The Interaction of Hepatic Cytochrome P-450 with Organic Solvents THE EFFECT OF ORGANIC SOLVENTS ON APPARENT SPECTRAL BINDING CONSTANTS FOR HYDROCARBON SUBSTRATES*

Studies have been undertaken to explain the ob-sewed variation of the apparent association constant for water-insoluble substrates, which were diluted in common organic solvents, as a direct function of the solvent/solute ratio. By the use of suitable equations, the solvents methanol, ethanol, propanol, and acetone are shown to interact with hydrocarbon substrates in a competitive manner in PB-treated male rats, with the solvent producing a type I spectral component. Such solvents are shown to elicit, in addition to the type I component, a modified type I[ component. In untreated rats, ethanol does not produce a type I component, and also does not affect the apparent association constant for the hydrocarbon substrates when used as a solvent for those substrates. All perturbations of the enzyme which cause a change in the apparent association con- stant of the substrate also cause a quantitatively similar change in the apparent association of the solvent for the enzyme. A sex difference, with respect to com- petitive solvent binding, is also observed. Cytochrome female multiple range were

The addition of compounds such as aniline, aliphatic amines, and imidazole result in a type I1 spectrum when added to cytochrome P-450. Type I1 compounds produce difference spectra with peaks ranging from 425 to 435 nm and troughs from 390 to 410 nm (2, 3). The spectra produced by type I1 compounds are not symmetrical; type I1 spectra exhibit a broad trough. The lack of symmetry has been attributed to the presence of a type I component in the type 11 response (3).
The modified type I1 spectral change is produced by alcohols, ketones, and certain drugs (e.g. phenacetin) and is characterized by a peak at 420 nm and a trough at 385 nm, the apparent mirror image of the type I change. Schenkman et al. (4) investigated the significance of the modified type I1 spectrum and concluded that these changes were the result of the interaction of MII' compounds at both the type I site and another site, apart from the type I1 site. Yoshida and Kumaoka (5) also studied the significance of the MI1 spectrum and concluded that the spectrum is the result of a type I interaction and a second interaction at the type I1 site. An additional hypothesis has recently been proposed (6) with respect to the action of MI1 compounds. It is based on the effect of the MI1 compounds (alcohols and ketones) on the dielectric constant of the aqueous phase of the microsomal preparation. As a MI1 compound is added to the microsomal preparation, a decrease in dielectric constant of the aqueous phase should be observed which would effectively "leach" endogenous substrate from the type I site of cytochrome P-450 into the aqueous phase (6).
When working with water-insoluble substrates such as hydrocarbons, steroids, and hydrophobic drugs, minute quantities of such compounds must be added in order to obtain an accurate value for an association constant. This problem is generally circumvented by dilution of such substrates in organic solvents. Since most organic solvents have been shown to elicit a MI1 spectral change, an accurate determination of the association constant for the type I substrate requires knowledge of the mode of interaction of the solvent with the enzyme. It appears that the MI1 spectral change is the least understood of the above types of spectra. Previous studies from this laboratory (7,8) have demonstrated that solvents such as methanol, ethanol, and acetone can have a profound effect upon the apparent association constant for the type I substrate ethylbenzene. This present paper is concerned with the determination of a reliable value for the association constant for water-insoluble type I substrates, by correcting for the effect of the solvent employed. In addition, quantitative evidence for the mechanism of interaction of MI1 compounds is presented, using a series of equations developed here. The tions were added to the sample cuvette, with equivalent aliquots of water being added to the reference cuvette. After each addition, the resulting spectra were recorded between 450 and 370 nm subsequent to mixing of the contents of the cuvettes. The total volume of the cuvettes was changed by less than 2% (v/v). The magnitude of the spectral change was measured between the peak and trough of the difference spectrum. For type I compounds, the peak and trough generally were found at 386 nm and 421 nm, respectively. The reciprocal of the magnitude of the spectral change was plotted against the reciprocal of the total concentration of substrate after each addition (a spectral analog of the Lineweaver-Burk plot for a rate process). Linear plots were obtained as described by Schenkman et al. (2). The apparent spectral association constant (&appf was calculated as the x intercept from the above described double reciprocal plot. The maximal spectral change (AA,,,) was calculated from the reciprocal of the y intercept. For compounds which were added in organic solvent, the experiments were done by means of one of the following methods.
Procedure I For determinations of spectral binding constants for water-insoluble substrates, a method analogous to that for water-soluble substrates was utilized. In this case, the microsomal preparation was added to the reference and sample cuvettes as described above. Stock solutions of a water-insoluble substrate were prepared by dilution in an organic solvent (e.g. methanol, ethanol, propanol, or acetone). After a base-line for the microsomal preparation was recorded, microliter increments of substrate solutions were added to the sample cuvette with equivalent volumes of the appropriate organic solvent being added to the reference cuvette. A11 other conditions were the same as described above.
Procedure II According to this method, two separate experiments are mathematically added as shown in Fig. 6. Initially, two cuvettes containing 3 ml of the microsomal preparation were placed in the reference and sample c o m p~e n~, respectively. After a base-Iine was obtained, microliter increments of the hydrophobic (ethylbenzene) substrate solution, diluted with organic solvent, were added to the sample cuvette, while equivalent volumes of water were added to the reference. Upon addition of each increment, the spectral change was determined.
Next, a volume of 9 ml of a fresh microsomal suspension was added to a glass-stoppered centrifuge tube. While blending on a Vortex mixer, 75 i Jof ethylbenzene (pure liquid) were slowly added to the suspension. After addition, the tube was closed and the preparation was blended vigorously on a Vortex mixer. Then, 3 ml of this solution were added to both the reference and sample cuvettes. A base-line was determined, and microliter aliquots of the organic solvent used for the addition of the hydrophobic substrate were added to the reference cuvette while equivalent volumes of water were added to the sample. The absorbances after each addition were recorded and added to the absorbances obtained with the previous preparation as described in Fig. 6. The reasoning behind these manipulations will be explained under "Results." High ~i c r o s o m e Concentrations-en microsome concentrations exceeded 5 mg of microsomal protein/ml, determination of difference spectra became difficult due to the high degree of turbidity of the preparation. To circumvent this problem, cuvettes with an 0.45 cm path length were utilized, to which 1.5 ml of the microsomal preparations were added.

