Biochemical Evidence for the Involvement of Tyrosine in Epoxide Activation during the Catalytic Cycle of Epoxide Hydrolase*

Epoxide hydrolases (EH) catalyze the hydrolysis of epoxides and arene oxides to their corresponding diols. The crystal structure of murine soluble EH suggests that Tyr465and Tyr381 act as acid catalysts, activating the epoxide ring and facilitating the formation of a covalent intermediate between the epoxide and the enzyme. To explore the role of these two residues, mutant enzymes were produced and the mechanism of action was analyzed. Enzyme assays on a series of substrates confirm that both Tyr465 and Tyr381 are required for full catalytic activity. The kinetics of chalcone oxide hydrolysis show that mutation of Tyr465 and Tyr381 decreases the rate of binding and the formation of an intermediate, suggesting that both tyrosines polarize the epoxide moiety to facilitate ring opening. These two tyrosines are, however, not implicated in the hydrolysis of the covalent intermediate. Sequence comparisons showed that Tyr465 is conserved in microsomal EHs. The substitution of analogous Tyr374 with phenylalanine in the human microsomal EH dramatically decreases the rate of hydrolysis ofcis-stilbene oxide. These results suggest that these tyrosines perform a significant mechanistic role in the substrate activation by EHs.

Epoxide hydrolases (EH, 1 EC3.3.2.3) hydrolyze epoxides and arene oxides to their corresponding diols (1). These enzymes are widely distributed among many species, including bacteria, fungi, plants, insects, and mammals (2)(3)(4)(5). In mammals, there are two major classes of EH with broad and complementary substrate selectivity, soluble EH (sEH) and microsomal EH (mEH) (6). sEH participates not only in xenobiotic detoxification but also endogenous lipid metabolism, acting on epoxides of linoleic acid (leukotoxin and isoleukotoxin) (7) and arachidonic acid (cis-epoxyeicosatrienoic acids) (8). Elevated titers of linoleate and arachidonic acid diols are, respectively, thought to be associated with the inflammatory disorder known as acute respiratory distress syndrome (7) and pregnancy-induced hypertension (9). Inhibition of epoxide hydration may accordingly have a therapeutic value for these two serious disorders.
Alternately, mEH appears to be mainly involved in the metabolism of xenobiotic epoxides (6). A protein-reactive and cytotoxic epoxide, naphthalene epoxide for example, is converted to the less toxic diol by this enzyme (10). The mEH is also related to activation of other arene oxides such as 7,12-dimethylbenzanthracene, a member of the polycyclic aromatic hydrocarbon class of chemical carcinogens (11). To understand xenobiotic toxicity, metabolic aberrations associated with pathological disorders, and to develop possible therapies against these, it is important to elucidate the molecular basis of EH catalysis.
The EHs belong to the ␣/␤ hydrolase fold family. These enzymes characteristically employ a two-step mechanism in which a catalytic nucleophile of the enzymes attacks a polarized electrophilic substrate, and the covalent intermediate is subsequently hydrolyzed (Fig. 1) (6). The mechanism of murine sEH has been mainly elucidated from a series of experiments utilizing heavy isotopes, protein mass spectrometry, and sitedirected mutagenesis (12,13). They indicated that Asp 333 acts as a catalytic nucleophile and that a water molecule is activated by the nearby His 523 and Asp 495 pair (Fig. 1). This mechanism was extended to other EHs (2, 14 -18). However, one or more additional amino acids are likely involved in the catalytic cycle especially in the activation of the epoxide ring (5), based on the mechanism of the haloalkane dehalogenase, HLD1, from Xantobacter autotrophicus G10, a related ␣/␤ hydrolase (6). Kinetics of the hydrolysis of chalcone oxide by sEH support this hypothesis. Thus a generalized scheme is postulated in which one or more amino acid(s) may polarize the epoxide oxygen by an acid-like mechanism, weakening the C-O epoxide bond and facilitating the attack on the carbon of the epoxide ring by a nucleophile, such as the conjugate base, Asp 333 (19). Recently, the crystal structure of murine sEH has been determined at 2.8 Å resolution (20). The structure supports the previously proposed mechanism and suggests that Tyr 465 and Tyr 381 are the possible acid catalysts that activate the epoxide ring (Fig. 1). Additionally, analysis of the crystal structure of EH from Agrobacterium radiobacter, AD1, likewise suggested that Tyr 152 -Tyr 215 in this prokaryotic enzyme may have the same function (21). Thus, biochemical evidence is required to verify the mechanism of epoxide activation and to complement structure-based prediction.
