Glutathione modulates ryanodine receptor from skeletal muscle sarcoplasmic reticulum. Evidence for redox regulation of the Ca2+ release mechanism.

In this report, we demonstrate the ability of the cellular thiol glutathione to modulate the ryanodine receptor from skeletal muscle sarcoplasmic reticulum. Reduced glutathione (GSH) inhibited Ca2+-stimulated [3H]ryanodine binding to the sarcoplasmic reticulum and inhibited the single-channel gating activity of the reconstituted Ca2+ release channel. The effects of GSH on both the [3H]ryanodine binding and single-channel measurements were dose-dependent, exhibiting an IC50 of ∼2.4 mM in binding experiments. Scatchard analysis demonstrated that GSH decreased the binding affinity of ryanodine for its receptor (increased Kd) and lowered the maximal binding occupancy (Bmax). In addition, GSH did not modify the Ca2+ dependence of [3H]ryanodine binding. In single-channel experiments, GSH (5-10 mM), added to the cis side of the bilayer lipid membrane, lowered the open probability (Po) of a Ca2+ (50 μM)-stimulated Ca2+ channel without modifying the single-channel conductance. Subsequent perfusion of the cis chamber with an identical buffer, containing 50 μM Ca2+ without GSH, re-established Ca2+-stimulated channel gating. GSH did not inhibit channel activity when added to the trans side of the bilayer lipid membrane. Similar to GSH, the thiol-reducing agents dithiothreitol and β-mercaptoethanol also inhibited high affinity [3H]ryanodine binding to sarcoplasmic reticulum membranes. In contrast to GSH, glutathione disulfide (GSSG) was a potent stimulator of high affinity [3H]ryanodine binding and it also stimulated the activity of the reconstituted single Ca2+ release channel. These results provide direct evidence that glutathione interacts with reactive thiols associated with the Ca2+ release channel/ryanodine receptor complex, which are located on the cytoplasmic face of the SR, and support previous observations (Liu, G, Abramson, J. J., Zable, A. C., and Pessah, I. N. (1994) Mol. Pharmacol. 45, 189-200) that reactive thiols may be involved in the gating of the Ca2+ release channel.

In muscle cells, cytosolic Ca 2ϩ levels are regulated by the intramuscular organelle, the sarcoplasmic reticulum (SR) 1 (1)(2)(3). Following the arrival of an action potential at the surface membrane and subsequent depolarization of the transverse tubule, the SR releases its lumenal store of Ca 2ϩ through the Ca 2ϩ release channel (CRC)/ryanodine receptor (RyR), thus triggering the contraction process. The interaction between the action potential at the transverse tubule and the release of Ca 2ϩ from the SR has been termed excitation-contraction coupling (ECC). In skeletal muscle, the molecular mechanism underlying ECC has remained unclear. Following excitation, resting Ca 2ϩ levels are reestablished through the active transport of the Ca 2ϩ back into the lumen of the SR by the Ca 2ϩ ,Mg 2ϩ -ATPase.
A number of thiol reagents act as potent stimulators of the SR Ca 2ϩ release channel. These compounds include heavy metals (4,5), Cu 2ϩ /cysteine (6), reactive disulfides (7), phthalocyanine dyes (8), anthraquinones (9), porphyrins (10), thimerosal (11), and the reactive oxygen species, H 2 O 2 (12,13). Recently Lui et al. (14,15) have described the presence of a discrete class of highly reactive thiols associated with the SR CRCs and other junctionally related proteins, which were labeled by the fluorogenic coumaryl maleimide, CPM. During activation of Ca 2ϩ release by a large class of non-thiol channel stimulators, a high molecular weight complex was formed. The addition of channel inhibitors resulted in the reduction of the disulfides formed, the dissociation of key SR proteins and the exposure of hyperreactive thiols (14,15). Based on the results derived from fluorescence assays, ion flux measurements, single-channel experiments, and SDS-gel electrophoresis, the authors concluded that thiol oxidation-reduction chemistry plays a critical role in the channel gating of the SR CRC⅐RyR complex.
Endogenous and exogenous redox agents have been observed to have profound effects on a wide range of ion channel systems. In addition to the SR Ca 2ϩ release channel, ion channels as varied as excitatory amino acid receptors, inositol 1,4,5triphosphate (IP 3 )-gated Ca 2ϩ release channels, and even certain K ϩ channels have been demonstrated to be modulated by redox agents. For example, the N-methyl-D-aspartate-sensitive excitatory amino acid receptor has been shown to be modulated by both thiol oxidants and reductants (16).
