Intracellular Calcium Translocation: Mechanism of Activation by Guanine Nucleotides and Inositol Phosphates

It is now well recognized that Ca²⁺ plays a pivotal regulatory role within cells, both as an intracellular mediator of receptor-activated signalling, and in the control of a multitude of cellular processes notable among which is the secretory event. The recent elucidation of the mechanisms coupling cell-surface receptors to Ca²⁺ mobilization in cells, based on the early observations of Hokin & Hokin (1953), has now established in principle the relationship between receptor-induced phosphoinositide breakdown and inositol phosphate-mediated Ca²⁺ release (Berridge & Irvine, 1984; Gill, 1985; Majerus et al. 1986; Berridge, 1987). In spite of the fact that much is now known about the phosphoinositide signalling pathway, it should be noted that the regulation of Ca²⁺ within cells involves a complex set of events. Thus, Ca²⁺ signalling occurs through the subtle alteration of one or more of an array of distinct transport mechanisms, located in a number of discrete organelles, and influenced by numerous intracellular regulatory systems. It is the purpose of this chapter to review some intriguing recent developments concerning the control of intracellular Ca²⁺ movements and their possible relationship to what has been ascertained on the processes that mediate Ca²⁺ signalling events within cells. In the first section, certain of the characteristics of Ca²⁺ regulatory organelles and their role in Ca²⁺ signalling are considered.

The transfer of Ca²⁺ across membranes within cells is controlled by a number of distinct classes of active or passive transport mechanisms (see Carafoli, 1987). The cytosol of most mammalian cells contains approximately 0·l μmoll⁻¹ free Ca²⁺ under resting conditions, compared with the low millimolar free Ca²⁺ concentration outside cells. This 10000-fold gradient of free [Ca²⁺] across the plasma membrane is actively maintained via ATP-dependent Ca²⁺ pumping, and perhaps also via the Na⁺/Ca²⁺ exchanger (Gill, 1982a). Ca²⁺ translocation via voltage sensitive Ca²⁺ channels is a well-established route of entry of extracellular Ca²⁺ into excitable cells and perhaps many other cell types (Miller, 1987). Moreover, it is clear now that activation of such channels can be finely controlled by intracellular messenger-mediated phosphorylation events (Tsien et al. 1986; Miller, 1987). In addition, many have considered that Ca²⁺ entry across the plasma membrane may be directly mediated by activation of channels distinct from voltage-sensitive Ca²⁺ channels (Gill, 1982a; Tsien et al. 1986; Miller, 1987). The existence and characterization of such channels has not been conclusively described. However, it seems clear that at least the prolonged responses to many Ca²⁺-coupled receptors are dependent on external Ca²⁺ and may involve entry of Ca²⁺ across the plasma membrane (Putney, 1986), as discussed later.

It has become increasingly clear that, in addition to the plasma membrane, internal organelles also play an important role in the maintenance of cytosolic [Ca²⁺], Mitochondria are known actively to accumulate Ca²⁺ (see Hansford, 1985) via a process dependent on the membrane potential existing across the internal membrane. However, from most observations it appears that mitochondria can only accumulate Ca²⁺ when free Ca²⁺ levels are high, that is, at or above1–10μmoll⁻¹; thus it is unlikely that they contribute directly either to the maintenance of physiological cytosolic Ca²⁺ levels or to the induction of Cam signalling events within cells. In contrast, it appears certain that other Cam accumulating organelles within cells are active in both respects. Thus, endoplasmic reticulum (ER) in a variety of cell types has been observed to sequester large quantities of Ca²⁺ (Henkart, Reese & Brinley, 1978; McGraw, Somlyo & Blaustein, 1980; Wakasugi et al. 1982; Burton & Laveri, 1985). Using permeabilized nonmuscle cells, it is clear from a number of different studies that nonmitochondrial organelle(s) exist which accumulate Ca²⁺via high-affinity (ATP + Mg²⁺)-dependent Ca²⁺ pumping activity (see, for example, Burgess et al. 1983; Gill & Chueh, 1985). Such internal Ca²⁺ pumps are analogous in function to those of the plasma membrane. However, a number of features distinguish the internal and plasma membrane pumping activities (Gill & Chueh, 1985). Interestingly, these distinguishing characteristics are remarkably consistent with those features which serve to distinguish sarcolemmal and sarcoplasmic reticulum (SR) Ca²⁺ pumps in muscle tissue (Carafoli, 1987). Thus, it has been suggested that ER in nonmuscle cells may fulfil at least some of the specialized Ca²⁺ regulatory functions ascribed to SR in muscle. However, although analogies exist with respect to Ca²⁺ accumulation, it is becoming increasingly apparent that the Ca²⁺ release mechanisms of SR and ER are quite distinct. It should also be noted that whereas the SR is a structurally identifiable organelle with a clearly defined Ca²⁺-regulatory function, the role of ER in Ca²⁺ signalling within nonmuscle tissues is considerably more tenuous. Thus, the involvement of ER in Cam mobilizing events is concluded from indirect evidence with, as yet, no proven localization of these mechanisms to this specific organelle. Indeed, recent evidence presented by Volpe et al. (1988) suggests that Ca²⁺-accumulating organelles which are distinct from ER may be involved in Ca²⁺-regulatory responses in cells. These organelles have been termed ‘calciosomes’ and their existence and function are described in detail in the chapter by Pozzan in this volume (Pozzan et al. 1988). In spite of the imprecise identity of Ca²⁺-releasing organelles, ER is frequently referred to as being the organelle from which Ca²⁺ release occurs in response to inositol phosphates, the actions of which are discussed next.

