Purification and Preliminary Characterization of a Serine Hydrolase Involved in the Microbial Degradation of Polychlorinated Biphenyls*

2-Hydroxy-6-oxo-6-phenylhexa-2,4-dienoate (6-phenyl-HODA) hydrolase (BphD), an enzyme of the biphenyl biodegradation pathway encoded by the bphD gene ofBurkholderia cepacia LB400, was hyperexpressed and purified to apparent homogeneity. SDS-polyacrylamide gel electrophoresis confirmed that BphD has a subunit molecular mass of 32 kDa, while gel filtration demonstrated that it is a homotetramer of molecular weight 122,000. The enzyme hydrolyzed 6-phenyl-HODA with ak cat of 5.0 (± 0.07) s−1 and ak cat/K m of 2.0 (± 0.08) × 107 m −1 s−1 (100 mm phosphate, pH 7.5, 25 °C). The specificity of BphD for other 2-hydroxy-6-oxohexa-2,4-dienoates (HODAs) decreased markedly with the size of the C6 substituent; 6-methyl-HODA, themeta cleavage product of 3-methylcatechol, was hydrolyzed approximately 2300 times less specifically than 6-phenyl-HODA. By comparison, the homologous hydrolase from the toluene degradation pathway, TodF, showed highest specificity for 6-methyl- and 6-ethyl-HODA (k cat/K m of 2.0 (± 0.05) × 106 m −1s−1 and 9.0 (± 0.5) × 106 m −1 s−1, respectively). TodF showed no detectable activity toward 6-phenyl-HODA and 6-tert-butyl-HODA. Neither BphD nor TodF hydrolyzed 5-methyl-HODA efficiently. The k cat of BphD determined by monitoring product formation was about half that determined by monitoring substrate disappearance, suggesting that some uncoupling of substrate utilization and product formation occurs during the enzyme catalyzed reaction. Crystals of BphD were obtained using ammonium sulfate combined with polyethylene glycol 400 as the precipitant. Diffraction was observed to a resolution of at least 1.9 Å, and the evaluation of self-rotation functions confirmed 222 (D2) molecular symmetry.

Microbial pathways responsible for the degradation of aromatic compounds have been intensively investigated due to their importance in maintaining the global carbon cycle as well as their potential usefulness in bioremediation (1,2). Characterizing the enzymes of these pathways is essential to our understanding of the biotransformation process, providing a basis for improving existing pathways and for engineering novel degradative capabilities for otherwise recalcitrant xenobiotics. This is exemplified by the bph pathway, which is responsible for the degradation of biphenyl via a catecholic intermediate ( Fig. 1) and is thus typical of many aerobic pathways that catabolize aromatic compounds (1). Engineering the specificities of the pertinent enzymes of the bph pathway will facilitate the development of strains able to more effectively degrade polychlorinated biphenyls (PCBs) 1 (3).
BphD (EC 3.7.1.8), a hydrolase encoded by bphD and the fourth enzyme of the bph pathway, catalyzes the hydrolysis of 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoates (6-phenyl-HODA) at a carbon-carbon bond yielding 2-hydroxypenta-2,4-dienoate and benzoate (4); certain chlorinated 2-hydroxy-6-oxohexa-2,4dienoates (HODAs) are similarly cleaved. At least two studies have suggested that the substrate specificity of BphD is a key determinant of the overall selectivity of the pathway with respect to the degradation of various aromatic compounds and their halogenated derivatives (5,6). For example, incubation of biphenyl-degrading strains with certain PCB congeners resulted in the accumulation of the corresponding chlorinated 6-phenyl-HODAs (5), indicating that the specificity of BphD in the tested strains prevented further degradation of the PCB. In another study, hybrid pathways were constructed using the bph and tod pathways (6). The latter is responsible for the degradation of toluene in Pseudomonas putida F1 and is very similar to the bph pathway (7). In these studies, it was demonstrated that the degradation of toluene by the bph pathway was blocked by BphD while the degradation of biphenyl by the tod pathway was blocked by TodF (EC 3.7.1.9), the homologous hydrolase from this pathway. It was noted that the sequences of TodF and BphD are less similar (35% identity) than the sequences of the preceding enzymes in the pathways (53 to 65% identity) (6).
