Mechanism of soluble epoxide hydrolase. Formation of an alpha-hydroxy ester-enzyme intermediate through Asp-333.

18O-Labeled epoxides of trans-1,3-diphenylpropene oxide (tDPPO) and cis-9,10-epoxystearic acid were synthesized and used to determine the regioselectivity of sEH. The nucleophilic nature of sEH catalysis was demonstrated by comparing the enzymatic and nonenzymatic hydrolysis products of tDPPO. The results from single turnover experiments with greater or equal molar equivalents of sEH:substrate were consistent with the existence of a stable intermediate formed by a nucleophilic amino acid attacking the epoxide group. Tryptic digestion of sEH previously subjected to multiple turnovers with tDPPO in H218O resulted in the isolation and purification of a tryptic fragment containing Asp-333. Electrospray mass spectrometry of this fragment conclusively illustrated the incorporation of 18O. After complete digestion of the latter peptide it was shown that Asp-333 of sEH exhibited an increased mass. The attack by Asp-333 initiates enzymatic activity, leading to the formation of an α-hydroxyacyl-enzyme intermediate. Hydrolysis of the acyl-enzyme occurs by the addition of an activated water to the carbonyl carbon of the ester bond, after which the resultant tetrahedral intermediate collapses, yielding the active enzyme and the diol product.

Mammalian epoxide hydrolases (E.C. 3.3.2.3) have been implicated in the metabolism of epoxide containing xenobiotics, many of which are believed to be mutagenic and/or carcinogenic (1)(2)(3). During the past 20 years the mammalian microsomal epoxide hydrolase (mEH) 1 has received a great deal of attention partly due to its higher selectivity for cyclic and arene oxides (4 -6). A great deal of work has provided a clear picture of the regio-, stereo-, and enantiospecificity of mEH (7)(8)(9)(10). Most researchers agree that mEH hydrolyzes the epoxide via an anti (commonly referred to as trans) opening of the oxirane, where attack of the nucleophile usually takes place at the least sterically hindered carbon. These results are substantiated by substituent effects on the rate of hydrolysis as investigated by Dansette et al. (11) and kinetic solvent isotope studies as reported by Armstrong et al. (12). By the use of single turnover experiments in H 2 18 O (13), Lacourciere and Armstrong have also postulated that the hydrolysis of epoxides by mEH proceeds via the intermediary of a nucleophilic amino acid, yielding an acyl-enzyme intermediate, which is hydrolyzed further to release the diol product and the active enzyme. This is in contrast with the previously more accepted hypothesis in which an activated water molecule was thought to be responsible for direct attack on the epoxide.
Much less work has been completed on sEH, but the preliminary data closely parallel those for mEH. Because of the critical role of sEHs in the metabolism of xenobiotics and their possible involvement in the biosynthesis of metabolites (14)(15)(16)18), 2 a detailed understanding of the catalytic mechanism of sEH is imperative. The anti opening of epoxides has been demonstrated on a select few substrates along with H 2 18 O studies, which have shown that nucleophilic attack occurs on the less hindered carbon (7,19,20). Recently, Hammock et al. (21) were able to trap a postulated acylenzyme intermediate, suggesting the possible involvement of a nucleophilic amino acid in the hydration of epoxides by sEH. Pinot et al. (22) were able to isolate a tryptic fragment of sEH containing the postulated acylated amino acid. The latter discovery is very significant in light of studies that have compared the sequence homology of several EHs. These studies have shown significant sequence similarity among EHs and a bacterial haloalkane dehalogenase (DhlA, from the bacterium Xanthobacter autotrophicus), which, among other regions of homology, possesses a conserved histidine and aspartic acid tandem (23)(24)(25). DhlA belongs to a group of enzymes known as the ␣/␤ hydrolase fold family, all of which contain a catalytic triad consisting of a nucleophile, a histidine, and an acid (26). If EHs, and particularly sEH in our case, belong to the ␣/␤ hydrolase fold family of enzymes then it can be inferred that the conserved aspartic acid is responsible for the nucleophilic attack on the epoxide, resulting in the acyl-enzyme intermediate observed by Hammock et al. (21).
Herein we detail experiments associated with the mechanism of sEH. Results of regioselectivity of hydration, direction of oxirane opening, mechanism of hydration, and identification of the putative catalytic amino acid are reported. Also, possible mechanisms of inhibitory action by chalcone oxides and glycidols are discussed.
