Reconstituted Liver Microsomal Enzyme System That Hydroxylates Drugs, Other Foreign Compounds, and Endogenous Substrates SPECIFICITIES OF THE CYTOCHROME P450 FRACTIONS FROM COSTROL AXD PHENOBARBITAL-TREATED RATS

SUMMARY The substrate specificities of the liver microsomal cytochrome P450 fractions from control and phenobarbital-treated immature, male rats were studied with the reconstituted hydroxylation system. In the presence of fixed amounts of the reductase and lipid fractions and various amounts of hemoprotein, the catalytic activities of the cytochrome P450 fractions from control and phenobarbital-treated rats were compared with a variety of substrates. Both fractions were about equally active for the N-demethyla-tion of ethyhnorphine and the hydroxylation of testosterone at positions 6/3 and 7~2. However, the cytochrome P450 fraction from phenobarbital-treated rats was much more active than the cytochrome P450 fraction from control rats for the N-demethylation of benzphetamine and chlorcyclizine, the oxidation of pentobarbital, and the hydroxylation of testosterone at position 16a. For the hydroxylation of 3,4-benzpyrene, the cytochrome P450 fraction from control rats was slightly more active than the P450 fraction

FIY)W the Department of Biochemistry an.d Drug Metabolism, Hofmann-La Roche Inc., Nutley, New Jersey 07110 SUMMARY The substrate specificities of the liver microsomal cytochrome P450 fractions from control and phenobarbitaltreated immature, male rats were studied with the reconstituted hydroxylation system.
In the presence of fixed amounts of the reductase and lipid fractions and various amounts of hemoprotein, the catalytic activities of the cytochrome P450 fractions from control and phenobarbitaltreated rats were compared with a variety of substrates. Both fractions were about equally active for the N-demethylation of ethyhnorphine and the hydroxylation of testosterone at positions 6/3 and 7~2. However, the cytochrome P450 fraction from phenobarbital-treated rats was much more active than the cytochrome P450 fraction from control rats for the N-demethylation of benzphetamine and chlorcyclizine, the oxidation of pentobarbital, and the hydroxylation of testosterone at position 16a. For the hydroxylation of 3,4-benzpyrene, the cytochrome P450 fraction from control rats was slightly more active than the P450 fraction from phenobarbital-treated rats. These results suggest that the cytochrome P450 from control rats is catalytically different from the cytochrome P450 from phenobarbital-treated rats. The cytochrome P448 fraction from rats treated with 3-methylcholanthrene was also assayed for its ability to metabolize various substrates.
The substrate specificity of the cytochrome P448 fraction differed from both cytochrome P450 fractions.
The osidat#ive metabolism of drugs and steroids is catalyzed by the microsomal enzyme system which has been resolved into three components (l-3) : a CO-binding hemoprotein (cytochrome P450), an SADPH-depcudent rcductase, and a lipid identified as phosphat,idplcholine (4). The administration of phenobarbital to rats increases the concentration of hemoprotein which is spectrally identical with t'he cytochrome P450 found in untreated animals, whereas the administration of 3-methylcholan-threne to rats induces the formation of a spectrally distinct hemoprotein, cytochrome P448l (8-10).
Using the reconstituted hydroxylation system from liver microsomes (11,12), we have recently shown that the enzyme systems prepared from PB2or 3-MC-treated immature rats exhibit different substrate specificities, and that such specificities reside primarily in the cytochrome fraction, rather than in the reductase or lipid fraction of liver microsomes (3,13). These studies indicate that the cytochrome P450 fraction from rats treated with PB is catalytically different from the cytochrome P448 fraction from rats treated with 3-,%X.
On the other hand, it has been generally assumed that the cytochrome 1'450 in untreated rats is catalytically identical with the cgtochrome P450 in PB-treated rats. Since, in rats, the metabolism of most substrates is enhanced by PB treatment, only quantitative but not qualitative differences between microsomes from control animals and microsomes from PB-treated animals have generally been noted. In addition, a direct comparison of the rates of metabolism of various substrates from control and PB-treated rats is difficult to make due to the differences in levels of both cytochrome P450 and reductase in these microsomal preparations.