In this study, site-directed mutagenesis was utilized to explore the role of Tyr 465  of murine sEH. The sEH variants were subjected to a series of enzyme assays and kinetic studies to assess their impact on epoxide activation. Additionally, this work was extended to Tyr 374 of human mEH. Interpreted in view of the EH crystal structure, these results clearly support the role of active site tyrosine residues in epoxide activation by eukaryotic sEH and mEH.
Expression of Wild-type, Mutant Murine sEH, and Human mEH-Enzymes were produced using BAC-TO-BAC Baculovirus Expression System (Life Technologies, Inc.). Cells from Trichoplusia ni (T. ni High5) (500 ml, 5 ϫ 10 5 cells/ml) were infected with virus solution at a multiplicity of infection of 0.1. Three days postinfection, the cells were resuspended in 20 ml of 0.1 M sodium phosphate buffer (pH 7.4) (buffer A) containing 1 mM phenylmethylsulfonyl fluoride, EDTA, and dithiothreitol, and homogenized with a Polytron. The crude extract was centrifuged at 12,000 ϫ g for 20 min. The supernatant was centrifuged again at 100,000 ϫ g for 1 h. For murine sEH, the resulting supernatant (cytosol fraction) was stored at Ϫ80°C, and for human mEH, the pellet (microsomal fraction) was resuspended in 3 ml of buffer A and stored at Ϫ80°C.
Protein Analysis-Protein concentrations were determined with the Pierce BCA assay (Pierce) using bovine serum albumin as a standard. SDS-polyacrylamide gel electrophoresis was performed according to Laemmli (28) using a 10% resolving gel. Assays for murine sEH, employing Western blot techniques, were performed using the correspond-ing polyclonal antibody as described previously (12). The polyclonal antibody for human mEH was a gift of Drs. Franz Oesch and Michael Arand (Mainz, Germany). Estimation of protein bands was carried out with the Scion Image software package (Scion, Frederick, MD).
Enzyme Assay-The sEH enzyme activity of the sEH wild type and its mutants was determined using five different substrates (compounds 1-5). These assays were performed as described previously: NEP2C (1) (19), tDPPO (2) (23), tSO (3) (24), [ 14 C]ESA (4) (23), and JH III (5) (23,29). The assays were run at 30°C for 1-30 min depending on the substrate and the mutant used. To compare specific activities between wild-type and the mutant enzymes, identical concentrations of enzymes were used as judged by Western blot ([E] ϭ 80 nM for 1, 3, 4, 5 and 3 nM for 2). For the human microsomal EH, cSO (compound 16) was used as substrate. Assay was performed as described (24) using 1 g of the microsomal fraction. The wild-type enzyme was incubated at 30°C for 1 min, while the Y374F mutant was incubated for 30 min. Cytosol fractions of control cells have no detectable epoxide hydrolase activities above background (30).
Kinetic Assay Conditions-The kinetic constants were determined using the two-step inhibition model previously described for EH inhibition, where chalcone oxides and NEP2C were used as substrate (19). Inhibitor concentrations between 0 and 2.5 M were used for the wildtype enzyme, whereas concentrations between 0 and 50 M were used for the two mutant enzymes. To achieve homogeneity in the results and to be able to compare them between the wild-type and the two mutants, Y381F and Y465F, identical concentrations of enzymes (80 nM) were used as determined by Western blot. Measurements were made in the presence of excess of substrate ([S] ϭ 40 M). Activity measurements at 0 -30 s post inhibitor introduction allow determination of the initial rate of enzyme inhibitor complex formation () as described previously (19). Under these conditions, less than 10% of the inhibitor was bound to the enzyme. The enzyme-inhibitor dissociation constant (K d ) and the rate of enzyme-inhibitor covalent complex formation (k 2 ) were calculated from plots of versus [I] (19). The rates of dealkylation and product release (k 3 ) were calculated from enzyme activities in 1:1 enzyme:inhibitor mixtures, at times varying between 1 and 20 min post inhibitor introduction as described previously (19). The decomposition rate of the inhibitors in the absence of the enzyme was found to be negligible (Ͻ1%) over the period studied. The accuracy of the applied kinetic model was assessed by comparing theoretical and measured results for , producing correlation factors greater than 0.95. Based on experimental inhibitor and substrate concentrations, a maximal error in observed K d of ϳ5% can be attributed to incomplete dissociation of EI. Moreover, the k 2 and k 3 values obtained indicated a 7% maximal error in k 2 because of the EC degradation.
