Transcriptional Activation of Quinoline Degradation Operons of Pseudomonas putida 86 by the AraC/XylS-Type Regulator OxoS and Cross-Regulation of the PqorM Promoter by XylS

Abstract

The quinoline-degradative gene cluster (oxoO, open reading frames 1 to 6 [ORF1 to -6], qorMSL, ORF7 to -9, oxoR) of Pseudomonas putida 86 consists of several overlapping operons controlled in response to quinoline by the master promoter PoxoO and internal promoters Porf3, PqorM, and PoxoR. ORF7 to -9, presumed to be important for maturation of the molybdenum hydroxylase quinoline 2-oxidoreductase, are also weakly transcribed independently of quinoline. Expression of the oxoS gene, located upstream of oxoO, is not influenced by the carbon source. OxoS shows 26% amino acid sequence identity to XylS, the transcriptional regulator of the meta pathway promoter Pm of TOL plasmid pWW0, and is required for quinoline-dependent transcription from PoxoO, Porf3, PqorM, and PoxoR. 5′ deletion analysis of PoxoO and PqorM suggested that a 5′-TGCPuCT-N₃-GGGATA-3′ motif, which resembles the distal 5′-TGCA-N₆-GGNTA-3′ half-site of the tandem XylS binding site, is essential for oxoS-dependent transcriptional activation. PqorM, which shows similarity to the tandem XylS recognition site of Pm, was cross-activated by the xylS gene product in response to benzoate. The distal half-site of PqorM is necessary, but probably not sufficient, for transcriptional activation by XylS. Despite conservation in PoxoO of a distal 5′-TGCA-N₆-GGNTA-3′ sequence, cross-activation of PoxoO by XylS and benzoate was not observed. The oxoS gene product in the presence of quinoline weakly stimulated transcription from the Pm promoter. Involvement of an XylS-type protein in the regulation of genes encoding synthesis of a molybdenum hydroxylase is without precedent and may reflect the evolutionary origin of this pathway in the metabolism of aromatic compounds.

Many N-heteroaromatic compounds are susceptible to microbial degradation (see reference 12 and references therein). Gene clusters, coding for catabolic enzymes and accessory proteins, were described for the degradation of purine (43), nicotine (19), s-triazines (29), and carbazole (49); for the “upper part” of the anthranilate pathway of quinaldine degradation (36); and for genes involved in quinoline degradation (3). Whereas the regulation of operons encoding the degradation of aromatic compounds like toluene and benzoate is well understood (reference 38 and references therein), comparatively little information is available on the regulation of the catabolism of N-heterocyclic compounds. In Bacillus subtilis, genes involved in purine degradation are regulated by the pathway-specific PucR protein; PucR and its homologues were proposed to constitute a separate family of transcriptional regulators (43). AtzR, a LysR-type regulator, is the main activator of the cyanuric acid degradation operon of Pseudomonas sp. strain ADP, which encodes hydrolytic enzymes involved in cleavage and mineralization of the s-triazine ring (16). In the degradation of carbazole by Pseudomonas resinovorans strain CA10 (pCAR1), the antABC operon coding for anthranilate 1,2-dioxygenase, as well as the car operon encoding conversion of carbazole to anthranilate, is regulated by the AraC/XylS-type transcriptional activator AntR in response to anthranilate; induction of car gene expression by anthranilate and AntR is presumed to be a consequence of transposition of a copy of Pant to a location upstream of the car gene cluster (49).

Pseudomonas putida 86 has the ability to grow on quinoline as a sole source of carbon and energy. Quinoline 2-oxidoreductase (Qor), which in the first step of the catabolic pathway catalyzes the hydroxylation at position 2 of the quinoline ring, belongs to the xanthine oxidase family of molybdenum enzymes; it contains the molybdopterin cytosine dinucleotide form of the molybdenum pyranopterin cofactor (MCD), two [2Fe2S] clusters, and FAD (2, 37). The crystal structure of Qor, as well as that of the oxygenase component of the second enzyme, 2(1H)-quinolinone 8-monooxygenase (OxoOR), was solved recently (2, 31).

A gene cluster of P. putida 86 (EMBL/GenBank accession number AJ583091) contains, besides genes coding for the first two enzymes of the degradation pathway (qorMSL, oxoO, and oxoR), open reading frames 1 and 2 (ORF1 and -2), presumed to code for an α/β-hydrolase fold protein and an amidase, and genes coding for accessory proteins presumed to be involved in the maturation of Qor (Fig. 1A). By analogy to XdhC from Rhodobacter capsulatus, the protein encoded by ORF3 might be a chaperone facilitating the insertion of the MCD cofactor into the apoprotein, or it might be involved indirectly in sulfuration of the cofactor, or both (27, 28). The deduced ORF4 protein contains an N-terminal region which resembles MoeC from Clostridium perfringens and contains a predicted molybdopterin binding domain and a C-terminal region similar to that of MobA, which catalyzes molybdopterin guanine dinucleotide formation from Mo-molybdopterin and GTP (46). ORF4 thus may code for a bifunctional molybdopterin-binding protein/nucleotidyltransferase, which forms the cytosine dinucleotide form of the molybdenum cofactor. Apart from a possible role in the maturation of Qor, the IscS- and IscU-like proteins encoded by ORF5 and -6 may additionally or alternatively be involved in the assembly of other iron-sulfur proteins, e.g., OxoOR. Putative ORF7, -8, and -9, located downstream of qorMSL, code for homologues of CoxG, CoxD, and CoxE, respectively; their gene products have been hypothesized to also take part in the maturation of the molybdenum enzyme Qor (3). Expression of all these genes was shown to be induced in the presence of quinoline and to require oxoS, which codes for an AraC/XylS-type transcriptional regulator (EMBL/GenBank accession number AJ617683). In contrast to ORF1 to -6, qorMSL, and the oxoO and oxoR genes, weak transcription of ORF7 to -9 was also observed in the absence of quinoline and independently of oxoS (3).

