Rabbit skeletal muscle microsomes contain two distinct ...

Eur. J. Biochem. 172, 233-238 (1988) 0 FEBS 1988

Rabbit skeletal muscle microsomes contain two distinct phenylalkylamine-binding sites Wolfgang WERNET, Manfred SIEBER, Wolfgang LANDGRAF and Franz HOFMANN Physiologische Chemie, Medizinische Fakultat, Universitat des Saarlandes, Homburg/Saar (Received October 16,1987) - EJB 87 1159

Lu49888, a photoaffinity analog of verapamil, was used to identify specific binding sites for phenylalkylamines of calcium channels present in rabbit skeletal muscle microsomes. Direct binding equilibrium measurements and displacement curves of Lu49888 by its non-radioactive analog yielded an apparent single class of binding sites with Kd and B,,, values of 16.5 nM and 7.5 pmol/mg respectively. Lu49888 was specifically incorporated into three proteins of apparently 165 kDa, 55 kDa and 33 kDa. Incorporation into the 55-kDa protein was blocked by 10 - 50-fold higher concentrations of unlabeled phenylalkylamines compared to incorporation into the 165kDa protein, suggesting that the 165-kDa and 55-kDa proteins contain a high and a low-affinity verapamilbinding site respectively. The photoaffinity-labeled proteins were solubilized by 1% digitonin or 1% Chaps in roughly equal amounts. The 165-kDa protein bound to wheat-germ-agglutinin(WGA) - Sepharose and sedimented in sucrose density gradients with the same constant as the purified dihydropyridine receptor, which has been reconstituted to a functional calcium channel. The 55-kDa membrane protein did not bind to the WGASepharose column and sedimented in sucrose density gradients with a lower s value than the 165-kDa protein. The 165-kDa but not the 55-kDa membrane protein was specifically labeled by azidopine, the photoaffinity analogue of dihydropyridines. The 55-kDa protein of the purified dihydropyridine receptor was not significantly labeled by Lu49888 showing that the 55-kDa protein of the membrane is unrelated to the purified high-affinity dihydropyridine receptor.

channel protein has not been identified unequivocally but recent evidence suggests that the 165-kDa protein contains also the calcium-conducting unit. These interpretations are challenged by in vitro experiments which suggested that vertebrate skeletal muscle t-tubule contains a voltage sensor [12 - 141and two calcium channels [15,16]. The voltage sensor is inhibited by low concentrations of dihydropyridines and phenylalkylamines whereas rather high concentrations of phenylalkylamines, i. e. concentrations above micromolar, inhibit one of the two calcium channels [12, 13, 161. Furthermore, in vivo labeling experiments suggested that the density of dihydropyridine-binding sites exceeds the density of the calcium channel at least tenfold [17]. These experiments suggested, therefore, that skeletal muscle microsomal membranes might contain high and lowaffinity binding sites for phenylalkylamines. This possibility was strengthened by previous results which showed that various tissues contain high and low-affinCorrespondence to F. Hofmann, Physiologische Chemie, Medizinische Fakultat, Universitat des Saarlandes, D-6650 Homburg/Saar, ity binding sites for phenylalkylamines [18 - 201. Binding to the low-affinity site present in cardiac microsomal membranes Federal Republic of Germany was not affected by dihydropyridines and was apparently Abbreviations. Devapamil, (-)demethoxyverapamil, (-)D888, 2,7-dimethyl-3-(3,4-dimethoxyphenyl)-3-cyan-7-aza-9-(3-methoxyspecific for phenylalkylamines [18]. In addition, specific phenyl) nonan; (+)PN200-110, isopropyl-4-(2,1,3-benzoxadiazol- labeling of several membrane proteins by the azido analogue 4-yl) - 1,4-dihydro - 2,6 - dimethyl- 5 - methoxycarbonylpyridine - 3-car - of verapamil has been reported [9,10]. We therefore attempted boxylate; Lu47781, (f)-5-[(3-azidophenethyl)-methylamino]-2-(3,4, to identify the low-affinity verapamil-binding site in skeletal 5-trimethoxyphenyl)-2-isopropylvaleronitrile;Lu49888 is the tritiated congener of (-)-5-[(3-azidophenethyl)-methylamino]-2-(3,4,5- muscle microsomes. It is shown that this site is confined to a trimethoxyphenyl)-2-isopropylvaleronitrile; WGA, wheat germ ag- 55-kDa peptide which is unrelated to the 165-kDa high-affnity calcium-channel-blocker-binding proteins. glutinin.

