Enzymatic hydration of leukotriene A4. Purification and characterization of a novel epoxide hydrolase from human erythrocytes.

Human erythrocytes contained a soluble cytosolic epoxide hydrolase for stereospecific enzymatic hydration of leukotriene A4 into leukotriene B4. The enzyme was purified 1100-fold, to apparent electrophoretic homogeneity, by conventional DEAE-Sephacel fractionation followed by high performance anion exchange and chromatofocusing procedures. Its characteristics include a molecular weight of 54,000 +/- 1,000, an isoelectric point 4.9 +/- 0.2, a Km apparent from 7 to 36 microM for enzymatic hydration of leukotriene A4, and a pH optimum ranging from 7 to 8. The enzyme was partially inactivated by its initial exposure to leukotriene A4. There was slow but detectable enzymatic hydration (pmol/min/mg) of certain arachidonic acid epoxides including (+/-)-14,15-oxido-5,8-11-eicosatrienoic acid and (+/-)-11,12-oxido-5,8,14-eicosatrienoic acid, but not others, including 5,6-oxido-8,11,14-eicosatrienoic acid. Human erythrocyte epoxide hydrolase did not hydrate either styrene oxide or trans-stilbene oxide. In terms of its physical properties and substrate preference for leukotriene A4, the erythrocyte enzyme differs from previously described versions of epoxide hydrolase. Human erythrocytes represent a novel source for an extrahepatic, cytosolic epoxide hydrolase with a potential physiological role.

Among those cells that can convert LTA, into LTB,, human erythrocytes attract attention from two perspectives. First, in contrast to other cells, particularly leukocytes (6, 7, 14-17), erythrocytes have seldom been attributed any capacity for eicosanoid biosynthesis until recently (18,19). Furthermore, erythrocytes have not been recognized as a source of extrahepatic epoxide hydrolase activity for hydration of xenobiotic oxiranes (2).
In view of its unexpected presence and its uncertain role in human erythrocytes, we have purified and partially characterized an epoxide hydrolase that converts LTA, into LTB,. This enzyme differs in several respects from previously identified forms of epoxide hydrolase (1-3,20-24).
Instrumental Analysis-A Lambda 5 UV/VIS recording spectrophotometer (Perkin-Elmer), a model llOA positive displacement pump (Altex Beckman), a Spectromonitor D@ variable wavelength detector (Milton Roy), and a fast protein liquid chromatography (FPLC) system (Pharmacia, Uppsala, Sweden) were used for spectrophotometry and quantitative or preparative chromatographic procedures.
Enzyme Preparation-Purified erythrocyte suspensions were prepared as described (31). After depletion of leukocytes (55 x 104/ml) and platelets (55 X 106/ml) by filtration through cellulose columns, the resulting suspension (50 ml) contained 2-3 X lo9 human eryth-rocytes/ml. Suspended erythrocytes were centrifuged at 200 X g for 20 min, and the supernatant fluid was discarded. Erythrocytes were then lysed hypotonically by resuspension in sterile water (50 ml). After restoration of isotonic conditions, the cytosolic fraction was separated from the stroma by centrifugation at 100,000 X g for 1 h. The 100,000 X g supernatant was filtered through a 0.45-pm Millipore filter and dialyzed twice against 4 liters of 0.01 M Tris, pH 8.0, for 24 h at 4 "C. The cell pellet was washed three times by centrifugation and resuspension in 0.01 M Tris, pH 8.0. Crude lysate containing membranes and cytosol, the post 100,000 x g supernatant fluid, and the post 100,000 X g cell pellet were incubated at 37°C with LT& (10 PM) to determine the subcellular distribution of epoxide hydrolase in human erythrocytes.
Determination of Enzyme Activity-Erythrocyte epoxide hydrolase activity was determined by measuring the conversion of LTA4 into its enzymatic hydration product, LTB4. For routine assays [3H]LT& Li (10 nmol) dissolved in tetrahydrofuran (2.5-5 pl) was added to solutions of cytosolic enzyme (1.0 ml, 40-50 mg/ml protein) or to partially purified enzyme (1.0 ml, 2-3 mg/ml protein) in 0.01 M Tris, pH 8.0. After incubation for 1 min at 37 "C, residual substrate was hydrolyzed by quenching with 9 ml of 0.9% NaCl (w/v), pH 5.5. Prostaglandin B, (1 pg) was added as an internal standard, and the solution was percolated through a Sep-Pak CIS cartridge. Absorbed material was removed by sequential elution with 10 ml each of water, hexane, and ethyl acetate (32). The ethyl acetate fraction contained >95% of the adsorbed radioactive material including LTB4, prostaglandin B1, and nonenzymatic hydration products. After evaporation of the solvent, the LTB4 content in the residue was quantified by RP-HPLC (6, 33).