IC 50 Assay-IC 50 values were determined using tDPPO (compound 2) as a substrate. Enzymes (3 nM) were incubated with inhibitors for 5 min at 30°C prior to substrate introduction ([S] ϭ 50 M). Assays were run for 1 min for wild-type and were extended to 15 min for Y381F and Y465F. IC 50 was calculated by linear regression of five datum points with a minimum of two points of either side of IC 50 . The results were generated from at least three separate runs each in triplicate to obtain the standard deviation.
Graphic Analysis-Images used in this study were rendered and displayed in the Swiss-Pdb Viewer (31-33) using coordinates supplied by Argiriadi et al. (26). Additional graphic enhancements were performed in the Photoshop graphic programming environment.
FIG. 1. Two step mechanism of murine sEH. In the first step, a reactive nucleophile Asp 333 attacks the epoxide ring leading to the ring opening and the formation of hydroxyalkyl enzyme. In the second step, the covalent intermediate is hydrolyzed by a water molecule activated by a His 523 -Asp 495 pair. Tyr 381 and Tyr 465 are proposed to polarize the epoxide moiety to facilitate the attack by nucleophilic Asp 333 (20,26).

RESULTS
Expression and Activities of Tyrosine Mutants of Murine sEH-To explore the role of each of the tyrosines of murine sEH, we prepared four different mutant constructs, Y381F, Y465F, Y465A, and Y381F/Y465F. The enzymes were subsequently produced using baculovirus expression system as described under "Experimental Procedures." In the purification of murine sEH, the ligand ( Fig. 2A) interacts with the putative hydrophobic pocket near the active site (12,34). Wild-type and mutant protein were detected at similar expression level in the cytosol fraction and not detected in the flow through fraction (Fig. 2C) indicating that the mutant enzyme probably maintained its structural integrity. The mutant enzymes were eluted in low yield with 1 mM 4-fluorochalcone oxide (FCO), a selective inhibitor for sEH, indicating probable lower affinity for the inhibitor (Fig. 2, D-E). Total protein eluted by FCO followed by SDS (Fig. 2E) was similar for both wild and mutant enzymes, indicating same overall binding for each.
The specific activities of the wild-type and mutant enzymes in the hydrolysis of five substrates are summarized in Table I. These substrates are classified, and ordered from the more reactive (NEP2C (compound 1)) to the less reactive (JH III (compound 5)). The specific activities of wild-type enzyme were similar to previously published results (19,23). NEP2C and JH III are known to be hydrolyzed by esterases; however, the esterase-dependent hydrolysis was not detectable in these experiments (data not shown). In mutant enzymes, the activities were lower than those of the wild-type. Substrates containing less reactive epoxides resulted in larger changes in activity between the wild-type and the mutant enzymes, suggesting that hydrolysis of less reactive epoxides is dependent on activation by tyrosine(s). The Y381F/Y465F mutant demonstrated no detectable catalytic activity with any of the substrates used. These results suggest that both Tyr 381 and Tyr 465 are required for full activity of the murine sEH. This is consistent with the results of analogous experiments with bacterial EH showing that active site tyrosines Tyr 152 and Tyr 215 are important for catalysis (35).
Kinetics of Inhibition-To further investigate the role of the two residues Tyr 381 and Tyr 465 in the catalytic mechanism, we determined the kinetic constants of their inhibition by chalcone oxides. Chalcone oxides are in fact poor EH substrates that inhibit the enzyme by forming a stable covalent intermediate (19). Their action is described by the following equation.