The skeletal muscle t-tubule contains a high density of binding sites for phenylalkylamines and dihydropyridines [l31. Radioactive analogues of these calcium-channel blockers bind with high affinity, i.e. in the nanomolar range, to these sites. Binding to the high-affinity dihydropyridine site is regulated allosterically by binding to other sites [3]. These highaffinity sites have been purified from rabbit skeletal muscle membranes [4 - 81. The purified preparation contains three peptides of 165 kDa, 55 kDa and 32 kDa in a constant ratio [lo]. In addition a 330/28-kDa protein contaminates this preparation to a variable degree [lo]. Photoaffinity labeling experiments suggest that the 165-kDa protein contains the highaffinity, allosterically regulated binding sites for dihydropyridines, phenylalkylamines and benzothiazepines [2, 6, 8 101. The purified receptor preparation has been reconstituted also to a functional calcium channel which shows many properties of an L-type calcium channel [ll]. The nature of the

234 MATERIALS AND METHODS Materials (f)[3H]Azidopine (40 Ci/mmol) and [3H]nitrendipine (40 Ci/mmol) were purchased from Amersham. [3H]Lu49888 (83 Ci/mmol) and [3H]devapamil (83 Ci/mmol) were kindly provided by Knoll AG (Ludwigshafen). All other materials used were of the highest purity available.

kDa

Preparation of microsomal membranes and purijkation of the dihydropyridine receptor Microsomal membranes were prepared from rabbit white skeletal muscle according to [5]. Solubilization and purification of the dihydropyridine receptor were carried out as described [lo]. The purification of the photolabeled phenylalkylamine receptor was carried out by a two-step procedure which included a wheat-germ-agglutinin(WGA) - Sepharose column chromatography followed by a sucrose density gradient [lo]. Photoaffinity labeling of the calcium-channeE-blocker-bin~ingsites Microsomal membranes of rabbit skeletal muscle (1 12 pg, 8 - 12 pmol nitrendipine-binding sites/pg protein) were incubated in the dark for 90 min at 4°C in 150 pl 50 mM Tris/ HC1 buffer, pH 8.0, containing 1 mM iodoacetamide, 0.1 mM phenylmethylsulfonyl fluoride, 1 mM orthophenanthroline, 1 pM pepstatin A, 1 pg/ml leupeptin, 1 pg/ml antipain, 1 mM benzamidine and 40 - 150 nM [3H]Lu49888 in the absence or presence of indicated concentrations of devapamil. After incubation samples were exposed at 4°C for 4 min to ultraviolet light source (200 W) at a distance of 10 cm. The photolyzed samples were centrifuged. The pellets were resuspended in 1% SDS sample buffer [22] containing 2 mM dithiothreitol and were boiled for 4 min. The denatured samples (56 pg protein/slot) were separated on a 7.5% polyacrylamide gel prepared according to Laemmli [22]. Gels were stained after electrophoresis with silver or Coomassie blue. Stained or unstained, individual gel lanes were cut into 2-mm or 3-mm slices. These were dissolved in HzOz (30%) for 15 h at 45°C and thereafter counted in a Triton-X-100/ toluene-based scintillator. The following proteins were used for calibration of the gel: myosin, 200 kDa; p-galactosidase, 116 kDa; phosphorylase, 97 kDa; bovine serum albumin, 66 kDa; ovalbumin, 45 kDa and carbonic anhydrase, 31 kDa. Protein was determined according to Bradford [23]. Photoaffinity labeling of the dihydropyridine site was carried out as described for the phenylalkylamine site except that the membranes were incubated in 50 mM Tris/HCl, pH 7.5, with (f)[3H]azidopine (30 nM) in the presence or absence of unlabeled (+)PN200-110 (30 pM). The purified dihydropyridine receptor (2 pg, corresponding to 24 nM binding sites) was incubated with 100- 500 nM [3H]Lu49888 with or without a thousandfold excess of unlabeled devapamil in a total volume of 150 pl. The sample was photolyzed as described above. Thereafter 50 pl fourfold concentrated SDS sample buffer was added. The samples were further processed as described above. Concentration of compounds added for photoaffinity labeling experiments are always total concentrations added. Binding assays Binding assays with [3H]devapamil (0.5 - 1 nM), [3H]Lu49888(0.1 - 10 nM) were performed as described [24]

