INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

2-Hydroxybiphenyl has been widely used in disinfectant and preservative formulations, as an intermediate in the synthesis of dyes, resins, and rubbers (71), and as a fungicide to control postharvest diseases of various fruits (15). From 1915 to 1978, 2-hydroxybiphenyl and 4-hydroxybiphenyl appeared as major by-products of the industrial synthesis of phenol. In the United States, dumping of the by-products on the production sites led to groundwater and surface water contamination with hydroxybiphenyls at nearby locations (71). 2-Hydroxybiphenyl is also formed during microbial desulfurization of dibenzothiophene in fossil fuels (26, 29).

Different bacterial strains are able to use hydroxybiphenyls as sole carbon and energy sources (23, 30). One of these strains, Pseudomonas azelaica HBP1, degrades 2-hydroxybiphenyl and 2,2'-dihydroxybiphenyl through a meta-cleavage pathway (30, 31). The initial metabolism of these compounds is catalyzed by enzymes encoded by the hbpCAD genes (57) (Fig. 1). HbpA, a flavin adenine dinucleotide-containing 2-hydroxybiphenyl-3-monooxygenase, catalyzes the NADH-dependent ortho hydroxylation of 2-hydroxy- and 2,2'-dihydroxybiphenyl to 2,3-dihydroxy- and 2,2',3-trihydroxybiphenyl, respectively (30, 31, 68). Next, HbpC, a 2,3-dihydroxybiphenyl-1,2-dioxygenase, catalyzes the meta cleavage, resulting in 2-hydroxy-6-oxo-6-phenyl-2,4-hexadienoic acid and 2-hydroxy-6-oxo-6-(2-hydroxyphenyl)-2,4-hexadienoic acid, respectively (31, 57). The last two compounds are hydrolyzed by the meta-cleavage product hydrolase HbpD to 2-hydroxy-2,4-pentadienoic acid and either benzoic acid or salicylic acid (31). Benzoic acid and salicylic acid are further converted by benzoate 1,2-dioxygenase and salicylate monooxygenase, respectively, to catechol, which is the substrate for the lower meta-cleavage pathway (30, 31).

View larger version (22K): [in this window] [in a new window]   FIG. 1.   (A) Genetic organization of the hbp genes in P. azelaica HBP1. The orientation and sizes of the genes are indicated by arrows; the solid line represents noncoding DNA. (B) Pathway for the initial metabolism of 2-hydroxybiphenyl and 2,2'-dihydroxybiphenyl in P. azelaica HBP1. The responsible enzymes are indicated below each conversion step. The names for the different compounds are as follows (in the case where the initial substrate is 2,2'-dihydroxybiphenyl, names are given in brackets): I, 2-hydroxybiphenyl (2,2'-dihydroxybiphenyl); II, 2,3-dihydroxybiphenyl (2,2',3-trihydroxybiphenyl); III, 2-hydroxy-6-oxo-6-phenyl-2,4-hexadienoic acid [2-hydroxy-6-oxo-6-(2-hydroxyphenyl)-2,4-hexadienoic acid]; IV, benzoic acid (salicylic acid); V, 2-hydroxy-2,4-pentadienoic acid. (C) Pathway for the initial metabolism of 2-propylphenol in P. azelaica HBP1 Prp. The responsible enzymes are the same as used for the initial metabolism of 2-hydroxybiphenyl and 2,2'-dihydroxybiphenyl. The names for the different compounds are as follows: I, 2-propylphenol; II, 3-propylcatechol; III, 2-hydroxy-6-oxo-2,4-nonadienoic acid; IV, butyric acid; V, 2-hydroxy-2,4-pentadienoic acid.

Although the breakdown of 2-hydroxybiphenyl in P. azelaica HBP1 had been well characterized on the biochemical level, little is known about the regulation of this pathway. Activity measurements in cell extracts revealed that the hbpCAD genes are specifically expressed in the presence of 2-hydroxybiphenyl or 2,2'-dihydroxybiphenyl (30, 31). From DNA sequence information of regions near the hbpCAD genes, hints about the possible location of a regulatory gene upstream of hbpC were obtained (57). The amino acid sequence encoded by this open reading frame (ORF) showed homology to members of the NtrC family of prokaryotic transcriptional activators. Members of this family activate gene expression in concert with the RNA polymerase holoenzyme containing the ⁵⁴ subunit (RNAP-⁵⁴) (reviewed in references 34 and 40). RNAP-⁵⁴ is unable to catalyze the isomerization to open transcriptional complexes after having formed stable closed complexes with distinct promoters containing a 24(GG)/12(GC) motif. The process can proceed only when it is coupled to the ATPase activity of an NtrC-type regulator which contacts RNAP-⁵⁴ (34); triggering the ATPase activity in such regulators depends on biochemical or physiological stimuli (40).

The intricate mechanism by which members of the NtrC family activate gene transcription is reflected in their modular design. The amino-terminal A domain is the signal receptor module, the central C domain has an ATPase activity and is responsible for contacting promoter-bound RNAP-⁵⁴, and the carboxy-terminal D domain contains a DNA binding motif (40). On the basis of sequence similarities within the A domain, different subclasses within the NtrC family can be distinguished, reflecting the different mechanisms underlying activity modulation of these regulators (58). In the subclass containing the response regulators of two-component systems, signal transduction is performed by a separate sensor histidine protein kinase mediating the phosphorylation of a conserved Asp residue in the A domain (reviewed in reference 64). Another subclass is reserved for NifA and analogous proteins. Members of this subclass are specifically inhibited by a second protein factor, such as NifL for NifA (5). The third group we will refer to as the XylR/DmpR subclass, after the two first-described and best-understood members forming this group (25, 59). Regulators of this subclass are activated by direct interaction with an effector molecule, without involvement of a sensor kinase, which is normally the substrate molecule for the catabolic pathways they control (58). In XylR and DmpR the A domain acts as a specific interdomain inhibitor which occludes the otherwise constitutive ATPase activity of the central C domain (18, 42, 47). In the current model of activation, the binding of an effector molecule leads to a conformational change in the A domain which is transmitted through a short flexible interdomain linker hinge region, the Q linker (72), in such a way that the ATPase activity of the C domain is being derepressed (58).

