Alkali metal ion dependence of inositol 1,4,5-trisphosphate-induced calcium release from rat cerebellar microsomes.

The effects of the alkali metal ions Na+, K+, Rb+, and Cs+ on ATP-dependent Ca2+ uptake, [3H]Inositol 1,4,5-trisphosphate (InsP3) binding, and quantal InsP3-induced Ca2+ release were investigated using rat cerebellar microsomes. Both the ion species and concentration affected the ability of the microsomes to support Ca2+ uptake with K+ being mot effective (3.8 nmol of Ca2+/min/mg at 100 mM K+). The order of efficacy of the other ions was as follows: K+ > Na+ > Rb+ = Cs+ >> Li+. The binding of [3H]InsP3 to cerebellar microsomes was, however, affected little by the presence of these ions. All these alkali metal ions (except Li+) supported InsP3-induced Ca2+ release at concentrations above 25 mM; however, the extent of Ca2+ release (expressed as a percent Ca2+ release compared with that released by the ionophore A23187) was dependent upon the ion species present. Again K+ was more potent than the other ions at facilitating InsP3-induced Ca2+ release (order of efficacy: K+ > Rb+ > Na+ > Cs+), although the concentration of InsP3 required to induce half-maximal Ca2+ release (IC50) was not significantly altered. Over the ion concentration range tested (25-100 mM), the extent of InsP3-induced Ca2+ release with both K+ and Rb+ increased in a linear fashion, while Na+ showed only a slight increase and Cs+ showed no increase over this range. The effect of K+ concentration on quantal Ca2+ release was to alter the extent of release rather than the IC50 InsP3 concentration. Using stopped-flow techniques, the effects of InsP3 and K+ concentrations on the kinetics of InsP3-induced Ca2+ release were shown to exhibit a monoexponential process in this microsomal preparation. The rate constants for Ca2+ release increased with InsP3 concentration (0.11 s-1 at 0.02 microM InsP3 to 0.5 s-1 at 40 microM InsP3); however, the relationship between the fractional extent of release and rate constants for release did not change in a similar way with InsP3 concentration. Although the fractional extent of Ca2+ release increased with K+ concentration, the rate constants for release over this K+ concentration range were unaffected. This observation leads us to question the role of K+ as a counter ion required for Ca2+ release, and we therefore postulate a role for K+ (and the other alkali metal ions) as a "co-factor" required for channel opening.

Inositol 1,4,5-trisphosphate (InsP 3 ) 1 is a second messenger which activates a Ca 2ϩ channel usually located within specific intracellular Ca 2ϩ stores of many cell types (1). Several studies have so far demonstrated that in order for InsP 3 -induced Ca 2ϩ release to occur univalent metal ions such as K ϩ are required (2)(3)(4)(5). An early proposal suggested that InsP 3 -induced Ca 2ϩ efflux from Ca 2ϩ stores is electrogenic and can only occur if accompanied by a K ϩ influx to counterbalance Ca 2ϩ movement, thus alleviating a buildup in membrane potential (2,3). The inhibition of InsP 3 -induced Ca 2ϩ release by K ϩ channel blockers such as tetraalkylammonium cations and nonyltriethylammonium has indicated that InsP 3 -sensitive Ca 2ϩ stores must also contain K ϩ channels (2). However, in a more detailed study where an extensive range of K ϩ channel blockers was used, it was concluded that the inhibition of InsP 3 -induced Ca 2ϩ release by these blockers was likely to be due to a direct interaction with the InsP 3 -sensitive Ca 2ϩ channel itself rather than inhibition of any putative K ϩ channels, since the addition of K ϩ ionophores valinomycin or gramicidin D, which would supply an alternative route for K ϩ influx, failed to reverse the inhibition of InsP 3 -induced Ca 2ϩ release by these K ϩ channel blockers (6,7).
