Journal of Neurochemistry Raven Press, Ltd., New York 0 1992 International Society for Neurochemistry
Short Review
Biochemical, Physiological, and Pathological Aspects of the Peripheral Benzodiazepine Receptor Moshe Gavish, Yeshayahu Katz, *Shalom Bar-Ami, and TRonit Weizman Rappapport Family Institute for Research in the Medical Sciences and Department of Pharmacology, Bruce Rappaport Faculty of Medicine, Technion-IsraelInstitute of Technology;*Department of Obstetrics and Gynecology, Rambam Medical Center, Haifa; TTeI Aviv Community Mental Health Center; and TSackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
Benzodiazepines (BZs) have anxiolytic, anticonvulsant, muscle-relaxant, and hypnotic properties. It has been suggested that these therapeutic effects are mediated via specific BZ receptors located in the CNS and coupled with y-aminobutyric acid (GABA) receptors and C1- channels (Tallman et al., 1980). It is surprising that high densities of [3H]diazepambinding sites have also been found in the kidney (Braestrup and Squires, 1977). [3H]Diazepam and [3H]Ro 5-4864 (4’-chlorodiazepam) have been found to label, with high affinity, recognition sites in various peripheral tissues, as well as in nonneuronal brain tissue. These membranal sites have been named “peripheral BZ binding sites.” Several studies have indicated that these sites have a physiological role, and therefore they have been termed “peripheral BZ receptors” (PBR) (Verma and Snyder, 1989). PBR differ from central-type BZ receptors (CBR) in their lack of coupling to GABA receptors and in their ligand specificity. CBR exhibit high affinity to clonazepam but not to Ro 5-4864 or PK 11195 (an isoquinoline carboxamide derivative). The reverse is true with regard to PBR, which exhibit high affinity to Ro 5-4864 and PK 11 195 but low affinity to clonazepam. The present review will deal with various aspects of PBR on the pharmacological, biochemical, endocrinological, and pathological level. SUBCELLULAR LOCALlZATlON OF PBR The subcellular localization of PBR in the adrenal gland ofthe rat has been studied (Anholt et al., 1986b).
It was found that the autoradiographic pattern of [3H]PK 11 195 sites in tissue sections of adrenal gland is similar to the histochemical distribution of cytochrome oxidase and monoamine oxidase (MAO). Subcellular fractionation of rat adrenal gland homogenates indicates that there is a correlation between [3H]PK 1 1 195 binding in the various fractions (nuclear, mitochondrial, microsomal, and soluble) and cytochrome oxidase activity-a marker for the mitochondria. Such a correlation has not been found with markers for the nuclei, lysosomes, peroxisomes,endoplasmic reticulum, plasma membrane, or cytoplasm. These results support the assumption that PBR are associated with the mitochondrial fraction (Anholt et al., 1986b). Similar observations have been made in other tissues, such as testis, lung, kidney, heart, skeletal muscle, liver, and brain (Antkiewicz-Michaluk et al., 1988a). Subcellular fractionation of these tissues revealed that PK 11 195 binding sites subfractionate in a manner nearly identical to that of the mitochondrial enzyme succinate dehydrogenase (Antkiewicz-Michaluket al., 1988a). Titration of the isolated mitochondria with digitonin causes the release of PBR and MAO, but not of cytochrome oxidase (Anholt et al., 1986b).This observation indicates that PBR are associated with the mitochondrial outer membrane. The fact that PBR density is increased in the mitochondrial outer membrane compared with the value in intact mitochondria (Anholt et al., 19866) supports this hypothesis. However, another study demonstrated PBR in red blood cells, which lack mitochondria (Olson et al., 1988). This finding indicates that PBR are also localized in a nonmitochondrial fraction.
~
amidopropyl)dimethylammonio]- 1-propane sulfonate;Ez ,estradiol178; GABA, y-aminobutyric acid; MAO, monoamine oxidase; MES, maximal electroshock MR, Maudsley reactive; P4,progesterone: PBR, peripheral benzodiazepine receptor(s); SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.
