Cytochrome P-450 of adrenal mitochondria. Spin states as detected by difference spectroscopy.

Adrenal mitochondrial cytochrome P-450 which functions in cholesterol side chain cleavage (P-450scc) exhibited type I (lambdamax 385, lambdamin 420 nm) and inverse type I (lambdamin 385, lambdamax 420 nm) difference spectra with several steroids. The magnitude and type of response were dependent on the particular steroid and on the extent to which cholesterol was bound to the cytochrome in the intact mitochondrion. the inverse type I difference spectrum induced by 3beta-hydroxy-pregn-5-ene-20-one (pregnenolone) was dependent on the proportion of high spin cholesterol-cytochrome P-450scc complexes. With rat adrenal mitochondria cholest-5-ene-3beta, 20alpha-diol (20alpha-hydroxycholesterol) invariably induced a smaller inverse type I response and, under conditions where cytochrome P-450scc was nearly free of cholesterol, even produced a small type I response. Two distinct steroid binding sites on cytochrome P-450scc were detected by, respectively, the slow type I response to cholest-5-ene-3beta, 25-diol (25-hydroxycholesterol) and the rapid type I response to a subsequent addition of cholest-5-ene-3beta, 20alpha, 22 R-triol (20alpha, 22R-dihydroxycholesterol). The relative proportions of the spectral responses to these steroids were dependent on the previous extent of adrenal activation by adrenocorticotropic hormone (ACTH), because this stimulatory process altered the combination of mitochondrial cholesterol with cytochrome P-450scc. It is proposed that the two steroid binding sites on cytochrome P-450scc interact with steroids in the following way: site I binds cholesterol, 25-hydroxycholesterol, and 20alpha, 22R-dihydroxycholesterol with formation of a partially high spin cytochrome; site II binds both pregnenolone and 20alpha-OH cholesterol resulting in a low spin cytochrome. Interactions between sites I and II are not competitive, and occupancy of site II ensures a low spin state irrespective of the occupancy of site I. A second mode of interaction by 20alpha, 22R-dihydroxycholesterol stabilizes a high spin cytochrome and is competitive with site II binding by 20alpha-hydroxycholesterol or pregnenolone. Formation of a maximally high spin cytochrome follows occupancy by 20alpha, 22R-dihydroxycholesterol at both sites.


SUMMARY
Adrenal mitochondrial cytochrome P-450 which functions in cholesterol side chain cleavage (P-450,,,) exhibited type 1 omx 385, Xmin 420 nm) and inverse type I (Xmin 385, x max 420 nm) difference spectra with several steroids. The magnitude and type of response were dependent on the particular steroid and on the extent to which cholesterol was bound to the cytochrome in the intact mitochondrion. The inverse type I difference spectrum induced by 3/3-hydroxypregn-5-ene-20-one (pregnenolone) was dependent on the proportion of high spin cholesterol-cytochrome P-450,,, complexes.
The relative proportions of the spectral responses to these steroids were dependent on the previous extent of adrenal activation by adrenocorticotropic hormone (ACTH), because this stimulatory process altered the combination of mitochondrial cholesterol with cytochrome P-450,,,.
It is proposed that the two steroid binding sites on cytochrome P-450,,, interact with steroids in the following way: site I binds cholesterol, 25 The conversion of cholesterol to 3fi-hydroxypregn-5-ene-20.one appears to be the rate-determining step in steroidogenesis in the adrenal cortex (1,2) and probably in other steroid-producing tissues. The cholesterol side chain cleavage step is activated by the action of adrenocorticotropic hormone upon the adrenal cortex through a series of steps mediated by cyclic adenosine 3': 5'monophosphate (2)(3)(4). This activation of cholesterol side chain cleavage is retained by intact adrenal mitochondria after isolation (5, 6) but is lost upon sonic disruption (7,8), and is not observed when either cholest&ene-3fi,2Ocr-diol (9) or cholest-5ene-3p, 25-diol (8) is the substrate of the side chain cleavage sys-. tern. Activation of steroidogenesis by ACTHr is blocked by the presence of puromycin (10) or cycloheximide (ll), both inhibitors of protein synthesis (12). There is evidence that ACTH induces the synthesis of a labile cholesterol carrier protein (12).
