Calcium ion as intracellular messenger and cellular toxin.

Ca2+ serves a nearly universal intracellular messenger function in cell activation, but excess Ca2+ is also a cellular toxin. The possibility of Ca2+ intoxication is minimized by an elaborate autoregulatory system in which changes in Ca2+ influx rate across the plasma membrane are rapidly compensated for by parallel changes in Ca2+ efflux rate. By this mean, cellular Ca2+ homestasis is maintained so that minimal changes in total cell calcium and cytosolic Ca2+ concentration occur during sustained Ca2(+)-mediated responses. Rather than a sustained increase in cytosolic Ca2+ concentration, it is the localized cycling of Ca2+ across the plasma membrane that is the critically important Ca2+ messenger during the sustained phase of cellular responses mediated via surface receptors linked to the hydrolysis of phosphatidylinositol 4,5-bisphosphate (PIP2). PIP2 hydrolysis gives rise to inositol(1,4,5)trisphosphate (IP3) and diacylglycerol (DAG). The IP3 acts to release Ca2+ from an intracellular pool, thereby causing a transient rise in cytosolic Ca2+ concentration. This transient Ca2+ signal activates calmodulin-dependent protein kinases transiently, and hence, causes the transient phosphorylation of a subset of cellular proteins that mediate the initial phase of the response. The DAG brings about the association of protein kinase C (PKC) with the plasma membrane where a receptor-mediated increase in Ca2+ cycling across the membrane regulates PKC activity. The sustained phosphorylation of a second subset of proteins by PKC mediates the sustained phase of the response. Hence, Ca2+ serves as a messenger during both phases of the cellular response, but its cellular sites of action, its mechanisms of generation, and its molecular targets differ during the initial and sustained phases of the response.(ABSTRACT TRUNCATED AT 250 WORDS)


Introduction
When Prometheus stole fire from the gods and presented it to man, it was a gift of immense value (1). It provided heat against the cold, light against the dark, and a means of communication. Yet, with this gift came also the destructive power of uncontrolled fire with its ability to imperil man's very existence.
This myth of the gift of fire to man can serve as the life motif for the evolutionary gift of Ca2' to the cell.
This ion has become the means of regulating heat production and of facilitating intracellular communication. Yet, with this gift has also come a destructive potential: excess calcium ions are cellular toxins (2). Because of the potential for toxicity, an elaborate system has been developed to ensure the maintenance of cellular Ca2' homeostasis over a wide range of circumstances (3). This creates a seeming paradox: changes in Ca2" ion concentration are thought to serve a nearly universal messenger function in cells that possess very effective mechanisms by which to minimize any change in intracellular Ca2+ ion concentration. A resolution of this seeming paradox has come from the recognition that the messenger function of Ca2' is expressed in quite subtle ways within both a spatial and temporal context (4). The purpose of the present article is to review our present views of the mechanisms by which cellular Ca2' homeostasis is achieved and then to consider the messenger function of Ca2' within this context.

Evolutionary Relationship between Ca2+ and Cell Function
Before considering the mechanisms by which cellular Ca2' homeostasis is maintained, it is worth considering the evolutionary aspects of cellular Ca2' metabolism. In so doing, it is instructive to begin by considering how mammalian cells have come to maintain such an enormous (10,000-fold) Ca2' concentration gradient across their plasma membranes. Recent R8RASMUSSEN ET AL. studies of the nature of sedimentary rocks formed during different geologic eras have led to new insights into the nature and composition of the primordial sea (5). It appears likely that when life first evolved, the pH of the sea was high (circa pH 8) and the concentration of Ca2', low (10-' M-10 M). In this low Ca2' environment, the basic life processes involving nucleic acids and, particularly, nucleotides (for example, ATP) could develop. These nucleotides are phosphorylated organic anions, and as such, their Ca2" salts are sparingly soluble. On the other hand, a crucial aspect of geologic evolution was the gradual increase in the Ca2+ content of the sea (6,7). This increase posed a threat to the most fundamental of life processes: those involving the participation of phosphorylated compounds, e.g., DNA and RNA synthesis and the basic energy-transducing systems, because these compounds would precipitate or form stable Ca2' complexes in a higher Ca2' environment (8). This threat was met by the development of a surface membrane having a low Ca2' permeability and possessing energy-dependent mechanisms (pumps) by which Ca2' could be expelled from the cell interior. By these means, it was possible to maintain a low intracellular Ca2' (circa 0.1 gM). By the time multicellular life evolved, the [Ca2'] in the sea was close to 1 mM (9). This later precambrian sea was incorporated into the multicellular organism as the extracellular fluid. The heritage of this evolutionary sequence was a cell with a plasma membrane displaying a limited permeability to Ca2' and possessing very efficient mechanisms for pumping out any Ca2+ that did leak into the cell.  (11,12): a Ca2+/2H+-ATPase or Ca2+ pump driven by the hydrolysis of ATP and a 3Na+/Ca2' exchanger driven by the Na+ electrochemical gradient across the plasma membrane. As very little is known about the regulation of the exchanger in situ, the ensuing discussion will focus on the crucial role of the Ca2' pump in maintaining cellular Ca2+ homeostasis even though it is clear that the 3Na+/Ca2' exchanger plays an equally important role in maintaining cellular Ca2+ homeostasis in cells such as the cardiac myocyte.
