Structural and Functional Characterization of Inositol 1,4,5 ...

Vol. 266, No. 2, Issue of January 15, pp. 1109-1116,1991 Printed in U. S.A .

THEJOURNAL OF BIOLOGICAL CHEMISTRY 01991 by The American Society for Biochemistry and Molecular Biology, Inc.

Structural and Functional Characterization of Inositol 1,4,5-Trisphosphate Receptor Channel fromMouse Cerebellum* (Received for publication, July 2, 1990)

Nobuaki MaedaS, Takashi Kawasaki& Shinji Nakadez, Nobutaka Yokotag, Takahisa TaguchiQ, Michiki KasaiQ, and Katsuhiko MikoshibaSV 11 From the $Divisionof Regulation of Macromolecular Function, Institutefor Protein Research, Osaka University, Osaka565, the §Department of Biophysical Engineering, Faculty of Engineering Science, Osaka university, Osaka560, and the 7lNational Institute for B&c Biology, Okazaki 444, Japan

The cerebellar inositol 1,4,5-trisphosphate(InsP3) was found to be highly enriched in Purkinje cells (2, 3). The receptor is a high molecular weight glycoprotein abun- cerebellar InsP3 receptor protein is a high molecular weight dantly expressed in Purkinje cells. The subunit struc- glycoprotein, which has been referred to as the Pdo0 protein ture of the InsP3 receptor protein was examined by (3-9). This protein is phosphorylated by CAMP-dependent cross-linking experiments. Agarose-polyacrylamide proteinkinaseand by Ca2+/calmodulin-dependent protein gel electrophoresis of the cross-linked materials dem- kinase I1 (4,5, 10). The binding specificity of various inositol onstrated that the cerebellar InsPs receptor protein is phosphates to this protein is comparable to that in InsP3composed of four noncovalently bound identical sub- sensitive calciummobilization (3).Three groupshave reunits each with a M, of 320,000 in both purified and microsome-bound states. Chromatography of the pu- ported the immunohistochemicallocalization of InsPBreceprified receptor on a calmodulin-Sepharose column tor in the Purkinje cells (6, 11-13). Although there are some demonstrated a Ca2+-dependent interaction of the discrepancies amongthesereports,itiscertainthatthis InsP3 receptor with calmodulin. Photoaffinity labeling protein is accumulated in smooth endoplasmic reticulum in of of the cerebellar microsomal fraction with [ ~ t - ~ ~ P ] 8the - dendrites, cell bodies, axons,andsynapticboutons azidoadenosine 5’-triphosphate revealed the presence Purkinje cells (6,ll-13). Although the InsP3receptor protein of ATP-binding site in the InsP3 receptor. Scatchard is highly expressed in the cerebellar Purkinje cells, immunoanalysis of the purified InsP3 receptor revealed the blot analysis withmonoclonal antibodies against thecerebelB,,, and K , values for ATP binding of 2.3 pmol/pg and lar InsP3 receptor and in situ hybridization indicated that this 17 p ~ respectively. , Reconstitution of the purified protein is distributed ubiquitously in various tissues such as InsPs receptor into the planar lipid bilayer indicated in the cerebrum, spinal cord, thymus, spleen, liver, heart, and channel activity in the purified receptor. It exhibited smooth muscles (7, 14). Furthermore, Ferris et al. (15) has a calcium conductance (26 pS in 53 mM Ca2+) and reported that the cerebellar InsP3 receptor protein itself is sodium conductance (21 pS in 100-500 mM asymmet- responsibleforcalciumfluxin reconstituted lipid vesicles. ric Na+ solutions) with permeability ratios of P ~ J P TThese ~ ~ facts ~ suggest that the InsP3 receptor protein purified = 6.3 and P N J P ~=, 5.4. The purified channel was from the cerebellum is an InsP3-sensitive calcium channel activated with submillimolar ATP in the presence of InsP3 and modified to reach a large conductance state. present in various types of cells. Recently, we succeeded in the cDNA cloning of cerebellar InsP3 receptor protein (16) and also in the functional expression of the receptor in L cells by transfecting a plasmid vector containing this cDNA (17). Analysis of the deduced amino acid sequence of this protein Inositol 1,4,5-trisphosphate (InsP3)’ is a well known second revealed that it is comprised of 2749 amino acids and that messenger, which causes the liberation of calcium ions from there are several membrane spanning domains near the C intracellular storage sites (1).The calcium release has been terminus (16). In addition, it has been found that the cerethought to be mediated by a calcium channel directly coupled bellar InsP3 receptor protein has some sequence homology t o the InsP3-specificreceptor. Recently, an InsP3 receptor with the skeletalmuscle ryanodine receptor protein(16). The protein was purified from mouse and rat cerebella, where it skeletal muscle ryanodine receptor is a calcium release chan* This study was supported by grants from Special Coordination nel in themuscle sarcoplasmic reticulum and isconsidered to Funds of the Science and Technology Agency of the Japanese Gov- be involved in excitation-contraction coupling (18, 19). It is ernment and from the Japanese Ministry of Education, Science, and important to note that thesequence homology is focused on Culture. The costs of publication of this article were defrayed in part part of the putative membrane-spanning domain and is not by the payment of page charges. This article must therefore be hereby so remarkable in other regions (16, 20). In various types of marked “advertisement” in accordance with 18 U.S.C. Section 1734 cell, the intracellular calcium concentration often oscillates solely to indicate this fact. 11 To whom correspondence should be addressed Div. of Regulation and insome cases, thecalcium signals are initiateda t discrete the of waves. These phenomof Macromolecular Function, Institute for Protein Research, Osaka sites and then propagate in form ena are best explained by the interactions between InsP3University, 3-2 Yamadaoka, Suita, Osaka 565, Japan. The abbreviations used are: InsP3, inositol 1,4,5-trisphosphate; induced Ca2+release and Ca2’-induced Ca2+ release mechaPMSF, phenylmethylsulfonyl fluoride; EGTA, [ethylenebis(oxyethyl- nisms (1).Ryanodine-sensitive calcium channels seem to be enenitriloltetraacetic acid; CHAPS, 3-[(3-cholamidopropyI)diresponsible for the Ca2+-inducedcalcium release (21, 22). In methylammonio]-1-propanesulfonicacid; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; PEG, polyethylene fact, a relativelylarge amount of high affinity ryanodine glycol; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; binding sites has been found in rat brain microsomes (23). AMP-PCP, P,r-methyleneadenosine 5’-triphosphate. Therefore, it is quite important to characterize an isolated

