154
Biochimica et Biophysica Acta, 1033 (1990) 154-161
Elsevier BBAGEN 23253
High-affinity binding sites for oxygenated sterols in rat liver microsomes: possible identity with antiestrogen binding sites * Peter L.H. Hwang Department of Physiology, National University of Singapore, Singapore (Republic of Singapore)
(Received 11 August 1989)
Key words: Oxygenated sterol; Oxysterol binding protein; Antiestrogen; Antiestrogen binding site
Oxygenated derivatives of cholesterol are known to exhibit a number of biological activities including the inhibition of cholesterol biosynthesis and of cell proliferation, but their mechanism of action remains unclear. Previous studies have identified a cytosolic protein which binds 25-hydroxycholesterol, as well as several other oxysterols, with high affinity, possibly mediating some of their effects. We now report the existence of a high-affinity oxysterol binding site in rat liver microsomes which is distinct from the cytosolic binding protein. Among the oxygenated sterols examined, 5a-cholestan-3fl-oi-7-one (7-ketocholestanol) had the highest affinity for this microsomal binding site ( K d -- 2.7 nM). Using 7-keto[3H]cholestanol as the radioactive ligand, we found that binding of this oxysterol to the microsomai binding site was saturable and reversible and was displaceable by the following oxysterols in descending order of potency: 7-ketocholestanol > 6-ketocholestanol > 7fl-hydroxycholesterol = 7-ketocholesterol > cholesten-3fl,5a,6fl-trioi = 7ahydroxycholesterol > 4-cholesten-3-one. All other sterols studied, including, notably, 25-hydroxycholesterol, had little or no inhibitory effect on 7-keto[3H]cholestanol binding. Additional studies revealed that the microsomal oxysterol binding site was probably identical to the antiestrogen binding site described by other workers. First, saturation analysis and kinetic studies demonstrated that the antiestrogen tamoxifen competed directly with 7-keto[3Hlcholestanol for the same binding site. Second, the ability of different oxysterois and antiestrogens to inhibit 7-keto[3Hlcholestanoi binding to the microsomai binding site paralleled their ability to inhibit [3H]tamoxifen binding to the antiestrogen binding site. Third, the tissue distribution of binding sites for 7-keto[3H]cholestanol was similar to that of the antiestrogen binding site. We conclude that: (1) in rat liver microsomes there are high-affinity oxysterol binding sites whose ligand specificity is different from that of the cytosolic oxysterol binding protein; and (2) the microsomal oxysterol binding site is probably identical to the antiestrogen binding site. The biological significance of these observations remains to be explored.
Introduction Oxygenated derivatives of cholesterol have long been known to exhibit a number of biological activities in-
* A preliminary report of some of the findings described here was presented at the 71st Annual Meeting of the Endocrine Society (U.S.A.), June 21-24, 1989, Seattle, Washington, U.S.A. Abbreviations: tamoxifen, 1-[4-(3-dimethylaminoethyoxy)phenyl]1,2-diphenylbut-l(Z)-ene; clomiphene, 1-[4-(2-diethylaminoethoxy) phenyl]-l,2-diphenyl-2-chloroethylene; nafoxidine, 1-{2-[4-(3,4-dihydro-6-methoxy-2-phenyl-l-naphthyl)-phenoxy]ethyl)pyrrolidine hydrochloride; C1 628, a-(4-pyrrolidinoethoxy)phenyl-4-methoxy-a'nitrostilbene; HMG-CoA reductase, 3-hydroxy-3-methylglutaryl coenzyme A reductase (EC 1.1.1.34). Correspondence: P.L.H. Hwang, Department of Physiology, National University of Singapore, 10 Kent Ridge Crescent, Singapore 0511, Republic of Singapore.
cluding, among others, the inhibition of cholesterol biosynthesis and of cell proliferation [1,2]. These biological effects, however, are generally observed in cultured cells and it remains unclear whether oxygenated sterols have any physiological role in vivo. Nevertheless, some of the in vitro effects of oxysterols have been extensively studied in recent years. (The term 'oxysterol' as used in this paper refers to any compound with the following characteristics: a cyclopentanoperhydrophenanthrene nucleus, a hydrocarbon side-chain at C17, a hydroxyl group at C3 and one or more additional oxygen functions attached to the nucleus or side-chain.) A number of oxysterols have been found to be potent inhibitors of cholesterol biosynthesis. Indeed, on a molar basis they are much more potent than cholesterol itself in inhibiting the activity of the enzyme 3-hydroxy-3methylglutaryl coenzyme A ( H M G - C o A ) reductase ((S)-Mevalonate : N A D P + oxidoreductase (CoA-acylating), EC 1.1.1.34) which is the rate-limiting enzyme in
0304-4165/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
155 cholesterol biosynthesis [3-6]. This observation has led to the proposal that oxygenated sterols, rather than cholesterol, might perhaps serve as the physiological regulators of cholesterol biosynthesis [2]. Although this hypothesis needs substantiation, an observation which may provide some support for it is the demonstration by Kandutsch and colleagues [7,8], as well as by other investigators [9-11], that a number of cell types and tissues contain an intracellular protein which binds oxygenated sterols, but not cholesterol, with high affinity. Taylor et al. [12] further demonstrated that the binding affinities of a large number of oxysterols for this protein appeared to parallel their ability to suppress HMG-CoA reductase activity in cultured mouse fibroblasts, suggesting that this oxysterol binding protein might mediate the inhibitory action of oxysterols on cholesterol biosynthesis. This oxysterol binding protein has been partly characterized in several laboratories [9,10,13,14]. There appears to be general agreement with respect to some of its properties although its precise physiological role remains to be defined. It binds 25-hydroxycholesterol and some other oxysterols with high affinity but does not bind cholesterol or the steroid hormones. The equilibrium dissociation constant (Kd) of thiS binding site for 25-hydroxycholesterol has generally been reported to be about 10 - 9 M, except in Chinese hamster ovary cells where the K a was found to be 1.4 x 10 -7 M [15]. With 25-hydrocholesterol bound to it, this binding protein has a molecular mass of 160 or 169 kDa, but it appears to be considerably larger (236 kDa) in the unliganded state and considerably smaller (97-98 kDa) when exposed to acid or urea, findings suggestive of the existence of subunits [14]. The oxysterol binding protein has very recently been purified to apparent homogeneity by Dawson et al. [16]. These investigators reported that this binding protein contained a doublet of peptides with molecular weights of 101 000 and 96000 as determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis but migrated as a complex with an apparent molecular weight of 280 000 on gel filtration. Its intracellular location remains to be systematically examined, but in all the studies reported to date it has been found in the cytosolic fraction of the cells and tissue studied; its possible existence in other subcellular fractions requires further study. Murphy et al. [17] first demonstrated that two oxygenated derivatives of cholesterol, 7-ketocholesterol and 4-cholesten-3-one, inhibited the binding of [3 H]tamoxifen to antiestrogen binding sites in cockerel liver, suggesting that oxysterols might bind to the antiestrogen binding sites which were first described by Sutherland and colleagues [18]. However, the binding affinity of 7-ketocholesterol and 4-cholesten-3-one for the antiestrogen binding site appeared to be very low, being less than 0.5 and 0.01%, respectively, of that of
tamoxifen. Our laboratory [19] subsequently examined a large series of oxygenated sterols for their ability to inhibit [3H]tamoxifen binding to rat liver microsomes. These studies showed that a number of oxygenated sterols were considerably more active than 7-ketocholesterol and 4-cholesten-3-one in inhibiting [3H]tamoxifen binding. The most potent of these was 5acholestan-3fl-ol-7-one or 7-ketocholestanol which was estimated to be some 70- and 3000-times more effective on a molar basis than 7-ketocholesterol and 4-cholesten3-one, respectively, in inhibiting [3H]tamoxifen binding. These and other observations [19] suggest that rat liver microsomes may also contain high-affinity binding sites for oxysterols. The evidence, however, remains indirect. Using 7-keto[3H]cholestanol as the radioactive ligand and an assay procedure developed in our laboratory, we now demonstrate directly that, indeed, this oxysterol, as well as a number of other oxysterols, binds to the microsomal fraction of rat liver with high affinity. We also show that this microsomal oxysterol binding site is clearly different from the cytosolic oxysterol binding protein reported by other workers and that it is probably identical to the antiestrogen binding site. Materials and Methods
Chemicals [N-methyl-3H]Tamoxifen (spec. act. 77 Ci or 2.85 TBq/mmol) was obtained from Amersham International and stored in ethanol at - 2 0 °C protected from light. Cholest[5,6- 3H]an-3/3-ol-7-one (7-keto[3 H]cholestanol) was custom prepared by Amersham International (spec. act. 15 C i / m m o l or 0.56 TBq/mmol). Before use it was purified by thin-layer chromatography on silica gel (AL SIL G; Whatman) using sequentially the following three solvent systems: hexane/diethylether/ acetic acid (70 : 30 : 1, v/v), chloroform/methanol (19:1, v/v) and chloroform. [1,2(n)-3H]Cholesterol (spec. act. 58 C i / m m o l or 2.1 TBq/mmol) and 25-hydroxy[26,27-3H]cholesterol (spec. act. 87 C i / m m o l or 3.2 TBq/mmol) were from New England Nuclear. Nafoxidine hydrochloride and CI 628 were graciously provided by The Upjohn Company. Tamoxifen citrate, clomiphene citrate, 4-cholesten-3-one, 25-hydroxycholesterol, 20c~-hydroxycholesterol and steroid hormones were obtained from Sigma Chemical Co. All other sterols were purchased from Steraloids. Other chemicals were of analytical grade and obtained from conventional sources.