Considerable advances in the understanding of the nature of Ca²⁺-signalling events within cells have been derived from elucidation of the pathways for metabolism and action of the inositol phosphates derived from receptor-mediated phospholipase C activation (Berridge, 1987). An overview of the role of phosphoinositide metabolism in signal transduction and the control of secretion is given in the chapter by Putney in this volume (Putney, 1988). It is currently held that an important direct product of phosphoinositide breakdown is inositol 1,4,5-trisphosphate (together with its 1,2-cyclic derivative), and that this molecule has proven effectiveness in releasing intracellular Ca²⁺ in a large variety of cells. The metabolism of this product is complex. Although it is not the purpose of the present chapter to describe the intricate processes involved in formation and breakdown of each of the products, brief mention of the major derivatives is given here since certain of these may also have roles in modifying Ca²⁺ movements in cells. Ins(l,4,5)P₃ undergoes either phosphorylation or dephosphorylation. 5′-Phosphatase activity in cells cleaves InsP₃ to the less active Ins(l,4)P₂ product. Alternatively, 3′-kinase activity can phosphorylate InsP₃ to produce inositol 1,3,4,5-tetrakisphosphate (InsP₄), which is itself a substrate for the 5′-phospha-tase, producing in this case inositol 1,3,4-trisphosphate. Whereas the latter molecule has very much less Ca²⁺-releasing activity than Ins(l,4,5)P₃, the InsP₄ molecule has been reported to exert indirect effects on Ca²⁺ mobilization (Irvine & Moor, 1986,1987; Morris, Gallacher, Irvine & Petersen, 1987). Thus InsP₄ was observed by Irvine & Moor (1986, 1987) to induce Ca²⁺-mediated effects in oocytes; these effects appear to be dependent on the presence of InsP₃ and also to require external Ca²⁺. Interpretation of the results may imply that InsP₄ induces the entry of Ca²⁺ into the InsP₃-releasable pool, perhaps from outside the cell (Michell, 1986; Irvine & Moor, 1987). In a recent report, Morris et al. (1987) described a similar synergism between the effects of InsP₃ and InsP₄ on activation of K⁺ channels in lacrimal gland; similar conclusions on the possible permissive effect of InsP₄ on the action of InsP₃-mediated Ca²⁺ mobilization were presented. More direct synergistic effects of InsP₃ and InsP₄ on Ca²⁺ have been reported by Spat et al. (1987). Thus, it was observed that the extent of InsP₃-mediated Ca²⁺ release from liver microsomal membrane vesicles was significantly increased in the presence of InsP₄. At present, although it seems likely that InsP₄ does exert effects, it is unclear whether it may directly control Ca²⁺ fluxes, whether it modifies the InsP₃-induced release process, or whether it has indirect effects through alteration of the metabolism of InsP₃, for example by competing with InsP₃ at the 5′-phosphatase level.

Studies by Muallem, Schoeffield, Pandol & Sachs (1985) suggest that the action of InsP₃ on the release of Ca²⁺ from what is believed to be ER occurs via a process that resembles activation of a channel. This conclusion has been drawn from a number of observations including the remarkably temperature-insensitive activation of Ca²⁺ release in response to InsP₃ (Smith, Smith & Higgins, 1985; Chueh & Gill, 1986). Direct electrophysiological evidence for an InsP₃-activated channel has not yet been published; however, promising results have been discussed and more definitive studies are expected. Studies using labelled InsP₃ have identified a binding site for InsP₃ within cells, with kinetics and specificity similar to that for activation of Ca²⁺ release (Baukal et al. 1985; Spat et al. 1986; Worley et al. 1987). The isolation of an InsP₃-binding protein, which was purified by heparin affinity chromatography, has recently been reported by Supattapone, Worley, Baraban & Snyder (1988). Indeed, our own recent evidence (Ghosh et al. 1988), which shows a profound antagonistic effect of heparin on the action of InsP₃ on Ca²⁺ release from within cells, strongly suggests that the binding protein isolated by Supattapone et al. is the physiological receptor for InsP₃. Thus, whereas studies on the. molecular structure and mechanism of the site of action of InsP₃ are in their infancy, it is likely that much will come to light in the near future.

There have been a number of recent observations on a guanine-nucleotide-activated process that appears directly to activate profound and rapid movements of Ca²⁺ within many different types of cells. Below is a description of the effects of GTP on Ca²⁺ movements, their relationship to the actions of InsP₃, and the possible mechanism of activation of GTP-induced Ca²⁺ translocation. In this section we will consider the characteristics of the fluxes of Ca²⁺ activated by GTP.

During some of the earlier experiments on the action of InsP₃ in inducing Ca²⁺ release, permeabilized cell systems of several different types were found to be particularly useful for observing the effects of InsP₃ (Streb, Irvine, Berridge & Schulz, 1983; Burgess et al. 1984). In contrast, isolated microsomal membrane fractions presented some problems in permitting observations on the effects of InsP₃ (Dawson & Irvine, 1984). Such difficulties probably reflected either the lability of the InsP₃-activated release process under lengthy vesicle purification procedures, and/or a low yield of intact vesicles derived from the InsP₃-sensitive intracellular organelle. Dawson and his colleagues were approaching this problem using liver microsomes in which they had observed small effects of InsP₃ (Dawson & Irvine, 1984). In attempting to augment this response, Dawson (1985) observed that GTP enhanced the effectiveness of InsP₃, and that this effect was promoted by polyethylene glycol. Undertaking similar experiments with microsomes isolated from cultured N1E-115 neuroblastoma cells, we observed a rather different response (Ueda, Chueh, Noel & Gill, 1986). With these microsomes, addition of InsP₃ effected release of a small fraction (approximately 10%) of releasable Ca²⁺. When GTP and InsP₃ were added simultaneously, a much larger release of Ca²⁺ was observed. However, in contrast to the results of Dawson, it was observed that GTP alone was highly effective in releasing Ca²⁺ (Ueda et al. 1986). The effect of GTP was rapid and profound, more than 50% of total accumulated Ca²⁺ being released from the microsomal membrane vesicles within a few seconds. As described below, the nucleotide specificity and sensitivity of the GTP effect were remarkable. The high GTP-sensitivity was considered possible since during their isolation the microsomes had undergone considerable washing and hence were largely devoid of endogenous nucleotides. With this in mind, it was reasoned that the permeabilized cell preparations used extensively in prior Ca²⁺ flux analyses (Gill & Chueh, 1985), having been subjected to fewer washing procedures, would be a less suitable preparation on which to observe GTP-induced Ca²⁺ fluxes. However, this prediction was incorrect, and in fact the permeabilized cell preparations became the system of choice on which most of the characteristics of GTP-activated Ca²⁺ movements were determined. Using permeabilized N1E-115 neuroblastoma cells loaded with Ca²⁺ to equilibrium, the EC₅₀ for GTP was abserved to be below 1μmoll⁻¹, GTP releasing between 50 and 70% of accumulated Ca²⁺ within 30s (Fig. 1A). The effect GTP was observed to be almost as rapid as that of the ionophore A23187, although the extent of release was not as complete, an observation that suggested heterogeneity of Ca²⁺-accumulating compartments (see below).

