Enzymic conversion of 11,12-leukotriene A4 to 11,12-dihydroxy-5,14-cis-7,9-trans-eicosatetraenoic acid. Purification of an epoxide hydrolase from the guinea pig liver cytosol.

(11S,12S)-Epoxy-5,14-cis-7,9-trans-eicosatetraenoic acid (11,12-leukotriene A4) was nonenzymically converted to seven compounds: two diastereomers of (12S)-hydroxyeicosatetraeno-delta-lactones (major products), two diastereomers of (5,12S)-dihydroxyeicosatetraenoic acid and three stereoisomers of (11,12S)-dihydroxyeicosatetraenoic acid. Among these compounds, (11R,12S)-dihydroxy-5,14-cis-7,9-trans-eicosatetraenoic acid proved to be the only enzymic product. This hydrolysis activity was present in the cytosol fractions of various tissues of guinea pig such as liver, adrenal gland, small intestine, and brain. We purified the epoxide hydrolase to an apparent homogeneity from the guinea pig liver. The enzyme had a molecular weight of 60,000 as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and an isoelectric point of 7.3. The partial amino acid sequence was different from that of the microsomal enzyme. Km and Vmax values for 11,12-leukotriene A4 were 18 microM and 2.4 mumol/min/mg protein, respectively. These results indicate that 11,12-dihydroxyeicosatetraenoic acid is enzymically synthesized from 11,12-leukotriene A4 by the action of the cytosolic epoxide hydrolase in vitro.

We report herein that 11,12-LT& is nonenzymically decomposed to at least seven compounds, and that one stereoisomer of 11,12-diHETE is enzymically synthesized by epoxide hydrolase (EC 3.3.2.3) present in the various tissues of the guinea pig. This enzyme was purified from the guinea pig liver to an apparent homogeneity, and various properties were examined.

RESULTS
Enzymic Synthesis of C-3 from 11,12-LTA4-When 11,12-L T G was incubated in the standard reaction mixture without enzyme, seven compounds ( Fig. L4, peaks C-1, -2, -3 (two components), -4, -5, and -6) were observed and all showed characteristic UV spectra of conjugated triene with Amax at 268 nm. Under present assay conditions, C-5 and C-6 were major products. When the guinea pig liver cytosol was added to the above reaction mixture, a prominent increase in C-3 was observed with a concomitant decrease in both C-5 and C-6 ( Fig. 1B). AS shown in Fig. 2, the 11,12-diHETE (C-3, see below) formation was increased dose-dependently on the enzyme amount, whereas the formations of both C-5 and C-6 were decreased. Thus, C-3 was enzymically synthesized from 11,12-LT&, and C-5 and C-6 were nonenzymic products.
Structural Identification of Products Derived from 11,lZ-LTA4-The methyl esters of C-1 to -4 were obtained by the acid hydrolysis of ll,lZ-LT& methyl ester (34), but those of C-5 and C-6 were not observed. C-1 and C-2 have already been identified (34) as the diastereomers of (5,12S)-dihydroxy-6,8,10-tram-l4-cis-eicosatetraenoic acid ((5,12S)-di-HETE). C-3 in Fig. 1   A, 11,12-LTA4 (1.5 nmol) was incubated in 100 mM Tris-HC1 buffer, pH 7.3, containing 10 mg/ml bovine serum albumin at 37 "C for 1 min, and solution A was added to terminate the reaction according to "Experimental Procedures." Prostaglandin B2 was added as an internal standard (IS). B, 11,lZ-LTAI was incubated with guinea pig liver cytosol (0.6 mg of protein) at 37 "C for 1 min. Peaks C-l to C-6 and * showed UV absorption of the conjugated triene. The peak * had the same retention time as C-1 methyl ester, possibly derived from unsaponified substrate (11,12-LTA4 methyl ester). C-2 methyl ester coeluted with C-5. FIG. 2. Dependence of C-3 formation on the enzyme amount. The cytosol was dialyzed against Buffer A for 12 h. 11,12-LTA4 (1.5 nmol) was incubated with the varying concentrations of guinea pig liver cytosol at 37 "C for 1 min. The products were calculated as described under "Experimental Procedures." with different C values of 23.6 for C-3 and 24.4 for C-3'. Both showed identical mass spectra with the structure of 11,12-diHETE (34). C-4 was identified as an isomer of 11,12-diHETE (data not shown).