Unlike the primitive cell in which the plasma membrane is the major regulator of Ca2+ metabolism, the highly developed mammalian cell has a considerably more complex system for regulating cellular Ca21 metabolism and for employing Ca2+ as an intracellular messenger. In addition to the plasma membrane, there are two other membrane-limited intracellular organelles which play important roles in Ca2' homeostasis: the mitochondria, and the sarcoplasmic reticulum or calciosome. Both of these intracellular membranes possess pump-leak pathways that are directionally oriented, so there is a net leak of Ca2+ from the matrix space of the particular organelle to the cytosol and an energy-dependent uptake that actively removes Ca2' from the cytosol and translocates it into the matrix space of that organelle.
In cellular Ca2' metabolism, these two intracellular organelles play quite different roles. These roles appear to be based on a) differences in the affinities of the respective active transport systems for Ca2', b) the difference in the mechanisms by which they store Ca2', and c) the mechanism by which Ca2' is released from these storage sites.
The Ca2' pump in the sarcoplasmic reticulum (SR) has a Km for Ca2' transport that is similar to that of the plasma membrane Ca2' pump (in the range of 0.3-0.9 FM). Because Ca2' is stored in the SR by binding to a specific calcium-binding protein, calsequestrin, the size of the storage Ca2' pool is determined by the extent of the specialized reticulum containing this protein and the quantity of this protein within this compartment (13). In contrast, the Km for active Ca2' uptake by the mitochondria is in the range of 0.9 to 1.6 gM, and the Ca2+ stored in the mitochondrial matrix space is in the form of a nonionic complex with phosphate and ATP. There is also a difference in the mechanisms by which events at the plasma membrane can lead to the rapid release of Ca21 from the calciosome (14,15) or sarcoplasmic reticulum, but not from the mitochondria.
From a functional point of view, the pool in the calciosome (or SR) serves as a source of trigger Ca2+ for stimulus-response coupling. This pool is most highly developed in skeletal muscle where it serves both as the source and sink for the Ca2+ involved in the regulation of contraction. The Ca2' in this pool is released during the initiation of the response and reaccumulated during the termination of the response. Because of the geometry of these large multinucleated cells and the organization of the SR that is in close proximity to the contractile elements, this pool is little influenced by the flux of Ca2+ across the plasma membrane. As a consequence, when it is released from the sarcoplasmic reticulum, it has very little influence on the Ca2' concentration just beneath the plasma membrane. Hence, during a normal contraction-relaxation cycle, very little of the Ca2+ released from the SR is pumped out of the cell (10,16).
In nonmuscle cells or in the small uninuclear cells found in mammalian smooth muscle, this Ca2' trigger pool serves a similar function but is relatively small; also it appears to be located close to the plasma membrane. Once Ca2+ is released from this pool in response to an appropriate stimulus, its location near the plasma membrane allows it to induce an increase in [Ca2`],m, thereby activating the Ca2' pump in this membrane. Thus, significant amounts of the released Ca2' are pumped out of the cell when the cell is activated (10).
