Three-Component O-Demethylase System Essential for Catabolism of a Lignin-Derived Biphenyl Compound in Sphingobium sp. Strain SYK-6

Abstract

Sphingobium sp. strain SYK-6 is able to assimilate lignin-derived biaryls, including a biphenyl compound, 5,5′-dehydrodivanillate (DDVA). Previously, ligXa (SLG_07770), which is similar to the gene encoding oxygenase components of Rieske-type nonheme iron aromatic-ring-hydroxylating oxygenases, was identified to be essential for the conversion of DDVA; however, the genes encoding electron transfer components remained unknown. Disruption of putative electron transfer component genes scattered through the SYK-6 genome indicated that SLG_08500 and SLG_21200, which showed approximately 60% amino acid sequence identities with ferredoxin and ferredoxin reductase of dicamba O-demethylase, were essential for the normal growth of SYK-6 on DDVA. LigXa and the gene products of SLG_08500 (LigXc) and SLG_21200 (LigXd) were purified and were estimated to be a trimer, a monomer, and a monomer, respectively. LigXd contains FAD as the prosthetic group and showed much higher reductase activity toward 2,6-dichlorophenolindophenol with NADH than with NADPH. A mixture of purified LigXa, LigXc, and LigXd converted DDVA into 2,2′,3-trihydroxy-3′-methoxy-5,5′-dicarboxybiphenyl in the presence of NADH, indicating that DDVA O-demethylase is a three-component monooxygenase. This enzyme requires Fe(II) for its activity and is highly specific for DDVA, with a K_m value of 63.5 μM and k_cat of 6.1 s⁻¹. Genome searches in six other sphingomonads revealed genes similar to ligXc and ligXd (>58% amino acid sequence identities) with a limited number of electron transfer component genes, yet a number of diverse oxygenase component genes were found. This fact implies that these few electron transfer components are able to interact with numerous oxygenase components and the conserved LigXc and LigXd orthologs are important in sphingomonads.

INTRODUCTION

Lignin, which accounts for 15 to 40% of the chemical components in plant cell walls (1), is the most abundant aromatic substance in nature; hence, its degradation is crucial for the earth's carbon cycle. It is thought that lignin peroxidase, manganese peroxidase, versatile peroxidase, and laccase, secreted by white rot fungi, attack mainly native lignin (2, 3). The lignin oligomers produced are further degraded and mineralized by bacteria (4). Since native lignins consist of various intermonomer linkages such as β-aryl ether, phenylcoumaran, biphenyl, and pinoresinol (5), bacteria appear to have a wide variety of enzymatic systems to degrade lignin oligomers. In regard to the catabolism of lignin-derived biaryls and monoaryls by bacteria, the catabolic pathways and catabolism genes for β-aryl ether, biphenyl, ferulate, vanillin, and syringate have been extensively characterized in an alphaproteobacterium, Sphingobium sp. strain SYK-6, which was isolated from a pond for the treatment of waste liquor from a kraft pulp mill (6, 7).

Among the intermonomer linkages in lignin, the biphenyl structure was estimated to account for approximately 5 to 7% in a Norway spruce milled-wood lignin (8–10). This structure is considered to be recalcitrant because of its C-C bond between the benzene rings. To date, the SYK-6 catabolic pathway of a biphenyl compound, 5,5′-dehydrodivanillate (DDVA), and almost all the genes involved in the conversion of DDVA into vanillate have been characterized (Fig. 1A). In SYK-6 cells, DDVA is initially O demethylated to generate 2,2′,3-trihydroxy-3′-methoxy-5,5′-dicarboxybiphenyl (OH-DDVA) (6, 11), and one of the two aromatic rings of OH-DDVA is cleaved by OH-DDVA dioxygenase, LigZ (12). The resulting meta-ring cleavage product is hydrolyzed by LigY (13), and then the product, 5-carboxyvanillate (5CVA), is converted by LigW and LigW2 decarboxylases into vanillate (14, 15), which is then further degraded to pyruvate and oxaloacetate via the protocatechuate (PCA) 4,5-cleavage pathway (7, 16). As mentioned above, all the genes involved in the degradation of OH-DDVA have been characterized; however, the enzyme genes for the first step of DDVA degradation, O demethylation, are still unknown.

FIG 1.

Catabolic pathway of 5,5′-dehydrodivanillate and its catabolism genes in Sphingobium sp. strain SYK-6. (A) Diagram of the DDVA catabolic pathway in SYK-6. Enzymes: LigXa, oxygenase component of DDVA O-demethylase; LigZ, OH-DDVA dioxygenase; LigY, meta-cleavage compound hydrolase; LigW and LigW2, 5-CVA decarboxylase; LigM, vanillate/3-O-methylgallate O-demethylase; LigA and LigB, small and large subunits of PCA 4,5-dioxygenase; LigC, 4-carboxy-2-hydroxymuconate-6-semialdehyde dehydrogenase; LigI, 2-pyrone-4,6-dicarboxylate hydrolase; LigU, 4-oxalomesaconate tautomerase; LigJ, 4-oxalomesaconate hydratase; LigK, 4-carboxy-4-hydroxy-2-oxoadipate aldolase/oxaloacetate decarboxylase. Abbreviations: DDVA, 5,5′-dehydrodivanillate; OH-DDVA, 2,2′,3-trihydroxy-3′-methoxy-5,5′-dicarboxybiphenyl; 5CVA, 5-carboxyvanillate; PCA, protocatechuate. (B) Locations of the genes involved in the DDVA catabolism in the SYK-6 genome. SLG_08500 (ligXc) and SLG_21200 (ligXd), characterized in this study, encode ferredoxin and ferredoxin reductase, respectively, of the components of DDVA O-demethylase. Each gene is indicated by an arrow. (C) Proposed reaction catalyzed by DDVA O-demethylase.

In a previous study, ligXa, which complemented the growth deficiency on DDVA of an SYK-6 mutant, was identified (11). The deduced amino acid sequence of ligXa showed 26 to 27% identities with the sequences of oxygenase component genes of phenoxybenzoate dioxygenase of Pseudomonas pseudoalcaligenes POB310 (pobA) (17), 3-chlorobenzoate-3,4-dioxygenase of Comamonas testosteroni BR60 (cbaA) (18), and phthalate 4,5-dioxygenase of Pseudomonas putida (pht3) (19). Based on Batie's classification of Rieske nonheme iron aromatic ring-hydroxylating oxygenases (RHOs) (20), all these dioxygenases belong to class IA, which consists of an oxygenase component containing an iron-binding site and a Rieske-type [2Fe-2S] cluster and a reductase containing a flavin and a [2Fe-2S] redox center. Therefore, electron transfer components (ETCs) are necessary for DDVA O-demethylase in addition to LigXa. To date, there have been no reports on O-demethylases for biphenyl compounds. However, a class IA dioxygenase, VanAB, which consists of oxygenase and reductase components, is known to be involved in O demethylation of vanillate in Pseudomonas, Acinetobacter, Streptomyces, and Rhodococcus (21–26). In contrast, SYK-6 employs a different type of vanillate and syringate O-demethylases, LigM and DesA, which catalyze the transfers of methyl moieties of the methoxyl groups of vanillate and syringate, respectively, to tetrahydrofolate (H₄folate) (27, 28). On the basis of the amino acid sequence similarity of LigXa with class IA oxygenases, a reductase component gene was expected to exist proximal to ligXa since in general vanA and vanB are tandemly arranged. The ligXa gene is located approximately 3.3 kb downstream of ligZ-orf1-orf2-ligY and located approximately 6.6 kb downstream of ligW and 570 kb upstream of ligW2 (Fig. 1B) (29). However, genes encoding ETCs were not found in the vicinity of ligXa and other DDVA catabolism genes.

