- Email: [email protected]
Human liver microsomal steroid metabolism: Identification of the major microsomal steroid hormone 6β-hydroxylase cytochrome P-450 enzyme
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 263, No. 2, June, pp. 424-436,198s
Human Liver Microsomal Steroid Metabolism: Identification of the Major Microsomal Steroid Hormone G@Hydroxylase Cytochrome P-450 Enzyme’ DAVID
J. WAXMAN,**’
CYNTHIA ATTISANO,* F. PETER DAVID P. LAPENSON*
GUENGERICH,?
AND
*Depart-t of Biological Chemistry and ikiolecular Pharmxzcobg~ and Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115 and +Department of Biochenzistry and Center in Molecular Toxicology, Vanderbilt University Sch.ool of Medicine, Nashville, Tennessee $7282 Received
December
10, 198’7, and in revised
form
February
22,1933
Cytochrome P-450-dependent steroid hormone metabolism was studied in isolated human liver microsomal fractions. S/3hydroxylation was shown to be the major route of NADPH-dependent oxidative metabolism (375% of total hydroxylated metabolites) with each of three steroid substrates, testosterone, androstenedione, and progesterone. With testosterone, 2/3 and 15p hydroxylation also occurred, proceeding at -10% and 3-4% the rate of microsomal S/3hydroxylation, respectively, in each of the liver samples examined. Rates for the three steroid G/3-hydroxylase activities were highly correlated with each other (T = 0.95-0.97 for 25 individual microsomal preparations), suggesting that a single human liver P-450 enzyme is the principal microsomal 6p-hydroxylase catalyst with all three steroid substrates. Steroid G/3-hydroxylase rates correlated well with the specific content of human P-45&F (r = 0.69-0.83) and with its associated nifedipine oxidase activity (r = 0.80), but not with the rates for debrisoquine 4-hydroxylase, phenacetin 0-deethylase, or S-mephenytoin 4-hydroxylase activities or the specific contents of their respective associated P-450 forms in these same liver microsomes (r < 0.2). These correlative observations were supported by the selective inhibition of human liver microsomal 6@hydroxylation by antibody raised to either human P-45ONr or a rat homolog, P-450 PB-2a. Anti-P-450nr also inhibited human microsomal testosterone 28 and 158 hydroxylation in parallel to the 6@-hydroxylation reaction. This antibody also inhibited rat P-450 Ba-dependent steroid hormone 6@hydroxylation in uninduced adult male rat liver microsomes but not the steroid ICY,16a, or 7a hydroxylation reactions catalyzed by other rat P-450 forms. Finally, steroid 6/3 hydroxylation catalyzed by either human or rat liver microsomes was selectively inhibited by NADPH-dependent complexation of the macrolide antibiotic triacetyloleandomycin, a reaction that is characteristic of members of the P-450~~ gene subfamily (P-450 IIIA subfamily). These observations establish that P-450NF or a closely related enzyme is the major catalyst of steroid hormone 6s hydroxylation in human liver microsomes, and furthermore suggest that steroid 6/3 hydroxylation may provide a useful, noninvasive monitor for the monooxygenase activity of this hepatic P-450 form. o 1988 Academic press, I”~.
Steroid hormones are hydroxylated hepatic cytochrome P-450 (P-45O)3
‘Supported in part by Grants DK33765 (D.J.W.) and CA 30907 (F.P.G.) from the National Institutes of Health. ‘To whom correspondence should be addressed at: Dana-Farber Cancer Institute, JF-525, 44 Binney Street, Boston, MA 02115. 0003-9861/88 Copyright All rights
$3.00
0 1988 by Academic Press, Inc. of reproduction in any form reserved.
s Abbreviations used: TAO, triacetyloleandomycin; body. 424
P-450, cytochrome MAb, monoclonal
by enP-450; anti-
HUMAN
LIVER
STEROID
HORMONE
zymes in a highly regioselective and stereospecific manner. Studies carried out in the rat have established that P-450-dependent hydroxylation of these endogenous substrates can occur at multiple sites, with the relative rates of hydroxylation at each site dependent on the steroid and on the monooxygenase induction status of the animal (1,2). Steroids such as testosterone and androstenedione are hydroxylated by purified P-450 enzymes to yield patterns of metabolites that are often characteristic of individual P-450 forms (3-5). These studies have not only aided in the characterization of P-450 enzymes that exhibit similar spectral, immunochemical, and enzymatic properties using drug substrates, but have also led to the identification of specific rat hepatic P-450 catalysts of microsomal steroid hydroxylase activities subject to developmental control and hormonal regulation. Rat P-450 form 2c4 was thus shown to correspond to the major steroid hormone 16a-hydroxylase induced at puberty in male but not female rats (6, 7, 7a), P-450 2a to the adult male-specific G/3-hydroxylase that is developmentally suppressed in
4 The nomenclature used in this study to designate rat P-450 enzymes is that described previously (6). For a more complete listing of other enzyme designations see reviews by Waxman (19) and Guengerich (44). In addition, the term P-450 2a is used to designate the constitutive, hormonally regulated rat steroid Go-hydroxylase, whose regulation is consistent with that determined using the cDNA clone PCN2 (20). P-450 PB-2a, isolated from phenobarbital-induced adult male rats (6), appears to be immunochemically related to the phenobarbital and/or synthetic steroid inducible form(s) termed p, PCN-E, PCNl, PCNa, or PCNb by other investigators (11,13, 20,27). Recent studies (46) suggest that the inducible proteins PB-2a, PCN-E, and PCNb may all correspond to the same P-450 form, while the preparations designated p, PCNl, and PCNa may all correspond to a second inducible form. These two groups of inducible P-4509 are generically refered to as “immunoreactive P-450 PB-2a” in the current study. The relationship between the two purified human members of this gene family, P-450~~ (13) and P-450 HLp (11) is not known with certainty. The genes for these rat and human P-450s all appear to be in the P-450 IIIA gene subfamily (28).
GB-HYDROXYLASE
P-450
425
maturing female rats (8), and P-450 2d to the adult female-specific steroid disulfate 15@hydroxylase (9). Multiple hepatic P-450 enzymes have been purified from human liver and characterized extensively using xenobiotic substrates (e.g., (10-14)). These studies have been carried out with the primary goal of identifying those forms that contribute to the genetically polymorphic oxidation of drugs such as debrisoquine and mephenytoin. Much less is known, however, about the possible participation of these or other human P-450s in the metabolism of endogenous substrates such as steroid hormones. In the case of testosterone, polyclonal antibody inhibition experiments suggest that the human liver microsomal Go-hydroxylase reaction (15) may be catalyzed by an enzyme designated P-45ONF (13). However, in view of the very low testosterone 6P-hydroxylase activity of purified P-450~~ (13) these observations must be viewed as tentative in the absence of additional experimental data. The present study was therefore undertaken to characterize in greater detail the pathways of NADPH-dependent testosterone metabolism catalyzed by human liver microsomes, and to examine the possible contributions of P-450NF and other human liver P-450 enzymes to the metabolism of testosterone, as well as androstenedione and progesterone. Results described in this report establish that 6/3 hydroxylation is the major pathway of oxidative metabolism for the three steroid hormones, with the three respective 60 hydroxylation rates shown to be highly correlated with each other in a collection of 25 individual human liver microsomes. In the case of testosterone, hydroxylation is also demonstrated to occur at the 2p and 15/3 positions, at rates that correlate closely with the G/3-hydroxylase reactions. Finally, antibody inhibition experiments, measurements of microsomal P-450~~ specific contents and nifedipine oxidase rates, and TAO complexation experiments are presented in support of the conclusion that P-450N~ is the major 6@hydroxylase for three different steroid hormone substrates in human liver microsomes.
426
WAXMAN METHODS
ikficroaomes. Human liver microsomes were prepared from organ donors who met accidental deaths. Tissues were removed and perfused with chilled buffers within 15-30 min of death. Samples were divided into small pieces (l-6 cm3) which were frozen by immersion in liquid nitrogen and stored at -70°C. Samples were initially homogenized in 0.1 M Tris acetate (pH 7.4) containing 0.1 M KCl, 1 mM EDTA, 20 pM butylated hydroxytoluene, and 0.4 mM phenylmethylsulfonyl fluoride, with microsome isolation as described previously (10). Under these conditions, catalytically active microsomes suitable for the isolation of active P-450 enzymes are obtained (e.g., (16)). Stemid hydroxglase assays. Microsomal fractions [O.l to 0.2 mg microsomal protein/ml of 100 mM potassium phosphate buffer (pH 7.4) containing 0.1 mM EDTA and 1% glycerol (v/v)] were assayed for steroid hydroxylase activities (lo- to 20-min incubations) using as substrate [4-“Cltestosterone, [4-i4C]androstenedione, or [4-“Clprogesterone (7-10 mCi/ mmol; 25 PM, final concentration). Hydroxylated metabolites were resolved by silica gel thin-layer chromatography and then quantitated by scintillation counting as described in detail (3,6). Antibody inhibitions of steroid hydroxylase activities were determined following preincubation of microsomes with isolated IgG fractions for 45 min at 20 to 22°C. Metabolite identijcations. Steroid metabolites were identified by their cochromatography with standard hydroxylated steroids on silica gel thin layer plates developed in multiple solvent systems (3). Identifications were made with metabolites formed by human liver microsome preparations HL35 and HL37. 6fi-Hydroxytestosterone formed by the human liver microsomes was identified by chromatography in solvents AA, AB, and C, 6@-hydroxyandrostenedione in solvents AA, HH, and I, and 6B-hydroxyprogesterone in solvents K, L, and M. In addition, 26-hydroxytestosterone was identified in solvents AA and BB, while 15/3-hydroxytestosterone was identified in solvents AB and BB. [See Ref. (3) for additional details relating to product identifications]. Solvent compositions used in these experiments were: solvent A = CHzClz/acetone (4/l); solvent B = CHC&/ethyl acetate/ethanol (4/l/0.7); solvent C = CHzClJacetone (915); solvent H = CHClJethyl acetate (l/2); solvent I = ethyl acetate/n-hexane/acetic acid (19/5/l); solvent K = CHClJethanol (9/l); solvent L = ethyl acetate/nhexane/acetic acid (15/4/l); and solvent M = diethylether/acetone (4/l). Each of the microsomal steroid metabolites identified by these methods was also distinguished on the thin-layer plates from the available monohydroxy steroids listed previously (3, 6) as well as from 7/3-hydroxytestosterone and 156hydroxytestosterone, kindly provided by Searle, Inc.,
ET AL. and 15cY-hydroxyprogesterone, obtained from the Steroid Reference Collection. Other methods. P-450 concentrations, protein concentrations, metabolism of ‘I-ethoxycoumarin (17), nifedipine (13), debrisoquine, phenacetin (16), mephenytoin (14), and TAO complexation (11) were measured as reported in the indicated references. Immunoquantitation of microsomal P-450s was carried out by Western blotting of liver microsomes with polyclonal antibodies raised to the human P-450 forms designated P-450Ds, P-450pA (16), P-450~~ (13), and P-45O~ap.~(14). Western blotting was also carried out using MAb C9, raised to rat P-450 PB-2a (Ref. (18); P-450 antigen used to prepare that antibody was previously designated 2a/PCN-E) and kindly provided by Dr. S. S. Park of the NCI. This MAb was shown to cross-react on Western blots with a single polypeptide present in human liver microsomes that exhibited the same electrophoretic mobility as P-45&~. P-450N~ levels determined with this anti-rat P-450 monoclonal antibody were highly correlated with the levels determined using the anti-human P-450~~ polyclonal antibodies (r = 0.83, n = 23). Correlation coefficients of r > 0.65 are highly significant (P < 0.001) for comparisons of N = 20 to 25 independent samples. RESULTS
Testosterone metabolism catalyzed by human liver microsomal fra&on.s. Shown in Fig. 1 is a representative chromatogram of the NADPH-dependent [‘“Cltestosterone metabolism profile exhibited by liver microsomes prepared. from 6 individual human livers. Similar metabolite patterns were observed for 20 other human liver microsomal samples (not shown). In each case a single prominent hydroxylation product was formed, This metabolite pattern contrasted with the more complex profile for rat hepatic microsomes, where at least four major products are observed (Fig. 1, first lane). The major human liver microsomal testosterone metabolite was distinguished from 15 available monohydroxytestosterones and then identified as 6p-hydroxytestosterone by its cochromatography with authentic standard in three chromatography solvent systems (see Methods). Three minor human liver microsomal testosterone metabolites were respectively identified as 2P-hydroxytestosterone, 15/?hydroxytestosterone, and androstenedione using similar analytical methods.
