Characterization of Hybrid Toluate and Benzoate Dioxygenases

Abstract

Toluate dioxygenase of Pseudomonas putida mt-2 (TADO_mt2) and benzoate dioxygenase of Acinetobacter calcoaceticus ADP1 (BADO_ADP1) catalyze the 1,2-dihydroxylation of different ranges of benzoates. The catalytic component of these enzymes is an oxygenase consisting of two subunits. To investigate the structural determinants of substrate specificity in these ring-hydroxylating dioxygenases, hybrid oxygenases consisting of the α subunit of one enzyme and the β subunit of the other were prepared, and their respective specificities were compared to those of the parent enzymes. Reconstituted BADO_ADP1 utilized four of the seven tested benzoates in the following order of apparent specificity: benzoate > 3-methylbenzoate > 3-chlorobenzoate > 2-methylbenzoate. This is a significantly narrower apparent specificity than for TADO_mt2 (3-methylbenzoate > benzoate ∼ 3-chlorobenzoate > 4-methylbenzoate ∼ 4-chlorobenzoate ≫ 2-methylbenzoate ∼ 2-chlorobenzoate [Y. Ge, F. H. Vaillancourt, N. Y. Agar, and L. D. Eltis, J. Bacteriol. 184:4096-4103, 2002]). The apparent substrate specificity of the α_Bβ_T hybrid oxygenase for these benzoates corresponded to that of BADO_ADP1, the parent from which the α subunit originated. In contrast, the apparent substrate specificity of the α_Tβ_B hybrid oxygenase differed slightly from that of TADO_mt2 (3-chlorobenzoate > 3-methylbenzoate > benzoate ∼ 4-methylbenzoate > 4-chlorobenzoate > 2-methylbenzoate > 2-chlorobenzoate). Moreover, the α_Tβ_B hybrid catalyzed the 1,6-dihydroxylation of 2-methylbenzoate, not the 1,2-dihydroxylation catalyzed by the TADO_mt2 parent. Finally, the turnover of this ortho-substituted benzoate was much better coupled to O₂ utilization in the hybrid than in the parent. Overall, these results support the notion that the α subunit harbors the principal determinants of specificity in ring-hydroxylating dioxygenases. However, they also demonstrate that the β subunit contributes significantly to the enzyme's function.

Bacterial ring-hydroxylating dioxygenases are involved in the aerobic catabolism of a wide variety of aromatic compounds and thus play a critical role in the global carbon cycle. These multicomponent enzymes catalyze the NAD(P)H-dependent dihydroxylation of the aromatic ring, incorporating both atoms of dioxygen into the product (14, 21). Ring-hydroxylating dioxygenases have applications in the biodegradation of pollutants, many of which are aromatic compounds (49). In addition, due to their regio- and enantiospecificity, these enzymes are of burgeoning importance in generating synthons for the pharmaceutical and chemical industries (7, 22).

Ring-hydroxylating dioxygenases typically consist of a hexameric oxygenase (ISP) of (αβ)₃ configuration, a reductase (RED), and sometimes a ferredoxin. The α subunit of the ISP contains a Rieske-type Fe₂S₂ center and an active-site mononuclear iron. RED and the ferredoxin, if present, transfer reducing equivalents from NAD(P)H to the ISP. Due to their importance and potential applications, considerable effort has been focused on identifying the specificity determinants of these enzymes. Structural studies of naphthalene dioxygenase (NDO) (9) and biphenyl dioxygenase (BPDO) (12) indicate that the substrate-binding pocket is contained entirely within the C terminus of the ISP α subunit. To functionally evaluate specificity determinants, hybrid ISPs consisting of the α subunit of one enzyme and the β subunit of a related enzyme have been studied. Such experiments with NDO, BPDO, and related enzymes indicate that the α subunit is responsible for substrate preference (2, 5, 31, 40, 41). Directed mutagenesis and gene-shuffling approaches have further indicated that the α subunit harbors the principal determinants of substrate preference (3, 32, 36, 42, 43, 48). In each of these studies, enzyme function was evaluated solely in terms of substrate preference, in part due to the limited solubility of the substrates. Moreover, this preference was evaluated by using whole-cell biotransformation. Interestingly, some studies using purified hybrid ISPs indicate that the β subunit can influence the substrate preference (28, 33). This is consistent with structural data indicating that the β subunit interacts with the α subunit close to the active site of the enzyme (9, 12).

Toluate dioxygenase of Pseudomonas putida mt-2 (TADO_mt2) (EC 1.14.12.-) (27) and benzoate dioxygenase of Acinetobacter calcoaceticus ADP1 (BADO_ADP1) (EC 1.14.12.10) are group II dioxygenases (37) (class IB according to the system of Batie et al. [4]) that catalyze the dihydroxylation of benzoates (Fig. 1). The α and β subunits of TADO_mt2 are encoded by xylXY, respectively, and those of BADO_ADP1 are encoded by benAB, respectively (39). The RED components of TADO_mt2 and BADO_ADP1 are encoded by xylZ and benC, respectively (25, 38). The ISP components of TADO_mt2 (ISP_TADO or α_Tβ_T) and BADO_ADP1 (ISP_BADO or α_Bβ_B) share approximately 62% sequence identity yet transform different ranges of substituted benzoates. Thus, TADO_mt2 transforms a wide range of substituted benzoates (54) and shows highest specificity for 3-methylbenazoate (20). In contrast, BADO transforms a much narrower range of substrates (53). The different specificities of these related enzymes and the solubility of their substrates allow kinetics studies to be performed with a wider range of substrate concentrations, thereby facilitating a more thorough investigation of the structural determinants of function in this important class of enzymes.

FIG. 1.

The reaction catalyzed by TADO_mt2 and BADO_ADP1. These ring-hydroxylating dioxygenases initiate the catabolism of substituted benzoates, catalyzing their transformation to the corresponding cis-1,2-dihydroxycyclohexadienes. Each enzyme consists of two components: an ISP, encoded by xylXY or benAB, and a RED, encoded by xylZ or benC.