Glutathione is one of the most abundant low molecular weight peptides in eukaryotic cells and the most prevalent intracellular thiol. Depending on the cell type, glutathione levels have been estimated to range from 1 to 10 mM (17,18). In the cell, glutathione acts as both a reducing agent and an antioxidant. Among its many physiological roles, glutathione * This work was supported by a grant from the American Heart Association Oregon Affiliate (to J. J. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
plays an important role in the protein folding process and serves to protect intracellular constituents from oxidation by scavenging reactive oxygen species produced during normal cell metabolism (19). In the cell, glutathione is present in both the reduced (GSH) and oxidized (GSSG or glutathione disulfide) form.
Glutathione disulfide is capable of undergoing thiol-disulfide exchange with reactive protein thiol residues to form a mixed disulfide complex, or it can completely oxidize endogenous sulfhydryls to form disulfides (20). In this capacity, GSSG has been demonstrated to modulate a number of ion channel systems. Gilbert et al. (21) observed that GSSG inhibited NMDA and glycine evoked [Ca 2ϩ ] i increases in embryonic rat neurons. This demonstrated that GSSG interacted with the NMDA receptor-associated redox site. Furthermore, Renard-Rooney et al. (22) has shown that IP 3 receptors are sensitized to IP 3 , in the presence of GSSG. These reports, along with others, demonstrate the ability of GSSG alone to regulate the function of redox-sensitive protein complexes.
The present investigation describes the interaction between glutathione, both the reduced and oxidized forms, as well as other thiol-reducing agents, and key reactive thiol groups associated with the SR Ca 2ϩ release mechanism. It is demonstrated that the thiol-reducing agents dithiothreitol (DTT), GSH, and ␤-mercaptoethanol (BME) inhibited, whereas GSSG stimulated SR CRC activity. These results were observed in [ 3 H]ryanodine binding assays as well as in single-channel experiments. The evidence in this report demonstrates that ligand-gated channel activity associated with the RyR 1 complex in skeletal muscle is subject to both inhibition via thiol reduction and activation by thiol oxidation. Furthermore, at physiologically relevant levels, GSH and GSSG modulate SR Ca 2ϩ release channel and [ 3 H]ryanodine binding activity.

EXPERIMENTAL PROCEDURES
Preparation of SR Vesicles-For all studies, crude sarcoplasmic reticulum vesicles were prepared from rabbit hind leg and back white skeletal muscle according to the method of MacLennan (23). The protein concentration was determined by absorption spectroscopy (24). Prior to use, all SR preparations were suspended at 15-25 mg/ml in buffer containing 100 mM KCl, 20 mM HEPES, pH 7.0 (KOH), and stored at 60 K, in liquid nitrogen. The same SR preparations were used for both the [ 3 H]ryanodine binding experiments and the single-channel measurements. However, in the bilayer studies, SR vesicles were suspended at 1.5 mg/ml in buffer containing 100 mM KCl, 0.3 M sucrose, 20 mM HEPES, at pH 7.0 and allowed to equilibrate overnight at ϳ5°C. The samples were then stored in liquid nitrogen.
Preparation of Glutathione Stock Solutions-To avoid potential artifacts due to alterations of pH in the presence of unbuffered glutathione, all glutathione stock solutions were adjusted to pH 7.0. GSH stock solutions were prepared at 1 M, with pH levels adjusted by the addition of KOH, in degased (bubbled with N 2 for 20 min), deionized H 2 O and stored in liquid nitrogen until use. GSSG stock solutions were prepared at 0.5 M in deionized H 2 O, pH adjusted to 7.0 with HCl, and stored at 5°C. All glutathione stock solutions were prepared fresh at least once per week. As a further safeguard, pH measurements of assay buffers were made in the presence of GSH and/or GSSG, at 25 and 34°C prior to use. In all cases, addition of either or both forms of glutathione did not result in an alteration in pH.
[ 3 H]Ryanodine Binding Assays-Detailed methods for measuring high affinity [ 3 H]ryanodine binding have been described elsewhere (25). Briefly, SR membranes (0.5 mg/ml) were incubated at 34°C for 4.5 h in a medium containing 250 mM KCl, 15 mM NaCl, 7.5 or 15 nM [ 3 H]ryanodine, and 20 mM HEPES, at pH 7.1 (KOH). Depending on the conditions of the assay, various channel modifiers were present during the incubation procedure. The binding reaction was quenched by rapid filtration through a Whatman GF/B glass fiber filter, which was then rinsed three times with 3 ml of standard buffer. The filters were placed in polytubes (VWR), filled with 2.5 ml of scintillation mixture (ICN, CytoScint), shaken overnight, vortex-mixed, and counted the following day. The experiments were typically repeated at least twice on two different SR preparations. Nonspecific binding was measured in the presence of a 200-fold excess of unlabeled ryanodine. For details of individual experiments, refer to figure captions.