HUMAN
LIVER
STEROID
HORMONE
427
P-450
Ratios of the three principal microsomal hydroxy testosterone metabolites were similar for each of the human liver preparations [6@/2j3/15@ = l/0.10/0.03-0.041, suggesting that they might be formed by the same P-450 enzyme (also see Table VA, below). Androstenedione formation was not observed in the absence of liver microsomes or NADPH and reflects oxidation of the substrate’s 17@hydroxy group to a ketone. Oxidation of testosterone to androstenedione has previously been observed with rat liver microsomes (e.g., Ref. (2)) as well as with several purified P-450s isolated from rat hepatic tissue (3-5). Metabolite quantitation revealed that G/3-hydroxytestosterone generally comprised from 75 to 85% of the total hydroxylated product in these human liver microsomal incubations (Table I). Although steroid 6@-hydroxylation is sexually differentiated in the rat (8), it does not appear sex-dependent in humans, since equally efficient 6P-hydroxylation was observed with several preparations of male and female human liver microsomes (Table I and data not shown).
x I
Z.aT 68,
HVMAN
6@-HYDROXYLASE
MICROSOMES
FIG. 1. Metabolism of testosterone by human liver microsomes. Shown is an autoradiograph of a silica gel thin-layer chromatogram resolving testosterone and metabolites formed by liver microsomes from six individual humans (samples HL23 through HL37). The chromatogram was developed in solvents AB and the metabolites were quantitated as described under Methods. Sample at left: metabolites from untreated adult male rat liver microsomes. Identified human liver microsomal testosterone metabolites are marked on the right and rat microsomal metabolites on the left. T, testosterone; 6j3-T, 6@-hydroxytestosterone, etc.; T,5or, Sa-dihydrotestosterone; A, androstenedione; X, unidentified.
Androstenedione tabolism. Similar
and progesterone
me-
approaches were used to characterize the NADPH-dependent metabolism of androstenedione and proges-
TABLE
I
TESTOSTERONEMETABOLISMCATALYZEDBYHUMANLIVERMICROSOMES'
Sex
HL23 HL25 HL32 HL33 HL35 HL37
Female Male Female Male Male Male
GPOH-T
1.48 0.72 0.09 0.31 0.43 0.69
2j3OH-T
0.16 0.06 0.01 0.03 0.05 0.07
15@OH-T Androstenedioneb (nmol product/min/mg protein) 0.06 0.02 10.005 0.01 0.02 0.03
0.05 0.05 0.02 0.10 0.04 0.03
T,5a”
Total activityd
0.06 0.05 0.02 0.09 0.04 0.05
1.81 0.90 0.14 0.54 0.58 0.87
a Shown are the rates of testosterone metabolite formation catalyzed by human liver microsomes prepared from donors HL23-I-IL37 and determined as described in Fig. 1 and Methods. T, testosterone. b Resulting from a P-450-catalyzed C17-oxidation. c 5cY-Androstane-3-one-17j3-ol and formed by microsomal steroid A4-3-one Sa-reductase. d Sum of the activities of the major metabolites indicated. Other metabolites (e.g., 16aOH-T) were generally formed at rates CO.01 nmol/min/mg.
428
WAXMAN
terone by human liver microsomes. With these steroids as well, 6@ hydroxylation was found to be the major route of oxidative metabolism (Fig. 2, Table II). Microsomal G/3-hydroxylase activities were highly correlated for the three steroid substrates (testosterone, androstenedione, progesterone) in the 25 human liver samples examined (r = 0.95-0.9’7) (Table III), suggesting that a single P-450 form catalyzes all three reactions. 6p hydroxylation rates were generally highest with testosterone, with activity ratios of 6@OH-testosterone/6POH-androstenedione = 1.85 f 0.58 and GPOH-testosterone/G/IOH-progesterone = 1.32 f 0.40 (means f SD) obtained for the 25 individual microsomes. Correlation between human liver steroid 6@-hydroxylase activity and P-450~~ specific content. Immunoreactive rat P-450
PB-2a has been shown to catalyze the majority of steroid hormone 6/3 hydroxylation in both induced and uninduced rat liver microsomes (8). Human liver P-450s homologous to rat P-450 PB-2a have been
ET AL.
identified and designated P-450N~ (nifedipine oxidase P-450; Ref. (13)) and HLp (ll), suggesting that these P-450s might mediate the corresponding 6s hydroxylation reactions in human liver microsomes. Consistent with this proposal, microsomal steroid 6p hydroxylation rates correlated significantly with P-450xr-dependent (13) microsomal nifedipine oxidase activity (r = 0.76-0.81) (Table III), as well as with immunochemically estimated P-450NF levels, determined using either polyclonal anti-P-450Nr antibodies (r = 0.69) or a monoclonal antibody raised to rat P-450 PB-2a (see Methods) (r = 0.83) (Table IV). Plots of microsomal testosterone 6@-hydroxylase activity versus P-450NF content yielded a slope of approximately 5 nmol GPOH-testosterone/min/nmol microsomal P-450NF (not shown), suggesting that the microsomal turnover number of this human 6P-hydroxylase P-450 is similar to that of the constitutive rat liver microsomal 6P-hydroxylase, P-450 2a (8). The y-intercept of 0.08 nmol 6B-hydroxy
Rf - 1.0
P, 5aA,50:STERONE-
ANDROSTENEOIONE-
Apolor
Tesiosteronex-
c
GPOH-Px-
6pOH-A-
-0.25 I6aOH-P-
NO $5
HL 25
HL 32
HL 33
HL 35
HL 37
NO pS
HL 25
HL 32
HL 33
HL 35
HL 37
FIG. 2. Metabolism of androstenedione (A) and progesterone (B) by human liver microsomes. Shown are autoradiograms as in Fig. 1. The chromatogram in A was developed in solvents HH and that in B in solvent I (see Methods). A, androstenedione; A,5a, 5a-dihydroandrostanedione; X, unidentified, P, progesterone; P,5o, 5a-dihydroprogesterone. “Apolar” corresponds to the region of migration of 2cy-hydroxy-, 17a-hydroxy-, 2Oa-hydroxy-, and 20j3-hydroxyprogesterones (see Ref. (6)). The microsomal metabolite migrating just below androstenedione (Rf -0.54, A) was not identified.
HUMAN TABLE
LIVER
STEROID
HORMONE
II
ANDROSTENEDIONE AND PROGESTERONE METABOLISM CATALYZED BY HUMAN LIVER MICROSOMES A. Androstenedione Testosterone* A,5ac G@OH-A” (nmol product/min/mg protein) HL25 HL32 HL33 HL35 HL37
0.31 0.04 0.20 0.26 0.46
0.17 0.26 0.33 0.14 0.18
0.07 0.05 0.15 0.07 0.06
B. Progesterone 6j30H-P”
HL25 HL32 HL33 HL35 HL37
0.34 0.03 0.16 0.51 0.74
16aOH-P Unknown’ (nmol product/min/mg 0.05 0.005 0.01 0.05 0.10
0.02 0.05 0.05 0.03 0.04
Apola# protein) 0.12 0.05 0.16 0.16 0.13
P,50ia
0.11 0.11 0.19 0.13 0.21
“66-Hydroxyandrostenedione and Gp-hydroxyprogesterone identified as described under Methods. Other hydroxylated metabolites (including 21-hydroxyprogesterone) were generally formed at rates GO.03 nmol/min/mg. P, progesterone. b Formed by reduction of the substrate’s 17-keto group to a 17/3-hydroxyl. ’ 5a-Androstane-3,17-dione and formed by microsomal steroid A’-bone 5u-reductase. d Mixture of at least two progesterone metabolites migrating on silica gel (Fig. 2B) in the region where 2aOH-P, 20oOH-P, BOBOH-P, and 17aOH-P chromatograph. ’ 5a-Pregnane-3,20-dione and formed by microsomal steroid A’-3-one 5a-reductase. ‘Metabolite X in Fig. 2B.
metabolite/min/mg microsomes suggests that relatively little 68 hydroxylation is carried out by independently regulated P-450 enzymes. Consistent with this conclusion, microsomal testosterone S/3 hydroxylation did not correlate with the specific contents of three human liver P-450 enzymes catalyzing polymorphic drug oxidation (P-45Ons, P-450PA, and P-45oyp-1) (Table IV), or with their associated enzyme activities (debrisoquine 4-hydroxylase, phenacetin 0-deethylase, and S-mephenytoin 4-hydroxylase, respectively) (Table III). It should be noted, however, that although P-450~~ levels correlated well with microsomal nifedipine oxidase activity (r = 0.79-0.81), specific contents of
6&HYDROXYLASE
P-450
429
the other three human P-450 forms did not correlate with their associated microsomal activities (r < 0; Table IV). Moreover, steroid hormone 6@hydroxylase activity also correlated with microsomal ‘7ethoxycoumarin 0-deethylase activity (r = 0.6-0.7; Table III), which is P-45ONr independent (Fig. 3; see below), as well as with NADPH P-450 reductase-catalyzed cytochrome c reduction (r - 0.6, Table III) and with total spectral P-450 content (r = 0.54; Table IV). Correlative relationships such as these should therefore be interpreted with some caution. Immunoinhibition and cross-reactivity studies of human and rat steroid hormone 6@-hydroxylases. The findings described
above suggest that P-45ONr (or a closely related P-450 form) may be the principal catalyst of steroid hormone SD hydroxylation in human liver microsomes. This possibility is supported by antibody inhibition experiments, which revealed that anti-P-45ONr IgG can inhibit ~80% of human liver microsomal6fl hydroxylation of testosterone, progesterone (Fig. 3), and androstenedione (not shown) under conditions where microsomal 7-ethoxycoumarin 0-deethylation is 120% inhibited. Anti-P-450Nr IgG also inhibited human liver microsomal testosterone 2/3 hydroxylation and testosterone 15p hydroxylation in parallel to testosterone 6/3 hydroxylation (Table VA), suggesting that these other hydroxylase activities are also P-45ONr dependent in human liver microsomes. Human liver steroid 6p hydroxylation was also inhibitable by antibody to rat P-450 PB-2a, but not by antibody to several other rat P-450 forms (Table VB), providing further evidence for immunochemical cross-reactivity between the human and rat liver steroid G/3-hydroxylases. Finally, anti-P-45ONr selectively inhibited P-450 Za-dependent testosterone 6fl hydroxylation in uninduced rat liver microsomes, but not testosterone 16a or 2a hydroxylation (both P-450 2c dependent), 7a hydroxylation (P-450 3 dependent), or 5~ reduction (a P-450-independent microsomal activity) (Fig. 4A). Rat liver microsomal testosterone 6/3 hydroxylation was also selectively inhibited by
430
WAXMAN
ET AL.
TABLE III CORRELATION
BETWEEN STEROID HORMONE 6#?-HYDROXYLASE, NIFEDIPINE OTHER HUMAN LIVER MONOOXYGENASE ACTIVITIES’
P-450 reaction T-6@-Hydroxylase A-G/3-Hydroxylase P-Go-Hydroxylase Nifedipine oxidase Debrisoquine 4-hydroxylase Phenacetin O-deethylase S-Mephenytoin 4-hydroxylase 7-Ethoxycoumarin O-deethylase Cytochrome c reductase
Activityb (means + SD) 0.50 0.30 0.39 1.78 0.058 0.21 0.082 0.19 123
+ + + f f f f + f
T-6@-OHase
0.46 (25) 0.31 (25) 0.37 (25) 1.84 (25) 0.031 (14) 0.13 (18) 0.089 (23) 0.10 (24) 58 (25)
OXIDASE,
AND
A-G/3-OHase P-6fl-OHase (T values)
1.0 0.95 0.97 0.76 0.04 0.20 -0.32 0.62 0.58
1.0 0.96 0.81 0.13 0.48 -0.27 0.58 0.62
1.0 0.76 0.15 0.03 -0.35 0.68 0.63
Nifedipine oxidase
1.0 0.07 0.14 -0.10 0.39 0.41
DShown are the correlation coefficients for comparisons of microsomal testosterone (T), androstenedione (A), and progesterone (P) G/3-hydroxylase and nifedipine oxidase with other microsomal activities. b Activities are expressed as nmol product/min/mg microsomes for N independent samples (value of N in parentheses).
anti-P-450 PB-2a (Fig. 4B), as we have reported previously (8). Influence of TAO complexation. Rat P-450~ and human P-450 HLp are both characterized by their ability to metabo-
lize the macrolide antibiotic TAO to form a stable, TAO-metabolite complex (X,,, 457 nm) in an NADPH-dependent reaction (11). TAO complexation is accompanied by inhibition of microsomal activities cata-
TABLE IV CORRELATION
BETWEEN IMMUNOCHEMICALLY LEVELS AND THEIR ASSOCIATED
DETERMINED MICROSOMAL
HUMAN LIVER MICROSOMAL P-450 ENZYME MONOOXYGENASE ACTIVITIES~
P-450& Polyclonal Testosterone Go-hydroxylase Nifedipine oxidase Debrisoquine 4-hydroxylase Phenacetin 0-deethylase S-Mephenytoin 4-hydroxylase ‘I-Ethoxycoumarin 0-deethylase Specific contentC
0.69 (25) 0.79 (25) -0.10 (14) 0.17 (18) -0.11 (23) 0.31 (24) 70 t 91 (25)
Monoclonal 0.83 (23) 0.81 (23) 0.03 (13) -0.22 (16) -0.20 (21) 0.50 (22) 100d + 98 (23)
P-450~~ P-45op* (T values (N)) -0.10 0.24 -0.27 -0.13 -0.13 -0.23 52 +
(16) (16) (13) (15) (16) (15) 29
(16)
P-450Mpml Total P-450
-0.19 (14) -0.26 (21) 0.24 (14) -0.35 (21) -0.08 (14) -0.17 (14) -0.01 (14) -0.18 (15) -0.17 (14) -0.13 (19) 0.04 (13) -0.42 (20) 52 + 19 100d f 28 (14) (21)
0.54 (25) 0.72 (25) 0.16 (14) 0.22 (18) -0.16 (23) 0.54 (24) 366 + 166 cm
a Correlation coefficients based on N independent comparisons, with the value of N shown in parentheses. b Correlation coefficients based on P-450N~ levels determined using either polyclonal or monoclonal antibodies, as described under Methods. “Specific contents of the indicated P-450 enzymes or of total spectrally measured P-450 (last column), expressed as pmol P-450/mg microsomal protein. Values shown are means + SD for N microsomal samples (value of N in parentheses on bottom line). d Relative P-450 enzyme contents, with the mean value set to 100.