From studies using permeabilized hepatocytes, Joseph and Williamson (3) showed that the chloride salts of the alkali metals K ϩ , Na ϩ , Rb ϩ , and Cs ϩ were essentially similar in supporting InsP 3 -induced Ca 2ϩ release. Their study also showed that salt concentration affected Ca 2ϩ release with optimal concentration being ϳ40 mM; however, the release process was not affected by changes in osmotic strength (3).
The fact that the purified cerebellar InsP 3 -sensitive Ca 2ϩ channel alone could also support Ca 2ϩ movements when reconstituted into sealed liposomes and exposed to InsP 3 (8 -10) must indicate that separate K ϩ channels are not required for Ca 2ϩ release, although the possibility exists that the InsP 3sensitive Ca 2ϩ channel may also have intrinsic K ϩ channel activity (11,12). One alternative possibility could be that alkali metal ions may stimulate the InsP 3 -sensitive Ca 2ϩ channel by acting as "co-factors" stimulating particular steps in the mechanism of channel opening instead of/or as well as acting as a counter-ion (13). This hypothesis therefore needs further investigation.
It is clear that alkali metal ions such as K ϩ are essential for InsP 3 -induced Ca 2ϩ efflux, but it is not clear as yet how these ions cause this activation. In this study we undertake a detailed investigation to characterize the effects of alkali metal ions on the cerebellar InsP 3 -sensitive Ca 2ϩ channel, by investigating their effects on [ 3 H]InsP 3 binding and the kinetics of quantal InsP 3 -induced Ca 2ϩ release.

MATERIALS AND METHODS
Fluo-3 was obtained from Sigma, InsP 3 from Calbiochem, and [ 3 H]InsP 3 from DuPont NEN. All alkali metal chlorides were obtained from Aldrich.
Rat cerebellar microsomes were prepared essentially as described in Refs. 14 and 15 with minor modifications. Briefly, 15-20 cerebella were minced and homogenized in 10 volumes of cold buffer (0.32 M sucrose, 5 mM Hepes, 0.1 mM phenylmethylsulfonyl fluoride, 0.03 mM benzami-dine, 5 M leupeptin A, and 10 M pepstatin, pH 7.2). The homogenate was centrifuged at 500 ϫ g for 10 min. The resulting pellet was homogenized in 5 volumes of buffer and again centrifuged at 500 ϫ g for 10 min. The supernatants were pooled and the mitochondria removed by centrifugation at 10,000 ϫ g for 20 min. The microsomal pellet was obtained by centrifuging the remaining supernatant for 1 h at 100,000 ϫ g. The pellet was resuspended in the Hepes-sucrose buffer to a concentration of ϳ15 mg/ml, snap-frozen in liquid nitrogen, and stored at Ϫ70°C.
Ca 2ϩ uptake and InsP 3 -induced release from cerebellar microsomes were measured using fluo-3 as described elsewhere (16), with some modifications. Rat cerebellar microsomes (0.3 mg/ml) were suspended in a buffer containing Tris phosphate (40 mM), creatine kinase (10 g/ml), phosphocreatine (10 mM, Tris salt), fluo-3 (250 nM), and the appropriate concentration of alkai metal ion as the chloride salt, pH 7.2, at 37°C. Ca 2ϩ uptake was initiated by the addition of 1.5 mM Mg-ATP and the fluorescence change monitored on a Perkin-Elmer LS-50B spectrofluorimeter, with excitation at 506 nm and detecting the emission at 526 nm. After ATP-dependent Ca 2ϩ loading, further Ca 2ϩ uptake was inhibited by the addition of between 0.2 and 0.5 mM sodium orthovanadate (which inhibited Ͼ90% of the Ca 2ϩ pumps (7)) and InsP 3 at the appropriate concentration was added. Total Ca 2ϩ accumulated within the microsomes was measured by permeabilization with Ca 2ϩ ionophore A23187 (12.5 g/ml).
