HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Laporte, R. Articles by Laher, I. Search for Related Content PubMed PubMed Citation Articles by Laporte, R. Articles by Laher, I.

Article

Pharmacological Modulation of Sarcoplasmic Reticulum Function in Smooth Muscle

Ferring Research Institute, Inc., Ferring Pharmaceuticals, San Diego, California (R.L.); and Department of Pharmacology and Therapeutics, University of British Columbia, Vancouver, Canada (A.H., I.L.)

Abstract

I. Smooth Muscle Cell Ca2+ Handling and Role of the Sarcoplasmic Reticulum

A. Ca2+ Handling

B. Sarcoplasmic Reticulum

C. Capacitative Ca2+ Entry

D. Sarcoplasmic Reticulum Fractions and Interaction with Mitochondria

E. Ca2+ Storage by the Sarcoplasmic Reticulum

F. Estimates of Ca2+ Content in the Sarcoplasmic Reticulum

G. Ca2+ Uptake and Release by the Sarcoplasmic Reticulum

II. Physiological and Pharmacological Agents

A. Ca2+ Pump (Sarco/Endoplasmic Reticulum Ca2+-ATPase)

1. Thapsigargin.

2. Cyclopiazonic Acid.

3. 2,5-Di-(tert-butyl)-1,4-benzohydroquinone.

B. Ca2+-Gated Ca2+ Release Channel/Ryanodine Receptor

1. Cyclic ADP-Ribose and Analogs.

2. Caffeine and 9-Methyl-7-bromoeudistomin D.

3. Ryanodine.

4. Procaine.

5. Ruthenium Red.

C. Inositol 1,4,5-Triphospate-Gated Ca2+ Release Channel/Inositol 1,4,5-Triphospate Receptor

1. Inositol 1,4,5-Triphosphate.

2. Adenophostins.

3. Xestospongins.

4. 2-Aminoethoxy-Diphenylborate.

5. mAb18A10 and Other Neutralizing Anti-Inositol 1,4,5-Triphosphate Receptor Antibodies.

6. Heparin.

III. Conclusions and Perspectives

The sarco/endoplasmic reticulum (SR/ER) is the primary storage and release site of intracellular calcium (Ca²⁺) in many excitable cells. The SR is a tubular network, which in smooth muscle (SM) cells distributes close to cellular periphery (superficial SR) and in deeper aspects of the cell (deep SR). Recent attention has focused on the regulation of cell function by the superficial SR, which can act as a buffer and also as a regulator of membrane channels and transporters. Ca²⁺ is released from the SR via two types of ionic channels [ryanodine- and inositol 1,4,5-trisphosphate-gated], whereas accumulation from thecytoplasm occurs exclusively by an energy-dependent sarco-endoplasmic reticulum Ca²⁺-ATPase pump (SERCA). Within the SR, Ca²⁺ is bound to various storage proteins. Emerging evidence also suggests that the perinuclear portion of the SR may play an important role in nuclear transcription. In this review, we detail the pharmacology of agents that alter the functions of Ca²⁺ release channels and of SERCA. We describe their use and selectivity and indicate the concentrations used in investigating various SM preparations. Important aspects of cell regulation and excitation-contractile activity coupling in SM have been uncovered through the use of such activators and inhibitors of processes that determine SR function. Likewise, they were instrumental in the recent finding of an interaction of the SR with other cellular organelles such as mitochondria. Thus, an appreciation of the pharmacology and selectivity of agents that interfere with SR function in SM has greatly assisted in unveiling the multifaceted nature of the SR.

I. Smooth Muscle Cell Ca²⁺ Handling and Role of the Sarcoplasmic Reticulum

A. Ca²⁺ Handling

The most abundant cation in the vertebrate body is calcium (Ca²⁺), where in humans it amounts to 20 to 30 g/kg body weight. There are large reservoirs of Ca²⁺ in the form of depots in bone, which are available to the body for cellular processes. Despite the availability of such a large, extracellular source of Ca²⁺, cells have an organized internal store of Ca²⁺ that is readily available for rapid release as needed upon membrane excitation. The identification and universal acceptance of the SR¹ as a Ca²⁺ store and sink is a relatively late event in smooth muscle research and begs the question of its utility in establishing a maintained cellular response in the face of a very large, inwardly facing gradient of accessible Ca²⁺ that is of infinite abundance relative to enzymatic requirements.

As Pozzan et al. (1994) and others have argued, the diffusion of Ca²⁺ within the cell is not an unimpeded process, so the presence of various immobile binding sites for Ca²⁺ imposes severe constraints on its ability to reach a rapid and useful cellular concentration at critical sites within the cell, especially in the case of cells with relatively large volumes. The intracellular diffusion coefficient for Ca²⁺ in water is 7 x 10^-6 cm²/s, which is more than 10 times that in cytoplasmic extracts containing intracellular Ca²⁺ buffers (Allbritton et al., 1992). This impediment to the movement of Ca²⁺, coupled with the presence of intracellular structures close to the plasma membrane, ensures local areas of high Ca²⁺ concentrations, which in turn may lead to a regenerative release of Ca²⁺ from the SR. In careful mathematical modeling of Ca²⁺ movement in cells using physiological constraints, Kargacin (1994) calculated that the concentration of Ca²⁺ below the plasma membrane reaches 8 µM during the Ca²⁺-induced Ca²⁺ release (CICR) process.

The release of Ca²⁺ from internal stores ensures a more diffuse and timely increase in cytoplasmic Ca²⁺ that is coordinated with the ensuing entry of extracellular Ca²⁺. Through such an intricate dependence of the timely movement of Ca²⁺ from within the cell and from extracellular sources, the cell is able to initiate Ca²⁺ oscillators that act as regional "switches" and "relays" within cytoplasmic domains (Bootman et al., 2002). For instance, in the smooth muscle cell, Ca²⁺-mediated contractile regulation by extracellular Ca²⁺ influx and/or SR Ca²⁺ release could take place principally by two mechanisms. The first is by direct activation of the contractile apparatus through global cytoplasmic free Ca²⁺ concentration ([Ca²⁺]_cyt) increase throughout the cytoplasm (Sanders, 2001). A second pathway is via indirect regulation of plasma membrane excitability by an increase in [Ca²⁺]_cyt spatially localized to a narrow gap (20-40 nm in depth) between the plasma membrane and the superficially located SR, termed the plasma membrane-SR junctional space (Lee et al., 2002a) (see Section I.A.) (Fig. 1). [Ca²⁺]_cyt increase in this junctional space may also be the source of cell-wide [Ca²⁺]_cyt oscillations and waves coupled either to contractile activity or relaxation (Pabelick et al., 2001b; Lee et al., 2002a).

View larger version (49K): [in this window] [in a new window]   FIG. 1. Schematic transversal section of a smooth muscle cell illustrating the main topological relationships between SR and key ultrastructural elements. Sarcolemma = plasma membrane, subsarcolemmal = subplasmalemmal.