In addition to bphD, which has been cloned and sequenced from a number of sources (8 -14), and todF (15), the genes encoding homologous enzymes involved in the catabolism of a variety of other aromatic compounds, including 3-(3-hydroxyphenyl) propionate (16) and xylenes (17) have been cloned. These enzymes (EC 3.7.1.-) hydrolyze vinylogous 1,5-diketones formed by the dioxygenative meta-ring cleavage of arenes and constitute a class of ␤-ketolases of potential interest in organic synthesis (18). The encoded enzymes all share conserved serine, histidine, and aspartate residues, reminiscent of the cata-lytic triad present in a number of other hydrolytic enzymes. Mutagenesis studies on XylF, the hydrolase of the TOL pathway, have confirmed that these residues are critical for catalysis (19). Sequence comparisons indicate that these enzymes belong to the family of proteins that have the ␣/␤ hydrolase fold (19,20).
Although the importance of BphD has been recognized, the biochemical characterization of the enzyme is not extensive. Indeed, only hydrolases that catalyze the fragmentation of the ring fission products of monocyclic substrates such as 3-methylcatechol (21,22) and 2,3-dihydroxyphenylpropionate (23) have thus far been purified to homogeneity and characterized with respect to their substrate specificities. The hydrolase responsible for the latter reaction, MhpC, has also been subject to steady-state and stopped-flow kinetic studies, which have provided insight into the mechanism of carbon-carbon bond cleavage catalyzed by this class of enzymes (23,24). Hydrolases involved in biphenyl degradation have been purified from Pseudomonas cruciviae and Rhodococcus sp. RHA1 (6,25), whereas hydrolases involved in dibenzofuran degradation have been isolated from Sphingomonas sp. strain RW1 (26). However, quantitative data on the substrate specificity of BphD are unavailable, limiting the inference of structure:function relationships among the different hydrolases.
We report here the overexpression, purification, and characterization of BphD from Burkholderia cepacia LB400, a soil isolate that degrades biphenyl and transforms a wide range of polychlorinated biphenyls (27,28). The specificity of the enzyme for a range of HODAs was established by steady-state kinetics and was compared with that of TodF. The extinction coefficients of these HODAs are reported. The observed specificities of the hydrolases, and differences therein, are discussed in terms of structural and mechanistic factors. Conditions for the crystallization of BphD were established, and an analysis of preliminary x-ray diffraction data from these crystals is reported.

MATERIALS AND METHODS
Chemicals-Catechol, 3-methylcatechol and 4-methylcatechol were purchased from Aldrich. 2,3-Dihydroxybiphenyl, 3-ethylcatechol, 3-isopropylcatechol, and 3-tert-butylcatechol were prepared in one of our laboratories (V. S.); their synthesis will be described elsewhere. Restriction enzymes and PfuI polymerase were from Amersham Pharmacia Biotech and Stratagene, respectively. PEG 400 and ammonium sulfate used in crystallization were from Fluka. All other chemicals were of analytical grade.
DNA Manipulation and Amplification-DNA was purified, digested and ligated using standard protocols (37). The bphD gene was amplified by the polymerase chain reaction (PCR) using two synthetic oligonucleotides with the respective sequences 5Ј-CCCGCATATGACCGCACT-CACCGAA-3Ј and 5Ј-CCGGAAGCTTGCCCCTTACGCGTGC-3Ј as primers. These primers introduced NdeI and HindIII restriction sites at the 5Ј and 3Ј ends, respectively, of the amplified product, facilitating subsequent cloning of the gene. The PCR reaction contained 0.2 mg of the template plasmid DNA, 0.75 units of Pfu DNA polymerase (Stratagene), 20 nM amounts of each dNTP, and 100 pmol of each primer in a final volume of 100 l. Twenty temperature cycles were performed as follows: 95°C for 1 min, 48°C for 1 min, and 72°C for 2 min. The PCR product was purified using the QIAquick PCR purification kit (Qiagen).
