S-Glutathionylation Decreases Mg2+ Inhibition and S-Nitrosylation Enhances Ca2+ Activation of RyR1 Channels*

We have analyzed the effects of the endogenous redoxactive agents S-nitrosoglutathione and glutathione disulfide, and the NO donor NOR-3, on calcium release kinetics mediated by ryanodine receptor channels. Incubation of triad-enriched sarcoplasmic reticulum vesicles isolated from mammalian skeletal muscle with these three agents elicits different responses. Glutathione disulfide significantly reduces the inhibitory effect of Mg2+ without altering Ca2+ activation of release kinetics, whereas NOR-3 enhances Ca2+ activation of release kinetics without altering Mg2+ inhibition. Incubation with S-nitrosoglutathione produces both effects; it significantly enhances Ca2+ activation of release kinetics and diminishes the inhibitory effect of Mg2+ on this process. Triad incubation with [35S]nitrosoglutathione at pCa 5 promoted 35S incorporation into 2.5 cysteine residues per channel monomer; this incorporation decreased significantly at pCa 9. These findings indicate that S-nitrosoglutathione supports S-glutathionylation as well as the reported S-nitrosylation of ryanodine receptor channels (Sun, J., Xu, L., Eu, J. P., Stamler, J. S., and Meissner, G. (2003) J. Biol. Chem. 278, 8184–8189). The combined results suggest that S-glutathionylation of specific cysteine residues can modulate channel inhibition by Mg2+, whereas S-nitrosylation of different cysteines can modulate the activation of the channel by Ca2+. Possible physiological and pathological implications of the activation of skeletal Ca2+ release channels by endogenous redox species are discussed.

. Not surprisingly, these Ca 2ϩ release channels are extensively regulated by a variety of endogenous ions and molecules, as well as through interactions with other proteins (3,(5)(6)(7)(8).
We have previously reported that thimerosal enhances single RyR channel activity both by increasing channel activation by micromolar [Ca 2ϩ ] and decreasing inhibition by 0.5 mM [Ca 2ϩ ] (23). Thimerosal, in a concentration-and time-dependent manner, also enhances CICR rates from SR vesicles from mammalian skeletal muscle and decreases or abolishes the inhibition of CICR by 1 mM [Mg 2ϩ ] (20). Thimerosal, however, is an exogenous organomercapturial compound that modifies SH residues most likely via S-alkylation (52). If endogenous GSNO had the potential to produce similar modifications of RyR channel activity as thimerosal, a significant enhancement of the CICR process would be expected in response to GSNOinduced endogenous redox modification of the channel protein. Therefore, we assessed the ability of GSNO to modify Ca 2ϩ release kinetics and determined whether, in addition to its reported S-nitrosylating activity (42), GSNO also induced Sglutathionylation of RyR1 channels. We found that GSNO produced similar activation of CICR as thimerosal, and promoted calcium-dependent S-glutathionylation of the RyR channel protein. In addition, we tested the effects on CICR of GSSG and of the NO donor NOR-3, as pure S-glutathionylating or S-nitrosylating agents, respectively. The present results suggest that S-glutathionylation modifies specific cysteine residues and, in doing so, either directly or indirectly modulates channel inhibition by Mg 2ϩ whereas S-nitrosylation modifies other cysteines, thereby altering activation by Ca 2ϩ . Possible physiological and pathological implications of these findings are discussed.

EXPERIMENTAL PROCEDURES
Materials-All reagents used were of analytical grade. Protease inhibitors (leupeptin, pepstatin A, benzamidine, and phenylmethylsulfonyl fluoride) and bovine serum albumin were obtained from Sigma. Calcium Green-2 and Calcium Green-5N were obtained from Molecular Probes, Inc. (Eugene, OR). The redox reagents GSH, GSSG, DTT, and NOR-3 were purchased from Calbiochem (La Jolla, CA). When stored, fresh GSH solutions were bubbled with N 2 and frozen at Ϫ20°C. To avoid oxidation of GSH, aliquots were thawed only once and the unused portion was discarded. [ 35 S]GSH and [ 3 H]ryanodine were purchased from PerkinElmer Life Sciences. Primary antibodies against RyR1 (MA3-925) or triadin (MA3-931) and the secondary antibody (SA1-100) were purchased from Affinity BioReagents, Inc. (Golden, CO).