Fluorescence intensity was related to [Ca 2ϩ ] by the following equation given in Ref. 16, where K d is the dissociation constant for Ca 2ϩ binding to fluo-3 at pH 7.2 and 37°C. F is the fluorescence intensity of the sample and F min and F max are the fluorescence intensities of the sample in 1 mM EGTA and 2.5 mM CaCl 2 , respectively. Under standard conditions (100 mM KCl, pH 7.2 and 37°C) we have shown the dissociation constant for Ca 2ϩ binding to fluo-3 to be 900 nM (17); however, both the alkali metal ion present and its concentration also affect this dissociation constant. The dissociation constants for fluo-3 binding to Ca 2ϩ were therefore measured in the appropriate alkali metal salt at a variety of concentrations by monitoring the change in fluorescence in a 10 mM Hepes/Tris buffer, pH 7.2, 37°C using 250 nM fluo-3 and varying the free Ca 2ϩ concentration by known concentrations EGTA and CaCl 2 as calculated by the "ION" computer program developed by Fabiato (18). It must be noted that the pH of the EGTA and CaCl 2 solutions were adjusted with Aristar Tris (from BDH) to avoid alkali metal ion contamination.
The rapid measurements of InsP 3 -induced Ca 2ϩ release were monitored using an Applied Photophysics stopped-flow spectrofluorimeter (model SX 17 MV), exciting the sample at 505 nm and measuring the emission above 515 nm using a cut-off filter. The microsomes were in the same buffer as described previously and Ca 2ϩ accumulation was followed on a conventional spectrofluorimeter. Once sufficient Ca 2ϩ loading had occurred, further accumulation was inhibited with orthovanadate and microsomal/fluo-3 suspension added to syringe A of the stopped-flow apparatus. Syringe B was filled with InsP 3 at 10 times the experimental concentration required as the mixing ratio of syringe A to B was 10:1, to avoid introducing substantial Ca 2ϩ contamination when mixing the microsomes/fluo-3 suspension with InsP 3 . The fluorescence data were initially adjusted by comparing the changes on the stopped-flow apparatus with identical experiments undertaken on a conventional fluorimeter, such that these traces could be related to fractional Ca 2ϩ release. These traces were then analyzed using nonlinear regression analysis programs supplied by Applied Photophysics. The time courses for Ca 2ϩ release for the microsomal preparation used in this study could be fitted well to a monoexponential process using the following equation, where A is the fractional amount or extent of release and k is the rate constant which defines this release process. Maximal amount of Ca 2ϩ release (1.0) was defined as that released by 40 M InsP 3 . Over the Ca 2ϩ concentration range in which InsP 3 -induced release was monitored, the fluorescence changes, when related to Ca 2ϩ concentrations, were around the K d value for Ca 2ϩ binding to fluo-3. Over this range of fluorescence is linearly related to Ca 2ϩ concentration (linear regression coefficient r Ͼ 0.99). The binding of [ 3 H]InsP 3 to cerebellar microsomes was carried out as described in Refs. 14 -16. 0.5 mg of rat cerebellar microsomes was suspended in 0.5 ml of buffer containing 50 mM Tris/HCl, pH 8.3, 1 mM EDTA, with the appropriate concentration of alkali metal chloride salt and doped with 0.02 Ci of [ 3 H]InsP 3 . Specific binding was measured at 40 nM InsP 3 (the K d value for InsP 3 binding cerebellar microsomes under our experimental conditions) and nonspecific binding measured in the presence of 10 M excess cold InsP 3 . After the addition of the microsomes to the buffer, the mixture was allowed to incubate at 4°C for 15 min. Bound InsP 3 was separated from free by centrifugation at 15,000 ϫ g for 20 min and the pellets then solubilized in 0.5 ml of Solvable (DuPont), added to Ultima flow scintillant and the radioactivity determined by liquid scintillation spectrometry.