Hill Analysis of Binding Data-Dose-dependence binding data, in the presence of channel activators (GSSG and Ca 2ϩ ) were fit to a Hill equation of the form: where B is the amount of ryanodine bound (pmol/mg), [A] is the concentration of activator, B max is the maximal binding in the presence of activator, K d is the apparent affinity of the activator site for A, and the apparent Hill coefficient, n, is a measure of the degree of cooperativity for activation of [ 3 H]ryanodine binding (35). Data were fit using nonlinear regression curve fitting routine (Sigma Plot for Windows version 1.02, Jandel Scientific). Determination of n and seed values (B max and K d ) for the nonlinear regression were determined by linear regression of Hill plots (ln(B/(B max Ϫ B)) versus ln([A])) of data between 20 and 80% of B max . The initial B max for the Hill plots was determined directly from the raw data. Using values from the resultant nonlinear regression, the EC 50 was calculated from the relationship, EC 50 ϭ (K d ) 1/n .
Dose-response curves in the presence of the reducing agents were fit, using nonlinear regression, to a modified (pseudo) Hill equation of the form: where B is the amount of ryanodine bound (pmol/mg), [I] is the concentration of inhibitor, B max is the maximal binding in the absence of inhibitor, K d is the apparent affinity of the inhibitory site for the inhibitor, and n is the pseudo-Hill coefficient (or logit slope). The pseudo-Hill coefficient and seed values for the nonlinear curve fit were obtained directly from linear regression of logit-log plots of ln((B max Ϫ B)/B) versus ln([I]) (36). Using values from the resultant nonlinear regression, the IC 50 was calculated from the relationship, IC 50 ϭ (K d ) 1/n .
Single Ca 2ϩ Channel Analysis-Ca 2ϩ release channel reconstitution into a bilayer membrane was carried out by the addition of SR vesicles to the cis side of a planar bilayer lipid membrane. Bilayers, made with a 5:3 mixture of phosphatidylethanolamine and phosphatidylserine at 50 mg/ml in decane, were formed across a 150-m hole drilled in a polystyrene cup separating two chambers of 0.7 ml each. The cis chamber contained 500 mM CsCl (or CsMS), 100 M CaCl 2 , 20 mM HEPES, pH 7.0, while the trans side contained 100 mM CsCl (or CsMS), 20 mM HEPES pH 7.0. SR vesicles, suspended in 0.3 M sucrose were added to the cis side at a final concentration of 5-20 g/ml. Following the fusion of a single vesicle, 150 M EGTA, pH 7.0, was added to the cis chamber to stop further fusions. The cis chamber was then perfused with an identical buffer containing no added Ca 2ϩ or EGTA. Channel activity was then measured at a holding potential of ϩ25 mV with respect to the trans (ground) side. A Warner Instruments Bilayer Clamp Amplifier (model BC-525A) was used to amplify picoampere currents. The data were processed with an Instratech Digital Data Recorder (model VR-10), stored unfiltered on a VCR tape and subsequently analyzed for channel activity. For analysis, the data were passed through a Krohn-Hite low pass filter (model 3202) at 1.5 kHz, digitized with a Scientific Solutions analog to digital converter and analyzed using the pCLAMP software package (version 5.5, Axon Instruments, Burlingame, CA). See figure captions for specific experimental conditions. Materials-All reagents were analytical grade. HEPES was obtained from Research Organics (Cincinnati, OH). [ 3 H]Ryanodine was purchased from DuPont NEN, and ryanodine-dehydroryanodine was purchased from Agrisystems Int. (Windy Gap, PA). All other chemicals were obtained from Sigma.

RESULTS
The measurement of high affinity ryanodine binding to sarcoplasmic reticulum vesicles has been demonstrated to be an effective probe of Ca 2ϩ release channel activity (25). High affinity ryanodine binding to SR membranes increases under conditions in which Ca 2ϩ release channels are activated, whereas channel inhibitors diminish binding. With few exceptions, Ca 2ϩ release channel agonists stimulate ryanodine binding to RyRs, while compounds that inhibit channel activity decrease binding. A notable exception is the effect of Ag ϩ , which acts as a potent SR Ca 2ϩ -releasing agent at micromolar concentrations (4,5), but decreases ryanodine binding by rapidly displacing bound ryanodine from its receptor (25).
