3,6-Dichlorosalicylate Catabolism Is Initiated by the DsmABC Cytochrome P450 Monooxygenase System in Rhizorhabdus dicambivorans Ndbn-20

ABSTRACT

The degradation of the herbicide dicamba is initiated by demethylation to form 3,6-dichlorosalicylate (3,6-DCSA) in Rhizorhabdus dicambivorans Ndbn-20. In the present study, a 3,6-DCSA degradation-deficient mutant, Ndbn-20m, was screened. A cluster, dsmR1DABCEFGR2, was lost in this mutant. The cluster consisted of nine genes, all of which were apparently induced by 3,6-DCSA. DsmA shared 30 to 36% identity with the monooxygenase components of reported three-component cytochrome P450 systems and formed a monophyletic branch in the phylogenetic tree. DsmB and DsmC were most closely related to the reported [2Fe-2S] ferredoxin and ferredoxin reductase, respectively. The disruption of dsmA in strain Ndbn-20 resulted in inactive 3,6-DCSA degradation. When dsmABC, but not dsmA alone, was introduced into mutant Ndbn-20m and Sphingobium quisquiliarum DC-2 (which is unable to degrade salicylate and its derivatives), they acquired the ability to hydroxylate 3,6-DCSA. Single-crystal X-ray diffraction demonstrated that the DsmABC-catalyzed hydroxylation occurred at the C-5 position of 3,6-DCSA, generating 3,6-dichlorogentisate (3,6-DCGA). In addition, DsmD shared 51% identity with GtdA (a gentisate and 3,6-DCGA 1,2-dioxygenase) from Sphingomonas sp. strain RW5. However, unlike GtdA, the purified DsmD catalyzed the cleavage of gentisate and 3-chlorogentisate but not 6-chlorogentisate or 3,6-DCGA in vitro. Based on the bioinformatic analysis and gene function studies, a possible catabolic pathway of dicamba in R. dicambivorans Ndbn-20 was proposed.

IMPORTANCE Dicamba is widely used to control a variety of broadleaf weeds and is a promising target herbicide for the engineering of herbicide-resistant crops. The catabolism of dicamba has thus received increasing attention. Bacteria mineralize dicamba initially via demethylation, generating 3,6-dichlorosalicylate. However, the catabolism of 3,6-dichlorosalicylate remains unknown. In this study, we cloned a gene cluster, dsmR1DABCEFGR2, involved in 3,6-dichlorosalicylate degradation from R. dicambivorans Ndbn-20, demonstrated that the cytochrome P450 monooxygenase system DsmABC was responsible for the 5-hydroxylation of 3,6-dichlorosalicylate, and proposed a dicamba catabolic pathway. This study provides a basis to elucidate the catabolism of dicamba and has benefits for the ecotoxicological study of dicamba. Furthermore, the hydroxylation of salicylate has been previously reported to be catalyzed by single-component flavoprotein or three-component Rieske non-heme iron oxygenase, whereas DsmABC was the only cytochrome P450 monooxygenase system hydroxylating salicylate and its methyl- or chloro-substituted derivatives.

INTRODUCTION

Dicamba (3,6-dichloro-2-methoxybenzoate) is an old herbicide that has been widely used to control a variety of broadleaf weeds since 1962 (1). Due to its low cost, broad weed control spectrum, low mammalian toxicity, and effective control of glyphosate-resistant weeds (2), dicamba is considered an ideal target herbicide for the engineering of herbicide-resistant transgenic crops (3). Monsanto has successfully constructed dicamba-resistant soybean and cotton, and these transgenic crops have been commercially planted since 2015 (3). Thus, the catabolism and ecotoxicology of dicamba and its metabolites are of great interest. Microbial metabolism plays a crucial role in the dissipation of dicamba in the environment (4, 5). So far, many dicamba-degrading isolates have been reported, including Stenotrophomonas (formerly Pseudomonas) maltophilia DI-6 (4), Sphingomonas sp. strain RW5 (6), Sphingobium sp. strain Ndbn-10, Rhizorhabdus dicambivorans (formerly Sphingomonas sp. strain) Ndbn-20 (7, 8), and Moorella thermoacetica (5). In all of these isolates, the initial step of dicamba degradation was demethylation, generating the herbicidally inactive metabolite 3,6-dichlorosalicylate (3,6-DCSA). Three types of dicamba demethylases have been identified: the Rieske non-heme iron oxygenase (RHO)-type monooxygenase dicamba monooxygenase (DMO) in S. maltophilia DI-6 (9, 10), the tetrahydrofolate (THF)-dependent dicamba methyltransferase Dmt in R. dicambivorans Ndbn-20 (8), and the three-component THF-dependent methyltransferase Mtv in the anaerobic strain Moorella thermoacetica (11). However, few studies addressed the further catabolism of 3,6-DCSA. Previously, Cork and Krueger proposed that the initial step of 3,6-DCSA degradation in S. maltophilia DI-6 was hydroxylation based on metabolite identification by thin-layer chromatography (TLC) and high-pressure liquid chromatography (HPLC) (12). However, the initial hydroxylation position, the whole catabolic pathway of 3,6-DCSA, and the genes involved have not been studied.

Salicylate and its derivatives are important plant hormones, drugs, chemical and pharmaceutical intermediates, and environmental pollutants (13–15), as well as key intermediates in the catabolism of aromatic compounds (16, 17). Their degradation and transformation have thus received considerable attention. Three salicylate microbial degradation pathways have been identified to date. In Pseudomonas putida PpG7 or Sphingomonas sp. strain CHY-1, salicylate is decarboxylatively hydroxylated to catechol by salicylate 1-hydroxylase NahG (18) or PhnII (19). NahG is a single-component flavoprotein from P. putida PpG7 (18, 20), and PhnII is a three-component RHO-type monooxygenase from Sphingomonas sp. strain CHY-1 (19). In the naphthalene-utilizing strain Ralstonia sp. strain U2, salicylate is 5-hydroxylated to gentisate (2,5-dihydroxybenzoate) by a three-component RHO monooxygenase, NagGHAaAb (21, 22). In Streptomyces sp. strain WA46, salicylate is transformed to salicylyl coenzyme A (salicylyl-CoA), which is then converted to gentisyl-CoA; finally, gentisyl-CoA undergoes spontaneous cleavage, generating gentisate and CoA (23).