The enzymic product, C-3, was identified by gas chromatography-mass spectrometry and proton-NMR. C-3 was converted to a trimethylsilyl ether methyl ester derivative.  ,34). The geometry of double bonds was assigned as 5,14-cis-7,9-trans by proton NMR (Table I). Although the stereochemistry of CI1 and CI2 was not fully identified, C-3 is assigned to be (11R,12S)-dihydroxy-5,14-cis-7,9-truns-eicosatetraenoic acid (see "Discussion"). 11,12-LT& was incubated in the standard reaction mixture without enzyme for 1 min, and the pH of the mixture was brought to 12 with 1 N KOH. The peaks of C-5 and C-6 disappeared with simultaneous increases of C-1 and C-2. C-5 was purified by reversed-phase HPLC, and the solvent was removed by lyophilization. It was subjected to fast atom bombardment-mass spectrometry. Molecular ions were observed at 319 and 320, with glycerol and [2Hs]glycer~l as matrix, respectively. These data indicated that C-5 has one hydroxy moiety with an atomic composition of Cz~H3003. By IR spectrometry, an absorption at 1,730 cm" was observed. C-6 gave essentially identical results as described for C-5. All these results strongly suggest that both C-5 and C-6 have a 6lactone structure of 5,12-diHETE (C-1 and C-2). The mechanism of the &lactone formation will be discussed later.
Catalytic Actiuities-The purified enzyme catalyzed the formation of 11J2-diHETE. The time course was linear for at least 5 min. An optimum pH for the reaction was around 7.3. N-Ethylmaleimide and p-hydroxymercuribenzoic acid inhibited the enzyme activity with ICs0 values of 5 and 0.06 mM, respectively. It acted on LT& and 14,15-LTA4 in addition to 11,12-LT&. Among them, 11,12-LTA4 was the best substrate in terms of both K,,, and Vmax values (Table IV).
The purified enzyme, also acted on the stable xenobiotic epoxides (2,3-disubstituted oxirans) to yield corresponding trans-opening hydrolysis products (1,2-glycols). The purified enzyme did not synthesize any detectable amount of the cisopening hydrolysis products. The Vmax value for trans-(3methylstyrene oxide was 1.2 times higher than that for 11,12-LTA4 (Table IV). However, the K,,, value (833 PM) for trans-P-methylstyrene oxide was more than 1 order of magnitude higher than that for 11,12-LTG (18 PM) ( Table IV). The 2,3diphenyl-substituted oxirans, cis-and trans-stilbene oxides, were poorer substrates than 11,12-LTG (Table V). The trunsisomer was hydrolyzed 3.2 times faster than the cis-isomer by the purified enzyme. Chalcone oxide (20 PM) showed a potent inhibitory effect on the hydrolysis of 11,12-LTA4 and the xenobiotic 2,3-disubstituted oxirans, while it showed only a little effect on the hydrolysis of styrene oxide by the liver cytosol.

DISCUSSION
The natural occurrence of 11,12-diHETE has been reported (1,19), but the pathway of its formation and the absolute configuration of this compound have not so far been reported. The present study clearly showed that 11,12-diHETE was enzymically synthesized from 11,12-LT& by the action of the cytosolic epoxide hydrolase. The absolute configuration of the enzymic product was assigned as followed. The gas chromatography-mass spectrometry revealed that it was 11,12-dihydroxy-5,7,9,14-eicosatetraenoic acid (11,12-diHETE). By the NMR spectrum, the configuration of double bonds was determined as 5,14-cis-7,9-truns (Table I). The chirality of C11 and Clz was not determined. Similar to epoxide hydrolases from the mouse liver (39), the purified enzyme from the guinea pig liver catalyzed the opening of various trans-disubstituted epoxides exclusively to erythro-glycols (Table V). It is assumable, therefore, that the stereochemistry of the vicinal diol in 11,12-diHETE has also erythro configuration. Because of the proximity to the A' double bond, Cll should be more susceptible than Clz to a nucleophilic attack of water, and the orientation of the hydroxyl group at C12 will, therefore, be retained from 11,12-LTA4. Under the same assumption, Haeggstrom et al. (40) proposed the formation of (5S,6R)-diHETE by the enzymic hydrolysis of LTG. Taken together,  we concluded that the structure of the enzymically synthesized product is (11R,12S)-dihydroxy-5,14-cis-7,9-truns-eicosatetraenoic acid. It was demonstrated in the present study that 11,12-diHETE, a previously reported compound (1,19), was synthesized from 11,12-LTA4 by the epoxide hydrolase in the guinea pig liver cytosol.
Cytosolic epoxide hydrolases have been purified from livers of rabbit (36), mouse (37,38,41) and human (42,43). Although a similar enzyme activity was demonstrated in the guinea pig liver (23, 44), the enzyme was not heretofore purified and information on the molecular properties was unavailable. As different from other cytosolic epoxide hydrolases, the guinea pig enzyme was extremely hydrophobic. We succeeded in purification of the enzyme to homogeneity by the use of a detergent. The amino acid sequence of the purified enzyme (peak 16,
The M , on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (60,000) and the amino acid compositions (Table 111, in the Miniprint) are similar to those reported for rabbit (36) and mouse (37) enzymes. However, the guinea pig enzyme differs from other enzymes significantly in that it is a monomeric protein (Mr 47,000 on Superose 12 column chromatography), while enzymes from rabbit (36), mouse (37), and human (42) were reported to be a dimer (Mr > 110,000).
