Regulation of skeletal muscle Ca2+ release channel (ryanodine receptor) by Ca2+ and monovalent cations and anions.

The effects of ionic composition and strength on rabbit skeletal muscle Ca2+ release channel (ryanodine receptor) activity were investigated in vesicle-45Ca2+ flux, single channel and [3H]ryanodine binding measurements. In <0.01 μM Ca2+ media, the highest 45Ca2+ efflux rate was measured in 0.25 M choline-Cl medium followed by 0.25 M KCl, choline 4-morpholineethanesulfonic acid (Mes), potassium 1,4-piperazinediethanesulfonic acid (Pipes), and K-Mes medium. In all five media, the 45Ca2+ efflux rates were increased when the free [Ca2+] was raised from <0.01 μM to 20 μM and decreased as the free [Ca2+] was further increased to 1 mM. An increase in [KCl] augmented Ca2+-gated single channel activity and [3H]ryanodine binding. In [3H]ryanodine binding measurements, bell-shaped Ca2+ activation/inactivation curves were obtained in media containing different monovalent cations (Li+, Na+, K+, Cs+, and choline+) and anions (Cl−, Mes−, and Pipes−). In choline-Cl medium, substantial levels of [3H]ryanodine binding were observed at [Ca2+] <0.01 μM. Replacement of Cl− by Mes− or Pipes− reduced [3H]ryanodine binding levels at all [Ca2+]. In all media, the Ca2+-dependence of [3H]ryanodine binding could be well described assuming that the skeletal muscle ryanodine receptor possesses cooperatively interacting high-affinity Ca2+ activation and low-affinity Ca2+ inactivation sites. AMP primarily affected [3H]ryanodine binding by decreasing the apparent affinity of the Ca2+ inactivation site(s) for Ca2+, while caffeine increased the apparent affinity of the Ca2+ activation site for Ca2+. Competition studies indicated that ionic composition affected Ca2+-dependent receptor activity by at least three different mechanisms: (i) competitive binding of Mg2+ and monovalent cations to the Ca2+ activation sites, (ii) binding of divalent cations to the Ca2+ inactivation sites, and (iii) binding of anions to specific anion regulatory sites.

In skeletal muscle, an intracellular Ca 2ϩ conducting channel releases Ca 2ϩ from the sarcoplasmic reticulum (SR) 1 in response to an action potential, to bring about muscle contraction (1)(2)(3). The Ca 2ϩ release channels are also known as ryanodine receptors (RyR) because they can bind the plant alkaloid ryanodine with high affinity and specificity. The skeletal muscle RyR has been purified as a 30 S protein complex comprising four large (ryanodine receptor, M r 565,000) and four small (FK506-binding protein, M r 12,000) subunits, and shown to be regulated by various endogenous and exogenous effector molecules including Ca 2ϩ , Mg 2ϩ , ATP, calmodulin, caffeine, and ryanodine (4 -6).
Skeletal muscle RyR activity is affected by the ionic strength and composition of the assay media. An increase in KCl or NaCl concentration stimulates Ca 2ϩ release from SR vesicles and increases [ 3 H]ryanodine binding (7)(8)(9)(10)(11). A stimulation of [ 3 H]ryanodine binding (9) and slowing of single channel gating (12) by sucrose in the presence of salt suggests that the osmolarity and viscosity of the assay media may play a role in determining channel activity. Anions often classified as chaotropic ions (Cl0 4 Ϫ , SCN Ϫ , I Ϫ , NO 3 Ϫ ) (13,14) and inorganic phosphate anions (15) stimulate Ca 2ϩ release channel activity and [ 3 H]ryanodine binding, whereas replacement of Cl Ϫ by gluconate Ϫ decreases SR Ca 2ϩ release and [ 3 H]ryanodine binding (13). These results suggest that monovalent cations and anions as well as osmolarity or viscosity may modulate skeletal muscle RyR activity. However, the mechanism(s) by which these ions affect the SR Ca 2ϩ release channel have remained unclear.
Here, we describe the effects of monovalent cations and anions on 45 Ca 2ϩ efflux from and [ 3 H]ryanodine binding to rabbit skeletal muscle SR vesicles. The effects of ionic strength were also determined in single channel measurements. Our results indicate that RyR activity may be affected by the binding of cations to Ca 2ϩ regulatory sites and anions to anion regulatory sites, and that there is a strong functional interaction between the two classes of regulatory sites. 45 Ca 2ϩ from ICN Biomedicals. Unlabeled ryanodine was obtained from Calbiochem, and leupeptin and Pefabloc (a protease inhibitor) from Boehringer Mannheim. All other chemicals were of analytical grade.