In the present study, BADO_ADP1 was overexpressed and purified by using approaches similar to those developed for TADO_mt2. Hybrids ISPs consisting of the α subunit of one enzyme and the β subunit of the other were expressed and purified, and their respective specificities for a range of substituted benzoates were compared to those of the parent enzymes. The coupling of substrate utilization in the hybrid enzymes was also investigated. The contributions of the different subunits to the activities of these enzymes are discussed.

MATERIALS AND METHODS

Reagents.

Benzoates were from Fluka Chimie. Catechols and catalase were from Sigma-Aldrich Fine Chemicals. Restriction enzymes, T4 ligase, and chromatography resins were from Amersham Biosciences. PfuI DNA polymerase was from Stratagene. Nickel-nitrilotriacetic acid Superflow resin was purchased from Qiagen. Oligonucleotides were purchased from Invitrogen Life Technologies. All other chemicals were of analytical grade. Buffers were prepared with water purified on a Barnstead (San Diego, Calif.) NANOpure UV apparatus to a resistivity of greater than 17 M × Ω · cm.

Strains and plasmids.

The strains and plasmids used in this work are listed in Table 1. Escherichia coli and P. putida strains were grown at 37 and 30°C, respectively. The medium was supplemented with 100 μg of ampicillin per ml and/or 10 μg of tetracycline per ml, as appropriate. Strains used in the propagation of DNA were grown in Luria-Bertani broth. Strains used in the expression of protein were grown in Terrific Broth (1) supplemented at 10 ml/liter with an HCl-solubilized mineral solution (20, 34, 50). One liter of medium in a 2-liter flask was inoculated with 10 ml of an overnight culture. Each of the four ISPs (α_Tβ_T, α_Bβ_B, α_Bβ_T, and α_Tβ_B) was expressed in strain P. putida CL01 containing pVLTXYZ1, pVLTAB1, pVLTAY1, and pJBXB1, respectively. When cultures reached an optical density at 600 nm of 0.6, isopropyl-1-thio-β-d-galactopyranoside (IPTG) and 3-methylbenzoate were added to final concentrations of 0.1 and 1 mM, respectively. This amount of 3-methylbenzoate was added a second time 3 h after the initial induction. The cultures were incubated for an additional 20 h before harvesting. RED_TADO and RED_BADO were expressed under similar conditions by using P. putida KT2442 containing pVLTZ1 and pVLTC1, respectively. Cultures were induced with 0.5 mM IPTG and incubated for an additional 20 h before harvesting. Cell pellets were frozen at −80°C until further use.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant genotype or propertiesa	Reference
Strains
E. coli DH5α	F−endA1 hsdR17 (rk− mk+) supE44 thi-1 λ−recA1 gyrA96 relA1 deoR Δ(lacZYA-argF) U169 φ80dlacZΔM15	23
P. putida KT2442	Prototrophic; hsdR Rif	35
P. putida CL01	P. putida KT2442 carrying xylS; Rif Pipr	20
Plasmids
pIB1354	pUC19 containing the benABC genes; Apr	39
pVLT31	Broad-host-range expression vector; RSF 1010, Ptac promoter, lacIq, Tcr	15
pLEHP20	Plac promoter, His tag expression vector, Apr	17
pEMRBS	pEMBL18 carrying Pm promoter and the phage T7 gene 10 leader sequence; Apr	F. H. Vaillancourt and L. D. Eltis, unpublished data
pEMBL18	Plac promoter; Apr	16
pJB655	Pm promoter, containing xylS; Apr	6
pEMXYZ1	Plac promoter, pEMBL18 carrying xylXYZ; Apr	20
pEMXB1	Plac promoter, pEMBL18 carrying xylX benB; Apr	This study
pEMA1	pEMRBS carrying first part of benA; Apr	This study
pEMAB1	pEMRBS carrying benAB; Apr	This study
pEMAY1	pEMRBS carrying benA xylY; Apr	This study
pVLTXYZ1	pVLT31 carrying xylXYZ; Tcr	20
pVLTC1	pVLT31 carrying benC with His tag; Tcr	This study
pVLTZ1	pVLT31 carrying xylZ with His tag; Tcr	20
pJBXB1	Pm promoter; Apr carrying xyIX and benB	This study
pVLTAB1	pVLT31 carrying benAB; Tcr	This study
pVLTAY1	pVLT31 carrying benA and xyIY; Tcr	This study

^aAp^r, ampicillin resistance; Pip^r, piperacillin resistance; Rif, rifampin resistance; Tc^r, tetracycline resistance.

DNA manipulation.

DNA was propagated by using E. coli DH5α. DNA was purified, digested, and ligated by standard protocols (46). PCR amplification was performed with a Thermolyne model DB66P25 thermocycler (Barnstead). Oligonucleotide-directed mutagenesis was performed by overlap extension PCR (47) and derivatives of pEMBL18 as the template DNA. Clones constructed by using PCR products were sequenced to verify the integrity of the sequence. Cells were transformed with plasmids via electroporation with a Gene Pulser transfection apparatus and pulse controller (Bio-Rad, Hercules, Calif.) according to the instructions of the manufacturer. Nucleotide sequencing was performed with the ABI Dye-Deoxy terminator protocol and an ABI model 373 Stretch DNA sequencer at the Nucleic Acid Analysis Unit at Université Laval.

Construction of expression vectors.