Membrane Preparations-Triad-enriched SR vesicles (from now on triads) were isolated from rabbit fast skeletal muscle in the presence of a combination of protease inhibitors, as described previously (53). To avoid spontaneous oxidation, all membrane fractions were rapidly frozen and kept under liquid N 2 . Experiments were performed within 24 -48 h of membrane isolation to minimize possible oxidation during storage. Before use, the functionality of each preparation was verified. Native vesicles (0.2 mg/ml) that, following addition of 5 mM Mg-ATP, took longer than 2 min at 25°C to actively decrease extravesicular free [Ca 2ϩ ] from 25 M to Ͻ0.2 M (measured with Calcium Green-2) were discarded. It is possible that, because of the steps taken to avoid oxidation, release rate constants reported in this work for native triads were somewhat lower than previously reported values (20). Protein concentration was determined according to Hartree (54) using bovine serum albumin as standard.
[ 3 H]Ryanodine Binding-[ 3 H]Ryanodine binding to triads was determined essentially as described (55). The composition of the solution was (in mM): 500 KCl, 0.5 adenosine 5Ј-(␤,␥-imino)triphosphate, 20 MOPS-Tris, pH 7.2, pCa 5. Total binding was routinely measured in the presence of 10 nM [ 3 H]ryanodine and nonspecific binding in the additional presence of 10 M ryanodine. The isolated triad preparations displayed, on average, B max values of ryanodine binding of 18 pmol/mg of protein.
Synthesis of GSNO-The synthesis of GSNO was performed as described (56). Briefly, equimolar quantities of GSH and NaNO 2 were incubated in 0.75 N HCl at room temperature. After 5 min of incubation, the solution was neutralized by addition of solid Trizma (Tris base). Concentrations of GSNO were determined from its absorbance, using an extinction coefficient of 767 M Ϫ1 cm Ϫ1 at 334 nm (44). Routinely, yields of GSNO were in the 90 -95% range. Synthesis of [ 35 S]GSNO was performed as above, using [ 35 S]GSH at a specific activity of 600 mCi/ mmol, after extraction with ethyl acetate of the DTT present in this reagent.
Ca 2ϩ Release Kinetics-Ca 2ϩ release kinetics was measured in a SX.18MV fluorescence stopped flow spectrometer from Applied Photophysics Ltd. (Leatherhead, United Kingdom) as described (20,57). Vesicles were actively loaded with Ca 2ϩ before eliciting Ca 2ϩ release.
For this purpose triads (1 mg/ml) were incubated for 20 min at 25°C in a solution containing (in mM): 0.05 CaCl 2 , 100 KCl, 5 ATP, 5 MgCl 2 , 10 phosphocreatine plus 15 units/ml creatine kinase, 20 imidazole-MOPS, pH 7.2. Calcium release was initiated by mixing 1 volume of Ca 2ϩloaded triads with 10 volumes of releasing solution. Releasing solutions were designed to produce after mixing pCa 5 (unless indicated otherwise), 0.6 -1.5 mM free [ATP], and variable free [Mg 2ϩ ], from 25 M (the lowest value attainable after active loading) to 900 M. The increase in extravesicular [Ca 2ϩ ] was determined by measuring the fluorescence of the Ca 2ϩ indicator Calcium Green-5N (20). The free [Ca 2ϩ ], free [Mg 2ϩ ] and free ATP concentration of releasing solutions were calculated with the WinMaxC program using the constants provided in the file cmc1002e.tcm (www.stanford.edu/ϳcpatton/winmaxc2.html).
To study the effect of SH redox modification on calcium release, kinetics, prior to active loading triads, were incubated with redox agents: GSSG (1 mM for 20 min at 25°C), GSNO (100 -500 M for 20 min at 25°C), or NOR-3 (50 M for 10 min at 25°C). In all cases, control vesicles were incubated in the same conditions. To ensure maximal calcium loading after SH modification, which slowed down Ca 2ϩ uptake rates, the time needed to actively reduce extravesicular free [Ca 2ϩ ] from 50 M to Ͻ0.2 M was determined in each case as described in detail elsewhere (57).