In this study, in order to thoroughly characterize DDVA O-demethylase of SYK-6, we searched for possible genes encoding ETCs of DDVA O-demethylase in the SYK-6 genome (29). Based on the gene disruption analyses and in vitro reconstitution of the DDVA O-demethylase activity with the products of the candidate genes, the essential ETC genes responsible for the DDVA conversion were identified. Using purified components, DDVA O-demethylase was enzymatically characterized.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

The strains and plasmids used in this study are listed in Table 1. Sphingobium sp. strain SYK-6 and its mutants were grown in lysogeny broth (LB; 10 g/liter Bacto tryptone, 5 g/liter yeast extract, and 5 g/liter NaCl) and Wx minimal medium (30) containing 10 mM vanillate or 5 mM DDVA at 30°C. When necessary, 50 mg of kanamycin (Km)/liter, 12.5 mg of tetracycline (Tet)/liter, and 300 mg of carbenicillin/liter were added to the cultures. Escherichia coli strains were grown in LB at 37°C. For cultures of cells carrying antibiotic resistance markers, the media for E. coli transformants were supplemented with 100 mg of ampicillin/liter, 25 mg of Km/liter, or 12.5 mg of Tet/liter.

TABLE 1.

Bacterial strains and plasmids used in this study

Strains and plasmids	Relevant characteristic(s)a	Source or reference
Strains
Sphingobium sp.
SYK-6	Wild type; Nalr Smr	68
SME049	SYK-6 derivative; ligXa::kan; Nalr Smr Kmr	This study
SME051	SYK-6 derivative; SLG_18830::kan; Nalr Smr Kmr	This study
SME052	SYK-6 derivative; ligXd::tet; Nalr Smr Tcr	This study
SME053	SME051 derivative; ligXd::tet; Nalr Smr Kmr Tcr	This study
SME073	SYK-6 derivative; ligXc::tet; Nalr Smr Tcr	This study
SME074	SYK-6 derivative; SLG_18840::kan; Nalr Smr Kmr	This study
SME075	SME073 derivative; SLG_18840::kan; Nalr Smr Tcr Kmr	This study
Escherichia coli
JM109	recA1 supE44 endA1 hsdR17(rK− mK+) gyrA96 relA1 thi-1 Δ(lac-proAB) [F′ traD36 proAB+ lacIq lacZΔM15]	69
BL21(DE3)	F− ompT hsdSB(rB− mB−) gal dcm (DE3); T7 RNA polymerase gene under the control of the lacUV promoter	70
Plasmids
pT7Blue	Cloning vector; Apr	Novagen
pIK03	Kanamycin cassette; Apr Kmr	27
pKRP12	Tetracycline cassette; Apr Tcr	71
pK18mobsacB	oriT sacB Kmr	47
pET-21a(+)	Expression vector, T7 promoter, Apr	Novagen
pET-16b	Expression vector, T7 promoter, Apr	Novagen
pJB864	RK2 broad-host-range expression vector; Apr Cbr Pm xylS	50
pT7XaU	pT7Blue with a 1.0-kb PCR-amplified fragment carrying upstream region of ligXa	This study
pT7XaD	pT7Blue with a 0.8-kb PCR-amplified fragment carrying downstream region of ligXa	This study
pT7XaUD	pT7XaU with a 0.8-kb EcoRV-XbaI fragment of pT7XaD	This study
pT7XaUDK	pT7XaUD with a 1.3-kb EcoRV fragment carrying kan of pIK03 into the same site	This study
pKXaK	pK18mobsacB with a 3.0-kb EcoRI-XbaI fragment of pT7XaUDK	This study
pT71883U	pT7Blue with a 0.4-kb PCR-amplified fragment carrying upstream region of SLG_18830	This study
pT71883D	pT7Blue with a 0.4-kb PCR-amplified fragment carrying downstream region of SLG_18830	This study
pT71883UD	pT71883U with a 0.4-kb EcoRV-XbaI fragment of pT71883D	This study
pK1883	pK18mobsacB with a 0.8-kb EcoRI-XbaI fragment of pT71883UD	This study
pK1883K	pK1883 with a 1.3-kb EcoRV fragment carrying kan of pIK03 into the same site	This study
pT7XdU	pT7Blue with a 1.2-kb PCR-amplified fragment carrying upstream region of ligXd	This study
pT7XdD	pT7Blue with a 1.0-kb PCR-amplified fragment carrying downstream region of ligXd	This study
pT7XdUD	pT7XdU with a 1.0-kb EcoRV-SalI fragment of pT7XdD	This study
pKXd	pK18mobsacB with a 1.9-kb EcoRI-XbaI fragment of pT7XdUD	This study
pKXdT	pKXd with a 1.9-kb SmaI fragment carrying tet of pKRP12 into the EcoRV site	This study
pT7XcU	pT7Blue with a 0.6-kb PCR-amplified fragment carrying upstream region of ligXc	This study
pT7XcD	pT7Blue with a 0.8-kb PCR-amplified fragment carrying downstream region of ligXc	This study
pT7XcUD	pT7XcU with a 0.8-kb EcoRV-HindIII fragment of pT7XcD	This study
pT7XcUDT	pT7XcUD with a 1.9-kb blunt-ended HindIII fragment carrying tet of pKRP12 into the EcoRV site	This study
pKXcT	pK18mobsacB with a 3.3-kb EcoRI-HindIII fragment of pT7XcUDT	This study
pT71884U	pT7Blue with a 0.9-kb PCR-amplified fragment carrying upstream region of SLG_18840	This study
pT71884D	pT7Blue with a 0.7-kb PCR-amplified fragment carrying downstream region of SLG_18840	This study
pT71884UD	pT71884U with a 0.7-kb EcoRV-XbaI fragment of pT71884D	This study
pT71884UDK	pT71884UD with a 1.3-kb EcoRV fragment carrying kan of pIK03 into the same site	This study
pK1884K	pK18mobsacB with a 3.3-kb BamHI-XbaI fragment of pT71884UDK	This study
pTXa	pT7Blue with a 1.3-kb PCR-amplified fragment carrying ligXa	This study
pET21Xa	pET-21a(+) with a 1.3-kb NdeI-BamHI fragment carrying ligXa of pTXa	This study
pJBXa	pJB864 with a 1.4-kb KpnI-HindIII fragment carrying ligXa of pET21Xa	This study
pTXd	pT7Blue with a 1.3-kb PCR-amplified fragment carrying ligXd	This study
pET21Xd	pET-21a(+) with a 1.3-kb NdeI-BamHI fragment carrying ligXd of pTXd	This study
pJBXd	pJB864 with a 1.4-kb KpnI-HindIII fragment carrying ligXd of pET21Xd	This study
pTXc	pT7Blue with a 0.4-kb PCR-amplified fragment carrying ligXc	This study
pET16Xc	pET-16b with a 0.4-kb NdeI-BamHI fragment carrying ligXc of pTXc	This study
pJBXc	pJB864 with a 0.5-kb XbaI (blunt-ended)-BamHI fragment carrying His tag-fused ligXc of pET16Xc into the HindIII (blunt-ended)-BamHI site	This study
pET16Xa	pET-16b with a 1.3-kb NdeI-BamHI fragment carrying ligXa of pTXa	This study
pET16Xd	pET-16b with a 1.3-kb NdeI-BamHI fragment carrying ligXd of pTXd	This study

