Jul 18, 1983 - Key words: calcium channel/diltiazem/nimodipine/oligomers/ target size diltiazem, which increases the density of channels having an.
The EMBO Journal Vol.2 No.10 pp.1729- 1732, 1983
Calcium channels: evidence for oligomeric nature by target size analysis
David R.Ferry', Alexandra Gollt and Hartmut Glossmann* Rudolf Buchheim-Institut fur Pharmakologie, Justus-Liebig-Universitat, Frankfurter Strasse 107, D-6300 Giessen, FRG Communicated by R.Rott Received on 9 June 1983; revised on 18 July 1983
Radiation inactivation was employed to measure the molecular size of calcium channels in guinea-pig skeletal muscle membranes, labelled by the potent 1,4-dihydropyridine calcium antagonist [3H]nimodipine. The molecular size was decreased when the membranes were preincubated and assayed with d-cis-diltiazem, a calcium channel blocker, which is structurally unrelated to the 1,4-dihydropyridines. d-cis-Diltiazem, which is a positive heterotropic regulator of 1,4-dihydropyridine calcium channel binding in vitro, reduced the molecular size from 178 000 to 111 500. l-cisDiltiazem, the diastereoisomer, which is devoid of calcium antagonistic action, did not decrease the molecular size of the 1,4-dihydropyridine binding site. Neither diastereoisomer affected the molecular size of the membrane-bound acetylcholinesterase, indicating that a stereospecific interaction with the calcium channel structure is the basis for these observations. It is concluded that this decrease in size is indicative of the oligomeric nature of the calcium channel and that calcium channel blockers, acting via different, but interacting drug receptor sites, induce different conformations of the channel structure, resulting in altered conductivity for ions. Key words: calcium channel/diltiazem/nimodipine/oligomers/ target size Introduction Voltage-operated calcium channels which are blocked by the 1,4-dihydropyridine calcium antagonist nifedipine (Stefani and Chiriandini, 1982) and the structurally unrelated compound d-cis-diltiazem (Gonzales-Serratos et al., 1982) have been identified in skeletal muscle by electrophysiological recording techniques. Recently the calcium channels of skeletal muscle membranes have been directly labelled in the guinea-pig with [3H] nimodipine (Ferry and Glossmann, 1982a) and in the rabbit with [3H]nitrendipine (Fosset et al., 1983). Subcellular distribution studies of the [3H]nimodipine labelled calcium channels in guinea-pig skeletal muscle revealed a copurification with plasma membrane (Na +, K + -ATPase activity and [3H]ouabain binding) but not with sarcoplasmic reticulum markers (Ca2 +-ATPase activity). Electron microscopic examination of the sucrose gradient purified fraction, which was most enriched in [3H]nimodipine binding sites, revealed that it was mainly of t-tubular origin (Glossmann et al., 1983). This finding is irn accord with electrophysiological evidence for an almost exclusive t-tubular localization of skeletal muscle calcium channels (Stefani and Chiriandini, 1982). [3H]Nimodipine binding to calcium channels in guinea-pig skeletal muscle membranes is stimulated by d-cis'Part of the theses of D.R.F. and A.G. *To whom reprint requests should be sent. © IRL Press Limited, Oxford, England.
diltiazem, which increases the density of channels having an equilibrium dissociation constant (KD) of 2 nM (at 37°C) and alters the kinetics of the 1,4-dihydropyridine binding site (Ferry and Glossmann, 1982a; Glossmann et al., 1983). Similar allosteric interactions between different drug receptor sites on channel structures have been described for the 'yaminobutyric acid/benzodiazepine receptor chloride ionophore complex (Tallmann et al., 1978) and the voltageoperated sodium channel (Catterall, 1977). Using radiation inactivation as a probe we have investigated whether the allosteric regulation by d-cis-diltiazem is accompanied by structural changes of the [3H]nimodipine labelled calcium channel.