MATERIALS AND METHODS (CDCl 3 ) as solvent and tetramethylsilane as the internal standard. GC/MS analyses were performed using a VG Trio-2 spectrometer equipped with a VG11-250 data system. Electron ionization (70 eV) was employed. GC separation was performed on a Hewlett Packard 5890A gas chromatograph fitted with a DB-5 (30-m) capillary column directly attached to the mass spectrometer. High performance liquid chromatography (HPLC) analysis employed a Perkin Elmer system consisting of a series 410 BIO pump and a Rheodyne injector. UV absorbance was monitored at 215 nm using a Perkin Elmer LC-235 diode array detector. HPLC separation was accomplished by the use of a Vydac C4 (4.6 mm ϫ 25 cm) column. A microbore HPLC model UMA 600 (Michrom BioResources, Inc., Pleasanton, CA) equipped with a C18 preinjection peptide trap and a 5-, 300-Å pore size, 1 mm ϫ 5 cm column packed with Vydac C4 particles (Michrom BioResources, Inc.) was used to concentrate tryptic digest peaks obtained from the analytical HPLC. UV absorbance was monitored at 210 nm. Electrospray mass spectra of peptides were generated in positive ion mode using a VG/Fisons Quattro-BQ triple quadrupole mass spectrometer (VG Biotech, Altrincham, UK). Peptide sequencing was performed at the Protein Structure Laboratory (University of California, Davis) by automated Edman degradation, utilizing an Applied Biosystems gas phase sequenator (Applied Biosystems, Foster City, CA) coupled to an Applied Biosystems 120A PTH analyzer and M900 data system.
Enzyme Preparation and Specific Activity Measurement-Recombinant murine sEH was produced in the baculovirus expression system as previously reported (22). The enzyme was purified from cell lysate by affinity chromatography as described by Wixtrom et al. (30). Specific activity of the purified enzyme was measured using 3 H-trans-stilbene oxide as described previously (31). Protein concentrations were determined with BCA reagent from Pierce using bovine serum albumin (BSA) as standard.
HPLC Analysis of tDPPO Hydrolysis Products-A 100 mM stock solution of tDPPO was prepared in dimethyl formamide. One l of the stock tDPPO solution was added to 100 l of sodium phosphate buffer (100 mM, pH 7.4) containing either 100 g/ml BSA or 100 g/ml BSA and 3.9 g/ml sEH. Also, one l of the stock tDPPO solution was added to a 5% solution of perchloric acid (100 l) in order to generate erythroand threo-diols of tDPPO. The latter two products are generated because acidic hydrolysis of the epoxide leads to a stabilized benzylic carbocation that can be attacked by water from either face of its trigonal planar geometry. The latter solutions were incubated in a 37°C water bath, and 20 l fractions were withdrawn at specific times for direct HPLC injection. The erythro-and threo-diols and the epoxide were separated on a reverse phase C18 column and monitored spectrophotometrically at 215 nm with a diode array detector. The solvent system used was as follows: 70:30::H 2 O:acetonitrile from 0 to 12 min to resolve the two diols, followed by a step gradient at 12 min to 45:55::H 2 O:acetonitrile to elute the epoxide. An aqueous solution of threo-1,3-diphenyl-1,2-propanediol was also prepared (100 M final concentration) and analyzed under the same HPLC conditions in order to identify the threo-diol peak from the mixture of threo-and erythro-diols generated by acidic hydrolysis of tDPPO. The specific activity of the sEH used in this experiment was measured with 3 H-trans-stilbene oxide and found to be 895 nmol/min/mg protein.
GC/MS Analysis of Hydrolysis Products of tDPPO and cis-9,10-Epoxystearic Acid-Stock solutions of 16 O/ 18 O-labeled tDPPO, 16 O/ 18 Olabeled cis-9,10-epoxystearic acid, and 16 O/ 18 O-labeled methyl cis-9,10epoxystearate (100 mM) were prepared in dimethyl formamide. Enzymatic hydrolysis of the latter substrates was achieved by the addition of 1 l of substrate (stock) to 100 l of sodium phosphate buffer (100 mM, pH 7.4) containing 100 g/ml BSA and 3.9 g/ml sEH. Controls were also run with buffer containing only 100 g/ml BSA to ensure that no hydrolysis had taken place due to the interaction of substrates with buffer and BSA. All enzymatic and control experiments were incubated in a 37°C water bath for 30 min, after which saturated NaCl (100 l) was added to quench the hydrolysis. The specific activity of the sEH used in this study was measured by 3 H-trans-stilbene oxide assay to be 916 nmol/min/mg protein.
Acidic hydrolyses of the same epoxides were performed by introducing 2 l of the substrates (stock) into 5% perchloric acid solution (100 l). After 30 min of incubation at 37°C in a water bath, saturated NaCl (100 l) and enough 10% KOH was added to neutralize the solution. Basic hydrolysis of 16 O/ 18 O-labeled tDPPO was effected by the addition of 2 l of each epoxide to a 10% KOH solution (100 l) followed by incubation for 20 h at room temperature, at which time saturated NaCl (100 L) was added and the reaction mixture was neutralized with 12 M HCl.
