Local intracellular Ca2+transients, termed Ca2+ sparks, are caused by the coordinated opening of a cluster of ryanodine-sensitive Ca2+ release channels in the sarcoplasmic reticulum of smooth muscle cells. Ca2+ sparks are activated by Ca2+ entry through dihydropyridine-sensitive voltage-dependent Ca2+ channels, although the precise mechanisms of communication of Ca2+ entry to Ca2+ spark activation are not clear in smooth muscle. Ca2+ sparks act as a positive-feedback element to increase smooth muscle contractility, directly by contributing to the global cytoplasmic Ca2+ concentration ([Ca2+]) and indirectly by increasing Ca2+ entry through membrane potential depolarization, caused by activation of Ca2+ spark-activated Cl− channels. Ca2+ sparks also have a profound negative-feedback effect on contractility by decreasing Ca2+ entry through membrane potential hyperpolarization, caused by activation of large-conductance, Ca2+-sensitive K+ channels. In this review, the roles of Ca2+sparks in positive- and negative-feedback regulation of smooth muscle function are explored. We also propose that frequency and amplitude modulation of Ca2+ sparks by contractile and relaxant agents is an important mechanism to regulate smooth muscle function.
- ryanodine-sensitive calcium-release channel
- voltage-dependent calcium channel
- calcium-sensitive potassium channel
- calcium-activated chloride channel
- sarcoplasmic reticulum
calcium ions directly regulate the contraction of smooth muscle through the activation of Ca2+/calmodulin-dependent myosin light-chain kinase (MLCK). For this to occur, an elevation of Ca2+ must be broadly or globally distributed throughout the cytoplasm. The view that a spatially homogeneous elevation in cytoplasmic Ca2+ is necessary for effective Ca2+ signaling has been radically altered by the discovery of local Ca2+ transients (“Ca2+ sparks”) in cardiac (37), skeletal (99,193), and smooth (140) muscle cells. Ca2+ sparks are caused by the opening of ryanodine-sensitive Ca2+-release (RyR) channels in the sarcoplasmic reticulum (SR). A single Ca2+spark is capable of producing a very high (10–100 μM) local (∼1% of the cell volume) increase in [Ca2+], while increasing the global intracellular concentration of Ca2+([Ca2+]i) by <2 nM (85, 140). A Ca2+ spark, by virtue of its high local [Ca2+] elevation, has the potential to modulate Ca2+-dependent processes that are not responsive to global increases in [Ca2+]i .
Early evidence for local Ca2+ release in cells came from measurements of spontaneous transient currents through Ca2+-sensitive K+ channels in frog sympathetic ganglia neurons (26, 125, 126). Subsequently, Benham & Bolton (13) measured transient outward currents through large-conductance Ca2+-sensitive K+(BKCa) channels (referred to as “spontaneous transient outward currents” or STOCs) in isolated single cells of longitudinal smooth muscle from rabbit jejunum and ear artery (13). STOCs, which are caused by Ca2+ sparks (140), have since been described in a wide variety of types of smooth muscle (13, 19, 43, 75, 79, 82, 89, 95,137, 140, 143, 145, 157, 163, 167, 217-219). The existence of STOCs suggested the possibility that subsarcolemmal [Ca2+] is different from global cytoplasmic [Ca2+] and that, under certain circumstances, subsarcolemmal [Ca2+] could be significantly higher than global [Ca2+]i. Further support for this idea has come from studies that demonstrated a dissociation between global [Ca2+]iand BKCa channel currents (56, 181). Also, Yamaguchi (215) provided direct evidence, using a modified confocal microscope, for an isoproterenol-induced increase in peripheral [Ca2+], which was ryanodine sensitive and distinct from [Ca2+] changes that occur in the cytosol.
The impact of a Ca2+ spark depends on a number of key factors: 1) the distance of the Ca2+ spark site from a Ca2+-sensitive target, 2) the number of target molecules affected by a Ca2+ spark, 3) the Ca2+ sensitivity of the target protein, and 4) the nature of the Ca2+-sensitive target. The physiological outcome of the Ca2+ spark depends on the [Ca2+] that reaches the target. Ca2+ reaches very high levels (10–100 μM) very close (e.g., 20 nm) to the release site (Ref. 153; see also Ref. 139). However, at several hundred nanometers from a Ca2+ spark release site, Ca2+ approaches equilibrium with mobile buffers (including fluorescent indicators), and the local elevation in [Ca2+] by a spark is less pronounced (< 500 nM; see Ref. 139). Molecular targets with low affinity for Ca2+, such as BKCa channels, are uniquely suited to respond to these very high local [Ca2+], since these channels require micromolar Ca2+ for significant levels of activity under physiological conditions (153). A single Ca2+ spark has a tremendous impact on cell membrane potential [up to 20 mV hyperpolarization of single cells (52)] through activation of BKCachannels (140). In contrast, some Ca2+/calmodulin (CaM)-dependent targets [e.g., MLCK, CaM kinase, small-conductance Ca2+-sensitive K+(SKCa) channels, nitric oxide synthase (NOS), calcineurin, Ca2+-activated chloride (ClCa) channels] are well suited to respond to changes in global Ca2+(0.1–1 μM), since CaM is saturated by [Ca2+] below 1 μM.
The goal of this review is to provide up-to-date information on RyR channels and Ca2+ sparks in smooth muscle and to propose the following: 1) voltage-dependent Ca2+ channels, Ca2+ sparks (RyR channels), and BKCa channels act as a functional unit to regulate smooth muscle function (Figs.1 and 2);2) communication between voltage-dependent Ca2+channels and RyR channels is a critical element in the control of global [Ca2+]i (Figs. 1 and 2);3) Ca2+ sparks are involved in both positive- and negative-feedback regulation of global [Ca2+]i (Figs. 1 and 2); 4) frequency and amplitude modulation (FM and AM) of Ca2+sparks by relaxant and contractile agents are mechanisms to regulate smooth muscle function. Unitary Ca2+ release may also occur due to inositol trisphosphate (IP3) activation of IP3 receptors in the SR [termed “blips” and “puffs” (22, 216)]. In this review, the term ‘Ca2+ sparks’ refers to Ca2+ release through RyR channels.