RESULTS
Before using fluo-3 to measure InsP 3 -induced Ca 2ϩ release in the presence of different alkali metal ions at different concentrations, the effects of these ions on the dissociation constant of Ca 2ϩ binding to fluo-3 were determined. Fig. 1 shows that the K d value for Ca 2ϩ binding to fluo-3 is dependent upon the ionic concentration and type of metal ion present. The K d value varies from 230 nM in the absence of added ions to 900 nM in the presence of 100 mM KCl, the latter being identical to the previously determined value (17). Both K ϩ , Na ϩ , and Li ϩ ions affected the K d for Ca 2ϩ binding to fluo-3 in a similar fashion. The change in K d for Ca 2ϩ binding to fluo-3 in the presence of Cs ϩ and Rb ϩ at concentrations up to 100 mM salt was substantially lower than the values obtained with K ϩ (varing from 230 nM to 550 nM). In all subsequent experiments the changes in Ca 2ϩ concentrations were calculated using the appropriate K d for the metal ion and concentration used in each experiment. However, in buffers containing cerebellar microsomes, no additional affect on the K d for Ca 2ϩ binding to fluo-3 was observed.
Prior to monitoring the effects of alkali metal ions on InsP 3induced Ca 2ϩ release, the microsomes were first loaded with Ca 2ϩ by activating the microsomal Ca 2ϩ -pump (Ca 2ϩ -ATPase) with ATP. If no alkali metal ions were present in the assay buffer, little or no Ca 2ϩ uptake could be measured (Ͻ0.1 nmol of Ca 2ϩ /min/mg), also extremely poor Ca 2ϩ uptake was also observed in the presence 100 mM Li ϩ (ϳ0.2 nmol of Ca 2ϩ /min/ mg). In our system K ϩ was the most effective alkali metal ion (3.8 Ϯ 0.2 nmol of Ca 2ϩ /min/mg at 100 mM K ϩ ); however, all other alkali metal ions used (except Li ϩ ) could support Ca 2ϩpump activity to a level which could sufficiently load the microsomes with Ca 2ϩ prior to performing release experiments (Na ϩ , 3.1 Ϯ 0.2; Rb ϩ , 1.9 Ϯ 0.3; Cs ϩ , 2.0 Ϯ 0.2 nmol of Ca 2ϩ / min/mg at 100 mM alkali metal ion, respectively). We also ensured that in all experiments Ca 2ϩ accumulation into the microsomal vesicles was allowed to reach similar levels, before further uptake was inhibited by the addition of up to 0.5 mM orthovanadate. Fig. 2 shows the effects of the alkali metal ions Na ϩ , K ϩ , Rb ϩ , and Cs ϩ (all at 100 mM concentration) on quantal InsP 3induced Ca 2ϩ release. It is clear that the amount of InsP 3induced Ca 2ϩ release (measured as a percent of that releasable with A23187) is dependent upon the type of metal ion present, with K ϩ able to support the greatest amount of release (15.7% Ca 2ϩ release at maximal InsP 3 concentration). Rb ϩ was the next most potent ion (causing 11.8% InsP 3 -induced Ca 2ϩ release), while Na ϩ and Cs ϩ supported lower levels of Ca 2ϩ release (9.3 and 6.7% release, respectively). Although the concentration of InsP 3 required to cause half-maximal InsP 3 -induced Ca 2ϩ release (IC 50 ) varied between 1.0 and 2.0 M for the metal ions tested at 100 mM (Na ϩ , , the standard errors for the IC 50 values were such that no significance could be placed on these small variations. Fig. 2 also shows that in this preparation the concentration of InsP 3 required to reach maximal release differed with the type of metal ion present. Both Kϩ and Rb ϩ required ϳ10 M InsP 3 to reach maximal levels of Ca 2ϩ release, while Na ϩ and Cs ϩ appear to require lower InsP 3 concentrations (ϳ3 M) to attain their maximal levels. Table I shows that different alkali metal ions do not significantly alter the affinity of the receptor for InsP 3 , since little or no effect was observed on the amount of [ 3 H]InsP 3 bound to the cerebellar membranes in the presence of these ions when meas-ured using 40 nM InsP 3 (the K d value for InsP 3 binding to cerebellar microsomes under our experimental conditions). There was also little effect of metal ion concentration upon [ 3 H]InsP 3 binding, as measured using K ϩ . A small decrease in the amount of InsP 3 bound was observed, however, in the absence of