I.  INTRODUCTION

Top Previous Next References

Cytosolic Ca²⁺ plays a key role in intracellular signaling in virtually all types of animal cells. In many instances, as exemplified by cardiac muscle contraction and the transient release of neurotransmitter substances, the elevated cytosolic free Ca²⁺ concentration ([Ca²⁺]_i) may be required for only brief periods of time.¹ The Ca²⁺ must then be rapidly removed from the cytosol. In other situations, such as during the maintenance of tonic tension in smooth muscle, elevated [Ca²⁺]_i may be required for long periods of time. Consequently, [Ca²⁺]_i 1) must be precisely controlled, 2) must be able to be increased and decreased rapidly to provide transient signals, and 3) must be modulated so that steady-state [Ca²⁺]_i can be altered to fit cell needs. Moreover, in at least some cells (neurons are a good example), there must be spatial as well as temporal variation in [Ca²⁺]_i so that multiple Ca²⁺-regulated processes within these cells are not all indiscriminately activated simultaneously. Furthermore, it follows that abnormal cell Ca²⁺ regulation is likely to have profound pathophysiological consequences. For these several reasons, much interest has focused on the regulation of [Ca²⁺]_i in all types of cells.

Calcium is uniquely suited for these signaling functions, due, in part, to its abundance (Ca²⁺ is the most prevalent cation in the body) as well as to its unusual chemistry (the combination of ionic radius, charge density, and coordination number) (981). Furthermore, the resting (or "background") intracellular free Ca²⁺ concentration {[Ca²⁺]_i(rest)} is sufficiently low (on the order of 10⁷ M; Refs. 89, 109, 156, 939) so that a large signal-to-background ratio {[Ca²⁺]_i/[Ca²⁺]_i(rest)} can be generated with only a relatively small increment in the absolute amount of Ca²⁺ added to the cytosol (89). This "signal Ca²⁺" ([Ca²⁺]_i) can come either from the extracellular fluid, via voltage-gated or receptor-operated Ca²⁺-selective channels located in the plasma membrane (PM) or from intracellular stores [especially endoplasmic reticulum (ER) or, in muscle, the sarcoplasmic reticulum (SR)]. Calcium is released from these intracellular stores by Ca²⁺-activated or inositol trisphosphate-activated mechanisms (63, 64, 112, 883). Moreover, the amount of Ca²⁺ sequestered in the stores, and thus potentially available for release during cell activation, is largely dependent on [Ca²⁺]_i(rest). An additional critical feature of Ca²⁺ metabolism is its slow diffusion in the cytosol (102, 419, 543). This is largely a consequence of Ca²⁺ buffering by cytoplasmic proteins such as parvalbumin and vitamin D-dependent Ca²⁺-binding protein (92, 93, 156, 994) and Ca²⁺ sequestration in intracellular organelles such as the SR/ER (105, 106, 647, 739) and mitochondria (277, 404, 459, but see Refs. 157, 882). Therefore, it is possible to generate local increases in [Ca²⁺]_i, and local responses, without having to raise [Ca²⁺]_i throughout the cell. This is energetically important because only limited energy expenditure is then required to remove this "local" Ca²⁺ from the cytosol.

Calcium-dependent signaling is mediated by specific receptors for Ca²⁺ such as calmodulin, troponin C, protein kinase C (PKC), and synaptotagmin. Cells must also have mechanisms to remove Ca²⁺ from the cytosol. Two important mechanisms are the ATP-driven Ca²⁺ pumps in the ER (or SR) for resequestering Ca²⁺, and in the PM, for extruding Ca²⁺ and helping to maintain steady Ca²⁺ balance. These ATP-driven Ca²⁺ pumps bind Ca²⁺ with high affinity but have relatively low turnover rates (i.e., the number of Ca²⁺ transported per carrier per unit time), typically ~100 s¹ (155, 156, 160, 240, 242). Their capacity to transport Ca²⁺ depends on both the turnover rate and the density of transport molecules in the membrane.

The PM of most animal cells also contain another Ca²⁺ transport system, the Na⁺/Ca²⁺ exchanger, that operates in parallel with the Ca²⁺-selective channels and the ATP-driven Ca²⁺ pump. The Na⁺/Ca²⁺ exchanger can move Ca²⁺ either into or out of cells, depending on the net electrochemical driving force on the exchanger. Interestingly, this driving force, and thus the net Ca²⁺ movement mediated by the exchanger, may change direction during a cell's activity cycle, when the membrane potential (V_M) varies (e.g., during an action potential) and/or when the cytosolic Na⁺ concentration ([Na⁺]_i) or [Ca²⁺]_i is altered. As described in section IIIB6, the Na⁺/Ca²⁺ exchanger has about a 10-fold lower affinity for Ca²⁺ but 10- to 50-fold higher turnover rate than the ATP-driven Ca²⁺ pumps.

Sodium-dependent Ca²⁺ transport has also been described in other organelles, including mitochondria (156, 214, 372) and secretory vesicles (451, 538, 539, 724, 804). This review is limited, however, to an examination of the properties of the PM Na⁺/Ca²⁺ exchanger.

A puzzling feature of cell Ca²⁺ regulation concerns the apparent redundancy of the PM Ca²⁺ transport systems, since the Na⁺/Ca²⁺ exchanger operates in parallel with both the Ca²⁺-selective channels and the ATP-driven pump. As we shall see, however, the exchanger has several unique features that may help to explain the need for this redundancy as well as the exchanger's utility. For example, the high maximal turnover rate of the Na⁺/Ca²⁺ exchanger may be important in cells that need to extrude a lot of Ca²⁺ rapidly (as do cardiac muscle cells; see sect. VIA).² Indeed, the rate of Ca²⁺ mediated by the exchanger increases and decreases as [Ca²⁺]_i is raised and lowered, respectively, during the cell activity cycle so that its activity is modulated to meet cellular demands for rapid transport of Ca²⁺ and regulation of [Ca²⁺]_i. Also, the electrochemical driving force on the exchanger in excitable cells is usually altered markedly during the cell cycle. Therefore, in some instances (such as during the plateau of the cardiac action potential), the decrease in the driving force will reduce the net rate of Ca²⁺ extrusion compared with the rate expected at 80 mV. Thus the exchanger may help to maintain [Ca²⁺]_i at a relatively elevated level for an extended period of time. In addition, the efflux of Ca²⁺ via the exchanger is associated with an inward current (see sect. IIID) that will tend to prolong the cardiac action potential (367). The time course of these effects will depend on the time course of the action potential and on that of the Ca²⁺ transient. During quiescent periods, the electrochemical driving force (_Na/Ca; see below) on the exchanger can be modulated by altering V_M or [Na⁺]_i. Changes in these parameters may therefore be used to modulate quasi-steady-state [Ca²⁺]_i as well as the amount of Ca²⁺ sequestered in the intracellular stores. This may be particularly important in cells that must sustain Ca²⁺-dependent activities (such as tonic tension in certain smooth muscle cells). One feature of the exchanger that has recently come to light and that may contribute in important ways to its physiological role, is its localization in restricted portions of the PM of at least some types of cells. As discussed in section VA, certain transport proteins are not universally distributed in the PM but are localized in particular domains of the PM.

The Na⁺/Ca²⁺ exchanger was first described in the late 1960s. Since then, it has been identified in most types of cells in higher animals in which assessment has been carried out; the human erythrocyte is a noteworthy exception. The properties of the exchanger appear to be very similar in most of these cells. The exchanger in vertebrate photoreceptors has some unusual ion coupling features, however, and is the product of a separate gene (see sect. IV). Therefore, after a brief historical note, we review the general properties of the exchangers found in most cells. The various modes of transport mediated by the Na⁺/Ca²⁺ exchanger are described, and a summary of the exchanger's thermodynamic and kinetic properties is presented in the context of overall cell Ca²⁺ homeostasis. We then discuss the molecular structure of the exchanger and its pharmacology. The final portion of this article is devoted to a discussion of the physiological roles of the Na⁺/Ca²⁺ exchanger (and possible pathophysiological implications) in various tissues in which the exchanger has been identified.

A number of reviews published during the past several years have covered selected aspects of the Na⁺/Ca²⁺ exchanger and its physiological role in specific tissues. Some of the more extensive reviews published within the last decade include those by DiPolo and Beaugé (250, 254) on excitable cells, by Philipson and Nicoll (721, 722) on the molecular and kinetic properties of the exchanger, by Reeves and co-workers (759-761) and Khananshvili (489) on the cardiac exchanger, by Sheu and Blaustein (845) on the cardiac and vascular smooth muscle exchangers, and by Schnetkamp (832) and Lagnado and McNaughton (547) on the vertebrate photoreceptor exchanger. In addition, many aspects have been summarized in the proceedings of three international conferences on the Na⁺/Ca²⁺ exchanger (13, 97, 415). The literature on the Na⁺/Ca²⁺ exchanger has exploded in the last few years. Major advances have been made in exploring the molecular biology of the exchanger. Moreover, there has been renewed emphasis on exchanger kinetics, mechanism, and regulation. A new comprehensive review now seems warranted to summarize and integrate the most recent literature and latest views on the properties, physiological roles, and possible pathophysiological significance of the Na⁺/Ca²⁺ exchanger. In many instances, it is not yet clear how the various observations fit together, and there are still numerous controversies. Nevertheless, we have endeavored to present a full spectrum of new concepts and observations in the context of earlier work that established the basic framework. A number of areas that are in need of further investigation are identified. We have tried to provide the reader with a current perspective that we hope will not be misleading. Perhaps our review will encourage others to prepare reviews on more specialized aspects of Na⁺/Ca²⁺ exchange that have been touched on only superficially in this very broad review.

Our review of the literature was completed in 1998. Because of the continually expanding literature on the subject of Na⁺/Ca²⁺ exchange, we are able to provide only selected citations to some of the vast literature (especially on Na⁺/Ca²⁺ exchange in the heart, about which much has been written). We regret omissions of specific citations, including some "first reports," and the citation to several reviews instead. Nevertheless, we hope that we have provided a useful perspective of the current status of many aspects of Na⁺/Ca²⁺ exchange.

	II.  HISTORICAL BACKGROUND

Top Previous Next References

In the 1880s, Ringer (785, 786) demonstrated that Ca²⁺ in the bathing medium were required for cardiac muscle contraction. Nearly 40 years later, Daly and Clark (221) drew attention to the fact that the effect of extracellular Na⁺ removal "shows a striking resemblance to the effect of strophanthidin" in its ability to enhance cardiac muscle contraction. Then, in 1948, these two observations were brought together, when Wilbrandt and Koller (980) demonstrated an interaction between Na⁺ and Ca²⁺ at the PM. They found that the strength of contraction of the frog's heart was directly related to the ratio [Ca²⁺]_o/([Na⁺]_o)², where subscript o refers to extracellular. Ten years later, Luttgau and Niedergerke (611) showed that, even in depolarized frog cardiac muscle, removal of external Na⁺ induced contractures, and in 1963, Niedergerke (690) demonstrated that frog ventricles gained Ca²⁺ when exposed to Na⁺-depleted solutions. These ventricles promptly lost Ca²⁺ and relaxed when they were returned to control (Na⁺-rich) Ringer solution. On the basis of these observations, Niedergerke (690) postulated that Ca²⁺ enters the muscle fibers via a carrier mechanism for which external Na⁺ and Ca²⁺ compete, presumably in the ratio 2 Na⁺:1 Ca²⁺.

Also in the mid 1960s, attention was directed toward the possible role of cell Na⁺ accumulation to explain the positive inotropic responses associated with cardiotonic steroids (775) and the staircase phenomenon (treppe) in the heart (550); however, no specific mechanisms were proposed. Subsequently, Langer (551) suggested that Na⁺ and Ca²⁺ might compete at intracellular membrane sites (on the sarcotubular system).

Competition between Na⁺ and Ca²⁺ at the PM was also observed in both twitch and slow skeletal muscle in the frog. In the rectus abdominis, a slow muscle, reduction of [Na⁺]_o caused sustained contractures (818). Cosmos and Harris (207) found that frog twitch muscle fibers gained Ca²⁺ when incubated in Na⁺-depleted media. This led them to conclude that in this tissue, too, Na⁺ and Ca²⁺ compete for entry sites. With hindsight, it is also possible to explain some of their other observations in terms of an Na⁺/Ca²⁺ exchange mechanism, namely, conditions that induced net Na⁺ gain (removal of external K⁺, cooling, or treatment with ouabain) also promoted net Ca²⁺ gain in frog muscle. They used radioactive tracer labeled Sr²⁺ (⁸⁹Sr) as a substitute for Ca²⁺ in efflux experiments; after ⁸⁹Sr loading, external Na⁺ promoted (net) ⁸⁹Sr efflux from this tissue.

Similarly, reports of apparent competition between Na⁺ and Ca²⁺ in mammalian smooth muscles were published in the 1960s. The key evidence was that the Ca²⁺ content of guinea pig tenia coli (362) and rabbit aorta (136) increased when [Na⁺]_o was lowered. Even earlier, Friedman and co-workers (323, 324) had suggested that blood pressure was, in part, dependent on "sodium transfer systems" that govern the influx and efflux of Na⁺ in vascular smooth muscle cells, but the importance of Ca²⁺ was not yet recognized.

These findings were reinterpreted and extended in the late 1960s by investigators in three different laboratories: Reuter and co-workers in Germany and Switzerland (783, 352); Baker, Blaustein, and Hodgkin and their co-workers in England (34, 35, 102); and Martin and DeLuca in the United States (625). While studying different tissues, they independently and simultaneously reached the conclusion that a coupled countertransport mechanism, involving the exchange of Na⁺ for Ca²⁺ (i.e., "Na⁺/Ca²⁺ exchange"), rather than just a competition between Ca²⁺ and Na⁺ for binding sites, may be involved in Ca²⁺ transport across the PM in mammalian cardiac muscle (352, 783), in invertebrate neurons (34, 35, 102), and in mammalian small intestinal epithelium (625). It is noteworthy that during the 1950s and 1960s experiments were also being carried out on the transport of small solutes (sugars and amino acids) that revealed a coupling between the fluxes of these other solutes and of Na⁺ (181, 182, 211, 784, 836). In these latter cases, Na⁺ was cotransported with the sugars or amino acids. This coupling of the Na⁺ electrochemical gradient (145) to the transport of other solutes (see sect. IIIB) is often referred to as "secondary active transport." It is widely recognized as a general mechanism that has been adapted in a variety of ways for the transport, against a concentration or electrochemical gradient (see sect. IIIB), of numerous ions and other solutes in virtually all types of animals cells.

In studies on guinea pig, sheep, and calf heart, Reuter and Seitz (782, 783) showed that Ca²⁺ efflux from ⁴⁵Ca-loaded auricular and ventricular muscle was largely dependent on external Na⁺ and Ca²⁺. They interpreted these observations as evidence for a Na⁺/Ca²⁺ exchange mechanism (see Fig. 1). Subsequently, Glitsch et al. (352) found that Ca²⁺ influx in mammalian cardiac muscle was proportional to ([Na⁺]_i)² over a wide range of [Na⁺]_i (see Ref. 85).

View larger version (17K): [in this window] [in a new window]   Fig. 1. Effects of removing extracellular Ca2+ concentration ([Ca2+]o) and extracellular Na+ concentration ([Na+]o) on Ca2+ efflux from a 45Ca-loaded guinea pig auricle (top) and a sheep ventricular trabeculum (bottom). Ordinate shows 45Ca efflux rate constant. Concentrations of Na+ and Ca2+ present in various external solutions at various times are indicated between the two curves; in Na+-free solutions, NaCl was replaced by isosmotic sucrose. Numbers just above each curve indicate fractional rates of 45Ca efflux in various solutions, relative to that in control Tyrode solution which contained 1.8 mM Ca2+ and 149 mM Na+. Note that most of Ca2+ efflux into Ca2+-free Tyrode is dependent on external Na+. [From Reuter and Seitz (783).]

At about the same time, Baker et al. (35) observed an external Ca²⁺-dependent, ouabain-insensitive Na⁺ efflux from squid axons (Figs. 2 and 3). In addition, Ca²⁺ influx appeared to be a function of the [Na⁺]_o/[Na⁺]_i ratio; increasing [Na⁺]_i and/or decreasing [Na⁺]_o (Fig. 3) promoted Ca²⁺ influx in this preparation (444). Similar observations were also made on crab nerve (34). The fact that the Ca²⁺ efflux from ⁴⁵Ca-injected squid axons was, in part, dependent on external Na⁺ (102) fostered the idea that Ca²⁺ extrusion involved a Na⁺/Ca²⁺ exchange mechanism. In sum, these Ca²⁺ influx and efflux data seemed consistent with a transport mechanism that could move Ca²⁺ in either direction across the PM, in exchange for Na⁺; net Ca²⁺ movement (and the Ca²⁺ gradient) appeared to depend on the prevailing Na⁺ electrochemical gradient _Na (35). The Na⁺/Ca²⁺ exchange systems in squid axons and mammalian cardiac muscle appeared to be similar, if not identical. Therefore, these findings led directly to the suggestion that the positive inotropic action of cardiac glycosides was due to a Na⁺/Ca²⁺ exchanger-mediated rise in cell Ca²⁺, as a result of Na⁺ pump inhibition and the consequent reduction in _Na across the sarcolemma (35 and see sect. VIA7). One feature that has challenged this hypothesis is the question of whether the very limited Na⁺ pump inhibition produced by low (nanomolar) concentrations of cardiotonic steroids is sufficient to account for the observed enhancement of contraction (see sect. VIA7).

View larger version (15K): [in this window] [in a new window]   Fig. 2. Effect of external cation concentration on Na+ efflux from squid giant axons bathed in media containing 105 M ouabain in presence (open symbols) and absence (closed symbols) of Ca2+. A: mean data from 5 experiments. B: external Ca2+-dependent Na+ efflux difference curves (curve 1 minus curve 2 and curve 3 minus curve 4 from A). Ordinates indicate [Na+]o or [Li+]o; at reduced alkali metal ion concentrations ([M+]o), monovalent cation was replaced by isosmotic dextrose. Curves in B were calculated from a model described in original article (35). These curves show that low [Na+]o and [Li+]o stimulate ouabain-insensitive, external Ca2+-dependent Na+ efflux and that high [Na+]o (>100 mM), but not high [Li+]o, inhibits this flux. [From Baker et al. (35).]

View larger version (15K): [in this window] [in a new window]   Fig. 3. Ca2+ influx (left) and external Ca2+ (Cao)-dependent Na+ efflux (right) in squid giant axons, graphed as a function of [Na+]o. At reduced [Na+]o, Na+ was replaced by isosmotic dextrose (dextrose-Na curves, ) or Li+ (Li-Na curves, ). All external solutions were K+ free and contained 105 M ouabain. Na+ efflux data (right) are shown as either efflux rate contants (left ordinate scale) or calculated Cao-dependent effluxes (right ordinate scale). Na+ efflux curves were calculated from a model described in original article (35). Both Ca2+ influx (left) and ouabain-insensitive, Cao-dependent Na+ efflux (right) are activated by low [Na+]o and inhibited by high [Na+]o. [From Baker et al. (35).]

Martin and DeLuca (625) examined the effect of Na⁺ in the mucosal and serosal fluids on the transepithelial transport of Ca²⁺ in rat duodenum. They found that Na⁺ was required in the serosal fluid to effect the net transfer of Ca²⁺ from mucosa to serosa. They therefore suggested that the _Na across the basolateral membrane of the columnar epithelial cells powered the extrusion of Ca²⁺ from the cytoplasm into the serosal medium via a Na⁺/Ca²⁺ exchange mechanism.

The regulation of the exchanger by ATP-dependent phosphorylation and by internal Ca²⁺ was first hinted at by the observations of Baker and Glitsch (36). Subsequently, DiPolo (237, 238) clearly demonstrated the activation of exchanger-mediated Ca²⁺ efflux by ATP, and Blaustein (88) showed that the effect was due to a shift in the cation activation kinetics. DiPolo (239, 244; see also Ref. 37) also made the unanticipated (and counterintuitive) observation that the Ca²⁺ entry (and Na⁺ exit) mode of exchange exhibited an absolute requirement for internal Ca²⁺. This observation presages the finding that intracellular Ca²⁺ catalyzes the turnover rate of the the Na⁺/Ca²⁺ exchanger (see sect. IIE4).

Early observations that the Ca²⁺ efflux was activated by the cooperative action of more than two Na⁺ provided a clue that the coupling ratio of the exchanger was >2 Na⁺:1 Ca²⁺ (85). This led to the view that the exchange was voltage sensitive (108, 670). Subsequently, Hume and Uehara (438, 439) observed "creep currents" in cardiac muscle, which they attributed to a rheogenic Na⁺/Ca²⁺ exchange. Kimura, Noma, and co-workers (514, 515) then clearly identified a current attributable to the Na⁺/Ca²⁺ exchanger in patch-clamped cardiac myocytes (Fig. 4). With the development of the giant membrane patch method by Hilgemann and co-workers (199, 405), rapid and accurate measurement of a current attributable to Na⁺/Ca²⁺ exchange in cardiac myocyte membranes could be made.

View larger version (22K): [in this window] [in a new window]   Fig. 4. Activation of Cao-dependent outward (Na+/Ca2+ exchange) current by Cao in a single patch-clamped guinea pig cardiac myocyte. A: voltage-clamp records showing voltage (top traces) and current (bottom traces). Ramp pulses from a holding potential of 30 to +60 mV, and then to 120 mV, at a ramp speed of 0.2 V/s, were given every 10-15 s. In first panel, patch pipette (internal solution) contained 30 mM Na+ and 42 mM EGTA (no free Ca2+), and external solution was also Ca2+ free; effect of [Ca2+]o = 1 mM (added during period indicated by bar below current trace) is shown. In second and third panels, pipette solution contained 73 nM free Ca2+; third panel shows effect of adding 1 mM free Ca2+ to external solution. External solutions all contained 1 mM BaCl2 and 2 mM CsCl to block K+ channels, 2 µM methoxyverapamil to block Ca2+ channels, and 20 µM ouabain to block Na+ pump. Pipette solutions all contained 126 mM CsCl to reduce current through K+ channels. B and C: current-voltage (I-V) curves were obtained from descending portions (+60 to 120 mV) of ramp pulses. In B, curves were obtained before (a) and during (b) superfusion with solution containing 1 mM Ca2+ under internal Ca2+-free conditions. There is no difference between these curves (i.e., no Cao-activated outward current in absence of internal Ca2+). In C, I-V relations were obtained with internal Ca2+ (Cai) equal to 73 nM before (c) and during (d) superfusion with solution containing 1 mM Ca2+. In this case, there is a Cao-activated outward current (d minus c) that is also dependent on internal Ca2+; current also depended on internal Na+. [From Kimura et al. (515).]

The apparent stoichiometry (coupling ratio) of the exchanger was established on the basis of a large variety of tracer flux and electrophysiological experiments in various tissues. These included cardiac muscle (285, 514, 726, 762, 765), squid axons (249), and barnacle muscle (757, 754). In all of these tissues, a coupling ratio of 3 Na⁺:1 Ca²⁺ was observed.