1109

1110

Characterization of the Cerebellar Imp3Receptor Channel

polyacrylamide and 0.5% agarose slab gel was carried out by the method of Kiehm et al. (24). The gels were subjected to immunoblotting as described elsewhere (7) or were stained in 0.05% Coomassie Brilliant Blue R-250 in acetic acid/isopropanol/water (1:4:10) overnight and destained in 7% acetic acid. Photoaffinity Labeling-Microsomal fraction was washed oncewith the solution containing 0.15 M NaCl, 1 mM EDTA, 0.1 mM 2mercaptoethanol, and 50 mM Tris-HCI, pH 7.4, and then resuspended in the same buffer at theprotein concentration of 20 mg/ml. Aliquots (25 pl) were mixed with the same volume of the solution containing 5"triphosphate (3 pCi), 1 mM EDTA, 20 p~ [~~-~~P]8-azidoadenosine 0.1 mM 2-mercaptoethanol, 0.15 M NaCl, 5 mM MgCl,, and 50 mM Tris-HC1, pH 7.4. The solutions were incubated for 1 min at 0 "C and then irradiated for 5 min with a UVM57 lamp at 5-cm distance at 0 "c.300 p1 of the solutions containing 10 mM dithiothreitol, 0.5 M NaCl, 1 mM EDTA, and 50 mM Tris-HC1, pH 7.4 were added and the samples were centrifuged. The resultant pellets were solubilized in EXPERIMENTALPROCEDURES 2% SDS, 5% 2-mercaptoethanol, 10% glycerol,0.0625 M Tris-HC1, Materials-InsP3 was obtained from Dojindo Lab. (Japan) and pH 6.8, and then were boiled for 3 min. The samples were subjected ['H]InsP3 (17.0 Ci/mmol) was obtained from Du Pont-New England to immunoprecipitation or to the 7.5% SDS-PAGE in the buffer Nuclear. [ C Y - ~ ~ P J (>400 A T P Ci/mmol) was obtained from Amersham system of Laemmli (25). Corp. and [ol-32P]8-azidoadenosine5"triphosphate was obtained from Immunoprecipitation-20 p1 of the samples were diluted with 1 ml ICN Biomedicals Inc. Bis(sulfosuccinimidyl)suberate,ethylene gly- of 0.1% bovine serum albumin, 0.15 M NaC1, 5 mM EDTA, 5 mM N colbis(succinimidylsuccinate), ethylene glycolbis(sulfosuccinimidy1- ethylmaleimide, 0.1 mM phenylmethylsulfonyl fluoride, 10 p~ pepsuccinate), disuccinimidyltartarate, disulfosuccinimidyltartarate, N - statin A, 0.2% Triton X-100, 10 mM sodium phosphate, pH 7.2. 10 pl succinimidyl(4-azidophenyl)-1,3'-dithiopropionate were obtained of the 18AlO monoclonal antibody solutions (concentration of antifromPierce Chemical Co. Calmodulin-Sepharose was obtained from body is about 100 pg/ml) wereadded to thesamples, which were then Pharmacia LKB Biotechnology Inc. Rabbit anti-rat IgG antibody incubated for 1 h a t 4 "C. 10 pl of the rabbit anti-rat IgG (Fc regionwas purchased from Jackson Immunoresearch Lab. Phosphatidyleth- specific) was added and the solutions were incubated for 1 h a t 4 "C. anolamine and phosphatidylserine were purchased from Sigma, All 100 pl of the 10% (w/v) suspension of Pansorbin (Calbiochem) was of the other chemicals were of the highest purity commercially avail- added to the sample. The sample tubes were gently rotated for 30 min a t 4 "C and then the Pansorbin particles were washedthree times able. Preparation of Membranes-Adult ddY mice were anaesthetized with 1 ml of 0.01% bovine serum albumin, 0.5% Triton X-100, 0.15 and thenkilled by decapitation and the cerebella were dissected. The M NaC1,10 mM sodium phosphate, pH 7.2. The Pansorbin pellets tissues were mixed with 9 volumes of the solution containing 0.32 M weremixed with 20 pl of 4% SDS, 10% 2-mercaptoethanol, 20% glycerol, 0.125M Tris-HC1, pH 6.8, and thenwere heated in a boiling sucrose, 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride, 10 p~ leupeptin, 10 p M pepstatin A, 1 mM 2-mercaptoethanol, and 5 mM water bath for 3 min. After centrifugation, the supernatants were Tris-HC1, pH 7.4, and were homogenized in a glass-Teflon Potter subjected to the 7.5% SDS-PAGE in the buffer system of Laemmli homogenizer with 10 strokes at 850 rpm. The suspension was centri- (25). The gel was dried and exposed to Kodak x-ray film at -80 "C. Measurement of [ C Y - ~ ~ P I A Binding-Measurement TP of ATP bindfuged at 12,000 X g for 15 min at 2 "C, and the pellet was washed again under the same conditions. The combined supernatants were ing was performed with PEG precipitation method. 100 p1 of the centrifuged at 105,000 X g for 60 min at 2 "C to precipitate the samples containing 2 pgof InsP3 receptor protein, 0.15 M NaC1, 1 microsomal fraction. The microsomal fraction was resuspended in mM EDTA, 1 mM 2-mercaptoethanol, and 50 mM Tris-HC1, pH 7.4, the buffer containing 0.1 M sucrose, 1 mM EGTA, 0.2 mM 2-mercap- were mixed with 2 pl of 25 p M [a-32P]ATP(1 pCi/tube) and were toethanol, and 10 mM Tris-HC1, pH 7.4, and thealiquots were frozen incubated for 10 min at 0 "C. 2 pl of 50 mg/ml human IgG and 100 pl of 30% PEG 6000, 0.15 M NaCl, 1 mM 2-mercaptoethanol, 50 mM in liquid N, and then stored a t -80 "C. Purification of ZnsP3 ReceptorProtein-InsP3 receptor protein was Tris-HC1, pH 7.4, were added to the samples, which were incubated purified from ddY mousecerebella according to themethod described for 5 min at 0 "C. Nonspecific binding was measured in the presence elsewhere (3) with minor modifications. The modified point is that of 1 mM cold ATP. After centrifugation at 10,000 X g for 5 min, the 10% glycerol wasadded to all of the solutions. And for the preparation precipitates were washed with 200 p1 of 15% PEG 6000,0.15 M NaCl, of the samples used for planar bilayer measurement, the buffer for 1 mM 2-mercaptoethanol, 0.5 mM EDTA, 50 mM Tris-HC1, pH 7.4. hydroxylapatite chromatography was replaced by 1%CHAPS, 0.5 M After brief centrifugation, the precipitates were dissolved in Protosol NaCI, 1 mM 2-mercaptoethanol, 25 or 150 mM sodium phosphate, pH (Du Pont-New England Nuclear) and the radioactivities were measured with a liquid scintillation counter. 8.0, to remove Triton X-100. The samples were frozen in liquid N, Measurement of fHIZnsP3 Binding-['H]InsPn binding assay was and then stored a t -80 "C. Cross-linking of Microsomal Fractions-The microsomal fraction performed with PEG precipitation method as described elsewhere (3) (final 12 nM) was added to the sample was washed with 50 mM sodium phosphate, pH 8.0, three times and with modification. [3H]In~P3 then resuspended in the same buffer at a protein concentration of 4 (50 pl) containing 0.6 pg of InsPs receptor, 0.15 M NaC1,l mM EDTA, mg/ml. Aliquots (0.2 ml) of the suspension were treated with various 1 mM 2-mercaptoethanol, and 50 mM Tris-HCI, pH 7.4. The sample concentrations of cross-linkers. The reactions were allowedto proceed was incubated for 10 min at 0 "C and then mixed with 2 pl of 50 mg/ for 30 min at 0 "C and were then quenched by addition of 40 pl of 1 ml human IgG. 50 pl of the solution containing 30% PEG 6000, 0.15 M NaCl, 1 mM 2-mercaptoethanol, 50 mM Tris-HC1, pH 7.4, were M ammonium acetate. After centrifugation, the membrane pellets were solubilizedin 1%SDS, 1 mM EDTA, 5% 2-mercaptoethanol, 10 added to thesample, which was incubated for 5 min a t 0 "C. NonspemM Tris-HC1, pH 8.0, 10% glycerol (agarose-PAGE sample buffer) cific binding was measured in the presence of 10 p M cold InsPB. After and were heated in a boiling water bath for 3 min. The samples were centrifugation at 10,000 X g for 5 min, the precipitate was dissolved in Protosol and the radioactivity was measured with a liquid scintilanalyzed with agarose-PAGE and immunoblotting. Cross-linkingof Purified ZmP3Receptor-The purified InsP3 recep- lation counter. MonoclonalAntibody-A monoclonal antibody against InsPs receptor was dialyzed against 0.2% Triton X-100, 1mM 2-mercaptoethanol, 5 mM sodium phosphate, pH 8.0. Aliquots (50 pl) of the solution were tor, 18A10 was prepared as described elsewhere (7). The hybridoma treated with various concentrations of bis(sulfosuccinimidy1)suberate cells were cultured in Nissui SFMlOl medium. The culture supernaat theprotein concentration of 0.1 mg/ml. The reactions were allowed tant was concentrated with an Amicon YM-10 membrane. The conto proceed for 30 min at room temperature and then were quenched centrated culture supernatant was dialyzed against phosphate-buffby the addition of 10 pl of 50 mM Tris-HC1, pH 8.0. The same volume ered saline and thenapplied to a Bio-Gel HPHT column equilibrated of the solutions containing 2% SDS,2 mM EDTA, 10% 2-mercapto- in 10 mM sodium phosphate, pH 6.8. The proteins were eluted with ethanol, 20 mM Tris-HC1, pH 8.0, and 20% glycerol were added to linear gradient of sodium phosphate (10-250 mM, pH 6.8). IgG the samples. The solutions were heated in a boiling water bath for 3 fractions were collected and concentrated with an Amicon YM-10 membrane. After dialysis against phosphate-buffered saline, the anmin and were subjected to agarose-PAGE. Agarose-PAGE and Immunoblotting-Electrophoresis in a 1.75% tibody solution was stored at -80 "C.

