Molecular Basis for the Stabilization and Inhibition of 2,3-Dihydroxybiphenyl 1,2-Dioxygenase by t -Butanol*

The steady-state cleavage of catechols by 2,3-dihy-droxybiphenyl 1,2-dioxygenase (DHBD), the extradiol dioxygenase of the biphenyl biodegradation pathway, was investigated using a highly active, anaerobically purified preparation of enzyme. The kinetic data obtained using 2,3-dihydroxybiphenyl (DHB) fit a compulsory order ternary complex mechanism in which substrate inhibition occurs. The K m for dioxygen was 1280 6 70 m M , which is at least 2 orders of magnitude higher than that reported for catechol 2,3-dioxygenases. K m and K d for DHB were 22 6 2 and 8 6 1 m M , respectively.

The microbial degradation of aromatic compounds constitutes an essential link in the global carbon cycle. The aerobic degradation of aromatic compounds such as toluene, naphthalene, and biphenyl generally proceeds via a catecholic catabolite with hydroxyl substituents on two adjacent carbon atoms. This catecholic compound is cleaved by a dioxygenase from one of two very different classes. Intradiol dioxygenases utilize non-heme ferric iron to cleave the aromatic nucleus ortho to (between) the hydroxyl substituents whereas extradiol dioxygenases utilize non-heme ferrous iron to cleave the aromatic nucleus meta (adjacent) to the hydroxyl substituents. The mechanism of intradiol dioxygenases is better understood due to their greater stability, favorable properties for spectroscopic examination, and the accessibility of catalytic intermediates (1,2). Interest in extradiol dioxygenases is nonetheless considerable, not only because of their general metabolic significance and catalytic properties, but also because of the potential exploitation of these enzymes in the degradation of environmental pollutants such as polychlorinated biphenyls.
2,3-Dihydroxybiphenyl 1,2-dioxygenase (DHBD) 1 is a component of the aerobic biphenyl degradation pathway of a number of microorganisms and cleaves 2,3-dihydroxybiphenyl (DHB) in an extradiol fashion as shown in Scheme 1. Crystallographic studies of DHBD from Burkholderia cepacia LB400 (3) and Pseudomonas sp. strain KKS102 (4) have provided details of the active site Fe(II) environment and conserved active site residues, and have contributed to our understanding of the evolution of these enzymes (5,6). Phylogenetic analyses indicate the existence of two evolutionarily independent types of extradiol dioxygenases (5). However, the catalytic strategy utilized by these two types of enzymes appears to be very similar and the emerging mechanism has been proposed based on studies of both types of enzymes. Spectroscopic and biochemical studies (7)(8)(9)(10)(11) suggest a mechanism in which the catechol first binds to the active site Fe(II) in a bidentate manner as a monoanion. Subsequent O 2 binding to the Fe(II) with charge transfer from the bound catecholic monoanion results in the generation of an iron-bound superoxide-like moiety and an iron-bound, activated catechol. It has been proposed that reaction of these two species forms a bridged peroxy intermediate, which undergoes Criegee rearrangement to give an unsaturated lactone intermediate. This lactone is finally hydrolyzed to yield the reaction product. Several steps in this mechanism have yet to be substantiated, and the catalytic roles of conserved active site residues remain to be fully elucidated.
With the development of structural data, DHBD has become an attractive system for experiments that will further our understanding of extradiol dioxygenase function. Such studies depend upon the availability of homogeneous, active preparations of the enzyme. The stability of a number of extradiol dioxygenase preparations has been improved through the inclusion of organic additives in solutions of the enzymes. While this strategy has been in use for at least 30 years (12), the mechanism of stabilization has yet to be reported. Even with this precaution, the best reported preparations of DHBD contain at most 50% of their complement of active site Fe(II) (13).
We report improved techniques for purifying and manipulating DHBD of B. cepacia LB400. Steady-state kinetic studies were conducted to determine the specificity of the enzyme for different aromatic substrates and O 2 . In addition, the inhibition of the enzyme by different organic solvents was investigated. The results of the specificity and inhibition studies were evaluated in terms of crystallographically determined structures of catechol, 3-methylcatechol, and DHB complexes.

MATERIALS AND METHODS
Strains, Media and Growth-DHBD was hyperexpressed in P. putida KT2442 freshly transformed with pLEBD4, a broad host range expression plasmid containing the structural gene encoding the dioxygenase, as described previously (16). For DHBD expression, the strain was grown in Luria broth containing a potassium phosphate buffer described for Terrific Broth (17)  Cultures were grown at 30°C and 250 rpm. One liter of media in a 2-liter flask was inoculated with 4 ml of an overnight culture. When the O.D. at 600 nm of the culture reached 0.5, isopropyl-1-thio-␤-D-galactopyranoside was added to a final concentration of 0.25 mM and the culture was incubated for an additional 18 h before harvesting.