In this study, we cloned a gene cluster (dsmR1DABCEFGR2) involved in 3,6-DCSA degradation in R. dicambivorans Ndbn-20. dsmABC was proved to encode the three components of a cytochrome P450 monooxygenase system responsible for the 5-hydroxylation of 3,6-DCSA to 3,6-dichlorogentisate (3,6-DCGA). In addition, the catalytic activity of a putative gentisate 1,2-dioxygenase encoded by dsmD was also studied, and a dicamba catabolic pathway is proposed.

RESULTS

Screening of the 3,6-DCSA degradation-deficient mutant Ndbn-20m.

To study the catabolic pathway of dicamba and the genes involved in this pathway in R. dicambivorans Ndbn-20, we attempted to detect the intermediate metabolites by HPLC and liquid chromatography-mass spectrometry (LC/MS). However, only 3,6-DCSA was detected in the culture during dicamba degradation by R. dicambivorans Ndbn-20 (Fig. 1A), whereas no metabolite was detected when 3,6-DCSA was used as the carbon source (data not shown). The reason might be that the transformation of 3,6-DCSA was the rate-limiting step in dicamba catabolism and that other metabolites were quickly transformed to an undetectable level. In previous reports, many genes or gene clusters responsible for the degradation of xenobiotic compounds were located on mobile genetic elements or catabolic plasmids and were thus prone to be lost without selective pressure from substrates (24–29). Therefore, in this study, we attempted to obtain a 3,6-DCSA degradation-deficient mutant by growing Ndbn-20 cells on 1/5 Luria-Bertani (LB) agar without the addition of 3,6-DCSA. After approximately 30 continuous transfers, we acquired a 3,6-DCSA degradation-deficient mutant, Ndbn-20m. This mutant lost the ability to degrade 3,6-DCSA but could still transform dicamba to 3,6-DCSA (Fig. 1B). These results indicated that mutant Ndbn-20m had lost at least the genes responsible for the initial step of 3,6-DCSA catabolism.

FIG 1.

HPLC analysis of the metabolites generated during dicamba degradation by R. dicambivorans Ndbn-20 (A) and its mutant Ndbn-20m (B). The culture was sampled at 0 h, 24 h, and 48 h of incubation. The detection wavelength used was 275 nm. The retention time (11.95 min) of the metabolite was identical to that of the standard, 3,6-DCSA.

Prediction of genes involved in 3,6-DCSA catabolism based on comparative genomic analysis.

To investigate which genes were lost in mutant Ndbn-20m, we sequenced the complete genome of wild-type strain Ndbn-20 and the draft genome of mutant Ndbn-20m. The complete genome of strain Ndbn-20 was 5.54 Mb and consisted of one circular chromosome of 5.03 Mb and four circular plasmids of 40.9 kb, 53.6 kb, 163.6 kb, and 256.4 kb. The draft genome of mutant Ndbn-20m contained 96 contigs, constituting a total size of 5.25 Mb. Pairwise comparison of the mutant genome with that of the wild-type strain revealed that a 64.1-kb DNA fragment was absent in mutant Ndbn-20m. This fragment was located on the chromosome from bp 2717190 to 2781331. The absence of the 64.1-kb fragment in the mutant was further confirmed by PCR. Open reading frame (ORF) analysis revealed that many conjugative transfer genes and one insertion sequence (IS) element existed in the 64.1-kb fragment (data not shown), which likely contributed to its instability in R. dicambivorans Ndbn-20.

A gene cluster designated dsm (3,6-DCSA monooxygenase) was found in the 64.1-kb fragment. The cluster contained nine putative genes; dsmD, dsmA, dsmB, dsmC, dsmE, dsmF, dsmG, and dsmR1 were orientated in the same direction, whereas dsmR2 was in the opposite orientation (Fig. 2A). DsmA shared 30 to 36% sequence identity with several cytochrome P450 monooxygenases (Table 1). Sequence alignment of DsmA with the related cytochrome P450 monooxygenases revealed that DsmA contains the highly conserved, functionally relevant regions found in P450s, which include the acid/alcohol residue pair (Glu260/Thr261), EXXR motif, heme propionate group-binding domain (His106, Arg309, and His365), and proximal heme thiolate ligand (Cys367) (30–32) (see Fig. S1 in the supplemental material). Upstream of dsmA, there were two adjacent genes, dsmB and dsmC. DsmB shared 57% identity with the [2Fe-2S] ferredoxin FdII from Caulobacter crescentus CB15 (33), and DsmC shared 44% identity with ThcD, the reductase component of the cytochrome P450 monooxygenase system ThcBCD from Rhodococcus erythropolis NI86/21 (34). Based on the above analysis, the three consecutive genes dsmA, dsmB, and dsmC were predicted to encode a three-component cytochrome P450 monooxygenase system. DsmD shared 51% identity with the dioxygenase GtdA, which catalyzed the cleavage of gentisate and 3,6-dichlorogentisate (3,6-DCGA), from Sphingomonas sp. strain RW5. DsmE shared 52% identity with the fumarylpyruvate hydrolase NagK from Ralstonia sp. strain U2 (35). DsmF was most similar (38% identity) to MhbT (3-hydroxybenzoate transporter), a major facilitator superfamily (MFS) transporter of Klebsiella oxytoca M5a1 (36). DsmG shared 39% identity with a maleylacetate reductase from Alcaligenes eutrophus JMP134 (37). The genes dsmR1 and dsmR2 putatively encoded the LuxR family transcriptional regulator and IclR family transcriptional regulator, respectively (Table 1). Considering that the initial step in degradation of salicylate and its derivatives is typically hydroxylation (12, 21) and that the dsm cluster was lost in the 3,6-DCSA degradation-deficient mutant Ndbn-20m, we predicted that the dsm cluster was involved in the degradation of 3,6-DCSA and that dsmABC was responsible for the hydroxylation of 3,6-DCSA.

FIG 2.

Proposed catabolism of dicamba in R. dicambivorans Ndbn-20. (A) Organization of genes involved in the catabolic pathway of dicamba. Arrows indicate the size and transcriptional direction of each gene. Lines below the gene cluster show the locations and sizes of the PCR fragments in panel B. (B) Agarose gel electrophoresis of RT-PCR products. + and − indicate the cells grown with 3,6-DCSA and glucose, respectively. (C) Proposed dicamba catabolic pathway in R. dicambivorans Ndbn-20. Dmt66, dicamba methyltransferase; DsmABC, cytochrome P450-type 3,6-DCSA 5-hydroxylase; GtdA, 3,6-dichlorogentisate (3,6-DCGA) 1,2-dioxygenase; DsmD, 3-chlorogentisate 1,2-dioxygenase; DsmG, maleylacetate reductase; DsmE, fumarylpyruvate hydrolase or maleylpyruvate hydrolase. The functions of Dmt66 and GtdA are cited from references 8 and 6, respectively. The activities of DsmABC and DsmD were confirmed in this study. The activities of DsmG and DsmE were predicted from homology comparison with known genes.