Expression vectors for BADO_ADP1 and hybrid ISPs were designed based on the systems that yielded the best expression of the TADO_mt2 components (20). To express α_Bβ_B, a vector was constructed by initially introducing an NcoI site at the start codon of benA in pIB1354. The initial portion of benA was then cloned as a 640-bp NcoI-HincII fragment into pEMRBS. This yielded pEMA1, in which the start codon of benA is immediately downstream of the P_m promoter and the ribosome binding site of the phage T7 gene leader sequence. The intact benA together with benB was reconstructed by inserting a 1.7-kb HincII fragment from pIB1354 into pEMA1, yielding pEMAB1. Finally, a 2.4-kb SacI-HindIII fragment containing the benAB genes together with the P_m promoter and the ribosome binding site was excised from pEMAB1 and cloned into pVLT31, yielding pVLTAB1.

To express α_Tβ_B, a BsmI site was introduced 2 bp downstream of xylX in pEMXYZ1 by directed mutagenesis with primer Bsm1XY (CTTCGTAGGAAGCATTCATTTACACGCCC [an introduced BsmI site is underlined]). A 1-kb BsmI-HincII fragment carrying benB was then excised from pIB1354 and used to replace the corresponding fragment in pEMXYZ1, yielding pEMXB1. This strategy positioned benB immediately downstream of xylX, analogous to the position of xylY in the TOL operon, without changing the sequence of XylX or BenB. Finally, a 2.5-kb XbaI-KpnI fragment from pEMXB1 containing the P_m promoter, xylX, benB, and part of benC was subcloned into pJB655 to form pJBXB1. This last cloning step took advantage of the XbaI site in the P_m promoter.

To express α_Bβ_T, xylY was amplified by PCR with primers Ay-1 (GGGAATTCGAATGCTACTATCTCCTACGAA [an introduced BsmI site is underlined]) and Ay-2 (CCCCGAAGCTTAAGTCAGTGGCAACC [an introduced HindIII site is underlined]). The resulting 500-bp fragment was digested with BsmI and HindIII and cloned into pEMAB1, yielding pEMAY1. Finally, a 2.3-kb SacI-HindIII fragment containing the benA and xylY genes together with the P_m promoter was excised from pEMAY1 and cloned into pVLT31, yielding pVLTAY1. This strategy resulted in the addition of two amino acids at the N terminus of XylY, asparagine and alanine, which are the second and third amino acids of BenB, respectively. This was done to preserve the same 4-bp overlap between benA and xylY that exists between benA and benB (26). Inspection of the sequence alignment and the structure of BPDO (12) indicates that the addition of these two amino acids to the N terminus of XylY should not change the properties of XylY.

RED_BADO was expressed as a His-tagged protein by using a system analogous to that for RED_TADO (20). Accordingly, benC was amplified by PCR with primers BCfor (CGCCCTGCAGTCACGACGTTG [a PstI site is underlined]) and BCrev (GCCCGCTAGCTTATATTTGAATAGG [an NheI site is underlined) from pIB1354. The resulting 1.2-kb fragment was digested with NheI and PstI and ligated into appropriately digested pLEHP20 following a sequence encoding a six-histidine tag. The His-tagged BenC encoded by this construct contains all of the residues of the wild-type protein except for the initial methionine. The gene encoding the fusion protein was then cloned into pVLT31 as an XbaI-PstI fragment, yielding pVLTC1.

Purification of proteins.

The dioxygenase components were purified anaerobically from cell pellets essentially as previously described for the TADO_mt2 components (20). Accordingly, the wild-type and hybrid ISPs were purified by using anion-exchange, gel filtration, and hydrophobic interaction chromatographies. The His-tagged REDs (ht-REDs) were purified by using immobilized metal affinity chromatography. Anaerobically prepared buffers used in the purification of ISPs contained 10% glycerol, 0.25 mM Fe(NH₄)₂(SO₄)₂, and 2 mM dithiothreitol to maximize the specific activity of the preparation. Protein-containing fractions were concentrated by ultrafiltration with an Amicon stirred cell equipped with a YM10 filter (Millipore, Nepean, Ontario, Canada). Preparations of purified protein were flash frozen in liquid nitrogen as beads and stored at −80°C until further use. Protein stored in this manner exhibited no loss of activity over 6 months.

Determination of protein purity and concentration.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed on a Bio-Rad Mini-Protean II apparatus with Coomassie blue staining according to established procedures (1). Protein concentrations were determined by the Bradford method (8) with bovine serum albumin as a standard. ISP concentrations were calculated by using the extinction coefficients presented in Results.

Determination of iron and sulfur contents.

Iron and sulfur concentrations were determined colorimetrically with Ferene S (24) and N,N-dimethyl-p-phenylene-diamine (11), respectively. Samples were manipulated in gas-tight cuvettes. All assays were performed in duplicate. The correlation coefficients of the standard curves were at least 0.98.

Steady-state kinetic studies.

The dioxygenase-catalyzed reactions were monitored polarographically following the consumption of O₂ by using a Clarke-type oxygen electrode (model 5301; Yellow Springs Instruments, Yellow Springs, Ohio), a thermojacketed respiration chamber, an O₂ meter, and a microcomputer as described previously (20, 50). The full scale was established by using buffer equilibrated with 5, 20, 50, or 100% O₂ (see below), depending on the experiment. The oxygen electrode was zeroed and calibrated as described previously (20). Initial velocities were determined from progress curves by analyzing the data with Microsoft Excel.

The standard activity assay was performed in a total volume of 1.4 ml of air-saturated 100 mM phosphate buffer (pH 7.0; 25 ± 1°C) (∼290 μM dissolved O₂) containing 430 μM NADH and either 100 μM 3-methylbenzoate (for TADO_mt2) or 100 μM benzoate (for BADO_ADP1). RED was added to a final concentration of 2.0 μM, and the background was recorded. The reaction was initiated by injecting the appropriate ISP into the reaction chamber to a final concentration of 0.37 μM. One unit of enzymatic activity was defined as the quantity of enzyme required to consume 1 μmol of O₂/min.

Reaction buffers containing different concentrations of dissolved O₂ were prepared by using humidified mixtures of O₂ and N₂ gases and were transferred to the gas-flushed reaction chamber by using a gas-tight syringe as described previously (20, 50).

