Cytochrome P-450 enzyme-specific control of the regio- and enantiofacial selectivity of the microsomal arachidonic acid epoxygenase.

Chiral analysis of the rat liver microsomal arachidonic acid epoxygenase metabolites shows enantioselective formation of 8,9-, 11,12-, and 14,15-cis-epoxyeicosatrienoic acids in an approximately 2:1, 4:1, and 2:1 ratio of antipodes, respectively. Animal treatment with the cytochrome P-450 inducer phenobarbital increased the overall enantiofacial selectivity of the microsomal epoxygenase and caused a concomitant inversion in the absolute configurations of its metabolites. These effects of phenobarbital were time-dependent and temporally linked to increases in the concentration of microsomal cytochrome P-450 enzymes. Reconstitution of the epoxygenase reaction utilizing several purified cytochrome P-450 demonstrated that the asymmetry of epoxidation is under cytochrome P-450 enzyme control. These results established that the chirality of the hepatic arachidonic acid epoxygenase is under regulatory control and confirm cytochromes P-450 IIB1 and IIB2 as two of the endogenous epoxygenases induced in vivo by phenobarbital.


Chiral
analysis of the rat liver microsomal arachidonic acid epoxygenase metabolites shows enantioselective formation of 8,9-, 11,12-, and 14,16-cis-epoxyeicosatrienoic acids in an approximately 2: 1,4: 1, and 2: 1 ratio of antipodes, respectively. Animal treatment with the cytochrome P-450 inducer phenobarbital increased the overall enantiofacial selectivity of the microsomal epoxygenase and caused a concomitant inversion in the absolute configurations of its metabolites. These effects of phenobarbital were time-dependent and temporally linked to increases in the concentration of microsomal cytochrome P-450 enzymes.
Reconstitution of the epoxygenase reaction utilizing several purified cytochromes P-450 demonstrated that the asymmetry of epoxidation is under cytochrome P-450 enzyme control.
These results established that the chirality of the hepatic arachidonic acid epoxygenase is under regulatory control and confirm cytochromes P-450 IIBl and IIB2 as two of the endogenous epoxygenases induced in vivo by phenobarbital.
The biological significance of arachidonic acid as the precursor for several physiologically important oxygenated metabolites is well established (1,2). Upon release from cellular glycerolipid stores, the fatty acid is metabolized by cyclooxygenase or lipoxygenases to a variety of lipid mediators of cell and tissue function (1). The initial report by our group (3) and others (4, 5) of a role for cytochrome P-450 (P-450)',' in the in vitro catalysis of arachidonic acid oxygenation has stimulated interest in its potential physiological role as a participant in the endogenous arachidonic acid cascade (1,2,6). The microsomal cytochrome P-450 arachidonic acid epox- ygenase catalyzes the NADPH-dependent epoxidation of the fatty acid to 5,6-, 8,9-, 11,12-, and 14,15-epoxyeicosatrienoic acids (EETs) (7). The EETs exhibit several potent in vitro biological activities (1, 2). Based on the documentation of the EETs as endogenous constituents of rat liver (8,9) and of rabbit kidney (10) and human urine (ll), we proposed the epoxygenase reaction as an additional member of the arachidonate cascade (8). Recently, this view has been substantiated by the high chirality of endogenous hepatic EETs (6). A functional role for P-450 as the endogenous epoxygenase was suggested by induction studies demonstrating a remarkable in uivo control of EET regio-and enantioselectivity by changes in tissue P-450 enzyme composition (6). These studies suggested a hitherto unrecognized function for this ubiquitous and important group of heme proteins in the control of cell and tissue homeostasis. As an important tool for studies dealing with the molecular characterization, regulation, and physiological significance of arachidonic acid epoxygenase, we examined the stereochemical characteristics of its products and demonstrate herein that their chirality is under P-450 enzyme control. and NADPH-cytochrome P-450 reductase were purified to homogeneity as described (16,19). The metabolism of arachidonic acid by purified P-450s was reconstituted as described (12). Briefly, P-450, NADPH-cytochrome P-450 reductase, and cytochrome b6 (1 PM each, final concentration) were mixed in the presence of sonicated ~-adilauroyl-sn-glycero-3-phosphocholine (200 pg/ml). After lo-15 min at 5 "C, the mixture was diluted &fold with 50 rnM Tris-Cl buffer (pH 7.5) containing 150 mM KCI, 10 mM MgC&, 8 mM sodium isocitrate, 0.5 IU isocitrate dehydrogenase, and 10% (v/v) gIycero1. Addition of [I-W]sodium arachidonate (47 pCi/,umol, 50 PM final concentration) was followed by initiation with NADPH (1 mM final concentration). Incubations were done under air with constant mixing at 30 "C. At different time periods, samples of the reaction mixtures were extracted with ethyl ether and the organic soluble products dried under an argon stream.