HUMAN
LIVER
STEROID
HORMONE
CONTROL
50
25
L
I m g anti-P-450nF
I
2 IgG/mg
3 mwx.~mes
G,+HYDROXYLASE
P-450
431
lyzed by these and perhaps other members of the P-450 IIIA gene subfamily. Consistent with this possibility, human liver microsomal androstenedione Sj!l hydroxylation was inhibited significantly (-60%) when the microsomes were incubated with TAO + NADPH under conditions that are optimal for TAO-metabolite complex formation (Table VI). 6/3 hydroxylation catalyzed by dexamethasone-induced rat liver microsomes was also inhibited by TAO complexation. By contrast, little effect was observed on the rat P-450 Zc-dependent steroid 16a hydroxylation reaction catalyzed by these same liver microsomes (Table VI). DISCUSSION
I m g antl-P-450,FIgG/mg
2
3 microsomes
FIG. 3. Immunoinhibition of human liver microsomal steroid hormone Go-hydroxylase and ‘I-ethoxycoumarin O-deethylase activities. (A) Shown are the effects of rabbit anti-P-450xr IgG on testosterone 6@-hydroxylase catalyzed by three different human liver microsomal preparations. Uninhibited rates for HL25, HL35, and HL39 were 1.13, 0.38, and 1.38 nmol/6&hydroxy testosterone/min/mg, respectively. Control IgG, effects of nonimmune rabbit IgG on G/3-hydroxylase activity of microsomes HL39. Anti-P-450ur inhibition of ‘I-ethoxycoumarin Odeethylase activity (7-EC; uninhibited rate = 0.50 nmol/min/mg) catalyzed by HL39 is shown for comparison. (IgG controls for this inhibition of ‘I-ethoxycoumarin activity ranged from 97 to 109% of control; not shown.) (8) Shown are the effects of anti-P-45Onr IgG on progesterone Go-hydroxylase and ‘I-ethoxycoumarin 0-deethylase activities of
Cytochrome P-450-dependent hydroxylation appears to be a major route of oxidative inactivation of steroid hormones in mammalian liver. Extensive studies carried out in the rat have demonstrated that steroids, including androgens and progestins, are metabolized in the liver at several sites (e.g., 2cq SD, ‘7a, 16a, and 16p) in reactions catalyzed by multiple P-450 enzymes subject to complex hormonal regulation and developmental control (reviewed in Ref. (19)). Metabolism at one of these sites, the 6/3 position, is catalyzed by the male-specific constitutive P-450 2a, and probably also (8, 20) by one or more of its close structural homologs, the phenobarbital- and synthetic steroid-inducible form(s) designated4 PB-2a, p, PCN-E, etc. In contrast to the multiple hydroxylation pathways observed in rats, human liver microsomes were found to hydroxylate principally at one site, 6& in the present study. Rates for microsomal S/3 hydroxylation of testosterone, androstenedione and progesterone were highly correlated with each other (T z 0.95), as well as with the microsomal specific content of human P-450NFand its associated nifedipine oxidase activity (r - 0.7-0.85). Antibodies to
HL25. Uninhibited rates were 0.41 nmol 6fi-hydroxy progesterone/min/mg and 0.52 nmol 7-hydroxycoumarin/min/mg.
432
WAXMAN
ET
TABLE ANTIBODY
INHIBITION
OF HUMAN
LIVER
A. Inhibition
AL.
V
MICROSOMAL
TESTOSTERONE
P-450NF
by anti-human Testosterone
mg anti-P-45O&mg HL37 microsomes 0 0.65 1.3 2.0 2.6
GPOH-T
26OH-T
(640) 84 20 21 16
(84) 69 19 18 13
B. Inhibition Antibody” Anti-P-450xr Anti-P-450Nr Anti-P-450 Anti-P-450 Anti-P-450 Anti-P-450 Anti-P-450
PB-2a PB-2a 2c PB-1 BNF-B
by anti-rat
mg IgG/mg HL35 microsomes 0 1.3 5 5 10 5 5 5
HYDROXYLATION”
hydroxylation
activity* 15flOH-T
16aOH-T
(27)
(7) 117 114 114 157
85 30 26 19
P-450s GBOH-T (nmol/min/mg) 0.38 0.07 0.07 0.16 0.09 0.41 0.36 0.43
Relative activity (% of control) =lOO 19 19 43 25 107 95 112
a Shown in A are the effects of increasing amounts of anti-P-450m IgG on testosterone hydroxylation catalyzed by human liver microsomes HL37. Shown in B are the effects of various anti-rat P-450 antibodies on testosterone Go-hydroxylation catalyzed by human liver microsomes HL35, with the inhibition by the antihuman P-450~~ shown for comparison. Incubations with antibody and catalytic assays were carried out as described under Methods. * Shown are the rates of testosterone hydroxylation at the indicated positions expressed as a percentage of the uninhibited controls. Activity values for the uninhibited controls are presented in parentheses in units of pmol metabolite/min/mg microsomal protein. ‘Rabbit polyclonal antibodies to rat P-450s were prepared as described in Ref. (8).
P-&ONr or to rat P-450 PB-2a selectively inhibited these 6/3 hydroxylation reactions, providing further evidence that P-450nr (or a closely related P-450 form) is the major catalyst of steroid hormone 68 hydroxylation in human liver microsomes. Antibodies to P-4&r were also inhibitory to P-450 Za-dependent 6p hydroxylation in uninduced rat liver microsomes, consistent with the observed immunochemical homology between the rat and human members of this P-450 gene subfamily (11, 13), designated IIIA (28). Finally, complexation of the macrolide antibiotic TAO, an NADPH-dependent reaction characteristic of both human and rat members of P-450 subfamily IIIA (ll), resulted in the selective inhibition of steroid hormone 6@
hydroxylation catalyzed by both human and rat liver microsomes. That this inhibition was only -60% complete may suggest heterogeneity of these 6fi-hydroxylase P-450s with respect to their TAO complexation activities. An alternative but less likely possibility is that the TAOcomplexed P-450s retain partial enzymatic activity. Interestingly, TAO complexation in vivo also inhibits only -6O-70% of rat liver microsomal steroid S/3 hydroxylation (23). Attempts to study directly the substrate specificities of the rat and human members of P-450 family III in purified, reconstituted systems have met with only limited success due to the inactivation of these P-450 forms during microsome solu-
HUMAN
LIVER
anli
STEROID
HORMONE
P-450m+,
FIG. 4. Selectivity of inhibition of adult male rat liver microsomal testosterone 6j3-hydroxylase activity by anti-P-450nr (A: 6.5 mg IgG/mg microsomes) and by anti-P-450 PB-2a (B: 5 mg IgG/mg microsomes). Uninhibited catalytic rates (nmol/min/mg microsomes) were 0.49 (16aOH-testosterone), 0.74 (i’mOH-testosterone), 0.43 (GOOH-testosterone), 0.52 (BaOH-testosterone), and 0.95 (5a-dihydrotestosterone). Anti-P-450 PB-2a elevated testosterone 5a-reduction to 138% of control.
bilization and enzyme purification (8, 13, 21,27). The biochemical basis for this loss of activity, one that is not generally observed with other hepatic microsomal P-45Os, remains to be established. Induction and other indirect studies suggest that one or more of the inducible rat P-450s in this family also contributes significantly to microsomal conversion of testosterone to A’-testosterone, product of a steroid C&C6 desaturation reaction (22), in addition to the formation of three other minor &face hydroxylated testosterone metabolites, 28-, 15/3-, and 18-hydroxytestosterone (23,24,45). At least two of these metabolites (2@- and 15@hydroxytestosterone) also appear to be formed by human liver P-450NF, as suggested by their correlation with microsomal 6@ hydroxylation rates and their specific inhibi-
G/3-HYDROXYLASE
P-450
433
tion by anti-P-450nr antibodies. These observations, taken together with the report by Guzelian and co-workers that the HLp protein is associated with high erythromycin demethylase and TAO complexation activities (ll), both characteristic of rat P-450~ and the corresponding rabbit P-450 form 3c (25), indicate that substrate specificities are at least partially conserved across species lines within this P-450 gene family. Some differences in substrate specificities are, however, apparent since immunoreactive P-450 PB-2a is the major catalyst of mephenytoin 4hydroxylation in the rat (26), whereas this same reaction is not catalyzed by P-450NF, but rather by immunoreactive P-45O~p.~in human liver microsomes (14). At least two of the rat P-450s in the P-450 IIIA gene subfamily are independently regulated, with P-450 2a adult male-specific and developmentally suppressed in maturing female rats (8, 20), and P-450 PB-2a and related forms highly inducible by synthetic steroids and/or macrolide antibiotics, independent of sex (e.g., (8, 20, 25, 27)). Earlier reports have indicated that human P-450 HLp is subject to the steroid and macrolide antibiotic induction characteristic of immunoreactive rat P-450 PB-2a (ll), suggesting that this human P-450 may be orthologous to one of the inducible rat P-450 forms, rather than to the constitutive rat P-450 2a. Similarly, the highest P-450~~ and steroid 6P-hydroxylase levels determined in the current study were catalyzed by liver microsomes obtained from a patient treated with dexamethasone (Table VI and data not shown). Moreover, sequence comparisons reveal that the P-450Nr-related cDNAs characterized (29, 30) are more highly homologous to an inducible rat P-450 member (PCNl) than to the constitutive rat P-450 member (PCNZ) of this gene family, particularly in the COOH-terminal third of the molecule (i.e., 82% homology to clone PCNl versus 74% homology to clone PCN2 for amino acid residues 342-485; E. J. Holsztynska and D. J. Waxman, unpublished observations). Human liver microsomes might, however, also contain a P-450NF-related enzyme
434
WAXMAN
ET AL.
TABLE EFFECTS
OF TAO
COMPLEXATION
VI
ON MICROSOMAL
ANDROSTENEDIONE
HYDROXYLATIO~~
Hydroxylase activity (nmol metabolite/min/mg HLm, microsomesb
protein)
Rat liver microsomes”
TAO complexation mixture
6j30H-A (% )d
GPOH-A (% )
16aOH-A (% )
Complete -NADPH
0.41 (41) 1.01 (100)
1.31 (45) 2.88 (100)
0.39 (98) 0.40 (loo)
a Human and rat liver microsomes (1 mg/ml) were preincubated for 5 min at 37°C in 0.1 M KPi (pH 7.4), 20% glycerol, 0.1 mM EDTA in the presence of 20 pM TAO and 1.7% dimetbyl sulfoxide. NADPH (1 mM) was then added, and the samples were incubated at 37°C for an additional 45 min to effect TAO complexation. Aliquota were diluted lo-fold into steroid hydroxylase assay mixtures containing fresh NADPH (1 ml@) and androstenedione metabolism was then assayed as described under Methods. Maximal inhibition of microsomal 6@-hydroxylation (-60% inhibition; this table) was obtained within 20-30 min. Increasing the concentration of TAO to 66 jiM did not result in further inhibition. b Isolated from a patient administered dexamethasone, and kindly provided by Dr. P. S. Guzelian, Medical College of Virginia. ‘Pooled microsomes isolated from adult male rats induced with dexamethasone (four daily ip injections in corn oil at 106 mg/kg). d Activities expressed as a percentage of the control complexation samples (-NADPH during the 45 min TAO complexation reaction).