Analysis of steady-state data.

Steady-state kinetic data were analyzed by using an equation that describes a compulsory-order ternary-complex mechanism (13) in which the binding of the aromatic substrate, A, precedes that of O₂:

In this equation, K_mA represents the K_m for the aromatic substrate, represents the K_m for O₂, and K_dA represents the dissociation constant for the aromatic substrate. The steady-state kinetic parameters were evaluated by using the program LEONORA (13). They are apparent, as they depend on the concentration of RED (20).

Coupling measurements.

Coupling experiments for all substrates were carried out under the same conditions as for the standard activity assay except that the RED, ISP, and substrate concentrations were 4, 1.8, and 215 μM, respectively. Reactions were initiated by adding ISP and quenched by diluting 200 μl of the reaction mixture with 400 μl of methanol 3 min after the initiation of the reaction or when O₂ consumption stopped. Oxygen consumption was monitored with the O₂ electrode. The consumption of aromatic substrate was determined by HPLC measurements. The amount of substrate remaining was determined from the area of the absorbance peak at 280 nm at the respective retention time. Standard curves were obtained from treating the pure substituted-benzoate solution in the same way as reaction mixtures at different known concentrations. The amount of hydrogen peroxide was determined by measuring the amount of O₂ released upon the addition of catalase as described previously (20).

Identification of reaction products.

The products of dioxygenase-catalyzed reactions were identified in reactions performed and quenched as described for the coupling assay except that the reaction mixture also contained 1.2 μM XylL and 128 μM NAD⁺ to transform cis-diols to the corresponding catechol. Substituted catechols were identified in reaction mixtures by comparing the retention times and absorption spectra of HPLC peaks with those for commercially available catechols. In some instances, product identification was confirmed by further treating reaction mixtures with catechol 2,3-dioxygenase and comparing the absorption spectra of the cleavage products with those of known compounds.

HPLC measurements.

HPLC measurements were performed with a Millennium³² system (Waters Corporation, Milford, Mass.), including a Waters 2996 photodiode array detector and an Alliance Waters 2695 separations module equipped with a C₁₈ reverse-phase octyldecyl silane hypersil column (4 by 125 mm). All components were interfaced with a Milliennium³² Client/Server. Samples of 10 μl were injected, and the column was operated at a flow rate of 1 ml/min. Benzoates, cis-diols, and catechols were resolved by using a 4:1 (vol/vol) mixture of 0.3% H₃PO₄ and acetonitrile, a 9:1 (vol/vol) mixture of H₂O and acetonitrile, and an 4:1 (vol/vol) mixture of H₂O and acetonitrile, respectively.

UV-visible absorption spectroscopy.

Absorption spectra were recorded with a Varian Cary 1E spectrophotometer equipped with a thermojacketed cuvette holder maintained at 25°C. The spectrophotometer was interfaced to a microcomputer and controlled by Cary WinUV software (version 2.00). Samples contained 1.2 μM protein in 25 mM HEPES buffer, pH 7.3. Spectra of anaerobic samples were recorded with a 3-ml gas-tight cuvette (Hellma, Concord, Ontario, Canada). Oxidized samples were prepared in a glove box by adding several grains of K₃Fe(CN)₆ to the protein sample and passing the sample through a small desalting column (0.7 by 6 cm; Bio-Gel P6 DG) equilibrated with the buffer of choice.

RESULTS

Expression and purification.

Wild-type and hybrid ISPs were expressed at relatively high levels in P. putida CL01 by using the constructs and protocols described in Material and Methods; each ISP constituted more than 20% of the total cellular protein (results not shown). Approximately 32 to 100 mg of each ISP was purified from 24 g (wet weight) of cells (Table 2). The final preparation of each ISP was judged to be greater than 90% pure based on a Coomassie blue-stained denaturing gel (Fig. 2). Elution of α_Bβ_B and hybrid ISPs from the gel filtration column was consistent with each ISP having a molecular mass of approximately 215 kDa (20).

TABLE 2.

Purification of ISPs

ISP	Protein yield (mg)	Sp act (U/mg)a	Fe content (Fe/α3β3)	S content (S/α3β3)	A280/A323
					Oxidized	Reduced
αBβB	75	5.0	11	6.0	6.1	12.7
αTβT	39	3.8	16	5.8	4.5	6.7
αTβB	100	2.3	15	5.9	4.5	6.7
αBβT	32	3.1	12	5.9	6.1	12.7

^aActivity units are defined in Materials and Methods. Specific activities were obtained with 3-methylbenzoate for α_Tβ_T and α_Tβ_B and with benzoate for α_Bβ_B and α_Bβ_T.

FIG. 2.

Denaturing gels of purified preparations of TADO_mt2 and BADO_ADP1 components. The first four gels were loaded with molecular weight standards (lanes M) and 5 μg of the indicated wild-type or hybrid ISP preparation. The α and β subunits are indicated. The last two gels were loaded with molecular weight standards and 5 μg of indicated RED preparation.

RED_BADO was expressed as a His-tagged protein in P. putida KT2442 at a level similar to that reported for ht-RED_TADO (20). ht-RED_BADO was purified anaerobically in a single step by using immobilized metal affinity chromatography as described for ht-RED_TADO, thereby preventing loss of the flavin adenine dinucleotide (20). Approximately 60 mg of ht-RED_BADO was purified to greater than 99% apparent homogeneity from 4 liters of culture (Fig. 2). The His tag was not removed, as previous studies with RED_TADO indicated that its presence did not influence dioxygenase activity.

Characterization of metallocenters.

The iron and sulfur contents of purified ISPs were determined with anaerobically desalted samples of protein. The values indicate that each ISP contained full complements of FeS and mononuclear iron prosthetic groups (Table 2). While all ISPs seemed to contain adventitiously bound iron, ISPs containing the α subunit of TADO_mt2 (α_Tβ_T and α_Tβ_B) contained more.

