Inositol phosphate formation and its relationship to calcium signaling.

The activation of a variety of cell surface receptors results in a biphasic increase in the cytoplasmic Ca2+ concentration due to the release or mobilization of Ca2+ from intracellular stores and to the entry of Ca2+ from the extracellular space. It is well established that phosphatidylinositol 4,5-bisphosphate hydrolysis is responsible for the changes in Ca2+ homeostasis. Stimulation of Ca2(+)-mobilizing receptors also results in the phospholipase C-catalyzed hydrolysis of the minor plasma membrane phospholipid, phosphatidylinositol 4,5-bisphosphate, with the concomitant formation of inositol (1,4,5) trisphosphate [1,4,5)IP3) and diacylglycerol. Analogous to the adenylyl cyclase signaling system, receptor-mediated stimulation of phospholipase C also appears to be mediated by one or more intermediary guanine nucleotide-dependent regulatory proteins. There is strong evidence that (1,4,5)IP3 stimulates Ca2+ release from intracellular stores. The Ca2(+)-releasing actions of (1,4,5)IP3 are terminated by its metabolism through two distinct pathways. (1,4,5)IP3 is dephosphorylated by a 5-phosphatase to inositol (1,4) bisphosphate; alternatively, (1,4,5)IP3 can be phosphorylated to inositol (1,3,4,5) tetrakisphosphate by a 3-kinase. Whereas the mechanism of Ca2+ mobilization is understood, the precise mechanisms involved in Ca2+ entry are not known. A recent proposal that (1,4,5)IP3 secondarily elicits Ca2+ entry by emptying an intracellular Ca2+ pool will be considered. This review summarizes our current understanding of the mechanisms by which inositol phosphates regulate cytoplasmic Ca2+ concentrations.


Introduction
An examination of the relationship of phosphoinositide turnover to Ca2' signaling began in the early 1950s with the observation by Mabel and Lowell Hokin that the muscarinic cholinergic receptor agonist, acetylcholine, selectively increased the incorporation of 32Pi into two minor plasma membrane phospholipids, phosphatidylinositol (PI) and phosphatidic acid (1). However, it was not until some 20 years later that Michell, noting that the receptors that stimulated phosphoinositide turnover also activated Ca2'-dependent processes in the cell, proposed that receptor-stimulated phosphoinositide turnover results in a cellular Ca2' response (2). PI is sequentially phosphorylated by kinases in the cell to phosphatidylinositol-4-phosphate (PIP) and phosphatidylinositol 4,5-bisphosphate (PIP2). Indeed, we now know that stimulation by any of a number of Ca2'mobilizing receptor agonists initially results in the phospholipase C-catalyzed hydrolysis of PIP2 (3). Furthermore, the Hokins' initial observation of the incor-stimulated PIP2 hydrolysis, breakdown of the inositol phosphates to inositol, and resynthesis of PI. Phospholipase C-catalyzed PIP2 hydrolysis results in the formation of the water-soluble inositol (1,4,5) trisphosphate [(1,4,5)IP3] and the lipid-soluble diacylglycerol. Berridge proposed that (1,4,5)IP3 was the intracellular messenger that stimulated Ca2" release from intracellular stores (4). Soon thereafter, the predicted effect of (1,4,5)IP3 on Ca21 mobilization was demonstrated; micromolar concentrations of (1,4,5)IP3 rapidly released Ca2' from a nonmitochondrial store in permeabilized pancreatic acinar cells (5). This result quickly was confirmed in other tissues in a number of laboratories (6,7). Thus, the evidence is convincing that (1,4,5)IP3, generated upon activation of Ca21-mobilizing receptors, releases Ca2" from intracellular stores.
Meanwhile, a parallel story evolved in Nishizuka's laboratory (8)(9)(10), which demonstrated that the other product of PIP2 hydrolysis, diacylglycerol, also was a potent intracellular messenger. Diacylglycerol remains in the plasma membrane to activate a ubiquitous protein kinase, designated as C-kinase by Nishizuka. This review briefly summarizes our current understanding of the mechanisms by which inositol phosphates regulate cellular Ca21 metabolism. The mechanisms involved in receptor activation of phospholipase C, as well as the metabolic pathways by which the inositol phosphates are interconverted, are discussed. Finally, proposed mechanisms by which inositol phosphates elicit intracellular Ca2" release and Ca2" entry from the extracellular space are described.
