INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The pulping of wood to extract fibers used to make paper produces toxic wastewaters (23). The majority of the acute toxicity can be attributed to resin acids found in these wastewaters (18). Resin acids, a class of diterpenoids found in wood extractives, can be grouped into abietanes and pimeranes. Abietane-type acids have an isopropyl chain at the C-13 carbon atom, while pimerane-type acids have methyl and vinyl substituents at this position. Although these compounds are abundant in nature, they are problematic in the pulp and paper industry because of their unnaturally high concentrations in wastewaters. Therefore, resin acids must be removed from wastewater prior to its discharge to the environment. Biological treatment using aerated lagoons is the most common method used for detoxifying the wastewater (19). Several bacteria isolated from such biotreatment systems have been reported to use resin acids as growth substrates, and this catabolic activity appears to be widespread (22, 25).

Pseudomonas abietaniphila BKME-9, a bacterium isolated from a pulp and paper wastewater treatment system, is able to use dehydroabietic acid, an abietane-type resin acid, as a sole source of carbon and reductant (4). A novel three-component aromatic-ring-hydroxylating dioxygenase from this strain has been cloned and expressed (21). In strain BKME-9, dehydroabietic acid is first oxidized by an unidentified enzyme at the C-7 position to yield 7-oxodehydroabietic acid (Fig. 1) (6, 21). The next step in the catabolic pathway involves the DitA aromatic-ring-hydroxylating dioxygenase. This dioxygenase consists of a putative [4Fe-4S] or [3Fe-4S]-type ferredoxin (DitA3) encoded by a gene located 9.2 kb from genes encoding the  and  subunits of the catalytic oxygenase component (DitA1A2) (Fig. 2). DitA activity was reconstituted in Escherichia coli and was found to dihydroxylate 7-oxodehydroabietic acid to 7-oxo-11,12-dihydroxy-8,13-abietadien acid. The following enzymatic steps in the dehydroabietic acid degradation pathway have not been characterized in P. abietaniphila BKME-9. However, Biellmann et al. (6) purified the 7-oxo-11,12-diol metabolite of dehydroabietic acid from a Pseudomonas sp. culture supplemented with the metabolic inhibitor/chelator ,'-dipyridyl. In a similar study, Flavobacterium resinovorum oxidized dehydroabietic acid at C-7 and C-3, followed by decarboxylation at C-4 prior to aromatic-ring hydroxylation (5).

View larger version (10K): [in this window] [in a new window]   FIG. 1.   Proposed pathway for abietanic diterpenoid degradation by P. abietaniphila BKME-9. Chemical designations: I, abietic acid; II, palustric acid; III, dehydroabietic acid; IV, 7-hydroxydehydroabietic acid; V, 7-oxodehydroabietic acid; VI, 7-oxopalustric acid; VII, 7-oxo-11,12-dihydroxy-8,13-abietadien acid; VIII, 7-oxo-11,12-dihydroxydehydroabietic acid. Dashed arrows represent hypothetical steps in the pathway.

In a previous study, we characterized the DitA dioxygenase encoded by a DNA fragment cloned from strain BKME-9 (21). A transposon (Tn5) insertion in the gene encoding the  subunit of the dioxygenase (ditA1) resulted in a mutant that had lost the capacity to grow on dehydroabietic acid as well as on the nonaromatic diterpenoid abietic acid. In the course of isolating the gene encoding the ferredoxin component of the dioxygenase (ditA3), we cloned and sequenced several open reading frames (ORFs) adjacent to the ditA1A2 and ditA3 genes. In this study, we further elucidated abietane degradation by BKME-9 and tested the hypothesis that this strain uses a convergent biodegradation pathway that aromatizes nonaromatic abietanes to dehydroabietic acid or 7-oxodehydroabietic acid prior to aromatic-ring degradation.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and culture conditions. The bacterial strains and plasmids used in this study are listed in Table 1. E. coli was cultured on Luria-Bertani medium, and P. abietaniphila BKME-9 was cultured on tryptic soy broth, or mineral medium, as previously described (24). Mutants of strain BKME-9 were cultured with 4 µg of gentamicin and/or 30 µg of kanamycin per ml, and P. abietaniphila strains harboring derivatives of pUCP26 were cultured with 2 µg of tetracycline per ml.