[perhaps equivalent to the human fetal liver P-450 HFLa (31)] that is expressed constitutively and perhaps regulated by endogenous hormones in a manner similar to rat P-450 2a. Interestingly, both monoclonal and polyclonal anti-P-450 PB-2a antibodies were observed to cross-react with an -51-kDa human liver microsomal peptide electrophoretically distinct from P-450NF (il& - 50,500) (D. P. Lapenson and D. J. Waxman, unpublished experiments). This cross-reactive protein was, however, present in only 6 of the 25 human liver samples examined and did not correlate with the sex of the patient, with the levels of P-45ONr, or with miCrOSOma1 S&hydroxylase activity. It remains to be established whether this protein corresponds to an additional member of the human P-450 IIIA gene subfamily. Metabolism of the calcium antagonist and vasodilator nifedipine has been reported to exhibit polymorphism among individual humans (32). Nifedipine oxidation iS P-450~~ dependent in human liver microsomes, and the rate of microsomal nifedipine oxidation correlates well with the specific content of P-45ONr ((13); Table Iv). Immunoreactive P-45ONr levels appear to reflect the recent history of expo-
sure of patients to drugs known to induce immunoreactive rat P-450 PB-2a (ll), suggesting that the observed bimodal distribution of nifedipine oxidase activity in human populations might reflect exposure to inducing agents (environmental and/or endogenous) prior to the nifedipine challenge, rather than a genetically determined polymorphism for the level or activity of this P-450 form. Alternatively, the inducibility of immunoreactive P-450~~ by synthetic steroids and other drugs might be genetically polymorphic in human populations, much as PI-450 induction is genetically controlled in the mouse (33). Although good correlations between microsomal P-45ONr levels and its associated steroid hormone 6@-hydroxylase and nifedipine oxidase activities were obtained in this study, no correlation was observed between immunochemically estimated microsomal contents of P-45Ons, P-454A, or P-450~~~~and their respective associated drug hydroxylase activities, in agreement with earlier reports (12, 14, 16). Although these data would seem to indicate that the respective P-450 forms are functionally altered or inactive in poor metabolizers, the reactivity of the antibodies employed
HUMAN
LIVER
STEROID
HORMONE
in these analyses with apocytochrome and/or with closely related P-450 forms (12,14) leaves open the possibility that the defect is at the level of P-450 enzyme expression. In fact, recent experiments have revealed splicing defects in the gene for a debrisoquine hydroxylating human P-450, designated P-450dbl (34), indicating that the clinical polymorphism in debrisoquine oxidation is most likely due to the absence of this specific P-450 form. P-450nr-dependent 66 hydroxylation appears to be the major route of microsomal oxidative metabolism for the three steroids examined in the current study. These findings are consistent with an earlier report of significant progesterone SPhydroxylase activity in human liver microsomes (15) as well as the observation that 6fl-hydroxylated steroids constitute a prominent class of unconjugated urinary steroid metabolites in man (36). The possible contribution of P-45hF and its rat homologs to the drug and steroid-inducible 6/3 hydroxylation of other physiologically important steroids, including cortisol (37) and various bile acids (38), remains to be established. The absence of significant progesterone 21-hydroxylase activity in the human liver preparations examined in this study is in agreement with an earlier report (39), and indicates that the liver is not likely to be a significant source for deoxycorticosterone formation in man. The absence of significant androgen 16a-, ‘7a-, and 16@hydroxylase activity in the human liver microsomes suggests that P-450 enzymes orthologous to the rat P-450 forms 2c, 3, and PB-4, respectively, might not be expressed at significant levels in human liver. However, Western blotting of human liver microsomes using polyclonal antibodies raised to these rat P-450 forms has revealed at least one immunoreactive polypeptide of M, -5O,OOO60,000 in each case (D. P. Lapenson and D. J. Waxman, unpublished experiments). The human liver polypeptides thus identified may correspond, however, to nonorthologous members of the same or perhaps even another P-450 gene subfamily within the same family. Alternatively, the substrate specificities of these corresponding human and rat P-450 forms
Go-HYDROXYLASE
435
P-450
might not be conserved using steroid substrates. In this context it should be noted that rabbit P-450 LM2 does not exhibit (40) the highly characteristic steroid IS@hydroxylase activity catalyzed by its rat ortholog, P-450 PB-4 (3,5). Steroid 6/3 hydroxylation has been proposed as a useful, noninvasive monitor of hepatic drug metabolism activity in man (41). The current study would suggest, however, that the level of urinary 6/3-hydroxylated steroids is likely to reflect the activity of only one or a limited number of closely related P-450 enzymes. Although these urinary steroid metabolites are therefore unlikely to provide any information on the capacity for hepatic metabolism of drugs falling under the debrisoquine, phenacetin, mephenytoin, or other polymorphisms, they are likely to be useful indicators for hepatic metabolism of drugs and other foreign compounds principally metabolized by P-450~~) including many dihydropyridine calcium channel blockers, quinidine, erythromycin, benzphetamine, and aldrin (11, 13, 42, 43). Future study is required to ascertain the possible consequences of the low hepatic steroid hormone 6P-hydroxylase activity that is likely to be associated with the reduced nifedipine oxidase activity observed in about one out of six individuals. Note added in proof. After
submission
of this
manuscript, Kawano et al (J. Biochem (Tokyo)
102,
493-501 (1987)) reported the isolation of a human liver P-450 designated P-450 (human-l) that appears closely related to human P-45hF and can catalyze testosterone Go-hydroxylation in reconstituted systems. Antibody inhibition experiments and correlation data were presented to document the role of this P-450 in human liver microsomal testosterone Sfihydroxylation, in agreement with the studies described in this manuscript. REFERENCES 1. CONNEY, A. H., AND KLUTCH, Chem 238,1611-1617. 2. VAN
DER HOEVEN,
T. (1981)
A. (1963)
Biochem
J. Biol.
Biophys.
Res. Commun. 100,1285-1291. 3. WAXMAN,
D. J., Ko, A., AND WALSH,
C. (1933)
J.
J. B. (1983)
J.
BioL Chem 258,11937-1194’7. 4. CHENG,
K.-C.,
AND
SCHENKMAN,
BioL Chem 253,11738-11’744. 5. WOOD, A. W., RYAN, D. E., THOMAS, P. E., AND LEVIN, W. (1933) J. BioL Chem 258,8839-8847.
436
WAXMAN
6. WAXMAN, D. J. (1984) J. Biol. Chem. 259, 1548115490. 7. MORGAN, E. T., MACGEOCH, C., AND GUSTAFSSON, J.-A. (1985) J. Biol. C&m. 260,11895-11898. ?a.KATo, R., YAMAZOE, Y., SHIMADA, M., MURAYAMA, N., AND KAMATAKI, T. (1986) J. Bie them 100,895-902. 8. WAXMAN, D. J., DANNAN, G. A., AND GUENGERICH, F. P. (1985) Biochemistry 24,4409-4417. 9. MACGEOCH, C., MORGAN, E. T., HALPERT, J., AND GUSTAFSSON, J.-A. (1984) J. Bid Chem. 259, 15433-15439. 10. WANG, P. P., BEAUNE, P., KAMINSKY, L. S., DANNAN, G. A., KADLUBAR, F. F., LARREY, D., AND GUENGERICH, F. P. (1983) Biochemistry 22, 5375-5383. 11. WATKINS, P. B., WRIGHTON, S. A., MAUREL, P., SCHUETZ, E. G., MENDEZ-PICON, G., PARKER, G. A., AND GUZELIAN, P. S. (1985) Proc. Nat1
Ad
Sci USA 82,6310-6314.
12. GUT, J., CATIN, T., DAYER, P., KRONBACH, T., ZANGER, U., AND MEYER, U. A. (1985) J. Biol Chem 261,11734-11’743. 13. GUENGERICH, F. P., MARTIN, M. V., BEAUNE, P. H., KREMERS, P., WOLFF, T., AND WAXMAN, D. J. (1986) J. Biol &em. 261,5051-5060. 14. SHIMADA, T., MISONO, K. S., AND GUENGERICH, F. P. (1986) J. Biol. Chem 261,909-921. 15. KREMERS, P., BEAUNE, P., CRESTEIL, T., DE GRAEVE, J., COLUMELLI, S., LEROUX, J.-P., AND GIELEN, J. E. (1981) Eur. J. Biochem. 118, 599-606. 16. DISTLERATH, L. M., REILLY, P. E. B., MARTIN, M. V., DAVIS, G. G., WILKINSON, G. R., AND GUENGERICH, F. P. (1985) J. Biol. Chem 260.
9057-9067. 17. WAXMAN,
D. J., AND WALSH,
C. (1982)
J.
Biol
Chem 257,10446-10457. 18. PARK, S. S., WAXMAN, D. J., MILLER, H., ROBINSON, R., ATTISANO, C., GUENGERICH, F. P., AND GELBOIN, H. V. (1986) Biochem Pharmacol35, 2859-286’7. 19. WAXMAN, D. J. (1988) Biochem Phnrmad 37, 71-84. 20. GONZALEZ, F. J., SONG, B.-J., AND HARDWICK, J. P. (1986) MoL Cell Bid 6,2969-2976. 21. ELSHOURBAGY, N. A., AND GUZELIAN, P. S. (1980) J. Biol Chem 255,1279-1285. 22. NAGATA, K., LIBERATO, D. J., GILLETTE, J. R., AND SASAME, H. A. (1986) Drug. Metab. Dispos. 14,559-565. 23. SONDERFAN, A. J., ARLOTTO, M. P., DUTTON, D. R., MCMILLEN, S. K., AND PARKINSON, A. (1987)
Arch. B&hem.
Biophys. 255,27-41.
24. DECKER, C., SUGIYAMA, K., UNDERWOOD, M., AND CORREIA, M. A. (1986) Biochem Biophqs. Res. Commun. 136.1162-1169. 25. WRIGHTON, S. A., SCHUETZ, E. G., WATKINS, P. B., MAUREL, P., BARWICK, J., BAILEY, B. S., HARTLE, H. T., YOUNG, B., AND GUZELIAN, P. (1985)
Mol. PhwmacoL
28,312-321.
ET AL. 26. SHIMADA, T., AND GUENGERICH, F. P. (1985) MoL Pharmacd 28.215-219. 27. GRAVES, P. E., KAMINSKY, L. S., AND HALPERT, J.
(1987) B&hem&q
26,3887-3894.
28. NEBERT, D., ADESNIK, M., COON, M., ESTABROOK, R., GONZALEZ, F., GUENGERICH, P., G~SALUS, I., JOHNSON, E., KEMPER, B., LEVIN, W., PHILLIPS, I., SATO, R., AND WATERMAN, M. (1987) DNA 6,1-11. 29. BEAUNE, P. H., UMBENHAUER, D. R., BORK, R. W., LLOYD, R. S., AND GUENGERICH, F. P. (1986) Proc Nat1 Acad Sci. USA 83,8064-8068. 30. MOLOWA, D. T., SCHUETZ, E. G., WRIGHTON, S. A., WATKINS, P. B., KREMERS, P., MENDEZ-PICON, G., PARKER, G. A., AND GUZELIAN, P. S. (1986) Proc. Nat1 Acad Sci. USA 83,5311-5315. 31. KITADA, M., KAMATAKI, T., ITAHASHI, K., RIKIHISA, T., AND KANAKUBO, Y. (1987) Biochem
Pharrnacol
36,453-456.
32. KLEINBLOESEM, C. H., VAN BRUMMELEN, P., FABER, H., DANHOF, M., VERMEULEN, N. P. E., AND BREIMER, D. D. (1984) B&hem Pharmaco1 33,3721-3724. 33. WHITLOCK, J. P., JR. (1986) Annu. Rev. Phxwma-
co1 Toxic01 26,333-369. 34. GONZALEZ, F., SKODA, R. C., KIMURA, S., UMENO, M., ZANGER, U. M., NEBERT, D. W., GELBOIN, H. V., HARDWICK, J. P., AND MEYER, U. A. (1988) Nature (London) 331,442-446. 35. HOSTETLER, K. A., WRIGHTON, S. A., KREMERS, P., AND GUZELIAN, P. S. (1987) Biochem J. 245, 27-33. 36. BURSTEIN, S., KIMBALL, H. L., KLAIBER, E. L., AND GUT, M. (1967) J. Clin Erw!wrinol Metab. 27,491. 37. KATZ, F. H., LIPMAN, M. M., FRANTZ, A. G., AND JAILER, J. W. (1962) J. Clin En&wind Metab. 22,71-77. 38. LAMBIOTTE, M., AND THIERRY, N. (1980) J. BioL Chem 255,11324-11331. 39. WINKEL, C. A., SIMPSON, E. R., MILEWICH, L., AND MACDONALD, P. C. (1980) Proc Nat1 Ad Sci
USA 77,7069-7073. 40. KOOP, D. R., PERSSON, A. V., AND COON, M. J. (1981) J. Biol Chem 256,10704-10711. 41. PARK, B. K. (1981) Brit. .I Clin Pharrnacd 12, 97-102. 42. GUENGERICH, F. P., MULLER-ENOCH, D., AND BLAIR, I. A. (1986) Mol. Pharrnad 30,287~295. 43. BOCKER,
R. G., AND GUENGERICH,
F. P. (1986)
J.
Med Chem 29,1596-1603. 44. GUENGERICH, F. P. (1987) in “Mammalian Cytochrome P-450” (Guengerich, F. P., Ed.), pp. l-54, CRC Press, Boca Raton, FL. 45. LEVIN, W., THOMAS, P. E., RYAN, D. E., AND WOOD, A. W. (1987) Arch. Biochem. Biophys. 258.630-635. 46. HALPERT, J. R. (1988) Arch. Biochem Biophys, in press.