any added metal ion (12.3 pmol/mg) compared with 100 mM K ϩ (14.7 pmol/mg). Fig. 3 shows the effects of alkali metal ion concentrations on InsP 3 -induced Ca 2ϩ release measured at 20 M InsP 3 . InsP 3induced Ca 2ϩ release increases in an essentially linear relationship with increasing K ϩ and Rb ϩ concentration over the range 25-100 mM. However, Cs ϩ has effectively reached the maximum level of InsP 3 -induced Ca 2ϩ release by 25 mM as this release remains constant over the whole concentration range tested. A small rise was observed with Na ϩ on InsP 3 -induced Ca 2ϩ release, but this appears to saturate at between 75 and 100 mM concentration. Fig. 3 shows that although the alkali metal ions K ϩ and Rb ϩ support higher levels of InsP 3 -induced Ca 2ϩ release than the other ions when measured at 100 mM ion concentration, this difference is in fact slightly reversed at the lower ion concentrations. Fig. 4 shows the effects of varying K ϩ concentration on quantal InsP 3 -induced calcium release. Increasing the K ϩ concentration increases the percent Ca 2ϩ released by InsP 3 . There also appears to be a small decrease in the IC 50 values of InsP 3 concentrations required for Ca 2ϩ release with increasing K ϩ concentration (2.0 Ϯ 0.5 M at 25 mM K ϩ , 1.9 Ϯ 0.3 M at 50 mM K ϩ , 1.4 Ϯ 0.4 M at 75 mM K ϩ , and 1.3 Ϯ 0.3 M at 100 mM K ϩ concentrations, respectively), but again when the standard errors for these values are taken into account, this appears to have little significance.
Several studies looking at the time course for Ca 2ϩ release induced by InsP 3 using permeabilized cells have shown it to be biphasic in nature, comprising a fast and slow component (19,20). Here an investigation of rapid InsP 3 -induced Ca 2ϩ release from rat cerebellar microsomes was undertaken. Fig. 5A shows the effects of increasing InsP 3 concentration from 0.02 to 40 M on Ca 2ϩ release measured using a stopped-flow spectrofluorimeter at 37°C and 100 mM KCl. The time courses for Ca 2ϩ release were plotted as fractional InsP 3 -induced Ca 2ϩ release, where maximal release was set to the percent Ca 2ϩ release observed at 40 M InsP 3 . As shown in Fig. 5A the Ca 2ϩ release data can be simply fitted to a monoexponential process (solid line). However, the data presented here could also equally well be fitted to a biexponential processes comprising two rate constants and two amplitudes where the values are similar in both cases. As a monoexponential equation could be used to fit all experimental conditions described here (i.e. varying InsP 3 and K ϩ concentrations), our analysis was confined to using the simplest mathematical function describing this process. At low InsP 3 concentrations maximal release is reached between 10 and 15 s after addition, while at high concentrations maximal release is reached after about 5 s. Fig. 5B shows that the rates and amplitudes of InsP 3 -induced Ca 2ϩ release are dependent upon the InsP 3 concentration added. The rate constants determined here appear to be 5-10-fold lower than previously re-FIG. 2. The effect of alkali metal ions on quantal InsP 3 -induced Ca 2ϩ release. Each curve represents the effect of 100 mM: q, K ϩ ; f, Na ϩ ; , Rb ϩ ; and , Cs ϩ on InsP 3 -induced Ca 2ϩ release measured as a percent of Ca 2ϩ released by InsP 3 (0.01-20 M) compared with that released by A23187 (12.5 g/ml). The data are presented as the mean Ϯ S.E. of three or more determinations.  (19,20), but are considerably faster than those obtained for the purified cerebellar InsP 3 receptor reconstituted into liposomes (8 -10). Fig.  5B also shows a variation in the relationship between the amplitudes and the rate constants with InsP 3 concentration. In the microsomal preparation used for this part of the study, the maximum amount of Ca 2ϩ release (amplitude) required ϳ1 M InsP 3 ; however, the rate constant for this process still had not reached its maximum level by 40 M InsP 3 . Fig. 6A shows the InsP 3 -induced Ca 2ϩ release time course (using 1 M InsP 3 ) at different K ϩ concentrations (25-100 mM). In some of the time courses in this figure a split time base was used to enhance the amount of data points collected within the first 2 s. A 1-ms time filter was also used to to reduce the signal-to-noise ratio. Again all the data could be fitted to a monoexponential equation. Fig. 6B shows that although the amplitude (fractional amount of Ca 2ϩ release) increases in a linear relationship with K ϩ concentration (see also Fig. 2), the rate constants for Ca 2ϩ release are essentially unaffected. DISCUSSION Fluo-3 is a commonly used fluorophore for measuring Ca 2ϩ fluxes in intact and permeabilized cells as well as subcellular fractions (14 -16, 22). The affinity of this dye for Ca 2ϩ is greatly dependent upon both ion species and ion concentration present in the medium. Most workers in the field using fluo-3 tend to use either 400 nM (21) or 900 nM (16,17) as the dissociation constant for Ca 2ϩ binding in order to calculate changes in Ca 2ϩ concentration. However, as illustrated, here the K d values can vary substantially from 225 to 900 nM dependent upon the ion present and its concentration and as such will affect the calculated free Ca 2ϩ concentrations. As pH and Mg 2ϩ concentration also affect the affinity of fluo-3 for Ca 2ϩ , 2 we must stress the importance of using the appropriate K d value for fluo-3 depending on the experimental conditions used. These observations should also serve as a warning to experimentalists attempting to draw conclusions from small differences in the calibrated Ca 2ϩ concentrations inside cells.
The rate of ATP-dependent Ca 2ϩ uptake into cerebellar microsomes is dependent upon the ion species and concentration. K ϩ was the best ion for Ca 2ϩ uptake being twice as effective as Rb ϩ at the same concentration. Li ϩ was extremely poor at eliciting uptake (approximately 5% of the rate as that for K ϩ ), and therefore no InsP 3 -induced Ca 2ϩ release experiments were undertaken with this ion. We noted that at least 20 mM alkali metal ion concentration was required to attain a sufficient level of Ca 2ϩ uptake into the microsomes in order for Ca 2ϩ release experiments to be undertaken. These results are comparable with those observed by Muallem et al. (2) using rat liver microsomes and probably relate to the fact that alkali metal ions, in particular K ϩ , stimulate the microsomal Ca 2ϩ -ATPase. K ϩ ions have been shown to stimulate the sarcoplasmic reticulum SERCA1 isoform of the Ca 2ϩ -ATPase by increasing the rate of the dephosphorylation step (E2P Ϫ E2) (23). Since the kinetic properties of the endoplasmic reticulum Ca 2ϩ -ATPase are similar to the SR type (24), it is likely that K ϩ stimulation of ATP-dependent Ca 2ϩ uptake by the microsomal Ca 2ϩ -ATPase is by a similar mechanism.
K ϩ is also most effective at stimulating InsP 3 -induced Ca 2ϩ release when measured at 100 mM concentration. However, the potency of K ϩ compared with the other ions tested was diminished at concentrations below 50 mM. The stimulation of InsP 3induced Ca 2ϩ by K ϩ increased linearly with concentration up to 100 mM, which was the maximum concentration tested in this study. This observation directly contrasts with the study of Joseph and Williamson (3), which showed that in rat permeabilized hepatocytes K ϩ stimulated InsP 3 -induced Ca 2ϩ release optimally at 40 mM, while at higher concentration Ca 2ϩ , release was inhibited. The only significant effects of the metal ion species and ion concentration on quantal InsP 3 -induced Ca 2ϩ release was on the percent or extent of release, as the IC 50 values for Ca 2ϩ release with InsP 3 concentration and [ 3 H]InsP 3 binding levels were affected little. The fact that here we observe little effect of K ϩ on [ 3 H]InsP 3 binding, while in the paper by Hannaert-Merah et al. (28) shows a 2-3-fold decrease in affinity with K ϩ , most probably reflects differences in experimental conditions used in both studies (i.e. binding studies were undertaken at 4°C rather than 20°C).
