Novel 2,4-Dichlorophenoxyacetic Acid Degradation Genes from Oligotrophic Bradyrhizobium sp. Strain HW13 Isolated from a Pristine Environment

Abstract

The tfd genes of Ralstonia eutropha JMP134 are the only well-characterized set of genes responsible for 2,4-dichlorophenoxyacetic acid (2,4-D) degradation among 2,4-D-degrading bacteria. A new family of 2,4-D degradation genes, cadRABKC, was cloned and characterized from Bradyrhizobium sp. strain HW13, a strain that was isolated from a buried Hawaiian soil that has never experienced anthropogenic chemicals. The cadR gene was inferred to encode an AraC/XylS type of transcriptional regulator from its deduced amino acid sequence. The cadABC genes were predicted to encode 2,4-D oxygenase subunits from their deduced amino acid sequences that showed 46, 44, and 37% identities with the TftA and TftB subunits of 2,4,5-trichlorophenoxyacetic acid (2,4,5-T) oxygenase of Burkholderia cepacia AC1100 and with a putative ferredoxin, ThcC, of Rhodococcus erythropolis NI86/21, respectively. They are thoroughly different from the 2,4-D dioxygenase gene, tfdA, of R. eutropha JMP134. The cadK gene was presumed to encode a 2,4-D transport protein from its deduced amino acid sequence that showed 60% identity with the 2,4-D transporter, TfdK, of strain JMP134. Sinorhizobium meliloti Rm1021 cells containing cadRABKC transformed several phenoxyacetic acids, including 2,4-D and 2,4,5-T, to corresponding phenol derivatives. Frameshift mutations indicated that each of the cadRABC genes was essential for 2,4-D conversion in strain Rm1021 but that cadK was not. Five 2,4-D degraders, including Bradyrhizobium and Sphingomonas strains, were found to have cadA gene homologs, suggesting that these 2,4-D degraders share 2,4-D degradation genes similar to those of strain HW13 cadABC.

2,4-Dichlorophenoxyacetic acid (2,4-D) is a manufactured herbicide that has been widely used for the control of broadleaf weeds since its introduction in the 1940s. Many 2,4-D-degrading microorganisms have been isolated from agricultural, urban, and industrial soils and sediments (2, 3, 9, 22, 30, 50), and the catabolic pathway of 2,4-D mineralization in Ralstonia eutropha JMP134 has been extensively characterized (8–10, 14, 19, 25, 26, 32–35, 38, 41–43, 48). In JMP134, 2,4-D is transformed to 2,4-dichlorophenol (2,4-DCP) by α-ketoglutarate-dependent 2,4-D dioxygenase encoded by tfdA, and 2,4-DCP is subsequently hydroxylated by 2,4-DCP hydroxylase encoded by tfdB to form 3,5-dichlorocatechol (3,5-DCC). 3,5-DCC is further metabolized through an intradiol ring cleavage pathway encoded by tfdCDEF (Fig. 1). These genes are located on plasmid pJP4.

FIG. 1.

Proposed 2,4-D degradation pathway encoded by tfd genes in R. eutropha JMP134. The TfdA enzyme, α-ketoglutarate-dependent 2,4-D dioxygenase, catalyzes the simultaneous oxidation of both 2,4-D and α-ketoglutarate to generate 2,4-DCP. 2,4-DCP is transformed to 3,5-DCC by TfdB, and 3,5-DCC is further metabolized through an intradiol ring cleavage pathway encoded by tfdCDEF. TCA, tricarboxylic acid cycle.

Most 2,4-D-degrading bacteria isolated from human-disturbed sites contain tfdA gene homologs. They include various copiotrophic, fast-growing genera in the β and γ subdivisions of the Proteobacteria and have been classified as class I 2,4-D degraders (24). Ka et al. reported another group of 2,4-D degraders (class II) that were also isolated from disturbed sites but have neither tfdA gene homologs nor α-ketoglutarate-dependent 2,4-D dioxygenase activity (21–23). This group is composed of copiotrophic, fast-growing strains in the α subdivision of the Proteobacteria, mostly belonging to the genus Sphingomonas. Fulthorpe et al. (17) and Kamagata et al. (24) isolated 2,4-D degraders from noncontaminated, pristine soils, degraders which have neither tfdA gene homologs nor α-ketoglutarate-dependent 2,4-D dioxygenase activity and, in contrast to those of the other two classes, grow slowly. This group of 2,4-D degraders (class III) is affiliated with the Bradyrhizobium-Agromyces-Nitrobacter-Afipia cluster (A. Saitou, H. Mitsui, and T. Hattori, Abstr. 11th Meet. Jpn. Soc. Microb. Ecol., p. 26, 1995) of oligotrophic bacteria in the α subdivision of the Proteobacteria. The existence of three distinct ecological and genetic classes of 2,4-D degraders indicates a diversity of 2,4-D degradation genes and perhaps of pathways among 2,4-D degraders. However, the 2,4-D degradation genes of class II and III 2,4-D degraders have not been characterized.