FIG. 1.

The qor region of Pseudomonas putida 86, presumed promoters, and transcriptional start sites. (A) Genetic map of the qor region. The gene oxoS, marked black, codes for a putative XylS-type transcriptional regulator; qorMSL codes for the subunits of Qor; oxoO and oxoR encode the oxygenase and reductase component, respectively, of OxoOR. For a description of ORF3 to -9, see the introduction. Potential quinoline-dependent promoters of oxoO, ORF3, qorM, and oxoR are designated PoxoO, Porf3, PqorM, and PoxoR, respectively; promoters of oxoS and ORF7 which are not influenced by quinoline are marked PoxoS and Porf7. (B) Nucleotide sequence of the oxoS promoter region. The transcriptional initiation site is shown in boldface, and putative −10 and −35 elements of the σ⁷⁰ recognition site are underlined. (C) Nucleotide sequences of the regions of the quinoline-dependent promoters of oxoO, ORF3, qorM, and oxoR. Putative binding sites for an XylS-like transcriptional activator as deduced from sequence similarity to the XylS binding sites (17) are boxed and designated I and II. Another apparently conserved motif is shown in boldface italics. Transcriptional start sites (3) are shown in boldface type, and putative −10 elements of σ³⁸ recognition sites are indicated. (D) Nucleotide sequence of the promoter region of ORF7. Putative transcriptional initiation sites as suggested by primer extension analysis are shown in boldface. For the start site that is proximal to ORF7, a putative extended −10 element (22, 33) of a σ⁷⁰ recognition site is indicated by its consensus sequence above Porf7. Sequences similar to the osmotic shock σ³⁸ recognition site (26) and to another possible σ⁷⁰ recognition site are marked by their consensus sequences below the Porf7 sequence.

Proteins of the AraC/XylS family are characterized by significant homology of their C-terminal regions (14). It was shown for XylS that this conserved C-terminal domain mediates DNA binding and transcriptional activation, whereas the N-terminal region provides effector responsiveness to the protein (21). The C-terminal region of OxoS (amino acids [aa] 143 to 244) contains a predicted α-helix-turn-α-helix DNA binding motif and shows 34% identity and 48% similarity to the corresponding region of XylS (aa 214 to 315). The N-terminal part of OxoS (aa 1 to 142), which may carry effector-binding and regulatory functions, is shorter than that of XylS (aa 1 to 213), and the sequences share only 19% identity and 30% similarity.

XylS in the presence of (alkyl)benzoate acts as transcriptional activator of the meta-cleavage operon of the TOL plasmid pWW0 (14). The target site for XylS binding has been described as two imperfect 5′-TGCA-N₆-GGNTA-3′ direct repeats, separated by 6 bp and located between −70 and −35 in the Pm promoter (17). Upstream of qorM of P. putida 86, a DNA region that resembles the consensus sequence of this XylS tandem motif was identified; however, the two half-sites are separated by only 5 instead of 6 bp (3). Remarkably, conserved 5′-TGCPu-N₆-GGATA-3′ single motifs are located upstream of oxoO, ORF3, and oxoR (Fig. 1C). The positions of these single motifs roughly correspond to the position of the distal half-site of the XylS tandem motif.

In this work, the requirement for oxoS in quinoline-dependent transcription of PoxoO, Porf3, PqorM, and PoxoR was verified, and additional quinoline-independent promoters were detected for oxoS and ORF7. The activities of all promoters of the gene cluster were determined, and the TGCPu-N₆-GGATA-like motif of PoxoO and the corresponding distal motif of the PqorM promoter were shown to be essential for quinoline-dependent transcriptional activation by OxoS. Additionally, we compared the specificities of OxoS and XylS as regulators of PoxoO, PqorM, and Pm and determined whether these regulators in the presence of their specific effectors are able to cross-activate each other's promoters.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth of strains.

The bacterial strains and plasmids used are indicated in Table S1 in the supplemental material. P. putida 86 (44) was grown at 30°C in mineral salts medium (48) with 4 mM quinoline, 16.8 mM succinate, or 8 mM benzoate. For growth on succinate or benzoate, the medium was supplemented with (NH₄)₂SO₄ (1 g/liter). The carbon sources were added repeatedly during growth. During growth on quinoline, the conversion to 2(1H)-quinolinone was followed by recording UV-visible spectra in the range of 250 to 400 nm. The decrease of absorbance at 299 nm and 323 nm was indicative of consumption of quinoline and 2(1H)-quinolinone, respectively. Escherichia coli DH5α (50) and E. coli S17-1 (45) were grown at 37°C in Luria-Bertani (LB) broth (41). When necessary, ampicillin (100 μg/ml for E. coli strains, 500 μg/ml for P. putida strains), chloramphenicol (50 μg/ml), kanamycin (50 μg/ml), or gentamicin (10 μg/ml) was added to the cultures.

DNA techniques.