Additional properties of the exchanger were elucidated as a result of the application of other novel methods: ion-selective Ca²⁺ and Na⁺ electrodes (846, 847), release of caged Ca²⁺ (695), and electrical measurements of transport in reconstituted proteoliposomes fused to a planar lipid bilayer (289).

Experiments on vertebrate photoreceptors by Yau and Nakatani (997) demonstrated that these cells, too, also contain a PM Na⁺/Ca²⁺ exchanger. The studies of Schnetkamp et al. (830) and Cervetto, McNaughton, and co-workers (165, 643), however, revealed that the vertebrate photoreceptor exchanger, unlike the cardiac/neuronal Na⁺/Ca²⁺ exchanger (261, 996), also has an absolute requirement for K⁺ on the same side as Ca²⁺ to effect Ca²⁺ transport. The coupling ratio of this vertebrate photoreceptor exchanger was then found to be 4 Na⁺:1 Ca²⁺ + 1 K⁺ (165).

Twenty years ago, Reeves and Sutko (764) launched the biochemical approach to the exchanger with their seminal studies of Na⁺/Ca²⁺ exchange in cardiac sarcolemmal vesicles. This led to the partial purification of the cardiac exchanger by Philipson et al. (719) and to the eventual cloning of this exchanger by Nicoll et al. (688). The exchanger from bovine rod outer segments was purified by Cook and co-workers (206, 767) (and see also Ref. 575), and subsequently cloned by Reilander et al. (769). The photoreceptor exchanger and the cardiac/neuronal exchanger have very different primary sequences, and it seems clear that these two types of exchangers evolved independently. This encourages speculation about the physiological need for such exchangers and the process of evolution, a topic that is discussed in section IVI1.

Very recently, Tsoi et al. (940) cloned a novel exchanger from rat brain. The exchanger expressed from this clone is very homologous to the photoreceptor exchanger, and Ca²⁺ transport mediated by this brain exchanger, like the photoreceptor exchanger, also requires K⁺ (see also Ref. 219). Surprisingly, as discussed in sections VA and VIE, both the K⁺-independent (cardiac-type) and K⁺-dependent (vertebrate photoreceptor-type) exchangers appear to be coexpressed in some of the same neurons (465, 940). These observations open important new avenues for exploration, just as they raise critical questions about our previously conceived notions about the physiological role(s) of Na⁺/Ca²⁺ exchange.

	III.  PROPERTIES OF THE SODIUM/CALCIUM EXCHANGER

Top Previous Next References

The observations on Na⁺/Ca²⁺ exchange in nerve, muscle, and epithelia made during the late 1960s provided a starting point and a paradigm for numerous studies on Ca²⁺ transport that have been carried out during the past three decades. A Na⁺/Ca²⁺ exchanger has now been identified in most types of tissues, in both vertebrates and invertebrates. Here we describe the properties of the very widely distributed "cardiac/neuronal" type of exchanger that is found in most tissues. Some of the unusual properties of the vertebrate photoreceptor exchanger, already alluded to above, are also discussed.

Initial studies in several preparations indicated that the Na⁺/Ca²⁺ exchanger could move net Ca²⁺ either out of or into cells depending on the prevailing electrochemical driving forces (i.e.., the Na⁺ and Ca²⁺ concentration gradients and the V_M). These two modes of Na⁺/Ca²⁺ exchange (see Figs. 1 and 3) will be referred to, respectively, as "Ca²⁺ exit mode" and "Ca²⁺ entry mode" exchange. The often used terms forward mode and reverse mode will be avoided because the latter may be construed, incorrectly, as an abnormal mode of exchange. Indeed, a very important part of the exchanger's function may be to promote Ca²⁺ entry (and to reduce Ca²⁺ exit) under some normal physiological conditions. A clear-cut example is in the erythrocytes of certain mammals such as dogs and ferrets (see sect. VIF1), where the Na⁺/Ca²⁺ exchanger plays an important role in the extrusion of Na⁺.

The Ca²⁺ entry mode of exchange is usually identified as an internal Na⁺ (Na_i)-dependent Ca²⁺ influx and an external Ca²⁺ (Ca_o)-dependent, ouabain-insensitive Na⁺ efflux. A critical feature of the (net) Ca²⁺ influx mode of operation (and also the Ca²⁺ efflux mode) is that it is dependent on (nontransported) intracellular Ca²⁺ (37, 239, 244, 246, 258, 406, 514, 515, 757). The [Ca²⁺]_i required for half-maximal activation (i.e., K_0.5 for intracellular Ca²⁺) of Ca²⁺ influx mode exchange in squid axons (239, 244, 254) and cardiac myocyte giant patches (406, 412; but see Ref. 514) is on the order of 1 µM under physiological conditions. Consequently, only a small fraction of the exchangers are active at the normal resting [Ca²⁺]_i (~100 nM) in most cells. In contrast, the exchanger is fully activated when [Ca²⁺]_i is in the low micromolar range (i.e., at about the [Ca²⁺]_i expected during peak activity in many types of excitable and secretory cells). The kinetic and physiological consequences of this regulation by internal Ca²⁺ are discussed in section IIIE4.

Another surprising feature of Ca²⁺ influx mode exchange (see Figs. 2 and 3) is its activation by low concentrations of external alkali metal ions, including Na⁺ (35, 98, 314, 336); the activating monovalent cations are not transported (57). High [Na⁺]_o reduces the exchanger-mediated Ca²⁺ influx, however. The kinetics of this activation by alkali metal ions and the biphasic actions of external Na⁺ are also described in section IIIE4.

This mode of exchange is defined operationally as an external Na⁺ (Na_o)-dependent Ca²⁺ efflux and an internal Ca²⁺ (Ca_i)-dependent ouabain- and tetrodotoxin (TTX)-insensitive Na⁺ influx. Like the Na_i-dependent Ca²⁺ influx, the Na_o-dependent Ca²⁺ efflux is activated by internal alkali metal ions and is markedly inhibited by high concentrations of internal Na⁺ (88) (but see Refs. 249, 254). The Ca²⁺ efflux mode of exchange also appears to be activated by nontransported intracellular Ca²⁺ (412) (and see sect. IIIE4), even though the K_0.5 for regulation and for transport may be similar. In contrast to Ca²⁺ entry mode exchange, however, this Ca²⁺ efflux does not require extracellular Ca²⁺ for activation (88, 249); this is clear functional evidence that the exchanger is asymmetric.

The hydrolysis of ATP is not required to power net Ca²⁺ extrusion mediated by the Na⁺/Ca²⁺ exchanger, and Na⁺/Ca²⁺ exchange can be demonstrated in ATP-depleted cells (88, 102, 244). Nevertheless, as discussed in section IIIE1, cytosolic ATP does substantially alter the exchanger kinetics (88), presumably by phosphorylating critical sites on the exchanger molecules (237, 238). This is further evidence that the exchanger is asymmetric. The asymmetry apparently enables the exchanger to function optimally because of the differences in the affinities for Ca²⁺ on the two sides of the membrane and the modulation of the transport kinetics and turnover by various ligands.

The identification of both Ca²⁺ entry and Ca²⁺ exit modes of exchange demonstrated that the exchanger can mediate the bidirectional movements of both Na⁺ and Ca²⁺ across the PM. This raised the question of whether the exchanger could also mediate the coupled exchange of internal Ca²⁺ for external Ca²⁺ (i.e., Ca²⁺/Ca²⁺ exchange) and/or internal Na⁺ for external Na⁺ (Na⁺/Na⁺ exchange) (see Fig. 5). The presence of these two modes of exchange, while perhaps not functionally important in terms of net ion translocation, could provide critical insight into the transport mechanism.

View larger version (18K): [in this window] [in a new window]   Fig. 5. Effects of ionized Cai and ATP on ouabain- and bumetanide-insensitive Na+ efflux from an internally dialized squid giant axon. Concentrations of certain key components in internal (int) and external (ext) solutions are shown at top and are in mM, except for Cai, which is in µM. Arrows indicate changes in internal solutions; different symbols indicate different external solutions. All external solutions contained 300 nM tetrodotoxin (blocks Na+ channels), 0.5 mM ouabain (blocks Na+ pump), 10 µM bumetanide (blocks Na+/K+/Cl cotransport), and 1 mM cyanide (blocks mitochondrial respiration). Cai-activated increase in Na+ efflux into 0Na-10Ca (Nao-free, 10 mM Ca) medium under ATP-free conditions corresponds to activation of Ca2+ entry mode Na+/Ca2+ exchange by Cai. In absence of external Ca2+ (0Ca, open squares) and presence of 80 µM Cai, addition of Nao activated an Na+ efflux (Na+/Na+ exchange) that was not observed in absence of Cai (solid circles at start of experiment). This efflux was augmented by internal ATP, as was Na+/Ca2+ exchange (half-filled circles at right). [From DiPolo and Beaugé (246).]

The Ca²⁺/Ca²⁺ exchange is usually identified as a trans-Ca²⁺-dependent Ca²⁺ flux (i.e., Ca_o-dependent Ca²⁺ efflux and Ca_i-dependent Ca²⁺ influx), where trans refers to ions acting on opposite sides of the PM. On the basis of this criterion, a Ca²⁺/Ca²⁺ exchange that appears to be mediated by the Na⁺/Ca²⁺ exchanger has been identified in various types of cells. A few examples are squid axons (35, 88, 107, 249, 254), barnacle muscle (686, 754, 757, 800), mammalian neurons (98), and cardiac muscle (763, 863).

Two lines of evidence indicate that these Ca²⁺ fluxes are mediated by the Na⁺/Ca²⁺ exchanger. One stems from the observations that the Ca_i-dependent Ca²⁺ influx is activated by external (i.e., on the same side of the PM, or cis to Ca²⁺) alkali metal ions and is inhibited by external Na⁺ in much the same way as is Ca²⁺ entry mode exchange (35, 107, 108). In this case, however, internal alkali metal ions (e.g., K⁺ or Li⁺) are also required for activation of the exchange, and high [Na⁺]_i is also markedly inhibitory (88). Second, the Na⁺/Ca²⁺ exchange and Ca²⁺/Ca²⁺ exchange do not appear to be independent because they cannot both operate simultaneously at maximal rates. Inhibition of Ca²⁺/Ca²⁺ exchange by external Na⁺ is associated with a shift from Ca²⁺/Ca²⁺ exchange to Ca²⁺ exit mode exchange (88). Conversely, inhibition by internal Na⁺ (107, 138) is likely due to a shift in the operation from Ca²⁺/Ca²⁺ exchange to Ca²⁺ entry mode exchange.

The Ca²⁺/Ca²⁺ exchange mode exhibits a coupling ratio of 1 Ca²⁺:1 Ca²⁺ (107) yet is (surprisingly) voltage sensitive (249) but electroneutral (562). There is no net transfer of charge during Ca²⁺/Ca²⁺ exchange, in contrast to the situation during Na⁺/Ca²⁺ exchange (see below). Whether or not the activating alkali metal ions are also transported during Ca²⁺/Ca²⁺ exchange is not known. However, the fact that cis-activating alkali metal ions (other than Na⁺) do not appear to be transported during Na⁺/Ca²⁺ exchange (57, 996) (and see below) suggests that they may not be transported during Ca²⁺/Ca²⁺ exchange.

Adenosine 5'-triphosphate also modulates the kinetics of Ca²⁺/Ca²⁺ exchange (686, 249). The Ca²⁺/Ca²⁺ mode of exchange can proceed at a very high rate in ATP-depleted cells (686). Adenosine 5'-triphosphate appears to inhibit the Ca²⁺/Ca²⁺ exchange mode, probably by increasing the affinity of the exchanger for Na⁺ (88).

In sum, all of these studies are consistent with the view that the Na⁺/Ca²⁺ exchanger can also mediate Ca²⁺/Ca²⁺ exchange.

Evidence for coupled Na⁺ influx and Na⁺ efflux mediated by the Na⁺/Ca²⁺ exchanger was first obtained in cardiac sarcolemmal preparations by Reeves and Sutko (764, 759). They described a Na⁺ efflux that was dependent on external Na⁺; activation was sigmoid, with a K_0.5 of 15 mM. DiPolo and Beaugé (246) subsequently identified, in squid axons, a ouabain-insensitive component of Na_o-dependent Na⁺ efflux that corresponds to a Na⁺/Na⁺ exchange (see Fig. 5) mediated by the Na⁺/Ca²⁺ exchanger. This view is based on the evidence that the Na_o-dependent Na⁺ efflux, like the Ca²⁺ entry mode of exchange, is 1) dependent on internal Ca²⁺ for activation, as is the Ca²⁺ entry mode of exchange; 2) activated by internal alkalinization; 3) inhibited by internal Mg²⁺; 4) activated by internal ATP and by adenosine 5'-O-(3-thiotriphosphate) which is a substrate for kinases but not ATPases (370, 793); and 5) inhibited by external Ca²⁺ under conditions that promote Ca²⁺ entry mode exchange (i.e., external Ca²⁺ shifts the exchanger from Na⁺/Na⁺ exchange mode to Ca²⁺ entry mode). A Ca_i-independent component of Na_o-dependent Na⁺ efflux was also identified; this flux had different properties and did not appear to be mediated by the Na⁺/Ca²⁺ exchanger (246).

A few attempts have been made to study half-cycles or single complete cycles of the Na⁺/Ca²⁺ exchanger to identify and characterize intermediate steps in the exchange cycle (407, 414, 693, 694). Nevertheless, there is no evidence that the exchanger can mediate significant net efflux of Ca²⁺ (or Na⁺) in the absence of external Na⁺ or Ca²⁺, or net influx of Ca²⁺ (or Na⁺) in the absence of internal Na⁺ or Ca²⁺. Thus it appears that the coupled countertransport is an obligatory feature of the mechanism: the exchanger does not mediate "uncoupled" transport of either Na⁺ or Ca²⁺. In other words, unloaded forms of the exchanger molecules apparently do not cycle or cycle very slowly.

B.  Energetics and Kinetics of Na⁺/Ca²⁺ Exchange

1.  Thermodynamic considerations

In the resting cells of higher animals, [Na⁺]_i and [Ca²⁺]_i are on the order of 10 mM and 0.1 µM (for present purposes, we refer to concentrations rather than activities; but see Ref. 85). The extracellular Na⁺ and free Ca²⁺ concentrations ([Na⁺]_o and [Ca²⁺]_o, respectively) are usually on the order of 450 mM and 3-4 mM in marine invertebrates (85) and 120-145 mM and 1-2 mM in most vertebrates and other invertebrates, respectively. Thus, with V_M values on the order of 80 to 60 mV, there is a large, inwardly directed electrochemical gradient for Na⁺ (_Na) and a very large, inwardly directed electrochemical gradient for Ca²⁺ (_Ca). Both ions tend to move into the cell passively and must be extruded by energy-dependent mechanisms. The two electrochemical gradients are defined, respectively, as

(1)

(1A)

and

(1`)

(1A`)

where E_Na and E_Ca are, respectively, the equilibrium potentials for Na⁺ and Ca²⁺, and R, T, and F are the gas constant, absolute temperature and Faraday's number, respectively.

For a transport system in which the flux of one Ca²⁺ is coupled to the counterflow of n Na⁺, the equilibrium relationship between the two electrochemical gradients is

(2)

If the fluxes of one Ca²⁺ and one K⁺ are coupled to the counterflow of n Na⁺, as in vertebrate photoreceptors, the equilibrium relationship is given by

(2A)

where _K is the electrochemical gradient for K⁺, defined as

From the definitions of the electrochemical gradients, Equations 2 and 2A can be expanded to, respectively

(3)

and

(3A)

The coupling ratio (n) is obviously a key parameter in Equation 3. We shall adopt the terminology of Läuger (559) and shall refer to the ratio of the number of transported Na⁺ to the number of transported Ca²⁺ as the "coupling ratio" (n_T). This may be different from the ratio of the number of binding sites for transported Na⁺ and transported Ca²⁺ on the exchanger which, following Läuger (559), we shall refer to as the "stoichiometry" (n_B). Numerous attempts have been made to determine the value of n_T or of the stoichiometry, including several direct measurements of the coupled fluxes. The critical problem in direct measurements of the coupling ratio is to account for all of the exchanger-mediated fluxes of Na⁺ and Ca²⁺ and to exclude all other, parallel movements of these two ions (i.e., those not mediated by the exchanger). Without a specific inhibitor of the Na⁺/Ca²⁺ exchanger, it is difficult to ascertain that all fluxes through other pathways are excluded.

The Hill coefficient corresponding to the cooperativity between external Na⁺ in activating Ca²⁺ efflux was used initially to estimate the coupling between Na⁺ and Ca²⁺ (35, 85, 800). The first direct measurement of the coupling ratio (i.e., measurement of the coupled fluxes of Na⁺ and Ca²⁺) was made in internally dialyzed, ATP-depleted squid axons. The ratio of the Ca_i-dependent Na⁺ influx to the Na_o-dependent Ca²⁺ efflux (i.e., "Ca²⁺ influx mode exchange) was 3.1:1 (107). In internally perfused, ATP-fueled barnacle muscle cells, the ratio of the Ca_o-dependent (Ca_i-activated) Na⁺ efflux to the Na_i-dependent (Ca_i-activated) Ca²⁺ influx, over a wide range of [Na⁺]_i, was 2.8-3.2:1, with an average of 3.1:1, as shown in Figure 6 (754, 757). In these barnacle muscle experiments, the exchanger-mediated fluxes were defined not only by the counter-ion dependence, but also, independently, by the activation by nontransported internal Ca²⁺ (Figs. 5 and 6, and see sect. IIIE); thus ion movements mediated by other transport mechanisms could be excluded.

View larger version (27K): [in this window] [in a new window]   Fig. 6. Effects of internal Na+ (Nai) on Cao-dependent Na+ efflux (; right ordinate scale) and Nai-dependent Ca2+ influx (; left ordinate scale) in internally perfused barnacle muscle cells. External solution in all experiments was Na+-free medium (Na+ replaced by isosmotic Tris) containing 0.1 mM ouabain. Intracellular Ca2+ concentration ([Ca2+]i) was either 0.01 µM (open symbols) or 1.0 µM (solid symbols). Note that ordinate scale for Ca2+ influx is expanded 3-fold, relative to scale for Na+ efflux. Solid line fits Hill equation with a Hill coefficient of 3.0, a dissociation constant for Nai of 30 mM, and maximal fluxes of 20.5 pmol·cm2·s1 for Nai-dependent Ca2+ influx and 61.5 pmol·cm2·s1 for Cao-dependent Na+ efflux. Broken line is best fit to data, with a Hill coefficient of 3.7, a dissociation constant for Nai of 30.0 mM, and a calculated JNa/Ca(Max) of 62.4 pmol·cm2·s1. These data indicate that coupling ratio of Ca2+ influx mode of Na+/Ca2+ exchange is 3 Na+:1 Ca2+. [From Rasgado-Flores et al. (757).]

Measurements of coupled Na⁺ and Ca²⁺ fluxes, from "initial flux rates," in cardiac sarcolemmal vesicles also yielded a coupling ratio of 3 Na⁺:1 Ca²⁺ (101), whereas n was reported to have a value of 4 in dog heart sarcolemmal vesicles (568) and in rat brain plasmalemmal vesicles (47, 48). In cultured rat heart cells, the ratio of the Ca_o-dependent Na⁺ efflux to the Na_i-dependent Ca²⁺ uptake was 3.1:1 (965, 966). Measurements of net Ca_o-dependent Na⁺ efflux and Ca²⁺ uptake in cardiotonic steroid-treated rabbit ventricle also yielded a Na⁺-to-Ca²⁺ flux ratio of 3:1 (131), although this was not based on initial rate data.

These direct flux measurements only address the coupling ratio between Na⁺ and Ca²⁺; they ignore possible coupled (Na⁺/Ca²⁺ exchanger-mediated) movements of other ions such as H⁺, K⁺, or Cl. However, if more than two Na⁺ are exchanged for one Ca²⁺, and the net charge is not counterbalanced by the coupled movements of other ions, then the exchange should be voltage sensitive and should involve net charge movement that ought to be detectable as an exchanger-generated current (i.e., the flux should be rheogenic).

Indirect methods for determining the coupling ratio of the Na⁺/Ca²⁺ exchanger are generally model dependent (290). Several elegant experiments were designed to determine the exchanger coupling ratio on the assumption that the only ions transported by the exchanger are Na⁺ and Ca²⁺ and that the Ca²⁺ is not accompanied by a proton, for example, to compensate for the net charge carried by Na⁺ if more than two Na⁺ are exchanged for one Ca²⁺. Reeves and Hale (762) used a null-point method (based on Eq. 2) to establish the Na⁺ and Ca²⁺ concentration gradient and V_M under conditions in which there was no net exchanger-mediated flux of Ca²⁺ into bovine cardiac sarcolemmal vesicles; their results yielded a coupling ratio of 3:1 (Fig. 7).

View larger version (30K): [in this window] [in a new window]   Fig. 7. Intravesicular Ca2+ concentrations and Na+/Ca2+ coupling ratio (n) determinations in cardiac plasmalemmal vesicles. A: [Ca2+]i graphed as a function of incubation time for membrane potential (VM) = Na+ reversal potential (ENa) = 0 mV () and for VM = 30.71 mV ( and ) with ENa = 0, 4.0, 10.6, 15.8, 19.7, and 21.1 mV. BD: data from A obtained at 1 s (B), 3 s (C), and 5 s (D) are shown as plots of Ca2+ versus ENa. For conditions where Ca2+ reversal potential (ECa) = 0 (i.e., at Ca2+ = 0), coupling ratio (n) can be calculated from the relationship n = 2/(1 ENa/VM), because ENa then exactly compensates for VM so there is no net loss or gain of Ca2+ via Na+/Ca2+ exchanger. This is based on components of the thermodynamic driving force for exchanger (see Eqs. 2 and 6'): Na/Ca = n × Na  Ca = (n  2) VM  nENa + 2ECa. Calculated values for n (based on intersections of regression lines with points of 0 net Ca2+ flux) were 3.10, 3.02, and 2.99. [From Reeves and Hale (762).]

Kimura et al. (514) determined the reversal potential (E_Na/Ca; see Eqs. 5 and 6) for the exchanger-mediated inward and outward currents in single, internally perfused guinea pig cardiac myocytes. They found that E_Na/Ca was very close to the value expected for a coupling ratio of 3 and very different from that predicted for n = 4 (Fig. 8). This result was confirmed and extended by Ehara et al. (285).

View larger version (16K): [in this window] [in a new window]   Fig. 8. I-V curves of Cao-induced INa/Ca in guinea pig cardiac muscle. A: currents induced by [Ca2+]o = 1 and 4 mM. Ni2+ (5 mM) completely blocked INa/Ca; thus curve obtained with 2 mM Ca2+ + 5 mM Ni2+ shows only residual current. B: I-V curves of Ni2+-sensitive (INa/Ca) currents. Reversal potentials (ENa/Ca) for these currents, about 13 mV for [Ca2+]o = 1 mM, and 53 mV for [Ca2+]o = 4 mM, are very close to those expected for an exchanger with a coupling ratio of 3 Na+:1 Ca2+, and very different from values expected for a 4:1 coupling ratio. [From Kimura et al. (515).]

