Chem. Res. Toxicol. 1997, 10, 767-771
767
Mechanism of Cytochrome P450-Catalyzed Aromatic Hydroxylation of Estrogens Stephen F. Sarabia, Bao Ting Zhu,‡ Takao Kurosawa,† Masahiko Tohma,† and Joachim G. Liehr* Department of Pharmacology and Toxicology, The University of Texas Medical Branch, Galveston, Texas 77555-1031, and Faculty of Pharmaceutical Sciences, Health Sciences University of Hokkaido, Ishikari-Tobetsu, Hokkaido 061-02, Japan Received February 14, 1997X
The mechanism of aromatic hydroxylation of estrogens by cytochrome P450 enzymes has been examined by comparing the oxidation of estrone with that of substrates carrying additional aromaticity such as equilenin and the structural analog 2-naphthol. Hamster liver microsomes preferentially catalyzed the conversion of estrone to 2-hydroxyestrone (Km ) 30 and 25 µM and Vmax ) 1497 and 900 pmol (mg of protein)-1 min-1 for 2- and 4-hydroxyestrone formation, respectively). In contrast, equilenin was hydroxylated exclusively at C-4 of the steroid ring system and 2-naphthol at the corresponding C-1 position (Km ) 67 and 42 µM and Vmax ) 2083 and 3226 pmol (mg of protein)-1 min-1 for 4-hydroxyequilenin and 1,2-dihydroxynaphthalene formation, respectively). This shift in the specificity of hydroxylation was due to the introduction of additional aromaticity at ring B of equilenin, because hamster liver microsomes are known not to contain any estrogen-4-hydroxylase, only estrogen-2-hydroxylase activity catalyzed by cytochrome P450 3A family enzymes. The exclusive 4-hydroxylation of equilenin is proposed to be due to a preferred delocalization of the naphthoxy radical, an intermediate in the hydroxylation, to C-4, whereas delocalization to C-2 requires additional activation energy and is energetically not favored. Based on these electronic considerations, a mechanism of aromatic hydroxylation of estrogens is proposed which features hydrogen abstraction from the phenolic hydroxy group, electron delocalization of the phenoxy radical to a carbon-centered radical, and subsequent formation of catechol metabolites by hydroxy radical addition at C-2 or C-4 depending on steric or electronic constraints.
Introduction In most species, steroidal estrogens are metabolized primarily to 2-hydroxylated catecholestrogens (1). For instance, cytochrome P450 (P450)1 3A enzymes of human or hamster liver convert estradiol to mainly 2-hydroxyestradiol accompanied by small amounts of 4-hydroxyestradiol (2, 3). The concomitant formation of 4-hydroxyestradiol in this metabolic conversion is due to a lack of specificity of the hepatic estrogen 2-hydroxylase(s), because the formation of both catechols responds coordinately to the same inhibitors (4). In extrahepatic tissues, P450 1A enzymes also catalyze this metabolic reaction (5, 6). In several organs, a new family of enzymes, P450 1B, has been found to convert estrogens specifically to 4-hydroxylated catechol metabolites (79). The exact mechanism of P450 oxidation of estrogens and the forces that determine the stereochemistry of 2or 4-hydroxylation of estrogens have not been well defined. Several mechanisms of aromatic hydroxylation of estrogens or phenols in general have been advanced previously. The intermediate formation of epoxides has * To whom correspondence and reprint requests should be addressed. † Health Sciences University of Hokkaido. ‡ Present address: Laboratory for Cancer Research, Department of Chemical Biology, Rutgers-The State University of New Jersey, Piscataway, NJ 08855-0789. X Abstract published in Advance ACS Abstracts, June 15, 1997. 1 Abbreviations: cytochrome P450, P450; 4-(2-hydroxyethyl)-1piperazineethanesulfonic acid, Hepes; heptafluorobutyric anhydride, HFBA; N,O-bis(trimethylsilyl)trifluoroacetamide, BSTFA.
S0893-228x(97)00021-0 CCC: $14.00
been postulated by analogy to the hydroxylation of a variety of aromatic hydrocarbons (Figure 1, pathway A) (10, 11). These epoxides were thought to rearrange to phenols by analogy to phenolic hydrocarbon metabolite formation (10-12). In support of this proposal, synthetic estrogen epoxyenones have been prepared, which are unstable and rearrange in aqueous media to 2- or 4-hydroxylated estrogens (10). The direct insertion of oxygen into the carbon-hydrogen bond at the site of hydroxylation has also been entertained as a more common mechanism of hydroxylation of aromatic systems (Figure 1, pathway B) (13). More recently, P450 catalysis has been postulated to involve an initial abstraction of a hydrogen atom from the hydroxyl group of the phenol and subsequent electron spin delocalization to a carboncentered radical ortho to the phenolic substituent (14, 15). Recombination of this postulated radical intermediate with a hydroxyl radical has been thought to result in catechol product formation (Figure 1, pathway C). One strategy for the elucidation of the mechanism of P450 catalysis is to examine the regiospecificity of hydroxylation of estrogen substrates or analogs with additional aromaticity in ring B influencing the site of oxidation and then to compare results with the preferential 2-hydroxylation of estrone, a monoaromatic (ring A) system (Figure 2). The aromaticity of the B ring of equilenin is expected to affect the site of hydroxylation of this estrogen if P450 catalysis involves radical intermediates. In contrast, mechanisms of P450 catalysis via epoxidation or direct insertion (Figure 1, pathways A and B, respectively) are expected to form both 2- and 4-hydroxyequilenin in a comparable ratio as with estrone, © 1997 American Chemical Society
768 Chem. Res. Toxicol., Vol. 10, No. 7, 1997
Sarabia et al.