The SR has many structural features of the endoplasmic reticulum (ER), sharing many intracellular chaper-one proteins and other histological features. The existence of a SERCA and of Ca²⁺ release channels reveals that the SR is, in fact, a region of the ER that specializes in Ca²⁺ homeostasis, predominantly Ca²⁺ release and uptake. Thus, the SR is in reality an intracellular networking system whereby different cytoplasmic domains are connected in an intercommunicating and interdependent manner. In reality, the SR and ER are constantly multitasking—some portions of the structures are involved in protein trafficking, while other portions are being earmarked for Ca²⁺ signaling (Berridge, 2002). These portions of the SR/ER are constantly remodeling, with this process being driven by stability of the proteins and the Ca²⁺ load (Berridge, 2002). The SR is also contiguous with the nuclear envelope and thus may play a role in Ca²⁺-dependent gene regulation via Ca²⁺-dependent transcription factors such as cAMP responsive element-binding proteins (CREB) and nuclear factor for activated T cells (NFAT) (Cartin et al., 2000; Gomez et al., 2003; Hill-Eubanks et al., 2003). InsP₃R are also found in the SR membrane and may regulate the opening of the nuclear pores (Stehno-Bittel et al., 1995).

The release and accumulation of intracellular Ca²⁺ regulates many aspects of cellular functions, including hormone and neurotransmitter release, endothelial secretion, muscle contractile activity, cell division, growth and migration, and apoptosis. It is, in fact, difficult to imagine an aspect of mammalian cell function that is not in some way regulated by the availability of Ca²⁺. With this in mind, it becomes apparent why there has been a burgeoning global interest in understanding how cells regulate all aspects of Ca²⁺ availability. Our interest in cellular Ca²⁺ handling owes much to the seminal observations of Ringer in 1883, who first demonstrated the absolute need for Ca²⁺ in muscle excitation-contractile activity coupling. The impetus for the search for drugs that modify the cellular responses to Ca²⁺ comes largely from the need to treat a variety of diseases thought to involve abnormal handling of Ca²⁺. Thus, an extensive array of pharmacological tools altering Ca²⁺ entry, release, sensitization, and extrusion has been created. In this review, we summarize the pharmacology of drugs frequently used to modulate intracellular Ca^{2 +} sequestration and release in smooth muscle.

Smooth muscle cells are spindle-shaped, with the widest part being 2 to 5 µm, and the length ranging up to 500 µm in visceral muscle and to 150 µm in vascular tissues. The plasma membrane and SR come into close contact with the SR membrane running parallel with the plasma membrane for distances of 1 µm or more (Devine et al., 1972; Gabella, 1983). The geometrical shape of smooth muscle cells ensures a disproportionate ratio of membrane (PM, sarcoplasmic membrane) surface/cytoplasm (cytosol, sarcoplasm) ratio. For example, visceral smooth muscle cells have a volume of 2500 to 3000 µm³ and a cell surface area of 5000 µm³ (not allowing for caveolae), producing an approximated cell surface/volume ratio of 1.5 µm^-1, which is equivalent to that of erythrocytes. For a comprehensive review of the ultrastructural features of smooth muscle, see the overview by Gabella (1983).

In vascular smooth muscle cells, there are 2.7 µm² of cell surface for every cubic micrometer of cell volume. This large cell surface/cell volume ratio in smooth muscle favors exquisite regulation of cell surface processes by instantaneous changes in intracellular composition, such as the regulation by (presumably) spontaneously released Ca²⁺ from the SR, known as "Ca²⁺ sparks", of Ca²⁺-activated K⁺ (K_Ca) (see reviews by Nelson et al., 1995; Jaggar et al., 2000; Wellman and Nelson, 2003) and chloride (Cl_Ca) channels (Kotlikoff and Wang, 1998). Earlier studies by Benham and Bolton (1986) and by Stehno-Bittel and Sturek (1992) lead to the suggestion that the frequently observed spontaneous K⁺ currents recorded in smooth muscle (Benham and Bolton, 1986; Ohya et al., 1987; Desilets et al., 1989; Hume and Leblanc, 1989) occur in regions of the cell where the SR and the PM are closely apposed. This Ca²⁺ release from the SR, presumed to occur spontaneously, occurs in close proximity to K_Ca channels and reaches a [Ca²⁺]_cyt of 10 to 100 µM and an average size of 13 µm², which covers 1% of the 1300 µm² of the smooth muscle membrane (Perez et al., 1999). This release of Ca²⁺ occurs via Ca²⁺-gated channel/ryanodine receptor (RyR) channels, likely the RyR2 subtype with an ancillary role for RyR3 (Lohn et al., 2001). However, the global change in cytoplasmic Ca²⁺ due to the spontaneous release of Ca²⁺ sparks from the SR is less than 2 nM (Jaggar et al., 2000). Ca²⁺ sparks originating from the SR occur with a relatively low frequency of 1 Hz (allowing for a tonic hyperpolarization throughout the electrically coupled smooth muscle), and the spreading distance in smooth muscle is 1.0 to 2.5 µm (Jaggar et al., 2000). Thus, there is the appearance of a specialized subsarcolemmal signaling space where high local concentrations of Ca²⁺ 10 µM) exist in microdomains without significant impact on global cytosolic Ca²⁺.

B. Sarcoplasmic Reticulum

Unlike other membrane systems, such as the mitochondrial inner membrane, there is no potential difference across the SR membrane. The concentration of Ca²⁺ in the extracellular space is between 1 and 10 mM, whereas the [Ca²⁺]_cyt in the cytoplasm is in the order of 0.1 µM, thus creating a large inwardly directed electrochemical gradient forcing Ca²⁺ entry across the plasma membrane. The cell has several mechanisms for maintaining a low [Ca²⁺]_cyt, which at the same time also ensures that the appropriate transient peak levels of Ca²⁺ are reached during activation. Notable among these are active processes such as Ca²⁺ extrusion across the plasma membrane into the extracellular reservoir by the plasma membrane Ca²⁺-ATPase pump (PMCA) and also accumulation of the ion into the SR by the SERCA. Although these two pumps essentially accomplish the same effect of rapidly reducing [Ca²⁺]_cyt levels, they have different physicochemical properties and regulatory mechanisms (Grover and Khan, 1992; Raeymaekers and Wuytack, 1993). Using indirect methods in a large vein, Nazer and van Breemen (1998) concluded that nearly half the cytoplasmic free Ca²⁺ load is extruded via the PMCA, with an equal role for the SERCA and the plasma membrane Na⁺-Ca²⁺ exchanger in removal of the remainder. Much is known about the molecular identity of the Na⁺-Ca²⁺ exchanger in smooth muscle (Nakasaki et al., 1993; Juhaszova et al., 1996), although details of its isoform distribution and functionality awaits further characterization.

Some investigators have suggested that the plasma membrane pathways (PMCA and Na⁺-Ca²⁺ exchanger) account for only 20 to 40% of Ca²⁺ removal (Cooney et al., 1991), whereas others have concluded that the PMCA removes only 10 to 20% of Ca²⁺ from the cell (Kargacin and Fay, 1991). However, in a resistance artery from the brain, Kamishima and McCarron (1998) proposed roles only for the Ca-ATPase pumps (SERCA and PMCA), and not the Na⁺-Ca²⁺ exchanger, in removal of free Ca²⁺ from the cytoplasm. In keeping with this, Kargacin and Kargacin (1995) proposed that the SERCA pumps are likely to make the largest single contribution to Ca²⁺ removal and can reduce Ca²⁺ at a rate of 60 to 80% the rate of Ca²⁺ removal seen in cells during a Ca²⁺ transient. Thus, there exists some uncertainty about the precise roles and relative importance of the extrusion mechanisms for cytoplasmic Ca²⁺. In part, this is related to the lack of specific inhibitors of the various processes that govern Ca²⁺ homeostasis. It is likely that many of the quantitative estimates described in the literature regarding Ca²⁺ pool sizes, diffusion rates, and other physicochemical parameters exhibit variations in estimates due largely to differences in methods, species, and tissues.

