Structural and Functional Characterization of an Inositol Polyphosphate Receptor from Cerebellum*

An inositol polyphosphate receptor has been purified from bovine cerebellum which consists of three different polypeptides with M, of 111,000, 102,000, and 62,000. Negative staining electron microscopy reveals globular-like structures 10-13 nm in diameter. The receptor has a Stokes radius of 400,000 daltons as determined by molecular sieve high performance liquid chromatography. The receptor preparation binds ino- sitol 1,3,4,5-tetrakisphosphate, inositol hexaphos-phate (or phytol), and inositol 1,4,5-trisphosphate (IP4, IPS, and IPS, respectively) with submicromolar affinity (0.19, 0.15, and 0.54 p ~ , respectively) at conditions approximating physiological ionic strength and pH. The purified receptor preparation, when reconstituted into planar bilayers, displays ion channel activity, preferentially permeable to K+. Permeability ratios of the channel are P ~ / P N ~ + -5 and PK+/PcI -19. In symmetrical 100 mM KCl, the channel is characterized by long open times (minutes) with a conductance of 7.2 picosiemens. The channel is selectively modulated by IP4. That is, at 1 p~ IP4, the mean open time decreased substantially to rapid flicker

$ Investigator of the American Heart Association (Tennessee Affiliate). second messenger to mobilize calcium from intracellular stores. By this mechanism, the generation of IP3 activates numerous Ca2+-dependent processes (1,2).
Characterization of the molecular machinery involved in the IP3-dependent Ca2+ release process has recently been achieved. An IP3 receptor has been isolated from both cerebellum (3,4) and smooth muscle (5). Comparison of the smooth muscle IP3 receptor with the IP3 receptor from cerebellum indicates that the two receptors are similar both structurally and functionally (6). Functional identity of the IP3 receptor as an intracellular Ca2+ release channel has been indicated by incorporation of the cerebellum IP3 receptor into phospholipid vesicles (7,8) and, more directly, by reconsititution of the smooth muscle IPS receptor into planar lipid bilayers and observing IP3 activated Ca2+ channels (9,10).
IP3 generated by the IP3 cascade is converted to other inositol polyphosphates by specific kinases and phosphatases. Some of these inositol polyphosphates have also been suggested to be involved in intracellular signaling (1, 2, 11). Inositol 1,3,4,5-tetraphosphate (IP4), a metabolite of IP3, has been postulated to act as an intracellular messenger (11)(12)(13). In this study, we describe the purification and characterization of a receptor from cerebellum which bound IP,. After isolation, we found that it bound also IPS and IP3. Reconstitution of this inositol polyphosphate receptor preparation into planar bilayers displays potassium channel activity which is selectively modulated by IP,. A preliminary communication has appeared (14).

Purification of a n Inositol Polyphosphute Receptor Methodology
Unless indicated otherwise, all solutions contained protease inhibitors obtained from Sigma (pepstatin A at 0.5 pg/ml, leupeptin at 0.5 pg/ml, and aprotinin at 0.21 pglml), and the pH of solutions was adjusted at room temperature. All purification steps and binding assays were performed at 4 "C.