In this paper, we characterize the hbpR gene, demonstrate that HbpR is necessary for transcriptional activation of the 2-hydroxybiphenyl pathway genes, and analyze its phylogenetic relation to members of the XylR/DmpR subclass within the NtrC family. The capability of different monoaromatic and biaromatic compounds to act as effectors for HbpR-mediated transcription activation is investigated. Finally, the effect of a spontaneous point mutation in a conserved residue within the C-terminal region of the A domain of HbpR on HbpR-mediated transcription activation is illustrated.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The bacterial strains used in this study are presented in Table 1. P. azelaica HBP1 is able to use 2-hydroxybiphenyl and 2,2'-dihydroxybiphenyl as the sole source of carbon and energy (30). P. azelaica HBP1 Prp is a spontaneous mutant of strain HBP1 that is capable of growing with 2-propylphenol (32). The plasmids used in this study are listed in Table 2, and the structures of the most relevant constructed plasmids are shown in Fig. 2. Plasmid pJAMA8 (Fig. 2) was constructed to make transcriptional fusions with the luxAB genes of Vibrio harveyi (11, 27).

View larger version (38K): [in this window] [in a new window]   FIG. 2.   Restriction maps of the most relevant plasmids used in this study. All plasmids were constructed as described in Materials and Methods. At the top, the relative location of the hbpR and hbpC genes on the chromosome of P. azelaica HBP1 is depicted. All plasmid constructs are shown relative to this DNA fragment. For each plasmid the vector from which it was derived is indicated together with the abbreviation and the orientation of vector-situated promoters. All relevant restriction sites are shown except for the 2.1-kb SfiI fragment in pHYBP104, pHYBP127, and pHYBP129 containing the res-npt-res element (33). Restriction sites in parentheses were destroyed during cloning. PCR-introduced restriction sites are indicated by asterisks. Abbreviations: `hbpR', internal region of the hbpR ORF; hbpR, hbpR containing a frameshift mutation; hbpR-T613C, hbpR gene isolated from strain HBP1 Prp; A and B, indicate the 3' and the 5' region of `hbpR', which are separated by the Smr marker in pHYBP121, respectively; ATG, start codon; (ATG), internal start codon in hbpR; RBS, ribosome binding site; STOP, stop codon(s); PhbpC, 54-dependent promoter of the hbpC gene (M. C. M. Jaspers, unpublished data); PhbpR, putative 70-dependent promoter of the hbpR gene. Arrows indicate the direction of transcription of the corresponding genes; shaded boxes indicate structural genes or fragments thereof; hatched boxes indicate the hbpR-hbpC intergenic region; checked boxes indicate resolution (res) sites of RP4; black boxes indicate the 19-bp I and O ends of Tn5 (not drawn to scale); white boxes indicate noncoding DNA; black solid lines indicate vector-derived DNA; black triangles indicate the orientation of vector-situated promoters; white triangles indicate the orientation of putative promoters in the hbpR-hbpC intergenic region; the grey triangle (in pHYBP133) indicates the introduced frameshift mutation in the hbpR ORF; hairpin-like structures represent transcriptional terminators.

Media and growth conditions. Escherichia coli strains were grown at 37°C on Luria-Bertani medium (54). P. azelaica strains were grown at 30°C on Pseudomonas mineral medium (MM) (21), containing 10 mM succinate, 500 mg of 2-hydroxybiphenyl per liter (2.9 mM), or 250 mg of 2-propylphenol per liter (1.8 mM). When required, the media were supplemented with the following antibiotics at the indicated concentrations: ampicillin, 100 µg/ml (E. coli); chloramphenicol, 25 µg/ml (E. coli); kanamycin, 50 µg/ml (E. coli and P. azelaica); rifampin, 50 µg/ml (E. coli); streptomycin, 50 µg/ml (E. coli) or 200 µg/ml (P. azelaica); and tetracycline, 10 µg/ml (E. coli) or 30 µg/ml (P. azelaica). When necessary, the media were supplemented with 0.004% 5-bromo-4-chloro-3-indolyl--D-galactopyranoside (X-Gal) or 1.0 mM isopropyl--D-thiogalactopyranoside (IPTG).

Preparation of cell extract and enzyme activity measurements. Cell extracts were prepared from 50-ml cultures of P. azelaica HBP1, HBP1-Prp, HBP121, HBP127, and HBP129 pregrown for 24 h in MM containing 10 mM succinate and then exposed for 10 h to 2-hydroxybiphenyl (1.2 mM), 2-propylphenol (1.2 mM), or succinate (1.5 mM), as described previously (30, 32). The activity of 2-hydroxybiphenyl 3-monooxygenase (HbpA) was measured in cell extracts by monitoring the disappearance of NADH at 340 nm as described elsewhere (30). The activities of 2,3-dihydroxybiphenyl dioxygenase (HbpC) and 2-hydroxy-6-oxo-6-phenyl-2,4-hexadienoic acid hydrolase (HbpD) were measured in cell extracts by monitoring 2-hydroxy-6-oxo-6-phenyl-2,4-hexadienoic acid (the meta-cleavage product) formation and disappearance at 434 nm, respectively, as described previously (31). Protein concentrations were determined by the method of Bradford (8) with bovine serum albumin as a standard.