The release of Ca²⁺ activated by guanine nucleotides observed using either permeabilized cells (Gill, Ueda, Chueh & Noel, 1986) or microsomes derived from cells (Ueda et al. 1986) has remarkably high sensitivity to GTP. The K_m for GTP measured in permeabilized N1E-115 cells is 0·75 μmol 1⁻¹. The effect also has very considerable nucleotide-specificity. Release was not observed with GMP, cyclic GMP, (either 2′,3′ or 3′,5′), or with the nonhydrolysable analogues of GTP, GTPγS or GppNHp (see Fig. 1). The latter is an important observation since it suggests a divergence in guanine nucleotide-specificity from that of the known G-proteins which are known to be much more effectively stimulated by nonhydrolysable GTP analogues. Other nucleoside triphosphates including ITP, UTP and CTP have no effect on Ca²⁺ movements, these nucleotides being largely ineffective even when added at concentrations up to 1 mmol I⁻¹ (Gill et al. 1986). Submicromolar GTP concentrations function to release Ca²⁺ in the presence of millimolar ATP concentrations (required to maintain constant Ca²⁺ pumping activity), indicating the exceptional specificity of the GTP-activated release process. It was observed that GDP does induce Ca²⁺ release, but only after a significant lag of about 30 s (Fig. 1A); thereafter it releases Ca²⁺ to approximately the same extent as GTP. Results clearly indicate that this effect results from conversion of GDP to GTP via nucleoside diphosphokinase (NDPK) activity (Ueda et al. 1986; Gill et al. 1986). Thus, the effect of GDP is blocked by ADP (Fig. 1B) which effectively competes for the nucleoside diphosphate site on NDPK (Kimura & Shimada, 1983). In fact GDP itself does not induce Ca²⁺ release; thus, GDP/8S (which is not easily phosphorylated to GTP/3S by NDPK) has no effect on Ca²⁺ release (Fig.1B). Moreover, not only is GDP without Ca²⁺-releasing effects of its own, but it actually blocks the action of GTP, as shown in Fig. 1C; (note that, at 100μmoll⁻¹, GDP saturates NDPK activity and remains present for a longer period to compete with GTP). Further experimentation (in the presence of high [ADP] to prevent conversion of GDP to GTP) revealed that the inhibitory effect of GDP was competitive with respect to GTP with a K_i) of approximately 3 μmol 1⁻¹ (Gill et al. 1986); GTPγS also blocks the effect of GTP, but rather surprisingly, GppNHp does not (Fig. 1C). This differential inhibitory action of the nonhydrolysable analogues has been a useful criterion for defining the specificity of the GTP-activated process and is referred to again later. The lack of direct action of GTPγS and its inhibitory effect on the action of GTP are evidence that GTP hydrolysis is required for the activation of Ca²⁺ release. In fact, a very slow release activated by GTPγS (Chueh & Gill, 1986) may be consistent with slow cleavage of the phosphorothioate residue which is known to occur (Eckstein, 1985). Further evidence for a GTP hydrolytic process being involved in activation of Ca²⁺ release derives from the competitive effect of GDP which indicates that either GTP or GDP can bind to the same site; presumably, the inhibitory effect of GDP arises through prevention of GDP dissociation after hydrolysis of GTP at the Ca²⁺ release-activating site.

Since in our early studies the observed effect of GTP on release of Ca²⁺ from the N1E-115 neuroblastoma cells was so profound, it was important to establish whether this effect was perhaps an anomaly restricted to this particular cell line used. Using a quite unrelated cell type, the DDTjMF-2 smooth muscle cell line derived from hamster vas deferens (Norris, Gorski & Kohler, 1974), experiments suggested this was not the case. Thus, a sensitive, specific and substantial GTP-dependent release of Ca²⁺ was observed using permeabilized DDTjMF-2 cells loaded with Ca²⁺, with pronounced effectiveness of as low as 0·1 μmol 1⁻¹ GTP in the presence of 1 mmol I⁻¹ ATP (Chueh et al. 1987). In addition to the DDT₁MF-2 cell line, we have measured almost identical effects of GTP on Ca²⁺ release using permeabilized cells from the rat BC₃H-1 smooth muscle cell line and from the human WI-38 normal embryonic lung fibroblast cell line. Using microsomal membrane vesicle fractions prepared from DDT₁MF-2 cells by methods similar to those described for N1E-115 cell-derived microsomes (Ueda et al. 1986), we have observed GTP effects on Ca²⁺ release almost identical to those seen with permeabilized cells. Furthermore, using microsomes derived from guinea pig parotid gland, Henne & Soling (1986) have observed very similar effects on release of accumulated Ca²⁺ induced by GTP. The observations of Jean & Klee (1986) on GTP- and InsP₃-mediated Ca²⁺ release from microsomes derived from NG108-15 neuroblastoma X glioma hybrid cells are also consistent with our findings.

The GTP-induced Ca²⁺ release process is specific to a nonmitochondrial Cab sequestering organelle, which may be ER or a subfraction thereof (we frequently refer to it as being ER simply for convenience). Importantly, rather clear experiments demonstrate that no effects of guanine nucleotides or InsP₃ can be observed on Ca²⁺ fluxes across mitochondrial or plasma membranes (Ueda et al. 1986; Chueh et al. 1987). The observation that less than 100% of Ca²⁺ release from ER is effected by GTP or InsP₃ suggests that only a sub compartment of ER contains the activatable efflux mechanisms. Although we have no direct proof that ER is a source of GTP-releasable Ca²⁺, interpretation of the effects of oxalate (described later), a known permeator of the ER membrane (Gill & Chueh, 1985), may indicate that ER is indeed a site of action of both GTP and InsP₃ (Chueh et al. 1987; Mullaney, Chueh, Ghosh & Gill, 1987; Mullaney, Yu, Ghosh & Gill, 1988). Moreover, we now know that GTP indeed modifies the movements of Ca²⁺ associated with the InsP₃-releasable Ca²⁺ pool and hence that GTP and InsP₃ can act on the same Ca²⁺ pool, as described below.

