Mutation of Tyrosine 383 in Leukotriene A4Hydrolase Allows Conversion of Leukotriene A4 into 5S,6S-Dihydroxy-7,9-trans-11,14-cis-eicosatetraenoic Acid

Leukotriene A4 hydrolase is a bifunctional zinc metalloenzyme that catalyzes the final step in the biosynthesis of the proinflammatory mediator leukotriene B4. In previous studies with site-directed mutagenesis on mouse leukotriene A4 hydrolase, we have identified Tyr-383 as a catalytic amino acid involved in the peptidase reaction. Further characterization of the mutants in position 383 revealed that [Y383H], [Y383F], and [Y383Q] leukotriene A4hydrolases catalyzed hydrolysis of leukotriene A4 into a novel enzymatic metabolite. From analysis by high performance liquid chromatography, gas chromatography/mass spectrometry of material generated in the presence of H2 16O or H2 18O, steric analysis of the hydroxyl groups, treatment with soybean lipoxygenase, and comparison with a synthetic standard, the novel metabolite was assigned the structure 5S,6S-dihydroxy-7,9-trans-11,14-cis-eicosatetraenoic acid (5S,6S-DHETE). The kinetic parameters for the formation of 5S,6S-DHETE and leukotriene B4 were found to be similar. Also, both activities were susceptible to suicide inactivation and were equally sensitive to inhibition by bestatin. Moreover, from the stereochemical configuration of the vicinal diol, it could be inferred that 5S,6S-DHETE is formed via an SN1 mechanism involving a carbocation intermediate, which in turn indicates that enzymatic hydrolysis of leukotriene A4 into leukotriene B4 follows the same mechanism. Inasmuch as soluble epoxide hydrolase utilizes leukotriene A4 as substrate to produce 5S,6R-DHETE, our results also suggest a functional relationship between leukotriene A4 hydrolase and xenobiotic epoxide hydrolases.

Leukotriene A 4 hydrolase is a bifunctional zinc metalloenzyme that catalyzes the final step in the biosynthesis of the proinflammatory mediator leukotriene B 4 . In previous studies with site-directed mutagenesis on mouse leukotriene A 4 16 O or H 2 18 O, steric analysis of the hydroxyl groups, treatment with soybean lipoxygenase, and comparison with a synthetic standard, the novel metabolite was assigned the structure 5S,6S-dihydroxy-7,9-trans-11,14-cis-eicosatetraenoic acid (5S,6S-DHETE). The kinetic parameters for the formation of 5S,6S-DHETE and leukotriene B 4 were found to be similar. Also, both activities were susceptible to suicide inactivation and were equally sensitive to inhibition by bestatin. Moreover, from the stereochemical configuration of the vicinal diol, it could be inferred that 5S,6S-DHETE is formed via an S N 1 mechanism involving a carbocation intermediate, which in turn indicates that enzymatic hydrolysis of leukotriene A 4 into leukotriene B 4 follows the same mechanism. Inasmuch as soluble epoxide hydrolase utilizes leukotriene A 4 as substrate to produce 5S,6R-DHETE, our results also suggest a functional relationship between leukotriene A 4 hydrolase and xenobiotic epoxide hydrolases.
Sequence comparisons of LTA 4 hydrolase with other zinc metalloenzymes, e.g. aminopeptidase M and thermolysin, led to the identification of a zinc-binding motif (3,4). Subsequently, the enzyme was shown to contain one catalytic zinc atom complexed to His-295, His-299, and Glu-318 (5)(6)(7). Furthermore, LTA 4 hydrolase was found to exhibit an aminopeptidase activity toward synthetic substrates (6,8). Although a physiological substrate has not yet been found, LTA 4 hydrolase has been shown to efficiently hydrolyze several arginyl tri-and dipeptides (9).