Identification of RyR1 and Triadin by Western Blot-Protein samples, denatured at 100°C for 5 min, were separated by SDS-PAGE as detailed below. After electrophoresis, proteins were transferred to polyvinylidene difluoride membranes and probed with primary antibodies against RyR1 or triadin, diluted 1:1000 in phosphate-buffered saline. After washing, membranes were incubated with secondary antibodies conjugated to horseradish peroxidase (1:5000 dilution in phosphatebuffered saline). Protein-antibody reactions were detected with an ECL kit (Amersham Biosciences, Uppsala, Sweden).
Radioactive ]. The amount of CaCl 2 , EGTA, and MgCl 2 needed for a given condition was calculated with the WinMaxC software as above. Incorporation of 35 S radioactivity into RyR1 channels was studied either in sucrose gradient fractions enriched in the channel protein or directly following electrophoresis of labeled triads in non-reducing SDS-containing polyacrylamide gels. For sucrose gradient RyR1 channel purification (58), SR vesicles incubated with [ 35 S]GSNO were sedimented at 45,000 ϫ g for 30 min (4°C). The resulting pellets were resuspended at 2 mg/ml and incubated with 0.5% Triton X-100, 1 mM CaCl 2 , 20 mM imidazole-MOPS, pH 7.2, for 20 min at room temperature under constant stirring, and sedimented at 45,000 ϫ g for 30 min at 4°C. The pellet was resuspended at 2 mg/ml in 3 mM EGTA, 20 mM MOPS-Tris, pH 9.0, incubated for 10 min at 4°C, and sedimented at 45,000 ϫ g for 30 min at 4°C. This procedure allows recovery of a fraction enriched in RyR channels, essentially devoid of Ca 2ϩ -ATPase and calsequestrin. The RyR1-enriched fraction was resuspended at 1 mg/ml in solubilization buffer solution, containing (in mM): 1000 NaCl, 0.2 CaCl 2 . 0.1 EGTA, 20 MOPS-Tris, pH 7.2, plus 2% CHAPS and 1% L-␣-phosphatidylcholine. This mixture was incubated for 60 min at 25°C, followed by 60 min at 4°C, layered onto a 5-20% linear sucrose gradient, and centrifuged at 120,000 ϫ g for 17 h at 4°C. The sucrose solutions used to generate the gradients were made in solubilization buffer plus 1% CHAPS and 0.5% L-␣-phosphatidylcholine. To identify channel-containing fractions, parallel sucrose gradients were run with RyR1 channels solubilized from triads incubated with non-radioactive 4 mM GSNO and 2 nM [ 3 H]ryanodine. Gradient fractions, collected in 1-ml aliquots, were analyzed for protein content (60) and for 35 S or 3 H radioactivity in a liquid scintillation counter. The protein composition of the fractions was analyzed by electrophoresis in 6% polyacrylamide-SDS gels under reducing conditions. Alternatively, the 35 S radioactivity incorporated into the RyR1 channel protein was determined in RyR1-containing gel fractions separated by electrophoresis under non-reducing conditions. To this purpose, triads incubated with [ 35 S]GSNO were sedimented at 178,000 ϫ g for 5 min at room temperature in a Beckman Airfuge. The resulting pellets were resuspended in non-reducing sample buffer (6 M urea, 1% SDS, 0,02% bromphenol blue, 48 mM NaH 2 PO 4 , 170 mM Na 2 HPO 4 ). Samples, denatured at 100°C for 5 min, were separated in polyacrylamide gradient gels (3.5-8%) using the Tris-acetate buffer system (Novex NuPAGE®, Invitrogen, Carlsbad, CA). Gels, stained with Coomassie Brilliant Blue, were cut in 2-mm slices, the slices were trypsinized, and their radioactivity was determined in a liquid scintillation counter. Alternatively, gels were dried and phosphorimaging was performed using a Phosphor Screen CP (Eastman Kodak Co.) and the Molecular Imager FX system (Bio-Rad). Screens were scanned, and the images were quantified using the Quantity One software (Bio-Rad).