Genomic DNA of P. putida 86 was isolated according to the method of Davis et al. (8). Plasmid DNA was prepared with the E.Z.N.A. Plasmid Mini Kit I (peqLab, Erlangen, Germany). Gel extraction of DNA fragments was performed with the Perfectprep Gel cleanup kit from Eppendorf (Hamburg, Germany). Digestion with restriction endonucleases, ligation, and agarose gel electrophoresis were carried out using standard protocols (41). The sequences of the constructed hybrid plasmids were confirmed by DNA sequencing (MWG-Biotech, Ebersberg, Germany); in-house sequencing according to the didesoxy chain termination method was performed using a Li-Cor model 4000L sequencer. Competent cells of E. coli were prepared and transformed using the CaCl₂ procedure as described by Hanahan (18). Conjugations of E. coli S17-1 (donor) harboring hybrid plasmids and P. putida 86 (44) or P. putida 86-1 (13) were performed by filter mating as described previously (13).

Construction of plasmids.

All primers used for PCR are listed in Table S2 in the supplemental material. To study the activity of PoxoO, a 400-bp PCR fragment that includes 144 bp of oxoS, the intergenic region between oxoS and oxoO, and the first seven triplets of oxoO was cloned into the HindIII/XhoI sites of pPR9TT (42) in frame with the lacZ gene, generating pPR9TT::PoxoO. To generate pPR9TT::Porf3, a 523-bp PCR fragment comprising 303 bp of ORF2, the intergenic region between ORF2 and ORF3, and the first seven triplets of ORF3 was cloned as described for PoxoO. For the construction of pPR9TT::PqorM, a 637-bp PCR fragment that included 399 bp of ORF6, the intergenic region between ORF6 and qorM, and the first eight triplets of qorM was cloned into the PstI/SalI sites of pPR9TT in frame with the lacZ gene. A 587-bp PCR fragment comprising 270 bp of ORF9, the intergenic region between ORF9 and oxoR, and seven triplets of oxoR was cloned as described for PoxoO to generate pPR9TT::PoxoR. To generate pPR9TT::Porf7, a 462-bp PCR fragment comprising 259 bp of the qorL gene, the intergenic region between qorL and ORF7, and the first nine triplets of ORF7 was cloned as described for PoxoO. For the construction of pPR9TT::PoxoS, a 475-bp PCR fragment that included the upstream region of oxoS and the first 11 triplets of oxoS was cloned as described for PoxoO. To study the Pm promoter, a 143-bp PCR fragment that included the XylS binding region (17, 23) and the first five triplets of the xylX gene was amplified using pJB653 (1) as a template. The PCR fragment was cloned as described for PoxoO, generating pPR9TT::Pm.

The PoxoO deletion series (see Fig. 4A) was constructed as follows. PoxoO promoter fragments were amplified by PCR using appropriate 5′ primers and the 3′ primer PToxoOr1 to construct pPR9TT::PoxoO2 and pPR9TT:PoxoO3. PToxoOr2 was the 3′ primer for PCR to construct pPR9TT::PoxoO4 and pPR9TT::PoxoO5. The PCR products were cloned into the HindIII/XhoI sites of pPR9TT in frame with the lacZ gene. PqorM promoter fragments (see Fig. 4B) for the construction of pPR9TT::PqorM2, pPR9TT::PqorM3, and pPR9TT::PqorM4 were amplified by PCR using the 3′ primer PTqorMr1 and appropriate 5′ primers. For amplification of promoter fragments to generate pPR9TT::PqorM5 and pPR9TT::PqorM6, PTqorMr2 was used as a 3′ primer. The products were cloned into the PstI/SalI sites of pPR9TT in frame with the lacZ gene.

FIG. 4.

Deletion analysis of the 5′ region upstream of the (A) PoxoO promoter and (B) PqorM promoter. The DNA region fused to the lacZ gene is shown at the left of each panel (black bars). The numbers indicate the positions of the 5′ termini of the transcriptional fusions relative to the transcription start points (+1) of oxoO and qorM. Filled boxes in the intergenic regions between oxoS and oxoO (A) and ORF6 and qorM (B) indicate “box I” as defined in the legend to Fig. 1C. P. putida 86-1 was transformed by the pPR9TT-based reporter plasmids. Growth of cells and addition of quinoline were performed as described in the legend to Fig. 3. β-Galactosidase activities of the cells grown on succinate and induced with 1 mM quinoline are shown at the right of each panel. β-Galactosidase activities were assayed after 5 hours of incubation with quinoline. The error bars indicate standard deviations.

A 1,048-bp PCR product containing the oxoS gene and 314 bp of its upstream promoter region was cloned into the XbaI site of pBBR1MCS-4 (24) to generate pBBR1MCS-4::oxoSa. For amplification of xylS and 214 bp of its upstream region containing both the Ps1 and Ps2 promoter (15), pJB653 was used as a template. The 1,185-bp PCR product was cloned into the XbaI site of pBBR1MCS-4 to generate pBBR1MCS-4::xylSa. To put the genes encoding transcriptional regulators under the control of the lac promoter, a 772-bp PCR product comprising oxoS and its presumed Shine-Dalgarno sequence and a 1,005-bp PCR product containing the xylS gene and its Shine-Dalgarno sequence were cloned into the XbaI/PstI sites of pBBR1MCS-4 to construct pBBR1MCS-4::oxoSb and pBBR1MCS-4::xylSb, respectively.

Isolation of RNA, primer extension, and reverse transcription (RT)-PCR analysis.