In this report, we describe the cloning and characterization of the new 2,4-D catabolic genes cadRABKC, which are responsible for the degradation of 2,4-D in a class III 2,4-D degrader, Bradyrhizobium sp. strain HW13. This strain was isolated from a buried, pristine soil from within Hawaii Volcanoes National Park, Hawaii. This soil was covered 4,800 years ago by a lava flow which has separated that soil from human impact since before the colonization of Hawaii (24). We also examined the distribution of cadA gene homologs among other class II and class III 2,4-D degraders. Our results reveal a second family of genes capable of 2,4-D degradation and show that these genes exist in communities previously unexposed to manufactured chlorinated chemicals.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

The plasmids and bacterial strains used are listed in Table 1. Bradyrhizobium strains were grown at 30°C in CPTYM (1.25 g of Bacto Tryptone, 1.25 g of Bacto Peptone, 1.25 g of Casamino Acids, 2.5 g of yeast extract, 1.5 g of mannitol, 0.03 g of MgSO₄ · 7H₂O, and 0.0035 g of CaCl₂ · 2H₂O, each per liter). Minimal salt medium W (W medium) (37) was used for the growth on 2,4-D or other phenoxyacetic acid substrates. Sphingomonas strains, Sinorhizobium meliloti Rm1021 (12), and Pseudomonas putida PpY101 (13) were used as host strains for the strain HW13 gene library and were grown at 30°C in Luria-Bertani (LB) medium. Escherichia coli HB101 was grown in LB medium at 37°C. Antibiotics were used at final concentrations of 50 μg/ml for ampicillin, 10 μg/ml for tetracycline, 50 μg/ml for kanamycin, and 100 μg/ml for nalidixic acid.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristic(s)	Reference or origin
Strains
Bradyrhizobium sp. strains
HW13	Wild type, class III 2,4-D degrader	24
HWK12	Wild type, class III 2,4-D degrader	24
BTH	Wild type, class III 2,4-D degrader	24
Sphingomonas sp. strains
B6-5	2,4-D conversion ability lost, originally isolated as a class II 2,4-D degrader	16
B6-10	Wild type, class II 2,4-D degrader	16
TFD26	Wild type, class II 2,4-D degrader	16
TFD44	Wild type, class II 2,4-D degrader	16
Ralstonia eutropha JMP134	Wild type, class I 2,4-D degrader	9
Sinorhizobium meliloti Rm1021	Host strain; Nalr Smr	12
Sinorhizobium meliloti 22-10	Rm1021 harboring p22-10	This study
Pseudomonas putida PpY101	Host strain; met Nalr	13
Escherichia coli HB101	supE44 hsdS20(rB− mB−) recA13 ara-14 proA2 lacY1 galK2 rpsL20 xyl-5 mtl-1 leuB6 thi-1	Takara Shuzo
Plasmids
pVK100	Cosmid vector; Kmr Tcr RK2 ori Mob+	29
p22-10	2,4-D conversion-positive clone of HW13 gene library in pVK100	This study
pVBC	pVK100 with 6-kb BglII-HindIII fragment of HW13 carrying cadRABKC	This study
pVBC1	cadR frameshift mutant of pVBC	This study
pVBC2	cadA frameshift mutant of pVBC	This study
pVBC3	cadB frameshift mutant of pVBC	This study
pVBC4	cadK frameshift mutant of pVBC	This study
pVBCΔ5	cadC deletion mutant of pVBC	This study
pBluescript II KS(+)	Cloning vector; Apr
pBHS1	pBluescript II KS with 3.6-kb HincII-SpeI fragment of HW13 carrying cadABK; direction of cadABK is identical to that of the lac promoter of pBluescript II KS	This study
pRK2013	Helper plasmid for triparental mating; Kmr Tra+	45

DNA and RNA manipulations, nucleotide sequencing, and sequence analysis.

DNA and RNA manipulations were carried out essentially as described elsewhere (1, 46). A Kilosequence kit (Takara Shuzo, Kyoto, Japan) was used to construct a series of deletion derivatives whose nucleotide sequences were determined by the dideoxy termination method with an ALFexpress DNA sequencer (Amersham Pharmacia Biotech, Piscataway, N.J.). The nucleotide sequence analysis was carried out with GeneWorks software (Intelligenetics, Inc., Mountain View, Calif.) and the FASTA and ClustalW programs provided by the National Institute of Genetics, Shizuoka, Japan. Southern and Northern slot blot hybridization analyses were performed as described previously (52). Total RNAs were extracted from HW13 cells grown on CPTYM followed by incubation in W medium containing 1 mM 2,4-D or in another potentially inducing substrate. RNA was blotted onto a nylon membrane, and hybridization was performed with a DNA probe specific to cadA. In the case of strain Rm1021 harboring plasmid pVBC, cells grown in LB medium were incubated in W medium containing 1 mM substrate. A frameshift mutant of the cadR gene in pVBC was constructed by digestion with BamHI followed by treatment with T4 DNA polymerase and ligation. The resultant plasmid, pVBC1, was recovered from a transformant, and the absence of a BamHI site in pVBC1 was verified. Frameshift mutants of cadA (pVBC2), cadB (pVBC3), and cadK (pVBC4) were constructed in the same manner; Sse8387I, XhoI, and EcoRI were used for the restriction enzyme cleavages respectively.