The side chain cleavage of cholesterol in the adrenal cortex is dependent upon a form of mitochondrial cytochrome P-450 (13), which has been partially separated from another cytochrome P-450, which functions in steroid lip-hydroxylation (14,15). The terms cytochrome P-450,,, and cytochrome P-450,,0 have been adopted for these forms of the cytochrome. The binding of specific steroids (16) and amines (17) produces optical difference spectra which can be related to changes in the spin state of the various P-450 cytochromes as observed by EPR spectroscopy (18,19). A combination of optical difference spectroscopy and EPR spectroscopy has shown that the action of ACTH upon rat adrenals causes a 2-to 3-fold increase in the proportion of cytochrome P-450,,, which is bound by cholesterol in isolated mitochondria (20).
Kinetic studies with acetone-extracted adrenal mitochondria have indicated that 20a-and 22R-hydroxycholesterols and cholest&ene-3fl, 2Oar, 22R-trio1 are converted rapidly to pregnenolone by the side chain cleavage system. However, the majority of the cholesterol side chain cleavage reaction passes without significant liberation of intermediates directly to pregnenolone (13, 21). Spectral changes produced by the interactions of these Female Wistar or Holtzman rats (150 to 200 g) were subjected to ether stress or were given injections intraperitoneally with cycloheximide (10 mg/rat in 0.5 ml of water) as previously described (0). The rats were killed 10 to 15 min after the initiation of ether stress or the cycloheximide injection. In a few experiments adrenals were removed from live animals under ether anesthesia but otherwise adrenals were removed rapidly after decapitation.
After removal of excess fat, the adrenal glands were placed in 0.25 M sucrose and mitochondria were prepared as previously described (6). Spectral measurements were generally carried out with mitochondria suspended at 0.2 to 0.4 mg/ml in 0.25 M sucrose, 20 mM KCl, 15 mM triethanolamine hydrochloride, 10 rnM potassium phosphate, and 5 mM MgClz (pH 7.0 unless stated otherwise). Steroids were added in acetone solution and optical changes were recorded with an Aminco-Chance spectrophotometer operated in the dual wavelength mode. Proteins were measured by a modified biuret method in which material that remained insoluble at the end of the normal procedure was extracted with chloroform. 25 from the adrenals of ether-stressed rats until no further response was obtained (fop left). The downward deflection represents an inverse type I response which has heen confirmed by direct observations of difference spectra both here and in other laboratories (23,25). A subsequent addition of pregnenolone produced a further inverse type I response. iThen deoxycorticosterone was added after both steroids an upward type I response was observed which was the same as was observed in the absence of 20a-OH cholesterol and pregnenolone. When the rats had been given injections of cyclohesimide prior to ether stress (top right), a selective decrease in the response to 20cr-OH cholesterol was observed. The lower part of the figure shows four experiments in which 25-OH cholesterol and 20,22R \vere added to two separate suspensions of mitochondria from each of the two groups of rats. In each case a type I response was obtained although the magnitude and time course of the response varied significantly.
We examined the characteristics of these spectral changes in further detail.
Binding of 25.IIZldroxllckolesterol and 20200(, 22R-Dihydroxyclzolesferol-The spectral response to saturating concentrations of 20,22R was 5 to 10 times larger than the spectral response to saturating concentrations of 25.OH cholesterol (Fig. 1). The majority of the spectral change which was induced by 20,22R Teas complete within seconds, but an additional 10 to 20% required 15 min for completion. The rate and magnitude of the slow response to 20,22R were similar to those found for type I binding  (using Ac390-r20 = 130 rnM-l cm-l (ZG)].  ol-steroids such as pregnenolone or methylandrostenediol (Fig.  2). By assuming direct competition between methylandrostenediol and 20,22R, the apparent binding constant for the interaction of methylandrostenediol with this low spin cytochrome P-450,,c was determined to be 0.4 PM. The addition of methylandrostenediol to adrenal mitochondria induced an inverse type I difference spectrum (X,,, 420, Xmin 385 nm). By means of this spectral change the affinity of methylandrostenediol for the high spin form of cytochrome P-450,,, (Fig. 2B)3 was compared with the affinity for the low spin form which binds 20,22R. Within experimental error, these two binding constants were the same (0.4 and 0.45 PM). AA,,, for the inverse type I response induced by methylandrostenediol was greatly enhanced by the presence of 20,22R although the binding constant was increased (Fig.  2B) to an extent which was consistent with full competition between the two steroids.