One of the advantages of using the reconstituted system to study the substrate specificities of cytochrome 1'450 from control and PB-treated rats is that it is possible to vary the hemoprotein fraction in the presence of fixed amounts of the lipid and reductase fractions.
Thus, it is possible to compare the relative catalytic activity of the cytochrome P450 fractions from control and PB-treated rats toward a variety of substrates.
If the two hemoprotein fractions are identica1, then the rate of metabolism of various substrates should be the same when equal concentrattions of hemoproteins arc used. Using fixed amounts of lipid and reductase and varying both the source and the con- centration of the hemoprotein, we have studied the metabolism of a number of substrates by the reconstituted system. The results presented in this paper show that the P450 fractions from control and PB-treated rats are about equally active for the metabolism of some substrates, but the PB-P4501 fraction is far more effective than the control-P450 fraction for the metabolism of other substrates. The results suggest that cytochrome P450 from control rats has different catalytic activity than cytochrome P450 from PB-treated rats.

METHODS
Male Long-Evans rats (from Blue Spruce Farms, Altamont, New York) weighing 50 to 60 g were treated intraperitoneally with PB or 3-MC as previously described (14). Control rats received no treatment. The partial purification of the cytochrome P450 fractions from control and PB-treated rats and the cytochrome P448 fraction from a-MC-treated rats has recently been described (15). Hemoprotein concentrations were determined by the reduced CO difference spectra (16) using the same extinction coefficient (91 rnM-1 cm-l) for all three preparations. When the heme content of these preparations was determined by the pyridine hemochromogen method (16), the extinction coefficient for A*xu~c, or L&49,1 in the reduced CO difference spectra was found to be within 10% of 91 rnK1 cm-' (15). The concentrations of the partially purified PB-P450 and P448 ranged from 4.5 to 7.5 nmoles of cytochrome per mg of protein; preparations with a concentration between 4.5 and 5.0 nmoles of cytochrome per mg of protein were used for the present studies. Control-P450 had a concentration of 2.4 nmoles of cytochrome per mg of protein. All three hemoprotein preparations were purified a-fold with respect to microsomal protein and 20-fold with respect to microsomal phospholipid. The reductase and lipid fractions were prepared from PB-treated rats as previously described (3,17).
Hydroxylation of Testosterone-The hemoprotein fractions prepared from control, PB-, and 3-MC-treated rats showed different specificities for the hydroxylation of test,osterone at positions S/I, 7a, and 16a! (Fig. 1). All three fractions were equally active for 6/%hydroxylation at all concentrations tested. The cytochrome P448 fraction was more active than the other two fractions for 7c+hydroxylation, while the PB-P450 fraction was far more active than the control-P450 or P448 fraction for the hydroxylation of testosterone at position 16or. Since the activities were determined in the presence of fised amounts of the reductase and lipid fractions and only the cytochrome fractions were varied, the differences in activity of the controLP450, PB-P450, and P448 fractions for the hydroxylation of testosterone at positions S/3, 7cu, and 16ar reflect differences in catalytic activities of the three hemoprotein fractions.
N-Demethylattin of Benzphetamine and Ethylmorphiw-Previous studies have shown that the PB-P450 fraction was much more active than the P448 fraction for the N-demethylation of benzphetamine (3). The results plotted in Fig. 2 confirm our earlier findings and also show that the control-P450 fraction was considerably less active than the PB-P450 fraction, but consistently more active than the cytochrome P448 fraction for benzphetamine N-demethylation. In contrast, all three hemoprotein fractions were approximately equally active for the N-demethylation of ethylmorphine.
The hydroxylation of pentobarbital (18) and 3,4-benzpyrene (3,19), the N-demethylation of chlorcyclizine (3), and the hydroxylation of testosterone at positions S/3, 7a, and 16a! (3) were determined according to previously published procedures. The N-demethylation of benzphetamine and ethylmorphine were assayed by following the rate of NADPH oxidation (1). The rate of substrate-dependent NADPH oxidation has previously been shown to be equal to the rate of formaldehyde formation (2).  Table I shows that the PB-P450 fraction was considerably more active than the control-P450 fraction for pentobarbital hydroxylation, while the control-P450 fraction was slightly, but consistently, more effective than the PB-P450 fraction for the hydroxylation of 3,4-benzpyrene at all concentrations tested. For the N-demethylation of chlorcyclizine, the PB-P450 fraction was more active than the control-P450 fraction. As was reported previously (3, u)), the cytochrome P448 fraction was very active for 3,4-benzpyrene hydroxylation, very poor for pentobarbital hydroxylation, but moderately active for chlorcyclizine N-demethylation.