Human Liver Microsomal Steroid Metabolism: Identification of the Major Microsomal Steroid Hormone G@Hydroxylase Cytochrome P-450 Enzyme’ DAVID
J. WAXMAN,**’
CYNTHIA ATTISANO,* F. PETER DAVID P. LAPENSON*
GUENGERICH,?
AND
*Depart-t of Biological Chemistry and ikiolecular Pharmxzcobg~ and Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115 and +Department of Biochenzistry and Center in Molecular Toxicology, Vanderbilt University Sch.ool of Medicine, Nashville, Tennessee $7282 Received
December
10, 198’7, and in revised
form
February
22,1933
Cytochrome P-450-dependent steroid hormone metabolism was studied in isolated human liver microsomal fractions. S/3hydroxylation was shown to be the major route of NADPH-dependent oxidative metabolism (375% of total hydroxylated metabolites) with each of three steroid substrates, testosterone, androstenedione, and progesterone. With testosterone, 2/3 and 15p hydroxylation also occurred, proceeding at -10% and 3-4% the rate of microsomal S/3hydroxylation, respectively, in each of the liver samples examined. Rates for the three steroid G/3-hydroxylase activities were highly correlated with each other (T = 0.95-0.97 for 25 individual microsomal preparations), suggesting that a single human liver P-450 enzyme is the principal microsomal 6p-hydroxylase catalyst with all three steroid substrates. Steroid G/3-hydroxylase rates correlated well with the specific content of human P-45&F (r = 0.69-0.83) and with its associated nifedipine oxidase activity (r = 0.80), but not with the rates for debrisoquine 4-hydroxylase, phenacetin 0-deethylase, or S-mephenytoin 4-hydroxylase activities or the specific contents of their respective associated P-450 forms in these same liver microsomes (r < 0.2). These correlative observations were supported by the selective inhibition of human liver microsomal 6@hydroxylation by antibody raised to either human P-45ONr or a rat homolog, P-450 PB-2a. Anti-P-450nr also inhibited human microsomal testosterone 28 and 158 hydroxylation in parallel to the 6@-hydroxylation reaction. This antibody also inhibited rat P-450 Ba-dependent steroid hormone 6@hydroxylation in uninduced adult male rat liver microsomes but not the steroid ICY,16a, or 7a hydroxylation reactions catalyzed by other rat P-450 forms. Finally, steroid 6/3 hydroxylation catalyzed by either human or rat liver microsomes was selectively inhibited by NADPH-dependent complexation of the macrolide antibiotic triacetyloleandomycin, a reaction that is characteristic of members of the P-450~~ gene subfamily (P-450 IIIA subfamily). These observations establish that P-450NF or a closely related enzyme is the major catalyst of steroid hormone 6s hydroxylation in human liver microsomes, and furthermore suggest that steroid 6/3 hydroxylation may provide a useful, noninvasive monitor for the monooxygenase activity of this hepatic P-450 form. o 1988 Academic press, I”~.
Steroid hormones are hydroxylated hepatic cytochrome P-450 (P-45O)3
‘Supported in part by Grants DK33765 (D.J.W.) and CA 30907 (F.P.G.) from the National Institutes of Health. ‘To whom correspondence should be addressed at: Dana-Farber Cancer Institute, JF-525, 44 Binney Street, Boston, MA 02115. 0003-9861/88 Copyright All rights
$3.00
0 1988 by Academic Press, Inc. of reproduction in any form reserved.
s Abbreviations used: TAO, triacetyloleandomycin; body. 424
P-450, cytochrome MAb, monoclonal
by enP-450; anti-
HUMAN
LIVER
STEROID
HORMONE
zymes in a highly regioselective and stereospecific manner. Studies carried out in the rat have established that P-450-dependent hydroxylation of these endogenous substrates can occur at multiple sites, with the relative rates of hydroxylation at each site dependent on the steroid and on the monooxygenase induction status of the animal (1,2). Steroids such as testosterone and androstenedione are hydroxylated by purified P-450 enzymes to yield patterns of metabolites that are often characteristic of individual P-450 forms (3-5). These studies have not only aided in the characterization of P-450 enzymes that exhibit similar spectral, immunochemical, and enzymatic properties using drug substrates, but have also led to the identification of specific rat hepatic P-450 catalysts of microsomal steroid hydroxylase activities subject to developmental control and hormonal regulation. Rat P-450 form 2c4 was thus shown to correspond to the major steroid hormone 16a-hydroxylase induced at puberty in male but not female rats (6, 7, 7a), P-450 2a to the adult male-specific G/3-hydroxylase that is developmentally suppressed in
4 The nomenclature used in this study to designate rat P-450 enzymes is that described previously (6). For a more complete listing of other enzyme designations see reviews by Waxman (19) and Guengerich (44). In addition, the term P-450 2a is used to designate the constitutive, hormonally regulated rat steroid Go-hydroxylase, whose regulation is consistent with that determined using the cDNA clone PCN2 (20). P-450 PB-2a, isolated from phenobarbital-induced adult male rats (6), appears to be immunochemically related to the phenobarbital and/or synthetic steroid inducible form(s) termed p, PCN-E, PCNl, PCNa, or PCNb by other investigators (11,13, 20,27). Recent studies (46) suggest that the inducible proteins PB-2a, PCN-E, and PCNb may all correspond to the same P-450 form, while the preparations designated p, PCNl, and PCNa may all correspond to a second inducible form. These two groups of inducible P-4509 are generically refered to as “immunoreactive P-450 PB-2a” in the current study. The relationship between the two purified human members of this gene family, P-450~~ (13) and P-450 HLp (11) is not known with certainty. The genes for these rat and human P-450s all appear to be in the P-450 IIIA gene subfamily (28).
GB-HYDROXYLASE
P-450
425
maturing female rats (8), and P-450 2d to the adult female-specific steroid disulfate 15@hydroxylase (9). Multiple hepatic P-450 enzymes have been purified from human liver and characterized extensively using xenobiotic substrates (e.g., (10-14)). These studies have been carried out with the primary goal of identifying those forms that contribute to the genetically polymorphic oxidation of drugs such as debrisoquine and mephenytoin. Much less is known, however, about the possible participation of these or other human P-450s in the metabolism of endogenous substrates such as steroid hormones. In the case of testosterone, polyclonal antibody inhibition experiments suggest that the human liver microsomal Go-hydroxylase reaction (15) may be catalyzed by an enzyme designated P-45ONF (13). However, in view of the very low testosterone 6P-hydroxylase activity of purified P-450~~ (13) these observations must be viewed as tentative in the absence of additional experimental data. The present study was therefore undertaken to characterize in greater detail the pathways of NADPH-dependent testosterone metabolism catalyzed by human liver microsomes, and to examine the possible contributions of P-450NF and other human liver P-450 enzymes to the metabolism of testosterone, as well as androstenedione and progesterone. Results described in this report establish that 6/3 hydroxylation is the major pathway of oxidative metabolism for the three steroid hormones, with the three respective 60 hydroxylation rates shown to be highly correlated with each other in a collection of 25 individual human liver microsomes. In the case of testosterone, hydroxylation is also demonstrated to occur at the 2p and 15/3 positions, at rates that correlate closely with the G/3-hydroxylase reactions. Finally, antibody inhibition experiments, measurements of microsomal P-450~~ specific contents and nifedipine oxidase rates, and TAO complexation experiments are presented in support of the conclusion that P-450N~ is the major 6@hydroxylase for three different steroid hormone substrates in human liver microsomes.
426
WAXMAN METHODS
ikficroaomes. Human liver microsomes were prepared from organ donors who met accidental deaths. Tissues were removed and perfused with chilled buffers within 15-30 min of death. Samples were divided into small pieces (l-6 cm3) which were frozen by immersion in liquid nitrogen and stored at -70°C. Samples were initially homogenized in 0.1 M Tris acetate (pH 7.4) containing 0.1 M KCl, 1 mM EDTA, 20 pM butylated hydroxytoluene, and 0.4 mM phenylmethylsulfonyl fluoride, with microsome isolation as described previously (10). Under these conditions, catalytically active microsomes suitable for the isolation of active P-450 enzymes are obtained (e.g., (16)). Stemid hydroxglase assays. Microsomal fractions [O.l to 0.2 mg microsomal protein/ml of 100 mM potassium phosphate buffer (pH 7.4) containing 0.1 mM EDTA and 1% glycerol (v/v)] were assayed for steroid hydroxylase activities (lo- to 20-min incubations) using as substrate [4-“Cltestosterone, [4-i4C]androstenedione, or [4-“Clprogesterone (7-10 mCi/ mmol; 25 PM, final concentration). Hydroxylated metabolites were resolved by silica gel thin-layer chromatography and then quantitated by scintillation counting as described in detail (3,6). Antibody inhibitions of steroid hydroxylase activities were determined following preincubation of microsomes with isolated IgG fractions for 45 min at 20 to 22°C. Metabolite identijcations. Steroid metabolites were identified by their cochromatography with standard hydroxylated steroids on silica gel thin layer plates developed in multiple solvent systems (3). Identifications were made with metabolites formed by human liver microsome preparations HL35 and HL37. 6fi-Hydroxytestosterone formed by the human liver microsomes was identified by chromatography in solvents AA, AB, and C, 6@-hydroxyandrostenedione in solvents AA, HH, and I, and 6B-hydroxyprogesterone in solvents K, L, and M. In addition, 26-hydroxytestosterone was identified in solvents AA and BB, while 15/3-hydroxytestosterone was identified in solvents AB and BB. [See Ref. (3) for additional details relating to product identifications]. Solvent compositions used in these experiments were: solvent A = CHzClz/acetone (4/l); solvent B = CHC&/ethyl acetate/ethanol (4/l/0.7); solvent C = CHzClJacetone (915); solvent H = CHClJethyl acetate (l/2); solvent I = ethyl acetate/n-hexane/acetic acid (19/5/l); solvent K = CHClJethanol (9/l); solvent L = ethyl acetate/nhexane/acetic acid (15/4/l); and solvent M = diethylether/acetone (4/l). Each of the microsomal steroid metabolites identified by these methods was also distinguished on the thin-layer plates from the available monohydroxy steroids listed previously (3, 6) as well as from 7/3-hydroxytestosterone and 156hydroxytestosterone, kindly provided by Searle, Inc.,
ET AL. and 15cY-hydroxyprogesterone, obtained from the Steroid Reference Collection. Other methods. P-450 concentrations, protein concentrations, metabolism of ‘I-ethoxycoumarin (17), nifedipine (13), debrisoquine, phenacetin (16), mephenytoin (14), and TAO complexation (11) were measured as reported in the indicated references. Immunoquantitation of microsomal P-450s was carried out by Western blotting of liver microsomes with polyclonal antibodies raised to the human P-450 forms designated P-450Ds, P-450pA (16), P-450~~ (13), and P-45O~ap.~(14). Western blotting was also carried out using MAb C9, raised to rat P-450 PB-2a (Ref. (18); P-450 antigen used to prepare that antibody was previously designated 2a/PCN-E) and kindly provided by Dr. S. S. Park of the NCI. This MAb was shown to cross-react on Western blots with a single polypeptide present in human liver microsomes that exhibited the same electrophoretic mobility as P-45&~. P-450N~ levels determined with this anti-rat P-450 monoclonal antibody were highly correlated with the levels determined using the anti-human P-450~~ polyclonal antibodies (r = 0.83, n = 23). Correlation coefficients of r > 0.65 are highly significant (P < 0.001) for comparisons of N = 20 to 25 independent samples. RESULTS
Testosterone metabolism catalyzed by human liver microsomal fra&on.s. Shown in Fig. 1 is a representative chromatogram of the NADPH-dependent [‘“Cltestosterone metabolism profile exhibited by liver microsomes prepared. from 6 individual human livers. Similar metabolite patterns were observed for 20 other human liver microsomal samples (not shown). In each case a single prominent hydroxylation product was formed, This metabolite pattern contrasted with the more complex profile for rat hepatic microsomes, where at least four major products are observed (Fig. 1, first lane). The major human liver microsomal testosterone metabolite was distinguished from 15 available monohydroxytestosterones and then identified as 6p-hydroxytestosterone by its cochromatography with authentic standard in three chromatography solvent systems (see Methods). Three minor human liver microsomal testosterone metabolites were respectively identified as 2P-hydroxytestosterone, 15/?hydroxytestosterone, and androstenedione using similar analytical methods.
