Characterization of Inositol 1,4,5-Trisphosphate Receptors and Calcium Mobilization in a Hepatic Plasma Membrane Fraction*

The distribution of hepatic binding sites for the cal- cium-mobilizing second messenger, inositol 1,4,5-tris-phosphate (Ips), was analyzed in subcellular fractions of the rat liver by binding studies with ['"P]IP3 and compared with the Ca2+ release elicited by IPS in each fraction. Three major subcellular fractions enriched in plasma membrane, mitochondria, and endoplasmic re- ticulum were characterized for their 5'-nucleotidase, glucose-6-phosphatase, succinate reductase, and an- giotensin I1 binding activities. The fraction enriched in plasma membrane showed 7- and 20-fold increases in IPS binding capacity over those enriched in endoplasmic reticulum and mitochondria, respectively, and contained a single class of high-affinity binding sites with Kd of 1.7 f 1.0 nM and concentration of 239 f 91 fmol/mg protein. IPS binding reached equilibrium in 30 min at 0 "C, and the half-time of dissociation was about 15 min. The specificity of the IP3 binding sites was indicated by their markedly lower affinities for inositol 1-phosphate, phytic acid, fructose 1,6-bisphos-phate, 2,3-bisphosphoglycerate, and inositol 1,3,4,5- tetrakisphosphate. and Incubations were performed, unless specified, for 30 min at 0 "C in a final volume of 500 pl with ["P]IP3 (-20,000 cpm - 100 fmol - 0.3 nM). Nonspecific binding was determined in the presence of 1 p~ IP3. Incubations were ter- minated by filtration through presoaked glass-fiber filters (Whatman GFP) and rapid washing with 2.5 ml of ice-cold incubation medium. The receptor-bound radioactivity was analyzed by liquid scintillation spectrometry. Ca'+-mobilization Studies-Subcellular fractions (3-10 mg of protein) were incubated in Ca'+-mobilization buffer (Buffer A comple- mented with an ATP-regenerating system: 10 mM creatine phosphate and 20 IU/ml creatine phosphokinase). The free Ca" concentration of the medium was monitored using Fura-2 free acid (3 phi) in a Perkin-Elmer LS-5 fluorescence spectrophotometer; the excitation wavelength was 340 nm (slit 10 nm) and the emission was recorded at 500 nm (slit 10 nm). Ca2+ uptake and release were measured under constant agitation at 37 "C in a final volume of 2 ml. Oligomycin (2.5 pg/ml) was added to block mitochondrial ATPase. ATP, IP3, and other effectors were added in small volumes. Each record was cali-brated by the addition of known amounts of Ca'+ (CaC03) to the mixture. The actual Ca2+ concentration of the medium was calculated from the maximal fluorescence (Fmm) and autofluorescence (Fmin) values obtained by adding excess Ca2+ and Mn2+, respectively, after with inomycin

The distribution of hepatic binding sites for the calcium-mobilizing second messenger, inositol 1,4,5-trisphosphate (Ips), was analyzed in subcellular fractions of the rat liver by binding studies with ['"P]IP3 and compared with the Ca2+ release elicited by IPS in each fraction. Three major subcellular fractions enriched in plasma membrane, mitochondria, and endoplasmic reticulum were characterized for their 5'-nucleotidase, glucose-6-phosphatase, succinate reductase, and angiotensin I1 binding activities. The fraction enriched in plasma membrane showed 7and 20-fold increases in IPS binding capacity over those enriched in endoplasmic reticulum and mitochondria, respectively, and contained a single class of high-affinity binding sites with Kd of 1.7 f 1.0 nM and concentration of 239 f 91 fmol/mg protein. IPS binding reached equilibrium in 30 min at 0 "C, and the half-time of dissociation was about 15 min. The specificity of the IP3 binding sites was indicated by their markedly lower affinities for inositol 1-phosphate, phytic acid, fructose 1,6-bisphosphate, 2,3-bisphosphoglycerate, and inositol 1,3,4,5tetrakisphosphate. The Ca"+-releasing activity of IP3 in the subcellular fractions was monitored with the fluorescent indicator, Fura-2.