^aAbbreviations: Nal^r, Sm^r, Km^r, Tc^r, Ap^r, and Cb^r, resistance to nalidixic acid, streptomycin, kanamycin, tetracycline, ampicillin, and carbenicillin, respectively.

Preparation of substrates.

DDVA and OH-DDVA were chemically synthesized as described previously (12). Vanillate, syringate, vanillin, syringaldehyde, acetovanillone, ferulate, and dicamba were purchased from Tokyo Chemical Co. Ltd. or Wako Pure Chemical Ind. Ltd.

Sequence analysis.

Sequence analysis was performed with the MacVector program (MacVector, Inc.). Sequence similarity searches were carried out using the BLAST program (31). For BLAST searches for ETC genes in the SYK-6 genome (AP012222 and AP012223), we used amino acid sequences of reductase of phthalate dioxygenase from Burkholderia cepacia DBO1 (OphA1_DBO1; class IA) (32), reductase of benzoate 1,2-dioxygenase from Acinetobacter baylyi ADP1 (BenC_ADP1; class IB) (33), ferredoxin reductase and ferredoxin of carbazole 1,9a-dioxygenase from Novosphingobium sp. strain KA1 (CarAdI_KA1 and CarAcI_KA1; class IIA) (34), ferredoxin reductase and ferredoxin of biphenyl dioxygenase from Acidovorax sp. strain KKS102 (BphA4_KKS102 and BphA3_KKS102; class IIB) (35, 36), and ferredoxin reductase and ferredoxin of naphthalene 1,2-dioxygenase from Pseudomonas sp. strain NCIB 9816-4 (NahAa_NCIB 9816-4 and NahAb_NCIB 9816-4; class III) (37). Pairwise and multiple alignments were performed with the EMBOSS alignment tool (38) and the ClustalW2 program (39), respectively. Phylogenetic trees were generated using the FigTree program (http://tree.bio.ed.ac.uk/software/figtree/). In order to examine the presence of orthologs of ligXc and ligXd together with other ETCs in six other completely sequenced sphingomonad genomes, i.e., Novosphingobium aromaticivorans DSM 12444 (GenBank accession numbers CP000248, CP000676, and CP000677), Novosphingobium sp. strain PP1Y (40), Sphingobium chlorophenolicum L-1 (41), Sphingobium japonicum UT26S (42), Sphingomonas wittichii RW1 (43), and Sphingopyxis alaskensis RB2256 (44), Conserved Domain Architecture Retrieval Tools (CDART) (45) were used for exploring the genes that encode putative ferredoxins containing the fer2 domain (cd00207), putative ferredoxin reductases containing a set of pyr_redox (cl15766) and Reductase_C (cl20683) domains, and putative reductases containing an FNR-like domain (cl06868) with or without the fer2 domain. The UCSF Chimera software (46) was employed to compare the similarity of the amino acid sequences between the predicted surfaces of Arx (3LXF) and putative sphingomonad ferredoxins (see Fig. S10 in the supplemental material), and between the predicted surfaces of ArR (3LXD) and putative sphingomonad ferredoxin reductases (see Fig. S10).

Construction of mutants.

For construction of disruption plasmids for ligXa, SLG_18830, SLG_21200, SLG_08500, and SLG_18840 (Table 1), upstream and downstream regions of each gene were PCR amplified using the primer sets shown in Table 2. After determining the nucleotide sequences of the PCR-amplified fragments with a CEQ2000XL genetic analysis system (Beckman Coulter, Inc.), these fragments were cloned into pK18mobsacB (47) with the insertion of a Km resistance gene or a Tet resistance gene. The resulting plasmids were independently introduced into SYK-6 cells by electroporation, and candidate mutants were isolated as described previously (48, 49). The disruption of each gene was examined by Southern hybridization analysis using the digoxigenin system (Roche). Complementary plasmids pJBXa, pJBXd, and pJBXc (Table 1) constructed using pJB864 (50) were independently introduced into cells of strains SME049 (ligXa mutant strain), SME052 (SLG_21200 mutant strain), and SME073 (SLG_08500 mutant strain), respectively, by electroporation. The growth of the resulting transformant cells in Wx medium containing 5 mM DDVA was examined.

TABLE 2.

Primer sequences used for construction of plasmids

Plasmid	Primer	Sequence (5′ to 3′)a
pT7XaU	ligX-AF	GAATTCTACGAGCGTGGAAATCAGGG (EcoRI)
	ligX-AR	GATATCATAGACAACAAACAAGGTCAAGCC (EcoRV)
pT7XaD	ligX-BF	GATATCGACGATGAGACGATGATGAGC (EcoRV)
	ligX-BR	TCTAGACCTCTGTTGCCTTGGACTCTTG (XbaI)
pT71883U	113fragAF	GAATTCGCCGACATCGTCATTGTTGG (EcoRI)
	113fragAR	GATATCGTCATCATCTGGTCCACATCTGC (EcoRV)
pT71883D	113fragBF	GATATCGACTGTGTCAATCAGGTCAAGGAC (EcoRV)
	113fragBR	TCTAGACCGAAACCAAAAGAGCCGTAG (XbaI)
pT7XdU	377fragA2F	TCTAGACTTCGTCAGCGACACACTG (XbaI)
	377fragA2R	GATATCGATAATCCTTGGAGAGGGGC (EcoRV)
pT7XdD	377fragB2F	GATATCTTATGATGCCGTGCCGTG (EcoRV)
	377fragB2R	GTCGACCAGAGCAGGAAGAGAATGG (SalI)
pT7XcU	Xc1UF	GAATTCAACAGCAGGTAGATTTCCGC (EcoRI)
	Xc1UR	GATATCCGAATGGCTTCCATGACC (EcoRV)
pT7XcD	Xc1DF	GATATCATCTGCTCGACAGCTCCG (EcoRV)
	Xc1DR	AAGCTTGGATCGATGCTGGACGG (HindIII)
pT71884U	Xc2UF	GGATCCGATGGTCTTTTCCGTTTTCG (BamHI)
	Xc2UR	GATATCCATGATCGCGGATAACTTCC (EcoRV)
pT71884D	Xc2DF	GATATCATCGCTCCGGAAGACTAAGG (EcoRV)
	Xc2DR	TCTAGAGGGAATAACCTCTCCGTTCG (XbaI)
pTXa	XaEXU	CATATGCTGTCCGCAGAGC (NdeI)
	XaEXD	CCAGTCTGATGCCGAGAGA
pTXc	Xc1EXU	CATATGGCGCAGCTGAAGGTCG (NdeI)
	Xc1EXD	ACGCGCGTGGACTGAATCG
pTXd	Xd1EXU	CATATGCCGCATTTTGATTGCC (NdeI)
	Xd1EXD	GGACGGGCGATGGGACCTTA

