INTRODUCTION

The muscle relaxant dantrolene is a potent and specific inhibitor of skeletal muscle excitation-contraction (E-C)¹ coupling (1). Dantrolene (~10 µM) reduces skeletal muscle twitch force by approximately 75% (2) and shifts the sensitivity of contractile activation to higher voltages (3, 4), these effects being attributed to a partial block by dantrolene of Ca²⁺ release from the sarcoplasmic reticulum (SR) (1, 5, 6). In contrast to these pronounced effects on skeletal muscle, effects of dantrolene on cardiac muscle contractility are mild or absent (1, 5, 7). Clinically, dantrolene has proven effective in the treatment of malignant hyperthermia (MH), a potentially fatal genetic disorder of skeletal muscle E-C coupling in which exposure to volatile anesthetics triggers uncontrolled SR Ca²⁺ release, muscle contracture, and accelerated metabolism (8).

The molecular basis of the action of dantrolene remains undefined but is generally presumed to involve either direct or indirect inhibitory effects on ryanodine receptor (RYR) Ca²⁺ channels. To date, three RYR isoforms have been identified in mammalian tissues and are termed RYR1, RYR2, and RYR3 (9). In skeletal muscle, the RYR1 isoform is the major pathway for SR Ca²⁺ release during E-C coupling, and defects in these channels have been linked to MH susceptibility in pigs and in certain human families (8, 10). Moreover, recent evidence indicates that high-affinity [³H]dantrolene and [³H]ryanodine binding sites are localized to the same or closely associated skeletal muscle SR membrane fractions (11, 12). RYR1 channels thus constitute a likely target for the physiologic and therapeutic actions of dantrolene on skeletal muscle Ca²⁺ regulation. Nonetheless, the precise mechanism by which dantrolene may affect the activation of RYR1 channels has remained unclear. In particular, it is not yet clear how dantrolene may alter RYR1 activation by the physiologic effectors of these channels that have been identified in studies using isolated SR vesicles preparations (13). In addition, whether the effects of dantrolene may reflect a direct interaction with the RYR1 channel complex itself or rather require dantrolene binding to a separate and as yet unidentified regulatory molecule in the SR membrane remains controversial (12, 14, 15). Finally, whether the effects of dantrolene may be restricted to the RYR1 channel isoform or may extend to other intracellular Ca²⁺ release channel isoforms is also in question (16). The answers to these questions may be important not only in defining the mechanism of action of dantrolene but also in clarifying those aspects of RYR channel regulation that may be specific to skeletal muscle E-C coupling and that may be altered in MH.

In this study, we have examined the effects of dantrolene on the ⁴⁵Ca²⁺ release and [³H]ryanodine binding activity of isolated SR vesicles prepared from MH-susceptible (MHS) and normal pig skeletal muscle. The effects of dantrolene on cardiac SR vesicle ⁴⁵Ca²⁺ release and [³H]ryanodine binding were also examined. Our results demonstrate specific inhibitory effects of dantrolene on the functional activity of skeletal muscle RYR1 channels that are consistent with the effects of dantrolene on E-C coupling in intact muscle and suggest a mechanism of action in which a direct, high-affinity interaction of dantrolene with the RYR1 channel complex may limit the activation of this channel by calmodulin (CaM) and Ca²⁺.

EXPERIMENTAL PROCEDURES

Pigs homozygous for either the MHS or normal RYR1 allele were obtained from the University of Minnesota Experimental Farm and genotyped on the basis of the Arg⁶¹⁵  Cys MHS mutation (10). ⁴⁵Ca²⁺ and [³H]ryanodine were purchased from NEN Life Science Products (Boston, MA). Unlabeled ryanodine was from Calbiochem (La Jolla, CA). Dantrolene (1-[[5-(p-nitrophenyl)furfurylidine]amino]hydantoin sodium), AMPPCP (a nonhydrolyzable ATP analog), and porcine brain CaM were from Sigma. Azumolene and aminodantrolene were manufactured by Proctor & Gamble (Norwich, NY) and were kindly provided by the laboratories of Drs. E. Gallant (University of Minnesota) and J. Parness (University of Medicine and Dentistry of New Jersey). Stock solutions of dantrolene, azumolene, and aminodantrolene (typically 1 mM) were prepared in 70% methanol for use on the same day.