The enzymatic and chemical hydrolysis products were extracted with similar efficiency by the addition of 450 l of ethyl acetate. Each test tube was vortexed vigorously and centrifuged in order to clearly separate the organic and aqueous phases. The top layer (ethyl acetate, 400 l) was withdrawn and placed in a conical reaction vessel, isooctane (5 l) was added as a trap solvent, and the solvent was evaporated under nitrogen flow. Each sample was derivatized 30 min prior to GC/MS analysis with BSTFA (50 l). The latter experiments were performed in quadruplicate. The enzymatic hydrolyses were performed with three separate enzyme preparations. It should be noted that [ 3 H]tDPPO was synthesized for substrate activity studies 3 and was utilized here to measure the efficiency of ethyl acetate extractions of the epoxide and its corresponding diol. It was found that the ethyl acetate extracts contained Ͼ99.9% of the [ 3 H]diol and 100% of the [ 3 H]epoxide that were originally incubated in the assay buffer.
Single Turnover Experiments-The following ratios of sEH: 16  To initiate enzymatic hydrolysis the cold substrate solutions were added and thoroughly mixed with the enzymatic solution at 0°C with a pipetter and incubated for 10 min in a 37°C water bath. This was sufficient for total conversion of epoxide to diol. The hydrolysis was quenched with the addition of saturated NaCl (100 l), and the buffer solution was extracted with ethyl acetate (450 l). The biphasic solution was vortexed vigorously and centrifuged. The top layer (ethyl acetate, 400 l) was withdrawn and placed in a conical reaction vessel, isooctane (5 l) was added, and the solvent was evaporated under nitrogen flow. The cis-9,10-epoxystearic acid hydrolysis products were dissolved in 50 l of methanol and were methylated with (trimethylsilyl)diazomethane (50 l, 2 M in hexanes). After 10 min the esterified products were again dried under nitrogen flow. Each sample was derivatized 30 min prior to GC/MS analysis with BSTFA (10 l). The diol was the sole metabolite detected chromatographically, and assignment of its structure was supported by co-chromatography and by full spectrum GC/MS. To achieve the highest degree of sensitivity and precision in the single turnover experiments the mass spectrometer was programmed to monitor only diagnostic fragments for each substrate (single ion monitoring). For the derivatized diol of tDPPO the masses monitored were m/z 179, 181, 193, and 195, and for the derivatized diol of cis-9,10-epoxystearic acid they were m/z 215, 217, 259, and 261. These correspond to fragments generated by cleavage of the C-C bond between the derivatized diol oxygens and can be used to establish the extent of labeling at each position.
Tryptic Digestion of 16  incubated with 100 mM tDPPO (4 l) in a 37°C water bath for 1 h, at which time the enzymatic solution was charged again with 100 mM tDPPO (4 l) and incubated for another hour (50 theoretical turnovers). The reaction was quenched by precipitating the protein with the addition of 100% trichloroacetic acid (22 l). The precipitated protein was pelleted by centrifuging the Eppendorf tube for 4 min at 16,000 ϫ g at 4°C. The pellet was washed twice with cold acetone (200 l) and resuspended in double distilled water (200 l) containing 1% acetic acid and then reprecipitated and washed as described above. The 1.0 mg of precipitated enzyme was dissolved in 250 l of 8 M urea, 500 mM sodium phosphate buffer (pH 7.4). The urea solution was divided into 10 separate Eppendorf tubes for digestion on a smaller scale. To each Eppendorf tube 45 mM dithiothreitol (5 l) was added, the tube was vortexed, and the mixture was incubated at 50°C for 15 min. After a 15-min incubation at room temperature, distilled water (60 l) and 5 g of trypsin (1 mg/ml) were added and vortexed. The tryptic digestion was allowed to proceed for 2 h at 37°C and was stopped either by freezing or by directly injecting into a reverse phase HPLC. The above procedure was duplicated with 1.0 mg of lyophilized sEH, which was rehydrated in H 2 16 O and incubated with excess tDPPO to achieve multiple turnovers. O-treated sEH were subjected to reverse phase HPLC analysis. Fragments were eluted with the following acetonitrile/H 2 O gradient: 0 -2 min at 100% A, changed to 75% B from 2 to 112 min, up to 100% B in 3 min, followed by 15 min at 100% B (A and B were 0.06% trifluoroacetic acid, H 2 O and 0.052% trifluoroacetic acid, 80% acetonitrile, H 2 O, respectively). The flow rate was set at 0.5 ml/min, and the peptide elution was monitored by UV absorbance at 215 and 280 nm. The elution profiles of the H 2 16 O-and H 2 18 O-treated sEH fragments were virtually identical. Nine peaks eluting from 84 up to 121 min were collected. The region containing these nine peaks was chosen for collection of the tryptic fragment containing the postulated catalytic Asp-333 by calculating the relative hydrophobicity of each theoretical tryptic fragment. It was found that T26, the tryptic fragment that contains Asp-333 (T is the designation for tryptic, and 26 refers to the 26th theoretical fragment obtained from the digestion of sEH by trypsin, counting from the N terminus) had the second highest hydrophobic score. Therefore, it was expected that T26 would elute toward the end of the elution profile. The program utilized for the hydrophobicity calculations was the Fisons Biopolymer analysis routine in the VG MassLynx software system. The pooled individual peaks from multiple HPLC injections (2-4 ml) were diluted with an equal volume of doubly distilled water and were subjected to microbore HPLC. The large volume of eluant for each peak was slowly loaded onto a peptide trap (C18 cartridge), which was installed inline on the microbore HPLC's injection port. The microbore HPLC was programmed to produce the following gradient: 0 -30 min at 50% A, changed to 100% B (A and B were 0.06% trifluoroacetic acid, H 2 O and 0.052% trifluoroacetic acid, 80% acetonitrile, H 2 O, respectively). The flow rate was set at 100 l/min. The eluting fragments were monitored by UV absorbance at 215 nm, and the peaks were collected in Eppendorf tubes (ϳ50 l total volume, Ͼ100fold reduction in volume).