Relative Fraction Fig. 1. Photoaffinity labeling of the membrane-bound receptor by varying concentrations of r3H]Lu49888. Aliquots of the skeletal muscle microsomal membranes (112 pg) were incubated with (A) 40 nM [3H]Lu49888 in the absence ( 0 )and presence (0)of 20 pM devapamil and (B) in the presence of 80 nM ( 0 )and 100 nM (0) [3H]Lu49888. From each sample 56 pg protein were layered on individual gel lanes. The further proceedings were as described in Materials and Methods

in the presence of 10 pM CaC12 with and without 1 mM MgC12. Unlabeled devapamil and Lu47781 (the non-radioactive (+) congener of Lu49888) were added from 1 nM to 100 pM. Binding of dihydropyridines was carried out as described previously [24]. Binding parameters were calculated either by conventional methods or by the Ligand program 1251. RESULTS Previously it was shown that Lu49888, a photoaffinity analog of verapamil, labeled specifically the 165-kDa peptide of the purified dihydropyridine receptor [9, 101 and 165-kDa, 55-kDa and 33-kDa peptides in rabbit [lo] and guinea-pig [9] skeletal muscle membranes. Incubation of microsomal membranes with 40 nM Lu49888 (all concentrations refer to the initial concentration added) resulted in the specific incorporation of radioactivity into 165-kDa, 55-kDa and 33-kDa peptides (Fig. 1A). In addition a 98-kDa peptide was labeled in a non-saturable manner. Specific incorporation into the 165-kDa and 55-kDa peptides increased when the concentration of Lu49888 was increased to 80nM and 100nM (Fig. 1B). The total amount of radioactivity associated with these peptides increased with the addition of higher concentrations of Lu49888. The relative distribution of radioactivity between the proteins after SDS gel electrophoresis suggests that the 55-kDa peptide is labeled to a higher extent than the 165-kDa protein when the concentration of the photoaffinity analog was increased from 40 nM to 100 nM. This indicates that the 165-kDa and 55-kDa peptide bound Lu49888 with different affinity. Incorporation into both peptides was specific and was prevented by unlabeled phenylalkylamines

235 kDa

-

O 9

m

0)

P cn

7-

c

c

s

5-

3-9

-10 1-

0.3

0.6

-7

-6

[Lu 408881 free

.. I

0'

-8

0.9

Fig. 3. Binding isotherm of [3H]Lu49888 to membranes. Microsomal membranes (0.2 mg) were incubated in the dark at 4°C with 1 mM Mg2+ in the presence of either 0.1-10 nM [3H]Lu49888 (0)or in Fig. 2. [3H]Lu49888 is incorporated with different affinities into the the presence of 1 nM [3H]Lu49888 and varying concentrations of 165-kDa and the 55-kDa protein. (A) Incubation of skeletal muscle Lu47781 (a).Non-specific binding was determined in the presence microsomal membranes (56 pg) with 40 nM [3H]Lu49888.(B) As (A) of 2 pM devapamil. After 60 min membranes were precipitated by in the presence of 0.4 FM (a), 4 pM (0) and 40 pM (0) poly(ethyleneglyco1) and processed as described in Materials and (-)gallopamil Methods. Inset A shows the same data after Scatchard transformation. Inset B shows the Scatchard transformation of a binding experiment carried out in the absence of MgZ Relative Fraction