DNA manipulations and sequence data analysis. The Tn5 mutagenesis, the construction and screening of the cosmid genomic DNA library from P. abietaniphila BKME-9, the subcloning, and the DNA sequencing methodologies were previously described (21). Figure 2 is a schematic representation of the subcloning strategy. Nucleotide sequence analysis was done with Clone Manager for Windows (version 4.01), and PCR primers were designed with Primer Designer (version 2.0). ORFs were analyzed for similarity to GenBank database entries by using the BLASTX and BLASTP programs (1) available on the National Center for Biotechnology Information server via the Internet. Deduced protein sequences were aligned with ClustalX. Searches for PROSITE protein signature consensus sequences and prediction of transmembrane regions were done with the ProScan and TMpred programs available on the Internet server of the Swiss Institute for Experimental Cancer Research.

View larger version (29K): [in this window] [in a new window]   FIG. 2.   Physical map of the dit gene cluster indicating fragments cloned from the P. abietaniphila BKME-9 gene library cosmid pLC12. Solid rectangles, fragments of DNA subcloned for homologous recombination. Arrows indicate the locations of the various insertions, with the endonuclease and base pair positions on the kilobase scale shown. ND, location of nested deletion of clones used in subcloning the DNA fragment. FPCR, fragment of DNA cloned by PCR.

Insertional inactivation and construction of transcriptional fusions of dit genes. The pEX100T gene replacement vector containing the sacB counterselectable marker was used for the insertional inactivation of dit genes (34). Overhanging ends of DNA fragments required for insertional inactivation were removed with Klenow polymerase or mung bean exonuclease and subcloned into the unique SmaI site of pEX100T. The locations of the fragments and the restriction enzymes used to subclone the genes of interest into pEX100T are shown in Fig. 2. To create xylE transcriptional insertions/fusions, the xylE-Gm^r cassette isolated from pX1918G (without transcriptional terminator) or pX1918GT (with transcriptional terminator) were inserted into the genes in unique restriction endonuclease sites identified in Fig. 2. The cassette containing no terminator was inserted only in those genes suspected of being in a polycistronic transcript (i.e., ditA3, ditB, ditD, and ditH), in order to minimize downstream polar effects. The Km^r cassette isolated from a SmaI digest of pUC4-KIXX was used to disrupt the putative regulatory gene ditR, to allow double mutations in strains carrying the xylE-Gm^r fusions. The putative extradiol cleavage dioxygenase gene ditC, was inactivated by insertion of the Gm^r gene (without xylE) isolated from a BamHI digest of pUCGm, to avoid possible complementation of the ditC mutation by catechol 2,3-dioxygenase (C23O). As several attempts to clone ditF using restriction enzymes failed, it was PCR amplified from plasmid pVM5 using primers SCP-1 (5'-TCGAGGATGTCTGGCTG-3') and SCP-2 (5'-GCTGAGCAAGGTGCTGT-3') and was cloned into the SmaI site of pEX100T. Homologous recombination of the mutated alleles into strain BKME-9 was accomplished by conjugation followed by a two-step selection method, as previously described (21). The gene replacements were confirmed by colony PCR with 17-mer primers at an annealing temperature of 58°C (47).

Phenotypic characterization of dit mutant strains. Mutants of strain BKME-9 were characterized for their ability to grow on 0.1 g of either dehydroabietic, abietic, palustric, or 7-oxodehydroabietic acid/liter or on 1 g of pyruvate/liter in a mineral medium. The medium was inoculated (0.1%) with a culture grown overnight on pyruvate and monitored for growth for 3 days. The turbidity caused by the precipitated resin acids in the media made accurate measurement of growth by optical density difficult; therefore, growth was determined by microscopic examinations with comparisons to positive (the wild-type strain, BKME-9) and negative (no substrate) controls. Mutant strains were also tested for the ability to oxidize dehydroabietic acid and accumulate pathway intermediates in cell suspension assays, which were performed as previously described (21). Cell suspensions were monitored by UV-visible light absorption spectroscopy and by gas chromatography as previously described (24), except that samples were acidified with 1 drop of 1 N HCl prior to ethyl acetate extraction in order to improve the recovery of acidic metabolites.