Electrospray Mass Spectrometry of Tryptic Fragments-Electrospray mass spectra of the concentrated peptides were generated using a H 2 O:acetonitrile:formic acid::49:50:1 mobile phase. An Isco LC-500 syringe pump delivered the mobile phase at 5 l/min. Solutions of purified peptides were analyzed by direct flow injection using an injection volume of 10 l. Spectra were obtained in positive ion mode using a capillary voltage of ϩ3.5 kV, a cone voltage of 50 V, and a source temperature of 65°C. Spectra were scanned over the range of 200 -2,000 m/z at a rate of 20 s/scan. Twenty scans were combined using the VG MCA acquisition mode. Molecular weights were determined by using the maximum entropy deconvolution algorithm (MaxEnt TM ) to transform the range of 500-1500 m/z to give a true mass scale spectrum. Mixed polyethylene glycols were used for mass calibration.
Enzymatic Hydrolysis of T26 and GC/MS Analysis-Immobilized protease Sg (250-l suspension) was pelleted by centrifuging at 16,000 ϫ g for 10 min. The aqueous solution was withdrawn, and the pellet was resuspended in 250 l of NH 4 CO 3 buffer (50 mM, pH 8.0) containing 5 mM CaCl 2 , centrifuged, and resuspended again. The latter wash was repeated three times, after which the protease, suspended in the latter buffer, was added to lyophilized T26 isolated from the tryptic digestion of H 2 16 O-and H 2 18 O-treated sEH. The latter reaction was incubated at 37°C for 18 h, at which time the immobilized protease Sg was pelleted by a 30-min centrifugation at 16,000 ϫ g. The NH 4 CO 3 /CaCl 2 buffer containing the hydrolyzed amino acids was removed from the pellet, and the solution was frozen and lyophilized. The residue was derivat-ized with MTBSTFA (20 l) 1 h prior to GC/MS analysis. Controls for the detection of derivatized aspartic acid and glutamic acid were also prepared by derivatization of each amino acid with MTBSTFA. Samples were analyzed using a DB-5 (15-m) capillary column (0.25-mm inner diameter, 0.25-m film) using splitless injection and 70 eV electron ionization. Helium was used as carrier gas at a flow rate of 35 cm/s at 150°C, and the injector was held at 285°C. The column temperature was programmed from 150°C (1-min hold) to 275°C at 10°C/min. These chromatographic conditions have been found to give complete separation of tert-butyldimethylsilyl (

General Base Catalysis of sEH Demonstrated with tDPPO-
Early in our investigation of sEH we set out to determine the mechanism by which the epoxide is opened, i.e. A1 (S N 1 type) versus A2 (S N 2 type) (32). As can be seen in Fig. 1a, acid hydrolysis of tDPPO led to the expected equal amounts of erythro-and threo-diols (t R ϭ 6.92 and t R ϭ 8.08 min) due to the trigonal planar geometry of the resultant benzylic carbocation transition state, which allows attack of water to occur from either face of the carbocation. Osmium tetroxide cis-dihydroxylation of trans-1,3-diphenylpropene results in formation of the threo-diol, exclusively (Fig. 1b, t R ϭ 8.15 min). Therefore, the erythro-diol was identified as the peak at 6.92 min. Fig. 1c illustrates the incubation of sEH with tDPPO after 5 min. As the epoxide (t R ϭ 18.47 min) is turned over by the enzyme only the erythro-peak (t R ϭ 6.57 min) is formed, and after 40 min (Fig. 1d) all the epoxide has been converted to one single diastereoisomer. The anti (A2 type) opening of trans-epoxides exclusively produces erythro-diols, which was the only diastereoisomer observed in the hydrolysis of tDPPO. The anti opening of cis-9,10-epoxystearic acid with sEH has been previously established by Gill and Hammock (33) by demonstrating that threo-9,10-dihydroxystearic acid was the sole enzymatic product.