Isolation of Bovine Cerebellum Microsomes'
Twelve 5-g portions of frozen bovine cerebellum were each thawed for 10 min in 250 ml of ice-cold buffer A (50 mM Tris-C1 (pH 7.15), 1 mM EDTA, 1 mM DTT) contained in Beckman JA-10 centrifuge bottles and homogenized for 15 s at setting 6 using a polytron homogenizer (PT 20 ). The homogenate5 were then centrifuged for 1 h at 14,000 rpm using a Beckman JA-14 rotor. The clear supernatants were discarded and the pellets resuspended in 250 ml buffer A containing 150 mM NaC1. The resuspensions were homogenized for 10 s at setting 6 using the polytron and centrifuged as described above. The pellets were resuspended to a final volume of 250 ml in buffer A and stored at -80 "C. Approximately 2400 mg of microsomal protein was obtained from 60 g wet weight of cerebellum. For competition studies of [3H]IP4 binding to purified receptor, 0.2 or 1.0 pg of receptor protein was assayed in 100 pl of binding medium composed of 25 mM Na-Hepes pH 7.1 (pH 7.4 on ice), 1 mM EDTA, 1 mM DTT, 1% CHAPS, 5 mg/ml equine IgG (Sigma), 10 nM [3H]IP4 (17 Ci/mmol) in the presence of 0-3.1 p~ unlabeled ligand and +lo0 mM KCl. The receptor was coprecipitated with IgG using 6% (w/w) polyethylene glycol and pelleted by centrifugation as described previously (5). Nonspecific counts were determined by omission of the purified receptor from the binding medium. Competition curves were generated by fitting the data to a generalized ligand binding equation: y = a / [ l + (n/b)'] by nonlinear regression (Marquart-Levenberg algorithm), where a is maximal binding (normalized to loo%), b is the and c is the slope factor where a value of 1 indicates normal hyperbolic binding while values greater or less than one suggest either positive or negative cooperativity, respectively (see Ref. 29).
For Scatchard analyses, [3H]IP3, [3H]IP4, and [3H]IP6 stock solutions were either used as obtained from Du Pont-New England Nuclear (0.59, 0.59, and 0.80 p~, respectively) or were supplemented with "cold" IP3, IP,, and IP,, respectively, to give stock concentrations of 5.9 p M for IP3 and IP, (-2000-2200 cpm/pmol, - 1.7 Ci/mmol), or 4.0 p~ for IPS (-2400 cpm/pmol, -2.4 Ci/mmol). The binding studies were then carried out as a function of ligand concentration by addition of increasing amounts of radioligand solution. The concentrations of the respective ligands used in all binding assays were calculated from information provided by the suppliers of radioactive and nonradioactive inositol phosphates.
(wet weight) of cerebellum were thawed and mixed with 1500 ml of Solubilization was initiated by addition of 1000 ml buffer B containing 60 g of CHAPS (final concentration = 2%, w/v). After gentle stirring for 1 h, the suspension was centrifuged for 1.5 h at 14,000 rpm using a Beckman JA-14 rotor. The pellets were discarded, and the supernatant was used for purification of the inositol polyphosphate receptor.
Purification of the Receptor-The supernatant was adjusted to 200 mM with NaCl and gently mixed with 85 ml of heparin-agarose (which had been equilibrated with H,O). After 2.5 h, the gel was collected by vacuum filtration, washed with 400 ml of buffer B containing 1% (w/ v) CHAPS and 200 mM NaCl, and poured into a 2.5-cm inner diameter ' The term microsomes is used in this study to refer to the pellet obtained from cerebellum homogenate by sedimentation at 19,000 X g. , for 1 h. column. The receptor was eluted with the same buffer containing 800 mM NaCl and 8-ml fractions were collected. The five peak fractions (Nos. [12][13][14][15][16] containing IP4 binding activity were pooled. The pooled fractions from the heparin-agarose column were dialyzed overnight against 1100 ml of buffer C (20 mM Tris-C1 (pH 7.4), 1 mM DTT, and 1 mM EDTA containing 1.0% (w/v) CHAPS). The dialyzate was diluted with an equal volume of buffer C and centrifuged for 30 min at 20,000 rpm using a Beckman Ti 70 rotor to remove aggregated protein. The supernatant was loaded onto a 15 ml column of DEAE-Sepharose-GB (1.5 cm inner diameter) at a flow rate of 0.5 ml/min, which had been prequilibrated with buffer C. Elution was achieved with a 150-ml gradient of 0-250 mM NaCl in buffer C, 7.5-ml fractions were collected. Peak fractions containing IP, binding activity (fractions [11][12][13][14] were pooled, frozen in liquid nitrogen, and stored at After thawing, the pooled