In virtually all of the experiments carried out at the start of this study, stress pretreatment of female Wistar rats produced little change in the 25-OH cholesterol t,ype I difference spectra ( In all of the experiments, the specific 20,22R type I response was scarcely affected by pretreatment. Binding by %?OZOa-Hydroxycholesterol-The inverse type I response to 2Ocr-OH cholesterol was less than that induced by pregnenolone or methylandrostendiol (Fig. 1, Table II). An additional spectral response to pregnenolone was observed after the completion of the response to 20~OH cholesterol which is referred to as the residual pregnenolone response. The total of the inverse 3  type I response to 2Oa(-OH cholesterol and the residual pregnenolone response was equal or somewhat smaller than the response to pregnenolone alone (Table III). After stress treatment of the rats the inverse type I response with 20~OH cholesterol increased 3 to 5 times relative to values obtained after cycloheximide treatment, whereas there was only a small decrease in the residual pregnenolone response (Table II, Fig. 3A). The dependence of the inverse type I spectral change on steroid concentration was similar when pregnenolone was added alone or after 20~OH cholesterol (Fig. 3B).
The competition between 20,22R and SOWOH cholesterol in binding cytochrome P-450,,, was then examined. The presence of 9 PM 20,22R substantially enhanced the inverse type I response to 20a-OH cholesterol but increased the binding constant from 1.3 PM to 100 PM (Fig. 4A). The increased inverse type I response apparently arises from a complete reversal by 20a-OH cholesterol of the type I change which had been induced by 20,22R. Data represented in Fig. 4A suggests that, 20~OH cholesterol and 20,22R compete directly for a binding site. By using the above binding constant for 20,22R (0.45 PM), the binding constant of 20~OH cholesterol for this form of cytochrome P-450,,, in the absence of 20,22R was calculated to be 2 to 2.5 MM. Thus, the interaction of 20a-OH cholesterol with this low spin form is somewhat weaker than the inverse type I binding to high spin cytochrome P-450se0. There was no high affinity component of 20cu-OH cholesterol binding in the presence of 20,22R which might correspond to the inverse type I binding of 20a-OH cholesterol to the endogenous high spin complex of cytochrome P-450,,,. Thus 20,22R hinders the interaction of 20~OH cholesterol with endogenous high spin cytochrome P-450,,, (Fig. 4) Fig. 1) at a protein concentration of 1.2 mg/ml. ever, the inverse type I binding const.ant for 20~OH cholesterol was, in this case, unaffected by the second steroid (Fig. 4B) E$ect of Changes in $-Although there is a general decrease in the total inverse type I spectral response to pregnenolone between pH 6 and pH 8 (8), this encompasses appreciable changes in the relative magnitudes of the 20~OH cholesterol and the residual pregnenolone spectral responses (Table III). In bovine adrenal mit,ochondria, this change associated with an increase in the 20,22R type I difference spectrum with increased pH (19). In contrast to the case with bovine adrenal mitochondria, the magnitude of the 20,22R induced type I change in rat adrenal mitochondria, was pH-independent ( Diference Spectra in Phosphate-free Media-When the normal sucrose medium was used without phosphate ions the inverse type I responses produced by 20a-OH cholesterol and by a subsequent addition of pregnenolone were both decreased substantially (Table II). This was observed with adrenal mitochondria which were isolated from either stressed or cycloheximide-treated rats. On the other hand, the type I spectral response to 25-OH cholesterol was higher in the absence of phosphate. There was no effect of phosphate ions on the type I spectral response produced by addition of 20,22R after 25.OH cholesterol. After cycloheximide treatment of the rats the effect of removing phosphate from the medium was even sufficient to effect a type I response to 20a-OH cholesterol (Fig. 5). A type I response of rat adrenal mitochondria to 20cr-OH cholesterol has recently been reported by Bell et al. (25). Addition of pregnenolone either directly to these adrenal mitochondria, or after the addition of 20a-OH cholesterol produced similar inverse type I spectral responses (Fig. 5).