Results
[3H]Nimodipine binds with a KD of 2 nM to guinea-pig skeletal muscle membrane calcium channels at 37°C (Figure 1). The binding is stimulated by d-cis-diltiazem with a 500o effective dose (ED50) of 1 AM. The biologically inactive diastereoisomer l-cis-diltiazem (Nagao et al., 1972) does not stimulate [3H]nimodipine binding, but inhibits with a 50% inhibitory concentration (IC50) of 20-40 itM. As can be seen in Figure la the stimulation of [3H]nimodipine binding is due to an increase in the maximum density of binding sites with a KD of 2 nM for [3H]nimodipine. The 'up-regulation' of [3H]nimodipine binding is consistently seen in different membrane preparations but is not
I
~~~~~~~~~~~~~
Uh.
E51 i120 FREE
[3H] NIMODIPINE[oglo,nM] d,or I -cis DILTIAZEM[-Ilog10M
Fig. 1. Panel A: saturation isotherms of [3H]nimodipine in the presence of 10 AM d-is-diltiazem (0) and the absence of d-cis-diltiazem (0). Points are means from duplicate determinations with the non-saturable binding (which is close to the filter blank) subtracted. The data points were fitted to the general dose-response equation (DeLean et al., 1978) which gave the following parameter estimates (- asymptotic standard deviation): In the absence of d-is-diltiazem KD = 3.3 4 0.9 nM, slope factor = 0.99 i 0.1, Bma = 196 + 27 pM; in the presence of 10 1tM d-is-diltiazem KD = 2 + 0.2 nM, slope factor = 1.1 t 0.06, Bnuu = 380 4 15 pM. The protein concentration in both experiments was 48 Ag/ml. Panel B: dose-effect curves of d-cis-diltiazem (-) and 1-cis-diltiazem (O). Each point is the mean value from two or three experiments which varied by 10Go. In these experiments the protein concentration was 30-40 /Ag/ml and the [3H]nimodipine concentration 0.8-1.5 nM. Bo is the calcium channel-bound [3H]nimodipine in the absence of added drug and B the calcium channelbound [3H]nimodipine in the presence of drug. The following parameters were calculated by fitting the data points to the general dose response equation (i asymptotic standard deviation): d-cis-diltiazem ED50 = I1 0.3 jtM, slope factor = 0.8 : 0.2; 1-cis-diltiazem IC50 = 24 k 8 yM, slope factor = 0.6 0.1.
1729
D.R.Ferry, A.Goll and H.Glossmann
10
IC)~~~~~~~~~~
-05
-~2F
x
0
0 Z
~ ~~~~~~1020
30
I
C
02 4 6 8 31
3
15
25
DOSE [M rad] Fig. 2. Radiation-inactivation of the [3H]nimodipine-labelled calcium channel. A: radiation-inactivation of the [3H]nimodipine-labelled calcium channel pre-incubated and assayed in 10 LM d-s-diltiazem: (-), pre-incubated and assayed in 10 AM I-s-diltiazem: (A), pre-incubated and assayed in buffer A (0). The membrane samples that have received the doses of radiation as indicated, were assayed with 1.2 nM [3H]nimodipine at a protein concentration of 20-35 Ag/ml. The data points are fitted to the equation y = y = ae- bx where y is bound ligand at the given dose 'x' of radiation, a is the y-axis intercept and b is the constant of decay (Mrad -1). For all three sets of data the correlation coefficient is 2 0.94. Lines are best fits by the least squares principle. Insert B: as in A but plotted as the natural logarithm of binding at a given dose of irradiation; (A) binding in nonirradiated controls; (A.), against the dose of irradiation. The D37 - values doses of the [3H]nimodipine-labelled calcium channel are: for preincubation with buffer A, 10.9 Mrad; for pre-incubation with 10 ItM l-cisdiltiazem, 10.8 Mrad; and for pre-incubation with 10 AM d-cis-diltiazem, 16.8 Mrad. Insert C: plot of equilibrium dissociation constants of [3H]nimodipine (nM) for residual calcium channels versus dose of radiation. In the presence of 10 AM d-cis-diltiazem (U); in the absence of d-i-diltiazem
(0).