All three fractions showed ATP-dependent Ca2+ uptake and rapidly released Ca2+ in response in IPS. The fraction enriched in plasma membrane was the most active in this regard, releasing 174 f 67 pmol Ca2'/mg of protein compared to 46 f 10 and 48 f 7 pmol/mg protein for the fractions enriched in endoplasmic reticulum and mitochondria, respectively. These data suggest that the [""PIIP, binding sites represent specific intracellular receptors through which IPS mobilizes Ca2+ from a storage site associated (or co-purifying) with the plasma membrane of the rat liver. It is likely that a specialized vesicular system (to which IP, can bind and trigger the release of Ca2+) is located in close proximity with the plasma membrane and is thus adjacent to the site at which IP, is produced during stimulation of the hepatocyte by Ca"+-mobilizing hormones. stimulated phosphodiesteratic cleavage of phosphatidylinosi-to1 4,5-bisphosphate is a major intracellular messenger which serves to increase the cytosolic Ca2+ concentration during hormone action. Recently, IP, has been shown to release calcium from non-mitochondrial stores in a wide variety of cells including hepatocytes (1)(2)(3)(4)(5), smooth muscle cells (6)(7)(8)(9), neuronal cells (10, l l ) , pituitary cells (12,13), and many others (for review, see Ref. 14). IP, acts through specific intracellular receptors that have been demonstrated in the adrenal cortex (15,16), macrophages (17), liver (18), neutrophils (19), brain (20) and anterior pituitary gland (21) by direct ligand-binding studies. An important question about the mechanism of action of IP3 is the nature and location of the pool from which Ca2+ is mobilized. Currently, the endoplasmic reticulum is believed to be the major source for release of calcium from intracellular stores. However, IPS may also promote calcium mobilization and/or calcium influx at the level of the plasma membrane during its action on the redistribution of intracellular calcium. We have recently observed that IP3 binding sites are relatively abundant in a plasma membrane-enriched preparation, consistent with an action of IP3 at this location (16). Since liver is readily fractionated into enriched organelles, and has been widely used for the study of calcium-mobilizing hormones (22-26 ; for review, see Ref. 27), it was employed for a more detailed analysis of the subcellular distribution of IPS binding sites. In this study, the fraction enriched in plasma membrane was found to be highly enriched in IP, receptors with binding properties similar to those recently described in adrenal cortex (16) and anterior pituitary gland (21). Significantly, the abundance of IP, receptor sites in the plasma membrane fraction was correlated with high IP3-induced Ca2+ release in this fraction. These results indicate that the vesicular system to which IP, binds and releases Ca2+ is closely associated with the plasma membrane and co-purifies with it during subcellular fractionation procedures.

IP3 Receptors and Calcium Release in Liver Fractions
through cheesecloth, and centrifugation at 1500 X g for 20 min, the pellet was resuspended in homogenization buffer and adjusted to 44% in sucrose. From that point, the plasma membrane fraction was purified according to the method of Neville (29) up to step 11. This fraction was washed and resuspended in Buffer A (the homogenization buffer without EGTA) at a concentration of 20-30 mg of protein/ ml (30). The mitochondrial fraction was prepared by centrifugation of the 1,500 X g supernatant at 8,000 X g for 10 min. The pellet was then resuspended in Buffer A, sedimented at 8,000 X g for 10 min and taken up at a high protein concentration (-30 mg/ml) in Buffer A. The microsomal fraction was prepared by centrifugation of the 8,000 X g supernatant at 35,000 X g for 20 min. This pellet was resuspended in Buffer A, recentrifuged at 35,000 X g for 20 min and taken up in Buffer A. The subcellular fractions were used freshly for Ca2+-mobilization studies or were frozen in small aliquots for binding or enzymatic analysis.
Churacterization of SubceUular Fractions-Glucose-6-phosphatase was measured at 37 "C by the method of Salomon et al. (31), 5'nucleotidase by the method of Bramley and Ryan (32). and succinatereductase according to Pennington (33). lZ6I-Angiotensin I1 binding was performed for 30 min at room temperature in a buffer containing: 20 mM Tris/HCl, pH 7.4, 100 mM NaCl, 5 mM MgCl,, 0.2% bovine serum albumin, and 0.5 mM dithiothreitol. The bound and free ligand were separated by vacuum filtration through glass-fiber (GF/c) membranes.
IPS Binding Studies--Rat liver fractions containing 200-OC3 pg of protein were incubated in a medium of the following composition: 25 mM NazHPO,, pH 7.4, 100 mM KC1, 20 mM NaCl, 0.1% bovine serum albumin, and 1 mM EDTA. Incubations were performed, unless otherwise specified, for 30 min at 0 "C in a final volume of 500 pl with ["P]IP3 (-20,000 cpm -100 fmol -0.3 nM). Nonspecific binding was determined in the presence of 1 p~ IP3. Incubations were terminated by filtration through presoaked glass-fiber filters (Whatman G F P ) and rapid washing with 2.5 ml of ice-cold incubation medium. The receptor-bound radioactivity was analyzed by liquid scintillation spectrometry.