InsP3 receptor channel and compare it with the ryanodine receptor channel in order to elucidate the calcium signaling mechanism. In this study, we investigated the structural aspects of the cerebellar InsP3 receptor protein and furtherperformed single channel recordings of this protein by planar lipid bilayer experiment. We found that the cerebellar InsPB receptor channel (i) has a homotetrameric subunit structure, (ii)binds adenine nucleotides and calmodulin, (iii) conducts calcium and sodium ions, and that (iv) the channel subconductance state is regulated by ATP in the presence of InsP3. These findings reveal remarkable similarities in the structural and functional aspects of InsP3 andryanodine receptor channels.


1111

Calmodulin-Sepharose Chromatography-Columns containing 350 p1 of calmodulin-Sepharose 4B or Sepharose 4B were equilibrated in a buffer containing 0.2% Triton X-100.0.2 M NaCI, 1 mM 2-mercap-

toethanol, 1.5 mM CaC12and 20 mM Tris-HCI, pH 7.4. 300 pl of the samples were applied and the columns were washed with the same buffer. Bound proteins were then eluted with a buffer containing 0.2% Triton X-100, 0.2 M NaCI, 1 mM 2-mercaptoethanol, 2 mM EGTA, and 20 mM Tris-HCI, pH 7.4. Planar Lipid Bilayer Measurements-The purified cerebellar InsP3 receptor protein was incorporated into Muellar-Rudin planar bilayer membrane. Planar lipid bilayer was formed from a mixture of phosphatidylethanolamine/phosphatidylserine in a 2:l weight ratio (20 mg/ml in decane). Observation of the CaZ+current was performed according to the method of Smith et al. (19) with modification (26). 2-3pgof the cerebellar InsPa receptor protein was added to one chamber, designated cis, composed of 1 mM CaCI2, 2 mM EGTA, 0.125 M Tris, 0.25 M Hepes, pH 7.4. The other chamber, designated trans, contained 53 mM Ca(OH)2,0.25 M Hepes, pH 7.4. When Na+ current was observed, the cis chamber contained 0.1 M NaCI, 1 mM CaCI2,2 mM EGTA, 5 mM Tris-Hepes, pH7.4, and the trans chamber contained 0.5 M NaCI, 5 mM Tris-Hepes, pH 7.4. The trans chamber was defined as electrical ground. Data were recorded on video tape using a modified audio processor and a video cassette recorder. All experiments were performed at room temperature. RESULTS