RESULTS
Fast Release Kinetic Measurements-The time course of CICR, measured in the presence of 1.5 mM free ATP, from native and SH-modified triads is illustrated in Fig. 1. In native triads Ca 2ϩ release followed a single exponential time course with an average rate constant( (k) value of 11.7 s Ϫ1 at Ͻ25 M free [Mg 2ϩ ] Table I). The experimental record obtained in one preparation that displayed a k value of 11.4 s Ϫ1 is illustrated in Fig. 1, panel A. In agreement with previous results (20), increasing free [Mg 2ϩ ] to 400 M produced a strong inhibition of CICR ( Fig. 1, panel B), decreasing k in this case to 1.6 s Ϫ1 , with an average value of 1.4 s Ϫ1 (Table I). At free [Mg 2ϩ ] Ͻ 25 M, the same triad preparation incubated with 1 mM GSSG exhibited a k value similar to that for controls, 12 s Ϫ1 (Fig. 1, panel  C); the average k value was 12.1 s Ϫ1 (Table I). However, in 400 M free [Mg 2ϩ ], GSSG-treated vesicles had a 3-fold higher k value than control, 4.6 s Ϫ1 (Fig. 1, panel D), with an average value of 5.0 s Ϫ1 (Table I). At Ͻ25 M free [Mg 2ϩ ], triads incubated with 100 M GSNO displayed a higher k value than control vesicles, 17.2 s Ϫ1 (Fig. 1, panel E), with an average value of 15.8 s Ϫ1 (Table I). In 400 M free [Mg 2ϩ ], these same triads also displayed higher k values than controls, 8.2 s Ϫ1 (Fig.  1, panel F), with an average k value of 8.9 s Ϫ1 (Table I). Incubation with the NO donor NOR-3 produced at Ͻ25 M free [Mg 2ϩ ] a k value of 21 s Ϫ1 (Fig. 1, panel G), with an average value of 25.2 s Ϫ1 , which is 2-fold higher than control (Table I). However, when release was measured in 400 M free [Mg ϩ2 ], triads incubated with NOR-3 displayed a low k value of 1.9 s Ϫ1 (Fig. 1, panel H), with an average value of 2.3 s Ϫ1 (Table I).
These results indicate that, in the incubation conditions used in this work, GSNO and NOR-3, but not GSSG, stimulated CICR when measured at Ͻ25 M free [Mg 2ϩ ]. These results suggest that GSNO and NOR-3 promoted through S-nitrosylation an increase in the Ca 2ϩ sensitivity of CICR. In addition, GSSG and GSNO, but not NOR-3, relieved the strong inhibitory effect on k exerted by Mg ϩ2 .
To test the hypothesis that S-nitrosylation increases the a The error range of only two determinations is given.
activation of CICR by Ca 2ϩ , we measured CICR kinetics at pCa 6, 5.6, and 5 in triads incubated with NOR-3. A comparison of the k values obtained in control and NOR-3 incubated triads is given in Fig. 2. Triads incubated with NOR-3 displayed higher k values than controls at pCa 6, 5.6, and 5, with the highest stimulation at pCa 6 ( Fig. 2, hatched columns) indicating that NOR-3 increased significantly the affinity of the release channels for Ca 2ϩ .
The inhibitory effects of free [Mg 2ϩ ] on release rate constants for native and redox modified triads are compared in Fig. 3. In agreement with previous results (20), increasing free [Mg 2ϩ ] inhibited calcium release rate constants in native triads with K 0.5 ϭ 67 Ϯ 15 M (Fig. 3A). Incubation with either GSSG (Fig.  3B) or GSNO (Fig. 3C) resulted in an important reduction of the inhibitory effect of Mg 2ϩ . Triads incubated with 1 mM GSSG had a K 0.5 value for Mg 2ϩ inhibition of 275 Ϯ 52 M, which increased to Ͼ2 mM for triads incubated with 5 mM GSSG (Fig. 3B). The corresponding K 0.5 value for triads incubated with 100 M GSNO was 400 Ϯ 96 M (Fig. 3C). In contrast, release rate constants in triads incubated with NOR-3 were strongly inhibited by increasing free [Mg 2ϩ ], with a K 0.5 ϭ 60 Ϯ 18 M (Fig. 3D). This value is comparable to the K 0.5 value of 67 Ϯ 15 M determined in native vesicles. Incubation of triads with GSH did not modify calcium release rate constants or K 0.5 for Mg 2ϩ inhibition when compared with control triads (data not shown).