TABLE 1.

Deduced functions of ORFs within or near the dsm cluster

Gene name	Database	Homologous protein, accession no., and source	% identity
orf1	NR	Autoinducer synthase, WP_072598505.1, Sphingomonas sp. strain JJ-A5	76
	Swiss-Prot	Acyl-homoserine-lactone synthase, P58584.1, Ralstonia solanacearum GMI1000	34
dsmR1	NR	LuxR family transcriptional regulator, WP_083831291.1, Sphingomonas sp. strain KC8	65
	Swiss-Prot	Transcriptional activator protein BjaR1, Q89VI3.1, Bradyrhizobium diazoefficiens USDA 110	27
dsmD	NR	Gentisate 1,2-dioxygenase, WP_068086548.1, Novosphingobium rosa NBRC15208	52
	Swiss-Prot	Gentisate 1,2-dioxygenase, O86041.2, Ralstonia sp. strain U2	33
dsmA	NR	Cytochrome P450, WP_067292530.1, Sulfitobacter sp. strain EhC04	55
	Swiss-Prot	Cytochrome P450, CYP107DY1, D5E3H2.1, Bacillus megaterium QM B1551	36
dsmB	NR	[2Fe-2S] ferredoxin, WP_056731560.1, Phenylobacterium sp. strain Root700	61
	Swiss-Prot	[2Fe-2S] ferredoxin, P37098.1, Caulobacter crescentus CB15	57
dsmC	NR	Hypothetical, protein WP_054538362.1, Confluentimicrobium sp. strain EMB200-NS6	53
	Swiss-Prot	Rhodocoxin reductase ThcD, P43494.2, Rhodococcus sp. strain NI86/21	44
dsmE	NR	Fumarylacetoacetase, WP_054763928.1, Brevundimonas sp. strain DS20	59
	Swiss-Prot	Fumarylpyruvate hydrolase NagK, O86042.1, Ralstonia sp. strain U2	52
dsmF	NR	MFS transporter, WP_054763929.1, Brevundimonas sp. strain DS20	58
	Swiss-Prot	3-Hydroxybenzoate transporter MhbT, Q5EXK5.1, Klebsiella pneumoniae M5a1	38
dsmG	NR	Maleylacetate reductase, WP_054763930.1, Brevundimonas sp. strain DS20	59
	Swiss-Prot	Maleylacetate reductase, P27137.1, Alcaligenes eutrophus JMP134	39
dsmR2	NR	IclR family transcriptional regulator, WP_054763933.1, Brevundimonas sp. strain DS20	44
	Swiss-Prot	HTH-type transcriptional repressor AllR, Q765S0.1, Klebsiella pneumoniae NTUH-K2044	26

Transcriptional analysis of the dsm cluster.

To investigate whether these genes were in an operon and induced by 3,6-DCSA, we performed reverse transcription-PCR (RT-PCR) and RT-quantitative PCR (RT-qPCR) using mRNA derived from strain Ndbn-20 grown with glucose or 3,6-DCSA. The RT-PCR results showed that dsmD, dsmA, dsmB, dsmC, dsmE, dsmF, dsmG, and dsmR1 were organized in an operon, whereas dsmR2 was individually transcribed (Fig. 2B). the RT-qPCR results showed that all genes in the dsm cluster were apparently induced by 3,6-DCSA. The transcription levels of dsmD, dsmA, dsmB, dsmC, dsmE, dsmF, and dsmG in 3,6-DCSA-cultured cells were increased by 80- to 568-fold relative to those in glucose-cultured cells, whereas the transcription levels of dsmR1 and dsmR2 were increased by 23- and 15-fold (Fig. 3). The transcriptional results further indicated that the dsm cluster was involved in the degradation of 3,6-DCSA.

FIG 3.

Transcriptional levels of the dsm genes in R. dicambivorans Ndbn-20. mRNA expression levels of the nine genes were estimated using RT-qPCR and the 2^−ΔΔCT method. The 16S rRNA gene was used as the reference gene. The data were derived from three independent measurements, and the error bars indicate standard deviations.

Disruption and functional complementation of dsmA in mutants Ndbn-20m and Ndbn-20 ΔdsmA.

To confirm that dsmA was responsible for the hydroxylation of 3,6-DCSA in R. dicambivorans Ndbn-20, dsmA was disrupted from R. dicambivorans Ndbn-20 using a homologous recombination technique, generating the mutant Ndbn-20 ΔdsmA. Similar to the case for mutant Ndbn-20m, Ndbn-20 ΔdsmA could not degrade 3,6-DCSA (Fig. 4) but could still transform dicamba to 3,6-DCSA (data not shown). This result demonstrated that dsmA was essential for the catabolism of 3,6-DCSA in R. dicambivorans Ndbn-20.

FIG 4.

Time course of 3,6-DCSA degradation by wild-type Ndbn-20, mutants Ndbn-20m and Ndbn-20 ΔdsmA, and recombinants Ndbn-20m-pBBRdsmA, Ndbn-20m-pBBRdsmABC, Ndbn-20 ΔdsmA-pBBRdsmA, and Ndbn-20 ΔdsmA-pBBRdsmABC. The data were derived from three independent measurements, and the error bars indicate standard deviations.

To investigate whether dsmABC could restore the ability of mutants Ndbn-20m and Ndbn-20 ΔdsmA to degrade 3,6-DCSA, fragments containing dsmA or dsmABC were amplified from the genome of Ndbn-20 and ligated into the broad-host-range plasmid pBBR1MCS-2 to generate the recombinant plasmids pBBRdsmA and pBBRdsmABC, respectively. The two recombinant plasmids were then introduced into the two mutants. The whole-cell transformation results showed that both recombinants, Ndbn-20 ΔdsmA-pBBRdsmABC and Ndbn-20 ΔdsmA-pBBRdsmA, regained the 3,6-DCSA-degrading ability, and there was no significant difference between the 3,6-DCSA degradation rates of the two recombinants and the wild-type strain Ndbn-20 (Fig. 4). During the degradation process, metabolite accumulation was not observed. These results further demonstrated that the P450 monooxygenase system encoded by dsmABC was responsible for the degradation of 3,6-DCSA.

