INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Chlorobenzoates (CBAs) constitute a favorable model for studying the molecular mechanisms of degradation of halogenated aromatic compounds. The most-extensively studied halobenzoate degraders are those bacteria that possess a modified chlorocatechol ortho-cleavage pathway. In these cases, halobenzoate is oxidized to the corresponding chlorocatechol, which is funneled into a modified ortho-cleavage route in which a fortuitous removal of halogen occurs. The genes for this pathway have been isolated from a number of strains and used to construct recombinant pathways for degradation of different halogenated aromatic xenobiotics (9, 41, 58, 59). The typical problem in the construction of a polychlorinated biphenyl-degrading microorganism by combining the modified ortho-cleavage and biphenyl oxidation pathways is the incompatibility of the meta and ortho pathways. The simultaneous functioning of these pathways usually creates suicide products (57); e.g., the meta fission of 3-chlorocatechol produces an acylchloride, which irreversibly inactivates (phenyl)catechol 2,3-dioxygenases such as 2,3-dihydroxybiphenyl 2,3-dioxygenase (4, 9). Hence, alternative strategies, such as the use of CBA dehalogenases, which remove chlorine prior to the oxidation of the aromatic ring, would appear to be useful for avoiding this incompatibility.

Oxygenolytic dehalogenation of 2-CBA was implicated in a number of Pseudomonas strains (23, 24, 28, 61, 72, 82). Dihydroxylation is frequently used by microbes as an initial step in the aerobic attack of aromatic compounds. The multicomponent nonheme iron dioxygenase systems catalyzing the dihydroxylation typically consist of two or three proteins that comprise a short electron transfer chain, mobilizing electrons from NADH, via flavin and 2Fe-2S redox centers, to the site of dioxygen activation (12, 47). The three-component system typically consists of an NADH: acceptor reductase component containing flavin adenine dinucleotide, a chloroplast-type 2Fe-2S ferredoxin, and a Rieske-type 2Fe-2S iron-sulfur protein (ISP) that is the terminal oxygenase component. In two-component systems, reductase and ferredoxin components are combined in the same protein. The two-component CBA 1,2-dioxygenase from 2-CBA-grown Burkholderia sp. strain 2CBS (29) is similar to the two-component plasmid-borne toluate 1,2-dioxygenase from Pseudomonas putida mt-2 (41) and the two-component benzoate 1,2-dioxygenases from P. putida C-1 (78) and Pseudomonas sp. strain B13 (35). This enzyme catalyzes the double hydroxylation of 2-halobenzoate with concomitant release of halide, carbon dioxide, and the nonchlorinated catechol (29). The plasmid-borne genes cbdABC encoding this two-component enzyme complex were isolated and sequenced (33). The cbdABC sequences showed similarity to the Acinetobacter calcoaceticus benABC genes, encoding benzoate 1,2-dioxygenase (52% identity of the deduced amino acid sequences), and to the P. putida mt-2 xylXYZ genes, encoding toluate 1,2-dioxygenase (51% identity).

Pseudomonas aeruginosa 142, used in the present study, differs from strain 2CBS in its ability to grow on both 2-CBA and 2,4-dichlorobenzoate (2,4-dCBA), its ability to oxidate and dehalogenate all ortho-halogenated dCBAs and triCBAs (61, 62), and its possession of a three-component dioxygenase (62) rather than a two-component enzyme (29). The ortho-CBA 1,2-dioxygenase activity requires molecular oxygen, NADH, and Fe(II) and results in conversion of 2-CBA to catechol and dCBAs to their respective chlorocatechols (Fig. 1). No dehydrogenase activity was required for conversion of CBA to catechol, in accord with spontaneous resolution of the 2-halo-3,5-cyclohexadiene-1,2-diol-1-carboxylic acid intermediate to catechol (29, 32, 33). The oxygenation (dehalogenation) reaction is followed by a separate catechol ortho-cleavage pathway (modified chlorocatechol ortho-cleavage pathway for dCBAs and tri-CBAs). The ortho-halobenzoate activity was resolved into three protein fractions, and a 13-kDa protein resembling a Rieske-type 2Fe-2S ferredoxin was purified and characterized (62).