Regioselectivity of sEH with tDPPO and cis-9,10-Epoxystearic Acid-Next, we turned our attention to the regioselectivity of epoxide opening with sEH. The oxygen-labeled [ 18 O]tDPPO and [ 18 O]cis-9,10-epoxystearic acid were synthesized as described previously (27). [ 18 O]Epoxystearic acid also shows some regioselectivity of epoxide opening with sEH. Acidic hydrolysis of [ 18 O]cis-9,10epoxystearic acid results in a 1:1 distribution of carbon 9-and carbon 10-labeled diol, a testament to the chemical equivalence of the two epoxide carbons. However, as depicted in Fig. 2, 68.5% Ϯ 1.3% of the enzymatic attack occurs at carbon 10. The 18 O label distribution following enzymatic hydrolysis of [ 18 O] methyl cis-9,10-epoxystearate is identical with the distribution following enzymatic opening of its corresponding free acid, suggesting that the anchoring of the carboxylic acid or ester on the enzyme is not the driving force for the observed regioselectivity.
Single Turnover Experiments with sEH- Fig. 3 illustrates the difference in products obtained in a single turnover experiment based on the two differing mechanisms proposed. Various concentrations of sEH:substrate, ranging from 100 to 0.01: 1 molar equivalents were incubated utilizing tDPPO and cis-9,10-epoxystearic acid as substrates. The sEH was lyophilized just prior to analysis, and it was rehydrated in H 2 18 O. The products of enzymatic hydration generated at 37°C for 10 min were extracted and derivatized with BSTFA as described above. During initial single turnover experiments it was noticed that tDPPO's high rate of hydrolysis led to complete consumption of substrate by a fraction of the enzyme population and yielded inconclusive data. However, after insuring rapid mixing of diluted substrate (in H 2 18 O) at O°C with hydrated enzyme (in H 2 18 O) and subsequent incubation at 37°C, the results obtained conclusively supported a nucleophilic attack of enzyme upon tDPPO, which was followed by the hydrolysis of the resultant acyl-enzyme intermediate.
The latter result was substantiated further by single turnover hydrolysis and analysis of cis-9,10-epoxystearic acid in H 2 18 O. The K m and V max for epoxy stearic acid and tDPPO were found to be 11.0 M and 3460 nmol/min/mg protein, and 2.80 O, the resulting diol would incorporate 18 O oxygen after one turnover. Mechanism 2 presupposes that a nucleophilic amino acid such as a carboxylate anion opens the epoxide, which would lead to the formation of an ␣-hydroxyester-enzyme intermediate. Subsequently, an activated water hydrolyzes the ester, which releases the diol product, and the enzyme is regenerated. If performed in H 2 18 O, the resulting diol would not incorporate 18 O oxygen after one turnover. Acidic activation of the epoxide oxygen, which could enhance catalytic activity, is consistent with both mechanisms.
M and 26200 nmol/min/mg protein, respectively. 3 Since the oxidized fatty acid turned over more slowly than tDPPO, a smaller fraction of the enzyme population experienced multiple turnover. This hypothesis is supported by the better fit of the cis-9,10-epoxystearic acid data to what is expected theoretically as compared with the data obtained for tDPPO (Fig. 4).
From our previous observations the expected site of 18 O insertion within tDPPO is at the benzylic carbon. The amount of 18 O label incorporated in the product was monitored by GC/MS for each experiment. As can be seen in Fig. 4, equal or greater molar equivalence of sEH (i.e. Յ1 theoretical turnover) led to minimal 18 O incorporation from the H 2 18 O solvent. The same pattern was observed with cis-9,10-epoxystearic acid (Fig. 4). These data suggest 16 O incorporation from enzyme onto the epoxide leading to the formation of an acyl-enzyme intermediate. 18

O Labeling of Catalytic Amino Acid and Isolation of Labeled Tryptic Fragment-
The nucleophilic amino acid responsible for the initial attack of sEH onto epoxides was labeled with 16  O incubated sEH with base lineresolved separation of most fragments. Theoretical computer calculation of possible tryptic fragments assigned a relatively high hydrophobic score for the tryptic fragment T26, which contains Asp-333 (T26 refers to the 26th theoretical fragment obtained from the tryptic digestion of sEH by counting from the N terminus). Therefore, the last nine major peaks in the reverse phase separation of the tryptic fragments were collected (Fig. 5). Table I lists the molecular weights obtained by electrospray mass spectrometry for the nine isolated fragments for the H 2 16 O and H 2 18 O incubated enzymes, along with the theoretical masses expected for each fragment. Peak III could not be assigned because it was a combination of many coeluting substances. The mass accuracy for peak I was also poor due to the low signal-to-noise ratio, which impedes very accurate mass assignments. However, the remaining peaks were assigned very easily with a very high degree of confidence due to excellent signal-to-noise ratio. Peaks II, VI, VII, and VIII do not exhibit any increase in mass between sEH turned over in H 2 16 O versus sEH turned over in H 2 18 O. Conversely, peaks IV, V a , V b , IX a , and IX b exhibit an average of 3.6 Da increase due to the incorporation of two 18 O oxygen atoms for the multiply turned over sEH in H 2 18 O. Peak IV was identified as the tryptic fragment T26 and is the common peptide stretch among the