Bridge et al. (132) compared the intergal of the inward, nifedipine-sensitive Ca²⁺ current evoked by depolarization in cardiac myocytes (i.e., net entering Ca²⁺, which should equal the amount of Ca²⁺ that must be extruded during recovery) with the integral of the inward Na⁺/Ca²⁺ exchange current after repolarization (Fig. 9). The integral of Ca²⁺ current (I_Ca) was twice as large as the integral of I_Na/Ca. This, too, is consistent with a 3:1 coupling ratio.

View larger version (14K): [in this window] [in a new window]   Fig. 9. A: relationship between integral of inward Ca2+ current (ICa; abcissa) and integral of inward Na+/Ca2+ exchange current (INa/Ca; ordinate) in guinea pig cardiac myocytes. Data fit regression line that corresponds to an exchanger coupling ratio of n = 3, but not n = 4 or n = 2. B: integral of ICa was obtained by integrating 10 µM nifedipine-sensitive current which was assumed to reflect Ca2+ entry. C: INa/Ca was integrated for 2 s beneath dotted baseline. This included all of transient exchange and was assumed to reflect Ca2+ extrusion. [From Bridge et al. (132).]

Yasui and Kimura (996) (and see Ref. 283) tested the possibility that K⁺ might be transported by the cardiac exchanger, as in the vertebrate photoreceptor (see below). But, again, E_Na/Ca fit a coupling ratio of 3 Na⁺:1 Ca²⁺, irrespective of the external K⁺ concentration.

The exchanger in squid axons also has a reversal potential close to that predicted for a 3 Na⁺:1 Ca²⁺ coupling ratio (249). Furthermore, neither the squid axon exchanger (204, 249) nor the barnacle muscle exchanger (755) requires external K⁺.

Taken together, these numerous measurements of the coupling ratio, using several different methods, all lead to the same conclusion: the cardiac/neuronal type Na⁺/Ca²⁺ exchanger has a coupling ratio of 3 Na⁺:1 Ca²⁺ when operating in either the Ca²⁺ influx or Ca²⁺ efflux mode.

Two groups (420, 546, 997) measured the net charge carried by the Na⁺/Ca²⁺ exchanger in amphibian rod photoreceptor outer segments (ROS) during net extrusion of Ca²⁺ when V_M and the internal and external ion concentrations were varied. In both studies, the charge movement associated with the Na_o-dependent efflux of Ca²⁺ suggested that one of the Na⁺ entered as an uncompensated positive charge. They concluded that Na⁺/Ca²⁺ exchange in ROS was consistent with a fixed coupling ratio of 3 Na⁺:1 Ca²⁺. Schnetkamp et al. (830) (see also Refs. 320, 575, 687, 831), however, observed that Na⁺/Ca²⁺ exchange in ROS was halted in the absence of K⁺, and they concluded that the exchanger in ROS also transports K⁺. Cervetto et al. (165) (see also Ref. 548) then showed that E_Na/Ca in the amphibian photoreceptor ROS is consistent with a coupling ratio of 4 Na⁺ to 1 Ca²⁺ and 1 K⁺ (Fig. 10). With hindsight, this difference in coupling ratio between the exchanger in ROS and that present in many other tissues does not seem surprising. The [Na⁺]_i in vertebrate photoreceptors is much higher than in many other cells, being on the order of 40 mM. This high [Na⁺]_i is due to the "dark current" (inward Na⁺ current observed in the absence of light). Thus there is insufficient energy in ROS _Na for an exchanger with a coupling ratio of 3 Na⁺:1 Ca²⁺ to be able to extrude Ca²⁺ against the large Ca²⁺ gradient. In contrast, invertebrate photoreceptors employ a different type of phototransduction mechanism, and they do not have a dark current or a relatively high [Na⁺]_i. The exchanger in invertebrate photoreceptors therefore does not require K⁺ and appears to be more like the cardiac/neuronal type exchanger, with a coupling ratio of 3 Na⁺:1 Ca²⁺ (22, 225, 657, 659) (and see sect. VII).

View larger version (23K): [in this window] [in a new window]   Fig. 10. Na+/Ca2+ exchanger currents in salamander rod outer segments. A: currents measured during exposure to a solution containing 0.5 mM K+, 1 mM Ca2+, and indicated Na+ concentrations; after 20 s, solution was switched to one containing 110 mM Na+ and 5 mM K+. B: relation between charge flow on return to second solution and [Na+]o during first 20-s period. Intersection of the horizontal broken line and the continuous diagonal line extrapolated from the data indicates that there is no (net) charge transfer at [Na+]o = 62 mM. Arrows show predicted crossing points for exchange stoichiometries of n Na+:1 K+, where n = 3, 4, or 5. C: data from a similar experiment in which [Ca2+]o and [Na+]o were changed simultaneously. In this case, rod outer segment (ROS) was maintained in a solution containing 110 mM Na+, 5 mM Ca2+, and 5 mM K+; [Ca2+]o was reduced to 1 mM, and [Na+]o was changed to concentration shown on abscissa for 100 s. Ordinate shows charge transferred upon return to original solution. Continuous line, which was drawn by eye, crosses dashed line (no net charge transfer) at [Na+]o = 72 mM. Arrows show expected crossing points for exchange stoichiometries of n Na+:1 Ca2+, where n = 3, 4, or 5. D: effect of a simultaneous change in VM and [K+]o on charge transfer by ROS. ROS was maintained in a solution containing 110 mM Na+, 1 mM Ca2+, and 5 mM K+, and [K+]o was reduced to 2 mM for 40 s; depolarizing pulses from a holding potential of 14 mV coincided with these solution changes. Graph shows charge transfer (ordinate) as a function of pulse amplitude (abscissa). Arrow indicates change in VM required to compensate for a 2.5-fold change in [K+]o if one charge exchanges for 1 K+. The only coupling ratio consistent with all of these data is 4 Na+:(1 Ca2+ + 1 K+). [From Cervetto et al. (165).]

Recently, a K⁺-dependent Na⁺/Ca²⁺ exchanger was identified in human platelets (517). Whether this exchanger mediates a 4 Na⁺:(1 Ca²⁺ + 1 K⁺) exchange (i.e., 4 Na⁺ are exchanged for 1 Ca²⁺ + 1 Na⁺), and whether it has sequence homology to the vertebrate photoreceptor exchanger is not yet known. Also, transcripts of a clone that is homologus to the bovine ROS exchanger are expressed in rat brain; the protein encoded by this cDNA is 45% identical to the bovine ROS exchanger (940). The expressed protein mediates K⁺-dependent Na⁺/Ca²⁺ exchange.

As discussed in section IVI, there is little amino acid sequence homology between the vertebrate ROS exchanger and the cardiac/neuronal type Na⁺/Ca²⁺ exchanger. Thus the ROS Na⁺/Ca²⁺ exchanger probably evolved de novo, and it appears that there are two different molecules with very similar functions. This implies that there was evolutionary (and physiological) pressure to develop a transport system of this type (i.e., an Na⁺/Ca²⁺ exchanger, in contrast to an ATP-driven Ca²⁺ pump) to fit the needs of the vertebrate photoreceptor. Detailed structural comparison of the cardiac-type and ROS-type Na⁺/Ca²⁺ exchangers should provide novel information about the nature of the Na⁺ and Ca²⁺ binding sites on these exchangers.

The sum of these varied types of observations, carried out on numerous different tissues and species under a wide variety of conditions, indicates that the Na⁺/Ca²⁺ exchanger in most tissues (but not ROS) 1) likely has a fixed coupling ratio of 3 Na⁺:1 Ca²⁺ and 2) does not normally transport other ions along with the Na⁺ and Ca²⁺. Thus Equation 3 can be rewritten as

(4)

We may then ask how well the equilibrium conditions defined by Equations 3 and 4 fit with experimental observations of Na⁺ and Ca²⁺ distribution and V_M. In other words, if [Ca²⁺]_o, [Na⁺]_o, [Na⁺]_i, and V_M are all known, can Equation 4 be used to predict [Ca²⁺]_i? Alternatively, if [Ca²⁺]_i is known, as well, does the calculated coupling ratio n (Eq. 3) equal 3 (i.e., can we confirm Eq. 4)?

A good illustration of the application of these equations comes from the work of Sheu and Fozzard (846, 847) (see also Ref. 32). When the ion concentration and V_M values obtained from their ion-selective electrode studies of resting (quiescent) cardiac muscle cells were inserted into Equation 3, Sheu and Fozzard (846) calculated a value of n = 2.5, i.e., their measured [Ca²⁺]_i was higher than the value predicted if n is actually equal to 3 (Eq. 4). This means that the Na⁺/Ca²⁺ exchanger was not driving [Ca²⁺]_i down to the value expected from the energy available in _Na; that is, entry of Ca²⁺ via other pathways and relatively slow extrusion via the Na⁺/Ca²⁺ exchanger (because of the low [Ca²⁺]_i; see below) were sufficient to maintain [Ca²⁺]_i slightly above the level predicted from E_Na/Ca. When the Na⁺ pump was inhibited with ouabain, however, causing [Na⁺]_i and [Ca²⁺]_i to rise (presumably thereby saturating more of the exchanger's intracellular binding sites; see next section), insertion of the new, elevated [Na⁺]_i and [Ca²⁺]_i values into Equation 3 fit with a value of n = 3 (Fig. 11). These observations are, nevertheless, consistent with the idea that the Na⁺/Ca²⁺ exchanger has a fixed coupling ratio of 3. When [Na⁺]_i and [Ca²⁺]_i are low, and the exchanger's ion binding sites are unsaturated so that exchanger net transport is low, other mechanisms will play an important, but not exclusive, role in the control of [Ca²⁺]_i. Indeed, the exchanger may even drive Ca²⁺ into the cells, depending on the polarity of the term V_M  E_Na/Ca (see Eq. 5). Alternatively, we must consider the possibility that some of the exchanger molecules may be operating in the Ca²⁺ efflux mode while others are mediating Ca²⁺/Ca²⁺ exchange. Under these circumstances, the equilibrium conditions required for the application of Equation 4 will not prevail. The apparent coupling ratio calculated from Equation 3 will then be lower than the exchanger's true coupling ratio [see Heinz and Geck (398) regarding coupling efficiency]. However, when [Na⁺]_i and [Ca²⁺]_i are high enough to saturate many of the binding sites on the exchanger so that turnover is high, the exchanger may play a dominant role in controlling [Ca²⁺]_i in cardiac myocytes. This leads us into a discussion of the kinetics of Na⁺/Ca²⁺ exchange.

View larger version (17K): [in this window] [in a new window]   Fig. 11. Interrelationship between Na+ and Ca2+ electrochemical gradients (Na and Ca) in cardiac ventricular muscle. Graph shows relationship between electrochemical gradients for Ca2+ [2(ENa  VM)] and Na+ (ENa  VM) when ventricular myocytes are treated with 5 µM ouabain. In all cases, control data (in absence of ouabain) are at top right ends of graph; each symbol represents a different cell. After Na+ pump inhibiton, values for both electrochemical gradients declined progressively, as indicated by each of lines connecting symbols. Broken lines represent predicted coupling ratio (n) relationships between Na and Ca for values of n = 2, 3, and 4. Data best fit relationship for a 3 Na+:1 Ca2+ exchange. [From Sheu and Fozzard (847).]

As mentioned above, Ca²⁺ can be transported both inwardly and outwardly via the Na⁺/Ca²⁺ exchanger. Net Ca²⁺ flux mediated by the exchanger [J_Ca(Na/Ca)] is expressed in units of moles/(area × time). The flux is governed by the difference between V_M and the reversal potential of the exchanger, E_Na/Ca, a thermodynamic term, as well as by the kinetic parameters that control the rate of exchange

(5)

where

(6)

and the equilibrium potentials for Na²⁺ and Ca²⁺, E_Na and E_Ca, respectively, are

(6A)

and

(6B)

Likewise, the net Na⁺ flux mediated by the exchanger [J_Na(Na/Ca)] is given by

(5`)

The variable k_Na/Ca in Equations 5 and 5' is a complex kinetic parameter that is analogous to a conductance density (i.e., a conductance per unit area). This variable depends on 1) the number of carriers, 2) the fractional saturation of the carrier binding sites by the activating ions and transported ions, 3) V_M, 4) ATP, and 5) other factors that may modulate the exchanger including phosphorylation and dephosphorylation (see sect. IIIE).

With a coupling ratio of 3, the reversal potential of the exchanger, E_Na/Ca, is given by (from Eq. 3)

(6`)

Moreover, with this coupling ratio, the Na⁺/Ca²⁺ exchanger mediates the transfer of net charge, and thereby generates an ionic current. Na⁺/Ca²⁺ exchange is therefore a rheogenic (current-generating) process, and its activity is now often measured as a current (e.g., see sect. IIID). Thus it is appropriate to write an equation equivalent to the flux equations (i.e., Eqs. 5 and 5') to describe the net current density mediated by the exchanger (I_Na/Ca)

(7)

I_Na/Ca is the current per unit area (or current density); it corresponds to the net charge transfer that results from the sum of the Ca²⁺ and Na⁺ counterfluxes [i.e., J_Ca(Na/Ca) + J_Na(Na/Ca)]. The kinetic term g_Na/Ca is therefore equivalent to a conductance per unit area (or conductance density). It corresponds to the slope of the relationship between the electrical driving force (V_M  E_Na/Ca) and the current density I_Na/Ca. Like k_Na/Ca in Equations 5 and 5', g_Na/Ca is a nonlinear, complex term, and both are influenced by the same kinetic and regulatory parameters.

In contrast to its role in primary active transport systems such as the PM or SR/ER Ca²⁺ pumps, ATP does not directly fuel the Na⁺/Ca²⁺ exchanger. ATP may, however, provide the energy for the phosphorylation of the exchanger. This phosphorylation apparently alters the kinetics of binding and translocation of the transported ions in cardiac muscle (410, 412) and in squid axons (88, 245), the two tissues that have been studied most extensively. Under normal physiological conditions, in ATP-replete cells, the exchanger molecules appear to be at least partially phosphorylated; this phosphorylated form exhibits a high affinity for external Na⁺ and internal Ca²⁺ (88, 243, 245).

A particularly important and paradoxical finding is the activation of the Ca²⁺ entry mode of exchange by internal Ca²⁺ (243, 244, 246, 757). As illustrated for barnacle muscle (Fig. 6) and cardiac muscle (Fig. 4), there is an absolute requirement for internal Ca²⁺. Indeed, no internal Na⁺-dependent, external Ca²⁺-activated outward current is observed in the absence of internal Ca²⁺. The apparent half-maximal activation of the Na⁺/Ca²⁺ exchanger by internal Ca²⁺ is observed with a [Ca²⁺]_i in the dynamic physiological range [between 0.1 and 1.0 µM, for example, in squid axons (246), barnacle muscle (757), and cardiac muscle (412)]. As we shall see, this property plays a very important role in the physiological activity of the exchanger.

As indicated in section IIIA, in which we briefly described the four modes of operation, the Na⁺/Ca²⁺ exchanger contains several sites for Na⁺ and for Ca²⁺ binding. Some of the bound ions are transported, whereas others serve an activating role and are not transported. To avoid confusion and simplify the ensuing discussion, these Na⁺ and Ca²⁺ binding sites are identified, and some of their properties are indicated in Table 1.

The exchanger's kinetics have been examined in numerous different preparations, under a large variety of conditions (not all of which are published in full detail); nevertheless, not all kinetic parameters for all modes of exchange have yet been characterized. Table 2 contains a summary of the ranges of experimentally determined values for the kinetic "constants" from frequently studied preparations with well-defined "sidedness" (especially squid axons, barnacle muscle, and mammalian heart and nerve preparations). In vesicular preparations, the presence of mixtures of inside-out and right side-out vesicles presents ambiguities of data intepretation in terms of membrane sidedness. Therefore, kinetic data from PM vesicle studies are omitted.

There is no a priori reason to expect that all Na⁺/Ca²⁺ exchangers have identical kinetics; indeed, as we shall see, the exchangers from different species may have some surprisingly different kinetic or mechanistic properties. Furthermore, it is important to note that some of the kinetic parameters are modulated by ATP and membrane potential (244, 246, 254, 631) as well as by a variety of other factors that will be discussed in more detail in section III, D and E. The kinetic constants (ion K_0.5 values = ion concentrations at half-maximal activation or inhibition) listed here are all based on transport studies (measurements of ionic fluxes or ionic currents). These K_0.5 values include the effects of catalytic modulation that prevailed during the measurements. The maximum flux (J_max) values are not listed because they may vary greatly as a result of exchanger modulation. It is beyond the scope of this review to list all of the measured kinetic constants under all conditions that have thusfar been studied; the interested reader should consult the review by DiPolo and Beaugé (254) as well as the original literature cited in the text, tables, and figures.

A) CA²⁺ entry mode exchange. External Ca²⁺ activation of the exchanger functioning in the Ca²⁺ entry mode is hyperbolic. The half-maximum activation (K_0.5 at site C2_o; see Table 1) is observed at about [Ca²⁺]_o of 2-7 mM in marine invertebrates incubated in Na⁺-free media (88, 107). In standard Na⁺-containing media, the apparent affinity for (transported) Ca_o is greatly reduced (K_0.5 = 50-100 mM); this presumably is a reflection of the antagonism between Na_o and Ca_o (e.g., see Fig. 3B). Mammalian heart (631) and brain (314, 809) exhibit a much higher affinity for Ca_o (K_0.5 = 0.1-0.3 mM). Antagonism between Na_o and Ca_o is also observed in these mammalian preparations (Table 2).

As noted above, the Ca²⁺ entry mode of operation of the exchanger is also totally dependent on internal Ca²⁺ (239, 246, 757). No Ca_o-dependent Na⁺ efflux or Na_i-dependent Ca²⁺ influx is observed when [Ca²⁺]_i <10⁸ M (e.g., see Figs. 4 and 6), and the rate of transport therefore depends directly on this activating Ca_i. The K_0.5 at this Ca²⁺ binding site (C1_i) is ~0.6-2 µM internal Ca²⁺ in ATP-fueled cells. This internal Ca²⁺ is not transported by the exchanger, however, because little or no Ca_o-dependent Ca²⁺ efflux (Ca²⁺/Ca²⁺ exchange) is observed when external Na⁺ is replaced by a monovalent cation that is not an alkali metal ion and [Na⁺]_i is high (i.e., under the conditions favoring Ca²⁺ entry mode exchange). The affinity for internal activating Ca²⁺ is reduced by a factor of ~10 in ATP-depleted cells (88, 246); this may indicate that the binding of Ca²⁺ to site C1_i is modulated by phosphorylation. It is important to note that there is no evidence for activation by Ca_o (i.e., there is no "C1_o site" in our notation) analogous to the activation by Ca_i (at site C1_i; see Table 1).

The activation of Ca²⁺ influx by Na_i is a sigmoid function of [Na⁺]_i; the K_0.5 (at site N2_i) is [Na⁺]_i equals 15-25 mM in mammalian cardiac muscle and 30-60 mM in squid axons and barnacle muscle fibers (see Table 2 for references). The apparent affinity for Na_i is increased by the activating Ca_i (239). As shown in Figure 6, the [Na⁺]_i activation curve for Ca²⁺ influx in barnacle muscle has a Hill coefficient of 3. This implies that Ca²⁺ efflux depends on the cooperative action of at least 3 Na⁺, which is consistent with the coupling ratio of 3 Na⁺:1 Ca²⁺. Here, and in the ensuing discussion, we present mean K_0.5 values for Na⁺ that correspond to the half-maximal activation (or inhibition) by the several Na⁺ acting cooperatively, but the situation is probably much more complex, as discussed below. It is not clear whether all three activating Na⁺ bind to site N2_i in this case, while only two Na⁺ bind to the site when internal Na⁺ inhibits Ca²⁺ exit mode exchange (see below). For simplicity (and in the absence of relevant data), we will assume that N1_i binds a single alkali metal ion (see below), that N2_i can bind only two Na⁺, and that exchanger can cycle when both sites are loaded with (a total of three) Na⁺.

Raising [Ca²⁺]_i would be expected to inhibit the Ca²⁺ entry mode of exchange by shifting the exchanger to the Ca²⁺ efflux mode or to Ca²⁺/Ca²⁺ (self) exchange (see below). This effect of Ca_i has not been studied explicitly, however.

Calcium entry mode exchange is activated (at site N1_o) by nontransported external alkali metal ions (57) including Na⁺ (K_0.5 ~50-75 mM in squid axons and crab nerves and ~2 mM in synaptosomes). Monovalent cations that are not alkali metal ions, such as choline and tetramethylammonium (TMA), are ineffective at activating Ca²⁺ entry (35), presumably because they cannot bind to site N1_o. High [Na⁺]_o, however, inhibits the exchange; the K_0.5 (or IC₅₀) for Na_o is 100-200 mM in squid axons and 45-70 mM in mammalian tissue (35, 314, 336). The inhibition is directly proportional to ([Na⁺]_o)², which indicates that two Na⁺ act cooperatively to effect the inhibition (314), presumably at site N2_o. The biphasic effect of external Na⁺, viz. activation by low [Na⁺]_o and inhibition by high [Na⁺]_o, hints at the very complex kinetics. The available data (see Ref. 314) indicate that a single Na⁺ or other alkali metal ion binds to the activating (relatively nonselective) site, N1_o, whereas two Na⁺ interfere with Ca²⁺ binding by, themselves, binding to the second (Na⁺-selective) site, N2_o. The latter site presumably accommodates only two Na⁺. How this site relates to the external site that binds Ca²⁺ (C2_o) is unclear. One possibility is that the same site can bind either two Na⁺ and, perhaps with some conformational rearrangement (to account for noncompetitive inhibition), one Ca²⁺; in other words, sites N2_o and C2_o may simply be slightly different conformations of the same ion binding site. Thus inhibition of Ca²⁺ entry by Na_o may reflect a shift in exchanger mode to Ca²⁺ exit mode or to Na⁺/Na⁺ self-exchange (see below).

B) CA²⁺ exit mode exchange. In ATP-fueled cells, the Ca²⁺ exit mode of exchange is activated by internal Ca²⁺ with a K_0.5 in the range of [Ca²⁺]_i equal to 0.6-6 µM in marine invertebrates as well as in mammals (Table 2). In ATP-depleted squid axons, the K_0.5 is increased to 8-15 µM. Molecular studies indicate that when the Ca_i regulatory site (C1_i) is mutated, the affinity for Ca_i at the transport site (C2_i) is greatly reduced (634, 635). Thus it seems likely that two Ca²⁺ may act cooperatively to promote Ca²⁺ efflux and that one of these Ca²⁺ (at the lower affinity site, C2_i) is transported. The other Ca²⁺ (at the higher affinity site, C1_i) may simply serve in a catalytic role, similar to the role played by internal Ca²⁺ during Ca²⁺ entry mode exchange; in other words, the C1_i sites for Ca²⁺ influx and efflux may be one and the same.