One of the most important areas of investigation concerns determination of the nature of the Ca²⁺ translocation process activated by GTP. With regard to this mechanism, either of two distinct possibilities appeared likely: first, GTP could activate a channel process to permit the flow of Ca²⁺ out of the organelle(s) into which Ca²⁺ is sequestered; second, GTP could activate a fusion between organelle membranes resulting in the release or transfer of Ca²⁺. In the latter case, it would be very unlikely that such a process would be reversible, that is, that the two fused membranes could be returned to the unfused state with the same original enclosed volume. Recently, we reported that GDP at least partially reverses the pride effectiveness of GTP suggesting some degree of reversibility of the action of GTP (Gill et al. 1986). Since then, a more definitive indication of the reversibility of the effect of GTP has come from a simpler study involving washing of cells after GTP-activation (Chueh et al. 1987). Thus, it has been observed that after activation of the GTP-dependent Ca²⁺ release process (with up to 100μmoll⁻¹ GTP), the effectiveness of GTP can be substantially (more than 70%) reversed by simple washing of the GTP-treated permeabilized cells with GTP-free medium. In such experiments, cells that had been treated with GTP under conditions that activate Ca²⁺ release were thoroughly washed; after this treatment Ca²⁺ uptake proceeded to an extent approaching that of untreated cells, that is, the ability of ER to accumulate Ca²⁺ was largely restored. Moreover, such GTP-pretreated, washed cells responded again to a further application of GTP, indicating that the release process can be reactivated by GTP. It would be difficult to reconcile this reversibility with a membrane fusion process activated by GTP; in other words, the effects of a direct membrane fusion event would be unlikely to be reversed by washing and result in the restoration of almost normal Ca²⁺ retention, as observed. It should be noted, however, that structural and biophysical measurements undertaken by Dawson and coworkers suggest that fusion of membranes can follow GTP treatment of microsomal vesicles (Dawson, Hills & Comerford, 1987; Comerford & Dawson, 1988). At present this question is unresolved.

Electron microscopic analysis of membrane vesicles treated with GTP has suggested that the action of GTP, although not necessarily involving membrane fusion, may be promoted by close association between membranes. It is now well established that the effects of GTP on Ca²⁺ release are promoted by 1–3% polyethylene glycol (PEG) (Chueh & Gill, 1986; Ueda et al. 1986; Gill et al. 1986). Thus, although in the absence of PEG, GTP induces a significant release of Ca²⁺, this effect is substantially increased in the presence of PEG. The effect of PEG is to increase both the sensitivity to GTP and the maximal release induced by it. Although PEG is a known fusogen when present above 25% w/v (Hui, Isac, Boni & Sen, 1985), we believe that the effect of PEG in enhancing Ca²⁺ release is unlikely to involve membrane fusion. Thus, our recent studies have analysed by electron microscopy the appearance of isolated microsomal membrane vesicles derived from N1E-115 cells after GTP-treatment with or without PEG (Chueh et al. 1987). We observed that GTP was without any effect on vesicle appearance, whereas 3% PEG induced a very clear coalescence of vesicles into tightly associated conglomerates with very few free or unattached vesicles. The effect of PEG was not visibly altered by GTP. It may therefore be concluded that GTP itself does not induce any observable alteration in vesicle structure or association. However, the striking effectiveness of PEG is good evidence to suggest that the effect of GTP in inducing Ca²⁺ movements is promoted by a condition that clearly increases close associations between membranes. This may be an important clue to be action of GTP, as discussed in detail below. Thus, we consider that close association between membranes might be sufficient to permit the GTP-induced event which could involve formation of some type of junctional process between membranes, perhaps permitting the flow of Ca²⁺; thereafter, it is possible that under certain conditions membrane fusion may occur.

A further major problem to be addressed is the relationship between the actions of InsP₃ and GTP, and whether the processes activated by each agent involve any common mechanism. As described in a recent report, a number of clear distinctions exist between the actions of InsP₃ and GTP on Ca²⁺ release (Chueh & Gill, 1986). First, InsP₃-mediated release is unaffected by either GDP or GTPγS, both of which block the action of GTP on Ca²⁺ release, as described above. Second, PEG, which considerably promotes GTP-activated release (as described above), does not alter the action of InsP₃; indeed, the lack of effect of PEG on InsP₃-induced Ca²⁺ release suggests that InsP₃ functions via a mechanism that does not require close membrane interactions. A third distinction between the actions of InsP₃ and GTP is the temperature-dependency of their effects. Thus, the effect of InsP₃ is remarkably insensitive to temperature changes, the rate of InsP₃-induced Ca²⁺ release being reduced by only 20% when the temperature is decreased from 37°C to 4°C; this contrasts with the complete abolition of the effectiveness of GTP at the lower temperature (Chueh & Gill, 1986). The latter result is consistent with GTP activating release via a process involving an enzymic step, perhaps an enzymic hydrolysis of GTP, whereas the action of InsP₃ is unlikely to involve an enzymic step. (As discussed above, this temperature independence of the action of InsP₃ is highly suggestive of a process involving direct activation of a channel.) A fourth major distinction between the actions of InsP₃ and GTP concerns their Ca²⁺-dependency. Thus, InsP₃-induced Ca²⁺ release, in contrast to that induced by GTP, is modified by the free Ca²⁺ concentration. Ca²⁺ uptake and release were normally measured at a free Ca²⁺ concentration of 0·1 μmol I⁻¹. When the free Ca²⁺ concentration is increased to 1μmoll⁻¹, the effect of InsP₃ is reduced by 50%; at 10μmoll⁻¹ free Ca²⁺ the action of InsP₃ is completely abolished. In contrast, GTP induces identical fractional Ca²⁺ release over this entire range of free Ca²⁺ concentration.

The inhibition of InsP₃-mediated Ca²⁺ release with levels of Ca²⁺ above the physiological resting concentration (1μmoll⁻¹) is a significant observation indicating that the InsP₃ release process is under negative feedback control from the level of Ca²⁺, a potentially important regulatory response (Chueh & Gill, 1986). Interestingly, recent work from Worley et al. (1987) indicates that binding of labelled InsP₃ to its putative membrane receptor has almost identical Cab sensitivity, suggesting that the feedback effect may exist at the InsP₃ binding step. This also provides evidence that the InsP₃ binding site identified by Worley et al (1987) is the site of action of InsP₃. Much more compelling evidence to link to binding and action of InsP₃ has recently arisen from analysis of the effects of the glycosaminoglycan, heparin, which has been shown not only potently to inhibit InsP₃ binding (Worley et al. 1987) but also to bind to and provide a high degree of purification of a specific lnsP₃-binding protein, as recently described by Supattapone, Worley, Baraban & Snyder (1988). In very recent experiments we have observed that heparin is a powerful antagonist of the action of InsP₃ in inducing Ca²⁺ release from either permeabilized cells or isolated membrane vesicles (Ghosh et al. 1988). Thus, heparin blocks InsP₃-induced Ca²⁺ release with a K_t of 3nmoll^−J, suggesting a much higher affinity for the site than any known inositol phosphate. Moreover, heparin was shown to inhibit competitively the action of InsP₃, and also to reverse the InsP₃-activated Ca²⁺ release and permit immediate re-uptake of Ca²⁺. As shown in Fig. 2, the effect of heparin was highly specific towards the action of InsP₃. Thus, heparin altered neither Ca²⁺ pumping activity for the equilibrium uptake level (hence heparin did not alter any passive Ca²⁺ fluxes that contribute to the attainment of equilibrium). There was also no effect of heparin on the releasability of Ca²⁺ in response to the ionophore A23187, indicating that heparin did not change the state of accumulated Ca²⁺. Importantly, GTP-activated Ca²⁺ release was not affected by heparin. In other experiments, even heparin concentrations as high as 100μgml⁻¹ were without effect on the action of GTP. This is yet further convincing evidence for the distinction between the mechanisms of Ca²⁺ release activated by InsP₃ and GTP. The reversible and competitive effect of heparin on the action of InsP₃ indicates that when heparin displaces the InsP₃ molecule from its site of action the release process is immediately terminated, suggesting that activation of the putative InsP₃-responsive Ca²⁺ channel is intimately related to occupation of the InsP₃-binding site. This conclusion supports the prior available evidence mentioned above suggesting direct channel activation by InsP₃, in contrast to the action of GTP which involves a quite distinct process.