P. putida 86 was grown on quinoline, succinate, or benzoate and harvested in the exponential growth phase (optical density at 600 nm [OD₆₀₀], about 1.5). Total RNA was isolated using the Nucleo Spin RNAII kit from Macherey-Nagel (Düren, Germany). Residual DNA was removed by digestion with 10 U RNase-free DNase (Promega, Madison, WI) in the presence of RNase inhibitor (20 U; MBI Fermentas, Vilnius, Lithuania) for 30 min at 37°C; afterward, the RNA was repurified. The primers were 5′ end labeled with the fluorescent dye IRD800 (MWG-Biotech, Ebersberg, Germany), and extension reactions were carried out as previously described (3), using 5 μg and 10 μg of RNA as templates for amplification of the transcripts of ORF7 and oxoS, respectively. The samples were analyzed on a polyacrylamide gel (SequaGel XR; National Diagnostics, Atlanta, GA) on a Li-Cor DNA sequencer (model 4000L). A sample of a sequencing reaction that used the same labeled primer and the CycleReader Auto Sequencing Kit (MBI Fermentas, Vilnius, Lithuania) was run alongside to determine the size of the primer extension product.

RT-PCR analyses were carried out as previously described (3) with 2 or 5 μg RNA as a template. cDNA products were amplified by PCR using 1 U Taq DNA polymerase (Promega), 20 pmol of the respective forward and reverse primers, and 1 μl of the product of the RT reaction as templates.

β-Galactosidase assays.

For all P. putida or E. coli clones, 1 ml of each cell suspension was harvested by centrifugation and resuspended in 800 μl of Z buffer (60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCl, 1 mM MgSO₄, 50 mM 2-mercaptoethanol). The cells were permeabilized using one drop each of chloroform and 0.1% sodium dodecyl sulfate. β-Galactosidase activity was measured as described by Miller (32) and was expressed in Miller units (MU). Three to five independently grown cultures were assayed in triplicate, and the values were averaged.

Enzyme assays.

Cells were suspended in 100 mM Tris-HCl buffer, pH 8.5, and disrupted by sonication at 4°C. Cell debris was removed by centrifugation. The activity of Qor was determined spectrophotometrically by measuring the substrate-dependent reduction of the artificial electron acceptor iodonitrotetrazolium chloride as described previously (48). One unit of Qor activity was defined as the amount of enzyme that reduces 1 μmol iodonitrotetrazolium chloride per minute at 25°C. Protein concentrations were estimated by the method of Bradford as modified by Zor and Selinger (52), using bovine serum albumin as a standard protein.

RESULTS

Transcription of oxoS independent of the carbon source.

RT-PCR analysis of oxoS and qorMSL was performed with RNA from P. putida 86 grown on benzoate, quinoline, or succinate. Whereas expression of the qorMSL genes strictly depended on the presence of quinoline, transcription of oxoS was not influenced by the carbon source used for growth (Fig. 2). Note that larger amounts of total RNA had to be used in the RT-PCR to detect the oxoS transcript than for detection of the messenger of qorMSL, presumably due to weak expression of oxoS or instability of the mRNA of oxoS. Primer extension analysis to map the 5′ end of the oxoS transcript indicated a putative transcription initiation point at position −35 from the oxoS start codon. Independently of the carbon source used for growth, the same primer extension product was formed in comparable amounts (data not shown). Putative −10 and −35 sequences identified upstream of the proposed transcriptional start site of oxoS showed only low similarity to the consensus sequences of the −10 and −35 elements of a σ⁷⁰-dependent promoter (Fig. 1B); however, homologies to recognition sites of other σ factors were not found in the region upstream of the transcriptional start site of oxoS. To determine the activity of the postulated promoter of oxoS, transcriptional fusion analysis was performed using P. putida 86-1(pPR9TT::PoxoS). The β-galactosidase activities (1.8 to 1.5 MU) were somewhat higher than those observed for P. putida 86-1(pPR9TT) (1 MU) and were not influenced by the presence of quinoline in the growth medium. The poor conservation of the −10 and −35 elements, especially the −10 hexamer (Fig. 1B), may account for the failure to detect significant promoter activity, since a general rule holds that the greater the similarity of these regions to their consensus sequence, the better the promoter functions (9). Taken together, the results tentatively suggest that oxoS is transcribed from a very weak σ⁷⁰-dependent promoter and that transcription might be constitutive, analogous to the expression at a low level of xylS from the Ps2 promoter (15).

FIG. 2.

RT-PCR analysis of oxoS and qorMSL. As a template, total RNA from P. putida 86 grown on benzoate (lanes 2 and 3), quinoline (lanes 4 and 5), or succinate (lanes 6 and 7) was used. In lanes 1, PCR was performed with DNA as a template (positive control). RNA (5 μg and 2 μg) was used in the RT reaction for analysis of the transcripts of oxoS and qorMSL, respectively. Lanes 3, 5, and 7 are negative controls, where reverse transcriptase was omitted in the RT reaction. Lane M, size marker (1.031, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, and 0.1 kb).

Activities of putative promoters of the qor gene cluster in response to quinoline and requirement of OxoS for transcriptional activation.

Possible transcription start sites of oxoO, ORF3, qorM, and oxoR had been identified by primer extension in a previous study (3). A lacZ reporter system was used to examine whether the upstream regions of these genes (Fig. 1C) indeed represent functional promoter regions involved in induction of transcription by quinoline. All the lacZ fusions were induced in P. putida 86-1 by the presence of quinoline in the growth medium. Five hours after addition of quinoline to the cultures, β-galactosidase activities were 15- to 54-fold higher in cells grown on succinate in the presence of quinoline than in succinate-grown cells (Fig. 3). The time courses of β-galactosidase activities in cells grown on succinate in the absence of quinoline suggest that the basal level of promoter activity is higher in cells in the early exponential growth phase than in late-exponential- or even stationary-phase cells (Fig. 3). Quinoline-dependent transcription was strongest for the promoter located upstream of oxoO, coding for the oxygenase component of OxoOR. As transcription from the oxoO promoter was previously proposed to result in a polycistronic mRNA which may cover the complete gene cluster (3), it seems plausible that PoxoO is the “master promoter.” However, as full-length mRNA may not be available in sufficient amounts due to instability, additional internal promoters should be advantageous. β-Galactosidase expression from the quinoline-responsive promoter Porf3 was about 4- to 15-fold lower than expression from PoxoO (Fig. 3A and B). The time courses of quinoline-dependent transcriptional activation of Porf3 and PqorM were similar, which is consistent with the presumed role of the gene products of ORF3 to -6 in the assembly of Qor encoded by qorMSL. However, it is noteworthy that transcription from the presumed “tandem-motif promoter” PqorM (Fig. 1C) was comparatively weak under the conditions tested (Fig. 3C). Considerable quinoline-dependent induction of PoxoR was observed after 5 hours of induction with quinoline (Fig. 3D). Such a relatively strong oxoR promoter may be important to ensure the synthesis of balanced amounts of OxoO and OxoR protein components, which together form the functional oxygenase.