Purification of BphD-Buffers containing 20 mM sodium HEPES, pH 7.5, were used throughout the purification procedure. Chromatography was performed on a Ä KTA Explorer (Amersham Pharmacia Biotech, Baie d'Urfe, Quebec, Canada). The cell pellet (15 g) was resuspended in buffer and disrupted by a single passage through a French Press (Aminco) at an operating pressure of 20,000 p.s.i. The cell debris was removed by ultracentrifugation at 50,000 rpm for 60 min in a T1250 rotor (DuPont Instruments). The clear supernatant fluid was filtered through a 0.45-mm filter and divided into two portions, each of which was loaded onto a Source TM 15Q (Amersham Pharmacia Biotech) anion exchange column (2 ϫ 9 cm), which had been preequilibrated with buffer containing 0.05 M NaCl. Bound proteins were eluted at a flow rate of 20 ml/min with a linear gradient of NaCl from 0.05 to 0.2 M over 15 column volumes. Fractions containing the enzyme activity, which eluted at 0.12 M NaCl, were concentrated to 2 ml by ultrafiltration using a stirred cell equipped with a YM10 filter (Amicon). The preparation was loaded onto a HiLoad 26/60 Superdex 200 gel filtration column (Amersham Pharmacia Biotech) that had been preequilibrated with buffer containing 0.15 M NaCl. The column was eluted with the same buffer operating at a flow rate of 3 ml/min. Fractions containing activity were pooled, concentrated to 5 ml and mixed with an equal volume of buffer containing 20% saturated ammonium sulfate. This solution was then loaded onto a Phenyl Sepharose (Amersham Pharmacia Biotech) column (1 ϫ 9 cm), preequilibrated with buffer containing 10% ammonium sulfate. The enzyme eluted at about 9% saturated ammonium sulfate in a decreasing ammonium sulfate concentration gradient (10% to 0% over 10 column volumes) at a flow rate of 1 ml/min. Fractions containing the enzyme were pooled, concentrated by ultrafiltration and washed repeatedly with fresh buffer to remove ammonium sulfate from the solution. The concentrated enzyme was stored frozen in aliquots at Ϫ80°C.
Purification of Ring-cleavage Dioxygenases-2,3-dihydroxybiphenyl dioxygenase (BphC) of B. cepacia LB400 was purified anaerobically as described previously. 2 Catechol 2,3-dioxygenase (C23O) was purified anaerobically from E. coli JM101 containing the plasmid pAW31 (39) using essentially the same technique 2 with the following modifications. The cell disruption buffer contained 10 mM Tris-HCl buffer, pH 7.5, 1 mM magnesium chloride, 1 mM calcium chloride, 10% isopropanol, and 0.1 mg/ml DNase. Buffer A, used for the two chromatographic steps, contained 10 mM Tris-HCl, pH 7.5, 10% isopropanol, 2 mM dithiothreitol, and 0.2 mM ferrous ammonium sulfate. C23O was eluted from the Mono Q column (Amersham Pharmacia Biotech) using a linear gradient of 0.18 -0.32 M NaCl in buffer A. C23O was eluted from the gel filtration column using buffer A containing 0.15 M NaCl. C23O-containing fractions were concentrated to 4 ml and exchanged into 10 mM Tris-HCl, pH 7.5, 10% butanol using a Bio-Gel P-6 OG (Bio-Rad) column. Fractions containing activity were frozen in liquid nitrogen and stored at Ϫ80°C.

Determination of Protein Purity and Concentration-SDS-PAGE was performed on a Bio-Rad MiniPROTEAN II apparatus and stained with
Coomassie Blue according to established procedures (40). Protein concentrations were determined by the bicinchoninic acid protein assay reagent kit (Pierce) using bovine serum albumin as standard.