Several of the distinctions between the actions of GTP and InsP₃ (other than the effect of heparin) have also been reported by Henne & Söling (1986) using either liver- or parotid-derived microsomes, and by Jean & Klee (1986) using microsomes derived from NG108-15 neuroblastoma X glioma hybrid cells. It is concluded that the rapidity, relative temperature insensitivity and reversibility of InsP₃-induced Ca²⁺ release are all consistent with its probable direct activation of a Ca²⁺ channel, a conclusion in agreement with the observations of others (Muallem et al. 1985; Smith et al. 1985). In contrast, GTP appears to effect release by a temperature-sensitive process which probably involves the enzymic hydrolysis of the terminal phosphate from GTP.

Both the InsP₃- and GTP-induced Ca²⁺ release processes function on a similar intracellular Ca²⁺-sequestering compartment. Yet, the size of the releasable pools of Ca²⁺ are distinct. In the N1E-115 cell line, for example, the pool of Ca²⁺ released by GTP is approximately twice the size of the InsP₃-releasable pool, as shown in Fig. 3. Thus, using permeabilized N1E-115 cells, following maximal Ca²⁺ release by GTP, InsP₃ is ineffective in releasing further Ca²⁺ (Fig. 3B); however, following maximal release by InsP₃ (approximately 30% of accumulated Ca²⁺), GTP does effect a further release of Ca²⁺ (Fig. 3A), in fact, down to the level GTP could induce when added alone (that is, approximately 60% of accumulated Ca²⁺). These results suggest that three compartments exist; one sensitive to both GTP and InsP₃, another releasing Ca²⁺ only in response to GTP, and a third not releasing Ca²⁺ in response to either agent. Thus, it is apparent that although the GTP-releasable pool differs from the InsP₃-releasable pool in being larger, at least a significant proportion of accumulated Ca²⁺ lies within a pool which can be released by either of the two agents. In other words, it appears that all the Ca²⁺ within the InsP₃-sensitive Ca²⁺ pool is also releasable by the GTP-activated process, even if additional GTP-releasable Ca²⁺ also exists. This implies of probable proximal relationship between the InsP₃- and GTP-activated Ca²⁺ release processes, and permits us to consider the existence of possible coupling events linking their modes of action.

Several parameters of the GTP-activated process have together suggested to us a model which may explain the translocation of Ca²⁺ which is observed. Before considering this model, another important result must be considered. We recently observed that GTP can induce an entirely opposite effect on Ca²⁺ movements in the presence of oxalate; that is, GTP induces uptake as opposed to release of Ca²⁺ when oxalate is present. Although this observation at first appeared anomalous, it has provided an important piece of evidence in formulating our model for the action of GTP.

From the evidence described above, it was suggested that a simple GTP-mediated membrane fusion event was not entirely consistent with the observed release of Ca²⁺ induced by GTP. However, to investigate this problem iments were designed to determine whether Ca²⁺ precipitated with oxalate could be released from within permeabilized cells upon application of GTP. As shown in our previous studies (Gill & Chueh, 1985) and established in many different cell types (Henkart et al. 1978; McGraw et al. 1980; Wakasugi et al. 1982; Burton & Laveri, 1985), the ER is permeable to anions including oxalate and phosphate which can diffuse into the ER lumen and hence promote a large increment in Ca²⁺ uptake due to formation of insoluble Ca²⁺/oxalate or phosphate complexes. To investigate further how GTP activates Ca²⁺ release, we tested to see if oxalate-precipitated Ca²⁺ within ER could be released by GTP; a negative result would again militate against a simple membrane fusion event accounting for release and would instead argue in favour of a more selective channel mechanism, through which precipitated Ca²⁺ would not be expected to pass. However, as shown in Fig. 4, a marked increase in Ca²⁺ uptake was observed in the presence of oxalate, a remarkable and entirely opposite effect to that observed in the absence of oxalate. The effect is observed with concentrations of oxalate (2 mmol 1⁻¹) that have very little effect on uptake of Ca²⁺ in the absence of GTP (Fig. 4C). When oxalate is present at a concentration inducing linear uptake of Ca²⁺ (Fig. 4E), GTP still activates an additional increase in the rate of uptake. This phenomenon is not restricted to particular cell types, thus an identical effect of GTP on Ca²⁺ uptake in the presence of oxalate was observed using either N1E-115 neuroblastoma or DDT₁MF-2 smooth muscle cells.

Considering the paradoxically opposite effects of GTP in the presence and absence of oxalate, it was important to establish whether the two actions of GTP are mediated via the same or different mechanisms. It is now clear from a large number of observations that a single GTP-activated mechanism mediates both effects (Mullaney et al. 1987). For example, the GTP-dependence of Ca²⁺ uptake induced in the presence of oxalate is almost identical to that of the release induced without oxalate. Thus, the K_m for GTP for Ca²⁺ uptake (with oxalate) is 0·9μmoll⁻¹, which is very close to the value of 0·75μmol 1⁻¹ derived from Ca²⁺ release data, as described above. Further studies have revealed that the uptake of Ca²⁺ induced by GTP in the presence of oxalate is promoted by PEG in a manner very similar to GTP-activated Ca²⁺ release without oxalate (Mullaney et al. 1987). Thus, although both GTP-activated release and uptake are observable in the absence of PEG, both effects are considerably augmented in the presence of 3% PEG. In addition to these similarities between Ca²⁺ uptake and release induced by GTP, the nucleotide specificity profiles of the two processes closely coincide. A particularly important observation in this regard is that both effects of GTP show the same differential specificity towards the actions of nonhydrolysable GTP Inalogues, as shown in Fig. 5. Thus, GTP-activated Ca²⁺ uptake in the presence of oxalate is activated by neither GTPγS (Fig. 5A) nor GppNHp (Fig. 5B).