An important initial step in elucidating the contribution and role of the various modalities of Ca²⁺ extrusion mechanisms is the description of caloxin, a peptide that inhibits Ca²⁺ extrusion by the PMCA (Chaudhary et al., 2001; Holmes et al., 2003). The search for a selective and specific inhibitor for the PMCA is considered by many to be enigmatic; the limitation of caloxin [IC₅₀ value 0.4-1 mM in red cell leaky ghosts, which mainly express PMCA4 (Holmes et al., 2003)] is that it is a peptide and thus not amenable to routine use. However, it produces an endothelium-dependent relaxation in intact rings of rat aorta (0.34 mM) (Chaudhary et al., 2001). Caloxin does not inhibit Mg²⁺-ATPase or Na⁺/K⁺-ATPase (Holmes et al., 2003) and has been shown to also inhibit PMCA activity in human mesenchymal stem cells (Kawano et al., 2003) and human embryonic kidney (HEK) cells (De Luisi and Hofer, 2003). A third and relatively novel family of Ca²⁺ pumps is the Ca²⁺/Mn²⁺ ATPases, which occur predominantly in the Golgi compartment of eukaryotic cells (reviewed by Wuytack et al., 2002, 2003).

The rise of intracellular Ca²⁺ is vital to cell function, but Ca²⁺ availability must occur rapidly and in sufficient concentration at required intracellular targets. This is accomplished by Ca²⁺ release from the SR either as a regenerative release of Ca²⁺ via CICR occurring through activation of RyR, or as InsP₃-induced Ca²⁺ release (IICR, occurring through activation of InsP₃R/Ca²⁺ release channel) (see Section I.G.2.). The RyR was originally shown to bind [³H]ryanodine (see Section II.B.3.a.), an agent that was then known to alter SR Ca²⁺ release events in skeletal muscle. Within the SR, a fraction of Ca²⁺ is bound to various Ca²⁺ storage proteins such as calsequestrin. (see Section I.E.). Once released, the diffusion rate of Ca²⁺ in the cytoplasm is limited by the presence of various high-affinity binding proteins, so that the rate of Ca²⁺ diffusion in the cytoplasm (10-100 µm²/s) is 3 to 30% the rate in free solution, which occurs at 320 µm²/s (Tsien and Tsien, 1990).

The description of an intracellular network of membrane systems that was later to be named SR was first described in skeletal muscle more than 100 years ago (Veratti, 1902). It was appreciated that muscle had an intrinsic "relaxing activity", which was later ascribed to the ability of SR membranes to accumulate Ca²⁺ at the expense of ATP hydrolysis (Ebashi and Lipmann, 1962). The SR is a system of anastomosing intracellular membranes organized into tubules that occupy between 1.5 to 7.5% of the cell volume, with greater volumes being present for instance in smooth muscle cells from large, conduit type arteries (Devine et al., 1972). This distribution of SR and its implications in arteries of varying sizes has not been revisited since the initial description by Devine et al.; it is likely that these estimates may be revised with the use of more modern techniques and the patterns of SR distribution better defined with regard to cell structure and colocalization with cellular elements. This represents a daunting task since we now know that the complex architecture of the SR is made functionally complicated by the presence of regions that specialize in Ca²⁺ uptake and release, with other adjoining parts involved in the assembly or degradation of proteins. The variable volume of SR in smooth muscle cells may also be a reflection of cell synthetic activity, so that large arteries, which synthesize more extracellular proteins than smaller diameter arteries, have a correspondingly larger SR volume (Somlyo, 1980), although this supposition has not been rigorously examined. Also, the SR volume of smooth muscle cells, at least in the uterus, is increased by estrogen and during pregnancy (Shoenberg, 1958; Ross and Klebanoff, 1971). Thus, proliferative, developing smooth muscle cells tend to have more SR (Campbell et al., 1971), as is the case for some injured and hypertensive vascular smooth muscle cells (Raeymaekers and Wuytack, 1993).

The SR contains several ionic species and of note is the following: 1) the Na⁺ and Cl^- concentrations are similar to those present in the cytoplasm, indicating that the SR is not effectively in contact with the extracellular space, 2) the Ca²⁺ concentration in the SR measured by electron probe analysis is 30 to 50 mmol/kg dry weight, and 3) the Ca²⁺ concentration near the plasma membrane is not uniformly distributed, with areas of low (1 mmol/kg dry weight) and high ("hot spots") concentrations. There is a concordance of junctional elements of the SR (those portions close to the plasma membrane) and areas of Ca²⁺ hot spots. It is estimated that the average number of SR elements indicated by hot spots that lie within 50 nm from the plasma membrane is between 3 and 5 per cell (Bond et al., 1984b). One consequence of this is that sufficiently high local concentrations (in mM) of Ca²⁺ are reached near the plasma membrane, allowing for local, intracellular regulation of K⁺ (Nelson and Quayle, 1995) and Ca²⁺ channels (Huang et al., 1989) in smooth muscle.

That the SR acts as a sink for Ca²⁺, i.e., can actively accumulate Ca²⁺, was initially demonstrated by Somlyo and Somlyo (1971). These investigators took advantage of the fact that the SR can accumulate Sr²⁺ and Ca²⁺ by the same transport mechanisms so that the SR became loaded to a greater extent with Sr²⁺ after incubation in a depolarizing solution. Based on the use of electron probe analysis, Somlyo and Somlyo (1971) estimated that the Ca²⁺ content of the central SR increases 3- to 4-fold following depolarization of smooth muscle. When Casteels and Droogmans (1981) used ⁴⁵Ca²⁺ to determine the content of the SR (junctional, central), they estimated it to be 60 µM/kg wet weight. Using the same technique, they also showed that 1) the maximal rate of filling of Ca²⁺ under normal conditions of external ionic composition (1.5 mM Ca²⁺) is nearly 70 µmol/kg, and 2) the permeability of Ca²⁺ across the plasma membrane is regulated by the extent of filling of the agonist-sensitive pool of SR Ca²⁺ (Casteels and Droogmans, 1981). This was the forerunner of what would later be described as "capacitative Ca²⁺ entry", as described by Putney (1986).

C. Capacitative Ca²⁺ Entry

In addition to functioning as a store and a sink for Ca²⁺, the SR is also at the origin of two cell signaling processes: 1) the SR generates Ca²⁺ sparks that, in smooth muscle cells, regulate the plasma membrane electrical potential through modulation of K_Ca channels (Nelson et al., 1995; Jaggar et al., 2000), and 2) the content of the SR determines entry of extracellular Ca²⁺ through "store-operated calcium entry" (Casteels and Droogmans, 1981) or "capacitative calcium entry (CCE)" (Putney, 1986; Putney et al., 2001). Thus, both Ca²⁺ sparks and CCE act to regulate Ca²⁺ entry, albeit through separate but related mechanisms.

