Purification and partial characterization of the extradiol dioxygenase, 2′-carboxy-2,3-dihydroxybiphenyl 1,2-dioxygenase, in the fluorene degradation pathway from Rhodococcus sp. strain DFA3

Abstract

Type II extradiol dioxygenase, 2′-carboxy-2,3-dihydroxybiphenyl 1,2-dioxygenase (FlnD1D2) involved in the fluorene degradation pathway of Rhodococcus sp. DFA3 was purified to homogeneity from a heterologously expressing Escherichia coli. Gel filtration chromatography and SDS-PAGE suggested that FlnD1D2 is an α₄β₄ heterooctamer and that the molecular masses of these subunits are 30 and 9.9 kDa, respectively. The optimum pH and temperature for enzyme activity were 8.0 and 30 °C, respectively. Assessment of metal ion effects suggested that exogenously supplied Fe²⁺ increases enzyme activity 3.2-fold. FlnD1D2 catalyzed meta-cleavage of 2′-carboxy-2,3-dihydroxybiphenyl homologous compounds, but not single-ring catecholic compounds. The K_m and k_cat/K_m values of FlnD1D2 for 2,3-dihidroxybiphenyl were 97.2 μM and 1.5 × 10⁻² μM⁻¹sec⁻¹, and for 2,2′,3-trihydroxybiphenyl, they were 168.0 μM and 0.5 × 10⁻² μM⁻¹sec⁻¹, respectively. A phylogenetic tree of the large and small subunits of type II extradiol dioxygenases suggested that FlnD1D2 constitutes a novel subgroup among heterooligomeric type II extradiol dioxygenases.

Graphical abstract

Type II extradiol dioxygenases (EDOs) are classified into 5 subgroups, and FlnD1D2 constitutes the distinct group from previously investigated EDOs.

Key words:

• bacteria
• characterization
• extradiol dioxygenase
• fluorene
• Rhodococcus

Aromatic ring cleavage is a crucial step in the bacterial aerobic degradation of various aromatic compounds. Most aromatic ring cleavage enzymes can be classified into two groups, intradiol and extradiol dioxygenases, based on the catalytic position of the dihydroxylated aromatic compound.¹⁾ Intradiol dioxygenases, which contain a non-heme Fe(III) ion at the active site, catalyze aromatic ring cleavage at the C-C bond between the vicinal hydroxyl groups (ortho-cleavage). In contrast, extradiol dioxygenases contain a non-heme Fe(II) ion in the active site and catalyze aromatic ring cleavage at the C-C bond adjacent to the vicinal hydroxyl groups (meta-cleavage). Extradiol dioxygenases are classified into three types based on evolutionarily independent families.²⁾ The majority of extradiol dioxygenases are classified as type I, which belong to the vicinal oxygen chelate superfamily. Type II extradiol dioxygenases do not belong to the vicinal oxygen chelate family and have a homooligomeric or heterooligomeric composition. Type III extradiol dioxygenases belong to the cupin superfamily. Several extradiol dioxygenase crystal structures have been reported. For example, the three-dimensional structures of 2,3-dihydroxybiphenyl 1,2-dioxygenases (BphCs) from Burkholderia cepacia (type I)³⁾ and Acidvorax sp. KKS102 (type I),⁴⁾ catechol 2,3-dioxygenase from Pseudomonas putida mt-2 (type I),⁵⁾ protocatechuate 4,5-dioxygenase (LigAB) from Sphingobium sp. SYK-6 (type II),⁶⁾ and gentisate 1,2-dioxygenase (GtdA) from Escherichia coli O157 (type III)⁷⁾ have been reported. Although different folds were reported for each enzyme type, all extradiol dioxygenases show similar active sites and catalytic mechanisms, suggesting that these enzymes evolved separately.²⁾