The UV-visible spectra of α_Bβ_B, α_Tβ_B, and α_Bβ_T were similar to that of α_Tβ_T (20), absorbing maximally at 280, 323, and 455 nm. As observed for α_Tβ_T (20), the anaerobic addition of sodium hydrosulfite to preparations of these ISPs as purified did not affect their spectra. This indicates that as purified, the Rieske-type FeS cluster of each ISP was fully reduced. As reported for α_Tβ_T, the FeS cluster could be oxidized in samples of each ISP by either exposure to air for 20 min or treatment with a slight excess of K₃Fe(CN)₆ (results not shown). The R value (A₂₈₀/A₃₂₃) of each oxidized ISP is lower than that of the corresponding reduced form (Table 2), principally because the absorption bands of the oxidized cluster are more intense than those of the reduced cluster. Interestingly, the R value of purified α_Tβ_T was significantly lower than that of α_Bβ_B (Table 2). Similarly, preparations of purified α_Tβ_T had a more intense brown color than preparations of α_Bβ_B of similar concentration. Indeed, the extinction coefficients of reduced α_Tβ_T and α_Bβ_B at 323 nm were 83.42 and 40.55 cm⁻¹ mM⁻¹, respectively (25 mM HEPES buffer [pH 7.3], 10% glycerol, 25°C), based on the sulfur content of the ISP. These values are comparable to those reported for oxidized 2-halobenzoate 1,2-dioxygenase (34 cm⁻¹ mM⁻¹) (18) and for BADO_C1 from Pseudomonas arvilla C-1 (58.6 cm⁻¹ mM⁻¹) (53) based on protein concentration. However, the absorption of the oxidized ISP is generally 50% higher than that of the reduced protein, indicating that the present preparations of α_Tβ_T and α_Bβ_B probably contain a higher proportion of their Fe₂S₂ cluster. Finally, the R value of each hybrid ISP corresponded to that of the parent from which the α subunit originated. This is consistent with structural data showing that the Fe₂S₂ cluster in ring-hydroxylating dioxygenases is contained entirely with the α subunit (30).

The stabilities of the wild-type and hybrid ISPs were compared by monitoring the absorption at 455 nm (A₄₅₅) of oxidized samples incubated at room temperature. After 48 h, the A₄₅₅ of aerobic, oxidized samples of α_Tβ_T, α_Tβ_B, α_Bβ_B, and α_Bβ_T decreased by 6, 12, 26, and 58%, respectively. These data indicate that exchanging the β subunit decreased the stability of the ISP with respect to that of the parent enzyme from which the α subunit originated. However, the change in A₄₅₅ was insignificant over the time course of a kinetic experiment. Finally, anaerobically incubated ISP samples were approximately five times more stable than the aerobic samples (data not shown).

Anaerobically purified preparations of ht-RED_BADO and ht-RED_TADO each absorbed maximally at 270, 340, and 455 nm and had essentially identical R values (A₂₇₀/A₄₅₅) of 3.1. Exposure of samples of RED to air resulted in an increase of the R value to a maximum of 5.9, which was observed at 20 min. Oxidized samples of RED could be reduced with a small molar excess of NADH (results not shown).

In vitro reconstitution of dioxygenase activity.

Dioxygenase activity was reconstituted in vitro under the standard conditions described for TADO_mt2 (molar ratio of RED to ISP of approximately 5.4:1, 0.1 M ionic strength phosphate buffer [pH 7.0], 25°C) and monitored by using an oxygraph assay (20). The specific activity of each reconstituted dioxygenase was determined by using the derivative of benzoate for which it showed the highest apparent specificity (i.e., 3-methylbenzoate for α_Tβ_T and α_Tβ_B and benzoate for α_Bβ_B and α_Bβ_T [see below]) and by using the RED of the dioxygenase from which the α subunit originated (i.e., ht-RED_TADO for α_Tβ_T and α_Tβ_B and ht-RED_BADO for α_Bβ_B and α_Bβ_T). Under these conditions, the specific activities of the dioxygenases reconstituted with α_Tβ_T, α_Bβ_B, α_Tβ_B, and α_Bβ_T were 3.8, 5.0, 2.3, and 3.1 U/mg, respectively.

Steady-state kinetics.

Steady-state kinetics studies were performed with a variety of substituted benzoates to determine the apparent specificities of wild-type and hybrid dioxygenases.

BADO_ADP1 had a narrower specificity than TADO_mt2, transforming only four of seven selected substrates (Table 3). The best substrate for BADO_ADP1 was benzoate (K_mA = 26 ± 1 μM, k_cat = 8.6 ± 0.1 s⁻¹, and = 53 ± 2 μM), and the enzyme utilized substituted benzoates in the following order of apparent specificity: benzoate > 3-methylbenzoate > 3-chlorobenzoate > 2-methylbenzoate. In contrast, TADO_mt2 utilized substituted benzoates in the following order of apparent specificity: 3-methylbenzoate > benzoate ∼ 3-chlorobenzoate > 4-methylbenzoate ∼ 4-chlorobenzoate ≫ 2-methylbenzoate ∼ 2-chlorobenzoate (20). Thus, BADO_ADP1 and TADO_mt2 differed significantly in their abilities to utilize para-substituted benzoates.

TABLE 3.