Tissue preparation For the characterization of oxysterol binding sites a 20000 X g supernatant prepared from the livers of female Sprague-Dawley rats was used. Livers removed from 250-300 g rats were homogenized in 12 vol. of ice-cold buffer (10 mM Tris-HCl, 1.5 mM EDTA, 10%
156 ( v / v ) glycerol, p H 7.5 at 4 ° C ) with an Ultra-Turrax homogenizer. The homogenate was centrifuged at 20 000 x g for 30 min and the supernatant was stored at - 70 o C. Initially we further centrifuged the 20 000 × g supernatant at 100000 × g for 60 min and carried out binding studies on the precipitated microsomal fraction. It was subsequently observed that almost all the binding sites in the 20 000 × g supernatant were recovered in the 100 000 x g precipitate with very little binding activity remaining in the cytosol. Subsequent quantitation of microsomal binding sites was accordingly carried out in the 20 000 x g supernatant. A number of other organs were processed by an identical procedure for the quantitation of oxysterol binding sites. The liver 20000 x g supernatant used in ligand binding studies was diluted 3- to 5-fold with homogenizing buffer before analysis (final protein concentration, 1.5-3 m g / m l ) while the supernatants from the other organs were used without further dilution.
ligand. Separation of bound from free ligand was achieved by the procedure outlined in Table I. Specifically bound ligand was the difference between binding obtained in the absence and presence of a 250-fold molar excess of unlabeled ligand. Other details of the procedure were described previously [22].
Ligand binding studies
Other procedures
The procedure which our laboratory has developed for demonstrating specific binding of 7-keto[ 3H]cholestanol to rat liver microsomes is outlined in Table I. In assays using [3H]tamoxifen as the radioactive ligand, separation of bound from free ligand was carried out either by the addition of activated charcoal as detailed previously [21] or following the procedure given in Table I. Both procedures yielded similar results when [3H]tamoxifen was the radioactive ligand, but the charcoal technique could not be used in assays with 7keto[3H]cholestanol (see below). Saturation analysis with 7-keto[3H]cholestanol or [3H]tamoxifen was carried out by incubating the 20000 x g supernatant with increasing amounts of radioactive
Protein was determined by the method of Lowry et al. [23] using bovine serum albumin as the standard. Other procedures were carried out as described in the appropriate Figure legends and Tables.
TABLE I
Procedure for demonstrating specific binding of 7-keto[ 3H] cholestanol to rat liver 20000 x g supernatant 1. The 20000 x g rat liver supernatant (0.5 ml, protein concentration 1.5-3 m g / m l ) is incubated with 7-keto[3H]cholestanol (5000070000 c p m in 20 btl ethanol) in the absence or presence of competing ligands (dissolved in 25/~1 of ethanol). 2. Following incubation at 4 ° C for 16 h, 10 ~1 of 0.5 M calcium chloride in water is added. The assay tubes are mixed and allowed to stand at 4 ° C for 10-15 rain. 3. The precipitate formed is sedimented by centrifugation at 1500 x g for 20 min. The supernatant is discarded. 4. The precipitate is resuspended in 2 ml of 25 m M Tris-HCl (pH 7.4) containing 0.4% ( w / v ) Tween 80, 0.1% ( w / v ) bovine serum albumin and 10 m M calcium chloride. The assay tubes are then centrifuged again at 1 5 0 0 x g for 20 min. The supernatants are discarded. 5. Step 4 is repeated. 6. The washed precipitate is transferred to scintillation vials and counted as previously described [21].
Dissociation kinetics For the dissociation studies, the 20 000 x g rat liver supernatant was incubated for 16 h at 4 ° C with approx. 10 nM 7-keto[3H]cholestanol in the presence of 0, 0.625, 1.25 or 2.5 nM of tamoxifen. At zero time a 250-fold molar excess of non-radioactive 7-ketocholestanol was added and the assay was terminated at different time intervals. Total and specific binding were determined for each time point as described above. The dissociation rate constant was determined from the plot of the specifically bound 7-keto[3H]cholestanol against time as described previously [22].
Results
The procedure outlined in Table I was found, in our hands, to be a simple and reproducible method for demonstrating specific binding of 7-keto[ 3H]cholestanol to microsomal fractions of different tissues. We initially attempted to demonstrate 'specific' binding of 7keto[SH]cholestanol using activated charcoal to separate bound from free ligand. This was unsuccessful in that we could not show displacement by excess unlabeled 7-ketocholestanol. Indeed, using the charcoal tecnique, about 80% of the added radioactivity appeared in the bound fraction and the binding was not saturable. We subsequently observed that displaceable binding could be demonstrated if we washed the microsomes on cellulose nitrate filters (0.2 ~m Whatman) with detergent. Part of the radioactivity could be easily washed off while a substantial fraction remained, presumably because of binding to sites with higher affinity. The radioactivity which resisted stripping by detergent displayed characteristics of specific binding in that it was displaceable by excess unlabeled 7-ketocholestanol and exhibited saturability. The detergent wash step when carried out on paper filters was laborious and made it difficult to process a large number of samples simultaneously. Accordingly, an attempt was made to precipitate the microsomes by the addition of calcium chloride [20]. This permitted the sedimentation of the microsomal binding sites by means of low-speed centrifugation (1500 X g for 20 min), mak-
157 ing it possible to carry out detergent washes on many samples simultaneously. The addition of calcium chloride at the end of incubation did not appear to interfere significantly with 7-keto[3H]cholestanol binding, since the results were comparable to or better than those obtained when paper filters were used for the detergent wash step. After examining several detergents for their effectiveness in revealing specifically bound counts, it was observed that two 2-ml washes with 0.4% (w/v) Tween 80 gave consistently reproducible results with low background; accordingly, the procedure given in Table I was followed in all our subsequent studies. Using this procedure, about 25-40% of the added radioactivity could be shown to be specifically bound. The ability of a number of sterols and nonsteroidal antiestrogens to inhibit 7-keto[3H]cholestanol binding to rat liver microsomes was studied (Fig. 1). Binding of 7-keto[3H]cholestanol was inhibited not only by unlabeled 7-ketocholestanol in a dose-dependent manner, but also by 6-ketocholestanol, 7-ketocholesterol and 4-cholesten-3-one. Cholesterol, 20a-hydroxycholesterol and, notably, 25-hydroxycholesterol had little or no inhibitory effect on 7-keto[3H]cholestanol binding to the microsomal binding site. By contrast, tamoxifen was considerably more potent than 7-ketocholestanol in inhibiting 7-keto[3H]cholestanol binding; the molar concentration of tamoxifen needed to reduce specific binding by 50% varied between 10 and 20% of that of 7-ketocholestanol in different experiments, suggesting that the affinity of tamoxifen for this site might be substantially higher than that of 7-ketocholestanol, although other interpretations are possible. Other nonsteroidal antiestrogens tested (clomiphene, C1 628 and
E
100 E o
75 g 5O
o
25
u
~, Y ~--
lO-1
t00
iOi
tO2
t03
t04
105
Concentration of competing ligand (nM)
Fig. I. Inhibition of 7-keto[ 3H]cholestanol binding to rat liver binding
sites by various compounds. Binding assays were set up as described in the Materials and Methods section in the presence of increasing concentrations of non-radioactive 7-ketocholestanol (©), 6-ketocholestanol (O), 7-ketocholesterol (O), 4-cholesten-3-one (a), cholesterol (D), 20ct-hydroxycholesterol (11), 25-hydroxycholesterol (zx) and tamoxifen (O). Results are expressed as percentage of maximum binding.
2
so
o
0
.
1
.
2
.
.
3
.
4
5
logioICso wiLh 7-keto[3H]cho]estano: as the radioactive ligand
Fig. 2. The lCh0 values of oxysterols and antiestrogens with either 7-keto[ 3H]cholestanol or [3 H]tamoxifen as radioactive ligand. Binding assays were set up with either radioactive ligand in the presence of varying concentrations of the following compounds: tamoxifen (1); nafoxidine (2); clomiphene (3); C1 628 (4); 7-ketocholestanol (5); 6-ketocholestanol (6); 7-ketocholesterol (7); 7fl-hydroxycholesterol (8); cholestan-3fl,ha,6fl-triol (9); 7a-hydroxycholesterol (10); and 4cholesten-3-one (11). The concentrations of these compounds (in nM) required to reduce specific binding of the radioactive ligand by 50% (IC50) were then read from the inhibition curves. The IC50 values obtained with 7-keto[3H]cholestanol as the radiolabeled ligand were plotted against the corresponding IC50 values with [3H]tamoxifen as the radioactive ligand. Note that logarithmic scales were used for both axes in order to accommodate the wide range of IC50 values obtained.