Preparation of Substrates and Determination of Their Extinction Coefficients-Stock solutions of HODAs in 100 mM ionic strength of potassium phosphate, pH 7.5, were prepared immediately prior to use. 6-Phenyl-HODA was prepared by dissolving 2,3-dihydroxybiphenyl in a small volume of 95% ethanol and diluting to the desired volume with buffer (ethanol constituted less than 0.2% of the final solution). To this solution was added 22 units of BphC. 6-Ethyl-HODA, 6-isopropyl-HODA, and 6-tert-butyl-HODA were prepared in a similar manner except that 3-ethylcatechol, 3-isopropylcatechol, and 3-tert-butylcatechol, respectively, were dissolved directly in buffer. Similarly, HODA, 6-methyl-HODA, and 5-methyl-HODA were prepared from catechol, 3-methylcatechol, and 4-methylcatechol, respectively, using C23O in place of BphC.
Extinction coefficients of the yellow-colored HODAs were determined by adding the ring-cleavage enzyme to solutions containing weighed amounts of catechol and determining the resultant absorbance spectrophotometrically. Stoichiometric cleavage of the catechol was verified by following the consumption of dioxygen using a Clark-type polarographic oxygen electrode (Yellow Springs Instruments model 5301). In the preparation of stock solutions of HODAs for the hydrolase assays, complete conversion was verified by ensuring: (i) that there was no further increase in absorbance at the wavelength of maximum absorbance when fresh ring-cleavage enzyme was added to a diluted sample; and (ii) that the absorbance expected from the amount of catechol present and the determined extinction coefficient was obtained. HODAs used for enzyme assays were freshly prepared, stored on ice, and used within 8 h.
Kinetic Measurements-Enzymatic activity was measured by following the consumption of the yellow substrate on a Varian Cary 3 spectrophotometer equipped with a thermojacketed cuvette holder. The spectrophotometer was interfaced to a microcomputer and controlled by Cary OS/2 multi-tasking software. The amount of enzyme used in each assay was adjusted so that the progress curve was linear for at least 2 min. Initial velocities were determined from a least squares analysis of the progress curves using the kinetics module of the Cary software.
The standard activity assay was performed in a total volume of 1.0 ml of 100 mM ionic strength potassium phosphate, pH 7.5, 25.0 Ϯ 0.1°C containing 10 M 6-phenyl-HODA. The reaction was initiated by adding between 5 and 10 l of an appropriately diluted enzyme preparation to the reaction cuvette. A reaction mixture prepared without the hydrolytic enzyme served as a reference. The reaction was monitored at 434 nm. One unit of enzymatic activity is defined as the quantity of enzyme required to consume 1 mol of 6-phenyl-HODA/min.
Specificity experiments were carried out in a total volume of 1.0 ml of 100 mM ionic strength potassium phosphate, pH 7.5, 25.0 Ϯ 0.1°C. Reactions were monitored at the wavelengths indicated in Table II. Initial velocities determined at different substrate concentrations were fitted to the Michaelis-Menten equation using the least squares and dynamic weighting options of LEONORA (41).
For the substrate 6-phenyl-HODA, enzyme activity was also monitored by the formation of the product spectrophotometrically at 225 nm. The absorbance of the product solution could be largely attributed to the formation of benzoate, which has a maximum absorbance at this wavelength, with some contribution from the other product, 2-hydroxypent-2,4-dienoate, which absorbs maximally at 270 nm. As 6-phenyl-HODA also absorbs at 225 nm, a differential extinction coefficient was used to calculate the activity of the enzyme. This coefficient was obtained from the difference in the absorbance of a sample before and after the addition of sufficient BphD to ensure rapid and complete hydrolysis.
Crystallization of BphD-An initial screen for possible crystallization conditions was performed at 18°C using the hanging drop vapor diffusion method (42). A variety of experiments were performed using the Crystal Screen Kits I and II (Hampton Research) essentially according to the manufacturer. A more precise search involved simultaneous variation of the concentration of ammonium sulfate (from 1.6 M to 2.2 M at 0.2 M increments), and the pH of buffer (from 6.0 to 9.0 at 0.5 increments), with or without the presence of PEG 400. Crystals used for diffraction analysis were obtained at 20°C by the use of 4-l drops containing BphD (9.6 mg/ml) mixed in a 1:1 ratio with reservoir solution and suspended over a 1-ml well solution containing 0.1 M sodium Hepes buffer, pH 8.0, 2.0 M ammonium sulfate, and 3% v/v PEG 400.