The low affinity of the Na⁺/Ca²⁺ exchanger for Ca_i, relative to the resting [Ca²⁺]_i, which in most cells is on the order of 100 nM, raises an important question about how the exchanger can extrude net Ca²⁺ (see Ref. 722). Indeed, the PM Ca²⁺ pump (PMCA pump) has about a 10-fold higher affinity for Ca_i than does the exchanger (239). Several factors enter into the answer. 1) The exchanger has at least a 10-fold higher turnover rate than the PMCA pump (where turnover relates to the cycling rate of the exchanger molecles). 2) In many tissues (e.g., cardiac muscle), the exchanger has a larger "capacity" or maximum velocity of Ca²⁺ transport than the PMCA pump (where capacity corresponds to the product of the pump or exchanger density per unit membrane area and the turnover rate). 3) In many tissues (e.g., cardiac muscle), the exchanger plays a major role in the extrusion of Ca²⁺ following cell activation and the substantial elevation of [Ca²⁺]_i. Under these circumstances the PMCA pump may be saturated with Ca²⁺, and its capacity to extrude Ca²⁺ rapidly may be limited, while the exchanger may still have a large reserve capacity. One other possibility is that local conditions may maintain [Ca²⁺]_i at a somewhat elevated level in the vicinity of the exchanger molecules; this will be discussed further in section VIA7.

The activation of the exchanger by ATP, and the apparent increase in affinity for Ca²⁺ at site C2_i (and perhaps also at site C1_i) involves a phosphorylation reaction because only MgATP and hydrolyzable analogs of ATP are effective (237, 243, 252, 253). CrATP, which is a potent inhibitor of most protein kinases, blocks the stimulation of the exchanger by MgATP. Nevertheless, CrATP does not affect the maximum velocity of the exchanger that is achieved at saturating [Ca²⁺]_i (252).

Internal alkali metal ions (acting at site N1_i) activate this mode of exchange (88). As is the case for Na_o at the extracellular surface, however, internal Na⁺ has a biphasic effect and, at high concentration, also inhibits (by binding at site N2_i) the Ca²⁺ exit mode of exchange (88). The Na_i apparently interferes, in a noncompetitive manner, with the action of internal Ca²⁺ (107, 357, 776) (but see below) and shifts the exchanger into the Ca²⁺ entry mode or Na⁺/Na⁺ self-exchange mode. Thus raising [Na⁺]_i causes the apparent affinity for internal Ca²⁺ to decrease (107, 138, 776). These interactions between Na_i and Ca_i occur at the transport site (sites N2_i and C2_i). As is true also for the external ion binding sites during Ca²⁺ entry mode exchange (see above), the precise mechanism of the interaction between internal Na⁺ and Ca²⁺ is unclear. Here, too, we may speculate that the two ionic species bind to different conformations (i.e., N2_i and C2_i) of the same site.

In contrast to these interactions between Na⁺ and Ca²⁺ at the intracellular transport site(s), there is evidence that high [Na⁺]_i does not have an inhibitory effect on the exchanger kinetics. Although Ca²⁺ exit mode exchange and Ca²⁺/Ca²⁺ exchange are both completely inhibited at [Na⁺]_i of 140 mM, increasing [Na⁺]_i even to 350 mM continues to increase Ca²⁺ entry mode exchange (251, 253).

The Ca²⁺ exit mode of exchange obviously depends on Na_o; the Na⁺ activation curve is sigmoid with a Hill coefficient of ~3 in most instances. Half-maximal activation occurs with [Na⁺]_o equal to 50-80 mM in ATP-fueled squid axons and with substantially higher [Na⁺]_o (~110-140 mM) in ATP-depleted axons; in other words, internal ATP also increases the affinity of the exchanger for external Na⁺ (88). The nature of the sites to which the external three Na⁺ bind is also unclear. We speculate, as above, that one Na⁺ binds to N1_o and two Na⁺ bind to N2_o; then when all three ions are bound, the exchanger is able to cycle.

The interaction between ATP and Na_i is also noteworthy. At very low (nominal "0") [Na⁺]_i, ATP has little or no activating effect on the Ca²⁺ efflux, whereas at [Na⁺]_i = 40-60 mM, ATP activates the Ca²⁺ efflux (243). Whether ATP activates the Ca²⁺ efflux mode by directly or indirectly increasing the affinity for Ca_i and/or reducing the affinity for Na_i is not yet clear.

As expected from the aforementioned kinetic properties of an exchanger that can operate to either extrude Ca²⁺ or mediate Ca²⁺ entry in exchange for Na⁺, the concentrations of both Na⁺ and Ca²⁺ on both sides of the PM play key roles in determining whether there is net exit or net entry of Ca²⁺. Thus it is not surprising that external Ca²⁺ inhibits the Ca²⁺ efflux mode of exchange because binding of external Ca²⁺ at site C2_o promotes either Ca²⁺ entry mode exchange or Ca²⁺/Ca²⁺ (self) exchange. The K_0.5 (or IC₅₀) for Ca_o (at site C2_o) is not known precisely, but is >1 mM in barnacle muscle fibers (29) and <1 mM in synaptosomes (809).

Before leaving this discussion of the kinetics of Ca²⁺ entry and Ca²⁺ exit mode exchanges, it is important to mention, briefly, the topic of voltage sensitivty, although this will be extensively discussed in section IIID. Both of these net Ca²⁺ transport modes are voltage sensitive and rheogenic. The Ca²⁺ entry mode is associated with an outward current and is augmented by depolarization and attenuated by hyperpolarization (661, 631). Conversely, the Ca²⁺ exit mode of exchange is associated with an inward current and is augmented by hyperpolarization and attenuated by depolarization (514, 515, 631). Furthermore, several of the ion dissociation constants (or K_0.5 values), such as those for Ca_i, are influenced by the membrane potential (631).

C) CA²⁺/ca²⁺ exchange. Ca²⁺/Ca²⁺ exchange obviously requires both external and internal Ca²⁺. This mode of operation of the exchanger also has an absolute requirement for internal as well as external activating alkali metal ions (88). In general, the kinetics of the ion interactions at the two surfaces of the exchanger are comparable to those observed for the net Ca²⁺ transport modes (Table 2). Thus, in squid axons and giant barnacle muscle fibers, the K_0.5 value for Ca_o ranges between 1 and 3 mM in the presence of activating alkali metal ions (88, 107, 108, 686). The K_0.5 value for Ca_o is somewhat lower in mammalian preparations, 0.2-0.8 mM (98; and Blaustein, unpublished data), and is comparable to the K_0.5 for Ca_o in Ca²⁺ influx mode exchange (see above). Both Sr²⁺ and Ba²⁺ can substitute for external Ca²⁺ in activating the exchanger-mediated Ca²⁺ efflux (88, 203).

The K_0.5 values for Ca_i are similar those for Ca²⁺ exit mode exchange, ~2.5 µM in ATP-replete squid axons and 10-15 µM in ATP-depleted axons. These values likely reflect Ca²⁺ binding to site C2_i. Whether or not Ca_i also binds to site C1_i (the regulatory site) is unknown and is difficult to measure. Nevertheless, it seems likely that occupation of site C1_i is required for Ca²⁺/Ca²⁺ exchange because it is necessary for all the other modes of operation of the exchanger.

The kinetics of activation of Ca²⁺/Ca²⁺ exchange by external Li⁺ (Li_o) are similar to those of activation of the Ca²⁺ entry mode of exchange. The K_0.5 value for Li_o is 50-100 mM in squid axons and ~30 mM in synaptosomes (98, 108). The kinetics of activation by internal alkali metal ions have not been studied.

Ca²⁺/Ca²⁺ (self) exchange is electroneutral and involves the transfer of equal numbers of Ca²⁺ in both directions across the PM. Nevertheless, studies on marine invertebrate preparations (but not mammalian preparations; see below) demonstrate that it is voltage sensitive (12, 249, 254, 756; D. W. Hilgemann, personal communication). Depolarization reduces Ca²⁺/Ca²⁺ exchange, whereas hyperpolarization augments it. The mechanism that underlies this voltage sensitivity is not clear, but these observations suggest that at least one of the Ca²⁺ binding or Ca²⁺ translocation steps is rate limiting. In this context, it may be relevant that the K_0.5 for Ca_i is influenced by the V_M (631).

D) NA⁺/na⁺ exchange. The kinetics of Na⁺/Na⁺ exchange have been less extensively studied. In the Na⁺/Na⁺ self-exchange mode, the squid axon exchanger K_0.5 for Na_o is comparable to the K_0.5 for Na_o activation in the Ca²⁺ exit mode, 120-130 mM for Na_o. Although the K_0.5 for Na_i in this (Na⁺/Na⁺ exchange) mode has not been determined, a value on the order of 10-20 mM is expected because the Na_i sites are saturated by 100 M Na⁺ (246). Like all other modes of operation of the exchanger, including the Ca²⁺ influx mode, Na⁺/Na⁺ self-exchange exhibits an absolute requirement for Ca_i (at site C1_i; see Fig. 5), with a K_0.5 of ~10 µM Ca²⁺ in the absence of ATP (244, 246). Also, like the other modes of exchange, Na⁺/Na⁺ exchange is inhibited by Mg_i with a K_0.5 (IC₅₀) of ~1 mM Mg²⁺.

Na⁺/Na⁺ exchange, like Ca²⁺/Ca²⁺ exchange, is electroneutral and presumably involves the transfer of equal numbers of Na⁺ in the two directions across the PM. Unlike Ca²⁺/Ca²⁺ exchange, however, Na⁺/Na⁺ exchange in marine invertebrate preparations is not voltage sensitive (249, 756). The significance of this observation is not yet clear. It may indicate that the Na⁺-loaded form of the exchanger does not bear a net charge. This, however, conflicts with kinetic models of the exchanger operation that are based on data from mammalian cardiac muscle (631) because the data indicate that the Na⁺-bound form of the cardiac Na⁺/Ca²⁺ exchanger is voltage sensitive (631, 693). As discussed in section IID, there may be some surprising species differences.

In sum, the aforementioned kinetic properties of the Na⁺/Ca²⁺ exchanger indicate that the mode of operation of the exchanger depends critically on these kinetic properties and on the prevailing ion concentrations in the cytosol and the extracellular fluid (ECF). Thus, whether the exchanger, at the cytoplasmic surface, loads with Na⁺ or Ca²⁺ depends primarily on [Na⁺]_i, the K_0.5 for Na_i, [Ca²⁺]_i, and the K_0.5 for Ca_i at site C2_i. Likewise, whether the exchanger, at the extracellular surface, loads with Na⁺ or Ca²⁺ depends primarily on [Na⁺]_o, the K_0.5 for Na_o (at sites N1_o and N2_o), [Ca²⁺]_o, and the K_0.5 for Ca_o at site C2_o. These K_0.5 values, as well as the J_max values, depend, in part, on "catalytic" modulation by a variety of factors, including Ca_i and ATP, that will be discussed in section IIIE. In addition, Ca²⁺/Ca²⁺ exchange requires activation by both internal and external alkali metal ions and thus depends on both the alkali metal ion concentrations and the K_0.5 values for these ions at the the two faces of the membrane (sites N1_o and N1_i). Whether the exchanger cycles, in any of its modes of operation, depends critically on activation by Ca_i at site C1_i.

Several possible models of Na⁺/Ca²⁺ exchange can be envisioned [as discusssed by Milanick and Frame (655), we use the term mechanism to refer to the physical model and the term kinetics to refer to the mathematical behavior]. 1) One simple model, analogous to that for the Na⁺ pump (803), is that the exchanger may bind only one species of transported ion at a time. For example, a Ca²⁺ may bind to the exchanger at the cytoplasmic face, be transported to the ECF, and may dissociate before Na⁺ (or another Ca²⁺) is bound (Fig. 12). This is the so-called "consecutive" or "Ping-Pong" mechanism (see Ref. 494). In this case, there is no intermediate conformation of the exchanger in which transported Ca²⁺ and Na⁺ are bound simultaneously. 2) A second possibility, the "simultaneous" exchange mechanism, involves the simultaneous binding of, for example, 3 Na⁺ at the external site and 1 Ca²⁺ at the cytoplasmic face. Then, with all sites occupied, the exchanger could undergo a conformational change so that the Ca²⁺ is discharged to the ECF and the Na⁺ are released to the cytosol. 3) A variant of the consecutive type of mechanism [termed "sequential by Milanick and Frame (655)] involves an unusual binding sequence. For example, cytosolic Ca²⁺ may be bound at the internal face and be translocated to the external face of the molecule by a conformational change, but external Na⁺ (or perhaps another alkali metal ion) may need to be bound to promote the dissociation of the Ca⁺ into the extracellular fluid (655). (We might speculate that this could be the role of the activating alkali metal ion binding; however, such a possibility has not, to our knowledge, been tested.) Thus this sequential mechanism differs from the consecutive mechanism in that at least one conformation of the exchanger has both Na⁺ (i.e., at least 1 Na⁺) and Ca²⁺ bound simultaneously, even though only one species is translocated at a time.

View larger version (20K): [in this window] [in a new window]   Fig. 12. Consecutive kinetic model of Na+/Ca2+ exchanger. Kinetic rate constants are not indicated (see text). In text, we discuss evidence that certain ions and other regulatory factors may modulate exchanger kinetics (maximum flux and/or ion concentration at half-maximal activation) (blue). Figure shows two different modulated states of exchanger. Only extracellular steps of altered (second, modulated state) are shown in figure (for clarity), but intracellular steps are also presumed to be identical before and after modulation of exchanger operation. The only constraint for an altered kinetic state is that modulation, per se, does not change thermodynamic boundary conditions.

An early suggestion was that Na⁺/Ca²⁺ exchange proceeded via a simultaneous countertransport of Na⁺ and Ca²⁺ (88; see also Refs. 580, 581). This view was based on two key observations. First, the kinetics of activation of Ca²⁺ efflux by external Na⁺ in squid axons appeared to be unaffected by the fractional saturation of the internal site with Ca²⁺. Second, because the same exchanger mediates both Na⁺/Ca²⁺ exchange and alkali metal ion-activated Ca²⁺/Ca²⁺ exchange (see above), it appeared that both the internal and external sites must be loaded simultaneously to effect any ion translocation. Otherwise, we might expect a large component of uncoupled (unidirectional) exchanger-mediated transport, which is not observed. However, this rather narrow view did not take into account the possibilities that either unloaded exchanger molecules cannot cycle (a necessary element of a consecutive or sequential mechanism) or that dissociation of one species (e.g., Ca²⁺) requires the binding of the second ion species (Na⁺, in this example, or another Ca²⁺ in the case of Ca²⁺/Ca²⁺ exchange). The latter is the type of sequential mechanism proposed by Milanick and co-workers (653-656) to explain their data from ferret red blood cells.

Subsequently, Läuger (558) discussed kinetic methods to distinguish between simultaneous and consecutive mechanisms of exchange. Khananshvili (488) then applied the principles enunciated by Läuger to a study of Na⁺/Ca²⁺ exchange in cardiac sarcolemmal vesicles. In particular, Khananshvili applied Läuger's "zero-trans" conditions in which Ca²⁺ was absent on one side of the membrane and Na⁺ was absent on the other side. He showed that the J_max-to-Michaelis constant (K_m) ratio did not change over a wide range of [Ca²⁺]_o and [Na⁺]_i values, when J_max and K_m both varied markedly. Accordingly, he concluded that these results fit Läuger's conditions for a consecutive exchange mechanism.

Milanick and co-workers (653-656) examined the kinetics of Na⁺/Ca²⁺ exchange in ferret erythrocytes. They estimated the J_max/K_m directly from Ca²⁺ influx measurements (as a function of [Na⁺]_i) at very low [Ca²⁺]_o (K_m), where the Michaelis-Menten equation reduces to J = (J_max × [Ca²⁺]_o)/K_m. In contrast to Kanashvili's results, these data showed that J was dependent on [Na⁺]_i, which seems inconsistent with a consecutive mechanism. This led to the conclusion that the exchanger mechanism involves Ca²⁺ binding and translocation but that at least one Na⁺ must then bind before the Ca²⁺ dissociates; in other words, a sequential mechanism. It is possible that a model such as this may resolve some of the controversy between consecutive and simultaneous exchange mechanisms. Some of the controversy may, however, be the result of differences in the experimental methods used to determine the transport mechanism. Alternatively, the precise kinetic mechanism may depend on the particular exchanger gene or splice variant (see sect. IV, D and E), as exemplified by the Cl/HCO₃ exchanger, in which one isoform apparently operates by a consecutive mechanism (327, 327a) and another by a simultaneous mechanism (778).

Reeves and Sutko (766) systematically examined the kinetics of Ca²⁺ fluxes in isolated cardiac sarcolemmal vesicles in terms of the interactions between Na⁺ and Ca²⁺. Their data led them to conclude that the exchanger coupling ratio of 3 Na⁺:1 Ca²⁺ could be explained by the presence of two types of cation binding sites: one site to which a single Na⁺ binds, and a second site to which either one Ca²⁺ or one or two Na⁺ bind. Sodium competitively inhibited the binding of Ca²⁺ to the latter site. Unfortunately, the ambiguity about vesicle sidedness leaves some uncertainty about how to compare these results with subsequent studies in which sidedness was better defined.

The Na⁺ and Ca²⁺ interaction in cardiac myocytes was later examined with electrophysiological techniques by Miura and Kimura (661). They observed competitive binding of Na⁺ and Ca²⁺ at the external surface of the exchanger, a result consistent with that of Reeves and Sutko (766). In squid axons and other neuronal preparations, too, the interaction between external Na⁺ and Ca²⁺ has been described as "competitive" (34, 35, 244), although the precise kinetics were not specifically addressed. In contrast, the inhibition of Ca²⁺ efflux by internal Na⁺ in squid axons appeared to be noncompetitive over a large range of [Ca²⁺]_i (138, 776, 777).

The interactions between Na⁺ and Ca²⁺ at the external surface of the plasmalemma were reexamined by Fontana and colleagues (314) in isolated nerve terminals ("synaptosomes"). They studied the apparent affinity for external Ca²⁺ at various (fixed) levels of [Na⁺]_o and the apparent inhibitory constant for external Na⁺ at various (fixed) levels of [Ca²⁺]_o. The results of both types of experiments yielded clear evidence of a noncompetitive type of inhibition of Ca²⁺ influx by external Na⁺, in which two Na⁺ acted cooperatively. The inhibition by Na_o was directly proportional to ([Na⁺]_o)², and Na_o did not affect the affinity for Ca_o, nor did Ca_o affect the affinity for Na_o. We have no ready explanation for the difference between these results and the aforementioned studies which suggest competitive inhibition of Ca²⁺ entry by external Na⁺. This dilemma may require molecular structural data for resolution.

Matsuoka and Hilgemann (631) explored the interaction between Na⁺ and Ca²⁺ at the cytoplasmic surface of cardiac sarcolemmal giant patches. They observed a mixed type of interaction (competitive and noncompetitive) between these two ions.

This diversity of results seems unsatisfactory. However, the likely nature of the binding sites for Na⁺ and Ca²⁺ requires some comment. On theoretical grounds, even though these two ions have virtually identical crystal radii (~1 Å; Ref. 710), the nature of the sites to which they bind are likely to be very different (287, 288, 937). Furthermore, if two Na⁺ act cooperatively to interfere with the binding of one Ca²⁺, it is, nevertheless, unrealistic to expect the two Na⁺ to occupy (i.e., to be accommodated in precisely) the same site conformation as the one Ca²⁺. One possibility is that the sites are relatively "rigid" and that Na⁺ and Ca²⁺ occupy different sites that are mutually exclusive (so that only one or the other site can be occupied). Alternatively, the ion binding site may be able to fluctuate between two slightly different conformations so that it can accomodate, preferentially, either two Na⁺ (in one conformation) or one Ca²⁺ (in the alternative conformation). These structural features might, depending on circumstances, give rise to either competitive or noncompetitive kinetics. Another possibility, mentioned in section IIIC1, is that different Na⁺/Ca²⁺ exchanger gene products or splice variants (see sect. IV, D and E) might exhibit different kinetic behaviors.

The exchanger, like most integral membrane proteins, is asymmetric (also see sect. IIIA). Nevertheless, one could imagine that the exchanger's ion binding sites at the internal and external faces of the PM might be comparable in terms of affinity. Indeed, this appears to be the case for the Na⁺ binding sites under "normal" physiological conditions (see Table 2). In squid axons [reviewed and summarized by DiPolo and Beaugé (254)], the internal Na⁺ activates Ca²⁺ efflux with a K_m(Na) of 40-60 mM, and external Na⁺ activates Ca²⁺ influx with a K_m(Na) of 50-80 mM. Furthermore, the K_i(Na) values for inhibition of Ca²⁺ efflux and influx by, respectively, internal and external Na⁺ are both ~30 mM. Comparable values for these parameters are also obtained in vertebrate tissues. For example, in cardiac myocytes, K_m(Na) on both sides is ~30 mM (717, 661), and K_i(Na) at the external surface is ~44 mM (661). In rat brain synaptosomes, the K_m(Na) values for activation by internal and external Na⁺ are, respectively, ~20 mM (314) and 34 mM (809), and K_i(Na) at the external surface is ~50-60 mM (314).

In contrast, the K_m(Ca) values for transported Ca²⁺ at the two surfaces of the PM are very different. In the squid, K_m(Ca) at the internal surface is ~1 µM in fueled axons but ranges between 2 and 50 mM at the external surface (88, 254). In guinea pig cardiac myocytes, the comparable values are 0.6 µM and 1.4 mM (661). Also, in rat brain synaptosomes, the K_m(Ca) at the external surface is 0.23 mM, while Ca²⁺ is extruded via the exchanger when [Ca²⁺]_i is <0.5 µM (314). Thus the apparent affinities for Ca²⁺ at the inward and outward facing Ca²⁺ binding sites differ by a factor of ~10³.

An important, unresolved question is, How can we explain this asymmetrical binding of transported Ca²⁺ at the two sides of the PM if we assume that there is only a single binding site for transported Ca²⁺? One possibility is that the exchanger undergoes a spontaneous conformational change that alters the affinity of the binding site when the Ca²⁺ is translocated across the membrane.