+

(Fig.2). However, whereas the labeling of the 165-kDa peptide was already suppressed by the lowest concentration of (-)gallopamil, 100-fold higher concentrations were required to block the labeling of the 55-kDa protein. Incorporation into the 55-kDa protein was inhibited half-maximally by 100 nM devapamil (not shown). About tenfold higher concentrations of (-)gallopamil were needed to reach the same inhibition as that obtained with unlabeled devapamil. In addition, a weak inhibition of incorporation was observed in the presence of the dihydropyridine (+)PN200-110 (not shown). Equilibrium binding experiments with the same membrane preparations yielded apparently a single class of binding sites with a Kd of 16.5 nM and a B,,, of 7.5 pmol/mg protein (Fig.3 inset A). These membranes contained the same density of the highaffinity nitrendipine-binding site. Using low concentrations of Lu49888 a second low-density binding site was detectable with an apparent Kd value of 2.7 nM and B,,, values between 0.5 pmol/mg and 1 pmol/mg. Since divalent cations apparently inhibit the binding of phenylalkylamines to high-affinity sites [26, 271, these experiments were repeated in the absence of added magnesium (Fig. 3 inset B). These data confirm the presence of a high and a low-affinity binding site with apparent Kd values between 0.3-0.9 nM and 10-23 nM respectively. The higher-affinity site was not detectable in regular binding assays as were those shown in Fig. 3A, presumably because of the presence of the relatively large concentration of the lower-affinity site under these conditions. Magnesium is known to inhibit binding to the high-affinity site [27]. The 55-kDa peptide was not labeled by the photoaffinity analog for dihydropyridines, (+)azidopine (Fig. 4). However, in agreement with previous results, azidopine was incorporated specifically into the 165-kDa protein suggesting that this protein is the high-affinity receptor for organic calciumchannel blockers.

kDa

a'I X

G I

S

Q

,

,

0.3

0.6

,I 0.9

Relative Fraction

Fig. 4. Photoaffinity labeling of skeletal muscle microsomal membranes (56 p g ) with (k)[3H]azidopine (30 n M ) in the absence (a) and presence (0) of 30 p M unlabeled (.t)PN200-110

These results suggested that Lu49888 binds specifically to a 55-kDa protein. It was possible that the 55-kDa peptide was identical with the 55-kDa peptide of the purified dihydropyridine receptor [5,10], although labeling of the 55-kDa peptide was not observed previously. In these experiments equal concentrations of receptor and ligand were used to ensure labeling of the high-affinity binding site. These experiments were repeated using 3.6 pmol pure dihydropyridine receptor and 15 pmol Lu49888. As is evident from Fig. 5 only the 165-kDa high affinity receptor protein was labeled under these conditions. This finding was further supported by experiments in which the microsomal membranes were first labeled with Lu49888 and then solubilized by digitonin [5, 101. The

236

5 10 15 Fraction Number

Relative Fraction

u 0.3

0

0.6

0.9

Relative Fraction

Fig. 5. Photoaffinity labeling of the puriJied skeletal muscle dihydropyridine receptor. Pure receptor ( 2 pg) was photolabeled in the presence of 100 nM [3H]Lu49888and in the absence ( 0 )and presence (0)of 25 pM devapamil

0

X

R

"