FIG. 3.

β-Galactosidase activities and growth of P. putida 86-1 containing (A) pPR9TT::PoxoO, (B) pPR9TT::Porf3, (C) pPR9TT::PqorM, or (D) pPR9TT::PoxoR. Cells were grown in mineral salts medium with (NH₄)₂SO₄ and succinate. At an OD₆₀₀ of 0.2, the cultures were split and quinoline (1 mM) was added to one sample of each set. β-Galactosidase activities were assayed as described in Materials and Methods for cells grown on succinate (white bars) or succinate plus 1 mM quinoline (hatched bars). Growth on succinate (□), or on succinate plus 1 mM quinoline (▪) is shown as OD₆₀₀. The error bars indicate standard deviations.

To demonstrate the role of OxoS in the expression of the various lacZ fusions, all reporter plasmids were transferred into the OxoS⁻ mutant P. putida 86-1 oxoS1::nptII oxoS2::aacC1. The β-galactosidase activities in all clones were below 4.2 MU, irrespective of the presence or absence of quinoline in the growth medium (Table 1), confirming that the oxoS gene product is required for transcriptional activation of the oxoO, ORF3, qorM, and oxoR promoters. The observation of significant β-galactosidase activities in succinate-grown P. putida 86-1 harboring fusion plasmids (Fig. 3), but not in the corresponding clones of P. putida 86-1 OxoS⁻, suggests that the oxoS product is able to mediate some transcriptional activation even in the absence of quinoline. A similar behavior has been reported for XylS, as high concentrations of XylS protein are sufficient for activation of Pm in the absence of its effector (14).

TABLE 1.

β-Galactosidase activities expressed from promoter-lacZ fusions in the OxoS⁻ mutant of P. putida 86-1^a

P-lacZ fusion plasmid	β-Galactosidase activity (MU)
	With quinoline	Without quinoline
pPR9TT::PoxoO	4.16 ± 0.40	3.41 ± 0.45
pPR9TT::Porf3	0.77 ± 0.16	0.44 ± 0.44
pPR9TT::PqorM	1.53 ± 0.11	1.06 ± 0.1
pPR9TT::PoxoR	3.07 ± 0.19	2.69 ± 0.29

^aCells of P. putida 86-1 oxoS1::nptII oxoS2::aacC1 containing derivatives of pPR9TT were grown at 30°C in mineral salts medium (48) with (NH₄)₂SO₄ (1g/liter), using succinate (16.8 mM) as a carbon source. At an OD₆₀₀ of 0.2, the cultures were split and quinoline (1 mM) was added to one sample of each set. The cultures were grown for 5 more hours, and β-galactosidase activities were measured as described in Materials and Methods. The β-galactosidase activities of P. putida 86-1 oxoS1::nptII oxoS2::aacC1(pPR9TT) grown in the presence and absence of quinoline were 1.06 ± 0.1 and 0.98 ± 0.08 MU, respectively.

DNA region required for the quinoline-dependent expression of oxoO and qorM.

In order to identify the cis-acting sequence required for transcriptional activation of quinoline-dependent promoters by OxoS, a deletion series of the region spanning nucleotides −330 to −67 of the oxoO upstream region and a series within nucleotides −561 to −63 of the qorM promoter were constructed in pPR9TT. The resulting lacZ fusion constructs were transferred into P. putida 86-1. As shown in Fig. 4, both the oxoO and qorM promoters became unresponsive when the deletion reached the region designated “box I” (Fig. 1C), suggesting that the conserved motif 5′-TGCPuCT-N₃-GGGATA-3′ is essential for transcriptional activation. Regions upstream of this site do not appear to significantly influence quinoline-dependent transcriptional activation (Fig. 4). The striking similarity of “box I” and the distal 5′-TGCA-N₆-GGNTA-3′ half-site of XylS binding strongly suggests that “box I” is part of a recognition site for the XylS-type transcriptional regulator OxoS. However, further conserved nucleotides downstream of this motif might contribute to OxoS binding, such as the conserved GGAGPyG in the region of “box II” (Fig. 1C).

Transcriptional activation of PqorM, but not PoxoO, by XylS and benzoate and activity of Qor in P. putida 86-1 OxoS⁻ complemented with oxoS or xylS.