HUMAN
LIVER
STEROID
HORMONE
427
P-450
Ratios of the three principal microsomal hydroxy testosterone metabolites were similar for each of the human liver preparations [6@/2j3/15@ = l/0.10/0.03-0.041, suggesting that they might be formed by the same P-450 enzyme (also see Table VA, below). Androstenedione formation was not observed in the absence of liver microsomes or NADPH and reflects oxidation of the substrate’s 17@hydroxy group to a ketone. Oxidation of testosterone to androstenedione has previously been observed with rat liver microsomes (e.g., Ref. (2)) as well as with several purified P-450s isolated from rat hepatic tissue (3-5). Metabolite quantitation revealed that G/3-hydroxytestosterone generally comprised from 75 to 85% of the total hydroxylated product in these human liver microsomal incubations (Table I). Although steroid 6@-hydroxylation is sexually differentiated in the rat (8), it does not appear sex-dependent in humans, since equally efficient 6P-hydroxylation was observed with several preparations of male and female human liver microsomes (Table I and data not shown).
x I
Z.aT 68,
HVMAN
6@-HYDROXYLASE
MICROSOMES
FIG. 1. Metabolism of testosterone by human liver microsomes. Shown is an autoradiograph of a silica gel thin-layer chromatogram resolving testosterone and metabolites formed by liver microsomes from six individual humans (samples HL23 through HL37). The chromatogram was developed in solvents AB and the metabolites were quantitated as described under Methods. Sample at left: metabolites from untreated adult male rat liver microsomes. Identified human liver microsomal testosterone metabolites are marked on the right and rat microsomal metabolites on the left. T, testosterone; 6j3-T, 6@-hydroxytestosterone, etc.; T,5or, Sa-dihydrotestosterone; A, androstenedione; X, unidentified.
Androstenedione tabolism. Similar
and progesterone
me-
approaches were used to characterize the NADPH-dependent metabolism of androstenedione and proges-
TABLE
I
TESTOSTERONEMETABOLISMCATALYZEDBYHUMANLIVERMICROSOMES'
Sex
HL23 HL25 HL32 HL33 HL35 HL37
Female Male Female Male Male Male
GPOH-T
1.48 0.72 0.09 0.31 0.43 0.69
2j3OH-T
0.16 0.06 0.01 0.03 0.05 0.07
15@OH-T Androstenedioneb (nmol product/min/mg protein) 0.06 0.02 10.005 0.01 0.02 0.03
0.05 0.05 0.02 0.10 0.04 0.03
T,5a”
Total activityd
0.06 0.05 0.02 0.09 0.04 0.05
1.81 0.90 0.14 0.54 0.58 0.87
a Shown are the rates of testosterone metabolite formation catalyzed by human liver microsomes prepared from donors HL23-I-IL37 and determined as described in Fig. 1 and Methods. T, testosterone. b Resulting from a P-450-catalyzed C17-oxidation. c 5cY-Androstane-3-one-17j3-ol and formed by microsomal steroid A4-3-one Sa-reductase. d Sum of the activities of the major metabolites indicated. Other metabolites (e.g., 16aOH-T) were generally formed at rates CO.01 nmol/min/mg.
428
WAXMAN
terone by human liver microsomes. With these steroids as well, 6@ hydroxylation was found to be the major route of oxidative metabolism (Fig. 2, Table II). Microsomal G/3-hydroxylase activities were highly correlated for the three steroid substrates (testosterone, androstenedione, progesterone) in the 25 human liver samples examined (r = 0.95-0.9’7) (Table III), suggesting that a single P-450 form catalyzes all three reactions. 6p hydroxylation rates were generally highest with testosterone, with activity ratios of 6@OH-testosterone/6POH-androstenedione = 1.85 f 0.58 and GPOH-testosterone/G/IOH-progesterone = 1.32 f 0.40 (means f SD) obtained for the 25 individual microsomes. Correlation between human liver steroid 6@-hydroxylase activity and P-450~~ specific content. Immunoreactive rat P-450
PB-2a has been shown to catalyze the majority of steroid hormone 6/3 hydroxylation in both induced and uninduced rat liver microsomes (8). Human liver P-450s homologous to rat P-450 PB-2a have been
ET AL.
identified and designated P-450N~ (nifedipine oxidase P-450; Ref. (13)) and HLp (ll), suggesting that these P-450s might mediate the corresponding 6s hydroxylation reactions in human liver microsomes. Consistent with this proposal, microsomal steroid 6p hydroxylation rates correlated significantly with P-450xr-dependent (13) microsomal nifedipine oxidase activity (r = 0.76-0.81) (Table III), as well as with immunochemically estimated P-450NF levels, determined using either polyclonal anti-P-450Nr antibodies (r = 0.69) or a monoclonal antibody raised to rat P-450 PB-2a (see Methods) (r = 0.83) (Table IV). Plots of microsomal testosterone 6@-hydroxylase activity versus P-450NF content yielded a slope of approximately 5 nmol GPOH-testosterone/min/nmol microsomal P-450NF (not shown), suggesting that the microsomal turnover number of this human 6P-hydroxylase P-450 is similar to that of the constitutive rat liver microsomal 6P-hydroxylase, P-450 2a (8). The y-intercept of 0.08 nmol 6B-hydroxy
Rf - 1.0
P, 5aA,50:STERONE-
ANDROSTENEOIONE-
Apolor
Tesiosteronex-
c
GPOH-Px-
6pOH-A-
-0.25 I6aOH-P-
NO $5
HL 25
HL 32
HL 33
HL 35
HL 37
NO pS
HL 25
HL 32
HL 33
HL 35
HL 37
FIG. 2. Metabolism of androstenedione (A) and progesterone (B) by human liver microsomes. Shown are autoradiograms as in Fig. 1. The chromatogram in A was developed in solvents HH and that in B in solvent I (see Methods). A, androstenedione; A,5a, 5a-dihydroandrostanedione; X, unidentified, P, progesterone; P,5o, 5a-dihydroprogesterone. “Apolar” corresponds to the region of migration of 2cy-hydroxy-, 17a-hydroxy-, 2Oa-hydroxy-, and 20j3-hydroxyprogesterones (see Ref. (6)). The microsomal metabolite migrating just below androstenedione (Rf -0.54, A) was not identified.
HUMAN TABLE
LIVER
STEROID
HORMONE
II
ANDROSTENEDIONE AND PROGESTERONE METABOLISM CATALYZED BY HUMAN LIVER MICROSOMES A. Androstenedione Testosterone* A,5ac G@OH-A” (nmol product/min/mg protein) HL25 HL32 HL33 HL35 HL37
0.31 0.04 0.20 0.26 0.46
0.17 0.26 0.33 0.14 0.18
0.07 0.05 0.15 0.07 0.06
B. Progesterone 6j30H-P”
HL25 HL32 HL33 HL35 HL37
0.34 0.03 0.16 0.51 0.74
16aOH-P Unknown’ (nmol product/min/mg 0.05 0.005 0.01 0.05 0.10
0.02 0.05 0.05 0.03 0.04
Apola# protein) 0.12 0.05 0.16 0.16 0.13
P,50ia
0.11 0.11 0.19 0.13 0.21
“66-Hydroxyandrostenedione and Gp-hydroxyprogesterone identified as described under Methods. Other hydroxylated metabolites (including 21-hydroxyprogesterone) were generally formed at rates GO.03 nmol/min/mg. P, progesterone. b Formed by reduction of the substrate’s 17-keto group to a 17/3-hydroxyl. ’ 5a-Androstane-3,17-dione and formed by microsomal steroid A’-bone 5u-reductase. d Mixture of at least two progesterone metabolites migrating on silica gel (Fig. 2B) in the region where 2aOH-P, 20oOH-P, BOBOH-P, and 17aOH-P chromatograph. ’ 5a-Pregnane-3,20-dione and formed by microsomal steroid A’-3-one 5a-reductase. ‘Metabolite X in Fig. 2B.
metabolite/min/mg microsomes suggests that relatively little 68 hydroxylation is carried out by independently regulated P-450 enzymes. Consistent with this conclusion, microsomal testosterone S/3 hydroxylation did not correlate with the specific contents of three human liver P-450 enzymes catalyzing polymorphic drug oxidation (P-45Ons, P-450PA, and P-45oyp-1) (Table IV), or with their associated enzyme activities (debrisoquine 4-hydroxylase, phenacetin 0-deethylase, and S-mephenytoin 4-hydroxylase, respectively) (Table III). It should be noted, however, that although P-450~~ levels correlated well with microsomal nifedipine oxidase activity (r = 0.79-0.81), specific contents of
6&HYDROXYLASE
P-450
429
the other three human P-450 forms did not correlate with their associated microsomal activities (r < 0; Table IV). Moreover, steroid hormone 6@hydroxylase activity also correlated with microsomal ‘7ethoxycoumarin 0-deethylase activity (r = 0.6-0.7; Table III), which is P-45ONr independent (Fig. 3; see below), as well as with NADPH P-450 reductase-catalyzed cytochrome c reduction (r - 0.6, Table III) and with total spectral P-450 content (r = 0.54; Table IV). Correlative relationships such as these should therefore be interpreted with some caution. Immunoinhibition and cross-reactivity studies of human and rat steroid hormone 6@-hydroxylases. The findings described
above suggest that P-45ONr (or a closely related P-450 form) may be the principal catalyst of steroid hormone SD hydroxylation in human liver microsomes. This possibility is supported by antibody inhibition experiments, which revealed that anti-P-45ONr IgG can inhibit ~80% of human liver microsomal6fl hydroxylation of testosterone, progesterone (Fig. 3), and androstenedione (not shown) under conditions where microsomal 7-ethoxycoumarin 0-deethylation is 120% inhibited. Anti-P-450Nr IgG also inhibited human liver microsomal testosterone 2/3 hydroxylation and testosterone 15p hydroxylation in parallel to testosterone 6/3 hydroxylation (Table VA), suggesting that these other hydroxylase activities are also P-45ONr dependent in human liver microsomes. Human liver steroid 6p hydroxylation was also inhibitable by antibody to rat P-450 PB-2a, but not by antibody to several other rat P-450 forms (Table VB), providing further evidence for immunochemical cross-reactivity between the human and rat liver steroid G/3-hydroxylases. Finally, anti-P-45ONr selectively inhibited P-450 Za-dependent testosterone 6fl hydroxylation in uninduced rat liver microsomes, but not testosterone 16a or 2a hydroxylation (both P-450 2c dependent), 7a hydroxylation (P-450 3 dependent), or 5~ reduction (a P-450-independent microsomal activity) (Fig. 4A). Rat liver microsomal testosterone 6/3 hydroxylation was also selectively inhibited by
430
WAXMAN
ET AL.
TABLE III CORRELATION
BETWEEN STEROID HORMONE 6#?-HYDROXYLASE, NIFEDIPINE OTHER HUMAN LIVER MONOOXYGENASE ACTIVITIES’
P-450 reaction T-6@-Hydroxylase A-G/3-Hydroxylase P-Go-Hydroxylase Nifedipine oxidase Debrisoquine 4-hydroxylase Phenacetin O-deethylase S-Mephenytoin 4-hydroxylase 7-Ethoxycoumarin O-deethylase Cytochrome c reductase
Activityb (means + SD) 0.50 0.30 0.39 1.78 0.058 0.21 0.082 0.19 123
+ + + f f f f + f
T-6@-OHase
0.46 (25) 0.31 (25) 0.37 (25) 1.84 (25) 0.031 (14) 0.13 (18) 0.089 (23) 0.10 (24) 58 (25)
OXIDASE,
AND
A-G/3-OHase P-6fl-OHase (T values)
1.0 0.95 0.97 0.76 0.04 0.20 -0.32 0.62 0.58
1.0 0.96 0.81 0.13 0.48 -0.27 0.58 0.62
1.0 0.76 0.15 0.03 -0.35 0.68 0.63
Nifedipine oxidase
1.0 0.07 0.14 -0.10 0.39 0.41
DShown are the correlation coefficients for comparisons of microsomal testosterone (T), androstenedione (A), and progesterone (P) G/3-hydroxylase and nifedipine oxidase with other microsomal activities. b Activities are expressed as nmol product/min/mg microsomes for N independent samples (value of N in parentheses).
anti-P-450 PB-2a (Fig. 4B), as we have reported previously (8). Influence of TAO complexation. Rat P-450~ and human P-450 HLp are both characterized by their ability to metabo-
lize the macrolide antibiotic TAO to form a stable, TAO-metabolite complex (X,,, 457 nm) in an NADPH-dependent reaction (11). TAO complexation is accompanied by inhibition of microsomal activities cata-
TABLE IV CORRELATION
BETWEEN IMMUNOCHEMICALLY LEVELS AND THEIR ASSOCIATED
DETERMINED MICROSOMAL
HUMAN LIVER MICROSOMAL P-450 ENZYME MONOOXYGENASE ACTIVITIES~
P-450& Polyclonal Testosterone Go-hydroxylase Nifedipine oxidase Debrisoquine 4-hydroxylase Phenacetin 0-deethylase S-Mephenytoin 4-hydroxylase ‘I-Ethoxycoumarin 0-deethylase Specific contentC
0.69 (25) 0.79 (25) -0.10 (14) 0.17 (18) -0.11 (23) 0.31 (24) 70 t 91 (25)
Monoclonal 0.83 (23) 0.81 (23) 0.03 (13) -0.22 (16) -0.20 (21) 0.50 (22) 100d + 98 (23)
P-450~~ P-45op* (T values (N)) -0.10 0.24 -0.27 -0.13 -0.13 -0.23 52 +
(16) (16) (13) (15) (16) (15) 29
(16)
P-450Mpml Total P-450
-0.19 (14) -0.26 (21) 0.24 (14) -0.35 (21) -0.08 (14) -0.17 (14) -0.01 (14) -0.18 (15) -0.17 (14) -0.13 (19) 0.04 (13) -0.42 (20) 52 + 19 100d f 28 (14) (21)
0.54 (25) 0.72 (25) 0.16 (14) 0.22 (18) -0.16 (23) 0.54 (24) 366 + 166 cm
a Correlation coefficients based on N independent comparisons, with the value of N shown in parentheses. b Correlation coefficients based on P-450N~ levels determined using either polyclonal or monoclonal antibodies, as described under Methods. “Specific contents of the indicated P-450 enzymes or of total spectrally measured P-450 (last column), expressed as pmol P-450/mg microsomal protein. Values shown are means + SD for N microsomal samples (value of N in parentheses on bottom line). d Relative P-450 enzyme contents, with the mean value set to 100.