RYANODINE RECEPTORS IN SMOOTH MUSCLE
Smooth muscle SR elements are very close (within 20 nm) to the plasma membrane (Fig. 3 A) (45). RyR channels have been localized to both the peripheral (Fig. 3 A) and central SR (45, 61, 118). Dihydropyridine receptors (voltage-dependent Ca2+ channels) and RyR channels colocalize in smooth muscle (33, 61), suggesting an intimate communication between these proteins in the plasma membrane and the SR. A strong spatial association has also been shown for the Na+-K+ pump and the Na+/Ca2+ exchanger, and the Na+/Ca2+ exchanger and Ca2+released from the SR (18, 136). This close spatial localization between RyR channels and the plasma membrane suggests a special communication between the SR and sarcolemmal ion channel and ion transport systems.
RyR channels are found in all muscle types and many other cell types, including those from mammals, birds, amphibians, reptiles, fish, insects, crustaceans, and molluscs (for a review see Ref. 183). RyR channels were first identified in insect and invertebrate striated muscles, due to the action of the plant alkaloid, ryanodine, a potent channel antagonist (see Ref. 183). Presently, three molecularly distinct subtypes of RyR channels (RyR1, RyR2, RyR3) have been identified and cloned, with each thought to exist as a homotetramer in the SR membrane. RyR1 is found primarily in skeletal muscle, whereas RyR2 and RyR3 are predominantly found in cardiac tissue and brain, respectively. mRNA transcripts for RyR1, RyR2, and RyR3 have been identified in smooth muscle preparations (see Refs. 67, 90, 117, 142). However, these results should be interpreted with some caution, since the detection of RyR1–3 transcripts may reflect contamination from cells other than smooth muscle cells (e.g., endothelium, neurons) and does not necessarily reflect the level of functional protein.
A major breakthrough in the investigation of the functional properties of RyR channels involved the incorporation of RyR channels from cardiac and skeletal muscle SR microsomes into planar lipid bilayers (176). One difficulty with isolating RyR channels from smooth muscle is that smooth muscle SR membranes cannot be readily separated from other membranes. As an alternative approach, RyR channels have been partially purified from aortic smooth muscle (71) (Table1) and from toad stomach smooth muscle (214) and incorporated into microsomes and then into planar lipid bilayers. The molecular and functional properties of the three cloned RyR channel subtypes and the smooth muscle RyR channel are summarized in Table 1. RyR channels from smooth muscle are activated by micromolar cytoplasmic [Ca2+] (Fig.4) and by caffeine and are inhibited by Mg2+ and ruthenium red (71), similar to RyR2 and RyR3. The single-channel conductance of the smooth muscle RyR channels is also similar to the cardiac (RyR2) and skeletal muscle (RyR1) channels. However, the density of RyR channels in smooth muscle is considerably lower than in other tissues (Table 1).
It has been suggested that RyR3 may be the prominent ryanodine receptor isoform in smooth muscle (67). However, a possible role of RyR3 in smooth muscle is unclear, since mice lacking RyR3 do not have impaired smooth muscle function, and smooth muscle cells from these animals respond normally to caffeine and norepinephrine (187). Indeed, RyR3-deficient mice mature to adulthood, with the only discernable abnormality being increased motor activity (187).
A number of factors suggest that the predominant RyR channel subtype in smooth muscle is similar to that in cardiac muscle (RyR2). For example, the pharmacological and biophysical characteristics of RyR channels isolated from native smooth muscle cells are similar to those of RyR2 channels incorporated into lipid bilayers (71, 214) (Table 1). Ca2+ sparks that occur in cardiac cells and smooth muscle cells have similar biophysical and pharmacological characteristics (e.g., see Refs. 37 and 140; Table 2). Furthermore, the ryanodine-binding receptors purified from toad stomach, and from cerebral artery smooth muscle cells, cross-react with monoclonal antibodies for RyR2 (Fig. 3 B) (61, 214). Therefore, the RyR channel responsible for Ca2+ release from the SR in smooth muscle appears to be similar to the RyR2 subtype, although an involvement of other RyR channel subtypes is also possible.
CA2+ SPARK PROPERTIES IN SMOOTH MUSCLE
Ca2+ sparks have been observed in cardiac (37), skeletal (99, 193), and smooth (140) muscle cells. These measurements have been made using laser scanning confocal microscopy and the fluorescent Ca2+ indicator, fluo 3. A number of lines of evidence indicate that Ca2+ sparks are caused by the opening of RyR channels in the SR, including modulation by ryanodine, voltage, Ca2+, and caffeine, and location of events (30, 37, 99,121, 140, 193). Table 2 summarizes the properties of Ca2+sparks in cardiac, skeletal, and smooth muscle cells. Ca2+sparks recorded in skeletal muscle are approximately one-third the size of those observed in cardiac and smooth muscle (Table 2), although the single-channel currents through skeletal and cardiac and smooth muscle RyR channels incorporated into bilayers are about the same (Table 1). It has been postulated that this difference in Ca2+ spark amplitude is due to the number of RyR channels contributing to the Ca2+ spark or to different RyR channel kinetics (99, 131).
Ca2+ sparks in smooth muscle cells were first described in myocytes from rat cerebral arteries [Figs.5 and 6 (140), see also Refs. 19, 61, 85, 86, 153, 157]. Subsequently, Ca2+ sparks have been measured in smooth muscle cells from coronary arteries (86, 157), mesenteric artery (132), rat portal vein (9), guinea pig (218) and porcine (172) trachea, guinea pig ileum (63), guinea pig urinary bladder (83), guinea pig vas deferens (83), and toad stomach (219). The properties of Ca2+ sparks appear to be similar in these different smooth muscle preparations (Table3). Figure 5 illustrates recordings of Ca2+ sparks in the smooth muscle cells of an intact pressurized (60 mmHg) cerebral artery. In arterial smooth muscle, Ca2+ sparks have a rise time of ∼20 ms, a half time of decay of 50–60 ms (Figs. 5 and 6), and a spatial spread (full width at the half-maximum amplitude) of 2.4 μm (19, 85, 140, 153). These unitary events occur most frequently close to the cell membrane (140), although events away from the cell membrane have been reported in smooth muscle cells from portal vein (9) and ileum (63). Ca2+ sparks occur in intact, nonpressurized (61, 85) and pressurized cerebral (Fig. 5) and mesenteric arteries (132), with an apparent frequency of ∼1 ⋅ s−1 ⋅ cell −1 at physiological membrane potentials (−60 to −40 mV). These results support the idea that [Ca2+]i is dynamically and heterogeneously distributed in intact tonic smooth muscle.