A) MONOVALENT CATIONS: LI, CHOLINE, TMA, AND K⁺. Another interesting feature of the Na⁺/Ca²⁺ exchanger is the action of small monovalent cations. The initial experiments that were carried out about 20-30 years ago have raised questions about transport kinetics that are still unexplained. The basic observations focus on the self-exchange modes of the Na⁺/Ca²⁺ exchanger, i.e., Ca²⁺/Ca²⁺ exchange or Na⁺/Na⁺ exchange. The experiments on Ca²⁺/Ca²⁺ exchange are carried out in the absence of Na⁺ to block net Na⁺/Ca²⁺ transport via the exchanger. In the absence of Na⁺, Ca²⁺ isotope can be transported across the membrane in which the Na⁺/Ca²⁺ exchanger resides even though there is no net Ca²⁺ transported. Thus the unidirectional Ca²⁺ flux reflects the components of the exchanger transport reaction that involve Ca²⁺ binding to the Na⁺/Ca²⁺ exchanger on one side of the membrane, Ca²⁺ translocation by the Na⁺/Ca²⁺ exchanger across the membrane, and Ca²⁺ unbinding (dissociation) from the Na⁺/Ca²⁺ exchanger on the other side of the membrane. The unidirectional flux of Ca²⁺ can be followed because ⁴⁵Ca²⁺ is added to one side of the membrane only; one can then follow the appearance of ⁴⁵Ca²⁺ on the other side of the membrane. The model in Figure 12 illustrates the elements of the Na⁺/Ca²⁺ exchanger transport reactions noted here. To the extent that monovalent cations can alter this isotopic flux of Ca²⁺, as noted above, we have ignored this property in the model of the transport reactions shown in Figure 12. Although it is easy to change our model, the problems presented by the experiments that illustrate the action of nontransported monovalent cations described below indicate that it is not yet clear how they relate to the other data on ion binding and translocation.

B) CA²⁺/CA²⁺ exchange and the action of CIS- AND TRANS-MONOVALENT CATIONS THAT ARE NOT TRANSPORTED. In squid axons, ⁴⁵Ca²⁺ efflux was at the "background" level when all of the extracellular Na⁺ was replaced by choline (88). When the extracellular Na⁺ was replaced by Li⁺, the ⁴⁵Ca²⁺ efflux was greatly increased, an effect that was also observed in barnacle muscle (686). Neither Li⁺ nor choline⁺ is transported by the Na⁺/Ca²⁺ exchanger (35, 88, 251, 422, 423). The problem presented by these results is, How does the identity of the nontransported monovalent cation (Li⁺ or choline⁺) act to alter Ca²⁺/Ca²⁺ exchange? At which step in the hypothesized reaction sequence do these nontransported monovalent cations act? They work from the trans-side, i.e., the side opposite to the side from which the Ca²⁺ is transported. Thus the effects described above focus on ⁴⁵Ca²⁺ efflux and the effects that extracellular Li⁺ and extracellular choline⁺ have. In addition, these ⁴⁵Ca²⁺ efflux experiments demonstrate the requirement for nontransported monovalent cation on the inside or cis-side of the membrane (i.e., cis with respect to the ⁴⁵Ca²⁺). The aforementioned experiments were carried out with intracellular K⁺ as the cis-monovalent cation. When the cis-monovalent cation was changed to Li⁺, the ⁴⁵Ca²⁺ efflux was significantly reduced (by ~30%). When the cis-monovalent cation was changed to TMA (a nonalkali metal cation), the ⁴⁵Ca²⁺ efflux was reduced by ~80% (compared with K⁺ as the cis-monovalent cation).

Clearly, our understanding of the actions of the nontransported cis- and trans-monovalent cations is limited. Knowledge of the mechanism(s) of action of these monovalent ions would undoubtedly help us to understand how the Na⁺/Ca²⁺ exchanger works at the molecular level. The problem is even more complicated, however. Effects of the internal monovalent cations were also observed when the Na⁺/Ca²⁺ exchanger operated in either the Ca²⁺ influx or Ca²⁺ efflux mode, but the effects were qualitatively quite different (88). When cis (internal) Li⁺ was present, the Na_o-dependent Ca²⁺ efflux was largest, being about twice that observed when cis-K⁺ was present, and almost four times that observed when cis-TMA⁺ was present. Recent experiments, although not as complete as the earlier results (88), are consistent with these basic observations insofar as the results overlap (99, 314, 335). We are thus left with a modeling dilemma, How can one explain the actions of the cis- and trans-monovalent cations that are not transported? One model of such interactions was given by Doering and Lederer (267), who discussed the way nontransported protons could inhibit the Na⁺/Ca²⁺ exchanger (see sect. IIIE4). In this model, intracellular protons inhibit transport by altering the exchanger kinetics: intracellular protons were hypothesized to bind to the Na⁺/Ca²⁺ exchanger only when Na⁺ are already bound. Then the Na⁺ cannot unbind until the protons first dissociate. Thus intracellular protons are believed to "trap" the exchanger in a largely nonfunctional conformation.

Although all rheogenic (current generating) transport mechanisms must be sensitive to voltage, not all voltage-sensitive processes must be rheogenic or electrogenic (i.e., many nonelectrogenic transport processes are voltage sensitive). Many early reports did not seem to take this simple notion into account and thus tended to infer rheogenic (or electrogenic) properties of the Na⁺/Ca²⁺ exchanger from the observed voltage-dependent properties [reviewed in Eisner and Lederer (290)]. From the beginning, two experimental issues have continued to confound investigations of both the rheogenic and the voltage-sensitive properties of the Na⁺/Ca²⁺ exchanger. The first problem, which persists even today, is the absence of a specific high-affinity blocker of Na⁺/Ca²⁺ exchanger function. This means that simple voltage-dependent difference currents (sensitive to this as yet unidentified blocker) cannot be used to describe, unambiguously, current or flux through the Na⁺/Ca²⁺ exchanger in any of its modes of operations (i.e., Ca²⁺/Ca²⁺ or Na⁺/Na⁺ self-exchange, or Ca²⁺ exit mode or entry mode Na⁺/Ca²⁺ exchange). The second problem is the confusing and multifaceted overlapping Na⁺ and Ca²⁺ fluxes in biological tissues. For example, changes in [Ca²⁺]_i can affect not only the Na⁺/Ca²⁺ exchanger, but also Ca²⁺ channels, Ca²⁺-activated channels (including channels that conduct Ca²⁺), and enzymes that modify channels and transporters (including the Na⁺/Ca²⁺ exchanger). Sodium also has diverse direct and indirect effects on transporters, ion channels, enzymes and, of course, several effects on the Na⁺/Ca²⁺ exchanger.

The Na⁺/Ca²⁺ exchanger was first identified in tracer flux experiments on squid giant axons (35, 102) and guinea pig and sheep heart muscle (783). In those experiments, isotopes of Ca²⁺ and Na⁺ were used to monitor the unidirectional movement of these ions (Figs. 1-3). By measuring the Na_o-dependent Ca²⁺ efflux and the Ca_i-dependent Na⁺ influx, Baker et al. (35) estimated that the coupling ratio of the transport was about 3 Na⁺ to 1 Ca²⁺. This early estimate of the coupling ratio was uncertain, in part because there was considerable scatter in the data [see sect. IIIB2 and the review by Eisner and Lederer (290), regarding the determination of the coupling ratio]. Nevertheless, these early experiments provided the first suggestion that the exchanger might be rheogenic, and further determinations of the coupling ratio (see sect. IIIB2) helped to substantiate this view.

The first clear measurement of the Na⁺/Ca²⁺ exchanger current was made by Kimura et al. (515). These experiments (Figs. 4 and 13), carried out in guinea pig cardiac myocytes, provided evidence that the Na⁺/Ca²⁺ exchanger actually produced measurable membrane current. Although such results were predicted by a 3:1 stoichiometry (35) or a 4:1 stoichiometry (666, 669), they would not be consistent with a 2:1 coupling ratio (783). The measured Na⁺/Ca²⁺ exchanger current thus helped to resolve an important issue and provided direct evidence that the Na⁺/Ca²⁺ exchanger coupling ratio was >2:1 (see also Refs. 844, 846, 847). The data of Kimura et al. (515) were also consistent with the Ca_i-activated inward current in cardiac muscle that had been attributed to the Na⁺/Ca²⁺ exchanger (189, 193, 436, 438, 439, 594, 645, 650). The presence of Ca_i-activated inward current in heart cells due to a Ca_i-activated cation channel (146, 286, 285, 477, 691), however, complicated many experiments (292, 294, 566, 594). Independent measurements on the cardiac Na⁺/Ca²⁺ exchanger, using an equilibrium (null point) method (Fig. 7) (762), however, provided compelling evidence for a 3:1 coupling ratio (290, 844), as noted above (also see Figs. 6, 8, 9, and 11).

View larger version (18K): [in this window] [in a new window]   Fig. 13. Currents generated by rheogenic Na+/Ca2+ exchanger. AC: outward INa/Ca activated by increases in extracellular Ca2+ in presence of 30 mM Nai (C; ramp currents e minus c or d) but not in its absence (B; ramp current b minus a). [From Kimura et al. (515).] D and E: inward INa/Ca activated by step increases in [Ca2+]i (with flash photolysis of caged Cai) at [Na+]o = 0, 72, and 145 mM (D). Na+ concentration at half-maximal activation for Nao is 73 mM, and Hill coefficient is 2.8. [From Niggli and Lederer (695).]

Quantitative measurements of the current carried by the Na⁺/Ca²⁺ exchanger became more manageable with the development of the "giant patch" technique (Fig. 14) (199, 405, 406). This technique enabled investigators to measure the Na⁺/Ca²⁺ exchanger current (and the current from other rheogenic or electrogenic transporters) in many cell types including heart cells, transfected cells, and frog oocytes. Because the maximum turnover rate of the Na⁺/Ca²⁺ exchanger is much lower than that of single channels (by a factor of ~10³), the electrical signal (i.e., the current generated by the exchanger) is much smaller and, thus, very difficult to measure with standard patch-clamp techniques. The giant patch has a surface area >100 times greater than that of the standard patch but retains the giga-ohm seal. Thus it is the increase in surface area in the giant patch that provides the increased signal needed to measure Na⁺/Ca²⁺ exchanger current with a patch clamp. The giant patch-clamp technique has enabled electrophysiological investigations of native, cloned, mutated, and chimeric Na⁺/Ca²⁺ exchanger proteins. It is now the preferred method for many types of studies of Na⁺/Ca²⁺ exchanger electrophysiology. Electroneutral fluxes of the kind associated with Na⁺/Ca²⁺ exchanger "self-exchange" (i.e., Ca²⁺/Ca²⁺ exchange or Na⁺/Na⁺ exchange) cannot, however, be measured with this method. Such self-exchange reactions represent reversible partial reactions of the Na⁺/Ca²⁺ exchanger protein; they can be observed in the absence of the counter-ion (i.e., Ca²⁺/Ca²⁺ exchange can be measured as the unidirectional isotopic Ca²⁺ flux that is observed in the absence of intracellular and extracellular Na⁺). Without additional criteria, however (such as activation by internal Ca²⁺, or inhibition by cis-Na⁺ on the side from which the Ca²⁺ is flowing or, even better, by an as yet unidentified specific inhibitor), one cannot be certain that all of the unidirectional Ca²⁺ flux is mediated by the exchanger (see below).

View larger version (13K): [in this window] [in a new window]   Fig. 14. Modulation of Na+/Ca2+ exchanger current measured with Hilgemann "giant" patch technique. A: activation of outward Na+/Ca2+ exchanger current by high [Na+]i also reveals inactivation, i.e., current rapidly rises to a peak and then, over course of about a minute, gradually declines (inactivates) to a new steady level. This current largely depends on regulatory [Ca2+]i, but kinetics of [Ca2+]i effects are slow, i.e., current falls and rises over tens of seconds when Cai is removed or added back. B: removing intracellular loop removes both Na+ inactivation and Ca2+ regulation. C: removing only a portion of intracellular loop removes Ca2+ regulation but does not alter Na+ inactivation. [From Matsuoka et al. (635).]

The voltage dependence of a current or ion flux through a channel or a transporter represents a "kinetic" property of the protein. Traditionally, one expects the flux to be predicted by the Nernst-Planck current equation [see Hille (416)] which describes the voltage dependence of current through a resistive pore with assumptions of a constant electrical field. Such assumptions are, however, frequently inadequate to describe the current flow through single channels. Electrogenic or rheogenic transporters do not even have such a simple model, and it is therefore useful to discuss some simple "expectations" before reviewing the experimental findings.

A) TRAPPING OF THE EXCHANGER. At extremes of voltage, electrically charged exchangers and pumps will tend to be trapped in specific conformations as a consequence of energetic constraints. Suppose the "naked" exchanger (i.e., with neither Na⁺ nor Ca²⁺ bound) carries an average net charge of 2.5 charges (693, 695). Under these circumstances, and with the assumption of a consecutive transport mechanism (see sect. IIIC1 and Fig. 12), as suggested by electrical evidence (411, 414, 693, 694) and flux evidence (488), the conformation of the Na⁺/Ca²⁺ exchanger ("X"; see Fig. 12) with 3 Na⁺ bound, Na₃X, would have a net charge of +0.5, and the conformation with 1 Ca²⁺ bound, CaX, would have a net charge of 0.5. An assumption frequently made for consecutive transport schemes (see Fig. 12) is that only the fully occupied conformations cross the membrane; thus a charged conformation must cross the voltage field whenever Ca²⁺ or 3 Na⁺ are transported by the Na⁺/Ca²⁺ exchanger (and see Ref. 631). Although there is quantitative uncertainty about the apparent charge associated with each of these hypothesized entities, CaX and Na₃X, it is believed that they have a net negative charge and a net positive charge, respectively (Hilgemann, personal communication). With these assumptions about the charge on CaX and Na₃X, it is useful to discuss the voltage-dependent self-exchange data.

B) NA⁺/NA⁺ EXCHANGE. At extremely positive potentials, the positively charged 3NaX⁺ conformation should be trapped in the extracellular conformation because, for energetic reasons, it cannot cross the voltage field. At extremely negative potentials, it should be trapped in the intracellular conformation because, again, it cannot cross the voltage field for energetic reasons. Thus, with the consideration of Na⁺/Na⁺ exchange only, the kinetics of isotopic transport should increase from zero at the two extremes of potential (see Fig. 15). The shape of the curve describing the voltage dependence of unidirectional Na⁺/Na⁺ exchange transport of Na⁺ should presumably look like a trapezoid. The curve should rise from zero at very negative potentials, display a (possibly sloped) plateau region in the mid-voltage range, and then decrease to zero at positive potentials. The fact that Na⁺/Na⁺ exchange in marine invertebrate tissues is voltage insensitive (249, 756) raises the possibility that the Na⁺-loaded exchanger in these preparations does not bear a net charge and that these considerations therefore do not apply here. This is obviously different from the vertebrate cardiac exchanger, in which the Na⁺-bound form of the exchanger clearly is voltage sensitive (414, 693).

View larger version (16K): [in this window] [in a new window]   Fig. 15. Theoretical effects of charged forms of Na+/Ca2+ exchanger on voltage sensitivity of Na+/Ca2+ exchanger: exchanger "trapping." A: schematic diagram of Na+/Ca2+ exchanger that transports n Na+ in exchange for 1 Ca2+. Rate constants are noted af, x, and y. Rate constants for translocation across membrane (g, h, j, k) are described by a simple Eyring barrier. For simplicity in this model, the "naked" exchanger (i.e., without translocated ions bound) has a net charge of 1, although more recent data suggest that this number is larger (see text). B: net transport of Ca2+ is plotted as a function of membrane potential for an exchanger where n = 3. Zero-flux potential is given by ENa/Ca = 3ENa  2ECa. A small depolarization from ENa/Ca must produce net Ca2+ entry and net Na+ efflux just as a small hyperpolarization from ENa/Ca must produce net Ca2+ efflux and net Na+ influx. At potentials more distant from ENa/Ca, magnitude and shape of current-flux curve will depend on kinetics of transporter and not on its thermodynamics. In this model, depolarization from ENa/Ca will increase rate constants h and k and decrease g and j. At sufficiently positive potentials, j and g will become very small and rate limiting, and this leads to regions vii and viii of curve. We attribute this region of curve to transporter trapping. Similarly at sufficiently negative potentials h and k will become very small and rate limiting and explain regions ii and i of curve, also due to transporter trapping. [From Eisner and Lederer (290).]

C) CA²⁺/CA²⁺ exchange. Similar considerations should apply to the negatively charged CaX conformation. At extremely positive potentials, the CaX conformation should be trapped in the intracellular conformation, and at extremely negative potentials, CaX should be trapped in the extracellular conformation because, in either situation, it cannot cross the voltage field for energetic reasons. Thus the curve that describes the voltage dependence of unidirectional Ca²⁺/Ca²⁺ exchange transport of Ca²⁺ should presumably also look like a trapezoid (as for Na⁺/Na⁺ exchange). It should rise from zero at very negative potentials, display a flat region, and then decrease at more positive potentials (see Fig. 15).

D) SHAPE OF THE "FLAT REGION" OF THE CURVES THAT ILLUSTRATE THE VOLTAGE DEPENDENCE OF SELF-EXCHANGE. There is no a priori reason to favor any particular shape for the region of the curve that describes the voltage dependence of Ca²⁺/Ca²⁺ or Na⁺/Na⁺ exchange. If the various conformations of the Na⁺/Ca²⁺ exchanger protein do not change with voltage over a range of potentials, then the transport would be voltage independent in this voltage range. For the purposes of this discussion, however, we will assume that the CaX conformation carries a net negative charge; the electrical field will then tend to favor the location of CaX at one side of the membrane or the other. Under these circumstances, the exchanger-mediated current will exhibit a maximal plateau region in the mid-voltage range and will slope downward at the negative and positive ends of the curve. Nevertheless, parts of the protein other than the ion binding sites may also carry net charge. If these portions of the protein are located in the voltage field, or if they move into the field, they may thereby be influenced by this field. Should these regions affect the kinetics of ion translocation, they may influence the voltage dependence of self-exchange.

E) VOLTAGE-DEPENDENT ION BINDING. In the simple models described here, we assumed that binding of transported cations to the Na⁺/Ca²⁺ exchanger occurs outside the voltage field. Under these circumstances, no voltage dependence of the ion binding steps should affect the kinetics of self-exchange. Should binding occur within the voltage field, however, the availability of ions will be sensitive to voltage. This, too, should impart some voltage dependence to the transport (337, 559, 748). If a cation binds to the unloaded or empty (X) conformation of the Na⁺/Ca²⁺ exchanger within the voltage field on the intracellular side, cation binding should be favored as the potential becomes more positive. Conversely, if a cation binds to the X conformation within the voltage field on the extracellular side, cation binding would be favored as the potential becomes more negative.

F) A PROBLEM FOR STUDIES OF SELF-EXCHANGE (NA⁺/NA⁺ OR CA²⁺/CA²⁺ EXCHANGE). Because self-exchange produces no net transport, it is assumed that the unidirectional Na⁺/Na⁺ exchange fluxes (i.e., Na⁺ influx and Na⁺ efflux) are identical for any set of ionic conditions. From a theoretical point of view this presents no problem. From an experimental perspective, however, one never knows for sure whether the unidirectional flux that is measured in an experiment is contaminated by Na⁺ flux from a source other than the Na⁺/Ca²⁺ exchanger operating in the Na⁺/Na⁺ (self) exchange mode. Influx and efflux are rarely measured under identical conditions, and there are no good inhibitors of the Na⁺/Ca²⁺ exchanger. Thus uncertainty about whether the two unidirectional fluxes are truly equal in magnitude continues to cloud the interpretation of many fine experiments.

G) MEASURED VOLTAGE DEPENDENCE OF CA²⁺/ca²⁺ exchange. Measurements of Ca_o-dependent Ca²⁺ efflux have been used to examine the voltage dependence of Ca²⁺/Ca²⁺ exchange. As an electroneutral transport process, voltage-clamp and patch-clamp methods offer little, except under special circumstances (see Refs. 693, 694). Thus virtually all of the relevant work has been carried out using tracer flux measurements. In squid axons (249), Ca_o-dependent Ca²⁺ efflux decreased with depolarization; the decline with voltage was linear over the range of potentials from 40 to +40 mV. The Ca²⁺ efflux was compared with the baseline efflux at 0 mV. This appearance of a voltage-dependent component of Ca²⁺ efflux that decreases with depolarization might be consistent with increased trapping, in the extracellular conformation, of the negatively charged Ca²⁺-bound form of the Na⁺/Ca²⁺ exchanger (CaX) at positive potentials. But, as noted above, other explanations are also available. At the very least, we can conclude that there is a charge on the protein that slows the kinetics of Ca²⁺ transport out of the cell as the membrane potential gets more positive. This voltage dependence is consistent with experimental observations in guinea pig cardiac ventricular myocytes (693) and in barnacle muscle (756).

It is interesting that the voltage dependence of the Na_o-dependent Ca²⁺ efflux in squid axon was similar to that of the Ca_o-dependent Ca²⁺ efflux (i.e., the Ca²⁺ efflux declined with depolarization; Ref. 249). The mechanism that underlies this result is incompletely understood. Depolarization might increase trapping of CaX in the intracellular conformation, increase trapping of Na₃X in the extracellular conformation, and/or reduce binding of extracellular Na⁺ (if there is a high resistance access channel; Refs. 337, 559, 748).

H) MEASURED VOLTAGE DEPENDENCE OF NA⁺/na⁺ exchange. According to the "consensus models" of Na⁺/Ca²⁺ exchanger function, the exchanger molecules with 3 Na⁺ bound (i.e., Na₃X) bear a net positive charge. Thus trapping of the Na₃X conformation would be expected at extremely negative and extremely positive potentials. Over a potential range from 60 to +40 mV, however, no change in the Na_o dependence of Na⁺ efflux was observed in either squid axons (249) or barnacle muscle fibers (756). This could simply mean that the voltage range in which trapping or voltage-dependent binding occurs was not tested in these experiments or that the voltage-dependent properties of the invertebrate and vertebrate exchangers are different, or that some other aspect of the Na_o-dependent Na⁺ efflux is rate limiting. At first consideration, these results do not seem consistent with the "E8" model (a model with 8 kinetically defined states) of Matsuoka and Hilgemann (631) or other transport models that attribute at least some of the voltage dependence of the Na⁺/Ca²⁺ exchanger to a voltage-dependent Na⁺ translocation step (see, for example, Refs. 414, 693). It would therefore be helpful to understand why this Na⁺/Na⁺ exchange, apparently mediated by the Na⁺/Ca²⁺ exchanger, is voltage insensitive, when we would expect to find a voltage-sensitive step associated with either Na⁺ translocation or Na⁺ binding.

I) ACTION OF -CHYMOTRYPSIN ON SELF-EXCHANGE MEDIATED BY THE NA⁺/ca²⁺ exchanger. Hilgemann and co-workers (634, 635) used the giant patch technique to examine the kinetics of the current carried by the Na⁺/Ca²⁺ exchanger (I_Na/Ca) in patches from mammalian cardiac myocytes (Fig. 14). They observed that I_Na/Ca was modulated by [Na⁺]_i and [ATP]_i and that this modulation could be removed by exposing the intracellular aspect of the giant patch to a low concentration of the proteolytic enzyme -chymotrypsin. This mild proteolytic treatment did not appear to alter the transport capabilities of the Na⁺/Ca²⁺ exchanger but did appear to "deregulate" this protein (406, 723). It is noteworthy that removal of most of the large intracellular loop of the Na⁺/Ca²⁺ exchanger by molecular biological techniques produced a similarly deregulated Na⁺/Ca²⁺ exchanger (634, 635). Therefore, it seems logical to suggest that -chymotrypsin may act by chemically removing the intracellular loop (or a critical portion of this loop) as well. With the assumption that this explanation is correct, however, it is not entirely clear why the remaining halves of the Na⁺/Ca²⁺ exchanger protein continue to function (perhaps the 2 halves are so well attracted to one another, and so well fitted together that they do not drift apart). Thus we are left with an enigma about the actual tertiary structure of the Na⁺/Ca²⁺ exchanger protein.