observed when binding assays are performed at 4°C (data not shown). This suggests that at 37°C the [3H]nimodipine binding component of the calcium channel exists in two different states. One state has a very low affinity for [3H]nimodipine and can be only detected by radioligand binding experiments in the presence of a positive heterotropic allosteric regulator such as d-cis-diltiazem. The second state has a KD of 2 nM for [3H]nimodipine and is observed in the absence of the regulator. The highest [3H]nimodipine concentration in the saturation experiment shown in Figure 1 was 10.5 nM. As no deviation from a monophasic saturation isotherm to a homogeneous, non-interacting population of binding sites is observed the KD of [3H]nimodipine for the low affinity state of the channel must be 2100 nM. Figure 2a shows the decay of calcium channel binding of [3H]nimodipine for three pre-incubation conditions of the membranes as a function of radiation dosage; the linear transformations are shown in Figure 2b. Under all three preincubation conditions the decay of the [3H]nimodipine labelled calcium channel is monoexponential, demonstrating the homogeneity of target size. When membranes were pre1730
20
30
DOSE [M rad]
25
0 1 2 345678910
10
Fig. 3. Radiation inactivation of acetylcholinesterase activity. An example of a single, typical experiment is shown. Membranes were pre-incubated in buffer A (0), in 10 zM d-cis-diltiazem (O), or in 10 M l-cis-diltiazem (A). The ln of (activity at a given dose of irradiation)/(activity in nonirradiated controls) is plotted against radiation dose. In this experiment the D37s of acetylcholinesterase in membranes pre-incubated in buffer were 25.60 Mrad, pre-incubated with 10 ILM d-is-diltiazem 25.25 Mrad and preincubated with l-cis-diltiazem 25.99 Mrad. The line is the best fit to the pooled data points for the three pre-incubation conditions. Table I. Pre-incubation condition Buffer d-cis-diltiazem
1-cis-diltiazem
[3H]Nimo- D37Mrad 16.1 + 0.95 10.1 0.4 11.3 L 0.6 mol.wt. 178000 ± 7200 111500 ± 7200a 158000 + 8000 dipine labelled calcium channel AcetylD37Mrad 27.4 + 1.9 30.9 + 2.8 26.0 +1.2 choline mol.wt. 65500 4500 58000 5400 68900 4 3200 esterase
D37-values and corresponding relative mol. wts. of the [3H]nimodipine-
labelled calcium channels and acetylcholinesterase. All numbers are arithmetic means, the standard error of the mean for three independent irradiations; aindicates p 0.003 with respect to control (buffer preincubated).
incubated in buffer, the dose of radiation reducing biological activity to 37% of the starting value (D37) of the [3H]nimodipine labelled calcium channels was 10.1 Mrad, corresponding to a mol. wt. of 178 000. 1-cis-Diltiazem pre-incubation did not signficantly alter the D37 of the [3H]nimodipine labelled calcium channel, but d-cis-diltiazem pre-incubation increased the D37 by 6 Mrad corresponding to loss in mol. wt. of 67 000 mass units. Irradiation of membranes at doses of radiation of up to 25 Mrad did not alter the KD of [3H]nimodipine for the residual saturable binding sites (Figure 2c). In contrast to the mol. wt. regulation of the [3H]nimodipine-labelled calcium channel by d-cis-diltiazem, the mol. wt. of the endogeneous acetylcholinesterase was identical for the three different pre-incubation conditions (see Figure 3 and Table I).