A Ca2+ Spark is due to the Simultaneous Activation of a Cluster of RyR Channels
Ca2+ sparks appear to represent the near-simultaneous activation of a cluster, or plaque, of RyR channels. The flux of Ca2+ from a spark site is substantial and corresponds to a Ca2+ current of 4 pA for ∼10 ms (37). With physiological levels of SR luminal [Ca2+], the unitary current of a single cardiac RyR channel is <0.6 pA (131). Therefore, a Ca2+ spark would correspond to the coordinated opening of at least 10 RyR channels (131). In cardiac muscle, the elementary physiological event appears to be Ca2+ sparks, since the global Ca2+ transient is composed of many Ca2+sparks (29). In a similar fashion, the miniature end plate potential represents the release of one presynaptic vesicle (the elementary physiological event), and not the release of one neurotransmitter molecule.
Fate of Ca2+ From a Ca2+ Spark
The decay of the Ca2+ spark may be determined by diffusion, uptake into the SR by the thapsigargin-sensitive Ca2+-ATPase, and extrusion from the cell through the plasmalemmal Ca2+-ATPase and Na+/Ca2+ exchanger. In cardiac muscle, Ca2+ spark decay reflects diffusion and SR Ca2+uptake, with a small contribution from Ca2+ extrusion (165). As measured with fluorescent dyes, the decay of Ca2+sparks in arterial smooth muscle is biexponential, with both rate constants (τ1 = 32 ms and τ2 = 275 ms) contributing equally to the decay (49 and 51%, respectively) (153). It is unclear why Ca2+ sparks demonstrate a biexponential decay pattern, but this may reflect limited space diffusion, fluorescent dye kinetics, or extrusion/uptake of Ca2+ by SR or sarcolemmal Ca2+ pumps. The faster of the two time constants for decay is similar to that which has been measured for Ca2+ sparks in heart muscle (62, 165), consistent with diffusion of Ca2+ in the cytoplasm. The contribution of Ca2+ extrusion/sequestration mechanisms to the decay of Ca2+ sparks in smooth muscle remains to be addressed. However, the decay of a whole cell Ca2+ transient, which reflects Ca2+ extrusion and SR Ca2+ uptake, is much slower [decay constant (τ) = 2.0–8.6 s (49, 207); see also Refs. 53-55, 57, 93, 94] than the decay of a Ca2+ spark. This observation suggests that Ca2+ spark decay is largely determined by diffusion (see Ref. 62).
The smooth muscle cell Ca2+ spark represents the spatially localized, unitary release of a significant amount of Ca2+very close to the plasma membrane, with essentially no effect on global [Ca2+]i or on contraction (140). The localized release of Ca2+ into a subsarcolemmal domain can thus represent a measurable element of the “superficial buffer barrier” (SBB) hypothesis (for review see Ref. 196). This hypothesis suggests that Ca2+ entering a smooth muscle cell is continuously buffered by the SR and is then discharged “vectorially” toward the plasma membrane without any effect on global [Ca2+]i. Subsequently, this Ca2+ is extruded from the cell via the sarcolemmal Na+/Ca2+ exchanger and Ca2+-ATPase. The anatomic requirements for a Ca2+ spark to activate BKCa channels are similar to those proposed for the SBB hypothesis. Additional studies on the decay of Ca2+ sparks should provide new insights into local Ca2+ extrusion and uptake mechanisms as well as the SBB hypothesis.
CA2+ SPARKS ARE INVOLVED IN BOTH POSITIVE AND NEGATIVE FEEDBACK REGULATION OF GLOBAL INTRACELLULAR [CA2+] IN SMOOTH MUSCLE
The role of local Ca2+ signaling is cell type dependent and should be viewed in the context of cell function. For example, cardiac muscle must contract forcefully and relax rapidly at high rates (1–7 Hz). In these cells, activation of voltage-dependent Ca2+ channels, which is synchronized by the action potential, leads to an immediate rise in Ca2+ influx, which activates, in a coordinated and rapid fashion, large numbers of Ca2+ sparks, which summate to cause the global [Ca2+]i transient. Thus Ca2+-induced Ca2+-release (CICR) through RyR channels acts as an amplifier for Ca2+ influx, contributing 90% of the cellular Ca2+ needed for contraction of heart muscle (Fig. 1). Although cardiac and smooth muscle share many of the same key components for excitation-contraction coupling (e.g., L-type Ca2+ channels and RyR channels), these elements are arranged to regulate function in very different ways. Importantly, cardiac muscle lacks BKCa channels. In smooth muscle, the triumvirate of dihydropyridine-sensitive voltage-dependent Ca2+ channels, RyR channels, and BKCa channels represent a functional unit to control the levels of [Ca2+]i (86).
Most types of smooth muscle are involved in maintaining tonic contraction. One important element in graded, tonically contracted smooth muscle is the negative-feedback control of membrane potential, and hence Ca2+ entry, through activation of BKCa channels (Figs. 1, 2, and 6; cf. Refs. 25 and 140). Some types of smooth muscle (e.g., urinary bladder) exhibit phasic contractions, suggesting that Ca2+ release through RyR channels may contribute directly to contraction, like cardiac muscle, and indirectly to relaxation through activation of BKCachannels (Fig. 6) (69). Furthermore, as discussed later, some types of smooth muscle have SKCa channels and ClCachannels (see Fig. 2 and Table 4), which will also contribute to the regulation of smooth muscle excitability.