Two possible functions have been attributed to the intramitochondrial Ca2' pool. On the one hand, this ion pool has been proposed to have a regulatory function in the control of mitochondrial metabolism.  (17). On the other hand, it has been proposed that the nonionic pool of calcium within the matrix space serves as a sink for excess Ca2' when the plasma membrane is unable to maintain Ca2' homeostasis. This latter role provides a means whereby the cell can sequester excessive Ca2" in times of a temporary imbalance between Ca2' influx and efflux across the plasma membrane. However, the Ca2' sequestering capacity of the mitochondria is finite.

Ca2"-Activated Response Elements
Before discussing the mechanisms by which Ca2" serves its intracellular messenger functions, an additional feature in the evolution of this system needs to be considered. Once Ca2' began to be used as an intracellular messenger, there was a need for specific intracellular receptors (proteins) that would receive this message and then couple it to specific cellular responses. There are two broad classes of Ca2' receptor proteins that arose early in evolution and are nearly universally expressed in animal cells. The first class are protein molecules that have no intrinsic enzymatic activity, but interact with other protein molecules which do. The second class are molecules which possess both Ca2'-binding sites and intrinsic enzymatic activity. The best characterized of the first class are calmodulin and troponin C (18); they are structurally related and possess four Ca2"-binding sites. In addition, they possess other sites that interact specifically with other proteins (response elements), thereby modifying their functions. Troponin C has a restricted cellular distribution being confined to skeletal and cardiac muscle where it interacts with and serves to couple the Ca2" message to the contractile proteins by facilitating a sequence of protein-protein interaction. Calmodulin (CaM), on the other hand, is nearly universally distributed throughout plant and animal cells. It participates in a wide variety of Ca2'-regulated intracellular processes. It brings about its effects in one of two ways. In both, the Ca2'-CaM complex interacts with another protein to modify its activity. In one case this is a protein which possesses an enzymatic activity other than protein kinase activity. In the other case the response element is one of a class of Ca2'-CaM-activated protein kinases. An example of the former class is the Ca2" pump or Ca2+/2H+-ATPase in the plasma membrane (19). The Ca2'-CaM complex binds directly to this enzyme to alter its activity by direct protein-protein interaction such that its Km for Ca2" is reduced and its Vmax for Ca2' transport increased. The latter class contains an increasing number of protein kinases with either very restricted or more universal specificity as regards their substrate proteins. Their importance lies in the increasing evidence that the regulation of the state of phosphorylation of intracellular proteins is a major mechanism by which changes in cell functions are brought about. In these cases, the Ca2'-CaM complex interacts with the protein kinase and serves as the regulatory subunit of this aggregate to bring about the activation of the particular kinase.
The protein kinase C (C-kinase) (20) is most important member of the class of enzymes that possess intrinsic enzymatic activity and are activated directly by the Ca2' ion without the partiepation of CaM or a calmodulinlike protein. This enzyme is activated directly by Ca2", but its sensitivity to Ca2+ activation depends on its association with lipids. In the nonactivated cell, C-kinase is confined largely to the cytosol.
In this state it is a relatively inactive Ca2"-insensitive enzyme. Upon activation of the cell, the C-kinase becomes associated with the plasma membrane as a consequence of both the increase in the DG content of this membrane and the Ca2" transient (see below). In this membrane-associated form, the C-kinase is a very active, Ca2+-sensitive protein kinase (21).

Classic Views of Ca2+ As Intracellular Messenger
Having discussed cellular Ca2`metabolism and Ca2"-activated response elements, it is now possible to consider the classic views of Ca2+ messenger function. These views were largely elaborated by an analysis of the events involved in neurosecretion at synapses, contraction of skeletal muscle, and contraction of heart muscle.
The first site where the messenger function of Ca2ẁ as defined was the presynaptic plasma membrane whose depolarization, as a result of an action potential, induces a brief opening of a particular class of voltage-dependent Ca2" channel (N channels). The resultant brief influx of Ca2" is thought to act in the cellular domain just beneath the plasma membrane to trigger the association of synaptic vesicles with the plasma membrane, resulting in the release of neurotransmitter (22). Only a very transient increase in Ca2+ influx is required. This brief increase in Ca2' influx is followed by a quiescent period during which the Ca2+ that entered the cell during its transient depolarization is pumped back out. Hence, Ca2+ homeostasis is maintained.