Recombinant Ndbn-20m-pBBRdsmABC, but not Ndbn-20m-pBBRdsmA, also reacquired the ability to degrade 3,6-DCSA. However, its 3,6-DCSA degradation rate was obviously lower than that of the wild-type strain (Fig. 4), possibly because recombinant Ndbn-20m-pBBRdsmABC lacked other genes in the dsm cluster that might be involved in the downstream catabolism of 3,6-DCSA. It is interesting that two metabolites with retention times of 3.63 min and 5.11 min were accumulated during the degradation process (Fig. 5A). MS analysis of metabolite 1 (3.63 min) showed a prominent deprotonated molecular ion peak at m/z 220.9 (M − H)⁻ with a fragment ion peak at m/z 176.9 (loss of a CO₂ group) (Fig. 5B). The molecular ion mass was equal to the theoretical molecular mass of the hydroxylation product of 3,6-DCSA (the molecular mass of 3,6-DCSA is 206 Da). This metabolite had two other peaks at m/z 222.9 and 224.9 because it contained two chlorine atoms, and chlorine has two isotopes, ³⁷Cl and ³⁵Cl, whose natural abundances are 24% and 76%, respectively. MS analysis of metabolite 2 (5.11 min) showed a deprotonated molecular ion peak at m/z 187.0 (M − H)⁻ (Fig. 5C), which was equal to the theoretical molecular mass of the hydroxylation product with a chlorine being replaced by a hydrogen. This result indicated that 3,6-DCSA was initially degraded through hydroxylation, followed by the elimination of a chlorine atom. Furthermore, our results also showed that these two intermediates gradually disappeared with prolonged incubation, indicating that they could be further transformed by Ndbn-20m-pBBRdsmABC (data not shown).

FIG 5.

UHPLC-MS spectra of the metabolites generated during 3,6-DCSA conversion by Ndbn-20m-pBBRdsmABC. (A) UHPLC spectra of the samples obtained at 24 h of incubation. (B and C) Mass spectra of metabolite 1 at 3.63 min (B) and metabolite 2 at 5.11 min (C).

Heterologous expression of dsmA, dsmB, and dsmC in Sphingobium quisquiliarum DC-2.

To further confirm the functions of genes dsmA, dsmB, and dsmC, the recombinant plasmids pBBRdsmA and pBBRdsmABC were transformed into Sphingobium quisquiliarum DC-2, which cannot degrade 3,6-DCSA and salicylate. Recombinant DC-2-pBBRdsmABC acquired the ability to transform 3,6-DCSA, and correspondingly, a hydroxylation product was generated (see Fig. S2 in the supplemental material). This hydroxylation product could not be further transformed by recombinant DC-2-pBBRdsmABC. Recombinant DC-2-pBBRdsmA could not transform 3,6-DCSA (data not shown). These results further demonstrated that dsmABC was responsible for the hydroxylation of 3,6-DCSA.

In addition to 3,6-DCSA, recombinant DC-2-pBBRdsmABC could also hydroxylate salicylate, 3-chlorosalicylate, 4-chlorosalicylate, 6-chlorosalicylate, 3-methylsalicylate, 4-methylsalicylate, and 6-methylsalicylate, indicating that DsmABC had a broad substrate spectrum. However, recombinant DC-2-pBBRdsmABC could not hydroxylate 3,5-DCSA and 5-chlorosalicylate, in which the hydrogen at C-5 has been replaced by a chlorine. The result indicated that the DsmABC-catalyzed hydroxylation might occur at the C-5 position of salicylate or 3,6-DCSA.

Determination of the hydroxylation position in 3,6-DCSA.

3,6-DCSA has two possible sites for hydroxylation (C-4 or C-5). To determine which carbon was hydroxylated, the hydroxylation product transformed by recombinant DC-2-pBBRdsmABC was extracted from the culture, purified using silica gel column chromatography, and subjected to single-crystal X-ray diffraction analysis. According to the acquired crystallographic parameters, bond length, bond angle, and torsion angle data (see Table S1 and Fig. S3 in the supplemental material), C-2 was connected to O-3, and C-5 was connected to O-4. O-3 and O-4 were each linked with an H atom, forming a hydroxyl group. These results clearly showed that the DsmABC-catalyzed hydroxylation occurred at the C-5 of 3,6-DCSA, forming 3,6-dichlorogentisate (3,6-DCGA).

Expression of dsmD and preliminary activity study of the purified DsmD.

DsmD was most related to gentisate and 3,6-DCGA 1,2-dioxygenase GtdA (6). To investigate whether DsmD could transform 3,6-DCGA, the dsmD gene was expressed in Escherichia coli BL21(DE3) using the pET24b(+) expression system. The His₆-tagged DsmD was purified to homogeneity using Co²⁺ chelate affinity chromatography (see Fig. S4 in the supplemental material). Enzymatic assay showed that the purified DsmD could transform gentisate (see Fig. S5A in the supplemental material) and 3-chlorogentisate (Fig. S5B) but not 6-chlorogentisate (Fig. S5C) or 3,6-DCGA (Fig. S5D).

DISCUSSION

3,6-DCSA, a chloro-substituted derivative of salicylate, is the bacterial demethylation product of the herbicide dicamba (3, 4, 5, 7, 8, 9, 12). However, the catabolic pathway of 3,6-DCSA and the genes and enzymes involved remain unknown. In this study, a 3,6-DCSA hydroxylation-deficient mutant, Ndbn-20m, was screened through continuous transfer on 1/5 LB agar. Comparison of the mutant genome with that of the wild-type strain revealed that the gene cluster dsm was lost in mutant Ndbn-20m. The dsm cluster consisted of nine genes, including three consecutive genes, dsmA, dsmB, and dsmC, encoding a three-component monooxygenase system. DsmABC was identified as a 3,6-DCSA 5-hydroxylase based on the following evidence. First, the disruption of dsmA abrogated the ability of strain Ndbn-20 to convert 3,6-DCSA. Second, the introduction of dsmABC into mutants Ndbn-20m and Ndbn-20 ΔdsmA recovered their ability to degrade 3,6-DCSA. In addition, Sphingobium quisquiliarum DC-2 in which dsmABC was introduced acquired the ability to 5-hydroxylate 3,6-DCSA. Finally, the transcription of dsmA, dsmB, and dsmC was significantly induced by 3,6-DCSA. Recombinants Ndbn-20m-pBBRdsmA and DC-2-pBBRdsmA, which lacked electron transport components DsmB and DsmC, could not hydroxylate 3,6-DCSA, indicating that DsmA had a strict specificity for the electron transport components.