Communication Between Voltage-Dependent Ca2+ Channels and RyR Channels is a Critical Element in the Control of Global Intracellular [Ca2+]
The first step in the flow of information following electrical excitation in heart muscle is an increase in Ca2+ influx through voltage-dependent Ca2+ channels. The communication between voltage-dependent Ca2+ channels and RyR channels is intimate, rapid, and bidirectional. RyR channels are positioned in junctional SR elements within short distances (∼20 nm) of voltage-dependent Ca2+ channels in the t-tubules of ventricular heart cells (for review see Ref. 51). RyR channels are activated by local Ca2+ shortly after they enter the cell through voltage-dependent Ca2+ channels (“local control”) (30, 36, 121). Local Ca2+ in the subsarcolemmal junctional space is likely to reach relatively high concentrations (>10 μM), corresponding to levels required for significant RyR channel activation. Thus local Ca2+ control of Ca2+ release can explain the graded and stable increases in Ca2+ release through RyR channels in response to Ca2+ influx (see Fig. 1) (30, 121, 178). The communication is bidirectional, since Ca2+ sparks can also induce inactivation of voltage-dependent Ca2+ channels (1, 65,170).
The communication between Ca2+ channels and RyR channels/Ca2+ sparks is less clear in smooth muscle. In smooth muscle, Ca2+ sparks are also regulated by Ca2+ influx through voltage-dependent Ca2+channels, since Ca2+ spark frequency and amplitude increase with activation of voltage-dependent Ca2+ channels in the smooth muscle cells of intact arteries (85). Membrane potential depolarization of intact cerebral arteries from about −60 to −40 mV, a membrane potential that would occur in pressurized arteries with tone, increases Ca2+ spark frequency about fivefold (85). The voltage dependence is similar to that observed in isolated cardiac myocytes over a similar voltage range (30, 164). This increase in Ca2+ spark frequency and amplitude depends on Ca2+ influx through voltage-dependent Ca2+channels, since it is reversed by diltiazem, a selective blocker of L-type Ca2+ channels (85). Additionally, in voltage-clamped rat portal vein myocytes, Ca2+ sparks have been observed to occur immediately following a depolarizing voltage step that elicits a Ca2+ current (9). Although these studies indicate that Ca2+ spark frequency is modulated by Ca2+ influx through voltage-dependent Ca2+channels, it is still unresolved whether local Ca2+regulates Ca2+ spark probability in smooth muscle.
Steady membrane depolarization of arterial smooth muscle increases voltage-dependent Ca2+ channel open probability (e.g., see Ref. 162), global [Ca2+]i, and presumably SR [Ca2+]. An elevation in local [Ca2+] from Ca2+ channels, global [Ca2+]i, or SR [Ca2+] could each contribute to the observed depolarization-induced increase in Ca2+ spark frequency and amplitude (Figs. 1 and 7). Indeed, recent evidence indicates that Ca2+ spark and STOC frequency in smooth muscle cells depends steeply on the SR Ca2+ load (219). Figure 7 illustrates modulation of the Ca2+dependence of Ca2+ spark frequency by SR Ca2+load and other factors (see fm and am of ca2+ sparks: physiological implications).
Ca2+ release from RyR channels can inactivate voltage-dependent Ca2+ channels in cardiac myocytes (23,134, 170), which would serve as another negative-feedback mechanism to regulate Ca2+ entry. The possibility that a similar mechanism plays a role in smooth muscle has not been extensively examined (although see Ref. 97). Indirect evidence from measurements of arterial wall [Ca2+]i and diameter in pressurized arteries suggests that this pathway is not significant (103), at least in the absence of receptor agonists.
Direct Contribution of Ca2+Sparks to Cytoplasmic [Ca2+]
In cardiac muscle, Ca2+ release through RyR channels (CICR) contributes most (∼90%) of the Ca2+ for contraction. In tonic smooth muscle (e.g., arteries), in the absence of smooth muscle constrictors, global [Ca2+]i is largely determined by steady-state Ca2+ entry through voltage-dependent Ca2+ channels, and Ca2+release likely contributes little to global [Ca2+]i (102, 103, 140). Ca2+ spark frequency is low in intact arteries [∼1 Ca2+ spark ⋅ s−1 ⋅ cell −1(Fig. 5)] (19, 61, 85, 140, 153, 157), and blocking Ca2+ sparks or RyR channels (in the presence of blockers of BKCa channels) has little direct effect on global [Ca2+]i and arterial tone (102,103, 140). However, it is conceivable that, under conditions of elevated arterial tone (e.g., in hypertension), Ca2+ sparks might have a significant impact on global [Ca2+]i.
In smooth muscle types that exhibit phasic contractions (e.g., urinary bladder), Ca2+ release through RyR channels may contribute Ca2+ directly for contraction. Unlike tonic smooth muscle, phasic smooth muscle exhibits action potentials, which activate voltage-dependent Ca2+ channels to trigger Ca2+release. Ca2+ release through RyR channels can contribute as much as 70% of the increment in global Ca2+ (30% to Ca2+ influx) in isolated voltage-clamped urinary bladder myocytes (54). The excitability and duration of action potentials in phasic smooth muscle are also regulated in a negative-feedback manner by BKCa and SKCa channels (69, 70). Therefore, blocking RyR channels affects both positive- and negative-feedback regulation of [Ca2+]i in these types of smooth muscle. The precise contribution of Ca2+release through RyR channels to global [Ca2+]i in intact phasic smooth muscle remains to be determined.