fractions were loaded at 0.5 ml/min onto a second heparin agarose column (3 ml, 1.0 cm inner diameter) which had been equilibrated in buffer D (50 mM Tris-C1 (pH 8.31, 1 mM EDTA, 1 mM DTT) containing 1.0% CHAPS. Elution of the receptor was achieved with a 40-ml gradient of 0.2-0.8 M NaCl in the same buffer, and fractions of 2 ml were collected. Fractions 8-15 containing the receptor were pooled and adjusted to 2.5 M NaC1. The sample was then applied onto a 2.0-ml column of phenyl-Sepharose 4B which had been equilibrated with buffer D containing 2 M NaCl and 1.0% CHAPS. The column was washed with 6 ml of the same buffer containing 2 M NaCl and the receptor obtained by sequential elution with 4 ml each of 1, 2,3, and 4% (w/v) CHAPS in buffer D (see Fig. 3B). Fractions (2 ml) containing IP4 binding activity, which eluted at CHAPS concentrations of 2% or more, were pooled. The pooled fractions from the phenyl-Sepharose column were concentrated (Centricon-30) to 1-2 ml and loaded onto a TSK gel G3000SW column (60 cm X 21 mm) and eluted at 1.5 ml/min with buffer E (50 mM imidazole (pH 6.9), 100 mM NaC1, 1% CHAPS, 1 mM EDTA, 1 mM DTT). Fractions of 2.4 ml were collected and elution was monitored by AZm. Peak fractions (43-46) which contained the IP4 binding activity, were pooled, concentrated (Centricon C-30), and stored frozen at -80 "C. In practice, a single preparation was performed over a period of five days.

Reconstitution into Planar Bilayers and Channel Measurements Preparation of Lipidlprotein and Lipid
Vesicle-The lipid used throughout this study was a lipid mixture: soybean lipid (Sigma, type 11-S, acetone-washed) with cholesterol (Fluka) in a weight ratio of 24:l. Lipid vesicles and lipid/protein vesicles were prepared in either 100 mM KC1 or 100 mM NaCl buffered to pH = 7.4 by 10 mM Hepes-Tris. Adjustments to other salt concentrations used, i.e. 500 mM KC1 or additional CaC12, were done after bilayer formation. Lipid vesicles were prepared by resuspending a thin film of 10 mg of lipid dried under nitrogen in 10 ml of buffer (15). Protein/lipid vesicles preparation: 10 pl of receptor protein (0.54 mg/ml), 6 mg/ml CHAPS, and 3 mg/ml soybean lipid were suspended in 1 ml of buffer for 10 min with occasional vortexing. Then Bio-Beads (Bio-Rad) were added (0.23 g/ml), and the flask rotated for 1 h, followed by a second 1-h rotation with fresh 0.33 g/ml Bio-Beads. The sample was stored on ice until use (up to 3 h).
Planar Lipidlprotein Bilayers Formation-Planar bilayers were formed from vesicles of defined lipid protein ratio, according to the septum-supported vesicle-derived bilayer technique (15). Briefly, planar bilayers (aperture size of 0.18 mm) are formed by apposition of two vesicle derived monolayers. Both cis (protein-containing side) and trans chambers contain 1 ml of solution. Five (Figs. 10 and 12) and 2.5 pl (Fig. 11) of fresh protein lipid vesicle preparation was diluted into 2 ml of fresh lipid vesicle preparation to provide a working suspension of proteolipid vesicles. Using the molecular weights, 400,000 for IP, receptor and 750 for soybean phospholipid, the two above protein concentrations correspond to final molar ratios of protein/lipid in vesicles of 2.2 and 1.1. lo-', respectively, as given in the figure legends.
IP4 (Boehringer Mannheim) was applied in 1-pl aliquots directly to the membrane on the cis side via a steel tube. The tube, of inner diameter 0.85 mm, was adjusted with its end to the membrane (center to center) at a distance of 0.15 mm. It could be removed for refilling and accurately replaced to the same position. Using valinomycin (concentration of 0.66 ng/ml, 1 p1 applied) for calibration, the time for half-maximal response of induced current was 6 + 1 s. The response showed a plateau between 30 s and 3 min followed by a slow decay (50% after 12 k 1 min) due to diffusion loss of the applied

Purification of an Inositol Polyphosphate Receptor 3475
valinomycin into the bulk solution (10). Electrical Measurements and Data Processing-Electrical contact with the solution was made via Ag/AgCl electrodes. Voltage is expressed as the voltage applied to the cis solution. The voltage signal across the feedback resistance (10'' ohms) of the current-measuring operational amplifier was filtered at 3 kHz and stored on a pulse code-modulated audio tape recorder modified to accept dc signals.