Chemicals-Catechol, 3-methylcatechol, and protocatechuate were purchased from Aldrich (Mississauga, Ontario, Canada). DHB and 3-ethylcatechol were a kind gift from Victor Snieckus. For crystallographic experiments, DHB was obtained from Wako. Ferene-S was purchased from ICN Biomedicals Inc. (Aurora, OH). All other chemicals were of analytical grade.
Protein Purification-The protein was purified following a modified protocol (13). After disruption of the cells, all procedures were performed under an inert atmosphere unless otherwise specified. Chromatography was performed on an Ä KTA Explorer (Amersham Pharmacia Biotech, Baie d'Urfé, P.Q., Canada). This instrument was installed next to a Labmaster Model 100 glovebox (MBraun, Peabody, MA) equipped with 1/16 and 3/8 inch Swagelok feed-throughs as well as a potted RS232 feed-through. In this configuration, anaerobic buffers could be delivered to the Ä KTA Explorer and the column eluate could be directed to a fraction collector in the glovebox. Buffers were made from water purified on a Barnstead NANOpure UV apparatus to a resistivity of greater than 17.5 M⍀cm. Buffer A was 10 mM Tris, pH 7.5, containing 10% t-butanol, 2 mM dithiothreitol, and 0.25 mM ferrous ammonium sulfate. Buffer B was buffer A containing 1.0 M NaCl. Buffer C was 10 mM Tris, pH 8.0, containing 10% t-butanol, 33 mM ammonium sulfate, 2 mM dithiothreitol, and 0.25 mM ferrous ammonium sulfate. Buffers were sparged with N 2 and equilibrated in the glovebox for at least 24 h prior to the addition of t-butanol, dithiothreitol, and ferrous ammonium sulfate.
Cells from 3 liters of culture were harvested by centrifugation. The cell pellet was resuspended in 15 ml of 10 mM Tris, pH 8.0, containing 20% t-butanol, 1 mM MgCl 2 , 1 mM CaCl 2 , and 0.1 mg/ml DNase I. The cells were disrupted by two successive passages through a French Press (Spectronic Instruments Inc., Rochester, NY) using an operating pres-sure of 20,000 p.s.i. The cell debris was removed by ultracentrifugation in gas-tight tubes at 50,000 rpm for 60 min in a T1250 rotor (DuPont Instruments, Wilmington, DE). The clear supernatant fluid was carefully decanted and referred to as the raw extract.
The raw extract (approximately 20 ml) was divided into four equal portions, each of which was loaded onto a HR10/10 Mono Q anion exchange column equilibrated with buffer A. The column was operated at a flow rate of 3.0 ml/min. The enzyme activity was eluted with a 66-ml linear gradient of 8 -19% buffer B. Fractions of 2.5 ml were collected. Fractions containing activity from the four runs were concentrated to 4 ml with a stirred cell concentrator equipped with a YM10 membrane (Amicon, Oakville, Ontario, Canada) and loaded onto a HiLoad 26/60 Superdex 200 column equilibrated with buffer C. The column was eluted with buffer C at a flow rate of 2.0 ml/min. Fractions (5 ml) exhibiting activities of more than 1500 units/ml were combined, concentrated to 25-35 mg/ml protein, and flash frozen as beads in liquid N 2 . Purified DHBD was stored in buffer C at Ϫ80°C for several months without significant loss of activity.
Handling of DHBD Samples-Exogenous iron was anaerobically removed from small samples of DHBD prior to kinetic experiments by gel filtration chromatography. Typically, 100 -200 l of purified DHBD was thawed in the glovebox and applied to a 0.7 ϫ 5-cm column of Bio-Gel P6 DG (Bio-Rad, Mississauga, Ont., Canada) equilibrated with stabilization buffer lacking bovine serum albumin (see below). The sample of DHBD was further diluted with stabilization buffer as appropriate.
The ability of different anaerobic buffers to stabilize the DHB cleaving activity of DHBD was investigated by monitoring the activity of a 25-50 g/ml solution of the enzyme over 24 h. Parameters that were varied include pH and the concentrations of ferrous ammonium sulfate, dithiothreitol, and bovine serum albumin. Acetone, ethanol, glycerol, isopropyl alcohol, and t-butanol were tested for their ability to stabilize the activity of DHBD.