Fluorene (FN) is one of the 16 polycyclic aromatic hydrocarbons (PAHs) labeled as a priority pollutant by the U.S. Environmental Protection Agency.⁸⁾ PAHs are toxic, mutagenic, and reside for a long period in the environment, because their fused ring structures have a stable planar structural moiety. Bacteria, which can degrade one or more PAHs, have been studied extensively for their ability to remove PAHs from polluted environments.^9–10) Our previous study, dibenzofuran (DF) and FN-utilizing bacteria, Terrabacter sp. strain DBF63 and Rhodococcus sp. DFA3, were isolated, and their DF/FN degradative dbf/fln gene clusters were cloned and analyzed.^11–13) flnD1D2 genes encoding 2′-carboxy-2,3-dihydroxybiphenyl (2′-carboxy-2,3-DHBP) 1,2-dioxygenase are contained within the dbf/fln gene cluster of Rhodococcus sp. strain DFA3. This enzyme catalyzes the meta-cleavage of 2′-carboxy-2,3-DHBP and requires both FlnD1 and FlnD2 proteins for activity. This enzyme is classified as a type II extradiol dioxygenase based on amino acid sequence comparisons to FlnD1, which is the catalytic subunit of the enzyme. Our previous study,^11,13) FlnD1D2, was shown to be an extradiol dioxygenase involved in the FN aerobic degradation pathway; however, detailed enzymatic characterization of FlnD1D2 has not been performed.

In the current study, the type II extradiol dioxygenase, FlnD1D2 from the DFA3 strain, a component of the FN degradation pathway, was heterogeneously expressed and purified, and basic structure and function analyses were performed.

Materials and methods

Bacterial strains and culture conditions in E. coli

DH5α (Toyobo Co., Ltd) and E. coli BL21(DE3) (Novagen) were used as hosts for plasmid construction and protein expression, respectively. E. coli strains were grown on Luria–Bertani (LB) medium¹⁴⁾ at 37 °C. Ampicillin (Ap) and kanamycin (km) at final concentrations of 100 and 50 μg/mL, respectively, were added when necessary.

DNA manipulation

Plasmid DNA extraction from E. coli was performed using the alkaline lysis method.¹⁵⁾ Restriction enzymes (Toyobo Co., Ltd. or Takara Bio Inc.) and Ligation High ver.2 (Toyobo Co., Ltd.) were used according to the manufacturers’ instructions. DNA fragments were extracted from agarose gels using a Fast Gene Gel/PCR Extraction Kit (NIPPON Genetics Co., Ltd.) according to the manufacturer’s instructions. PCR was performed using Ex Taq (Takara Bio Inc.) or KOD-Plus- (Toyobo Co., Ltd.) according to the manufacturers’ instructions.

Plasmid construction

Expression plasmids for six histidine-tagged (His-tagged) FlnD1D2 were constructed by PCR using pA3S5021¹³⁾ (pUC119 with an 1175-bp HindIII-EcoRI insert from DFA3 strain DNA containing flnD1D2) as a template. To amplify flnD1, the forward 5′-TCTAGAATAAGGAGGTGTTCATATGGGCAGGCTGGTAGGTGCGTAC-3′ (underlined, double-underlined, and boldfaced sequences are the XbaI site, NdeI site, and Shine–Dalgarno [SD] sequences, respectively) and the reverse 5′-CGGTTCCTCCTCGGGGGCCAACCGAGCCAG-3′ (the underlined and dotted-lined sequences are the EcoRI site and six histidines, respectively) primers were used. Similarly, to amplify flnD2, the forward 5′-GAATTCTAAGGAGGTGGCTTGATGAACTTACCCCTGGAC-3′ (the underlined and boldfaced sequences are the EcoRI site and SD sequence, respectively) and the reverse 5′-GTCGACTTACGTCCCTCCGTGGTTAAGTCA-3′ (SalI site is underlined) primers were used. Both PCR products were simultaneously ligated into the XbaI-SalI site of pUC119 (Takara Bio Inc.). The resultant plasmid was named pUFL301, and the PCR product sequences were confirmed by sequencing. Unfortunately, the 341st C base of the flnD1 sequence was found to be replaced by a T, and therefore PCR-based site-directed mutagenesis of pUFL301 was performed according to the protocol of the KOD -Plus- Mutagenesis Kit (Toyobo Co., Ltd.) protocol. Forward 5′- CGACCTCGCGTTCTCCGCCAACCCGAAGATCGACC-3′ and reverse 5′-GGTCGATCTTCGGGTTGGCGGAGAACGCGAGGTCG-3′ (double-underlined base indicates the mutation site in each primer sequence) primers were used for replacement. The PCR amplicon was treated with DpnI to cleave the template DNA and then purified. This constructed plasmid was named pUFL304, and the nucleotide sequence of the insert was sequenced. The NdeI-SalI fragments from pUFL304 were ligated into the corresponding pET-26b(+)(Novagen) sites, forming pET-DF as a result.