From data obtained by site-directed mutagenesis and biochemical analysis of purified recombinant proteins, Glu-296, a residue conserved within the zinc-binding motif, was shown to be catalytic in the peptidase reaction, where it presumably acts as a general base (10). Such a reaction mechanism postulates a proton transfer to the nitrogen of the peptide bond (11,12). Further sequence comparisons between LTA 4 hydrolase and aminopeptidase M led to the identification of a conserved proton donor motif, and tyrosine 383 was suggested as the putative proton donor in LTA 4 hydrolase (13). This finding prompted us to investigate the role of Tyr-383 for the two catalytic activities of LTA 4 hydrolase. Thus, the Tyr residue was exchanged for a Phe, His, or Gln residue in mouse LTA 4 hydrolase by site-directed mutagenesis (14). The resulting proteins, [Y383F], [Y383H], and [Y383Q]LTA 4 hydrolases, were expressed in Escherichia coli, purified to apparent homogeneity, and assayed for both activities. The results from this study showed that Tyr-383 is important for the peptidase activity, in which it may act as a proton donor. In contrast, Tyr-383 was not critical for the epoxide hydrolase activity, i.e. the conversion of LTA 4 into LTB 4 . However, a more detailed characterization of these mutants revealed an additional enzyme activity for the substrate LTA 4 , as described in the present report. Thus, mutants in position 383, in particular [Y383Q]LTA 4 hydrolase, converted LTA 4 into 5S,6S-DHETE, the stereochemistry of which implies an S N 1 mechanism involving a carbocation intermediate in its formation. Since the mutants produced both 5S,6S-DHETE and LTB 4 , this result indicates that the same mechanism applies to the enzymatic hydrolysis of LTA 4 into LTB 4 . Inasmuch as soluble epoxide hydrolase accepts LTA 4 as substrate and converts it into an epimeric 5,6-DHETE, our data also suggest a functional link between LTA 4 hydrolase and xenobiotic soluble epoxide hydrolase.
Construction of Mutants-The mutants were produced by site-directed mutagenesis of the recombinant plasmid pULTA4, an E. coli expression vector containing the entire coding sequence for mouse LTA 4 hydrolase (15). The mutagenesis was carried out by a method based on polymerase chain reaction, and as described previously, the entire cDNAs of all mutants were sequenced to confirm that no alterations other than the desired mutations had occurred (14).
Expression and Purification of Recombinant Enzymes-The mutant proteins were expressed in E. coli JM101 cells in M9 medium at ϩ37°C. After isopropyl-␤-D-thiogalactopyranoside induction, the incubation was continued for 2-3 h, and the cells were collected by centrifugation and purified by fast protein liquid chromatography as described (16). The purity of the final enzyme preparations was assessed by SDSpolyacrylamide gel electrophoresis and Western blot analysis, which revealed essentially homogeneous proteins. The protein concentration was determined by the Bradford method (17) using bovine serum albumin as standard.
Incubations-Typically, aliquots of 10 g of purified [Y383F], [Y383H], and [Y383Q]LTA 4 hydrolases in 100 l of 50 mM Tris-Cl, pH 8.0, were incubated with the substrate LTA 4 (5-100 M) at room temperature. After 15 s, the reaction was stopped by the addition of 2 volumes of methanol and a known amount of the internal standard prostaglandin B 1 (The Upjohn Co.). The samples were acidified to pH 3 with 0.1 M HCl and subjected to solid-phase extraction with Chromabond C 18 columns (Macherey Nagel). The product formation was analyzed by reverse-phase HPLC. For inhibition studies, the enzyme was preincubated for 40 min at room temperature with the inhibitor prior to activity determination.
Reverse-phase HPLC-For reverse-phase HPLC, a column packed with Novapak C 18 (4 m, 5 ϫ 100 mm, Radial-Pak, Waters Associates) was eluted with acetonitrile/methanol/water/acetic acid (30:35:35:0.01, by volume) at a flow rate of 1.2 ml/min. The ultraviolet detector monitored the eluate at 270 nm. Quantifications were made from area integrations using Baseline 810 computer software (Waters Associates) based on a standard curve of known amounts of prostaglandin B 1 and the compound to be measured. Quantification of 5S,6S-DHETE was carried out by UV spectrophotometry assuming an extinction coefficient of 40,000 M Ϫ1 cm Ϫ1 . For LTB 4 , an extinction coefficient of 56,000 M Ϫ1 cm Ϫ1 was used. Material eluting under individual peaks was collected, concentrated under a stream of nitrogen, and dissolved in methanol prior to further analysis.
Ultraviolet Spectrophotometry-Ultraviolet spectra were recorded on a Hewlett-Packard 8450 A spectrophotometer with methanol as solvent.
Incubations with Soybean Lipoxygenase-Compound V (1.6 g) was incubated with soybean lipoxygenase (2.2 g) in 500 l of 0.1 M sodium borate buffer, pH 9, at room temperature in a quartz cuvette. The reaction was monitored by repeated measurements of the UV absorbance at 274 nm (substrate) and 303 nm (product).