Address correspondence and reprint requests to Dr. M. Gavish at Department of Pharmacology, Bruce Rappaport Faculty of Medicine, Technion-Israel Institute of Technology, P.O. Box 9649, 3 1096 Haifa, Israel. Abbreviations used. ACTH, corticotropin; BZ, benzodiazepine; CBR, central-type benzodiazepine receptor(s); CHAPS, 3-[(3-chol-
I589
1590
M. GAVISH ET AL.
AGONIST AND ANTAGONIST FOR PBR Thermodynamic analysis of [3H]Ro 5-4864 and [3H]PK 1 1 195 binding to rat cardiac membranes indicates that [3H]Ro5-4864binding is enthalpy driven, whereas [3H]PK 1 1195 binding is entropy driven (Le Fur et al., 1983b). The two-step model proposed for @-receptors(Weiland et al., 1979)suggests that agonist binding is enthalpy driven, whereas antagonist binding is entropy driven. Hence, it seems, according to these results (Le Fur et al., 1983b), that Ro 5-4864and PK 1 1 195 are an agonist and an antagonist, respectively. of the PBR. However, this hypothesis was fundamentally proposed for @-receptorligands and has been shown not to be universal. The binding of these two ligands to kidney membranes before and after chemical modification of histidine in PBR by diethylpyrocarbonate also supports this hypothesis (Benavides et al., 1984~).[3H]Ro5-4864binding is not affected by such chemical modification, whereas [3H]PK 1 1 195 binding is decreased owing to a decrease in the receptor density without a change in the affinity. However, Ro 5-4864 protects the inactivation of [3H]PK 1 1 195 binding obtained by this chemical modification. These results suggest that Ro 5-4864 and PK 1 1 195 interact with different conformations of the receptor and that Ro 54864 and PK 1 1 195 may be an agonist and an antagonist, respectively, of the PBR. This hypothesis is further supported by the differential effect of detergents on [3H]Ro 5-4864and [3H]PK 1 1 195 binding to PBR (Awad and Gavish, 1988). The detergents Triton X100, 3-[(3-cholamidopropyl)dimethylammonio]1propane sulfonate (CHAPS). Tween 20, and deoxycholic acid at low concentrations cause a decrease of at least 50% in [3H]Ro 5-4864 binding to rat kidney membranes, whereas [3H]PK 1 1 195 binding remains unaffected (Awad and Gavish, 1988). Electrophysiological studies have shown that Ro 54864 decreases the duration of intracellular action potential and the contractility in guinea pig heart preparation. These effects are antagonized by PK 1 1 195 (Mestre et al., 1984).On the other hand, other studies have shown that Ro 5-4864 and PK 1 1 195 produce similar effects. One study showed that both Ro 5-4864 and PK 1 1 195 increase coronary flow in isolated retrograde-perfused LangendorfF rat heart preparations (Grupp et al.. 1987).Another study demonstrated that the secretion of progesterone (P4)and estradiol-170 (E2) from human term placental explants is increased by either Ro 5-4864 or PK 1 1 195. The effect of Ro 54864 on the secretion of P4 and E2 is not antagonized by PK 1 1 195. Clonazepam, a BZ ligand specific for CBR, has no effect on the secretion of either steroid (Barnea et al., 1989).