Previous studies have reported that the microbial degradation of salicylate and its derivatives was usually initiated by 5-hydroxylation (21, 38), decarboxylative 1-hydroxylation (18–20), and in conjunction with coenzyme A (23). Two salicylate 5-hydroxylases, NagGHAaAb and HybABCD, have been identified, both of which are three-component RHO-type monooxygenases (21, 22, 38). The 3,6-DCSA 5-hydroxylase presented in this study also consisted of three components: a terminal oxygenase encoded by dsmA, a [2Fe-2S] ferredoxin encoded by dsmB, and a ferredoxin reductase encoded by dsmC. However, DsmABC was identified as a cytochrome P450 monooxygenase system but not as a RHO-type monooxygenase. Cytochrome P450s are a widely distributed group of heme-containing monooxygenases that catalyze a wide range of oxidative reactions, such as hydroxylation, N-, O-, and S-dealkylation, dehalogenation, desulfuration, sulfoxidation, deamination, epoxidation, and peroxidation, and contribute to the biosynthesis of various bioactive compounds and the degradation of xenobiotic compounds (39–42). To the best of our knowledge, DsmABC is the only reported cytochrome P450 monooxygenase system that hydroxylates salicylate and its chloro- or methyl-substituted derivatives. Both DsmABC and NagGHAaAb could catalyze the 5-hydroxylation of the 3- and 4-substituted salicylates, such as 3-chlorosalicylate, 4-chlorosalicylate, 3-methylsalicylate, and 4-methylsalicylate. NagGHAaAb could also 6-hydroxylate 5-chlorosalicylate and 5-hydroxylate 5-methylsalicylate, whereas DsmABC could not.

In a phylogenetic tree constructed with the neighbor-joining (NJ) algorithm based on the related cytochrome P450 monooxygenases, DsmA was located within this NJ tree but formed a monophyletic branch (Fig. 6). Furthermore, it is interesting that DsmA was most related to some cytochrome P450 monooxygenases catalyzing the hydroxylation of antibiotics. However, its sequence identities with these cytochrome P450 monooxygenases were very low; e.g., it had only 36% identity with the mevastatin hydroxylase CYP107DY1 (43), 33% identity with the dihydrobacillaene hydroxylase PksS (44), and 30% identity with the 6-deoxyerythronolide B hydroxylase EryF (45). Furthermore, these antibiotics are macromolecular compounds with molecular weights 2- to 4-fold greater than that of 3,6-DCSA. Therefore, these cytochrome P450 monooxygenase systems structurally needed a larger substrate-binding pocket than DsmABC to bind these macromolecular antibiotics. Based on the above analyses, DsmABC was different from the previously reported cytochrome P450 monooxygenase systems.

FIG 6.

Phylogenetic tree constructed based on the alignment of DsmA with related cytochrome P450 monooxygenases. The multiple-alignment analysis was performed with ClustalX 2.1, and the phylogenetic tree was constructed by the neighbor-joining (NJ) algorithm using MEGA 7.0. Numbers at nodes are percent bootstrap values based on 1,000 resampled data sets; values below 50% are not indicated. Bar, 0.10 substitution per nucleotide position. The items are arranged in the following order: protein function, protein name, UniProtKB accession number, and strain name.

Our study revealed that the dsm cluster was absent in the 3,6-DCSA degradation-deficient mutant Ndbn-20m and that genes in the cluster were significantly induced by 3,6-DCSA, indicating that these genes might be involved in 3,6-DCSA catabolism. DsmD shared high identity with the 3,6-DCGA 1,2-dioxygenase GtdA (6); it was therefore predicted to be responsible for the cleavage of 3,6-DCGA in R. dicambivorans Ndbn-20. However, to our surprise, purified DsmD could not cleave 3,6-DCGA in vitro; furthermore, 3,6-DCGA was further degraded by recombinant Ndbn-20m-pBBRdsmABC, although dsmD was absent in this recombinant, indicating that there was another dioxygenase responsible for the cleavage of 3,6-DCGA in R. dicambivorans Ndbn-20. Bioinformatics analysis revealed that one gene (ORF3896) shared 100% identity with gtdA in the genome of R. dicambivorans Ndbn-20; thus, it is possible that gtdA, but not dsmD, was responsible for the 3,6-DCGA cleavage in the strain. DsmF was predicted to be an MFS-type transporter and thus might be responsible for the transmembrane transport of 3,6-DCSA. DsmE shared 52% and 26% identity with fumarylpyruvate hydrolase and maleylpyruvate hydrolase, respectively (35, 46). DsmG shared high identity with the maleylacetate reductase, which catalyzes the reduction of maleylacetate to 3-oxoadipate or the simultaneous 2-dehalogenation and reduction of 2-chloromaleylacetate to 3-oxoadipate (37), so DsmG therefore might be responsible for the dechloridation of 2-chloromaleylpyruvate, a possible downstream intermediate of 3,6-DCSA catabolism. Based on the above analysis, we propose the following catabolic pathway for dicamba in R. dicambivorans Ndbn-20 (Fig. 2C). Dicamba is demethylated by Dmt66, generating 3,6-DCSA, which is then 5-hydroxylated to 3,6-DCGA by DsmABC; 3,6-DCGA is subsequently ring cleaved by GtdA, resulting in a cleavage product that is dechlorinated by an unknown dehalogenase to generate 2-chloromaleylpyruvate; 2-chloromaleylpyruvate is further dechlorinated by DsmG to maleylpyruvate, which is transformed into fumarylpyruvate by an unknown isomerase; and fumarylpyruvate is hydrolyzed to fumarate and pyruvate by DsmE. Considering that DsmE shared low sequence identity with maleylpyruvate hydrolase, it is also possible that maleylpyruvate is directly hydrolyzed to maleate and pyruvate by DsmE. On the other hand, dsmD did not join in the above-proposed pathway. However, dsmD was located in the dsm cluster and was significantly induced by 3,6-DCSA, indicating that it may be involved in 3,6-DCSA catabolism in R. dicambivorans Ndbn-20. Furthermore, a dechlorination product (3-chlorogentisate or 6-chlorogentisate) of 3,6-DCGA was detected during 3,6-DCSA conversion by recombinant Ndbn-20m-pBBRdsmABC, and 3-chlorogentisate was the substrate of DsmD. These findings suggested another possible metabolic route for 3,6-DCGA to 2-chloromaleylpyruvate in R. dicambivorans Ndbn-20: 3,6-DCGA is first 6-dechlorinated to 3-chlorogentisate by an unknown dechlorinase, and 3-chlorogentisate is then cleaved to 2-chloromaleylpyruvate by DsmD (Fig. 2C).