Fraction Number

L

0.3

0.6

I

0.9 0.3 Relative Fraction

I

0.6

0.9

Fig. 7. Sucrose density gradient of photolabeled skeletal muscle microsomes. Microsomal membranes ( 2 mg) were photolabeled in the presence of 130 nM [3H]Lu49888 and in the presence (0)and absence (0) of 65 pM devapamil and solubilized by digitonin. The supernatants were layered on two 5-20% sucrose density gradients (A). The distribution of radioactivity after centrifugation is shown in (A). The pellets remaining after solubilization were denatured by SDS and proteins were separated on a SDS gel (3).SDS-PAGE of aliquots of the sucrose fractions 10 (C), 15 (D), 16 (E) and 17 (F) from each gradient (0,0 )is shown in (C-F). The fractions were concentrated individually before denaturation in SDS sample buffer

Relative Fraction

Fig. 6. Identification of the high-affinity phenylalkylamine-bindingsite. Skeletal muscle microsomal membranes (20 mg) were photolabeled with [3H]Lu49888 (100 nM), solubilized and passed over a WGASepharose column (A, B). The eluted fractions were then layered on a sucrose density gradient (C, D). (A) Elution profile of a WGASepharose column (3 x 1 cm). The arrows indicate the pooled fractions: (1) break-through fraction, (2) wash I, ( 3 ) wash 11. The fractions show the elution profile of the column with N-acetylglucosamine. (B) SDS-PAGE ofaliquots from the break-through fraction 1 (0)and the N-acetylglucosamine-elutedfractions 4 (0)and 6 (0).(C) Sucrose density gradient of the pooled N-acetylglucosamine-eluted fractions 3-6. (D) SDS-PAGE of an aliquot of fraction 10 of the sucrose gradient

solubilized proteins were chromatographed on a wheat germ column. Covalent incorporated radioactivity was eluted with N-acetylglucosamine (Fig. 6 A). SDS gel electrophoresis of the retained and non-retained fractions showed that the retained fractions contained the 165-kDa protein (Fig. 6B). The fraction not retained by the column contained radioactivity associated with the 98-kDa protein and some radioactivity associated with the 55-kDa protein. The fractions containing the 165-kDa peptide were concentrated and layered on a

sucrose density gradient. The incorporated radioactivity sedimented as a single peak with an apparent szo,wvalue of 20 S (Fig. 6 C). Further analysis of the radioactive fractions from the sucrose density gradient indicated that the radioactivity was associated only with the 165-kDa protein (Fig. 6D). These results show that the 165-kDa protein labeled in the intact membrane system by Lu49888 is the high-affinity receptor for organic calcium-channel blockers. They show further that the 55-kDa protein does not copurify with this receptor. Other evidence suggest that the 55-kDa protein was not derived proteolytically from the 165-kDa protein: both proteins were present in relatively constant amounts in all membrane preparations tested; variation of the protease inhibitors did not change the relative amount of the labeling of both proteins. The 55-kDa peptide was not recovered in high concentrations after the WGA- Sepharose column raising the possibility that this protein might have been poorly solubilized. SDS gel electrophoresis of solubilized membranes showed that about 50% of both the 165-kDa and 55-kDa protein was retained in the pellet (Fig. 7B). The solubilized proteins were separated further on a sucrose density gradient. The specifity of labeling was tested by also using membranes which were

237 labeled in the presence of 0.5 pM devapamil. Specifically incorporated radioactivity sedimented at several peaks in the sucrose gradient (Fig. 7 A). Individual fractions from these two gradients were concentrated and labeled proteins were separated by SDS gel electrophoresis (Fig. 7 C - F). Fraction 10, sedimenting with an apparent szo,wvalue around 20 S contained only radioactivity incorporated specifically into the 165-kDa protein (Fig.7C). The fractions 15, 16 and 17 contained radioactivity incorporated non-specifically into the 98-kDa protein and that incorporated specifically into the 55-kDa peptide (Fig. 7 D - F). Compared with fraction 15, fraction 17 contained more radioactivity incorporated specifically into the 55-kDa peptide than non-specifically into the 98-kDa proteins. The rather broad distribution of the radioactivity associated with the 55-kDa protein suggests that it was associated with a variable amount of other proteins. An apparent sz0,+,value of 10- 12 S was estimated from this distribution.