peptides in the peaks isolated that exhibit increased mass due to 18  O and subsequent HPLC purification after tryptic digestion were hydrolyzed completely with a nonspecific protease (immobilized protease Sg). After removal of the protease and lyophilization, the residue of both samples was derivatized with MTBSTFA and analyzed by GC/MS in the single ion monitoring mode. The fragments monitored were generated by the loss of a t-butyl group ([M Ϫ 57] ϩ ϭ A ϩ ). Table II  O]T26 are very similar (all background due to isotopic impurity), demonstrating that the mass increase within T26 is solely due to the involvement of Asp-333 in catalytic action, and that Glu-348 does not participate in the catalytic mechanism. The latter result clearly implicates Asp- FIG. 4. The graph illustrates single turnover experiments performed with various ratios of sEH:substrate. The theoretical line (q-q) represents the theoretical data based on mechanism 2. The data representing single turnover experiments performed with tDPPO (E⅐⅐⅐E) and cis-9,10-epoxystearic acid (ϫ⅐⅐⅐ϫ) clearly follows the pattern dictated by mechanism 2. The theoretical line (Ç-Ç) represents the theoretical data obtained based on mechanism 1. 2 18 O. Peaks I-IX were isolated and concentrated by microbore HPLC (Ͼ100 fold) as described under "Materials and Methods." 333 (the only aspartic acid in T26) as the nucleophilic amino acid responsible for the first step in the hydration of epoxides. DISCUSSION In this study we set out to examine the mechanism of epoxide hydration by sEH. Previous publications from this and other laboratories have reported some stereo-and regiochemical controls exerted by sEH. It is generally accepted that sEH opens epoxides by nucleophilic attack (as compared with a general acid catalysis) often at the least hindered carbon in an S N 2 fashion. This is normally accomplished by the back side attack of the epoxide carbon (anti opening) (20). Therefore, the initial stereochemistry of the epoxide dictates the stereochemical outcome of the diol product. trans-1,3-Diphenylpropene oxide was synthesized as a general substrate for sEH. The asymmetry about the epoxide allows regiochemical studies of nucleophilic attack. Also, the nature of the epoxide carbons vary greatly, one being benzylic, which greatly favors nucleophilic attack, and the other being homobenzylic, which is slightly less hindered. As can be seen from Fig. 1, sEH opens tDPPO via a back side displacement of the epoxide oxygen. This is evident from the exclusive production of erythro-diol from the trans-epoxide substrate. A completely acid-catalyzed mechanism would probably result in a mixture of threo-and erythro-diols since the carbocation involved in the mechanism would be greatly stabilized by the phenyl group and would allow water to attack the carbocation from either face due to its trigonal planar geometry. However, the latter data do not preclude the possibility of an acid assisted catalysis occurring in concert with a base catalyzed attack of the epoxide. Transient carbocations can retain their tetrahedral geometry, and therefore, block one face of nucleophilic water attack, which would result in the formation of either erythro-or threo-diol (depending on the stereochemistry of the parent epoxide) and not both.

FIG. 5. Reverse phase HPLC spectra for the tryptic digestion of sEH treated with excess tDPPO in H
Next, the regioselectivity of sEH with tDPPO, cis-9,10epoxystearic acid and methyl cis-9,10-epoxystearate was investigated. The [ 18 O]epoxides of the latter substrates were synthesized as described previously. The chemically (base-and acidcatalyzed) and enzymatically generated diols were analyzed by GC/MS to determine label incorporation into each position. In the case of sEH-hydrolyzed tDPPO, 97.1% of the attack occurred at the benzylic position (Fig. 2). The same pattern was observed for the base-hydrolyzed [ 18 O]tDPPO (97.0%). Even though the homobenzylic carbon is slightly less sterically hindered, the nucleophilic attack occurs at the benzylic carbon, which can better stabilize the transition state. As expected, the acid-catalyzed hydrolysis of [ 18 O]tDPPO led to products that exclusively exhibited attack of the benzylic position due to the stability of the benzylic carbocation.
We were also able to show that sEH is able to discriminate two chemically similar epoxide carbons. The enzymatic hydrolysis of [ 18 O]cis-9,10-epoxystearic acid led to a 68.5% attack of C-10 in favor of C-9 (Fig. 2). The methyl ester of [ 18 O]cis-9,10epoxystearic acid was also subjected to sEH hydrolysis with identical results, discounting the importance of the carboxylate anion of this fatty acid being anchored by the enzyme. The observed selectivity of sEH between the two nearly identical epoxide carbons can be due to several reasons. The accessibility of the nucleophilic amino acid to C-9 versus C-10 can account for the observed data. Another possibility is the different spatial orientations by which the fatty acid can be absorbed onto the catalytic site. This could lead to a different population of C-9 versus C-10 positioned closer to the catalytic amino acid for nucleophilic attack (if one spatial orientation is favored over others). Of course, the true explanation might be a combination of the latter two hypotheses. The regioselectivity data were instrumental in the calculations of single turnover data.