Experiments on Individual Animals-By exactly scaling down the isolation procedure, adrenal mitochondria were obtained from single rats. For the four animals in each group, the recovery of adrenal mitochondria was similar (3.4 to 4.3 mg/animal). Steroid spectral responses within each group differed by less than the experimental error, except for one animal within each group (marked A and B in Fig. 6) which gave high readings for all responses and may have had a higher specific content of cytochrome P-450. The effect of pretreatment of the rats on the direct inverse type I spectral response to pregnenolone was again considerably greater than the opposing change in 25

DISCUSSION
Cytochrome P-450,,, provides a likely regulation site for steroidogenesis in the adrenal cortex because it is the terminal oxidase for the sequence of oxygen-dependent steps in the ratelimiting conversion of cholesterol to pregnenolone (I, 13). The complexity of cholesterol binding to cytochrome P-450So which has been uncovered by studies of steroid binding (8,19) may be a key part of the mechanism by which ACTH regulates this enzyme system.
Certain steroids reversibly bind to adrenal mitochondrial cytochrome P-450 in the oxidized state so as to effect a spin state change, either from a low spin heme to a high spin heme (type I change) or in the reverse direction (inverse type I change). We have assumed that, for the associated spectral changes, AA (390 to 420 nm) depends only on the net number of cytochrome P-450 hemes which change spin state, i.e. apart from the direction of the change (type I or inverse type I) the extinction change is independent of the steroid and the form of cytochrome P-450. These adrenal mitochondria were sufficiently depleted of reducing equivalents to prevent significant metabolism of steroids which were bound to cytochrome P-450.
Type I spectral responses are induced by 25-OH cholesterol (8,24) and by 20,22R (19,21,23) which are both substrates of the side chain cleavage system (21). Thus, these steroids bind to low spin forms of cytochrome P-450,,, with the subsequent formation of a high spin form which is typical of a cytochrome P-450.substrate complex (26). The smallness of the type I response to 25-OH cholesterol was derived neither from an inhibition of the type I binding at higher concentrations of steroid nor from a mixture of type I and inverse type I spectral changes to 25.OH cholesterol. We consider that these distinct spectral responses derive from two different steroid binding sites which are associated with low spin forms of cytochrome P-450scc; one site binds both 25.OH cholesterol and 20,22R (slow, small type I response), whereas the second site binds only 20,22R (rapid, large type I response). The former spectral response is reversed by steroids such as pregnenolone and 20cz-OH cholesterol which elicit an inverse type I response but the interactions are not competitive (Fig. 4B). On the other hand, 20,22R and inverse type I steroids clearly compete for the second site (Fig. 2C).