^aEngineered restriction sites are indicated by bold type, and the corresponding restriction enzymes are shown in parentheses.

Expression of His tag-fused ligXa, ligXc, and ligXd in E. coli and purification.

The ligXa, ligXc, and ligXd genes were PCR amplified using primer sets (Table 2). After the confirmation of the nucleotide sequences, the resulting fragments were digested with NdeI-BamHI and then cloned into the same sites of pET-16b to obtain pET16Xa, pET16Xc, and pET16Xd (Table 1). These plasmids were independently introduced into E. coli BL21(DE3). The transformants were grown in 100 ml of LB medium at 30°C. When the absorbance at 600 nm of the cultures reached 0.5, expressions of the genes were induced for 4 h by adding 1 mM isopropyl-β-d-thiogalactopyranoside. Cells were then washed with 50 mM KH₂PO₄-KOH buffer containing 100 μM Fe(NH₄)₂(SO₄)₂ · 6H₂O and 2 mM cysteine hydrochloride (FE22 buffer, pH 6.5). Cells resuspended in the same buffer were broken by an ultrasonic disintegrator (UD-201; Tomy Seiko Co.). After centrifugation (19,000 × g for 10 min at 4°C), the supernatants were obtained as crude extracts. For purification, crude extracts were applied to His Spin Trap columns (GE Healthcare) previously equilibrated with FE22 buffer (pH 6.0). After centrifugation at 100 × g for 30 s, samples were washed with FE22 buffer (pH 6.0) containing 100 mM imidazole and 500 mM NaCl. Proteins were then eluted with FE22 buffer (pH 6.0) containing 500 mM imidazole and 500 mM NaCl, and the resultant fractions were subjected to desalting and centrifugal filtration with Amicon Ultra 30k (LigXa and LigXd) and 10k (LigXc) (Millipore). The protein concentration was determined by the Bradford method with bovine serum albumin as the standard (51).

Identification of the reaction product.

A mixture of purified LigXa (20 μg of protein), LigXc (20 μg of protein), and LigXd (20 μg of protein) was incubated with 100 μM DDVA in FE22 buffer (pH 6.5) in the presence and absence of NADH at 30°C. Portions of the reaction mixtures were periodically collected and analyzed by high-performance liquid chromatography (HPLC; Acquity UPLC system; Waters) coupled with an Acquity TQ detector (Waters) using a TSKgel ODS-140HTP column (2.1 by 100 mm; Tosoh) as described previously (52). The mobile phase of the HPLC system was composed of a mixture of water (90%) and acetonitrile (10%) containing formic acid (0.1%) at a flow rate of 0.5 ml/min. Compounds were detected at 223 nm. In the electrospray ionization-mass spectrometry (ESI-MS) analysis, mass spectra were obtained by using the negative-ion mode with the settings described in a previous study (53).

Determination of molecular mass.

The purity of the preparations was examined by sodium dodecyl sulfate–12% polyacrylamide gel (or 15% polyacrylamide gel for the analysis of LigXc) electrophoresis (SDS-PAGE). The molecular masses were determined by gel filtration chromatography using a Superdex200 10/300GL column (GE Healthcare) (54) and in vitro cross-linking (55). Native PAGE was performed using 7.5% polyacrylamide with a high-molecular-weight calibration kit for native electrophoresis (GE Healthcare).

Enzyme characterization of LigXd.

For identification of the flavin cofactor bound to LigXd, a 200-μl solution of purified LigXd (21 μM) in 50 mM KH₂PO₄-KOH buffer (pH 6.0) was incubated in boiling water for 5 min. The denatured protein was removed by centrifugation, and then the resulting supernatant was analyzed by HPLC. The mobile phase of the HPLC system was composed of a mixture of water (90%) and acetonitrile (10%) containing 0.1% formic acid at a flow rate of 0.3 ml/min. The flavin cofactors were detected at 267 nm.

UV-visible spectra of purified LigXd (1 mg of protein/ml) in 50 mM KH₂PO₄-KOH buffer (pH 6.0) in the presence and absence of 1 mM NADH were measured in a cell with a 1-cm light path length using a spectrophotometer (V-630BIO; Jasco Corporation).

The reductase activity of LigXd was evaluated by measuring the NAD(P)H-dependent reduction activity for 2,6-dichlorophenolindophenol (DCPIP). DCPIP (200 μM) was incubated in 50 mM KH₂PO₄-KOH buffer (pH 6.5) with LigXd (0.05 to 0.25 μg/ml) in the presence of NADH or NADPH (10 to 1,000 μM) at 30°C. The decrease in absorbance at 600 nm (DCPIP ε₆₀₀ = 8,900 M⁻¹ cm⁻¹) was monitored. One unit of enzyme activity was defined as the amount of enzyme required to reduce 1 μmol of DCPIP per min at 30°C. Specific activities were expressed in units per milligram of protein. The K_m and k_cat values were obtained from Hanes-Woolf plots and were expressed as the averages ± standard deviations from at least three independent experiments.

Assays for DDVA O-demethylase.

The enzyme reaction was typically carried out in a 200-μl reaction mixture containing FE22 buffer (pH 6.0), LigXa (0.2 μM, 10 μg/ml of protein), LigXc (6.0 μM, 83 μg/ml of protein), LigXd (0.4 μM, 19 μg/ml of protein), 100 μM DDVA, and 200 μM NADH. After the incubation (3 min for kinetic analysis and 5 min for the determination of optimal pH and temperature), the reaction was terminated by the addition of methanol (final concentration, 25%) and analyzed by HPLC as described above. One unit of enzyme activity was defined as the amount of enzyme that converted 1 μmol of DDVA per min at 30°C. Specific activities were expressed in units per milligram of LigXa protein. The optimal pH and temperature for DDVA O-demethylase were determined at pH and temperature ranges of 5.0 to 9.0 and 10 to 50°C using 50 mM GTAFE2 buffer, consisting of 50 mM 3,3-dimethylglutarate, 50 mM Tris, and 50 mM 2-amino-2-methyl-1,3-propanediol, 100 μM Fe(NH₄)₂(SO₄)₂ · 6H₂O, and 2 mM cysteine hydrochloride and FE22 buffer (pH 6.0), respectively. The K_m and k_cat values were obtained from Hanes-Woolf plots and were expressed as the averages ± standard deviations from at least three independent experiments. Kinetic parameters were determined using 10 to 500 μM DDVA and 500 μM NADH.