The effects of increasing [GSNO] and [GSSG] on the rate constant values k of CICR, measured at low or at high free [Mg 2ϩ ], are compared in Fig. 4. Near-maximal release rate constant stimulation was obtained at 100 M GSNO in release solution containing Ͻ25 M free [Mg 2ϩ ] (Fig. 4A, open squares). In contrast, when measured in Ͻ25 M free [Mg 2ϩ ], increasing [GSSG] up to 5 mM did not affect significantly release rate constants (Fig. 4B, open circles) To ascertain specific glutathionylation of RyR channels and to quantify this incorporation, it is essential to rule out the presence of other proteins that may contaminate RyR1-enriched gradient fractions. Accordingly, we carried out an additional set of experiments to detect 35 S incorporation into RyR1 channels separated by gel electrophoresis under non-reducing conditions. A Coomassie Blue-stained non-reducing gel of triads incubated with varying concentrations of [ 35 S]GSNO, from 12.5 M to 500 M, is shown in Fig. 6A. The corresponding phosphorimage is shown in Fig. 6B. Incubation of triads with 12.5-500 M [ 35 S]GSNO at pCa 5 produced increasing 35 S incorporation into RyR1 channels and other triadic proteins (Fig. 6B). Some radioactive material found at the top of the gel stained faintly with Coomassie Blue and may represent a minor component of aggregated proteins. The second band of radioactivity corresponded to the RyR1 monomer. Lower molecular weight bands had apparent molecular weights consistent with triadin and triadin oligomers. These assignments of band identities were confirmed with Western blots using specific antibodies, as illustrated in Fig. 7. Whereas the RyR1containing band did not shift under non-reducing conditions, the single triadin-containing band observed in reducing gels appeared as multiple higher molecular weights bands in nonreducing gels. However, the RyR1-containing protein band did not contain associated triadin after separation in non-reducing gels (Fig. 7). This migration pattern was not affected by incubation of triads with up to 1 mM GSNO (data not shown). The incorporation of radioactivity from [ 35 S]GSNO into RyR1 increased with increasing concentrations of [ 35 S]GSNO; halfmaximal incorporation was obtained at 157 Ϯ 29 M (Fig. 6C). The half-time for radioactive labeling of RyR1 channels was 8 min (data not shown). In the following experiments, we incubated triads with GSNO for 20 min as done in kinetic experiments.
To quantify the number of S-glutathionylated SH residues per channel, RyR1-containing bands were sliced from the gels and their 35 Table II. We found no differences in the radioactivity incorporated into RyR1 channels at pCa 5 or at pCa 7, either in the absence or in the presence of 0.8 mM free [Mg 2ϩ ]. Further decreasing free [Ca 2ϩ ] to pCa 9 reduced to 67% the incorporation of 35 S. The NO scavenger PTIO reduced the incorporation of 35 S into RyR channels; 50 or 25% of the incorporation observed at pCa 5 was attained in the presence of 100 or 300 M PTIO, respectively (Table II)

DISCUSSION
Recent reports reveal RyR channels as highly sensitive redox-sensing proteins (33,39,40,60). We have shown previously that thimerosal stimulates RyR1 single-channel activity and modifies the pattern of RyR1 activation by Ca 2ϩ , shifting to the left the Ca 2ϩ activation curve of channel open probability (23) even in purified RyR channels (61). We have also reported that incubation of triad vesicles from skeletal muscle with 250 M thimerosal enhances CICR kinetics when measured at low free [Mg 2ϩ ] (Ͻ 25 M) and decreases the inhibition of CICR exerted by 1 mM free [Mg 2ϩ ] (20). Furthermore, although incubation with 500 M thimerosal does not enhance CICR at low free [Mg 2ϩ ], it completely abolishes CICR inhibition by 1 mM free [Mg 2ϩ ] (20). These combined results suggest that thimerosal reacts with different SH residues of the channel protein to regulate either channel activation by Ca 2ϩ or inhibition by Mg 2ϩ .