In this study the rate of InsP 3 -induced Ca 2ϩ release from rat cerebellar microsomes was resolved using stopped-flow techniques and shown to be slower than earlier reports using permeabilized rat basophilic leukemia cells and rat hepatocytes 2 F. Michelangeli, unpublished observation.  (19,20), but considerably faster than the rates observed for the cerebellar InsP 3 receptor reconstituted into liposomes (8 -10). The differences in the rates of Ca 2ϩ release using different cell types may reflect differences in InsP 3 receptor isoforms present in these cells. From immunological studies using isoform specific antibodies, the cerebellum appears to express mainly the type I isoform, while hepatocytes express mainly the type II isoform (25). It is as yet unknown what isoforms are present in basophilic leukemia cells. The differences observed between the Ca 2ϩ release rates from cerebellar microsomes compared with cerebellar InsP 3 receptors reconstituted into liposomes may well reflect a difference in the receptor density in membrane vesicles between the two systems, which in turn may affect the rates of release (26). The difference in the rates observed here compared with those in other studies is unlikely to be due to any effects of Ca 2ϩ pumps, since Ͼ90% of the Ca 2ϩ pumps were inhibited prior to InsP 3 addition. We have calculated the rate constant for Ca 2ϩ uptake into the microsomes following orthovanadate inhibition to be Ͻ0.01 s Ϫ1 . Since the rate of InsP 3 -induced Ca 2ϩ release through the InsP 3 sensitive channel is much faster than the rate of Ca 2ϩ uptake in this system, especially after orthovanadate inhibition, the rate of InsP 3 -induced Ca 2ϩ release is affected little whether the Ca 2ϩ pumps were fully inhibited or 90% inhibited, since even at lowest InsP 3 concentration used the rate constant for Ca 2ϩ release was calculated to be Ͼ0.1 s Ϫ1 .
In this preparation the release process with low InsP 3 concentration reached completion after 10 s, whereas with high InsP 3 concentration (40 M) the maximal amount of release was reached after 5 s. The fact that the rates of InsP 3 -induced Ca 2ϩ increase with InsP 3 concentration can be used to argue against the model for quantal Ca 2ϩ release which assumes all-or-none release from Ca 2ϩ stores which have heterogeneous sensitivities to InsP 3 (5). (For a detailed description of the current models for quantal Ca 2ϩ see Ref. 12. As has already been made apparent by Hirose and Iino (26), this rather simplistic model would therefore imply that only the extent of Ca 2ϩ release would vary with InsP 3 concentration as this would reflect the number of stores recruited to release Ca 2ϩ , while the rate constants for Ca 2ϩ release should remain unaffected unless other factors such as variability in the channel density between distinct Ca 2ϩ stores also occurs. However, the fact that the rate of Ca 2ϩ release is more sensitive to InsP 3 concentration than the extent or amplitude of Ca 2ϩ release (illustrated by the fact that in Fig. 5B  release has reached a maximum by 1 M InsP 3 , while the rate constant still appears to increase beyond 40 M InsP 3 ) would also argue against a more elaborate version of the all-or-none model for quantal Ca 2ϩ release which assumes that the stores are not only heterogeneous with respect to their sensitivities to InsP 3 but also have heterogeneous receptor densities. In this case the concentration of InsP 3 required to reach both the maximum amount of Ca 2ϩ release and the maximum rate of Ca 2ϩ release should be the same.