The designation of CCE is by analogy with an electrical capacitor; a charged or full intracellular store prevents Ca²⁺ entry through this pathway, whereas a discharged or empty store facilitates Ca²⁺ entry and refilling of the store. The CCE model thus proposes that when the SR store is stimulated to discharge Ca²⁺, either by InsP₃-generating signal such as stimulation of G-protein coupled (G_q/G₁₁-) receptors or by receptor/InsP₃-independent means, such as through SERCA inhibition, there is a fall in the Ca²⁺ content of the SR, which then signals a novel Ca²⁺ pathway on the plasma membrane so that refilling of the store can occur. The refilling is rapid and allows for a constant internal store of Ca²⁺ to be available for smooth muscle oscillations and maintained tone. The pharmacology of CCE is in its infancy, with agents of generally low specificity being used (Putney et al., 2001). Notable among these are [with attendant K_i values (µM)]: Gd³⁺, <1; econozole, 2-10; miconazole, 1.0; flufenamic acid, 33; eicosatetraynoic acid, 4.0; SK&F 96365, 3-50; 2-aminoethoxydiphenylborate (2-APB), 30; LU52396, 2.0; and L-651,582, 1.2 (see Putney et al., 2001 for complete list). The lack of a specific inhibitor stems from a paucity of information regarding two key aspects of the model for CCE: 1) the molecular nature of the signaling between a depleted SR and the plasma membrane, and 2) the specific details of the membrane events (ion channels, receptor mechanisms) activated during the process. Not-withstanding these limitations, there has been progress made in unraveling details regarding CCE, and a clearer picture is starting to emerge, at least in some cell types.

In isolated portal vein smooth muscle cells, norepinephrine (NE) causes a transient increase in Ca²⁺ due to SR activation; this is rapidly followed by a more sustained increase in Ca²⁺ due to entry from extracellular sources (Pacaud et al., 1993). In rat aorta, inhibition of SERCA with low concentrations (1 µM) of cyclopiazonic acid (CPA) increases [Ca²⁺]_cyt without causing contractile activity, suggesting that increases in [Ca²⁺]_cyt in noncontractile compartments are being affected by CPA (Tosun et al., 1998). Low concentrations of CPA induce increases in [Ca²⁺]_cyt that are similar to those produced by KCl, but unlike the latter, fail to induce contraction. On the other hand, higher concentrations of CPA (10-20 µM) appear to increase [Ca²⁺]_cyt beyond the noncontractile compartment. Of interest is that the increase in [Ca²⁺]_cyt produced by low concentrations of CPA is insensitive to L-type voltage-gated Ca²⁺ channel (Ca_L) blockers, such as verapamil, but was inhibited by Ni²⁺ (Tosun et al., 1998). Vasodilation in rabbit and mouse aorta by nitric oxide (NO) has been linked to inhibition of CCE, possibly by the effect of NO in rapidly filling the SR store via stimulation of SERCA (Cohen et al., 1999).

A unique but poorly characterized current, Ca²⁺ release-activated Ca²⁺ current (I_CRAC), characterizes the depletion-activated entry of Ca²⁺. Hoth and Penner (1993) first identified a Ca²⁺ current activated by store depletion (in mast cells) and several aspects of this current were noted: It had a very low conductance (0.02 pS), high Ca²⁺ selectivity, inward rectification, inhibition by intracellular Ca²⁺, blockade by Ni⁺ and Cd²⁺, and lack of voltage-dependent gating (Hoth and Penner, 1993). A current with similar characteristics has also been reported in the mouse anococcygeus muscle, where it is termed I_DOC and has a unitary conductance of 10 pS (Wayman et al., 1998). CCE has been described in a number of smooth muscles (Ito et al., 2000; Weirich et al., 2001; Young et al., 2001); and these findings have been reviewed recently (Gibson et al., 1998; Albert and Large, 2002; McFadzean and Gibson, 2002). There is some evidence to support the notion that store depletion activates tyrosine kinases to signal Ca²⁺ entry in a number of cell types including smooth muscle (Pacaud and Bolton, 1991a; Wijetunge et al., 1992; Doi et al., 2000) and endothelial cells (Jacob, 1990; Sharma and Davis, 1996). In rat aortic smooth muscle cells, depletion of the SR with thapsigargin or bradykinin stimulates phospholipase D to generate phosphatidic acid, which enhances sustained Ca²⁺ entry (Walter et al., 2000). Pharmacological evidence in support of CCE in maintaining the basal tone in resistance arteries has been documented in the rat cremaster arterioles (Potocnik and Hill, 2001). Whereas the CCE was not altered by disruption of the cytoskeleton in cremaster arterioles (Potocnik and Hill, 2001), there is marked inhibition of CCE when actin filaments are disrupted in endothelial cells (Bishara et al., 2002).

Although the molecular identity of the channel responsible for CCE remains elusive, there is persuasive evidence that it has considerable homology with Drosophila transient receptor potential channel (TRP) proteins, which are involved in phototransduction in the fruit fly where they mediate CCE. TRP channels are a large family (at least 20 genes) of plasma membrane, nonselective cationic channels. This latter feature of being nonselective makes them depolarizing agents, and their Ca²⁺ permeability suggests a role in intracellular Ca²⁺ signaling. The TRP channels have six transmembrane segments and are activated by products of G-protein coupled receptor (GPCR) stimulation. The properties of TRP have recently been reviewed (Zhu and Birnbaumer, 1998; Nilius, 2003). TRP shares sequence homology with voltage- and second messenger-gated Ca²⁺ channels, and this homology has a long evolutionary history, being also present in Caenorhabditis elegans. However, a concern is that TRP channels are not highly selective for Ca²⁺, and the available evidence indicates that none of the TRP proteins induce a conductance with the known properties of CCE (Vennekens et al., 2002). The TRP(C3) channel may be directly linked to InsP₃R activation, whereby InsP₃R activation leads to entry of extracellular Ca²⁺ (Boulay et al., 1999). It is likely then that CCE is not mediated by a single TRP protein but is possibly occurring through a channel formed by a multimeric structure containing various combinations of TRP and TRP-like proteins (Montell, 1997). The vagaries of the CCE have found detractors who present evidence that key elements of store-operated Ca²⁺ entry are incompatible with generating oscillatory [Ca²⁺]_cyt signals (Shuttleworth, 1999) or that depletion of the internal stores for Ca²⁺ does not always lead to CCE (Haller et al., 1996).

D. Sarcoplasmic Reticulum Fractions and Interaction with Mitochondria

Popescu and Diculescu (1975) segregated smooth muscle SR into three regions: peripheral, deep, and central. In general, the peripheral SR is located close to the plasma membrane and sometimes in apposition with the caveolae (equivalent to junctional SR described above). This SR element is in contact with the deep SR positioned near the myofilaments and is in continuity with the central SR deeper within the cell and associated with the nuclear membrane (Forbes et al., 1977, 1979). Notable details of the analogy of smooth muscle SR to that of striated muscle have been painstakingly pointed out by Forbes et al. (1979) who described as "peripheral SR" the collection of saccules, tubules, and cisternae lying in close apposition (gap of 10-20 nm) to the inner plasmalemmal surface. One functional implication of the peripheral SR is that it dampens the impact of the basal Ca²⁺ entry by acting as a "superficial buffer barrier" (SBB), which then reinforces this buffering capacity by causing a vectorial extrusion of Ca²⁺ to the extracellular space (van Breemen et al., 1995) (Fig. 2). Another implication of the close apposition of the SR and plasma membrane is that sufficiently high concentrations of local are reached to activate the Na⁺-Ca²⁺ exchanger and also to activate spontaneous outward K⁺ currents. This was recorded by Tomita and Bulbring (1969) and later visualized and described in detail by Nelson et al. (1995).

View larger version (18K): [in this window] [in a new window]   FIG. 2. Preferential sequestration by the peripheral SR of the extracellular Ca2+ entering the cell in the restricted subsarcolemmal space as proposed in the superficial buffer barrier model. 1, extracellular Ca2+ enters the vascular smooth muscle cells through voltage-dependent or -independent Ca2+ channels. 2, it accumulates in the restricted subsarcolemmal space. 3, it is taken up by the peripheral SR through its SERCA. Sarcolemma = plasma membrane, subsarcolemmal = subplasmalemmal, myoplasmic = cytoplasmic.