Expression and purification in E. coli

BL21(DE3) carrying pET-DF was cultured in 2.5 l LB medium supplemented with km and isopropyl β-d-thiogalactopyranoside at final concentrations of 50 μg/mL and 0.1 mM, respectively, at 30 °C for 12 h. All purification procedures were carried out at 4 °C. The cells were harvested by centrifugation at 6000 g for 10 min, washed twice with buffer A (20 mM, Tris–HCl [pH 7.5] containing 0.5 M NaCl and 10% glycerol), and suspended in 120 mⅬ buffer A. The crude extract was prepared by sonication and centrifugation at 25,000 g for 1 h. Column chromatography was performed on an ÄKTA FPLC instrument (GE Healthcare), following the manufacturer’s recommendations. The enzyme was purified using a HiTrap Chelating HP column (GE Healthcare) with buffers A and B (20 mM Tris–HCl [pH 7.5] containing 0.5 M NaCl, 300 mM imidazole, and 10% glycerol). After washing the column with 60 mM imidazole (80% buffer A and 20% buffer B), the enzyme was eluted with 150 mM imidazole (50% buffer A and 50% buffer B) using a step-wise elution method. Enzyme fractions were collected, concentrated to 1 mⅬ, and buffer-exchanged with buffer C (20 mM Tris–HCl [pH 7.5] containing 0.2 M NaCl and 10% glycerol) using the Vivaspin 20 (10,000 MWCO PES, Millipore). The resulting concentrated preparation was applied to the HiLoad 26/600 Superdex-200 pg (GE Healthcare) with buffer C. Enzyme fractions were collected and concentrated to 1 mⅬ. Protein concentrations were determined using a protein assay kit (Bio-Rad Laboratories, Inc.)¹⁶⁾ with bovine serum albumin as a standard.

Molecular mass determination

The molecular mass of FlnD1D2 was estimated by gel filtration chromatography using the Superdex-200 with buffer C. Ferritin (440 kDa), aldolase (158 kDa), conalbumin (75 kDa), ovalbumin (44 kDa), and ribonuclease A (13.7 kDa) (Gel Filtration Calibration Kit LMW and Gel Filtration Calibration Kit HMW, GE Healthcare) were used as calibration standards. The molecular mass of each subunit was determined by 16% Tricine SDS-PAGE¹⁷⁾ followed by staining with SimplyBlue SafeStain (Invitrogen^TM).

Enzyme activity assay

Enzyme activity was determined using 2,3-DHBP as the substrate, and monitoring the A₄₃₄ increase with a spectrophotometer (JASCO Corporation). The wavelength and the corresponding molar extinction coefficient of the meta-cleavage product of 2,3-DHBP, used by Happe et al. (1993)¹⁸⁾ (λ_max= 434 nm, ε = 13,200 M⁻¹ cm⁻¹), were employed. The enzyme reaction was performed in 1 mⅬ of 20 mM Tris–HCl buffer (pH 8.0) at 30 °C containing 2,3-DHBP at a final concentration of 100 μM. The reaction was started by the addition of 10–20 μg enzyme. One unit of enzyme activity was defined as the amount of enzyme necessary to convert 1 μmol substrate/min at 25 °C.

Effects of metal ions on enzyme activity

The effects of metal ions on enzyme activity were determined by adding various metal salts to the reaction mixtures. The effect of EDTA was also assessed, because it may remove the metal from the active center of the enzyme. The enzyme was incubated in 20 mM Tris–HCl buffer (pH 7.5) at 4 °C containing a final concentration of 2 mM metal solution (FeCl₂4H₂O, FeCl₃6H₂O, CaCl₂, CoCl₂6H₂O, CuCl₂2H₂O, MgCl₂6H₂O, MnCl₂4H₂O, or ZnCl₂6H₂O) or EDTA for 30 min. After incubation, enzyme activity was assessed as described above.