X-ray Data Collection and Analysis-Because attempts to transfer crystals to cryo-stabilization solutions failed, the crystals were mounted by standard procedures in quartz or glass capillaries. X-ray diffraction patterns were obtained by the use of Cu-K␣ radiation from a Rigaku rotating anode x-ray generator operated at 50 kV and 85-100 mA and equipped with focusing mirror optics (Molecular Structures Corp.). Diffraction patterns were recorded at ϳ20°C by the oscillation method and the use of R-axis IIC or R-axis IV imaging plate detector systems (Molecular Structures Corp.). The typical oscillation range and exposure time per frame were 1°and 10 min; the typical crystal to detector distance was 15 cm. The patterns were analyzed, and the intensity measurements were scaled, post-refined, and merged by the use of the HKL software package (43). Rotation functions were calculated with the GRLF program (44).

Construction of the Expression
Vector-Amplification of the bphD gene from PstI-digested plasmid pDD5301 yielded a DNA fragment of about 860 base pairs. This fragment was digested with HindIII and NdeI, extracted from agarose gel after electrophoresis, purified using QIAEX II gel extraction kit (Qiagen), and then ligated into appropriately digested pT7-7 plasmid to produce the plasmid pSS74. The entire sequence of the cloned gene was determined to ensure that no PCR-induced errors had been introduced. The bphD gene, together with an upstream ribosomal binding site of T7 phage gene 10 protein of the T7-7 vector, was isolated from pSS74 as an XbaI/HindIII fragment and inserted into two plasmids: pEMBL18 (35); and pVLT31 (34), a broad host range expression vector. These constructs were designated pSS184 and pSS314, respectively.
Expression, Purification, and Evaluation of Molecular Mass-Among the systems tested for the expression of BphD, the highest level of recombinant hydrolase was obtained from E. coli DH5␣ and P. putida KT2442 containing the plasmid pSS314. In both cases, BphD constitutes about 20% of the total cellular protein. However, in the pseudomonad, the addition of isopropyl-1-thio-␤-D-galacatopyranoside to a final concentration of 0.5 mM is required to produce the level of expression achieved in E. coli DH5␣ without the addition of of isopropyl-1-thio-␤-D-galacatopyranoside. BphD was therefore purified from E. coli DH5␣ containing pSS314. A total of 28 mg of BphD was purified from 15 g of cells (wet weight) using anion-exchange, hydrophobic interaction and gel filtration chromatographies as summarized in Table I. To produce the cells, 8 liters of medium was inoculated with 12 ml of an overnight culture and incubated for 24 h before harvesting. The final preparation of BphD was judged to be greater than 99% pure based on a Coomassie Blue-stained denaturing gel ( Fig. 2A). A solution of 0.1 mg/ml enzyme was stable at room temperature for at least 24 h, with no detectable loss of activity. The TodF enzyme was expressed in P. putida KT2442 using a construct analogous to pSS314, with bphD replaced by todF. Enzyme of purity greater than 99%, as judged by SDS-PAGE, was obtained using the purification procedure developed for the BphD enzyme (see Fig. 2B).
The subunit molecular mass of BphD was estimated to be 32.0 kDa by its behavior during SDS-PAGE in comparison with a calibration curve developed from standard molecular weight markers (Fig. 3A). The relative molecular weight of the native protein was 122,000, as estimated by calibrated gel filtration chomatography (Fig. 3B).
Steady-state Kinetic Analysis-Prior to kinetic analysis, the wavelength of maximal absorbance and the extinction coefficient at this wavelength were determined for each HODA under assay conditions; the results are provided in Table II. The spectra of HODA were observed to depend not only on pH but also on the nature of the buffer (results not shown).
BphD was found to obey classic Michaelis-Menten kinetics for all of the substrates tested here. A plot of initial velocity of substrate consumption versus 6-phenyl-HODA concentration is shown in Fig. 4.