nafoxidine) were also considerably more effective than 7-ketocholestanol in inhibiting binding and their displacement curves (not shown) were similar to that of tamoxifen. Other oxysterols which inhibited 7-keto[3H] cholestanol binding (not shown in Fig. 1) included 7c~-hydroxycholesterol, 7fl-hydroxycholesterol and cholestan-3fl,5a,6fl-triol, whereas the following compounds had minimal or no inhibitory effect at a concentration of 50/~M: 5a-cholestane, 5a-cholestan-3ct-ol, 5a-cholestan-3fl-ol, 5a-cholestan-3-one, 5c~-cholestan3fl,6a-diol and 5a-cholestan-3fl,6fl-diol. Among the oxysterols which inhibited 7-keto[3H]cholestanol binding, the order of inhibitory potency was as follows: 7-ketocholestanol > 6-ketocholestanol > 7fl-hydroxycholesterol = 7-ketocholesterol > cholestan-3fl,5a,6fl-triol = 7a-hydroxycholesterol > 4-cholesten-3-one. None of the steroid hormones examined, including cortisol, testosterone, estradiol, corticosterone, deoxycorticosterone, dehydroepiandrosterone and 4-androsten-3,17-dione inhibited 7-keto[3H]cholestanol binding at concentrations up to 50/tM. When [3H]tamoxifen was used as the radioactive ligand instead of 7-keto[3H]cholestanol, essentially the same order of relative inhibitory activity was obtained for the compounds tested. This observation is presented graphically in Fig. 2 which plots the logarithms of the IC50 (concentration of ligand in nM required to reduce
158 the specific binding of radioactive ligand by 50%) for 11 different compounds when [3H]tamoxifen was used as the radioactive ligand against the IC50 values of the same compounds with 7-keto[3H]cholestanol as the labeled ligand. The relationship is a linear one, indicating that the binding sites for [3H]tamoxifen and 7keto[3H]cholestanol exhibit very similar binding affinities for these compounds, a finding consistent with the possibility that the binding site for [3H]tamoxifen may be the same as that for 7-keto[3H]cholestanol. Fig. 3 shows the Scatchard analysis of 7-keto [3H]cholestanol and [3H]tamoxifen binding to the rat liver microsomal binding site. The plot reveals a single class of binding sites for 7-keto[3H]cholestanol with an apparent equilibrium dissociation constant (Kd) of 2.7 nM. The number of binding sites amounted to 790 f m o l / m g protein. A similar Scatchard analysis using the same rat liver microsomal preparation, but with [3H]tamoxifen as the radioactive ligand, revealed a K d of 2.0 nM and a binding site concentration of 839 f m o l / m g protein. It would appear that the affinity of tamoxifen for the binding site was only slightly higher than that of 7-ketocholestanol, whereas the binding site concentrations for the two ligands were very similar. Fig. 4 shows the effect of various concentrations of tamoxifen on 7-keto[3H]cholestanol binding to the microsomal binding site. The presence of tamoxifen did not significantly affect the number of binding sites (horizontal intercept), but clearly decreased the binding affinity (slope) in a dose-dependent manner. In the presence of 0, 1, 4, 10, 16 and 20 nM tamoxifen, the K d increased from 4.4 nM to 4.9, 6.8, 12.7, 17.2 and 20.4 nM, respectively. If tamoxifen competes with 7keto[3H]cholestanol for the same binding site, then the change in the apparent dissociation constant should fit the expression K~ = K a ( l +
I/Ki)
1.0
oo . • o g
0.5-
o.o o.o
o.5 t .o 1.5 2.o 3H-ligand bound (nN}
Fig. 3. Scatchard analysis of 7-keto[3HJcholestanol (©) and [3H]tamoxifen (O) binding to rat liver microsomal binding sites. Saturation analysis was carried out in duplicate as described in the Materials and Methods section. The straight lines were obtained by linear regressionanalysis.
2:V
1.4 12
1
0
0
~ 15
~
~1
0
5
I0
15
20
go o.4 0.2 0.0
2
4
Bound 7-ket0 [3H]ch01estan01 InN) Fig. 4. Scatchard analysis of 7-keto[ 3H]cho]estanol binding to rat liver binding sites in the presence of varying concentrations of tamo×ifen. Saturation analysis was set up in duplicate as described in the Materials and Methods section in the presence of 0 (©), I (e), 4 (z~), 10 (A), 16 (O) and 20 (e) nM tamoxifen. Inset: the prevailing
tamoxifen concentration was plotted against the apparent equilibrium dissociation constant K~ (nM). All lines were obtained by linear regression analysis.
where K d is the dissociation constant for 7keto[3H]cholestanol in the absence of tamoxifen, K d is the apparent dissociation constant for 7-keto[3H]cho lestanol in the presence of tamoxifen, I is the concentration of tamoxifen and Ki is the dissociation constant for tamoxifen. This relationship predicts that, if tamoxifen competes with 7-keto[3H]cholestanol for the same binding site, Kd should vary linear with the prevailing tamoxifen concentration (I). This was found to be the case (Fig. 4, inset). Additional support for the possibility that tamoxifen competes directly with 7-keto[3H]cholestanol for the same binding site comes from kinetic studies in which the effect of various concentrations of tamoxifen on the dissociation rate of 7-keto[3H]cholestanol from the microsomal binding site was studied. Fig. 5 shows the dissociation of bound 7-keto[3H]cholestanol from the microsomal binding site as a function of time in the presence of varying concentrations of tamoxifen. The time required for 50% of specifically bound 7keto[3H]cholestanol to dissociate ( q / 2 ) in the absence of tamoxifen was 123.8 rain with a corresponding dissociation rate constant of 9.33 × 10 -5 s a. In the presence of 0.625, 1.25 and 2.5 nM tamoxifen, the tl/2 values were 121.3, 122.6 and 126.4 min, respectively. The corresponding dissociation rate constants (which equal In 2/q/2 ) were 9.52, 9.42, and 9.14× 10 5 s-l, respectively. It is clear that tamoxifen has no significant effect on the dissociation rate of bound 7-keto[3H]cholestanol, a finding to be expected if tamoxifen and 7-ketocholestanol compete for the same binding site. Table II compares the concentrations of 7keto[3H]cholestanol binding sites with those of [3H]tamoxifen binding sites in the 20000 × g super-
159
-0.2
We have also carried out limited studies with [3H]cholesterol and 25-hydroxy[3H]cholesterol as the radioactive ligand. Neither of these compounds displayed specific binding to the rat liver microsomal fraction whether activated charcoal or the detergent wash procedure outlined in Table I was used to separate bound from free ligand.
-0.3
Discussion
o.I o.o x~
g 2
-0.1
u
2
~I l l -0.4 t~
~
q
-0.~
i
i
30
60
i
i
i
90 120 t50
Time {rain)
Fig. 5. dissociation of bound 7-keto[3H]cholestanol from rat liver binding sites. The procedure for the dissociation kinetics experiment was described in the Materials and Methods section. The amount of specifically bound 7-keto[3H]cholestanol in the absence (e), or presence of 0.625 (4), 1.25 ( , ) or 2.5 nM (11) tamoxifen is plotted against time. The results represent the mean of duplicate determinations. The straight lines and the half-times of dissociation used for calculating the dissociation rate constants were obtained by linear regression analysis.
natants of ten different organs of the rat. The concentrations of binding sites for the two ligands in general parallel each other, the highest concentrations being observed in the liver and the lowest in skeletal and cardiac muscle whether [3H]tamoxifen or 7-keto[3H] cholestanol was used as the radioactive ligand. This observation is again consistent with the possibility that the two binding sites are the same.