Procedure III
Determination of KZPp for hydrophobic substrates in the total absence of solvent. An aqueous KCl/Tris buffer will dissolve an amount of ~thylbenzene corresponding to a p p r o~a t e~y 2 ilil~ of the hydrocarbon. The absorbance at 260 n m was used to calculate the hydrocarbon concentration ( E = 0.258 M" cm"). Meanwhile, a microsomal preparation was concentrated to 14 mg of microsomal protein/&. Two ml of this solution was pipetted into each of 10 tubes. Volumes of 0.5, 1, 1.5, 2, 2.5, 3, 4, 5, 6, and 8 ml of the buffer containing the hydrocarbon were then added to tubes 1 to 10, respectively. Volumes of buffer without h y~~a r b o f f were added to each tube to bring the total volume to 10 ml and the mixtures were blended on a Vortex mixer. Each of these solutions were placed in the sample cuvette, while a microsomal suspension lacking hydrocarbon was added to the reference cuvette. The absorbance of each solution was determined and subtracted from the base-line (the preparation lacking ethylbenzene was added to both cuvettes).

Solvent Scans
A fresh microsomal preparation was added to both the reference and sample cuvettes (3 ml each). Microliter increments of solvent were added to the sample cuvette with equal volumes of water being added to the reference cuvette.
When a hydrocarbon-saturated preparation was required, the me tho do lo^ described under "Procedure II" was used to saturate the microsomes. Microliter increments of solvent were then added to the sample cuvette with equal volumes of water being added to the reference. Endogenous substrates were removed from the microsomal preparation by washing it with defatted bovine serum albumin according to the method of Powis et al. (11). Protein concentrations were d e t e~i n e d using the method of Lowry et al. (12). Statistical analyses using the Newman-Keuls multiple range test were utilized where appropriate.

Materials
Bovine serum albumin, steroid substrates, and Tris were obtained from Sigma. All hydrocarbons were purchased from Eastman. Hexobarbital and methohexital were purchased from Wmthrop Laboratories and Lay, respectively. Absolute ethanol was obtained from United States Industrial Chemicals Co., with all other chemicals being purchased from Fisher.

Effect of Solvent on the Binding of ~y d r o p~o b~e Sub-
strates-When substrate binding to aytochrome P-450 is measured using difference spectroscopy, compounds are commonly diluted in water; increments of substrate solution are added to the sample cuvette and an equivalent volume of water added to the reference ("Experimental Procedures"). Dilution of substrates prior to addition to the microsomal preparation is necessary due to the relatively high affiiities of substrates for cytochrome P-450; thus, dilution is necessary for an accurate determination of the association constant (KJ.
The study of binding of hydrocarbons and other hydrophobic substances poses a particular obstacle due to insolubility in water. To c~cumvent this problem, water-insoluble substrates are commonly diluted in organic solvents such as low molecular weight alcohols or ketones. The diluted substrate solution is added to the sample cuvette while an equal volume of the organic solvent is added to the reference in a manner analogous to that used for water-soluble compounds.
The addition of ethylbenzene (or the other steroid or hydrocarbon substrates used in this study) to a preparation of liver microsomes, results in a type I spectrum which can be monitored by difference spectroscopy. An example of the type

Hydrocarbon-Organic Solvent Interaction
with Cytochrome P-450 7215 I spectrum elicited by ethylbenzene is shown in Fig. 1.
The apparent binding constant (Ksapp) for a number of hydrophobic substances was determined in untreated and phenobarbital-pretreated male rats, using working solutions of different concentrations in ethanol. Each of these compounds, when added to a liver microsome preparation, elicited a type I change, which has been demonstrated to be indicative of substrate binding to cytochrome P-450 (3, 13). By plotting the reciprocal of the magnitude of the spectral change against the reciprocal of the substrate concentration, a linear relationship analogous to the Lineweaver-Burk plot for enzyme kinetics may be obtained and an association constant may be determined.
The results in Table I   for both an aliphatic and an aromatic hydrocarbon. In untreated rats, a change in the concentration of the hydrocarbon stock solutions did not affect iis"pp.
There are at least three possible explanations for this effect. First, the differences in apparent binding constants may be due to the presence of multiple forms of cytochrome P-450. Second, the organic solvent may, in some manner, affect the apparent association constant. Finally, the presence of multiple forms of cytochrome P-450 and the interaction of solvent may both contribute to the observed results. Since equal volumes of substrate solutions were added to the microsomal preparations, and the concentrations of those substrate solutions were different, each of the titrations involved substrate additions over different concentration ranges. For example, when an 81.7 mM stock solution of ethylbenzene was used, the substrate concentration varied from 0.14 mM to 1.08 mM; however, when a 20.4 mM stock solution was used, the substrate concen~ation varied from 0.034 mM to 0.27 m. If different forms of cytochrome P-450 possessing different aEnities for the substrate were present, results similar to those observed in Table I may be expected. If the change in binding constant (ethylbenzene, PB-treated rat) were due solely to the presence of multiple forms of cytochrome P-450, then addition of increasing volumes of the 20.4 mM stock solution to the microsomes to a find concentration of 0.66 mM, should cause a downward break in the double reciprocal plot demonstrating biphasic behavior as suggested by Klotz and Hunston (14,15). This should occur due to saturation of the high affinity binding site, permitting the subsequent titration of the binding site possessing the smaller affinity. Since a final concentration of 0.66 m ethylbenzene should be well into the titration of the second binding site (neither the 81.7 mM nor the 20.4 mM solutions alone demonstrated biphasic nature), biphasic results should then clearly be observed. However, if variation of K / P P is due to a contribution of the solvent rather than the presence of multiple forms of cytochrome P-450, changes in the ratio of solvent to substrate should cause a change in the apparent association constant.
To demonstrate that the change in x/pp observed for some substrates in PB-treated rats is due to changes in the concentration of the working solutions rather than the presence of multiple enzyme forms both capable of binding the same substrate with different affinities, increments of a dilute substrate solution were added to the microsomal preparation, then part way through the titration, a concentrated substrate solution (different substrate/solvent ratio) was used to complete the titration. The results are shown in Fig. 2