Substrate preference and metal dependency.

For determination of substrate preference of DDVA O-demethylase, 100 μM substrate (vanillin, vanillate, syringaldehyde, syringate, acetovanillone, ferulate, dicamba, and OH-DDVA) was incubated with a mixture of LigXa (0.2 μM), LigXc (6.0 μM), and LigXd (0.4 μM) in FE22 buffer (pH 6.0) in the presence of 200 μM NADH. After 3 min of incubation at 30°C, the decrease in the amount of substrate excluding OH-DDVA was determined by HPLC using the conditions described above. For the analysis of OH-DDVA, the mobile phase of HPLC system was composed of a mixture of water (95%) and acetonitrile (5%) containing formic acid (0.1%) at a flow rate of 0.5 ml/min. Vanillin, vanillate, syringaldehyde, syringate, acetovanillone, ferulate, dicamba, and OH-DDVA were detected at 230, 219, 309, 218, 229, 323, 202, and 223 nm, and their retention times were 2.7, 1.6, 3.2, 1.7, 3.7, 4.1, 3.7, and 7.0 min, respectively.

To determine the metal ion dependency of DDVA O-demethylase, LigXa was purified in the absence of Fe(II). LigXa purified in the absence of Fe(II) (0.2 μM) was mixed with LigXc (6.0 μM) and LigXd (0.4 μM) and then incubated for 3 min with 100 μM DDVA and 500 μM NADH in the presence and absence of 100 μM Fe(NH₄)₂(SO₄)₂, MnCl₂ · 4H₂O, MgCl₂ · H₂O, CuSO₄ · 5H₂O, CoSO₄ · 7H₂O, or ZnCl₂. The decrease in the amount of DDVA was measured by HPLC analysis.

RESULTS AND DISCUSSION

Confirmation of the involvement of ligXa in the catabolism of DDVA.

In order to confirm whether ligXa (SLG_07770), which was previously identified as the gene encoding an oxygenase component of DDVA O-demethylase, is actually involved in the conversion of DDVA, ligXa in SYK-6 was disrupted by the insertion of a kanamycin (Km) resistance gene. The resultant ligXa mutant, SME049, lost the ability to grow in Wx minimal medium containing 5 mM DDVA (Fig. 2A). In contrast, it grew normally in 10 mM vanillate (Fig. 2B). The growth defect of SME049 on DDVA was complemented by the introduction of pJBXa carrying ligXa (Fig. 2C). These results indicate that ligXa is essential for the catabolism of DDVA in SYK-6.

FIG 2.

Growth of ligXa mutant on DDVA and vanillate. (A) Growth of ligXa mutant (SME049) on DDVA. The cells of SYK-6 (open circles) and SME049 (closed circles) were incubated in Wx medium containing 5 mM DDVA. (B) Growth of SME049 on vanillate. The cells of SYK-6 (open circles) and SME049 (closed circles) were incubated in Wx medium containing 10 mM vanillate. (C) Complementation of SME049 with pJBXa carrying ligXa. The cells of SYK-6 harboring pJB864 (open circles), SME049 harboring pJB864 (triangles), and SME049 harboring pJBXa (closed circles) were incubated in Wx medium containing 5 mM DDVA. All the experiments were performed in triplicate, and the data are the averages ± standard deviations.

Genome search for candidate genes encoding ETCs of DDVA O-demethylase in Sphingobium sp. SYK-6.

BLAST searches revealed the presence of three putative ferredoxin reductase genes, SLG_04360, SLG_18830, and SLG_21200, which showed 19%, 59%, and 59% amino acid sequence identities with NahAa_NCIB9816-4, CarAdI_KA1, and CarAdI_KA1, respectively, and the presence of a putative flavodoxin reductase gene, SLG_p_00600, which showed 16% amino acid sequence identity with OphA1_DBO1 (Table 3). Three putative ferredoxin genes were also found, SLG_08500, SLG_13390, and SLG_18840, the deduced amino acid sequences of which showed 39%, 33%, and 39% identities with CarAcI_KA1 (Table 3). Two putative ferredoxin reductases (SLG_18830 and SLG_21200) and two putative ferredoxin genes (SLG_08500 and SLG_18840) similar to each other showed 58% and 64% amino acid sequence identities, while SLG_13390 showed 30% and 35% amino acid sequence identities with SLG_08500 and SLG_18840, respectively. A putative ferredoxin reductase gene, SLG_18830, is located immediately downstream of a putative ferredoxin gene, SLG_18840, implying that these genes may function together.

TABLE 3.

Putative ETC genes from seven completely sequenced sphingomonad genomes and percentages of identity with ligXc or ligXd at the amino acid sequence level^a

Domain	SYK-6		DSM 12444		PP1Y		L-1		UT26S		RW1		RB2256
	Gene	% ID	Gene	% ID	Gene	% ID	Gene	% ID	Gene	% ID	Gene	% ID	Gene	% ID
Identity with ligXc
fer2	SLG_08500 (LigXc)	100	Saro_0509	56.8	PP1Y_AT7331	60.0	Sphch_0633	33.0	SJA_C1-08030	64.8	Swit_0053	32.5	Sala_0795	31.9
	SLG_13390	30.4	Saro_1115	61.9	PP1Y_AT28162	27.3	Sphch_1567	65.7	SJA_C1-11940	15.7	Swit_0285	36.1	Sala_0845	61.0
	SLG_18840	63.8	Saro_1477 (Arx)	63.8	PP1Y_AT31173	33.9	Sphch_3064	59.8	SJA_C1-21730	33.0	Swit_0363	61.9
			Saro_2796	32.4	PP1Y_Mpl3486	33.6	Sphch_3575	60.0			Swit_0998	25.2
			Saro_2941	21.1			Sphch_3593	35.5			Swit_1759	35.8
			Saro_3658	32.7							Swit_1843	34.0
			Saro_3849	22.6							Swit_4893 (Fdx3)	33.0
											Swit_5088 (Fdx1)	31.8
Identity with ligXd
pyr_redox + Reductase_C	SLG_18830	58.1	Saro_0216 (ArR)	59.7	PP1Y_AT1991	57.6	Sphch_0328	61.9	SJA_C1-25250	29.3	Swit_0898	34.3	Sala_1971	61.4
	SLG_21200 (LigXd)	100	Saro_3838	36.9	PP1Y_AT15827	35.2	Sphch_3944	40.4	SJA_C1-26560	61.7	Swit_2055 (RedA1)	61.4
											Swit_4920 (RedA2)	57.4
FNR + fer2	SLG_p_00600	8.5	Saro_2763	12.5	PP1Y_AT15752	17.2	Sphch_2158	11.2	SJA_C1-05160	8.5	Swit_0757	12.7	Sala_2468	9.5
			Saro_3846	17.7	PP1Y_AT31343	11.7	Sphch_3460	15.6	SJA_C1-12030	13.2	Swit_3265	10.7
									SJA_P1-01380	13.8
FNR-like	SLG_04360	12.2	None		None		Sphch_1619	13.8	SJA_C1-10850	13.7	Swit_0104	6.1	Sala_0109	10.4
							Sphch_2141	9.5			Swit_4603	13.0	Sala_1032	2.7

^aStrains: SYK-6, Sphingobium sp. strain SYK-6; DSM 12444, Novosphingobium aromaticivorans DSM 12444; PP1Y, Novosphingobium sp. strain PP1Y; L-1, Sphingobium chlorophenolicum L-1; UT26S, Sphingobium japonicum UT26S; RW1, Sphingomonas wittichii RW1; RB2256, Sphingopyxis alaskensis RB2256. % ID, percent identity. The gene products are listed in parentheses.