Construction and transfer of HW13 gene library.

HW13 total DNA was partially digested with EcoRI, and the fragments obtained were inserted into a broad-host-range cosmid vector, pVK100 (29). The average size of the DNA inserts was approximately 25 kb. The resultant DNA library was introduced into E. coli HB101 by in vitro packaging and transfection by using Gigapack II plus gold packaging extract (Stratagene, La Jolla, Calif.). The gene library was conjugally transferred from E. coli to P. putida PpY101 and S. meliloti Rm1021 by triparental mating with a helper plasmid, pRK2013 (11, 45).

Screening for 2,4-D-converting clones.

Approximately 2,500 clones of each host strain harboring the HW13 gene library were screened as follows. Five clones were inoculated and cultured together in 3 ml of LB medium. The harvested cells were washed and incubated with shaking for 2 days at 30°C in 3 ml of W medium supplemented with 50 μM 2,4-D. A 1-ml aliquot was centrifuged (10,000 × g, 10 min), the supernatant was filtered through a membrane filter (pore size, 0.45 μm; Advantec, Tokyo, Japan), and the filtrate was analyzed for disappearance of 2,4-D by high-performance liquid chromatography (HPLC) as described below. When 2,4-D conversion activity was observed in the culture, the five individual clones were cultured separately, and each culture filtrate was analyzed for 2,4-D disappearance.

Detection of phenoxyacetic acid conversion activity. (i) 2,4-D conversion activity assay.

HPLC analysis of a culture filtrate was performed at room temperature with an Alliance 2690 system (Waters, Milford, Mass.) equipped with a TSKgel ODS-80TM column (6 mm [internal diameter] by 150 mm; Tosoh, Tokyo, Japan). The mobile phase was a mixture of water (49.5%), acetonitrile (49.5%), and phosphate (1.0%), run at a flow rate of 1.0 ml/min. 2,4-D and the metabolite, 2,4-DCP, were detected at 280 nm by a UV spectrophotometric detector. The metabolite was subjected to gas chromatography-mass spectrometry analysis as described previously (47).

(ii) Conversion activity of phenoxyacetic acids.

Conversion activity of phenoxyacetic acids, including 2,4-D, was examined as follows. Strain Rm1021 harboring pVBC was grown in 50 ml of LB medium with 50 μM 2,4-D. The cells were resuspended at an optical density at 660 nm (OD₆₆₀) of 2.0 in 10 ml of W medium containing 50 μM 2,4-D and other phenoxyacetic acids and were incubated with shaking for 3.5 h at 30°C. At appropriate intervals, a 1-ml aliquot was withdrawn and analyzed by HPLC as described above.

(iii) Conversion activity in crude cell extract.

Crude cell extract of strain Rm1021 harboring pVBC was prepared, and an enzyme reaction was performed as described by Danganan et al. (6). 2,4-D was added at a final concentration of 50 μM. The resulting reaction solution was subjected to HPLC analysis.

Detection of gene products.

The cadABK genes in plasmid pBSH1 were induced under the control of the lac promoter in E. coli HB101, and the whole cellular protein was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis as described previously (36). In the case of HB101 harboring pVBC, cells grown in LB medium containing 100 μM 2,4-D were incubated in W medium containing 100 μM 2,4-D for 3.5 h at 30°C, and their protein was subjected to SDS-PAGE analysis. The expression of the cad genes in strain Rm1021 was also examined. Rm1021 cells harboring pVBC were grown in LB medium containing 50 μM 2,4-D and were incubated in W medium containing 50 μM 2,4-D for 3.5 h at 30°C. The whole cellular protein was then subjected to SDS-PAGE.

Nucleotide sequence accession number.

The nucleotide sequence determined in this study has been deposited in the DDBJ, EMBL, and GenBank databases under accession no. AB062679.

RESULTS

Cloning of 2,4-D degradation genes of Bradyrhizobium sp. strain HW13.