To further characterize this low affinity IP, receptor, its purification was achieved as described under "Experimental Procedures" and summarized in Table I. We have designated this receptor the inositol polyphosphate receptor since it also binds IPS and IPS (see below). Microsomes from cerebellum  were solubilized with CHAPS, and the receptor was purified using a combination of column chromatography procedures. Approximately 80% of IP, binding activity was solubilized by extraction of cerebellum microsomes with 2% CHAPS. The [3H]IP4 binding activity was then concentrated and enriched by absorption on heparin-agarose and elution with buffer containing high salt (Table I). After dialysis to lower the NaCl concentration, the IP, binding activity was purified by sequential chromatography on DEAE-Sepharose 6B, a second heparin-agarose column, and phenyl-Sepharose 4B. The final step made use of a TSK gel G3000SW gel filtration column. The receptor is enriched approximately 140-fold from the solubilized microsomes, with an estimated recovery of 1%. A significant loss of IP, binding equivalents is sustained prior to the DEAE-Sepharose 6B step (Table I). This is attributable in part to co-precipitation of IP, binding activity (-50%) together with contaminating proteins during the dialysis step preceding the DEAE-Sepharose 6B chromatography.
The protein profile of the fractions obtained in the purification procedure was characterized by SDS-PAGE (Fig. 2). The (Fig. 4B, fractions 43-45), a constant ratio of the three polypeptide bands is observed, as determined by densitometry, which correlates with [3H]Ins(1,3,4,5)P4 binding. These findings suggest that the receptor is a heterooligomeric complex of the three bands. Pretreatment of the purified receptor with 2-mercaptoethanol and heating (100 "C, 1 min) did not change the electrophoretic profile (not shown).
The three bands characteristic of the receptor are co-enriched at each step in the purification (Fig. 2, lanes 2-6). The enrichment, observed in each of 10 preparations, correlates with the determinations of IP, binding activity (Table I) Molecular sieve HPLC reveals a molecular weight of -400,000 for the purified receptor in 1% CHAPS (Fig. 5). Negative staining electron microscopy reveals uniform globular-like structures of approximately 10-13 nm in diameter (Fig. 6).
Characterization of IPc Ips, and IPS Binding to the Purified Receptor-Binding of [3H]Ins(1,3,4,5)P4 to the purified receptor is not unique, since competition is observed not only with IP4, but also with IPS (Fig. 7A) and IP3 (Fig. 7B) (Fig. 7B). The estimated ICm value for Ins (1,4,5)P3 increases about 2.5-fold in going from the lower ionic strength to physiological ionic strength at pH 7. 4. The relatively higher ICso, values under conditions approximating physiological pH and ionic strength, indicate that in situ the receptor has weaker affinity for these inositol polyphosphate ligands. FIG. 6. Electron microscopy of inositol polyphosphate receptor. Electron micrograph of the purified receptor obtained after negative staining with uranyl acetate. The receptor has a diameter of 10-13 nm. Fig. 9. The Scatchard plot is consistent with a t least two classes of binding sites. The higher affinity site is likely referable to the high affinity IP3 receptor (3-6), whereas the lower affinity site corresponds, at least in part, to the receptor described in this study. Curve fitting of the binding data was performed to estimate the relative density of the two binding sites (Fig. 9, cf. legend). The results indicate that the higher affinity site (& = 2.1 nM) has a B,,, of -0.83 pmol/mg, whereas the lower affinity site ( K d - 1.4

p M ) has a Bmax of -19
pmol/mg. Thus, in cerebellum microsomes, we estimate the magnitude of the lower affinity binding site to be about an order of magnitude greater than the higher affinity binding site.