SDS-polyacrylamide gel electrophoresis was performed on a Bio-Rad MiniPROTEAN II apparatus and stained with Coomassie Blue according to established procedures (17). Protein concentrations were determined by the Bradford method (18). Iron concentrations were determined colorimetrically using Ferene S (19).
Kinetic Measurements-Enzymatic activity was measured by following the consumption of dioxygen using a Clark-type polarographic O 2 electrode (Yellow Springs Instruments Model 5301 (Yellow Springs, OH)). Reactions were performed in a thermojacketted Cameron Instrument Co. model RC1 respiration chamber (Port Aransas, TX) equipped with a Lauda Model RM6 circulating water bath. The electrode signal was amplified using a Cameron Instrument model OM200 O 2 meter and recorded on a microcomputer equipped with a PC-LPM-16 multifunction board and Virtual Bench Data Logger (National Instruments, Austin, TX). Data were recorded every 0.1 s. Initial velocities were determined from progress curves by analyzing the data using Microsoft Excel (Redmond, WA). The slope of the progress curve and the correlation coefficient of the slope were calculated for all consecutive 6-s intervals using the full set of 61 data points. The calculated slopes had correlation coefficients of at least 0.998.
Reaction buffers containing different concentrations of dissolved O 2 were prepared by vigorously bubbling them with humidified mixtures of O 2 and N 2 gases for at least 15 min prior to the experiment. The equilibrated buffer was transferred to the reaction chamber using a gas-tight syringe and the stopper was inserted into the reaction chamber. During this operation, the reaction chamber was flushed continuously with the humidified gas mixture. Ultra high purity O 2 (100% O 2 or 10% O 2 in N 2 ) and prepurified N 2 were mixed in the desired proportions with a stainless steel Concoa Model 561 gas proportioner. This mixture was humidified by bubbling it through a 5 ϫ 10-cm column of water. The concentration of dissolved O 2 in the reaction mixture was verified using the O 2 electrode.
The standard activity assay was performed in a total volume of 1.45 ml of air-saturated 20 mM HEPPS, 80 mM NaCl (I ϭ 0.1 M), pH 8.0, 25.0 Ϯ 0.1°C containing 80 M DHB. The reaction was initiated by injecting between 2 and 10 l of an appropiate dilution of enzyme preparation into the reaction chamber. Buffers were used within 24 h of preparation. Stock solutions were prepared fresh daily and stored under argon on ice. On each day kinetic assays were performed, the zero of the O 2 electrode was established by adding an excess of sodium hydrosulfite to the buffer in the reaction chamber. The O 2 electrode was calibrated using standard concentrations of protocatechuate 3,4-dioxygenase and protocatechuate (20) or DHB and an excess of DHBD. The amount of active DHBD in the assay was defined by the iron content of the injected purified enzyme solution. This quantity was used in the calculation of the specificity and catalytic constants. One unit of enzy-SCHEME 1 Stabilization and Inhibition of an Extradiol Dioxygenase matic activity was defined as the quantity of enzyme required to consume 1 mol of O 2 /min. Specificity, coupling, inactivation and inhibition experiments were carried out in 20 mM HEPPS, 80 mM NaCl, pH 8.0, 25.0 Ϯ 0.1°C. The coupling of catecholic substrate and O 2 consumption was investigated by monitoring the amount of O 2 consumed upon the addition of weighed amounts of catecholic substrate to the reaction mixture. An excess of DHBD was used in these experiments, whereas a limiting amount of DHBD was used in experiments designed to determine the partition ratio for a given catecholic substrate. The amount was chosen such that the enzyme was completely inactivated before the catecholic substrate or O 2 was completely consumed in the reaction mixture. The partition ratio was calculated from the ratio of the amount of O 2 consumed to the amount of active DHBD added to the assay.
In specificity experiments with catechol, 3-methylcatechol, 3-ethylcatechol, and DHB, concentrations of the aromatic substrates were varied from 0.2 times the determined K m to the maximum concentration possible without suicide inhibition affecting the initial velocities. Dissolved O 2 concentrations were varied from approximately 60 to 1300 M.
Inhibition experiments with t-butanol, isopropyl alcohol, acetone, ethanol, and glycerol were performed in air-saturated buffer (290 M dissolved O 2 ) using concentrations of the inhibitor up to 3 M unless otherwise noted. In these experiments, the catecholic substrate was used at concentrations below those at which substrate inhibition or suicide inhibition were observed.