Although BphD hydrolyzed all of the HODAs tested, the enzyme was most specific for 6-phenyl-HODA (k cat /K m of 2.0 (Ϯ 0.08) ϫ 10 7 M Ϫ1 s Ϫ1 , 100 mM phosphate, pH 7.5, 25°C) and its specificity decreased markedly with the size of the C6 substituent (Table III). Indeed, the k cat /K m for 6-methyl-HODA was smaller than the value for 6-phenyl-HODA by a factor of 2300. The dependence on substituent size is especially pronounced in the comparison of the 6-methyl and 6-ethyl substrates, where k cat /K m increased by a factor of 186 with the addition of a single methylene group. In contrast to the consistent trend in the dependence of k cat /K m on substituent size, the k cat of BphD was relatively independent of the C6 substituent for HODAs possessing a substituent larger than a methyl group at this position. It is also noteworthy that 5-methyl-HODA, the meta cleavage product of 4-methylcatechol, was hydrolyzed very inefficiently by BphD.
At saturating concentrations of 6-phenyl-HODA (10 M), the specific activity obtained for BphD depended on the assay strategy; the value determined from monitoring the disappearance of substrate at 434 nm was 9.5 units/mg, whereas that obtained by monitoring product formation at 225 nm was 4.4 units/mg. Furthermore, the k cat and K m determined by monitoring product formation are 2.4 Ϯ 0.05 s Ϫ1 and 0.39 Ϯ 0.05 M, respectively. It should be noted, however, that the low extinc-tion coefficients of the products limit the minimum concentration of 6-phenyl-HODA at which steady-state reactions can be performed to approximately the K m value.
In contrast to BphD, TodF showed highest specificity for HODAs with small substituents at C6 (Table IV). Interestingly, although TodF hydrolyzed 6-methyl-HODA and 6-ethyl-HODA with k cat values of 35 s Ϫ1 and 14 s Ϫ1 , respectively, the enzyme was determined to have a much lower K m for 6-ethyl-HODA and to be 4.5 times more specific for this substrate (k cat /K m of 9.0 ϫ 10 6 M Ϫ1 s Ϫ1 and 2.0 ϫ 10 6 M Ϫ1 s Ϫ1 , respectively). TodF also appears to have a narrower specificity than BphD, and did not detectably hydrolyze 6-tert-butyl-HODA or 6-phenyl-HODA. As was observed in BphD, TodF hydrolyzed 5-methyl-HODA very inefficiently.  A subsequent evaluation of the influence of pH and ammonium sulfate concentrations showed that a minimum concentration of 2.0 M ammonium sulfate was required for the formation of crystals. At pH 7.5 and above, small, very thin, rhombus-shaped crystals were produced. The addition of PEG 400 affected the crystal habit. Concentrations of PEG 400 below 2% yielded fewer but larger rhombus-shaped crystals that occurred as stacked plates. Increasing the PEG 400 concentration above 2% yielded individual crystals of adequate size for diffraction studies. Two growth habits were common: square and triangular (isoceles) prisms (see Fig. 5).
Diffraction Studies-Diffraction patterns obtained from sev-eral crystals of both prismatic habits were examined; no significant difference correlated with the habit was detected. Reflections were observed to at least 1.9-Å resolution, but none of the crystals produced a pattern consistent with a single lattice regardless of the absence or presence of visible flaws or multiple crystals. However, all diffraction patterns could be explained by nonmerohedrally twinned crystals. By selection of reflections from the dominant lattice during indexing and initial refinement of crystal and instrument parameters, it was possible to quantitatively evaluate the intensities. Statistics pertaining to data acquired from one crystal (0.4 ϫ 0.4 ϫ 0.08 mm) by rotation over a continuous range of 140°are reported in Table V. The diffraction patterns for this crystal and several others were best indexed by a triclinic lattice, implying space group P1. The cell parameters for this crystal were a ϭ 74.5 Å, b ϭ 86.4 Å, c ϭ 91.5 Å, ␣ ϭ 89.8°, ␤ ϭ 86.7°, and ␥ ϭ 87.6°; similar values were obtained for several crystals. The probable content of the unit cell is eight monomers, which corresponds to a volume-to-mass ratio, V M , of 2.3 Å 3 /Da.