The second system where the messenger function of Ca2' was defined was that of the skeletal muscle cell.
The depolarization of the plasma membrane in this cell is linked to the release of Ca2' from the SR pool at a specialized junction between the two membrane systems (the so-called T system). The precise manner in which the information is conveyed from one membrane to another remains a matter of discussion, but the consequence is clear: a sudden and transient release of Ca2' from the SR pool leads to a rise in [Ca2+], in the cell cytosol sufficient to activate contraction. This rise in [Ca2'], is transient because the Ca2+ is rapidly reaccumulated by the SR. As a result, the contractile response is also transient (23).
The third system, the mammalian heart, combines features of the other two. During each electrical systole, depolarization of the plasma membrane leads to an increase in Ca2+ influx via another type of voltagedependent Ca2+ channel (L channel). The bolus of Ca2+ which enters the cell is not sufficient by itself to trigger a contractile response. Rather, it acts as a signal to trigger the release of Ca2+ from the SR that causes a sufficient increase in [Ca2+], to initiate a contractile response. Because the Ca2" released is rapidly reaccumulated by the SR, the contractile response is a transient one because the rise in [Ca2+], is transient.
In addition, the Ca2+ that entered the cell during systole is pumped out of the cell during diastole so that cellular Ca2' homeostasis is maintained (24).
In spite of the different mechanisms by which the Ca21 signal is generated, all three systems possess a common attribute: the magnitude and duration of the cellular response is a direct function of the magnitude and duration of the Ca2+ message.
Calcium Toxicity due to Disorders of the Classic Pathways These homeostatic systems can break down and lead to serious and even lethal consequences. The pathophysiologic realty of cellular Ca2+ toxicity can be illustrated in skeletal and cardiac muscle, respectively, by considering the condition of malignant hyperthermia and Isoproterenol-induced cardiac myotoxicity.

Malignant Hyperthermia
Malignant hyperthermia is a disease that occurs in susceptible humans when these individuals are exposed to certain anesthetic agents such as halothane. This susceptibility is inherited in humans as also in an inbred strain of pigs (25). In these animals, halothane exposure first leads to an increase in metabolic rate and heat production in skeletal muscle that is followed, eventually, by irreversible muscle contracture and death. The intracellular site of halothane action is the SR. Under resting conditions the rate of Ca2' efflux from this organelle in these pigs' muscles is greater than that from the SR of normal pigs. This increase in Ca2' efflux rate leads to a slight increase (0.2-0.4 gM) in the [Ca2+]c of the muscle cell, as measured by a Ca2+-sensitive microelectrode. This rise is not sufficient to cause a contraction of the muscle, but it is sufficient to stimulate the ATP-dependent reuptake of Ca2+ by the SR. Hence, there is an increased rate of Ca2`cycling across the SR which, because of the increased rate of ATP hydrolysis, is associated with an increase in metabolic rate. When the animal is exposed to low doses of halothane, the Ca2+ efflux pathway is stimulated further. As a consequence, the [Ca2+] rises more and further stimulates the reuptake of Ca2 , i.e., there is a further increase in Ca2`cycling and an attendant increase in heat production. Initially, however, this rise in [Ca2+], is still insufficient to induce a contraction. Hence, in this early phase of malignant hyperthermia, the major observable effect is an increased rate of heat production and hyperthermia.
However, as the rate of Ca2+ efflux increases further, it eventually reaches a point where it exceeds the capacity of the reaccumulation pathway. Under these circumstances, the [Ca2"], rises further, and an irreversible contracture and death ensues.