Mol. wt. of calcium antagonist binding sites on the Ca2 I-chaanen
Discussion The major conclusion of this study is that the [3H]nimodipine-labelled calcium channel is a large structure composed of homologous or heterologous subunits. In the absence of the use-dependent calcium channel blocker d-cisdiltiazem (Hirth and Ehren, 1982), the in situ mol. wt. value is 178 000, which is 82 000 mass units smaller than the [3H]ethylene diamine tetrodotoxin labelled sodium channel of rat brain (Barhanin et al., 1983). The mol. wt. of the [3H]nimodipine-labelled calcium channel pre-incubated in buffer is close to the value reported for the [3H]nitrendipine-labelled calcium channel of rabbit skeletal muscle when irradiated in the lyophilized state (Norman et al., 1983). In the experiments described in this paper, d-cis-diltiazem treatment of the membranes decreases the in situ mol. wt. by 67 000 mass units, but 1-cis-diltiazem was inactive in this respect. The same reduction in calcium channel mol. wt. by d-cis-diltiazem pre-incubation has been reported for [3H]PN 200-110 (isopropyl 4-2,3,1 2,6-dimethyl-5-methoxybenzoxadiazol-4-yl)-1,4-dihydro carbonyl-pyridine 3-carboxylate) labelled calcium channels in guinea-pig skeletal muscle membranes (Goll et al., 1983). The experiments with l-cis-diltiazem serve as a dual control. First, they demonstrate that only the biologically active diastereoisomer, d-cis-diltiazem (Nagao et al., 1972) can regulate the calcium channel in situ structure; second, the mol. wt. regulation by d-cis-diltiazem cannot be due to non-specific artifacts such as 'radical capture'. The mol. wt. of acetylcholinesterase was identical under the three pre-incubation conditions of the membranes, further underlining the specificity of the d-cisdiltiazem regulation of the [3H]nimodipine-labelled calcium channel. In addition, the mol. wt. of the enzyme is in very good agreement with the data from other laboratories (e.g., Doble and Iversen, 1982) underlining the validity of our dosimetry. The phenomenon of channel mol. wt. regulation by drugs has previously been reported for the [3H]flunitrazepamlabelled chloride channel/benzodiazepine receptor ionophore of rat brain (Doble and Iversen, 1982). In this system, yaminobutyric acid acts as positive allosteric heterotropic regulator of [3H]diazepam binding to the chloride channel/ benzodiazepine receptor complex (Tallmann et al., 1978) (in an analogous manner to the action of d-cis-diltiazem on [3H]nimodipine binding to the calcium channel). Pre-incubation of rat brain membranes with 'y-aminobutyric acid reduces the mol. wt. value of the chloride channel/benzodiazepine receptor ionophore from 89 000 to 63 000 (Doble and Iversen, 1982). As a hypothesis, we would like to suggest that d-cisdiltiazem causes the dissociation of a subunit of the calcium channel with a mol. wt. of 67 000. The [3H]nimodipine binding component with a mol. wt. of 111 000 can then bind [3H]nimodipine with high affinity. The hypothetical states of the channel may be represented by
Ch° S=Ch' S d-cis-diltiazem. Ch" + S where Ch' is a channel with KD Z 2 nM for [3H]nimodipine, Cho is a channel with KD =1 /M for [3H]nimodipine, and S is a subunit (mol. wt. = 67 000) dissociable by d-cis-diltiazem. In the absence of d-cis-diltiazem and in native membranes, Ch'.S and Cho.S are in equilibrium at 370C with -30% in the Ch'.S state and 70'0/ in the Ch0.S state. The mol. wt. of Ch'.S is 178 000; Ch°.S cannot be easily assayed with direct binding studies at 37°C because its affinity for [3H]-
nimodipine is too low. Addition of d-cis-diltiazem converts all channels into the Ch" state (mol. wt. = 111 000) with high affinity for [3H]nimodipine. The Ch' .S and Ch" states behave differently with respect to [3H] nimodipine dissociation kinetics, i.e., [3H]nimodipine-Ch' .S complexes dissociate 3-fold faster than [3H]nimodipine-Ch" complexes and are distinguished by their different interactions with (-)D-600 (Ferry and Glossmann, 1982a). Our study cannot answer the question if the [3H]nimodipine and d-cis-diltiazem drug receptors within the calcium channel are located on the same channel component. The availability