The concept of the SBB as proposed by van Breemen (see van Breemen et al., 1995) is being challenged. Using more direct approaches, McCarron's data suggests a different model of SR-plasma membrane arrangement that essentially argues against the SBB model. In elegant experiments, Bradley et al. (2002) made two important findings: first, that Ca²⁺ accumulation by the SR proceeds even when Ca²⁺ influx ceases, implying that for this to be possible, the SR and plasma membrane need not necessarily be closely apposed; and second, close apposition of the SR and plasma membrane is instead essential for Ca²⁺ removal by the Na⁺-Ca²⁺ exchanger. Using an array of biophysical constants related to Ca²⁺ diffusion characteristics and microdomain constraints, Kargacin's group present intriguing data suggesting that the rate of Ca²⁺ uptake by the SR is insufficient to significantly alter the dynamics of a Ca²⁺ transient, either in its magnitude or spread (Bazzazi et al., 2003). Immunofluorescence imaging reveals no obvious differences in the density of Ca²⁺ pumps or phospholamban between the peripheral (superficial) SR and deep (central) SR.

In addition to the distinct distribution patterns of SR within the cell, there is also some evidence that SR fractions represent heterogeneous compartments of releasable Ca²⁺. In this regard, Golovina and Blaustein (1997) reported that, in mesenteric arteries, there are two or possibly three functionally distinct compartments: a region (55%) that empties and refills in a InsP₃- and RyR-independent manner (e.g., not affected by CPA and thapsigargin) and a second region (22%) that empties and fills only in response to caffeine (RyR pool), with other regions (16%) being responsive to CPA and caffeine. Consistent with this, there is a portion of the SR that appears to be more richly endowed with InsP₃R than RyR (Wibo and Godfraind, 1994). Other studies propose two stores of Ca²⁺ that are distinguished by their Ca²⁺ refilling sources: the store that expresses only RyR is filled by cytoplasmic Ca²⁺, whereas the store that expresses both RyR and InsP₃R is filled by extracellular Ca²⁺ (Flynn et al., 2001). In contrast to these studies, Itoh et al. (1983) provided compelling evidence that the Ca²⁺ stores associated with IICR (NE-induced) and CICR (caffeine-induced) are identical; additional support for this also comes from the findings of Leijten and van Breemen (1984) and more recently from Bradley et al. (2002), who depleted caffeine-sensitive Ca²⁺ stores through release with flash photolysis of caged InsP₃.

It has been known for some time that mitochondria frequently envelop the central and peripheral portions of the SR (Forbes et al., 1979). Furthermore, mitochondria isolated from smooth muscle cells actively sequester Ca²⁺, as demonstrated by Somlyo and Somlyo (1971). The mitochondria are also, in many instances, located near the caveolae (Somlyo, 1975; Forbes et al., 1979), and this raises the possibility of transfer of ions between these three structures. Some direct evidence for an important regulatory role of mitochondria on intracellular Ca²⁺ homeostasis were provided by Loew et al. (1994), Drummond and Tuft (1999), and McCarron and Muir (1999). Keeping with this role for mitochondrial interaction with the SR and plasma membrane with regard to Ca²⁺ cycling are the recent findings of Kamishima and Quayle (2002) and Szado et al. (2003). Evidence in freshly dissociated vascular (Drummond and Tuft, 1999; Gurney et al., 2000) and visceral (McCarron and Muir, 1999) smooth muscle cells also suggest a functional integration between SR Ca²⁺ release and mitochondrial Ca²⁺ uptake under physiological conditions. For instance, membrane potential, and thus Ca²⁺ entry, can be regulated by mitochondrial Ca²⁺ accumulation, since mitochondrial Ca²⁺ uptake has recently been shown to modulate Ca²⁺ spark activity in isolated rat cerebral artery smooth muscle cells (Cheranov and Jaggar, 2004). Furthermore, new evidence from rat ventricular cardiac muscle cells (permeabilized cells, microsomes, and RyR2 reconstituted into planar lipid bilayers) demonstrated that physiological concentrations of NADH inhibit CICR and that NADH oxidation, likely right at the SR membranes level, is tightly linked to and essential for this effect. This suggests that it is an important physiological negative feedback mechanism, coupling SR Ca²⁺ fluxes and mitochondrial energy production (Cherednichenko et al., 2004). In contrast, there appears to be no role for mitochondria in Ca²⁺ decay following agonist activation, as shown in rat myometrial smooth muscle cells (Shmigol et al., 1999). The dependence of Ca²⁺ entry from the extracellular space on the status of ER Ca²⁺ content and the influence of the mitochondria on this interaction (Parekh, 2003), as well as the transmission of InsP₃-generated Ca²⁺ signals to the mitochondria, have recently been reviewed (Hajnoczky et al., 2000a,b, 2002; Pacher et al., 2000; Szalai et al., 2000; Csordas and Hajnoczky, 2001, 2003).

E. Ca²⁺ Storage by the Sarcoplasmic Reticulum

Once the SR accumulates Ca²⁺, it is loosely bound and available for release; however, the divalent cation does not exist in an ionized form, since this would lead to inhibition of the SERCA, as mentioned above. Hence, there exist within the SR storage proteins capable of binding large quantities of Ca²⁺ in a complex that is readily available for dissociation when Ca²⁺ release is triggered. The total concentration of intracellular Ca²⁺ buffers in smooth muscle cells is estimated to be 200 to 300 µM (Bond et al., 1984a; Carafoli, 1987; Allbritton et al., 1992) or 500 µM (Daub and Ganitkevich, 2000). The storage proteins that are expressed at the highest levels in the SR are calsequestrin and calreticulin. These proteins have high capacity (25-50 mol/mol) and low affinity (1-4 mM) for Ca²⁺. Calsequestrin is the product of two different genes that have a 65% homology; smooth muscle is thought to express both isoforms but in much reduced quantities compared with other muscle types (Pozzan et al., 1994). On the other hand, calreticulin exists in multiple isoforms. There may be some structural similarity between these two Ca²⁺-binding proteins (calsequestrin and calreticulin), since there is antibody cross-reactivity. It is of note that once the SR accumulates Ca²⁺, there is preferential binding to proteins, such as calsequestrin, that are strategically located close to the Ca²⁺ release channels. In particular, calsequestrin is retained within the SR lumen by the presence of discrete string-like molecular anchors (Pozzan et al., 1994). In a detailed study of the vas deferens, Villa et al. (1993) provided intriguing evidence that the Ca²⁺ proteins are not uniformly distributed in the SR. They reported that the peripheral portion of SR is rich in calsequestrin (which, incidentally, is also enriched with InsP₃R), whereas calreticulin was more evenly distributed in the cell. This selective distribution of binding proteins is in agreement with the hot spots for Ca²⁺ stores reported earlier by Bond et al. (1984a). However, by combining immunogold labeling and immunohistochemical studies, Nixon et al. (1994) concluded that calsequestrin is absent in tonic muscles (aorta), although it was located in the superficial regions of cells from tonic smooth muscle (vas deferens), but the InsP₃R distribution was largely in the peripheral SR in both tissue types. There is some evidence that protein kinase C (PKC) can directly reduce the Ca²⁺ storage by thapsigargin-sensitive [i.e., involving SERCA (see Section II.A.1.b.)] mechanisms in cells from the rat aorta. This unloading of the SR content by PKC is suggested to be defective in hypertension (Neusser et al., 1993).