Several endogenous redox molecules, including H 2 O 2 , GSH, GSSG, GSNO, and CysNO, modify RyR channel activity in vitro (21,36,37,39,41,42). Consequently, we determined whether the intracellular redox agents GSNO and GSSG stimulate CICR kinetics mediated by skeletal RyR1 channels. We also assessed the effects of NOR-3, an NO donor that, in analogy with other NO donors (38), presumably promotes S-nitrosylation of RyR1 channels. In addition, we determined whether RyR1 channels undergo S-glutathionylation following incubation with [ 35 S]GSNO. Results obtained in kinetic experiments will be discussed first.
Differential Effects of GSNO, GSSG, and NOR-3 on CICR Kinetics-Our studies show that GSSG, an S-glutathionylating agent, and NOR-3, which S-nitrosylates protein SH residues, have different effects on CICR kinetics. We also found that GSNO produced effects similar to the combined effects of NOR-3 and GSSG, and modified CICR kinetics in a similar fashion as thimerosal (20), suggesting that GSNO and thimerosal may modify the same free SH residues. Both NOR-3 and GSNO, but not GSSG, stimulated CICR kinetics when measured in Ͻ25 M free [Mg 2ϩ ], as reflected in a 2-fold increase of Ca 2ϩ release rate constants k (see Table I). This stimulation suggests that NOR-3 and GSNO S-nitrosylate SH residues that either directly or indirectly participate in RyR1 channel activation by Ca 2ϩ . The marked stimulation of release rate constants observed at pCa 6 in triads incubated with NOR-3 support this proposal. Additionally, both GSSG and GSNO, but not NOR-3, reduced the strong inhibitory effect of [Mg 2ϩ ] on CICR, increasing significantly the free [Mg 2ϩ ] needed to inhibit k values by half. These findings suggest that, under our experimental conditions, GSSG and GSNO both S-glutathionylate SH residues and alter the low affinity Mg 2ϩ inhibitory site. Whereas, in addition to its S-glutathionylating properties, GSSG produces changes in redox status through changes in the GSSG/GSH ratio (40,62), our results clearly show that GSNO induces cysteine S-glutathionylation. Previous reports indicate that both skeletal and cardiac RyR channels are endogenously S-nitrosylated (33,63). Skeletal RyR1 channels can also be exogenously S-nitrosylated with GSNO, which enhances [ 3 H]ryanodine binding in a dose-dependent manner both at atmospheric and at reduced O 2 concentration (42). GSNO decreases RyR1 free sulfhydryl content, partly through S-nitrosylation of cysteines other than Cys-3635 and partly through other oxidation reactions (42).
S-Glutathionylation of Skeletal RyR Channels-Several recent reports have described covalent modification of protein SH residues by GSSG or GSNO through the formation of S-glutathionyl derivatives. The growing list of these proteins include glyceraldehyde-3-phosphate dehydrogenase, c-Jun, SERCA1, creatine kinase, H-Ras, protein phosphatase 2A, NFB, thioredoxin, tyrosine hydroxylase, mitochondrial complex I, and actin (47-49, 64 -72). Cells generate both GSSG and GSNO from GSH, which is present in mM concentrations in the cytoplasm (62). Oxidation of GSH generates GSSG, whereas GSNO is formed by reaction of GSH with NO (73,74). It has been proposed that GSNO, which has a longer half-life in cells than NO (75), serves as a storage molecule of likely physiological relevance because it can either act as an S-nitrosylating or an S-glutathionylating agent (46,76). However, except from our previous short report (37), S-glutathionylation of RyR1 channels has not been reported.