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Molecular mechanism and involvement of the ohbAB genes in oxygenolytic ortho dehalogenation of halobenzoates. The three-component 1,2-dioxygenase is responsible for oxygenation followed by fortuitous ortho dehalogenation of halobenzoates in P. aeruginosa 142 (62). The ohbAB genes isolated in the present work encode two subunits of the terminal oxygenase ISPOHB of the ortho-halobenzoate 1,2-dioxygenase and presumably require reductase and ferredoxin components of a heterologous host. The hypothetical dihydrodiol intermediate 2-chloro-(chloro)cyclohexadiene-1,2-diol-1-carboxylic acid is presumed to spontaneously lose carbon dioxide and halogenide (29, 62). FAD, flavin adenine dinucleotide; FMN, flavin mononucleotide.

Our major objective was to clone and characterize the genes controlling dehalogenation of ortho-halobenzoates in P. aeruginosa 142. In this paper, we report on the isolation and expression of the novel ohb genes in Escherichia coli, the construction of a functioning recombinant pathway for growth on 2-CBA in P. putida, nucleotide sequence determination and analysis of the ohb DNA region, and the identification and phylogenetic placement of the structural ohb genes.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. P. aeruginosa 142 was provided by I. I. Starovoitov (IBPhM, Russian Academy of Science, Puschino, Russia). This strain was isolated from polychlorinated biphenyl-contaminated soil in Moscow Region, Russia, and readily grew on both 2-CBA and 2,4-dCBA (61). E. coli laboratory strains used in this work were DH5F' (Bethesda Research Laboratories, Bethesda, Md.), JM109 (79), and minicell-producing strain 925 (F⁺ minA minB thr leu thi) (69). Pseudomonas strains utilized were P. putida mt-2 (PB2440 rm⁺) (3), P. putida KZ6R (Rif^r) (83), and Pseudomonas sp. strain B13 (18). Plasmid vector pSP329 (Tc^r), a derivative of low-copy-number, broad-host-range (bhr) plasmid RK2 (IncP), contains an HaeII fragment from pUC18 carrying multiple cloning sites and the lacZ -complementation gene block cloned into the plasmid pTJS75 (63). The plasmid pSP329 was a gift from Vladimir Ksenzenko (IBPhM). Other plasmids employed were E. coli vectors pUC19 (79) and BlueScript (Stratagene, La Jolla, Calif.). Plasmid pRK2013, a Km^r Tra⁺ ColE1 derivative of RK2 (17), was used as a helper in triparental matings.

Media and growth conditions. E. coli strains were maintained at 37°C on enriched Luria broth or Luria agar (48), minimal medium M9 (45), or twice-diluted C1-free medium K1 (83). Pseudomonas strains were routinely grown at 30°C in medium K1. Growth substrates were added at the following concentrations: Glucose, 0.2% (wt/vol); benzoate and 4-hydroxybenzoate, 3.0 mM; CBAs, 0.5 to 3 mM; catechol, 2 mM; and sodium acetate, 10 mM. Antibiotics were added as needed as follows (with concentrations in micrograms per milliliter): ampicillin, 30 to 300; tetracyclin, 15; kanamycin, 30; and rifampin, 50 to 200. Isopropyl--D-thiogalactoside (IPTG) and 5-bromo-4-chloro-3-indolyl--galactopyranoside (X-Gal) were added when necessary to a final concentration of 0.004% (wt/vol).

Isotopes, enzymes, and chemicals. L-[³⁵S]methionine was purchased from Amersham Life Science, Inc. (Arlington Heights, Ill.), and Na₂³⁵SO₄ was obtained from ICN Biochemicals, Inc. (Costa Mesa, Calif.). Enzymes and reagents were from New England Biolabs, Inc. (Beverly, Mass.), Boehringer GmbH (Mannheim, Germany), Gibco BRL (Gaithersburg, Md.), Sigma Chemical Co. (St. Louis, Mo.), and Merck (Darmstadt, Germany).

Detection of dehalogenation activity. C1 was detected as described elsewhere (5, 74). I was measured by a modification of a previously described method (7). Modifications included the use of medium M9 or twice-diluted K1 for bacterial growth, 2.5 M citric acid buffer (obtained by mixing 66 ml of 2.5 M citric acid, 35 ml of 2 M NH₄OH, and 8 g of NH₄H₂PO₄ for a final pH of 4), and 0.3 to 0.45% (wt/vol) Oxone solution in H₂O. The assay allowed detection of iodide at concentrations as low as 5 µM in K1 medium and 25 µM in M9 medium.

Conjugation. Plasmids pOD22 and pOD33 were transferred into Pseudomonas strains KZ6R and B13 by triparental matings (17). Following a 6- to 8-h incubation at 30°C, the conjugational mix was washed off the filter (BA85; 0.45 µm pore size, Schleicher & Schuell, Dassel, Germany) (21) and transconjugants were selected by plating on K1 agar containing benzoate and tetracycline (B13) or Luria agar containing tetracycline and rifampin (KZ6R). Fifty to 100 randomly chosen colonies were replica plated onto K1 agar containing the desired substrate.