Microsomal
EET Purification and Deriuatization-The reaction products generated by microsomal incubates or reconstituted P-450 systems were initially resolved by reverse phase HPLC on a 5-pm Dynamax Micro- Baker Chemical Co.) exactly as described (20). Absolute configurations were assigned by chromatographic comparisons with enantiomerically pure 14,15-EET-Me and 11,12-and 8,9-EET-PFB prepared by total chemical synthesis (21,22). The individual EET derivabives, dissolved in mobile phase, were injected onto the HPLC column and the eluents monitored at 210 nm (20). The optical antipodes of each EET derivative were individually collected from the HPLC column based on their UV elution protile, dried under an argon stream, and quantified by liquid scintillation spectrometry. Recovery of the radioactivity injected onto the Chiralcel columns was, in all cases, ~85% cm RESULTS AND DISCUSSION An essential requirement for the delineation of the potential functional significance and regulation of the arachidonic acid epoxygenase is the characterization of those P-450 forms involved in the bioactivation of the fatty acid. A detailed knowledge of the stereochemistry of the epoxygenase metabolites would be a unique aid for the characterization of the relevant enzyme forms (23)(24)(25). Consequently, we have utilized a simple nondestructive chromatographic method, recently developed in these laboratories, for the enantiomeric resolution of all four regioisomeric EETs (20) to profile the chirality of the epoxygenase metabolites produced by rat liver microsomes.
For these studies, the incubation conditions favored primary metabolism, i.e. the initial oxidation products do not undergo significant further metabolism (12). With low microsomal protein concentrations (~0.70 mg/ml) and short incubation times (510 min at 30 "C), enzymatic and/or chemical hydration of 8,9-, l&12-, and 14,15-EET was minimal and did not significantly alter EET concentration and/or recovery (12, 13). On the other hand, the proximity of the carboxylic acid to the 5,6-oxido ring facilitates spontaneous hydration and lactonization. Consequently, the labile 5,6-EET was recovered in poor and variable yields, and it was not further investigated. In control experiments, it was determined that neither substrate concentration (l-50 FM arachidonic acid) nor incubation time (2-10 min) had significant effects in EET stereochemical purity or absolute configuration. As previously reported, olefin epoxidation accounts for approximately 64% of the total arachidonic acid metabolism by uninduced rat liver microsomal fractions (26) ( Table I, Fig.  I). Generation of 11,12-EET, the predominant product of the microsomal epoxygenase, proceeded with high enantiofacial selectivity (81% (llR,12S)). On the other hand, the microsomal enzyme system displayed lesser stereoselectivity toward the 8,9-(68% (8R,9S)) and 14,15-olefins (67% (14S,15R)) ( Table I). Significantly, the in vitro stereospecificity of the microsomal system ( Table I) was opposite that of the in uiuo epoxygenase (6) and that of a purified rat liver P-450 (form IIBl) (25). These results clearly suggested the presence in the microsomal fractions of P-450 enzyme forms with distinct enantiofacial selectivities.
The rat liver endoplasmic reticulum houses a multiplicity Microsomal fractions (0.5 mg of protein/ml) were incubated for 5 min at 30 "C with [I-"Clarachidonic acid (100 PM) in the presence of NADPH (1 mM) and an NADPH-regenerating system. The ethyl ether soluble products were dissolved in ethanol and injected onto a 5-&m Tables I and II. When compared with uninduced controls (Table I), fi-NF treatment resulted in overall decreases in EET optical purity (Table II). These results suggested a net loss of unique hepatic P-450 epoxygenase enzyme(s) with distinct regio-and stereoselectivities.