Materials-[ 3 H]Ryanodine was purchased from DuPont NEN and
Preparation of SR Vesicles-"Heavy" SR vesicle fractions enriched in [ 3 H]ryanodine binding and Ca 2ϩ release channel activities were prepared in the presence of protease inhibitors (100 nM aprotinin, 1 M leupeptin, 1 M pepstatin, 1 mM benzamidine, 0.2 mM phenylmethylsulfonyl fluoride) as described (16). The maximum number of highaffinity [ 3 H]ryanodine-binding sites determined under optimal binding conditions (17) ranged from 11 to 23 pmol/mg protein, depending on the preparation. 45 Ca 2ϩ Efflux Measurements-SR vesicles (5-10 mg of protein/ml) were passively loaded for 60 min at 23°C with 2 mM 45 Ca 2ϩ in media containing 20 mM imidazole, pH 6.8, protease inhibitors (0.2 mM Pefabloc, 20 M leupeptin), and different salts as described (18). 45 Ca 2ϩ efflux was initiated by diluting vesicles 1:300 into efflux media that contained the salt used in the incubation step, and stopped by placing 0.4-ml aliquots at various times on a 0.45-m filter (type HA, Millipore). Filters were washed with a quench solution containing 20 mM imidazole, pH 6.8, the salt used in the incubation step, 10 mM Mg 2ϩ , 20 M ruthenium red, and 0.2 mM EGTA. Rapid 45 Ca 2ϩ efflux was determined with a Biologic Rapid Filtration system (Meylan, France). Aliquots of the passively loaded vesicles (about 10 g of protein) were placed on 0.65 m (type DA) Millipore filters. The filters were prewashed for 30 s with 3 ϫ 1 ml of a medium containing 20 mM imidazole, pH 6.8, 1 mM choline EGTA, 1 mM MgCl 2 , and the salt used in the incubation step. Vesicles on the filters were then washed for 0.05-3 s with release media containing 20 mM imidazole, pH 6.8, the salt used in the incubation step, and different concentrations of free Ca 2ϩ . Radioactivity remaining with the vesicles on the filters was determined by liquid scintillation counting. The time course of 45 Ca 2ϩ efflux from the Ca 2ϩ -permeable vesicle population was obtained by subtracting the amount not readily released (18).
Single Channel Measurements-Single channel measurements were performed by fusing proteoliposomes containing the purified skeletal muscle Ca 2ϩ release channel with Mueller-Rudin-type bilayers as described (19). Single channels were recorded in symmetric KCl buffers containing the additions indicated in the text. Electrical signals were filtered at 4 kHz, digitized at 20 kHz, and analyzed as described (19).
[ 3 H]Ryanodine Binding-Unless otherwise indicated, samples were incubated at 12°C with 1 nM [ 3 H]ryanodine in media containing 20 mM imidazole, pH 7.2, 0.2 mM Pefabloc, 20 M leupeptin, and the indicated salt, 0.45 mM 1,2-bis(2-aminophenoxy)ethanetetraacetic acid, 0.9 mM nitrilotriacetic acid, and Ca 2ϩ concentrations to yield the indicated free Ca 2ϩ concentrations. Nonspecific binding was determined using a 1000fold excess of unlabeled ryanodine. A relatively low incubation temperature of 12°C was used to minimize receptor inactivation during the binding reaction. At 12°C, an incubation time of 90 -120 h was generally sufficient to obtain close to maximum [ 3 H]ryanodine binding (see "Results"). After 90 -120 h, aliquots of the samples were diluted with 20 volumes of ice-cold water and placed on Whatman GF/B filters soaked with 2% polyethyleneimine. Filters were washed with three 5-ml volumes of ice-cold 0.1 M KCl, 1 mM K-Pipes, pH 7.0, medium, and the radioactivity remaining on the filters was determined by liquid scintillation counting to obtain bound [ 3 H]ryanodine.
Other Biochemical Assays-Protein concentrations were determined by the Lowry method using bovine serum albumin as the protein standard. Free Ca 2ϩ concentrations were obtained by including in the solutions the appropriate amounts of Ca 2ϩ and Ca 2ϩ chelators as determined using the stability constants and computer program published by Shoenmakers et al. (20). Free Ca 2ϩ concentrations of Ͼ1 M were verified with the use of a Ca 2ϩ selective electrode (World Precision Instruments, Inc., Sarasota, FL), except in the 1 M solutions in which the Ca 2ϩ electrode measurements were not possible at Ͻ10 M free Ca 2ϩ because of limited electrode selectivity.