Isoelectric point of the guinea pig enzyme was 7.3, whereas those of rabbit (36) and mouse (37) were 6.0 and 5.5, respectively. The N-terminal amino acid was blocked for the guinea pig enzyme, while it was reported to be serine for the rabbit enzyme (36). These discrepancies might be due either to the species heterogeneity or to the presence of different enzyme in the guinea pig liver. In fact, the antibody against mouse liver cytosolic epoxide hydrolase did not react immunologically with the cytosol of guinea pig (46). Though all these enzymes are collectively termed "xenobiotic epoxide hydrolase," the K,,, values for 11,12-LTA4 and other epoxide L T  -H'  were much lower than that for tram-P-methylstyrene oxide (833 p~) , the best xenobiotic substrate among those examined in the present study (Table V). This finding indicates that the cytosolic epoxide hydrolase acts on endogenous substrates more preferably than on xenobiotic epoxides. It is of interest whether or not the previously reported xenobiotic epoxide hydrolases also catalyze the facile conversion of 11,12-LT& in to 11,12-diHETE. In addition to the above enzymic product, at least seven dihydroxy acids were produced from 11,12-LT&. One of the important findings in the present study is isolation and structural identification of 12-HETE-6-lactones formed from 11,12-LTA4 by nonenzymic acid-catalyzed rearrangement. A possible mechanism of 12-HETE-&lactone formation is depicted in Fig. 5, Scheme l . The 12-hydroxy intermediate was formed, and the carboxyl moiety could attack the carbonium ion at C5 position. On the other hand, the 5-hydroxy intermediate was formed from LTA4 (Fig. 5, Scheme Z), and the carboxyl moiety was unable to attack the Cs position. This might explain the difference in acid-catalyzed reactions between LTA, (40) and 11,12-LTA4. These lactones are useful indicators to demonstrate the formation of 11,12-LT& from arachidonic acid.
In conclusion, 11,12-LTA4 is nonenzymically converted to 5, at least seven compounds (Fig. 5): two diastereomers of 12-HETE-&lactones (C-5, C-6), three stereoisomers of 11,12-diHETE (C-3, C-3', and C-4), and two diastereomers of (5,12S)-diHETE (C-1, . (11R,12S)-Dihydroxy-5,14-ck- 7,9-tram-eicosatetraenoic acid is synthesized by the action of the cytosolic epoxide hydrolase purified from the guinea pig liver (Fig. 5). Although the 12-lipoxygenase activity was reported in porcine liver (17), whether 11,12-LTA4 serves as a substrate of the hepatic epoxide hydrolase in uiuo remains unclear. It is possible to speculate that the enzyme is involved in the hydrolysis of 11,12-LTA4 which has been transferred from the circulating leukocytes/platelets. An analogous transport of LTA, was reported from neutrophils to erythrocyte (48), to arterial endothelial cells (49) and platelets (50). The acquisition of structural information of all these compounds makes feasible analysis of the formation of 11,12-LTA4 in vivo and of the biological significance of 12-lipoxygenase pathway. -Cis-stilbene oxide, from Wako Pure Chemical (Takyal. 11.12-LTAa and 14.15-reaction mlxture contained 0.1 M Trls-HC1 buffer, pH 7.3, 10 nq/ml bovine ~s s a v of enoxide hydrolase with 11.12-LTh as d substrate. The standard serum albumin and the enzyme Ln a total volume of 0.05 ml. After preincubation for 5 mln at 3 7 W . 30 M 11.12-LTA4 11.5 "mol d1smlved I" 1 mi" at 37%. the reaction was terrnlnated by the additmn Of 100 1 of u l of ethanol) was added to initlat2the Ieactlon. After incubation for 1 solurlon A l a c e t a n l t r i l e l m e t h a n o l / a c e t i c acid. 150:50:3 (v/v/ul:containing 0.3 nmol of PGB2 a 5 an internal standard). The mixture was kept at -2O0C for at least 30 m l n . fallowed by centrlfuqatlon at 10.000 x g for 10 m~n . TO 120 V I of the supernatant. 20 ill of BDlYtlOn B (0.15 9 EDTA Na2I vas added. and its allguot (50 "11 w a s analyzed by reversed-phase HPLC. The column lTSK ODS-80TH, 0.46 x 15 cm, ToSOh, Tokyo) vas equipped with a Jasco TRI-ROTER 6 system 1Japan S p e C t r~s~o p~C , Tokyo), and eluted vlth Solvent I 112) ~a c e t o n~t r~~e l m e t h a n o l / w a t e r / a c e t i c acid. 3:1:3:0.006, v l v l v / v . 0.05 % EDTA Na2, pH adlusted to 5.6 vlth annon~al at a flaw rate of 1 mllrnin at 3 5 T . The absorbance at 270 nm was monitored, and the amount of 11.12-dlHETE IC-3 ~n Fig. 1 ) formed was calculated from the peak area ratlo, Shlrnadzu, Kyotol. The Proteln concentratlon -1% determined by the method 11,12-dLHETE/PGB2. using a data process~nq system 1Chromatpac C -R 3 A , Of Lowry a 1291 ulth bovine serum album," as standard.