Analysis of Steady-state Data-Catechol 2,3-dioxygenase utilizes a compulsory order ternary complex mechanism in which the binding of catecholic substrate, A, precedes that of dioxygen (9). In addition, DHBD of B. cepacia LB400 is subject to substrate inhibition at high concentrations of catecholic substrate (13). A mechanism accounting for these observations is shown in Scheme 2 The steady-state rate equation derived from this mechanism using the King-Altman method (21) is given in Equation In Equation 1, K mA represents the K m for the catecholic substrate; K mO 2 , the K m for O 2 ; K dA , the dissociation constant for the catecholic substrate; and K iA , the inhibition constant for the catecholic substrate. K mA and K mO 2 are not classically defined Michaelis constants as the effect of substrate inhibition ([A] 2 ) is to make the rate approach zero at high concentrations of A. However, K mA and K mO 2 are derived from the same first-order rate constants as the equivalent steady-state constants for a compulsory order ternary complex mechanism in which substrate inhibition does not occur (e.g. K mO 2 ϭ (k -2 ϩ k 3 )/k 2 ). Furthermore, K mA and K mO 2 are the ratios between k cat and the specificity constants k A and For specificity experiments in which the concentrations of substrates A and O 2 were varied, the initial velocities were fit to Equation 1 or to the equivalent steady-state rate equation describing a compulsory order ternary complex mechanism without substrate inhibition. For specificity experiments in which the concentration of only one substrate was varied, the initial velocities obtained were fit to the Michaelis-Menten equation or to the equivalent equation describing a mechanism in which substrate inhibition occurs ( In this equation, K mA app and K iA app are the apparent K mA and K iA in air-saturated buffer.
Steady-state kinetic data obtained from experiments using organic additives as inhibitors, I, were fit to competitive, uncompetitive, and mixed inhibition equations ( In Equation 3, K ic represents the competitive inhibition constant and K iu app , the apparent uncompetitive inhibition constant. The latter is related to the true K iu by Equation 4. Data were fit to steady-state rate equations using the least squares and dynamic weighting options of LEONORA (22).
Preparation of Crystalline Enzyme-Substrate Complexes and Diffraction Experiments-All crystals were prepared and mounted under anaerobic conditions within a N 2 atmosphere glove box (Innovative Technology, Newburyport, MA) maintained at 2 ppm O 2 or less. Crystals of DHBD were grown in the presence of t-butanol and absence of substrates as described previously (3). Binary complexes were prepared by incubating crystals in small volumes (about 1 ml) of anaerobic solutions containing the substrate, 22% (w/v) PEG-4000, 15% (v/v) t-butanol, and 100 mM HEPES buffer at pH 7.5; for some experiments, 1 mM Fe(II)(NH 4 ) 2 (SO 4 ) 2 and 0.5 mM Na 2 S 2 O 4 were also present. For catechol, 3-methylcatechol, and DHB, the concentrations used were 16, 10, and 1 mM, respectively. Crystals were incubated for 12-24 h and mounted in quartz capillaries sealed with epoxy cement. By these procedures, the enzyme was maintained in the active, ferrous form. Diffraction patterns were measured at room temperature (ϳ20°C) by the rotation method with the use of an RAXIS IIc imaging plate diffractometer. X-rays were provided by a Rigaku RU200 x-ray generator equipped with a copper anode and focusing mirror optics (Molecular Structure Corp.) and operated at 50 kV and 80 -100 mA. The rotation angle and exposure time per image were typically 1°and 10 min, and the crystal-to-detector distance was 110 mm. The diffraction data were analyzed with the HKL package (23).
Structure Refinement and Other Crystallographic Procedures-The CCP4 package (24) was used for general crystallographic calculations, whereas the XPLOR system (25) was used for restrained (26) crystallographic refinement. Electron density analysis and model building were accomplished with the program O (27). Refinement of the 3-methylcatechol complex was initiated with a model derived from the refined structure of the substrate-free protein (3) by deleting the active site iron atom and its two water ligands as well as one additional water and a t-butanol molecule located within 6 Å of the iron. Following one cycle of XPLOR refinement (coordinates and individual B factors), F o Ϫ F c and 2F o Ϫ F c maps were analyzed, and the active site iron was added to the model. Protein-associated solvent atoms with B factors larger than 60 Å 2 were deleted, and a single solvent atom was added to account for a strong, positive feature ϳ9 Å from the iron. Following a second cycle of XPLOR refinement and map evaluation, 3-methylcatechol and two waters adjacent to the iron atom were added, and the previously added solvent atom was replaced with a t-butanol molecule. Four additional rounds of computational refinement and model building completed the refinement. Whenever the active site iron atom was included in the model, the iron-ligand bond distances were restrained to a target value of 2.2 Å with a weak force constant of 10 kcal⅐mol Ϫ1 ⅐Å Ϫ2 .