To determine the rate of metabolism of various substrates, fixed amounts of reductase and lipid, and variable amounts of control-P450 and PB-450 were incubated with substrates, buffers, and cofactors. Thus, reductase and lipid were in excess at low hemoprotein concentrations but limiting at high hemoprotein concentrations. For the purpose of comparison, cytochrome P448 was also assayed for its ability to metabolize various substrates. Although at low concentrations of hemoprotein (up to 9.2 nmole per ml), no catalytic activity was observed in the absence of added reductase, the hemoprotein fractions were not entirely free of reductase, and a small but measurable amount of metabolism was observed without the addition of reductase when a large amount of hemoprotein (0.5 to 1.0 nmole per ml) was used. Therefore, at the high concentrations of hemoprotein, the reaction rates were corrected by subtracting the activity obtained in the absence of reductase from the activity obtained in the presence of reductase. When this correction was made, hydroxylation activity increased with increasing concentrations of hemoprotein, reached a maximum, and plateaued. The reaction mixture was incubated at 37" for 10 min and radioactive 6&, 7cu-, and I&-hydroxytestosterone were separated and determined (3). Control, PB, and 3-MC refer to the hemoprotein fractions prepared from untreated, PB-treated, and 3-MC-treated rats, respectively. Both the reduct.ase and lipid fractions were prepared from PB-treated rats, 0.1 mg of lipid; 0.09 mg of reductase for pentobarbital and chlorcyclizine assays and 0.04 mg of reductase for 3.4-benzpyrene assay; and the indicated amounts of hemoprotein fractions. The reaction mixtures were incubated at 37" for 5,10, and 15 min for studies on the metabolism of 3,4-benzpyrene, pentobarbital, and chlorcyclizine, respectively. Both the reductase and lipid fractions were prepared from PB-treated rats. No hydroxylation activity was detected when hemoprotein was omitted from the reaction mixture. Although the control-P450 fraction was only slightly more active than the PB-P450 fraction for 3,4-benzpyrene hydroxylation, these two reactions could be differentially affected by 7,8benzoflavone and SKF-525A. As shown in Fig. 3, with the control-P450 fraction, the reaction was inhibited by 7, %benzoflavone while with the PB-P450 fraction, 3,4-benzpyrene hydroxylation was either slightly stimulated or slightly inhibited, de-  3. Effect of 7,%benzoflavone on 3,4-benzpyrene hydroxylation in t,he reconstituted systems containing the cytochrome P450 fraction from control or PB-treated rats. The assay conditions for the hydroxylation of 3,4-benzpyrene were the same as described in Table I. 7,%Benzoflavone dissolved in methanol was added (in 10 lLlj to the incubation mixture.
The concentrations of cytochrome P450 from control and PB-treated rats used were 0.16 and 0.15 nmole per ml, respectively.
The activities for 100% were 0.27 and 0.12 nmole/5 min for the systems containing control-P450 and PB-P450, respectively.
pending on the concentration of 7,8-benzoflavone added. These results were surprising since the hydroxylation of 3,4-benzpyrene was stimulated by 7 ,%benzoflavone in microsomes from both control and PB-treated rats (21). We have no explanation for this difference between the microsomal and reconstituted systems. In contrast, SKF-525A at a concentration of 4 X low3 M inhibited the PB-P45kupported hydroxylation of 3,4-benzpyrene by SO%, but only inhibited the control-P450-supported reaction by 15%.