We constructed a gene library of HW13 in E. coli HB101 by using a cosmid vector, pVK100, which can be mobilized from E. coli to other gram-negative host strains by triparental mating. When E. coli HB101 cells harboring the HW13 gene library were grown in LB medium and incubated in 2,4-D minimal medium, no 2,4-D-degrading activity was seen among approximately 2,500 clones. We then conjugally transferred the HW13 gene library by triparental mating to P. putida PpY101 (13), which was derived from the host strain of the toluene degradation plasmid, TOL, and hence was expected to express degradation genes more efficiently than E. coli. In this case, we also examined about 2,500 clones but did not detect any positive clones of P. putida PpY101. Since there could be some difficulty in expressing HW13 genes in these distantly related hosts, we employed a closely related host, S. meliloti strain Rm1021, for screening 2,4-D degradation genes of HW13. Sutton et al. (49) succeeded in inducing the expression of root curling genes of Bradyrhizobium japonicum by using S. meliloti as the host strain. The HW13 gene library was transferred by triparental mating to S. meliloti Rm1021, and the clones were screened for 2,4-D degradation activity. Among approximately 2,500 clones screened, one transconjugant, 22-10, showed 2,4-D degradation activity. 2,4-D disappeared in the culture broth of 22-10, a phenomenon which was accompanied by the appearance of a metabolite with a retention time identical to that of authentic 2,4-DCP. Gas chromatography-mass spectrometry analysis of the metabolite revealed a fragment spectrum that matched that of 2,4-DCP (data not shown). No other metabolite was detected in the culture broth, and further degradation of 2,4-DCP was not observed. These results indicate that clone 22-10 metabolized 2,4-D to 2,4-DCP, as has been found for strain JMP134.

A plasmid containing fragment 22-10 was extracted, recovered in E. coli, and designated p22-10. p22-10 has an insertion length of about 22 kb. Subclones of p22-10 were constructed by using vector pVK100 as shown in Fig. 2. These subclones were conjugally transferred to the host strain Rm1021, and 2,4-D degradation by each conjugant was examined. 2,4-D degradation activity was conferred by pVBGP and pVBC. The insertion fragment of pVBC completely overlapped that of pVBGP. The 0.4-kb left-terminal and 1.3-kb right-terminal deletions of the pVBC insert resulted in a loss of 2,4-D degradation activity, as shown in pVBM5 and pVE42, respectively, suggesting that most of the pVBC insert (a 6.0-kb BglII-HindIII fragment) is required for 2,4-D metabolism. E. coli HB101 and P. putida PpY101 containing pVBC, however, showed no 2,4-D degradation.

FIG. 2.

Subcloning and identification of 2,4-D degradation genes. The spanning region of inserts in subclones of p22-10 (top), the positions of frameshift mutations (vertical arrows), and the region of a deletion in mutant derivatives of pVBC (bottom) are illustrated. The 2,4-D degradation activity of each clone is summarized on the right. The ORFs containing cad genes are represented by large horizontal arrows at the bottom which correspond to the physical maps of mutant derivatives of pVBC.

We also examined the 2,4-D degradation activity of crude cell extracts prepared from pVBC-containing cells of E. coli HB101, P. putida PpY101, and S. meliloti Rm1021. None of these crude extracts exhibited 2,4-D degradation activity.

Nucleotide sequence analysis of the DNA insert in pVBC.

The entire 6-kb DNA insert of pVBC was sequenced and found to consist of 6,040 nucleotides. As shown in Fig. 2, it contained five open reading frames (ORFs) whose deduced amino acid sequences are similar to those of the gene products involved in the degradation of other aromatic chemicals (Table 2). The ORF1 product showed the highest identity with NitR, an AraC/XylS type of transcriptional regulator of Rhodococcus rhodochrous J1 (21%) (31). The ORF2 and ORF3 products showed 46 and 44% identities with TftA and TftB of Burkholderia cepacia AC1100, which correspond to large and small subunits of 2,4,5-trichlorophenoxyacetic acid (2,4,5-T) oxygenase, respectively (7). The ORF2 and ORF3 products also have identities exceeding 30% with the large and small subunits of benzoate dioxygenase, BenA and BenB, respectively, of Acinetobacter sp. strain ADP1 (40) and Rhodococcus sp. strain RHA1 (28). The ORF4 product showed 60% identity with the 2,4-D transport protein, TfdK, of pJP4 (35). The ORF5 product showed 37 and 35% identities with the [2Fe-2S] ferredoxin protein, ThcC, of thiocarbamate-herbicide-degrading Rhodococcus erythropolis NI86/21 (39) and with the putidaredoxin protein, CamB, of camphor-degrading P. putida ATCC 17453, respectively, proteins which mediate electron transfer from NADH to cytochrome P-450cam monooxygenase (44). Based on these results, we infer these ORFs to be chloroaryl ether degradation genes and designate ORF1 to ORF5 as cadR, cadA, cadB, cadK, and cadC, respectively.

TABLE 2.

Proteins homologous to ORFs found in pVBC (deduced amino acid sequence)