The extent to which cytochrome P-45O,o is bound by endogenous substrates in adrenal mitochondria has been estimated from the inverse type I spectral change which is induced by pregnenolone (6,8). The observation that the inverse type I response to 20a-OH cholesterol was always less than that of pregnenolone probably in part derives from a type I contribution to the ~OCX-OH cholesterol response. Thus, 20a-OH cholesterol but not pregnenolone can form a partially high spin complex with cytochrome P-450,,, which is predominantly low spin as a result of low cholesterol levels (25,27). We observed a type 1 response to 20cr-OH cholesterol with rat adrenal mitochondria, but only when the action of ACTH in viva was inhibited by cycloheximide and when the absence of phosphate ions in the medium had depressed the combination of cholesterol with cytochrome P-450,,,. Thus, the observed response to 20cu-OH cholesterol upon interaction with adrenal mitochondrial cytochrome P-450,,, contains a contribution from an inverse type I change with cholesterol-cytochrome P-450 complexes and an opposite contribution from a type I change with the "free" cytochrome. The residual response to pregnenolone after saturation of cytochrome P-450,,, with 2Ocr-OH cholesterol may be explained by the interaction of pregnenolone with high spin 20a-OH cholesterol-cytochrome P-450,,, complexes. The absence of an effect of 20a-OH cholesterol on the affinity of pregnenolone for cytochrome P-450scc in the residual response may be compared to the lack of competition between 25-OH cholesterol and pregnenolone. However, this may not be the full explanation of differences between pregnenolone and 20ar-OH cholesterol inverse type I responses. In Fig. 6, the residual pregnenolone response is insufficiently large to account for the difference between the inverse type I response to pregnenolone and the type I response to ~OCX-OH cholesterol. A possible explanation is that ZOa-OH cholesterol does not induce spectral changes in all of the cholesterol cytochrome P-45O,o complexes in rat adrenal mitochondria which respond to pregnenolone.
At this point it is pertinent to ask whether the various steroidinduced spectral changes account for all of the adrenal mito-  (19). Spectra were carried out in sucrose buffer (see Fig. 1) at 25" with mitochondria from the adrenals of stressed rats. chondrial cytochrome P-450. Table IV shows these values for adrenal mitochondria from stressed rats. The sum of the steroidinduced type I and inverse type I responses (AA (390 to 420 nm)) is 40% in excess of AA (450 to 390 nm) for the reduced CO difference spectrum. On the basis of the extinction coefficients for equivalent changes in purified cytochrame I'-4500AM of Pseudomonas pu2id~,~ these steroid-induced changes account for slightly more cytochrome P-450 than could be detected with the CO complex of the reduced cytochrome. Small negative contributions from CO complexes of hemoglobin and cytochrome oxidase may cause a small decrease in the latter (28).
High and low spin cytochrome P-450,,, and low spin cytochrome P-45O11b (deoxycorticosterone response) thus appear to account for virtually all the rat adrenal mitochondria cytochrome P-450. EPR spectra indicate that cytochrome P-450,,,5 has less than 10% of the high spin state in adrenal mitochondria from stressed rats5 These rat adrenal mitochondria carry out 1%hydroxylation of deoxycorticosterone at about one-half the rate of 1 lp-hydroxylation (29). Since steroid 18.hydroxylation also requires cytochrome P-450 (30), the type I response to deoxycorticosterone probably includes a contribution from the cytochrome which is involved in l%hydroxylation whether identical with or distinct from cytochrome P-45O11,3.
The EPR experiments of Cheng and Harding (27) indicate that, when 20a-OH cholesterol is added to bovine adrenal cytochrome P-450scc which is depleted of cholesterol, the proportion of the low spin state decreases 10 to 15%. It seems reasonable to assume that there was a concomitant 10 to 15% increase in the high spin state. 13~ contrast, EPR measurements in this laboratory (19) show that 20cu-OH cholesterol can produce at least 95% decrease in high spin state in preparations from bovine adrenals which contain sufficient cholesterol to saturate the cytochrome. These two experiments cannot be reconciled if only a single steroid binding site is considered for cytochrome P-450,,,.
These observations and many of the experiments presented in this paper can be explained by the existence of at least two steroid binding sites on cytochrome P-450,,,. Site I binds cholesterol and accounts for the slow type I responses to 25-OH cholesterol and 20,22R. Binding at this site induces a change in the heme interactions so that a partial formation of a high spin state ensues. The proportion of high spin state induced by cholesterol is higher at lower pH (31,32). Steroids which induce inverse type I spectral changes bind to site II and stabilize a low spin configuration of theheme. It is proposed that binding to site I exerts little influence on binding to site II so that both sites may be simultaneously occupied. The effect of 20a-OH cholesterol on the cytochrome can be explained if the spin state of the bisteroid complex is dominated by the site II interactions and is predominantly low spin (Fig. 7A). Changes in the spin state during a titration with 20o(-OH cholesterol which have been calculated from this model are shown in Fig. 7B. The 3 to 4y0 of high spin cytochrome P-450,,,6 which was observed on addition of 2Ocu-OH cholesterol to rat adrenal mitochondlia under conditions of depleted cholesterol can be fitted to this model if K (site I) is about equal to K (site II) (Fig. 7B). A higher proportion of high spin cytochrome P-450,,, (15%) may be induced in cholesterol-depleted beef adrenal mitochondria because, in this case, binding to site I elicits a larger change to high spin.