A similar analysis done with microsomal fractions isolated from the livers of PB-induced animals is illustrated in Fig. 1 and Table III. Parallel to a nearly 1.7-fold increase in the catalytic rate of the arachidonic acid epoxygenase (Tables I and III) (26), PB treatment was accompanied by (a) an overall increase in the optical purity of the epoxygenase products (Tables I and III) and (b) a remarkable inversion in the absolute configurations of 8,9-, 11,12-, and 14,15-EET (Table III). Importantly, the chirality of the EETs formed in vitro by the PB-induced microsomal arachidonic acid epoxygenase (Table III) is similar to that of the EETs present in uiuo in the livers of control and PB-induced animals (6).
The coexistence in the microsomal membrane of multiple P-450 forms with unique enantioselective properties is more  clearly illustrated when the effect of the inducers is analyzed in terms of the rate of formation of each individual EET enantiomer.
While P-NF decreased the rates of formation of both enantiomers of 8,9-, 11,12-, and 14,15-EET to similar extents (Fig. 2, (Fig. 2). Therefore, PB treatment changed the overall stereoselectivity of the microsomal epoxygenase, presumably by increasing the steady state concentrations of one or more P-450 forms with a high and distinct enantiofacial selectivity toward the 8,9-, 11,12-, and 14,15-olefins, while at the same time decreasing the specific content of those forms responsible for the formation of the opposite stereoisomers (Fig. 2, A-C).
To further elucidate the role of different forms of microso-ma1 P-450 on the in vitro chirality of EET formation, adult rats were treated with PB for 1, 2, 4, and 8 days, and the effects of PB treatment duration on the asymmetry of arachidonic acid epoxidation and on the concentrations of total and of P-450s IIBl and IIBZ were compared with that of nontreated animals. After an almost linear rise, the specific content of microsomal P-450 reached a maximum at the 4th day of PB treatment and remained essentially constant thereafter (0.9 and 2.4 nmol of P-450/mg of protein for control and PB microsomes, respectively) (Fig. 3A). However, inmunoblot analysis showed a rapid increase in the concentration of microsomal P-450s IIBl and IIBP, the major PB-inducible forms of rat liver P-450 (27), between the first and second days of PB treatment (66-fold increase over control values) (Fig. 3A). Moreover, analysis of the relationship between EET chirality and the duration of PB treatment demonstrated that a) the inversion of EET configuration described in Tables I  and II was time-dependent (Fig. 3, B-D), suggesting a correlation with changes in microsomal P-450 enzyme levels (Fig.  3A) rather than a direct effect of the barbiturate on the epoxygenase kinetic properties. b) The time course of the changes in the overall microsomal enantioselectivity was regioselective. Loss of stereoselectivity for 14,15-EET occurs after the first day of PB treatment ( Fig. 30) and between the 3rd and 4th day of treatment for 8,9-and 11,12-EET (Fig. 3, B and C). c) In contrast to the cyclooxygenase and lipoxygenase members of the arachidonate cascade, the enantioselectivity of the microsomal epoxygenase is variable and highly dependent on regulatory factors that control the inventory of P-450(s) present in the microsomal membrane. Thus, among the enzyme systems of the arachidonate cascade, the P-450 epoxygenase is unique in that its stereochemical selectivity is under regulatory control and can be experimentally altered by animal manipulation in uiuo (Fig. 3) (6, 27, 28). Fig. 4 shows the absolute configurations of the predominant enantiomers of 8,9-, 11,12-, and 14,15-EET generated by incubates containing microsomal fractions isolated from the livers of untreated animals or from animals treated with PB or &NF. Assuming a folded hairpin conformation (31,32) for the arachidonic acid molecule and a plane of symmetry established by coplanarity of its carbon atoms, the microsomal 100 80 enzymes oxygenate the 8,9-and 11,12-olefins with the same sidedness, opposite to that of the 14,15olefin. Animal treatment with PB changed those topographic features of the enzyme(s) active site critical for substrate binding without altering the above relationship, i.e. there was an apparent 180" rotation of the substrate molecule with respect to the heme-oxygenating locus.