Cross-linking of ZnsPaReceptorProtein-InsPa receptor protein was purified from mouse cerebella according to the methoddescribedelsewhere (3). The purified protein was cross-linked with bis(sulfosuccinimidy1)suberate. When the cross-linkedprotein was analyzed by SDS-polyacrylamidegel electrophoresis with the buffer system of Laemmli (25), the proteins did not enter the separation gel even with 3.75% polyacrylamide gel. Therefore, we used instead the agarosepolyacrylamide gelelectrophoresissystem describedby Kiehm et al. (24).Fig. L4 shows the separation of cross-linked products using this system. Four protein bands (I, 11,111, and IV) were clearlyseenin the cross-linked samples.Molecular weight calibration of these bands indicated that I, 11,111, and IV correspondedin approximate M , to 320,000, 650,000, 1,000,000,and 1,250,000, respectively (Fig. 1C).On the basis of the cDNA sequence, the M , of the cerebellar InsP3receptor was estimated to be 313,000 (16). An earlier estimation of the M , of the cerebellar InsP3 receptor was 250,000 (7), but in this case the M , was calibrated using the electrophoresis system of Laemmli (25). The higher molecular weight proteins tend to be moved aberrantly in this system. Thus, the M , of 320,000 seems to be a better estimation for the cerebellar InsP3 receptor monomer. Thus, I, 11,111, and IV correspond to the monomeric, dimeric,trimeric, and thetetrameric forms of the InsP3 receptor protein, respectively. Next, the microsomal fraction from mouse cerebella was cross-linked with disuccinimidyl tartarate (Fig. 2).The crosslinked products of the InsPa receptor were detected by immunoblotting method using a monoclonal antibody against the cerebellar InsP3 receptorprotein (18A10).The pattern of cross-linked products was the same as that of the purified receptor, and the tetrameric form was the major product at the cross-linker concentration of 10 mM (Fig.2f). Similar results were obtained with cross-linkers such as disulfosuccinimidyl tartarate, ethylene glycolbis(succinimidylsuccinate), and N-succinimidyl(4-azidophenyl)-1,3'-dithiopropionate. Destruction of Subunit Structure under Reducing and Nonreducing Conditions-When purified InsPa receptor protein was denatured by SDS in the absence of a reducing agent and was analyzed by agarose-PAGE, fourbands were observed as in the case of the cross-linked samples (Fig. 1B). The purified InsPs receptorwas incubated in 1% SDS solution under nonreducing condition for various periodsat 100 "C and the proteins were detected by immunoblotting after agarose-

I

0

I

1

2 3 4 5 Distance fmm Origin. em

8

FIG.1. Cross-linking of the purified cerebellar InsPs receptor protein. A, purified cerebellar InsP3 receptor (0.1 mg/ml) was cross-linked with bis(sulfosuccinimidy1)suberate BS? a t reagent concentrations of 0 (lane a), 0.125 (lane b), 0.25 (lane c ) , 0.5 (lane d), and 1.0 (lanee) mM, solubilized with agarose-PAGE sample buffer under reducing condition and subjected to agarose-PAGE. The proteins were stained with Coomassie Brilliant Blue R-250. B, purified cerebellar InsP3 receptor without cross-linking was treated in the solution containing 1% SDS, 1 mM EDTA, 10% glycerol, and 10 mM TrisHCI, pH 8.0 (nonreducing condition) for 1min a t 100 "C. The proteins were subjected to agarose-PAGE and stained with Coomassie Bril= 45,000), liant Blue R-250.C, the standard proteins, ovalbumin (M, = 92,5001, bovine serum albumin (Mr= 66,200),phosphorylase b (M, = 116,250),myosin (M, = 200,000),and monomer @-galactosidase(M, (M.= 179,000), dimer, and tetramer of a,-macroglobulin were subjected to agarose-PAGE. The molecular weights of these proteins were plotted against electrophoretic mobility. I, II, III, and IV, the positions of the monomeric, dimeric, trimeric, and tetrameric forms of the cerebellar InsP3 receptor protein, respectively.

FIG.2. Cross-linking of the cerebellar microsomal fraction. Microsomal fraction was prepared from the mouse cerebellum, and membrane suspensions (4 mg/ml) were cross-linked with disuccinimidyl tartarate a t reagent concentrations of 0 (lane a), 0.06 (laneb), 0.6 (lane c), 1.7 (lane d), 5 (lane e), and 10 (lanef) mM, solubilized, and subjected to agarose-PAGE. The cross-linked InsP3 receptor protein was visualized by immunoblotting with monoclonal antibody (18A10). I, ZI, III,and IV, the positions of monomeric, dimeric, trimeric, and tetrameric forms of the InsP3 receptor, respectively. PAGE (Fig. 3A). The amounts of tetramer decreased slowly and after a 10-min incubation, there was almost no tetramer band. Correspondingly, the amounts of dimeric and monomeric forms increased during incubation. In contrast, when the denaturation was performed inthe presence of 2-mercaptoethanol, destruction of the subunit structure was very rapid. Even after 30-s incubation at 100 "C, almost all of the protein was in monomeric form (Fig. 3B). Calmodulin Binding to the Cerebellar ZnsP, Receptor-Ii order to examine whetherthe cerebellar InsP3 receptor binds with calmodulin,the purified receptor protein was applied to

Characterization of the Cerebellar Imp3 Receptor Channel

1112

-

II

' -I

FIG.3. Destruction of the subunit structure of cerebellar InsP3 receptor under reducing and nonreducing conditions. The purified InsPs receptor was denatured for 30 s (lanes a ) , 1 min (lanes b), 3 min (lanes c), and 10 min (lanes d ) a t 100 "C in agarosePAGE sample buffer in the absence ( A ) or presence ( B ) of 2-mercaptoethanol. The proteins were subjected to agarose-PAGE and the InsPR receptor was detected by immunoblotting with monoclonal antibody (18A10). I, 11, I l l , and IV, the positions of monomeric, dimeric, trimeric, and tetrameric forms of InsPn receptor protein.

'

Fractlon -Number

."

%2SK -L

i

-It2 K

-45K

FIG.5. Photoaffinity labeling of cerebellar InsP3 receptor with [a-32P]8-azido-ATP.A, the cerebellar microsomal fraction was photoaffinity labeled with [a-'*P]8-azido-ATP in the absence (lane a ) or presence (lane b) of 1 mM ATP, or in the presence of 50 p~ InsPs (lane c). The membranes were solubilized with SDS-PAGE sample buffer and subjected to 7.5% SDS-PAGE. The gels weredried and then subjected to autoradiography. E, the cerebellar microsomal fraction was photoaffinity labeled as in A in the absence (a, b) or presence (c, d ) of 1 mM ATP, or in the presence of 50 p M InsPn (e, fl, and solubilized with SDS-PAGE sample buffer. The InsPI receptor was immunoprecipitated with 18A10 monoclonal antibody (a, c, e ) , subjected to 7.5% SDS-PAGE, and then to autoradiography. When rat IgG was used instead of 18A10 monoclonal antibody, no labeled product was precipitated (b, d, fl. Open triangles indicate positions of InsPRreceptor and the closed triangle indicates the labeled 100-kDa protein. The positions of molecular mass markers (in kilodaltons) are shown a t right.