C23O assays. For C23O activity assays, strain BKME-9 was grown on pyruvate to mid-log phase (optical density at 610 nm [OD₆₁₀], 0.6 to 0.7). Cultures were chilled on ice for 15 min, harvested, and washed twice in 10 mM KPO₄ buffer (pH 7.5) at 4°C. Antibiotics were excluded from the mineral medium, as they reduced the growth rate of the cultures and affected the specific C23O activity. The washed cells were suspended in 10 mM KPO₄ buffer (pH 7.5) and adjusted to a final OD₆₁₀ of 0.6 before induction with 0.5 mM dehydroabietic, abietic, 7-oxodehydroabietic, or isopimaric acid or pyruvate or with 0.1 mM 12,14-dichlorodehydroabietic acid, biphenyl, isopropylbenzene, or phenanthrene. These concentrations were far above the compounds' aqueous solubilities, with the exception of the concentrations of pyruvate and isopropylbenzene. The cell suspensions were incubated with the compounds for 8 h at 30°C on a rotary shaker at 150 rpm, after which C23O activity was measured. All C23O induction experiments were replicated with identical results. Triplicate enzyme assays were performed on whole cells in 1 ml of 10 mM KPO₄ buffer (pH 7.5). C23O activity was assayed spectrophotometrically at 30°C by measuring the formation of 2-hydroxymuconic semialdehyde at 375 nm ( = 44 mM¹ cm¹) for 1 min. Protein concentrations of cell suspensions were determined by using the micro-bicinchoninic acid protein assay kit (Sigma) and bovine serum albumin as the standard (39).

Nucleotide sequence accession number. The nucleotide sequence reported in this study has been submitted to GenBank under accession no. AF119621.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Genetic organization of the dit gene cluster and characterization of mutants. The mutagenesis of P. abietaniphila BKME-9 with the Tn5 transposon and the characterization of the diterpenoid aromatic-ring-hydroxylating dioxygenase (21) produced the cosmid clone pLC12, which contains the dit gene cluster described in this study. Subcloning of the pLC12 cosmid allowed the sequencing of a 16.7-kb DNA fragment containing 13 complete ORFs and 1 partial ORF (Fig. 2). This fragment was subcloned as two EcoRI fragments of 5.8 and 9.8 kb (pVM1 and pVM2), and DNA from pVM2 was further subcloned as 3.6-kb EcoRI-SmaI (pVM4) and 4.3-kb SmaI-SmaI (pVM5) fragments (Fig. 2). The complete sequence of ORF2 was obtained by sequencing directly from cosmid pLC12 with custom primers designed from known sequence.

Evidence of a convergent pathway for abietane degradation. Strains with mutations in two of the genes coding for the aromatic-ring-hydroxylating dioxygenase (ditA1 or ditA3) accumulated the pathway intermediate 7-oxodehydroabietic acid in cell suspension assays with dehydroabietic acid as the substrate (Fig. 3A) (21). Since the disruption of genes encoding this enzyme also resulted in the loss of growth on the nonaromatic substrate, abietic acid (21), we hypothesized that strain BKME-9 aromatizes abietic acid to dehydroabietic acid prior to aromatic-ring attack. To test this hypothesis, the ditA1::Tn5 mutant strain BKME-941 was incubated in cell suspension assays in the presence of abietic and palustric acids, two nonaromatic abietane diterpenoids. Gas chromatography (GC) analyses showed that 7-oxodehydroabietic acid was produced from both substrates and that dehydroabietic acid was produced from palustric acid (Fig. 3B and C). Although it is likely that dehydroabietic acid is an intermediate in the conversion of abietic acid to 7-oxodehydroabietic acid, dehydroabietic acid was not observed when abietic acid was used as the substrate for the mutant strain BKME-941 (Fig. 3B). Two unknown compounds, one from abietic acid and the other from palustric acid, were also observed. Although it was not identified with certainty, we suspect that unknown II (Fig. 3C) is the 7-oxo derivative of palustric acid. GC-mass spectrometry (MS) analysis of the culture supernatant showed that the mass spectrum of the methyl ester derivative of unknown II had a fragmentation pattern similar to that of 7-oxodehydroabietic acid with an additional mass of m/z 2 for the molecular ion at m/z 330 (19% relative intensity) and a base peak of m/z 255 (7-oxodehydroabietic acid has a molecular ion of m/z 328 and a base peak of m/z 253). This compound was therefore tentatively identified as 7-oxopalustric acid (Fig. 1, compound VI). The identity of unknown I could not be determined from its mass spectrum alone. Approximately 90% of the carbon was recovered as substrate or identifiable metabolites after 24 h of incubation with dehydroabietic or palustric acid. The carbon mass balance was poor for abietic acid, with approximately 40% recovery. However, abiotic controls also showed a loss of 22% of the abietic acid carbon, indicating that this compound was somewhat unstable under the assay conditions. The transformation of abietic and palustric acids to dehydroabietic acid or isomerization of palustric to abietic acid did not occur in the abiotic incubations of the substrates in mineral medium, despite the fact that these reactions have been previously reported to occur spontaneously under some conditions (40). The inability of the strain BKME-9 mutants to grow on nonaromatic compounds (Table 3) and the metabolites identified in cell suspension experiments (Fig. 3) demonstrated that the aromatization of abietic and palustric acids to dehydroabietic and/or 7-oxodehydroabietic acid is essential to this biodegradation pathway for abietane-type diterpenoids.