Electrotransformations. DNA transformation of E. coli and P. putida PB2440 cells was conducted by electroporation with an E. coli Gene Pulser (Bio-Rad Laboratories, Hercules, Calif.). Competent cells were prepared from early-log-phase cultures, grown in Luria broth, according to the protocol of Dower et al. (19), which is based on washing the cells three times with cold deionized H₂O (equal volume) and then concentrating them 200- to 400-fold in a 15% (wt/vol) glycerol solution. Cells were stored at 70°C in 50-µl aliquots.

DNA isolation and manipulation. Chromosomal DNA was isolated by the method of Marmur (46). Routine screening of strains for the presence of plasmid DNA, preparation of low-copy-number plasmid DNA, digestion of DNA with restriction endonucleases and exonuclease Bal31, treatment with alkaline phosphatase and the Klenow fragment of DNA polymerase, DNA ligation, and DNA electrophoresis in agarose gels were performed according to standard procedures (45). Recovery of DNA from agarose gels was performed according to the method of Dretzen et al. (20) or by using a DNA Cleanup kit (Promega, Madison, Wis.). Plasmid DNA for sequencing was purified by using a Promega Wizard 373 kit.

Construction of the gene library. Chromosomal DNA from strain 142 was partially digested with restriction endonuclease Sau3A. After separation in a 0.8% agarose gel, the fraction of fragments in the size range of 5 to 20 kb was recovered. This DNA was ligated to pSP329 DNA which had been linearized and dephosphorylated at its BamHI site. Following transformation of DH5F' cells, Tc^r white colonies were selected on Luria agar containing tetracycline, X-Gal, and IPTG. The recombinant colonies were replica plated onto M9 medium containing glucose as the growth substrate and in the presence of 2-CBA and 2,4-dCBA. After 2 days of incubation, colonies were analyzed for production of catechol by the p-toluidine test of Parke (54).

High-performance liquid chromatography analysis. Benzoates and their products were analyzed by isocratic reverse-phase chromatography on a 250- by 4-mm C₁₈ column (Hibar RT E; Merck). The eluent, a mixture of 0.1% H₃PO₄ and acetonitrile (at ratio of either 80/20 or 66/33, depending on the expected products), was applied at a flow rate of 1.5 ml/min. Compounds were detected by measuring UV absorbance at 230 nm. Products were identified by comparison of retention times with those of authentic standards.

Minicell assay. In the minicell assay, single colonies of fresh transformants of strain 925 were assayed as previously described (69, 74).

Induction study. Induction of degradative enzymes in strain 142 was assessed in labeling experiments using Na₂³⁵SO₄. Cells were serially transferred three times in twofold-diluted nutrient broth (Difco, Detroit, Mich.), harvested in early stationary phase, and washed in phosphate-buffered saline (PBS) (8.5 g of NaCl, 0.6 g of Na₂HPO₄ and 0.3 g of KH₂PO₄ per liter of H₂O; pH 7.0). Washed cells were resuspended in PBS, and after incubation at room temperature for 6 h (starvation), the cell suspension was amended with the substrate/inducer (20 µM) and isotope (0.4 µCi). Control cells were also suspended in PBS, but no inducer or isotope was added. After 68 h of incubation at room temperature, the cells were harvested, washed in PBS, and resuspended in deionized H₂O. Lysed samples containing 15 µg of total protein (44) were then separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (40). After being stained, the gels were exposed to Kodak X-OMAT AR film (Eastman Kodak, Rochester, N.Y.) for 5 days.

DNA sequencing and sequence analysis. Determination of the nucleotide sequence of the entire DNA insert from plasmid pOD33 was done by using deletional variants and internal primers synthesized at the Michigan State University Macromolecular Synthesis Facility. Each DNA strand was completely sequenced at least four times. Automated fluorescent sequencing was done at the Michigan State University DNA Sequencing Facility. Primers were designed by using the LASERGENE software package (DNASTAR Inc., Madison, Wis.). Primary sequence editing was performed with Sequencer V.3, (Gene Code Corporation, Madison, Wis.) and DNASTAR. The BLAST program (1), with BEAUTY postprocessing (77), was used for similarity searches of the nonredundant NCBI sequence database (National Center for Biotechnology Information, National Institutes of Health, Bethesda, Md.). An initial multiple alignment was designed by using the CLUSTAL W Program, version 1.7 (73), offered by BCM Search Launcher (Human Genome Center, Baylor College of Medicine, Houston, Tex.). This was used as a starting alignment to create a hidden Markov model-based alignment with the SAM-T98 program (38). Unrooted Fitch-Margolash dendrograms were derived from the alignment by using the SEQBOOT, PROTDIST, FITCH and CONSENSE programs of the PHYLIP package (27).