Optimal pH and temperature

The optimal pH (from 5.5 to 10.5) for enzyme activity was determined at 25 °C and assessed using 20 mM of the following buffers (pH 5.5–6.5, MES–NaOH buffer; pH 6.5–8.0, MOPS–NaOH buffer; pH 7.5–9.0, Tris–HCl buffer; pH 8.5–10.5, glycine–NaOH buffer). The optimal temperature was measured in 20 mM Tris–HCl buffer (pH 7.5) at incubation temperatures ranging from 5 to 45 °C. Enzyme activity was assayed as described above.

Determination of substrate specificity and kinetic analysis

Substrate specificity was determined by evaluating the relative activities for 2,3-DHBP, 2,2′,3-trihydroxybipenyl (2,2′,3-THBP), catechol, 3-methylcatechol, and 4-methylcatechol in 20 mM Tris–HCl buffer (pH 8.0) at 30 °C. The wavelengths and the molar extinction coefficients of the meta-cleavage products of the respective substrates were as follows according to Happe et al. (1993)¹⁸⁾: 2,3-DHBP, as described above; 2,2′,3-THBP, λ_max= 438 nm and ε = 22,100 M⁻¹ cm⁻¹; catechol, λ_max= 375 nm and ε = 33,400 M⁻¹ cm⁻¹; 3-methyl catechol, λ_max= 388 nm and ε = 13,800 M⁻¹ cm⁻¹; 4-methyl catechol, λ_max= 410 nm and ε = 34,000 M⁻¹ cm⁻¹. Kinetic parameters, K_m, V_max, k_cat, and k_cat/K_m, were determined from Line–Burk plots for eight different substrate concentrations in 6.25–400 μM. Enzyme activity was assayed as described above after reactivation of the enzyme. Average values of triplicate enzyme activity determinations for each substrate concentration were used for kinetic analysis. Reactivation was achieved by exposure to 2 mM FeCl₂4H₂O for 5 min at 4 °C.

Phylogenetic analysis

Amino acid sequences of type II extradiol dioxygenases were retrieved from the NCBI protein database by BLAST search. Amino acid sequence alignment and phylogenetic analysis were performed using CLC Genomic Workbench 7.5 software (CLC bio) as follows. First, type II extradiol dioxygenases whose crystal structures had been determined (LigAB,⁶⁾ DesB,¹⁹⁾ CnbCaCb,²⁰⁾ and CarBaBb [our unpublished results]) were aligned. The resulting alignment was modified manually based on structural features including conserved α-helices and β-sheets. Next, the alignment was performed for other type II extradiol dioxygenases using the modified alignment as a template. In phylogenetic analyses, the neighbor-joining method was used with 1,000 bootstraps.

Results and discussion

Expression and purification

His-tagged FlnD1 and native FlnD2 were expressed from the single pET expression vector, pET-DF, in E. coli BL21(DE3). The His-tagged FlnD1D2 complex was purified by metal chelate affinity chromatography and gel filtration chromatography; the relevant details of the purification steps are shown in Table 1. The specific activity of the purified enzyme was 144.2 U/mg. By a two-step purification method, His-tagged FlnD1D2 was purified 126.8-fold, and 2.02 mg purified enzyme was obtained from 1 l E. coli culture. SDS-PAGE analysis of the FlnD1D2 samples obtained at each purification step is shown in Fig. 1. After gel filtration chromatography, SDS-PAGE revealed two bands, representing His-tagged FlnD1 (upper) and native FlnD2 (lower), with molecular masses of 30 and 9.9 kDa, respectively. The native molecular mass of His-tagged FlnD1D2 was estimated to be 180 kDa by gel filtration chromatography. Two tertiary structures have been reported for type I extradiol dioxygenases: homo-octamer for BphC³⁾ and homotetramer for catechol 2,3-dioxygenase.⁵⁾ For heterooligomeric extradiol dioxygenase type II enzymes, α₂β₂ heterotetramers were reported for protocatechuate 4,5-dioxygenase (LigAB),⁶⁾ 2′-aminobiphenyl-2,3-diol 1,2-dioxygenase (CarBaBb),^21–22) and 2-aminophenol 1,6-dioxygenase (CnbCaCb).²⁰⁾ The results obtained from the gel filtration chromatography suggested that FlnD1D2 has an α₄β₄ heterooctameric structure, indicating that this enzyme has a novel, functionally stable tertiary structure in solution.