Isoproterenol-induced Cardiac Muscle Toxicity
When rats are treated with large doses of Isoproterenol for a period of hours, they develop irreversible cardiac muscle dysfunction and die. Death is due to Ca2' intoxication of the myocytes that develops as a consequence of an imbalance between the influx and efflux of Ca2' across the plasma membrane of these cells (26). Isoproterenol is a beta agonist that interacts with catecholamine receptors on the heart muscle cells causing, thereby, an activation of adenylate cyclase and a rise in the cAMP content of these cells. As a consequence, both the heart rate and the force of contraction of each heart beat increase. The major result of cAMP action is the activation of the cAMP-dependent protein kinase. This enzyme catalyzes the phosphorylation of Ca2' channels in the plasma membrane (L channels). Consequently, during each electrical systole, there is an increased rate of Ca2' influx into the cell, and hence a larger Ca2" signal to trigger Ca2" release from the SR. Consequently, the intracellular Ca2" transient is higher and the force of contraction greater. The rise in cAMP also leads to a greater rate of Ca2" efflux across the plasma membrane via both the Ca2" pump and the 3Na+/Ca2+ exchanger. Hence, during diastole there is a greater efflux of Ca2+ from the cell. However, as the heart rate increases, the period of diastole is shortened considerably more than the period of systole. This means that at high rates of beating not all of the Ca21 that has entered during a particular systole can be pumped back out of the cell during the subsequent diastole. A net accumulation of Ca2+ occurs during each cardiac cycle. Nonetheless, because this extra Ca2+ is sequestered within intracellular organelles, the heart continues to beat effectively, using transient rises and falls of [Ca2+]c to regulate its function.
At first, this extra Ca21 is stored in the SR and provides for a larger pool to be released at each systole. However, the storage capacity of this SR pool is quite limited and is soon saturated. The mitochondrial matrix compartment then becomes the major storage site for the excessive Ca2+. Because the capacity of this storage site is relatively large, the heart can be steadily gaining total cell calcium and yet continue to display a normal Ca2+-dependent contraction and relaxation cycle. Eventually, however, the capacity of the mitochondrial storage site is exceeded, and the [Ca2+]c then rises sufficiently to interfere with relaxation.
Also, the increase in [Ca2+], activates proteases and phospholipases so that cell destruction and death ensue. However, if Verapamil, a Ca2' channel blocker that prevents Ca2+ influx, is given simultaneously with Isoproterenol, the net accumulation of Ca2' by the myocytes is blocked. As a consequence, the sequence of events outlined above is prevented. These results indicate that maintenance of cellular Ca2" homeostasis is essential for the maintenance of normal cell function when cells employ extracellular Ca2" as an intracellular messenger.

Messenger System
In contrast to these three classic systems where there is a direct relationship between the amplitude and duration of the Ca2' message and the strength and duration of the response, there is a large number of cellular responses for which this direct relationship appears not to hold (10). These responses have in common the following characteristics: a) they are usually linked to the turnover of polyphosphoinositides; b) they are often continued responses to the sustained presence of an agonist; c) they are associated with a transient increase in [Ca2'],, lasting only 1 to 2 min even in the continued presence of the agonist; d) they display a sustained increase in the rate of Ca2ì nflux across the plasma membrane; e) they are not associated with an increase in total cellular Ca2" content, despite the sustained increase in Ca2" influx; and J) they are mediated by the temporal and spatial activation of more than one type of Ca2"-dependent protein kinase. Phosphoinositide Metabolism and Ca2F or our purposes, the important facts are that agonist-receptor interaction leads, via a G protein (27), to the activation of a specific phospholipase C that catalyzes the hydrolysis of phosphatidylinositol 4,5bisphosphate to generate two intracellular messengers, 1,2-diacylglycerol and inositol 1,4,5trisphosphate (28). The former is lipid soluble and remains within the plasma membrane to bring about the protein kinase C activation. Inositol 1,4,5trisphosphate is water soluble and functions in the cytosolic domain to initiate the release of Ca2" from the specialized component of the endoplasmic reticulum, the calciosome. Whether there are other receptor-linked pathways by which 1,2-diacylglycerol is generated, or whether additional inositol phosphates such as inositol 1,3,4,5-tetrakisphosphate and inositol 1,3,4-trisphosphate have messenger functions remain matters of debate.

Agonist-induced Responses
Methods available to measure changes in total cell Ca2', Ca2" influx rate and intracellular free Ca2" concentration have made is possible to understand the subtleties of the Ca2' messenger function. Such studies have led to an appreciation of the fact that there are temporally and spatially distinct phases of these Ca2'-mediated responses.