Data Analysis-The Ca 2ϩ dependence of [ 3 H]ryanodine binding was analyzed assuming that the skeletal muscle Ca 2ϩ release channel possesses cooperatively interacting high-affinity Ca 2ϩ activation and lowaffinity Ca 2ϩ inactivation binding sites. A simple scheme used to describe the Ca 2ϩ dependence of channel activity in media containing inorganic monovalent cations was, R L | ; M Ca 2ϩ A Ca L | ; mM Ca 2ϩ Ca I Ca SCHEME 1 In the above scheme (Scheme 1), the Ca 2ϩ release channel is assumed to have high-affinity Ca 2ϩ activation and low-affinity Ca 2ϩ inactivation sites. The channel is present in its closed Ca 2ϩ -free form, designated R at [Ca 2ϩ ] Ͻ 0.1 M, and its Ca 2ϩ -activated (A Ca ) and Ca 2ϩ -inactivated ( Ca I Ca ) forms at M and mM Ca 2ϩ concentrations, respectively. The tetrameric Ca 2ϩ release channel contains cooperatively interacting Ca 2ϩ activation sites and Ca 2ϩ inactivation sites (see "Results"), however, only one Ca 2ϩ activation and one Ca 2ϩ inactivation site are shown.
Ryanodine binding (and, by extension, channel activity) was fitted by the product of an activation and an inactivation variable, each related to [Ca 2ϩ ] by the Hill formalism, where which formalizes the assumption that choline ϩ is a weak, noncooperative Ca 2ϩ agonist of the Ca 2ϩ release channel.
In the competition studies, [ 3 H]ryanodine binding was fitted with the equations, where B is the [ 3 H]ryanodine binding value at a given [Ca 2ϩ ], B o the binding maximum in the absence of the inhibitor (I), K Ca the Ca 2ϩ activation constant, and K i the inhibition constant of the inhibitor. Results are given as means Ϯ S.D. with the number of experiments in parentheses. Unless otherwise indicated, significance of differences of data was analyzed with Student's unpaired t test. Differences were regarded to be statistically significant at p Ͻ 0.05.

RESULTS
SR Vesicle-45 Ca 2ϩ Efflux Measurements-In preliminary experiments, the effects of ionic composition on Ca 2ϩ release channel activity were assessed in SR vesicle-45 Ca 2ϩ efflux measurements. Fig. 1   Pipes Ϫ as an anion. In the presence of 5 mM Mg 2ϩ at Ͻ0.01 M Ca 2ϩ , a time of 100 s or more was required for the vesicles to release half their 45 Ca 2ϩ stores. Omission of Mg 2ϩ from the low Ca 2ϩ media resulted in a significant increase in the 45 Ca 2ϩ efflux rates. The highest rate was measured in choline-Cl medium followed by KCl, choline-Mes, K-Pipes, and K-Mes medium. In all five media, the 45 Ca 2ϩ efflux rates were increased when the free [Ca 2ϩ ] was raised from Ͻ0.01 to 20 M, and decreased as the free [Ca 2ϩ ] was further raised to 1 mM. In agreement with previous vesicle ion flux measurements (10,11,13,15,18,21,22), these results suggest that the Ca 2ϩ release channel is activated by micromolar concentrations of Ca 2ϩ , inhibited by millimolar concentrations of Ca 2ϩ , and furthermore, that the channel's activity is profoundly affected by the ionic composition of the Ca 2ϩ efflux media.
Ca   (Table I) shows a qualitatively similar dependence on [Ca 2ϩ ], thus supporting the idea that under the above ionic conditions [ 3 H]ryanodine binding correlated well with channel activity.  Table II shows the averaged Hill constants and coefficients of several experiments. The data suggest that changes in the apparent affinity as well as cooperativity of the Ca 2ϩ -activating and Ca 2ϩ -inactivating sites contribute to the different levels of [ 3 H]ryanodine binding observed in Fig. 2, A and B. The significance of the changes evidenced in Table II will be discussed specifically for each intervention.
Scatchard analysis indicated the presence of a single highaffinity [ 3 H]ryanodine-binding site (not shown). Changes in binding affinity (K D ) without major changes in B max value were observed in KCl, K-Mes, and choline-Mes media (all at 20 M Ca 2ϩ ) and choline-Cl medium (at Ͻ0.01 and 20 M Ca 2ϩ ) (Table  III). These results suggest that the different binding values of  was maintained at 20 M, as at this level of free Ca 2ϩ close to maximum ryanodine binding was observed (Fig. 2). Inspection of the four current traces of were obtained when channels were recorded at ϩ40 or Ϫ40 mV holding potential (Fig. 3B). These results suggest that skeletal muscle Ca 2ϩ release channel activity is highly sensitive to the ionic strength of the recording solutions. Fig. 4A (Table II). A linear correlation coefficient of 0.91 (n ϭ 24) indicates that the increase in K i was highly significant. Two important differences were, however, that substantial levels of [ 3 H]ryanodine binding were measured at [Ca 2ϩ ] Ͻ10 Ϫ8 M in the choline-Cl but not KCl media, and second that the apparent Ca 2ϩ affinity of the Ca 2ϩ activation sites monotonously increased as the [choline-Cl] was raised from 0.1 to 1.0 M. A linear correlation coefficient of 0.78 (n ϭ 24) indicates that the increase in affinity was highly significant.