The effects of thiol-reducing agents on high affinity ryanodine binding to SR membranes were examined in Fig. 1. In this figure, Ca 2ϩ (50 M)-stimulated ryanodine binding was inhibited by increasing concentrations of reducing agent (DTT, GSH, or BME). For all three reducing agents, binding was decreased from a starting value of ϳ2.1-0.6 pmol/mg, a decrease of 70%. DTT was the most effective channel inhibitor, followed by BME then GSH. The IC 50 values were calculated to be 0.1, 0.4, and 2.4 mM, respectively. The binding data were fit to a modified Hill equation, as described under "Experimental Procedures." The results of the Hill analyses are summarized in Table I. As shown in this table, the inhibition of alkaloid binding due to either DTT or GSH exhibited positive cooperativity (n Ͼ 1). The logit slope for BME inhibition of ryanodine binding was near unity. Furthermore, as observed in Fig. 2, both in the presence of inhibitors ( Fig. 2A) or channel activators (Fig. 2B), reduced glutathione inhibited high affinity [ 3 H]ryanodine binding. Similar results were obtained with DTT or BME (data not shown).
GSH levels as a function of time were measured using absorbance spectroscopy to verify that significant oxidation of GSH was not occurring during the course of the ryanodine binding assays. In these measurements, GSH was suspended at 0.1, 1.0, 5.0, and 10 mM in standard binding buffer and incubated at 34°C in either the absence or presence of SR (0.5 mg/ml). At each time point, the GSH was diluted to 50 M, based on the original concentration, into a similar buffer containing 1 mM DTNB, pH 7.1. The absorbance increase due to the interaction between DTNB and free sulhydryl groups on GSH or thiols associated with the SR was monitored at 412 nm (data not shown). To calibrate the DTNB-SH response, measurement of absorbance as a function of either freshly prepared GSH or BME were performed, in the concentration range between 10 and 80 M, where the response was linear. From these dose-response (calibration) curves, the extinction coefficient was determined to be in the range of 13,200 -13,600 M Ϫ1 cm Ϫ1 . In the absence of SR vesicles, following a 4.5-h incubation at 34°C, the [GSH] dropped significantly from an original concentration of 100 M to 35 M, a decrease of 65%. At 100 M, two-thirds of the original GSH was oxidized to GSSG during the incubation period. At GSH levels of 1.0, 5.0, and 10.0 mM, the percentage of GSH oxidized after incubation at 34°C for 4.5 h was approximately 20%, 13%, and 9%, respectively. However, in the presence of SR, an insignificant decrease in GSH levels was observed at all concentrations measured. Following a 4.5-h incubation at 34°C, GSH levels were never less than 95% of the original starting concentration of GSH (0.1, 1.0, 5.0, and 10.0 mM). No significant decrease in GSH levels was observed during the time course of the ryanodine binding assays presented in this report.
In contrast to the reducing agents, the oxidized form of glutathione was a potent stimulator of ryanodine binding. As shown in Fig. 1, glutathione disulfide increased binding from 2.1 to 3.3 pmol/mg, roughly 60%. Hill analysis yielded an EC 50 of ϳ90 M and an apparent Hill coefficient (n) of ϳ0.8 (see Table I).
The effects of channel inhibitors and activators on GSSGstimulated ryanodine binding are examined in Fig. 3. In Fig.  3A, ryanodine binding was measured with and without GSSG (2 mM) in the presence of the known channel inhibitors ruthenium red and Mg 2ϩ . Ruthenium red (20 M) completely inhibited ryanodine binding in both the presence and absence of GSSG, whereas 1 mM Mg 2ϩ decreased binding for control and GSSG treated vesicles 75% and 60%, respectively. The observation that oxidation stimulated channel activity is inhibited by known channel antagonists has been similarly observed in previous reports (6 -11).
Ryanodine binding was measured in the presence of known channel activators, with or without GSSG, in Fig. 3B. In the absence of GSSG, both caffeine (5 mM) and cAMP (1 mM) stimulated binding significantly, when compared to control conditions. Caffeine stimulated ryanodine binding by 37%, while the presence of cAMP increased binding by 95%. GSSG was shown to increase ryanodine binding under all three activating conditions. However, the percent increase of binding due to the agonists was attenuated in the presence of GSSG. Furthermore, comparison of experiments carried out in the absence and presence of GSSG, under activating conditions, demonstrates that GSSG increased binding is diminished as the level of control binding increased.
