Multiple classes of sulfhydryls modulate the skeletal muscle Ca2+ release channel.

Two sulfhydryl reagents, N-ethylmaleimide (NEM), an alkylating agent, and diamide, an oxidizing agent, were examined for effects on the skeletal muscle Ca2+ release channel. NEM incubated with the channel for increasing periods of time displays three distinct phases in its functional effects on the channel reconstituted into planar lipid bilayers; first it inhibits, then it activates, and finally it again inhibits channel activity. NEM also shows a three-phase effect on the binding of [3H]ryanodine by first decreasing binding (phase 1), followed by a recovery of the binding (phase 2), and then a final phase of inhibition (phase 3). In contrast, diamide 1) activates the channel, 2) enhances [3H]ryanodine binding, 3) cross-links subunits within the Ca2+ release channel tetramer, and 4) protects against phase 1 inhibition by NEM. All diamide effects can be reversed by the reducing agent, dithiothreitol. Diamide induces intersubunit dimer formation of both the full-length 565-kDa subunit of the channel and the 400-kDa generated by endogenous calpain digestion, suggesting that the cross-link does not involve sulfhydryls within the N-terminal 170-kDa fragment of the protein. NEM under phase 1 conditions blocks the formation of the intersubunit cross-links by diamide. In addition, single channels activated by diamide are further activated by the addition of NEM. Diamide either cross-links phase 1 sulfhydryls or causes a conformational change in the Ca2+ release channel which leads to inaccessibility of phase 1 sulfhydryls to NEM alkylation. The data presented here lay the groundwork for mapping the location of one of the sites of subunit-subunit contact in the Ca2+ release channel tetramer and for identifying the functionally important sulfhydryls of this protein.

The Ca 2ϩ release channel of skeletal muscle sarcoplasmic reticulum is a homotetramer with a subunit molecular mass of 565 kDa (1). The channel opens in response to a signal from the transverse tubules, which is triggered by membrane depolarization (2), and the resulting flux of Ca 2ϩ from the sarcoplasmic reticulum initiates the sequence of events that leads to muscle contraction. Several laboratories have clearly demonstrated that the activity of the Ca 2ϩ release channel is modulated by oxidation-reduction reactions (3)(4)(5)(6)(7)(8)(9)(10)(11). Oxidizing agents that stimulate Ca 2ϩ release from the sarcoplasmic reticulum include H 2 O 2 (3), 2,2Ј-dithiodipyridine, 4,4Ј-dithiodipyridine (4), Cu 2ϩ phthalocyanine dyes (5), anthraquinone doxorubicin (6,7), and thimerosal (8). Heavy metals such as Hg 2ϩ , Ag ϩ , Cu 2ϩ , Cd 2ϩ , and Zn 2ϩ have also been reported to induce Ca 2ϩ release, either by directly interacting with a sulfhydryl or by causing oxidation (9). These studies led Abramson and Salama (10) to propose a model for redox modulation of the channel that involves three different sulfhydryl groups that exist in close proximity and that can form mixed disulfides to open or close the channel. The evidence that the channel can be altered by oxidation is conclusive (3)(4)(5)(6)(7)(8)(9)(10)(11); the question is whether this oxidation plays a physiological role in skeletal muscle. It is not yet known if disulfide interchange or oxidation-reduction of sulfhydryls on the Ca 2ϩ release channel contribute to normal excitation-contraction coupling, but an increasing body of evidence suggests that such a mechanism could be an important modulatory element. Under basal conditions, unfatigued skeletal muscle produces reactive oxygen species (12) and nitric oxide (NO) derivatives (13,14) that have been shown to modulate excitation-contraction coupling (12,13). Strenuous contractile activity increases reactive oxidant production (14 -17), which contributes to fatigue of both isolated muscle preparations (15,16,18,19) and human muscle in vivo (20). Redox modulation of the ryanodine-binding protein has been proposed as a common mechanism for these effects (21).