Table 1. Purification of FlnD1D2.

Purification step	Total protein (mg)	Total activity^a (U)	Specific activity^a (U/mg)	Yield (%)	Purification (fold)
Crude extract	770	876	1.14	100	1.00
HiTrap Chelating HP	22.0	597.7	27.2	68.2	23.9
HiLoad Superdex 200	2.0	296.8	144.2	33.3	126.8

Note:

^a One unit of enzyme activity was defined as the amount of enzyme necessary to convert 1 µmol substrate/min.

Fig. 1. SDS-PAGE analysis of FlnD1 and FlnD2 at each purification step.

Notes: Lane M, SeeBlue Plus 2 standard (Invitrogen); lane 1, crude extract; lane 2, after HiTrap Chelating HP; lane 3, after HiLoad Superdex 200.

Effects of metal ions on activity

Extradiol dioxygenase has a non-heme Fe²⁺ as an active center. However, some extradiol dioxygenases can replace Fe²⁺ with another metal ion.^23–24) To clarify whether FlnD1D2 can accept another metal ion into its active site, relative activities was measured. The values were expressed as the percentage of that for untreated enzyme, which was set as 100%. As shown in Table 2, various metal ions decreased enzyme activity 5–68.9%, while the addition of Fe²⁺ increased enzyme activity 3.2-fold. This increase in activity suggests that the active site Fe²⁺ was lacked during the purification process, and thus exogenously supplemented Fe²⁺ filled the vacant active site. In fact, when EDTA was supplied as a chelating agent, enzyme activity was approximately half that of the untreated enzyme (Table 2). These results also confirmed that FlnD1D2 needs Fe²⁺ as the active center metal ion.

Table 2. Effect of metal ions on enzyme activity.

Chemicals	Relative activity^a (%)
EDTA	52.1 ± 8.1
Fe^2+	318.3 ± 23.2
Fe^3+	68.9 ± 4.3
Ca^2+	13.7 ± 0.8
Co^2+	6.9 ± 1.0
Cu^2+	6.2 ± 1.5
Mg^2+	50.3 ± 6.4
Mn^2+	5.3 ± 0.9
Ni^2+	28.4 ± 6.0

Note:

^a Relative activity was expressed as the percentage of untreated enzyme activity, taken to be 100%.

Optimal pH and temperature

As shown in Fig. 2, FlnD1D2 showed relatively high activity in the pH range of 7.5 to 9.0. Maximum activity was detected at pH 8.0–8.5. At pH 8.5 and higher, the FlnD1D2 enzyme aggregated soon after incubation started, although activity was detected up to pH 9.5. When the incubation temperature was increased, the activity of FlnD1D2 increased until peaking at 35 °C (Fig. 3), above which enzyme activity decreased rapidly. The optimal pH and temperature determined through these experiments are similar to those of 2′-aminobiphenyl-2, 3-diol 1,2-dioxygenase (CarBaBb) isolated from Pseudomonas resinovorans CA10.²¹⁾

Fig. 2. Optimal pH of FlnD1D2.

Notes: An enzyme assay was performed at 25 °C in 20 mM of the following buffers (×: MES–NaOH buffer, ○: MOPS–NaOH buffer, ▲: Tris–HCl buffer, □: Glycine–NaOH buffer). The bars indicate the standard deviation for the three samples.

Fig. 3. FlnD1D2 optimal temperature.

Notes: Enzyme assays were performed at various temperatures in 20 mM Tris–HCl buffer (pH 7.5). The bars indicate the standard deviation for the three samples.