The nonlinear regression of the initial rate of EC formation () versus inhibitor concentration ([I]) permits the calculation of K d and k 2 (19). The slope of the EC complex decomposition, ln(A 0 Ϫ A) versus time, permits the calculation of k 3 (19). Derived kinetic constants are listed in Table II. Both mutations of Tyr 465 and of Tyr 381 by phenylalanine resulted in increases in the K d values up to 50-fold, with the exception of compound 7 and the Tyr 381 mutant (Table II). These results demonstrate that both tyrosines are implicated in the binding of the substrate. The change in K d is well correlated (r 2 ϭ 0.91) with the values of the para-substitution of chalcone oxides for Y465F, whereas no relationships were found for Y381F. The observed increased K d in mutant enzymes reflect the smaller interaction between the enzyme and the inhibitor. The linear relation observed between K d and for Y465F highlights the possible bond between Tyr 465 and epoxide moiety, as stated (26).
The rate of alkylation (k 2 ) is also greatly altered in both mutations (Table II). The values of k 2 for Y465F decreased increasingly from 2-to 30-fold for compounds 11 to 6, respectively. For Y381F such trends in the decreased k 2 were observed only from compound 9 to 6, whereas compounds 11 to 9 showed a 7-fold decrease in their k 2 values. The Hammett plot (Fig. 3) showing the logarithm of the relative rates ( 4ϪR k 2 / 4Ϫ H k 2 ) versus the constant ϩ for each 4-position substituent illustrates this influence. These results were correlated with ϩ for the three enzymes, indicating that development of a relative positive charge at the reactive center is important for the alkylation step. For the wild-type enzyme, a linear relationship (r 2 ϭ 0.93) with a slope of Ϫ0.56 was obtained, indicating a push-pull mechanism, in which the epoxide oxygen is activated by protonation facilitating a nucleophilic attack on the carbon of the epoxide ring by the Asp 333 carboxylate anion (19). A linear relationship (r 2 ϭ 0.91) was obtained for Y465F, indicating a consistent mechanism operating throughout the chalcone oxide series. However, the sign of the slope (ϩ 0.27) is inverted from the slope of the of the wild-type enzyme, indicating a different mechanism. This value is very close to the value (ϩ0.32) found for a general basic mechanism of opening of similar epoxides (36). This result strongly suggests that a simple nucleophilic mechanism is implied in the action of Y465F and that Tyr 465 is directly related to the polarization of the epoxide moiety. A bell shaped relationship was obtained for Y381F, indicating a change of mechanism operating throughout the chalcone oxide series. A slope (ϩ0.26) similar to one of Y465F is obtained for the electron donating para-substitutions (6 to 9), whereas a slope (Ϫ0.51) similar to the one for wild type is obtained for electron withdrawing para-substitutions (9 to 11). Therefore, Tyr 381 participates in the polarization of the epoxide moiety. However, its role is less clear than that of Tyr 465 .
The rate of dealkylation k 3 was less influence by either of the tyrosine mutations; the decreased k 3 between 1-and 3-fold are observed for Y465F and Y381F compared with the wild type  (Table II). Moreover, no relation was found between the intensity of the change in k 3 and the nature of the para-substitution for Y465F and Y381F. These results indicate that neither tyrosine directly influences the hydrolysis of the enzyme-inhibitor covalent intermediate. This agrees with prediction from the crystal structure (20). Additionally, changes in specific activities for compounds 1-5 probably came from the changes in K d and k 2 , because k 3 was unchanged for compounds 6-11.

Inhibition of Tyrosine Mutants by Ureas and Carbamate-
The crystal structure of sEH-N-cyclohexyl-NЈ-decyl urea (compound 14) complex shows that Tyr 465 and Tyr 381 provide hydrogen bond interaction with the carbonyl group of the urea (26). To evaluate this apparent binding trend with the tyrosines, IC 50 values were determined for wild type and for both tyrosine mutants (Table III). Substitution of Tyr 381 by phenylalanine resulted in an enzyme with 8 -88-fold higher IC 50 values for compounds tested. Replacement of Tyr 465 by phenylalanine increased IC 50 by 2-13-fold. These results suggest that both Tyr 381 and Tyr 465 interact with these inhibitors. Particularly Tyr 381 seems more important than Tyr 465 for binding with compounds 13 and 14. Tyr 465 may contribute less to the free energy of binding for these inhibitors than Tyr 381 .