For enzyme kinetic experiments, the substrate concentration was varied from 10 to 100 p~ by incubating enzyme (0.5 ml, 2-3 mg/ml) with 5-50 nmol of LTA, dissolved in 2.5-5 pl of tetrahydrofuran. For kinetics at 37"C, the enzyme reaction was terminated at 1 min by quenching with 5 ml of 0.9% NaCl, pH 5.5, to destroy unreacted substrate. Enzymatic kinetics were also determined at 25 and 4°C. At these temperatures, in some experiments the amount of intact LTA, was determined after a 1-min incubation to assure that it was present in sufficient excess. Intact LTA4 was quantitated by RP-HPLC as previously described (26).
DEAE Anion Exchange Chromatography-After dialysis against 0.01 M Tris, pH 8.0, approximately 10-20 ml (40-50 mg/ml protein) of the 100,000 X g supernatant fluid was applied to a 2.5 X 20-cm DEAE-Sephacel (Pharmacia) anion exchange column equilibrated with 0.01 M Tris, pH 8.0. Nonbound material was removed by elution with 0.01 M Tris, pH 8.0. The material that bound to the column, including hemoglobin and epoxide hydrolase, was eluted at 0.5 ml/ min with a linear gradient of 250 ml each of 0.01 M Tris, pH 8.0, and 0.01 M Tris, pH 8.0, containing 1 M NaCl. The absorbance at 415 and 280 nm was monitored in 10-ml fractions, and their protein content was determined spectrophotometrically (34). Fractions were concentrated 2-fold by ultrafiltration through Centricon 30 microconcentrators (Amicon) and assayed for epoxide hydrolase activity as described above. Active fractions were pooled and dialyzed against 0.01 M Tris, pH 8.0, for use in subsequent experiments.
Substrate Specificity-Certain epoxide metabolites of arachidonic acid (35,36) and selected xenobiotic epoxides were also evaluated as substrates for human erythrocyte epoxide hydrolase. Enzymatic hydration of [3H]trans-stilbene oxide and [3H]styrene oxide was measured as described (37, 38). Enzymatic hydration of 5,6-EET, 11,12-EET, and 14,15-EET was determined by an adaptation of RP-HPLC procedures (35,39). Samples (1.0 ml) from enzymatic reaction mixtures containing 5-300 nmol of substrate were quenched with 0.9% NaCl (4.0 ml), pH 5.5, and extracted with ethyl acetate (3 X 5 ml). The organic layers were evaporated under nitrogen, and residual 5,6-EET, 11,12-EET, or 14,15-EET and their corresponding hydration products were converted into UV absorbing naphthacyl esters, A,, Characterization of Epoxide Hydrolase-Cytosolic epoxide hydrolase purified by DEAE-Sephacel anion exchange chromatography was used to characterize the enzymatic hydration of LT&. The pH dependence was determined with enzyme dialyzed against 0.01 M Tris-HC1 buffers ranging from pH 6.0 to 9.0. The protein concentration dependence was determined with geometrically diluted enzyme preparations containing 0-15 mg of protein/ml in 0.01 M Tris-HC1, pH 8. Suicide substrate inactivation was monitored by determining cumulative LTB4 formation following three successive additions of LTA, (10 nmol) to a single incubation mixture (1.0 ml, 3 mg of protein/ml) after 2-min intervals. Stability was determined by comparing the initial enzymatic activity with that obtained after its incubation at 37 "C for 48 and 72 h. To characterize enzymatic inhibition, solutions (1.0 ml) containing partially purified enzyme were incubated with selected compounds (10-100 pM) for 2 min at 37°C prior to addition of substrate (10 pM LTAJ. LTB4 production was quantitated as described above. Isolation of Epoxide Hydrolase by FPLC-Pooled, dialyzed DEAE-Sephacel fractions (5-10 mg of protein) were applied to an FPLC Mono-Q@ anion exchange column (HR 5/5) equilibrated with 0.01 M Tris, pH 8.0. Bound proteins were eluted at 1 ml/min with a linear gradient from 0 to 0.5 M NaCl in 0.01 M Tris, pH 8.0, for 20 min. The absorbance of the eluent was monitored continuously at 280 nm. Fractions were concentrated to 0.5 ml by ultrafiltration, and aliquots (200 pl) were assayed for epoxide hydrolase activity. Active fractions from the Mono-& purification step were purified further by FPLC chromatofocusing. Approximately 5-10 mg of protein was applied to an FPLC Mono-P@ column equilibrated with 0.025 M bis-Tris, pH 6.4. Bound proteins were eluted at 1.0 ml/min with 10% (v/v) Polybuffer 74 (Pharmacia), pH 4.7. The pH in 1.0 ml fractions was determined with a glass electrode. Fractions corresponding to specific peaks were pooled, concentrated by ultrafiltration, and reconstituted in 0.01 M Tris, pH 8.0. Aliquots (200 pl) were then assayed for epoxide hydrolase activity.