Skeletal muscle SR vesicles were isolated from the longissimus dorsi muscle of MHS and normal pigs as described previously (17). Briefly, vesicles obtained by differential ultracentrifugation of a muscle homogenate were extracted with 0.6 M KCl and subsequently fractionated on discontinuous sucrose gradients. The terminal-cisternae-derived (i.e. "heavy") SR vesicles that band at the 36-40% interface were collected and stored frozen at 70 °C. Cardiac muscle SR vesicles were isolated from porcine ventricular tissue (17). Following homogenization in 10 mM NaHCO₃, membranes were extracted in 0.6 M KCl, 20 mM Tris, pH 6.8, and then resuspended in 10% sucrose and stored frozen at 70 °C. All isolation buffers contained a mixture of protease inhibitors (aprotinin, leupeptin, and phenylmethylsulfonyl fluoride).

SR vesicles (10-15 mg of protein/ml) were passively loaded for 2-3 h at 36 °C in media containing 150 mM potassium propionate, 15 mM Pipes, pH 7.0, 5 mM ⁴⁵Ca²⁺, 1 µM CaM, 1 µg/ml aprotinin, 1 µg/ml leupeptin, and either 10 µM dantrolene or an equivalent volume of methanol vehicle (final concentration 0.7%). Dantrolene did not affect the amount of ⁴⁵Ca²⁺ loaded (42 ± 3.6 nmol/mg for skeletal SR vesicles, n = 6; 36 ± 6.3 nmol/mg for cardiac SR vesicles, n = 4). ⁴⁵Ca²⁺-loaded vesicles (2 µl) were placed on the side of a polystyrene tube containing 200 µl of a Ca²⁺ release medium (150 mM potassium propionate, 15 mM Pipes, pH 7.0, 8.6 mM EGTA, 3 mM AMPPCP, 3 mM MgCl₂ (free Mg²⁺ ~0.45 mM), 2 mM calcium acctate₂ (free Ca²⁺, ~100 nM), 1 µM CaM, 0.7% methanol, and 10 µM dantrolene, as indicated. Ca²⁺ release was initiated by rapid mixing and stopped at the indicated times (1-10 s using a metronome) by rapid dilution into 12 ml of a release-inhibiting medium (150 mM potassium propionate, 15 mM Pipes, pH 7.0, 10 mM EGTA, 5 mM MgCl₂, and 20 µM ruthenium red), followed immediately by filtration onto 0.45-µm Millipore HA membranes. The fraction of total loaded ⁴⁵Ca²⁺ that was not released after 10-s incubations in release media containing Ca²⁺ ionophore was considered background and was subtracted from all determinations. Estimates of the time required for vesicles to release half of their ⁴⁵Ca²⁺ contents (t_1/2) were based on fits to the equation, R = R_max × t/(t_1/2 + t), where R is Ca²⁺ released, R_max is the maximal Ca²⁺ release, and t is time.

SR vesicles were preincubated at 36 °C for 10 min in media containing dantrolene (0-30 µM, as indicated), 150 mM potassium propionate, 15 mM Pipes, pH 7.0, 3 mM AMPPCP, 2.7 mM MgCl₂ (free Mg²⁺, ~0.35 mM), 1 µM CaM, and a calcium acetate₂-EGTA buffer set to give the desired Ca²⁺ concentration (100 nM, except where otherwise indicated). Following addition of [³H]ryanodine (100 nM), SR vesicles were incubated for 90 min at 36 °C, collected on Whatman GF/B filters, and washed with 8 ml of ice-cold 150 mM KCl. Nonspecific binding was measured in the presence of 20 µM nonradioactive ryanodine. Data are expressed as percentages of the maximal [³H]ryanodine binding capacity of SR vesicle preparations as determined within each experiment in media containing 450 mM KCl, 10 mM ATP and 10 µM Ca²⁺ (12.2 ± 1.8 pmol/mg for skeletal muscle SR, n = 8; 4.0 ± 0.9 for cardiac SR, n = 3). Determinations of K_i values for the inhibition of [³H]ryanodine binding by dantrolene and dantrolene analogues were based on fits to the Hill equation.