Lineweaver-Burk analysis was performed to determine whether or not GSSG stimulation and GSH inhibition of ryanodine binding were due to an interaction with the same critical thiols/redox sites associated with the RyR (Fig. 4). GSSG doseresponse curves were generated in the presence of various amounts of GSH. The resultant data were plotted as 1/(ryanodine bound) as a function of 1/[GSSG]. The linear regression of the resulting Lineweaver-Burk plot, for the three concentrations of GSH, intersect at a y intercept of 0.26 Ϯ 0.04 mg/pmol.
Oxidizing agent EC 50 Apparent Hill coefficient This suggests that the mechanisms of GSSG oxidation and GSH reduction interact via a competitive interaction. It is likely that GSH and GSSG compete for the same reactive thiol sites, reducing endogenous disulfides or oxidizing free -SH groups, depending on the thiol status of the channel. From the y intercept, the theoretical B max for oxidation-stimulated binding was determined to be 3.8 Ϯ 0.6 pmol/mg. To better describe their effects on the RyR 1 , Scatchard analyses were performed in the presence of the reducing agents (GSH, BME, and DTT), in Fig. 5 (A and B), and oxidized glutathione (GSSG), in Fig. 6 (A and B). In Fig. 5A, ryanodine saturation binding experiments were performed in the absence or presence of reducing agents. The raw data were plotted as bound/free versus bound ryanodine (Fig. 5B). Linear regression of the resultant Scatchard plots, in Fig. 5B, show that the reducing agents decreased the binding affinity of the alkaloid for its receptor, increasing the K d by 2-5-fold. DTT (2 mM), demonstrated the most potent effect on the binding affinity, followed by BME, then GSH. In addition to decreasing the binding affinity, the maximum number of binding sites (or B max ) decreased slightly for each of the reducing agents, from 8.7 to 8.0 (for GSH), 7.7 (for BME), or 7.2 (for DTT) pmol/mg. The decrease in high affinity ryanodine binding induced by the addition of reducing agents (Fig. 1) is caused by an increase in the K d and a slight decrease in the B max for ryanodine binding.
Whereas the reducing agents increased the K d and decreased the B max , GSSG had an opposite effect. In Fig. 6A, ryanodine saturation binding curves were generated in the absence or presence of GSSG. The binding data were then transformed into the Scatchard plot in Fig. 6B. In Fig. 6B, it is shown that 0.5 mM [GSSG] lowered the K d from 11.5 to 9.7 nM, almost 20%, while the B max increased slightly to 9.5 pmol/mg. In the presence of 2 mM GSSG, the K d decreased nearly 2-fold, while the B max increased to 10.5 pmol/mg (nearly 25%). The increase in ryanodine binding observed in Fig. 1, is primarily caused by an approximate 2-fold decrease in the K d for ryanodine binding 2 pmol/mg, respectively) of GSH. In B, 1 mM cAMP or 5 mM caffeine increased binding compared from control (3.3 pmol/mg) to 4.9 or 6.1 pmol/mg, respectively. SR incubated in the presence of GSH exhibited decreased binding, to 2.5 or 3.5 pmol/mg (compared to 1.5 pmol/mg in absence of activators), respectively. Data are the average of two experiments performed in duplicate. The fractional error for each measurement was less than 5%.

FIG. 3. GSSG stimulates [ 3 H]ryanodine binding in the presence of Ca 2؉ channel modulators.
Equilibrium binding experiments, in the presence of known channel inhibitors (A) or activators (B), were carried out in the absence or presence of 2 mM GSSG. Standard assay conditions were the same as described in Fig. 1, with 15 nM [ 3 H]ryanodine and 50 M Ca 2ϩ at pH 7.1. In A, binding was increased from 3.0 (untreated) to 5.0 pmol/mg when incubated with 2 mM GSSG. Channel inhibitors, 20 M ruthenium red or 1 mM Mg 2ϩ , decreased binding in the presence (to Ͻ0.1 or 2.0 pmol/mg, respectively) or absence (to Ͻ0.1 or 0.8 pmol/mg, respectively) of 2 mM GSSG. In B, 1 mM cAMP or 5 mM caffeine increased binding, compared to control values, in the presence (5.5 or 6.6 pmol/mg, respectively) and absence (4.1 or 5.9 pmol/mg, respectively) of GSSG. Data are the averages of experiments performed in duplicate. The fractional error for each measurement was less than 5%. (Fig. 6). A summary of the Scatchard analyses is presented in Table II.