DISCUSSION In agreement with earlier reports [9, 101 the verapamil analog Lu49888 labeled specifically two distinct membrane proteins in rabbit skeletal muscle: a 165-kDa and a 55-kDa protein. The 165-kDa protein has been identified as the highaffinity receptor for calcium-channel blockers by several criteria. (a) The molecular mass of the protein labeled in the membrane is identical with that of the purified receptor. (b) The protein prelabeled in the membrane copurifies with the calcium-channel-blocker-binding protein. (c) The 165-kDa protein is labeled also by the dihydropyridine analogue azidopine. These results confirm previous work on the molecular nature of the dihydropyridine receptor [2-6, 9, lo]. It has been shown that this protein contains all the regulatory sites known to exist on a voltage-regulated L-type calcium channel [lo]. Therefore the 165-kDa protein may be part of the calcium channel. This is supported by the recent cloning of this protein [28], which showed that this protein has remarkable sequence homologies with the sodium channel. Purified preparations of this protein have been reconstituted to a functional calcium channel [I 11. The 55-kDa protein is unrelated to the calcium antagonist receptor by several criteria. (a) This protein is not derived proteolytically from the larger 165-kDa protein during incubation. (b) It binds specifically the verapamil photoaffinity analogue, but with a lower affinity than the 165-kDa receptor protein. (c) It is not labeled by the dihydropyridine azidopine. (d) It does not copurify with the high-affinity 165-kDa receptor protein. (e) It does not comigrate with the high-affinity dihydropyridine/phenylalkylamine receptor in sucrose density gradients, but sediments around 12 s. A phenylalkylamine receptor, which sedimented with a similar constant, has been identified previously [29]. It has a lower affinity for verapamil analogues than the 165-kDa protein. The identity of this second phenylalkylamine receptor is not clear. Several studies showed [18-20, 301 that various tissues contain a low-affinity binding site for phenylalkylamines, which is not regulated allosterically by dihydropyridines. Such a low-affinity site has not been noted in skeletal muscle membranes in equilibrium binding experiments [2, 3, 5, 9, 10, 311. However, very recently Reynolds and colleagues [31] showed in kinetic experiments that skeletal muscle membranes contain a high and a low-affinity binding site for devapamil. These two sites were not detected in Scatchard

plots. In this study a higher and a lower-affinity site were detected depending on the assay condition. It is therefore quite reasonable that photoaffinity labeling of membranes combined with SDS gel separation of the labeled proteins will detect more receptor proteins than classical equilibrium binding experiments. The physiological function of this second receptor is not clear although other tissues may contain the same low-affinity receptor protein. Skeletal muscle has apparently high and low-affinity receptors for calcium-channel blockers. The highaffinity receptor has been identified as the so-called voltage sensor, which may function as a calcium channel [12-14, 17, 281. The 165-kDa high-affinity binding protein may be identical with this L-type calcium channel. In addition, skeletal muscle contains other calcium channel types [15, 161, one of which is blocked by rather high concentrations of phenylalkylamines and dihydropyridines [16]. A speculative suggestion is that the 55-kDa receptor protein is related to this channel. However, other possibilities have to be considered. As shown already for certain dihydropyridines [26, 321 multiple receptors may exist for phenylalkylamines which are unrelated to the L-type calcium channel. We thank Mrs Poesch for typing the manuscript and Mrs Siepmann for the graphical work. We also thank Dr Traut, Dr Seitz and Ing. Lietz from Knoll AG (Ludwigshafen), for providing us the photoaffinity analog Lu49888. This work was supported by grants of the Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie.