Identification of the ferredoxin and ferredoxin reductase genes involved in the growth on DDVA.

To examine the involvement of the putative ferredoxin reductase genes in the growth of SYK-6 on DDVA, SLG_18830 and SLG_21200 were disrupted in SYK-6 by the insertion of a Km resistance gene and a Tet resistance gene, respectively. The ability of an SLG_18830 mutant (SME051) and an SLG_21200 mutant (SME052) to grow in Wx medium containing 5 mM DDVA was evaluated. The results showed that SME051 cells grew normally on DDVA and entered the stationary phase at ca. 20 h; however, the growth of SME052 cells in the same medium was not observed until ca. 80 h of incubation (Fig. 3A). Furthermore, the SLG_18830 SLG_21200 double mutant (SME053) completely lost the ability to grow on DDVA (Fig. 3A). The growth defect of SME052 on DDVA was complemented by the introduction of pJBXd carrying SLG_21200 (Fig. 3B). These results indicate that SLG_21200 is essential for the normal growth of SYK-6 on DDVA. We designated SLG_21200 ligXd.

FIG 3.

Growth of mutants of putative ferredoxin reductase genes and ferredoxin genes on DDVA. (A) Growth of the mutants of putative ferredoxin reductase genes on DDVA. The cells of SYK-6 (open circles), SLG_18830 mutant (SME051, triangles), SLG_21200 mutant (SME052, closed circles), and SLG_18830 SLG_21200 double mutant (SME053, squares) were incubated in Wx medium containing 5 mM DDVA. (B) Complementation of SME052 with pJBXd carrying SLG_21200 (ligXd). The cells of SYK-6 harboring pJB864 (open circles), SME052 harboring pJB864 (triangles), and SME052 harboring pJBXd (closed circles) were incubated in Wx medium containing 5 mM DDVA. (C) Growth of the mutants of putative ferredoxin genes on DDVA. The cells of SYK-6 (open circles), SLG_08500 mutant (SME073; closed circles), SLG_18840 mutant (SME074; triangles), and SLG_08500 SLG_18840 double mutant (SME075; squares) strains were incubated in Wx medium containing 5 mM DDVA. (D) Complementation of SME073 with pJBXc carrying SLG_08500 (ligXc). The cells of SYK-6 harboring pJB864 (open circles), SME073 harboring pJB864 (triangles), and SME073 harboring pJBXc (closed circles) were incubated in Wx medium containing 5 mM DDVA. All the experiments were performed in triplicate, and the data are the averages ± standard deviations.

Similarly, the putative ferredoxin genes SLG_08500 and SLG_18840 in SYK-6 were disrupted. Disruption of SLG_18840 had no effect on the growth of SYK-6 in Wx medium containing 5 mM DDVA (Fig. 3C). However, the growth of SLG_08500 mutant (SME073) cells on DDVA was not observed until ca. 80 h of incubation, while the wild-type cells entered the stationary phase at ca. 20 h (Fig. 3C). An SLG_08500 SLG_18840 double mutant (SME075) completely lost the ability to grow on DDVA. The growth defect of SME073 on DDVA was complemented by the introduction of pJBXc carrying SLG_08500 (Fig. 3D). These results indicate that SLG_08500 is essential for the normal growth of SYK-6 on DDVA. According to these results, we designated SLG_08500 ligXc.

Our preliminary experiment indicated that the growth defect of SME073 on DDVA was complemented by an introduction of pJB864 carrying SLG_18840 (data not shown), indicating that SLG_18840 has a potential to function as a ferredoxin component of DDVA O-demethylase. This fact also suggests that SLG_18840 and probably SLG_18830 were not fully expressed in SYK-6 cells when grown on DDVA.

Reconstitution of DDVA O-demethylase activity with the gene products of ligXa, ligXc, and ligXd.

SDS-PAGE showed the production of 49-kDa, 13-kDa, and 48-kDa proteins in E. coli BL21(DE3) cells carrying ligXa, ligXc, and ligXd (Fig. 4A and B). These sizes are close to the predicted molecular masses of His tag-fused LigXa (51,253 Da), LigXc (13,842 Da), and LigXd (46,710 Da). LigXa, LigXc, and LigXd were purified to near homogeneity by Ni affinity chromatography (Fig. 4C). The solutions of purified LigXa and LigXc exhibited a brown-red color, suggesting the presence of an iron-sulfur cluster in these proteins. On the other hand, the solution of purified LigXd showed a brown-yellow color. This color appeared to be derived from flavin.

FIG 4.

Expression of ligXa, ligXc, and ligXd in E. coli and purification of the gene products. Proteins were separated on an SDS–12% (A and C) and SDS–15% (B) polyacrylamide gels and stained with Coomassie brilliant blue. Molecular masses are given on the left. (A) Lanes: M, molecular mass markers; 1, crude extract of E. coli BL21(DE3) harboring pET-16b (10 μg of protein); 2, crude extract of E. coli BL21(DE3) harboring pET16Xa (10 μg of protein); 3, crude extract of E. coli BL21(DE3) harboring pET16Xd (10 μg of protein). (B) Lanes: M, molecular mass markers; 1, crude extract of E. coli BL21(DE3) harboring pET-16b (10 μg of protein); 2, crude extract of E. coli BL21(DE3) harboring pET16Xc (10 μg of protein). (C) Lanes: M, molecular mass markers; 1, Ni-Sepharose fraction of LigXa (5.0 μg of protein); 2, Ni-Sepharose fraction of LigXc (5.0 μg of protein); 3, Ni-Sepharose fraction of LigXd (5.0 μg of protein).