RYR1 was isolated on sucrose gradients from Chaps-solubilized skeletal muscle SR essentially as described (18). For reconstitution of the isolated RYR channel complex into liposomes, gradient fractions containing RYR protein were pooled, concentrated, and diluted (0.1 mg SR protein/ml) into 50 mM NaCl, 50 mM KCl, 20 mM Tris, pH 7.4, 8 mg/ml phosphatidylcholine, 1% Chaps, and dialyzed for 48 h at 4 °C against 50 mM NaCl, 50 mM KCl, 20 mM Tris, pH 7.4, 2 mM -mercaptoethanol, 1 mM EGTA, 1 µg/ml aprotinin, 1 µg/ml leupeptin. Reconstituted RYR1-containing proteoliposomes were collected by centrifugation at 150,000 × g for 30 min and stored frozen at 70 °C. Prior to use, proteoliposome vesicles were slowly thawed, resuspended in 150 mM potassium propionate, 15 mM Pipes, pH 7.0, then sonicated 4 min in a bath sonicator. Proteoliposomes prepared from 2 MHS and 2 normal animals exhibited a mean [³H]ryanodine binding capacity of 154 ± 23 pmol/mg protein.

RESULTS

To investigate the effect of dantrolene on the Ca²⁺ permeability of skeletal muscle SR, ⁴⁵Ca²⁺ release from passively loaded SR vesicles was examined at 36 °C in a medium that contained putative physiological effectors of RYR channels (adenine nucleotide, Mg²⁺, Ca²⁺, and CaM, as described under "Experimental Procedures"). The effects of dantrolene were examined at a concentration (10 µM) that approximates therapeutic plasma levels (2) and is known to effectively inhibit the contractility of intact skeletal muscle fiber preparations (1).

In the absence of dantrolene, MHS SR vesicles (Fig. 1A) exhibited an increased rate of ⁴⁵Ca²⁺ release in comparison to normal SR vesicles (Fig. 1B). The t_1/2 for ⁴⁵Ca²⁺ release from MHS SR vesicles was thus only one fifth of that from normal SR vesicles (Table I), consistent with the well-documented increased MHS RYR1 channel activity under a variety of experimental conditions (8). In the presence of dantrolene, the rate of ⁴⁵Ca²⁺ release from both MHS and normal SR vesicles was decreased (t_1/2 for release increased ~3.5-fold for both MHS and normal SR, Table I). ⁴⁵Ca²⁺ release studies were also performed at 19 °C as previous studies have indicated that the effects of dantrolene on intact mammalian fiber preparations are absent below 20 °C (4, 19, 20). At 19 °C, the rate of ⁴⁵Ca²⁺ release from MHS SR vesicles was decreased and was similar in the presence and absence of dantrolene (Fig. 1C and Table I). These ⁴⁵Ca²⁺ release studies thus indicate that the temperature-dependent inhibitory effect of dantrolene on SR Ca²⁺ release that has been demonstrated in studies using intact muscle fiber preparations is also evident in studies using isolated SR vesicles.