BINDING OF Ro 5-4864 AND PK 11195 TO PBR IN VARIOUS SPECIES It has been established that various BZs, such as diazepam, flunitrazepam, and Ro 5-4864,but not clonJ. Neurochem., Vol. 58. No. 5 , 1992
azepam, label PBR in the brain and peripheral organs and cells of various mammals, such as the rat -and guinea pig (Braestrup and Squires, 1977;Wang et al., 1980; Regan et al., 1981; Schoemaker et al., 1981; Weissman et al., 1984).The non-BZ ligand PK 1 1 195 also exhibits binding affinity to PBR (Le Fur et al., 1983~).However, the affinity of Ro 5-4864 for cow brain is 200 times lower than its affinity for rat brain (Benavideset al., 1985b).Moreover, the potency of Ro 5-4864 in inhibiting [3H]PK 1 1 195 binding to cat brain sections and membranes is 140 times lower than that of PK 1 1 195 (Benavides et al., 1984b),whereas in rat brain membranes Ro 5-4864 and PK 1 1195 exhibit nanomolar affinity (Benavideset al., 19833).These results suggest that PBR in various mammals are not identical. Marked differences in PBR density and distribution in the brain of various species have been shown using an autoradiographic technique (Cymerman et al., 1986). The binding of [3H]PK 1 1 195 and [3H]Ro 5-4864 to membrane preparations from cerebral cortex and peripheral tissues of various species has been studied (Awad and Gavish, 1987). [3H]PK 1 1 195 binds with high affinity to rat and calf cerebral cortex and kidney membranes. [3H]Ro5-4864also successfully labels rat cerebral cortex and kidney membranes, but in calf cerebral cortex and kidney membranes, its binding is negligible. Displacement studies have shown that unlabeled Ro 5-4864,diazepam, and flunitrazepam are much more potent in displacing [3H]PK 1 1 195 from rat cerebral cortex and kidney membranes than from calf tissues. The potency of unlabeled Ro 5-4864in displacing [3H]PK 1 1195 from the cerebral cortex of various other species has also been tested, and the rank order of potency is rat = guinea pig > cat = dog > rabbit > calf. Analysis of these displacement curves reveals that Ro 5-4864 binds to two populations of binding sites from rat and calf kidney and from rat, guinea pig, rabbit, and calf cerebral cortex, but to a single population of binding sites from cat and dog cerebral cortex. The differences in binding characteristics of Ro 54864 versus PK 1 1 195 in various species may be related to their functions as agonist and antagonist, respectively (Le Fur et al., 1983b;Benavides et al., 1984~; Mestre et al., 1985). These differences may be related to differences in PBR structure. Similar differences in the binding affinity of [3H]Ro 5-4864and [3H]PK 1 1 195 have also been obtained in rat and calf pineal gland (Basile et al., 1986). PBR species differences and heterogeneity observed in membrane-bound PBR are retained in the soluble state and are probably attributable to variations in the molecular structure of PBR rather than to differences in the membrane environment (Awad and Gavish, 1989~). The specific binding o ~ ' [ ~ H ] P1K1 195 and [3H]Ro 5-4864to various human tissues has been determined to elaborate whether binding characteristics of PBR located in human tissues are-similar to those of PBR located in calf or rat tissues (Awad and Gavish, 1991).