In order to verify that the DDVA O-demethylase activity is actually reconstituted with LigXa, LigXc, and LigXd, 100 μM DDVA was incubated with the mixture of these purified components (20 μg each of protein/ml) in the presence of 200 μM NADH at 30°C for 30 min. HPLC analysis of the reaction mixture indicated a decrease in the amount of DDVA and the production of a new peak (compound I) with a retention time of 2.0 min (Fig. 5A and B). Negative ESI-MS analysis of compound I showed a major fragment at m/z 319 (Fig. 5C). Since the retention time and the m/z value of the deprotonated ion of compound I corresponded to those of the authentic OH-DDVA (Fig. 5D and E), it was concluded that DDVA was converted to OH-DDVA by mixing LigXa, LigXc, and LigXd in the presence of NADH. On the other hand, the same enzyme mixture did not transform DDVA in the absence of NADH (Fig. 5F). All these results indicate that DDVA O-demethylase is a three-component monooxygenase system, which consists of an oxygenase component, LigXa, a ferredoxin, LigXc, and a ferredoxin reductase, LigXd, and NADH is required for this reaction (Fig. 1C). A three-component O-demethylase had been reported only for dicamba O-demethylase (56–58). This enzyme consists of an oxygenase component, DdmC, a ferredoxin, DdmB, and a ferredoxin reductase, DdmA1 or DdmA2 (DdmA1 and DdmA2 are 99% identical to each other), of Stenotrophomonas (formerly Pseudomonas) maltophilia DI-6. The deduced amino acid sequences of ligXc (105 amino acids [aa]) and ligXd (412 aa) exhibited high similarity (61% and 58% identities) with those of ddmB (105 aa) and ddmA1 (408 aa), whereas the deduced amino acid sequence of ligXa (424 aa) showed only 16% identity with that of ddmC (339 aa). Based on the similarity of LigXc and LigXd with CarAcI and CarAdI of Novosphingobium sp. KA1 (34), these ETCs are basically categorized into class IIA RHO, in which ferredoxins and ferredoxin reductases are significantly similar to putidaredoxin and putidaredoxin reductase (59, 60).

FIG 5.

Reconstitution of DDVA O-demethylase activity with LigXa, LigXc, and LigXd. DDVA (100 μM) was incubated with a mixture of purified LigXa, LigXc, and LigXd (20 μg each of protein/ml) in the presence or absence of 200 μM NADH. Portions of the reaction mixtures were collected at the start and at 30 min and analyzed by HPLC and liquid chromatography (LC)-MS. (A and B) HPLC chromatograms of the reaction mixtures incubated in the presence of NADH at the start and at 30 min, respectively. (C) Negative-ion ESI-MS spectrum of compound I. (D) HPLC chromatogram of the authentic OH-DDVA. (E) Negative-ion ESI-MS spectrum of OH-DDVA. (F) HPLC chromatogram of the reaction mixture incubated in the absence of NADH at 30 min.

Molecular masses of LigXa, LigXc, and LigXd.

Gel filtration chromatography demonstrated that the molecular masses of LigXc and LigXd were 18 kDa and 54 kDa, respectively (see Fig. S1 in the supplemental material). These results indicated that both LigXc and LigXd are monomeric proteins. Although LigXa was analyzed by gel filtration chromatography, the molecular mass of LigXa was not determined due to the instability of its retention time. Therefore, an in vitro cross-linking experiment was performed. SDS-PAGE of the cross-linked LigXc and LigXd suggested that these proteins are monomeric (see Fig. S2A and B in the supplemental material). On the other hand, a band at ca. 149 kDa was observed for the cross-linked LigXa (see Fig. S2C). In addition, native-PAGE of LigXa showed a band at ca. 174 kDa (see Fig. S2D). These results strongly suggest that LigXa is a homotrimer.

Enzymatic properties of LigXd.

In general, the reductase components of RHOs contain either FAD or FMN as a prosthetic group, and the deduced amino acid sequence of LigXd contains a FAD binding domain (pfam00070). In order to identify the flavin cofactor in LigXd, a solution of LigXd was heat treated, and the resulting supernatant was analyzed by HPLC. Based on the comparison of the retention times of authentic FAD and FMN, FAD was identified as the prosthetic group of LigXd (see Fig. S3 in the supplemental material). The amount of FAD included in LigXd was estimated to be approximately 0.86 mol of FAD/mol of LigXd. This result suggests that LigXd binds one FAD molecule per monomer of LigXd.

UV-visible spectrum of LigXd showed absorption maxima at 384 nm, 454 nm, and 482 nm (see Fig. S4 in the supplemental material). These absorption maxima disappeared when 1 mM NADH was added to the LigXd solution, suggesting that LigXd was reduced by NADH (see Fig. S4).

In order to determine the coenzyme requirement of LigXd, the kinetic parameters of LigXd for NADH and NADPH were determined by the DCPIP reductase assay. The K_m value for NADH was estimated to be 56.0 ± 6.3 μM, and k_cat with NADH was 538 ± 12 s⁻¹ (see Fig. S5A and B in the supplemental material). On the other hand, the kinetic parameters of LigXd for NADPH were unable to be determined because the activity was not saturated at concentrations of up to 1 mM NADPH (see Fig. S5C). The DCPIP reductase activities of LigXd using NADH (0.01 to 1.0 mM) were 3.4- to 120-fold higher than those using the same concentrations of NADPH. These results indicate that the physiological electron donor of LigXd is NADH.

Optimal mixing ratio of LigXa, LigXc, and LigXd.

To examine the enzyme properties of DDVA O-demethylase, the optimal mixing ratio of LigXa, LigXc, and LigXd was determined. When the concentrations of LigXa (0.2 μM monomer) and LigXd (0.2 μM) were maintained, the LigXc concentration was varied from 10- to 40-fold higher than that of LigXa. In these reactions, a >20-fold-higher molar concentration of LigXc than LigXa was necessary for maximum activities (see Fig. S6A in the supplemental material). Next, the concentrations of LigXa (0.2 μM monomer) and LigXc (6.0 μM) were maintained, and the LigXd concentration was varied from 1.0- to 3.5-fold higher than that of LigXa. As a result, maximum activity was obtained when a 2-fold-higher molar concentration of LigXd than LigXa was added to the reaction mixture (see Fig. S6B). Based on these results, the optimum LigXc concentration was determined when the concentrations of LigXa (0.2 μM monomer) and LigXd (0.4 μM) were maintained (see Fig. S6C). In these reactions, maximum activities were obtained when using a >30-fold-higher concentration of LigXc than LigXa. These results indicate that the optimal mixing ratio of LigXa (monomer), LigXc, and LigXd was 1:30:2.

Enzyme properties of DDVA O-demethylase.

When DDVA O-demethylase activity was measured at different temperatures at pH 6.0, the highest activities (approximately 0.87 U/mg) were obtained at 25 to 30°C (see Fig. S7A in the supplemental material). When DDVA O-demethylase activity was measured at different pH ranges at 30°C, the highest activity was observed at pH 6.0 (1.5 U/mg) (see Fig. S7B). Another peak was seen at pH 8.5; this may be caused by the different pH optima for each component.