The self-exchange modes of operation of the Na⁺/Ca²⁺ exchanger have also been examined after treament with -chymotrypsin. Rasgado-Flores and colleagues (756, 925) reported that Na⁺/Na⁺ exchange was unaltered by this protease but that the voltage dependence of Ca²⁺/Ca²⁺ exchange was dramatically changed in barnacle muscle fibers. Instead of declining with depolarization (i.e., at more positive internal potentials), as observed in the mammalian heart (see above), following -chymotrypsin exposure, the Ca_o-dependent Ca²⁺ efflux in barnacle muscle increased with depolarization. This finding suggests that -chymotrypsin alters a part of the protein that controls/influences a voltage-dependent process. Because there is no evidence that this treatment alters any thermodynamic process, and because Ca²⁺/Ca²⁺ exchange is electroneutral, the results imply that Ca²⁺ binding or translocation step kinetics are altered. The questions that this raises are as follows: Where is this part of the Na⁺/Ca²⁺ exchanger protein? What does this region of the exchanger normally do? Perhaps comparison of the molecular structures of the vertebrate and invertebrate exchangers will give us some clues.

J) QUESTION FOR FUTURE STUDY. Examination of the Na⁺/Ca²⁺ exchanger by examining self-exchange has raised an exciting new question. How is it possible for Na⁺/Na⁺ self-exchange to be voltage independent when it has been argued that this same membrane-crossing transition by Na₃X is one of the most significant voltage-sensitive transport reactions of the Na⁺/Ca²⁺ exchanger? One possibility is that the Na⁺-dependent transition is not voltage dependent. Should this explanation be correct, we would need to reexamine our current models of Na⁺/Ca²⁺ exchanger behavior. Another possibility is that, during Na⁺/Na⁺ exchange, the Na⁺ translocation step is no longer rate limiting (as may occur, for example, if the Na⁺ binding/unbinding step becomes rate limiting and takes place outside the electric field). In any event, it will be interesting and instructive to elucidate the molecular basis for the observed phylogenetic differences in the voltage sensitivity of the Na⁺- and Ca²⁺-bound forms of the Na⁺/Ca²⁺ exchanger.

In an early study, Baker and Glitsch (36) provided the first evidence that the exchanger might be activated by ATP. This was further documented by DiPolo (236-238) who, in a series of elegant experiments on dialyzed squid axons, demonstrated that ATP and analogs with a hydrolyzable terminal phosphate promoted Na_o-dependent Ca²⁺ efflux. This ATP-dependent mechanism was subsequently shown to be mediated by a Ca_i-dependent phosphorylation (245). This activation by ATP appears to involve an increase in the affinity for internal Ca²⁺ and external Na⁺ (88) and a decrease in inhibition by internal Na⁺ (238); indeed, in nominally Na_i-free conditions, ATP does not enhance exchanger-mediated Ca²⁺ efflux. CrATP, which serves as a "substrate" for most types of kinases, and inhibits them, blocks the activation of Na⁺/Ca²⁺ exchange by MgATP; CrATP does not, however, interfere with the exchange activity observed in the absence of ATP (252). Recently, DiPolo et al. (259) reported that a novel (but as yet, unsequenced) 13-kDa peptide is required for the modulation of the squid axon Na⁺/Ca²⁺ exchanger by MgATP. Whether this ATP-dependent modulation of the exchanger is mediated by PKC or protein kinase A (PKA), or is affected by a change in phosphatase activity (see sect. III, E2 and E3), is unknown. Microinjection of exogenous alkaline phosphatase into squid axons, however, reverses the activation of the exchanger by ATP (257).

Phosphoarginine (Pa), but not phosphocreatine, was also found to activate Na⁺/Ca²⁺ exchange in squid axons by a mechanism that depends on intracellular Mg²⁺ and Ca²⁺ (256). The activation by Pa is mediated by a different mechanism than that of ATP (257). Activation by Pa does not affect the affinities for the transported ions. Phosphoarginine promotes Na²⁺/Ca²⁺ exchange, but not Ca²⁺/Ca²⁺ exchange. The action of Pa is not blocked by CrATP, but it is reversed by microinjected alkaline phosphatase; thus the activation by ATP and Pa both appear to involve phosphorylation-dephosphorylation reactions (257).

Using giant membrane patches from cardiac myocytes, Hilgemann et al. (413) showed that the exchanger is inactivated by intracellular Na⁺ (Fig. 14A). This inactivation could be relieved by adding MgATP to the solution bathing the cytoplasmic face of the membrane. Indeed, this may be the explanation for DiPolo's (238) observation that that ATP has no effect on Ca²⁺ efflux in Na_i-free media.

Despite the evidence that some type of phosphorylation is required for activation by ATP, it should be clear that, in contrast to the PMCA pump, the exchanger does not use energy directly from the hydrolysis of ATP to transport Ca²⁺ (i.e., a phosphorylation is not required for each cycle of the exchanger). Indeed, the Na⁺/Ca²⁺ exchanger can operate in the virtual absence of ATP (88, 238, 686). Recent observations have now begun to clarify the mechanism of this regulation.

Several investigators (198, 410, 412, 608, 957) have shown that acidic lipids, including phosphatidylserine and some inositol phosphates [phosphatidylinositol 4-phosphate and phosphatidylinositol 4,5-bisphosphate (PIP₂)], promote Na⁺/Ca²⁺ exchange. Furthermore, Mene et al. (648) suggested that activation of phospholipase C by vasoconstrictors stimulates Na⁺/Ca²⁺ exchange (presumably via the phosphoinositide cascade) in human mesangial cells. Recently, Hilgemann and Ball (409) examined the effects of some of these lipids in greater detail in cardiac myocyte patches. They showed that the effect of ATP was mediated by PIP₂ and was inhibited by a PIP₂-specific phospholipase C (PLC); high [Ca²⁺]_i reversed the effect of ATP, perhaps by activating an endogenous Ca²⁺-dependent PLC. Thus a high [Ca²⁺]_i might tend to inhibit Ca²⁺ extrusion by promoting inactivation of the exchanger (a positive feedback effect on [Ca²⁺]_i). These results illustrate the complexity of regulation of the exchanger by phospholipids and Ca²⁺.

The effects of phorbol esters (which activate PKC) and cyclic nucleotides on Na⁺/Ca²⁺ exchange vary from tissue to tissue, and there are some conflicting results even in similar tissues. For example, the exchanger in cultured rat aortic myocytes was augmented by brief exposure to phorbol esters such as phorbol 12-myristate 13-acetate (PMA), presumably via activation of PKC (959, 960), since it was associated with phosphorylation of the exchanger (448). Furthermore, the activating effect of platelet-derived growth factor BB (PDGF-BB) was also associated with a phosphorylation of the exchanger; this effect of PDGF-BB was prevented by prior prolonged incubation with PMA (which downregulates PKC). Also, in rabbit renal arterioles, the exchanger was activated by PMA, and this effect was blocked by PKC inhibition (315).

Along a similar line, Khoyi et al. (497, 498) observed that the exchanger in aortic myocytes was augmented by phorbol 12,13-dibutyrate (PDBu), but not by the inactive analog 4-phorbol 12,13-didecanoate (4-PDD). They also found that the exchanger was stimulated by norepinephrine, and this effect was inhibited by 1-(5-isoquinolinylsufonyl)-2-methylpiperazaine (H-7), a relatively nonselective inhibitor of PKC. They therefore suggested that the action of norepinephrine, like that of PDBu, was mediated by PKC, even though norepineprine usually acts through PKA-dependent mechanisms. Their data do not, however, rule out the possibility that the exchanger in these cells was modulated by both PKC (PDBu) and PKA (norepinephrine; see sect. IIIE3).

Studies in cardiac muscle indicate that Na⁺/Ca²⁺ exchange is activated by a pathway involving PKC. Ballard and colleagues (41, 42) found that PMA and the agonists phenylephrine, angiotensin II, and endothelin-1, which augment PKC activity via a G protein (G_q)-mediated mechanism, enhanced Na⁺-dependent Ca²⁺ uptake into cardiac sarcolemmal vesicles. This transport activity was also augmented by the nonhydrolyzable GTP analog 5'-guanylyl amidophosphate.

In the central nervous system, too, there is evidence that stimulation of PKC increases the activity of the exchanger. External Na⁺-dependent Ca²⁺ efflux and Na_i-dependent Ca²⁺ influx in rat brain synaptosomes were both greatly augmented by PDBu, but not by 4-PDD (99). Also, PDBu, but not 4-PDD, promoted phosphorylation of the exchanger. Furthermore, this phosporylation, as well as the augmentation of the Ca²⁺ fluxes by PDBu, were blocked by H-7. Pharmacological studies on adrenal medullary cells also suggest that the exchanger is modulated by PKC, but not PKA (446).

In two other studies, however, activation of PKC by phorbol esters was found to decrease exchanger activity. Phorbol 12-myristate 13-acetate inhibited both basal and vasoconstrictor-stimulated, exchanger-mediated Ca²⁺ influx in human mesangial cells (648). Downregulation of PKC by 24-h pretreatment with PMA did not, however, prevent activation of the exchanger by angiotensin II. Thus the stimulation by the vasoconstrictor involves a PKC-independent pathway, possibly mediated by phospholipase C (see above).

Phorbol 12-myristate 13-acetate, but not the inactive 4-PMA, also was found to inhibit exchanger expression and function in a monkey kidney cell line (873). In this case, however, the effects of PMA were induced with a half-time (t_1/2) of 19 h, which raises the possibility that the reduction of activity was actually associated with downregulation of PKC rather than an acute inhibitory effect mediated by the activation of PKC.

There are several possible explanations for the inconsistent actions of PKC reported in the literature: 1) the NCX genes or splice variants activated by phosphorylation are different in the different cell types, 2) the different NCX genes or splice variants of the NCX genes (see sect. IV, D and E) behave differently after phosphorylation by PKC, 3) the initial experimental conditions are different, 4) at least some of the PKC effects are mediated by other proteins, and these intermediary proteins are different in the different experimental preparations.

Both basal and vasoconstrictor-stimulated Na⁺/Ca²⁺ exchange activity in human mesangial cells were acutely inhibited by receptor-mediated and by forskolin-induced activation of adenylyl cyclase activity and by dibutyryl cAMP (649). Activation of guanylate cyclase as well as treatment with dibutyryl cGMP had similar effects. In contrast, however, Furukawa and co-workers (332, 333) reported that cGMP augments Na_i-dependent Ca²⁺ influx in rat aortic myocytes. Also, in rat brain synaptosomes and cultured astrocytes, 8-bromo-cGMP augmented Na⁺/Ca²⁺ exchange, as did nitric oxide donors (presumably via a cGMP-mediated mechanism) (26). A recent report (394) suggests that the neuronal isoform of the exchanger (sse sect. IVE) is preferentially increased by PKA activation, when both neuronal and glial isoforms are studied in a Xenopus oocyte expression system.

In a study on immortalized rat arterial myocytes, however, chronic activation of adenylyl cyclase with forskolin or IBMX downregulated the exchanger (874). In the heart, -adrenergic agonists potentiated the force of contraction that was mediated by ₁-adrenoceptors and activation of adenylate cyclase (938). Fan et al. (303) have suggested that activation of ₁-receptors also, via the adenylate cyclase/cAMP cascade, may downregulate the exchanger (i.e., may reduce the exchanger-mediated current) and that this, in turn, may help to augment contraction by reducing Ca²⁺ efflux. This effect was blocked by the ₁-antagonist propranolol and was mimicked by forskolin, cAMP, and theophylline. Whether this is a direct effect, however, or whether the well-known cAMP-induced decline in [Ca²⁺]_i contributes to this downregulation of the Na⁺/Ca²⁺ exchanger is unresolved.

Calcium reabsorption across kidney distal tubule cells (see sect. VIG) involves, in part, passive Ca²⁺ entry across the apical membrane and, in part, extrusion by a Na⁺/Ca²⁺ exchanger located in the basolateral membrane (78). This reabsorption is stimulated by parathyroid hormone, which appears to augment the exchanger via a cAMP-mediated mechanism (117, 384, 735).

In sum, the observations described in the preceding paragraphs indicate that PKC and PKA contribute to the modulation of the Na⁺/Ca²⁺ exchanger in a variety of tissues. It appears that phosphorylation/dephosphorylation of the exchanger is involved in at least some of these modulatory processes. What is not clear, however, is precisely how these multiple and complex regulatory mechanisms influence net Na⁺/Ca²⁺ exchanger activity and contribute to cellular Ca²⁺ homeostasis. What is now needed is a systematic study of the influence of up- and downregulation of the Na⁺/Ca²⁺ exchanger, perhaps with the judicious use of selective kinase and phosphatase inhibitors, on Ca²⁺ homeostasis and physiology in intact cells and tissues. A major problem in the interpretation of such data will, however, hinge on the fact that PKA and PKC cascades influence many cellular processes that may both directly (e.g., A-kinase activation increases Ca²⁺ influx in heart cells by increasing Ca²⁺ entry via the L-type Ca²⁺ channel phosphorylation) and indirectly affect Ca²⁺ homeostasis and Ca²⁺-dependent processes such as contraction, secretion, and cell division. Here it is also important to emphasize the fact that _Na, which drives the exchanger, also may be influenced by PKA- and PKC-mediated phosphorylations that affect the activity of the Na⁺ pump (58, 72, 114). Thus it is also essential to know how much the exchanger (in isolation) is directly affected by these various regulatory mechanisms.

The consecutive transport model shown in Figure 12 (as well as other models) implies that the kinetics of transport or the turnover rate of the Na⁺/Ca²⁺ exchanger will be affected by the availability of Na⁺ and Ca²⁺ on both sides of the plasma membrane. In more concrete terms, even the "unidirectional" translocation of an ion will be affected by the concentration of that ion. If Ca²⁺ , for example, is to be translocated from inside to outside, the initial requirement must be that the Na⁺/Ca²⁺ exchanger be available so that intracellular Ca²⁺ can bind to the exchanger. A translocation step is then required to move the bound Ca²⁺ from inside to outside, followed by a Ca²⁺ unbinding step at the outside. This "vectoral" translocation involves a minimum of four kinetic steps. The rate of this unidirectional movement clearly depends on 1) the rate at which an "available-for-binding" state of the Na⁺/Ca²⁺ exchanger protein appears at an intracellular site; 2) the rate of Ca²⁺ binding to the Na⁺/Ca²⁺ exchanger protein "X" at the intracellular site (C2), which depends on [Ca²⁺]_i; 3) the rate of translocation of the bound Ca²⁺ (as CaX); and 4) the rate of Ca²⁺ dissociation at the extracellular surface (which should not depend on [Ca²⁺]_o). There is a similar unidirectional movement of Ca²⁺ into the cell that depends on the four kinetic steps acting in reverse (see Fig. 12). If these unidirectional fluxes of Ca²⁺ are equal, then there is no net (exchanger-mediated) transport of Ca²⁺.

A similar and symmetric consideration applies to the unidirectional movements of Na⁺ into and out of the cell. The unidirectional movements of Ca²⁺ or of Na⁺ reflect the kinetic features of the Na⁺/Ca²⁺ exchanger, while the net movements of Ca²⁺ and of Na⁺ are dependent on thermodynamic considerations. The two are clearly and necessarily interconnected. Net transport of Na⁺ and Ca²⁺ requires unequal, exchanger-mediated unidirectional fluxes of each ion species; the ratio of the inequalities of Na⁺ and Ca²⁺ transport (i.e., the ratio of the net Na⁺ and net Ca²⁺ fluxes) then provides a measure of the average coupling ratio. Because available evidence suggests that the transporter cannot "translocate" the ion binding sites in an empty state, and that the transport coupling ratio is fixed, the unidirectional transport inequality provides information on the net movement of ions by the Na⁺/Ca²⁺ exchanger.

Here, we use the term catalytic to refer to factors (other than the concentration dependence of the binding of transported ions) that affect the kinetics of the Na⁺/Ca²⁺ exchanger. Thus, for example, virtually all phosphorylation of the exchanger by PKA, PKC, or calmodulin kinases affect transport by catalytic mechanisms. Clearly, it should be possible to determine how these catalytic factors affect the Na⁺/Ca²⁺ exchanger-mediated net transport of ions by measuring the unidirectional ion movements. Particularly noteworthy (and unusual) features of catalytic regulation of the exchanger are the regulatory roles of Na⁺ and Ca²⁺, the transported ion species.

A) CATALYTIC MODULATION FOLLOWING -CHYMOTRYPSIN EXPOSURE TO THE INTRACELLULAR SIDE OF THE NA⁺/ca²⁺ exchanger protein ("deregulation"). A surprising observation was reported by Hilgemann et al. in 1992 (413), shortly after he first described the "giant" patch-clamp method (405, 406). They found that there was a Na_i- and time-dependent reduction in outward Na⁺/Ca²⁺ exchanger current (i.e., the I_Na/Ca associated with the Ca²⁺ entry mode). This reduction (Fig. 14A), called Na_i-dependent inactivation (see below), appeared to arise from an alteration in the kinetics of Na⁺/Ca²⁺ exchange, and not from a change in the thermodynamic "driving force" on the exchanger. Both the proton-dependent inhibition of Na⁺/Ca²⁺ exchange (see below) and the Ca_i-dependent activation of Na⁺/Ca²⁺ exchange (see below) also result from changes in the kinetics of the Na⁺/Ca²⁺ exchanger, rather than from changes in the energetics of transport. Thermodynamic considerations alone would suggest that an increase in [Na⁺]_i should increase rather than decrease the Na⁺/Ca²⁺ exchanger-mediated outward current. Likewise, thermodynamic considerations alone would also suggest that increasing [Ca²⁺]_i should decrease the Na⁺/Ca²⁺ exchange outward current, rather than increase it. Thermodynamically, protons should not affect the Na⁺/Ca²⁺ exchanger net transport because protons are not transported by this exchanger. Consequently, alterations in the Na⁺/Ca²⁺ exchanger turnover rate induced by changes in the intracellular concentrations of Na⁺, Ca²⁺, and H⁺ appear to depend on kinetic modulation of Na⁺/Ca²⁺ exchanger function and thus have been referred to as catalytic actions. This terminology is particularly appropriate for both Na⁺ and Ca²⁺ because these ions obviously also contribute to the energetics of transport.

Several observations suggest that the putative intracellular loop region of the Na⁺/Ca²⁺ exchanger is responsible, at least in part, for kinetic modulation of the Na⁺/Ca²⁺ exchanger transport by Na⁺, Ca²⁺, and protons. Following the initial report of activation of the Na⁺/Ca²⁺ exchanger by -chymotrypsin in cardiac vesicles (406, 635), Matsuoka and Hilgemann (632) showed that exposure of the intracellular surface of the giant patch to the protease, -chymotrypsin, removed the modulatory action of Ca²⁺. A similar result was obtained in internally perfused giant barnacle muscle fibers (299, 755). Subsequent studies revealed that -chymotrypsin also removed the modulatory action of H⁺ (266, 267). It has not yet been proven that treatment with -chymotrypsin removes the intracellular loop (or a large portion of it) from the remaining part of the Na⁺/Ca²⁺ exchanger protein. Nevertheless, several sites in the loop are candidate cleavage sites for proteolysis induced by the introduction of -chymotrypsin into the cytosol. Furthermore, Philipson and co-workers (634, 635) showed that removal or modification of the intracellular loop by mutagenesis also removed the catalytic modulation of the Na⁺/Ca²⁺ exchanger by Ca²⁺. Thus there is strong evidence that the intracellular loop plays an important role in regulating the kinetics of the Na⁺/Ca²⁺ exchanger.

The aforementioned studies focus primarily on the activity of the Na⁺/Ca²⁺ exchanger while it is operating in the Ca²⁺ entry mode and while it is extruding net Na⁺ rather than net Ca²⁺ (i.e., when it is producing a net outward current). Rasgado-Flores and colleagues (756, 299) have shown, however, that -chymotrypsin deregulation affects all modes of transport mediated by the exchanger in giant barnacle muscle fibers. This includes the "normal" (Ca²⁺ exit) mode of operation, in which net Ca²⁺ is transported out of the intracellular compartment, as well as both self-exchange modes.

B) CATALYTIC MODULATION BY [NA⁺]_i (na_i-dependent inactivation). A large patch of cardiac sarcolemma can be used with electrophysiological methods to investigate ion transport mediated by the Na⁺/Ca²⁺ exchanger (i.e., the giant patch-clamp method) (405, 406). A tight seal is made between a wide-tipped (e.g., 10 µm diameter) glass pipette and a bleb of sarcolemma. By pulling the pipette away from the cell, the membrane sticking to the glass is ripped away from the cell. This exposes the intracellular surface to the bathing solution while the extracellular surface of the membrane is exposed to the contents of the pipette.

In this configuration, when the bath [Na⁺] (= "cytoplasmic" Na⁺, or "Na_i") was elevated, "outward" current carried by the Na⁺/Ca²⁺ exchanger increased. Surprisingly, however, when [Na⁺] in the bathing medium was increased in a stepwise manner, outward current increased almost instantly, but the increase was only transient. Following the peak, the current declined in magnitude over time, suggesting that there was a time-dependent inactivation (Fig. 14A). This time-dependent inactivation of outward I_Na/Ca was termed Na_i-dependent inactivation because it was produced by elevating [Na⁺]_i (413, 632). Inactivation of I_Na/Ca is analogous to inactivation of ionic current flowing through channels and may depend on the rearrangement of a protein or part of a protein. For example, voltage-dependent inactivation of Na⁺ or Ca²⁺ channels is believed to depend on part of the channel protein simultaneously rearranging itself spatially to plug the channel pore (i.e., the ball and chain model; see Hille, Ref. 416). As suggested above, the putative intracellular loop region of the Na⁺/Ca²⁺ exchanger is a good candidate to play a key role in underlying Na_i-dependent inactivation. Indeed, intracellular application of -chymotrypsin, which may clip off all or parts of the cytoplasmic loop of the exchanger, removes Na_i-dependent inactivation (632). The molecular mechanisms that underlie ion transport and exchange are not established, however, and we do not know how the intracellular loop participates in the Na_i-dependent inactivation. We speculate that two independent processes take place simultaneously, but at different rates. One process (the more rapid) involves Na⁺ binding to its transport site on the Na⁺/Ca²⁺ exchanger to participate in Na⁺/Ca²⁺ exchange. The other (slower) process involves Na⁺ binding to a region of the intracellular loop, thereby causing the loop to interact with the part of the protein that controls transport kinetics. In the case of Na⁺ binding to the exchanger inhibitory peptide (XIP) region of the intracellular loop (633), the effect is to reduce the outward I_Na/Ca. The implication is that a rise in [Na⁺]_i has both a kinetic effect, viz. a decrease in Ca²⁺ influx, and an opposite thermodynamic effect, viz. a rise in [Ca²⁺]_i as a result of the change in the Na⁺ gradient; the former (kinetic) effect should only be observed with very high [Na⁺]_i.