To determine whether the similarity of the proposed cis-acting sequences of quinoline-dependent promoters and Pm is sufficient to effect cross-activation by XylS, the activities of PqorM and PoxoO in response to XylS and benzoate were monitored using lacZ reporter systems. Complementation of the OxoS⁻ mutant of P. putida 86-1 with oxoS expressed from its own promoter on pBBR1MCS-4::oxoSa restored quinoline-dependent transcriptional activation of PoxoO (Table 2). However, transcription of PoxoO in the absence of quinoline was about 19% of quinoline-induced transcription, again suggesting that OxoS, even in the absence of effector, acts as a weak transcriptional activator. The gene product of archetypal xylS, expressed on pBBR1MCS-4::xylSa from its own promoter(s), is not able to mediate noticeable effector-dependent transcriptional activation of PoxoO (Table 2). This observation suggests that despite conservation of the distal TGCA-N₆-GGNTA sequence (“box I” in Fig. 1C), absence of the proximal sequence of the archetypal XylS target site in PoxoO prevents productive binding of effector-activated XylS. However, note that the β-galactosidase activities in benzoate- and quinoline-treated cells, as well as in cells grown on succinate, were higher than the observed activity of about 4 MU in P. putida 86-1 OxoS⁻(pPR9TT::PoxoO) (Table 1). As this effector-independent increase in reporter gene expression is correlated with the presence of pBBR1MCS-4::xylSa, we may hypothesize that XylS protein weakly stimulates transcription from PoxoO.

TABLE 2.

β-Galactosidase activities expressed from promoter-lacZ fusions in the OxoS⁻ mutant of P. putida 86-1 complemented with oxoS or xylS^a

P-lacZ fusion plasmid	Regulator	β-Galactosidase activity (MU)
		Benzoate	Quinoline	None
pPR9TT::PoxoO	OxoS	504 ± 49	2,873 ± 274	569 ± 78
	XylS	20 ± 0.56	16 ± 3	17 ± 0.96
pPR9TT::PqorM	OxoS	12 ± 1.77	117 ± 18	11 ± 0.92
	XylS	19 ± 1.01	3 ± 0.66	4 ± 0.6
pPR9TT::PqorM5	OxoS	9 ± 2.55	143 ± 10	10 ± 2.36
	XylS	13 ± 0.71	2 ± 0.22	3 ± 0.87
pPR9TT::PqorM6	OxoS	6 ± 1.75	7 ± 0.88	7 ± 1.85
	XylS	2 ± 0.49	2 ± 0.6	2 ± 0.22

^aCells were grown at 30°C in mineral salts medium (48) with (NH₄)₂SO₄ (1g/liter), using succinate (16.8 mM) as a carbon source. At an OD₆₀₀ of 0.2, the cultures were split in three portions, and quinoline (1 mM) or benzoate (1 mM) was added to two samples of each set, as indicated. The cultures were grown for five more hours, and β-galactosidase activities were measured as described in Materials and Methods. The transcriptional regulators XylS and OxoS were provided from the plasmids pBBR1MCS4::xylSa and pBBR1MCS4::oxoSa, respectively.

In P. putida 86-1 OxoS⁻(pPR9TT::PqorM, pBBR1MCS-4::oxoSa) activation by quinoline resulted in β-galactosidase activities of about 117 MU, showing complementation of the chromosomal oxoS knockout by plasmid-borne oxoS. Cross-activation of PqorM by the xylS product did not occur in the presence of succinate or quinoline, but a significant increase in β-galactosidase activity was observed when P. putida 86-1 OxoS⁻(pBBR1MCS-4::xylSa) containing pPR9TT::PqorM or pPR9TT::PqorM5 (Fig. 4B) was grown in the presence of benzoate (Table 2). As the response of the PqorM promoter was shifted to the effector benzoate of the XylS protein, we suggest that activated XylS indeed recognizes and binds to a cis-acting sequence preceding qorM. In contrast, P. putida 86-1 OxoS⁻(pBBR1MCS-4::xylSa) containing pPR9TT::PqorM6, which lacks the “box I” motif of the promoter region (Fig. 4), showed negligible β-galactosidase activities under all conditions (Table 2), indicating that XylS binding to PqorM, like OxoS binding, requires the distal half-site. However, the observed inability of XylS to activate PoxoO suggests that such a distal half-site is necessary but not sufficient for XylS binding. As the proximal motif of the archetypal XylS binding site was shown to be most important for transcriptional stimulation by effector-activated XylS (17), we may speculate that despite its imperfect conservation (TGCG-N₆-GGAGT instead of TGCA-N₆-GGNTA), and despite the distance of 5 instead of 6 bp between the presumed half-sites, the proximal motif of PqorM might play an important role in XylS-mediated (effector-dependent) transcriptional activation of PqorM.

It is remarkable that in the absence of the XylS effector benzoate, the PqorM promoter appears to be less responsive to XylS than the PoxoO promoter. Construction and analysis of a series of PqorM mutant promoters would be required to analyze whether this is due to the imperfect conservation of the distal submotif (TGCG instead of the TGCA of the XylS binding site) or other features of the promoter sequence, such as the distance of the presumed −10 element from the transcriptional start site, or both.

Consistent with the results from analysis of transcriptional activation, the specific activity of Qor observed in crude extracts of the OxoS⁻ mutant of P. putida 86-1 was very low and independent of the presence of quinoline (Table 3). Complementation with oxoS, expressed from its own promoter on pBBR1MCS4, restored the inducibility of enzyme synthesis by quinoline; however, Qor activity was lower than in extracts of wild-type cells. Complementation of the oxoS knockout mutant with xylS, likewise transcribed from its own promoter(s), led to significant Qor activity in benzoate-grown cells (Table 3). Thus, as a physiological consequence of the replacement of oxoS by xylS, quinoline conversion became responsive to benzoate instead of quinoline.

TABLE 3.