Nucleotide sequence accession number. The nucleotide sequence of the ohb DNA region has been deposited with GenBank (accession no. AF121970).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation and expression of the ohb genes in E. coli cells. Among 3,700 recombinant clones of the library of total DNA from P. aeruginosa 142, two positive colonies were identified by the p-toluidine method. Plasmids from these clones, designated as pOD22 and pOD33, contained overlapping DNA inserts of about 8 and 6 kb in size, respectively (Fig. 2). Strain pOD22 was characterized by 90% segregational instability, whereas plasmid pOD33 was stable and therefore was the plasmid primarily used in further studies. Strains DH5F'(pOD22) and DH5F'(pOD33) stoichiometrically converted 2-CBA into catechol and 2,4-dCBA and 2,5-dCBA into 4-chlorocatechol, when grown in M9 medium in the presence of different CBAs (Table 1), and released iodide into the medium when grown in the presence of 2-iodobenzoate (2-IBA). Release of chloride into K1 medium also was detected. We concluded that these recombinant plasmids contained the ortho-halobenzoate 1,2-dioxygenase genes (ohb) of strain 142 and that these genes were expressed in E. coli.

View larger version (29K): [in this window] [in a new window]   FIG. 2.   Physical map and organization of the ohb DNA region. Physical maps of the 6.052-kb DNA insert of plasmid pOD33 and its deletion variants are presented. The overlapping region of the DNA insert of plasmid pOD22 is shown at the top. Plasmids pK33A16, pK33H7, and pK33A20 were constructed with the BlueScript vector; for the other variants, vector pSP329 was used. Phenotypes are indicated to the right of the plasmid names. The DNA insert of plasmid pOD33 has been sequenced by using the deletion variants and internal primers. The locations and orientations of ORF1 (top), ORF2 (ohbR), ORF3 (ohbA), ORF4 (ohbB), ORF5 (ohbC), and ORF6 (tnpA) are shown in the lower part of the figure, along with the calculated molecular masses of the corresponding polypeptides. Restriction sites are shown, with their locations indicated in parentheses.

View this table: [in this window] [in a new window]   TABLE 1.   Conversion of CBAs by recombinant E. coli DH5F'(pOD33) and DH5F'(pOD22)

Physical mapping of the ohb DNA region. Attempts to reclone the DNA inserts from plasmids pOD22 and pOD33 into multicopy E. coli vector plasmids pUC19, pGEM, and BlueScript failed due to the instability of the resulting constructs in both the DH5F' and JM109 strains. Further cloning experiments to locate and isolate the functional ohb region were done with vector pSP329. Some nonfunctional deletion subclones were produced with vectors BlueScript SK(+) and KS(+). The overlapping region of the DNA inserts in plasmids pOD22 and pOD33 (Fig. 2, positions 1.1 to 6 kb on the map of pOD33) was approximately 4.9 kb in size and presumably contained all of the information necessary to control halobenzoate oxidation in E. coli cells. This was confirmed by subcloning restriction fragments of the DNA insert from plasmid pOD33. The 5.3-kb EcoRI-HindIII and 4.75-kb KpnI-HindIII fragments conferred on the host the ability to oxidize 2-CBA, as determined with p-toluidine (Fig. 2, plasmids p33E10 and p33K21, respectively), whereas the 3.35-kb SalI fragment (Fig. 2, plasmid p33G1) proved to be nonfunctional. Exonuclease Bal31-derived deletions allowed the determination of the location of the functionally active ohb gene region within the 3.7-kb DNA fragment (Fig. 2, plasmid pE43).

Expression of the ohb genes in Pseudomonas cells. The recombinant plasmids pOD22 and pOD33 were introduced into Pseudomonas sp. strains PB2440, KZ6R, and B13 by either conjugation or transformation. Each of these strains is capable of degrading catechol via the ortho-cleavage pathway. Strain B13 additionally harbors the modified ortho pathway for oxidation of chlorocatechol (18). Recombinants were plated onto solidified K1 medium containing 2-CBA or 2,4-dCBA as the sole carbon source and incubated for up to 12 weeks with no specific growth being observed. However, cells of PB2440(pOD33) and KZ6R(pOD33) released small amounts of iodide (10 to 20 µM) when grown in the presence of 2-IBA (200 µM) in K1 mineral medium containing glucose or acetate; this suggested that the expression of the ohb genes was insufficient to allow growth on the halobenzoate.