SDS-PAGE-Polyacrylamide electrophoresis was performed according to the method of Laemmli (41), as modified by Kelly and Luttges (42) using a Bio-Rad Protean@ dual slab gel electrophoresis apparatus and a concave gradient of 7.5-20% acrylamide. All protein samples contained 10 mM dithiothreitol (Bio-Rad). SDS-PAGE molecular weight markers were obtained from Bio-Rad (No. 161-0304). Protein bands were stained with Coomassie Brilliant Blue R-250 (Bio-Rad).
Their specific activity was 0.044 f 0.026 nmol of LTBJminl mg. Enzymatic hydration of LTA, occurred instantaneously; LTB4 formation was maximal within 1 min and its concentration remained constant for 10 min. For instance, at 0.5, 1.0, and 2.0 min after addition of LTA4, enzymatic reaction mixtures contained 0.85 & 0.2, 1.07 f 0.1, and 0.97 f 0.08 nmol of LTB4/ml, respectively. Formation of LTB, at 37T, 1.53 nmol/ml, slightly exceeded that at 23T, 1.3 nmol/ml. The epoxide hydrolase present in erythrocyte cytosol was stable. It retained 106 k 10 and 108 * 10% of its original enzymatic activity after incubation at 37°C for 48 and 72 h, respectively. However, incubation for 1 h at 56 "C eliminated detectable epoxide hydrolase activity.
In contrast to the cytosolic fraction, the erythrocyte stroma did not convert LTA, into LTB,; only nonenzymatic hydration products were formed (Fig. 1).
Partial Purification and Characterization-Fractionation of dialyzed 100,000 x g supernatant fluid by anion exchange chromatography on DEAE-Sephacel revealed that the epoxide hydrolase activity was distinct from the major hemoglobin peak (Fig. 2). The specific activity of DEAE-purified enzyme increased 5-fold to 0.36 nmol of LTB4/min/mg (Table I). According to Lineweaver-Burk analysis, the apparent K, for erythrocyte epoxide hydrolase ranged from 7 to 36 p~, and the V,,, ranged from 0.29 to 1.43 nmol of LTB,/ml/min for DEAE-purified enzyme obtained from three blood donors. These values are based on LTB, formation after 1 min at 37°C. Under these conditions, the amount of intact, residual substrate may have declined considerably by accompanying, nonenzymatic hydration. Consequently, kinetics were also determined at 25 and 4 "C. At these temperatures, intact substrate was still present in excess after a 1-min incubation.