Effects of AMP and Caffeine on Ca 2ϩ -dependence of [ 3 H]Ryanodine Binding-Ca 2ϩ -gated Ca 2ϩ release channel activity is affected by various endogenous and exogenous effectors such as adenine nucleotides and caffeine (4 -6). In this study, we used AMP rather than ATP or a nonhydrolyzable ATP analog because AMP, in contrast to adenine triphosphates, binds Ca 2ϩ with a negligible affinity. Fig. 5 shows that the addition of AMP to 0.25 M KCl medium resulted in an increase in [ 3 H]ryanodine binding. This increase could be accounted for by a small (not significant) increase in the apparent affinity of the receptor activation sites and 3-4-fold (significant) decrease in the apparent affinity of the inactivation sites for Ca 2ϩ (Fig. 5, Table  II). Caffeine (20 mM) shifted the Ca 2ϩ activation curve to the left, by increasing the apparent affinity of the Ca 2ϩ activation and Ca 2ϩ -inactivation sites by a factor of 15 and 1.7, respectively (Fig. 5, Table II). An additional effect of caffeine was to decrease the cooperativity of Ca 2ϩ activation and inactivation.
Interaction of Mg 2ϩ and Monovalent Cations with High-af-  Table II. finity Ca 2ϩ Activation Sites-We considered the possibility that monovalent cations inhibit [ 3 H]ryanodine binding by competing with Ca 2ϩ for the high-affinity Ca 2ϩ activation sites. Initially, we tested the effects of Mg 2ϩ , which is known to inhibit Ca 2ϩ release channel activity by interacting with the Ca 2ϩ activation sites (5). In these studies, we took advantage of the observation that significant levels of [ 3 H]ryanodine binding were observed in choline-Cl media containing submicromolar [Ca 2ϩ ] (Fig. 4B). It could be argued that the effects of the monovalent cations are best studied in the absence of another monovalent cation. However, at submicromolar [Ca 2ϩ ] the presence of an activating anion (Ն0.25 M Cl Ϫ ) is required to observe satisfactory levels of [ 3 H]ryanodine binding (Fig. 2).
The inhibitory effects of Mg 2ϩ (Fig. 6) and K ϩ (Fig. 7) were assessed in 0.5 M choline-Cl media at different free Ca 2ϩ concentrations that were expected to partially activate [ 3 H]ryanodine binding but to have only negligible inhibitory effects. To interpret the [ 3 H]ryanodine binding data, three simple alternative types of inhibition were considered, namely that Mg 2ϩ and K ϩ were competitive, noncompetitive, and uncompetitive inhibitors. The binding data could not be fitted assuming non-or uncompetitive inhibition (not shown) but could be well fitted when it was assumed that Mg 2ϩ (Fig. 6) and K ϩ (Fig. 7) inhibited [ 3 H]ryanodine binding by a competitive mechanism (formalized by Equations 3 and 4 under "Experimental Procedures"), according to which the two cations bind to the Ca 2ϩ activation site, but fail to activate the channel. Table IV summarizes the derived Ca 2ϩ activation and inhibition constants and coefficients for Mg 2ϩ and four monovalent cations (Li ϩ , Na ϩ , K ϩ , and Cs ϩ ). Similar Ca 2ϩ activation constants and Hill coefficients were obtained in all five media. Among the cations tested, Mg 2ϩ was most effective in inhibiting [ 3 H]ryanodine binding (K i ϭ 0.013 mM). For the monovalent cations, the order of effectiveness was Li ϩ Ͼ Na ϩ Ͼ K ϩ Ͼ Cs ϩ . Hill coefficients of ϳ1.6 suggested that the monovalent cations inhibited the channel by a cooperative interaction involving at least two cations. The higher affinity of Na ϩ for the Ca 2ϩ activating site(s) provided at least a partial explanation for the observation that higher [Ca 2ϩ ] were required to half-maximally activate [ 3 H]ryanodine binding in NaCl than in KCl or CsCl media (Table II). The inhibitory effects of Mg 2ϩ and the monovalent cations were also tested in 0.5 M choline-Cl media containing 5 mM AMP. Table IV shows that the addition of AMP resulted in a 1.1-2.2fold increase in the affinity of the Ca 2ϩ activation sites for Ca 2ϩ . No major changes in the inhibition constants and coefficients were observed.