1591
PERIPHERAL BENZODIAZEPINE RECEPTORS
[3H]PK 11195 bound with high affinity (-2 nM) to human cerebral cortex, kidney, and colon membranes; the specific binding of [3H]Ro 5-4864, on the other hand, was barely detectable (nonspecific binding was -90% of the total binding). Furthermore, unlabeled PK 1 1 195 was two orders of magnitude more potent than unlabeled Ro 5-4864 in displacing 13H]PK1 I 195 specific binding from human cerebral cortex and kidney membranes (Table 1). These results are in accord with those of calf tissues but differ from those of rat tissues. The differences may have importance for the evolution of these receptors in various species. Preliminary results in our laboratory indicate that the affinity differences are not accompanied by alterations in the molecular weight of the protein that binds isoquinoline carboxamide derivatives. MOLECULAR PROPERTIES OF PBR
Purification of PBR will contribute to understanding their function. Purification of intact PBR requires a suitable detergent that will not cause major damage to the activity. CBR have been solubilized successfullyby various detergents; PBR, on the other hand, are more sensitive to detergent treatment and usually undergo denaturation at detergent concentrations lower than those commonly used (Anholt et al., 1986a;Awad and Gavish, 1988). Digitonin has been found to be the most suitable for solubilization of PBR from rat kidney and adrenal gland (Benavides et al., 1985a; Gavish and Fares, 19853). Triton X-100 was found suitable for PBR solubilization from rat kidney only after the detergent concentration was lowered by Bio-Beads (Martini et al., 1983). Sodium cholate also was found appropriate for PBR solubilization from rat kidney, but only after addition of soybean lipids (Anholt et al., 1986~).PBR have also been solubilized from cat cerebral cortex by CHAPS in the presence of NaCl (Awad and Gavish, 19893). The molecular mass of PBR solubilized from rat heart by digitonin in the presence of the detergent was 220 kDa (Doble et al., 1985). Such a high molecular weight is due to digitonin-PBR micelles. However, this molecular weight was determined
by radiation inactivation, which often yields erratic values and is dependent on protein complexes rather than on micelles. Partial purification of PBR containing both the site for isoquinoline carboxamide and that for BZs has been reported by McEnery et al. (1989), who used a hydroxylapatite column in the presence of an alkyl glucoside. Using a fast-protein liquid chromatography sizing column, they found the molecular mass of PBR to be 52 kDa. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) revealed two bands with molecular masses of 18 and 30 kDa, which contained the sites for isoquinoline carboxamide and BZs, respectively (McEnery et al., 1989). Another approach to receptor purification is covalent attachment of the radioactive ligand to the binding site, and then it is possible to follow the radioactivity during the various steps of the purification process. [3H]PK 14105 binds with high affinity to PBR and labels the site covalently after ultraviolet irradiation of the PK 14105 receptor complex (Doble et al., 1987).The molecular mass of the subunit containing the isoquinoline carboxamide site, as detected by this technique followed by SDS-PAGE, is -17-18.5 kDa (Doble et al., 1987; Antkiewicz-Michaluk et al., 1988b; Skowronski et al., 1988). [3H]Flunitrazepam also photolabels PBR, but the molecular mass of this subunit is -30-35 kDa (Trifiletti et al., 1986; Snyder et al., 1987). The thiocyanate derivative of Ro 5-4864, [3HJAHN086, also covalently labels a protein of 30 kDa in the rat pineal gland (McCabe et al., 1989). Hence, the PBR may be a protein complex containing two subunits: one labeled by BZs, with a molecular mass of -30-35 kDa, and a second one, labeled by isoquinoline carboxamide, with a molecular mass of -17-18.5 kDa (Snyder et al., 1989). The site for [3H]PK 11195 has been purified to apparent homogeneity (Antkiewicz-Michaluk et al., 19886). Rat adrenal gland mitochondrial membranes have been covalently labeled with [3H]PK 14105 and solubilized by digitonin. The soluble PBR-[3H]PK 14105 complex has been purified by ion-exchange chromatography and reverse-phase HPLC. SDS-PAGE
TABLE 1. Inhibition of 13HjPK 11195 speciftc binding to rut, human, and catfcerebrul cortex and kidney membranes by PK 1I1 95, Ro 5-4864. and clonazevam
Cerebral cortex
Kidney
Inhibitor
Rat
Human
Calf
Rat
Human
Calf