C) CATALYTIC MODULATION BY [CA²⁺]_i. DiPolo (239) made the counterintuitive observation that the squid axon exchanger has an absolute requirement for intracellular Ca²⁺ while operating in the Ca²⁺ entry mode. This modulatory effect of intracellular Ca²⁺ could explain the inhibition of exchanger-mediated Ca²⁺ entry by intracellular Ca²⁺ chelators such as quin 2 (10) and EGTA (859). Also, removal of intravesicular Ca²⁺ by preincubating cardiac sarcolemmal vesicles with EGTA also reduces Na_i-dependent Ca²⁺ uptake (763). This catalytic effect of internal Ca²⁺ does not involve PKC because specific calmodulin inhibitors have little effect on the exchanger-mediated current in cardiac myocytes (512).

The giant patch-clamp method enables the study of the Ca²⁺ entry mode of exchange with electrophysiological methods. This permits one to investigate the modulation of transport function of the Na⁺/Ca²⁺ exchanger by catalytic amounts of [Ca²⁺]_i. If intracellular Ca²⁺ activates this Ca²⁺ entry mode of exchange, the kinetic and thermodynamic effects of a rise in [Ca²⁺]_i should be exerted in opposite directions. Thus activation of net influx of Ca²⁺ or net extrusion of Na⁺ by the Na⁺/Ca²⁺ exchanger, induced by elevation of [Ca²⁺]_i, must be brought about by a modulatory or catalytic action of the Ca²⁺. With the use of this method of investigation, the Na⁺/Ca²⁺ exchanger in the heart is half-maximally activated at a [Ca²⁺]_i level of ~1 µM (412). In squid axons (243, 244) and barnacle muscle (753, 757), tracer flux studies yielded similar results (see Table 2). In fact, the Ca²⁺ entry mode of exchange was abolished when [Ca²⁺]_i was reduced to <0.1 µM.

There also is strong evidence that intracellular Ca²⁺ can augment the turnover rate of the Na⁺/Ca²⁺ exchanger when it is operating in the Ca²⁺ exit mode (412, 661). This catalytic effect of intracellular Ca²⁺ could, because it is activated by a rise in [Ca²⁺]_i, be confused with an increase in the thermodynamic driving force favoring the extrusion of Ca²⁺. It is difficult to separate the two effects experimentally, and there continues to be uncertainty about the exact level of [Ca²⁺]_i needed to produce the catalytic activation of Na⁺/Ca²⁺ exchanger transport. Reported K_0.5 values range from ~10-50 nM (661) to on the order of 1 µM (247, 248, 258, 412).

With the use of a Ca²⁺ overlay analysis of the Na⁺/Ca²⁺ exchanger protein on a Western blot, it has been possible to identify part of the intracellular loop in NCX1 as the site of the Ca²⁺-protein interaction (579). The conclusion from these experiments has been supported by deletion mutagenesis studies which suggest that residues 371 to 508 are necessary for Ca²⁺ binding. Two highly acidic regions (residues 446-454 and 499-510), each containing three consecutive aspartic acid residues, appear to be particularly important. When Ca²⁺ binds to this region, the rate of Na⁺/Ca²⁺ exchange is accelerated.

Because the steady-state actions of cytosolic Na⁺ and Ca²⁺ have opposite catalytic effects on the Na⁺/Ca²⁺ exchanger, with Na_i decreasing transport rate and Ca_i increasing it, one may speculate that one of two basic mechanisms is operating. One possibility is that the baseline function of the Na⁺/Ca²⁺ exchanger involves tonic but incomplete autoinhibition in which some or all of the intracellular loop participates. If this is the case, then increasing [Na⁺]_i augments that autoinhibition to produce the Na_i-dependent inactivation. Conversely, increasing [Ca²⁺]_i should decrease the autoinhibition and thereby increase the rate of Na⁺/Ca²⁺ exchange. The other possibility is that the intracellular loop is involved in autostimulation of the exchanger. If this is the case, then increasing [Na⁺]_i should lead to the reduction of the autostimulation of Na⁺/Ca²⁺ exchange. Conversely, increasing [Ca²⁺]_i should augment the autostimulation and thereby increase Na⁺/Ca²⁺ exchange. Because treatment with -chymotrypsin increases the Na⁺/Ca²⁺ exchanger current, and because we presume that this proteolysis removes the intracellular loop, we favor the autoinhibition hypothesis. The alternative hypothesis would have been supported if -chymotrypsin treatment had decreased the rate of Na⁺/Ca²⁺ exchange.

The functional significance of the different splice variants of the NCX1 Na⁺/Ca²⁺ exchanger (527) is only now beginning to be studied. Recent preliminary experiments suggest that the cardiac and the kidney isoforms of the Na⁺/Ca²⁺ exchanger behave differently (795). The differences noted include different responses to PKA-dependent phosphorylation, to [Ca²⁺]_i, and to V_M. How these reported differences in exon A-containing isoforms (e.g., heart) and exon B-containing isoforms (e.g., kidney) affect cellular physiology remains unknown. The main structural differences in these splice variants occur in the large putatively cytoplasmic loop. No clear mechanistic model has emerged that allows one to understand how the slice variants could bring about these reported differences. Nevertheless, distant relatives of the mammalian Na⁺/Ca²⁺ exchanger such as the Drosophila Na⁺/Ca²⁺ exchanger appear to have splice variants in the same intracellular loop region (796). Furthermore, recent findings (431, 796) suggest that intracellular Ca²⁺ regulation by the Drosophila Na⁺/Ca²⁺ exchanger differs from that observed in the mammalian Na⁺/Ca²⁺ exchanger. Distinctive regulation of Drosophila isoforms (796) by diverse modulators has not been demonstrated.

The speed at which the catalytic actions of Na_i and Ca_i occur may have important physiological consequences. When [Na⁺]_i increases, net transport by the Na⁺/Ca²⁺ exchanger (Ca²⁺ entry and Na⁺ exit) normally increases rapidly (see Fig. 14A). This occurs because the net transport of Na⁺ out of the intracellular compartment is then favored thermodynamically and is no longer substrate limited. Following this increased net outward Na⁺ transport, however, the exchanger spontaneously inactivates (Na_i-dependent inactivation; t_1/2 = 1.5-2 s at 37°C; see Fig. 14A and Refs. 413, 632). Catalytic activation by Ca_i is fairly rapid; the component that is independent of Na_i activates with a t_1/2 of 100 ms at 37°C (412; Hilgemann, personal communication). Even during a single cardiac [Ca²⁺]_i transient lasting ~100-200 ms or more, catalytic activation by [Ca²⁺]_i will occur. Importantly, Ca_i-dependent activation decays slowly upon removal of Ca_i (Fig. 14A; t_1/2 is on the order of seconds under physiological conditions; see Ref. 412). These catalytic actions of [Ca²⁺]_i on the Na⁺/Ca²⁺ exchanger turnover rate mean that as the time-averaged [Ca²⁺]_i rises, the Na⁺/Ca²⁺ exchanger becomes more responsive to changes in [Ca²⁺]_i.

The functional importance of catalytic activation of the Na⁺/Ca²⁺ exchanger transport by increased [Ca²⁺]_i appears to make sense in cells where Ca²⁺ signaling depends on the frequency and amplitude of [Ca²⁺]_i transients rather than on the mean averaged [Ca²⁺]_i such as heart cells and neurons {see discussions on local and global [Ca²⁺]_i signaling and on the frequency and amplitude modulation of this signaling (65, 113)}. In these cells a premium is placed on the restoration of ion balance even when Ca²⁺ entry is increased. The catalytic activation of the Na⁺/Ca²⁺ exchanger will support this process, since this is a means of increasing the turnover rate of the Na⁺/Ca²⁺ exchanger when it operates in the Ca²⁺ exit mode. The Na⁺/Ca²⁺ exchanger under physiological conditions works in the Ca²⁺ efflux mode almost all of the time in most cells. The slow net catalytic effect of the Na⁺/Ca²⁺ exchanger to increased [Ca²⁺]_i serves the purpose of integrating the effects of [Ca²⁺]_i transients on the Na⁺/Ca²⁺ exchanger. The Na_i-dependent inactivation may reflect molecular and biophysical processes of the Na⁺/Ca²⁺ exchanger; it may not play an important functional role because normal [Na⁺]_i is low (5-10 mM) and is not likely to effect much inactivation. Thus the Na_i-dependent inactivation may be a valuable tool for use when exploring the molecular operation of the Na⁺/Ca²⁺ exchanger but may not be an important physiological modulator. Nevertheless, insofar as Na_i-dependent inactivation is turned on, it will tend to reduce the turnover rate of the Na⁺/Ca²⁺ exchanger and thus will tend to increase global averaged [Ca²⁺]_i. Interestingly, under the conditions of elevated [Na⁺]_i, the Ca²⁺- dependent and Na⁺-dependent catalytic processes would appear to work against each other.

A further complication is the evidence that, in some cells, the Na⁺/Ca²⁺ exchanger may operate on Na⁺ and Ca²⁺ in a tiny restricted subplasmalemmal volume of cytosol where the concentrations of these ions may be quite different from those in "bulk" cytosol (103, 328, 618, 895a; see sect. VIA7). Under these circumstances, the kinetic modulation of the Na⁺/Ca²⁺ exchanger could not be predicted from the dynamic changes in bulk [Na⁺]_i and [Ca²⁺]_i.

D) CATALYTIC MODULATION OF THE NA⁺/ca²⁺ exchanger by protons. The giant patch-clamp method was also used to examine the modulation of the Na⁺/Ca²⁺ exchanger by intracellular pH (266, 267). The basic protocol was to produce a large outward Na⁺/Ca²⁺ exchanger current (I_Na/Ca), similar to those described above, by changing the bath [Na⁺] (i.e., [Na⁺]_i) from a low level (e.g., 0 mM) to a high level (e.g., 60 mM Na⁺). Intracellular pH was changed at various times, under these conditions. The data (Fig. 16) confirmed earlier reports (241, 718) that protons inhibited the Na⁺/Ca²⁺ exchanger, but there were two unexpected results. The first was that the normal intracellular pH (~7.2) was in the middle of the curve that relates exchanger activity to pH_i (Fig. 16B). This means that any variation of intracellular pH may have a great impact on Na⁺/Ca²⁺ exchanger by either inhibiting transport (at more acidic pH) or augmenting Na⁺/Ca²⁺ exchanger transport (at more basic pH). The maximum rate of Na⁺/Ca²⁺ exchange could thereby range between zero and double the rate at normal pH. Becasue -chymotrypsin did not augment Na⁺/Ca²⁺ exchange at high (basic) intracellular pH (Fig. 16A), these results suggest that protons may promote autoinhibition of the Na⁺/Ca²⁺ exchanger by interacting with one or more regions of the intracellular loop. As noted above, the Ca_i and Na_i catalytic effects also appear to depend on interactions of the intracellular loop with the transmembrane regions of the Na⁺/Ca²⁺ exchanger.

View larger version (14K): [in this window] [in a new window]   Fig. 16. Effect of protons on INa/Ca. A: Na+ inactivation depends on intracellular pH (pHi); it is observed at pHi 6.4, but not at pHi 7.2. B: inactivation by Nai is removed by treatment with chymotrypsin (which presumably removes intracellular loop enzymatically). C: magnitude of steady Na+/Ca2+ exchanger current depends directly on pHi (); inactivation by Nai is abolished at high pHi. This dependence on pHi is removed by treatment with chymotrypsin. [From Doering and Lederer (267).]

The second surprising result was that the proton-dependent modulation of the Na⁺/Ca²⁺ exchanger involves Na⁺-dependent modulation of the Na⁺/Ca²⁺ exchanger. In the absence of intracellular Na⁺, protons had little effect, whereas, at high (basic) pH, intracellular Na⁺ did not produce inactivation. These findings suggest that both Na⁺ and protons are needed to produce the autoinhibition of the exchanger by the intracellular loop. This second result thus further reinforces the working hypothesis that has evolved from the work of several groups (266, 267, 413, 590). According to this hypothesis, the intracellular loop region of the Na⁺/Ca²⁺ exchanger is critically involved with modulation of the Na⁺/Ca²⁺ exchanger function.

E) MODULATION OF THE NA⁺/ca²⁺ exchanger by temperature. The isoform of the Na⁺/Ca²⁺ exchanger that is expressed in the heart is very sensitive to temperature at temperatures near body temperature, with a Q₁₀ between 3 and 4 in guinea pig cardiac myocytes (693). Using the giant patch method, Kappl and Hartung (475) obtained a slightly lower value (2.6) for a temperature of 21-38°C and a very high Q₁₀ (6.7) for temperatures below 21°C (475). This wide range of values for the measured Q₁₀ is expected since at lower temperatures, the Na⁺/Ca²⁺ exchanger function falls to zero (see Ref. 693). Clearly, this zero-function temperature (T_ZF) represents a discontinuity for Q₁₀ measurements and varies with species and condition. Measuring Q₁₀ close to T_ZF may account for the wide variations in measured values. The observations noted above have generally been supported by recent investigations by others (386, 495, 924, 969) in heart muscle cells, and Khananshvili et al. (495) found a decrease in the Q₁₀ at temperatures above 30°C (495). Again, this is not an unexpected result, given the proximity to T_ZF. The Q₁₀ for Na⁺/Ca²⁺ exchange in rat brain is ~3.0 at 30-35°C (606). As suggested above, it is likely that some of this marked temperature sensitivity depends on which gene encodes for the expressed Na⁺/Ca²⁺ exchanger, or on which splice variant of the Na⁺/Ca²⁺ exchanger is expressed as well as on the physiological body temperature. This likelihood is supported by the fact that, in the giant axon of the squid and muscle of the lobster, cold water invertebrates, the Na⁺/Ca²⁺ exchanger is used to extrude Ca²⁺ at temperatures at which the guinea pig heart Na⁺/Ca²⁺ exchanger activity is largely inhibited (693). Moreover, most studies on Na⁺/Ca²⁺ exchangers from invertebrates and fishes reveal Q₁₀ values on the order of 1.2-1.8 (289, 464, 921, 924); the frog heart exchanger also has a lower Q₁₀ than those from dogs and rabbits (71). Ruscak et al. (799), however, obtained a Q₁₀ of 3.2 for the exchanger from crayfish muscle. An understanding of the temperature dependence of the different gene products may provide valuable information on the relationship between the structure of the Na⁺/Ca²⁺ exchanger and its function.

A convenient as well as important way to examine the physiological role of a transport system is to block the system with a highly selective inhibitor. Good examples are the use of ouabain to inhibit the Na⁺ pump and TTX to inhibit the voltage-gated Na⁺ channel in excitable cells. Even somewhat less selective inhibitors have been used effectively to identify and characterize some transport systems when the cell type being studied contains, almost exclusively, just one of the transport proteins affected by the inhibitor. A case in point is amiloride, which has been used to identify and characterize the Na⁺ channel present in high-resistance ("tight") epithelia such as the apical membranes of mammalian renal cortical collecting tubules (59).

The situation is not so simple for the Na⁺/Ca²⁺ exchanger. Regrettably, there are no inhibitors of the Na⁺/Ca²⁺ exchanger that are sufficiently selective that they can be used routinely to characterize the exchanger in intact cells and tissues unless the exchanger is more-or-less isolated from other Na⁺ or Ca²⁺ transport systems or current carriers. Another complicating factor is that the exchanger mediates both Ca²⁺ influx and Ca²⁺ efflux. It is possible that the kinetics of these two modes of exchange may be differently affected by inorganic cations. For example, some divalent cations, when added to the extracellular fluid, may interfere with the binding of Ca²⁺ at the extracellular surface of the PM, but may also, themselves, be transported by the exchanger so that they promote Ca²⁺ extrusion (see sect. IIIF5).

The subject of Na⁺/Ca²⁺ exchanger inhibitors was extensively reviewed a decade ago (474). Several new developments have taken place since then. The XIP was introduced in 1991 (587). A new nonpeptidic inhibitor, an isothiourea derivative, was also described recently (449, 970). Using another approach that followed the cloning of the exchanger, several laboratories employed antisense oligodeoxynucleotides (AS-oligos) to knock out the exchanger by selectively inhibiting expression of the protein (83, 596, 865-867, 907).

Numerous structurally distinct organic molecules have been tested for their ability to inhibit the Na⁺/Ca²⁺ exchanger. Some of the most effective and potent compounds identified to date are certain amiloride analogs (472, 474, 520, 854, 861). Amiloride, an acylguanidine diuretic, is, itself, a weak inhibitor of the exchanger (IC₅₀ ~1 mM), but it is a much more potent blocker of certain epithelial Na⁺ channels (59). In contrast, hydrophobic substitutions at the acylguanidinium nitrogen (e.g., with a benzyl group) or at the 5-position amino group of the pyrazine ring, substantially increase the potency for block of the exchanger. Two examples are "benzamil," with an IC₅₀ of 100 µM, and N⁵-(2,4-dimethylbenzyl)amiloride, with a K_i of 10 µM. Unfortunately, many of these compounds have very limited solubility in aqueous solutions. The major problem with them, however, is their lack of specificity. For example, N⁵-substituted analogs are also potent inhibitors of Na⁺/H⁺ exchange, while guanidino N-substituted analogs are potent blockers of epithelial Na⁺ channels. A further problem with using amiloride analogs to study Ca²⁺ homeostasis is that some of the more effective inhibitors of Na⁺/Ca²⁺ exchange, such as 3,4-dichlorobenzamil, also block voltage-gated Ca²⁺ channels even more potently than they block the exchanger, at least in some preparations (75, 304, 505, 506). Consequently, interpretation of physiological/pharmacological studies on the action of amiloride analogs in intact cells or tissues (142, 223, 224, 342, 504, 609, 727, 869) or even in vesicles (825) must be viewed with considerable caution. Nevertheless, a number of investigators have used amiloride analogs such as 3,4-dichlorobenzamil to identify Na⁺/Ca²⁺ exchanger-mediated currents when other ion channels were already blocked (593, 693).

The mechanism of block of the exchanger by amiloride analogs is complex (861). At low concentrations they bind preferentially to an Na⁺ binding site, whereas, at higher concentrations, they also interact at another site to which Na⁺, Ca²⁺, and K⁺ also bind.

Several antiarrhythmic agents such as quinacrine (797, 75) and bepridil (a substituted pyrrolidine ethanamine) (474) also inhibit the Na⁺/Ca²⁺ exchanger. These molecules, too, are nonselective; they inhibit other transport processes as well, at concentrations similar to those needed inhibit the exchanger.

Various local anesthetics (e.g., tetracaine and dibucaine; Ref. 651), general anesthetics (e.g., halothane; Refs. 53, 391, 392), and Ca²⁺ entry blockers (457) have also been shown to inhibit the Na⁺/Ca²⁺ exchanger. They do so, however, at concentrations that may be considerably higher than the concentrations at which they exert their primary actions. For example, the Ca²⁺ channel blocker verapamil inhibits the Na⁺/Ca²⁺ exchanger (297), but the effect is a very weak one, compared with its ability to block L-type Ca²⁺ channels (457). Thus the main problem is the overzealous use of this compound in experimental situations where interpretation of its action as a relatively selective block of Ca²⁺ channels may be compromised because of secondary effects such as block of the Na⁺/Ca²⁺ exchanger.

A number of other molecules such as ascorbic acid (629, 909), chlorpromazine (161) and other anti-depressants (560), quinacrine (which also activates the exchanger under some circumstances) (228), quinidine (230), neomycin (152), polymyxin B (723), and the amino oxidase inhibitor harmaline (898), also are weak inhibitors of the Na⁺/Ca²⁺ exchanger. Again, however, these molecules all have other actions that can be expected to overshadow their effects on the Na⁺/Ca²⁺ exchanger except under very selective circumstances.

Recently, a new isothiourea derivative was reported to inhibit the Ca²⁺ entry mode of the exchanger preferentially (449, 970). This agent, 2-[2-[4-(4-nitrobenzyloxy)phenyl]ethyl]isothiourea methanesulfonate (compound number 7943), inhibited, with high affinity (IC₅₀ =1-3 µM), Na_i-dependent ⁴⁵Ca²⁺ influx (and the concomitant rise in [Ca²⁺]_i) in vascular smooth muscle, cardiac sarcolemmal vesicles, and fibroblasts transfected with Na⁺/Ca²⁺ exchanger cDNA (449). In contrast, the IC₅₀ for inhibition of the Na_o-dependent Ca²⁺ efflux and Na_o-dependent decline in [Ca²⁺]_i was ~30 µM. Similarly, compound number 7943 inhibited outward I_Na/Ca (corresponding to Ca²⁺ entry mode exchange) with an IC₅₀ of ~0.3 µM, but inward I_Na/Ca (corresponding to Ca²⁺ exit mode exchange) with an IC₅₀ of ~17 µM (970). In other words, it is apparently ~50-fold less effective in blocking the Ca²⁺ influx mode of exchange than the efflux mode. The significance of this preferential block of one mode of net Ca²⁺ transport is difficult to understand based on current models of the Na⁺/Ca²⁺ exchanger mechanism. Furthermore, this agent is not very selective, in that it also blocks Na⁺ channels, Ca²⁺ channels, and certain K⁺ channels with IC₅₀ values of 7-14 µM (970).

When they cloned and sequenced the cardiac Na⁺/Ca²⁺ exchanger, Philipson and colleagues (587, 687) observed that the large intracellular loop of the exchanger contained a segment of interspersed hydrophobic and basic residues reminiscent of a calmodulin binding site. To examine the possibility that this site might be autoinhibitory, they synthesized a peptide with this sequence ("exchanger inhibitory peptide," XIP) and showed that it did, indeed, inhibit Na⁺/Ca²⁺ exchange (587). Other studies, on various preparations (179, 255, 519, 635), have confirmed this effect. Use of XIP as an exchanger inhibitor in appropriate isolated systems may provide insight into the modulatory role of the intracellular loop (377, 396). The physiological consequences of XIP action may be difficult to interpret, however, because XIP can be expected to bind indiscriminately to other calmodulin binding proteins and to interfere with the actions of endogenous calmodulin in intact cells.

An opiate-like binding site has also been functionally identified in the cardiac Na⁺/Ca²⁺ exchanger. The molluscan peptide Phe-Met-Arg-Phe-amide (FMRF-amide) is known to bind to sites with an opiate binding site sequence. FMRF-amide has now been shown to inhibit the exchanger in cardiac sarcolemmal vesicles (492, 493), in internally dialyzed squid giant axons (255), and in pancreatic -cells (951).