View larger version (21K): [in this window] [in a new window]   FIG. 3.   GC-FID analysis of the dehydroabietic (A), abietic (B), and palustric (C) acid biotransformation products of P. abietaniphila BKME-941 (ditA1::Tn5) at 0 and 24 h of incubation.

Analysis of ditA1 and ditA3 gene expression and inducer specificity. The expression of the genes encoding the  subunit and ferredoxin components of the diterpenoid dioxygenase was analyzed by measuring C23O activities of ditA1::xylE (strain BKME-92) and ditA3::xylE (strain BKME-91) transcriptional fusion strains. Cultures of the wild-type strain, BKME-9, grown on pyruvate and subsequently induced with each of five diterpenoids showed no endogenous C23O activity, which demonstrated that xylE was an adequate reporter gene for induction studies of this pathway. Uninduced cultures of the reporter strains BKME-91 and BKME-92 showed basal levels of C23O expression (Fig. 4), indicating that the expression of ditA1 and ditA3 was leaky. The specificity of induction of these genes was investigated by using dehydroabietic, abietic, and 7-oxodehydroabietic acids, which are growth substrates for strain BKME-9, as well as 12,14-dichlorodehydroabietic and isopimaric acids, which are not metabolized by BKME-9 (25). Surprisingly, all five diterpenoids induced ditA1 and ditA3 expression (Fig. 4). Since we have demonstrated that abietic and dehydroabietic acids are transformed to 7-oxodehydroabietic acid by ditA1 and ditA3 mutants (Fig. 3), it is plausible that only 7-oxodehydroabietic acid, the substrate for the dioxygenase, is the inducer and not abietic or dehydroabietic acids. However, considering the high C23O levels observed with the nonmetabolizable diterpenoids 12,14-dichlorodehydroabietic and isopimaric acids, we believe that ditA1 and ditA3 expression was directly inducible by abietic and dehydroabietic acids. Induction of the aromatic-ring-hydroxylating dioxygenase appears to require diterpenoids, as the aromatic compounds biphenyl, isopropylbenzene, and phenanthrene, which are not growth substrates for strain BKME-9 (25), did not induce ditA1 or ditA3 expression (Fig. 4).

View larger version (46K): [in this window] [in a new window]   FIG. 4.   Expression of ditA1 and ditA3 in response to various diterpenoids and aromatic compounds. Compounds were added at 0.5 mM (dehydroabietic acid, abietic acid, 7-oxodehydroabietic acid, and isopimaric acid) and at 0.1 mM (dichlorodehydroabietic acid, biphenyl, isopropylbenzene and phenanthrene). Activity values are means of three enzyme assays ± standard deviations.

Regulation of ditA3 expression by DitR. Located between ditA3 and ditA1, 4.4 kb downstream of the former gene and 3.8 kb upstream of the latter gene, an ORF was identified with similarity to regulatory proteins (Table 2). This ORF, designated ditR, is located 79 nucleotides downstream of ditE. Comparison of the deduced amino acid sequence of DitR to the GenBank database entries revealed low sequence identity to IclR-type transcription regulators (41) (Table 2). In addition to the sequence similarity, we also identified a potential helix-turn-helix (HTH) DNA binding motif by using the method of Dodd and Egan (10), further suggesting that ditR encodes a regulatory protein. Like regulators of the IclR-family, the HTH motif (53-SVDLARVLGINPSTCFNILR-71) of DitR is located in the N-terminal region of the deduced protein. To determine if DitR regulates the expression of the ferredoxin gene, the ditR gene was inactivated in the C23O reporter strain, BKME-91. A Km^r cassette was inserted in ditR of BKME-91, and the selection of Gm^r Km^r double mutants yielded strain BKME-912 (ditA3::xylE-Gm^r ditR::Km^r). C23O assays of strain BKME-912 induced with 7-oxodehydroabietic acid showed that the expression level of ditA3 was similar to noninduced levels (Fig. 5). Electroporation of pVM220, a ditR-containing plasmid, into strain BKME-912 restored the transcriptional regulation of ditA3, as shown by its wild-type expression level (Fig. 5). Thus, DitR positively regulates expression of ditA3, and a ditR knockout mutant can be complemented in trans.