When the microsomes are extracted with CHAPS as described under "Experimental Procedures," only a relatively small fraction, approximately 10% of the total IP3 binding, is referable to the high affinity IP3 receptor as monitored by its sedimentation into the lower regions of sucrose density gradients (5). The high affinity Ins(1,4,5)P3 receptor is not eluted from the DEAE-Sepharose 6B column under the conditions used to elute the inositol polyphosphate receptor.
Channel Characteristics of the Purified Inositol Polyphosphate Receptor-The purified receptor preparation was reconstituted into lipid vesicles from which planar membranes were formed (see "Experimental Procedures"). Ion channel activi-  Table  11. The percentage of bound [3H]IP4 relative to total radiolabel ranged from 10 (at low ionic strength) to 3% (in isotonic salt). Nonspecific binding varied from 3% (low ionic strength) to less than 1% (isotonic salt). Binding assays, determination of IC, values and slope factors were carried out as shown in Fig. 7, A and B (see "Experimental Procedures"), on two receptor preparations. The binding assay was performed a t pH 7.4 in a medium containing low salt (25 mM NaHEPES, pH 7.4) or supplemented with 100 mM KC1 ("isotonic salt"). The values in parenthesis represent the range of IC, values and slope factors determined for each preparation. A slope factor of 1.0 indicates simple hyperbolic binding; a slope factor of less than one indicates negative cooperativity (29  1.30) ties were clearly observed in 45 out of 60 membranes and were dependent on the presence of the receptor, in contrast to lipid membrane controls which showed no channel activity a t all. Fig. 10 shows bursts of channel events of about 10-s duration which are typical for a KCl/NaCl gradient (100/100 mM) with positive potential applied to the cis side containing KC1 (50 mV in Fig. 10). The upper burst was observed just binding to the purified IPa/IPs receptor. Specific (duplicate samples) and nonspecific binding were determined at pH 8.9 in 50 mM Tris-C1 and 5 mM NaCl using -1. 25   Scatchard analysis of binding data to the purified receptor Binding was carried out at pH 8.9 in a medium containing 50 mM Tris-C1 and 5 mM NaCl, see Fig. 8. For IP, and Ips, nonspecific binding was equivalent to -2% total ligand. For IPS, -3% of total ligand. IP3 and IP, concentrations ranged from 18 to 590 nM. IPS concentrations ranged from 12 to 400 nM. Specific activities: IPS, 2200 cpm/pmol; IP4, 2200 cpm/pmok IPS, 2400 cpm/pmol.  Fig. 11). Under these conditions, the channel showed a constant conductance of 7.2 f 0.8 pS in the range of -100 to +lo0 mV applied potential and no significant dependence of mean open time with applied voltage (data not shown). IP4 was found to close the channel. The long mean open time of the channel in symmetrical 100 mM KC1 was used to study effects of inositol polyphosphates on this channel activity. A key finding is illustrated in Fig. 11. It shows a continuous record where channel activity was completely blocked by 10 p~ IP4 (Ins(1,3,4,5)IP4 when applied close to the membrane (cis side with 1 pl of 10 PM IP4, using a precisely readjustable tube, see "Experimental Procedures")). Upon tube removal and 30-s stirring (dilution of IP4 to -10 nM), channel activity reappeared. Channel blockage by 10 p~ IP4 and reopening upon IP4 dilution was repeated three times; Complete blockage by 10 p~ IP, was observed in four out of four planar bilayer membranes (Fig. 11). Application of 10 pM IP3 (Ins(1,4,5)P3) or 10 pM IPS at conditions of Fig. 11 did not activate additional channels or block the observed channels. In the presence of 1 p~ IP, (conditions of Fig. 11) the long lasting open channel events were not completely blocked, but convert to flicker behavior, i.e. repetitive transitions between open and closed states, albeit similarly long lasting trains of rapid closing and reopening events (1-10 min). Such activity is exemplified in Fig. 12a, where it has been employed to determine the permeability ratio PK+/PC)-of the channel. The data were obtained in the presence of 1 PM IP4 and a KC1 gradient (500/100 mM). Open channel currents of the traces in Fig. 12a are plotted against applied voltage ( Traces were filtered with 100 Hz. "flicker" activity at 1 p~ IP4 and the same reversal potential. For the channel currents observed in a KCl/NaCl gradient (100/100 mM) as shown in Fig. 10 for +50 mV applied to the KC1 side, a reversal potential of -35 to -40 mV was obtained.