The present results, which show that incubation of skeletal triads with [ 35 S]GSNO resulted in 35 S incorporation into RyR1 channels, constitute a direct demonstration of S-glutathionylation of RyR1 channel SH residues. To our knowledge, this is the first report to describe S-glutathionylation of an ion channel. Incubation with 500 M GSNO resulted in near-maximal Sglutathionylation, with values of 180 pmol/mg of triad protein (see "Results"). To convert this number into incorporation per RyR1 channel monomer, we considered the average maximal density of [ 3 H]ryanodine binding sites of our triad preparations, 18 pmol/mg of protein. On the assumption that each RyR1 homotetramer binds only one molecule of [ 3 H]ryanodine (77,78), this B max value yields 18 pmol of RyR1 homotetramer/mg of protein. Accordingly, the 180 pmol of SH residues S-glutathionylated/mg of triad protein obtained following incubation with 500 M GSNO would correspond to 10 SH residues S-glutathionylated per tetramer, or 2.5 residues/monomer.
Near-maximal effects on release rate constants measured in 0.3 mM free [Mg 2ϩ ] were obtained with 500 M GSNO (Fig. 4B). This finding suggests that S-glutathionylation of ϳSH residues per RyR1 monomer (Fig. 6C) is enough to decrease Mg 2ϩ inhibition of CICR. On the other hand, near-maximal effects on CICR kinetics measured in Ͻ25 M free [Mg 2ϩ ] were already observed at 100 M GSNO (Fig. 4A). Whereas we did not determine the number of SH residues S-nitrosylated by GSNO, we estimate from published data (42) that S-nitrosylation of no more than 2 SH residues/RyR1 monomer is responsible for the enhancement of channel activity by Ca 2ϩ . The present results, however, do not rule out the possibility that redox modifications of RyR1 channel associated proteins, such as triadin, may contribute to the effects of GSNO on CICR kinetics. Our previous findings (61) indicate that redox modification of purified RyR channels with thimerosal makes single channels more responsive to activation by micromolar [Ca 2ϩ ] and less susceptible to inhibition by 0.5 mM [Ca 2ϩ ], which at this high concentration occupies the Mg 2ϩ inhibitory site. On this basis, we propose that the stimulation of CICR kinetics by GSNO reported in this work is caused by redox modifications of the RyR1 channel itself.
The same extent of S-glutathionylation was obtained when triads were incubated at pCa7 or at pCa 5; addition of 0.8 mM free [Mg 2ϩ ] did not affect 35 S labeling. However, S-glutathionylation decreased Mg 2ϩ inhibition of CICR. These results suggest that Mg 2ϩ binding to its low affinity inhibitory site does not interfere with the ability of GSNO to react with the SH residues involved in controlling the affinity of the site for Mg 2ϩ . In contrast, covalent attachment of the glutathionyl group to these SH residues would hinder the accessibility of Mg 2ϩ to this site, either because these SH residues are directly present in the site or because, as a consequence of S-glutathionylation, they induce a channel conformational change that hinders Mg 2ϩ access.  The above findings suggest that under resting conditions, i.e. pCa 7 plus 0.8 mM [Mg 2ϩ ], skeletal muscle RyR1 channels can readily undergo S-glutathionyl incorporation. Only a substantial decrease in free [Ca 2ϩ ] to pCa 9 produced a significant decrease of label incorporation. Therefore, we propose that very low free [Ca 2ϩ ] hinders the accessibility of hyperreactive SH groups to GSNO presumably by inducing RyR1 channel conformational changes. Similar effects of low free [Ca 2ϩ ] have been reported for other RyR channel redox-active modulators (22,40,79).
Comparison between S-Glutathionylation and S-Nitrosylation of RyR Channels-Previous reports indicate that skeletal muscle RyR1 channels are activated by NO-induced S-nitrosylation of a single SH residue at reduced but not at atmospheric O 2 concentrations (33). The S-nitrosylation reaction was specific for RyR1 channels among SR proteins, and its effects depended on the presence of calmodulin. Moreover, NO and NO-donors S-nitrosylate Cys3635 (38,42) whereas GSNO-induced S-nitrosylation, is calmodulin-independent, takes place either at reduced or at atmospheric O 2 concentrations and does not involve Cys3635 (42). Interestingly, incubation with GSNO removes more free SH residues from skeletal muscle RyR1 channels than the residues actually S-nitrosylated (42), indicating that GSNO induces other SH modifications. Here, we report as a novel finding S-glutathionyl modification by GSNO of RyR1 channel SH residues as well as the functional consequences of GSNO-induced redox modification. If GSNO also induced S-nitrosylation and S-glutathionylation of cardiac RyR channels, this double feature would explain why GSNO removes more free SH residues from purified cardiac RyR channels than those actually S-nitrosylated (63).