For instance, enzyme reaction mixtures contained (mean * S.D., n = 5 ) 81 k 7,75 f 6, and 60 * 8% of the initially added LTG, intact, after incubation at 4 "C for 1, 2, and 5 min, respectively. Enzyme reaction mixtures contained 52 f 7% of the initially added LTA,, intact, after incubation at 25 "C for 1 min. At 25 "C, the apparent K,,, ranged from 9 to 30 pM; at 4 "C, K, apparent ranged from 19 to 36 p~. Suicide Inactivation by LTA4-At 25 .and 4 "C, LTB4 formation was maximal within 2 min even though the enzyme reaction mixtures contained a sufficient excess of LTA4 to sustain enzymatic hydration for a longer time. This suggested that epoxide hydrolase was inactivated by its reaction with LTA,. This was substantiated by demonstrating that accumulation of LTB, was not proportional to successive, supplemental additions of LTA, to enzyme. For example, an initial incubation of DEAE-Sephacel-purified enzyme (1.0 ml, 3 mg) with 10 nmol of LTA, produced 2.2 nmol of LTB4/ml. After two successive additions of LTA, (10 nmol), at 2-min intervals, cumulative LTB, production increased to 2.4 and 3.0 nmol/ml, respectively. Therefore, LTB4 production per incremental addition of LTA, corresponded to 2.2, 0.2, and 0.6 nmol of LTBJml/lO nmol of LT& or increases of 9 and 25% per respective addition. There was no evidence of product inhibition by LTB4 (10 p~) or its nonenzymatic 5,12-dihydrodiol isomers (10-100 p~) . Furthermore, when fresh enzyme (1.0 ml, 3 mg) was mixed with enzyme (1.0 ml, 3 mg) inacti-vated by LTA,, the fresh enzyme could synthesize LTB,. Addition of LTA, (10 nmol) to such a mixture restored cumulative LTB, production from 1.3 to 2.9 nmol/ml. Activity associated with addition of fresh enzyme was consistent with suicide inactivation, but not product inhibition, since the incubation mixture contained LTB, and nonenzymatic hydration products derived from prior exposure to substrate. Substrate Specificity-Several other epoxides were evaluated as substrates. There was slow, but detectable enzymatic hydration of certain epoxide metabolites of arachidonic acid such as 14,15-EET and 11,12-EET but not others, such as 5,6-EET. Epoxide hydrolase in the 100,000 X g supernatant fluid transformed 14,15-EET (50 p~) and 11,12-EET (50 p~) into their corresponding vicinal diols at respective rates of 1.78 f 0.57 pmol/min/mg (mean f S.E., n = 6) and 0.59 & 0.11 pmol/min/mg (n = 4). Initial hydration rates were linearly proportional to the protein content in the incubation mixtures. The plot of protein concentration (abscissa, 0-42 mg/ml) versus 14,15-EET hydration rate (ordinate, 0-10.1 nmol/h/ml) was linear with a correlation coefficient 0.99 and a slope of 0.24 nmol/h/mg. For 11,12-EET, the plot was linear with a correlation coefficient of 0.99 and a slope of 0.031 nmol/h/mg. The initial hydration velocity was constant for at least 15 h. Nonenzymatic hydration in 0.01 M Tris, pH 8.0, at 37°C was negligible: 0.0029 and 0.0014 h-l, respectively, for 14,15-EET and 11,12-EET. Enzymatic hydration of 5,6-EET (50 p~) was indistinguishable from its nonenzymatic hydration which occurred spontaneously at 0.33 f 0.05 nmol/min/ ml (mean f S.D., n = 4) in 0.01 M Tris, pH 8.0, at 25 "C. According to Lineweaver-Burk analysis, 14,15-EET had a K, apparent of 20 p~ and Vmax of 2.3 pmol/min/mg. Accurate determination of the apparent K, for 11,12-EET was not possible due to its slow hydration rate. At various stages of purification, there was no detectable enzymatic hydration of [3H]trans-stilbene oxide or [3H]styrene oxide, indicating that these common substrates for hepatic epoxide hydrolase (43) were poor substrates for erythrocyte epoxide hydrolase.
Purification to Electrophoretic Homogeneity-The epoxide hydrolase present in DEAE-Sephacel fractions was purified to electrophoretic homogeneity by FPLC anion exchange chromatography on a Mono-Q (HR 5/5) column (Fig. 3) followed by FPLC chromatofocusing on a Mono-P (HR 5/20) column (Fig. 4). SDS-PAGE of the active Mono-P fraction contained a single band with a PI 4.9 f 0.2, a specific activity of 83.5 nmol/min/mg, and M, 54,000 k 1,000 (Fig. 5). Table   I summarizes the purification scheme.