HUMAN
LIVER
STEROID
HORMONE
CONTROL
50
25
L
I m g anti-P-450nF
I
2 IgG/mg
3 mwx.~mes
G,+HYDROXYLASE
P-450
431
lyzed by these and perhaps other members of the P-450 IIIA gene subfamily. Consistent with this possibility, human liver microsomal androstenedione Sj!l hydroxylation was inhibited significantly (-60%) when the microsomes were incubated with TAO + NADPH under conditions that are optimal for TAO-metabolite complex formation (Table VI). 6/3 hydroxylation catalyzed by dexamethasone-induced rat liver microsomes was also inhibited by TAO complexation. By contrast, little effect was observed on the rat P-450 Zc-dependent steroid 16a hydroxylation reaction catalyzed by these same liver microsomes (Table VI). DISCUSSION
I m g antl-P-450,FIgG/mg
2
3 microsomes
FIG. 3. Immunoinhibition of human liver microsomal steroid hormone Go-hydroxylase and ‘I-ethoxycoumarin O-deethylase activities. (A) Shown are the effects of rabbit anti-P-450xr IgG on testosterone 6@-hydroxylase catalyzed by three different human liver microsomal preparations. Uninhibited rates for HL25, HL35, and HL39 were 1.13, 0.38, and 1.38 nmol/6&hydroxy testosterone/min/mg, respectively. Control IgG, effects of nonimmune rabbit IgG on G/3-hydroxylase activity of microsomes HL39. Anti-P-450ur inhibition of ‘I-ethoxycoumarin Odeethylase activity (7-EC; uninhibited rate = 0.50 nmol/min/mg) catalyzed by HL39 is shown for comparison. (IgG controls for this inhibition of ‘I-ethoxycoumarin activity ranged from 97 to 109% of control; not shown.) (8) Shown are the effects of anti-P-45Onr IgG on progesterone Go-hydroxylase and ‘I-ethoxycoumarin 0-deethylase activities of
Cytochrome P-450-dependent hydroxylation appears to be a major route of oxidative inactivation of steroid hormones in mammalian liver. Extensive studies carried out in the rat have demonstrated that steroids, including androgens and progestins, are metabolized in the liver at several sites (e.g., 2cq SD, ‘7a, 16a, and 16p) in reactions catalyzed by multiple P-450 enzymes subject to complex hormonal regulation and developmental control (reviewed in Ref. (19)). Metabolism at one of these sites, the 6/3 position, is catalyzed by the male-specific constitutive P-450 2a, and probably also (8, 20) by one or more of its close structural homologs, the phenobarbital- and synthetic steroid-inducible form(s) designated4 PB-2a, p, PCN-E, etc. In contrast to the multiple hydroxylation pathways observed in rats, human liver microsomes were found to hydroxylate principally at one site, 6& in the present study. Rates for microsomal S/3 hydroxylation of testosterone, androstenedione and progesterone were highly correlated with each other (T z 0.95), as well as with the microsomal specific content of human P-450NFand its associated nifedipine oxidase activity (r - 0.7-0.85). Antibodies to
HL25. Uninhibited rates were 0.41 nmol 6fi-hydroxy progesterone/min/mg and 0.52 nmol 7-hydroxycoumarin/min/mg.
432
WAXMAN
ET
TABLE ANTIBODY
INHIBITION
OF HUMAN
LIVER
A. Inhibition
AL.
V
MICROSOMAL
TESTOSTERONE
P-450NF
by anti-human Testosterone
mg anti-P-45O&mg HL37 microsomes 0 0.65 1.3 2.0 2.6
GPOH-T
26OH-T
(640) 84 20 21 16
(84) 69 19 18 13
B. Inhibition Antibody” Anti-P-450xr Anti-P-450Nr Anti-P-450 Anti-P-450 Anti-P-450 Anti-P-450 Anti-P-450
PB-2a PB-2a 2c PB-1 BNF-B
by anti-rat
mg IgG/mg HL35 microsomes 0 1.3 5 5 10 5 5 5
HYDROXYLATION”
hydroxylation
activity* 15flOH-T
16aOH-T
(27)
(7) 117 114 114 157
85 30 26 19
P-450s GBOH-T (nmol/min/mg) 0.38 0.07 0.07 0.16 0.09 0.41 0.36 0.43
Relative activity (% of control) =lOO 19 19 43 25 107 95 112
a Shown in A are the effects of increasing amounts of anti-P-450m IgG on testosterone hydroxylation catalyzed by human liver microsomes HL37. Shown in B are the effects of various anti-rat P-450 antibodies on testosterone Go-hydroxylation catalyzed by human liver microsomes HL35, with the inhibition by the antihuman P-450~~ shown for comparison. Incubations with antibody and catalytic assays were carried out as described under Methods. * Shown are the rates of testosterone hydroxylation at the indicated positions expressed as a percentage of the uninhibited controls. Activity values for the uninhibited controls are presented in parentheses in units of pmol metabolite/min/mg microsomal protein. ‘Rabbit polyclonal antibodies to rat P-450s were prepared as described in Ref. (8).
P-&ONr or to rat P-450 PB-2a selectively inhibited these 6/3 hydroxylation reactions, providing further evidence that P-450nr (or a closely related P-450 form) is the major catalyst of steroid hormone 68 hydroxylation in human liver microsomes. Antibodies to P-4&r were also inhibitory to P-450 Za-dependent 6p hydroxylation in uninduced rat liver microsomes, consistent with the observed immunochemical homology between the rat and human members of this P-450 gene subfamily (11, 13), designated IIIA (28). Finally, complexation of the macrolide antibiotic TAO, an NADPH-dependent reaction characteristic of both human and rat members of P-450 subfamily IIIA (ll), resulted in the selective inhibition of steroid hormone 6@
hydroxylation catalyzed by both human and rat liver microsomes. That this inhibition was only -60% complete may suggest heterogeneity of these 6fi-hydroxylase P-450s with respect to their TAO complexation activities. An alternative but less likely possibility is that the TAOcomplexed P-450s retain partial enzymatic activity. Interestingly, TAO complexation in vivo also inhibits only -6O-70% of rat liver microsomal steroid S/3 hydroxylation (23). Attempts to study directly the substrate specificities of the rat and human members of P-450 family III in purified, reconstituted systems have met with only limited success due to the inactivation of these P-450 forms during microsome solu-
HUMAN
LIVER
anli
STEROID
HORMONE
P-450m+,
FIG. 4. Selectivity of inhibition of adult male rat liver microsomal testosterone 6j3-hydroxylase activity by anti-P-450nr (A: 6.5 mg IgG/mg microsomes) and by anti-P-450 PB-2a (B: 5 mg IgG/mg microsomes). Uninhibited catalytic rates (nmol/min/mg microsomes) were 0.49 (16aOH-testosterone), 0.74 (i’mOH-testosterone), 0.43 (GOOH-testosterone), 0.52 (BaOH-testosterone), and 0.95 (5a-dihydrotestosterone). Anti-P-450 PB-2a elevated testosterone 5a-reduction to 138% of control.
bilization and enzyme purification (8, 13, 21,27). The biochemical basis for this loss of activity, one that is not generally observed with other hepatic microsomal P-45Os, remains to be established. Induction and other indirect studies suggest that one or more of the inducible rat P-450s in this family also contributes significantly to microsomal conversion of testosterone to A’-testosterone, product of a steroid C&C6 desaturation reaction (22), in addition to the formation of three other minor &face hydroxylated testosterone metabolites, 28-, 15/3-, and 18-hydroxytestosterone (23,24,45). At least two of these metabolites (2@- and 15@hydroxytestosterone) also appear to be formed by human liver P-450NF, as suggested by their correlation with microsomal 6@ hydroxylation rates and their specific inhibi-
G/3-HYDROXYLASE
P-450
433
tion by anti-P-450nr antibodies. These observations, taken together with the report by Guzelian and co-workers that the HLp protein is associated with high erythromycin demethylase and TAO complexation activities (ll), both characteristic of rat P-450~ and the corresponding rabbit P-450 form 3c (25), indicate that substrate specificities are at least partially conserved across species lines within this P-450 gene family. Some differences in substrate specificities are, however, apparent since immunoreactive P-450 PB-2a is the major catalyst of mephenytoin 4hydroxylation in the rat (26), whereas this same reaction is not catalyzed by P-450NF, but rather by immunoreactive P-45O~p.~in human liver microsomes (14). At least two of the rat P-450s in the P-450 IIIA gene subfamily are independently regulated, with P-450 2a adult male-specific and developmentally suppressed in maturing female rats (8, 20), and P-450 PB-2a and related forms highly inducible by synthetic steroids and/or macrolide antibiotics, independent of sex (e.g., (8, 20, 25, 27)). Earlier reports have indicated that human P-450 HLp is subject to the steroid and macrolide antibiotic induction characteristic of immunoreactive rat P-450 PB-2a (ll), suggesting that this human P-450 may be orthologous to one of the inducible rat P-450 forms, rather than to the constitutive rat P-450 2a. Similarly, the highest P-450~~ and steroid 6P-hydroxylase levels determined in the current study were catalyzed by liver microsomes obtained from a patient treated with dexamethasone (Table VI and data not shown). Moreover, sequence comparisons reveal that the P-450Nr-related cDNAs characterized (29, 30) are more highly homologous to an inducible rat P-450 member (PCNl) than to the constitutive rat P-450 member (PCNZ) of this gene family, particularly in the COOH-terminal third of the molecule (i.e., 82% homology to clone PCNl versus 74% homology to clone PCN2 for amino acid residues 342-485; E. J. Holsztynska and D. J. Waxman, unpublished observations). Human liver microsomes might, however, also contain a P-450NF-related enzyme
434
WAXMAN
ET AL.
TABLE EFFECTS
OF TAO
COMPLEXATION
VI
ON MICROSOMAL
ANDROSTENEDIONE
HYDROXYLATIO~~
Hydroxylase activity (nmol metabolite/min/mg HLm, microsomesb
protein)
Rat liver microsomes”
TAO complexation mixture
6j30H-A (% )d
GPOH-A (% )
16aOH-A (% )
Complete -NADPH
0.41 (41) 1.01 (100)
1.31 (45) 2.88 (100)
0.39 (98) 0.40 (loo)
a Human and rat liver microsomes (1 mg/ml) were preincubated for 5 min at 37°C in 0.1 M KPi (pH 7.4), 20% glycerol, 0.1 mM EDTA in the presence of 20 pM TAO and 1.7% dimetbyl sulfoxide. NADPH (1 mM) was then added, and the samples were incubated at 37°C for an additional 45 min to effect TAO complexation. Aliquota were diluted lo-fold into steroid hydroxylase assay mixtures containing fresh NADPH (1 ml@) and androstenedione metabolism was then assayed as described under Methods. Maximal inhibition of microsomal 6@-hydroxylation (-60% inhibition; this table) was obtained within 20-30 min. Increasing the concentration of TAO to 66 jiM did not result in further inhibition. b Isolated from a patient administered dexamethasone, and kindly provided by Dr. P. S. Guzelian, Medical College of Virginia. ‘Pooled microsomes isolated from adult male rats induced with dexamethasone (four daily ip injections in corn oil at 106 mg/kg). d Activities expressed as a percentage of the control complexation samples (-NADPH during the 45 min TAO complexation reaction).