TABLE II Concentration of binding sites for [ 3H]tamoxifen and 7-keto[ ~H]choles tanol in the 20000 × g supernatant of different organs in the rat
The preparation of the 20000× g supernatants of different organs and the measurement of 7-keto{ 3H]cholestanol binding site concentrations were carried out as described in the Materials and Methods section. Scatchard analysis was employed to determine the binding site concentration in the liver. In other organs, estimates of binding site concentration were made by means of single saturating dose analysis. Results represent the mean of duplicate determinations and are expressed in f m o l / m g protein. Organ
[ 3H]Tamoxifen binding site concentration
7-Keto[ 3H]cholestanol binding site concentration
Liver Adrenal Ovaries Kidneys Brain Uterus Lung Spleen Muscle Heart
800 110 90 60 43 40 30 12 10 3
1060 110 145 65 58 54 27 10 11 4
The studies of Murphy et al. [17] first suggested that oxygenated derivatives of cholesterol might bind to cockerel liver antiestrogen binding sites, albeit with very low affinity. The experiments reported here clearly indicate that using the experimental procedure described in Table I high-affinity binding sites for oxysterols could be shown to exist in the microsomal fraction of rat liver. They also show that these microsomal binding sites are clearly distinguishable from the cytosolic oxysterol binding protein described by other workers and are probably identical to the antiestrogen binding sites. In the studies described here, we had found it impossible to show specific binding of 7-keto[3H]cholestanol by means of conventional techniques such as charcoal adsorption. The major problem, well recognized by investigators studying the binding of 25-hydroxy[3H] cholesterol to the cytosolic oxysterol binding protein [8], appeared to be the presence in tissue preparations of other binding sites which are not saturable. In the presence of these nonsaturable binding sites, it is difficult to show inhibition of radioactive ligand binding by excess unlabeled ligand, since any displaced radioactive ligand would be taken up promptly by the nonsaturable binding sites and would, therefore, appear to be 'bound'. The procedure outlined in Table 1 proved critical for revealing specific binding sites for 7keto[3H]cholestanol. The detergent washes appeared to remove selectively most of the radioactivity bound to the non-saturable sites, thus permitting quantitation of the radioactive ligand remaining bound to the saturable binding sites which were detergent-resistant presumably because they were of higher affinity. Activated charcoal, by contrast, was unable to strip the radioactive ligand off the non-saturable sites. Detergent treatment may have greater general applicability and should probably be explored further as a means of reducing interference by non-saturable binding sites in ligand binding studies. Interference by non-saturable binding sites in the rat liver preparation used in our binding studies also made it difficult to compare the binding affinities of oxysterols with those of non-steroidal antiestrogens. In the binding studies similar to those depicted in Fig. 1, for example, we consistently observed that tamoxifen was 5-10-times more potent than unlabeled 7-ketocholestanol in inhibiting 7-keto[3H]cholestanol binding. However, given the fact that 7-ketocholestanol, but not tamoxifen,
160 binds to a very significant extent to the non-saturable binding sites present in the 20000 x g tissue supernatant (unpublished observation), it could be argued that not all the 7-ketocholestanol added was available for competition for the high-affinity saturable binding site, leading to a spuriously low estimate of its binding affinity. Its true affinity for the binding site might in fact be closer to that of tamoxifen, as was suggested by the Scatchard analysis shown in Fig. 3. The studies reported here also provide evidence that the high-affinity microsomal binding site for 7keto[ 3H]cholestanol is also the antiestrogen binding site. First, the relative binding affinities of the non-steroidal antiestrogens and oxysterols examined were similar, irrespective of whether the radioactive ligand was 7-keto[3H]cholestanol or [3H]tamoxifen (Fig. 2). Second, saturation analysis (Fig. 4) and kinetic studies (Fig. 5) clearly suggested that 7-ketocholestanol and tamoxifen competed for the same binding site. Furthermore, the tissue distribution of antiestrogen binding sites was very similar to that of binding sites for 7-ketocholestanol (Table II). Taken together, these findings strongly suggest that the high-affinity binding site for 7-ketocholestanol is the same as the antiestrogen binding site. Definitive proof of this point must await the isolation and characterization of the binding protein. The identification of the antiestrogen binding site as a high-affinity binding site for oxygenated sterols raises a number of questions. The first concerns the relationship of this binding site to the cytosolic oxysterol binding protein described earlier by other workers [7-11,13,14,16]. The latter displays a number of features which clearly distinguish it from the antiestrogen binding site. First, it appears to be a cytosolic protein, whereas the antiestrogen binding site is localized predominantly to the microsomal fraction [24]. Second, it binds 25-hydroxycholesterol and 20a-hydroxycholesterol (as well as some other oxysterols) with high affinity [9,12-14.16], whereas the antiestrogen binding site appears to have little or no affinity for these two sterols (Fig. 1). We have also not been able to detect any specific binding of 25-hydroxy[3H]cholesterol in the rat liver microsomal fraction by either the procedure given in Table I or by charcoal adsorption (result not shown). It would appear, therefore, that the microsomal and the cytosolic oxysterol binding sites have very different ligand specificities. In this regard, it is of interest to note that the cytosolic binding protein has been reported by some investigators [9] to exhibit much higher affinities for oxysterols with oxygen functions on the side chain than for oxysterols which are oxygenated at C7. The microsomal binding site, on the other hand, appears to have diametrically opposite ligand specificities (Fig. 1). Indeed, one may speculate that different classes of oxysterols may have different intracellular binding sites, although whether this has any functional
correlate requires further study. In any case, it is clear that the microsomal oxysterol binding site and the cytosolic oxysterol binding protein are different not only in their intracellular distribution but also in their ligand specificities. The second question relates to the possible biological significance of the high-affinity binding of oxysterols by antiestrogen binding sites which are known to be ubiquitously distributed. They bind non-steroidal antiestrogens and many other pharmacological agents with high affinity (for a review see Ref. 25). Currently little is known about their physiological function. The demonstration that they bind oxysterols with high affinity raises the possibility that some of the known biological effects of oxysterols could be mediated by the antiestrogen binding sites, just as the cytosolic oxysterol binding protein is believed by some investigators to mediate the inhibition of cholesterol biosynthesis by oxysterols [13,14]. In this connection, it is intriguing to note that the non-steroidal antiestrogens have been reported to inhibit cholesterol biosynthesis [26,27] and cell proliferation [28] by mechanisms which are independent of the estrogen receptor. Both of these effects are shared by the oxysterols. The possibility that both the antiestrogens and oxysterols exert these effects through the antiestrogen binding site deserves further examination. Thirdly, although the nature of the endogenous ligand for antiestrogen binding sites remains elusive, the observation that some oxysterols bind to this site with high affinity raises the possibility that the endogenous ligand, if one exists, could be an oxysterol or a related compound. Currently there is no evidence that 7-ketocholestanol plays any regulatory role in vivo and the biological significance of its binding to the antiestrogen binding site remains to be explored. Nevertheless, the identification of oxysterols which bind to the antiestrogen binding site with high affinity should facilitate further studies on the possible physiological role of antiestrogen binding sites and the mechanisms underlying the biological effects of oxysterols. It would be of particular interest to determine, for example, whether the microsomal oxysterol binding site plays a role in the sterol-enhanced degradation of HMG-CoA reductase [29], an enzyme which is located in the microsomes and which regulates cholesterol biosynthesis.
Acknowledgements This study was supported by the National University of Singapore and the Shaw Foundation. The excellent technical assistance of Miss Chiew Yoke Eng and the superb secretarial help of Miss Asha Das are gratefully acknowledged. I thank Dr. O.L. Kon for helpful discussions.
161
References 1 Beck, J.P. and Crastes de Paulet, A. (eds.) (1988) Biological Activities of Oxygenated Sterols, Colloques Inserm Volume 166, Institut National de la Sant6 et de la Recherche M6dicale, Paris. 2 Kandutsch, A.A., Chen, H.W. and Heiniger, H.-J. (1978) Science 201,498-501. 3 Brown, M.S. and Goldstein, J.L. (1974) J. Biol. Chem. 249, 7306-7314. 4 Kandutsch, A.A. and Chen, H.W. (1973) J. Biol. Chem. 248, 8408-8417 5 Kandutsch, A.A. and Chen, H.W. (1974) J. Biol. Chem. 249, 6057-6061. 6 Chen, H.W., Kandutsch, A.A. and Waymouth, C. (1974) Nature 251,419-421. 7 Kandutsch, A.A., Chen, H.W. and Shown, E.P. (1977) Proc. Natl. Acad. Sci. USA 74, 2500-2506. 8 Kandutsch, A.A. and Thompson, E.B. (1980) J. Biol. Chem. 255, 10813-10826. 9 Beseme, F., Astruc, M.E., Defay, R. and Crastes de Paulet, A. (1987) FEBS Lett. 210, 97-103. 10 Beseme, F., Astruc, M.E., Defay, R., Descomps, B. and Crastes de Paulet, A. (1986) Biochim. Biophys. Acta 886, 96-108. 11 Sinensky, M. and Mueller, G. (1981) Arch. Biochem. Biophys. 209, 314-320. 12 Taylor, F.R., Saucier, S.E., Shown, E.P., Parish, E.J. and Kandutsch, A.A. (1984) J. Biol. Chem. 259, 12382-12387. 13 Taylor, F.R. and Kandutsch, A.A. (1987) in Sterol/Steroid Hormone Mechanism of Action (Kumar, R. and Spelsberg, T., eds.), pp. 395-407, Martinus Nijhoff, Boston.
14 Taylor, F.R. and Kandutsch, A.A. (1985) Chem. Phys. Lipids 38, 187-194. 15 Sinensky, M. and Mueller, G. (1981) Arch. Biochem. Biophys. 209, 314-320. 16 Dawson, P.A., Van der Westhuyzen, D.R., Goldstein, J.L. and Brown, M.S. (1989) J. Biol. Chem. 264, 9046-9052. 17 Murphy, P.R., Breckenridge, W.C. and Lazier, C.B. (1985) Biochem. Biophys. Res. Commun. 127, 786-792. 18 Sutherland, R.L., Murphy, L.C., Foo, M.S., Green, M.D., Whybourne, A.M. and Krozowski, Z.S. (1980) Nature 288, 273-275. 19 Hwang, P.L.H. and Matin, A. (1989) J. Lipid Res. 30, 239-245. 20 Kamat, V.B. and Wallach, D.F.H. (1965) Science 148, 1343-1345. 21 Hwang, P.L.H. (1986) Biochem. J. 237, 749-755. 22 Hwang, P.L.H. (1987) Biochem. J. 243, 359-364. 23 Lowry, O.H., Roseborough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275. 24 Sudo, K., Monsma, F.J. and Katzenellenbogen, B.S. (1983) Endocrinology 112, 425-434. 25 Watts, C.K.W. and Sutherland, R.L. (1987) Mol. Pharmacol. 31, 541-551. 26 Cypriani, B., Tabacik, C. and Descomps, B. (1988) Biochim. Biophys. Acta 972, 167-178. 27 Cypriani, B., Tabacik, C., Descomps, B. and Crastes de Paulet, A. (1988) J. Steroid Biochem. 31, 763-771. 28 Sutherland, R.L., Watts, C.K.W. and Ruenitz, P.C. (1986) Biochem. Biophys. Res. Commun. 140, 523-520. 29 Gil, G., Faust, J.R., Chin, D.J., Goldstein, J.L. and Brown, M.S. (1985) Cell 41, 249-258.