DISCUSSION
An epoxide hydrolase involved in leukotriene metabolism has not been purified or characterized extensively. Consequently, it is uncertain to what extent it resembles or differs from multiple forms of the hepatic enzyme involved in xenobiotic metabolism. The latter have been characterized sufficiently to facilitate a comparison (1)(2)(3)(20)(21)(22)(23)(24)  epoxide hydrolase for efficient enzymatic hydration of LTA,. This enzyme differs in several respects from known microsomal or cytosolic forms of hepatic epoxide hydrolase. Its molecular weight (Mr 54,000) is similar to values reported for microsomal (M, 50,000-53,000) or cytosolic (M, 58,000) hepatic enzyme (22,24,45,46). However, its stability, with no evident inactivation after 72 h at 37 "C, exceeds that reported for cytosolic (25-40% inactivation after 7 h at 37 "C) (23) or microsomal (10% inactivation after 24 h at 25 "C) hepatic enzymes (45). Its substrate specificity for LT& and its inability to catalyze hydration of either styrene oxide or transstilbene oxide distinguish it from any known form of hepatic epoxide hydrolase. The latter efficiently hydrate one and often both of these xenobiotic substrates (1)(2)(3)37,38). Furthermore, human erythrocyte cytosolic epoxide hydrolase hydrates 14,15-EET and 11,lP-EET at rates (pmol/min/mg) that differ by 106-fold from hydration rates (pmol/min/mg) reported for hepatic cytosolic enzyme (39). Such differences in substrate specificity have been useful for categorizing different hepatic epoxide hydrolases (43). Conversion of LTA, to LTB, by erythrocyte epoxide hydrolase was unaffected by several compounds that can inhibit the corresponding hepatic enzyme (44). The pH optimum (7.0-8.0) and PI (4.9 f 0.2) for the erythrocytes enzyme concur with values reported for hepatic, cytosolic enzyme (pH optimum 7.0 (43), PI 5.1-6.1 (24)), but differ from the pH optimum (8.9-9.4) for hepatic microsomal enzyme (45). It is noteworthy that human erythrocytes also contain a unique, extrahepatic form of glutathione S-transferase (47), a functionally related enzyme.
Values for K,,, and Vmax can be derived from Lineweaver-Burk analyses; however, it is necessary to stress that enzymatic hydration of LTA, conforms imperfectly to Michaelis-Menten kinetics for two reasons. First, spontaneous, nonenzymatic hydration of LTA, accompanies its enzymatic hydration. At 37 "C, rapid depletion of LTA, by both processes reduces the interval during which enzymatic hydration velocity remains constant. At 4 and 25 "C, this problem is less significant. Second, suicide inactivation by LTA, also limits the duration of constant enzymatic velocity. Since LTB, formation was maximal within 1 min, the values for initial reaction rate used in Lineweaver-Burk plots were based on LTB, formation in that span. The initial rate of enzymatic hydration may have been somewhat higher; therefore, values for Vmax are a conservative estimate. The values obtained for K,,,, 7-36 p~, are plausible from a physiological perspective, and they were independent of temperature in the kinetic experiments. Stable substrates, such as 14,15-EET or 14,15-EETE with negligible rates of spontaneous hydrolysis, with no evident suicide inactivation effects, and with persistent initial hydration velocity, conformed closely to Michaelis-Menten kinetics.
It is interesting to note that several enzymatic processes involved in biosynthesis and metabolism of eicosanoids are similar to those involved in activation or detoxification of xenobiotics. In addition to the epoxide hydrolase system, glutathione S-transferase may conjugate either eicosanoids or xenobiotics, and the cyclooxygenase-peroxidase enzyme complex may also catalyze cooxidation of xenobiotics (48,49). The electrophilic and hydrophobic nature of many eicosanoids or xenobiotics probably accounts for this uniformity. Purification and biochemical characterization of functionally discrete enzymes is warranted to determine if pharmacological goals, such as selective modulation of eicosanoid biosynthesis, are attainable without corresponding impairment of fundamental detoxification processes.
In summary, human erythrocytes are a novel extrahepatic source for epoxide hydrolase. In terms of either eicosanoid biosynthesis or xenobiotic metabolism, there was no cellular metabolic precedent to anticipate its presence in erythrocytes, until recently (18). Its capacity for conversion of a naturally occurring epoxide into a biologically active dihydrodiol, LTB,, suggests that this may be an important function, under certain circumstances. Further investigation is warranted in this regard. In view of the losses during isolation and purification and the established multiplicity of hepatic forms of epoxide hydrolase, we cannot exclude that additional forms of erythrocyte epoxide hydrolase may exist. The erythrocyte enzyme resembles in some, but not all, respects an LTA4 hydrolase isolated from human leukocytes (50).