the molecular aspect of bph gene regulation is as follows. In KKS102 cells, the BphS protein binds to its binding sites and inhibits transcription from the pE promoter. In this repressed state, binding of BphS to BS I plays a central role, and BphS protein bound to the other sites helps to stabilize the BphS protein bound at BS I. When the cells encounter biphenyl, biphenyl is converted to HOPDA by bph gene products that are somehow maintained at the basal level, leading to dissociation of the BphS protein from the operator DNA and to subsequent active transcription initiation at the pE promoter. The increase in bph gene products results in further elevation of the HOPDA concentration. In conclusion, the BphS protein is a key component of the molecular switch regulating expression of the bph operon in KKS102. It regulates the pE promoter that is essential for expression of the bph operon.