View larger version (16K): [in this window] [in a new window]   FIG. 3.   Growth of the recombinant P. putida PB2440(pE43) on 2-CBA. Batch cultures were grown aerobically in mineral medium K1 with 2-CBA as the sole source of carbon at concentrations of 1, 1.5, and 2 mM. Aliquots were taken at certain time points to measure optical densities (OD) at 600 nm (solid lines) and 2-CBA concentrations (broken lines).

View larger version (17K): [in this window] [in a new window]   FIG. 4.   Oxidation of 2-CBA, 2,4-dCBA, 2,5-dCBA, and 2,6-dCBA by resting cells of the recombinant strain P. putida PB2440(pE43). 2-CBA grown cells were harvested and then resuspended in fresh K1 medium, which was amended with different CBAs. The reaction mixtures were incubated at room temperature with aeration. Triplicate samples were taken at certain time points to measure the concentrations of CBA and (chloro)catechol.

The ohb-encoded proteins. As shown in Fig. 5, minicells containing plasmid pOD33 (lane 2), p33K21 (lane 3), pOD22 (lane 4), or pE43 (lane 5) synthesized three new polypeptides with apparent molecular masses of 48, 47, and 22 kDa. No specific bands in addition to those featured for the vector pSP329 (lane 1) were found for the deletion variants (data not shown) except for pE33-1 (lane 6). In the last case, both bands (48 and 47 kDa) were replaced by a band of 24 kDa, possibly a remnant of one of the former polypeptides. These polypeptides could be products of overlapping genes, or they could have resulted from partial proteolysis. Sequencing results (see below) confirmed the former suggestion.

View larger version (65K): [in this window] [in a new window]   FIG. 5.   Synthesis of plasmid-encoded proteins in E. coli minicells. Shown is an autoradiogram of [35S]methionine-labeled proteins synthesized in strain 925 minicells containing plasmids. Lanes: 1, pSP329; 2, pOD33; 3, p33K21; 4, pOD22; 5, pE43; 6, pE33-1. Samples were prepared according to the method of Platt (56), and 20-µl aliquots of each were loaded onto SDS-12.5% polyacrylamide gels. Molecular mass standards (SDS-7; Sigma) were as follows: bovine serum albumin (66.2 kDa), hen egg white ovalbumin (45 kDa), glyceraldehyde-3-phosphate dehydrogenase (36 kDa), carbonic anhydrase (29 kDa), trypsinogen (24 kDa), trypsin inhibitor (20.1 kDa), and -lactalbumin (14.2 kDa).

View larger version (105K): [in this window] [in a new window]   FIG. 6.   Induction of protein synthesis in P. aeruginosa 142 by (chloro)benzoates. Shown is an autoradiogram of 35S-labeled proteins synthesized in strain 142 upon induction by potential substrates/inducers. Lanes: 1, no substrate; 2, 2,4-dCBA; 3, 2,5-dCBA; 4, 2-CBA; 5, benzoate; 6, 4-chlorocatechol; 7, catechol; 8, succinate. Each sample loaded onto the SDS-10% polyacrylamide gel contained 37.5 µg of total protein. Molecular mass standards (Bio-Rad) were as follows: rabbit muscle phosphorylase B (96 kDa), bovine serum albumin (66.2 kDa), hen egg white ovalbumin (45 kDa), bovine carbonic anhydrase (31 kDa), and soybean trypsin inhibitor (21.5 kDa).

Determination and analysis of the nucleotide sequence of the ohb DNA region. The nucleotide sequence of the 6,052-bp DNA insert of plasmid pOD33 contains six open reading frames (ORFs) in both orientations (Fig. 2 and 7). The ORF designated ohbA has two possible translation start sites, at positions 2636 and 2783. However, the first ATG codon is not preceded by a putative Shine-Dalgarno site, whereas an extended 5'-AAGAGGAGGGAGAG-3' sequence is located upstream of the ATG codon, at position 2783. The deduced molecular masses for the ohbA, ohbB, and ohbC gene products were in accord with those of polypeptides synthesized in E. coli (Fig. 5) and in parental strain 142 (Fig. 6). The ohbB and ohbC genes are located within nearly the same DNA locus but are transcribed in opposite directions, in agreement with data for subclone pE33-1 (Fig. 5).