[View Larger Version of this Image (18K GIF file)]

Table I. Effect of dantrolene on SR 45Ca2+ release 45Ca2+ release was determined as described under "Experimental Procedures" in the absence or presence of 10 µM dantrolene. Times at which SR vesicles released half of their Ca2+ stores are based on fits to the data presented in Figs. 1 and 6 (means ± S.E. of three separate determinations). MHS 36 °C Normal 36 °C MHS 19 °C Cardiac 36 °C 45Ca2+ release, t1/2 (s) Control 1.0  ± 0.1 5.0  ± 1.3 13  ± 0.7 58  ± 21 Dantrolene 3.7  ± 0.4 16  ± 1.4 16  ± 3.0 57  ± 8.5

Ryanodine binds with high-affinity to the open state of RYR channels and changes in [³H]ryanodine binding that occur in the presence of RYR effectors provide a useful index of changes in RYR channel activity (21). The effects of dantrolene on SR vesicle [³H]ryanodine binding were examined to determine if the demonstrated effects on SR Ca²⁺ release were attributable to specific effects on RYR1 channels and to obtain detailed information on the concentration-dependence of dantrolene inhibition. These experiments were performed in the same medium used in the ⁴⁵Ca²⁺ release studies except that the concentration of MgCl₂ was lowered from 3.0 to 2.7 mM to increase the SR vesicle [³H]ryanodine binding signal.

In the absence of dantrolene, the [³H]ryanodine binding activity of MHS SR vesicles was approximately 4-fold greater than that of normal SR vesicles (Fig. 2). [³H]Ryanodine binding to both MHS and normal SR vesicles was inhibited by dantrolene, consistent with an inhibitory effect of the drug on the functional activity of RYR1 channels. The concentration-dependence of dantrolene inhibition was monophasic, with maximal inhibition at dantrolene concentrations in the therapeutic range (3-30 µM). Calculated K_i values for dantrolene inhibition of [³H]ryanodine binding were similar for MHS (130 ± 32 nM) and normal (150 ± 18 nM) SR vesicles, suggesting that sensitivity to dantrolene was not affected by the MHS RYR1 mutation (Fig. 2, inset).

[View Larger Version of this Image (22K GIF file)]

To investigate the pharmacological specificity of dantrolene inhibition, the effects of the dantrolene analogues aminodantrolene and azumolene were also evaluated. [³H]Ryanodine binding to MHS SR vesicles was inhibited by both analogues (Fig. 3). However, inhibition by aminodantrolene required concentrations greater than 3 µM, whereas azumolene was a much more potent inhibitor (K_i = 84 ± 24 nM). The data in Figs. 2 and 3 thus indicate that dantrolene and its analogues inhibited skeletal muscle RYR activity with distinct potencies (azumolene  dantrolene  aminodantrolene). The maximal inhibitory effect of the different analogues was similar, however, and the extent of inhibition was never complete (~40% of activity in the absence of drug). In addition, in media containing dantrolene (2 µM), azumolene (30 nM to 30 µM) did not further inhibit SR vesicle [³H]ryanodine binding (Fig. 3, filled circles). These results indicate that dantrolene and its analogues may act at a common saturable site and that saturation of this site does not fully inhibit RYR1 channel activity.

[View Larger Version of this Image (22K GIF file)]

Previous studies have demonstrated that MHS RYR1 channels exhibit an increased activation by Ca²⁺ (22) and CaM (23), suggesting that an altered regulation of RYR1 channels by these physiologic effectors may underlie defects in Ca²⁺ homeostasis and E-C coupling in MHS muscle. We therefore investigated the effect of dantrolene on Ca²⁺- and CaM-activated [³H]ryanodine binding to MHS and normal SR vesicles.

The Ca²⁺ dependence of [³H]ryanodine binding to MHS and normal SR vesicles was determined in media containing 1 µM CaM (Fig. 4A). Ca²⁺ activation of [³H]ryanodine binding was biphasic, and the maximal extent of Ca²⁺ activation was approximately 2.3-fold greater for MHS than for normal SR vesicles. In addition, the apparent Ca²⁺ sensitivity of binding to MHS SR vesicles was shifted to lower Ca²⁺ concentrations; thus the half-maximally activating Ca²⁺ concentration (K_a) for MHS SR vesicles was approximately one-third of that for normal SR vesicles (Table II). In the presence of 10 µM dantrolene, the maximal Ca²⁺-activated [³H]ryanodine binding to both MHS and normal SR was decreased (Fig. 4A) and the K_a for Ca²⁺ was increased 2.4-3-fold (Table II).