Gas Chromatography/Mass Spectrometry-GC/MS was performed with a Hewlett-Packard Model 5970B mass selective detector connected to a Hewlett-Packard Model 5890 gas chromatograph equipped with a phenylmethylsilicone capillary column (12-m length, 0.33-m film thickness). Helium at a flow rate of 38 cm/s was used as a carrier gas. Injections were made in the split mode at an injector temperature of 200°C. The initial column temperature was 120°C and was raised at 10°C/min until 240°C.

Incubation of Mutant Enzymes with LTA 4 -Reverse-phase
HPLC analysis of products formed when mutant enzymes, i.e. mouse [Y383F], [Y383H], and [Y383Q]LTA 4 hydrolases, were incubated with LTA 4 revealed five peaks (I-V) originating from the substrate (Fig. 1). Peaks I and II corresponded to the epimers at C-12 of 5S,12-dihydroxy-6,8,10-trans-14-cis-eicosatetraenoic acid, nonenzymatic hydrolysis products of LTA 4 , whereas peak III eluted with the expected enzymatic product, LTB 4 . Peak IV represented the least hydrophobic isomer of the two nonenzymatically formed 5,6-DHETEs, both of which are normally formed in small and equal amounts in aqueous solutions (20). The large peak V, on the other hand, had a retention time similar to that of the most hydrophobic isomer of the two 5,6-DHETEs, formed by spontaneous hydrolysis. Notably, in incubations of the wild-type enzyme and the control mutant [G386A]LTA 4 hydrolase, only minute amounts of peaks IV and V were detected (data not shown). The retention time in reverse-phase HPLC suggested that compound V could be an isomer of 5,6-DHETE. It has previously been reported that soluble epoxide hydrolase converts LTA 4 into 5S,6R-DHETE. For comparison, a sample of soluble epoxide hydrolase was incubated with LTA 4 to produce 5S,6R-DHETE (21), which did not, however, elute with compound V, but, as previously reported, with the more hydrophilic nonenzymatic isomer (Fig. 1).
Structure of Compound V-The material eluting under peak V was collected and subjected to UV spectrophotometry, which revealed a conjugated triene spectrum (in methanol) typical of leukotrienes, with max ϭ 274 nm and shoulders at 263 and 285 nm (Fig. 2). These values differ by 1-2 nm from previously published data for nonenzymatically formed or enzymatically formed isomers of 5,6-DHETE (20,21), which most likely reflects differences in the experimental conditions and the instrumentation rather than structural differences of the compounds.  (Fig. 2), compatible with the presence of a tetraunsaturated C 20 fatty acid with hydroxyl groups at C-5 and C-6.
To determine the origin of the oxygen in the hydroxyl groups of compound V, incubations of [Y383Q]LTA 4 hydrolase (6.9 M) with LTA 4 (69 M) were carried out in 50 mM Tris-Cl, pH 8, containing 84% H 2 18 O. Compound V was collected and subjected to GC/MS analysis, which revealed that the ion at m/z 291 (M Ϫ 203), which carries the C-6 hydroxyl group, was greatly reduced in intensity and was accompanied by an abundant ion at m/z 293. In contrast, the ion at m/z 203, containing the C-5 hydroxyl group, appeared at the same ratio versus the ion at m/z 205, as in the spectrum of compound V generated in H 2 16 O. Furthermore, selected ion monitoring of m/z 463 and . Hence, the hydroxyl group at C-6 was derived from water. The stereochemistry of the vicinal diol in compound V was established by a comparative GC/MS analysis with authentic methyl erythro-and threo-5,6-dihydroxyeicosanoates. Thus, a sample of compound V was methylated and hydrogenated. The trimethylsilyl ether derivative of the resulting methyl 5,6-dihydroxyeicosanoate was subjected to GC/MS analysis. The retention time was 15.9 min, and the mass spectrum showed prominent ions at m/z 471, 299, and 203. This retention time and mass spectrum were identical to those of the trimethylsilyl ether derivative of authentic methyl threo-5,6-dihydroxyeicosanoate. Since H 2 18 O was incorporated at C-6, the hydroxyl group at C-5 will retain its S-configuration from LTA 4 , and the hydroxyl group at C-6 must be in the S-configuration. As expected, the trimethylsilyl ether derivative of a hydrogenated and methylated sample of synthetic 5S,6S-dihydroxy-7,9trans-11,14-cis-eicosatetraenoic acid showed a peak at 15.9 min, corresponding to the trimethylsilyl ether derivative of methyl threo-5,6-dihydroxyeicosanoate. Hence, the stereochemistry of the vicinal diol in compound V is 5S,6S.