Substrate specificity and kinetic studies

The preparation of 2′-carboxy-2,3-DHBP was attempted initially from 9-fluorenone using the resting cell reaction of E. coli cells expressing 9-fluorenone 1,1a-dioxygenase and 1,1a-dihydroxy-1-hydro-9-fluorenone dehydrogenase, because 2′-carboxy-2,3-DHBP is the original substrate of FlnD1D2. However, 2′-carboxy-2,3-DHBP is unstable, because it readily forms a lactone ring.¹¹⁾ Although part of the substrate presumably existed as 2′-carboxy-2,3-DHBP, the exact substrate concentration could not be measured. Therefore, stable and available compounds, including 2′-carboxy-2,3-DHBP-related compounds (2,3-DHBP and 2,2′,3-THBP) and catecholic compounds (catechol, 3-methylcatechol, and 4-methylcatechol), were used for substrate specificity analysis.

As shown in Table 3, FlnD1D2 showed higher activity toward biphenylic compounds such as 2,3-DHBP and 2,2′,3-THBP. FlnD1D2 showed little or negligible activity toward single catecholic compounds such as catechol, 3-methylcatechol, and 4-methylcatechol. This result suggests that FlnD1D2 is specific for biphenylic compounds. Kinetic studies were performed only for the biphenyl compounds with relatively high activities. Relevant kinetic parameter details are shown in Table 4. The K_m and k_cat/K_m values of FlnD1D2 were 97.2 μM and 1.5 × 10⁻² μM⁻¹·sec⁻¹ for 2,3-DHBP and 168.0 μM and 0.5 × 10⁻² μM⁻¹·sec⁻¹ for 2,2′,3-THBP, respectively. Previous research has suggested that the enzyme affinity for the above two substrates is similar, but that the catalytic efficiency of FlnD1D2 is 3.0-fold higher for 2,3-DHBP than for 2,2′,3-THBP. The tendency for the lower activity of 2′-substituted 2,3-DHBPs was also found in CarBaBb from P. resinovorans CA10²¹⁾ and 2,2′,3-trihydroxybiphenyl dioxygenase (DbfB) from Sphingomonas sp. RW11,⁸⁾ which catalyzes 2′-substituted 2,3-DHBP as an original substrate.

Table 3. Substrate specificity of FlnD1D2.

Substrate	Relative activity^a (%)
2,3-DHBP	100 ± 11.5
2,2′,3-THBP	52.2 ± 3.8
Catechol	0.1 ± 0.02
3-methylcatechol	1.4 ± 0.13
4-methylcatechol	0.5 ± 0.05

Note:

^a Relative activity was expressed as the percentage of enzyme activity using 2,3-DHBP, taken to be 100%.

Table 4. Kinetic parameters of FlnD1D2.

Substrate	Vmax (μM/min)	Km (μM)	Kcat (sec^−1)	Kcat/Km (μM^−1 sec^−1)
2,3-DHBP	9.6	97.2	14.8	1.5 × 10^−2
2,2′,3-THBP	5.8	168.0	9.0	0.5 × 10^−2