Plots of self-rotation functions calculated in polar coordinates (, , ) with the rotation angle () set to 180°, using various subsets of the structure factor amplitudes, were consistent with the presence in the unit cell of two tetramers with 222 (D2) point group symmetry. The plots display three somewhat elongated maxima mutually separated by 90°; an exemplary plot is provided in Fig. 6. Each of the maxima is consistent a pair of unresolved peaks, where one peak in each pair is associated with one of the tetramers. Thus, the two-fold axes of the two tetramers have similar orientations with respect to the cell axes. Plots of rotation functions calculated with ϭ 72°, 90°, or 120°showed only diffuse features.

DISCUSSION
The current work describes an efficient expression system for BphD of B. cepacia LB400, a serine hydrolase involved in the microbial catabolism of biphenyl and some PCB congeners. Rapid purification yielded 28 mg of a highly homogeneous preparation of the protein from 8 liter of cell culture. The relative molecular weight of the subunit as determined by SDS-PAGE corresponds well with the molecular mass predicted from the DNA sequence (286 amino acid residues, 32.0 kDa; Ref. 10). The molecular weight of the native protein, as determined by gel filtration, is 122,000, suggesting that the enzyme is a tetramer of four identical subunits. This conclusion is consistent with the diffraction data acquired from crystals of   BphD, which indicate the existence of tetramers of 222 symmetry. Molecular mass determinations of HODA degrading hydrolases from P. putida NCIB 10015 and 9865 have also yielded results consistent with tetrameric structures (6). The BphD enzyme from Rhodococcus sp. RHA1 was found to be octameric (43), whereas it was concluded that MhpC (23) and XylF (22) are dimeric. Two hydrolases purified from Sphingomonas sp. strain RW1 were found to be monomeric (26). As with other catabolic enzymes, it appears that there is considerable variation in the oligomeric structures of enzymes in this class, but the variations are not correlated with substrate specificity.
Consistent with the occurrence of BphD in a pathway re-sponsible for the catabolism of biphenyl, this enzyme hydrolyzed 6-phenyl-HODA more efficiently than a variety of other substituted HODAs. Although the specificity of BphD decreased with the size of the 6-substitutent, the enzyme exhibited a low activity toward 6-methyl-HODA, the meta-cleavage product of 3-methylcatechol. This is in apparent contrast to the BphD of another biphenyl degrader, P. pseudoalcaligenes KF707, which shares 97.2% sequence identity to the BphD of B. cepacia LB400. Although it has been reported that the former is inactive toward 6-methyl-HODA, those studies were based on whole cell assays, which probably contained lower amounts of BphD (5). Interestingly, the large decrease in specificity of BphD for 6-ethyl-versus 6-methyl-HODA is correlated with the specificity of BphC, the preceeding enzyme of the pathway. When BphC's specificity was examined with a series of 3-alkylcatechols, the precursors of the 6-alkyl-HODAs used in this study, 3-ethylcatechol was strongly preferred over 3-methylcatechol. 2 The specificity of TodF for HODAs possessing small substituents at the 6-position is also consistent with the context in which this enzyme evolved, as this hydrolase is part of a pathway responsible for the catabolism of toluene (7). In this context, it is not immediately apparent why TodF has evolved to be more specific for 6-ethyl-HODA than 6-methyl-HODA. XylF, the hydrolase of the TOL pathway of P. putida mt-2, which is responsible for the degradation of toluene and m-and p-xylene, is also specific for HODAs bearing small alkyl substituents at the 6-position (22). Under conditions similar to those used in the present study (0.1 M sodium phosphate, pH 7.5), XylF has a lower K m for 6-ethyl-HODA than 6-methyl-HODA, although the difference is only a factor of 2, substantially less than the factor of 10 observed for TodF. While catalytic constants were not determined in that study, the reported V max values indicate that XylF is slightly more specific for 6-methyl-HODA than 6-ethyl- ϭ 0°), respectively. The rotation function was calculated for directions separated by 3°in and over the