The ability of voltage-dependent Ca2+ influx to stimulate Ca2+ release through RyR channels appears to vary considerably among different types of smooth muscle. All types of smooth muscle are capable of releasing significant amounts of Ca2+ through RyR channels, since caffeine, an activator of RyR channels (Table 1), can induce global ryanodine-sensitive Ca2+ transients in virtually all smooth muscle types (cf. Refs. 12, 39, 49, 54, 61). Yet, influx through voltage-dependent Ca2+ channels can induce ryanodine-sensitive elevations in global [Ca2+]i in some (54, 94), but not all (49, 128), types of smooth muscle. The lack of effective CICR through RyR channels suggests that, in smooth muscle, many of the RyR channels are spatially distant from the high local [Ca2+] caused by the opening of voltage-dependent Ca2+ channels. However, even in cell types (e.g., urinary bladder) that exhibit significant ryanodine-sensitive global Ca2+ transients, the situation is very different from cardiac muscle. CICR, when it does occur in smooth muscle, is very sluggish, developing slowly over hundreds of milliseconds (93, 94). In contrast, the close proximity of cardiac Ca2+ channels and RyR channels enables rapid and effective communication, i.e., Ca2+ influx stimulates Ca2+ release (Ca2+ sparks) within milliseconds (30, 121). Therefore, the failure of voltage-dependent Ca2+influx to induce substantial Ca2+ transients does not seem to reflect a lack of releasable Ca2+ through RyR channels, but may reflect a physical separation of voltage-dependent Ca2+ channels from a majority of the RyR channels. Communication between Ca2+ channels and RyR channels in smooth muscle may be analogous to that in atrial muscle where the t-tubular network is virtually absent. Atrial muscle has been found to contain “feet-like” structures similar in appearance to those in smooth muscle (45) that occur close to the plasmalemma (50). Electrical stimulation of atrial myocytes has been demonstrated to elicit Ca2+ sparks immediately beneath the plasmalemma, which then trigger Ca2+ sparks in more central cellular regions to elicit a global Ca2+ transient (15, 80).
In heart, the Ca2+ transient is composed of the simultaneous activation of many Ca2+ sparks. However, in smooth muscle, the contribution of Ca2+ sparks to ryanodine-sensitive global Ca2+ transients is not known. It is also conceivable that, in smooth muscle, Ca2+ spark sites communicate only with local targets (e.g., BKCachannels), and ryanodine-sensitive global Ca2+ transients are caused by Ca2+ release through RyR channels, which are not in Ca2+ spark sites (these channels may be located in the central SR rather than at peripheral subplasmalemmal locations). The precise regulation of Ca2+ sparks and RyR channels by local Ca2+ influx and the direct contribution of Ca2+ release to contraction remain to be elucidated.
Negative Feedback Regulation of Global [Ca2+]iThrough Ca2+ Spark Activation of BKCa Channels in Smooth Muscle
BKCa channels have an estimated single-channel conductance of ∼80 pS at −40 mV with a physiological K+gradient and occur at a density of ∼1–4 channels/μm2 (Table 4) (13, 173, 190). Nelson and colleagues (140) proposed that Ca2+ sparks activate a number of nearby BKCa channels to cause a macroscopic BKCa channel current, which had been previously described as a STOC (140). Simultaneous optical and electrical measurements in isolated smooth muscle cells support the idea that Ca2+sparks activate STOCs (133, 153, 218). Quantitative analysis of such simultaneous measurements indicates that virtually all Ca2+sparks are associated with STOCs in arterial myocytes (Fig. 6,A and B) (153). These measurements also indicate that Ca2+ sparks increase the open probability of BKCa channels ∼106-fold, which is consistent with subsarcolemmal activator [Ca2+] in the order of 10–100 μM (153). STOCs rise and decay quite rapidly [time to peak, 10–20 ms; half time of decay, 9 ms (140,153)], and are consistent with the simultaneous activation of at least 15 BKCa channels in a membrane surface area of ∼15 μm2 (∼1% of the surface area of an arterial myocyte). In contrast, the Ca2+ spark, as measured with the fluorescent indicator fluo 3, rises from baseline to peak in ∼20 ms but decays more slowly [half time of decay 50–60 ms (85,140, 153)] (Tables 2 and 3). These results indicate that the RyR channels that cause Ca2+ sparks are located very close to the sarcolemmal BKCa channels. Disparities in the peak [Ca2+] of a Ca2+ spark measured with fluo 3 (up to 0.5 μM) and the Ca2+requirement for an associated increase in BKCa channel activity (local Ca2+ may actually reach levels of 10–100 μM), as well as kinetic differences between Ca2+ sparks and STOCs, suggest that the fluorescent Ca2+ indicator is unable to accurately track Ca2+ in the subsarcolemmal space. This is consistent with the known chemical properties of fluo 3.
Ca2+ sparks cause up to 20 mV hyperpolarizations of isolated arterial myocytes through activation of BKCachannels (Fig. 8, Aand B, and Table 4) (52). Figure 8 A illustrates a simultaneous measurement of Ca2+ sparks and associated membrane potential hyperpolarizations in an isolated cerebral artery myocyte. Smooth muscle cells are electrically coupled in the intact tissue, such that asynchronized hyperpolarizations that are caused by Ca2+ sparks should summate and lead to a graded level of hyperpolarization (see Fig. 8, B and C) (48). Ca2+ spark-induced membrane potential hyperpolarizations contribute significantly to the membrane potential of isolated smooth muscle cells. For example, in Fig. 8 B, Ca2+spark-induced hyperpolarizations caused an average membrane potential change of ∼15 mV. Ryanodine caused a 15-mV depolarization by inhibiting the transient hyperpolarizations (Fig. 8 B). In support of this mechanism, inhibition of Ca2+ sparks or BKCa channels causes a membrane potential depolarization of ∼10 mV in pressurized cerebral arteries (Fig. 8 C) (103), which leads to an elevation in global [Ca2+]i and vasoconstriction (61,103, 140, 157). The effects of blocking Ca2+ sparks and BKCa channels on membrane potential, arterial wall [Ca2+]i, and diameter are not additive, supporting the idea that Ca2+ sparks regulate BKCa channel activity (61, 103, 140, 157). These results support the concept that Ca2+ sparks, in part, regulate vascular tone, in a negative feedback manner, through activation of BKCa channels (19, 48, 61, 85, 86, 103, 140, 153, 157).