Apparent steady-state kinetic parameters of ISPs for selected substituted benzoates^a

ISP	Substrate	KmA (μM)	KmO2 μM	kcat (s−1)	kcat/KmA (104 M−1 s−1)	kcat/KmO2 (104 M−1 s−1)
αBβB	Benzoate	26 (1)	53 (2)	8.6 (0.1)	33 (3)	16 (1)
	2-Methylbenzoate	1,401 (48)	247 (20)	4.2 (0.2)	0.3 (0.1)	1.7 (0.9)
	3-Methylbenzoate	89 (6)	129 (11)	7.0 (0.5)	8 (2)	5.4 (0.4)
	3-Chlorobenzoate	113 (8)	128 (11)	3.0 (0.1)	3 (1)	2.3 (0.5)
αTβTc	Benzoate	19 (4)	13 (2)	2.8 (0.1)	15 (3)	22 (2)
	2-Methylbenzoate	250 (65)	46 (10)	2.3 (0.13)b	0.9 (0.2)b	4.9 (0.9)
	3-Methylbenzoate	9.1 (1.3)	16 (2)	3.9 (0.2)	43 (4)	25 (2)
	4-Methylbenzoate	28 (6)	8.4 (1.5)	2.5 (0.1)	9 (2)	30 (4)
	2-Chlorobenzoate	144 (90)	92 (22)	1.9 (0.1)b	13 (0.8)b	2.1 (0.4)
	3-Chlorobenzoate	26 (3)	19 (6)	3.1 (0.1)	15 (1)	21 (1)
	4-Chlorobenzoate	41 (9)	12 (2)	3.0 (0.1)	8 (1)	25 (3)

^aExperiments were performed with 0.1 M sodium phosphate (pH 7.0 at 25°C) containing 430 μM NADH. Standard errors are given in parentheses. ht-REDs were paired with the α subunit of the ISP. The data sets used to calculate the parameters for different substrates contain 80 to 110 points. BADO displayed no detectable activity in the presence of either 4-methylbenzoate, 2-chlorobenzoate, or 4-chlorobenzoate.

^bValues were based on O₂ consumption to be consistent with the other values in the table. Values calculated based on benzoate consumption were at least one order of magnitude lower.

^cValues published in reference 20.

The nature of the substituted benzoate influenced the ability of BADO_ADP1 to utilize O₂ (Table 3). This was also observed for TADO_mt2 (20). However, in the case of BADO_ADP1, the correlates with the specificity of the enzyme for the substituted benzoate. Thus, the of BADO_ADP1 was lowest in the presence of benzoate (the enzyme's preferred substrate), was 2 times higher in the presence of 3-methylbenzoate and 3-chlorobenzoate, and was highest in the presence of 2-methylbenzoate (the worst substrate tested).

The apparent steady-state kinetic parameters of a given wild-type ISP were relatively unaffected when the dioxygenase was reconstituted with each of the two ht-REDs. Thus, the specific activities of α_Tβ_T and α_Bβ_B were both slightly higher in the presence of ht-RED_TADO (Table 4), and small differences in the apparent k_cat and K_m were observed. However, the apparent specificity of each ISP for 3-methylbenzoate and benzoate was unaffected by the identity of RED (results not shown). As noted in studies with TADO_mt2, the level of RED affects both the K_m and the k_cat (20). Subsequent experiments with hybrid ISPs were performed with the RED of the dioxygenase from which the α subunit originated. This choice was guided in part by the observation that the α subunit of toluene dioxygenase could be reduced in the absence of the β subunit with NADH and catalytic amounts of this enzyme's RED and ferredoxin (29).

TABLE 4.

Activities of ISPs with different reductases^a

ISP	RED	Substrate	Sp act (U/mg)b
αBβB	TADOmt2	Benzoate	14.5 (0.8)
		3-Methylbenzoate	10.2 (0.9)
	BADOADP1	Benzoate	8.2 (0.4)
		3-Methylbenzoate	7.6 (0.3)
αTβT	TADOmt2	Benzoate	3.0 (0.2)
		3-Methylbenzoate	3.2 (0.1)
	BADOADP1	Benzoate	1.7 (0.1)
		3-Methylbenzoate	2.8 (0.2)

^aExperiments were performed with 100 mM sodium phosphate (pH 7.0, 25°C) containing 430 μM NADH.

^bStandard errors are in parentheses.

The apparent substrate specificities of α_Tβ_B and α_Bβ_T for substituted benzoates were determined in air-saturated buffer (Table 5). The apparent substrate specificity of α_Bβ_T corresponded to that of BADO_ADP1, the parent from which the α subunit originated. In contrast, α_TβB differed slightly from TADO_mt2 in that it had greatest apparent specificity for 3-chlorobenzoate (3-chlorobenzoate > 3-methylbenzoate > benzoate ∼ 4-methylbenzoate > 4-chlorobenzoate > 2-methylbenzoate > 2-chlorobenzoate). In general, the K_mA of each hybrid ISP for a given benzoate was 2- to 10-fold higher than that of the corresponding parent ISP.

TABLE 5.

Apparent steady-state kinetic parameters of ISPs for selected substituted benzoates in air-saturated buffer

ISP	Substrate	KmA (μM)	kcat (s−1)	kcat/KmA (104 M−1 s−1)
αBβB	Benzoate	38 (1)	16.3 (0.3)	43 (1)
	2-Methylbenzoate	2,890 (240)	2.1 (0.2)	0.07 (0.01)
	3-Methylbenzoate	49 (3)	5.3 (0.2)	11 (2)
	3-Chlorobenzoate	94 (8)	2.2 (0.1)	2.3 (0.3)
αTβTc	Benzoate	15 (1)	2.4 (0.1)	16 (1)
	2-Methylbenzoate	590 (60)	1.6 (0.1)b	0.3 (0.02)b
	3-Methylbenzoate	5.3 (0.3)	2.9 (0.04)	46 (3)
	4-Methylbenzoate	43 (3)	2.25 (0.06)	5 (1)
	2-Chlorobenzoate	1,200 (240)	1.4 (0.1)b	0.1 (0.01)b
	3-Chlorobenzoate	22 (2)	1.03 (0.01)	4.6 (0.5)
	4-Chlorobenzoate	83 (7)	0.86 (0.03)	1 (0.1)
αTβB	Benzoate	81 (7)	1.3 (0.12)	1.6 (0.1)
	2-Methylbenzoate	410 (124)	0.3 (0.02)	0.09 (0.01)
	3-Methylbenzoate	38 (4)	1.4 (0.06)	3.6 (0.2)
	4-Methylbenzoate	73 (1)	1.2 (0.16)	1.7 (0.2)
	2-Chlorobenzoate	1,200 (22)	0.3 (0.02)	0.02 (0.005)
	3-Chlorobenzoate	20 (3)	1.3 (0.05)	7 (1)
	4-Chlorobenzoate	458 (96)	1.47 (0.09)	0.3 (0.06)
αBβT	Benzoate	110 (20)	4.0 (0.2)	3.6 (0.3)
	2-Methylbenzoate	5,130 (390)	2.7 (0.2)	0.05 (0.01)
	3-Methylbenzoate	178 (2)	2.4 (0.3)	1.3 (0.1)
	3-Chlorobenzoate	281 (38)	3.1 (0.1)	1.1 (0.1)