[View Larger Version of this Image (23K GIF file)]

Table II. Effect of dantrolene on SR [3H]ryanodine binding [3H]Ryanodine binding was determined as described under "Experimental Procedures" in the absence or presence of 10 µM dantrolene. Half-maximally activating concentrations of Ca2+ and calmodulin are based on fits of the data presented in Figs. 4 and 6 to the Hill equation. ND, not determined due to insufficient activation by CaM. MHS Normal Cardiac Ka Ca2+ (nM) Control 120  ± 50 330  ± 62 980  ± 180 Dantrolene 300  ± 76 790  ± 82 1200  ± 340 Ka CaM (nM) Control 12  ± 2.7 16  ± 3.4 ND Dantrolene 16  ± 5.7 ND ND

The CaM dependence of activation of MHS and normal SR vesicle [³H]ryanodine binding was examined in media containing 100 nM Ca²⁺ (Fig. 4B). CaM activation of SR [³H]ryanodine binding was monophasic and was maximal at physiologic CaM concentrations (~1 µM) (24). MHS and normal SR vesicles exhibited a marked difference in the extent of CaM activation; thus, maximal CaM-activated [³H]ryanodine binding to MHS SR vesicles was approximately 3.7-fold greater than to normal SR vesicles (Fig. 4B). In contrast, the half-maximally activating CaM concentration for MHS and normal SR vesicles CaM were similar (K_a ~15 nM, Table II). Dantrolene (10 µM) markedly reduced the maximal CaM activation of both MHS and normal SR vesicle [³H]ryanodine binding (Fig. 4B); however, the K_a for CaM activation was unaffected by dantrolene (Table II).

To determine if dantrolene inhibition of RYR1 channels in these skeletal muscle SR vesicle preparations may be attributable to a direct action of dantrolene on the RYR1 channel complex, RYR1 protein was isolated from solubilized skeletal muscle SR vesicles and reconstituted into liposomes. Sucrose gradient fractions in which RYR1 was the major protein present were identified by electrophoretic analysis (Fig. 5A), and RYR1-containing proteoliposomes prepared from these fractions exhibited a greater than 12-fold enrichment of [³H]ryanodine binding capacity in comparison to skeletal muscle SR vesicle preparations ("Experimental Procedures").

[View Larger Version of this Image (28K GIF file)]

As reported previously (18), following isolation and reconstitution, the MHS RYR1 retained an increased [³H]ryanodine binding activity relative to the normal RYR1 (Fig. 5B). [³H]Ryanodine binding to both MHS and normal RYR1-containing proteoliposomes was completely inhibited by the RYR channel blocker ruthenium red (10 µM). Dantrolene (10 µM) also inhibited proteoliposome [³H]ryanodine binding (p < 0.005 for both MHS and normal), although as in the SR vesicle studies (Fig. 2), inhibition by dantrolene was incomplete. Furthermore, a lower dantrolene concentration (0.2 µM) was approximately half as effective as 10 µM dantrolene in inhibiting proteoliposome [³H]ryanodine binding (Fig. 5B). Thus the isolated, reconstituted skeletal muscle RYR1 protein displayed a sensitivity to inhibition by dantrolene that was similar to that of RYR1 in native SR vesicle preparations (Fig. 2).