During the refinement, residual positive difference peaks in the vicinity of the active site were recognized as features representing the active site structure of the substrate-free state. That is, it was apparent that the crystal contained active sites in two states: the substrate-free form and the ES complex. The structure at the active site was therefore modeled by two sets of atoms at partial occupancy (50% each). Refinement of the enzyme-catechol complex followed a similar course and required similar modeling of the active site. The refinement of the enzyme-DHB complex is described elsewhere. 2 The procedures were similar, but the substrate-binding site was fully occupied by DHB so that the use of two models of the active site was not necessary.

RESULTS
Relevant details of the purification are shown in Table I. The enzyme was estimated to be greater than 99% pure as judged by SDS-polyacrylamide gel electrophoresis followed by Coomassie Blue staining (results not shown). This is comparable to the purity obtained previously (13). Anaerobically purified DHBD had a specific activity of 430 units/mg, which was over twice that reported previously (13).
The activity of a 25-50 g/ml solution of DHBD was stabilized by the presence of dithiothreitol and bovine serum albu-2 S. Han, L. D. Eltis, and J. T. Bolin, manuscript in preparation.

SCHEME 2
Stabilization and Inhibition of an Extradiol Dioxygenase min, but not ferrous ammonium sulfate. The activity was most stable at pH 8.0; deviations of 0.25 pH units decreased this stability. In the presence of 1 M t-butanol or isopropyl alcohol, approximately 24% of the activity remained after 24 h, as compared with 19% in the presence of 1 M ethanol or glycerol and 13% in the absence of organic compounds (see data on inhibition below). For steady-state kinetic assays, DHBD was diluted in 20 mM HEPPS, 80 mM NaCl, pH 8.0, 10% t-butanol, 2 mM dithiothreitol, 0.1 mg/ml bovine serum albumin and stored on ice under an inert atmosphere. Under these conditions, more than 90% of the initial activity of the diluted preparation remained after 6 h.
The coupling of catecholic substrate and O 2 utilization by DHBD was investigated in 20 mM HEPPS, 80 mM NaCl, pH 8.0 (25°C), using an O 2 electrode that had been calibrated with protocatechuate and protocatechuate 3,4-dioxygenase (20). Within experimental error, the amount of O 2 consumed corresponded to the amount of catechol or DHB added to the reaction mixture (results not shown). It was concluded that the utilization of catecholic substrate and O 2 by DHBD were tightly coupled under these conditions. In subsequent experiments, the O 2 electrode was calibrated with DHB and DHBD.
In experiments performed in air-saturated buffer using low concentrations of DHBD, the velocity of the reaction decreased to zero prior to the complete consumption of catecholic substrate or O 2 . This inactivation of the enzyme was faster in the presence of catechol than DHB. These observations indicated that DHBD was subject to irreversible, suicide inhibition as has been described for catechol 2,3-dioxygenase (14,28). In air-saturated buffer, the partition ratio of DHBD increased with the size of the substituent at C-3 of the catechol, varying from 1,230 for catechol to 84,900 for DHB (Table III).
High concentrations of catecholic substrate were observed to reduce the initial velocity of O 2 consumption. To assess whether this decrease in the initial velocity was due to inactivation of DHBD via suicide inhibition, the substrate concentrations and reaction velocities were calculated 10 -30 s after the initiation of the reaction. Reaction velocities were calculated from 61 data points over a 6-s interval. Substrate concentrations were based on initial concentrations and the depletion of O 2 . Measured reaction velocities were compared with those calculated using steady-state equations, best-fit parameters, and the estimated substrate concentrations. In experiments in which the catechol or 3-methylcatechol concentration exceeded 750 or 1400 M, respectively, the measured velocity after 10 s was less than 90% the calculated velocity. The difference between the measured and calculated velocities increased as the reaction progressed. These observations indicated that suicide inhibition significantly affected the initial velocity under these conditions and no initial velocities obtained from such experiments were further analyzed. In contrast, suicide inhibition affected the initial velocities of reactions performed with DHB and 3-ethylcatechol to a much lesser extent. Using 4 mM 3-ethylcatechol, the measured velocity after 10 s was within 90% of the calculated velocity. Using 800 M DHB, the measured velocity after 30 s was within 95% of the calculated velocity.