Proteins other than the Ca 2ϩ release channel may contribute to its modulation by oxidizing agents. Pessah and co-workers (22,23) have suggested that oxidation may involve the crosslinking of triadin to the Ca 2ϩ release channel. The functional role of triadin in skeletal muscle and its relationship to the Ca 2ϩ release channel have remained elusive. Caswell and coworkers (24) have suggested that it is involved in coupling of the Ca 2ϩ release channel to the t-tubule voltage sensor. Others (25) on the basis of its putative arrangement in the membrane have argued against a role for triadin in connecting the voltage sensor to the Ca 2ϩ release channel. There is, however, general agreement that this protein interacts with the Ca 2ϩ release channel (24,26). The functional significance of this interaction is not yet known.
In addition to modulation by oxidation-reduction, the Ca 2ϩ release channel is also sensitive to reagents that react with free sulfhydryls but do not form disulfide bonds. N-Ethylmaleimide (NEM), 1 a sulfhydryl alkylating agent, activates Ca 2ϩ release at low concentrations, while it inhibits release at higher concentrations (27). 2 NEM, however, induces Ca 2ϩ release with a slower onset than the heavy metals, a finding that has been interpreted to mean that the sulfhydryls responsible for the observed effects on Ca 2ϩ release are in a hydrophilic environment (9). Quinn and Ehrlich (28) reported that modification of cysteines on the Ca 2ϩ release channel by methiosulfonate compounds reduces the conductance of the channel. The reaction occurs only when the channel is in the open state. These compounds reduce the conductance in multiple steps until complete closure of the channel is obtained, suggesting a reactive sulfhydryl in the ion conducting pathway. The multiple effects of sulfhydryl reagents on the activity of the Ca 2ϩ release channel raise the question of whether all of these reagents are reacting with the same sulfhydryls or whether alteration of multiple classes of sulfhydryls can alter the function of the channel.
Each subunit of the Ca 2ϩ release channel tetramer has 100 cysteines (1). In this report we attempt to distinguish between classes of functionally important sulfhydryls. We show that the alkylating reagent NEM has three distinct functional effects on the channel and the oxidizing agent, diamide (29), produces intersubunit cross-links within the tetrameric channel. Additional information about the location of the cross-links is obtained using a membrane preparation that has both the fulllength 565-kDa subunit and the 400-kDa fragment generated by the action of endogenous calpain.
Sarcoplasmic Reticulum (SR) Membrane Preparation-SR membranes were prepared from rabbit backstrap and hindleg skeletal muscle and were purified by using sucrose gradient centrifugation as described elsewhere (30,31). Protein was estimated by the method of Lowry et al. (32), using BSA as standard.
Cross-linking-For binding experiments, membranes in buffer I (300 mM NaCl, 100 M CaCl 2 , 50 mM MOPS (pH7.4)) were incubated with 100 to 500 M diamide for 30 min at 4°C or for 10 min at room temperature. Diamide was removed either by dilution (10-fold) or by pelleting and washing the membranes in a Beckman Airfuge by centrifuging 4 min at 30 p.s.i. For bilayer experiments the sample was incubated for 10 min at room temperature with 500 M diamide in buffer II (225 mM CsSO 3 CH 3 , 10 M CaCl 2 , 10 mM MOPS, pH 7.4). Membranes cross-linked in buffer I or II were washed and assayed for [ 3 H]ryanodine binding and for changes in gel patterns by SDS-PAGE. All [ 3 H]ryanodine binding assays were done in buffer I.
SDS-PAGE Electrophoresis in One and Two Dimensions-SR membranes (20 -40 g) were cross-linked with diamide for 10 min at room temperature (23°C) or 30 min on ice. The samples were treated with 5 mM NEM for 20 min at room temperature before solubilization in sample buffer. Electrophoresis on 5% SDS-PAGE was performed for the first dimension. Electrophoresis was continued for 15 min after the dye front ran off gel. The lanes were excised and used for second dimension electrophoresis. The excised gel strips were treated with 50 mM DTT in sample buffer without SDS and bromphenol blue for 60 min at room temperature (23°C). The gel strips were loaded on the second dimension gel (5% SDS-PAGE). The gap was sealed with melted agarose (1% agarose, 2% SDS, 50 mM DTT). After electrophoresis, the second dimension gel was silver-stained.