Interaction of Divalent Cations with Low-affinity Inhibitory Sites-The decline of SR Ca 2ϩ release activity and [ 3 H]ryanodine binding at elevated [Ca 2ϩ ] indicates that the Ca 2ϩ release channel possesses low affinity inactivation sites (Figs. 1, 2, and  4). The divalent cation specificity of these sites was tested in media that contained 0.1 M KCl, 0.5 M KCl, or 0.5 M choline-Cl, 5 mM AMP, a close to maximally activating [Ca 2ϩ ] (Fig. 4; 20 M at 0.1 M and 50 M at 0.5 M), and different concentrations of Mg 2ϩ , Ca 2ϩ , Sr 2ϩ , and Ba 2ϩ . Essentially identical inhibition patterns were obtained for the four divalent cations (Fig. 8) Fig. 8 suggest that the low-affinity inhibitory ryanodine  (Fig. 10). In media containing 20 M Ca 2ϩ (Fig. 9), we used Pipes Ϫ rather than Mes Ϫ as the buffer anion because [Ca 2ϩ ] binds Mes Ϫ with K D ϳ 0.2 M. In the experiments using low concentrations of Ca 2ϩ (Fig. 10), we preferred to use Mes Ϫ because it is more fully present in its anionic form at pH 7.2. In Fig. 9, the effects of Pipes Ϫ on [ 3 H]ryanodine binding were examined in the presence of K ϩ because in the presence of K-Pipes but not choline-Pipes a nearly complete inhibition of [ 3 H]ryanodine binding could be observed at micromolar [Ca 2ϩ ] (not shown). The binding data could be reasonably well fitted assuming that Pipes Ϫ inhibited [ 3 H]ryanodine binding by competing with Cl Ϫ for an anion regulatory site. The derived Hill activation (for Cl Ϫ ) and inhibition (for Pipes Ϫ ) constants and coefficients are shown in Table IV. Qualitatively similar data were obtained with Mes Ϫ as the competing buffer (not shown).
In Fig. 10 In Fig. 10, A and B, the continuous lines were obtained assuming competitive inhibition, with Mes Ϫ inhibiting Ca 2ϩ binding to the Ca 2ϩ activation site(s). A good fit was obtained at the elevated [Ca 2ϩ ], whereas the data at the lower [Ca 2ϩ ] deviated by a factor of up to 1.3 from the calculated values. Table IV summarizes the derived Hill constants and coefficients. No reasonable fits were obtained when it was assumed that Mes Ϫ was a noncompetitive or uncompetitive inhibitor. In the presence of 5 mM AMP, a 1.8-fold decrease in the activation constant was obtained without a change in the Ca 2ϩ inactivation constant (Table IV). DISCUSSION The goal of the present study was to characterize the action of monovalent cations and anions on the RyR/Ca 2ϩ release channel of rabbit skeletal muscle. Among the various endogenous effectors of the RyR, Ca 2ϩ is widely accepted to play a pivotal role. This study shows that inorganic monovalent cations affect RyR activity by competitive binding to the receptor's Ca 2ϩ activation sites. Second, our results indicate that anion-  Table IV. specific binding sites play an important role in regulating RyR activity by modifying the apparent Ca 2ϩ affinity of the receptor's Ca 2ϩ regulatory sites.
The effects of ionic composition and ionic strength on skeletal muscle Ca 2ϩ release channel activity were monitored with [ 3 H]ryanodine binding, SR vesicle-45 Ca 2ϩ flux, and single channel measurements. Although multiple ryanodine-binding sites and a complex interaction of ryanodine with these sites have been reported (17,(23)(24)(25), the binding kinetics are relatively straightforward when low ryanodine concentrations are used, concentrations that limit binding to a single high-affinity receptor site. Ryanodine is generally thought to preferentially bind to the open channel and, as observed in the present study, binding is thought to be affected by Ca 2ϩ and other effectors similarly as SR Ca 2ϩ release or single channel activities. However, it is unlikely that ryanodine binding and channel activity are regulated in exactly the same way, because of the different time scales on which the channel gates (s to ms) and binds [ 3 H]ryanodine (minute to hour). Although [ 3 H]ryanodine binding provides less direct information on channel activity than single channel measurements, we chose to rely mostly on [ 3 H]ryanodine binding measurements because they allowed us to examine various ionic conditions. The regulation of the skeletal muscle Ca 2ϩ release channel was examined in the presence of nM to mM [Ca 2ϩ ] in media containing different mono-and divalent cations and anions. Assuming that the binding of Ca 2ϩ to high affinity sites (K a Ͻ 1 M) activates the channel, while binding of Ca 2ϩ to separate low affinity sites (K i Ͼ 50 M) inactivates it (10, 11, 13, 15, 18, 21, 22, this study), the results of our experiments can be described by expanding the scheme (Scheme 1) shown under "Experimental Procedures" as follows, x 2ϩI* Ca SCHEME 2 In the above scheme (Scheme 2), it is assumed that the RyR/ Ca 2ϩ release channel