Recently, Khananshvili et al. (494) synthesized a series of positively charge hexapeptides (Arg₂, Phe₂, Cys₂, containing an internal disulfide bond) with a terminal Arg-Phe-amide to target the internal opiate-like binding site on the exchanger. A subsequent study revealed that one of these molecules, FRCRCF-amide, when dialyzed into cardiac myocytes at a concentration of 1 µM or more, completely and selectively inhibited I_Na/Ca [418; and see Khananshvili and colleagues (490, 491)]. This agent had no effect on L-type Ca²⁺ current or on K⁺ currents.

A) LANTHANIDES. A variety of divalent and trivalent cations have long been employed as substitutes or inhibitors for various Ca²⁺-dependent processes because these ions, with a high density of positive charge, often bind to sites that also bind Ca²⁺. Therefore, it is not surprising that La³⁺ was tested for its effects on Na⁺/Ca²⁺ exchange beginning with some of the earliest studies on this transporter (35). Numerous reports indicate that La³⁺ inhibits Na⁺/Ca²⁺ exchange in a large variety of cell types, including nerve and various kinds of muscle (35, 139, 153, 289, 347, 443, 514, 544, 868, 936). With the exception of a few studies on isolated or reconstituted vesicles (289, 347, 936), the La³⁺ was applied only in the extracellular fluid. However, La³⁺ can apparently enter cells, e.g., as a surrogate for Ca²⁺ on the Na⁺/Ca²⁺ exchanger (712, 849). Thus, in at least some instances, it is possible that La³⁺ may be acting at the inside of the plasma membrane. More importantly, Katzung et al. (481) first showed that external La³⁺ had a relatively low affinity for the cardiac Na⁺/Ca²⁺ exchanger. This has also been observed in vascular smooth muscle, where the IC₅₀ exceeds 0.5 mM (333, 849). Because external La³⁺ is known to inhibit the plasmalemmal Ca²⁺ pump (and voltage-gated Ca²⁺ channels) with much higher affinity (744, 814), it is possible to use low-dose La³⁺ (<0.5 mM) to inhibit these other Ca²⁺ transport systems selectively (333, 849, 866) so that the function of the exchanger can be examined in relative isolation. Nevertheless, several investigators have used La³⁺ to block (and identify) exchanger-mediated ionic currents in electrophysiological experiments (153, 514).

Trosper and Philipson (936) studied the effects of several other lanthanides (Nd³⁺, Tm³⁺, and Y³⁺) on Na⁺/Ca²⁺ exchange in cardiac sarcolemmal vesicles; these ions were all less-effective inhibitors than La³⁺. Also, Gd³⁺ does not inhibit Na⁺/Ca²⁺ exchange, although it blocks the Ca²⁺ channels in nerve terminals.

B) NI²⁺ and other divalent cations. Many divalent cations bind to Ca²⁺ binding sites on proteins and have thus been employed as either surrogate ions for Ca²⁺ or as inhibitors of Ca²⁺-dependent processes. Indeed, some ions may be transported by the Na⁺/Ca²⁺ exchanger and may also inhibit the exchanger, as in the case of high La³⁺ concentrations (849).

Among the alkaline earth ions, Mg²⁺ is reported to be a very weak inhibitor (243, 513, 868) that does not appear to be transported by the exchanger (227, 936). Indeed, in at least some cell types, Mg²⁺ is transported by an independent, Na⁺-coupled countertransport (Na⁺/Mg²⁺ exchange) system (361, 247, 918).

In the heart, for example, external Mg²⁺ inhibits the exchanger with an IC₅₀ of ~12.5 mM (513). Physiological levels of intracellular Mg²⁺ (~4 mM in squid axons) inhibit both the Ca²⁺ influx and Ca²⁺ efflux modes of the squid axon exchanger by ~50% (138, 243, 246). Other evidence, however, indicates that external Mg²⁺ does not inhibit the squid axon exchanger (254).

In cultured arterial smooth muscle, it has been reported that Mg²⁺ inhibits Ca²⁺ entry with a K_i of 93 µM (868). In contrast, preliminary results indicate that the Na_o-dependent extrusion of Ca²⁺ in these cells is negligibly affected by 10 mM extracellular Mg²⁺ (A. Arnon and M. P. Blaustein, unpublished data; and see Refs. 849, 866).

Unlike Mg²⁺, Sr²⁺ and, to a lesser extent, Ba²⁺ are both transported by the Na⁺/Ca²⁺ exchanger (88, 203, 800; see also Refs. 85, 207); however, the maximal rate of transport of these ions by the exchanger is much slower than the rate of transport of Ca²⁺ (88). Furthermore, both Sr²⁺ and Ba²⁺ also interfere with Ca²⁺ transport mediated by the exchanger (88, 800, 936).

Cd²⁺ (417, 936), Co²⁺ (405), Mn²⁺ (9, 800, 936), Zn²⁺ (200), and Ni²⁺ (514) all inhibit Na⁺/Ca²⁺ exchange, but none of these ions is a very specific inhibitor. Moreover, at least Ni²⁺, Mn²⁺, and Cd²⁺ may be transported (as surrogates for Ca²⁺) by the exchanger (316, 655).

Ni²⁺ is an inhibitor of Na⁺/Ca²⁺ exchange that has often been used to block, and thereby identify, Na⁺/Ca²⁺ exchanger-mediated currents in electrophysiological experiments (328, 514, 618, 693, 695, 699). The concentrations of Ni²⁺ required are on the order of 2-5 mM so that the Ni²⁺ might be expected to have other effects unless the exchanger current is examined in relative "isolation" (i.e., after other Ni²⁺-sensitive ionic currents are inhibited by different agents).

It should be possible to inhibit Na⁺/Ca²⁺ exchange with appropriate antibodies raised against the Na⁺/Ca²⁺ exchanger. Unfortunately, however, almost all of the readily accessible epitopes are located in the large cytoplasmic domains of the exchanger; polyclonal antibodies appear to bind relatively poorly to the extracellular domains of the exchanger (460, 610). One monoclonal antibody was originally reported to block Na⁺/Ca²⁺ exchanger function, but this could not be confirmed (474), and there have been no further reports concerning block of Na⁺/Ca²⁺ exchanger function by antibodies.

One other way to interfere with the function of the Na⁺/Ca²⁺ exchanger is to inhibit the expression of the exchanger protein. Recently, several groups employed AS-oligos directed against the exchanger mRNA to this end. Lipp et al. (596) targeted a single 19-mer nonchimeric phosphorothioated oligo to the conserved 3'-untranslated region of the NCX1 gene mRNA. They reported that their AS-oligo, at a rather high concentration (3 µM), knocked down 80% of exchanger function (measured as a reduction of [Ca²⁺]_i after release of "caged" Ca²⁺) in cardiac myocytes (reversibly) within 24 h. Takahashi and colleagues (83, 907) also targeted a single nonchimeric phosphorothioated 20-mer AS-oligo to a region near the untranslated 3'-end of the NCX1 exchanger mRNA. They treated cardiac myocytes with 10 µM AS-oligo in the presence of lipofectamine (used to promote AS-oligo uptake) and reported that the Ca²⁺ influx from Na⁺-free medium was reduced by 20-30%, and beat frequency was increased in 30% of treated cells. The integrity of other physiological properties was not shown in either study, nor was knockdown of the exchanger protein demonstrated biochemically. The specificity of the knock down (as opposed to a nonspecific effect of these AS-oligos) may be questioned in these cases because both groups used relatively high AS-oligo concentrations and reported substantial knock down within 24 h; this would imply an exchanger protein half-life of <12 h, which seems very short for an integral membrane transport protein.

Slodzinski and colleagues (865-867) targeted a tandem pair of short (15- and 17-mer) chimeric phosphorothiated AS-oligos to the region around the start codon of the NCX1 gene mRNA. Low concentrations (0.5 µM) of these AS-oligos markedly and reversibly knocked down Na⁺/Ca²⁺ exchanger function (measured as Na_i-dependent Ca²⁺ entry and Na_o-dependent Ca²⁺ exit) in cultured arterial smooth muscle cells, while sparing other membrane transport functions (866, 867). During fast (every 3 min) but not slow (every 15 min) repetitive stimulation with serotonin, the recovery of [Ca²⁺]_i to control levels was greatly slowed in the AS-oligo-treated arterial myocytes. These AS-oligos did not, however, interfere with the growth and division of the arterial myocytes. These same chimeric oligos also knocked down exchanger expression and function in ~60% of treated neonatal rat cardiac myocytes, but the effect was apparent only after 4 days of incubation with the AS-oligos, which is consistent with the measured exchanger half-life of ~33 h in these cells (865). In contrast to the results reported by Takahashi et al. (907), the spontaneous activity was inhibited in the 60% of the AS-oligo-treated cells in which the exchanger function was knocked down (865).

In a study on cultured rat astrocytees, Takuma and colleagues (630) employed antisense strategy to knock down exchanger function. This attenuated reperfusion injury (manifested by elevated cytosolic [Ca²⁺]) and reduced cell toxicity. The implication is that the cells gained Na⁺ when external Ca²⁺ was withdrawn; then, upon reintroduction of external Ca²⁺, the usual rapid Ca²⁺ entry was apparently mediated by the Na⁺/Ca²⁺ exchanger, since it was reduced in AS-oligo-treated cells.

The effects of AS-oligos have also been studied in rat pancreatic -cells (951a). The results of Na⁺/Ca²⁺ exchanger knock-down with the AS-oligos indicate that the exchanger mediates Ca²⁺ entry in response to depolarization and is responsible for up to 70% of Ca²⁺ removal following cell activity.

	IV.  MOLECULAR BIOLOGY OF THE SODIUM/CALCIUM EXCHANGER

Top Previous Next References

In 1990, the Na⁺/Ca²⁺ exchanger gene was first cloned from dog heart muscle by Nicoll et al. (688). This work was a turning point in Na⁺/Ca²⁺ exchanger research. All of the molecular biology research work described in the present review was influenced by this work. As indicated below, this seminal study has made possible a wide range of physiological studies, including investigations into issues of structure and function.

The Na⁺/Ca²⁺ exchangers comprise a family of three genes (to date): NCX1 (21), NCX2 (85) and NCX3 (689). These genes have similar hydropathy patterns, suggesting similar overall structure and moderate amounts of sequence identity. Genes homologous to NCX1 have been characterized in diverse mammalian species including human (525, 530), rabbit (527), cattle (4), guinea pig (943), mouse (507), and rat (331, 606) as well as other species. Interestingly, genes similar to the Na⁺/Ca²⁺ exchanger gene family have also been cloned from nonmammalian species such as Drosophila (796, 840), Xenopus (450), Caenorhabditis elegans (nematode; genebank accession number Z70312) (982), and Loligo (squid; genebank accession number U93214) (397).

The successful cloning of the Na⁺/Ca²⁺ exchanger has stimulated numerous investigations and has spawned specific findings that bear on diverse biochemical, physiological, biophysical, and molecular aspects of the exchanger's properties. Because of this recent explosion of studies, this section of the review is necessarily limited to a brief overview of the work on the molecular biology of the Na⁺/Ca²⁺ exchanger. Additional relevant discussion is distributed throughout the other sections.

Although one of the two initial identifications and descriptions of the Na⁺/Ca²⁺ exchanger was carried out in mammalian heart (783), the physiological importance and multiple roles of the Na⁺/Ca²⁺ exchanger have only been realized more recently. Much of that appreciation arises from the now classic transport studies of the Na⁺/Ca²⁺ exchanger in cardiac sarcolemmal vesicles. These early studies demonstrated that, in the heart, there was rapid and massive Na_i-dependent Ca²⁺ movement, that the coupling ratio was 3 Na⁺:1 Ca²⁺, and that the transport was probably rheogenic (electrogenic) (70, 161, 440, 726, 762, 764, 765). This work laid the foundation for the protein purification (719) that made the cloning of the Na⁺/Ca²⁺ exchanger possible.

Preparation of cardiac sarcolemmal vesicles, which led to the eventual protein purification, was carried out as follows (264, 720). Heart tissue was homogenized and purified to obtain sarcolemmal vesicles. These vesicles were purified with respect to the Na⁺-K⁺-ATPase (Na⁺ pump), a robust and omnipresent PM protein (726, 764, 765). Because the Na⁺/Ca²⁺ exchanger is also predominantly a PM protein (172, 199, 318, 501), this purification strategy was valuable. The first strong indication of the rheogenicity (726) and the coupling ratio (765) of the cardiac Na⁺/Ca²⁺ exchanger was based on vesicle experiments using this purification method. That the Na⁺/Ca²⁺ exchanger was present in the cardiac PM vesicle preparations was demonstrated by examining the Na_i-dependent Ca²⁺ influx in the vesicles that was stimulated by lowering [Na⁺]_o (70, 726, 764, 765).

Cloning of the Na⁺/Ca²⁺ exchanger was testament to both thoughtful and careful work combined with persistence and the clever application of traditional and novel methods. The strategy used by Nicoll et al. (688) involved purification of the protein. This key step led to the identification of polypeptides with molecular masses 70, 120, and 160 kDa (719). Polyclonal antibodies were raised against these exchanger proteins and were used to screen a cardiac expression library. A clone that coded for a protein fragment that cross-reacted with these antibodies was initially identified, and this clone was used for further library screenings. A full-length clone was identified and used to produce cRNA that, when injected into Xenopus oocytes, induced the expression of Na⁺/Ca²⁺ exchanger activity (688).

The canine cardiac Na⁺/Ca²⁺ exchanger clone (of the NCX1 gene) was shown to express a functional protein by injecting cRNA into Xenopus oocytes and measuring transport function in the injected oocytes. Nicoll et al. (688) showed that Na_i-dependent (low [Na⁺]_o stimulated) Ca²⁺ influx could be measured in oocytes injected with the cRNA for the Na⁺/Ca²⁺ exchanger but not in water-injected controls (688). In those experiments, the Xenopus oocytes were first preloaded with Na⁺ by incubating the oocytes with a cardiotonic steroid to inhibit the Na⁺-K⁺-ATPase before changing extracellular Na⁺. In contrast, when the human Na⁺/Ca²⁺ exchanger was cloned (525, 530), function was first demonstrated using an electrophysiologic measurement of Na⁺/Ca²⁺ exchanger current in HEK 293 cells and in COS cells transfected with a vector containing the cDNA encoding the Na⁺/Ca²⁺ exchanger (525). These same cells were also shown to contain Na⁺/Ca²⁺ exchanger protein on the plasmalemmal membrane by examining net Na_o-sensitive Ca²⁺ movement in these cells by confocal [Ca²⁺]_i imaging (525). Importantly, the giant patch-clamp method has been used effectively to examine the functional expression of the Na⁺/Ca²⁺ exchanger genes by electrophysiological methods in Xenopus oocytes that had been injected with cRNA (199, 408, 414, 587, 636).

The canine NCX1 cDNA open reading frame is composed of a sequence of 2,910 nucleotides from the full-length canine cardiac Na⁺/Ca²⁺ exchanger clone. This provides the primary information from which to deduce the 970 amino acids that make up the protein (688). This protein (Fig. 17) has a deduced (calculated) molecular mass of 108 kDa. Although there are six possible N-linked glycosylation sites, a single site is glycosylated in the native dog protein (433).

View larger version (43K): [in this window] [in a new window]   Fig. 17. Primary amino acid sequence of canine Na+/Ca2+ exchanger protein. Cardiac Na+/Ca2+ exchanger gene was cloned (see text) from a dog heart cDNA library. Following a signal sequence, there are 11 transmembrane-spanning sequences (TM1TM11). A large intracellular loop follows first 5 transmembrane sequences and contains a "XIP" sequence at its beginning (see text and Fig. 18). Same overall organization of Na+/Ca2+ exchanger protein is found in other isoforms of NCX1 gene and in products of two other Na+/Ca2+ exchanger genes, NCX2 and NCX3. [From Nicoll et al. (688).]

The intracellular loop is thought to be involved in modulation of the Na⁺/Ca²⁺ exchanger function not only by intracellular kinases, but also by intracellular ions such as Ca²⁺, Na⁺, and H⁺, and perhaps other factors The entire intracellular loop can be deleted, leaving only sufficient residual amino acids to permit a connection between the region of five transmembrane crossings of the protein (at the amino-terminal end) and the region of six transmembrane crossings (at the carboxy-terminal end). This shortened protein is, nevertheless, a functional Na⁺/Ca²⁺ exchanger, as assessed by the giant membrane patch method (Fig. 14B; see Ref. 635). This result supports the notion that the intracellular loop region does not play a direct role in ion translocation. The treatment of giant patches taken from native cardiac myocyte preparations with the protease -chymotrypsin, which "derepresses" the Na⁺/Ca²⁺ exchanger (406, 632), is thought to operate much like the loop-deletion mutation described above. These findings have led to speculation that modulation of the Na⁺/Ca²⁺ exchanger by intracellular protons (266, 267), and Ca²⁺ and Na⁺ (632), may depend on the intracellular loop.

The Ca_i-dependent modulation of the Na⁺/Ca²⁺ exchanger has been investigated at the molecular level (578, 579). Using a ⁴⁵Ca²⁺ "overlay" method, Levitsky et al. (579) identified the region of high-affinity Ca²⁺ binding as the segment between amino acids 371 and 508; this is in the middle of the intracellular loop region of the protein. Within this span there are two highly acidic sequences, each containing three consecutive aspartic acid residues that, when mutated, lead to a markedly reduced Ca²⁺ affinity. This suggests that the Ca_i-dependent modulation of the Na⁺/Ca²⁺ exchanger kinetics (88, 96, 258, 265, 266, 406, 412, 661) may arise from the binding of Ca²⁺ to this domain of the exchanger protein.

At the 5'-end of the cDNA, a Kosak-like consensus sequence marks the initiation site of the open reading frame. There appears to be an initial leader peptide (M0), followed by five transmembrane regions (M1-M5), a large intracellular loop region, and six transmembrane regions (M6-M11) at the carboxy-terminal end (525, 688, 807). These distinct regions, suggested by hydrophobicity analysis, provide a putative model for the protein that places the amino terminus in the extracellular space and the large loop and the carboxy terminus in the intracellular domain (Fig. 18).

View larger version (33K): [in this window] [in a new window]   Fig. 18. Model of Na+/Ca2+ exchanger protein. Transmembrane-spanning regions are identified (light blue = PM), as are XIP region (red), Na+-inactivation region (green), Ca2+-binding regions (yellow), and alternatively spliced region (dark blue). [Modified from Nicoll et al. (688).]

It may be less surprising that in similar tissues but in different species, there are no clearly important differences in the gene structure for the relevant isoform of NCX1. This can be best appreciated in the heart, where the primary cardiac isoforms of NCX1 from a variety of mammalian species are almost identical, as shown in Figure 19A. In contrast, the signal sequences are only 60-75% identical (Fig. 19B; and see Refs. 284, 570, 839). The significance of the signal sequences is, however, unclear; several groups have shown that the signal peptide is not required to target the exchanger to the PM (330, 605, 807).

View larger version (48K): [in this window] [in a new window]   Fig. 19. A: phylogeny of several Na+/Ca2+ exchanger molecules. Rootless tree was constructed from deduced amino acid identity sequences from various Na+/Ca2+ exchangers, as indicated. [From Schulze et al. (838).] B: comparison of signal sequences from various mammalian Na+/Ca2+ exchangers. Arrow at top right of sequences indicates cleavage position of signal peptide; hydrophobic region is indicated by line above sequences. Charged or polar amino acids are underlined; highlighted amino acids are hydrophobic. Positions in which amino acids in all sequences are identical are indicated by an asterisk at bottom. Positions in which all of sequences have charged or hydrophobic amino acids are indicated by C or H, respectively, at bottom. [From Schulze and Lederer (839).]

The NCX1 gene from which this general structure was first identified (688) appears to be the model for all expressed Na⁺/Ca²⁺ exchanger genes except that of the vertebrate rod (NCKX1) and possibly the platelet (see above). The human NCX1 gene is found on chromosome 2p22-p23 (641, 848), but the locations of NCX2 and NCX3 are not yet known.

Consensus phosphorylation sites have been identified that suggest the Na⁺/Ca²⁺ exchanger may be a target for PKA and/or one of the members of the PKC family (D. Schulze, personal communication). Growing evidence of this expected modulation of the Na⁺/Ca²⁺ exchanger by phosphorylation/dephosphorylation is available (99, 447, 448, 873, 874; see sect. IIIE). However, a variety of seemingly inconsistent and conflicting physiological results have been obtained following PKA or PKC activation (for references, see sect. IIIE). The site(s) of this putative phosphorylation remains to be identified but is expected to be on the intracellular loop.

Calcium metabolism in different tissues is remarkably disparate. In the heart, for example, [Ca²⁺]_i transiently rises to a peak of 1-3 µM in 10-20 ms and declines with a t_1/2 of ~150 ms (146, 812, 977). In contrast, the [Ca²⁺]_i transients in synaptic boutons may be larger and faster than those in the heart; during a nerve action potential, [Ca²⁺]_i may rise to a peak of ~100 µM or more within a few hundred microseconds and may then decline with a t_1/2 in the range of a few milliseconds (877). A greater contrast is found in the connecting tubule of the nephron, where [Ca²⁺]_i is presumed to be quite constant over time.

These three tissues (heart, nerve, and kidney) are distinctly different in their Ca²⁺ signaling and in the time courses of their [Ca²⁺]_i transients, but they are similar in that each possesses a great deal of Na⁺/Ca²⁺ exchanger protein (318, 465, 501, 610, 613, 719, 746) and RNA from the NCX1 gene (525, 526). Because the Na⁺/Ca²⁺ exchanger is so well expressed in these tissues, and yet, Ca²⁺ signaling is so different, the exchanger itself might be expected to display large differences in the primary cDNA sequence as well as functional differences. This is not the case, however. Instead, there are tissue-specific isoforms (see below) that, with respect to the coded sequence, express at least 32 distinctive isoforms of NCX1 (838). The alternatively spliced NCX1 variants arise from diversity in a limited region of the intracellular loop (527). Thusfar, however, no distinctive functions of the different isoforms have been observed, although a brief report has suggested that the cardiac isoform, but not the renal isoform, is activated by PKA (394). Perhaps the alternatively spliced isoforms are important because of spatial targeting that permits the Na⁺/Ca²⁺ exchanger to be placed at an appropriate subcellular site (see sect. VA, below, showing the immunofluorescent localization). These possibilities still need to be examined experimentally.

The intron/exon structure of the gene NCX1 is presented in Figure 20 (527). The entire gene spans more than 200 kb (527, 532, 533, 826, 838) and includes a 5'-end alternatively spliced area before the start of the open reading frame (570). This region presumably plays a role in regulating expression of the different isoforms, although the mechanism of this hypothesized action is unknown. The major alternatively spliced region in the open reading frame is found in the putative intracellular loop region of the protein (see below and Fig. 18). Following the stop codon there are conserved sequences across isoforms and species (532, 533). This conservation of sequence beyond the stop signal presumably influences and regulates expression, but the mechanism of action is not known.