Equilibrium each of aprotinin, leupeptin, and pepstatin A. Nonspecific binding was defined in the presence of either 10 M or 100 M ryanodine. Bound [ 3 H]ryanodine was separated from free by rapid filtration of the sample through Whatman GF/F glass fiber filters followed by five 3-ml washes with ice-cold wash buffer containing 0.3 M NaCl, 100 M CaCl 2 , and 10 mM MOPS (pH 7.4). The radioactivity bound to the filters was quantitated by liquid scintillation counting of the filters in 5 ml of Ultima Gold scintillant.
Bilayer Techniques-Planar bilayers consisting of 8:2 L-␣-phosphatidylethanolamine and L-␣-phosphatidylcholine were formed following the Mueller-Rudin procedure across a 100-m diameter aperture in Teflon cups as described previously (34). The mixture of phospholipids was dissolved in n-decane (Sigma) at a concentration of 50 mg/ml. Both chambers were filled with buffer solution (20 mM CsSO 3 CH 3 , 10 M CaCl 2 , 10 mM MOPS, pH 7.4). After bilayer formation, 5 l of SR membranes were added to the cis chamber to give a final protein concentration of 0.5 g/ml. The other side of the bilayer was defined as trans and is the ground. An osmotic gradient was formed between the cis and trans chambers by adding a concentrated salt solution (4 M CsSO 3 CH 3 , 10 M CaCl 2 , 10 mM MOPS, pH 7.4) to the cis chamber. Recording solutions contained 225 mM CsSO 3 CH 3 in both the cis and trans chambers. All additions were made to the cis chamber. Agar/KCl bridges were used to connect the chambers to Ag/AgCl electrodes immersed in 2 M KCl. The holding potential was ϩ40 mV. The data were filtered at 2.5 KHz and digitized at 10 KHz.
Data Analysis-[ 3 H]Ryanodine binding data were analyzed with nonlinear curve fitting using SigmaPlot (Jandel Scientific). Nonspecific binding was subtracted prior to analysis.
Single-channel recordings were analyzed using

NEM Has Multiple Effects on Both Channel Activity and [ 3 H]Ryanodine
Binding-NEM is an alkylating reagent, which reacts primarily with cysteine residues (35). The effect of NEM alkylation on the activity of the Ca 2ϩ release channel reconstituted into planar lipid bilayers is shown in Fig. 1A. Immediately after the addition of NEM, the channel was inhibited (phase 1). Continued exposure to the alkylating agent (3-10 min) produced a substantial activation (phase 2), but longer incubations led again to inhibition of channel activity (phase 3). This stepwise effect of NEM is shown in Fig NEM also has time-dependent effects on [ 3 H]ryanodine binding. The effect of reaction of the membranes with 5 mM NEM is shown in Fig. 2A. For binding experiments, NEM was reacted in binding buffer at 4°C and the alkylation was stopped by the addition of 10 mM DTT. Similar to the effects on channel activity, there were three distinct phases of the alkylation that altered [ 3 H]ryanodine binding. The three phases are an initial and rapid inhibition (phase 1), a recovery or enhancement of binding (phase 2), and a second inhibitory step (phase 3). The phase 1 decrease with 5 mM NEM reaches 49.1 Ϯ 4.4% (n ϭ 3) inhibition prior to the onset of phase 2. The time course of these three phases was dependent on the concentration of NEM (Fig.  2B). Lower concentrations of NEM (500 M) slow the onset of phases 2 and 3, allowing phase 1 to plateau at 54.6 Ϯ 5.7% (n ϭ 4) inhibition. If the sulfhydryls that have reacted at 500 M NEM are primarily phase 1 sulfhydryls, it should be possible to add 5 mM NEM to membranes at the plateau stage and recover the phase 2 enhancement of [ 3 H]ryanodine binding. This experiment is shown in Fig. 2C. The addition of 5 mM NEM 25 min after the 500 M NEM addition gave rise to the recovery of binding (phase 2) followed by phase 3 inhibition. The reaction with NEM under bilayer conditions was faster than that obtained using the protocol employed for Fig. 2 (A-C). For comparison, the effect on [ 3 H]ryanodine binding of the reaction of membranes with 500 M and 5 mM NEM in bilayer buffer and at room temperature is shown in Fig. 2D. The three phases were still seen, and the time of onset of each phase corresponded to those observed in the bilayer; however, the extent of inhibition was always less under bilayer conditions, possibly reflecting the difference in the redox state of the channel under these conditions. All membrane preparations that we tested showed the three-phase effect of NEM. Some variation in the magnitude of the three phases was, however, observed. Some air oxidation of sulfhydryls did occur, and the difference in the three phases may reflect differences in the oxidation state of the membranes.