Since the oxygenase components of RHOs generally contain Fe(II) at the catalytic site, FE22 buffer, which contains 100 μM Fe(II), was used for the purification of LigXa. To determine the actual metal ion dependency of DDVA O-demethylase, LigXa was purified in the absence of Fe(II), and the activity of DDVA O-demethylase using this LigXa was measured in the presence and absence of 100 μM Fe(II). DDVA O-demethylase showed specific activities of 1.5 U/mg and 0.29 U/mg, in the presence and absence of Fe(II), respectively. However, when the effects of other metal ions, including Mg(II), Ca(II), Cu(II), Co(II), Zn(II), and Mn(II), were tested, significant activation was not observed. These results strongly suggest that LigXa contains Fe(II) at its catalytic center.

Substrate preference of DDVA O-demethylase.

The mixture of LigXa, LigXc, and LigXd was incubated with 100 μM vanillin, vanillate, syringaldehyde, syringate, acetovanillone, ferulate, and dicamba (2-methoxy-3,6-dichlorobenzoate) to determine its substrate preference. No conversion was observed for the lignin-derived monoaryls and dicamba, while DDVA O-demethylase exhibited a specific activity of 1.5 U/mg toward DDVA. On the other hand, when 100 μM OH-DDVA was incubated with DDVA O-demethylase in the presence of NADH, a faint conversion was observed (0.06 U/mg). These results indicated that biphenyl carboxylic acids with methoxyl groups could be the substrates for DDVA O-demethylase; however, this enzyme is highly specific for DDVA. This substrate preference of DDVA O-demethylase supports our previous result; the ring cleavage of OH-DDVA catalyzed by LigZ is the major catabolic route for DDVA (Fig. 1A) (12, 15). Kinetic analysis of DDVA O-demethylase was performed in the presence of 500 μM NADH. The enzyme displayed Michaelis-Menten kinetics and showed apparent K_m values of 63.5 ± 4.9 μM for DDVA and k_cat of 6.1 ± 0.2 s⁻¹ (see Fig. S8 in the supplemental material).

Genome search for ETC genes in sphingomonads.

The searches using CDART indicated that six other sphingomonads whose complete genome sequences were determined, N. aromaticivorans DSM 12444, Novosphingobium sp. strain PP1Y, S. chlorophenolicum L-1, S. japonicum UT26S, S. wittichii RW1, and S. alaskensis RB2256, have two to eight ferredoxins, one to three ferredoxin reductases containing a set of pyr_redox and Reductase_C domains, and two to four reductases containing an FNR-like domain with or without a fer2 domain (Table 3). Interestingly, each strain has more than one ligXc ortholog and more than one ligXd ortholog, which have >58% amino acid sequence identities (Table 3; see also Fig. S9 in the supplemental material). Among these sphingomonad LigXc and LigXd orthologous genes, redA1 (Swit_2055) (61) and redA2 (Swit_4920) (62) were characterized as ferredoxin reductase genes of dioxin dioxygenase of S. wittichii RW1. RW1 dioxin dioxygenases are composed of large and small subunits of oxygenase components, DxnA1 and DxnA2, a ferredoxin, Fdx1 (Swit_5088) (61), or an alternative ferredoxin, Fdx3 (Swit_4893) (63), and RedA1 or RedA2. The gene products of Swit_5088 and Swit_4893, which are less similar to LigXc (32% to 33% identity), were identified as ferredoxin of dioxin dioxygenase instead of Swit_0363, which has 62% identity with LigXc. Another case is Saro_1477 ferredoxin (Arx) and Saro_0216 ferredoxin reductase (ArR) of N. aromaticivorans DSM 12444, which showed 64% and 60% identities with LigXc and LigXd, respectively. Interestingly, these gene products were characterized as ETCs of heme oxygenases, cytochrome P450s (64, 65). DSM 12444 possesses 16 P450 genes, and efficient monooxygenase activities toward terpenoid compounds have been reconstituted, with 5 of them using the gene products of Arx and ArR as ETCs. These facts may suggest that the ETCs similar to LigXc and LigXd are able to transfer electrons not only to oxygenase components of RHOs but also multiple P450 enzymes. The three-dimensional structures of Arx, ArR, and one of the P450 enzymes, CYP101D1, were determined. The results indicated that their general structural features are similar to those of putidaredoxin, putidaredoxin reductase, and CYP101A1 of the camphor hydroxylase system of P. putida (66). Significant differences are found in the proposed protein-protein interaction surfaces. There are also regions of positive charge on the possible interaction face of ArR and CYP101D1, and corresponding negatively charged regions on the surface of Arx. Analysis by UCSF Chimera software indicated that the amino acid residues of the regions of the LigXc and LigXd orthologs corresponding to the negatively and positively charged surface areas of Arx and ArR, respectively, were highly conserved, as shown in Fig. S10 in the supplemental material. These facts suggest that the LigXc and LigXd orthologs have protein-protein interaction surfaces similar to those of Arx and ArR.

Genome search for oxygenase component genes in sphingomonads.

A BLAST search using the NCBI nr (nonredundant) database indicated that Swit_3062 of S. wittichii RW1and SJA_C1-02290 of S. japonicum UT26S were the genes most similar to ligXa. Their amino acid sequence identities were 60% and 61%, respectively. High similarity of these genes with ligXa and the presence of ligXc and ligXd orthologs may suggest that RW1 and UT26S show DDVA O-demethylase activities. In addition, RW1 has genes highly similar to ligZ (Swit_3064, 64% identity) and ligY (Swit_3063, 70% identity), although no genes significantly similar to LigW were found. Analysis by CDART using the Rieske superfamily (cl00938) and RHO_alpha_C domain (cl14643) conserved in oxygenase components of RHOs revealed that the seven sphingomonads contain relatively numerous oxygenase component genes (number of genes: SYK-6, 11; DSM 12444, 24; PP1Y, 37; L-1, 7; UT26S, 8; RW1, 53; RB2256, 1) and P450 genes (number of genes: SYK-6, 1; DSM 12444, 16; PP1Y, 5; L-1, 15; UT26S, 3; RW1, 18; RB2256, 6). A phylogenetic tree of the deduced amino acid sequences of the oxygenase component genes of the seven sphingomonads showed that these genes, divided into the two groups already reported (67), were highly diversified, with previously characterized oxygenase components accounting for only a small fraction of the tree (see Fig. S11 in the supplemental material). SYK-6 has 11 putative oxygenase component genes and one P450 gene; however, only three ferredoxin and four reductase genes were detected in the genome (Table 3). RW1 has the largest number of putative oxygenase component genes (53 genes) and P450 genes (18 genes), though this strain has only eight putative ferredoxin genes and seven potential reductase genes (Table 3). These considerably lower numbers of ETC genes than oxygenase component genes suggest that these few ETCs are able to transfer the electron from NADH (NADPH) to a particular number of oxygenase components of RHOs and P450 enzymes. These compatible ETC genes may have contributed to expanding the substrate spectra of oxygenase systems in bacteria by acquiring new oxygenase component genes through gene duplications and gene transfer. The presence of highly conserved LigXc and LigXd orthologs in sphingomonads implies their importance as ETCs of the oxygenase systems in this group of bacteria.