Mutagenesis of Tyr 374 in Human mEH-Although human mEH has 21% sequence identity with murine sEH (5), the catalytic triad (Asp 226 , Glu 404 , and His 431 ) is conserved when compared with sEH (6). Sequence alignment shows that Tyr 374 in human mEH seems to be analogous to Tyr 465 in murine sEH (2). The tyrosine is conserved in all sequenced mEHs (20,21). As shown in Fig. 4A, the linear distance between these amino acids is similar in the primary structures of murine sEH and human mEH. Tyr 283 or Tyr 291 in human mEH could correspond to Tyr 381 in murine sEH. However, the tyrosines are not conserved in all mEHs (20,21). Thus only the Tyr 374 mutant was made to explore the role of the residue in this enzyme. As shown in Fig. 4B, wild-type and Y374F enzymes were expressed in the microsomal fraction at similar levels. Specific activity was determined for the hydrolysis of cis-stilbene oxide (compound 16), which is selective for mEH. Compared with wild-type enzyme, the Y374F mutant had dramatically decreased activity (Fig. 4D). The recent crystal structure of EH from Aspergillus niger, a soluble member of mEH, showed that Tyr 314 likely plays a significant role in epoxide activation (37). The authors suggested that Tyr 314 appears to be equivalent to Tyr 374 in human mEH based on the sequence comparison between these two enzymes. Taken together, mutation of Tyr 374 likely affects polarization of epoxide moiety rather than some other step in the catalytic cycle. This result suggests that involvement of tyrosine in the catalytic mechanism is common in both sEH and mEH. DISCUSSION In the present study, we examined the role of the tyrosines in the catalytic mechanism of EH. It had been proposed that the epoxide oxygen might be activated by a possible acid catalyst to TABLE I Specific activities of wild-type and mutant murine sEH All assays were performed on the cytosol fraction from baculovirus-infected T. ni cells. The reported results are the mean Ϯ S.D. (n ϭ 3). Levels of sEH protein were apparently identical in all assays based on Western blot. a n.d., nondetectable. weaken the C-O bond of the epoxide ring and facilitate the attack by a nucleophilic aspartate (6). However, a variety of early studies showed that the reaction intermediate lacked full carbocation character (reviewed in Ref. 6). Based on the crystal structure of murine sEH (20,26), we tested the hypothesis that Tyr 465 and Tyr 381 are involved in epoxide activation by examining the mutant enzymes.
The substitution of tyrosine with phenylalanine is not expected to create a significant conformational change, even locally, because the body of the tyrosine and phenylalanine are both partially buried and unchanged, whereas the changed region, the phenolic hydroxyl group, is in the solvent accessible region of the active site. We tested the hypothesis by showing that all mutant enzymes bind to the active site-directed affinity column with similar efficiency. This would not be expected if mutation caused a major alteration in the conformation of the enzyme. Our results are consistent with the prediction. The sEH variants bound on the affinity column were eluted by 1 mM FCO with lower recovery than the wild type. Because the affinity ligand is thought to interact with the putative hydrophobic tunnel close to the active site (34), the lower recovery in sEH variants might result from weaker competition with FCO (Fig. 2E). The larger K d found for Y465F and Y381F in compound 8 (Table II) supports the possibility. Based on the results in Table II, one could expect compound 7 to give better elution; however, the decreased solubility overcame increased affinity for the enzymes. Such an effect also was observed for the wild-type enzyme (34). Y381F/Y465F was eluted by 1 mM FCO with relatively higher yield than other mutants. In this case, the loss of two hydroxyl groups may be contributing to a weaker interaction between the enzyme and affinity ligand. No change was seen in the mobility of the wild-type and mutant enzymes on isoelectric focusing (data not shown). Taken together these data suggest that the mutations do not cause dramatic conformational changes in the active site. We hypothesize that the decreased specific activities in the mutants (Table I) (Fig. 5A). The binding nature suggests that the inhibitor mimics the transition state for epoxide ring opening in sEH catalysis (26). Our results are consistent with this structurebased prediction. Both mutations of the two tyrosines increased IC 50 values for ureas tested (Table III) Table II were plotted versus electronic parameter ϩ . The slope of the lines indicates the type of mechanism involved and its intensity.  Tyr 381 , and the substrate epoxide oxygen modeled into the sEH active site (26) is quite consistent with the preferred stereochemistry of such interactions reflected by the examples from the Cambridge Structural Data base. Epoxide-phenol hydrogen bond stereochemistry is also consistent with the preferred stereochemistry of hydrogen bonds to tyrosines in refined protein structures as outlined by Ippolito et al. (39). In the crystal structure of murine sEH, Tyr 465 is flanked by edge to face interactions with Trp 334 and Phe 385 (Fig. 5C). The interaction is thought to stabilize phenolate anion in the transition state (20). Similar interaction is found in the recent crystal structures of bacterial EH and fungus EH (21,37). Our biochemical results highlight the significance of Tyr 465 in the epoxide activation.