PK 1 1 195 RO 5-4864 Clonazepam
0.003
0.005 0.750 >10
0.004 9.850 110
0.0 18
0.008
0.027 >10
0.750 > 10
0.003 9.860 >10
0.007 >10
Incubations were performed for 60 min at 4°C with 1 nM ['HIPK I 1 195 in the absence or presence of various concentrations of unlabeled PK 1 1 195, Ro 5-4864, or clonazepam (final concentration, 10-'0-10-4M) as inhibitors. lCSo(concentration of unlabeled inhibitor required to inhibit 50% of specific ['HIPK 11 195 binding) values shown are the mean of three separate experiments with
Short Review
Biochemical, Physiological, and Pathological Aspects of the Peripheral Benzodiazepine Receptor Moshe Gavish, Yeshayahu Katz, *Shalom Bar-Ami, and TRonit Weizman Rappapport Family Institute for Research in the Medical Sciences and Department of Pharmacology, Bruce Rappaport Faculty of Medicine, Technion-IsraelInstitute of Technology;*Department of Obstetrics and Gynecology, Rambam Medical Center, Haifa; TTeI Aviv Community Mental Health Center; and TSackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
Benzodiazepines (BZs) have anxiolytic, anticonvulsant, muscle-relaxant, and hypnotic properties. It has been suggested that these therapeutic effects are mediated via specific BZ receptors located in the CNS and coupled with y-aminobutyric acid (GABA) receptors and C1- channels (Tallman et al., 1980). It is surprising that high densities of [3H]diazepambinding sites have also been found in the kidney (Braestrup and Squires, 1977). [3H]Diazepam and [3H]Ro 5-4864 (4’-chlorodiazepam) have been found to label, with high affinity, recognition sites in various peripheral tissues, as well as in nonneuronal brain tissue. These membranal sites have been named “peripheral BZ binding sites.” Several studies have indicated that these sites have a physiological role, and therefore they have been termed “peripheral BZ receptors” (PBR) (Verma and Snyder, 1989). PBR differ from central-type BZ receptors (CBR) in their lack of coupling to GABA receptors and in their ligand specificity. CBR exhibit high affinity to clonazepam but not to Ro 5-4864 or PK 11195 (an isoquinoline carboxamide derivative). The reverse is true with regard to PBR, which exhibit high affinity to Ro 5-4864 and PK 11 195 but low affinity to clonazepam. The present review will deal with various aspects of PBR on the pharmacological, biochemical, endocrinological, and pathological level. SUBCELLULAR LOCALlZATlON OF PBR The subcellular localization of PBR in the adrenal gland ofthe rat has been studied (Anholt et al., 1986b).
It was found that the autoradiographic pattern of [3H]PK 11 195 sites in tissue sections of adrenal gland is similar to the histochemical distribution of cytochrome oxidase and monoamine oxidase (MAO). Subcellular fractionation of rat adrenal gland homogenates indicates that there is a correlation between [3H]PK 1 1 195 binding in the various fractions (nuclear, mitochondrial, microsomal, and soluble) and cytochrome oxidase activity-a marker for the mitochondria. Such a correlation has not been found with markers for the nuclei, lysosomes, peroxisomes,endoplasmic reticulum, plasma membrane, or cytoplasm. These results support the assumption that PBR are associated with the mitochondrial fraction (Anholt et al., 1986b). Similar observations have been made in other tissues, such as testis, lung, kidney, heart, skeletal muscle, liver, and brain (Antkiewicz-Michaluk et al., 1988a). Subcellular fractionation of these tissues revealed that PK 11 195 binding sites subfractionate in a manner nearly identical to that of the mitochondrial enzyme succinate dehydrogenase (Antkiewicz-Michaluket al., 1988a). Titration of the isolated mitochondria with digitonin causes the release of PBR and MAO, but not of cytochrome oxidase (Anholt et al., 1986b).This observation indicates that PBR are associated with the mitochondrial outer membrane. The fact that PBR density is increased in the mitochondrial outer membrane compared with the value in intact mitochondria (Anholt et al., 19866) supports this hypothesis. However, another study demonstrated PBR in red blood cells, which lack mitochondria (Olson et al., 1988). This finding indicates that PBR are also localized in a nonmitochondrial fraction.
~
amidopropyl)dimethylammonio]- 1-propane sulfonate;Ez ,estradiol178; GABA, y-aminobutyric acid; MAO, monoamine oxidase; MES, maximal electroshock MR, Maudsley reactive; P4,progesterone: PBR, peripheral benzodiazepine receptor(s); SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.