Using the photoprotein aequorin as an intracellular indicator, it is possible to measure transient changes in intracellular free Ca2' in response to appropriate agonists (29). Thus, for example, when angiotensin II acts on adrenal glomerulosa cells, it stimulates the turnover of the inositol phospholipids; the generation of inositol 1,4,5-triphosphate and 1,2-diacylglycerol; and a transient rise in [Ca2'], from approximately 0.15 to 0.8 gM, which lasts only 1 to 2 min (30). Following this, the [Ca2'], returns to a value only slightly above the original basal value. Nevertheless, the increase in aldosterone secretion (the specific cellular response) occurs gradually over a period of 10 to 15 min, reaching a sustained plateau value that is maintained for an hour or more, as long as angiotensin II is present. The removal of angiotensin II leads to a prompt reversal of the enhanced rate of aldosterone secretion with little change in [Ca2'],.
Even though there is a dissociation between the magnitude and duration of the Ca2' transient and the magnitude and duration of the cellular response, there is, nevertheless, a continued need for Ca2" during the sustained phase of the cellular response. The removal of extracellular Ca2' or an inhibition of Ca2' influx across the plasma membrane by the addition of drugs that block Ca2' channels leads to a prompt inhibition of the cellular response (31). Furthermore, in order for the sustained response to occur, the agonist must induce a sustained increase in Ca2' influx rate.
During cell activation, total cell calcium falls (at the time when the Ca2+ transient appears), reaches a plateau, and during the sustained phase of the response is maintained at a level below that found in the resting cell (31). Hence, during the sustained phase of the response, total cell Ca2+ is changing very little in spite of the fact that the Ca2' influx rate is increased. Hence, the Ca2+ efflux rate across the plasma membrane must increase sufficiently to balance the increase in Ca2' influx rate. Thus, during the sustained phase of the response, there is an increase in Ca2' cycling across the plasma membrane even though [Ca2'], has changed very little from its basal value. These findings raise three important questions: Release of Ca2' from the calciosomes leads to a rise in [Ca2'] in the submembrane domain of the plasma membrane, as well as in the bulk cytosol. This increase in [Ca2']m leads to a stimulation of the Ca2" pump (33). As a consequence, much of the Ca2" released from the trigger pool is pumped out of the cell, thereby accounting for the net loss of total cell calcium. This mechanism is so efficient that a bolus equivalent to 20 jimole or more of Ca2" (per liter of cell H20) must be released in order to raise the cytosolic Ca2' concentration 1 ,uM (4).

Cellular and Molecular Mechanisms of Autoregulation
There are at lease two feedback mechanisms operating to maintain Ca2' homeostasis during the period when there is a sustained increase in the Ca2' influx rate. It is intuitively obvious that maintenance of homeostasis must involve, predominantly, the mechanisms by which the Ca2' efflux across the plasma membrane is controlled. As noted previously, there are two such efflux mechanisms: Na+/Ca2' countertransport via a 3Na+/Ca2+ exchanger and a 2H+/Ca2+ countertransport driven by the hydrolysis of ATP, i.e., a Ca2' pump.
Our discussion will be confined to considering the factors that regulate the pump. The most fundamental means for regulating the activity of this plasma membrane Ca2' pump is via Ca2+-calmodulin. Any increase in [Ca2'],m, whether due to an influx of Ca2+ across the plasma membrane or the release of Ca2+ from an intracellular pool, leads to an increase in the association of Ca2' CaM with the Ca2' pump protein. This increase results in an increase in Vmax of the Ca2t ransport rate and a decrease in the Km for Ca2+ (34).
Hence, any physiologically induced change in [Ca2']c leads to a prompt change in the activity of the pump. By this device, changes in both the amplitude and duration of the change in [Ca2+]c are minimized. Additionally, the Ca2'-dependent activation of protein kinase C leads to the phosphorylation of the Ca2' pump protein that causes an increase in Vmax of the transport system without altering its Km for Ca2+ (35).
Thus, during a sustained cellular response, two negative feedback loops link the activity of the Ca2' pump to changes in the Ca2+ influx rate.