ORF (protein encoded)	Homolog (% identity)	Origin (plasmid)	Function or enzyme	Accession no.
ORF1 (CadR)	NitR (21)	Rhodococcus rhodochrous J1	Regulator for nitrilase	D67026
	FeaR (18)	Escherichia coli K-12	Regulator for 2-phenylethylamine catabolism	AE000235
ORF2 (CadA)	TftA (46)	Burkholderia cepacia AC1100	2,4,5-T oxygenase large subunit	U11420
	BenA (34)	Acinetobacter sp. strain ADP1	Benzoate 1,2-dioxygenase terminal oxygenase large subunit	AF009224
	BenA (34)	Rhodococcus sp. strain RHA1	Benzoate 1,2-dioxygenase terminal oxygenase large subunit	AB055706
ORF3 (CadB)	TftB (44)	Burkholderia cepacia AC1100	2,4,5-T oxygenase small subunit	U11420
	BenB (32)	Acinetobacter sp. strain ADP1	Benzoate 1,2-dioxygenase terminal oxygenase small subunit	AF009224
	BenB (31)	Rhodococcus sp. strain RHA1	Benzoate 1,2-dioxygenase terminal oxygenase small subunit	AB055706
ORF4 (CadK)	TfdK (60)	Ralstonia eutropha JMP134(pJP4)	2,4-D transporter	U16782
	Pcak (28)	Acinetobacter sp. strain ADP1	4-Hydroxybenzoate transporter	L05770
	BenK (27)	Acinetobacter sp. strain ADP1	Benzoate transporter	AF009224
ORF5 (CadC)	ThcC (37)	Rhodococcus erythropolis NI86/21	Putative ferredoxin	U17130
	FdVI (37)	Rhodobacter capsulatus B10	Putative ferredoxin	Y11304
	CamB (35)	Pseudomonas putida ATCC 17453	Putidaredoxin for cytochrome P-450cam monooxygenase	J05406

Expression of cad genes in E. coli.

To identify the gene products in E. coli, a 3.6-kb HincII-SpeI fragment containing cadABK was subcloned into pBluescript II KS to construct pBHS1. The molecular sizes of the cadABK products were estimated to be 49.8, 19.8, and 48.9 kDa, respectively, on the basis of their nucleotide sequences. However, the induction of polypeptides with sizes corresponding to putative cadABK gene products was not observed (data not shown). Similar SDS-PAGE analysis was also performed with S. meliloti Rm1021 harboring pVBC, which showed 2,4-D degradation, but the polypeptides expected for cadRABKC gene products were not detected. We also could not detect these gene products in HB101 harboring pVBC (data not shown).

cadR, cadA, cadB, and cadC were essential for 2,4-D transformation in S. meliloti Rm1021.

To examine whether each of the cadRABKC genes is essential for 2,4-D conversion activity, frameshift and deletion mutants of pVBC were constructed as illustrated in Fig. 2. pVBC1, pVBC2, pVBC3, and pVBC4 contained frameshift mutations in cadR, cadA, cadB, and cadK, respectively, and pVBCΔ5 had a deletion of the 825-bp SpeI-HindIII fragment, including the entire cadC gene. These plasmids were introduced into S. meliloti Rm1021 by triparental mating. Each conjugant was grown in LB medium and incubated in W medium containing 50 μM 2,4-D to determine 2,4-D degradation activity. As summarized in Fig. 2, pVBC1, pVBC2, pVBC3, and pVBCΔ5 lost the ability to transform 2,4-D. In contrast, pVBC4 retained the ability to convert 2,4-D to 2,4-DCP. These results indicate that cadA, cadB, cadC, and cadR are essential for 2,4-D conversion in Rm1021 but that cadK is not.

Substrate preference of cad gene products in Rm1021.

In order to examine the substrate preference of cad gene products, the conversion rates of phenoxyacetic acids, including 2,4-D, phenoxyacetic acid (PAA), 2-chlorophenoxyacetic acid (2-CPAA), 4-chlorophenoxyacetic acid (4-CPAA), 2,3-dichlorophenoxyacetic acid (2,3-D), 3,4-dichlorophenoxyacetic acid (3,4-D), and 2,4,5-T, were examined by using resting cells of pVBC-containing Rm1021. Strain Rm1021, which contained the vector pVK100, showed no transformation activity against any of these substrates. The cells grown in LB medium with 50 μM 2,4-D were incubated in W medium with a 50 μM concentration of each of the phenoxyacetic acids mentioned above, and the concentrations of the phenoxyacetic acids were determined periodically. As shown in Fig. 3, PAA, 4-CPAA, 3,4-D, 2,4,5-T, and 2,4-D were transformed efficiently at the initial conversion rate of 10 to 12.5 μM/h at an OD₆₆₀. 2-CPAA was transformed less effectively at an initial rate of 6.3 μM/h at an OD₆₆₀. 2,3-D was not transformed at all. These results indicate that the cad gene products can metabolize a variety of phenoxyacetic acids. Benzoate was not metabolized, although cadAB have similarity to the benzoate dioxygenase genes benAB (data not shown).

FIG. 3.

Transformation of phenoxyacetic acids by Rm1021 harboring pVBC. The cells grown in the presence of 2,4-D were incubated in W medium containing a 50 μM concentration of each substrate. The remaining substrates were analyzed by HPLC. The data are the averages of three independent experiments. Standard deviations are indicated by error bars.

cadA gene transcription induction.

To determine the induction profiles of cadA gene transcription, Northern hybridization was performed with HW13 and pVBC-containing S. meliloti Rm1021. No cadA transcription was observed in HW13 cells or in pVBC-containing Rm1021 cells pregrown in CPTYM or in LB medium in the absence of phenoxyacetic acids. When phenoxyacetic acids were added, only 2,4-D and 4-CPAA induced cadA gene transcription, and this occurred in both strain HW13 and strain Rm1021 (Table 3). We also examined the growth of HW13 on these phenoxyacetic acids and found that the assimilation profiles were consistent with the cadA transcription profiles (Table 3). These results suggest that cad gene expression is strictly regulated at the transcriptional level.