7216
~y d r~~~r~o n -U r g a n i c Solvent ~nterac~ion with Cytochrome P-450 reciprocal plot was made to appear biphasic simply by changing the concentration of the working solution. If the concentration of the solution remained unchanged, a linear double reciprocal plot was obtained. Similar results were obtained with hexane and pregnenolone in PB-treated rats (data not shown). This effect could not be demonstrated with ethylbenzene and hexane in untreated rats, as would be expected in the absence of a "solvent effect." These results show that the organic solvent has an effect on the apparent association constant, and that the results obtained for ethylbenzene, hexane, and pregnenolone in the PBtreated male rat are not due to the presence of multiple forms of cytochrome P-450. This treatment does not rule out the existence of a multiplicity of binding sites for other substrates; however, these results demonstrate that for the binding of ethylbenzene, hexane, and pregnenolone, cytochrome P-450 from PB-treated male rats behaves as though only a single enzyme were involved. These results do not indicate that ethylbenzene, hexane, and pregnenolone bind to the same fraction of cytochrome P-450, nor do they suggest that these substrates bind to only one form of cytochrome P-450. Wowever, over the concentration range with which we are working, each of these substrates is apparently interacting with only one form of cytochrome P-450 distinguishable by these methods.
Since the effects demonstrated in Table I and Fig. 2 were shown to be caused by solvent, studies were initiated in an attempt to determine the true association constant for the hydrocarbon ethylbenzene, correcting for the effect of solvent. By using suitable equations, the organic solvent was treated simply as another substrate able to interact with cytochrome P-450 by competitive or noncompetitive mechanisms. The significance of the double reciprocal plots would be dissimilar to the classical interpretations of competitive and noncompetitive interactions because of two factors: 1) Both substrate and solvent may be able to bring about an absorbance change when formed in a complex with the enzyme, and 2) for each aliquot of substrate added, a proportional amount of solvent is also added; the solvent (inhibitor) concentration is not held constant as in the classical i n t e~r e~t i o n .
When two substrates ( A and 3 ) bind to an enzyme (E 1, the following general scheme may be applied.

E
The equilibria involved are: This scheme assumes that no higher complexes (eg. EA2 or EBy4) are formed other than those shown in the above scheme. It can be easily shown K B K ' A = K A K ' B . The following treatment has been derived for conditions where the enzyme substrate complexes result in a spectral change; however, these derivations also hold for rate processes where both substrates yield the same assayable product, and the condition of pseudoequilibrium prevails. The general equation (written in the reciprocal form for convenience) describing the dependence of the spectral change upon the concentrations of both A and B is as follows:

A A T~M (*BKB[B] + 'PAKA[B]N -k * A B~~~~~F A [ B ]~N ) [ E ] O
(2) It will be easily seen that the above equation does not predict a linear double reciprocal plot when one plots l / h A~~~i versus I/[S 1. This equation represents the general noncompetitive case where EB, EA, and EAB complexes all absorb. It can be shown that other noncompetitive mechanisms also yield nonlinear double reciprocal plots under these conditions (8).
Competitive Interactions-If the EB and EA complexes are mutually exclusive, thus prohibiting the formation of EA3, equation 3 can be reduced to a competitive interaction. In this case, K I A and 8'8 are both equal to zero, resulting in the following equation:

+ 'P*K*N)[E]~[B] ('PBKB + ' W L N ) [ E ] o
For this case, plotting l / h A~,~l uersus 1/[3] will yield a linear plot. If l / h A~~~l is set equal to zero, from the x intercept it follows that Z?/pp = (NEA + &I). Equation 3 is illuminating, in that it indicates that if one dissolves a substrate B in a solvent A (which competes for the enzyme binding site), a linear double reciprocal plot d l be obtained. Of course E: "", the association constant, and AA,,, will have different significances than would be the case if substrate B were added in water or a noninteracting organic solvent. By preparing solutions of different concentrations (different N values) for our experiments, the binding constants for both substrates can be determined. This can be accomplished by determining K2" at different values of N and plotting K:PP against N. This procedure will result in a linear plot with the y intercept equaling K B and a slope of K A . A plot of the reciprocal of the slope uersus N will also result in a linear relationship as follows: By setting l / . h A~~~l equal to zero, K$PP can be determined from the x intercept. The significance of this replot is the same as that from equation 3. A replot of l/slope versus N also yields a linear plot.

f4a) with y intercept of *&B[E]o and slope of \ke&[E]o.
A replot of l/AA,., (equation 4) uersus N now results in a relationship that is independent of N.
A special case of equation 3 is when the EA complex can form but does not absorb; therefore, * A equals zero. Equation 3 then simplifies to:

A.4 'Pke[E]&[B]
+ *B[EI&B (5) which also results in a linear relationship. When l/AA is set equal to zero, the x intercept has exactly the same significance as in equation 3 The interaction between hexobarbital and methohexital in binding to the type I site of liver microsomal cytochrome P-450. Solutions of varying hexobarbital/methohexital ratios were prepared. Each of these solutions were then used for the determination of Keapp by adding microliter increments to the microsomal preparation in the sample cuvette. Equivalent volumes of water were added to microsomes in the reference cuvette and changes in absorbances were measured. Addition of substrate solutions resulted in a type I spectral change with peak and trough at 386 and 421 nm, respectively. No changes in the wavelengths of the peaks and troughs were observed as the substrate ratios were varied. Phenobarbitaltreated male rat liver microsomes containing 2.4 mg of microsomal

N with a slope of &/&*~fE]o and intercept of lf*B[E]o. A
replot of the slope from equation 5 versus N will result in no dependence.
According to the above equations, nonlinear double reciprocal plots should be obtained with noncompetitive interactions. Since we do indeed obtain linear double reciprocal plots for ethylbenzene/ethanol binding in PB-treated male rats, the data suggest a competitive mechanism.
Barbiturate Competition-In order to confirm the validity of equation 3 (general competitive case), two barbiturates (methohexital and hexobarbital) were selected. These compounds were chosen because of differences in the association constants for each compound. The extinction coefficients (AA,.,) were similaf for both substrates. Structural similarity would suggest that both compounds bind to the same site in the microsomal preparation, and since the salts of both barbiturates are water-soluble, direct de~rmination of the association constants and AA,,, values for both substrates can be readily obtained in the absence of any solvent other than water for comparison with extrapolated values. of stock solutions of varying methohexital/hexobarbital ratios were added to the microsomal suspension and the reciprocal of the spectral change was plotted against the reciprocal of the methohexital concentration as described under "Experimental Procedures." The effect of varying quantities of hexobarbital on the apparent association constant for methohexital is clearly demonstrated from these results. As predicted by equation 3 and shown in Fig. 3b REpp for methohexital and slope of K$pp for hexobarbital. The extrapolated values for the association constants for methohexital (41,000 M-*) and hexobarbital (10,000 M-*) compare quite well with values obtained by conventional direct determination (42,000 M" and 9,000 M" for methohexital and hexobarbital, respectively).
Replotting l/slope against N also yields a linear relationship as predicted by equation 3 and shown in Fig. 3c Fig. 3, b and c, the theoretical curve for N