In this preparation of rat cerebellar microsomes we found that InsP 3 -induced Ca 2ϩ release could be fitted assuming a simple monoexponential process at all InsP 3 concentrations and K ϩ concentrations used; however, a biexponential process with similar rate components and amplitudes for the two components could also be fitted equally well. In investigations of the rate constants of InsP 3 -induced Ca 2ϩ release using different preparations of cerebellar microsomes, we have concluded that the rate constants are consistently lower than those previously reported in other studies, using permeabilized cells (19,20). However, they are consistently similar between cerebellar microsomal preparations (varying between 0.5 and 1.7 s Ϫ1 with 20 -40 M InsP 3 ). In some microsomal preparations InsP 3induced Ca 2ϩ release can only be successfully fitted to a biexponential process consisting of two independent monoexponential components (27), while other preparations, such as the one used in this study, can be fitted equally well to a monoexponential process. As yet the reason for this variation between preparations remains unknown, although since there is also a variation in the levels of Ca 2ϩ released by InsP 3 , the IC 50 values and the cooperativity of InsP 3 -induced Ca 2ϩ release between cerebellar microsome preparations, the variability may be due to subtle differences in membrane preparations. A related phenomenon was recently reported by Hannaert-Merah et al. (28), who showed that the kinetics of InsP 3 binding and dissociation to cerebellar microsomes was either monophasic or biphasic depending on the cerebellar microsome preparation used.
From the kinetic data presented here, the only affect of varying K ϩ concentrations seems to be on the extent of Ca 2ϩ release rather than on any effects on the rate constants for release. Since it has been reported previously that there is heterogeneity between InsP 3 -sensitive Ca 2ϩ stores (5), one plausible explanation for this observation might be that these stores are also heterogeneous in their sensitivities to K ϩ , such that some stores will be able to respond to InsP 3 and release Ca 2ϩ at low K ϩ concentration while other stores require higher K ϩ concentration before Ca 2ϩ release occurs. An alternative explanation, which at present cannot be ruled out, is the possibility that as InsP 3 -sensitive Ca 2ϩ channels slowly desensitize after being opened by the addition of InsP 3 (29), K ϩ could slow down this desensitization step, thus increasing the amount of Ca 2ϩ release without necessarily affecting the rate of release. It is unlikely that the effects of K ϩ we have observed on quantal InsP 3 -induced Ca 2ϩ release are due substantially to an increase in the rate of InsP 3 dissociation from its receptor as suggested by Hannaert-Merah et al. (28).
As K ϩ concentration does not affect the rate of InsP 3 -induced Ca 2ϩ release, this may have implications in assessing the possible role of K ϩ as a counter ion. Although some studies have tried to monitor changes in 86 Rb ϩ uptake into cerebellar microsomes upon addition of InsP 3 , such changes have not been detected (6). We have also tried to monitor changes in K ϩ uptake into microsomes upon exposure to InsP 3 , using flame spectrophotometry with no success. Since the cerebellar InsP 3 receptor can be purified and reconstituted into sealed vesicles and still retain Ca 2ϩ channel activity (8 -10), this must imply that if K ϩ ions are required as a counter ion during Ca 2ϩ release then the channel itself must be an antiporter allowing Ca 2ϩ to flow in one direction, while K ϩ moves in the opposite direction (11). Although several studies using electrophysiological approaches have shown the InsP 3 receptor to be weakly permeable to both K ϩ and Na ϩ (30, 31), they have only been shown to move in the same direction as Ca 2ϩ and not in the opposite direction as would be required here. If the InsP 3 receptor was an antiporter, then by analogy with other co-transporters such as the the Na ϩ /Ca 2ϩ exchanger (32,33), changing the concentration of one ion should have a direct effect on the rate at which the other ion is transported as long as neither are at saturating concentrations. However, the fact that changing the concentration of K ϩ has no effect on the rate of Ca 2ϩ release (measured at InsP 3 concentrations where the rate of Ca 2ϩ release is not maximal) must imply that K ϩ is unlikely to be acting as a counter ion for Ca 2ϩ release. This therefore leaves us to postulate a more direct role for K ϩ and the other alkali metal ions in affecting the mechanism of channel opening, possibly by acting as a co-factor. This possibility of the InsP 3 receptor containing a putative K ϩ binding site which affects channel function obviously requires further investigation.