TABLE 3.

Induction substrate specificity of cadABC

cadA gene in class II and class III 2,4-D degraders.

To examine the distribution of cad gene homologs among class II and class III 2,4-D degraders, Southern hybridization analysis was performed with the HW13 cadA gene probe on SalI digests of total DNAs prepared from a class I 2,4-D degrader, R. eutropha JMP134; class II 2,4-D degraders, Sphingomonas sp. strains B6-5, B6-10, TFD26, and TFD44 (16); and class III 2,4-D degraders, including Bradyrhizobium sp. strains HWK12 and BTH (24). As shown in Fig. 4, all of the class II and III 2,4-D degraders tested, except for the class II degrader B6-5, showed evidence of cadA gene homologs. Strain B6-5 originally grew on 2,4-D but had lost this ability before the hybridization experiments were performed. The class I 2,4-D degrader R. eutropha JMP134 did not show evidence of any cadA gene homologs. These results suggest that many class II and III 2,4-D degraders could employ cad gene homologs for 2,4-D degradation.

FIG. 4.

Southern hybridization analysis of 2,4-D degraders. Hybridization was performed with the HW13 cadA gene probe on SalI-digested total DNAs from 2,4-D degraders. Lane 1, S. meliloti Rm1021; lane 2, HW13; lanes 3 and 4, class III degraders (Bradyrhizobium sp. strains HWK12 and BTH); lanes 5 to 8, class II degraders (Sphingomonas sp. strains B6-5, B6-10, TFD26, and TFD44); lane 9, class I degrader (R. eutropha JMP134). Lanes M, DNA size marker.

DISCUSSION

New 2,4-D degradation genes have been cloned from Bradyrhizobium sp. strain HW13, which was recovered from a buried Hawaiian soil that has been isolated from 2,4-D and other anthropogenic chemicals by lava flows since before human colonization of the islands. These lava layers formed solid, 2- and 5-m-thick rock caps over the soil from which the strain was isolated. Furthermore, neither 2,4-D nor 2,4,5-T was ever used in this region of the park (T. Tunison, U.S. National Park Service, personal communication). The degradation genes were entirely different from the well-known tfd genes that encode the 2,4-D degradation pathway of R. eutropha JMP134, suggesting that at least the initial 2,4-D degradation step of HW13 has a totally distinct evolutionary origin from that of JMP134.

The new 2,4-D degradation genes were designated cadRABKC. With the exception of cadK, they have no similarity to tfd genes. The cadR gene may encode a positive transcriptional regulator, because its sequence is somewhat similar to that of nitR, which an AraC/XylS codes for type of positive transcriptional regulator (31), and the deletion and frameshift mutations within cadR resulted in the loss of 2,4-D conversion activity in S. meliloti Rm1021 (Fig. 2). This gene seems to be essential for the expression of cadABKC genes in Rm1021. The cadA and cadB genes are predicted to encode large and small subunits, respectively, of 2,4-D oxygenase, based on sequence similarity to the tftAB 2,4,5-T oxygenase genes of B. cepacia AC1100 (7) and the benAB benzoate dioxygenase genes of Acinetobacter sp. strain ADP1 (40) and Rhodococcus sp. strain RHA1 (28). The cadC gene is deduced to encode a ferredoxin subunit of 2,4-D oxygenase based on its identity with the camB ferredoxin (putidaredoxin) gene of P. putida ATCC 17453, which is involved in the oxidation of camphor. Frameshift mutations in each of the cadABC genes resulted in the loss of conversion activity of 2,4-D to 2,4-DCP, indicating that the respective gene encodes a functional subunit of a multicomponent 2,4-D oxygenase. The cadK gene is thought to encode a 2,4-D transporter because of its high sequence similarity to tfdK, the 2,4-D transporter gene of JMP134 (35). The frameshift mutation of cadK did not affect 2,4-D conversion activity expressed in Rm1021. However, this result does not establish whether tfdK is involved in 2,4-D transport, because transporters are not indispensable for 2,4-D and benzoate uptake. For example, growth of strain JMP134 on 2,4-D and strain ADP1 on benzoate was not prevented by mutations in tfdK (35) or the benzoate transporter gene, benK (4), respectively. Also, Rm1021 may itself have an equivalent 2,4-D transporter that conceals the defect of the cadK frameshift mutation in Rm1021. Resolving the role of cadK in 2,4-D degradation requires the disruption of cadK in HW13, but all of the attempts to disrupt cadK in HW13 were unsuccessful.

The cadABC-encoded 2,4-D oxygenase of HW13 seems to be included in the family of aromatic ring-hydroxylation dioxygenases, that is, Rieske nonheme iron oxygenases (18). The enzymes of this family are composed of a terminal oxygenase component and one or two electron transport components. In the case of HW13 2,4-D oxygenase, the terminal oxygenase and electron transport components are encoded by cadAB and cadC, respectively.