To obtain information about the location and geometry of the double bonds, a sample of compound V was subjected to soybean lipoxygenase conversion. 1.6 g of the compound was incubated with 2.2 g of soybean lipoxygenase in 500 l of 0.1 M sodium borate buffer, pH 9, at room temperature. The reaction was followed on a UV spectrophotometer. The triplet spectrum of compound V with an absorbance maximum at 275 nm (in buffer) was completely shifted within 20 min to another triplet spectrum with a maximum at 303 nm, typical for a conjugated tetraene and in accordance with the formation of 5S,6S-dihydroxy-15S-hydroperoxy-7,9,13-trans-11-cis-eicosatetraenoic acid (Fig. 3). Since soybean lipoxygenase requires a cis,cis-1,4-pentadiene structure for activity, we conclude that the ⌬ 11 -and ⌬ 14 -double bonds are both in cis-configuration.
Finally, compound V was shown to cochromatograph with a The time course for the reaction, measured as the increase in A 303 , was followed on a UV spectrophotometer. The 5,6-DHETE triplet spectrum with a peak maximum at 275 nm in buffer (upper left inset) was completely shifted within 20 min to another triplet spectrum with a peak maximum at 303 nm, typical for a conjugated tetraene (lower right inset). synthetic standard of 5S,6S-dihydroxy-7,9-trans-11,14-cis-eicosatetraenoic acid (Fig. 4). In addition, GC/MS analysis of the methyl ester trimethylsilyl ether derivative of the synthetic standard showed a practically identical chromatographic behavior and mass spectrum as compared with compound V (data not shown). Based on the data obtained from analysis by HPLC, UV spectrophotometry, GC/MS analysis of material generated in H 2 16 O and H 2 18 O, stereochemical analysis of the hydroxyl groups, soybean lipoxygenase conversion, and comparison with a synthetic standard, compound V is assigned the structure 5S,6S-dihydroxy-7,9-trans-11,14-cis-eicosatetraenoic acid.
Catalytic  (Table I). Also, the relative formation of 5S,6S-DHETE did not change when the concentration of substrate varied between 7 and 80 M. The wild-type enzyme and the control mutant [G386A]LTA 4 hydrolase produced only 3 and 4% 5S,6S-DHETE, reflecting nonenzymatic decomposition of LTA 4 . Since [Y383Q]LTA 4 hydrolase had the highest epoxide hydrolase activity, it was selected for further kinetic analysis. Apparent kinetic constants were determined for the formation of 5S,6S-DHETE and LTB 4 by [Y383Q]LTA 4 hydrolase as well as the formation of LTB 4 by the wild-type enzyme (Table II). The K m values for LTA 4 were determined to be 31 and 61 M for the formation of LTB 4 and 5S,6S-DHETE, respectively, by [Y383Q]LTA 4 hydrolase. Furthermore, the maximal initial reaction velocity (V max ) was higher for the formation of 5S,6S-DHETE as compared with LTB 4 (932 and 649 nmol mg Ϫ1 min Ϫ1 , respectively), which was also reflected in a higher turnover (k cat ) of LTA 4 into the former product. The specificity constant (k cat /K m ) was also higher (34 ϫ 10 3 s Ϫ1 M Ϫ1 ) for the conversion of LTA 4 into 5S,6S-DHETE, suggesting that the active site of [Y383Q]LTA 4 hydrolase is better suited for this reaction than for the hydrolysis of LTA 4 into LTB 4 (k cat /K m ϭ 12 ϫ 10 3 s Ϫ1 M Ϫ1 ). The values of K m , V max , k cat , and k cat /K m for the wild-type enzyme were determined to 5 M, 1030 nmol mg Ϫ1 min Ϫ1 , 1.18, and 236 ϫ 10 3 s Ϫ1 M Ϫ1 , respectively, in good agreement with previously published data (16). Thus, when compared with [Y383Q]LTA 4 hydrolase, the wild-type enzyme exhibited a higher affinity for LTA 4 (Table III). In parallel, a sample of the wild-type enzyme was treated in the same way. As expected, the wild-type enzyme lost 50 and 58% of its peptidase and epoxide hydrolase activities, respectively, after exposure to LTA 4 . Similarly, [Y383Q]LTA 4 hydrolase lost 50 and 58% of its ability to form LTB 4 and 5S,6S-DHETE, respectively. Although we did not demonstrate covalent binding between LTA 4 and [Y383Q]LTA 4 hydrolase, the loss of enzyme activities indicates that the wild-type and mutant enzymes are equally susceptible to suicide inactivation by LTA 4 .