Recombinant DNA techniques and DNA sequencing. Plasmid DNA isolations, restriction endonuclease digestions, ligations, transformations, and other DNA manipulations were carried out by well-established procedures (54). DNA amplification by PCR was performed with Taq DNA polymerase (Gibco BRL, Life Technologies, Inc., Gaithersburg, Md.) as described elsewhere (37).

Sequence analysis. Comparisons of sequence data with published sequences in the nonredundant (NR) protein sequence database were performed with version 2.0 of the BLAST program via the Internet at http://www.ncbi.nlm.nih.gov/BLAST (3). Pairwise comparisons between protein sequences were made with the BLAST 2 sequences algorithm. A clustal alignment of the predicted amino acid sequences of the HbpR protein with other proteins was made with the program MegAlign from the Lasergene package (DNASTAR, Inc.). Phylogenetic inferences were derived on a set of aligned amino acid sequences with significant homology to HbpR, prepared with the PILEUP routine of the Wisconsin sequence analysis software package 8.0 (Genetics Computer Group, Madison, Wis.). Bootstrapping (100 replicas), protein distance calculations (PROTDIST), maximum-parsimony analysis (PROTPARS), and neighbor-joining analysis (NEIGHBOR) were carried out with the routines in the program package PHYLIP (version 3.5c) (17).

Overexpression of hbpR in E. coli. For overexpression of the hbpR gene in E. coli, plasmid pHYBP132, in which the ATG triplet as present in the NcoI site of plasmid pET3d (66) was used as the start codon for hbpR, was constructed. To do this, a 481-bp fragment of hbpR was amplified by PCR on P. azelaica HBP1 total DNA to introduce an NcoI site at the start of hbpR, which was then cloned with the remaining part of hbpR in several steps into pET3d (pHYBP132 [Fig. 2]). Plasmid pHYBP133 is similar to pHYBP132 except for an interruption of the ORF of hbpR by a 4-bp deletion at the internal SphI site (giving the smaller ORF hbpR) (Fig. 2). To test HbpR expression, cultures of E. coli BL21(DE3)(pLysS) harboring plasmid pHYBP132 (hbpR), pHYBP133 (hbpR), or pET3d were induced with IPTG. Crude extracts were prepared and analyzed as described previously (50).

Disruption of the chromosomal hbpR gene copy and trans complementation. To disrupt the ORF of the hbpR gene on the P. azelaica chromosome, we used single recombination with a nonreplicating plasmid (pHYBP121) containing an internal fragment of hbpR, which was interrupted by an Sm^r marker (Fig. 2). Plasmid pHYBP121 was transferred from E. coli S17-1pir to P. azelaica in a biparental mating carried out on filter disks as described elsewhere (22). Selection for P. azelaica recombinants was done on MM plates supplemented with streptomycin and 10 mM succinate. A single recombinant, P. azelaica HBP121, was purified and verified by Southern hybridizations on total DNA.

Construction of a transcriptional fusion of the hbpRC intergenic region with luxAB. A 704-bp DNA fragment containing the intergenic region between the hbpR and hbpC genes, as well as the first 47 nucleotides of the hbpC coding region, was obtained by PCR on P. azelaica HBP1 total DNA with primers HBP-1 (5'-GCATGCGATGGTTCAGGTCCGG-3'; the SphI site is underlined) and HBP-2 (5'-TCTAGATCACTTACACCTAAAACG-3'; the XbaI site is underlined). This fragment was cloned into pT7Blue(R)-T, resulting in plasmid pHYBP100. The insert of pHYBP100 was sequenced and confirmed to be identical to the original sequence. The hbpR-hbpC intergenic region was recovered from pHYBP100 as a 0.7-kb SphI-XbaI fragment and ligated with pJAMA8 digested with SphI and XbaI (producing pHYBP103). The 3.1-kb NotI fragment of pHYBP103 containing the luxAB fusion was inserted into PCK218 (33) by replacing its 3.2-kb NotI fragment with the 3.1-kb NotI fragment of pHYBP103. This resulted in plasmid pHYBP104, which was used for chromosomal insertion of the luxAB-based transcriptional fusion in P. azelaica HBP1 or HBP1 Prp. Proper insertions were verified by Southern hybridization (strains HBP104 and HBP104Prp [Table 1]).

Luciferase assays. Activity of the hbpC promoter in E. coli and P. azelaica strains was analyzed by measuring luciferase activity. For reasons of convenience and reproducibility, frozen stocks were prepared from each strain to be tested. These stocks were prepared by adding dimethyl sulfoxide (DMSO) (5% [vol/vol]) to washed cultures which had grown to mid-log phase (optical density at 600 nm of 0.60). Stocks were stored at 80°C until further use and were used within 2 months, a period during which no decrease of the inducibility of the hbpC promoter could be observed.