The dsmA, dsmB, dsmC, and dsmABC genes were also cloned into the expression vector pET24b(+) or pET29a(+) to verify their biochemical functions. Regrettably, however, no activity was detected in the lysates of E. coli BL21(DE3) expressing dsmABC or in the mixture of lysates expressing dsmA, dsmB, and dsmC. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis indicated that the dsmA gene was expressed in E. coli BL21 as an insoluble product in the form of an inclusion body (data not shown). Therefore, the enzymatic characteristics of DsmABC remain to be resolved through optimizing the expression conditions for DsmA in the future.

MATERIALS AND METHODS

Chemicals and media.

Dicamba, gentisate, salicylate, and their derivatives were obtained from Molbase; all the chemicals were >98% purity. LB broth and LB agar were purchased from Difco Laboratories (Detroit, MI, USA). One-fifth LB was prepared by diluting LB broth 4-fold with distilled water. Mineral salt medium (MSM) contained (per liter) 1.3 g K₂HPO₄, 0.86 g KH₂PO₄, 0.66 g (NH₄)₂SO₄, 0.097 g MgSO₄, 0.025 g MnSO₄ · H₂O, 0.005 g FeSO₄ · 7H₂O, and 0.0013 g CaSO₄ · 6H₂O (pH 7.0).

Bacterial strains, plasmids, and culture conditions.

The bacterial strains and plasmids used in this study are listed in Table 2, and the primers are listed in Table 3. R. dicambivorans Ndbn-20 and its derivatives were grown aerobically in 1/5 LB broth or on 1/5 LB agar at 30°C, and all E. coli strains were grown aerobically at 37°C in LB broth or on LB agar. When necessary, 100 μg/ml of ampicillin, 100 μg/ml of streptomycin, 50 μg/ml of kanamycin, 30 μg/ml of chloramphenicol, or 50 μg/ml of gentamicin was added to the medium. The degradation of dicamba or 3,6-DCSA was performed in MSM supplemented with 1.0 mM dicamba or 3,6-DCSA according to the whole-cell transformation method described by Wang et al. in 2014 (24). The concentrations of the substrate and intermediate were determined by HPLC as described in “Analytical methods” below.

TABLE 2.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristicsa	Source or reference
Rhizorhabdus dicambivorans strains
Ndbn-20	Degrades dicamba; Smr	This study
Ndbn-20m	Ndbn-20 mutant, unable to degrade 3,6-DCSA	This study
Ndbn-20 ΔdsmA	Ndbn-20 mutant with dsmA gene disrupted	This study
Sphingobium quisquiliarum DC-2 (KACC 17149)	Unable to degrade dicamba, 3,6-DCSA, salicylic acid, and gentisate; Smr	29
E. coli strains
DH5α	Host strain for cloning plasmid	TaKaRa
BL21(DE3)	Host strain for expression plasmid	TaKaRa
HB101(pRK600)	Conjugation helper strain; Cmr	TaKaRa
Plasmids
pBBR1MCS-2	Broad-host-range plasmid; Kmr	56
pBBRdsmA	pBBR1MCS-2 derivative carrying dsmA; Kmr	This study
pBBRdsmABC	pBBR1MCS-2 derivative carrying dsmABC; Kmr	This study
pJQ200SK	Suicide recombinant plasmid; Gmr	55
pJQ200SKdsmA	pJQ200SK derivative carrying a fragment of dsmA; Gmr	This study
pET24b(+)	Expression plasmid; Kmr	Novagen
pET24bdsmD	pET24b(+) derivative carrying dsmD	This study
pMD19-T	T-A cloning plasmid; Ampr	TaKaRa

^aSm^r, streptomycin resistant; Cm^r, chloramphenicol resistant; Km^r, kanamycin resistant; Amp^r, ampicillin resistant, Gm^r, gentamicin resistant.

TABLE 3.