As indicated in Fig. 5A and B, the subunits of 2,4-D oxygenase encoded by cadAB could be classified in the benzoate subfamily of two-component ring-hydroxylation dioxygenases because of their identity with benAB-encoded subunits of benzoate dioxygenase. On the other hand, the cadC-encoded ferredoxin subunit of 2,4-D oxygenase seems to be a member of the P-450 ferredoxin subfamily, as illustrated in Fig. 5C. These results suggest that cadABC-encoded 2,4-D oxygenase employs very unique electron transfer components distinct from those of the two-component benzoate dioxygenases. Thus, the cadABC region encoding 2,4-D oxygenase is suggested to have three components. cadAB and cadC are thought to encode the terminal oxygenase and ferredoxin components, respectively. The gene encoding the ferredoxin reductase component remains unidentified. From our conjugation experiments, the cadABC genes were sufficient to confer 2,4-D conversion activity to Rm1021, suggesting that Rm1021 provided a protein equivalent to a ferredoxin reductase component for the cadABC-encoded 2,4-D oxygenase. In contrast, cadABC genes on the pVBC plasmid did not confer 2,4-D conversion activity to the E. coli or P. putida host strains. When we used E. coli and P. putida as host strains, no positive clones having 2,4-D conversion activity were obtained from the HW13 genomic library. The expression of cadABC genes in these hosts might be too poor to show 2,4-D conversion activity, or these hosts might not have equivalents to a ferredoxin reductase component for the gene of cadABC that encodes 2,4-D oxygenase. No promoter consensus sequences were found in the preceding upstream sequences of cadR and cadA, a finding that supports the former possibility. The latter possibility can be proposed based on the case of tftAB-encoded 2,4,5-T oxygenase. In contrast to HW13 cadABC, tftAB genes from B. cepacia AC1100 conferred 2,4,5-T conversion activity to Pseudomonas aeruginosa PAO1. Four polypeptides, including two oxygenase subunits encoded by tftAB and polypeptides of 15 and 29 kDa, are required for the activity of 2,4,5-T oxygenase in AC1100 (6, 7). The 15- and 29-kDa polypeptides are thought to constitute a reductase component. P. aeruginosa PAO1 seems to have the substitute for the reductase component of tftAB-encoded 2,4,5-T oxygenase. The putative molecular sizes of putidaredoxin and putidaredoxin reductase for cytochrome P-450cam monooxygenase encoded by camB and camA are estimated to be 11.5 and 45.5 kDa, respectively, based on their amino acid sequences (44). The predicted molecular size of cadC-encoded ferredoxin is 11.2 kDa, which is in good agreement with the size of camB-encoded putidaredoxin. Thus, the subunits of the tftAB-encoded 2,4,5-T oxygenase that are predicted to constitute a reductase component seem to be distinct from those of the cadABC-encoded 2,4-D oxygenase.

FIG. 5.

Phylogenetic trees of CadABC subunits of strain HW13 2,4-D oxygenase and the related oxygenase subunits. The phylogenetic trees of cadA product and large (α) subunits of aromatic ring hydroxylation dioxygenases (A), cadB product and small (β) subunits of aromatic ring hydroxylation dioxygenases (B), and cadC product and ferredoxin subunits of ring hydroxylation dioxygenases and cytochrome P-450cam monooxygenases (C) are shown. In panel C, ferredoxin reductase component protein sequences were used in the case of the two-component dioxygenase system (benzoate subfamily). The deduced amino acid sequences were aligned by using the ClustalW program. CadA, CadB, and CadC subunits of HW13 2,4-D oxygenase are boxed. The accession numbers are shown to the right of each enzyme. The representative substrate for each subfamily is indicated to the right of the figure.

Although the amino acid sequence identities of the cadAB and tftAB products are not very high (around 45%), the substrate preferences of cadAB-encoded 2,4-D and tftAB-encoded 2,4,5-T oxygenases on phenoxyacetic acids are very similar. Both oxygenases transformed a variety of phenoxyacetic acids to similar extents (6). Both of them readily degraded 4-CPAA, 2,4-D, and 2,4,5-T. In addition, cadAB-encoded 2,4-D oxygenase transformed 3,4-D. The chlorine substitution at the para position seems to be important for their catalytic activity against chlorinated phenoxyacetic acids. On the other hand, tftAB-encoded 2,4,5-T oxygenase transformed 2,4,5-T relatively more rapidly than it transformed 2,4-D. The degradation activities of 2,4,5-T and 2,4-D by cadAB-encoded 2,4-D oxygenase were almost the same. It is very likely that tftAB-encoded 2,4,5-T oxygenase is more specific to 2,4,5-T than is cadAB-encoded 2,4-D oxygenase, because tftAB-containing strain AC1100 was isolated after long-term selection for growth on 2,4,5-T in which a chemostat community of microorganisms from waste dump sites plus those harboring the degradative plasmids CAM, TOL, SAL, and pAC25 experienced a gradual shift from substrates for the degradative plasmids to 2,4,5-T, with the latter eventually becoming the sole carbon source (27). The substrate specificity on phenoxyacetic acids of tfdA-encoded dioxygenase of JMP134 apparently differs from those of cadAB-encoded 2,4-D and tftAB-encoded 2,4,5-T oxygenases. tfdA-encoded dioxygenase degraded 2,4-D at a high rate, degraded 4-CPAA less rapidly, and degraded 2,4,5-T and PAA very slowly (15). In contrast to the case for tfdA, whose encoded 2,4-D dioxygenase catalyzes simultaneous dioxygenation of 2,4-D and α-ketoglutarate, Xun and Wagnon proposed that tftAB-encoded 2,4,5-T oxygenase catalyzes monooxygenation of 2,4,5-T (51). On the basis of the similarities of amino acid sequences and substrate preferences between tftAB-encoded 2,4,5-T and cadAB-encoded 2,4-D oxygenases, we suggest that cadAB-encoded 2,4-D oxygenase also catalyzes the monooxygenation of 2,4-D.