From this value and from the permeability ratio PK+/PcIof -19 a value of PK+/PNa+ of -5 is obtained (17).
Attempts to determine a value for PK+/Pc~~+ failed thus far.
High Ca2+ concentrations (above 5 mM) applied to either side of the membrane abolished channel activity observed before a t symmetrical KC1 (conditions of Fig. 11). Rare channel events (at 50 mM CaClz at the trans side) indicated, however, that the reversal potential remained at 0 -+ 5 mV, which sets Ca2+ permeability of the channel to rather low values. In addition, Ca2+ (50 mM) appeared to reduce channel currents for applied voltages of either sign by about a factor of 3, which also rendered determination of reversal potentials difficult (data not shown).

DISCUSSION
We have purified an inositol polyphosphate receptor from bovine cerebellum which binds IP,, IP3, and IPS. The receptor preparation, when reconstituted into planar bilayers, displays channel activity. IP, induces rapid channel flicker at 1 p~ concentration and closes the channel at 10 p~. IPS or IPS are without effect under similar conditions. The receptor has been enriched about 140-fold in IP, binding from the starting cerebellum microsomes (Table I) where it appears to exist at higher concentration by an order of magnitude than the IP3 receptor described previously (3-5) ( Fig. 9). At conditions approximating physiological pH and ionic strength, the purified receptor has binding affinity (IC5o competition uers'sus [3H]IP4) in the range of -0.15, 0.2, and 0.5 pM for IPS, IP4, IP3, respectively. At lower ionic strength (25 mM), 5-10-fold higher affinity is observed for IP, and IPS and >2-fold for IP3 binding ( Fig. 7 and Table 11). Heparin markedly blocks the binding of IP3, IP4, and IPS; inositol 1phosphate and inositol 1,4-&phosphate do not effectively compete for binding of either IPS or IP,.
The purified receptor preparation consists of three bands with apparent molecular weights of 111,000, 102,000, and 52,000 as determined by SDS-PAGE (Fig. 2). The complex has an estimated stoichiometry of 1:3:2. This would be compatible with the minimal molecular weight of about 500,000, assuming that each of the peptides had the same extinction coefficient with Coomassie Blue and that the electrophoretic mobility obtained by SDS-PAGE provides reliable estimates of molecular weight. Our data suggest specific association of the three polypeptides as a receptor complex: 1) the coenrichment of the 111,000, 102,000, and 52,000 bands as observed by SDS-PAGE (Figs. 2 and 3); 2) the symmetry of the HPLC gel titration elution profile (Figs. 4 and 5); and 3) the uniform size of the particles observed by electron microscopy (Fig. 6).
The Stokes radius of the purified receptor solubilized in CHAPS was determined by HPLC gel filtration to be Purification of a n Inositol Polyphosphate Receptor -400,000 (Fig. 5). Negative staining electron microscopy reveals a globular structure with a diameter of 10-13 nm (Fig.  6), consistent with a molecular weight in the range of 400,000. A receptor with M, of 400,000, which binds one mole equivalent of ligand, has an expected BmaX of 2.5 nmol/mg. This is in the range of B,,, values determined for IP,, but is somewhat lower for IPS and IP, (Table 111).
The reconstitution experiments and channel measurements provide evidence that the purified receptor preparation is associated with ion channel activity. Channel currents were strictly dependent on the presence of the receptor preparation. We find that IP, induces channel flickering in the concentration range of the binding ( K d -1 p~) , with complete blockage at 10 p M IP4 concentration. For initial classification of this channel, we studied permeability ratios of physiological ions (K+, C1-, Na+, Ca2+), which indicate that the channel may be classified as a potassium channel, although residual Ca2+ permeability cannot be excluded, at present. The localization of the receptor in the cell remains to be determined. The K+ permeability and sensitivity of its open time to asymmetric Na+ and K+ across the membrane might indicate that the channel is of plasma membrane origin. In this first study, the channel characterization was limited to show that the receptor preparation forms an ion channel after reconstitution into planar bilayers. Although inactivation of the channel by IP, is reported here, the nature of activation of the channel, i.e. what turns it on and why it is selectively modulated by IP,, remains to be discerned.