Address correspondence and reprint requests to Dr. M. Gavish at Department of Pharmacology, Bruce Rappaport Faculty of Medicine, Technion-Israel Institute of Technology, P.O. Box 9649, 3 1096 Haifa, Israel. Abbreviations used. ACTH, corticotropin; BZ, benzodiazepine; CBR, central-type benzodiazepine receptor(s); CHAPS, 3-[(3-chol-
I589
1590
M. GAVISH ET AL.
AGONIST AND ANTAGONIST FOR PBR Thermodynamic analysis of [3H]Ro 5-4864 and [3H]PK 1 1 195 binding to rat cardiac membranes indicates that [3H]Ro5-4864binding is enthalpy driven, whereas [3H]PK 1 1195 binding is entropy driven (Le Fur et al., 1983b). The two-step model proposed for @-receptors(Weiland et al., 1979)suggests that agonist binding is enthalpy driven, whereas antagonist binding is entropy driven. Hence, it seems, according to these results (Le Fur et al., 1983b), that Ro 5-4864and PK 1 1 195 are an agonist and an antagonist, respectively. of the PBR. However, this hypothesis was fundamentally proposed for @-receptorligands and has been shown not to be universal. The binding of these two ligands to kidney membranes before and after chemical modification of histidine in PBR by diethylpyrocarbonate also supports this hypothesis (Benavides et al., 1984~).[3H]Ro5-4864binding is not affected by such chemical modification, whereas [3H]PK 1 1 195 binding is decreased owing to a decrease in the receptor density without a change in the affinity. However, Ro 5-4864 protects the inactivation of [3H]PK 1 1 195 binding obtained by this chemical modification. These results suggest that Ro 5-4864 and PK 1 1 195 interact with different conformations of the receptor and that Ro 54864 and PK 1 1 195 may be an agonist and an antagonist, respectively, of the PBR. This hypothesis is further supported by the differential effect of detergents on [3H]Ro 5-4864and [3H]PK 1 1 195 binding to PBR (Awad and Gavish, 1988). The detergents Triton X100, 3-[(3-cholamidopropyl)dimethylammonio]1propane sulfonate (CHAPS). Tween 20, and deoxycholic acid at low concentrations cause a decrease of at least 50% in [3H]Ro 5-4864 binding to rat kidney membranes, whereas [3H]PK 1 1 195 binding remains unaffected (Awad and Gavish, 1988). Electrophysiological studies have shown that Ro 54864 decreases the duration of intracellular action potential and the contractility in guinea pig heart preparation. These effects are antagonized by PK 1 1 195 (Mestre et al., 1984).On the other hand, other studies have shown that Ro 5-4864 and PK 1 1 195 produce similar effects. One study showed that both Ro 5-4864 and PK 1 1 195 increase coronary flow in isolated retrograde-perfused LangendorfF rat heart preparations (Grupp et al.. 1987).Another study demonstrated that the secretion of progesterone (P4)and estradiol-170 (E2) from human term placental explants is increased by either Ro 5-4864 or PK 1 1 195. The effect of Ro 54864 on the secretion of P4 and E2 is not antagonized by PK 1 1 195. Clonazepam, a BZ ligand specific for CBR, has no effect on the secretion of either steroid (Barnea et al., 1989).