of a high affinity label for the d-cis-diltiazem drug receptor site would allow this question to be directly addressed. The demonstrated biophysical states of calcium channels in situ may be complementary to the existing electrophysiological data, where calcium channels have been found in 'resting', open' and 'inactivated' states (Reuter, 1983). Although currently we cannot correlate the channel states identified in situ with those of electrophysiological studies, future work may lead to this goal. Our findings support the concept of different, but interacting drug receptor sites within the calcium channel (Glossmann et al., 1982; Ferry and Glossmann, 1982b) and may provide clues how different calcium channel drugs alter the conductivity of the ionic pore. Materials and methods Materials
[3H]Nimodipine (158 Ci/mmol) and unlabelled nimodipine were a gift from Bayer AG, Wuppertal, FRG. d- and l-cis-diltiazem were provided by the Godecke AG, Freiburg, FRG. Radiochromic dye films Qot no FWT-60-00) were from Far-West Technology Inc. Goleta, CA. Preparation of guinea-pig skeletal muscle membranes and irradiation of membranes Guinea-pig hind-limb skeletal muscle membranes were prepared as described by Ferry and Glossmann (1982a) and suspended in ice-cold 50 mM TrisHCI buffer (pH 7.4) supplemented with 0.1 mM phenylmethylsulfonylfluoride (Buffer A) at a protein concentration of 2-3 mg/ml. The membranes were then pretreated for 30 min at 37°C in either buffer A or in buffer A supplemented with 10 ytM d-cis-diltiazem or 10 AM l-cis-diltiazem. The pretreated membranes were flash-frozen in volumes of 2 ml in polypropylene vials by immersion in liquid nitrogen. The frozen membranes were then irradiated with high energy electrons (10 MeV) from a linear accelerator (CSF Thompson, Paris, France) situated at the Strahlenzentrum of the Justus-LiebigUniversitat in Giessen. 10-20 samples were irradiated simultaneously by a focused electron beam at dose rates of 1.5 - 2.5 Mrad/min in a sample holder which was circulated with liquid nitrogen to automatically maintain a temperature of - 1 10°C (Lubbecke et al., 1983; Ferry et al., 1983). Typically, the radiation intensity at the beam centre was 5- to 7-fold higher than at the beam periphery. For each sample position in the sample holder, doses were recorded with radiochromic dye films (Lubbecke et al., 1983; Kempner and Haigler, 1982). Control samples were stored and handled identically to irradiated samples until assay. The mol. wts. were calculated (Kepner and Macey, 1968) according to: Mol. wt. = fD4M l05 D37Mrad The temperature correction factorf is 2.8 at - 1 10°C (Kempner and Haigler, 1982), and D37 is the dose of radiation reducing biological activity to 37°lo of the starting value. Assay for calcium channel and acetylcholinesterase activity Control and irradiated membranes were rapidly thawed and diluted 20-fold with ice-cold buffer A. The assays for calcium channels were performed in a volume of 0.25 ml in buffer A (5-25 jg of membrane protein) or for diltiazem-pretreated membranes in buffer A with the corresponding diastereoisomer at a concentration of 10 1tM at 37°C with [3H]nimodipine under equilibrium binding conditions (30 min incubation) prior to collection of membrane-bound [3H]ligand on Whatman GF/C filters after rapid dilution
1731
D.R.Ferry, A.Goll and H.Glossmann and filtration with ice-cold buffer (Ferry and Glossmann, 1982a, 1982b). Binding of [3H]nimodipine to saturable sites was defined by inclusion of 1 /M unlabelled nimodipine. Saturation isotherms for [3H]nimodipine and the dose-response curves for d- and l-cis-diltiazem were analysed by iterative nonlinear curve fitting to the general dose-response relationship (DeLean et al., 1978). As internal standard for the radiation inactivation technique and the dosimetry, the activity of acetylcholinesterase was determined according to Ellman et al. (1961), with a Beckman DU 8 spectrophotometer and the Beckman Kinetic I or II program over 10 min in a final assay volume of 1.1 ml. Three different dilutions of the membranes were employed for each sample to test for linearity with respect to time and protein content.