F. Estimates of Ca²⁺ Content in the Sarcoplasmic Reticulum

There is a large variation in the estimate of Ca²⁺ concentration within mammalian intracellular stores, ranging from 1 µM to 5 mM, with these values being subject to the limitations of the method used [e.g., aequorin measures free Ca²⁺ in the SR, whereas electron probe X-ray microanalysis measures total Ca²⁺] and the cell type studied (Meldolesi and Pozzan, 1998). With the advent of more precise technology, other more direct measurements of the Ca²⁺ content of the SR have been made. For example, Ganitkevich and Hirche (1996) determined that the quantity of Ca²⁺ released by acetylcholine (Ach) is 680 attomoles in smooth muscle; this translates to 230 µmoles of total Ca²⁺ per liter of cytoplasmic volume. Assuming that this released Ca²⁺ is derived only from the SR, then the SR content would be 7.5 mM per liter SR. In permeabilized smooth muscle cells where uptake of the SR was monitored with fura-2 acid, it is estimated that the SERCA pumps can remove Ca²⁺ at a rate that is 45 to 75% the rate at which Ca²⁺ is removed from the cytoplasm of intact cells during transient Ca²⁺ signals (Kargacin and Kargacin, 1995). Kargacin and Kargacin (1995) calculated that the SR of a single smooth muscle cell could store more than 10 times the amount of Ca²⁺ required to generate a single transient contractile response.

The use of fluorescent indicators has revolutionized the study and understanding of intracellular Ca²⁺ regulation and has also allowed a more accurate quantification of intraluminal free Ca²⁺ in smooth muscle SR. However, an important limitation of this technique is that frequently these dyes, particularly the acetomethoxy ester derivative-type dyes, are prone to compartmentalization in various other intracellular organelles as well (Takahashi et al., 1999). Thus, Hofer and Schulz (1996) have determined that, at least in fibroblasts, only 88% of the loaded low-affinity Ca²⁺ indicator mag-fura-2 (furaptra) (Martinez-Zaguilan et al., 1998) is restricted to the thapsigargin-sensitive ER. In fact, potential contribution of a confounding signal stemming from its accumulation in mitochondria must be kept in mind (Gurney et al., 2000) (see Section I.D. for discussion of functional significance). Despite this limitation, the indicator has been used to monitor SR free Ca²⁺ changes in several intact smooth muscle preparations (Hirose and Iino, 1994; Sugiyama and Goldman, 1995; Steenbergen and Fay, 1996; Golovina and Blaustein, 1997; Gomez-Viquez et al., 2003), as has been mag-indo-1, another low-affinity Ca²⁺ indicator (Pesco et al., 2001). Indeed, although it is possible to use high-affinity Ca²⁺ indicators such as fluo-3 or fura-2, the use of these lower affinity Ca²⁺ indicators may be more suitable to monitor dynamic changes in Ca²⁺ stores. The use of aequorin targeted to the ER, as initially described Kendall et al. (1992), has been a significant but technically limiting advance (Alvarez and Montero, 2002) and has been use only in limited number of smooth muscle preparations (Szado et al., 2003).

G. Ca²⁺ Uptake and Release by the Sarcoplasmic Reticulum

It has been known for some time that there is a basal leak of Ca²⁺ from the cell: it is likely that under resting physiological conditions, all cells are exposed to some endogenous low-level stimulation by Ca²⁺-releasing agents, of which there are many. This will undoubtedly cause a basal leak of Ca²⁺ from the SR, which has been confirmed in a number of tissues by tracking increases in cytoplasmic Ca²⁺ or decreases in SR Ca²⁺ content after maximal inhibition of SERCA with pharmacological tools. This basal leak in various cell types is in the range of 20 to 200 µM/min (Camello et al., 2002; Lomax et al., 2002), and it was shown to be 22% per min in cultured smooth muscle cells (Missiaen et al., 1996). It is thus clear that, in the majority of cell types, this leak is of sufficient magnitude to deplete the SR in a few minutes and forms the largest source of Ca²⁺ efflux from the SR (Missiaen et al., 1996). To offset it, SERCA act to increase SR luminal Ca²⁺ by removing Ca²⁺ from the cytoplasm. Interestingly, pumping activity does not lead to the buildup of a membrane potential across the SR membrane during Ca²⁺ uptake (neither does it occur during Ca²⁺ release) because of the coupling of an efflux of H⁺ to this uptake (Inesi and Hill, 1983). There is also evidence for an inward movement Cl^- (Pollock et al., 1998) helping to maintain SR electrical neutrality during Ca²⁺ uptake.

Thus, the SR performs several functions that are directly aided by the presence of SERCA: 1) it acts as a reservoir of releasable Ca²⁺ (the concentration of Ca²⁺ in the SR lumen is nearly three times greater than in the cytoplasm), 2) it buffers the Ca²⁺ leak into the cell that is driven by the steep electrochemical gradient according to the superficial buffer barrier hypothesis proposed by van Breemen and colleagues (1995), 3) it sequesters Ca²⁺ to facilitate smooth muscle relaxation, 4) it provides a pool of Ca²⁺ that is ideally located to activate hyperpolarizing currents via stimulation of K_Ca channels, although the spontaneous release of discreet amounts of Ca²⁺ termed Ca²⁺ sparks (Nelson et al., 1995), and 5) it provides Ca²⁺ from InsP₃-sensitive stores to activate Cl_Ca currents and therefore sustain regenerative changes in membrane potential in gastric muscle (Hirst, 1999; Hirst and Edwards, 2001; Hirst et al., 2002).

1. Ca²⁺ Pump (Sarco/Endoplasmic Reticulum Ca²⁺-ATPase). The pumping of Ca²⁺ by the SERCA, which belongs to the P-type ion pumps family, is a cycle of chemical reactions leading to a series of conformational states divided into two main groups, termed E₁ and E₂, which are based on a general model describing P-type ATPases activity (the E₁/E₂ model) (Martonosi, 1996; Moller et al., 1996; Adebanjo et al., 1999; Lee, 2000). The E₁ conformations have a high affinity for Ca²⁺ (K_D = 10^-7 M) and can be phosphorylated by MgATP to form a high-energy phosphorylated intermediate, E₁ P. Furthermore, their Ca²⁺-binding sites are only accessible from the cytoplasm and not from the SR lumen. By contrast, the E₂ conformations have a lower affinity for Ca²⁺ (K_D = 10^-3 M) and can be phosphorylated by inorganic phosphate (P_i), in the absence of Ca²⁺, to form a low energy phosphorylated intermediate (E₂P), and their Ca²⁺ binding sites are only accessible from the SR lumen and not from the cytoplasm.