Ca'+-mobilization Studies-Subcellular fractions (3-10 mg of protein) were incubated in Ca'+-mobilization buffer (Buffer A complemented with an ATP-regenerating system: 10 mM creatine phosphate and 20 IU/ml creatine phosphokinase). The free Ca" concentration of the medium was monitored using Fura-2 free acid (3 p h i ) in a Perkin-Elmer LS-5 fluorescence spectrophotometer; the excitation wavelength was 340 nm (slit 10 nm) and the emission was recorded at 500 nm (slit 10 nm). Ca2+ uptake and release were measured under constant agitation at 37 "C in a final volume of 2 ml. Oligomycin (2.5 pg/ml) was added to block mitochondrial ATPase. ATP, IP3, and other effectors were added in small volumes. Each record was calibrated by the addition of known amounts of Ca'+ (CaC03) to the mixture. The actual Ca2+ concentration of the medium was calculated from the maximal fluorescence (Fmm) and autofluorescence (Fmin) values obtained by adding excess Ca2+ and Mn2+, respectively, after treatment with inomycin (1 p~) .

Characterization of Hepatic Subcellular
Fractions-Since modifications of the previously described procedures (4,29) were employed for separation of the liver subcellular fractions, enzymatic characterization of each preparation was performed. As shown in Table I, the plasma membrane fraction was the most enriched in 5'-nucleotidase activity (a marker for plasma membrane) and the least enriched in glucose-6phosphatase activity (a marker for endoplasmic reticulum). On the other hand, as expected, the microsomal fraction showed an exactly opposite pattern of enrichment. The mitochondrial fraction, although the most enriched in succinate reductase activity, showed substantial contamination with both glucose-6-phosphatase and 5"nucleotidase activities. The angiotensin I1 binding capacity, which provided a further marker of the plasma membrane, was consistent with the data obtained by enzyme assays.
The three subcellular fractions were analyzed for their IP3 binding properties. As shown in Fig. 1 (see also Table I), the plasma membrane fraction showed a much higher binding capacity for IPS (6 and 20 times higher than the fractions enriched in endoplasmic reticulum and mitochondria, respectively). The upper leftpanel shows typical displacement curves obtained during inhibition of tracer ["P]IP, binding by increasing concentrations of unlabeled IPS. In this experiment, the IP3 binding capacity of the plasma membrane fraction was 6.3 and 13 times higher than those of the fractions enriched in endoplasmic reticulum and mitochondria, respectively. In the upper right panel, the data from the same experiment are normalized to show the similar binding affinities of the three fractions for IPS. Scatchard analyses of the same data (lower panel) also indicate the similar binding affinities (0.8, 0.9, and 0.8 nM) and the marked difference between the binding capacities (146, 23, and 11 fmol/mg protein) of the plasma membrane, endoplasmic reticulum, and mitochondria-enriched fractions, respectively.
To define better the subcellular location of the binding sites for IPS, the plasma membrane fraction was further purified by application to a 13-37% sucrose gradient overlaying a 50% sucrose cushion. After centrifugation for 60 min at 550 X g, the material at the cushion interface was collected and washed in Buffer A. The 5"nucleotidase activity and IPS binding capacity of this more purified membrane fraction were further TABLE I Churacterization of subcellular fractions prepared from rat liver Angiotensin I1 binding was estimated by incubating 25 pg of each fraction with '"I-angiotensin I1 (200,000 cpm -60 fmol -0.5 nM) in the presence and absence of M unlabeled ligand. For the determination of 5"nucleotidase and glucose-6-phosphatase activities, 50 pg of each fraction was used; the amount of organic phosphate formed was measured by the method of Fiske and Subbarow (35). Mitochondrial succinate-tetrazolium reductase activity was evaluated with 50 pg of each fraction; formazan production was calculated from a value of 20.1 X lo3 for the molar extinction coefficient (490 nm) of the formazan dissolved in ethyl acetate (33). IP3 binding was evaluated by Scatchard analysis of competitive dose-displacement experiments using, respectively, 500 pg, 1 mg, and 2 mg of protein for the plasma membrane, the endoplasmic reticulum and the mitochondria-enriched fractions. These results represent the mean f standard deviation of triplicate determinations for the different activities in a single preparation. Similar results were obtained in five other preparations. Binding was reversible, and addition of M unlabeled IP3 was followed by rapid dissociation of the bound ligand with a half-time of about 15 min, according to a single exponential function as shown in the lower panel of Fig. 2. The association rate data fitted a second-order equation as shown in Fig. 3, upper panel. The association rate constant (k+l), calculated from the slope of the curve, was 2.1 X lo7 M" .min-'. The dissociation rate constant (12-J estimated by fitting the binding data to a first-order rate equation (Fig. 3, lower panel) was 0.02 min-'. The Kd calculated from the ratio of the rate constants for dissociation and association was 0.95 nM, in agreement with the value obtained from steady state binding experiments.