Biochimica et Biophysica Acta, 1033 (1990) 154-161
Elsevier BBAGEN 23253
High-affinity binding sites for oxygenated sterols in rat liver microsomes: possible identity with antiestrogen binding sites * Peter L.H. Hwang Department of Physiology, National University of Singapore, Singapore (Republic of Singapore)
(Received 11 August 1989)
Key words: Oxygenated sterol; Oxysterol binding protein; Antiestrogen; Antiestrogen binding site
Oxygenated derivatives of cholesterol are known to exhibit a number of biological activities including the inhibition of cholesterol biosynthesis and of cell proliferation, but their mechanism of action remains unclear. Previous studies have identified a cytosolic protein which binds 25-hydroxycholesterol, as well as several other oxysterols, with high affinity, possibly mediating some of their effects. We now report the existence of a high-affinity oxysterol binding site in rat liver microsomes which is distinct from the cytosolic binding protein. Among the oxygenated sterols examined, 5a-cholestan-3fl-oi-7-one (7-ketocholestanol) had the highest affinity for this microsomal binding site ( K d -- 2.7 nM). Using 7-keto[3H]cholestanol as the radioactive ligand, we found that binding of this oxysterol to the microsomai binding site was saturable and reversible and was displaceable by the following oxysterols in descending order of potency: 7-ketocholestanol > 6-ketocholestanol > 7fl-hydroxycholesterol = 7-ketocholesterol > cholesten-3fl,5a,6fl-trioi = 7ahydroxycholesterol > 4-cholesten-3-one. All other sterols studied, including, notably, 25-hydroxycholesterol, had little or no inhibitory effect on 7-keto[3H]cholestanol binding. Additional studies revealed that the microsomal oxysterol binding site was probably identical to the antiestrogen binding site described by other workers. First, saturation analysis and kinetic studies demonstrated that the antiestrogen tamoxifen competed directly with 7-keto[3Hlcholestanol for the same binding site. Second, the ability of different oxysterois and antiestrogens to inhibit 7-keto[3Hlcholestanoi binding to the microsomai binding site paralleled their ability to inhibit [3H]tamoxifen binding to the antiestrogen binding site. Third, the tissue distribution of binding sites for 7-keto[3H]cholestanol was similar to that of the antiestrogen binding site. We conclude that: (1) in rat liver microsomes there are high-affinity oxysterol binding sites whose ligand specificity is different from that of the cytosolic oxysterol binding protein; and (2) the microsomal oxysterol binding site is probably identical to the antiestrogen binding site. The biological significance of these observations remains to be explored.
Introduction Oxygenated derivatives of cholesterol have long been known to exhibit a number of biological activities in-
* A preliminary report of some of the findings described here was presented at the 71st Annual Meeting of the Endocrine Society (U.S.A.), June 21-24, 1989, Seattle, Washington, U.S.A. Abbreviations: tamoxifen, 1-[4-(3-dimethylaminoethyoxy)phenyl]1,2-diphenylbut-l(Z)-ene; clomiphene, 1-[4-(2-diethylaminoethoxy) phenyl]-l,2-diphenyl-2-chloroethylene; nafoxidine, 1-{2-[4-(3,4-dihydro-6-methoxy-2-phenyl-l-naphthyl)-phenoxy]ethyl)pyrrolidine hydrochloride; C1 628, a-(4-pyrrolidinoethoxy)phenyl-4-methoxy-a'nitrostilbene; HMG-CoA reductase, 3-hydroxy-3-methylglutaryl coenzyme A reductase (EC 1.1.1.34). Correspondence: P.L.H. Hwang, Department of Physiology, National University of Singapore, 10 Kent Ridge Crescent, Singapore 0511, Republic of Singapore.
cluding, among others, the inhibition of cholesterol biosynthesis and of cell proliferation [1,2]. These biological effects, however, are generally observed in cultured cells and it remains unclear whether oxygenated sterols have any physiological role in vivo. Nevertheless, some of the in vitro effects of oxysterols have been extensively studied in recent years. (The term 'oxysterol' as used in this paper refers to any compound with the following characteristics: a cyclopentanoperhydrophenanthrene nucleus, a hydrocarbon side-chain at C17, a hydroxyl group at C3 and one or more additional oxygen functions attached to the nucleus or side-chain.) A number of oxysterols have been found to be potent inhibitors of cholesterol biosynthesis. Indeed, on a molar basis they are much more potent than cholesterol itself in inhibiting the activity of the enzyme 3-hydroxy-3methylglutaryl coenzyme A ( H M G - C o A ) reductase ((S)-Mevalonate : N A D P + oxidoreductase (CoA-acylating), EC 1.1.1.34) which is the rate-limiting enzyme in
0304-4165/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
155 cholesterol biosynthesis [3-6]. This observation has led to the proposal that oxygenated sterols, rather than cholesterol, might perhaps serve as the physiological regulators of cholesterol biosynthesis [2]. Although this hypothesis needs substantiation, an observation which may provide some support for it is the demonstration by Kandutsch and colleagues [7,8], as well as by other investigators [9-11], that a number of cell types and tissues contain an intracellular protein which binds oxygenated sterols, but not cholesterol, with high affinity. Taylor et al. [12] further demonstrated that the binding affinities of a large number of oxysterols for this protein appeared to parallel their ability to suppress HMG-CoA reductase activity in cultured mouse fibroblasts, suggesting that this oxysterol binding protein might mediate the inhibitory action of oxysterols on cholesterol biosynthesis. This oxysterol binding protein has been partly characterized in several laboratories [9,10,13,14]. There appears to be general agreement with respect to some of its properties although its precise physiological role remains to be defined. It binds 25-hydroxycholesterol and some other oxysterols with high affinity but does not bind cholesterol or the steroid hormones. The equilibrium dissociation constant (Kd) of thiS binding site for 25-hydroxycholesterol has generally been reported to be about 10 - 9 M, except in Chinese hamster ovary cells where the K a was found to be 1.4 x 10 -7 M [15]. With 25-hydrocholesterol bound to it, this binding protein has a molecular mass of 160 or 169 kDa, but it appears to be considerably larger (236 kDa) in the unliganded state and considerably smaller (97-98 kDa) when exposed to acid or urea, findings suggestive of the existence of subunits [14]. The oxysterol binding protein has very recently been purified to apparent homogeneity by Dawson et al. [16]. These investigators reported that this binding protein contained a doublet of peptides with molecular weights of 101 000 and 96000 as determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis but migrated as a complex with an apparent molecular weight of 280 000 on gel filtration. Its intracellular location remains to be systematically examined, but in all the studies reported to date it has been found in the cytosolic fraction of the cells and tissue studied; its possible existence in other subcellular fractions requires further study. Murphy et al. [17] first demonstrated that two oxygenated derivatives of cholesterol, 7-ketocholesterol and 4-cholesten-3-one, inhibited the binding of [3 H]tamoxifen to antiestrogen binding sites in cockerel liver, suggesting that oxysterols might bind to the antiestrogen binding sites which were first described by Sutherland and colleagues [18]. However, the binding affinity of 7-ketocholesterol and 4-cholesten-3-one for the antiestrogen binding site appeared to be very low, being less than 0.5 and 0.01%, respectively, of that of
tamoxifen. Our laboratory [19] subsequently examined a large series of oxygenated sterols for their ability to inhibit [3H]tamoxifen binding to rat liver microsomes. These studies showed that a number of oxygenated sterols were considerably more active than 7-ketocholesterol and 4-cholesten-3-one in inhibiting [3H]tamoxifen binding. The most potent of these was 5acholestan-3fl-ol-7-one or 7-ketocholestanol which was estimated to be some 70- and 3000-times more effective on a molar basis than 7-ketocholesterol and 4-cholesten3-one, respectively, in inhibiting [3H]tamoxifen binding. These and other observations [19] suggest that rat liver microsomes may also contain high-affinity binding sites for oxysterols. The evidence, however, remains indirect. Using 7-keto[3H]cholestanol as the radioactive ligand and an assay procedure developed in our laboratory, we now demonstrate directly that, indeed, this oxysterol, as well as a number of other oxysterols, binds to the microsomal fraction of rat liver with high affinity. We also show that this microsomal oxysterol binding site is clearly different from the cytosolic oxysterol binding protein reported by other workers and that it is probably identical to the antiestrogen binding site. Materials and Methods
Chemicals [N-methyl-3H]Tamoxifen (spec. act. 77 Ci or 2.85 TBq/mmol) was obtained from Amersham International and stored in ethanol at - 2 0 °C protected from light. Cholest[5,6- 3H]an-3/3-ol-7-one (7-keto[3 H]cholestanol) was custom prepared by Amersham International (spec. act. 15 C i / m m o l or 0.56 TBq/mmol). Before use it was purified by thin-layer chromatography on silica gel (AL SIL G; Whatman) using sequentially the following three solvent systems: hexane/diethylether/ acetic acid (70 : 30 : 1, v/v), chloroform/methanol (19:1, v/v) and chloroform. [1,2(n)-3H]Cholesterol (spec. act. 58 C i / m m o l or 2.1 TBq/mmol) and 25-hydroxy[26,27-3H]cholesterol (spec. act. 87 C i / m m o l or 3.2 TBq/mmol) were from New England Nuclear. Nafoxidine hydrochloride and CI 628 were graciously provided by The Upjohn Company. Tamoxifen citrate, clomiphene citrate, 4-cholesten-3-one, 25-hydroxycholesterol, 20c~-hydroxycholesterol and steroid hormones were obtained from Sigma Chemical Co. All other sterols were purchased from Steraloids. Other chemicals were of analytical grade and obtained from conventional sources.