FIG. 4. Effect of ethanol
on the apparent association of ethylbenzene for the type I Bite in untreated and PB-treated male rats. Ethylbenzene was diluted to different stock concentrations in ethanol. -Each of these solutions were then used for the determination of as described under "Experimental Procedures" (procedure I). Substrate additions resulted in a type I change with peak and trough at 387 and 422 nm, respectively. Shifts in peaks or troughs were not observed as the ethylbenzene concentration was varied. Microsomes containing 3.0 mg of microsomal protein/ml were prepared from both untreated and PB-treated male rats. (a) Double reciprocal ptots for the apparent association of ethylbenzene to cytochrome P-450 in PB-treated rats. The hydrocarbon concentration l/AA,,, versus N can be plotted as shown in Fig. 3d. Experiments from untreated male rats using the same substrates gave qualitatively similar results (data not shown).
Effect of Ethanol on K s a p p for Ethylbenzene-The concentration of the stock solution of ethylbenzene diluted in ethanol, which is added to the microsomal suspension, has a profound effect on the apparent association constant obtained in PBtreated male rats as shown in Table I and Fig. 2.
From the characteristics of the double reciprocal substrate dependence plots and the secondary plots of K$pp, l/slope, and l/AA,, versus N, the mechanism of interaction between solvent and substrate with the enzyme may be determined, in addition to enabling determination of association constants and extinction coefficients for both substrate and solvent complexes. In a manner analogous to the barbiturate studies, as the value of N ([ethanol]/[ethylbenzene]) increased, the K$pp for ethylbenzene also increased. These results are shown in Fig. 4a. Replotting K$pp versus N resulted in a linear plot from which Ksapp for ethylbenzene can be obtained from the y intercept and for ethanol from the slope (Fig. 4b). The association constants for ethylbenzene and ethanol were 5,000 M" and 12 M..', respectively.
No solvent dependence was observed in the untreated male, indicating no demonstrable binding of solvent to the ethyl-

~y d r~c a r b o n -~r g a~i c Solvent Inter~ction with Cytochrome P-450 7219
benzene type I spectral site. The association constant for ethylbenzene in the untreated rat was 2,800 M-'.
Replotting the reciprocal of the slopes of the lines from Fig. 4a against N  In the untreated male, the slope of the line in Fig. 4c is essentially zero, which is consistent with the absence of an interaction of ethanol with the type I site. The extinction coefficient for ethylbenzene in the untreated rat was 0.06/mg of protein. The slope of the line for the PB-treated male is positive, suggesting that represents a type I response similar to \ks. Alcohols have been shown to possess a modified type I1 (also called a reverse type I) spectrum when added to liver microsomes (4). If this component were the competing species, then \ k A would be negative, rather than the positive result actually obtained. These spectral studies were done by adding ethylbenzene diluted in ethanol to the sample cuvette, and ethanol to the reference cuvette, a common procedure (6,(16)(17)(18)(19)(20). This method causes inconsistencies in the application of the abovementioned equations. These equations were derived in a form which assumes that both substrates were to be added to the sample cuvette and the solvent added to the reference, only if the solvent is a noninteracting one such as water. If substrate and solvent bind to separate n o~n t e r a c t~g sites, then adding solvent to the reference cuvette would be proper procedure and no solvent dependence would be observed. Such is the case with the untreated male rat. A method for correcting the procedure for phenobarbital-treated males follows.
The points from the l/AA,,, versus N plot in the phenobarbital-treated rat are consistent with the curve predicted by equation 3 (Fig. 4d). In the untreated male, l/AA,,, was independent of N as would be expected if ethanol did not interact with ethylbenzene at the type I site (not shown).
Similar results were obtained using either Sprague-Dawley or Wistar rats (data not shown for Sprague-Dawley). Thus, this characteristic is not peculiar to Wistar rats.
Dual Interaction of Ethanol in Phenobarbital-treated Rats-In the phenobarbital-treated male rat there appears to exist a more complex interaction between ethylbenzene, ethanol, and the enzyme than is observed in the untreated male. Ethanol exhibits a modified type I1 interaction in the phenobarbital-treated male, which can be observed by simply adding the alcohol compound to the microsomal preparation and recording the spectral change. This spectral change is also observed in the untreated rat and must be eliminated to obtain linear double reciprocal plots when these compounds are used as solvents for type I substrates. This end is commonly attained by adding ethanol to the reference cuvette. By analogy, the absorbance from the apparently noninteracting MI1 change in phenobarbital-treated rats must also be eliminated. The data from Fig. 4 suggest that in PB-treated rats, ethanol possesses a type I spectral change that appears to compete with ethylbenzene. Since this component is accounted for by equation 4, it should not be eliminated as is the MI1 component by the addition of solvent to the reference cuvette. Therefore, for the system with which we are concerned, there appear to be two sites. One site will bind solvent and exhibit a MI1 spectral change (Ez), the other site will bind the hydrocarbon and exhibit a type I spectral change ( E l ) . In the untreated male, ethanol does not bind to EL; however, ethanol apparently does bind to E , in the PB male. If this is indeed the case, ethanol spectra from the different animals should have different characteristics. In the untreated male, the addition of saturating amounts of the type I substrate ethylbenzene to both cuvettes should not affect the MI1 response to ethanol. In the PB-treated male, however, the MI1 change observed should be the sum of two spectral components, a type I and a type 11. Therefore, the addition of saturating amounts of the type I substrate to both cuvettes should increase the magnitude of the MI1 spectral change by effectively blocking any solvent interaction with the type I site. This is indeed the case, as clearly demonstrated in Fig. 5, b and d. This effect was observed over a wide range of ethanol concentrations. Fig. 5, u and c, demonstrates the lack of effect in the untreated rat.
The interaction of ethanol with the type I site in the PBtreated rat requires that the experimental procedure be redesigned to permit proper determination of the equilibrium constants. Fig. 6 describes the protocol and the species involved. Fig. 7 compares the results obtained with and without the above correction in PB males. A double reciprocal plot comparing corrected with uncorrected data is shown in Fig. 7a. The data clearly show that either method for determining gBpp results in approximately the same value for K B and K A (Fig. 7b). Introducing the correction causes a small decrease in both apparent association constants. Likewise, Fig. 7c shows a small increase in both \k [E]o values. Using the constants obtained from Fig. 7 , b and c, a theoretical curve can be calculated using equation 3. The points from the experimental determ~ations agree well with those predicted by the theoretical curve (Fig. 7 d ) .
Careful observation of Fig. 7a will reveal that the double reciprocal plots deviate from linearity at very high substrate concentrations. A situation that appears to be analogous to substrate inhibition for kinetic data occurs, in that the plots appear to be slightly concave upward. Introduction of the correction causes the disappearance of this deviation exemplifed by a general increase in the correlation coefficients for the double reciprocal plots.
An additional complication would be encountered if ethylbenzene binds to the MI1 site. Although this treatment does not account for such a possibility, it is generally thought that hydrocarbons are pure type I compounds; therefore, not exhibiting the type I1 response (5,21).
When comparing corrected versus uncorrected data, three characteristics may be observed 1) an increase in linearity in the double reciprocal plots, in addition to a displacement of the corrected curve to smaller l/AA values, 2) a decrease should be noted in K B and & of the K$pp versus N replots, and 3) the replot should be linear.
Equation 6 describes the substrate dependence of two competing substrates that are both able to bring about an absorbance change when only one of the substrates (substrate A) is added to the reference cuvette; this describes the uncorrected data (Fig. 4).  by equations 3 and 6 were compared with those obtained in Fig. 7. The results are shown in Fig. 8. The theoretical data shown in Fig. 8 show each of the characteristics demonstrated in Fig. 7, lending further support to the idea that ethanol is acting as a competitive substrate which is capable of eliciting a spectral change. The reason that generally satisfactory results were obtained without introduction of the correction is that if K A is relatively small and N is not too large, equation 6 becomes approximately equivalent to equation 3.
To support the validity of the extrapolation method, deter-  Table 11. The association constant for ethylbenzene was determined in the absence of solvent by taking advantage of the very limited solubility of the hydrocarbon in an aqueous buffer. An aqueous KCl/Tris buffer (0.15 M, 50 mM, pH 7.4) will dissolve ethylbenzene to a concentration of approximately 2 mM. The concentration of ethylbenzene presented to the microsomes was varied by mixing various ratios of buffer containing the hydrocarbon with buffer lacking the hydrocarbon and adding these solutions to microsomal preparations that were concentrated to 14 mg of microsomal proteinlml as described under "Experimental Procedures." The results clearly demonstrate good agreement between values obtained by the extrapolation method and those obtained in the absence of solvent for both E/pp and \k~[Efa. Large amounts of microsomal material were required in determining E/pp for ethylbenzene in the absence of solvent (many times that necessary for an extrapolation). Furthermore, direct addition of hydrocarbon is a tedious and time-consuming process which is impractical for the determination of large numbers of binding constants.
Effect  Fig. 9a clearly show that the binding of ethanol to the type I site in PB-treated males was not due to the presence of different levels of endogenous substrates in the microsomes of the PB uersus untreated rats. In the untreated male the extraction process caused essentially no change in the apparent association constant for ethylbenzene. Independence from solvent eiYects remained, Endogenous substrate extraction of microsomes from PB-treated male rats did not cause the disappearance of the solvent dependence. In fact, the association constant for ethanol as determined from the slope of the K s a p P versus N plot actually increased very slightly, as did that for ethyibenzene.
Using the data obtained from Fig. 9b, A A ,  endogenous substrates. The association constants for both ethanol and ethylbenzene increased 21% and 13%, respectively, when subjected to bovine serum albumin treatment.