In saturation studies (Fig. 4) The specificity of the IP, binding sites was analyzed in competitive binding experiments performed with inositol phosphates and related compounds. As shown in Fig. 5, phytic acid and inositol 1-phosphate showed extremely low affinities for the binding site (at least 10,000-fold less than the homologous ligand). Other compounds bearing phosphate groups on vicinal carbon atoms, such as fructose 1,6-bisphosphate and 2,3-bisphosphoglycerate (an inhibitor of the enzyme IP3 5'phosphatase) (36, 37), showed extremely low binding affinities. The most closely similar compound, inositol 1,3,4,5tetrakisphosphate, showed relatively low affinity (about 5% compared with IP,) for the binding site. The shapes of the displacement curves and their parallelism with the IPS binding-inhibition curve were consistent with competitive interaction of these compounds with the IPS binding site.
Ca2+-rekasing Activity of IP3-Ca2+ uptake and release in response to IP3 were determined in each of the subcellular fractions. The plasma membrane fraction showed ATP-dependent Ca2+ sequestering activity (which was inhibited by vanadate), and decreased the ambient Ca2+ concentration to about 500 nM (Fig. 6, panel A). This relatively high value should not be considered aa the "set point" but was probably due to the limited capacity of the vesicular system in comparison to the large amount of Ca2+ in the medium. When the ambient calcium concentration was reduced by adding a small amount of EGTA (-10 PM) or if the plasma membrane preparation was added in higher concentration (10 mg of protein), the ATP-dependent Ca2+ uptake process lowered the Ca2+ level to below 200 nM (data not shown).
Addition of IP, (2.5 PM) caused immediate release of 0.4 nmol of Ca2+ followed by slow re-uptake. Subsequent stimulation with IP, evoked a comparable response, although with repeated additions the response diminished in magnitude. Addition of 1 PM ionomycin immediately released the accumulated Ca2+, indicating the vesicular nature of the Ca2+ sequestering process. The fractions enriched in endoplasmic reticulum (Fig. 6, panel B ) and mitochondria (Fig. 6, paneE C) also showed high ATP-dependent Ca2+ sequestering activities, decreasing the ambient Ca2+ concentration to about 200 nM. However, these preparations showed smaller responses to IP3 (2.5 FM) stimulation, releasing 0.13 and 0.07 nmol of Ca2+, Again, since the plasma membrane preparation was the most active, this fraction was used to further characterize the action of IP, in the liver. The dose-response relationship between IPS and Ca2+ mobilization is shown in Fig. 7. The threshold response was elicited by 50 nM IP, and the EDso for Ca2+ mobilization was 257 f 61 nM (n = 4), comparable to values observed in other laboratories (1,4).
To determine whether the effect of IP, on Ca2+ release was influenced by GTP, as previously observed in liver microsomes (39), we also examined the influence of guanyl nucleotides on IP, action and Ca2+ release in the plasma membrane fraction. In our hands, GTP (100 p~) had no effect on the IP3-induced Ca2+ release and was unable by itself to release Ca2+ from the liver plasma membrane preparation (see Fig. 8,  panel A ) . However, when the incubation medium was complemented with 3% PEG (Fig. 8, panel B ) , GTP alone was found to release Ca2+. The amount of Ca2+ released (467 pmol/mg protein) by a maximal dose of GTP (100 p~) was larger than the amount released by IPS (250 pmol/mg protein in the experiment shown in panel A ) , although the rate of release was much slower (a few seconds to reach the peak response with IPS compared to about 5 min with GTP). The threshold response was elicited by 1 p~ GTP and the EDb0 for Ca2+ mobilization was about 7 p~ (not shown). Gpp(NH)p, a nonhydrolyzable analog of GTP, did not release Ca2+ at concentrations up to 250 p~, but was able to block the effect of low doses of GTP (not shown). In the presence of PEG, doses of GTP that elicited small increments in Ca2+ release did not influence the effects of low doses of IPS. As shown in Fig. 8 (panel C ) , 0.25 p~ IP3 released 120 pmol of Ca2+ in the presence of GTP (2 p~) and released 130 pmol of Ca2+ in its  absence. This observation, together with the finding that IP3 could still release substantial amounts of Ca2+ after GTP has produced its maximal effect (see panel B ) , suggests that IP, and GTP exert their actions through distinct mechanisms, as also indicated by other recent studies (11,40,41).