To determine if dantrolene inhibition was selective for the skeletal muscle RYR1 channel isoform, the effects of dantrolene on cardiac muscle SR vesicle ⁴⁵Ca²⁺ release and [³H]ryanodine binding were also examined. These experiments were performed in the same media as the corresponding skeletal muscle SR vesicle studies (Figs. 1 and 4) and indicated that under these conditions cardiac RYR2 and skeletal RYR1 channels may exhibit previously undescribed differences in their activation by Ca²⁺ and CaM. Accordingly, Fig. 6A shows that in the presence of 100 nM Ca²⁺, the rate of ⁴⁵Ca²⁺ release from cardiac SR vesicles was slow in comparison with that of skeletal muscle SR vesicles (Fig. 1 and Table I). Similarly, [³H]ryanodine binding to cardiac SR vesicle was not significantly activated at a Ca²⁺ concentration of 100 nM (Fig. 6B). Thus, in comparison with skeletal muscle SR vesicles (Fig. 4A), the apparent threshold for Ca²⁺ activation of cardiac SR vesicle [³H]ryanodine binding was higher (Fig. 6B), and the K_a for Ca²⁺ activation of binding was shifted to the right (K_a = 980 nM Ca²⁺ for cardiac SR versus 330 nM Ca²⁺ for normal skeletal muscle SR, Table II).

[View Larger Version of this Image (13K GIF file)]

Because CaM activation accounted for the major fraction of skeletal muscle SR vesicle [³H]ryanodine binding at submicromolar Ca²⁺ concentrations (Fig. 4), it is possible that CaM contributed to the observed differences in the Ca²⁺ activation of cardiac and skeletal muscle SR [³H]ryanodine binding. Consistent with this possibility, cardiac SR [³H]ryanodine binding was not significantly activated by CaM (Fig. 6C), in marked contrast to the activating effect of CaM on skeletal muscle SR vesicle [³H]ryanodine binding (Fig. 4B). Furthermore, neither cardiac SR vesicle ⁴⁵Ca²⁺ release nor cardiac SR vesicle [³H]ryanodine binding were significantly affected by dantrolene (Fig. 6).

DISCUSSION

Although numerous agents are known to affect the in vitro activity of RYR channels (25), the effects of dantrolene on these channels are of particular interest for at least two reasons. First, dantrolene is among the most potent and selective modulators of skeletal muscle E-C coupling, and its effects on the voltage-dependent activation of skeletal muscle contraction have been well described (1, 3, 4). Second, dantrolene is perhaps the only RYR modulator used clinically, and its efficacy in preventing and reversing the pathophysiology of MH in both patients and experimental animals are well established (1, 8). Thus it is considered that a more complete understanding of the molecular mechanism of dantrolene action may not only provide insights into the in vivo mechanisms controlling RYR1 channels during skeletal muscle E-C coupling but also help to clarify the ways in which these mechanisms may be altered in MH.

Notwithstanding the clear and specific effects of dantrolene on skeletal muscle Ca²⁺ regulation in patients and in intact muscle fiber preparations, attempts to define its effects on RYR1 channels in more isolated preparations have to date yielded results that are inconsistent and inconclusive. For example, whereas some investigators have reported no effect of dantrolene on SR vesicle ⁴⁵Ca²⁺ release and [³H]ryanodine binding (12, 26, 27), others have reported inhibitory effects but at dantrolene concentrations that have typically exceeded those required to inhibit the contractility of intact muscle (28-30). Most recently, it was reported that RYR1 channels in planar lipid bilayer were either activated or inhibited by dantrolene, depending on the dantrolene concentration (14). In interpreting these various reports, it therefore remains unclear if the described effects of dantrolene on RYR channels have provided a meaningful reflection of the physiologic and therapeutic actions of the drug in intact muscle. Consequently, the molecular mechanism and locus of dantrolene action have remained undefined.