Role of Ca2`during the Sustained Phase of the Cellular Response Given the fact that there are two distinct temporal phases of altered cellular Ca2+ metabolism during agonist action, it is not surprising that distinct phases of agonist-induced changes exist in protein kinase function. Evidence for such time-dependent activations of distinct protein kinases was found by analyzing the temporal patterns of agonist-induced changes in protein phosphorylation in angiotensin IImediated aldosterone secretion (36) and carbacholamine-induced smooth muscle contraction (37). The phosphoproteins that appear in conjunction with the [Ca2"], transient are thought to be substrates for CaM-dependent kinases; those which appear during the sustained phase in conjunction with the increased cycling of Ca2" across the plasma membrane are believe to arise as a consequence of protein kinase C activation. From these data a two-phase model of cell activation has been developed. During the first phase the important initiating event (occurring as a consequence of agonist-receptor interactions) is the IP3induced release of Ca2'; the result is a transient rise in [Ca2"], and the activation of CaM-dependent enzymes, including CaM-dependent protein kinases that phosphorylate a specific subset of cellular proteins responsible for initiating the cellular response.
During the second phase, the two important changes are the association of protein kinase C with the plasma membrane (which occurred during the initial phase as a consequence of the rise in the DG content of the plasma membrane and the [Ca2'], transient), and a sustained cycling of Ca2' across the plasma membrane. These two components are linked.
The rate of Ca2" cycling, or the submembrane Ca2' concentration [Ca2`]m determined by this cycling, controls the activity of the membrane-associated Ca2"-sensitive form of protein kinase C. The phosphorylation of a second subset of cell proteins by this kinase is the mechanism by which the response is sustained.
In this view, Ca21 is a critical messenger during both the initial and the sustained phase of the cellular response; however, the source of the messenger Ca2 , its site of action within the cell, and its molecular targets differ. During the initial phase, the Ca21 messenger arises from the intracellular trigger pool and acts within the total cytosolic domain to control the activities of CaM-regulated enzymes. During the sustained phase, the Ca2' messenger arises from the extracellular pool, acts in a restricted subcellular domain at the endoplasmic face of the plasma membrane, and along with receptor-generated diacylglycerol regulates the activity of protein kinase C.

Properties of the C-Kinase Transducer
In effect, during the process of cell activation, a new plasma membrane transducer has appeared. This transducer consists of two components: the membrane-associated, calcium-sensitive form of protein C kinase and the cycling of Ca2' across the cell membrane. This transducer has several interesting properties. First, under physiologic circumstances both components are necessary. During the initial phase of cell activation, there is a transient rise in [Ca2'], and a sustained or biphasic rise in the DG content of the plasma membrane. As a consequence, a percentage of the protein kinase C in the cytosol is translocated to the plasma membrane where the sustained increase in the DG content of the membrane assures its continued association (38). However, because [Ca2'], returns nearly to its basal value during this phase of the response, no additional C-kinase becomes associated with the membrane during the sustained phase of an agonist-induced response. Thus, it is postulated that the maximal capacity of the sustained response depends upon the amount of protein C kinase that becomes associated with the plasma membrane during the initial phase of the response.
The rate of Ca2" flux or Ca2" cycling, on the other hand, determines what percentage of this maximum is actually expressed in a given circumstance. Under normal physiologic conditions, the magnitude of the sustained response can be enhanced by increasing the rate of Ca2" influx. Hence, the membrane-associated protein C kinase is not normally maximally activated. Drugs or physiologic agents such as BAY K 8644, which do nothing more than increase or decrease Ca2+ flux across the membrane, can increase or decrease the magnitude of the response up to the limit imposed by the ability of the membrane-associated C-kinase molecules to respond to the Ca2' signal.

The C-Kinase Transducer and Cellular Memory
One unexpected feature of the sustained cellular responses mediated by this dual Ca2"-regulated information flow is the discovery that this control system displays a type of short-term memory (39). For example, when adrenal glomerulosa cells are exposed to angiotensin II for 30 min, the rate of aldosterone secretion reaches a plateau that will be maintained for several hours during continuous exposure to the angiotensin II. However, if after 30 min, the angiotensin II is removed, the rate of aldosterone secretion returns to the prestimulation value within 10 to 15 min. Likewise, the angiotensin II-induced increase in Ca2`influx rate is rapidly (within minutes) reversed.