Phylogenetic analysis

All heterooligomeric extradiol dioxygenases, including FlnD1D2, have been classified as type II extradiol dioxygenases. Because the three-dimensional structure of the type II enzymes LigAB,⁶⁾ DesB,¹⁹⁾ CnbCaCb,²⁰⁾ and CarBaBb (our unpublished results) have already been determined, we aligned their amino acid sequences based on their secondary structures. Using the resultant alignment as a template, amino acid sequence alignments of type II extradiol dioxygenases, including FlnD1D2 and the above-mentioned four enzymes, were performed. Phylogenetic trees of FlnD1 (large, catalytic subunit) and FlnD2 (small, possibly structural subunit) with the corresponding subunits of other type II extradiol dioxygenases are shown in Fig. 4(A) and (B), respectively. The phylogenetic trees of the large subunits suggested that type II extradiol dioxygenases can be grouped into several subgroups (Fig. 4(A)). Subgroup (a) consists of homooligomeric extradiol dioxygenase. The members of subgroup (b) are heterooligomeric, and these extradiol dioxygenases have two similar subunits (>24% identities).^25–27) In fact, both catalytic and non-catalytic subunits constitute a single subgroup (b) (Fig. 4(A)). Subgroups (c), (d), and (e) consist of heterooligomeric extradiol dioxygenases. PcmA from Arthrobacter keyseri 12B²⁸⁾ and DesB from Sphingobium sp. SYK-6¹⁹⁾ are homooligomeric extradiol dioxygenases, but these enzymes are composed of single polypeptides with two domains, N- and C-termini, that apparently correspond to large and small subunits of the protocatechuate 4,5-dioxygenase (data not shown). Furthermore, DesZ from Sphingobium sp. SYK-6²⁹⁾ and PhnC from Burkholderia sp. RP007³⁰⁾ are also homooligomeric extradiol dioxygenases, but these enzymes do not possess the region corresponding to the small subunit of the protocatechuate 4,5-dioxygenase. These homooligomeric enzymes appear to derive from heterooligomeric extradiol dioxygenases. As shown in Fig. 4(B), the branching of the large subunits corresponds to that of the small subunits [see subgroups (c), (d), and (e)]. Together with other extradiol dioxygenases involved in dibenzofuran/dibenzo-p-dioxin degradation, each FlnD1D2 subunit has a clear and independent lineage [subgroup (d)]. These results clearly suggest that FlnD1D2 is a novel heterooligomeric extradiol dioxygenase, which may be different from LigAB, DesB, CnbCaCb, and CarBaBb. In addition, gel filtration chromatography suggested that FlnD1D2 has a α₄β₄ tertiary structure, while CarBaBb and other type II extradiol dioxygenases have α₂β₂ structures. It will be interesting to clarify the formation of the novel α₄β₄ oligomer structure. To investigate these questions, further studies should focus on solving the crystal structure of FlnD1D2.

Fig 4. Phylogenetic trees of large (A) and small (B) subunits of type II extradiol dioxygenases.