range of 0 -180°for both variables. Diffraction data in the resolution range of 10.0 to 3.0 Å were used and the radius of integration was 20 Å. The lowest contour corresponds to two standard deviations from the mean of the calculated values, and the separation between contour levels is 0.5 standard deviations function. HODA. As TodF, XylF was unreactive toward 6-phenyl-HODA. The high specificity constant of BphD for 6-phenyl-HODA on the one hand, and TodF for 6-ethyl-HODA on the other, resides mainly in the low relative K m of these enzymes for their respective preferred substrates. It is likely that these enzymes have evolved substrate binding sites that are tailored to fit the substituent at the C6 of the HODA substrate. In BphD, such a site would be quite large to optimally bind the 6-phenyl substituent. Such a site could also accommodate smaller substrates, albeit with a lower affinity. The corresponding site in TodF is expected to be much smaller, so as to optimally bind an ethyl substituent. Such a site would not necessarily accommodate larger substituents, explaining TodF's lack of activity toward 6-tert-butyl and 6-phenyl-HODA.
BphD and TodF are both more than 60-fold less specific for 5-methyl-HODA than 6-methyl-HODA. 5-Methyl-HODA is an aldehyde that results from the meta-cleavage of 4-methylcatechol. The preference of BphD and TodF for ketones versus aldehydes, which has also been observed in two 6-methyl-HODA hydrolases (21) and XylF (22), is consistent with the proposal that aldehydes are degraded by a dehydrogenation/ decarboxylation branch of the pathway. This branch, which includes a dehydrogenase, a tautomerase, and a decarboxylase, results in the net production of one equivalent of NADH for each substrate equivalent, such as HODA and 5-methyl-HODA, transformed (45).
Given the energetic advantage of degrading aldehydes by dehydrogenation and decarboxylation, it is not surprising that HODA hydrolases have not evolved to bind 5-methyl substrates. However, mechanistic considerations suggest a further reason why 5-methyl-HODA would be a poorer substrate for these hydrolases. Recent evidence from deuterium isotope exchange and rapid kinetic experiments on MhpC has led to the proposal that HODA hydrolases catalyze an enol-keto tautomerization of the substrate prior to hydrolysis (23,24). The dienol would thus function as an electron sink, facilitating subsequent nucleophilic attack on the electrophilic C6. In this scheme, a methyl substituent at C5 would be expected to stabilize the C4-C5 double bond, and thus stabilize the enoltautomer with respect to the keto-form.
The observed difference in the kinetics of 6-phenyl-HODA consumption and product formation suggests that uncoupling of substrate utilization and product formation, as reported for MhpC (23,24) also occurs for BphD. This uncoupling, or "leakiness," has been explained by the dissociation of the keto intermediate from the enzyme prior to hydrolysis (24). In both BphD and MhpC, the k cat determined by following product formation was about half that determined by following substrate consumption. It is thus likely that the mechanism of BphD is essentially the same as that proposed for MhpC and that leakiness is a general phenomenon of HODA hydrolases. Improvement of the crystals of BphD should enable the elucidation of its three-dimensional 3D structure in the near future. This will provide a basis for identifying the structural determinants of substrate specificity in BphD and TodF. It will also greatly facilitate the assessment of the catalytic roles of residues involved in the proposed mechanism of carbon-carbon bond cleavage. A preliminary report of a structure for an octameric BphD from a Gram-positive bacterium, Rhodococcus RHA1, has recently been published (46). However, this enzyme shares only 32% sequence identity with BphD from B. cepacia LB400 and TodF, limiting the usefulness of the rhodococcal hydrolase in addressing detailed questions of structure:function relationships in these enzymes. This degree of identity is similar to that between these latter enzymes and a haloperoxidase for which a structure has been determined (47). The structures of the three enzymes should facilitate the evaluation of evolutionary relationships among these hydrolases.