Tissue preparation For the characterization of oxysterol binding sites a 20000 X g supernatant prepared from the livers of female Sprague-Dawley rats was used. Livers removed from 250-300 g rats were homogenized in 12 vol. of ice-cold buffer (10 mM Tris-HCl, 1.5 mM EDTA, 10%
156 ( v / v ) glycerol, p H 7.5 at 4 ° C ) with an Ultra-Turrax homogenizer. The homogenate was centrifuged at 20 000 x g for 30 min and the supernatant was stored at - 70 o C. Initially we further centrifuged the 20 000 × g supernatant at 100000 × g for 60 min and carried out binding studies on the precipitated microsomal fraction. It was subsequently observed that almost all the binding sites in the 20 000 × g supernatant were recovered in the 100 000 x g precipitate with very little binding activity remaining in the cytosol. Subsequent quantitation of microsomal binding sites was accordingly carried out in the 20 000 x g supernatant. A number of other organs were processed by an identical procedure for the quantitation of oxysterol binding sites. The liver 20000 x g supernatant used in ligand binding studies was diluted 3- to 5-fold with homogenizing buffer before analysis (final protein concentration, 1.5-3 m g / m l ) while the supernatants from the other organs were used without further dilution.
ligand. Separation of bound from free ligand was achieved by the procedure outlined in Table I. Specifically bound ligand was the difference between binding obtained in the absence and presence of a 250-fold molar excess of unlabeled ligand. Other details of the procedure were described previously [22].
Ligand binding studies
Other procedures
The procedure which our laboratory has developed for demonstrating specific binding of 7-keto[ 3H]cholestanol to rat liver microsomes is outlined in Table I. In assays using [3H]tamoxifen as the radioactive ligand, separation of bound from free ligand was carried out either by the addition of activated charcoal as detailed previously [21] or following the procedure given in Table I. Both procedures yielded similar results when [3H]tamoxifen was the radioactive ligand, but the charcoal technique could not be used in assays with 7keto[3H]cholestanol (see below). Saturation analysis with 7-keto[3H]cholestanol or [3H]tamoxifen was carried out by incubating the 20000 x g supernatant with increasing amounts of radioactive
Protein was determined by the method of Lowry et al. [23] using bovine serum albumin as the standard. Other procedures were carried out as described in the appropriate Figure legends and Tables.
TABLE I
Procedure for demonstrating specific binding of 7-keto[ 3H] cholestanol to rat liver 20000 x g supernatant 1. The 20000 x g rat liver supernatant (0.5 ml, protein concentration 1.5-3 m g / m l ) is incubated with 7-keto[3H]cholestanol (5000070000 c p m in 20 btl ethanol) in the absence or presence of competing ligands (dissolved in 25/~1 of ethanol). 2. Following incubation at 4 ° C for 16 h, 10 ~1 of 0.5 M calcium chloride in water is added. The assay tubes are mixed and allowed to stand at 4 ° C for 10-15 rain. 3. The precipitate formed is sedimented by centrifugation at 1500 x g for 20 min. The supernatant is discarded. 4. The precipitate is resuspended in 2 ml of 25 m M Tris-HCl (pH 7.4) containing 0.4% ( w / v ) Tween 80, 0.1% ( w / v ) bovine serum albumin and 10 m M calcium chloride. The assay tubes are then centrifuged again at 1 5 0 0 x g for 20 min. The supernatants are discarded. 5. Step 4 is repeated. 6. The washed precipitate is transferred to scintillation vials and counted as previously described [21].
Dissociation kinetics For the dissociation studies, the 20 000 x g rat liver supernatant was incubated for 16 h at 4 ° C with approx. 10 nM 7-keto[3H]cholestanol in the presence of 0, 0.625, 1.25 or 2.5 nM of tamoxifen. At zero time a 250-fold molar excess of non-radioactive 7-ketocholestanol was added and the assay was terminated at different time intervals. Total and specific binding were determined for each time point as described above. The dissociation rate constant was determined from the plot of the specifically bound 7-keto[3H]cholestanol against time as described previously [22].
Results
The procedure outlined in Table I was found, in our hands, to be a simple and reproducible method for demonstrating specific binding of 7-keto[ 3H]cholestanol to microsomal fractions of different tissues. We initially attempted to demonstrate 'specific' binding of 7keto[SH]cholestanol using activated charcoal to separate bound from free ligand. This was unsuccessful in that we could not show displacement by excess unlabeled 7-ketocholestanol. Indeed, using the charcoal tecnique, about 80% of the added radioactivity appeared in the bound fraction and the binding was not saturable. We subsequently observed that displaceable binding could be demonstrated if we washed the microsomes on cellulose nitrate filters (0.2 ~m Whatman) with detergent. Part of the radioactivity could be easily washed off while a substantial fraction remained, presumably because of binding to sites with higher affinity. The radioactivity which resisted stripping by detergent displayed characteristics of specific binding in that it was displaceable by excess unlabeled 7-ketocholestanol and exhibited saturability. The detergent wash step when carried out on paper filters was laborious and made it difficult to process a large number of samples simultaneously. Accordingly, an attempt was made to precipitate the microsomes by the addition of calcium chloride [20]. This permitted the sedimentation of the microsomal binding sites by means of low-speed centrifugation (1500 X g for 20 min), mak-
157 ing it possible to carry out detergent washes on many samples simultaneously. The addition of calcium chloride at the end of incubation did not appear to interfere significantly with 7-keto[3H]cholestanol binding, since the results were comparable to or better than those obtained when paper filters were used for the detergent wash step. After examining several detergents for their effectiveness in revealing specifically bound counts, it was observed that two 2-ml washes with 0.4% (w/v) Tween 80 gave consistently reproducible results with low background; accordingly, the procedure given in Table I was followed in all our subsequent studies. Using this procedure, about 25-40% of the added radioactivity could be shown to be specifically bound. The ability of a number of sterols and nonsteroidal antiestrogens to inhibit 7-keto[3H]cholestanol binding to rat liver microsomes was studied (Fig. 1). Binding of 7-keto[3H]cholestanol was inhibited not only by unlabeled 7-ketocholestanol in a dose-dependent manner, but also by 6-ketocholestanol, 7-ketocholesterol and 4-cholesten-3-one. Cholesterol, 20a-hydroxycholesterol and, notably, 25-hydroxycholesterol had little or no inhibitory effect on 7-keto[3H]cholestanol binding to the microsomal binding site. By contrast, tamoxifen was considerably more potent than 7-ketocholestanol in inhibiting 7-keto[3H]cholestanol binding; the molar concentration of tamoxifen needed to reduce specific binding by 50% varied between 10 and 20% of that of 7-ketocholestanol in different experiments, suggesting that the affinity of tamoxifen for this site might be substantially higher than that of 7-ketocholestanol, although other interpretations are possible. Other nonsteroidal antiestrogens tested (clomiphene, C1 628 and
E
100 E o
75 g 5O
o
25
u
~, Y ~--
lO-1
t00
iOi
tO2
t03
t04
105
Concentration of competing ligand (nM)
Fig. I. Inhibition of 7-keto[ 3H]cholestanol binding to rat liver binding
sites by various compounds. Binding assays were set up as described in the Materials and Methods section in the presence of increasing concentrations of non-radioactive 7-ketocholestanol (©), 6-ketocholestanol (O), 7-ketocholesterol (O), 4-cholesten-3-one (a), cholesterol (D), 20ct-hydroxycholesterol (11), 25-hydroxycholesterol (zx) and tamoxifen (O). Results are expressed as percentage of maximum binding.
2
so
o
0
.
1
.
2
.
.
3
.
4
5
logioICso wiLh 7-keto[3H]cho]estano: as the radioactive ligand
Fig. 2. The lCh0 values of oxysterols and antiestrogens with either 7-keto[ 3H]cholestanol or [3 H]tamoxifen as radioactive ligand. Binding assays were set up with either radioactive ligand in the presence of varying concentrations of the following compounds: tamoxifen (1); nafoxidine (2); clomiphene (3); C1 628 (4); 7-ketocholestanol (5); 6-ketocholestanol (6); 7-ketocholesterol (7); 7fl-hydroxycholesterol (8); cholestan-3fl,ha,6fl-triol (9); 7a-hydroxycholesterol (10); and 4cholesten-3-one (11). The concentrations of these compounds (in nM) required to reduce specific binding of the radioactive ligand by 50% (IC50) were then read from the inhibition curves. The IC50 values obtained with 7-keto[3H]cholestanol as the radiolabeled ligand were plotted against the corresponding IC50 values with [3H]tamoxifen as the radioactive ligand. Note that logarithmic scales were used for both axes in order to accommodate the wide range of IC50 values obtained.