When the initial rates of O 2 consumption obtained over a range of concentrations of DHB (1-800 M) and O 2 (55-1300 M) were fitted to Equation 1, random trends in the residuals were observed. Examples of the data and the quality of the fit are shown in Fig. 1. The values for the steady-state kinetic parameters obtained from this fit are provided in Table II. Similar values for K mA , K mO 2 , K dA , and k cat were obtained by fitting a subset of these data, corresponding to 1-90 M DHB, to an equation describing a compulsory order ternary complex mechanism without substrate inhibition. The data obtained using 3-ethylcatechol (10 -4200 M) and O 2 (55-1300 M) also yielded random trends in the residuals when fitted to Equation   not be varied over a sufficiently wide range. As discussed above, the useful range of concentration for 3-methylcatechol (50 -1400 M) was also limited by suicide inhibition. Apparent steady-state kinetic parameters for catechol and 3-methylcatechol were thus evaluated by fitting the data obtained in airsaturated buffer to the Michaelis-Menten equation (Table III). Fits of the same data to Equation 2 (with substrate inhibition) yielded essentially the same values for K mA app and k cat app , respectively. The values for DHB and 3-ethylcatechol in air-saturated buffer, evaluated by fitting the data to the substrate inhibition equation, are also provided in Table III and do not differ significantly from the values calculated from the parameters provided in Table II. Consistent with previous studies (13), the apparent specificity of DHBD in air-saturated buffer is 350-fold higher for DHB than for catechol. Although the K mO 2 of DHBD in the presence of catechol could not be evaluated, at a concentration of 1275 M catechol, the apparent constant, K mO 2 app was 1000 Ϯ 160 M.
Steady-state kinetic data obtained from the DHBD-catalyzed cleavage of DHB with t-butanol present as an inhibitor showed random trends in the residuals when fit to an equation describing mixed inhibition (Equation 3). When equations describing competitive or uncompetitive inhibition were used, larger residuals were obtained. The competitive component of this inhibition was stronger than the uncompetitive component (Fig.  2), even when the limiting concentration of O 2 is considered (Table IV). The relative ability of the tested organic molecules to inhibit the DHBD-catalyzed cleavage of DHB was t-butanol Ͼ isopropyl alcohol ϳ acetone Ͼ ethanol Ͼ Ͼ glycerol (Table IV).
Inspection of the quality of fits to equations describing competitive, uncompetitive, and mixed inhibition, respectively, indicated that the DHBD-catalyzed cleavage of catechol and 3-methylcatechol were inhibited by t-butanol in a competitive fashion. The inhibition constants were similar in magnitude to the competitive inhibition constants of the DHBD-catalyzed cleavage of DHB (Table IV). Nevertheless, weak uncompetitive inhibition of the DHBD cleavage of catechol and 3-methylcatechol could not be ruled out as the maximum concentration of t-butanol used in these experiments was 2 M and the maximum substrate concentration did not exceed three times their respective K mA aap values. Table V documents the extent and quality of the diffraction data. Models for the catechol, 3-methylcatechol, and DHB complexes were refined against diffraction data extending to 1.9-, 1.9-, and 2.0-Å resolution, respectively. Statistics pertinent to the quality of the final models are provided in Table VI. Fig. 3 illustrates the essential features of the active site in the substrate-free and substrate bound states of the enzyme. The iron is bound to the protein by the side chains of His-146, His-210, and Glu-260 in both states. In the substrate-free state, the iron ligation is five-coordinate, with sites trans to His-210 and Glu-260 occupied by water; the site trans to His-146 is unoccupied. A t-butanol molecule binds at the bottom of a channel that leads from the surface to the iron atom. Although the electron density for the t-butanol is strong, its features suggest rotationally disordered binding; van der Waals contacts with seven side chains are potentially involved in different orientations. The distance from the central carbon of tbutanol, C-1, to the iron atom is 6.0 Å.
The mode of (catecholic) substrate binding is similar in all three complexes and is consistent with the results reported for complexes of the ferric form of a homologous DHBD (4). One hydroxyl group of the substrate binds in the site trans to His-146, whereas the other binds trans to His-210, displacing a water ligand. The water trans to Glu-260 shifts by approximately 1 Å, but remains weakly associated with the iron in all complexes of the active, ferrous form of the enzyme (this work), but is not observed in complexes of the inactive, ferric form (4).