However, whereas 100μmoll⁻¹ GTPγS almost completely blocks the action of 5μmol 1⁻¹ GTP (Fig. 5A), GppNHp has almost no effect on the action of GTP (Fig. 5B). As mentioned above, it is assumed that the specificity between the two GTP analogues derives from the specificity of the nucleotide-binding site. The noneffectiveness of these analogues in promoting GTP-like effects is in clear distinction to the effects of guanine nucleotides on known G-protein activities (such as those modulating adenylate cyclase) where nonhydrolysable analogues are maximally or super effective. The similarity between the GTP-activated release and uptake processes is further exemplified by the effects of GDP on Ca²⁺ movements. Thus, GDP gives a full but delayed uptake response (Mullaney et al. 1987) which exactly coincides with its effect on release in the absence of oxalate (Gill et al. 1986; see Fig. 1). Moreover, as with release, GDP-mediated uptake is blocked by a high (1 mmol 1⁻¹) ADP concentration, indicating that its action arises from conversion to GTP via nucleoside diphosphokinase activity; in the presence of ADP, GDP blocks the action of GTP. GDPβS, which does not activate uptake, also blocks the action of GTP exactly as it does on GTP-induced Ca²⁺ release (see Fig. 1)

These data reveal almost complete correlation between parameters affecting GTP-activated uptake and release. A summary of these effects is given in Table 1. Such data provide very strong evidence suggesting that the same GTP-activated process mediates both uptake and release of Ca²⁺ in the presence and absence of oxalate, respectively. The only divergence between the two processes is the effectiveness of vanadate which blocks GTP-induced uptake but does not block GTP-activated release (Mullaney et al. 1987), indicating that GTP-activated uptake is dependent on the continuous action of the Ca²⁺ pump. However, as discussed below, the proposed model for the actions of GTP accounts for this difference.

From the above data, a number of clues can be derived which together have suggested to us a model invoking a GTP-mediated conveyance of Ca²⁺ across membranes and perhaps between organelles. Before discussing this conclusion, let us summarize this new information. First, we have observed that a discrete pool of GTP-releasable Ca²⁺ exists in cells, a pool that may incorporate within it a smaller InsP₃-releasable Ca²⁺ pool; yet, despite the overlap between pools, we have provided substantial evidence suggesting distinctions between the mechanisms of GTP and InsP₃ in activating Ca²⁺ release (Chueh & Gill, 1986). Second, since PEG promotes both the effects of GTP and a clearly observable membrane coalescence at the same concentration (1-3%) (Chueh et al. 1987), it is probable that activation of Ca²⁺ movements within cells is related to the occurrence of close appositions between membranes. Third, although direct fusion between membranes could account for some of the effects of GTP on Ca²⁺ movements, the observed reversibility of the effects of GTP together with the nonreleasability of oxalate-complexed Ca²⁺ by GTP would argue against a simple GTP-mediated fusion event between membrane surfaces as being the direct cause of Ca²⁺ movements. Fourth, there seems little doubt that the process of GTP-activated Ca²⁺ uptake in the presence of oxalate occurs via the same mechanism by which GTP activates release of Ca²⁺, in spite of the apparent opposite nature of these two GTP-mediated events.

This last piece of information appeared the most perplexing, yet ironically it may provide the most significant clue to the action of GTP. Thus, it is likely that oxalate promotes the uptake of Ca²⁺ into a discrete Ca²⁺-accumulating pool. As described above, it is well known that the ER membrane is permeable to anions including oxalate and phosphate. Hence passive entry of oxalate permits the formation of clearly observable insoluble complexes within the lumen of ER in cells; the entry of such anions may be mediated via a nonselective anion transporter activity analogous to that functioning in the SR membrane of muscle (see Martonosi, 1982). It is also apparent from our previous studies that, whereas Ca²⁺ accumulation in permeabilized cells and isolated microsomal membrane vesicles is oxalate-promoted, the accumulation of Ca²⁺ within purified inverted plasma membrane vesicles via the high-affinity plasma membrane Ca²⁺ pump is not enhanced by oxalate (Gill & Chueh, 1985). Since we have shown that these plasma membrane vesicles can indeed accumulate high intravesicular Ca²⁺ concentrations (Gill, Grollman & Kohn, 1981; Gill, 1982b; Gill, Chueh & Whitlow, 1984), more than sufficient to be precipitated in the presence of millimolar oxalate concentrations, we conclude that such membranes are largely impermeable to oxalate or phosphate. Thus, there is a good precedent for the existence of membranes through which passage of anions such as oxalate does not occur, and that organelle membranes may perhaps be distinguished according to their permeability to oxalate.

With the knowledge that distinct membranes exist which are differentially permeable to oxalate, we propose that in the presence of oxalate, GTP promotes uptake of Ca²⁺ as the result of a GTP-mediated movement of Ca²⁺ from a nonoxalate-permeable pool, which actively pumps Ca²⁺, to another Cab pumping pool which is freely permeable to oxalate. Thus, it is envisaged that GTP promotes a transmembrane conveyance of Ca²⁺ between such pools by activating some type of junctional process between the two membranes (see Fig. 6). Alternative schemes involving GTP-promoted oxalate-permeability or enhanced Ca²⁺ pumping are possible; but why then should an almost identical GTP-dependent process mediate movement (release) of Ca²⁺ in the absence of oxalate? In the model depicted in Fig. 6, the oxalate-permeable pool is very likely to be the ER or a subcompartment thereof; the nature of the putative nonoxalate-permeable pool is uncertain. Although the plasma membrane has been rendered permeable in our studies, it is possible that separate enclosed membranes derived from the plasma membrane might exist within the cell; such autonomous vesicles would be largely protected from the permeabilizing effects of saponin. The postulated process of junction formation between membranes would obviously be promoted by conditions that favour close appositions between membranes, as occurs in the presence of PEG. The action of GTP is envisaged as a necessary vector in either inducing the formation of junctions or activating the movement of Ca²⁺ through junctional processes arising by either random or PEG-promoted membrane interactions. Such transfer of Ca²⁺ would be activated by terminal phosphate hydrolysis from GTP; when GTP is washed away, the continued operation of such transfer would be terminated, as indicated by the reversibility experiments described above (Chueh et al. 1987).