After firmly establishing the regioselectivity and mode of attack with both tDPPO and cis-9,10-epoxystearic acid, the mechanism of action could be probed. Two scenarios could be envisioned. The first hypothesis, which has been the generally a The notation T followed by a number refers to the tryptic fragment obtained from the theoretical tryptic digestion of sEH by counting from the N terminus. Peaks containing multiple tryptic fragments (as indicated by hyphens) are the result of incomplete tryptic digestion and are covalently bound by peptide bonds. These are not separate peptides. The contents of each set of parentheses is one single peptide fragment consisting of one or more theoretical tryptic fragment(s).
b Number of Asp and Glu residues in each fragment. accepted theory, alleges an activated water (possibly by a histidine) delivery onto the epoxide carbon and subsequent protonation of the generated alkoxide to yield the product (Fig. 3,  mechanism 1). The second theory supposes that the side chain of a nucleophilic amino acid such as a carboxylate anion within sEH attacks and opens the epoxide. The resultant acyl-enzyme intermediate would then be hydrolyzed by an activated water (again postulated via a histidine) that would generate the native enzyme and release the diol product (Fig. 3, mechanism  2). Even though there has not been direct evidence for the second postulated mechanism, recent discoveries from the mechanism of DhlA and single turnover experiments performed with mEH by Lacourciere and Armstrong (13) raised the possibility that sEH could also follow the same mechanistic path (25,35). X-ray crystallographic study of DhlA with bound substrate has clearly shown the involvement of an aspartic acid as the nucleophile, which in the first step of enzymatic catalysis attacks the substrate leading to the formation of an acylenzyme intermediate (␣-hydroxyester-enzyme) (36). Also, single turnover experiments with mEH incubated in H 2 18 O have shown that the oxygen introduced within the epoxide is not 18 O-labeled during the first turnover. The latter data suggest that mEH also hydrolyzes epoxides via an acyl-enzyme intermediate (13). Fig. 3 illustrates the difference in products which would be obtained from a single turnover experiment performed in H 2 18 O based on the two postulated mechanisms. Since the first mechanism presupposes a direct attack of an activated water onto the epoxide, it follows that under any conditions of hydrolysis the 18 O would be incorporated within the product. Therefore, after a single turnover of sEH with tDPPO the diol product should be labeled with 18 O. However, if the hydrolysis follows the second mechanism suggested, then the oxygen incorporated during the first run of the enzyme is supplied by the enzyme. Since the amino acids of sEH contain only 16 O oxygens it is clear that the diol product obtained from the first turnover would contain only 16 O oxygen. The labeled oxygen from water would hydrolyze the acyl-enzyme intermediate and, therefore, incorporate itself within the enzyme. The enzyme would therefore contain a 1:1 16 O: 18 O labeled amino acid after one turnover. As multiple turnovers occur, the catalytic amino acid would be completely labeled with 18 O and would yield diol products identical to the first suggested mechanism, hence the importance of performing single turnover experiments.
As indicated from the data in Fig. 4, the lack of 18 O within the diol product for equimolar or greater sEH:substrate refutes the direct incorporation of 18 O from H 2 18 O into the epoxide. Care must be taken in such an experiment to insure each enzyme does not turn over more than one substrate by adding diluted substrate and mixing enzyme and substrate at cold temperatures. This is particularly important for substrates turned over with a high k cat such as tDPPO. Initially our results were inconclusive since we added concentrated substrate to the enzyme buffer solution. Based on our initial results it is safe to assume most of the substrate was consumed as the mixing occurred within pockets of high substrate concentration. This would lead to a population of enzyme experiencing multiple turnovers. Proper mixing resulted in the data in Fig.  4, which are consistent with expected results of the second mechanism involving the intermediary of an acyl-enzyme.
Results of single turnover experiments conclusively showed the involvement of a nucleophilic amino acid within sEH responsible for the di- 16 O oxygen observed in the product that had been obtained by hydrolysis of substrate in H 2 18 O. With the recent investigation into the homology of epoxide hydrolase genetic sequences (23,24) and their prob-able evolutionary connection with the ␣/␤ hydrolase fold family of enzymes (26), a conserved region containing an aspartic acid at position 333 of sEH has been identified. Consequently, the probable catalytic Asp-333 could be labeled with 18  O-treated groups exhibit molecular weights identical to those predicted theoretically. These observations indicate the absence of any labeled oxygen within those fragments. The difference between fragments VI and VIII is 2 Da, which could be attributed to disulfide bridge formation within T16. These data suggest that in the absence of specific catalysis there is no rapid exchange of water into carboxylic acid functionalities under the conditions used here.