Our preliminary results suggest that Tyr 374 in human mEH is analogous to Tyr 465 in murine sEH. The tyrosine is absolutely conserved in all EHs sequenced (20,21). Although it is not known if Tyr 374 forms edge to face interaction with other aromatic amino acids, sequence alignment shows that Trp 227 and Phe 290 in human mEH likely correspond to Trp 334 and Phe 385 in murine sEH. Additional experiments will be required to address the question if a corresponding second tyrosine exists for mEH. Tyr 283 and Tyr 291 in human mEH may be equivalent to Tyr 381 in murine sEH on the basis of distance comparison (Fig. 4A). However, both tyrosines are not conserved in all mEHs (20,21). Tyr 299 was suggested to be a possible second tyrosine in human mEH, based on the sequence comparison with fungus sEH. This fungus sEH belongs to the same class as mammalian mEH, although the sequence relationships were weak (37). There is another possibility that mEH may have only one tyrosine for activation. Genetic relationships suggest that mEH and sEH likely have diverged from common ancestor (5). Both enzymes are distantly related and have the same major amino acid side chain involved in hydrating epoxides. However, they are separated by a vast evolutionary distance (5) and have different substrate preference; sEH has an activity on less reactive epoxides such as arachidonic acid epoxides, and mEH on reactive epoxides such as arene oxides. The sEH is more active on epoxide substrates on linear system as one would anticipate from its active site being a tunnel rather than a groove (20).
EHs are members of the ␣/␤ hydrolase fold family-like esterases and haloalkane dehalogenases (5,40). EHs and esterases have a common mechanism in hydrating substrates but a different nucleophile, aspartate and serine, respectively. The D333S mutant of murine sEH has no activity on epoxides (12). It was hypothesized that the mutant enzyme might acquire esterase activity with serine as a nucleophile. However, it had no hydrolytic activity on several esters. 2 The ester will be trigonal moving to tetrahedral as a transition state and transient intermediate interacting with NH moiety of two glycines and one alanine (41,42). In contrast, the epoxide will be tetrahedral in the transition state interacting with OH moiety of two tyrosines. The absence of esterase activity in D333S can be explained in part by having an incorrect angle of the activating groups. Genetic relationships reveal that haloalkane dehalogenase and the sEH C-terminal catalytic domain are more related and might have been diverged from common ancestor (5). Both enzymes have a common nucleophilic aspartate (6). In haloalkane dehalogenase HLD1, however, two tryptophans (Trp 125 and Trp 175 ) are responsible for the activation of the halideleaving group (43). As shown in Fig. 5A, Trp 334 and Trp 524 conserved in the EH family are located near the active site in murine sEH. In the past, these two residues were proposed to activate the epoxide ring (14,15). However, the orientation, angle, and distance are not suitable to activate the substrate epoxide, supporting the previous mutagenesis results of these tryptophans (14,15). In addition, an epoxide likely needs more activation than a halide group does. These data suggest the possibility that EHs have acquired a unique mechanism for substrate activation through evolution. Alternately, lysine was 2 B. D. Hammock, unpublished data.  (38). The epoxide group of each pair is superimposed onto the reference epoxide shown while allowing the hydroxyl oxygen position to "ride" into place (light blue spheres). The hydroxyl oxygens of Tyr 465 and Tyr 381 in a model of the sEH-substrate complex are also superimposed (green spheres) and are well within the expected range of optimal hydrogen bond stereochemistry reflected by the examples from the data base. C, amino acid polarizing the substrate. Tyr 465 is flanked by edge-to-face interactions with Phe 385 and Trp 334 that would stabilize the intermediate phenolate anion (20). previously postulated to be a proton-donating group (2,5,44). This residue is, however, not present in the active site (Fig. 5A).
In conclusion, we demonstrated the involvement of tyrosine in the activation of epoxide in the catalytic cycle of EH. The activation mechanism is apparently conserved within EH family and is likely unique in the ␣/␤ hydrolase fold family.