DISCUSSION
The messenger action of IP, is expressed through its interaction with intracellular receptors that have recently been characterized in several different tissues. The intracellular location of this receptor is an important aspect of the mechanism of action of IP,. This question was addressed by Dawson and Imine (4) in the liver by studies which showed that IPS releases Ca2+ from a vesicular system other than mitochondria. Attempts to purify a microsomal fraction caused loss of the IPS effect, suggesting that the IP,-sensitive Ca2+ pool is a specialized part of the endoplasmic reticulum which co-purifies with heavier cellular organelles.  . (18) showed enrichment of IP, binding sites in the microsomal as compared to the cytosolic and mitochondrial fractions of rat liver, but the plasma membrane fraction was not examined. The present data clearly demonstrate that the plasma membrane fraction is highly enriched in IP, receptors with binding properties similar to those present in the adrenal cortex (16) and anterior pituitary gland (21). Our finding that the plasma membrane fraction contains the highest level of IP,-induced Ca2+ release activity also lends further significance to the presence of membrane binding sites for IP,.
Comparison of the subcellular distributions of IP, binding activity and IP3-induced calcium-release activity required analysis of each of these functions by procedures that differed widely in their requirements for optimal assay conditions. Also, the conditions required for the preparation of optimally responsive microsomes for Ca2+-release studies differ considerably from those employed for liver fractionation. The homogenization medium is iso-osmotic, containing Hepes buffer, EGTA, and dithiothreitol, whereas a hypotonic medium containing 1 mM NaHC0, is used for plasma membrane purification by the Neville procedure (29). Furthermore, the subcellular preparations are relatively unstable and rapidly lose their Ca2+ mobilizing properties. For these reasons, we employed methods that allow rapid fractionation of the liver tissue, and the purity of our routine subcellular fractions was somewhat less than that attainable by more rigorous fractionation schemes. However, when additional purification of the plasma membrane fraction was performed by sucrose density gradient centrifugation, concomitant enrichment of 5"nucleotidase activity and IPS binding capacity was again observed.
A current question of major interest is whether IPS serves to control Caz+ influx directly across the plasma membrane, and/or its release from a calcium-containing structure associated with the plasma membrane, as well as from the endoplasmic reticulum. IP, was recently found to activate Ca2+ channels in the plasma membrane of T-lymphocytes (42), consistent with Michell's (43) original proposal that phospholipid breakdown could regulate the influx of external Ca2+ through transmembrane channels. In addition to its possible entry through IP,-regulated calcium channels, our data show that Ca2+ is released by IP3 from vesicular structures after uptake by an ATP-dependent process (which is inhibited by vanadate and can be completely reversed upon addition of a calcium ionophore). Although such structures could represent inside-out vesicles originating from the plasma membrane during homogenization, this would not account for the Ca2+releasing activity of IP, in a wide variety of permeabilized cells (8,(44)(45)(46) including hepatocytes (1-3, 47) which do not contain inside-out vesicles. The most likely explanation is that IP, binds to and triggers the release of CaZ+ from a vesicular system that is closely associated with the plasma membrane, and which purifies with the membrane when subcellular fractions are prepared.
Several recent reports have provided evidence for a mechanism whereby direct entry of extracellular calcium into an IP3-sensitive pool contributes to receptor-operated calcium influx and is responsible for sustained calcium entry in agonist-stimulated cells (48)(49)(50)(51)(52). The IP3-sensitive pool, with its associated IPS receptors and calcium releasing mechanism, would thus be expected to be adjacent to the plasma membrane. At least three current models of receptor-activated calcium entry involve a close apposition between IP,-sensitive regions of endoplasmic reticulum and the plasma membrane. In Putney's (52) capacitative scheme, agonist-induced emptying of calcium from a component of the endoplasmic reticulum (the receptor-regulated calcium pool) serves as a signal for calcium entry by promoting its transport from the sub-plasmalemma1 space into the endoplasmic reticulum. A more complex model developed by Gill and colleagues (53) to explain the role of GTP in calcium movements involves the formation of junctional processes between adjacent membranes, including connections between the plasma membrane and the endoplasmic reticulum. Likewise, the model proposed by Irvine and Moor (54) for the conjoint actions of IP, and IP, on calcium entry also includes functional coupling between plasma membrane and endoplasmic reticulum, in this case promoted by Ins-1,3,4,5-P4, to permit direct entry of external calcium into the endoplasmic reticulum.