The constants \ k~[ E ] 0 and \~B [ E ]~
also varied in a similar manner when measured on a percentage basis as a result of bovine serum albumin treatment. Increases in extinction coefficients of 8 and 9% were observed for ethanol and ethylbenzene, respectively.
If a competitive interaction occurs between an added substrate and an endogenous substrate, it can be shown that a linear double reciprocal plot will be obtained and the apparent association constant has the following significance: which q u a n t i~t i v~l y indicates that the presence of an endogenous substrate causes a decrease in the apparent binding constant for an exogenous substrate.

Eswp for ethylbenzene in the absence of solvent in PB-treated male rat liver microsomes
The association constants and M,, for ethylbenzene were determined in PB-treated microsomes using two methods: Method 1, a direct determination using a solventless system, and Method 2, the extrapolation method involving the dual cuvette experiments described in Fig. 6. The equilibrium constants for ethanol obtained from solvent extrapolation are also presented. Microsome concentration was 2.7 mg of microsomal proteinlml.   Ethylbenzene Extrapolations in Various Solvents-Phenobarbital-treated male rats possess, in addition to a sensitivity to ethanol, a similar sensitivity to other small polar organic solvents. Table 111 summarizes the results from extrapolations for the substrate ethylbenzene. For each of the solvents tested, the data were consistent with equation 3 where the solvent binds in a competitive manner with the type I site of the enzyme and absorbs to give a type I change. As demonstrated in Table 111, the extrapolated values for the association constants for ethylbenzene and \kt$Elo were independent of the solvent used. However, the association constants and %..I[E 30 for the different solvents were shown to be variable.
A plot of the apparent free energy of binding for the various solvents against a number of carbon atoms in that solvent are shown in Fig. 11. A linear relationship was obtained with a slope of -0.6 kcal/mol/carbon atom added. A number of laboratories have demonstrated that substrate and inhibitor binding to enzymes results in such linear relationships (19,23-28) of approximately the same slope; partitioning processes in general have been shown to produce similar results (approximately -0.7 kcal/mol/carbon atom added). It is clear that the solvents behave in the same way as other ligands in regard to size dependence of binding. The size dependence of alcohol binding to cytochrome P-450 is only an approximation in that the values obtained for the individual alcohol binding constants are only as accurate as the solvent correction described in Fig. 6. A second inaccuracy presents itself in that a microsomal partitioning factor is present in this result, as described in the preceding section; however, since a microsomal concentration of roughly 2 mg of protein/& was used in the above experiments, this effect will be relatively small.

TABLE I11 EIfect of various solvents on K A a P p for ethylbenzene in untreated and PB-treated rats
The association constant and A A m a x for ethylbenzene was determined in microsomal preparations from untreated and PB-treated rats in a variety of organic solvents. Equilibrium constants for each of the solvents are also presented. Microsomal concentrations in untreated and PB-treated preparations were 2.6 mg of microsomal protein/ml and 3.0 mg of microsomal protein/ml, respectively. Also illustrated in Fig. 11 FIG. 11. Dependence of the free energy of binding of the type I and type 11 components for a series of alcohols with molecular size in PB-treated male rats. The free energy (AGO) of binding of the type I component (0) was determined from the solvent dependence of ethylbenzene as described in Table I11

Determination of ethanol solvent dependence with various substrates
Association constants and extinction coefficients for a number of different substrates were determined by the extrapolation method described under "Experimental Procedures" and Fig. 7. Studies were performed on a microsomal suspension from PB-treated male rats containing 2.8 mg of microsomal proteiniml. Substrate solutions were diluted to various con cent ratio^ in ethanol.  Table IV. Two hydrocarbons of similar structure were used, since it would be reasonable to assume that both substrates would bind to the same species of cytochrome P-450. The results demonstrate that these hydrocarbons possess greatly different affinities for cytochrome P-450; however, in each case the apparent binding constant for the solvent ethanol is the same.