BINDING OF Ro 5-4864 AND PK 11195 TO PBR IN VARIOUS SPECIES It has been established that various BZs, such as diazepam, flunitrazepam, and Ro 5-4864,but not clonJ. Neurochem., Vol. 58. No. 5 , 1992
azepam, label PBR in the brain and peripheral organs and cells of various mammals, such as the rat -and guinea pig (Braestrup and Squires, 1977;Wang et al., 1980; Regan et al., 1981; Schoemaker et al., 1981; Weissman et al., 1984).The non-BZ ligand PK 1 1 195 also exhibits binding affinity to PBR (Le Fur et al., 1983~).However, the affinity of Ro 5-4864 for cow brain is 200 times lower than its affinity for rat brain (Benavideset al., 1985b).Moreover, the potency of Ro 5-4864 in inhibiting [3H]PK 1 1 195 binding to cat brain sections and membranes is 140 times lower than that of PK 1 1 195 (Benavides et al., 1984b),whereas in rat brain membranes Ro 5-4864 and PK 1 1195 exhibit nanomolar affinity (Benavideset al., 19833).These results suggest that PBR in various mammals are not identical. Marked differences in PBR density and distribution in the brain of various species have been shown using an autoradiographic technique (Cymerman et al., 1986). The binding of [3H]PK 1 1 195 and [3H]Ro 5-4864 to membrane preparations from cerebral cortex and peripheral tissues of various species has been studied (Awad and Gavish, 1987). [3H]PK 1 1 195 binds with high affinity to rat and calf cerebral cortex and kidney membranes. [3H]Ro5-4864also successfully labels rat cerebral cortex and kidney membranes, but in calf cerebral cortex and kidney membranes, its binding is negligible. Displacement studies have shown that unlabeled Ro 5-4864,diazepam, and flunitrazepam are much more potent in displacing [3H]PK 1 1 195 from rat cerebral cortex and kidney membranes than from calf tissues. The potency of unlabeled Ro 5-4864in displacing [3H]PK 1 1195 from the cerebral cortex of various other species has also been tested, and the rank order of potency is rat = guinea pig > cat = dog > rabbit > calf. Analysis of these displacement curves reveals that Ro 5-4864 binds to two populations of binding sites from rat and calf kidney and from rat, guinea pig, rabbit, and calf cerebral cortex, but to a single population of binding sites from cat and dog cerebral cortex. The differences in binding characteristics of Ro 54864 versus PK 1 1 195 in various species may be related to their functions as agonist and antagonist, respectively (Le Fur et al., 1983b;Benavides et al., 1984~; Mestre et al., 1985). These differences may be related to differences in PBR structure. Similar differences in the binding affinity of [3H]Ro 5-4864and [3H]PK 1 1 195 have also been obtained in rat and calf pineal gland (Basile et al., 1986). PBR species differences and heterogeneity observed in membrane-bound PBR are retained in the soluble state and are probably attributable to variations in the molecular structure of PBR rather than to differences in the membrane environment (Awad and Gavish, 1989~). The specific binding o ~ ' [ ~ H ] P1K1 195 and [3H]Ro 5-4864to various human tissues has been determined to elaborate whether binding characteristics of PBR located in human tissues are-similar to those of PBR located in calf or rat tissues (Awad and Gavish, 1991).