View larger version (164K): [in this window] [in a new window]   FIG. 7.   Nucleotide sequence of the 6,052-bp ohb DNA region of P. aeruginosa 142. Shown are the six ORFs along with their deduced amino acid sequences. Single-letter symbols for amino acids are aligned with the second nucleotide of each codon. Conserved amino acid residues in sequences of OhbA, OhbB, OhbC, and OhbR are shown in boldface (for an explanation, see the text). Possible Shine-Dalgarno sites and a potential transcription terminator for the ohbB are underlined. Also shown are locations of some restriction sites (indicated above the sequence) and the IS1396-like sequence and bordering imperfect IRs (<, below sequence; >, above sequence).

View larger version (20K): [in this window] [in a new window]   FIG. 8.   Unrooted Fitch-Margolash dendrograms of large -ISP subunits of aromatic oxygenases and OhbB (A) and of small -ISP subunits and OhbA (B). The horizontal axes are scaled in terms of expected numbers of amino acid substitutions per site. The numbers on branches refer to the percent confidence estimated by bootstrap analysis (100 replications). The proteins are labeled by trivial abbreviations. (A) In descending order of overall sequence similarity, the proteins were the putative BphA1c and BphA1d from S. aromaticivorans F199 (GenBank accession no. AF079317), NagG of the salicylate-5-hydroxylase from Pseudomonas sp. strain U2 (31), putative OhbB from P. aeruginosa JB2 (GenBank accession no. AF087482), the hypothetical ORF2 of the aromatic dioxygenase from Burkholderia sp. strain DNT (70), the putative BphA1e from strain F199 (GenBank accession no. AF079317), the hypothetical ORFG1 protein of the aromatic dioxygenase from Sphingomonas sp. strain RW1 (2), the putative BphA1 from Pseudomonas sp. strain B4 (GenBank accession no. U95054), BphA of the biphenyl dioxygenase from Burkholderia sp. strain LB400 (25), TcbAa of the chlorobenzene dioxygenase from Pseudomonas sp. strain P51 (75), AntA of the anthranilate dioxygenase from Acinetobacter sp. strain ADP1 (11), and CbdC of the 2-halobenzoate 1,2-dioxygenase from B. cepacia 2CBS (33). (B) In descending order of overall sequence similarity, the proteins were the putative BphA2c and BphA2d from strain F199 (GenBank accession no. AF079317), the putative OhbC from strain JB2 (GenBank accession no. AF087482), NagH from strain U2 (31), the putative BphA2e from strain F199 (GenBank accession no. AF079317), the hypothetical ORFG6 from strain RW1 (2), CmtAc of the p-cumate dioxygenase from P. putida F1 (22), CarAb of the carbazole dioxygenase from Sphingomonas sp. strain CB3 (66), BenB of the benzoate 1,2-dioxygenase from strain ADP1 (formerly A. calcoaceticus) (51), AntB from strain ADP1 (11), the putative BphA2 from Rhodococcus erythropolis TA421 (DDBJ accession no. D88021), an unknown MtmX from Streptomyces argillaceus (43), and DxnA2 of the dioxin dioxygenase from strain RW1 (2).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Cloning and characterization of the ohb genes confirmed an oxygenolytic mechanism of ortho dehalogenation of chlorobenzoates in strain 142 that was implied previously (62, 65). However, our study showed that the ortho-halobenzoate 1,2-dioxygenase activity in recombinant E. coli and Pseudomonas strains is manifested by the ohbAB-encoded terminal oxygenase ISP_OHB alone, implicating utilization of electron transfer components provided by the recipient cell. Given the prolonged incubation required for initial growth of the PB2440(pE43) transformants on 2-CBA, it is possible that an unspecified host mutation, perhaps leading to constitutive or higher-level production of electron transfer components, complements ISP_OHB. The apparent inability of CBAs to induce synthesis of electron transfer components in strain 142 indirectly supports this inference. On the other hand, 2-CBA toxicity also appears to affect growth and rates of degradation in both recombinant PB2440(pE43) (present study) and parental strain 142 (61).

The optimum concentration of 2-CBA for growth was around 2.5 to 3.0 mM for parental strain 142 (61), compared to 1.5 mM for recombinant PB2440(pE43); this could arguably be due to cognate ferredoxin or reductase components in the parent. However, in later experiments we have expressed the ohb genes in biphenyl-degrading C. testosteroni VP44, and the resulting recombinant, VP44(pE43), readily mineralized 10 mM each 2CBA and 2-chlorobiphenyl (36).