View larger version (50K): [in this window] [in a new window]   FIG. 5.   Expression of C23O by the ditA3::xylE reporter strain BKME-91 (DitR+), by the ditA3::xylE ditR::Km double mutant strain BKME-912 (DitR), and by strain BKME-912 harboring a plasmid containing ditR (DitR pVM220). Cultures were either not induced (solid bars) or induced with 0.5 mM 7-oxodehydroabietic acid (shaded bars). Activity values are means of three enzyme assays ± standard deviations.

The extradiol ring cleavage dioxygenase DitC. The ditC gene encodes an extradiol ring cleavage dioxygenase. Cloning of a 1.2-kb fragment containing ditC into pEX100T, under the control of the lac promoter in the heterologous host E. coli, resulted in clones capable of the extradiol cleavage of 2,3-dihydroxybiphenyl (2,3-DHB), as indicated by the formation of yellow colonies in spray plate assays. However, this ring cleavage dioxygenase activity was not observed when the cosmid library clone pLC12 containing ditC was tested with 2,3-DHB. The latter result might be explained by a lack of ditC expression from its native promoter in E. coli or by an undetectable activity due to low cosmid copy number combined with low activity towards 2,3-DHB. Amino acid sequence analysis of DitC revealed that it is a two-domain type I extradiol dioxygenase which contains the PROSITE consensus sequence between residues 240 and 261. Phylogenetic analysis (data not shown) of DitC indicates that it belongs to the I.3 family of dioxygenases, which includes enzymes with a preference for bicyclic substrates (14). This classification of DitC also conforms to the phylogenetic scheme proposed by Eltis and Bolin (11), since DitC has around 30% identity with several enzymes of this family (Table 2). However, DitC appears to form a new subfamily, since it has less than 54% identity to all extradiol dioxygenases found in GenBank.