The refinements of the catechol and 3-methylcatechol complexes established that the crystals in each case contained molecules in both the substrate-free and substrate-bound states, as demonstrated by the electron density maps shown for the 3-methylcatechol complex in Fig. 4. This phenomenon was not observed for the preferred substrate DHB. 2 The maps very clearly demonstrate the significant overlap between the substrate-binding site and the site occupied by t-butanol in the substrate-free form of the enzyme. In fact, the refined position of the central atom of the t-butanol in the substrate-free form is within 1.1 Å of the position of the 3-methyl group of 3-methylcatechol and within 0.7 Å of the position at the center of the nonhydroxylated ring of DHB in the respective DHBD-substrate complexes.
The crystal structures of all three DHBD-substrate complexes demonstrate the binding of t-butanol in a location distinct from the site occupied in the substrate-free structure: the position of C-1 is shifted by more than 4.8 Å and it is more than 8.4 Å distant from the iron. In this auxiliary site t-butanol is in van der Waals contact with the non-hydroxylated ring of DHB, but not with any atom of catechol or 3-methylcatechol. For DHB, three (C to C) contacts at distances of 3.3, 3.8, and 4.2 Å are observed, whereas the shortest distances for catechol and 3-methylcatechol are 4.5 Å. DISCUSSION The described rapid, anaerobic purification of DHBD of B. cepacia LB400 yielded a preparation whose specific activity is at least twice that of aerobically purified preparations of the same enzyme (13) or a highly similar one (29). Furthermore,   3-ethylcatechol Experiments were performed using 20 mM HEPPS, 80 mM NaCl, pH 8.0, at 25°C. The values of the parameters and their standard errors (provided in parentheses) were calculated using the least squares and dynamic weighting options of LEONORA (22). The data sets used to calculate the parameters for DHB and 3-ethylcatechol contain 189 and 96 data points, respectively. the anaerobic removal of exogenous iron from preparations used for steady-state kinetic studies enabled the calculation of catalytic and specificity constants based on the iron content of these preparations. Electron paramagnetic resonance, magnetic circular dichroic, and x-ray absorption spectra indicated that DHBD samples prepared in this manner contain a single species of iron and that the properties of this species are those expected for a high spin Fe(II) in the active site of the enzyme (7). 3 The steady-state kinetic analysis revealed that the ability of DHBD from B. cepacia LB400 to utilize O 2 is strikingly different from that of evolutionarily related catechol 2,3-dioxygenases. Although the steady-state utilization of O 2 has not been investigated in many extradiol dioxygenases, the K mO 2 of DHBD is 2-3 orders of magnitude higher than those of a number of catechol 2,3-dioxygenases determined using catechol (0.7-10 M) (9,15,30). Interestingly, the K mO 2 of an extradioltype 2-aminophenol dioxygenase was reported to be 710 M (31). Significantly, the nature of the substituent at C-3 of the catechol does not appear to appreciably affect the K mO 2 of DHBD. Intradiol dioxygenases also have a wide range of K mO 2 .
The kinetic analysis further establishes that DHBD is subject to two forms of substrate inhibition: reversible substrate inhibition, and a mechanism-based inactivation or suicide inhibition. The latter has also been reported for xylE-encoded catechol 2,3-dioxygenase and seems to involve the oxidation of the active site iron to the catalytic incompetent ferric state (14,28). For both DHBD and catechol 2,3-dioxygenase, suicide inhibition is more marked for poorer substrates, suggesting that the substrate-binding pocket of these enzymes is tuned both to maximize specificity for a particular substrate and minimize unproductive oxidation of the iron during the cleavage of that substrate. Interestingly, the partition ratio of catechol 2,3dioxygenase for catechol is 1,400,000 (28), indicating that DHBD is much more susceptible than is catechol 2,3-dioxygenase to suicide inactivation by its putative preferred substrate. Despite the higher susceptibility of DHBD to suicide inactivation, the bph pathway apparently does not contain a homologue to XylT, the small ferredoxin of the TOL pathway, which serves to maintain the active site iron of catechol 2,3-dioxygenase in the ferrous state (34).