nafoxidine) were also considerably more effective than 7-ketocholestanol in inhibiting binding and their displacement curves (not shown) were similar to that of tamoxifen. Other oxysterols which inhibited 7-keto[3H] cholestanol binding (not shown in Fig. 1) included 7c~-hydroxycholesterol, 7fl-hydroxycholesterol and cholestan-3fl,5a,6fl-triol, whereas the following compounds had minimal or no inhibitory effect at a concentration of 50/~M: 5a-cholestane, 5a-cholestan-3ct-ol, 5a-cholestan-3fl-ol, 5a-cholestan-3-one, 5c~-cholestan3fl,6a-diol and 5a-cholestan-3fl,6fl-diol. Among the oxysterols which inhibited 7-keto[3H]cholestanol binding, the order of inhibitory potency was as follows: 7-ketocholestanol > 6-ketocholestanol > 7fl-hydroxycholesterol = 7-ketocholesterol > cholestan-3fl,5a,6fl-triol = 7a-hydroxycholesterol > 4-cholesten-3-one. None of the steroid hormones examined, including cortisol, testosterone, estradiol, corticosterone, deoxycorticosterone, dehydroepiandrosterone and 4-androsten-3,17-dione inhibited 7-keto[3H]cholestanol binding at concentrations up to 50/tM. When [3H]tamoxifen was used as the radioactive ligand instead of 7-keto[3H]cholestanol, essentially the same order of relative inhibitory activity was obtained for the compounds tested. This observation is presented graphically in Fig. 2 which plots the logarithms of the IC50 (concentration of ligand in nM required to reduce
158 the specific binding of radioactive ligand by 50%) for 11 different compounds when [3H]tamoxifen was used as the radioactive ligand against the IC50 values of the same compounds with 7-keto[3H]cholestanol as the labeled ligand. The relationship is a linear one, indicating that the binding sites for [3H]tamoxifen and 7keto[3H]cholestanol exhibit very similar binding affinities for these compounds, a finding consistent with the possibility that the binding site for [3H]tamoxifen may be the same as that for 7-keto[3H]cholestanol. Fig. 3 shows the Scatchard analysis of 7-keto [3H]cholestanol and [3H]tamoxifen binding to the rat liver microsomal binding site. The plot reveals a single class of binding sites for 7-keto[3H]cholestanol with an apparent equilibrium dissociation constant (Kd) of 2.7 nM. The number of binding sites amounted to 790 f m o l / m g protein. A similar Scatchard analysis using the same rat liver microsomal preparation, but with [3H]tamoxifen as the radioactive ligand, revealed a K d of 2.0 nM and a binding site concentration of 839 f m o l / m g protein. It would appear that the affinity of tamoxifen for the binding site was only slightly higher than that of 7-ketocholestanol, whereas the binding site concentrations for the two ligands were very similar. Fig. 4 shows the effect of various concentrations of tamoxifen on 7-keto[3H]cholestanol binding to the microsomal binding site. The presence of tamoxifen did not significantly affect the number of binding sites (horizontal intercept), but clearly decreased the binding affinity (slope) in a dose-dependent manner. In the presence of 0, 1, 4, 10, 16 and 20 nM tamoxifen, the K d increased from 4.4 nM to 4.9, 6.8, 12.7, 17.2 and 20.4 nM, respectively. If tamoxifen competes with 7keto[3H]cholestanol for the same binding site, then the change in the apparent dissociation constant should fit the expression K~ = K a ( l +
I/Ki)
1.0
oo . • o g
0.5-
o.o o.o
o.5 t .o 1.5 2.o 3H-ligand bound (nN}
Fig. 3. Scatchard analysis of 7-keto[3HJcholestanol (©) and [3H]tamoxifen (O) binding to rat liver microsomal binding sites. Saturation analysis was carried out in duplicate as described in the Materials and Methods section. The straight lines were obtained by linear regressionanalysis.
2:V
1.4 12
1
0
0
~ 15
~
~1
0
5
I0
15
20
go o.4 0.2 0.0
2
4
Bound 7-ket0 [3H]ch01estan01 InN) Fig. 4. Scatchard analysis of 7-keto[ 3H]cho]estanol binding to rat liver binding sites in the presence of varying concentrations of tamo×ifen. Saturation analysis was set up in duplicate as described in the Materials and Methods section in the presence of 0 (©), I (e), 4 (z~), 10 (A), 16 (O) and 20 (e) nM tamoxifen. Inset: the prevailing
tamoxifen concentration was plotted against the apparent equilibrium dissociation constant K~ (nM). All lines were obtained by linear regression analysis.
where K d is the dissociation constant for 7keto[3H]cholestanol in the absence of tamoxifen, K d is the apparent dissociation constant for 7-keto[3H]cho lestanol in the presence of tamoxifen, I is the concentration of tamoxifen and Ki is the dissociation constant for tamoxifen. This relationship predicts that, if tamoxifen competes with 7-keto[3H]cholestanol for the same binding site, Kd should vary linear with the prevailing tamoxifen concentration (I). This was found to be the case (Fig. 4, inset). Additional support for the possibility that tamoxifen competes directly with 7-keto[3H]cholestanol for the same binding site comes from kinetic studies in which the effect of various concentrations of tamoxifen on the dissociation rate of 7-keto[3H]cholestanol from the microsomal binding site was studied. Fig. 5 shows the dissociation of bound 7-keto[3H]cholestanol from the microsomal binding site as a function of time in the presence of varying concentrations of tamoxifen. The time required for 50% of specifically bound 7keto[3H]cholestanol to dissociate ( q / 2 ) in the absence of tamoxifen was 123.8 rain with a corresponding dissociation rate constant of 9.33 × 10 -5 s a. In the presence of 0.625, 1.25 and 2.5 nM tamoxifen, the tl/2 values were 121.3, 122.6 and 126.4 min, respectively. The corresponding dissociation rate constants (which equal In 2/q/2 ) were 9.52, 9.42, and 9.14× 10 5 s-l, respectively. It is clear that tamoxifen has no significant effect on the dissociation rate of bound 7-keto[3H]cholestanol, a finding to be expected if tamoxifen and 7-ketocholestanol compete for the same binding site. Table II compares the concentrations of 7keto[3H]cholestanol binding sites with those of [3H]tamoxifen binding sites in the 20000 × g super-
159
-0.2
We have also carried out limited studies with [3H]cholesterol and 25-hydroxy[3H]cholesterol as the radioactive ligand. Neither of these compounds displayed specific binding to the rat liver microsomal fraction whether activated charcoal or the detergent wash procedure outlined in Table I was used to separate bound from free ligand.
-0.3
Discussion
o.I o.o x~
g 2
-0.1
u
2
~I l l -0.4 t~
~
q
-0.~
i
i
30
60
i
i
i
90 120 t50
Time {rain)
Fig. 5. dissociation of bound 7-keto[3H]cholestanol from rat liver binding sites. The procedure for the dissociation kinetics experiment was described in the Materials and Methods section. The amount of specifically bound 7-keto[3H]cholestanol in the absence (e), or presence of 0.625 (4), 1.25 ( , ) or 2.5 nM (11) tamoxifen is plotted against time. The results represent the mean of duplicate determinations. The straight lines and the half-times of dissociation used for calculating the dissociation rate constants were obtained by linear regression analysis.
natants of ten different organs of the rat. The concentrations of binding sites for the two ligands in general parallel each other, the highest concentrations being observed in the liver and the lowest in skeletal and cardiac muscle whether [3H]tamoxifen or 7-keto[3H] cholestanol was used as the radioactive ligand. This observation is again consistent with the possibility that the two binding sites are the same.