Effect of Ethanol on the K$pp for Ethylbenzene in Female
Rats: Demonstration of Sex Differences-The results of extrapolations of ethylbenzene diluted in ethanol for both untreated and phenobarbital-treated females are shown in Table  V. The results are consistent with those predicted by equation 3. Table V shows the interaction of ethanol in the PB-treated female rat having a ftsapp for ethylbenzene and ethanol of 2,400 to 9, OOO M" and 9 M-', respectively. (The reason for the high degree of variability in the K$pp for ethylbenzene is unknown at this time, but may be due to the interaction of cyclic hormonal changes within the individual female animal with the drug effect.) When compared with untreated females, a 2fold increase in cytochrome P-450 levels was observed in PBtreated females. The X:pp for ethanol in the PB-treated female is significantly different from that found in the PBtreated male (Table V). Although a difference in the K $ P P between induced males and females could not be demonstrated for ethylbenzene binding, a sex difference in for ethanol binding does exist. Results from the untreated females when compared to untreated males (Fig. 4) also show a sex difference with respect to apparent ethanol binding to the type I site (Table V); ethanol does bind to the type I site in the untreated female; the association constant for ethanol being 5 M-'. Since ethanol dependence was not observed in the untreated male, this indicates a sex difference for ethanol binding to the type I site, despite the similarity in the association constant for ethylbenzene (Table V).