REFERENCES 1. Triggle, D. J. & Janis, R. A. (1987) Annu. Rev. Pharmacol. Toxicol. 27, 157-193. 2. Galizzi, J.-P., Borsotto, M., Barhanin, J., Fosset, M. & Lazdunski, M. (1986) J. Biol. Chem. 261, 1393-1397. 3. Glossmann, H., Ferry, D. R., Goll, A., Striessnig, J. & Zernig, G. (1985) Arzneim.-Forsch. 35, 1917-1935. 4. Curtis, B. M. & Catterall, W. A. (1984) Biochemistry 23, 21132118. 5. Flockerzi, V., Oeken, H.-J. & Hofmann, F. (1986) Eur. J . Biochem. 161,217-222. 6. Striessnig, J., Moosburger, K., Goll, A,, Ferry, D. A. & Glossmann, H. (1986) Eur. J . Biochem. 161,603-609. 7. Barhanin, J. Coppola, T., Schmid, A,, Borsotto, M. & Lazdunski, M. (1987) Eur. J . Biochem. 164, 525 - 531. 8. Leung, A. T., Imagawa, T. & Campbell, K. P. (1987) J . Biol. Chem. 262,7943 - 7946. 9. Striessnig, J., Knaus, H.-G., Grabner, M., Moosburger, K., Seitz, W., Lietz, H. & Glossmann, H. (1987) FEBS Lett. 212, 247253. 10. Sieber, M., Nastainczyk, W., Zubor, V., Wernet, W. & Hofmann, F. (1987) Eur. J . Biochem. 167, 117-122. 11. Flockerzi, V., Oeken, H.-J., Hofmann, F., Pelzer, D., CavaliC, A. & Trautwein, W. (1986) Nature (Lond.) 323, 66-68. 12. Lamb, G. D. (1986) J . Physiol. 376, 85-100. 13. Rios, E. & Brum, G. (1987) Nature (Lond.) 325, 717-720. 14. Berwe, D., Gottschalk, G. & Luttgau, H. Ch. (1987) J . Physiol. (Lond.) 385,693 - 707. 15. Cota, G. & Stefani, E. (1986) J . Physiol. 370, 151 - 163. 16. Palade, P. T. & Almers, W. (1985) Pflugers Arch. Eur. J . Physiol. 405,91-101. 17. Schwartz, L. M., McCleskey, E. W. & Almers, W. (1985) Nature (Lond.) 314,747-751. 18. Oeken, H.-J., von Nettelbladt, E., Zimmer, M., Flockerzi, V., Ruth, P. & Hofmann, F. (1986) Eur. J . Biochem. 156, 661667. 19. Reynolds, I. J., Gould, R. J. & Snyder, S. H. (1983) Eur. J . Pharmacol. 95, 319-321.

238 20. Bruns, C. & Marmt, D. (1985) Naunyn-Schmiedebergk Arch. Pharmacol. suppl. 329, R48. 22. Laemmli, U. K. (1970) Nature (Lond.) 227,680-685. 23. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. 24. Ruth, P., Flockerzi, V., Oeken, H.-J. & Hofmann, F. (1986) Eur. J . Biochem. 155,613-620. P. J. & Rodbart, D. (1980) Anal. Biochem. 107, 22025. Munson, _.. 239. 26. Ruth, P., Flockerzi, V., von Nettelbladt, E., Oeken, H.-J. & Hofmann, F. (1985) Eur. J. Biochem. 150,313-322. 27. Goll, A., Ferry, D. R. & Glossmann, H. (1984) Eur. J. Biochem. 141, 171-186.

28. Tanabe, T., Takeshima, H., Mikami, A,, Flockerzi,V., Takahashi, H., Kangawa, K., Kojima, M., Matsuo, H., Hirose, T. & Numa, S. (1987) Nature (Lond.) 328, 313-318. 29. Goll, A. & Ferry, D. (1985) Naunyn-Schmiedeberg’~Arch. Pharmacol. 329, suppl., R46. 30. Andrejauskas, E., Hertel, R. J. & Marmt, D. (1985) J . Biol. Chem. 260. 5411 -5414. 31. Reynolds, I. J., Snowman, A. M. & Snyder, S. H. (1986) J . Pharmacol. Exp. Ther. 237,131 - 738. 32. Striessnig, J., Zernig, G. & Glossmann, H. (1985) Eur. J. Biochem. 150,67 - 17.