While GSH and GSSG have been shown to modulate ryanodine binding, neither of the compounds significantly alters the Ca 2ϩ dependence of ryanodine binding. As shown in Fig. 7, although the overall binding was amplified by GSSG and diminished by GSH, the receptor sensitivity to Ca 2ϩ activation was retained. Hill analysis of the data in Fig. 7  free . The observation that GSH inhibition and GSSG stimulation of ryanodine binding does not alter the apparent sensitivity of the receptor to Ca 2ϩ activation suggests that neither the reduced or oxidized form of glutathione bind Ca 2ϩ . Direct measurements of free Ca 2ϩ levels, upon addition of GSH or GSSG, using a Ca 2ϩ -selective electrode (World Precision Instruments, Inc.) verified that at concentrations as high as 20 mM glutathione only a negligible amount of Ca 2ϩ was bound (less than 2%; data not shown). These measurements were performed in standard binding buffer in the absence of SR vesicles.
To better characterize the interaction between GSH, GSSG, and the SR, single-channel experiments were performed on SR Ca 2ϩ release channels reconstituted into an artificial lipid membrane. This experimental procedure allows the visualization of channel current fluctuations across the SR membrane and facilitates the direct observation of modifications to channel open probability (P o ) and conductance. Fig. 8 illustrates two Ca 2ϩ channels reconstituted into a planar artificial membrane. Using Cs ϩ as the carrier current in a 5 to 1 cis-trans gradient, channel gating was recorded as a function of time at a holding potential of ϩ25 mV with respect to the trans side of the channel. In trace i, channel open probability was very low following channel fusion (P o ϭ 0.05 for n ϭ 2 channels). In trace ii, 50 M Ca 2ϩ was added to the cis chamber and channel gating was stimulated, P o ϭ 0.75. Addition of 2 mM GSH to the cis chamber lowered the channel P o to 0.15, as shown in trace iii. Subsequent additions of 4 mM GSH aliquots to the cis chamber further reduced the channel open probabilities to 0.10 and Ͻ0.05, respectively. This experiment verifies the effects of GSH observed in Fig. 1. Furthermore, GSH inhibits Ca 2ϩ release channel activity by decreasing channel open probability without affecting unitary channel conductance (data not shown).
Additional single-channel experiments were performed to determine the sidedness of GSH inhibition, as shown in Fig. 9.
In the first trace of this figure (i), channel gating was recorded in 5:1 cesium methanesulfonate gradient. The single-channel open probability was initially very low, P o ϳ 0.01. Addition of 50 M cis Ca 2ϩ increased channel gating, P o ϭ 0.45, in trace ii. As shown in trace iii, addition of 5 mM GSH to the trans side was observed to have no effect on channel open probability. However, when 5 mM GSH was added to the cis side (trace iv), channel gating activity was lowered (P o ϭ 0.30). Raising the cis GSH concentration to 10 mM, in trace v, decreased the P o to . These findings suggest that treatment with GSSG may stimulate channel gating by increasing the on rate associated with channel activation, without affecting the off rate. Furthermore, the ability of GSSG to stimulate channel gating was observed to be highly Ca 2ϩ -dependent. GSSG was unable to stimulate single-channel activity at [Ca 2ϩ ] cis below 2 M. Moreover, if GSSG was observed to stimulate channel activity, and the Ca 2ϩ concentration was then decreased below 2 M, all channel activity was completely inhibited (data not shown). DISCUSSION In this report, the interactions between thiol-reducing agents, DTT, BME, the reduced and oxidized forms of glutathione (GSH and GSSG), and the SR Ca 2ϩ release mechanism have been examined. It has been demonstrated that thiol reductants decrease ryanodine binding stimulated by Ca 2ϩ , caffeine, and the adenine nucleotide, cAMP. Ca 2ϩ -stimulated binding was particularly susceptible to inhibition by thiol reduction. While the reducing agents inhibited ryanodine binding in experiments carried out at 50 M Ca 2ϩ , GSH did not appear to affect the sensitivity of ryanodine binding to Ca 2ϩ activation. The reducing agents both decreased the binding affinity of the SR membranes to the alkaloid and lowered the available number of receptor binding sites. Similar results were observed at the single-channel level, where GSH was shown to reversibly decrease channel open probability, without affecting unitary

TABLE II
Thiol reducing agents and GSSG modulate ryanodine binding kinetics of the skeletal muscle SR ryanodine receptor Experiments were carried out as described in Figs. 5 and 6. Scatchard plots were fit to a one-site binding model using linear regression. The slope of the regression yielded the K d (dissociation constant) and the x intercept was identified as the B max (maximum binding occupancy). Data shown are the mean Ϯ standard deviation. channel conductance. Furthermore, the apparent location of the GSH interaction site was determined to be on the cytosolic side of the membrane. This laboratory has investigated the role of thiol oxidation in the skeletal muscle SR Ca 2ϩ release mechanism. It has been previously demonstrated that thiol reactive compounds can initiate rapid Ca 2ϩ release from and stimulate ryanodine binding to SR vesicles via mercaptidation (4, 5), thiol-disulfide exchange (6,7), and direct thiol oxidation (8 -10, 12). In addition, it has recently been reported that alkylation of highly reactive thiols associated with the ryanodine receptor complex inhibited Ca 2ϩ induced Ca 2ϩ release from isolated SR vesicles (14). In this latter investigation, hyperreactive thiols were labeled with the fluorogenic maleimide, CPM, under conditions that promoted Ca 2ϩ channel closure. Under stimulatory conditions these reactive thiols were unavailable for CPM labeling. These observations strongly suggested that during stimulation of Ca 2ϩ release, highly reactive thiols were oxidized to disulfides while under inhibitory conditions these thiols existed in the reduced form.