The reversible substrate inhibition observed in DHBD has been reported for a number of other DHB-cleaving extradiol dioxygenases (13,(35)(36)(37)(38) as well as for a 2,3-dihydroxyphenyl propionate cleaving enzyme (39) although it has only rarely been reported for catechol 2,3-dioxygenases (40). However, for the other DHB cleaving enzymes, it is not clear what proportion of the decrease in the initial rate of DHB cleavage at high concentrations of DHB is due to reversible substrate inhibition and irreversible suicide inhibition, respectively. Notably, the initial rates of cleavage of substituted catechols by a DHBD of strain BN6 could not be fitted to substrate inhibition Equation 2 (38). DHBD of B. cepacia LB400 is clearly subject to both modes of inhibition by both DHB and 3-ethylcatechol. Because catechol and 3-methylcatechol are such potent suicide inhibitors of DHBD, it was not possible to determine whether these substrates also reversibly inhibit the enzyme. While the mechanism of reversible substrate inhibition is not clear, it is unlikely to involve negative cooperativity between the subunits of DHBD as it has also been reported for a monomeric enzyme (37). It is possible that DHB could occupy the auxiliary tbutanol-binding site observed in the DHBD-DHB complex, thereby inhibiting the enzyme in a similar manner (see below).
Previous studies have established that organic additives such as isopropyl alcohol and acetone stabilize the active site iron of extradiol dioxygenases (12,13). The crystallographic data from DHBD indicate that such small organic molecules can occupy the active site, close to the catalytic iron center (3). In this position, the organic molecule could stabilize the active site and/or protect the iron from direct access by oxidants or substrates. Consistent with this notion, t-butanol competitively inhibited the cleavage of DHB, 3-methylcatechol, and catechol by DHBD. Moreover, t-butanol and isopropyl alcohol, which inhibit DHBD more effectively than ethanol and glycerol, also stabilize the enzyme more effectively (Table IV). Interestingly, acetone competitively inhibits catechol 2,3-dioxygenase with K ic of 13 mM (12,15). This suggests that this enzyme has a much higher affinity for acetone than DHBD has for t-butanol, and may explain why the former is so much more stable in acetone-containing buffers (15) than is DHBD in t-butanolcontaining buffers. Inspection of the t-butanol-binding site in DHBD reveals that it is partly formed by nonconserved residues, suggesting that the best organic stabilizer will be dioxygenase-specific.
Finally, we note that the structural results are remarkably concordant with the specific types of inhibition manifested by t-butanol with the three substrates DHB, 3-methylcatechol, and catechol. The uncompetitive component of mixed inhibition observed with DHB as substrate correlates with the close interactions between DHB and a t-butanol molecule that binds at an auxiliary site in the enzyme-DHB complex. By binding in this position, t-butanol could uncompetitively inhibit DHB cleavage by decreasing the access of O 2 to the active site, perturbing the substrate orientation or conformation so as to lower the rate of ring cleavage, and/or interfering with product release. The absence of an uncompetitive component for the  a Calculated from K iu app using Equation 6. b Enzymatic activity remaining after 24 h in a solution of DHBD containing 1 M of the listed organic additive (see "Materials and Methods"). c No inhibition detected. Statistics are provided for the lowest resolution shell, the highest two resolution shells, and the full data set. b Obs., percentage of possible reflections measured at least once. The mean redundancy for each data set is Ն5.6 observations per reflection. a All models include some atoms modeled at partial occupancy. For the catechol (24 atoms) and 3-methylcatechol structures (25 atoms), these include the substrate, two molecules of t-butanol, five water oxygens, and one adventitiously bound iron atom located at a position remote from the active site. For the DHB model (100 atoms) all represent alternative conformations of protein residues or side chains.
FIG. 4. Electron density maps and models illustrating the structure of the DHBD:3-methylcatechol-bound and substrate-free forms of DHBD. Each part is a (divergent) stereo drawing prepared with the program MolView (42). The identical observed structure factors were used in all maps, which demonstrates the presence of both forms of the enzyme in the same crystal. All maps are at 1.9-Å resolution and are contoured at two standard deviations above the mean of the map. In the models, the carbon atoms are more darkly shaded than the nitrogen and oxygen atoms. A, F o Ϫ F c electron density representing the iron, 3-methylcatechol, and two water ligands. The F c 's and phases are from the structure of the substrate-free enzyme (3). The model is the initial fit to this density. B, residual F o Ϫ F c electron density following refinement of a model that included the iron, 3-methylcatechol, two water ligands, and a t-butanol bound in the auxiliary site, as shown. The F c 's and phases smaller substrates correlates with the larger catechol to tbutanol (auxiliary site) distances: the closest contact distance in either complex is 4.5 Å, which is much larger than the expected van der Waals contact distance, 3.8 Å (41). The crystallographic data thus suggest that the manifestation of uncompetitive inhibition with DHB as substrate, whatever the mechanism, depends on the close contact between t-butanol and DHB.