The pumping cycle (Fig. 3) is thus initiated by the binding of two Ca²⁺ ions to high-affinity binding sites on the cytoplasmic surface of the SERCA that is in an E₁ conformation with high-affinity Ca²⁺ binding sites (E₁·[Ca₂]), which is followed by the binding of a molecule of ATP (MgATP·E₁·[Ca₂]). This leads to autophosphorylation of the enzyme and release of MgADP resulting in an E₁P·[Ca₂] intermediate. The energy released from the high-energy phosphate bond leads to major conformational changes, through hinge-type or sliding motions, affecting the Ca²⁺ binding sites: the conversion from the high-energy E₁P to the low-energy E₂P (E₂P·[Ca₂]). The Ca²⁺ ions are then released to the SR lumen from the now low-affinity Ca²⁺ binding sites (E₂P). The cycle is then terminated by hydrolysis of the bound phosphate (E₂), countertransport of H⁺ ions to maintain electroneutrality of the SR membrane (although this countertransport of H⁺ does not completely balance the charge carried by Ca²⁺, making the Ca²⁺-ATPase electrogenic), and the conversion of E₂ into E₁ to reset the Ca²⁺ pumping cycle. Variants of this model have been proposed (Martonosi, 1996; Moller et al., 1996; Adebanjo et al., 1999; Lee, 2000) as well as an alternative model not based on the E₁/E₂ dichotomy (Jencks, 1992). As Ca²⁺ accumulates within the lumen, the rate of SERCA activity would be expected to slow down due to a negative feedback. To overcome this limitation, the SR is endowed with Ca²⁺ buffering proteins (see above). The stoichiometry of 1ATP:2Ca has been proven only for SERCA1; in tissues such as smooth muscle, there is only indirect evidence from Hill coefficients (Mg ATPase being 1 and 2 for Ca²⁺) for SERCA pump activation (Grover and Samson, 1986).

View larger version (15K): [in this window] [in a new window]   FIG. 3. SERCA pumping cycle. (): occlusion. See text for details.

The SERCA are 110-kDa proteins and are encoded by three genes: SERCA1 is largely expressed in fast-twitch skeletal muscle, with 1a in adult and 1b in neonatal muscle, SERCA2a is mainly in cardiac and slow-twitch muscle, and SERCA2b is mainly in smooth muscle and most nonmuscle cells, whereas SERCA3 is present on a more widespread basis (East, 2000; Sorrentino and Rizzuto, 2001). Smooth muscle mainly expresses the SERCA2b isoform (>70%), with the SERCA2a and SERCA3 isoforms forming the remainder of the SERCA population (Lytton et al., 1989; Wuytack et al., 1989; Eggermont et al., 1990; Amrani et al., 1995; Trepakova et al., 2000; Wu et al., 2001). Both isoforms are splice variants transcribed from the SERCA2 gene (Wu and Lytton, 1993). The SERCA2b has a higher affinity for Ca²⁺ than SERCA2a (Verboomen et al., 1992) and a lower turnover rate for both Ca²⁺ transport and ATP hydrolysis (Lytton et al., 1992). The maximal rate of Ca²⁺ uptake by smooth muscle SR (loaded with oxalate to create linear kinetics) is 100 nmol/kg/min, which is lower (by two times) than in striated muscle (Raeymaekers, 1982; Raeymaekers and Jones, 1986). This reduced rate of Ca²⁺ uptake is most likely due to the reduced density of SERCA in smooth muscle (Wuytack et al., 1989).

The activity of SERCA is largely regulated by phospholamban, a small protein of 52 amino acids that forms a homopentamer and is present in the smooth muscle SR membranes. Although it was shown to be expressed in the porcine gastric smooth muscle (Raeymaekers and Jones, 1986) and rabbit aorta (Cornwell et al., 1991), there is species and tissue variability in the amount of phospholamban expressed (Raeymaekers and Jones, 1986). There is also variability in the sensitivity of SERCA isoforms to this regulatory protein; although SERCA2b is regulated by both phospholamban and calmodulin kinase, SERCA3 is not regulated by either of these proteins. In addition, SERCA3 is more resistant to reactive oxygen species than SERCA2b (Grover and Samson, 1997).

Unphosphorylated phospholamban 1) lowers the apparent turnover rate (global V_max) of SERCA2 through interactions with its cytoplasmic domain (likely by decreasing the Ca²⁺ transport portion of the cycle E₁P·[Ca₂] ... E₂P (Hughes et al., 1994, 1996) and 2) lowers its apparent affinity for Ca²⁺, without affecting the true chemical affinity, through interactions with its transmembrane domain (James et al., 1989; Sasaki et al., 1992a; Cantilina et al., 1993). Its phosphorylation causes its dissociation from SERCA (Tada, 1992) and increases SERCA apparent affinity for Ca²⁺ by reducing the activation energy for a slow transition triggered by Ca²⁺ binding in the Ca²⁺ pumping cycle (Fig. 3), making the pump more "reactive" to cytoplasmic Ca²⁺ (Cantilina et al., 1993).

When phospholamban is phosphorylated, e.g., by cAMP or Ca²⁺/calmodulin kinase, SERCA activity is increased resulting in an enhanced uptake of Ca²⁺ by the SR, although it should be noted that the extent of cAMP-stimulated phospholamban activity is considerably lower than in cardiac tissue (Watras, 1988). Phospholamban is also an excellent substrate for PKG (Raeymaekers et al., 1988). cGMP is more effective in reducing cytoplasmic Ca²⁺ and is thus a potent mediator of smooth muscle relaxation (Felbel et al., 1988). Indeed, it is likely involved in NO-induced SERCA activation; NO activation of guanylate cyclase would increase cGMP concentration, which would activate SR membrane-located PKG, which would then phosphorylate phospholamban (Felbel et al., 1988; Raeymaekers et al., 1988; Twort and van Breemen, 1989; Cornwell et al., 1991; Karczewski et al., 1992; Andriantsitohaina et al., 1995). Several groups have also described a novel means of activation of SERCA by direct phosphorylation by a Ca²⁺-calmodulin kinase (CaM kinase)—for example in heart (Xu et al., 1993), HEK cells (Toyofuku et al., 1994), skeletal muscle (Hawkins et al., 1994), and coronary arteries (Grover et al., 1996).

2. Ca²⁺ Release Channels. As mentioned earlier, release of Ca²⁺ from the SR can occur via a basal leak that is removed from the cytoplasm via a concerted action of mitochondrial uptake, SERCA activity, and plasma membrane extrusion through the Na⁺-Ca²⁺ exchanger and PMCA. The most physiologically relevant SR Ca²⁺ release, however, occurs though activation of InsP₃R and RyR. Interestingly, in many smooth muscle preparations, the SR Ca²⁺ pools released through these channels overlap (Missiaen et al., 1992b). Furthermore, there could be functional interactions between SERCA and the SR Ca²⁺ release channels (Gomez-Viquez et al., 2003).

There are three genes that encode InsP₃R and also three genes that encode RyR, each generating a specific isoform. The sequences encoded by InsP₃R and RyR genes house a basic similarity in structure, sharing fragmental amino acid residue homology concentrated in the ligand-binding and Ca²⁺ channel domains, implying fundamental roles of these domains in the activity of the SR Ca²⁺ channels (Yoshida and Imai, 1997) and suggesting a common ancestral history (Sorrentino and Rizzuto, 2001). Each of these release channels is indeed configured in as a tetrameric formation, with the RyR being homotetrameric, the InsP₃R being heterotetrameric, and the molecular weights of purified InsP₃R (500 kDa) and RyR (300 kDa) indicating large protein structures (Furuichi et al., 1989, 1994). However, despite the above molecular similarities, the three-dimensional structure at 24-Å resolution of these two classes of channels is quite different (Jiang et al., 2002) (see the two following sections).