The interchangeability of electron transfer components between evolutionarily related primary dioxygenases was described previously (12, 31, 81). Utilization of a heterologous host's ferredoxin oxidoreductase by aromatic oxygenases was also reported earlier (67, 80, 81). The expression of Pseudomonas mendocina KR1 toluene-4-monooxygenase activity in E. coli and other Pseudomonas strains did not require the reductase component, TmoF, although the latter enhanced this activity by at least twofold (81). Similarly, Pseudomonas sp. strain U2 naphthalene dioxygenase ISP_NAG genes nagAc and nagAd along with ferredoxin gene nagAb allowed the oxidation of indole to indigo in E. coli, while reductase gene nagAa was apparently not required (31).

The ohbAB-encoded ISP_OHB of the ortho-halobenzoate 1,2-dioxygenase forms a cluster with several other terminal oxygenases that are distant from the rest of the members of the family of primary oxygenases (Fig. 8). In this cluster, the ISP_OHB from strain 142 and the putative ISP_BPH from strain F199 are more deeply branching and are outside a tight internal cluster formed by ISP_OHB from strain JB2 (GenBank accession no. AF087482), NagG from strain U2 (31), the ORF2 protein from strain DNT (70), NahG from strain GZ42 (84), and the truncated ORF2 protein from strain JS42 (53). Only the ISP_OHB from strain 142 (present study) and the ISP_SAL from strain U2 were assigned functions experimentally. Characteristically, these structurally related ISP_OHBs and the ISP_SAL were capable of promiscuous use of an energy supply system provided by the host bacterium. The ISP_SAL genes nagG and nagH alone enabled E. coli to convert salicylate to gentisate, albeit with a low level of activity. Addition of ferredoxin gene nagAb increased this activity, while reductase gene nagAa again was not required for salicylate oxidation in E. coli (31). Structural evidence indicated that a loose association with electron transfer components might be characteristic of this cluster. Indeed, the ohbAB from strain 142 and bphA1dA2d from strain F199 (GenBank accession no. AF079317) are not associated with any electron component genes, while another deeply branching ISP_BPH (bphA2cA1c) from the same strain, F199, is preceded by a Rieske 2Fe-2S-type ferredoxin gene, bphA3, in what could be viewed as an intermediate stage in gene assembly. Evolutionary relationships between tightly related genes from strains DNT, U2, and GZ42 were previously discussed (31). In all three strains was found the same dual oxygenase operon arrangement, reductase-ISP_X-ferredoxin-ISP_Y, in which two ISPs share the same set of electron transfer components. Fuenmayor et al. (31) concluded that the ISP_SAL genes nagG and nagH (ORF2 and ORFX in DNT; nahG and nahH in GZ42) were acquired as an insert in a preexisting naphthalene (nitrotoluene) degradation pathway. However, functional evidence found by Fuenmayor et al. (31) and structural evidence presented in the form of the recently released sequence of an ohb operon from strain JB2 (GenBank accession no. AF087482) suggest otherwise. In this JB2 ohb operon, reductase ohbA, ohbBC (ISP_OHB), and ferredoxin ohbD genes are in the same gene order and are highly similar to iso-functional genes from the nag, nah, and dnt operons. It appears that these dual oxygenase-containing operons might have been assembled from preexisting reductase-ISP_X (nahGH, nagGH, or ORF2-ORFX)-ferredoxin and independently evolved ISP_Y (nahAcAd, nagAcA, or dntAcAd) DNA regions. The unaccompanied ORFG1 protein from strain RW1 (2) and the product of the bphA1e-bphA2e genes (associated with a reductase gene [bphA4] from strain F199) (GenBank accession no. AF079317) showed pairwise levels of similarity to the OhbB from strain 142 in the range of 30%, intermediate between the similarity values seen in the OhbB cluster and the levels of similarity between OhbB and the BphA LB400(25)-BphA1 B4 (GenBank accession no. U95054)-TcbAa P51 (75) cluster. Interestingly, the three ISP_BPHs from strain F199, similar to the ISP_OHBs, are among six sets of putative aromatic oxygenase ISP genes dispersed within a 184-kb sequence of catabolic plasmid pNL1; however, only two ferredoxin genes (bphA3) and two ferredoxin oxidoreductase genes (bphA4) were annotated for the entire sequence. Overall, our sequence analysis suggested coevolution of the -ISP and -ISP genes, whereas electron transfer components might not be obligatory for these oxygenases and could have been independently acquired via horizontal gene transfer.