It has previously been proposed that during normal Ca 2ϩ channel activation, a high molecular weight disulfide-linked complex is formed between key SR proteins (15). If this hypothesis is correct, reduction of these disulfide linkages by treatment with reducing agents would be expected to inhibit Ca 2ϩ channel activity. In Figs. 1 and 2, it was observed that thiolreducing agents inhibited Ca 2ϩ , caffeine, and cAMP-stimulated ryanodine binding and single-channel activity (Figs. 8 and 9). Furthermore, the relative order of potency (DTT, BME, and GSH, respectively) agrees with predictions based on redox potential and accessibility arguments. The reduction potential for DTT, V redox , has been measured to be Ϫ330 mV (26), which is significantly lower that of either BME or GSH (V redox ϳ Ϫ250 mV) (34). Furthermore, although the reduction potentials of BME and GSH are comparable, the smaller size and lipophilic nature of BME suggests that it would probably have better accessibility to redox sites than the larger more hydrophilic GSH (31).
On a more general level, the results of this report suggest that physiologically controlled levels of GSH, within the muscle fiber, may serve to down-regulate the gating state of the SR CRC. Based on the data presented, such a system would assist in keeping SR Ca 2ϩ permeability low, particularly during the resting state of the cell. In addition, the effects of cellular GSH may act in unison with Mg 2ϩ levels to counteract the stimulatory effects of in situ adenine nucleotide levels under resting conditions. It is interesting that estimated cytosolic levels of GSH within skeletal muscle are maintained at about the concentration of the IC 50 for GSH inhibition of the RyR. Given this observation, it is likely that small perturbations of the cellular GSH level may have profound implications on RyR function and Ca 2ϩ homeostasis.
Although the distribution of GSH to GSSG varies by cell type, most of the total cellular glutathione exists as GSH (typically Ͼ 90%). Maintenance of the cytosolic redox environment, through the regulation of cellular GSSG to GSH levels, is obtained through a glutathione reductase pathway (18,19). In skeletal muscle, total glutathione levels have been estimated to be about 1-3 mM (27,30). In murine hybridoma CRL-1606 cells, the cytosolic GSH/GSSG ratio is equal to roughly 60:1, while in the endoplasmic reticulum the lumenal GSH/GSSG ratio is about 2:1 (28). It is likely that the ratios of GSSG to GSH in the SR are similar to that of endoplasmic reticulum, though these measurements have not been reported in muscle preparations. Oxidative stress, induced by oxygen-derived free radicals, can either directly oxidize thiols of key protein components (12,29) or cause a disruption in the normal cytosolic GSSG/GSH balance (i.e. oxidizing GSH to GSSG) (30). Disruptions of the redox balance of glutathione could thus have wide-ranging profound implications on the regulation of contractile function.
The observation that micromolar concentrations of GSSG stimulate Ca 2ϩ channel function may also be physiologically significant. Presumably, GSSG and the RyR interact via thiol disulfide exchange, resulting in the oxidation of critical RyR thiols. Reactive disulfides have previously been demonstrated to stimulate CRC function (7). The effects of GSSG on RyR activity contrasted with those of GSH demonstrate that both oxidation and reduction of protein thiols were observed to modulate channel function. In muscle fibers, while cytoplasmic levels of GSSG are kept at very low levels, in the range of 30 -50 M, GSSG is a by-product of GSH depletion during oxidative stress. In the presence of reactive oxygen species, decreased [GSH] accompanied by increased [GSSG] would be expected to increase the Ca 2ϩ permeability of the SR. Higher resting Ca 2ϩ concentrations could disrupt ECC signaling, distinct from the direct effects of the oxygen radicals on SR membrane proteins.