FIG.4. Calmodulin-Sepharose chromatography of the purified InsP3 receptor protein. Approximately 30 pg ofthe purified cerebellar InsPs receptor protein was applied to a column of calmodulin Sepharose 4B (0)and Sepharose 4B (0)pre-equilibrated with 0.2% Triton X-100, 0.2 M NaCl, 1 mM 2-mercaptoethanol, 1.5 mM CaCI2,and 20 mM Tris-HC1, pH 7.4. After washing, the receptor was eluted with EGTA (the position of the application of EGTA is indicated by an arrow).Aliquots were assayed for ['H]InsPa binding activity. Inset shows the result of SDS-PAGE of the fractions in calmodulin-Sepharose 4B chromatography (proteins were stained with Coomassie Brilliant Blue R-250).

a column of calmodulin-Sepharose (Fig. 4). Almost all of the InsP3receptor protein was retained on the calmodulin-Sepharose in the presence of Ca2+and was eluted by the application of EGTA. Onthe otherhand, almost allof the InsP3receptor protein was eluted in the void volume, when it was chromatographed on underivatized Sepharose 4B. However, a small portion was retained onSepharose 4B column and was eluted by EGTA. This may be due to the calcium-dependent aggregation of InsP3 receptor protein. The calmodulin binding activity of the InsP3receptor tends to be lost during storage. About half of the InsP3 receptor lost binding activity on the calmodulin-Sepharose column after storage a t -80 "C for about 6 months withoutany loss of the InsP3-bindingactivity. The effects of calmodulin on the binding of InsP3 to the purified InsP3 receptor and to the microsomal fraction from the mouse cerebellum were studied. However, calmodulin exerted no effect on InsPs binding in both cases even a t a concentration of 3 p ~ In. addition, the effects of calmodulin and calmodulin antagonist W7 were investigated on calcium release from the cerebellar microsomal fraction. However, no clear effect was observed when using a time scale of seconds. ATP-binding to the Cerebellar Imp3 Receptor-The microsomal fraction from the mouse cerebellum was incubated with [a-"'P]8-azidoadenosine 5"triphosphate and was then UVirradiated. Several proteins were specifically labeled in the presence of MgC12 (Fig. 5A). Among these, proteins with M, of 100,000 and 250,000 were labeled at a moderate density.

FIG.6. [a-"P]ATP binding to the purified InsP3 receptor. Purified InsP3 receptor (20 pgjml) was incubated for 10 min a t 0 "C with 0.5-32 p~ [(U-'~P]ATP. [a-R2P]ATP binding was measured as described under "Experimental Procedures," and nonspecific binding was measured in the presence of 1 mM cold ATP. Scatchard analysis of boundjfree ( B j F ) versus bound ( B pmoljpg of protein) yields apparent Bmaxand Kd values of2.3 pmoljpg of protein and 17 p ~ , respectively.

When photoaffinity labeling was performed in theabsence of M F , a 100-kDa protein was not labeled, whereas a 250-kDa protein was labeled as in the presence ofMg2+ (data not shown). The 100-kDa protein may be Ca2+-ATPase, which exists abundantly in the cerebellar Purkinje cells (27). Next, immunoprecipitation was performed to examine whether the 250-kDa protein is an InsPn receptor. As shown in Fig. 5B, the immunoprecipitated InsPs receptor was actually labeled with [a-32P]8-azidoadenosine5'-triphosphate, and thislabeling was diminished by the addition of unlabeled ATP. Scatchard analysis of [a-"PIATP binding to the purified InsPs receptor indicated that the Kd was 17 p~ and the Bmaxwas 2.3 pmol/pg of protein (Fig. 6). This BmaX value is almost the same as thatfor InsPs binding (3). The binding specificity was examined using the purified InsP3 receptor (Fig. 7). ATP and ADP were the most potent competitors for [L~-~'P]ATP binding, although ADP was less effective. AMP also inhibited [a-""PIATP binding but the

Characterization of the Cerebellar InsPa Receptor

Channel

1113

with 1% CHAPS for the planar lipid bilayer experiment. Control experiments indicated that thebuffer without protein and with 1%CHAPS had no effect on bilayer conductance. When the purified cerebellar InsP3 receptor was added to the cis chamber in the presence of4.8 pM Inspa, current fluctuations were observed after about 10 min, but almost no fluctuation was observed when the receptor protein was added in the absence of InsP3. Next, the purified receptor protein was added to the cis chamber, and after several minutes 4.8 p~ InsP, was added to the same chamber, and then current fluctuation was observed after approximately 5 min. Aswe FIG. 7. Inhibition of [(r-32P]ATPbinding to thepurified could not know the time of receptor incorporation into the InsP3 receptorby nucleotides. The purified cerebellarInsP, receptor was incubatedwith 0.5 pM [w3’P]ATP in the presence of various membrane, the time lag of channel opening after the binding concentrations of ATP (O),ADP (0),AMP (A), GTP (A), sodium of InsP3 to the receptor could not be determined in this pyrophosphate (U),and InsP3 (X). [w3’P]ATP binding was measured experiment. Fig. 9A shows the representative current fluctuabyPEG precipitation method as describedunder “Experimental tions in bi-ionic recording solutions composed of 125 mM Procedures.” Tris+ (cis) and 53 mM Ca2+(trans). While there are several subconductance states, thecurrent-voltage (I-V) relationship of the most frequently observed amplitude revealed a slope conductance of 26 pS and areversal potential of 26 mV (Fig. 9C). This corresponds to a permeability ratio of Pca/PT,i8 = 6.3. Next, the recordings were performed on the bilayer in asymmetric NaCl solution (cis:O.l M NaCl and trans:0.5 M NaC1) (Fig. 9B). The I-V relationship of the most common amplitude revealed a slope conductance of 21 pS and reversal a potential of 25 mV, corresponding to PN,/Pcl = 5.4 (Fig. 9C). Fig. 10 shows the effects of ATP on the current fluctuation. When 200 p~ ATP was added to the cis chamber after the Dr.* cenremtretiem , r Y initiation of InsP3-induced current fluctuation, no change was FIG. 8. Inhibition of [3H]InsPabinding to the purified InsPa observed (Fig. lob). However after the addition of 600 p~ receptor by nucleotides, pyrophosphate, and phosphate. The ATP, channelopenings with large conductance were observed purifiedcerebellar InsPs receptorwasincubated with 12 nM [3H] (Fig. 1Oc). Next, the concentration of ATP was increased to InsPa in the presence of the various concentrations of ATP (O),ADP 1 mM and subsequently the channel opening with large con(O), AMP (A), GTP (A), pyrophosphate (U),and phosphate (0).[3H] InsP3 bindingwasmeasured byPEG precipitation method as de- ductance disappeared and smaller conductance currents similar to those shown in Figs. 9B and 10a reappeared (Fig. 10d). scribed under “Experimental Procedures.” activity was 20 times weaker than that of ATP. GTP, InsP,, and sodium pyrophosphate were very weakor inactive at this site. In order to test whether this nucleotide binding site was involved in the phosphatase or kinase activities, we examined the ATPase, protein kinase and InsP3kinase activities of the purified InsPs receptor in thepresence or absence of Ca2+and calmodulin. However, none of these activities were detected in thepurified InsP3 receptor protein. Inhibition of fH]InsP3 Binding by Various Nucleotides, Pyrophosphate, and Phosphate-Willcocks et al. (28) reported that ATP andGTP inhibit[3H]InsP3 binding to the rat cerebellar membrane (IC50 = 0.5mM). We examined the effects of nucleotides on [3H]InsP3binding to the purified cerebellar InsP, receptor (Fig. 8). ATP inhibited [3H]InsP3 binding to the purified InsPB receptor protein with an IC50 value of 2 mM. ADP was less effective and possessed an IC5o value of 4.5 mM. AMP was a very weak inhibitor of binding and the IC50 value was in excess of 10 mM. GTP was as a potent inhibitor of binding as ATP. The inhibition curves of sodium pyrophosphate and sodium phosphate were nearly identical with those of ATP and AMP, respectively (Fig. 8). Reconstitution of the Purified InsP3 Receptor Protein into Planar Lipid Bilayers-In the previous study (3) and in the present biochemical experiments, InsP, receptor was purified in the presence of 0.2% Triton X-100. However, significant artifacts were observed when the samples containing Triton X-100 were applied to a planarlipid bilayer system. Therefore, in the last purification step of hydroxylapatite chromatography, Triton X-100 was completely washed out and replaced