In this study, we have further examined the effects of dantrolene on SR vesicle ⁴⁵Ca²⁺ release and [³H]ryanodine binding using media that mimicked physiologic conditions in regard to temperature, ionic composition, and concentrations of RYR effectors. These conditions revealed a selective, monophasic inhibition of skeletal muscle RYR1 channels by dantrolene concentrations in the therapeutic range. Accordingly, inhibition of skeletal muscle SR [³H]ryanodine binding by dantrolene was maximal at 3-30 µM, and the concentration dependence of inhibition suggested action at a discrete high-affinity dantrolene site (K_i ~150 nM) (Fig. 2). Physiologically active analogues of dantrolene also inhibited skeletal muscle SR vesicle [³H]ryanodine binding (Fig. 3), and the relative potency of the different analogues reflected their potency in inhibiting skeletal muscle contractility (31-33). In addition, dantrolene inhibition of skeletal muscle SR vesicle ⁴⁵Ca²⁺ release (Fig. 1) exhibited a temperature dependence that was in agreement with the temperature dependence of dantrolene inhibition demonstrated in muscle fiber preparations from pigs (19), rats (4), and guinea pigs (20). Our observations thus demonstrate inhibitory effects of dantrolene on skeletal muscle RYR1 channels that are consistent with the actions of dantrolene on Ca²⁺ regulation in intact skeletal muscle.

In contrast to these inhibitory effects on the functional activity of skeletal muscle RYR1 channels, dantrolene did not significantly affect cardiac SR vesicle ⁴⁵Ca²⁺ release or [³H]ryanodine binding (Fig. 6). These findings are again consistent with the actions of dantrolene in vivo in that any effects of dantrolene on cardiac contractility are mild in comparison with effects on skeletal muscle, require higher concentrations of dantrolene, and can be attributed to effects of dantrolene at sites other than the cardiac SR (1, 5, 7). Our findings further indicate that the selective effect of dantrolene on skeletal muscle E-C coupling in vivo likely reflects a selective action of dantrolene at skeletal (RYR1) as compared with cardiac (RYR2) isoforms of the Ca²⁺ release channel. Whether action at RYR1 channels may also account for the effects of dantrolene in certain non-muscle cells (e.g. central neurons, Refs. 16, 34) is not yet clear.

Parness and co-workers (11, 12) have identified a specific, high-affinity [³H]dantrolene binding site in skeletal muscle SR vesicle preparations and have suggested that this site may reside either on the RYR1 channel itself or on some separate but closely associated regulatory molecule in the SR membrane. The possibility that dantrolene may bind to a receptor distinct from the RYR1 was supported by data which indicated that SR vesicle [³H]dantrolene binding was unaffected by both pharmacological modulators of RYR channels and by the RYR1 mutation associated with porcine MH (11, 12). In addition, peaks of [³H]dantrolene and [³H]ryanodine binding observed following fractionation of unsolubilized SR vesicles were consistent with the possibility that receptors for dantrolene and ryanodine may be present on overlapping but distinct vesicle populations (12).

To determine if the inhibitory effects of dantrolene on RYR1 activity that we observed may depend on dantrolene binding to a non-RYR1 receptor, SR vesicles were solubilized and RYR1 protein was isolated on sucrose gradients. Following reconstitution into liposomes, the isolated RYR1 protein displayed a sensitivity to dantrolene comparable with that of RYR1 in native SR (Fig. 5). This result strongly suggests that the mechanism of dantrolene inhibition does not require dantrolene binding to a receptor that is distinct and easily separable from RYR1 but rather involves a direct interaction of dantrolene with the RYR1 channel complex.

Inhibition of RYR1 channels by dantrolene was associated with clear and pronounced effects on the activation of these channels by CaM. Dantrolene decreased the maximal extent of CaM-activated [³H]ryanodine binding by more than half (Fig. 4B) but did not affect the K_a for CaM activation (Table II). This result suggests a noncompetitive inhibition of CaM activation by dantrolene; that is, rather than competing at a CaM activation site on the RYR1 channel, dantrolene may allosterically reduce the fraction of channels that may be activated by CaM. Dantrolene also decreased the maximal extent of Ca²⁺-activated [³H]ryanodine binding (Fig. 4A) and shifted the apparent K_a for Ca²⁺ activation to higher Ca²⁺ concentrations (Table II). Notably, the effects of dantrolene on the parameters of RYR1 activation by both CaM and Ca²⁺ opposed the effects of the MHS RYR1 mutation on channel activation by these effectors. This suggests that dantrolene may inhibit RYR1 channels via a mechanism that selectively counteracts the functional consequences of the MHS Arg⁶¹⁵  Cys mutation.