In the last few years considerable advances have been made in the isolation and structural characterization of several forms of rat liver microsomal P-450 (27,28). Studies at the protein or gene level have documented that structurally distinct macromolecules provide the molecular basis for the catalytic heterogeneity of microsomal P-450 (27,28). The resolution of the microsomal electron transport chain into defined components that can be reconstituted into catalytically functional systems provides a unique tool for the analysis of the regio-and stereoselective properties of different heme protein forms. Therefore, the following forms of male rat liver P-450 were solubilized and purified from the livers of untreated (forms IICll and IIClZ), isosafrole (form IAZ), &NF (form IAl), or PB-treated animals (forms IIAl, IIBl and IIBZ). The regioselectivity of these different P-450 forms was then evaluated by reconstituting the arachidonic acid monooxygenase reaction in the presence of NADPH-cytochrome P-450 reductase, cytochrome b5 and NADPH. P-450 form IA1 actively catalyzes the hydroxylation of arachidonic acid at those carbons proximal to the methyl terminus (C1&,0 alcohols, 87% of total products) (30) ( Table IV). While P-450 IIBl and IIBZ metabolized the fatty acid at rates lower than IA1 (Table IV), these heme proteins were highly selective epoxygenases, generating EET mixtures as their only reaction products (Table IV and Fig. 3). P-450s IA2 and IICll displayed lesser regioselectivity generating mixtures of C&-C& alcohols, HETEs, and EETs (Table IV). Cytochromes IIAl and IICl2 did not catalyze arachidonic acid metabolism to a significant extent.
The epoxygenase metabolites generated by P-450s IA2, IAl, IICll, IIBl, and IIBP (21, 6, 62, 100, and 100% of the total reaction products, respectively) were resolved into the corresponding EET regioisomers by normal phase HPLC and then  Values shown are the means calculated from three different experiments with S.E. ~15% of the mean. Product distribution is expressed as percent of the total epoxygenase activity. Reaction rates are given in nanomoles of EETs formed per min/nmol of purified P-450 form at 30 "C. quantified by liquid scintillation spectrometry. A comparative analysis of the relative epoxygenase activities showed that none of the five reconstituted enzymes displayed selectivity for the epoxygenation of a single olefin with exclusion of the others. Thus, in addition to a general low reactivity at the 5,6-oletin, all forms catalyzed epoxidation at the 8,9-, 11,12-, and 14,15-olefins (Table V). P-450 IA2 and IA1 displayed the highest degree of regioselectivity with 11,12-and 14,15-EET accounting for 58% (form IA2) and 61% (form IAl) of their total epoxygenase activity (Table V). On the other hand, those P-450 forms with the highest epoxygenase turnover numbers, i.e. IICll, IIBl, and IIB2, catalyzed 8,9-, 11,12-, and 14,15-EET formation as nearly equimolar mixtures (Table V). The contribution of individual P-450 forms to the overall stereochemical properties of the microsomal arachidonic acid epoxygenase was studied by characterizing the chirality of the EETs generated by reconstituted systems containing P-4509 IA2, IAl, IICll, IIBl, and IIBB. To facilitate comparisons, the data in Tables VI-VIII  illustrate the stereochemistry  of  the metabolites  resulting from epoxidation  at the fatty acid  8,9-, 11,12-, and 14,15-olefins. P-450s IA2, IAl, IIBl, and IIB2  epoxidized  the 8,9-olefin  stereospecifically,  generating  (8S,9R)-EET with 93, 87, 86, and 90% purity, respectively (Table VI). Significantly, with these four heme proteins, substrate binding results in oxygen insertion at the same si,reface of the 8,9 double bond. On the other hand, catalysis by the rat liver constitutive P-450 IICll produced nearly racemic 8,9-EET (Table VI). With the exception of P-450 IICll, which formed 11,12-EET in a roughly 1:l ratio of antipodes, epoxidation of the 11,12 double bond also proceeded with high enantiofacial selectivity. The (IlS,lBR)-EET enantiomer was generated by P-450s IAl, IIBl, and IIB2 with ~84% purity (Table VII). Importantly, the P-450 IA2 arachidonic acid 11,12-epoxygenase was unique. This enzyme catalyzed highly asymmetric oxygenation at the opposite enantiotopic face of the 11,12-olefin (re,si-face) with (IIR,IPS)-EET formed in 95% purity (Table VII) %  IA2  20  63  37  IA1  10  85  15  IICll  120  49  51  IIBl  110  64  36  IIB2  30  65  35   by guest on March 24, 2020 http://www.jbc.org/ Downloaded from 6:l ratio of antipodes (Table VIII). On the other hand, forms IA2, IIBl, and IIB2 showed a generalized lower enantiofacial selectivity toward the 14,lSolefin when compared with the 8,9-and 11,12-double bonds (Table VIII). With the exclusion  of P-450 IICll, which produced nearly racemic mixtures of  14,15EET, all four forms catalyzed preferential oxygen addition to the re,si-face of the 14,15-double bond.3 To date, no crystaIlographic