Synthetic oligonucleotides and chemicals. Primers labeled with the fluorescent dye IRD-800 at the 5' end were purchased from MWG-BIOTECH GmbH (Ebersberg, Germany); all other primers were obtained from Microsynth GmbH (Balgach, Switzerland). 2,3-Dihydroxybiphenyl was obtained from Wako Chemicals GmbH (Neuss, Germany), 2-chlorobiphenyl was obtained from Johnson Matthey GmbH (Karlsruhe, Germany), and 3-hydroxybiphenyl was obtained from Eastman Kodak Co. (Rochester, New York). 2,2'-Dihydroxybiphenyl, 2,5-dihydroxybiphenyl, 4,4'-dihydroxybiphenyl, 2-ethylphenol, 2-hydroxydiphenylmethane, 3-methylcatechol, and 2-propylphenol were purchased from Aldrich-Chemie (Steinheim, Germany). All other organic chemicals were obtained from FLUKA Chemie AG (Buchs, Switzerland).

Nucleotide sequence accession number. The nucleotide sequence of hbpR has been deposited in the GenBank database under accession no. U73900.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

DNA sequence and overexpression of hbpR. Recently, the hbpCAD genes, which encode the enzymes responsible for the initial metabolism of 2-hydroxybiphenyl in P. azelaica HBP1, were cloned and sequenced (57). Analysis of the DNA region located upstream of the hbpC gene revealed a large ORF which was oriented in the opposite direction to the hbpCAD genes (Fig. 1A). The nucleotide sequence of this ORF (tentatively named hbpR [see below]) was 1,710 bp, corresponding to a protein of 570 amino acids (aa) with a predicted molecular mass of 63,004 Da. To confirm the integrity of the ORF and the actual size of the hbpR-encoded protein, the hbpR coding region was fused with the ATG start codon present on plasmid pET3d (pHYBP132) and overexpressed in E. coli BL21(DE3)(pLysS). A polypeptide of approximately 63 kDa could be seen on sodium dodecyl sulfate-polyacrylamide gel electrophoresis in crude extracts prepared from induced E. coli BL21(DE3)(pLysS) harboring plasmid pHYBP132 (Fig. 3, lane 2). This size corresponded to that predicted from the sequence of hbpR. The 63-kDa protein was absent in extracts from induced BL21 cultures containing pHYBP133, which showed a 31-kDa protein instead (lane 3). This shorter protein was the result of the introduced frameshift mutation at the unique internal SphI restriction site at nucleotide 520 in hbpR. This mutated hbpR gene (hbpR) would code for a shorter protein of 272 aa with a predicted molecular mass of 30,679 Da. Both the 63- and 31-kDa proteins were absent in crude extract prepared from E. coli BL21(DE3)(pLysS) carrying pET3d itself (Fig. 3, lane 1). A second specifically induced protein band of 26 kDa was observed in extracts of E. coli BL21(DE3)(pLysS) harboring pHYBP132 or pHYBP133 (lanes 2 and 3). This protein may have been produced from an internal ATG codon at nucleotide 1006 of hbpR (downstream of the introduced frameshift mutation) (Fig. 2).

View larger version (69K): [in this window] [in a new window]   FIG. 3.   Overexpression of the hbpR gene in E. coli. Shown is a Coomassie blue-stained sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel of crude extracts from E. coli BL21(DE3)(pLysS) cultures induced with IPTG and harboring pET3d (lane 1), pHYBP132 (intact hbpR; lane 2), or pHYBP133 (frameshift in hbpR, hbpR; lane 3). Lane 4 shows molecular markers with their masses indicated. The black triangles in lanes 2 and 3 indicate the full-length HbpR (63 kDa) and the out-of-frame truncated HbpR (31 kDa), respectively. The white triangles in lanes 2 and 3 point to the shorter (26-kDa) in-frame product possibly originating from an internal hbpR ATG codon. The indicated protein bands were not visible under noninduced conditions (data not shown).

HbpR is a member of the NtrC family of bacterial transcriptional activators. Comparison of the deduced amino acid sequence of the HbpR protein with other sequences in the nonredundant (NR) protein sequence database revealed that HbpR is homologous to members of the NtrC family of bacterial transcriptional activators. The homology of HbpR to most NtrC family members was restricted to the central C domain, with the exception of 13 proteins within the XylR/DmpR subclass that had an overall homology to HbpR (Fig. 4). Pairwise comparisons revealed that within this subclass, HbpR showed the highest overall homology to TbuT, the regulator of the toluene-3-monooxygenase operon in Burkholderia pickettii PKO1 (10) (now renamed Ralstonia pickettii PKO1 [73]), with 42% identity and 58% similarity in a 553-aa overlap. HbpR showed the lowest overall homology to XylR, the activator for both the upper-pathway genes and the regulatory gene xylS on plasmid pWW0 in Pseudomonas putida mt-2 (25), with 37% identity and 53% similarity in a 549-aa overlap. Considering the different domains, the extent of homology between HbpR and the other members of the XylR/DmpR subclass was higher for the C domain (48 to 56% identity, 65 to 71% similarity) and the D domain (35 to 57% identity, 67 to 76% similarity) than for the A domain (31 to 37% identity, 45 to 56% similarity).