[perhaps equivalent to the human fetal liver P-450 HFLa (31)] that is expressed constitutively and perhaps regulated by endogenous hormones in a manner similar to rat P-450 2a. Interestingly, both monoclonal and polyclonal anti-P-450 PB-2a antibodies were observed to cross-react with an -51-kDa human liver microsomal peptide electrophoretically distinct from P-450NF (il& - 50,500) (D. P. Lapenson and D. J. Waxman, unpublished experiments). This cross-reactive protein was, however, present in only 6 of the 25 human liver samples examined and did not correlate with the sex of the patient, with the levels of P-45ONr, or with miCrOSOma1 S&hydroxylase activity. It remains to be established whether this protein corresponds to an additional member of the human P-450 IIIA gene subfamily. Metabolism of the calcium antagonist and vasodilator nifedipine has been reported to exhibit polymorphism among individual humans (32). Nifedipine oxidation iS P-450~~ dependent in human liver microsomes, and the rate of microsomal nifedipine oxidation correlates well with the specific content of P-45ONr ((13); Table Iv). Immunoreactive P-45ONr levels appear to reflect the recent history of expo-
sure of patients to drugs known to induce immunoreactive rat P-450 PB-2a (ll), suggesting that the observed bimodal distribution of nifedipine oxidase activity in human populations might reflect exposure to inducing agents (environmental and/or endogenous) prior to the nifedipine challenge, rather than a genetically determined polymorphism for the level or activity of this P-450 form. Alternatively, the inducibility of immunoreactive P-450~~ by synthetic steroids and other drugs might be genetically polymorphic in human populations, much as PI-450 induction is genetically controlled in the mouse (33). Although good correlations between microsomal P-45ONr levels and its associated steroid hormone 6@-hydroxylase and nifedipine oxidase activities were obtained in this study, no correlation was observed between immunochemically estimated microsomal contents of P-45Ons, P-454A, or P-450~~~~and their respective associated drug hydroxylase activities, in agreement with earlier reports (12, 14, 16). Although these data would seem to indicate that the respective P-450 forms are functionally altered or inactive in poor metabolizers, the reactivity of the antibodies employed
HUMAN
LIVER
STEROID
HORMONE
in these analyses with apocytochrome and/or with closely related P-450 forms (12,14) leaves open the possibility that the defect is at the level of P-450 enzyme expression. In fact, recent experiments have revealed splicing defects in the gene for a debrisoquine hydroxylating human P-450, designated P-450dbl (34), indicating that the clinical polymorphism in debrisoquine oxidation is most likely due to the absence of this specific P-450 form. P-450nr-dependent 66 hydroxylation appears to be the major route of microsomal oxidative metabolism for the three steroids examined in the current study. These findings are consistent with an earlier report of significant progesterone SPhydroxylase activity in human liver microsomes (15) as well as the observation that 6fl-hydroxylated steroids constitute a prominent class of unconjugated urinary steroid metabolites in man (36). The possible contribution of P-45hF and its rat homologs to the drug and steroid-inducible 6/3 hydroxylation of other physiologically important steroids, including cortisol (37) and various bile acids (38), remains to be established. The absence of significant progesterone 21-hydroxylase activity in the human liver preparations examined in this study is in agreement with an earlier report (39), and indicates that the liver is not likely to be a significant source for deoxycorticosterone formation in man. The absence of significant androgen 16a-, ‘7a-, and 16@hydroxylase activity in the human liver microsomes suggests that P-450 enzymes orthologous to the rat P-450 forms 2c, 3, and PB-4, respectively, might not be expressed at significant levels in human liver. However, Western blotting of human liver microsomes using polyclonal antibodies raised to these rat P-450 forms has revealed at least one immunoreactive polypeptide of M, -5O,OOO60,000 in each case (D. P. Lapenson and D. J. Waxman, unpublished experiments). The human liver polypeptides thus identified may correspond, however, to nonorthologous members of the same or perhaps even another P-450 gene subfamily within the same family. Alternatively, the substrate specificities of these corresponding human and rat P-450 forms
Go-HYDROXYLASE
435
P-450
might not be conserved using steroid substrates. In this context it should be noted that rabbit P-450 LM2 does not exhibit (40) the highly characteristic steroid IS@hydroxylase activity catalyzed by its rat ortholog, P-450 PB-4 (3,5). Steroid 6/3 hydroxylation has been proposed as a useful, noninvasive monitor of hepatic drug metabolism activity in man (41). The current study would suggest, however, that the level of urinary 6/3-hydroxylated steroids is likely to reflect the activity of only one or a limited number of closely related P-450 enzymes. Although these urinary steroid metabolites are therefore unlikely to provide any information on the capacity for hepatic metabolism of drugs falling under the debrisoquine, phenacetin, mephenytoin, or other polymorphisms, they are likely to be useful indicators for hepatic metabolism of drugs and other foreign compounds principally metabolized by P-450~~) including many dihydropyridine calcium channel blockers, quinidine, erythromycin, benzphetamine, and aldrin (11, 13, 42, 43). Future study is required to ascertain the possible consequences of the low hepatic steroid hormone 6P-hydroxylase activity that is likely to be associated with the reduced nifedipine oxidase activity observed in about one out of six individuals. Note added in proof. After
submission
of this
manuscript, Kawano et al (J. Biochem (Tokyo)
102,
493-501 (1987)) reported the isolation of a human liver P-450 designated P-450 (human-l) that appears closely related to human P-45hF and can catalyze testosterone Go-hydroxylation in reconstituted systems. Antibody inhibition experiments and correlation data were presented to document the role of this P-450 in human liver microsomal testosterone Sfihydroxylation, in agreement with the studies described in this manuscript. REFERENCES 1. CONNEY, A. H., AND KLUTCH, Chem 238,1611-1617. 2. VAN
DER HOEVEN,
T. (1981)
A. (1963)
Biochem
J. Biol.
Biophys.
Res. Commun. 100,1285-1291. 3. WAXMAN,
D. J., Ko, A., AND WALSH,
C. (1933)
J.
J. B. (1983)
J.
BioL Chem 258,11937-1194’7. 4. CHENG,
K.-C.,
AND
SCHENKMAN,
BioL Chem 253,11738-11’744. 5. WOOD, A. W., RYAN, D. E., THOMAS, P. E., AND LEVIN, W. (1933) J. BioL Chem 258,8839-8847.
436
WAXMAN
6. WAXMAN, D. J. (1984) J. Biol. Chem. 259, 1548115490. 7. MORGAN, E. T., MACGEOCH, C., AND GUSTAFSSON, J.-A. (1985) J. Biol. C&m. 260,11895-11898. ?a.KATo, R., YAMAZOE, Y., SHIMADA, M., MURAYAMA, N., AND KAMATAKI, T. (1986) J. Bie them 100,895-902. 8. WAXMAN, D. J., DANNAN, G. A., AND GUENGERICH, F. P. (1985) Biochemistry 24,4409-4417. 9. MACGEOCH, C., MORGAN, E. T., HALPERT, J., AND GUSTAFSSON, J.-A. (1984) J. Bid Chem. 259, 15433-15439. 10. WANG, P. P., BEAUNE, P., KAMINSKY, L. S., DANNAN, G. A., KADLUBAR, F. F., LARREY, D., AND GUENGERICH, F. P. (1983) Biochemistry 22, 5375-5383. 11. WATKINS, P. B., WRIGHTON, S. A., MAUREL, P., SCHUETZ, E. G., MENDEZ-PICON, G., PARKER, G. A., AND GUZELIAN, P. S. (1985) Proc. Nat1
Ad
Sci USA 82,6310-6314.
12. GUT, J., CATIN, T., DAYER, P., KRONBACH, T., ZANGER, U., AND MEYER, U. A. (1985) J. Biol Chem 261,11734-11’743. 13. GUENGERICH, F. P., MARTIN, M. V., BEAUNE, P. H., KREMERS, P., WOLFF, T., AND WAXMAN, D. J. (1986) J. Biol &em. 261,5051-5060. 14. SHIMADA, T., MISONO, K. S., AND GUENGERICH, F. P. (1986) J. Biol. Chem 261,909-921. 15. KREMERS, P., BEAUNE, P., CRESTEIL, T., DE GRAEVE, J., COLUMELLI, S., LEROUX, J.-P., AND GIELEN, J. E. (1981) Eur. J. Biochem. 118, 599-606. 16. DISTLERATH, L. M., REILLY, P. E. B., MARTIN, M. V., DAVIS, G. G., WILKINSON, G. R., AND GUENGERICH, F. P. (1985) J. Biol. Chem 260.
9057-9067. 17. WAXMAN,
D. J., AND WALSH,
C. (1982)
J.
Biol
Chem 257,10446-10457. 18. PARK, S. S., WAXMAN, D. J., MILLER, H., ROBINSON, R., ATTISANO, C., GUENGERICH, F. P., AND GELBOIN, H. V. (1986) Biochem Pharmacol35, 2859-286’7. 19. WAXMAN, D. J. (1988) Biochem Phnrmad 37, 71-84. 20. GONZALEZ, F. J., SONG, B.-J., AND HARDWICK, J. P. (1986) MoL Cell Bid 6,2969-2976. 21. ELSHOURBAGY, N. A., AND GUZELIAN, P. S. (1980) J. Biol Chem 255,1279-1285. 22. NAGATA, K., LIBERATO, D. J., GILLETTE, J. R., AND SASAME, H. A. (1986) Drug. Metab. Dispos. 14,559-565. 23. SONDERFAN, A. J., ARLOTTO, M. P., DUTTON, D. R., MCMILLEN, S. K., AND PARKINSON, A. (1987)
Arch. B&hem.
Biophys. 255,27-41.
24. DECKER, C., SUGIYAMA, K., UNDERWOOD, M., AND CORREIA, M. A. (1986) Biochem Biophqs. Res. Commun. 136.1162-1169. 25. WRIGHTON, S. A., SCHUETZ, E. G., WATKINS, P. B., MAUREL, P., BARWICK, J., BAILEY, B. S., HARTLE, H. T., YOUNG, B., AND GUZELIAN, P. (1985)
Mol. PhwmacoL
28,312-321.
ET AL. 26. SHIMADA, T., AND GUENGERICH, F. P. (1985) MoL Pharmacd 28.215-219. 27. GRAVES, P. E., KAMINSKY, L. S., AND HALPERT, J.
(1987) B&hem&q
26,3887-3894.
28. NEBERT, D., ADESNIK, M., COON, M., ESTABROOK, R., GONZALEZ, F., GUENGERICH, P., G~SALUS, I., JOHNSON, E., KEMPER, B., LEVIN, W., PHILLIPS, I., SATO, R., AND WATERMAN, M. (1987) DNA 6,1-11. 29. BEAUNE, P. H., UMBENHAUER, D. R., BORK, R. W., LLOYD, R. S., AND GUENGERICH, F. P. (1986) Proc Nat1 Acad Sci. USA 83,8064-8068. 30. MOLOWA, D. T., SCHUETZ, E. G., WRIGHTON, S. A., WATKINS, P. B., KREMERS, P., MENDEZ-PICON, G., PARKER, G. A., AND GUZELIAN, P. S. (1986) Proc. Nat1 Acad Sci. USA 83,5311-5315. 31. KITADA, M., KAMATAKI, T., ITAHASHI, K., RIKIHISA, T., AND KANAKUBO, Y. (1987) Biochem
Pharrnacol
36,453-456.
32. KLEINBLOESEM, C. H., VAN BRUMMELEN, P., FABER, H., DANHOF, M., VERMEULEN, N. P. E., AND BREIMER, D. D. (1984) B&hem Pharmaco1 33,3721-3724. 33. WHITLOCK, J. P., JR. (1986) Annu. Rev. Phxwma-
co1 Toxic01 26,333-369. 34. GONZALEZ, F., SKODA, R. C., KIMURA, S., UMENO, M., ZANGER, U. M., NEBERT, D. W., GELBOIN, H. V., HARDWICK, J. P., AND MEYER, U. A. (1988) Nature (London) 331,442-446. 35. HOSTETLER, K. A., WRIGHTON, S. A., KREMERS, P., AND GUZELIAN, P. S. (1987) Biochem J. 245, 27-33. 36. BURSTEIN, S., KIMBALL, H. L., KLAIBER, E. L., AND GUT, M. (1967) J. Clin Erw!wrinol Metab. 27,491. 37. KATZ, F. H., LIPMAN, M. M., FRANTZ, A. G., AND JAILER, J. W. (1962) J. Clin En&wind Metab. 22,71-77. 38. LAMBIOTTE, M., AND THIERRY, N. (1980) J. BioL Chem 255,11324-11331. 39. WINKEL, C. A., SIMPSON, E. R., MILEWICH, L., AND MACDONALD, P. C. (1980) Proc Nat1 Ad Sci
USA 77,7069-7073. 40. KOOP, D. R., PERSSON, A. V., AND COON, M. J. (1981) J. Biol Chem 256,10704-10711. 41. PARK, B. K. (1981) Brit. .I Clin Pharrnacd 12, 97-102. 42. GUENGERICH, F. P., MULLER-ENOCH, D., AND BLAIR, I. A. (1986) Mol. Pharrnad 30,287~295. 43. BOCKER,
R. G., AND GUENGERICH,
F. P. (1986)
J.
Med Chem 29,1596-1603. 44. GUENGERICH, F. P. (1987) in “Mammalian Cytochrome P-450” (Guengerich, F. P., Ed.), pp. l-54, CRC Press, Boca Raton, FL. 45. LEVIN, W., THOMAS, P. E., RYAN, D. E., AND WOOD, A. W. (1987) Arch. Biochem. Biophys. 258.630-635. 46. HALPERT, J. R. (1988) Arch. Biochem Biophys, in press.