Fig. 7 incorporates the observed effects of dantrolene on RYR1 channels into a simple model of RYR1 activation during E-C coupling in MHS and normal muscle (36). The model postulates that RYR1 channels may exist in alternative closed states that differ in their sensitivity to activation by CaM and Ca²⁺. Dantrolene is postulated to increase the stability of the insensitive (Closed_i) state, thereby reproducing the observed decrease in the fraction of channels that may be activated by CaM and Ca²⁺ in the presence of dantrolene (i.e. noncompetitive inhibition, Fig. 4). Conversely, the MHS Arg⁶¹⁵  Cys mutation is postulated to increase the stability of the sensitive (Closed_s) state, thereby reproducing the observed increase in the fraction of MHS channels that may be activated by CaM and Ca²⁺ (Fig. 4; Refs. 22-23).

[View Larger Version of this Image (12K GIF file)]

The model further postulates that transitions between the Closed_i and Closed_s states may be controlled by depolarization (V) acting through the transverse tubule voltage sensor. In this way, the model reproduces the opposing effects of dantrolene and the Arg⁶¹⁵  Cys mutation on the voltage dependence of activation that have been documented in studies using intact muscle fibers (4, 35, 37, 38). Thus the model suggests that voltage and Ca²⁺/CaM may not activate RYR1 channels via readily separable mechanisms, but rather through a common mechanism, termed voltage-dependent Ca²⁺-induced Ca²⁺ release. The model thereby predicts that while RYR1 activation may be controlled by voltage, it may nevertheless display sensitivity to Ca²⁺ and to various effectors of Ca²⁺-induced Ca²⁺ release (39, 40). In such a mechanism, CaM may be a particularly critical effector, as CaM remains constitutively bound to RYR1 and effectively shifts the threshold for Ca²⁺ activation to resting Ca²⁺ concentrations (41). Thus CaM activation might contribute to a hallmark of skeletal muscle E-C coupling, i.e. rapid activation via a mechanism that operates in the absence of an initial increase in myoplasmic Ca²⁺ (42).

Previous work from our laboratory has suggested that, whereas RYR1 channels exhibit changeable Ca²⁺ activation properties, cardiac RYR2 channels appear to exhibit a more consistent Ca²⁺-activable state (43). In terms of the present model (Fig. 7), RYR2 channels are thus postulated to exhibit a more stable Closed_s state, and the lack of an effect of dantrolene on cardiac SR may be interpreted as a reflection of the increased stability of this state for the RYR2 isoform. Alternatively, the lack of an effect of dantrolene on cardiac SR may reflect the lack of a high-affinity dantrolene binding site on the RYR2 isoform. An additional observation from the present studies was that in contrast to RYR1 channels, RYR2 channels were not activated by physiological CaM (Fig. 6C). This apparent difference in regulation by CaM may therefore represent an important and previously undescribed functional difference between the RYR1 and RYR2 isoforms. Indeed, if CaM is necessary for RYR1 activation at resting Ca²⁺ concentrations (above), then the absence of CaM activation of RYR2 would be consistent with a hallmark of cardiac E-C coupling, i.e. activation via a mechanism that is strictly dependent on an initial increase in Ca²⁺.

In evaluating our model, it will be important to verify differences in the functional interaction of CaM with cardiac and skeletal muscle RYR isoforms. In addition, it will be important to compare the functional consequences of other human MHS RYR1 mutations with those of the Arg⁶¹⁵  Cys mutation (44), as the model predicts that rather than altering particular ligand binding sites, MH mutations may affect a global conformational transition in the RYR1 channel that is controlled by transverse tubule depolarization and results in increased channel activation by CaM and Ca²⁺. Finally, it will be useful to further define the molecular mechanism of action of RYR-isoform-specific modulators of E-C coupling, such as dantrolene.