We also investigated the sensitivity of the two catalytic activities of [Y383Q]LTA 4 hydrolase to inhibition by the competitive inhibitor bestatin. Thus, the concentrations of bestatin required for half-maximal inhibition (IC 50 ) of the formation of 5S,6S-DHETE and LTB 4 were 1.7 and 1.9 M, respectively (Table IV). To circumvent potential differences in inhibitor potency related to differences in substrate affinity, the K i values for inhibition of the two activities were calculated from Dixon plots. The   (Fig. 1). Structure of the Novel Enzymatic Product-Several different approaches were used to solve the structure of the novel enzymatic metabolite, termed compound V. Its UV spectrum and retention time in reverse-phase HPLC indicated that it could be an isomer of 5,6-DHETE, an assumption that was verified by GC/MS analysis (Fig. 2). The mass spectrum of compound V, generated in a buffer containing H 2 18 O, revealed that the hydroxyl group at C-6 originated from water. Since the epoxide oxygen of LTA 4 is in the S-configuration at C-5, the stereochemistry of the vicinal diol must be either 5S,6S or 5S,6R. The latter alternative seemed unlikely since compound V did not cochromatograph with 5S,6R-DHETE (Fig. 1), the enzymatic product obtained when soluble epoxide hydrolase is incubated with LTA 4 (21). Final proof for the stereochemistry of the diol was obtained by GC/MS analysis of the trimethylsilyl ether derivative of a hydrogenated and methylated sample of compound V, which displayed a retention time and fragmentation pattern identical to those of an authentic standard of methyl threo-5,6-dihydroxyeicosanoate, demonstrating that the hydroxyl groups at C-5 and C-6 must be in the S-configuration. Furthermore, evidence for the presence of two cis-double bonds at ⌬ 11 and ⌬ 14 was obtained by treatment of compound V with soybean lipoxygenase, which led to a 30-nm bathochromic shift of the UV spectrum into an absorbance profile typical of a conjugated tetraene (Fig. 3). Since soybean lipoxygenase requires a 1,4-cis-pentadiene structure for activity, it is reasonable to assume that 5S,6S-dihydroxy-7,9-trans-11,14-cis-eicosatetraenoic acid was converted into 5S,6S-dihydroxy-15Shydroperoxy-7,9,13-trans-11-cis-eicosatetraenoic acid. Finally, compound V was found to have the same retention time in reverse-phase HPLC as a synthetic standard of 5S,6S-dihydroxy-7,9-trans-11,14-cis-eicosatetraenoic acid (Fig. 4), and GC/MS analysis of the synthetic material gave virtually identical results as for compound V. Based on these analytical data and the known structure of the substrate LTA 4 , compound V was assigned the tentative structure 5S,6S-dihydroxy-7,9trans-11,14-cis-eicosatetraenoic acid (Fig. 5).  (Table I). For [Y383Q]LTA 4 hydrolase, the mutant with the highest specific epoxide hydrolase activity, the relative formation of 5S,6S-DHETE versus LTB 4 was 150%, and thus, 5S,6S-DHETE was the dominating product (Fig. 5). This is noticeable since 5,12-dihydroxy acids are by far the most abundant metabolites obtained via nonenzymatic hydrolysis as well as hydrolysis catalyzed by wild-type LTA 4 hydrolase, suggesting that the vicinal diols are thermodynamically unfavored products. Also, the relative formation of 5S,6S-DHETE by [Y383Q]LTA 4 hydrolase did not change when the concentration of substrate varied between 7 and 80 M.