Activities of Qor in crude extracts of wild-type P. putida 86, the P. putida 86-1 OxoS⁻ mutant, and P. putida 86-1 OxoS⁻ complemented with oxoS or xylS^a

Strain	Specific Qor activity (MU mg−1) after growth on:
	Succinate	Benzoate	Benzoate + quinoline	Quinoline
P. putida 86	1	1.5	188	210
P. putida 86-1 oxoS1::nptII oxoS2::aacC1	1	1.5	2	−b
P. putida 86-1 oxoS1::nptII oxoS2::aacC1(pBBR1MCS-4::oxoSa)	1	20	71	65
P. putida 86-1 oxoS1::nptII oxoS2::aacC1(pBBR1MCS-4::xylSa)	1	56	29	−

^aCells were grown at 30°C in mineral salts medium (48) on 16.8 mM succinate, 8 mM benzoate, or 4 mM quinoline. For growth on succinate or benzoate, the medium was supplemented with (NH₄)₂SO₄ (1 g/liter). The carbon sources were added repeatedly during growth. For growth on benzoate plus quinoline, 2 mM quinoline was added at an OD₆₀₀ of about 2.

^b−, no growth on quinoline as sole source of carbon and energy.

Transcriptional activation of the meta pathway promoter Pm by OxoS and quinoline.

In both P. putida 86-1(pPR9TT::Pm) and P. putida 86-1 OxoS⁻(pPR9TT::Pm) β-galactosidase activities of about 800 MU and 2,500 MU were observed when the cells were grown on quinoline and benzoate, respectively, suggesting that an endogenous benzoate-responsive XylS-like protein that activates Pm is present in P. putida 86. Regulation of Pm by the oxoS gene product thus had to be examined in a heterologous background. E. coli DH5α(pPR9TT::Pm) grown in LB showed considerable β-galactosidase activities, independently of the presence of benzoate or quinoline (Table 4). Additional presence of xylS (on pBBR1MCS-4::xylSb) increased the activity; however, activation of XylS by benzoate clearly resulted in further transcriptional activation (Table 4). Cross-activation of Pm by the oxoS gene product was monitored in E. coli DH5α(pPR9TT::Pm, pBBR1MCS-4::oxoSb). Transcription from Pm indeed was somewhat stimulated by quinoline (Table 4), indicating that effector-activated OxoS weakly complements the active form of XylS. Taken together, our data tentatively suggest that the effects of ligand binding on the properties of the regulators, and the mechanisms of transcriptional activation from the cognate promoters, are similar for OxoS and XylS.

TABLE 4.

β-Galactosidase activities expressed from Pm-lacZ fusion in E. coli DH5α in the presence of xylS or oxoS^a

Regulator	β-Galactosidase activity (MU)
	Benzoate	Quinoline	None
None	1,273 ± 47	1,197 ± 57	1,268 ± 84
XylS	6,024 ± 343	2,639 ± 156	2,488 ± 322
OxoS	1,285 ± 51	1,775 ± 182	1,216 ± 130

^aCells of E. coli DH5α(pPR9TT::Pm) clones were grown at 30°C in LB (41) supplemented with 1 mM isopropyl-β-d-thiogalactopyranoside to induce the lac promoter on pBBR1MCS-4. At an OD₆₀₀ of 0.25, the cultures were split in three portions, and quinoline (1 mM) or benzoate (1 mM) were added to two samples of each set, as indicated. The cultures were grown for five more hours, and β-galactosidase activities were measured as described in Materials and Methods. The transcriptional regulators XylS and OxoS were provided from plasmids pBBR1MCS4::xylSb and pBBR1MCS4::oxoSb, respectively.

Identification of quinoline- and OxoS-independent Porf7.

Strong transcription of ORF7 to -9 in the presence of quinoline is assumed to be driven from the quinoline-dependent PqorM promoter. However, our previous observation that some transcription of ORF7 to -9 occurs independently of quinoline and of oxoS (3) suggested that the upstream regions of these ORFs contain a separate promoter that is not part of the quinoline regulon. Primer extension analysis using total RNA from P. putida 86 as a template indicated a putative transcriptional start site located 84 nucleotides upstream of the start codon of ORF7; an additional minor start site was located 6 nucleotides further downstream (Fig. 1D). The result of the primer extension analysis was not influenced by the carbon source used for growth (data not shown). A poorly conserved −10 hexamer of a possible σ⁷⁰ recognition site was identified upstream of the start site proximal to ORF7; the sequence of the −35 region, however, is quite different from the −35 consensus sequence. Since an extended −10 element characterized by T and G at positions −15 and −14, respectively, was reported to significantly contribute to promoter strength (22, 33), the sequence TGT preceding the presumed −10 hexamer of the minor transcriptional start site might compensate for its nonconsensus −35 region (Fig. 1D). Remarkably, a sequence that is more similar to the “osmotic shock consensus sequence” GCGG-(15 to 16 bp)-CTAcacTt of the σ³⁸ recognition site (26) than to the −10 and −35 elements of the σ⁷⁰ binding site is located upstream of the putative major transcriptional initiation site (Fig. 1D). However, the distance of these elements from the presumed distal transcriptional start site is much larger than in known σ³⁸-dependent promoters (26).

Promoter activities of a Porf7-lacZ fusion in P. putida 86-1(pPR9TT::Porf7) were low (β-galactosidase activities of 9 to 13.5 MU) and independent of the presence of quinoline. Such quinoline-independent transcription from Porf7 might be a remnant of evolution if these genes were recruited from another background into this gene cluster, or it might be a necessity if the gene products are additionally involved in reactions other than quinoline catabolism.