View larger version (9983K): [in this window] [in a new window]   FIG. 4.   Alignment of the predicted amino acid sequences of the HbpR protein and the 13 most similar members of the NtrC family of transcriptional activators in the nonredundant (NR) protein sequence database. The alignment was made with the program MegAlign from the Lasergene package (DNASTAR, Inc.) and further improved manually. The ruler shown above the sequences corresponds to the amino acid sequence of the HbpR protein. Conserved amino acid residues in all proteins are indicated against a black background, whereas residues conserved in at least 10 of 14 proteins appear boxed. Gaps introduced to maximize the alignment are dashed. The regions corresponding to the different domains are boxed, and their names are displayed above the sequences. The borders of the A, C, and D domains in the different proteins were set according to those proposed for the XylR protein (25). The sections C1 to C7, shown below the sequences, correspond to the conserved regions within the C domain as described by Morett and Segovia (40). The Walker A motif within regions C1 (55, 69), involved in ATP binding, and a putative helix-turn-helix in the D domain (40) are indicated below the sequences. U represents hydrophobic residues. White triangles below the sequences mark fully conserved residues within the C and D domains of the XylR/DmpR subclass members but not fully conserved in an alignment made by Morett and Segovia, which included 27 members of the NtrC family, including XylR (40). Black triangles below the sequence mark positions which were not conserved in at least 10 of 14 proteins but appeared to be conserved in the alignment made by Morett and Segovia (40). Underlined numbers below the black triangles indicate which residues replaced the conserved ones: 1, F or Y instead of L or M; 2, A or V instead of H; 3, G, K or Q instead of only G: 4, A, G, I, S, or V instead of only G; 5, A, M, or V instead of I or V; 6, A or P instead of only P; 7, C, L, or M instead of L or M; 8, L, M or V instead of L or M. Encircled residues mark the positions of described point mutations within XylR/DmpR subclass members. The phenotypes of described point mutations is indicated above the sequence: black dots, altered effector specificity (ALT); open diamonds, constitutive or semiconstitutive transcription-promoting activity (CON); empty squares, abolition of transcription-promoting activity (NON); crosses, suppressor mutation of a semiconstitutive mutation (SUP); or a combination of these symbols. Numbers above these symbols are used to describe all known mutations for the corresponding position and their resulting phenotype(s): 1, DmpR-L83P, SUP selected for DmpR-E135A (42); 2, XylR-P85S, CON (13); 3, DmpR-F93L, SUP selected for DmpR-E135D (42); 4, DmpR-E133K, NON (61); 5, DmpR-E135A, ALT CON (61); DmpR-E135D, ALT CON (61); DmpR-E135K, ALT (46); DmpR-E135R, ALT (61); XylR-D135E, NON (53); XylR-D135N, CON (13); XylR-D135Q, CON (53); 6, DmpR-D140K, ALT CON (61); 7, XylR-E172K, ALT (12); 8, DmpR-R184W, ALT (12); 9, DmpR-K188E, SUP selected for DmpR-E135D (42); 10, HbpR-W205R, CON (this study); DmpR-W193R, SUP for DmpR-V276A (42); 11, DmpR-L219P, CON (61); XylR-V219DA220P (double mutation), CON (18); 12, XylR-V219DA220P (double mutation), CON (18); 13, DmpR-A241T, CON (61); 14, DmpR-G268S, NON (42); XylR-G268N, NON (48); 15, DmpR-V276A, CON (61); DmpR-V276G, CON (61); 16, DmpR-P297R, SUP selected for DmpR-E135D (42); 17, DmpR-F424L, CON (61); 18, XylR-R453H, NON (48). The database entries for the different proteins are given in the legend of Fig. 6.

Disruption of the hbpR gene in P. azelaica HBP1. Southern analysis of total DNA isolated from P. azelaica HBP1 with a probe against the hbpR gene indicated that hbpR was present in only one copy on the chromosome (data not shown). The hbpR gene was disrupted in strain HBP1 by single recombination with the introduced plasmid pHYBP121. Attempts to obtain double recombinants by plating out serial dilutions of cultures of single recombinants on MM containing streptomycin, followed by counterselection against the Tc^r marker, were not successful in our hands. However, every type of single and double recombination of the truncated hbpR gene on pHYBP121 with the chromosomal hbpR copy would lead to disruption of the hbpR ORF. Southern analysis of the total DNA from one purified recombinant, designated strain HBP121, showed that in this strain pHYBP121 had integrated in region B (Fig. 2). Therefore, it contained two partial copies of hbpR, the first copy lacking the 5' part and the second copy disrupted by the Sm^r marker (data not shown).

1

1

HbpR activates transcription from the hbpC promoter in the presence of 2-hydroxybiphenyl. To establish whether HbpR was indeed able to activate gene transcription from a promoter upstream of the hbpC gene and whether 2-hydroxybiphenyl was the necessary effector for activation, a transcriptional fusion between the hbpRC intergenic region and the luxAB genes of V. harveyi was inserted into the chromosome of P. azelaica HBP1. The resulting strain, P. azelaica HBP104, showed a 23-fold-increased bioluminescence after a 3-h incubation in the presence of 0.2 mM 2-hydroxybiphenyl (Fig. 5A). This indicated that the hbpRC intergenic region contained a regulatable promoter (called the hbpC promoter, P_hbpC) and again demonstrated the role of 2-hydroxybiphenyl as an effector. When the hbpR gene in strain HBP104 was disrupted by homologous recombination (P. azelaica HBP104121) with plasmid pHYBP121 at site A (Fig. 2 and data not shown), inducible expression from the hbpC promoter was abolished (Fig. 5A, inset). Adding 2-hydroxybiphenyl to strain HBP104121 even reduced the levels of observed bioluminescence by 37% after 3 h.