Physiological Implications-Redox modulation of cellular signal transduction processes has received increased attention during the last few years (46, 80 -83). The intracellular redox state is mainly controlled by the GSH/GSSG ratio (62). Nevertheless, redox species, and hence cellular redox state, may vary in response to physiological stimuli. Currently, S-nitrosothiols are regarded as intracellular nitric oxide transporters/donors and S-nitrosylating/thionylating agents (82,84). In particular, S-glutathionylation has been recently considered among the physiologically relevant redox modifications and, along with S-nitrosylation, may be responsible for cysteine/cystine homeostasis within cells (46). Recent reports using proteome analysis have shown that S-glutathionylation takes place as a posttranslational protein modification (50,51).
CICR mediated by RyR channels represents an important signal transduction amplification mechanism, by means of which a localized [Ca 2ϩ ] increase generated by Ca 2ϩ entry into a cell induces Ca 2ϩ release from intracellular stores and significantly enhances the initial Ca 2ϩ signal. Through CICR, Ca 2ϩ signals can be propagated from the cell periphery into the nucleus, where nuclear Ca 2ϩ can induce gene expression (1,2,85). Accordingly, endogenous activation of CICR by cellular redox agents such as NO, GSNO, GSSG, and H 2 O 2 , may have important physiological consequences and, if uncontrolled, is likely to bring about pathological conditions. Taken together, our results show that both GSSG and GSNO enhanced RyR1 channel-mediated CICR from mammalian skeletal muscle triads. In addition, our data show that Mg 2ϩ inhibition of RyR channel activity can be modified by GSSG or GSNO, but in the conditions used here not by a purely Snitrosylating agent such as NOR-3. Both GSSG and GSNO are generated in vivo, GSSG under oxidative stress conditions and GSNO as a physiological by-product of NO generation. Whereas GSSG may promote channel activation by S-glutathionylation, GSNO is likely to do so by a combination of S-glutathionylation and S-nitrosylation of the RyR1 channel protein.
The RyR1 channel isoform has a central role in excitationcontraction (E-C) coupling in skeletal muscle (3,(5)(6)(7)(8). Skeletal muscle E-C coupling requires direct or indirect interactions between the RyR1 channels of junctional sarcoplasmic reticulum and transverse tubule voltage sensors (86). Accordingly, endogenous redox modifications that affect RyR1 channel activity may have consequences on E-C coupling and muscle function. In particular, if the decrease of the strong inhibitory effect of Mg 2ϩ on CICR observed in vitro after RyR1 channel modification with GSNO or GSSG also took place in intact muscle fibers, the initial voltage-induced Ca 2ϩ release would be amplified by CICR, enhancing skeletal muscle contraction. In addition, GSNO-induced S-nitrosylation would also enhance CICR by increasing the Ca 2ϩ sensitivity of this process. Excessive production of endogenous GSSG or GSNO, such as might happen during oxidative stress conditions, may cause pathological states through uncontrollable stimulation of CICR. However, studies on redox regulation of E-C coupling in skinned or intact skeletal muscle fibers have yielded contradictory results. Addition of hydrogen peroxide stimulates caffeineinduced contraction in skinned fibers but does not affect action potential generated contraction (87). Clearly, a definite demonstration of the relevance of RyR1 redox modification in the context of physiological gating of the channel is still missing.
To conclude, the possible role of GSSG and GSNO as endogenous redox modulators of RyR channels should be addressed in other systems such as brain, where RyR channel-mediated Ca 2ϩ release is a requisite step in some types of neuronal plasticity (28).