DISCUSSION

The results presented here demonstrate that thecerebellar InsP, receptor is composed of four noncovalently bound identical subunits of M, 320,000, each of which contains an InsP3 and an ATP binding site. The purified InsP, receptor binds to calmodulin and forms a calcium-permeable channel with large conductance in the planar lipid bilayer experiment. InsP, opens the channel andthe open probability was greatly increased by ATP in thepresence of InsP,. Earlier estimations of the molecular weight of the cerebellar InsP3 receptor are controversial. SDS-PAGE analysis of the purified cerebellar InsP, receptor using the system of Laemmli (25) indicated that it is a 250-260-kDa protein (2, 3). On the other hand, Supattapone et al. (2) suggested that the native molecular weight of the cerebellar InsP, receptor is 1,000,000 on the basis of Sepharose S-400 chromatography, although this value may bean overestimate due to thebinding of Triton X-100 to the protein. In addition, the cDNA cloning of the mouse cerebellar InsP3 receptor has demonstrated that it is a 313-kDa polypeptide (16). Since higher molecular weight proteins tend to be moved aberrantly in the SDS-PAGE system of Laemmli (251, weused the agarose-PAGE system of Kiehm et al. (24) to estimate more precisely the molecular weight of the InsP, receptor. Agarose-PAGE analysis of the crosslinked InsP3 receptor revealed four distinctbands of M , 320,000, 650,000, 1,000,000and 1,250,000. No higher molecular weight cross-linked products were observed at any of the cross-linker concentrations used, nor with any of the crosslinkers tested. There was no difference in pattern of crosslinking between purified and microsome-bound receptors.

1114


A

L

L

/ b k k d " h

LL.

a -

a

:

A b~

b

A

= -

~

m- ."'

~

~ "

~

~

'

'

b I

B a

"

d

&

I ,

&

"

J4.

FIG. 10. ATP stimulates the InsPa-induced current. Na' currents wererecorded in the asymmetricalNaClsolutions. The cis chamber contained 0.1 M NaCl, 5 mM Tris-Hepes, pH 7.4, 0.1 p~ free Ca2+ and 0.5 pg/ml of the purified InsPBreceptor. The trans chamber contained 0.5 M NaCl and 5 mM Tris-Hepes,pH 7.4. InsPB (3.6 p ~ was ) added to the cis chamber ( a ) . Subsequently, 0.2 ( b ) ,0.6 (c), and 1 mM (d) ATP were added to the cis chamber, sequentially. The holding potential was 0 mV. Verticalcalibration:2 PA. Horizontal calibration: 5 s. Arrows on the left indicate the closed state.