View larger version (30K): [in this window] [in a new window]   FIG. 5.   HbpR-mediated in vivo transcriptional activation from PhbpC in a homologous (P. azelaica HBP104, HBP104121, HBP104 Prp) or heterologous (E. coli) host system. (A) Activation from PhbpC in P. azelaica HBP104 containing a functional hbpR gene. Induction took place with 2-hydroxybiphenyl (black circles), 2-propylphenol (black triangles), or DMSO only (white circles). Inset: Activation from PhbpC in P. azelaica HBP104121 containing a disrupted hbpR gene. Induction took place with 2-hydroxybiphenyl (black diamonds) or DMSO only (white diamonds). (B) Activation from PhbpC in P. azelaica HBP104 Prp containing the mutated hbpR gene (hbpR-T613C). Induction took place with 2-hydroxybiphenyl (black circles), 2-propylphenol (black triangles), or DMSO only (white circles). (C) Activation from PhbpC in E. coli with functional hbpR (pHYBP109, circles) or dysfunctional hbpR (pHYBP110, diamonds). Induction took place with 2-hydroxybiphenyl (black symbols) or DMSO only (white symbols). The concentrations of 2-hydroxybiphenyl and 2-propylphenol used in the assays were 0.2 mM. The error bars indicate the standard deviation in two independent experiments each carried out in triplicate. Structures of the relevant hbp configurations are depicted above the respective panels. RLU, relative light units.

HbpR possesses a limited effector range. Several compounds other than 2-hydroxybiphenyl, including different monoaromatics, hydroxybiphenyls, alkylphenols, and polycyclic aromatic compounds, were tested (at 0.2 mM) for their ability to activate luciferase expression from P_hbpC in P. azelaica HBP104 (Table 4). Significant induction was found only for a small group of structurally similar compounds, including 2-hydroxybiphenyl, 2,2'-dihydroxybiphenyl, 2-aminobiphenyl, and 2-hydroxydiphenylmethane (in order of the relative induction observed). Other compounds and monoaromatics did not yield significant induction. A few compounds even reduced luciferase activities to below the background observed without any effector. These included 2,3-dihydroxybiphenyl, 2,5-dihydroxybiphenyl, 3-methylcatechol, and 2,3-dihydroxybenzaldehyde.

Spontaneous mutation in hbpR results in 2-propylphenol utilization. A previously isolated spontaneous mutant of P. azelaica HBP1, designated strain HBP1 Prp, is capable of using the non-wild-type substrate 2-propylphenol as the sole source of carbon and energy by means of the 2-hydroxybiphenyl pathway enzymes (32) (Fig. 1C). The HbpA, HbpC, and HbpD enzymes in strain HBP1 Prp were expressed not only when grown in the presence of 2-propylphenol but also when grown in the presence of succinate (Table 3), which suggested that the strain carried a regulatory mutation. To analyze the hbpR gene of strain HBP1 Prp, it was retrieved from a partial gene library of strain HBP1 Prp chromosomal DNA in pUC18 by selection for HbpC activity in E. coli. In this way, a 6.8-kb EcoRI fragment, covering the complete hbpRC genes of the Prp mutant, was cloned. Complete double-stranded sequence analysis of the region from the start of hbpC to the 3'-end of hbpR revealed one single transition mutation, TC, at position 613 from the start of the hbpR gene. This changed codon 205 from TGG to CGG, resulting in a TrpArg substitution in the deduced amino acid sequence of HbpR (Fig. 2 and 4). Overexpression of hbpR-T613C in E. coli BL21(DE3)(pLysS) as before for wild-type hbpR resulted in a similar-size protein (63 kDa) (data not shown), indicating that the mutation caused no apparent instability of the protein in E. coli. Strain HBP121 (containing a knockout of the wild-type hbpR gene) could be complemented in trans with the hbpR-T613C gene by mini-Tn5 delivery from plasmid pHYBP129 and conferred on this strain (HBP129) the ability to grow with 2-hydroxybiphenyl (data not shown). This indicated that hbpR-T613C was producing an active HbpR protein. Strain HBP129, in contrast to strains HBP1 (wild-type) and HBP121, could indeed also use 2-propylphenol as the sole source of carbon and energy. The maximum specific growth rate on 2-propylphenol was 0.082 h¹ (t_d = 8.5 h) for strain HBP129, which was slightly reduced in comparison with the observed µ_max on 2-propylphenol for strain HBP1 Prp (µ_max = 0.11 h¹, t_d = 6.4 h). Similar to the findings with strain HBP1 Prp, HbpA, HbpC, and HbpD activities in strain HBP129 were expressed during growth not only on 2-propylphenol but also on succinate (Table 3), although all three enzyme activities were slightly more elevated in HBP1 Prp grown with 2-propylphenol. This showed that HbpR-W205R had the same effects in HBP1 as in the mutant HBP1 Prp and that this mutation (T613C) was solely responsible for the change in activity of HbpR. The same luciferase levels were measured from the P_hbpC-luxAB fusion in P. azelaica HBP104 Prp in the absence of inducer (Fig. 5B) as with strain HBP104 in the presence of 2-hydroxybiphenyl (Fig. 5A). This suggests strongly that HbpR-W205R mediates constitutive expression of the hbpC promoter. Luciferase expression in strain HBP104 Prp was even reduced in the presence of 2-hydroxybiphenyl or 2-propylphenol compared to a control reaction with DMSO only (Fig. 5B). This is probably the result of direct inhibition of the luciferase reaction by phenolic compounds (20), since this decrease was not observed from enzyme activity levels of HbpA, HbpC, or HbpD in strain HBP129 (Table 3).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study we characterized the P. azelaica HBP1 hbpR gene. The nucleotide sequence of the hbpR gene was determined, and the deduced protein product of 570 aa was identified by overproduction in E. coli as a protein with an apparent mass of 63 kDa. The role of the HbpR protein as a key element in the regulation of the hbpCAD genes was inferred from batch growth experiments with strains in which the hbpR gene was disrupted (as in strain HBP121) and complemented again (as in strain HBP127). Induction experiments with E. coli provided evidence that HbpR itself activated expression from the hbpC promoter. Since E. coli cannot degrade 2-hydroxybiphenyl, this also identified 2-hydroxybiphenyl as the actual inducer rather than a metabolite formed during its degradation. Analysis of a spontaneous mutant of P. azelaica HBP1 degrading 2-propylphenol identified one residue located near the C-terminal end of the A domain within HbpR, which might be important for maintaining C-domain repression.