While GSH and GSSG are shown to be potent modulators of RyR function, these compounds did not affect the [Ca 2ϩ ] dependence of high affinity ryanodine binding (Fig. 7). In the presence of millimolar GSH, maximum binding levels as a function of [Ca 2ϩ ] are markedly decreased, while the EC 50 for activation remained unchanged. Additionally, millimolar levels of GSSG increased peak binding levels with little effect on the [Ca 2ϩ ] EC 50 . This evidence suggests that glutathione may act as an amplifier for channel activity but does not serve as a primary trigger for Ca 2ϩ release. However, the glutathione balance within the cytosol, may serve to modulate Ca 2ϩ -stimulated channel activity following an EC coupling event.
The interaction between GSSG and RyR is not an overly surprising observation. In fact, a number of other groups have reported GSSG interaction with various membrane protein systems. For instance, Gilbert et al. reported that GSSG inhibited NMDA-and glycine-stimulated [Ca 2ϩ ] i levels in neurons (21), by interacting with an NMDA receptor-associated redox site. In addition, the authors noted that GSSG inhibited [Ca 2ϩ ] i changes produced by KCl depolarization, possibly through an interaction with N-and L-type Ca 2ϩ channels. In another report, Renard-Rooney et al. (22) observed that GSSG, under certain conditions, stimulated IP 3 binding to hepatic IP 3 receptors by increasing the available number of binding sites (B max ), while not affecting the binding affinity. IP 3 Rs share certain sequence homologies with RyRs, which include cysteine residues near the C terminus. It has been suggested that this common thiol-containing motif may play a common role in the redox regulation of the respective protein complexes (31). Furthermore, Park et al. (32) observed that K Ca channel activity, from either pulmonary or ear arterial smooth muscle cells, was increased by treatment with GSSG.
It has been reported previously that the cardiac RyR isoform (RyR 2 ), like RyR 1 , is subject to regulation via thiol oxidation and reduction. Boraso et al. (33) reported that SR isolated from cardiac muscle was susceptible to modulation by the reactive oxygen species peroxide. The authors demonstrated that millimolar levels of peroxide were capable of stimulating cardiac CRC gating in single-channel experiments. In addition, it was observed that Ca 2ϩ -stimulated RyR 2 channel gating was inhibited by 10 mM DTT. This latter report demonstrated that RyR 2 exhibits sensitivity to redox modulation similar to the skeletal isoform. As stated above, the skeletal RyR has also been shown to be sensitive to direct stimulation by peroxide (12). It is highly likely that both RyR 1 and RyR 2 are, to some degree, regulated by intracellular levels of GSH and GSSG. This evidence suggests that, in both cardiac and skeletal muscle, oxidative stress is induced by both a direct interaction between oxygen-derived radicals and key SR protein components, and by an alteration of normal myoplasmic GSH levels. Either of these mechanisms could lead to a disruption of normal cellular Ca 2ϩ homeostasis.
GSH appears to be protected from autooxidation in the presence, but not in the absence of the SR. Micromolar concentrations of heavy metals (Cu 2ϩ or Fe 3ϩ ) are known to catalyze the oxidation of GSH to GSSG. The SR may be preventing the oxidation of GSH by binding contaminating heavy metals, or the SR may contain an endogenous reductase that is capable of reducing GSSG back to GSH. This latter explanation is unlikely since no reduced substrate (i.e. NADH or NADPH) is present in these assays. Irrespective of the explanation, in the presence of SR vesicles, GSH levels remain constant during the 4.5 h that the SR was incubated with [ 3 H]ryanodine at 34°C.
In summary, we have demonstrated that thiol-reducing agents inhibit both ryanodine binding and single-channel gating stimulated by known channel agonists. These compounds appear to inhibit ryanodine binding primarily by decreasing the binding affinity of ryanodine for its receptor. At the singlechannel level, thiol-reducing agents were also shown to decrease the open channel probability (P o ). In addition, it was shown that GSSG stimulated ryanodine binding and singlechannel gating by directly oxidizing thiols associated with the channel complex. The interaction between GSSG stimulation and GSH inhibition modulate ryanodine binding via a competitive mechanism. One explanation for this observation is that there exists a discrete set of reactive thiol groups that are subject to both GSSG oxidation and GSH reduction, depending on the activation state of the channel. Single-channel measurements indicate that the thiol groups associated with GSH inhibition of the channel are located on the cytoplasmic face of the SR membrane. These observations support previous work from this laboratory, suggesting that oxidation-reduction of protein thiols within the ryanodine receptor complex plays an important role in SR Ca 2ϩ release channel function.