DISCUSSION
Since the amounts of the reductase and lipid were kept constant, it is apparent from the studies described above that the control-P450, PB-P450, and 3-MC-P448 fractions have different substrate specificities. The relative catalytic activities of the control-P450, PB-P450, and P448 fractions from immature, male rats for various substrates in the presence of reductase and lipid are summarized in Table II. The relative activities were calculated from either the maximal activity obtained with each substrate (where the hemoprotein concentration is in excess and the reductase and lipid are limiting) or the activity obtained with each substrate at 0.1 PM hemoprotein (where the hemoprotein concentration is limiting), setting the control-P450 value equal to 100%. From this table, it can be seen that the relative activities of the three hemoprotein fractions depended upon the particular substrate studied. The need to use a variety of substrates to establish differences in catalytic activity between various hemoprotein fractions is therefore evident. For some substrates, all three hemoprotein fractions were equally active, whereas for other substrates, either the PB-P450 or the P448 fraction was much more active. Although these results strongly suggest that the control-P450, PB-P450, and 3-MC-P448 have different catalytic activity, the possibility cannot be excluded that a substance other than P450 in the various partially purified preparations contributes to the different substrate specifkities of the three hemoprotein fractions.
The possibilky that several CO-binding pigments with different catalytic activities occur in liver microsomes has been one of the most extensively studied aspects of the microsomal mixed function oxidase system. Studies in recent years with microsomal suspensions have provided evidence to support the possibility that multiple forms of cytochrome P450 exist in liver microsomes (22)(23)(24)(25)(26)(27).
The hydrosylation of testosterone at positions 66, 'icr, and 16a in rat liver microsomes occurs by a cytochrome P450-dependent system, and the three hydroxylation reactions are under different regulatory control (22,23). In microsomes from control rats, the three hydroxylation reactions had different, patkerns of development with age. In addition, chlorthion strongly inhibited the 16oc-hydroxylation of testosterone but only slightly inhibited hydrosylation at positions 6/? and 7ar. Chronic t.reatment of rats with PB and 3-MC also increased the S/3-, 'iol-, and 16a-hydroxylations of testosterone by liver microsomes to ra,rying degrees. Although the hydroxylation of testosterone at' all three positions in microsomes from control rats was inhibited by CO, and the inhibition by CO of all three reactions was relieced maximally by monochromatic light at or near 450 nrn, t,he ratio of CO:02 needed for 50% inhibition of testosterone hydrosylation at positions 60, 7c~, and 16a was very different.
These results suggest that different forms of cytochrome P450 wit.11 different sensitivities to CO participate in the 6@-, 7cu-, aud 16a-hydroxylations of testosterone.
A number of other studies have also shown some differences between the liver microsomes isolated from untreated and PBtreated animals, but it is uncertain from these studies whether different cytochrome P45Os were responsible for the differences. For example, N-demethylation of metharbital in liver microsomes from PB-treated rats was 50-fold faster than the N-demethylation in microsomes from control animals (28). This increased metabolism could not be explained solely on the basis of increased levels of reductasc and 1'450. Guarino et al. (29) found that in PB-treated rats, the K, and K, (spectral dissociation constant) values for aniline in liver microsomes were significantly different from the K, and K, values for aniline in microsomes from control rats. Wiebel et al. (21) reported that acetone slight'ly stimulated 3,4-benzpyrene hydroxylation in microsomes from control rats, but inhibited the hydroxylation of 3.4-benzpyrenc in microsomes from PB-treated rats. J3ased on kinet,ic analysis, it has been suggested that different enzyme systems may be responsible for the metabolism of several drugs in microsomes from control and PB-treated rats (25,30,31).
The ratios of the specific activities for the N-demethylation of (+) and ( -) methylphenobarbital, and the metabolism of (+) and ( -) hcxobarbital by microsomes from control and PB-treated rats were found to be significantly different (26,27).
The hydroxylation of nortriptyline and desmethylimipramine in microsomes was significantlydecreased in rats treated with PB, even though cytochrome P450 content was increased (32). In addition, a small but significant change in half-life of cytochrome P450 after PB-treatment has been reported (33). The ratio of the 455 t.o 430 nm a.bsorption peaks in the ethyl isocyanide difference spectrum is slightly, but consistently lower in untreated rats than in PB-treated rats (34). These results and the data presented in t,he present paper suggest that multiple CO-binding hemoproteins-each with a, different substrate specificity-exist in liver microsomes, but a complete separation and purification of the various I'450 species are needed to firmly establish such a conclusion.

Bckrwwledgnzent-We
wish to thank Xtrs. Cat,hy Chvasta for her assistance in the preparation of this manuscript.