5S,6S-DHETE Is a Major Enzymatic Product Generated by
[Y383Q]LTA 4 Hydrolase Turns over LTA 4 More Efficiently than Does the Wild-type Enzyme-As previously reported, [Y383Q]LTA 4 hydrolase exhibits a higher K m for LTA 4 as compared with the wild-type enzyme, and at saturating concentrations of substrate, the specific epoxide hydrolase activity (considering only the formation of LTB 4 ) was estimated to ϳ60% of the control (14). However, if the formation of 5S,6S-DHETE is taken into account, the turnover of LTA 4 is in fact higher for [Y383Q]LTA 4 hydrolase, i.e. k cat ϭ 1.8 versus 1.2 s Ϫ1 for the wild-type enzyme (Table II). When the kinetic data for the formation of LTB 4 and 5S,6S-DHETE by [Y383Q]LTA 4 hydrolase were plotted separately, no significant differences were observed, except for a somewhat higher V max value for the vicinal diol formation, as expected. The kinetic data listed in Table II indicate a difference in K m for LTA 4 . However, due to the instability of LTA 4 , determinations of the Michaelis constant become very uncertain and difficult to evaluate. Thus, it is questionable whether a 2-fold increase in K m is really significant.
Putative Mechanism for the Formation of 5S,6S-DHETE-All mutants of Tyr-383 catalyzed the stereospecific addition of water at two positions, i.e. C-6 and C-12, in the substrate LTA 4 . This unusual catalytic behavior can be explained in several ways. The mutant enzymes may use two separate functional elements toward either of the two positions of the substrate, which in turn would require that the respective mutation "activate" an additional structure within the active site to perform a nucleophilic attack on the substrate (Fig. 6). Alternatively, one functional element can operate on both positions. The latter and perhaps more plausible option would be possible if the substrate could enter the active site in two opposite head- to-tail orientations such that either C-12 or C-6 is positioned close to the incoming water (Fig. 6). Precedents for such alternative modes of substrate binding have been reported in the literature, e.g. as a possible explanation for the positional specificity of 15-lipoxygenase toward isomers of polyunsaturated fatty acids (22). However, considering the formation of LTB 4 and 5S,6S-DHETE by [Y383Q]LTA 4 hydrolase, the similar K m values for LTA 4 , the equal sensitivity to inhibition by the competitive inhibitor bestatin, and the similar susceptibility to suicide inactivation observed for the two activities argue against this hypothesis and rather suggest that LTA 4 binds in only one way to the active site. Moreover, the fact that LTA 4 methyl ester was not converted into the methyl ester of 5S,6S-DHETE lends further support to this interpretation. An additional possibility would be a single binding mode for the substrate in an active site with one functional element controlling the stereospecific hydrolysis leading to both products. Although LTA 4 would bind in one orientation to the substrate-binding pocket, it may still have a conformation and position in space that make C-12 and C-6 available for attack by a single nucleophile (Fig. 6).
Evidence for the Presence of a Carbocation Intermediate in the Enzymatic Hydrolysis of LTA 4 -From experiments with H 2 18 O, it was inferred that the nucleophilic attack of water during the formation of 5S,6S-DHETE was directed toward C-6 according to an S N 1 or S N 2 reaction. Considering the S-configuration of the hydroxyl group at C-6, an S N 2 reaction, with concomitant chiral inversion, would not be possible since the epoxide oxygen of the substrate is already in the 6S-configuration. Thus, enzymatic hydrolysis of LTA 4 into 5S,6S-DHETE must occur via an S N 1 reaction involving a carbocation intermediate. This intermediate would be planar at C-6, allowing an enzyme-directed nucleophilic attack from either side of the molecule. Since [Y383F], [Y383H], and [Y383Q]LTA 4 hydrolases could produce not only 5S,6S-DHETE, but also LTB 4 , it seems very likely that hydrolysis of LTA 4 into LTB 4 proceeds according to the same mechanism. This conclusion was further corroborated by the fact that [Y383Q]LTA 4 hydrolase makes both products with indistinguishable reaction kinetics. Nevertheless, we cannot rule out the possibility that LTB 4 could be formed via an S N 2 or rather S N 2Ј reaction, an interpretation that seems unlikely, however, since it would imply that the mutants can operate simultaneously at C-6 and C-12 of LTA 4 via two distinct enzyme mechanisms. Hence, assuming that the mutations do not interfere with the fundamental enzyme mechanism for the epoxide hydrolysis, the formation of a vicinal diol with the stereochemical configuration 5S,6S by the mutant enzymes represents the first experimental evidence that the formation of LTB 4 by wild-type LTA 4 hydrolase follows an S N 1 mechanism involving a carbocation intermediate. In addition, this reaction mechanism would be in agreement with the mechanism for nonenzymatic hydrolysis of LTA 4 (2) and conforms to the general rule that enzymes reduce the activation energy for chemical reactions that also occur spontaneously (23).