1591
PERIPHERAL BENZODIAZEPINE RECEPTORS
[3H]PK 11195 bound with high affinity (-2 nM) to human cerebral cortex, kidney, and colon membranes; the specific binding of [3H]Ro 5-4864, on the other hand, was barely detectable (nonspecific binding was -90% of the total binding). Furthermore, unlabeled PK 1 1 195 was two orders of magnitude more potent than unlabeled Ro 5-4864 in displacing 13H]PK1 I 195 specific binding from human cerebral cortex and kidney membranes (Table 1). These results are in accord with those of calf tissues but differ from those of rat tissues. The differences may have importance for the evolution of these receptors in various species. Preliminary results in our laboratory indicate that the affinity differences are not accompanied by alterations in the molecular weight of the protein that binds isoquinoline carboxamide derivatives. MOLECULAR PROPERTIES OF PBR
Purification of PBR will contribute to understanding their function. Purification of intact PBR requires a suitable detergent that will not cause major damage to the activity. CBR have been solubilized successfullyby various detergents; PBR, on the other hand, are more sensitive to detergent treatment and usually undergo denaturation at detergent concentrations lower than those commonly used (Anholt et al., 1986a;Awad and Gavish, 1988). Digitonin has been found to be the most suitable for solubilization of PBR from rat kidney and adrenal gland (Benavides et al., 1985a; Gavish and Fares, 19853). Triton X-100 was found suitable for PBR solubilization from rat kidney only after the detergent concentration was lowered by Bio-Beads (Martini et al., 1983). Sodium cholate also was found appropriate for PBR solubilization from rat kidney, but only after addition of soybean lipids (Anholt et al., 1986~).PBR have also been solubilized from cat cerebral cortex by CHAPS in the presence of NaCl (Awad and Gavish, 19893). The molecular mass of PBR solubilized from rat heart by digitonin in the presence of the detergent was 220 kDa (Doble et al., 1985). Such a high molecular weight is due to digitonin-PBR micelles. However, this molecular weight was determined
by radiation inactivation, which often yields erratic values and is dependent on protein complexes rather than on micelles. Partial purification of PBR containing both the site for isoquinoline carboxamide and that for BZs has been reported by McEnery et al. (1989), who used a hydroxylapatite column in the presence of an alkyl glucoside. Using a fast-protein liquid chromatography sizing column, they found the molecular mass of PBR to be 52 kDa. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) revealed two bands with molecular masses of 18 and 30 kDa, which contained the sites for isoquinoline carboxamide and BZs, respectively (McEnery et al., 1989). Another approach to receptor purification is covalent attachment of the radioactive ligand to the binding site, and then it is possible to follow the radioactivity during the various steps of the purification process. [3H]PK 14105 binds with high affinity to PBR and labels the site covalently after ultraviolet irradiation of the PK 14105 receptor complex (Doble et al., 1987).The molecular mass of the subunit containing the isoquinoline carboxamide site, as detected by this technique followed by SDS-PAGE, is -17-18.5 kDa (Doble et al., 1987; Antkiewicz-Michaluk et al., 1988b; Skowronski et al., 1988). [3H]Flunitrazepam also photolabels PBR, but the molecular mass of this subunit is -30-35 kDa (Trifiletti et al., 1986; Snyder et al., 1987). The thiocyanate derivative of Ro 5-4864, [3HJAHN086, also covalently labels a protein of 30 kDa in the rat pineal gland (McCabe et al., 1989). Hence, the PBR may be a protein complex containing two subunits: one labeled by BZs, with a molecular mass of -30-35 kDa, and a second one, labeled by isoquinoline carboxamide, with a molecular mass of -17-18.5 kDa (Snyder et al., 1989). The site for [3H]PK 11195 has been purified to apparent homogeneity (Antkiewicz-Michaluk et al., 19886). Rat adrenal gland mitochondrial membranes have been covalently labeled with [3H]PK 14105 and solubilized by digitonin. The soluble PBR-[3H]PK 14105 complex has been purified by ion-exchange chromatography and reverse-phase HPLC. SDS-PAGE
TABLE 1. Inhibition of 13HjPK 11195 speciftc binding to rut, human, and catfcerebrul cortex and kidney membranes by PK 1I1 95, Ro 5-4864. and clonazevam
Cerebral cortex
Kidney
Inhibitor
Rat
Human
Calf
Rat
Human
Calf
PK 1 1 195 RO 5-4864 Clonazepam
0.003
0.005 0.750 >10
0.004 9.850 110
0.0 18
0.008
0.027 >10
0.750 > 10
0.003 9.860 >10
0.007 >10
Incubations were performed for 60 min at 4°C with 1 nM ['HIPK I 1 195 in the absence or presence of various concentrations of unlabeled PK 1 1 195, Ro 5-4864, or clonazepam (final concentration, 10-'0-10-4M) as inhibitors. lCSo(concentration of unlabeled inhibitor required to inhibit 50% of specific ['HIPK 11 195 binding) values shown are the mean of three separate experiments with