Receptor Activation of Phospholipase C The mechanisms by which cell surface receptors stimulate phospholipase C have been the focus of considerable research efforts. Much of our current understanding of these mechanisms has evolved from the well-characterized adenylyl cyclase signaling system, which converts ATP to the intracellular messenger, 3',5'-cyclic AMP. In the adenylyl cyclase system, cell surface receptors communicate with the adenylyl cyclase enzyme, located on the cytoplasmic face of the plasma membrane, through intermediary guanine nucleotide-dependent regulatory proteins (G-proteins) (11). Two different G-proteins, G. and Gi, link stimulatory and inhibitory receptors, respectively, to adenylyl cyclase. G. and G;, comprised of a, ,3, and y subunits, are two members of a family of heterotrimeric proteins whose function is regulated by guanine nucleotides. The a subunits of this family of G-proteins are heterogeneous, whereas the ,3 subunits are quite similar, if not identical (12). Activation of G. by hormones (in the presence of guanine nucleotides) or by guanine nucleotides alone results in the displacement, by GTP, of the GDP bound to the a subunit and subsequent dissociation of the GTP bound-a subunit from the 0/y subunits. This activated GTP-bound-a subunit stimulates the adenylyl cyclase enzyme. GTP hydrolysis to GDP by a GTPase (an inherent activity of the G-proteins) terminates cyclic AMP formation and is assumed to result in reassociation of the G-protein subunits.
Gi also undergoes subunit dissociation after incubation with guanine nucleotides. However, the precise mechanism by which Gi inhibits adenylyl cyclase is controversial. GTP-bound ai may directly inhibit adenylyl cyclase (13). Alternatively, inhibition of cyclic AMP formation may occur through the liberation of a stoichiometric excess of ,/y subunits that associate with the free a subunits of G., resulting in enzyme inhibition (14). Regardless of the precise mechanism of inhibition of adenylyl cyclase, it appears that Gprotein activation requires dissociation of the protein into separate a and P/y subunits.
The ability of the adenylyl cyclase-linked G-proteins to be purified, sequenced, and reconstituted into phospholipid vesicles was due to the capacity of these proteins to serve as substrates for covalent modification by bacterial toxins. That is, G. and Gi are ADPribosylated by cholera and pertussis toxin, respectively. If this toxin-catalyzed ADP-ribosylation is performed in the presence of 32P-NAD, 32P-ADPribose is covalently transferred to G. and Gi. The cholera toxin-catalyzed ADP-ribosylation of Gs inhibits its inherent GTPase activity to irreversibly activate Gs. On the other hand, pertussis toxin-catalyzed ADP-ribosylation of Gi inactivates the protein and blocks the receptor-mediated inhibition of adenylyl cyclase.
Receptor-mediated activation of phospholipase C also appears to involve an intermediary G-protein.
The earliest indication that receptors linked to phospholipase C might occur through a G-protein(s) analogous to Gs and Gi was the observation, in a number of tissues, that guanine nucleotides decreased the apparent affinity of agonists for receptors known to stimulate phospholipase C (15)(16)(17)(18)(19)(20)(21).
The subsequent observation that guanine nucleotide analogs potentiated the stimulatory actions of thrombin on diacylglycerol formation, protein phosphorylation, and serotonin secretion in permeabilized platelets provided more direct evidence of the involvement of a G-protein in the receptor activation of phospholipase C (22,23). Most recently, several laboratories have demonstrated a guanine nucleotidemediated activation of phospholipase C in membrane or permeable cell preparations from a number of cell types (24)(25)(26)(27)(28)(29). The predicted synergistic stimulation of PIP2 hydrolysis by agonists and guanine nucleotides also has been observed (29,30). Furthermore, activation of phospholipase C by guanine nucleotides shows the same relative sensitivity to guanine nucleotide analogs as was observed with the adenylyl cyclase signaling system (30).
Taken together, the guanine nucleotide dependence of receptor-stimulated phosphoinositide turnover in permeabilized cells or in membrane preparations, the hormonal stimulation of GTPase activity (19), and the guanine nucleotide regulation of agonist binding to the Ca21-mobilizing receptors suggest a striking similarity between the G-protein that links cell surface receptors to phospholipase C and the G-proteins that couple receptors to the activation and inhibition of adenylyl cyclase. However, unlike the adenylyl cyclase system, the precise identity of the G-protein(s) mediating phospholipase C activation is currently unknown.
Two general mechanisms have been described for receptor activation of phosphoinositide turnover. First, in some systems (including neutrophils and mast cells), receptor activation of phosphoinositide turnover is sensitive to inactivation by pertussis toxin (31)(32)(33). This result suggests that Gi or (more likely) a Gi-like protein regulates phospholipase C in these cells. However, stimulation of inositol phosphate formation by the majority of phospholipase C-linked receptors is insensitive to pertussis toxin (30,34,35). This result is consistent with the hypothesis that a guanine nucleotide-dependent regulatory protein that is similar, but not identical to the proteins that regulate adenylyl cyclase, links cell surface receptors to phospholipase C. This differential sensitivity to pertussis toxin may indicate that different G-proteins regulate phosphoinositide metabolism in different tissues.