However, if the same concentration of angiotensin II is applied 15 to 20 min after cessation of the previous exposure to angiotensin II, then the glomerulosa cells respond with a prompt increase in aldosterone secretory rate to a sustained plateau that is significantly higher than the rate seen in cells exposed continuously to the same concentration of angiotensin II. If this cycle is repeated, a third exposure to angiotensin II leads to an even greater sustained rate of aldosterone secretion, i.e., there is a cumulative memory. A similar phenomenon has been observed in tracheal and vascular smooth muscles when exposed intermittently to an appropriate agonist. Similarly, an exposure of islets of Langerhans to intermittent glucose pulses induces a similar kind of cumulative memory in terms of glucose-induced insulin secretion (40).
Hence, memory appears to be a common property of this control system.
The basis for this short-term memory in endocrine target cells remains to be fully defined. Nonetheless, it appears that protein kinase C plays a central role in this phenomenon. Our data are consistent with a model in which there are three states in which Ckinase may exist. In the nonactivated cell, this enzyme is located largely in the cytosol and exists in its Ca2"-insensitive state. Upon activation of the cell, a percentage of this cytosolic enzyme is translocated to the plasma membrane. The amount of enzyme translocated in this step depends upon the magnitude of the DG signal in the plasma membrane and the size of the [Ca2'], transient. Once the enzyme has associated with the membrane, it is transformed into its Ca2'-sensitive form; in this form its activity is regulated by the rate of Ca2' flux or cycling across the plasma membrane. It is postulated that with time this membrane-associated protein C-kinase undergoes some type of covalent modification that anchors the enzyme to the membrane. In this form the enzyme remains Ca2" sensitive; however, when agonist action is terminated and receptor activation ends, the anchored C-kinase leaves the membrane only slowly.
In contrast, the agonist-induced Ca2" influx rate reverses rapidly so that information flow through the C-kinase branch of the calcium messenger system ceases. Upon readdition of agonist, another bolus of protein C-kinase is translocated to the membrane. Hence, during the second exposure to agonist, the total amount of protein C-kinase associated with the membrane is greater than during the first exposure. As a consequence, the agonist-mediated sustained response is greater during the second exposure. This model appears to account for the short-term memory in endocrine systems and may also account for the phenomenon of associative learning in the marine snail Hermissenda (4) and in short memory in the mammalian nervous system.

The Protein C-Kinase Transducer and Tumor Promoters
Of particular importance to a consideration of toxicology is a group of drugs first identified as tumor promoters. These drugs, such as phorbol myristic acid, can replace DG as activators of protein kinase C (41). However, this class of compounds does more than simply activate protein kinase C. They act somewhat differently than does DG. First, their addition causes an association of protein C-kinase with the plasma membrane in the absence of a [Ca2+], transient; second, because of this relatively Ca2'-independent membrane association of protein C-kinase, a progressive accumulation of protein C-kinase within the membrane occurs; and third, the C-kinase that is associated with phorbol ester in the membrane is considerably more sensitive to activation by Ca2" than is the enzyme associated with DG in the membrane (42). Hence, this class of drugs can turn on the transducing system under circumstances where there is little or no change in cellular Ca2' metabolism. Because activation of the inositol phospholipid-calcium messenger system often serves as one of the inputs that mediate cell proliferation (21), these drugs can promote cell growth either appropriately, or more often inappropriately, in conjunction with other physiologic and/or environmental signals.

Conclusion
The highly sophisticated organization of the calcium messenger system attests to the fact that during evolution cells have devised a system by which Ca2" can serve as an intracellular messenger to mediate sustained cellular responses under conditions in which cellular Ca2" homeostasis is maintained. Nevertheless, the potential of cellular calcium intoxication persists. Circumstances that alter the properties of this calcium messenger system and, in particular, interfere with the autoregulatory system in the plasma membrane have the potential to lead to serious cellular dysfunction and death.
Supported by grant AM19813 from the National Institutes of Health.