The rapid binding of 20,22R which is competitive with the interaction of inverse type I steroids is assigned in the model either to site II or to a site whose conformation is allosterically inhibited by binding at site II. In binding at this site 20,22R further increases the proportion of high spin cytochrome P-450so. When both sites are occupied the magnitude of the steroid-induced difference spectra in relation to the reduced CO spectrum indicates that cytochrome P-450 see is almost entirely in a high spin state. According to data in Table IV binding to site I can contribute to about 40% of high spin cytochrome (inverse type I + 25.OH cholesterol-induced type I) whereas binding to site-II can contribute 60% of high spin cytochrome. When the proportion of cholesterol-cytochrome P-450,ce complexes is low in isolated mitochondria because of a lack or inhibition of ACTH activation, the type I response to 25.OH cholesterol is frequently not elevated to a corresponding extent. The 25.OH cholesterolcytochrome complex may be less in a high spin state than the 6 We have recently observed that 20a-HO-cholesterol used by us and that obtained from Ikapharm, Israel differ in their spectral responses with rat adrenal mitochondria even though the samples were not distinguishable by thin layer chromatography. Using the Ikapharm sample we obtained substantial type I spectral changes in agreement with Bell et al. (25) who used these snpplies. A possible explanation is that the Ikapharm material may contain a small orooortion of 20B-HO-cholesterol which has been reported to prod;ce-type I spectral changes with cytochrome P-250 in adrenal mitochondrial preparations (22).  7. A, a model for the binding of 20ol-hydrosycholesterol to cytochrome P-450,,, an d for accompanying spin state changes. The diagram represents the interaction of the steroid(s) at site I inducing a partial change to a high spin heme, and at site II when a low spin heme is retained. When both sites are occupied (bisteroid complex), a low spin heme is postulated. B, the dependence of the spin state on the concentration of 20o-hydroxycholesterol. Computations were made by applying the following assumptions to the above model: there is no competition between site I and II, i.e. K'I = KI, K'II = KII, binding to site I produces a complex with 407, of high spin heme (see text). The concentration of steroid is represented in units of KII. For the protein concentration which we have used KIT is approximately 1 FM. cholesterol complex. Changes in the extent to which site I ligands induce the high spin state could account for the restrained state which we have previously reported for cytochrome P-450,,, in rat adrenal mitochondria.
This model accounts for two important features of these experiments. Changes in the extent of cholesterol binding to the cytochrome caused by various pretreatments (ACTH, cyclohesimide, etc.) or the binding of 25.OH cholesterol, did not significantly affect the affinity of cytochrome P-450,,, for inverse type I steroids. The action of ACTH in viva or Ca*+-induced changes in mitochondria in titro (29, 33) can only enhance the binding of cholesterol to cytochrome P-45O,,o to a point where less than one-half of the cytochrome is in a high spin stat.e whereas the rapid response to 20,22R is unaffected by these changes. In these cases cholesterol binds only to site I.
The changes in the spectroscopic properties of adrenal mitochondrial cytochrome P-450,,, have been explained by supposing that there is a regulation of the capacity of extra-and intramitochondrial cholesterol to combine with the cytochrome (6,8). This supported and extended Garren's theory (12) that adrenal cholesterol is mobilized by a labile, ACTH-induced protein. Thus, the reduced level of this protein in cycloheximide-treated or quiescent rats restricts the distribution of cholesterol within adrenal mitochondria although not significantly changing the total mitochondrial cholesterol. This restriction may be correlated with the failure of cholesterol to combine with site I on cytochrome P-450,,, in isolated mitochondria when ACTH is at a low level or its action is inhibited.