may be present in states of different Ca 2ϩ binding affinities. At a low [Cl Ϫ ] or in the presence of a competing inhibitory anion (Y Ϫ ) Ca 2ϩ binds at Ͻ1 M free Ca 2ϩ to the Ca 2ϩ activation sites of a Ca 2ϩ -free RyR (R) to yield a Ca 2ϩ -activated receptor (A Ca ), and binds at Ͼ50 M free Ca 2ϩ to A Ca to yield a Ca 2ϩ -inactivated receptor ( X 2ϩ]I Ca * ). An increase in [Cl Ϫ ] results in receptor forms (R * , A Ca * , X 2ϩI Ca * ) that are characterized by an increased Ca 2ϩ affinity of the Ca 2ϩ activation sites and decreased Ca 2ϩ affinity of the Ca 2ϩ inactivation sites. Mg 2ϩ and monovalent cations (in parentheses) are competitive inhibitors that inhibit the formation of the A Ca and A Ca * receptor states by competing with Ca 2ϩ for the Ca 2ϩ activation sites. In the above scheme, in addition to Ca 2ϩ , Mg 2ϩ and other divalent cations (X 2ϩ ) inhibit the receptor by binding to the Ca 2ϩ inactivation sites. The scheme further proposes that buffer anions (Mes Ϫ and Pipes Ϫ ) deter the formation of the activated R * , A Ca * , and X 2ϩI Ca * receptor forms by competing with Cl Ϫ for anion regulatory site(s). As shown in this study, the Ca 2ϩ release channel contains cooperatively interacting Ca 2ϩ activation sites and Ca 2ϩ inactivation sites. The above scheme has been simplified by showing only one Ca 2ϩ activation and one Ca 2ϩ inactivation site each.
The effects of monovalent cations on the Ca 2ϩ dependence of [ 3 H]ryanodine binding were analyzed using the chloride salts of Li ϩ , Na ϩ , K ϩ , Cs ϩ , and choline ϩ . Ca 2ϩ activated [ 3 H]ryanodine binding by a cooperative interaction with the highest apparent affinity in choline-Cl medium followed by CsCl, KCl, NaCl, and LiCl medium. The studies showed that choline ϩ behaves like a weak Ca 2ϩ agonist of the channel, and inorganic monovalent cations lower the apparent Ca 2ϩ affinity by competitive binding to the Ca 2ϩ activation sites. Hill coefficients greater than 1 (Tables II and IV) suggest that Ca 2ϩ activates and inorganic monovalent ions inhibit the skeletal muscle Ca 2ϩ release channel by cooperative interactions. Recently, the effects of [KCl] on the Ca 2ϩ activation profile were also examined by determining the permeation of choline ϩ in light scattering measurements with SR vesicles present in choline-Cl media (11). At variance with the present study, an increase in [KCl] from 0 to 1 M shifted the Ca 2ϩ activation profile to higher Ca 2ϩ concentrations. The decreases in the apparent Ca 2ϩ affinities for both the Ca 2ϩ activation and Ca 2ϩ inactivation sites were  Table IV. explained by assuming competition between K ϩ and Ca 2ϩ at the Ca 2ϩ binding sites of the channel. The effects of [Cl Ϫ ] on channel activity were not considered by Kasai et al. (11). The present study shows that both the actions of the monovalent cations and anions need to be taken into account to understand the way in which the ionic milieu modulates activation and inactivation of the skeletal muscle Ca 2ϩ release channel by Ca 2ϩ .
In agreement with vesicle flux studies (18,22), a competitive binding to the high-affinity Ca 2ϩ activation sites was also observed for Mg 2ϩ . The inhibition constant for Mg 2ϩ was lower than those for the monovalent cations by more than 2 orders.
For K ϩ , the major monovalent cation in muscle, the inhibition constant determined was 3000-fold higher than for Mg 2ϩ . The intracellular free [K ϩ ] in skeletal muscle is about 100 times higher than that of Mg 2ϩ . Therefore, the channel's Ca 2ϩ activation sites are likely occupied to a greater extent by Mg 2ϩ than by K ϩ at rest. However, occupation of some sites by K ϩ may be of physiological importance because Ca 2ϩ may bind faster to channel sites occupied by K ϩ than sites occupied by more tightly bound Mg 2ϩ .
The interaction of di-and monovalent cations with the lowaffinity channel inactivation sites was less amenable to analysis because of their concurrent interaction with the Ca 2ϩ activation sites. The specificity of the inactivation sites with regard to Ca 2ϩ and monovalent cations was determined in 0.25 M Cl Ϫ media containing different monovalent cations (Fig. 2, A and  B). Analysis of [ 3 H]ryanodine binding data suggests that Ca 2ϩ binds with a higher apparent affinity to the inactivation sites in KCl and CsCl medium than in NaCl or choline-Cl medium. The Hill inactivation coefficients ranged from 1.3 in Na ϩ and K ϩ medium to 2.1 in choline-Cl medium (Table II), which suggests that the monovalent cations also affect the Ca 2ϩ binding cooperativity to the Ca 2ϩ inactivation sites. We conclude that monovalent cations affect by an as yet unidentified mechanism the interaction of the channel inactivation sites with Ca 2ϩ .