It has not been established whether distinct structural associations between the plasma membrane and endoplasmic reticulum, which could represent the membrane-associated vesicular system implicit in the above models and suggested by the present findings, are present in non-excitable cells. However, the presence of associations between the plasma membrane and sarcoplasmic reticulum in smooth muscle (55) and endoplasmic reticulum in platelets (56), as well as the recent proposal that specialized organelles (calciosomes ) are involved in calcium mobilization (57), suggest that discrete perimembrane structures may play a general role in the process of calcium entry and distribution within the cell. Such structures may be present in plasma membrane vesicle fractions from human platelets (56) and rat parotid acini (58), both of which have been shown to release Ca'+ in an IP,-sensitive manner.
The IP, binding sites observed in the plasma membrane fraction of the rat liver possess all the characteristics of a true receptor. IP, binds to a particulate component and the level of binding is proportional to the amount of membrane protein.
The binding is rapidly reversible as expected for a normal ligand-receptor interaction, and this would permit rapid termination of the response. The rapid rates of association and dissociation are consistent with the calcium mobilizing activity of IP,. The binding sites are saturable and binding of the ligand is competitive, being inhibited by increasing concentrations of unlabeled ligand. The binding is also highly specific for IP,; inositol-derived compounds and organic compounds bearing phosphoryl groups on vicinal carbons showed weaker affinities for the binding site. Finally, the important criterion of a functional response is demonstrated by the ability of IP, to release Caz+ from the same membrane preparation used for the binding experiments.
One question arising from our study is the high affinity of the binding site by comparison with the potency of IP, for releasing Ca*+. Such a discrepancy was also observed in the adrenal cortex (16) and anterior pituitary gland (21), and may be largely attributable to the divergent experimental conditions necessary for the two assays. In particular, the optimal conditions for IP, binding are very different from those within the cell. Also, the degradation of added IP, by the plasma membrane fraction at 37 "C is very rapid and could lower the actual IP3 concentration during Ca2+ release studies. Furthermore, ATP inhibits the binding of IP, by a mechanism that is not yet clear, and its presence at millimolar concentration in the Ca2' release assays could substantially lower the potency of IP,. Thus, the discrepancy probably results from the combination of higher apparent affinity in the binding assay, due to choice of optimal binding conditions, and lower apparent affinity in the Ca'+ release assays, due to rapid degradation of IP, at 37 "C in the presence of Mg2+ together with the competitive effect of ATP. The difference might also reflect decreased efficiency of the calcium-gating mechanism due to loss or impairment of a putative regulatory component during preparation of the subcellular fractions.
One such component could be a guanine nucleotide regulatory protein associated with the calcium release mechanism. However, the effect of GTP on Caz+ release from the liver plasma membrane fraction was independent of IP,, as previ-ously observed in the adrenal cortex' and the anterior pituitary gland (21). In several tissues, GTP seems to act through a mechanism distinct from IP3, and in parotid acini the GTPsensitive and IP,-sensitive calcium pools appear to reside in different regions of the endoplasmic reticulum (58). Thus, further studies are needed to clarify the importance of this effect of GTP in the regulation of Ca'+ mobilization during hormonal stimulation.
It should be noted that two previous studies using subcellular fractions of rat pancreas (59) and rat adipocytes (60) showed that IP, mobilized Ca2+ from endoplasmic reticulum but not from plasma membrane. Whether this is a reflection of differences between specific tissues or among fractionation procedures remains to be clarified. We are attempting to answer this question by analysis of IP3 binding in adrenal cortex and anterior pituitary plasma membranes to establish the intracellular site of action of IP, in these tissues. However, the presence of IPS binding and/or IPS-induced calcium mobilization in hepatic, platelet (56), and parotid acinar (58) plasma membrane fractions suggests that the IP,-sensitive calcium pool may be closely related to the plasma membrane in a wide variety of target cells that respond to calciummobilizing stimuli.