Possible Function of Tyr-383-We have recently identified a 21-residue peptide segment, denoted peptide K21, to which LTA 4 binds during suicide inactivation (24), and amino acid sequence analysis of a covalently modified form of peptide K21, isolated from human LTA 4 hydrolase inactivated by LTA 4 ethyl ester, indicated that Tyr-378 is a primary site for covalent binding of lipid to the protein. This conclusion was further corroborated by mutational analysis, which revealed that exchange of Tyr-378 for a Phe or Gln rendered the enzyme virtually resistant to mechanism-based inactivation (25). Furthermore, Tyr-378 appeared to be involved in the formation of the correct double bond geometry in LTB 4 since both of these mutants, human [Y378F] and [Y378Q]LTA 4 hydrolases, catalyzed the formation of a second enzymatic product from LTA 4 , viz. ⌬ 6 -trans-⌬ 8 -cis-LTB 4 (26). Notably, Tyr-383 is also located within peptide K21, separated from Tyr-378 by only four residues. Hence, each of the two tyrosines is involved in one of the two main catalytic features of LTA 4 hydrolase, i.e. stereospecific hydrolysis at C-12 and creation of a cis,trans,trans-conjugated double bond system in the product LTB 4 (cf. Fig. 5). Consequently, the results of the present study corroborate our previous conclusion that the proteolytic peptide fragment K21 contains active-site residues involved not only in suicide inactivation, but also in peptidolysis as well as epoxide hydrolysis. Depicted are models of the active site with an arginyl residue as a potential carboxylate recognition site and the substrate drawn as a carbocation with the positive charge delocalized over the conjugated triene system. Filled circles symbolize functional sites capable of stereospecific addition of a hydroxyl group, as indicated by the arrows. A, the substrate enters the active site in one orientation, and two separate functional entities would be responsible for the addition of a hydroxyl group to C-6 and C-12. B, the substrate could enter the active site in two opposite headto-tail orientations. A single site could account for the addition of a hydroxyl group at either C-6 or C-12, depending on the position of the substrate. C, shown are a single orientation of the substrate and a single functional element acting on both C-6 and C-12. For further details, see "Discussion." Notably, apart from the three zinc-binding ligands, Tyr-383 is the only amino acid residue identified thus far that seems to be involved in both the peptidase and epoxide hydrolase activities of LTA 4 hydrolase, thus bridging the two corresponding active sites. It should also be noted that the results concerning Tyr-378 were obtained with the human LTA 4 hydrolase, whereas the data in the present report were obtained with the mouse enzyme. However, since the human and mouse LTA 4 hydrolases are 93% identical at the amino acid level, it is reasonable to assume that results obtained by mutational analysis in one of the species pertain to the other.
Functional Relationship to Soluble Xenobiotic Epoxide Hydrolase-Apart from LTA 4 hydrolase, there are at least two other mammalian epoxide hydrolases that have been extensively characterized, i.e. soluble and microsomal epoxide hydrolases (27,28). Both of these enzymes are believed to be involved in the detoxification of potentially harmful xenobiotic epoxides, although several endogenous physiological substrates have been described, at least for soluble epoxide hydrolase (21, 29 -32). Recent work including computer-assisted sequence comparisons, x-ray crystallographic analysis of structurally related enzymes, and biochemical and mutational analyses has identified soluble and microsomal epoxide hydrolases as members of the ␣/␤-fold family of hydrolases (33)(34)(35)(36)(37)(38)(39). Concerning LTA 4 hydrolase, there is not enough structural or biochemical data available to conclusively determine whether or not it belongs to the same class of enzymes. However, LTA 4 has been shown to be an excellent substrate for soluble epoxide hydrolase (but not for microsomal epoxide hydrolase), which converts the allylic epoxide into 5S,6R-DHETE, i.e. an epimer at C-6 of the vicinal diol produced by [Y383Q]LTA 4 hydrolase (21). Hence, the subtle structural changes at the active site of [Y383Q]LTA 4 hydrolase shift the positional specificity of the stereospecific hydrolysis such that the mutant enzyme begins to mimic the action of soluble epoxide hydrolase. One may speculate that this functional resemblance, caused by a single amino acid change, is a sign of structural similarity between the active sites of soluble epoxide hydrolase and LTA 4 hydrolase.