Local accessory proteins tailor the functional properties of these channels within particular cells and subcellular domains (Mackrill et al., 1996). Some of these proteins modulate activity of all SR Ca²⁺ release channels, whereas others have classor even isoform-specific effects. Some proteins exert both direct and indirect regulation, sometimes with opposing effects, whereas others are themselves modulated by [Ca²⁺]_cyt changes, thus being part of feedback loops.

a. Ca²⁺-Gated Channel/Ryanodine Receptor. The RyR channel is activated when surrounding [Ca²⁺]_cyt increases sufficiently to trigger CICR (Fig. 4) (see below). The RyR is a homotetrameric protein approximately 2 MDa in molecular weight. In mammalian tissues (including smooth muscle), it forms a family of three isoforms, RyR1, RyR2, and RyR3, each encoded by a distinct gene, either ryr1 (initially cloned and sequenced from skeletal muscle), ryr2 (initially cloned and sequenced from cardiac muscle), and ryr3 (initially cloned and sequenced from the brain) (Sutko et al., 1997). Two alternatively splice variants of RyR1 and one variant of RyR2 have also been identified. RyR knockout mice have also been developed; in mice lacking RyR3, caffeine and NE maintain their contractile effects (Yamazawa et al., 1996), whereas Ca²⁺ activity is significantly increased (Lohn et al., 2001). RyR2 knockout is lethal due to cardiac malformation (Takeshima et al., 1998). An interesting approach was used by Drega et al. (2001) to minimize the function of RyR: Using organ culture techniques (4 days, 10-100 µM ryanodine), RyR protein was recovered but RyR were nonfunctional. An interesting finding of this study is that although intracellular stores recover following chronic ryanodine treatment, RyR activity is essential for Ca²⁺-spark activity but not for Ca²⁺ waves/oscillations (Dreja et al., 2001).

View larger version (25K): [in this window] [in a new window]   FIG. 4. CICR and Ca2+ sparks. 1, influx of extracellular Ca2+ occurs through opening of voltage-gated Ca2+ channels into both the junctional and the restricted subsarcolemmal spaces. 2, Ca2+ accumulation in these spaces activates Ca2+-gated Ca2+-release channel/RyR leading to the opening of the channel, a phenomenon termed CICR. 3, the resulting CICR into the junctional subsarcolemmal space (termed Ca2+ spark) raises the local Ca2+ concentration enough to activate large conductance Ca2+-activated K+ channel (BK) on the plasma membrane. 4, the resulting BK opening leads to K+ efflux, which in turn induces membrane hyperpolarization. 5, this hyperpolarization promotes voltage-gated Ca2+ channel closure reducing Ca2+ influx both in the junctional subsarcolemmal space (5a) and in the restricted subsarcolemmal space (5b). Sarcolemma = plasma membrane, subsarcolemmal = subplasmalemmal, myoplasmic = cytoplasmic.

Expression patterns of RyR subtypes show variable distribution, with RyR1 and RyR2 being present in skeletal muscle and cardiac muscle, respectively; RyR2 is the predominant isoform in neural tissue. RyR3 is expressed mainly in embryonic tissue, and levels decrease during development (Rossi and Sorrentino, 2002). In smooth muscle cells, RyR2 and RyR3 are the primary isoforms (Sanders, 2001). Interestingly, although all three RyR isoforms are present in vascular smooth muscle of neonatal mice, where the SR content appears normal, these RyR do not become fully functional until further maturity of the animals (Gollasch et al., 1998). There is also a regional variability in the expression combination of the three RyR isoforms in smooth muscle from different organs. RyR2 is required for generation of Ca²⁺ sparks, with either a minimal (Mironneau et al., 2001) or inhibitory (Lohn et al., 2001) contribution for RyR3.

The channel is an assembly of four RyR subunits (protomers) of the same isoform (thus an homotetramer) forming a central Ca²⁺-conducting pore, which has a diameter of 2 to 3 nm. RyR1 and RyR3 have been shown to differ significantly in vitro in terms of gating and activation. Topologically, various studies suggest that the RyR spans the membrane 4, 10, or 12 times (Michikawa et al., 1996), with highly conserved ionchannel-forming membrane-spanning regions that appear to be localized to the carboxyl terminal (20% of the protein), while the remaining amino-terminal region of the protein forms a large cytoplasmic foot domain that assumes a quatrefoil shape (Sutko et al., 1997; Welch et al., 1997; Wagenknecht and Samso, 2002). The RyR is anchored to the SR by interaction with the Ca²⁺ binding storage protein calsequestrin (see Section I.E.).

The distribution pattern of RyR follows that of the SR such that in where there is a patch distribution of SR in the cytoplasm, as in the guinea pig aorta, there is a sparse labeling with ryanodine markers, whereas tissues with a more prominent peripheral SR have a rich marking of RyR in the periphery—in other words, RyR distribution parallels that of the SR (Lesh et al., 1998). The physiological roles of RyR in smooth muscle cells are still being elucidated (Guerrero-Hernandez et al., 2002). Although its endogenous gating ligand is Ca²⁺ leading to CICR, as mentioned above, it appears that its basal Ca²⁺ sensitivity is in the micromolar range, a [Ca²⁺]_cyt not globally reached in the bulk of the cytoplasm (Sanders, 2001). Hence, RyR appears to be activated in proximity to the plasma membrane (e.g., on the SR face of the plasma membrane-SR junctional space) by extracellular Ca²⁺ influx through Ca_L to produce CICR of Ca²⁺ sparks, which in turn modulates plasma membrane excitability (e.g., hyperpolarization) through activation of small conductance K_Ca channels (SK), large conductance K_Ca (BK), and depolarization through Cl_Ca (Jaggar et al., 2000; Sanders, 2001). In cerebral artery smooth muscle cells, for instance, Ca²⁺ sparks-activated BK openings promote relaxation (Nelson et al., 1995). There may be a role, however, for regional vascular differences with regard to the role of Ca²⁺ sparks. For example, spontaneous transient outward currents (STOC), and hence Ca²⁺ sparks, are very active in fetal pulmonary arteries, and this activity diminishes with maturation (Pratusevich and Balke, 1996; Jaggar et al., 2000). This contrasts completely with the nearly 100-fold increase in Ca²⁺ spark activity and STOC frequency during maturation of systemic arterial cells (Gollasch et al., 1998). Thus, in pulmonary artery cells, endothelin, the potent endogenous vasoconstrictor, activates Ca²⁺ sparks by causing the cross-signaling of RyR and InsP₃R—in this case, increased smooth muscle Ca²⁺ activity is associated with constriction (Ge et al., 2003; Zhang et al., 2003). Another example of the physiological relevance of Ca²⁺ sparks is the finding that stretching of urinary bladder smooth muscle cells generates Ca²⁺ sparks from RyR sites on the SR (Ji et al., 2002).

The finer details of the organization of microdomains that underlie Ca²⁺ sparks are being unraveled with insightful experimental analysis. It is now apparent that K_Ca channels that underpin STOC are exposed to a mean Ca²⁺ concentration on the order of 10 µM during a Ca²⁺ spark (ZhuGe et al., 2002). The membrane area over which a concentration of 10 µM or more (range 12-21 µM) is achieved has an estimated radius of 15 to 30 nm, corresponding to an area that is a fraction of one square micron (0.07-0.28 µm²). It is also apparent that K_Ca channels are not uniformly distributed over the membrane but exist as clusters at sites of frequent discharge of Ca²⁺ sparks, where the K_Ca channels and RyR mediating Ca²⁺ sparks is in the order of 25 nM (ZhuGe et al., 2000, 2002). Indirect support for an intimate relationship between K_Ca channels and RyR comes from findings in smooth muscle-excised patches of portal vein (Xiong et al., 1992) and vas deferens (Ohi et al., 2001b), where there is evidence for clustering of K_Ca channels and an apparent attachment of SR membrane.<