IP3 has been implicated in intracellular Ca2+ mobilization in a diversity of cell types and has been shown to release Ca2+ from microsomes derived from cerebellum (18)(19)(20), smooth muscle (21), and platelets (22). A high affinity IPS receptor has been isolated previously from both cerebellum (3,4) and smooth muscle ( 5 ) . The cerebellum IPS receptor has been cloned and shown to have a protomer molecular weight of 313,000 (4,6, 25). The smooth muscle IPS receptor is structurally and functionally similar to the cerebellum receptor (6). The cerebellum IP3 receptor has been implicated as a Ca2+ channel, since IP3 enhances Ca2+ efflux from vesicles containing the reconstituted receptor (7,8). More recently, we have reconstituted the purified IPS receptor from smooth muscle into planar lipid bilayers and have directly demonstrated it to be an IP3 activated Ca2+ channel (10). Thus, the high affinity IP, receptor from cerebellum (3,4) and smooth muscle ( 5 ) is an IP3-activated channel responsible for mediating release of intracellular Ca2+.
The inositol polyphosphate receptor preparation described in this study is clearly different from the high affinity IPS receptors isolated previously from cerebellum (3,4) and smooth muscle ( 5 ) . The high affinity IP, receptor is a tetramer of a single polypeptide chain which has a protomer molecular weight of 313,000 (4,6,25). Morphologically, the IP3 receptor exhibits 4-fold symmetry, appearing as a pinwheel with four arms radiating from a central hub (5). The inositol polyphosphate receptor binds IP3 with lower affinity (-0.5 p~ uersus 1 1 0 nM) under comparable conditions (pH 8.9 and -50 mM salt) (this study and Ref. 5). The inositol polyphosphate receptor is a heterooligomer consisting of three polypeptides. It is spherical and much smaller in size (10-13 nm, Fig. 5) than the IP3 receptor which has dimensions of 25 X 25 x -10 nm (5). We estimate that in cerebellum, the inositol polyphosphate receptor is capable of binding about 20-fold more IP3 than the high affinity IP, receptor (Fig. 9).
Significantly, IP3 kinase and IP, phosphatase activities (either Ca2+/calmodulin-dependent or -independent) is not detected in the purified inositol polyphosphate receptor, although readily detected in the first heparin pooled fraction., Values for intracellular IP3 concentration in the literature vary widely from submicromolar basal levels to several micromolar or higher with activation (32). There is less data reported in terms of the IP, concentrations in tissues. Several micromolar of IP, or higher has been indicated which is higher concentration than the ICso for IP, binding (32).
Partial purification of a high affinity IP4 binding protein from cerebellum has previously been reported by Theibert et al. (26) and Donie et a1 (27). Both groups report binding specific for IP4 (& = 1-10 nM), but not for IP3. After our manuscript was submitted, the purification of a high affinity IP4 receptor and an IPS receptor was reported using an IP4 affinity column (30). The IPS receptor is a heterotrimer of similar subunit composition (115, 105, and 50 kDa) to that reported here for the inositol polyphosphate receptor and appears to have similar binding characteristics. Competitive binding studies using [,H]IP4 indicate that a number of inositol polyphosphates (IP,, IP,, IPS, and IPS) bind to the receptor. The reported binding of IPS is high affinity (ICS0 -12 nM) at low ionic strength. We find that binding is considerably weaker at isotonic salt concentration and that binding of IP, and IPS are not very different.
We have designated the receptor the inositol polyphosphate receptor to reflect is promiscuity of binding. Reconstitution studies with the inositol polyphosphate receptor preparation indicate an IP4-modulated K+ channel. The role of the receptor and its further characterization, especially the channel gating and its relevance to function, remain to be determined.