The divalent cation specificity of the channel inactivation sites was tested in 0.1 M KCl, 0.5 M KCl, and 0.5 M choline-Cl media in the presence of 5 mM AMP and a relatively high [Ca 2ϩ ] to minimize interaction of the other divalent cations with the Ca 2ϩ activation sites. In the three media, all four divalent cations tested (Mg 2ϩ , Ca 2ϩ , Sr 2ϩ , and Ba 2ϩ ) displayed a similar ability of inhibiting [ 3 H]ryanodine binding. These results are in agreement with vesicle-Ca 2ϩ flux measurements which provided evidence of a similar affinity of the channel inactivation sites for Ca 2ϩ and Mg 2ϩ (18).
The effects of monovalent anions on channel activity were investigated by determining the Ca 2ϩ dependence of [ 3 H]ryanodine binding in media of different anionic composition ( Fig.  2A) and concentration (Fig. 4, A and B), and in competition studies at submicromolar [Ca 2ϩ ] (Fig. 10) or close to fully activating [Ca 2ϩ ] (Fig. 9). In agreement with previous studies (7)(8)(9)(10)(11), an increase in salt concentration (KCl and choline-Cl) greatly increased the levels of [ 3 H]ryanodine binding in the presence of 0.1 M to 10 mM Ca 2ϩ . Analysis of these data (Table  Fig. 2A). Accordingly, Cl Ϫ appears to widen the "Ca 2ϩ window" of receptor activation, that is to allow a more complete occupation of the Ca 2ϩ activation sites by Ca 2ϩ before substantial Ca 2ϩ binding to the Ca 2ϩ inactivation sites occurs. A consequence of a widened Ca 2ϩ window was that increased affinities (Table III) and levels ( Fig. 2A) of [ 3 H]ryanodine binding could be observed in Cl Ϫ media. The initial decrease in apparent Ca 2ϩ affinity as the [KCl] was raised from 0.1 to 0.25 M can be explained assuming that in this concentration range K ϩ competes more strongly with Ca 2ϩ for the activation sites than Cl Ϫ increases the Ca 2ϩ affinity of the Ca 2ϩ activation sites.
A decrease in the vesicle Ca 2ϩ efflux rates and [ 3 H]ryanodine binding was seen when Cl Ϫ was replaced by Mes Ϫ or Pipes Ϫ . Substitution of Cl Ϫ by Mes Ϫ or Pipes Ϫ in choline ϩ medium resulted in ϳ2-fold decrease in the maximum level of [ 3 H]ryanodine binding. [ 3 H]Ryanodine binding decreased close to background levels when these experiments were done in K ϩ media. In these cases, a decrease in the apparent Ca 2ϩ affinity of the Ca 2ϩ activation sites and increase in the apparent Ca 2ϩ affinity of the inactivation sites resulted in a narrowing of the Ca 2ϩ window of receptor activation. In the above scheme these observations are taken into account by proposing that the binding of Cl Ϫ to anion regulatory sites mediates the transition of the RyR channel to a state of greater susceptibility to activation by Ca 2ϩ . The presence of anion regulatory sites was verified by showing a competitive inhibition of the Cl Ϫ -activated receptor by Pipes Ϫ . In these studies a free [Ca 2ϩ ] of 20 M was used to maintain the receptor in its different Ca 2ϩ -activated A Ca and A Ca * states. H]ryanodine binding was determined as described under "Experimental Procedures" in 0.5 M choline-Cl media containing the indicated concentrations of Mes Ϫ and free Ca 2ϩ . In A and B, the continuous lines were obtained with Equations 3 and 4, using a single set of parameters for all data. In the equations, the activating ion was Ca 2ϩ , and the inhibitor was Mes Ϫ . In B, data were plotted using a derived Hill inactivation coefficient of 0.84. Averaged Hill constants and coefficients of four separate experiments are shown in Table IV. A strong functional interaction between the Ca 2ϩ activation sites and anion regulatory sites was observed in choline-Cl media in the presence of submicromolar [Ca 2ϩ ] and using Mes Ϫ as the competing ion. To our surprise, we found that to a first approximation our data could be described by a competitive inhibition mechanism, with Mes Ϫ competing with Ca 2ϩ for the Ca 2ϩ activation sites. We consider it unlikely that Mes Ϫ competed with Ca 2ϩ by direct binding to the Ca 2ϩ activation sites. Two other more likely mechanisms would be an occlusion of the Ca 2ϩ activation site by the bulky Mes Ϫ or a protein conformational change that is caused by binding of the anion to a specific site and distorts the Ca 2ϩ activation site. Additional experiments will be required to characterize more fully the functional interaction between the anion regulatory and Ca 2ϩ activation channel sites.