Acknowledgements We are grateful for [3H]nimodipine and unlabelled nimodipine which were provided by the Bayer A.G., Wuppertal, FRG; d- and l-cis-Diltiazem were donated by Godecke A.G., Freiburg, FRG. Drs. E.-L.Sattler and G.Doell of the Strahlenzentrum of the Justus-Liebig Universitat of Giessen kindly operated the linear accelerator and painstakingly performed the dosimetry. B.Haberman, C.Auriga, I.Seidel and M.Rombusch provided excellent technical assistance. Research of the authors is supported by the Deutsche Forschungsgemeinschaft.
References Barhanin,J., Schmid,A., Lombert,A., Wheeler,K.P., Lazdunski,M. and Ellory,J.C. (1983) J. Biol. Chem., 258, 700-702. Catterall,W.A. (1977) J. Biol. Chem., 252, 8660-8668. DeLean,A., Munson,P.J. and Rodbard,D. (1978) Am. J. Physiol., 4, E97E102. Doble,A. and Iversen,L.L. (1982) Nature, 295, 522-523. Ellman,G., Courtney,D., Andres,V. and Featherstone,R. (1961) Biochem. Pharmacol., 7, 88-93. Ferry,D.R. and Glossmann,H. (1982a) FEBS Lett., 148, 331-337. Ferry,D.R. and Glossmann,H. (1982b) Naunyn-Schmiedeberg's Arch. Pharmacol., 321, 80-83. Ferry,D.R., Goll,A. and Glossmann,H. (1983) Naunyn-Schmiedeberg's Arch. Pharmacol., 323, 292-297. Fosset,M., Jaimovich,E., Delpont,E. and Lazdunski,M. (1983) Eur. J. Pharmacol., 86, 141-142. Glossmann,H., Ferry,D.R., Lubbecke,F., Mewes,R. and Hoffmann,F. (1982) Trends Pharmacol. Sci., 3, 431-437. Glossmann,H., Ferry,D.R. and Boschek,C.B. (1983) Naunyn-Schmiedeberg's Arch. Pharmacol., 323, 1-11. Goll,A., Ferry,D.R. and Glossmann,H. (1983) FEBS Lett., 157, 63-69. Gonzales-Serratos,H., Valle-Aguilera,R., Lathrop,D.A. and Garcia,M.del C. (1982) Nature, 298, 292-294. Hirth,C. and Ehren,N. (1982) Naunyn-Schmiedeberg's Arch. Pharmacol., 319, 139. Kempner,E.S. and Haigler,H.T. (1982) J. Biol. Chem., 257, 13297-13299. Kepner,G.R. and Macey,R.I. (1968) Biochim. Biophys. Acta, 163, 188-203. Lubbecke,F., Ferry,D.R., Glossmann,H., Sattler,E.-L. and Doell,G. (1983) Naunyn-Schmiedeberg's Arch. Pharmacol., 323, 96-100. Nagao,T., Sato,M., Iwasawa,Y., Takada,T., Ischida,R., Nakajima,H. and Kiyomoto,A. (1972) Japan J. Pharmacol., 22, 467-478. Norman,R.I., Borsotto,M., Fosset,M. and Lazdunski,M. (1983) Biochem. Biophys. Res. Commun., 111, 878-883. Reuter,H. (1983) Nature, 301, 569-574. Stefani,E. and Chiriandini,D.J. (1982) Annu. Rev. Physiology, 44, 357-372. Tallmann,J.F., Thomas,J.W. and Gallager,D.W. (1978) Nature, 274, 383-385.
1732