An obvious problem is how GTP-mediated Ca²⁺ release could be accounted for by the same model. It seems entirely possible that the same type of junctional connections could be formed between intact organelles such as ER and nonclosed membranes, perhaps the plasma membrane. In this case, transmembrane conveyance of Ca²⁺ would result in release of Ca²⁺ to the medium (see Fig. 6). If such a conveyance of Ca²⁺ to the outside could be mediated by GTP how could GTP induce a build-up of Ca²⁺ within the oxalate-permeable pool? If the hypothetical GTP-activated junctional processes transmit only small solutes between pools (as gap junctions between cells are known to do), then a precipitate of the Ca²⁺-oxalate complex would not be expected to be transferred. Thus, in the experiments described above, oxalate and Ca²⁺ have been permitted to equilibrate within the oxalate-permeable pool; addition of GTP may result in a substantial ‘injection’ of Ca²⁺ from a nonoxalate-permeable (but, nevertheless, Ca²⁺-pumping) pool; this Ca²⁺ would be immediately precipitated owing to the excess oxalate present. When GTP and oxalate are added simultaneously at the beginning of uptake, GTP causes a prolonged inhibition of Ca²⁺ uptake due to activation of the release process. With time and in the presence of sufficient oxalate, there is a gradual increase in uptake followed eventually by a sustained uptake which proceeds at a rate approaching the maximal rate of uptake observed when GTP is added after oxalate (see below). The initial phase of this type of response is presumably due to the continued release of Ca²⁺ to the exterior, thus preventing sufficient build-up of Ca²⁺ to that critical level at which precipitation with oxalate occurs. Irrespective of when oxalate is added, the Ca²⁺-conveyance model predicts that Ca²⁺-pumping activity is essential to sustain GTP-activated Ca²⁺ uptake in the presence of oxalate, a prediction clearly confirmed by the blocking action of vanadate.

A significant problem that has dominated the physiological implications of the GTP-activated Ca²⁺ movements we have described, is how the high levels of GTP within cells (0·1–0·3mmoll⁻¹) can be reconciled with the extreme sensitivity of the GTP-activated process; thus, it was argued that under physiological conditions, the intracellular pool acted upon by GTP would be permanently depleted (Baker, 1986). However, by implicating a transfer of Ca²⁺ only between actively pumping organelles (and possibly with the outside of the cell), there would not be any collapse of existing gradients. Thus, the release that is observed with GTP may only reflect an artificially imposed, diminished external Ca²⁺ level that is a consequence of using permeabilized cells. In other words, such release could actually represent reversed movement of Ca²⁺ through a system that normally exists to convey Ca²⁺ perhaps to replenish the intracellular pool. The implication here is that such interpool communication may normally exist between organelles in intact cells but be reversed when cells are broken and GTP washed away.

Alternatively, the functioning of such Ca²⁺-communication between organelles may be regulated in situ by another cytosolic factor.

With the above model in mind, perhaps the most relevant problem to be addressed was the relationship between the pools of Ca²⁺ modified by GTP and that Ca²⁺ pool sensitive to InsP₃. This area of investigation has produced some important results. One initial step was to ascertain whether InsP₃ releases Ca²⁺ from an oxalate-permeable or oxalate-impermeable pool. This question is largely answered by the data shown in Fig. 7. In permeabilized cells from the DDT₁MF-2 smooth muscle and the N1E-115 neuroblastoma cell lines, InsP₃ in the absence of oxalate reduces Ca²⁺ uptake by 50 and 30%, respectively (Fig. 7A,C), effects entirely consistent with the extent of Ca²⁺ release observed following InsP₃ addition to Ca²⁺-loaded cells, as described above. In the presence of oxalate a sustained increase in the rate of ATP-dependent Ca²⁺ accumulation is observed (Fig. 7B,D) consistent with formation of the insoluble Ca²⁺/oxalate complex and hence a reduced rate of Ca²⁺ efflux (Gill & Chueh, 1985; Mullaney et al. 1987). Importantly, 10μmoll⁻¹ InsP₃ (a maximally effective concentration) completely eliminates the increment in Ca²⁺ uptake induced by oxalate in permeabilized DDTjMF-2 cells (Fig. 7B) indicating that InsP₃ activates Ca²⁺ release from an oxalate-permeable pool. Although not completely abolishing oxalate-enhanced Ca²⁺ uptake, the effectiveness of InsP₃ is very similar using permeabilized N1E-115 cells (Fig. 7D); hence in these cells, whereas InsP₃ does release Ca²⁺ from an oxalate-permeable pool, a small fraction of this pool may be unresponsive to InsP₃. As stated above, it is well established that the ER membrane is permeable to anions including oxalate, hence permitting clearly observable precipitation of Ca²⁺ within the ER lumen when oxalate is presented intracellularly. Thus, these data, although not providing definitive proof, are consistent with the view that the source of InsP₃-mobilizable Ca²⁺ is the ER or at least a subcompartment thereof.

Although, as described above, there are clear distinctions between the mechanisms by which InsP₃ and GTP activate Ca²⁺ movements, the data shown in Fig. 8 clearly establish a link between the actions of the two effectors. When added from the start of uptake, GTP and InsP₃ inhibit the accumulation of Ca²⁺ in a nonadditive manner (Fig. 8A) consistent with the extent of release described above. In the presence of oxalate, the action of GTP is very different from that of InsP₃ (Fig. 8B). Thus, whereas InsP₃ merely inhibits accumulation, GTP shows a biphasic effect. This effect of GTP is interpreted to support further the model in Fig. 6 since it shows that the two opposing GTP-activated movements of Ca²⁺ directly compete for access of Ca²⁺ to a common compartment. Thus, although initially Ca²⁺ release occurs resulting from interactions between closed and open compartments, thereafter, as the threshold of accumulated Ca²⁺ reaches that precipitable by oxalate, release of complexed Ca²⁺ is prevented and Ca²⁺ continues to accumulate at a higher rate reflecting the combined pumping activity of intact pools between which Ca²⁺ movement has been activated by GTP. Most importantly, the GTP-induced enhanced Ca²⁺ uptake phase is almost completely abolished when InsP₃ is added together with GTP (Fig. 8B) indicating that InsP₃ releases Ca²⁺ from the same pool into which GTP activates Ca²⁺ accumulation These results obtained using permeabilized N1E-115 neuroblastoma cells have been repeated using permeabilized DDT₁MF-2 smooth muscle cells. It should be noted that InsP₃ does not block the effects of GTP per se, since Ca²⁺ accumulation is reduced to a level well below that induced by InsP₃; thus, it may be inferred that although InsP₃ prevents the additional accumulation of Ca²⁺ activated by GTP, it in fact permits the Ca²⁺-releasing effects of GTP to dominate. These results provide direct evidence for the operation of both GTP-and InsP₃-activatable Ca²⁺ transport mechanisms on the same pool of Ca²⁺. Most significantly, they suggest that loading of Ca²⁺ within the InsP₃-sensitive pool may be controlled by the GTP-activated Ca²⁺ translocation process.