^aExperiments were performed with 0.1 M air-saturated sodium phosphate (pH 7.0 at 25°C) containing 430 μM NADH. Standard errors are given in parentheses. ht-REDs were paired with the α subunit of the ISP. In these assays, α_Bβ_B and α_Bβ_T displayed no detectable activity in the presence of p-toluate, 2-chlorobenzoate, or 4-chlorobenzoate.

^bValues were based on O₂ consumption to be consistent with the other values in the table. Values calculated based on benzoate consumption were at least one order of magnitude lower.

^cValues published in reference 20.

Coupling of O₂ consumption to benzoate transformation.

In BADO_ADP1, substrate and O₂ consumption were well coupled for all benzoates used in this study (Table 6). This is in contrast to TADO_mt2, in which the utilization of ortho-substituted benzoates was very poorly coupled to O₂ consumption (20). Benzoate turnover was as well coupled to O₂ utilization in the α_BβT hybrid, as in BADO_ADP1. Unexpectedly, the turnover of ortho-substituted benzoates was much better coupled to O₂ utilization in the α_Tβ_B hybrid than in TADO_mt2.

TABLE 6.

Coupling of substrate utilization in TADO_mt2, BADO_ADP1, and their hybrids^a

Enzyme	Substrate	Substrate/O2 ratio	H2O2/O2 ratio
αBβB	Benzoate	1.02 (0.08)	0.01 (0.002)
	2-Methylbenzoate	1.01 (0.10)	0.04 (0.01)
	3-Methylbenzoate	1.04 (0.12)	0.02 (0.002)
	3-Chlorobenzoate	1.04 (0.10)	0.002 (0.0008)
αTβTb	Benzoate	0.96 (0.10)	0.03 (0.006)
	2-Methylbenzoate	0.10 (0.10)	0.27 (0.02)
	3-Methylbenzoate	0.92 (0.13)	0.01 (0.003)
	4-Methylbenzoate	0.90 (0.10)	0.03 (0.005)
	2-Chlorobenzoate	0.10 (0.09)	0.2 (0.05)
	3-Chlorobenzoate	0.94 (0.10)	0.002 (0.0005)
	4-Chlorobenzoate	0.93 (0.10)	0.03 (0.01)
αTβB	Benzoate	0.98 (0.10)	0.02 (0.003)
	2-Methylbenzoate	0.92 (0.10)	0.04 (0.002)
	3-Methylbenzoate	0.98 (0.13)	0.007 (0.001)
	4-Methylbenzoate	0.94 (0.10)	0.01 (0.003)
	2-Chlorobenzoate	0.90 (0.09)	0.05 (0.01)
	3-Chlorobenzoate	0.96 (0.10)	0.03 (0.0006)
	4-Chlorobenzoate	0.97 (0.10)	0.04 (0.007)
αBβT	Benzoate	1.02 (0.10)	0.03 (0.003)
	2-Methylbenzoate	1.01 (0.10)	0.02 (0.02)
	3-Methylbenzoate	1.04 (0.13)	0.01 (0.002)
	3-Chlorobenzoate	1.01 (0.10)	0.02 (0.0003)

^aExperiments were performed with air-saturated 100 mM sodium phosphate (pH 7.0 at 25°C) containing 430 μM NADH, 4 μM RED, 1.8 μM ISP, and 215 μM substrate. Standard errors are given in parentheses. The reductase in the assay was paired to the α subunit of the ISP.

^bValues published in reference 20.

Identification of reaction products.

A single reaction product was detected for each substrate turned over by BADO_ADP1, as observed for the TADO_mt2-catalyzed transformations (20). For benzoate and 3-methylbenzoate, these products eluted at 2.152 and 2.168 min, respectively, and absorbed maximally at 261.5 and 262.7 nm, respectively. When treated with XylL, the resultant compounds coeluted with catechol and 3-methylcatechol, respectively, in HPLC analyses, and the amount of catechol detected corresponded to the amount of benzoate transformed within experimental error. In the case of 2-methylbenzoate, the cis-diol coeluted with the cis-diol produced from benzoate and had the same absorption as the latter, again as observed for TADO_mt2. These results indicate that both TADO_mt2 and BADO_ADP1 exclusively catalyzed the 1,2-dihydroxylation of the tested benzoates.

In general, the hybrid ISPs transformed benzoates to the same products as the parental enzymes. The α_Tβ_B hybrid proved to be exceptional, as it transformed 2-methylbenzoate to a product whose retention time on the HPLC column (2.036 min) and absorption spectrum (λ_max = 262.7 min) did not correspond to those of any of the other cis-diols observed in this study, including that produced from 2-methylbenzoate by TADO_mt2 and BADO_ADP1. Transformation of this unknown cis-diol by XylL and catechol 2,3-dioxygenase yielded a yellow product whose spectrum was identical to that of the meta-cleavage product of 3-methylcatechol. TADO_mt2 transforms 3-methylbenzoate to 1,2-dihydroxy-3-methyl-cyclohexa-3,5-diene-carboxylate, which is also transformed to 3-methylcatechol (20). However, the cis-diol produced by the α_Tβ_B-catalyzed transformation of 2-methylbenzoate was clearly different. It was therefore concluded that the latter was 1,6-dihydroxy-2-methyl-cyclohexa-2,4-diene-carboxylate.