DISCUSSION
Data obtained from PB-treated male rats where ethylbenzene is diluted in ethanol are consistent with the equations describing the solvent (ethanol) as a competitive substrate. According to this treatment, the type I site is capable of binding either ethanol or hydrocarbon with both complexes (EA and EB) eliciting a spectral change. In addition, this competing solvent possesses a type I component which is superimposed on the MI1 spectrum in PB-treated male rats. Evidence for this interpretation is 2-fold. 1) The slope of the l/slope uersus N replot has a positive sign (Figs. 4c and 7c), indicating that the competing species (solvent) causes an absorbance that is in the same direction as that of the type I substrate. 2) Ethanol and other solvents are able, when added alone, to elicit a spectral change (4,17), which has been called either a modified type I1 or reverse type I spectrum. If the MI1 spectrum represents the sum of a type I and a type I1 component, then blocking type I binding of the solvent by addition of saturating quantities of a type E substrate (ethylbenzene) to both cuvettes will cause an increase in the magnitude of the MI1 change. This point has also been demonstrated by other investigators (4, 5 ) . This effect was only observed in the present work when a solvent dependence consistent with that presented in Fig. 7 was also present; the untreated male rat produced neither of these results.
The concept of dual interaction, where the MI1 spectral change is the result of two overlapping spectral components, has been proposed by both Schenhan's laboratory (3,4) and Yoshida and Kumaoka (5). Schenkman et al. used phenacetin as the MI1 compound and demonstrated that the magnitude of the phenacetin spectrum increased in untreated rat liver microsomes when a type I compound (aminopyrine or hexobarbital) was present in both the reference and sample cuvettes. Yoshida and Kumaoka (5) presented similar evidence with the MI1 compound butanol by demonstrating an increase in the magnitude of the butanol spectral change when the type I compound pentane was added to both sample and reference cuvettes containing liver microsomes from PBtreated male rats. The treatment presented here gives strong quantitative support to this proposed mechanism. The results presented in Table I and Fig. 2 demonstrate the large discrepancies that may occur as a result of the use of organic solvents to dilute those hydrophobic substrates. Although the magnitude of the effect may vary depending on the solvent used, it is by no means negligible. Therefore, in the study of sparingly water-soluble substances, the substrate must either be added without an organic solvent or the affinity of the solvent for the site in question must be determined so that K$pp in the absence of solvent can be calculated by extrapolation. Use of the former method (as described under "Experimental Procedures") is tedious and utilizes much larger quantities of microsomes when compared with the solvent extrapolation. In addition, the microsomes which are added to the reference and sample cuvettes must be diluted separately; therefore, small pipetting errors w i l l affect the base-line which cannot be directly determined. This w i l l increase the amount of scatter observed with the experimental points. Despite these shortcomings, it may be advantageous to use this method under certain circumstances.
Hydrocarbons, steroids, and water-insoluble drugs are commonly added to the cytochrome P-450 system using organic solvents as the vehicle. Results obtained from many of these studies may be seriously in error due to effects of the solvent. Therefore, if solvents are used as the vehicle for a substrate, at the very bast an extrapolation similar to that shown in Fig,  4 must be performed to determine the magnitude of any effect caused by the presence of solvent.
As shown in Fig. 2, if solvent interacts with the enzyme, then simply changing the ratio of solvent to substrate should cause a break in the linearity of the double reciprocal plot at the point where the substrate concentration of the stock solution is changed. This is a simple and rapid method for determination of the presence of a "solvent effect" of the type studied here. Such a procedure (Fig. 2) should be utilized when nonaqueous solvents must be used in order to determine whether or not a solvent has an effect on the formation of the enzyme-substrate complex.
As noted in Fig. 8a, deviation from linearity at very high substrate concentrations was observed using procedure I (uncorrected method) where solvent was simply added to the reference cuvette. When the uncorrected method was utilized, this deviation was characterized by a concave upward curvature in the double reciprocal plots. Guengrich (16) also noted such an effect with 7-ethoxycoumarin (acetone was used as the solvent) binding to purified cytochrome P-450 from PBtreated rats, but was unable to explain the cause of such a deviation. (The effect was attributed to nonspecific binding at high substrate concentrations.) Ow results clearly demonstrate an increase in linearity of the double reciprocal plots by redesigning the experiments as described in Fig. 6. The presence of the concave upward curvature with 7-ethoxycoumarin in acetone (Fig. 5, Ref. 16) on a pursed cytochrome P-450 preparation suggests that phenobarbital-induced rat P-450 B has the ability to bind small molecules such as acetone and presumably alcohols, and may be the same binding site with which we are dealing in our microsomal preparations.
Despite the ability of the correction procedure (Fig. 6) to increase the linearity of the double reciprocal plots, it must be stressed that the correction is only an approximation. The accuracy of the values obtained by this methodology is based on the ability of the blocking (type I) substrate to exclude solvent binding at the type I site. An increase in the accuracy of such a correction may be accomplished by using a substrate with a very large affinity for that type I site.
In this study, the suggestion has been made that in the PBtreated rat, the ethanol-induced MI1 spectrum i s the sum of both a type I and a MI1 component, where in the untreated rat, the ethanol-induced spectrum is simply the result of the MI1 component. The po~ibility s t 8 exists that ethanol can bind to a type I site of an isozyme in the untreated rat that was not detected by these methods; however, type I binding is not a necessary characteristic for the categorization of MI1 substrates. The presence of a type I component in MI1 substrates would be expected to depend not only on the substrate employed, but also on the isozymic profile of the tissue preparation.
Throughout this study, anything that was done to perturb the apparent type I binding of the hydrocarbon substrate also affected the apparent binding of the solvent in a similar manner. When the PB-pretreated microsomal preparation was treated with defatted bovine serum albumin in order to extract endogenous substrates, a small increase was observed in K:pp for ethylbenzene with a corresponding small increase in K2pp for ethanol. Results of this t-ype would be expected if the bovine serum albumin treatment removed a small quantity of an endogenous substrate (probably including phenobarbital) that was competing for the type I site. The presence of this "inhibitor" should cause a decrease in the apparent binding constant as previously predicted by equation 7. If the increase in K: pp for the hydrocarbon was simply due to the removal of a competitive inhibitor, an increase in K / p p for ethanol should also be observed, as demonstrated in Fig. 9. Parry et al. (22) have quantitatively treated the effect of membrane concentration on the association of substrates for sites on membrane associated enzymes. Ebel et aE. (6) demonstrated this effect by measuring the apparent binding of type I substrates to liver microsomal cytochrome P-450 at various microsomal concentratiom. According to this treatment, as the microsomal concentration increases, the reciprocal of the association constant (l/KZPP) also increases in a linear fashion. Therefore, if the solvent dependence observed in the PB-treated rat is due to competition for ethanol for the type I (ethylbenzene) binding site, then both ethylbenzene and ethanol should vary in a similar manner. According to Fig. 10, both ethylbenzene and ethanol follow the same kind of law with respect to the effect of microsome concentration, Again the solvent ethanol is acting as though it were simply an alternative substrate.
Further evidence that the observed solvent dependence is the result of Competition between ethanol and ethylbenzene for the type I site is demonstrated by the invariance of K:pp for ethanol when the affinities of different h y~o c~b o n substrates were measured (Table IV). In these studies, two hydrocarbons were chosen as substrates since structural similarity would suggest that these substrates would bind to the same enzyme site. If these hydrocarbons do bind to the same fraction of cytochrome P-450, and ethanol competes with one of the hydrocarbons, then ethanol should also compete with the other hydrocarbon and exhibit the same solvent dependence, since the solvent dependence simply reflects K:PP for ethanol. Deviations from these predicted results may occur. One possible case is where multiple binding sites are present. As shown by Klotz and Hunston (14,15), if two sites are present that bind the same compound (in this case, ethanol), the apparent association constant for the fist site titrated (high affinity site) equals -El, while the apparent association constant for the low affinity site equals --2~& / # 1 + Kz). If the latter site i s being studied, deviations would be expected. Another possible case is where the substrate and solvent are small enough that both compounds may simultaneously fit in the single binding site. In this case, the characteristics of the double reciprocal plots might be expected to fit either a noninteracting (simple one substrate case), or a noncompetitive situation.
The dependence of Kt (association constant for enzymecompetitive inhibitor complex) upon molecular size of the inhibiting molecule has been demonstrated for a-chymotrypsin (23,24), yeast alcohol dehydrogenase (26), and C-1 esterase (27). The inhibitors used in each were an isochemical series of aromatic hydrocarbons similar to that used here. This dependence was then compared with that obtained for the distribution of the same compounds between an organic solvent and water. Similar reports have been published with the cytochrome P-450 system where the size dependence of binding of a series of hydrocarbons (19), barbiturates (29), and alkylamines (28) Table 111 and Fig. 11 are consistent with those predicted.
It is interesting to note that the variation of AGO with size has a slope of approximately -0.6 kcal/mol/carbon atom increase in chain length for the alcohol series. This result is similar to the slope obtained for partitioning of the same compounds between octanol and water (approximately -0.7 kcal/mol/carbon atom). In this respect, the organic alcohols exhibit a linear size dependence and again behave as though they are simply alternative substrates. Although a small increase in the slope of the size dependence plot might be observed by extrapolating to zero microsome concentration, these results offer a very good estimate of the extrapolated values since the microsome concentration is relatively low in these studies (8). The dependence of AGO of binding for the alcohol series on the number of carbon atoms added compares well with the dependence of AGO of binding for a hydrocarbon series which binds to the same site in the PB-treated rat (8).
Since both sets of compounds (alcohols and hydrocarbons) share a basic structural feat,ure (alkane moiety), similar size dependences might be expected. These results indicate that hydrophobicity is a major factor in binding of these compounds to the type I site. However, these results do not indicate that the hydroxyl group is unimportant with respect to alcohol binding, only that the increase in K$pp observed by