Recently, several toxin-insensitive G-proteins have been identified and/or purified based on their capacity to bind radiolabeled guanine nucleotides with high affinity and on the inability of these proteins to serve as substrates for ADP-ribosylation by pertussis toxin (36)(37)(38). However, to date, the purification and successful reconstitution of a pertussis toxin-insensitive G-protein linked to phospholipase C has not been achieved.
In cholinergically stimulated pancreatic minilobules, Dixon and Hokin determined that the levels of (1,4,5)IP3 rapidly increased, then fell to a new elevated steady-state, whereas (cl:2,4,5)IP3 slowly accumulated (51). Furthermore, they suggested that (1,4,5)IP3 could be responsible for intracellular Ca21 release at short times of stimulation, whereas both (1,4,5)IP3 and (cl:2,4,5)IP3 might contribute equally to Ca2' release during prolonged stimulation. However, in most published reports of inositol phosphate formation in vivo, experiments are terminated by the addition of acid, which cleaves the 1:2 cyclic bond to yield a mixture of (1,4,5)IP3 and (2,4,5)IP3. To date, an accurate assessment of the amounts of (cl:2,4,5)IP3 formed in cells, as well as its contribution to the Ca21 response in a variety of tissues, has not been determined.

Inositol Phosphates and Ca2+ Release
It is well established that receptor-stimulated Ca2m obilization involves an initial release of Ca2' from intracellular stores, followed by Ca2" entry from the extracellular space (57,58). This biphasic Ca2r esponse can be measured either directly by the fluorescent Ca2`indicators, Quin-2 and Fura-2 (59), or indirectly by monitoring changes in the Ca2+-dependent processes [such as unidirectional 86Rb+ efflux through Ca2+-activated K+ channels (60)].
Specifically, the addition of a Ca2'-mobilizing agonist results in a rapid and transient increase in cytoplasmic Ca2 , which persists in the absence of extracellular Ca2+ and, therefore, appears to result from the release of an intracellular pool of Ca2+. A sustained phase then follows, which is dependent on the presence of extracellular Ca2+ and presumably reflects Ca2+ entry from the extracellular space. There is strong evidence to support the proposal that (1,4,5)IP3 mediates intracellular Ca2+ release (61).
The effects of (1,4,5)IP3 were examined in permeabilized cells and in subcellular fractions under experimental conditions that would selectively poison mitochondrial versus nonmitochondrial pools. These manipulations led a number of laboratories to conclude that (1,4,5)IP3 releases Ca2" from an intracellular pool that was insensitive to inhibitors of mitochondrial Ca2`uptake and, by default, was likely to be a component of the endoplasmic reticulum (5)(6)(7)62,63). However, it appears that only a portion of the Ca21 stored in the endoplasmic reticulum is released by (1,4,5)IP3. Approximately 30 to 40% of the exchangeable Ca2+ in the endoplasmic reticulum of permeabilized hepatocytes is releasable by (1,4,5)IP3 (6,64), which suggests that the remaining Ca2' is present in a pool not regulated by (1,4,5)IP3.
Recently, a capacitative mechanism was proposed in which (1,4,5)IP3 secondarily promotes Ca2+ entry (71), by emptying intracellular Ca21 stores. We know that the ability of a cell to respond to a second Ca2+_ mobilizing agonist, following the termination of the first stimulus, depends on the refilling of the IP3sensitive intracellular pool (60,72). Under resting conditions, this IP3-sensitive intracellular Ca2' store is resistant to depletion by extracellular chelating agents. However, upon depletion, refilling of this pool only occurs in the presence of extracellular Ca2+. This refilling process occurs after the termination of the first stimulus (and presumably in the absence of intracellular messengers, such as inositol phosphates). This intracellular pool appears to be in close apposition to the plasma membrane because the refilling process occurs with only a small increase in the cytoplasmic Ca2+ concentration (59). According to this capacitative model, emptying of the intracellular Ca2+ pool by IP3 permits the direct communication of this pool with the plasma membrane. In the presence of extracellular Ca2+, Ca2+ enters the cell through this interface and subsequently into they cytosol. When IP3 is degraded, extracellular Ca2+ continues to enter the cell via this interface until the intracellular Ca2+ pool is restored.

Conclusions
Virtually our entire understanding of the mechanisms by which the inositol phosphates regulate the Ca2"-signaling system has evolved within the last 5 years. Whereas (1,4,5)IP3 has been accepted as the signal that elicits intracellular Ca2' release, the second phase of the cellular Ca2' response, namely Ca2' entry, is less understood. Furthermore, the complexity of the metabolism of the inositol phosphates (with its alternative phosphorylation/dephosphorylation pathways) implies that additional inositol phosphates may be biologically active. It seems safe to predict that as our knowledge of the phosphoinositide/Ca2+signaling system increases, our unanswered questions will increase as well.