DISCUSSION

TADO_mt2 and BADO_ADP1 are group II aromatic ring-hydroxylating dioxygenases whose ISP and RED components share 62 and 53% sequence identity, respectively (26). Their relatively high degree of similarity and the solubility of their respective substrates make these enzymes a useful model for studies of specificity determinants in ring-hydroxylating dioxygenases. Since BADOs from different bacterial isolates have been characterized to different extents in the literature and since no specificity data exist for any of these, it was first necessary to determine the reactivity of BADO_ADP1 with substituted benzoates.

The activity of purified, reconstituted BADO_ADP1 reported here is consistent with that of BADO_C1, whose substrate preference was investigated by using 1 mM concentrations of different benzoates (benzoate > 3-methylbenzoate > 3-chlorobenzoate > 2-methylbenzoate) (53). Thus, BADOs from both strains have a preference for methyl versus chloro substituents and for substituents in the following positions: meta > ortho > para. BADO_B13, present in 3-chlorobenzoate-grown Pseudomonas sp. strain B-13, has this same preference for substituent position (44). However, the B-13 enzyme prefers chloro substituents over methyl substituents. The substrate preference of 2-halobenzoate 1,2-dioxygenase of Pseudomonas cepacia 2CBS, which shares 56% sequence identity with BADO_ADP1, is different again, preferentially utilizing chloro versus methyl substituents and benzoates substituted in the following positions: ortho > meta > para (18). Significant for the present study, the apparent substrate specificity of BADO_ADP1 was narrower than that of TADO_mt2 (20) and differed in its preference of ortho versus para substituted benzoates.

The relative interchangeability of RED_TADO and RED_BADO in the two dioxygenases is not surprising given their sequence identity. Single-turnover studies have established that the reductase component is required for product release from the ISP but not for substrate hydroxylation (51, 52). Indeed, some ring-hydroxylating enzymes appear to share the same reductase in vivo. Thus, in Ralstonia sp. strain U2, the respective ISPs of salicylate 5-hydroxylase, which transforms salicylate to gentisate, and NDO share a ferredoxin and reductase (55). Similarly, plasmid pNL1 of Sphingomonas aromaticivorans F199 carries genes encoding seven different ISPs but only two copies of reductase-encoding genes (45), although it is possible that other reductases are encoded elsewhere in the genome.

The apparent specificity of each of the hybrid enzymes, α_Tβ_B and α_Bβ_T, corresponded most closely to that of the parent from which the α subunit originated. This indicates that this subunit harbors the major determinants of specificity in these group II dioxygenases. This finding is consistent with the results of subunit-swapping experiments with group III and IV enzymes, including NDO (40, 41) and BPDO (2, 31). Directed mutagenesis and gene-shuffling experiments with these systems have further localized the major determinants of substrate preference in the group III and IV enzymes to the C-terminal domain of the α subunit (3, 32, 36, 42, 43, 48). The present study extends this finding, as it investigates substrate specificity in purified enzymes.

The present study nevertheless clearly indicates that the β subunit contributes to reactivity of the dioxygenase. Thus, although α_Tβ_B transformed the same broad range of substituted benzoates as TADO_mt2, the apparent specificities of these ISPs were slightly different despite the ISPs sharing the same α subunit. Moreover, α_Tβ_B transformed 2-methylbenzoate with a different regioselectivity and improved coupling compared to α_Tβ_T. The contribution of the β subunit to dioxygenase reactivity has been suggested in at least two studies of hybrid BPDOs. For example, replacement of the β subunit of toluene dioxygenase with that of a BPDO yielded an enzyme with improved trichloroethylene-transforming activity (19, 33). Similarly, the ability of purified hybrid BPDOs to transform polychlorinated biphenyls was determined to some extent by the β subunit (10, 28). Interestingly, the present studies are consistent with an early insertional mutagenesis study of TADO that demonstrated that disruption of the β subunit affects the enzyme's substrate preference (27).

The observation that the β subunit contributes to the reactivity of ring-hydroxylating dioxygenases is consistent with crystallographic structures of NDO (9) and BPDO (12). These structures demonstrate that while the mononuclear Fe(II) site and the substrate-binding pocket of the ISP are contained entirely within the C-terminal domains of the α subunits of these enzymes, the β subunit probably contributes to the structural integrity and function of the active site. Thus, in BPDO, the interface between the α and β subunits is extensive, constituting a buried surface of 3,360 Ų per αβ dimer. Much of this interface involves a central sheet of the β subunit and an extended helix of the α subunit. This extended helix contains Asp386, one of the Fe(II) ligands. Similar packing interactions in NDO lock the structurally analogous helix of this dioxygenase in position. Moreover, the structure of the BPDO-product complex suggests that the extended helix shifts by up to 1.4 Å during the catalytic cycle of the enzyme (C. L. Colbert and J. T. Bolin, personal communication). Thus, the structural data indicate that the close contacts between the α and β subunits in the vicinity of the active site could influence substrate specificity, the coupling of substrate utilization during the dynamic catalytic process, and the stability of the ISP.

The present study demonstrates the value of using highly active, purified enzyme preparations in investigating the structural determinants of substrate specificity. More specifically, the results demonstrate that it will be important not to overlook the role of the β subunit in generating ring-hydroxylating dioxygenases with useful activities, both to fine-tune these activities and to optimize the stability of variant enzymes (3, 42, 43, 48). Indeed, it is possible that subunit exchange occurred during the natural evolution of ring-hydroxylating dioxygenases to produce enzymes with novel, useful activities. Current efforts are being focused on understanding the structural basis of the specificities of different BADOs for chlorinated versus alkylated benzoates.