An alternative explanation for the anion-sensitivity of the SR Ca 2ϩ permeability has been given by Sukhareva et al. (26) who identified a nonselective Cl Ϫ and Ca 2ϩ conducting channel activity that displayed a pharmacology different in several respects from that of the skeletal muscle RyR. Replacement of methanesulfonate Ϫ by Cl Ϫ caused an increase in SR Ca 2ϩ permeability but not single Ca 2ϩ release channel open probability. These observations led Sukhareva et al. (26) to suggest that a separate, nonselective Cl Ϫ channel mediates the Cl Ϫ -dependent Ca 2ϩ release. Our results suggest that single Ca 2ϩ release channel and ryanodine binding activities are strongly dependent on Cl Ϫ concentration. Thus, it is possible to explain our SR permeability studies (Fig. 1, Table I) with the existence of one Cl Ϫ -dependent Ca 2ϩ release pathway in the SR membrane.
Our [ 3 H]ryanodine binding measurements confirm previous SR vesicle-ion flux, single channel and [ 3 H]ryanodine binding measurements, which showed that adenine nucleotides and caffeine activate the skeletal muscle Ca 2ϩ release channel (4,5). As observed in the present study, caffeine primarily activated the channel by increasing the apparent affinity of the Ca 2ϩ activation sites for Ca 2ϩ . In most previous studies, ATP or a nonhydrolyzable ATP analog (AMP-PCP or AMP-PNP) were used to study the effects of adenine nucleotides. In the present study, we limited the number of potential channel effector species by using AMP because this compound, in contrast to ATP and the ATP analogs, binds Ca 2ϩ with only a negligible affinity. We found that the apparent affinity of the Ca 2ϩ inactivation sites for Ca 2ϩ was lowered by greater than 3-fold by AMP, whereas only a modest increase (Ͻ2-fold) in the Ca 2ϩ affinity of the channel activation sites was observed (Tables II and IV). Interestingly, AMP did not substantially increase the affinity of the Ca 2ϩ activation sites for the competing cations (Mg 2ϩ , monovalent cations) (Table IV). Taken together, our results suggest that caffeine primarily activated the skeletal muscle Ca 2ϩ release channel by increasing the affinity of the channel's high-affinity Ca 2ϩ activation sites for Ca 2ϩ , whereas the primary effect of AMP was to decrease the Ca 2ϩ affinity of the low-affinity channel inactivation sites.
Identification of Ca 2ϩ -binding sites has been handicapped by the absence of clearly identifiable Ca 2ϩ binding motifs in the primary amino acid sequence of the rabbit skeletal muscle Ca 2ϩ release channel (5). However, some experimental evidence for the involvement of several channel protein regions in regulating Ca 2ϩ -dependent channel activity has been obtained. In malignant hyperthermia-susceptible pigs, the channel contains an arginine residue at position 615, which when mutated to cysteine, altered the Ca 2ϩ and caffeine sensitivity of the channel (27). Evidence for several Ca 2ϩ -sensitive regions has been obtained in 45 Ca 2ϩ and ruthenium red overlay studies with trpE fusion proteins (28). An antibody directed against one of these peptides (amino acid residues 4478 -4512) increased the Ca 2ϩ sensitivity of Ca 2ϩ release channels incorporated into planar lipid bilayers without affecting single channel conductance. However, it is unlikely that the antibody bound directly to a critical Ca 2ϩ activation site, because Ca 2ϩ was still able to activate the antibody-Ca 2ϩ release channel complex. Our characterization of the cation specificity of the Ca 2ϩ activation and Ca 2ϩ inactivation sites should help to identify these sites in future studies.
In this study, we chose nonphysiological concentrations to identify the principal mechanisms by which monovalent cations and anions regulate the channel. In resting frog skeletal muscle, the concentrations (all in mM) of the principal ionic species have been reported to be K ϩ (141), phosphocreatine (50), carnosine (19), amino acids (12), Na ϩ (9), MgATP 2Ϫ (6), Cl Ϫ (2), and Mg 2ϩ (0.8) (29). How these ionic species separately and in combination affect the function of the skeletal muscle Ca 2ϩ release channel remains to be explored in future studies. Moreover, all our experiments were done under steady-state conditions and therefore did not address the possibility that the rate of Ca 2ϩ application may influence the affinity constants (30).
In conclusion, the results of this study show that monovalent ions profoundly affect the regulation of the skeletal muscle Ca 2ϩ release channel by Ca 2ϩ , in a manner that can be accounted for as changes of the Ca 2ϩ binding affinities of the activation and inactivation channel sites.