Notes: The bootstrap values show more than 50% from 1,000 resamplings. The scale bar denotes 1.0 substitution per site. Proteins used in phylogenetic analyses shown in panel (A) are as follows (ID in parentheses indicate accession number of GenBank); FlnD1_DFA3, α subunit of 2′-carboxy-2,3-dihydroxybiphenyl 1,2-dioxygenase from Rhodococcus sp. DFA3 (AB181125.1); FlnD1_DBF63, α subunit of 2′-carboxy-2,3-dihydroxybiphenyl 1,2-dioxygenase from Terrabacter sp. DBF63 (AB095015.1); CarBb_CA10, β subunit of 2′-aminobiphenyl-2, 3-diol 1,2-dioxygenase from Pseudomonas resinovorans CA10 (D89064.1); CarBb_KA1, β subunit of 2′-aminobiphenyl-2, 3-diol 1,2-dioxygenase from Sphingomonas sp. KA1 (WP011607921.1); CarBb_OC9, β subunit of 2′-aminobiphenyl-2, 3-diol 1,2-dioxygenase from Kordiimonas sp. OC9 (BAH60849.1); PhnC_BR007, extradiol dioxygenase from Burkholderia sp. RP007 (AAD09870.1); LigB_SYK-6, β subunit of protocatechuate 4,5-dioxygenase from Sphingobium sp. SYK-6 (WP_014075576.1); FldU_LB126, α subunit of protocatechuate 4,5-dioxygenase from Sphingomonas sp. LB126 (CAB87561.1); Edi4_YK2, extradiol dioxygenase from Rhodococcus sp. YK2 (AB070456.1); HppB_PWD1, 2,3-dihydroxyphenylpropionate 1,2-dioxygenase from Rhodococcus globerulus PWD1 (AAB81314.1); OhpD_V49, 3-(2,3-dihydroxyphenyl) propionic acid dioxygenase from Rhodococcus sp.V49 (AAF81826.1); MhpB_TA441, 3-(2,3-dihydroxyphenyl)propionate 1,2-dioxygenase from Comamonas testosteroni TA441 (BAA82879.1); MhbB_K-12, 2,3-dihydroxyphenylpropionate 1,2-dioxygenase from Escherichia coli K-12 (BAA13053.1); MpcI_JMP222, catechol 2,3-dioxygenase from Alcaligenes eutrophus JMP222 (CAA36665.1); DesB_SYK-6, gallate dioxygenase from Sphingobium sp. SYK-6 (BAK65008.1); DesZ_SYK-6, 3-O-methylgallate 3,4-dioxygenase from Sphingobium sp. SYK-6 (BAK66578.1); PmdB_BR6020, β subunit of protocatechuate 4,5-dioxygenase from Comamonas testosteroni BR6020 (AAK73573.1); ProOb_NGJ1, β subunit of protocatechuate 4,5-dioxygenase from Pseudomonas ochraceae NGJ1 (BAD04058.1); PcmA_12B, protocatechuate 4,5-dioxygenase from Arthrobacter keyseri 12B (AAK16524.1); EdoD_l1, extradiol dioxygenase from Rhodococcus sp. I1 (CAA06875.1); CnbCa_CNB-1, α subunit of 2-amino-5-chlorophenol 1,6-dioxygenase from Comamonas testosteroni CNB-1 (YP001967697.1); CnbCb_CNB-1, β subunit of 2-amino-5-chlorophenol 1,6-dioxygenase from Comamonas testosteroni CNB-1 (YP001967698.1); NbzCa_HS12, α subunit of 2-aminophenol 1,6-dioxygenase from Pseudomonas putida HS12 (YP009076856.1); NbzCb_HS12, α subunit of 2-aminophenol 1,6-dioxygenase from Pseudomonas putida HS12 (YP009076857.1); AmnA_AP-3, α subunit of 2-aminophenol 1,6-dioxygenase from Pseudomonas sp. AP-3 (BAB03532.1); and AmnB_AP-3, β subunit of 2-aminophenol 1,6-dioxygenase from Pseudomonas sp. AP-3 (BAB03531.1). Similarly, proteins used for panel (B) are as follows (ID in parentheses indicate accession number of GenBank): FlnD2_DFA3, β subunit of 2′-carboxy-2,3-dihydroxybiphenyl 1,2-dioxygenase from Rhodococcus sp. DFA3 (BAD51801.1); ORF6_YK2, unnamed protein product from Rhodococcus sp. YK2 (BAC00807.1); CarBa_CA10, α subunit of 2′-aminobiphenyl-2,3-diol 1,2-dioxygenase from Pseudomonas resinovorans CA10 (NP758568.1); CarBa_KA1, α subunit of 2′-aminobiphenyl-2, 3-diol 1,2-dioxygenase from Sphingomonas sp. KA1 (BAB88911.1); CarBa_OC9, α subunit of 2′-aminobiphenyl-2, 3-diol 1,2-dioxygenase from Kordiimonas sp. OC9 (BAH60848.1); LigA_SYK-6, α subunit of protocatechuate 4,5-dioxygenase from Sphingobium sp. SYK-6 (BAK65926.1); FldV_LB126, β subunit of protocatechuate 4,5-dioxygenase from Sphingomonas sp. LB126 (CAB87562.1); PmdA_BR6020, α subunit of protocatechuate 4,5-dioxygenase from Comamonas testosteroni strain BR6020 (AAK73572.1); and ProOa_NGJ1, α subunit of protocatechuate 4,5-dioxygenase from Pseudomonas ochraceae NGJ1 (BAD04057.1).

Author contributions

Hideaki Nojiri and Kenichi Iwata designed the project. Tatsuro Kotake performed experiments and wrote the manuscript. Jun Matsuzawa, Chiho Suzuki-Minakuchi, and Kazunori Okada contributed analysis and discussion.

Supplemental material

The supplemental material for this paper is available at http://dx.doi.org/10.1080/09168451.2015.1123605.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

Abbreviations: Ap, ampicillin; DF, dibenzofuran; FN, fluorene; km, kanamycin; LB, Luria–Bertani; PAHs, polycyclic aromatic hydrocarbons; SD, Shine–Dalgarno; 2,2′,3-THBP, 2,2′,3-trihydroxybipenyl, 2,3-DHBP, 2,3-dihydroxybiphenyl; 2′-carboxy-2,3-DHBP, 2′-carboxy-2,3-dihydroxybiphenyl.