PCR primers used in this study

Primer	DNA sequence (5′→3′)a	Purpose
pBBR-dsmA-KpnI-F	GGGGTACCCCCATCCCCGAAAGCCAGTTCTGACAC	Amplification of a fragment containing dsmA for gene functional verification
pBBR-dsmA-EcoRI-R	CGGAATTCCGGGGCGTGTTTGATCGACGTAGCAG
pBBR-dsmABC-KpnI-F	GGGGTACCCCGCTGGGGAAGGTCTTGGTCGCAT	Amplification of a fragment containing dsmABC for gene functional verification
pBBR-dsmABC-EcoRI-R	CGGAATTCCGCAGGCTCCGGCTTTACGAAATTACCC
pJQ200SK-dsmA-F	CTTGATATCGAATTCCTGCAGCTGGCCAGCGGCAGTTTCAGCGTTC	Amplification of a fragment of dsmA for gene disruption
pJQ200SK-dsmA-R	GCTCTAGAACTAGTGGATCCCCATGGGAAATCGTCCGCGTTGAGG
pET-dsmD-NdeI-F	CGGGAATTCCATATGACTGCTACGTCGATCAAACAC	Amplification of dsmD for expression in E. coli BL21
pET-dsmD-XhoI-R	CCGCTCGAGCGCGCTACGCCAAAGGCCCAG
RT-16S-F	GGCGACGATCCATAGCTGGTCTGAG	Amplification of a 141-bp fragment of 16S rRNA by RT-qPCR
RT-16S-R	TTCATCACTCACGCGGCATTGCTG
RT-dsmA-F	TTCGTACTCAACCCCCATGTGTCGC	Amplification of a 172-bp fragment of dsmA by RT-qPCR
RT-dsmA-R	CGTCACCCTGCTTCTCGGCATACAC
RT-dsmB-F	GAGTCCGACGGAAATGTTCATGAGC	Amplification of a 154-bp fragment of dsmB by RT-qPCR
RT-dsmB-R	TTGCCCATCCCGCATCGACATAG
RT-dsmC-F	GCTCGGTGACGAACGAACGGTGC	Amplification of a 152-bp fragment of dsmC by RT-qPCR
RT-dsmC-R	CGTAGATGTCGGGATCAGACGTGCG
RT-dsmD-F	GCACTGGCACGATCATGGCAATAC	Amplification of a 168-bp fragment of dsmD by RT-qPCR
RT-dsmD-R	TGATTGCCAAAACGCACGAAGCC
RT-dsmE-F	GTGCCGATGGGCGACACGCTCTTC	Amplification of a 165-bp fragment of dsmE by RT-qPCR
RT-dsmE-R	GGTGAAAATGAGGTCGCCTTCGCCA
RT-dsmF-F	AGCGCCACAAAGCCAAGGTTCGTTC	Amplification of a 170-bp fragment of dsmF by RT-qPCR
RT-dsmF-R	GAAGAACACGAACCACAGCGTCAGC
RT-dsmG-F	GGCAAAGCCGCAGGTGGTGATCTAC	Amplification of a 180-bp fragment of dsmG by RT-qPCR
RT-dsmG-R	AGGCCGTCGAAAAAGGCTTCTATGC
RT-dsmR1-F	CTGCGCGATGCGACAGGGGAAAC	Amplification of a 196-bp fragment of dsmR1 by RT-qPCR
RT-dsmR1-R	CAGCCACATCCCTTATCTGCACGG
RT-dsmR2-F	GAAAAAGCCTCGATCATGGCCGGAC	Amplification of a 155-bp fragment of dsmR2 by RT-qPCR
RT-dsmR2-R	GCCGATAAGTTGTACCAGGCCCAGG
RT1-F	GACGGAAATCGCTAAAATCGCAGGC	Amplification of a 513-bp fragment of dsmR1 by RT-PCR
RT1-R	AAGCGTCCATTACGGGAGCAGCG
RT2-F	TCCAGATCGACATCATGGCGGTTTC	Amplification of 1,113 bp of the dsmR1-dsmG-spanning region
RT2-R	CAGGCCGTCGAAAAAGGCTTCTATG
RT3-F	GCACTTGAAGGCGGCTTTTATGGCG	Amplification of 967 bp of the dsmG-dsmF-spanning region
RT3-R	CTAATGAGCGGCGATGAGGCAGCAC
RT4-F	GTTCATAGCACTGCTCGGTTTCGCC	Amplification of 873 bp of the dsmF-dsmE-spanning region
RT4-R	CGACGAGATGTCACTGCTCTGCTTG
RT5-F	CGACGATGCCAAGCAGAGCAGTGAC	Amplification of 884 bp of the dsmE-dsmC-spanning region
RT5-R	ACTATAACGAGCTTGCTACCGGCGC
RT6-F	ATGGCAGCGTCATCCGCCTCGAATC	Amplification of 494 bp of the dsmC-dsmB-spanning region
RT6-R	TTGCCCATCCCGCATCGACATAGAC
RT7-F	GACCTGCCACGTCTATGTCGATGCG	Amplification of 757 bp of the dsmB-dsmA-spanning region
RT7-R	CCCCCAATGTGTGGAGGTAGTAGCC
RT8-F	GGTGTATGCCGAGAAGCAGGGTGAC	Amplification of 972 bp of the dsmA-dsmD-spanning region
RT8-R	CTCCCTCGATGATGAGACGCAGTGC
RT9-F	GGTTCTCGACACCATCTCCTGCTTC	Amplification of 578 bp of the dsmD-dsmR2-spanning region
RT9-R	AGGCACCATCGGCGAAAATACAAGC
RT10-F	TCGGCATCTTCGCATTTCAAGCAGC	Amplification of 673 bp of the dsmR2-orf1-spanning region
RT10-R	CAGAGTCGGGTGCTTTCCCATATCG
RT11-F	CCAGTCCATGTTCGCAGACCGCAAG	Amplification of a 486-bp fragment of orf1 by RT-PCR
RT11-R	CCGATATGCACCATGAAGGCTCCG

^aRestriction sites and overlapped bases are underlined.

Screening of the 3,6-DCSA degradation-deficient mutant.

A lawn of R. dicambivorans Ndbn-20 grown on 1/5 LB agar was streaked onto fresh 1/5 LB agar and incubated at 30°C for 4 days. After approximately 30 continuous transfers, the lawn was collected and suspended in MSM. After the suspended cells were serially diluted, they were spread onto 1/5 LB agar and incubated until colonies developed well. Each colony was then simultaneously inoculated using a sterilized toothpick onto 1/5 LB agar and MSM agar supplemented with 1.0 mM 3,6-DCSA. Colonies that grew on 1/5 LB agar but failed to grow on MSM agar with 3,6-DCSA were picked up, and their ability to degrade 3,6-DCSA in liquid MSM was determined.

Genome sequencing and bioinformatics analysis.

The draft genome of R. dicambivorans Ndbn-20 has been sequenced by Yao et al. (8). To sequence the complete genome of wild-type R. dicambivorans Ndbn-20 and the draft genome of mutant Ndbn-20m, the genomic DNAs of the two strains were extracted according to the method of Sambrook and Russell (47). The complete genome of Ndbn-20 was sequenced by single-molecule real-time (SMRT) technology (48). Low-quality reads were filtered by using SMRT Analysis 2.3.0, and the filtered reads were assembled to generate one contig without gaps (49). The draft genome of mutant Ndbn-20m was sequenced by massively parallel sequencing (MPS) Illumina technology, and the reads were assembled by SOAPdenovo (50). The functional annotation was accomplished by BLAST analysis in the nonredundant protein (NR), Swiss-Prot, KEGG, and COG databases. To identify the lost fragment, the draft genome of Ndbn-20m was compared with the complete genome of Ndbn-20 using the dot plot analysis in OMIGA 2.0 software (51) and the RAST online analysis tool with the sequence-based model (52). DNA and amino acid sequence identity searches were conducted using the BLASTN and BLASTP tools (https://blast.ncbi.nlm.nih.gov/Blast.cgi). For phylogenetic analysis, all protein sequences were aligned using ClustalX 2.1, and the phylogenetic tree was constructed by the NJ algorithm using MEGA 7.0 (53). Kimura's two-parameter model was used to calculate the distances, and bootstrap values were set as 1,000 replications for analysis.

RNA preparation and transcription analysis.

Ndbn-20 cells were cultured to mid-logarithmic phase in 1/5 LB broth, harvested by centrifugation, washed twice with fresh MSM, suspended in MSM containing 1.0 mM 3,6-DCSA or 1.0 mM glucose, and then incubated at 30°C. The cells were harvested until approximately 50% of the added 3,6-DCSA was degraded. Total RNA was extracted from the cells using a MiniBEST universal RNA extraction kit (TaKaRa) according to the manufacturer's protocols. RNA integrity was confirmed with a NanoDrop2000 spectrophotometer (Thermo Fisher Scientific). The RNA was incubated at 42°C with gDNA Eraser (TaKaRa) for 2 min to remove any genomic DNA contamination. Reverse transcription was performed with a PrimeScript reverse transcriptase kit (TaKaRa). The acquired cDNA was used for RT-PCR with the primers listed in Table 3. RT-qPCR was performed with a SYBR Premix Ex Taq RT-PCR kit (Tli RNaseH Plus; TaKaRa) on an Applied Biosystems 7300 real-time PCR system (Applied Biosystems, CA, USA). The 16S rRNA gene was set as an internal reference to evaluate the relative difference in integrity between individual RNA samples. Three independent biological replicates were assayed under each experimental condition. The 2^−ΔΔCT method was used to calculate relative expression levels (54).