Analysis of the other ORFs of the dit gene cluster. The two ORFs ditD and ditH encode putative proteins with weak sequence identity to HpcE and HpaG (28, 31) (Table 2). The putative stop codon of ditC is located 1 nucleotide upstream of the ditD start codon, and the predicted ditE start codon overlaps the putative stop codon of ditD by 4 nucleotides. Therefore, ditCDE are presumably cotranscribed. The coding sequences of ditH and ditG also overlap by 4 nucleotides, and these two genes are most likely cotranscribed. Three putative proteins encoded by ditB, ditG, and ditI showed similarity to proteins of the short-chain alcohol dehydrogenase/reductase (SDR) superfamily (15) (Table 2). The genes encoding the cleavage dioxygenase, DitC, and the ferredoxin, DitA3, are separated by ditB (Fig. 2). This location of ditB in the gene cluster suggests that it may encode a dihydrodiol dehydrogenase. However, its putative amino acid sequence shows little sequence similarity to known dihydrodiol dehydrogenases of aromatic degradation pathways (Table 2). A mutation in ditB did not impede dehydroabietic acid degradation (Table 3), and cell suspensions of the ditB knockout mutant strain did not accumulate pathway intermediates identifiable by GC-flame ionization detection (FID). In contrast, the ditI mutant accumulated one intermediate from 7-oxodehydroabietic acid. The mass spectrum (GC-MS) of this intermediate showed that it was not the dihydrodiol previously reported (6), but the structure of this intermediate was not elucidated. The analysis of the deduced amino acid sequences of ditE and ORF2 indicated similarity to membrane-bound permease proteins of the major facilitator superfamily (MFS) (Table 2). However, the two putative permeases showed no significant similarity to permeases of the aromatic acid:H⁺ symporter family, which are frequently associated with aromatic acid catabolic pathways (26). Computer-assisted transmembrane topology predictions with TMpred and hydropathy plots of the deduced amino acid sequences identified 12 potential transmembrane helices in the ORF2 gene product and only 11 in DitE. From the predicted topology of the permease-like ditE gene product and its genetic locus, we postulate that it may be involved in the transport of diterpenoids into the cell, but there is no further evidence for this. The ditF gene encodes a 397-amino-acid protein with similarity to 3-ketoacyl coenzyme A (CoA) thiolases and sterol carrier proteins. SCP-X are multifunctional eukaryotic proteins with thiolase activity encoded in the N-terminal domain, which also promote the exchange in vitro of a variety of lipids and sterols between membranes (35). The sterol carrier activity is encoded in the 143-amino-acid C-terminal domain of SCP-X (36), a domain that is not present in DitF.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We have previously shown that 7-oxodehydroabietic acid is produced from the incubation of dehydroabietic acid in dioxygenase-deficient strains of P. abietaniphila and that 7-oxodehydroabietic acid is the substrate for the diterpenoid dioxygenase DitA (21). The present study indicates that 7-oxodehydroabietic acid is also a key intermediate in the degradation of nonaromatic abietanes by strain BKME-9, since DitA strains lack the ability to grow on abietic and palustric acids and accumulate 7-oxodehydroabietic acid when incubated with those substrates. There are some precedents for microbial degradation of cycloalkanes via aromatization and subsequent aromatic-ring cleavage. Both quinate (a substituted cycloalkane) and shikimate (a substituted cycloalkene) can be mineralized via the aromatic intermediate protocatechuate (43), and this pathway appears to be widespread among diverse microorganisms. Cyclohexanecarboxylic acid can be mineralized via the aromatic intermediates p-hydroxybenzoic acid and protocatechuic acid (7, 16), and this pathway exists in diverse microorganisms but is used by a minority of those growing on cyclohexane carboxylic acid (42). Despite these precedents, microbial degradation of cycloalkanes via aromatic intermediates appears to be unusual.

Like BKME-9, most of the resin acid-degrading bacteria previously isolated with dehydroabietic acid as the sole organic substrate also have the ability to grow on nonaromatic abietanes but not on pimeranes (25). This is consistent with a common abietane-specific degradation pathway for all of these organisms. Interestingly, 7-oxodehydroabietic acid has been detected in effluent biotreatment systems (46), suggesting that the abietane pathway reported here is ubiquitous. In light of the results presented in this study, the presence of this compound in biotreatment systems might be used as an indicator of biomass inhibition or sludge health with respect to its ability to degrade resin acids. If a common pathway is used for the microbial degradation of abietanes, this may have an important implication for the pulp and paper industry. Any adverse condition inhibiting this pathway in a wastewater biotreatment system would prevent the degradation of all abietane-type diterpenoids. This may result in the accumulation of resin acids at levels that are toxic and above regulatory criteria.

Pairwise comparison of the deduced amino acid sequences encoded by dit genes to similar proteins in databases showed at most 41% residue identity, with most proteins sharing 30% identity or less (Table 2). This meant that we could not deduce the function of most genes from their sequences, as is often the case for gene clusters encoding aromatic degradation pathways. Furthermore, a comparison of the genetic organization of the gene clusters encoding aromatic hydrocarbon oxidation pathways to that of the dit cluster revealed no similarity in the order or relative location of homologous genes, which suggests that these clusters are not closely related. In most instances, the three or four genes encoding the ring-hydroxylating dioxygenases for aromatic hydrocarbons are located in one transcriptional unit, eliminating the need for the coordination of gene expression (44). The genes encoding the diterpenoid oxygenase subunits, ditA1 and ditA2, are located on a separate transcriptional unit from the genes encoding the electron transport components, ditA3 and its putative ferredoxin reductase gene (Fig. 2). Similar unlinked organization was recently reported for genes encoding the components of dibenzo-p-dioxin (2) and of naphthalene/phenanthrene dioxygenases (17). Both studies demonstrated the substrate-dependent expression of the oxygenase components but did not show coordinated or constitutive expression of the electron transport proteins. In the case of the diterpenoid dioxygenase DitA, we have demonstrated the simultaneous induction of both the ferredoxin and oxygenase components. The multicomponent alkane hydroxylase of an Acinetobacter sp. is another example where the gene encoding the catalytic hydroxylase component (alkM) is distant on the chromosome from genes encoding its electron transport proteins (rubA and rubB) (13). In this case, expression of the catalytic hydroxylase component is regulated by alkanes (29), but expression of the electron transport proteins is constitutive (12).