Positive Feedback Regulation of Global [Ca2+]iThrough Ca2+ Spark Activation of ClCa Channels in Smooth Muscle
Some types of smooth muscle, but not all, express ClCachannels. To date, ClCa currents have been described in smooth muscle cells from rat (206) and rabbit (68, 205) pulmonary artery, rabbit ear artery (3, 4), rabbit (114) and porcine (101) coronary artery, rat renal artery (64), human mesenteric artery (100,101), rat (A7r5) (156, 199) and pig (46) aorta, rabbit (28, 203, 204) and rat portal vein (151), guinea pig mesenteric vein (197), rat small intestine (84, 144), rabbit colon (182), canine (87, 89) and guinea pig (87-89) trachea, rabbit esophagus (2), guinea pig uterus (40), rat myometrium (8), and rat anococcygeus (27). ClCa channels are activated by increases in [Ca2+]i with a half-maximal activation at 365 nM and full activation around 600 nM (150) and have a single-channel conductance of 2.6 pS (Table 4) (100). The simultaneous activation of a number of ClCa channels underlies the spontaneous transient inward currents (“STICs”) first described in tracheal (87) and portal vein smooth muscle cells (203) (for a review see Ref. 116). STICs are activated by Ca2+ sparks (218) in a manner similar to that of STOCs. Interestingly, the same Ca2+ sparks that activate STICs also activate STOCs in these cells, indicating that ClCa channels and BKCa channels are colocalized in this preparation (tracheal smooth muscle) (218). Figure 9 reproduced from ZhuGe et al. (218) illustrates the biphasic nature of the current activated by a Ca2+ spark, which initially evokes an outward current due to the activation of BKCa channels, followed by a long-lasting inward current due to the activation of ClCa channels (218).
There are some notable differences in the properties of STOCs and STICs. First, the apparent number of channels involved in STOCs and STICs may be very different. The amplitudes of STICs and STOCs are similar (with similar driving forces for Cl− and K+; e.g., Fig. 9), even though the unitary conductance of the BKCa channel (80 pS) is ∼30-fold greater than the ClCa channel [2.6 pS (100); for reviews see Refs. 31,116, 141]. Therefore, a STIC appears to represent the activation of at least 600 ClCa channels by a Ca2+ spark. ClCa channels may be clustered above Ca2+ spark sites, since tracheal myocytes have <10,000 ClCachannels/cell (116), and a STIC represents the activation of at least 5% of the ClCa channels in 0.25% of the membrane surface area in this preparation. Second, the decay of STICs is much slower than the decay of STOCs and mirrors the decay of the fluorescence transient (see Figs. 6 and 9) (218).
The precise effect of the simultaneous activation of ClCaand BKCa channels by a Ca2+ spark on membrane potential would depend on a number of factors, including the intrinsic voltage dependence of ClCa (none) and BKCachannels [open probability changes about e-fold for 12 -14 mV (14, 115)] and the voltage and Ca2+ dependence of other conductances in the cell membrane (Table 4). It is curious that ClCa channels and STICs have only been reported in a handful of smooth muscle preparations [rabbit portal vein (77,78, 203), rabbit pulmonary artery (76), and canine and guinea pig tracheal myocytes (87, 218, and see Ref. 116)], suggesting that these channels subserve a functional role unique to these tissues. The coexistence of ClCa and BKCa channels in the sarcolemmal membrane adjacent to Ca2+ spark sites suggests another level of fine control of membrane excitability through Ca2+ sparks (see Figs. 2 and 9). Indeed, differential modulation of ClCa and BKCa channels by vasoactive substances would provide a means to alter the control of membrane potential by Ca2+ sparks. For example, vasoconstrictor-induced activation of ClCa channels and inhibition of BKCa channels (see Decrease in Ca2+ Spark Frequency by PKC) could be a mechanism to enhance membrane excitability.
The decay of STICs [t 1/2 = 65–182 ms (89, 218)] is much slower than the decay of STOCs [t 1/2 = 9–43 ms (153, 218)] and is similar to the decay of Ca2+ sparks as measured with the fluorescent indicator fluo 3 (see Figs. 6 and 9) (218). This observation is consistent with the idea that the BKCachannels (STOCs) are activated by high (10–100 μM) local Ca2+ (Fig. 6). The rapid decay of the STOC is determined by dissipation of this local Ca2+ following closure of the RyR channels. In contrast, ClCa channels are saturated with Ca2+ above 1 μM, and these currents do not decay until [Ca2+]i falls below 1 μM, which are concentrations measured accurately by fluo 3. Interestingly, whole cell ClCa currents decay much faster than whole cell Ca2+ transients, as the decay of the current is determined by CAM kinase-induced inactivation of ClCa channels (207). Therefore, the decay of a STIC appears to be determined by the decline in [Ca2+], whereas the decay of the whole cell ClCa currents is regulated by [Ca2+] and CAM kinase.
FM AND AM OF CA2+ SPARKS: PHYSIOLOGICAL IMPLICATIONS
A number of factors regulate the apparent cytoplasmic Ca2+sensitivity of RyR channels, including phosphorylation state, pharmacological agents, and the SR Ca2+ load (Fig. 7). Modulation of RyR channels is manifest as a change in Ca2+spark frequency and/or amplitude. Ca2+ spark amplitude should depend on the open time of the RyR channels and the Ca2+ load of the SR, i.e., the driving force for Ca2+ (165). Increasing SR Ca2+ load (219) or elevating cAMP or cGMP (157) leads to an increase in Ca2+spark frequency in smooth muscle, without elevating global [Ca2+]i. Activators of protein kinase C (PKC) have the opposite effect (19).
FM and AM of Ca2+ sparks by SR Ca2+ load
Increases in the cytoplasmic [Ca2+] increase the open probability of smooth muscle RyR channels (Fig. 4 and 7) through a direct effect on cytosolic Ca2+-activating sites on the RyR channel protein (71). SR luminal [Ca2+] also modulates RyR channel activity, producing profound effects on both the frequency and the amplitude of Ca2+ sparks in smooth muscle (219). Skeletal and cardiac muscle RyR channels, incorporated into planar lipid bilayers, are activated by increases in luminal [Ca2+] (for reviews see Refs. 42, 51, 184). It has been proposed that skeletal and cardiac muscle RyR channels have luminal Ca2+ binding sites that interact with regulatory sites located in the cytosol, since luminal Ca2+ will only activate RyR channels when cytosolic ATP is present (174, 175). Recent evidence indicates that luminal Ca2+ that passes through the RyR channel may activate by interaction with cytosolic binding sites (72, 192). Regardless of the mechanism, increasing SR luminal Ca2+ elevates RyR channel open probability in bilayers (42) and Ca2+ spark frequency in intact myocytes (219).