Disruption of dsmA in R. dicambivorans Ndbn-20 by homologous recombination.

The knockout plasmid pJQ200SKdsmA for gene disruption was constructed by fusing a 561-bp fragment of dsmA to the PstI-BamHI site of the suicide plasmid pJQ200SK (55) with one-step cloning kit (Vazyme Biotech, Nanjing, China). pJQ200SKdsmA then was introduced from E. coli DH5α into R. dicambivorans Ndbn-20 by triparental conjugative transfer with E. coli HB101(pRK600) as the helper. The candidate mutant was screened on 1/5 LB agar containing 100 μg/ml of streptomycin and 15 μg/ml of gentamicin. The disruption of dsmA in the mutant was verified by PCR and DNA sequencing. The resulting mutant was designated Ndbn-20 ΔdsmA.

Function study of dsmABC in Sphingobium quisquiliarum DC-2 and mutants Ndbn-20m and Ndbn-20 ΔdsmA.

A 1,378-bp DNA fragment containing dsmA and a 3,102-bp DNA fragment containing dsmABC were amplified from the genome of strain Ndbn-20 using the primers listed in Table 3. The two fragments were each ligated into the KpnI-EcoRI site of the broad-host-range plasmid pBBR1MCS-2 (56). The resulting plasmids, pBBRdsmA and pBBRdsmABC, were introduced into Sphingobium quisquiliarum DC-2, Ndbn-20m, and Ndbn-2 0ΔdsmA by triparental conjugative transfer. The abilities of these transformants to degrade salicylate and its derivatives (3,6-DCSA, 3,5-DCSA, 3-chlorosalicylate, 4-chlorosalicylate, 5-chlorosalicylate, 6-chlorosalicylate, 3-methylsalicylate, 4-methylsalicylate, and 6-methylsalicylate) were determined using the whole-cell transformation method as described above. The intermediate generated was identified by ultra-high-performance liquid chromatography–tandem mass spectrometry (UHPLC-MS) analysis as described below.

Analytical methods.

The culture samples were acidified to pH 2.0 with 0.5% HCl and extracted twice with ethyl acetate. The ethyl acetate solvent was then removed with a stream of nitrogen gas. The extracts were redissolved in methanol, filtered through a 0.22-μm Millipore membrane, and used for HPLC analysis on an UltiMate 3000 titanium system (Thermo Fisher Scientific) equipped with a C₁₈ reversed-phase column (4.6 by 250 mm, 5 μm; Agilent Technologies). The mobile phase was a mixture of water, acetonitrile, methanol, acetic acid, and phosphoric acid (58.4:31.7:7.5:2.4:0.2, vol/vol/vol/vol/vol) and flowed at 1 ml/min. A VWD-3100 single-wavelength detector was used to monitor the UV absorption; the detection wavelengths were 275 nm for dicamba, 290 nm for 3,6-DCGA, 319 nm for 3,6-DCSA, and 330 nm for salicylate and its derivatives. UHPLC-MS analysis was performed on a Hypersil Gold C₁₈ column (2.10 by 100 mm, 3 μm; Thermo Fisher Scientific). The mobile phase consisted of water (0.1% formic acid) (A) and acetonitrile (B) at a flow rate of 0.4 ml/min. A stepped solvent gradient was used as follows: 0 to 1.50 min, 10% B; 1.50 to 14.00 min, 10% to 95% B; 14.00 to 16.00 min, 95% B; 16.00 to 16.50 min, 95% to 10% B; and 16.50 min to 20.00 min, 10% B. Electrospray ionization (ESI)-MS spectra were recorded in the negative ionization mode on an LTQ Orbitrap XL mass spectrometer (Thermo Fisher Scientific) equipped with an ESI probe.

Determination of the DsmABC-catalyzed hydroxylation position in 3,6-DCSA by single-crystal X-ray diffraction.

Cells of recombinant DC-2-pBBRdsmABC were inoculated into MSM containing 1.0 mM 3,6-DCSA, and the culture was then incubated at 30°C until almost all the added 3,6-DCSA was transformed. The hydroxylation product was then extracted from the culture with ethyl acetate and purified using silica gel column chromatography (3 by 40 cm). The mobile phase consisted of chloroform-ethyl acetate-formic acid (10:8:1), and the eluate was collected in 10-ml fractions. In single-crystal X-ray diffraction analysis, single crystals of the hydroxylation product were obtained by slow evaporation of the solvent acetonitrile at room temperature. A suitable single crystal was carefully selected under an optical microscope and glued onto glass fibers. Single-crystal data were collected on a Bruker D8 diffractometer. The calculation was performed using the SHELXTL97 crystallographic software package, and the CIF file was checked using the free online checkCIF service provided by the International Union of Crystallography (http://checkcif.iucr.org/).

Heterologous expression of dsmD and purification of the His₆-tagged DsmD.

The gene dsmD was amplified from the genomic DNA of R. dicambivorans Ndbn-20 using the primers listed in Table 3 and PrimeSTAR GXL DNA polymerase. The PCR product was digested with NdeI-XhoI and ligated into the corresponding site of pET24b(+). The resulting plasmid, pET24bdsmD, was then introduced into E. coli BL21(DE3). E. coli BL21(DE3)-pET24bdsmD was grown in LB at 37°C to an optical density of 0.4 at 600 nm and then induced with 0.05 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at 16°C for an additional 8 h. Cell lysate was obtained by sonication, and the C-terminal His₆-tagged protein was purified using a 1-cm³ Co²⁺-charged resin column (HiTrap Talon crude; GE Healthcare Life Sciences) according to the method described by Cheng et al. (28). The protein concentration was determined by the Bradford method, with bovine serum albumin as the standard. The activity of DsmD was qualitatively determined by the UV scanning method as described by Zhou et al. (35).

Accession number(s).

The dsm cluster sequence, the chromosome genome of R. dicambivorans Ndbn-20, plasmids P1, P2, P3, P4, and the draft genome of mutant Ndbn-20 have been deposited in the GenBank database under accession numbers MG494382, CP023449, CP023450, CP023451, CP023452, CP023453 and NWUF00000000, respectively.