HbpR displayed significant amino acid sequence homology to members of the XylR/DmpR subclass within the NtrC family of bacterial transcriptional activators. The homology to other members of the NtrC family was restricted to the central C domain. Phylogenetic analysis suggested that the HbpR protein is a distinct member clustering outside most other known XylR/DmpR-type activators (Fig. 6). Nevertheless, HbpR seems to have the same mechanism of direct effector activation as the other members of the XylR/DmpR subclass. The possibility that HbpR is part of a two-component signal transduction system seems very unlikely for two reasons: (i) the A domain of HbpR has no significant homology to the A domains of response regulators within the NtrC family, and (ii) it is very unlikely that E. coli possesses a cognate histidine kinase reacting on 2-hydroxybiphenyl.

View larger version (14K): [in this window] [in a new window]   FIG. 6.   Single unrooted phylogenetic tree of the HbpR amino acid sequence with 13 members of the XylR/DmpR subgroup. The tree was obtained by protein distance calculations (PROTDIST, Kimura setting) and subsequent neighbor-joining analysis (NEIGHBOR; both programs from the software package Phylip version 3.5 [17]) on a subalignment of 616 aa including gaps. Values at the forks indicate the number of times that particular branching occurred among 100 resamplings of the data set (generated with bootstrapping) as determined by consensus neighbor-joining analysis (first value) or maximum-parsimony analysis (second value). Missing values indicate that particular branching to be absent in the consensus tree. The consensus tree was displayed by using TREEVIEW (45). The bar indicates 0.1 change per site. The origins of the different proteins are as follows: AphR from Comamonas testosteroni TA441 (GenBank accession no. AB006480) (4); BphR encoded on plasmid pNL1 from Sphingomonas aromaticivorans F199 (GenBank accession no. AF079317) (52); DmpR encoded on plasmid pVI150 from Pseudomonas sp. strain CF600 (GenBank accession no. X68033) (59); HbpR from P. azelaica HBP1 (GenBank accession no. U73900) (this study); MopR from Acinetobacter calcoaceticus NCIB8250 (GenBank accession no. Z69251) (56); PheR from P. putida BH (GenBank accession no. D63814); PhhR from P. putida P35X NCIB9869 (GenBank accession no. X79599) (43); PhlR encoded on plasmid pPGH1 from P. putida H (GenBank accession no. X91145) (41) in the figure indicated as PhlR (P.); PhlR from a Ralstonia eutropha JMP134 derivative (GenBank accession no. AF065891), in the figure indicated as PhlR (R.); PhnR from Burkholderia sp. strain RP007 (GenBank accession no. AF061751) (35); PoxR from R. eutropha E2 (GenBank accession no. AF026065) (24); TbuT from R. pickettii PKO1 (GenBank accession no. U72645) (10); TmbR from P. putida TMB (GenBank accession no. U41301) (16); XylR encoded on plasmid pWW0 from P. putida mt-2 (GenBank accession no. M20635) (25).

Most members of the XylR/DmpR type react on a variety of monoaromatic compounds. XylR, for example, recognizes not only the growth substrates toluene, m- and p-xylene, and benzyl alcohol but also several nongrowth compounds like trimethylbenzene, ethyl- and chlorotoluene, and p-chlorobenzaldehyde (1). DmpR reacts not only with the pathway substrates phenol and methylphenols but also with dimethyl-, ethyl-, and chlorophenols, benzyl alcohol, and salicylic acid (60). HbpR is unique in the sense of recognizing compounds with a biphenyl backbone but not monoaromatic structure as effectors. We found only four compounds acting as effectors (Table 4), which all contained a biaromatic ring structure plus a hydroxy or amino group at the ortho ring position, since biphenyl itself and 3-hydroxy, 4-hydroxy-, 4,4'-dihydroxy-, and 2-chlorobiphenyl did not cause activation of HbpR. However, 1-naphthol was not an effector for HbpR activation, indicating that the overall size or orientational flexibility between the two aromatic rings is important.

Interestingly, but not unusually for pathways of aromatic degradation, the substrate spectrum of the 2-hydroxybiphenyl catabolic enzymes is different from the effector spectrum of the cognate regulator HbpR. For example, HbpA, the first enzyme of the hydroxybiphenyl pathway, hydroxylates various molecules with a 2-hydroxyphenyl-R structure, with R being a hydrophobic group (e.g., methyl, ethyl, propyl, sec-butyl, phenyl, or 2-hydroxyphenyl) (30). However, 2-hydroxybiphenyl and 2,2'-dihydroxybiphenyl are the only two compounds recognized by HbpR which are also substrates for HbpA. 2-Aminobiphenyl and 2-hydroxydiphenylmethane, on the other hand, cannot serve as substrates for HbpA (57) but are (fortuitous) inducers of the hbpCAD system. A spontaneous mutant of strain HBP1, which seemed to have overcome the regulatory constraints set by wild-type HbpR and which could benefit from the flexible capabilities of the HbpA, HbpC, and HbpD enzymes to using 2-propylphenol as growth substrate as well as 2-hydroxybiphenyl, could be isolated (32). Our results presented here demonstrated that a single mutation in the hbpR regulatory gene was responsible for allowing the mutant strain to grow on 2-propylphenol. However, the mutant HbpR protein did not seem to have acquired any special new recognition specificities. Rather, it seemed to have been locked in a constitutively active form (Table 3 and Fig. 5B). From an evolutionary perspective, this is an effective way for bacteria to deal with potentially new growth substrates. The resulting amino acid substitution (W205R), interestingly, is at a (seemingly) conserved position among activators of the XylR/DmpR type (Fig. 4). For DmpR, one mutant with the same substitution at this position (DmpR-W193R) has been described (42). This mutant was originally selected as a suppressor mutation on DmpR-V276A but does not cause a constitutive phenotype as an isolated mutation. However, since the Trp-205 residue in HbpR is very close to the interdomain hinge or Q-linker region whereas Trp-193 in DmpR is not, substitutions at this position might have different effects on regulator activity.