DISCUSSION

The gene cluster of P. putida 86 involved in quinoline degradation is transcribed from several quinoline-dependent promoters (PoxoO, Porf3, PqorM, and PoxoR) in the presence of oxoS; the oxoS gene coding for a quinoline-responsive transcriptional activator of the XylS family is expressed independently of quinoline. Transcription from PoxoO, Porf3, and PqorM was previously proposed to result in multiple overlapping polycistronic mRNAs (3), suggesting that Porf3, PqorM, and even PoxoR are internal promoters of a large (>16 kb) polycistronic operon. Consistent with its presumed role as “master promoter” of quinoline catabolism, PoxoO is the strongest promoter of the gene cluster. The weaker internal promoters would have an important function when the primary transcript from PoxoO is easily interrupted or degraded. Instability of the full-length mRNA might indeed occur, as suggested by our previous failure to detect transcripts of defined length in Northern analyses (3). Internal promoters have been identified in operons of several other bacterial gene clusters. Apart from their role to ensure adequate expression of distal genes of the operon in long and unstable transcripts, multiple internal promoters may provide for differential expression of genes within an operon in response to different physiological states or growth conditions (20, 30, 51).

While polycistronic transcripts originating from PoxoO, or from the internal Porf3 or PqorM promoter, account for quinoline-dependent expression of ORF7 to -9, additional quinoline-independent promoters were detected upstream of ORF7. The proteins encoded by ORF7 to -9 have been hypothesized to be involved in the assembly or maturation of molybdenum hydroxylases (3). As P. putida 86 contains at least one other molybdenum enzyme, namely, xanthine dehydrogenase (35), a low level of expression of ORF7 to -9 independent of quinoline might be essential for housekeeping functions. Alternatively, quinoline-independent transcription from Porf7 might simply result from recruitment of ORF7 to -9 and their upstream sequence from another genetic background into this gene cluster.

To examine the specificity, or lack of it, for the interactions of the XylS-type regulators with their cognate promoters, we have compared the activities of XylS and OxoS as regulators of PoxoO/PqorM and Pm, respectively. In P. putida 86-1 OxoS⁻, the xylS gene product in the presence of its effector benzoate weakly stimulated the PqorM promoter, but not PoxoO, possibly due to closer similarity of PqorM to Pm, especially in the “box II” sequence in Fig. 1C. Very weak cross-activation of Pm by OxoS was observed in the presence of quinoline. Cross-activation has been studied intensely for XylR-type regulators and their σ⁵⁴-dependent promoters, which actually can be activated to various degrees by heterologous proteins. Cross-binding of DmpR of the phenol catabolic operon and XylR of the upper TOL (toluene) operon to each other's upstream activating sequences, or activation by XylR and HbpR of the promoters of 2-hydroxybiphenyl and the toluene operon, have been reported (11, 47). The promoters of gene clusters coding for two different toluene monooxygenases are both activated by the NtrC family regulators TbmR and TbuT (25). Cross-activation of two different pathways for toluene degradation can be mediated by the two-component signal transduction systems TmoS-TmoT and TodS-TodT in the presence of toluene (39). These examples illustrate that many upstream activating sequences within catabolic promoters are not strictly specific for their regulator, and many regulators show some promiscuity with respect to operator binding specificity. Such cross-binding of “low-specificity regulators” to “leaky” operator sequences has been proposed to play an important role in the evolution of new catabolic pathways, as it allows primitive regulation of genes acquired by horizontal gene transfer by preexisting regulators (4, 10).

The assembly and evolution of catabolic pathways is assumed to mainly result from a series of gene duplication events, followed by specialization. By concomitant duplication of the cis-acting elements, a duplicated gene should inherit the regulation from the original gene, and the existing promoter would evolve by adapting to another signal responsiveness (4). Regarding this hypothesis, the presence of a recognition site for an XylS-type transcriptional regulator upstream of the qorMSL genes is remarkable, as these transcriptional activators have not been previously described as being involved in transcriptional control of genes encoding molybdenum hydroxylases. In contrast, several other homologues of OxoS besides XylS are positive regulators of catabolic pathways that open up access to xenobiotic or aromatic carbon sources. Examples involve the ThcR protein presumed to be involved in regulation of S-ethyldipropylcarbamothioate (the herbicide EPTC) degradation by Rhodococcus erythropolis NI86/21 (34); the transcriptional regulator AntR of carbazole and anthranilate degradation by P. resinovorans CA10 (49); the EthR regulator of Rhodococcus ruber IFP2001, probably involved in ethyl tert-butyl ether degradation (6); AndR from Burkholderia cepacia DBO1, controlling expression of a three-component anthranilate 1,2-dioxygenase (5); IpbR, which might be involved in isopropylbenzene degradation by Pseudomonas sp. strain JR1 (TrEMBL accession number Q9KK00); and the BenR regulator of benzoate degradation by P. putida PRS2000 (7). Involvement of an XylS-type regulator in quinoline degradation might actually reflect the evolutionary origin of this pathway in the metabolism of aromatic carbon compounds. Aerobic bacterial degradation of aromatic compounds is typically initiated by hydroxylation reactions catalyzed by multicomponent oxygenases. Note that 2(1H)-quinolinone, the substrate of OxoOR, which is a member of the multicomponent aromatic-ring non-heme-iron oxygenases (40), also is an effector of OxoS (3). We may tentatively speculate that genes for the synthesis and maturation of the molybdenum enzyme Qor were assembled into an ancestral catabolic gene cluster (characterized by oxoS, oxoO, and oxoR) coding for the OxoS-controlled degradation of an aromatic compound. If indeed such patchwork assembly of a “molybdenum hydroxylase gene cluster” into an “aromatic-compound degradation gene cluster” occurred, the additional internal promoters Porf3 and PqorM might have evolved due to the necessity to ensure efficient substrate-controlled transcription of the large DNA region.