The proposed scheme of GTP-activated Ca²⁺ movements accounts for all the observed effects of GTP and oxalate on Ca²⁺ movements. Alternative schemes invoking direct effects of GTP on Ca²⁺ pumping or GTP-enhanced movements of oxalate are inherently unlikely since they do not account for rapid GTP-mediated Ca²⁺ release. In fact, GTP-mediated Ca²⁺ release occurs in the presence of vanadate and in the absence of ATP (Mullaney et al. 1987), that is, in the absence of any pumping activity. Also, in a recent report, Hamachi et al. (1987) described similar GTP-enhanced uptake of Ca²⁺ in the presence of oxalate. Although no explanation was offered for the effect, direct experiments revealed no effect of GTP on oxalate movements. Although recent work from Dawson and colleagues (Dawson et al. 1987; Comerford & Dawson, 1988) suggests membrane fusion may account for the effects of GTP, as stated above, the observations we have made on reversibility of the effects of GTP and on the electron microscopic structure of microsomal membrane vesicles treated with GTP, together argue against a simple membrane fusion process being activated by GTP. Although GTP hydrolysis is clearly implicated in the process of GTP-activated Ca²⁺ translocation (Chueh & Gill, 1986; Gill et al. 1986), it is presently unclear whether terminal phosphate is transferred to water (as in the case of a GTPase reaction), or whether a kinase-mediated mechanism transfers phosphate to another substrate molecule. Evidence for the former was recently presented by Nicchitta, Joseph & Williamson (1986), whereas a GTP-induced protein phosphorylation possibly associated with Ca²⁺ release was claimed by Dawson, Comerford & Fulton (1986).

Based on several important conclusions drawn from the data given in Figs 7 and 8, the scheme described above to account for the effects of GTP can be extended to encompass the action of InsP₃. First, from the data in Fig. 8A and data described earlier, it is apparent that the InsP₃-releasable Ca²⁺ pool is both smaller than and contained within the GTP-activatable pool. Second, based on the results shown in Fig. 7, the pool from which InsP₃ induces Ca²⁺ release is itself permeable to oxalate. Third, and most significant, this InsP₃-releasable Ca²⁺ pool is indeed the same pool that can be loaded with Ca²⁺via the GTP-induced Cam translocating process, as shown in Fig. 8B. These observations suggest to us that the InsP₃-releasable Ca²⁺ pool is the oxalate-permeable subcompartment of the GTP-activatable pool, as depicted in the model shown in Fig. 9. Thus, we assume that the efficient operation of the InsP₃-activated Ca²⁺ channel enhances efflux of Ca²⁺ from this pool effectively enough to prevent sufficient build-up of Ca²⁺ to reach the oxalate-precipitable threshold. Interestingly, when the experiment shown in Fig. 8B is conducted with 10 mmol I⁻¹ oxalate or higher (data not shown), some GTP-dependent build-up of Ca²⁺ does occur at later times in the presence of InsP₃, suggesting that by lowering the oxalate threshold, even the rapid release effected by InsP₃ is insufficient to prevent a significant build-up of Ca²⁺.

The direct reversal of the effect of GTP by InsP₃ provides a strong argument for considering that indeed both InsP₃ and GTP can act upon a common pool of Ca²⁺. Such conclusions are reminiscent of our earlier ‘flux reversal’ studies which provided direct proof for the coexistence of specific plasma membrane Ca²⁺ and Na⁺ flux mechanisms in a single population of synaptic membrane vesicles (Gill & al. 1981; Gill, 1982a,b). The most significant implication of the scheme shown in

Fig. 9 is that a close interrelationship probably exists between the actions of InsP₃ and GTP. We had previously speculated that this might be the case (Mullaney et al. 1987) but had no proof. The data presented in Figs 7 and 8 provide for the first time direct evidence that both InsP₃ and GTP can modify the same Ca²⁺ compartment in spite of their probable distinct mechanisms of action.

It is very possible that the GTP-regulated Ca²⁺-translocating process may control the size of the InsP₃-induced Ca²⁺ signal by permitting InsP₃ to release Ca²⁺ from a more extensive internal Ca²⁺ pool. Moreover, the same process may regulate the loading and/or replenishment of Ca²⁺ within the InsP₃-releasable pool. Such potential regulation derives much relevance from the considerable percent attention that has been directed towards the possible mechanisms by which the InsP₃-releasable Ca²⁺ pool may be replenished from the outside. Thus, Putney (1986) has suggested that external Ca²⁺ entry may be directed into this pool and hence account for the frequently observed prolonged responses to receptor-induced signals which are dependent on extracellular Ca²⁺. Recently, Irvine & Moor (1986, 1987) have presented experimental evidence suggesting the possible involvement of inositol 1,3,4,5-tetrakisphosphate (InsP₄) in inducing Ca²⁺ entry; in fact, their studies on activation of sea urchin eggs are consistent with the possibility that InsP₄ may promote entry of external Ca²⁺ into the InsP₃-releasable pool via a mechanism remarkably similar to the scheme described here for the movements of Ca²⁺ induced by GTP (Irvine & Moor, 1987). We are currently investigating whether this putative action of InsP₄ is related to GTP-activated Ca²⁺ movements and/or whether InsP₄ may modulate GTP-induced Ca²⁺ translocation.

We extend our sincerest gratitude to our former colleagues Dr Teruko Ueda and Dr Sheau-Huei Chueh for their very considerable contributions to the work described, and to Mark Noel and Cindy Ebert for their excellent technical assistance. This work was supported by grant number NS 19304 from the National Institutes of Health, and grant number DCB-8510225 from the National Science Foundation.