Analysis of thecDNA of the cerebellar InsPB receptor protein suggested that each subunit has several membranespanning domains near the C terminus (16). Scatchard plot analysis of InsP3 bindingsuggested that the cerebellar InsPB receptor has one high affinity binding site for each subunit -40 , -21 (3). Meyer et al. (29) has shown that the binding of at least three molecules of InsP3 is required for the opening of calcium-release channels in permeabilized rat basophilic leukewe have mia cells. LikeFerris et al. ( E ) , inthisstudy demonstrated that the InsP3 receptor protein itself constitutes a calcium-permeable channel. A simpleinterpretation of these results is as follows. The membrane-spanning domainsof the four subunits form a single centraltransmembrane pore. Binding of an InsP3molecule t o each subunit causesconformational change in the membrane-spanning domain in that subunit. After the bindingof three or four InsP3 molecules to -2.01 the subunits, an open ion channel isformed. Further morphoFIG. 9. Single channel current of InsPs receptor channel on the isolated InsPs receptor logical andkineticstudies recorded in planar lipid bilayer. The purified InsPBreceptor was channel are necessary for evaluation of this hypothesis. incorporated into planar lipid bilayer as described in "Experimental Smith et al. (30) and Suematsu et al. (31) have reported Procedures." A , Ca2' currents were recorded by the addition of 2 pg calcium release insmooth of the InsPs receptorto the cis chamber containing 125mM Tris, 250 that ATP stimulated InsP3-induced mM Hepes, and 0.1 PM free Ca2+,pH 7.4. The trans chamber was muscle cells. A photoaffinity labeling experiment directly composed of 53 mM Ca(OH)2and 250 mM Hepes, pH 7.4. No current proved that thecerebellar InsP3 receptor has an ATP binding fluctuation wasobservedbefore addition of Inspa ( a ) but channel site.The purified InsPBreceptorexhibited asingle ATP opening was observed after the addition of 4.8 p~ Inspa to the cis binding site witha Kd value of 17 p~ and a B,,, of 2.3 pmol/ chamber (b). The holding potential was -10 mV. B, Na+ currents were recorded in the asymmetric solutionsof NaC1. The cis chamber pg of protein. Scatchard analysisof [3H]InsP3 bindingto the contained 0.1 M NaCl, 0.1 p M free Caz+,and 5 mM Tris-Hepes, pH purified receptor yielded a B,,, value of 2.1pmol/pg of protein (3). Therefore, itseems that there is the same number of ATP 7.4, and the trans chamber contained 0.5 M NaCl and 5 mM TrisHepes, pH 7.4. Recordings are before ( a ) and after ( b )the addition of and InsP3 binding sites on the InsPsreceptor protein. If we 4.8 p M InsPs to the cis chamber. The holding potential was 0 mV. C, assume that the M , of the InsP3 receptor subunit is320,000 single channel current uersuS voltage relationships of Caz+( 0 )and and that the subunit has one binding site/molecule, the B,,, Na+ (W) currents are recorded. Slope conductances for Ca2+ and Na+ are 26 and 21 pS, respectively. The reversal potentials for Ca2+ and is calculated to be 3.1 pmol/pg of protein. Since the value is Na+currents are26 and 25 mV, respectively. Vertical calibration: 0.5 close to the above-mentionedmeasured values, it seems that PA. Horizontal calibration: 5 s. Arrows on the left indicate the closed there is one ATP binding site and one InsPa binding site on state. each receptor subunit. Anucleotide binding consensus sequence Gly-X-Gly-X-X-Gly was found intheN-terminal When the purified I m p 3 receptor was denatured by SDS cytoplasmic domain of the mouse cerebellar InsP3 receptor is selective under reducing condition and analyzed with agarose-PAGE, (amino acid residues 2016-2021) (16). The binding only a band of M, 320,000 was observed. When denaturation for adenine nucleotide and the affinity is in the order, ATP was performed under nonreducing condition,four bands were > ADP >> AMP. Planar lipid bilayer experiment indicated observed as in the caseof the cross-linkedsamples. However, that ATP promoted conductance of the reconstituted cerelonger incubation of the sample resulted in disappearanceof bellar InsPs receptor channel in the presence of InsP3, but et al. (32) the higher molecular weight bands and a corresponding in- ATP itself didnotopenthechannel.Ehrlich p~ AMP-PCP, a nonhydrolyzable analogue reported that 100 crease in the lower molecular weight bands. Based on these results, we propose that the native cerebellar InsPs receptor of ATP, increased 2-fold the Po of InsP3-gated channels of exists asa noncovalently coupledhomotetramer of M , 320,000 the aortic sarcoplasmic reticulum. Although the single channel conductance of the aortic InsP3-gated channel reported subunits.

n

Characterization of the Cerebellar ImpBReceptor Channel by them was significantly lower than that of the cerebellar InsP3 receptor channel, both channelsmay be regulated similarly by adenine nucleotides. Quite recently, Ferris et al. (33) reported that ATP stimulated InsP3-induced 45Caflux from the Inspa receptor-incorporated lipid vesicles. The stimulatory effect of ATP was highly concentration dependent.ATP increased InsP3-induced 45Caflux at 1-10 p ~but , this effect diminished between 0.1 and 1.0 mM (33). As the intracellular concentration of ATP is about 1 mM, they speculated that Ca2'-ATPase decreased the local ATP concentration to a submillimolar level and then the InsP3 receptor channel is activated (33). However, in our reconstitution experiment, ATP stimulated channel opening most effectively at a concentration of 0.6 mM. This discrepancy may be due to differences of the lipids used for reconstitution or in buffer composition. It is likely that thediminishment of enhancing effect of ATP at the higher concentration reported by Ferris et al. (33) is mediated by the inhibition of InsPa binding by ATP. ATP, ADP, and GTP inhibited [3H]InsP3 binding to the cerebellar InsP3 receptor at millimolar concentrations. This inhibitory effect seems to be mediated by pyrophosphate regions of the nucleotides, because the inhibition curve of pyrophosphate is superimposable on those of ATP and GTP. In addition, Palade et al. (34) reported that millimolar pyrophosphate inhibited InsP3-induced Ca2+ release from brain microsomes. Therefore, various nucleotides and pyrophosphates should be taken into account in assessing the physioreceptor protein. logical function of ATP binding to the InsP3 Thus themodel by Ferris et al. (33) seems to be oversimplified and many more investigations areessential to reveal the physiological significance of ATP binding. Ca2+inhibits InsP3-binding to the microsomal fraction of the cerebellum, but sensitivity to Ca'+ is lost in the purified receptor (2,3). Hill et al. (35) reported that calmodulin antagonists W7, W13, and CGS934313 inhibited InsP3-stimulated calcium mobilization in rat liver epithelial cells. CalmodulinSepharose 4B chromatography of the cerebellar Inspa receptor protein indicated that this protein binds to calmodulin dependent on Ca2+.However, the addition of calmodulin did not alter InsP3 binding to the cerebellar InsP3 receptor protein. As far as we examined, neither calmodulin nor calmodulin antagonist affected the InsP3-induced calcium release from the cerebellar microsomal fraction. However these analyses were performed at the time scale of seconds to minutes. Analysis with a time frame in the order of milliseconds may be necessary to reveal the effects of calmodulin. Another possibility is that calmodulin modulates the possible enzyme activity of theInsP3 receptor protein. Because theInsP3 receptor has aATP binding site, it may be a Ca'+ calmodulindependent ATPase, aprotein kinase, or an InsP3-kinase. However, none of these enzyme activities were detected with or without Ca"-calmodulin. Reconstitution of the purified InsP3receptor into theplanar lipid bilayer indicated that the receptor constitutes a cation selective channel that is opened by InsP3. Channelsdisplayed several subconductance states. When ATP was added in the presence of InsP3, thereappeared larger conductance currents. This may be due to the change in the full open conductance or the shift to a larger subconductance state. However, even in the absence of ATP, the same conductance levelwas occasionally observed. Thus itis likely that ATP modifies the channel to reach the larger subconductance state. Recently, the primary structures of cerebellar Inspa receptor andskeletal muscle ryanodine receptor have been reported and it was revealed that the cerebellar InsP3 receptor has fragmentary sequence homology with the ryanodine receptor