As can be seen from Otreated sEH increased 3.5 Da, therefore implicating T26 as the fragment containing the catalytic amino acid. Peaks IV, V a , V b , IX a , and IX b , which also contain the labeled oxygen are the result of incomplete tryptic digestion upstream and downstream from T26. The identity of T26 was further confirmed by amino acid sequencing of its N terminus (not shown), which clearly supported our assignment of the fragment. The reason an increase of 3.6 Da is observed as opposed to 4.0 Da (from the incorporation of two 18 O atoms) might be attributed to experimental error in molecular mass determination. However, close examination of the MaxEnt-transformed spectra reveals that the peak widths are broader for the 18 O-labeled peptides than for the unlabeled peptides. This suggests the presence of a small amount of peptide that contains only a single 18 O, and this would account for the molecular weights being slightly less than expected for incorporation of two 18  The presence of two 18 O units within T26 clearly indicates the involvement of an amino acid in the delivery and opening of tDPPO. The failure to see any increase in mass in peaks II, VI, VII, and VIII and the increase in 4 Da rather that 8 Da or more in peaks IV, V a , V b , IX a , and IX b support the hypothesis that there is minimal exchange of H 2 18 O with the carboxylic acids of aspartic and glutamic acids unless there is a catalytic involvement. Single turnover experiments with sEH preincubated at 37°C in H 2 18 O for 1 h yielded the exact same results as the normal single turnover experiment, discounting any catalytic exchange of 18 O into the catalytic amino acid without the presence of substrate. The cDNAs and predicted peptide sequence of T26 contains Asp-333 as the only aspartic acid within the fragment. As discussed previously, Asp-333 has been implicated as the catalytic residue by sequence homology to other EHs, and the ␣/␤ fold hydrolase family of enzymes. Based on the latter data, it seems very plausible that Asp-333 is the catalytic amino acid responsible for the initial step of hydrolysis. However, T26 contains a glutamic acid (Glu-348), which is not conserved but could conceivably be responsible for catalytic activity.
The complete digestion of T26 (from both H 2 16 O and H 2 18 O treated sEH) was achieved with a nonspecific protease (immobilized protease Sg). After derivatization of the hydrolyzed amino acid residues with MTBSTFA the samples were analyzed by GC/MS. Since there is only one aspartic acid in T26, and since the aspartic acid derivative obtained from the hydrolysis of the [H 2 18 O]T26 exhibited an increased mass (Table  II), it seems unequivocal that the aspartic acid in T26 is responsible for the nucleophilic attack onto tDPPO. Furthermore, it was shown that Glu-348 (the only other carboxylate containing amino acid in T26) did not incorporate any 18 O oxygen. The attack of Asp-333 initiates enzymatic activity, leading to the formation of an ␣-hydroxyester-enzyme intermediate. Hydrolysis of this acyl-enzyme is accomplished by the addition of an activated water to the carbonyl carbon of the ester bond, after which the resultant tetrahedral intermediate collapses, yielding the active enzyme and the diol product. The second step of the latter mechanism resembles the mechanism of most serine esterases and proteases.
A push-pull mechanism, where the epoxide oxygen is activated by protonation or hydrogen bonding that would weaken the C-O epoxide bond seems likely since the carboxylate anion (Asp-333) is the attacking nucleophile. The latter hypothesis fits well with our knowledge of carboxylate chemistry. Chemically, carboxylates are not considered strong nucleophiles that could easily open epoxide rings. However, the epoxide carbon can be chemically activated toward nucleophilic attack by weak nucleophiles such as carboxylate anions through the coordination of the epoxide oxygen with a lewis acid (37). Within the catalytic cavity of sEH, the epoxide oxygen can be activated by various amino acids through either proton donation or hydrogen bonding. The x-ray crystallographic solution of DhlA cocrystallized with substrate clearly demonstrated that two tryptophan residues are responsible for polarization of the halideleaving group (36). Genetic sequence homology between DhlA and sEH investigated by Arand et al. (23) has shown that one of the two tryptophans in DhlA is conserved in sEH. This latter postulated tryptophan (Trp-334) could activate the epoxide by hydrogen bonding with the oxirane oxygen, but it is not necessarily the only mechanism for epoxide activation by sEH. A push-pull mechanism has also been suggested previously based on the general structure of sEH inhibitors (34,38,39). The general structure of sEH inhibitors contains a carbonyl or hydroxy ␣ to the epoxide. It has been suggested that the presence of the ␣-hydroxy/carbonyl functionality interferes with the acidic activation of the epoxide by hydrogen bonding with the proton donor (17). However, in light of a two-step mechanism of action a detailed study is necessary to elucidate the inhibitory mechanism.
In conclusion, we report that we have been able to conclusively show that sEH effects hydrolysis of epoxides via an A2 type, back side nucleophilic attack of Asp-333. An isolatable acyl-enzyme intermediate is formed, which is hydrolyzed by an activated water, resulting in the regeneration of the active enzyme and the release of the diol product.