information is available for any microsomal P-450. Information about the topology of the heme environment and of the active site has been approached by indirect observations of substrate selectivity and regioand stereochemical product analysis. A role for the porphyrin propionic acid side chains in substrate anchoring via Hbonding or cation chelate has been proposed (33). Importantly, methyl arachidonate is not a substrate for P-450 epoxidation nor does the addition of the ester to a microsomal suspension produce any spectral manifestation of active site binding (30). Jerina et al. (34) have proposed the presence of a hydrophobic depression in the active site of P-450. The absolute orientation of the prosthetic heme group for the major PB-inducible form of rat liver P-450 has been reported (35). Additionally, it has been suggested that the active site has a lipophilic binding area over pyrrole ring C and that ring B is sterically encumbered (36). With the exception of P-450 IIC11," the stereochemical properties of the EETs formed by either microsomal or purified P-450s revealed an unprecedented high stereoselectivity for the oxidation of such an unbiased acyclic molecule. The above indicates that the active site molecular coordinates responsible for heme-fatty acid spatial orientation are remarkably rigid and structured. As discussed for the microsomal enzymes, purified P-450s IIBl, IIB2, and IA1 epoxidized the arachidonate 8,9-and 11,12olefins with the same enantiotopic sidedness and opposite to that of the 14,15-olefin (Fig. 4). Since in reconstituted systems the asymmetry of the epoxygenase reaction is under the control of a single protein catalyst, we propose, therefore, a very similar or common geometry for the substrate binding sites of cytochromes P-450 IIBl, IIB2, and IA1 and significantly different from that of form IA2 (Tables VI-VIII). The experimental data presented provide a coherent explanation for the effects of PB induction on the enantiofacial selectivity of the microsomal epoxygenase(s), i.e. the EET stereochemical inversion correlates with net increases in the specific contents of P-450s IIBl and IIB2. On the other hand, with the exception of the 11,12-epoxygenase activity of P-450 IA2, none of the purified enzymes studied appeared to contribute significantly to the chirality of the EETs generated by P-NF induced or non-induced microsomal fractions. These results suggest the presence in the microsomal membrane of yet unidentified P-450 epoxygenases with unique enantiofacial selectivities.
The biological relevance of the epoxygenase reaction has been established by the demonstration of the in uiuo biosynthetic origin of its reaction products (6). Thus, epoxidation of arachidonic acid represents a novel route for the oxidation of the fatty acid and, more importantly, for the generation of novel pools of membrane phospholipids containing esterified EETs moieties (9). Efforts to demonstrate enzymatic, NADPH-dependent, or NADPH-independent epoxidation of " The racemic nature of the products of the P-450 IICll reaction requires the caveat that the enzyme preparation was not analyzed for arachidonyl phosphatidylinositol or arachidonyl phosphatidylcholine by rat liver homogenates or microsomal or cytosolic fractions proved unsuccessful (37). The available evidence supports a sequential mechanism in which P-450 epoxidation of free arachidonic acid is followed by subsequent EET acylation. The in vitro stereochemical properties of the P-450 epoxygenases reported here suggest a role for P-450 forms IIBl and IIB2 in the enzymology of rat liver in uiuo EET formation (6). The demonstration of an enzymatic route for in uiuo formation of oxidized phospholipids has significant implications for cell membrane physiology and for a potentially decisive role for P-450 in its control. Alterations in the fatty acid composition of membrane phospholipids have important consequences for the structural integrity and the functional properties of cellular membranes (38). The oxidation of phospholipid-bound fatty acids has documented effects on membrane properties including changes in membrane ion permeability (39), alterations in the enzymatic activity of several membrane-bound enzymes (40,41), as well as changes in membrane fluidity and fusogenic properties (42, 43). From the foregoing evidence as well as published data (6,25), we conclude that the regio-and enantioselectivity of the arachidonic acid epoxygenase is cytochrome P-450 proteindependent. Additionally, the data show that, in contrast to the lipoxidase and cyclooxygenase members of the arachidonate cascade, the enantioselectivity of the epoxygenase is under in uiuo regulatory control. These observations as well as the documented presence in rat liver of unique phospholipids containing an EET-esterified moiety suggest important functional roles for these prominent heme proteins in controlling membrane physicochemical properties and thus function.