Figure 1. Mechanisms of cytochrome P450-mediated aromatic hydroxylation of estrone, 1. In pathway A, the epoxides 2 and 3 postulated previously (10, 11) are shown in brackets, because they have not been identified as intermediates in the P450-catalyzed oxidation. Intermediates in the direct insertion of oxygen into corresponding C-H bonds in the formation of 2- or 4-hydroxyestrone, 7 and 8, respectively, are not known (pathway B). In pathway C, the phenoxy radical 4 or its resonance structures 5 and 6 have been postulated as intermediates of aromatic hydroxylation (14, 15) but have not been identified.
Figure 2. Chemical structures of substrates used in this study. The identification of carbons and ring systems is consistent with estrogen nomenclature as used by Martucci and Fishman (1).
because ring B aromaticity may not affect either reaction. In addition to the aromatic hydroxylation of estrone and equilenin, we also used 2-naphthol (a structural analog of equilenin, shown in Figure 2) as substrate to help elucidate the catalytic mechanism of catechol estrogen formation.
Experimental Procedures Chemicals. Estrone, 2-hydroxyestrone, 4-hydroxyestrone, and 7-methylestradiol were purchased from Steraloids (Wilton, NH). Equilenin, 2-naphthol, 1,2- and 2,3-dihydroxynaphthalene, 2,3-dihydroxybenzoic acid, and Fremy’s Salt were obtained from Aldrich Chemical Co. (Milwaukee, WI). Heptafluorobutyric anhydride (HFBA), N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) with 1% trimethylchlorosilane, and pyridine were purchased from Pierce Chemical Co. (Rockford, IL). All other chemicals were obtained from Sigma Chemical Co. (St. Louis, MO). All glassware was silylated with 1% dimethylchlorosilane in toluene. Steroid Synthesis. 4-Hydroxyequilenin was synthesized from equilenin by a modification of the procedure of Teuber (16).
Briefly, equilenin was oxidized with Fremy’s salt to 3,4-equilenin quinone, which was isolated by TLC and reduced with ascorbic acid to 4-hydroxyequilenin. 2-Hydroxyequilenin was synthesized from nonsteroidal precursors and 2-fluoro-4-hydroxyestradiol from 2-fluoroestradiol as described previously (17, 18). The purity of all synthesized steroids was confirmed by gas chromatography-mass spectrometric analysis. Preparation of Microsomes. Six-week-old male Syrian hamsters were purchased from SASCO (Houston, TX). Liver tissue from four to six animals was pooled, and microsomes were prepared by differential centrifugations according to the method of Dignam and Strobel (19). Microsomal protein concentrations were determined using a bicinchoninic acid assay kit (Pierce). Microsomes were stored at -80 °C prior to use. Microsomal Incubations. As described previously (4, 18), the microsomal incubation mixture consisted of 0.5 mg/mL microsomal protein, various concentrations of estrone, equilenin, or 2-naphthol as substrate, 5 mM NADPH, and 5 mM ascorbic acid in a 100 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (Hepes)-Tris HCl buffer (pH 7.4). Incubations were carried out at 30 °C for 10-20 min and were arrested by placing the tubes on ice. An internal standard (i.e., 7-methylestradiol for estrone metabolites, 2-fluoro-4-hydroxyestradiol for equilenin metabolites, or 2,3-dihydroxybenzoic acid for 2-naphthol metabolites) was added to each reaction mixture. Catechol metabolites were extracted with ethyl acetate, and the organic extracts were dried in 1 mL reaction vials under a stream of nitrogen. Gas Chromatographic Analysis of Estrogen and Naphthol Metabolites. Catechol metabolites were assayed as described previously (20). Briefly, dried catechol estrogen extracts were redissolved in ethyl acetate and acylated with HFBA. Dried naphthol metabolite extracts were redissolved in pyridine and silylated with BSTFA containing 1% trimethylchlorosilane. Derivatized metabolites were separated on a DB17 capillary column (J&W Scientific, Folsom, CA). Chromatographic analyses were obtained with a Hewlett-Packard model 5890A gas chromatograph equipped with an electron-capture detector or a flame-ionization detector for catechol estrogen or dihydroxynaphthalene derivatives, respectively. Both detector systems were interfaced with Hewlett-Packard model 3393A
P450-Catalyzed Catechol Estrogen Formation
Chem. Res. Toxicol., Vol. 10, No. 7, 1997 769 Table 1. Kinetic Parameters of the NADPH-Dependent Hydroxylation of Estrone, Equilenin, and 2-Naphthol by Hamster Liver Microsomesa substrate
Km
Vmax
2-hydroxyestrogen estrone 30 1497 equilenin ND ND 2,3-dihydroxynaphthalene 2-naphthol ND ND
Km
Vmax
4-hydroxyestrogen 25 900 67 2083 1,2-dihydroxynaphthalene 42 3226
a Lineweaver-Burk plots of data presented in Figure 3 were fitted by linear regression to calculate Km and Vmax values, expressed as µM and pmol (mg of protein)-1 min-1, respectively (ND ) not detectable).
Table 2. Inhibition of the NADPH-Dependent Formation of 1,2-Dihydroxynaphthalene by Hamster Liver Microsomesa 1,2-dihydroxynaphthalene conditions control SKF-525A (100 µM) CO:O2 (100:0) boiled microsomes omission of NADPH
pmol (mg of protein)-1 min-1 % of control 1261 ( 209 533 ( 89* 109 ( 30*