In stomach smooth muscle cells isolated from the toad Bufo marinus, simultaneous measurements of SR [Ca2+], cytoplasmic [Ca2+], and Ca2+ sparks (or STOCs) have provided new insights into the relationship between SR Ca2+ load and Ca2+ spark properties (219). After emptying the SR with a bolus of caffeine, the frequency of Ca2+ sparks and STOCs increased steeply with the SR [Ca2+] (219), as has been reported in heart (30, 36, 164). Therefore, SR Ca2+ load appears to be an important regulator of Ca2+ spark frequency and amplitude in smooth muscle. However, the mechanism by which SR Ca2+regulates Ca2+ spark properties in smooth muscle remains to be elucidated.
Increase in Ca2+ Spark Frequency by cGMP and cAMP
Nitric oxide, a potent vasodilator produced by both the vascular endothelium and by clinically used nitrovasodilators, activates guanylyl cyclase, leading to increased production of cGMP and stimulation of cGMP-dependent protein kinase (PKG) (119). Other endogenous smooth muscle relaxants coupled to Gs proteins (e.g., calcitonin gene-related peptide, adenosine, β-adrenoceptor agonists) activate adenylyl cyclase and increase cAMP, resulting in the stimulation of cAMP-dependent protein kinase (PKA). Several mechanisms of action have been proposed for PKG and PKA, including increased Ca2+ uptake into the SR and activation of BKCachannels (158, 161, 186, 212). Indeed, cGMP- and cAMP-mediated relaxations in many types of smooth muscle are, in part, through activation of BKCa channels, causing membrane potential hyperpolarization, which closes voltage-dependent Ca2+channels and leads to smooth muscle relaxation (see Table5) (6, 7, 17, 32, 91, 92, 96, 107, 135,146, 152, 157-159, 168, 186, 189, 201, 212).
Smooth muscle relaxants that increase cGMP and cAMP have been shown to activate BKCa channels through direct phosphorylation effects on the channel protein (108, 161, 191, 208) and through elevation of Ca2+ spark frequency (86, 157). Agents that elevate cGMP or cAMP increase Ca2+ spark frequency, and therefore BKCa channel activity, threefold (86, 157). The direct effect of PKG and PKA on BKCa channel activity in intact cells appears to be much smaller (1.3-fold) than the activation caused by increased Ca2+ spark frequency. These results suggest that cyclic nucleotide-mediated smooth muscle relaxation may occur partially through an increase in the frequency of Ca2+ sparks and a subsequent increase in the activity of BKCa channels.
An increase in Ca2+ spark frequency following elevation of cGMP and cAMP may be due to phosphorylation of RyR channels, phospholamban, or an unknown regulatory protein (Figs. 7 and10). Phospholamban, when phosphorylated by PKG or PKA, dissociates from the SR Ca2+-ATPase, which leads to increased Ca2+ pumping and an elevated SR Ca2+ load, which increases Ca2+ spark frequency (165, 219). Recent results, using arteries from phospholamban-deficient mice, indicate that some of the frequency and amplitude modulation of Ca2+ sparks and STOCs by forskolin and sodium nitroprusside could occur through modulation of the SR Ca2+ load by phospholamban (47). RyR channels are also directly regulated by PKA (195), and this may contribute to the actions of cyclic nucleotides in smooth muscle. The emerging view of cyclic nucleotide action through BKCa channels is that the concerted action of PKG or PKA on phospholamban, RyR channels, and BKCa channels leads to an elevation of Ca2+ spark and STOC frequency and amplitude (Fig. 10).
Decrease in Ca2+ Spark Frequency by PKC
Vasoconstrictors [e.g., serotonin (58) or thromboxane A2 (194)] stimulate phospholipase C (PLC) through activation of an associated G protein (Gq). PLC activation induces the cleavage of membrane-associated phophatidylinositol 4,5-bisphosphate to diacylglycerol (DAG) and IP3. Each product of this pathway has its own distinct effect in the cell. IP3 induces the release of stored Ca2+ through binding to an IP3 receptor located on the membrane of intracellular stores, whereas DAG activates PKC. Ca2+release through IP3 receptors plays an important role to increase smooth muscle contractility, in particular to endogenous smooth muscle contractile agents (e.g., norepinephrine). PKC may regulate the contractility of smooth muscle cells in a variety of ways, including the phosphorylation of actin binding proteins (179) and ion channels (20, 21; for a review see Ref. 171) and increased Ca2+ sensitivity of contractile proteins (60, 147).
PKC may also contribute to the action of vasoconstrictors on smooth muscle through a novel mechanism, by decreasing the frequency of Ca2+sparks, which would lead to membrane depolarization and activation of voltage-dependent Ca2+ channels (Fig. 10) (19). Activators of PKC, phorbol 12-myristate 13-acetate and sn-1,2-dioctanoyl-glycerol, decreased Ca2+spark and STOC frequency by ∼60% in isolated cells from rat cerebral arteries (19). Activators of PKC also decreased STOC amplitude by ∼20%, which could be explained by a direct inhibition of BKCa channels (Fig. 10) (19). The inhibitory effect of PKC activators on Ca2+ spark frequency does not involve a reduction of SR Ca2+ load, suggesting that PKC directly inhibits RyR channels (Fig. 10) (19). The precise molecular mechanisms of PKC inhibition of Ca2+ sparks remain to be determined.
IP3 induced Ca2+ release has been described in terms of a hierarchical nomenclature based on the spatiotemporal characteristics of the observed events. Fundamental Ca2+release through a single IP3 receptor has been termed a blip [amplitude, 20–30 nM; spread, 1–3 μm; decay (t 1/2) = 45 ms (22)], whereas intermediate release due the recruitment of blips are termed Ca2+ puffs (22, 216). Propagating Ca2+ waves occur due to the summation of these smaller events. Although these elementary and intermediate events have not been described in smooth muscle cells, application of an agonist that elevates cytoplasmic levels of IP3 should activate fundamental Ca2+ release through IP3 receptors.