Phylogenetic analysis of the protein encoded by ditR indicates that it belongs to the IclR-like family of transcription regulators (Fig. 6). This family includes IclR, a repressor of the glyoxylate bypass operon in E. coli (41), GylR, an activator of the glycerol operon in Streptomyces coelicolor (38), and regulators of aromatic metabolism. The latter include PobR, an activator of the p-hydroxybenzoate hydroxylase enzyme PobA from an Acinetobacter sp. (9), and PcaR, an activator of protochatechuate degradation in Pseudomonas putida (30), as well as HppR and MhpR, regulators of 3-(3-hydroxyphenyl)propionic acid degradation in Rhodococcus globerulus and E. coli, respectively (3). Although we demonstrated that DitR positively regulates the expression of ditA3, the P. abietaniphila strain lacking a functional ditR maintained its ability to grow on abietanes (Table 3). This phenotype may be the result of the residual expression level of ditA3 (Fig. 5). In addition, the ditR mutant reporter strain BKME-912 reproducibly responded, albeit at low expression levels, to the inducer 7-oxodehydroabietic acid (Fig. 5). These observations suggest that expression of ditA3 may be controlled by at least one other mechanism. The examination of the nucleotide sequence upstream of ditA3 identified a putative ⁵⁴-promoter consensus sequence 89 bp upstream of the putative ATG start codon of ditA3. Although we have no data confirming the involvement of the alternative sigma factor, ⁵⁴, other than the consensus sequence, it is possible that the regulation of ditA3 expression might also involve a ⁵⁴-dependent regulator.

View larger version (15K): [in this window] [in a new window]   FIG. 6.   Phylogenetic tree of IclR-type transcription regulators. The unrooted tree was generated with sequences aligned with ClustalX and by using the PHYLIP protein distance and neighbor-joining methods. The numbers on the branches represent bootstrap values of 100 replicate analyses. Accession numbers are as follows: for HppR.Rglo, U89712; for YAGI.Ecol, P77300; for KdgR.Ecol, P37728; for GylR.Scoe, P15360; for IclR.Ecol, AE000475; for PcaR.Ropa, AF003947; for PcaU.Acin, L05770; for PobR.Acin, L05770; and for MphR.Ecol, P77569.

It was interesting to observe that structural analogues of dehydroabietic acid are gratuitous inducers of the dioxygenase genes. These analogues include isopimaric acid, a natural compound, and 12,14-dichlorodehydroabietic acid, a xenobiotic analogue resulting from pulp bleaching, neither of which is a substrate for the proposed pathway. This relaxed inducer specificity is common in the regulatory systems of catabolic pathways for the degradation of xenobiotic chemicals (8). However, diterpenes are natural compounds that have been present in nature for a long time. It might be hypothesized that since resin acid-degrading bacteria are unlikely to encounter an environment with only one species of diterpene, a lack of selective pressure to evolve a more specialized inducer response might have resulted in the response to a broad range of diterpenoid inducers.

Mutational analysis of several genes of the dit cluster failed to produce mutants that grow on dehydroabietic acid but do not grow on abietic and palustric acids (Table 3). Thus, the gene(s) encoding the enzyme(s) responsible for the aromatization of abietanes remains unidentified. The strain with a mutation in the gene encoding the extradiol cleavage dioxygenase produced a yellow supernatant from dehydroabietic acid in cell suspension assays. This yellow compound was not characterized, but we suspect that it was formed from the spontaneous oxidation of 7-oxo-11,12-dihydroxydehydroabietic acid to 7-oxodehydroabietic acid-11,12-quinone. The oxidation of the diol to a yellow compound would be similar to the previously described spontaneous oxidation of 1,2-dihydroxynaphthalene to 1,2-naphthaquinone (27).

The characterization of the abietane catabolic pathway of P. abietaniphila BKME-9 has significantly increased our understanding of the biodegradation of this industrially important and naturally abundant class of compounds. However, the present study also determined that the enzymes responsible for the early steps of the pathway were not located in the dit gene cluster. Therefore, work to identify the gene(s) encoding the enzymes for the conversion of abietic acid to the central metabolite, 7-oxodehydroabietic acid, is currently under way.