The DNA sequence of the hbpR gene predicts a similar subdomain (A, B, C, and D) structure for HbpR to that for the other XylR/DmpR-type activators (Fig. 4). The A domains of these regulators are presumed to be the receptor module, directly interacting with the effector molecule and transmitting this signal to the C domain, which carries the ATPase activity (58). The present understanding is that the A domain is an interdomain repressor, keeping the activity of the C domain low when no signal is present (18, 42, 44, 47). Analysis of several mutations, mostly in XylR or DmpR, has illustrated the current activation model (summarized in Fig. 4). For example, mutations which changed the effector specificity of XylR or DmpR were found exclusively in the A domain (Fig. 4). In addition, the isolated A domain of DmpR was shown to bind phenol in vitro (44). Several mutations point to the importance of a small region between residues 140 and 160 (Fig. 4) for effector binding. Since the effector spectrum of HbpR is so different from the other XylR/DmpR-type activators, it would be interesting to determine which amino acid residues allow the unique binding of biaromatic structures into (or onto) the HbpR protein. It is clearly too early to do so, but two features in the A domain attract attention. One of these is the presence of four extra amino acid residues in the HbpR sequence in the region between aa 110 and 130 compared to other XylR/DmpR members. The other feature is the considerably shorter C-terminal end of the A domain directly preceding the B-domain region (Fig. 4). This last feature is shared with the predicted sequence of BphR, a putative regulator which might be involved in regulating the expression of biphenyl degradation (52). Both HbpR and BphR also lack an otherwise conserved Gly residue (at position 199), which often points to structural conservation. The B domain of HbpR itself might still be called a `Q' linker (72), with four Gln residues, whereas those of PhlR (of Ralstonia), TbuT, and PhnR can hardly be named as such (Fig. 4).

There is presently no consensus on which residues within the A domain are functionally important for all XylR/DmpR members, since most studies have concentrated on either DmpR or XylR. For example, the mutant activator DmpR-E135D has semiconstitutive activity (61) whereas the Asp residue at that position in the wild-type XylR protein does not result in semi-constitutive activity (53). XylR-D135E on its turn is a null mutant (53). This suggests that structural information will be needed to understand the subtle differences in signal transmission between the A and C domains of the different XylR/DmpR-type activators.

The specific features of the C domain are much better understood, not least because of its stronger similarity to the NtrC family of transcription activators as a whole. For example, the locations of the seven conserved motifs (C1 through C7) within the C domain (40) can readily be detected on the aligned sequences of the XylR/DmpR members (Fig. 4). The C1 and C4 motifs were shown to be involved in ATP binding and/or hydrolysis (40, 42, 48, 69). The C3 motif might form the site contacting RNAP-⁵⁴, as recently demonstrated for the NtrC-type activator DctD (70), but it also plays a role in ATP hydrolysis (51). For the other motifs, no clear function has been demonstrated, although some (C6 and C7) also seem to be involved in ATP hydrolysis or binding (48, 51). In a number of conserved positions, the XylR/DmpR-type activators differ from the other NtrC family members (Fig. 4). For example, whereas most other NtrC-type proteins carry a His at position 302 within the C3 motif (of the HbpR numbering), most XylR/DmpR members have a Val. The same holds for other residues conserved within the NtrC family (40) but not for the XylR/DmpR members (Fig. 4). The opposite, i.e., conserved residues among XylR/DmpR members but not for NtrC as a whole, is also found (Fig. 4). Some of these substitutions seem to point at protein structure differences (e.g., conserved Pro or Gly residues), and the occurrence of five conserved residues within the XylR/DmpR subclass, found almost next to one another near the C5 motif, is very suggestive for a functional difference (Fig. 4). Such differences might be a useful starting point for future site-directed mutation studies.

The last domain of the XylR/DmpR-type activators is formed by the D domain, which is contacting the DNA at the upstream activating sequences. Preceding the D domain is a region of approximately 40 residues with little conservation among the XylR/DmpR members. Conserved within the D domain is the A(L)-X₉-AA-X₂-LG motif (40), proposed to correspond to the first `helix' and `turn' of a helix-turn-helix DNA binding motif. The RPXLAYRLXK region directly at the C-terminal part might then form the second recognition helix for XylR/DmpR members. In comparison to other NtrC members, two differences in conserved residues are visible (Fig. 4), which might point to differences in recognition specificity.

In conclusion, the discovery of a novel XylR/DmpR-type transcription activator with completely different effector specificities might open up new avenues to understanding the recognition potential of the A domain and the potential to evolve new recognition specificities. It will also be interesting to study the DNA sequence in this region and find if it is more prone to acquiring small changes.