Although our analysis was limited to a few sequences best matching that of ohbAB, in a larger picture, recent findings by other investigators also point to independent evolution of individual components of primary oxygenases. In B. cepacia DBO1, the terminal oxygenase gene (ophA2) was found to lie together with the dihydrodiol dehydrogenase gene (ophB) while reductase gene ophA1 was found 7 kb away (13). Genes dxnA1 and dxnA2, encoding dioxin dioxygenase ISP_DXN, were found on a DNA fragment that did not code for electron transport components in Sphingomonas sp. strain RW1, while the cognate ferredoxin (fdx1) and reductase (redA2) genes were isolated from two other, separate DNA regions (2). Similar to S. aromaticivorans F199, several other putative ISP genes were found dispersed on the RW1 genome (2). Those included ORFG1, which was found to be similar to OhbB (Fig. 8A); the pair ORFG5 (-ISP) and ORFG6 (-ISP), which demonstrated recognizable similarity to OhbB and OhbA (Fig. 8B); and ORFG4 (a putative -ISP), which alone controlled conversion of indole to indigo in E. coli (2).

Only six residues were conserved among the small subunits of primary dioxygenases (Fig. 7), reflecting the highly diverse nature of -ISP subunits, which might be involved in determining substrate specificity, perhaps along with the large -ISP subunit (12). Whereas small subunits of the ISPs from strains U2 and DNT showed (by BLAST search) recognizable similarity only to the members of the identified cluster, OhbA showed recognizable lengthwise similarity to 14 -ISPs for a variety of substrates, including mono- and polyaromatic compounds (Fig. 8B), possibly suggesting a broader substrate range for ISP_OHB. The ohb genes isolated in this study dehalogenated a variety of halobenzoates, and 2-CBA-grown cells of strain 142 oxidized 4- and 5-chloroanthranilate (62) as well as 2-trifluoromethylbenzoate and anthranilate (65). Comparative studies of the substrate ranges of the terminal oxygenases from the new cluster perhaps will be helpful in gaining a further understanding of the evolution of new degradative activities. The previous hypothesis that ortho-halobenzoate 1,2-dioxygenases and anthranilate 1,2-dioxygenases from different organisms should be closely related, if not the same (65), was not confirmed. Sequence analysis of the ISP_OHBs from strains 142 and JB2 and of the ISP_CBD from strain 2CBS (33) (Fig. 8A) did not reveal a specific relationship between these iso-functional genes. The large evolutionary distance separating the ohb and cbd genes suggests separate origins of the dehalogenation activities.

The strain 142 ohb genes are embedded in a transposon-like context, implying the likely involvement of horizontal gene transfer in the evolution of the ortho-halobenzoate 1,2-dioxygenase activity (Fig. 2 and 7). Insertion of the IS1396-like sequence adjacent to a potential transcription terminator (56) of ohbB is consistent with the acquisition of structural genes by IS elements via targeting their terminator or promoter regions (15, 16, 39, 52). Testing for transmissibility of the ohb DNA region was hindered by the failure to sustain the DNA insert of pOD33 on E. coli-specific vectors. However, the transposable nature of the dehalogenation activity in strain JB2 was previously demonstrated (55). IS1396 is a member of an insertion element family that is broadly but sparsely dispersed among genomes of gram-positive bacteria, cyanobacteria, and broad-host-range plasmids from gram-negative bacteria (14, 16, 39, 76). This, and the similarity of the putative Top protein to the plasmid RP4 topoisomerase (TraE) (42), could imply a plasmid origin for the ohb DNA region; however, plasmids have not been reported in strain 142 (62), and finding only two ohb clones among 3,700 recombinants was consistent with their being chromosome-borne genes. Comparison of nucleotide sequences from halobenzoate-degrading strains 142 and JB2 revealed that the IS1396-like sequence from strain 142 at positions 4673 to 5533 (Fig. 7) was 94% similar to the sequence at positions 8302 to 9161 upstream of the putative ohbABCD genes of strain JB2 (GenBank accession no. AF087482). In the latter, the remnant of IS1396-like sequence is separated from ohbB (an analog of the ohbB gene in strain 142) by an IS21-like sequence containing the tnpAB genes, by a LysR-type putative ohbR, and by the reductase ohbA gene. Although the dehalogenation genes from strain 142 and JB2 are clearly divergent (Fig. 8), it appears that the same IS1396-like insertion element might have been involved in assembly of the ortho-chlorobenzoate pathway in both strains.