The gene organization of the tftAB region in strain AC1100 is very different from that of cadAB. The cadAB region consists of cadRABKC, which encodes a positive regulator, oxygenase subunits, and a putative transporter. In contrast, the tftAB region contains only tftAB, which encodes oxygenase subunits (7, 20). DNA rearrangements may have occurred during the long-term selection for 2,4,5-T growth, a process that may have selected variants with the tftAB segment excised from the region that originally had had a gene organization similar to that of cadRABKC. The loss of a positive regulator gene equivalent to cadR might have been restored by the acquisition of an alternate constitutive promoter sequence or by some other means of removing the obstacle to improved growth on 2,4,5-T.

The deletion of cadR in a cad gene plasmid resulted in the loss of 2,4-D conversion activity in strain Rm1021. The transcription of cadA was induced only by 2,4-D and 4-CPAA in both strain Rm1021 and strain HW13. These results strongly suggest that cadR is primarily responsible for the transcription induction of cadA by 2,4-D and 4-CPAA in both Rm1021 and HW13. Of the several phenoxyacetic acids tested, 2,4-D and 4-CPAA were found to support the growth of HW13 and to induce cadA gene transcription, suggesting that the growth of HW13 on phenoxyacetic acids is governed by the induction of cadA transcription regulated by cadR.

Although the cadRABKC genes conferred 2,4-D conversion activity to S. meliloti Rm1021, we are not sure whether they are solely responsible for 2,4-D degradation in strain HW13. We made several attempts to inactivate the cadA gene by a gene replacement, but all of the inserts were localized outside cadA. However, the following evidence strongly suggests that at least the cadRABC genes are responsible for 2,4-D degradation in HW13. (i) Genes other than cadAB were not obtained from the HW13 gene library (which contained about 2,500 clones), which is enough to cover the HW13 chromosome with 99.9% probability. (ii) cadAB gene products showed good identity with tftAB-encoded 2,4,5-T oxygenase subunits that catalyze degradation of 2,4,5-T and 2,4-D. (iii) cadA transcription was induced by 2,4-D. (iv) The substrates which induced cadA transcription in HW13 were consistent with the substrates that supported growth. (v) HW13 does not have identifiable tfd gene homologs. (vi) cadA gene homologs were shared by class II and class III 2,4-D degraders that have no tfd gene homologs. In addition, we recently obtained two independent mutant strains of HW13 that had been irradiated with UV light, were deficient in their growth on 2,4-D, and accumulated 2,4-DCP (unpublished results). These results suggest that the conversion of 2,4-D to 2,4-DCP that can be conducted by cadABC is a part of the 2,4-D degradation pathway in HW13.

In this study, we isolated new 2,4-D oxygenase genes, cadRABKC, from a class III 2,4-D degrader, Bradyrhizobium sp. strain HW13. This set of 2,4-D oxygenase genes appears to be responsible for 2,4-D transformation in HW13 and may be involved in 2,4-D transformation in at least some class II and III 2,4-D degraders. If cadABC-encoded 2,4-D oxygenase is primarily responsible for 2,4-D catabolism in 2,4-D degraders of both class II and class III, then the difference between these classes may have originated not from the difference in 2,4-D degradation genes but from the difference in the degraders’ intrinsic growth rates. The Bradyrhizobium strain we studied is apparently able to call on an existing oxygenase complex to attack a new chlorinated chemical not experienced in its natural evolutionary history. This strain seems to have recruited and adapted the oxygenase components to catabolize 2,4-D in addition to having the existing chlorophenol catabolic pathway, as suggested in the pentachlorophenol degradation pathway (5). The unique ferredoxin component encoded by cadC and the terminal oxygenase component encoded by cadAB may suggest the independent recruitment of these two components from separate origins. The organism’s slow growth on 2,4-D suggests that its enzymes are not optimized for growth on this substrate, but it shows nonetheless that a native soil community can have genetic resources to degrade at least some anthropogenic chemicals.