TABLE II Concentration of binding sites for [ 3H]tamoxifen and 7-keto[ ~H]choles tanol in the 20000 × g supernatant of different organs in the rat
The preparation of the 20000× g supernatants of different organs and the measurement of 7-keto{ 3H]cholestanol binding site concentrations were carried out as described in the Materials and Methods section. Scatchard analysis was employed to determine the binding site concentration in the liver. In other organs, estimates of binding site concentration were made by means of single saturating dose analysis. Results represent the mean of duplicate determinations and are expressed in f m o l / m g protein. Organ
[ 3H]Tamoxifen binding site concentration
7-Keto[ 3H]cholestanol binding site concentration
Liver Adrenal Ovaries Kidneys Brain Uterus Lung Spleen Muscle Heart
800 110 90 60 43 40 30 12 10 3
1060 110 145 65 58 54 27 10 11 4
The studies of Murphy et al. [17] first suggested that oxygenated derivatives of cholesterol might bind to cockerel liver antiestrogen binding sites, albeit with very low affinity. The experiments reported here clearly indicate that using the experimental procedure described in Table I high-affinity binding sites for oxysterols could be shown to exist in the microsomal fraction of rat liver. They also show that these microsomal binding sites are clearly distinguishable from the cytosolic oxysterol binding protein described by other workers and are probably identical to the antiestrogen binding sites. In the studies described here, we had found it impossible to show specific binding of 7-keto[3H]cholestanol by means of conventional techniques such as charcoal adsorption. The major problem, well recognized by investigators studying the binding of 25-hydroxy[3H] cholesterol to the cytosolic oxysterol binding protein [8], appeared to be the presence in tissue preparations of other binding sites which are not saturable. In the presence of these nonsaturable binding sites, it is difficult to show inhibition of radioactive ligand binding by excess unlabeled ligand, since any displaced radioactive ligand would be taken up promptly by the nonsaturable binding sites and would, therefore, appear to be 'bound'. The procedure outlined in Table 1 proved critical for revealing specific binding sites for 7keto[3H]cholestanol. The detergent washes appeared to remove selectively most of the radioactivity bound to the non-saturable sites, thus permitting quantitation of the radioactive ligand remaining bound to the saturable binding sites which were detergent-resistant presumably because they were of higher affinity. Activated charcoal, by contrast, was unable to strip the radioactive ligand off the non-saturable sites. Detergent treatment may have greater general applicability and should probably be explored further as a means of reducing interference by non-saturable binding sites in ligand binding studies. Interference by non-saturable binding sites in the rat liver preparation used in our binding studies also made it difficult to compare the binding affinities of oxysterols with those of non-steroidal antiestrogens. In the binding studies similar to those depicted in Fig. 1, for example, we consistently observed that tamoxifen was 5-10-times more potent than unlabeled 7-ketocholestanol in inhibiting 7-keto[3H]cholestanol binding. However, given the fact that 7-ketocholestanol, but not tamoxifen,
160 binds to a very significant extent to the non-saturable binding sites present in the 20000 x g tissue supernatant (unpublished observation), it could be argued that not all the 7-ketocholestanol added was available for competition for the high-affinity saturable binding site, leading to a spuriously low estimate of its binding affinity. Its true affinity for the binding site might in fact be closer to that of tamoxifen, as was suggested by the Scatchard analysis shown in Fig. 3. The studies reported here also provide evidence that the high-affinity microsomal binding site for 7keto[ 3H]cholestanol is also the antiestrogen binding site. First, the relative binding affinities of the non-steroidal antiestrogens and oxysterols examined were similar, irrespective of whether the radioactive ligand was 7-keto[3H]cholestanol or [3H]tamoxifen (Fig. 2). Second, saturation analysis (Fig. 4) and kinetic studies (Fig. 5) clearly suggested that 7-ketocholestanol and tamoxifen competed for the same binding site. Furthermore, the tissue distribution of antiestrogen binding sites was very similar to that of binding sites for 7-ketocholestanol (Table II). Taken together, these findings strongly suggest that the high-affinity binding site for 7-ketocholestanol is the same as the antiestrogen binding site. Definitive proof of this point must await the isolation and characterization of the binding protein. The identification of the antiestrogen binding site as a high-affinity binding site for oxygenated sterols raises a number of questions. The first concerns the relationship of this binding site to the cytosolic oxysterol binding protein described earlier by other workers [7-11,13,14,16]. The latter displays a number of features which clearly distinguish it from the antiestrogen binding site. First, it appears to be a cytosolic protein, whereas the antiestrogen binding site is localized predominantly to the microsomal fraction [24]. Second, it binds 25-hydroxycholesterol and 20a-hydroxycholesterol (as well as some other oxysterols) with high affinity [9,12-14.16], whereas the antiestrogen binding site appears to have little or no affinity for these two sterols (Fig. 1). We have also not been able to detect any specific binding of 25-hydroxy[3H]cholesterol in the rat liver microsomal fraction by either the procedure given in Table I or by charcoal adsorption (result not shown). It would appear, therefore, that the microsomal and the cytosolic oxysterol binding sites have very different ligand specificities. In this regard, it is of interest to note that the cytosolic binding protein has been reported by some investigators [9] to exhibit much higher affinities for oxysterols with oxygen functions on the side chain than for oxysterols which are oxygenated at C7. The microsomal binding site, on the other hand, appears to have diametrically opposite ligand specificities (Fig. 1). Indeed, one may speculate that different classes of oxysterols may have different intracellular binding sites, although whether this has any functional
correlate requires further study. In any case, it is clear that the microsomal oxysterol binding site and the cytosolic oxysterol binding protein are different not only in their intracellular distribution but also in their ligand specificities. The second question relates to the possible biological significance of the high-affinity binding of oxysterols by antiestrogen binding sites which are known to be ubiquitously distributed. They bind non-steroidal antiestrogens and many other pharmacological agents with high affinity (for a review see Ref. 25). Currently little is known about their physiological function. The demonstration that they bind oxysterols with high affinity raises the possibility that some of the known biological effects of oxysterols could be mediated by the antiestrogen binding sites, just as the cytosolic oxysterol binding protein is believed by some investigators to mediate the inhibition of cholesterol biosynthesis by oxysterols [13,14]. In this connection, it is intriguing to note that the non-steroidal antiestrogens have been reported to inhibit cholesterol biosynthesis [26,27] and cell proliferation [28] by mechanisms which are independent of the estrogen receptor. Both of these effects are shared by the oxysterols. The possibility that both the antiestrogens and oxysterols exert these effects through the antiestrogen binding site deserves further examination. Thirdly, although the nature of the endogenous ligand for antiestrogen binding sites remains elusive, the observation that some oxysterols bind to this site with high affinity raises the possibility that the endogenous ligand, if one exists, could be an oxysterol or a related compound. Currently there is no evidence that 7-ketocholestanol plays any regulatory role in vivo and the biological significance of its binding to the antiestrogen binding site remains to be explored. Nevertheless, the identification of oxysterols which bind to the antiestrogen binding site with high affinity should facilitate further studies on the possible physiological role of antiestrogen binding sites and the mechanisms underlying the biological effects of oxysterols. It would be of particular interest to determine, for example, whether the microsomal oxysterol binding site plays a role in the sterol-enhanced degradation of HMG-CoA reductase [29], an enzyme which is located in the microsomes and which regulates cholesterol biosynthesis.
Acknowledgements This study was supported by the National University of Singapore and the Shaw Foundation. The excellent technical assistance of Miss Chiew Yoke Eng and the superb secretarial help of Miss Asha Das are gratefully acknowledged. I thank Dr. O.L. Kon for helpful discussions.
161
References 1 Beck, J.P. and Crastes de Paulet, A. (eds.) (1988) Biological Activities of Oxygenated Sterols, Colloques Inserm Volume 166, Institut National de la Sant6 et de la Recherche M6dicale, Paris. 2 Kandutsch, A.A., Chen, H.W. and Heiniger, H.-J. (1978) Science 201,498-501. 3 Brown, M.S. and Goldstein, J.L. (1974) J. Biol. Chem. 249, 7306-7314. 4 Kandutsch, A.A. and Chen, H.W. (1973) J. Biol. Chem. 248, 8408-8417 5 Kandutsch, A.A. and Chen, H.W. (1974) J. Biol. Chem. 249, 6057-6061. 6 Chen, H.W., Kandutsch, A.A. and Waymouth, C. (1974) Nature 251,419-421. 7 Kandutsch, A.A., Chen, H.W. and Shown, E.P. (1977) Proc. Natl. Acad. Sci. USA 74, 2500-2506. 8 Kandutsch, A.A. and Thompson, E.B. (1980) J. Biol. Chem. 255, 10813-10826. 9 Beseme, F., Astruc, M.E., Defay, R. and Crastes de Paulet, A. (1987) FEBS Lett. 210, 97-103. 10 Beseme, F., Astruc, M.E., Defay, R., Descomps, B. and Crastes de Paulet, A. (1986) Biochim. Biophys. Acta 886, 96-108. 11 Sinensky, M. and Mueller, G. (1981) Arch. Biochem. Biophys. 209, 314-320. 12 Taylor, F.R., Saucier, S.E., Shown, E.P., Parish, E.J. and Kandutsch, A.A. (1984) J. Biol. Chem. 259, 12382-12387. 13 Taylor, F.R. and Kandutsch, A.A. (1987) in Sterol/Steroid Hormone Mechanism of Action (Kumar, R. and Spelsberg, T., eds.), pp. 395-407, Martinus Nijhoff, Boston.
14 Taylor, F.R. and Kandutsch, A.A. (1985) Chem. Phys. Lipids 38, 187-194. 15 Sinensky, M. and Mueller, G. (1981) Arch. Biochem. Biophys. 209, 314-320. 16 Dawson, P.A., Van der Westhuyzen, D.R., Goldstein, J.L. and Brown, M.S. (1989) J. Biol. Chem. 264, 9046-9052. 17 Murphy, P.R., Breckenridge, W.C. and Lazier, C.B. (1985) Biochem. Biophys. Res. Commun. 127, 786-792. 18 Sutherland, R.L., Murphy, L.C., Foo, M.S., Green, M.D., Whybourne, A.M. and Krozowski, Z.S. (1980) Nature 288, 273-275. 19 Hwang, P.L.H. and Matin, A. (1989) J. Lipid Res. 30, 239-245. 20 Kamat, V.B. and Wallach, D.F.H. (1965) Science 148, 1343-1345. 21 Hwang, P.L.H. (1986) Biochem. J. 237, 749-755. 22 Hwang, P.L.H. (1987) Biochem. J. 243, 359-364. 23 Lowry, O.H., Roseborough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275. 24 Sudo, K., Monsma, F.J. and Katzenellenbogen, B.S. (1983) Endocrinology 112, 425-434. 25 Watts, C.K.W. and Sutherland, R.L. (1987) Mol. Pharmacol. 31, 541-551. 26 Cypriani, B., Tabacik, C. and Descomps, B. (1988) Biochim. Biophys. Acta 972, 167-178. 27 Cypriani, B., Tabacik, C., Descomps, B. and Crastes de Paulet, A. (1988) J. Steroid Biochem. 31, 763-771. 28 Sutherland, R.L., Watts, C.K.W. and Ruenitz, P.C. (1986) Biochem. Biophys. Res. Commun. 140, 523-520. 29 Gil, G., Faust, J.R., Chin, D.J., Goldstein, J.L. and Brown, M.S. (1985) Cell 41, 249-258.