~ydrocarbon-Organic Soluent Interaction with
Cytochrome P-450 adding a methylene group is unaffected by the presence of the hydroxy1 group. Therefore, the hydroxyl group apparently makes approximately the same contribution to the binding of the various alcohols to cytochrome P-450. It is interesting to note that a linear size dependence was also observed for the binding of the type I1 component of the MI1 spectrum, again suggesting substrate binding. The slope of this size dependence was -0.3 kcal/mol/carbon atom. Although quantitative data are not yet available, Yoshida and Kumaoka (5)  compound phenacetin would not affect the carbon monoxide spectrum when cytochrome P-450 was in the reduced form, while type I1 substrates did produce such an interaction. From these data, Schenkman et al. (4) suggested that the MI1 spectral change is the result of substrate binding to a unique site different from the type I1 (heme) site. Studies from our laboratory have demonstrated that methanol and ethanol, which are considered to be MI1 compounds, are able to decrease the magnitude of the carbon monoxide spectrum of reduced cytochrome P-450 (data not shown). The variability of the interaction between reduced GO spectra and different MI1 substrates (e.g. phenacetin and methanol) suggest a qualitative difference between MI1 binding of different substrates, and that some MI1 compounds (e.g. alcohols) may actually interact at the heme as do type I1 compounds. Since type I1 compounds affect CO spectra in a manner similar to the MI1 compounds ethanol and methanol, a similarity in mode of interaction between MI1 and type I1 compounds might be expected.
Jefcoate et al. (28) determined the size dependence for a series of alkyiamines (type I1 compounds), from which a slope of approximately -0.7 kcal/mol/carbon atom was obtained. If the type If component of the alcohol spectrum and that of the amine spectrum represent binding to the same site, quite different size dependences might be expected if the contribution of the amine functions and hydroxyl functions to the entropy of binding were different. Comparison of the size dependence of amine binding (-0.7 kcal/mol/carbon atom), as determined by Jefcoate, with that obtained for MI1 alcohol binding (-0.3 kcal/mol/carbon atom), as determined in this study, supports the above possibility. It must be stressed, however, that this suggestion is based on presumptive evidence. Strong support for this argument may be obtained by a quantitative demonstration of competition between the type I1 component of an amine and that of an alcohol. Use of the equations previously presented should be of particular value in this determination. These experiments are in progress.
A number of investigators have suggested that the MI1 spectral change is the result of the removal of endogenous substrate from the type I site (4, 6, 11). Ebel et al. (6) have refined this hypothesis by suggesting that the modified type I1 spectral change is due to a perturbation in the dielectric constant of the aqueous phase which would increase the relative solubility of substrates in that phase and thus essentially "leach" endogenous substrates from the type I site.
Although the evidence presented here cannot totally discount this possibility, preliminary experiments have demonstrated oniy a small decrease in the chemical potential of the substrate in the aqueous phase by the addition of ethanol to a final concentration of 225 mM (approximately 50 pi in 3 ml of aqueous buffer) (41). Studies by Vore et al. (31) conclusively demonstrated that butanol extraction of liver microsomes from PB-treated rats did not decrease the magnitude of the ethanol-or phenacetin-induced MI1 spectra. In fact, the mag-nitude of the MI1 spectral change was actually increased.
Butanol, added to a lyophilized microsomal pellet, was shown to remove a significant amount of endogenous substrates. If the MI1 spectrum is caused by "leaching" of endogenous substrates from the type I site, the magnitude of the MI1 spectrum in butanol-extracted microsomes should be decreased. From their results, Vore et al. (31) concluded that the MI1 spectrum is not due simply to displacement of endogenous substrates. Butanol treatment exposes the microsomes to a medium with a dielectric constant of approximately 17.1 (32). The dielectric of a solution to which ethanol has been added in concentrations commonly encountered in spectral studies varies from 80 (water) to approximately 78 (32). Therefore, it seems unlikely that "endogenous substrates" that could not be removed by butanol could then be "leached" from the enzyme by the small change in dielectric caused by ethanol addition to the aqueous suspension and in turn bring about a MI1 spectral change.
The possibility may arise that endogenous substrate could not be removed from the enzyme due to steric factors caused by protein conformation. If such a case occurs, however, the endogenous substrate should be considered to be acting in a sense as a prosthetic group ( i e . a functional part of the enzyme) and should not be thought of as an endogenous substrate. This laboratory has recently reported that a reverse type I spectral change can be observed by addition of a hydrophobic substance to a protoferriheme solution (33). Further studies have demonstrated that the heme monomer is the form responsible for this spectral change? It has been proposed that, by analogy, a hydrophobic substance (possibly phospholipid) may be in the vicinity of the heme in the enzyme and a change in the position of this moiety (conformational change) with respect to the heme could elicit the type I spectral change (33). Although the presence of a pros-

H y~r o c a r~o n -O r g~n i~
Solvent Interaction with Cytochrome F-450 7227 of cytochrome P-450. Substrates which bind to the same site should exhibit similar solvent dependences; however, a particular solvent dependence should not necessarily be assumed for structurally different compounds. The reasons for this are 2-fold: 1) two different substrates might not bind to the same form of cytochrome P-450, if this occurs, different solvent dependences may be expected. 2) One of the substrates may bind to more than one site. If this occurs, the apparent association constant for the second (low affinity) site would have a different significance, as explained previously (14, 15); therefore, the significance of the solvent dependence may also be affected. Van den Berg et al. (30) presented evidence supporting the existence of a sex difference in a type I binding component for butanol from mouse liver microsomes. These investigators demonstrated that liver microsomes from untreated female mice would bind butanol and elicit a type I component in addition to the MI1 spectral change, but liver microsomes from untreated male mice would not. Similar results were obtained in this study, with a smaller alcohol (ethanol) binding to the type I site in untreated female rats, but not in untreated male rats.
The binding of ethanol was found to be significantly different when comparing untreated with PB-treated female rats, although microsomes from female rats under both treatments were able to bind the solvent. It is interesting to note that a large variation was found in the association constant for ethylbenzene in the PB-treated female rat, but only a small variation was observed in the association constant for ethanol. The reason for this anomaly is at this time unexplained, but may be related to cyclic hormonal changes in the female rat. Studies are currently in progress to clarify this situation and to elucidate the mechanism involved in the sex-dependent hormonal regulation of solvent binding.
Mezey et al. (34) demonstrated that ethanol was oxidized by a liver microsomal component containing high concentrations of cytochrome P-450 and cytochrome c reductase, which were found to be required for maximal ethanol-oxidizing ability. Additional studies demonstrated inhibition of ethanoloxidizing capacity by the addition of inhibitors of drug metabolism such as SKF-525A, carbon monoxide, and cytochrome c (34)(35)(36). Microsomal ethanol-oxidizing activity was also shown to be induced by ethanol pretreatment (37,38). Despite this evidence, the involvement of cytochrome P-450 in microsomal ethanol oxidation has been questioned, based on evidence implicating contaminating catalase as the system involved with ethanol oxidation (35,39,40).
Since substrates for the cytochrome P-450 system have been shown to elicit a type I spectrum, such a spectral change may be expected with ethanol if it was metabolized by cytochrome P-450. This study clearly demonstrates that ethanol can produce a type I component to the MI1 spectrum in PBtreated male rats, supporting the possibility of mixed function oxidation of ethanol; however, the type I component of the ethanol spectrum was not detected in untreated male rats (Figs. 4 and 5). It is interesting to note that in all studies where cytochrome P-450-dependent oxidation of ethanol was discounted (35,39,40), microsomes from untreated male rats were used, but in the study where cytochrome P-450-dependent ethanol oxidation was most clearly demonstrated (34), PB-treated male rats were used. These results support the possibility of differential involvement of cytochrome P-450 in ethanol oxidation when comparing PB-induced and untreated male rats.