Vasoconstrictors have been shown to increase Ca2+ wave frequency and amplitude in the smooth muscle cells of rat tail artery, an effect that has been attributed to the activation of IP3receptors by IP3, and subsequent activation of RyR channels by Ca2+ released through IP3 receptors (81). IP3-induced Ca2+ release would increase the cytoplasmic [Ca2+] and would tend to increase Ca2+ spark frequency, particularly if RyR channels are situated close to IP3 receptors and sense the high local [Ca2+]. However, the reduction in SR Ca2+ load would tend to decrease Ca2+ spark frequency (219). The extent to which IP3-induced Ca2+ release increases or decreases Ca2+ spark activity (16, 52, 98, 106) would depend on a balance between an increase in Ca2+ near the cytoplasmic face of the RyR channels and the reduction in stored Ca2+ that would follow. It remains to be resolved how PKC and IP3 affect the cytoplasmic and SR luminal communication between RyR channels and IP3 receptors.
Ca2+ sparks occur in a wide variety of smooth muscle types and are caused by the activation of a cluster of RyR channels. In smooth muscle, the majority of Ca2+ sparks occur very close to the cell membrane, consistent with a role in signaling ion channels in the plasma membrane. The close proximity of Ca2+ spark sites to the plasma membrane support earlier studies indicating that, under certain circumstances, subsarcolemmal [Ca2+] could be significantly higher than global [Ca2+]i (215). Further support for this idea has come from a demonstrated dissociation between global [Ca2+]i and BKCachannel currents, which should be a measure of subsarcolemmal [Ca2+] (56, 153, 181).
Voltage-dependent Ca2+ channels, Ca2+ sparks (RyR channels), and BKCa channels act as a functional unit to regulate smooth muscle function (Figs. 1 and 2). In some cases ClCa channels may be regulated by Ca2+ sparks, adding another level of control to smooth muscle membrane potential (Fig. 2). Apamin-sensitive SKCa channels are also expressed in some types of smooth muscle (Table 4), and the effect of Ca2+ sparks on SKCa channels remains to be established.
Ca2+ entry through dihydropyridine-sensitive, voltage-dependent Ca2+ channels regulates Ca2+spark properties. The speed and efficacy of communication of voltage-dependent Ca2+ channels to activate Ca2+ sparks depends on proximity, and this issue is not resolved for smooth muscle. In contrast, it is clear that Ca2+ sparks in cardiac muscle are activated by local Ca2+ entry through voltage-dependent Ca2+channels (see Fig. 1). Ca2+ spark communication to BKCa channels is local, and the activator [Ca2+] for BKCa channels likely reaches 10–100 μM (153) within the 20-nm gap between the SR and plasma membrane. Activation of BKCa channels by Ca2+ sparks causes membrane potential hyperpolarization (e.g., Fig. 8), which closes voltage-dependent Ca2+channels, leading to a decrease in global intracellular [Ca2+] and smooth muscle relaxation (Figs. 1and 2) (102, 103, 140, 157). This negative-feedback pathway may be of particular importance for tonic smooth muscle.
The ability of Ca2+ entry through voltage-dependent Ca2+ channels to stimulate Ca2+ release through RyR channels and contribute to a global Ca2+ transient seems to be quite variable and is not consistent with local Ca2+ controlling Ca2+ release. The precise mechanisms and significance of Ca2+-induced Ca2+ release to global Ca2+ transients in smooth muscle are still unclear. It is likely that CICR does contribute Ca2+ for contraction in some types of phasic smooth muscle. Nonetheless, blocking Ca2+ release with ryanodine contracts most types of smooth muscle (10, 11, 38, 39, 44, 103, 105, 122, 140,155, 185, 209-211), suggesting that RyR channel-mediated negative-feedback control of [Ca2+]i dominates in smooth muscle (Fig. 2). It is possible that local Ca2+ release through RyR channels (Ca2+ sparks) provides a means for Ca2+ extrusion, thus decreasing cytoplasmic Ca2+ (the SBB hypothesis). RyR channels could exert negative-feedback control of [Ca2+]i through membrane potential hyperpolarization caused by activation of BKCa channels (Fig. 2) and SKCa channels and through inactivation of voltage-dependent Ca2+ channels (Fig. 2).
Modulation of Ca2+ spark frequency and amplitude by smooth muscle relaxants and constrictors appears to regulate smooth muscle membrane potential and hence contractility. PKA, PKG, and PKC have especially prominent roles in frequency and amplitude modulation of Ca2+ sparks. These kinases appear to exert coordinated and concerted actions on the key elements of Ca2+ signaling, including voltage-dependent Ca2+ channels, RyR channels, phospholamban, and BKCa channels (see Fig. 10). The exact molecular targets that determine the frequency and amplitude modulation of Ca2+ sparks by different kinases are still unresolved. However, it is clear that modulation of Ca2+ sparks provides a powerful means to regulate smooth muscle function.
We thank Adrian D. Bonev, Joseph Brayden, Laura Cartin, Del-Rae Eckman, Thomas J. Heppner, Gerald Herrera, David Hill-Eubanks, and George C. Wellman for helpful comments on the manuscript. We also thank the following persons who have contributed to research on smooth muscle Ca2+ sparks at the University of Vermont: Adrian D. Bonev, Heping Cheng, Maik Gollasch, Thomas J. Heppner, Gerald Herrera, Thomas Kleppisch, Harm J. Knot, J. B. Patlait, Guillermo J. Perez, Michael Rubart, L. Fernando Santana, Andrá S. Stevenson, and George C. Wellman. Unpublished observations have been kindly provided by Drs. George C. Wellman and L. Fernando Santana.
Address for reprint requests and other correspondence: M. T. Nelson, Dept. of Pharmacology, College of Medicine, The Univ. of Vermont, Burlington, VT 05405 (E-mail:).
This review was supported by National Institutes of Health Grants HL-44455, HL-51728, and DK-53832 (to M. T. Nelson) and HL-25675, HL-36974, and HL-61602 (to W. J. Lederer), by National Science Foundation Grants IBN-9631416 and BIR-9601682 (to M. T. Nelson), and by fellowships from the American Heart Association, New England (to J. H. Jaggar) and Vermont (to V. A. Porter) affiliates.
Present address of V. A. Porter: Dept. of Pediatrics, Univ. of Minnesota, 420 Delaware St. SE, Minneapolis, MN 55455.
- Copyright © 2000 the American Physiological Society