1115

protein (16, 20). The skeletal muscle ryanodine receptor is a calcium release channel of the sarcoplasmic reticulum, which is considered to play important roles in excitation-contraction coupling (18, 19, 37). Both proteins havelarge N-terminal regions located on the cytoplasmic side and short C-terminal regions containing several membrane-spanning domains. A remarkable similarity between the two proteins is found in two of the transmembrane segments and the successive Cterminal regions. In addition to the similarities in primary structure, thereis a strikingresemblance of subunit structures and functional aspects. Lai et al. (36) cross-linked the skeletal muscle ryanodine receptor with glutaraldehyde and suggested that thisprotein is composed of four identical subunits. Crosslinking of the cerebellar InsPa receptor indicated that this protein is also composed of four identical subunits. Furthermore, both proteins bind ATP and calmodulin, and conduct Ca2+ and Na' (36-38).However, important differences are present between the gating mechanisms of the two channels. For example, ryanodine did not affect InsP3 receptor channel activity.' And in contrast to the ryanodine receptor channel (37, 38), the single channel conductance of Inspa receptor for Na+ was smaller than thatfor Ca'+. This is striking difference between the ryanodine and InsP, receptor channels. Detailed comparison of the pharmacological properties of the both channels are now in progress in our laboratory. The present study demonstratedthe remarkable structural and functional similarities of InsPa and ryanodine receptor channels. Acknowledgment-We wish to thank Dr.Michio Niinobe foruseful discussions. REFERENCES 1. Berridge, M. J., and Irvine, R.F. (1989) Nature 341, 197-205 2. Supattapone, S., Worley, P. F., Baraban, J. M., and Snyder, S. H. (1988) J. Biol. Chem. 263, 1530-1534 3. Maeda, N., Niinobe, M., and Mikoshiba, K. (1990) EMBO J. 9, 61-67 4. Yamamoto, H., Maeda, N., Niinobe, M., Miyamoto, E., and Mikoshiba, K. (1989) J. Neurochem. 53,917-923 5. Mikoshiba, K., Okano, H., andTsukada, Y. (1985) Deu. Neurosci. 7, 179-187 6. Maeda, N., Niinobe,M.,Inoue, Y., and Mikoshiba, K. (1989) Deu. Biol. 133,67-76 7. Maeda, N., Niinobe, M., Nakahira, K., and Mikoshiba, K. (1988) J. Neurochem. 51, 1724-1730 8. Mallet, J., Huchet, M., Pougeois, R., and Changeux, J . P. (1976) Brain Res. 103, 291-312 9. Mikoshiba, K., Huchet, M., and Changeux, J. P. (1979) Deu. Neurosci. 2, 254-275 10. Supattapone, S., Danoff, S. K., Theibert, A., Joseph, S. K., Steiner, J., and Snyder, S. H. (1988) Proc. Natl. Acad. Sci. U. S. A. 85,8747-8750 11. Ross, C. A., Meldolesi, J., Milner, T . A., Satoh, T., Supattapone, S., and Snyder, S. H. (1989) Nature 339, 468-470 12. Mignery, G . A., Sudhof, T. C., Takei, K., and De Camili, P. (1989) Nature 342, 192-195 13. Otsu, H., Yamamoto,A., Maeda, N., Mikoshiba, K., and Tashiro, Y.(1990) Cell Struc. Funct. 15, 163-173 14. Furuichi, T., Shiota, C., and Mikoshiba, K. (1990) FEBS Lett. 267,8548 15. Ferris, C. D., Huganir, R. L., Supattapone, S., and Snyder, S. H. (1989) Nature 342,87-89 16. Furuichi, T., Yoshikawa, S., Miyawaki, A,, Wada, K., Maeda, N., and Mikoshiba, K. (1989) Nature 342, 32-38 17. Miyawaki, A., Furuichi, T., Maeda,N., and Mikoshiba, K. (1990) Neuron 5, 11-18 18. Campbell, K. P., Knudson, C. M., Imagawa,T.,Leung, A. T., Sutko, J. L., Kahl, S. D., Raab, C. R., and Madson, L. (1987) J.Biol. Chem. 262,6460-6463

T. Kawasaki and N. Maeda, unpublished observation.

1116


19. Smith, J. S., Imagawa, T., Ma, J., Fill, M., Campbell, K. P., and 30. Coronado, R. (1988)J. Gen. Physiol. 92, 1-26 20. Takeshima, H., Nishimura, S., Matsumoto, T., Ishida, H., Kan- 31. gawa, K., Minamino, N., Matsuo, H., Ueda, M., Hanaoka, M., Hirose, T., and Numa, S. (1989) Nature 339,439-445 21. Iino, M., Kobayashi, T., and Endo, M. (1988) Biochem. Biophys. 32. 33. Res. Commun. 152,417-422 22. Iino, M. (1989) J. Gen. Physiol. 9 4 , 363-383 34. 23. Ashley, R. H. (1989) J.Membr. Biol. 111,179-189 24. Kiehm, D. J., and Ji,T. H. (1977) J.Biol. Chem. 252,8524-8531 25. Laemmli, U. K. (1970) Nature 227,680-685 35. 26. Kawasaki, . T., and Kasai, M. (1989) J. Biochem. (Tokyo) 106, 401-405 36. 27. Kaprielian, Z., Campbell, A. M., and Fambrough, D.M. (1989) Mol. Brain Res. 6, 55-60 28. Willcocks, A. L., cooke, A. M., potter, B. V. L., and Nahorski, s. 37. K. (1987) Biochem. Biophys. Res.Commun. 1 4 6 , 1071-1078 29. Meyer, T., Holowka, D., and Stryer, L. (1988) Science 40, 653- 38. 656

Smith, J. B., Smith, L., and Higgins, B. L. (1985) J. Biol. Chem. 260914413-14416 Suematsu, E., Hirata, M., Sasaguri, T., Hashimoto, T., and Kuriyama, H. (1985) Comp. Biochem. Physiol. 82A, 645-649 Ehrlich, B. E., and Watras, J. (1988) Nature 336,583-586 Ferris, C.D., Huganir, R. L., and Snyder, S.H. (1990) Proc. Natl. Acad. Sci. U. S. A . 87, 2147-2151 Palade, P., Dettbarn, C., Alderson, B., and Volpe, P. (1989) Mol. Pharmacol. 36, 673-680 Hill, T. D., Campos-Gonzalez, R., Kindmark, H., and Boynton, A. L. (1988) J.Biol. Chem. 263, 16479-16484 Lai, F. A,, Misra, M., Xu, L., Smith, H. A., and Meissner, G. (1989) J. Biol. Chem. 264,16776-16785 Lai, F. A., Erickson, H. P., Roussean, E., Liu, Q.-Y., and Meissner, G. (1988) Nature 331, 315-319 Anderson, K., Lai, F. A., Liu, Q.-Y., Rousseau, E., Erickson, H. P., and Meissner, G. (1989) J. Biol. Chem. 264, 1329-1335