Diamide Treatment Increases Open Probability of the Channel, Enhances [ 3 H]Ryanodine Binding, and Forms Intersubunit
Cross-links-The sulfhydryl oxidizing agent, diamide, can also activate the channel reconstituted into planar lipid bilayers Treatment of membranes with diamide caused an alteration in the electrophoretic mobility of the 565-kDa subunit of the Ca 2ϩ release channel. In diamide cross-linking experiments with either membranes (Fig. 4A) or purified Ca 2ϩ release channel, 3 the 565-kDa band disappeared and a new band was seen with an electrophoretic mobility consistent with dimer formation. Approximate molecular mass values of the oligomers were determined using a 200-kDa myosin standard and the fulllength ryanodine-binding protein as a 565-kDa standard. Higher concentrations of diamide produced higher oligomers of the 565-kDa standard, which did not enter the gel. To demonstrate that the high molecular weight bands generated by diamide treatment are indeed dimers formed by intersubunit cross-linking and not by the cross-linking of the Ca 2ϩ release channel to other proteins, we performed two-dimensional electrophoresis, reducing the disulfides between the first and second dimensions (37). The two-dimensional SDS-PAGE is shown in Fig. 4B. The membranes used in the experiment shown in Fig. 4B had a significant amount of a 400-kDa fragment of the Ca 2ϩ release channel, which is derived from the 565-kDa band by calpain digestion (38,39). Both bands underwent crosslinking to form higher molecular weight complexes. Upon reduction prior to the second dimension (Fig. 4B), the high molecular weight bands decreased to 565-kDa and 400-kDa bands. Surprisingly, very little cross-linking was detected between the 400-kDa and the 565-kDa proteins. Instead both the 565-kDa and the 400-kDa proteins appeared to form only homodimers, suggesting that, when the 565-kDa is digested to the 400-kDa, the other subunits within the tetramer are also digested. Although only high molecular weight bands are shown in this gel, we have extensively searched for lower molecular weight bands such as triadin and have found no evidence that any other protein is involved in the diamide-induced cross-linking to the Ca 2ϩ release channel (data not shown).
To determine if the formation of intersubunit cross-links correlates with channel activation, membranes were treated with increasing concentrations of diamide. The effect of these treatments on [ 3 H]ryanodine binding is shown in Fig. 5A. Diamide treatment enhanced [ 3 H]ryanodine binding. To quantitate the formation of dimers, the Coomassie Brilliant Bluestained gel was scanned with a densitometer and the optical densities of the dimer and monomer were plotted as a function of diamide concentration (Fig. 5B). The disappearance of the 565-kDa band correlated with the appearance of the dimer and with enhanced [ 3 H]ryanodine binding. At higher concentrations of diamide (1 mM and higher), the intensity of the dimer also decreased and higher oligomers appeared to accumulate at the top of the gel. Formation of higher oligomers was accompanied by a decrease in [ 3 H]ryanodine binding. Channels pretreated with diamide and washed prior to incorporation into planar lipid bilayers showed a substantial activation compared to controls (Fig. 5C). In the experiment shown, the P o increased from 0.02 to 0.22 with 500 M diamide.

Diamide and NEM Can Be Used to Differentiate between
Classes of Sulfhydryls on the Ca 2ϩ Release Channel-To determine whether diamide and NEM react with the same sulfhydryls, we examined the effect of NEM on the diamide-activated channel. The single-channel records are shown in Fig. 6. The first tracing is the control (P o ϭ 0.03 Ϯ 0.01, n ϭ 3). Addition of 250 M diamide activated the channel (P o ϭ 0.12 Ϯ 0.03, n ϭ 3). The addition of 5 mM NEM to the diamide-modified channel stimulated channel activity further. The P o with NEM was 0.53 Ϯ 0.04 (n ϭ 3) within the first 5 min. This activation is not reversed by DTT (data not shown). NEM does not cause phase 1 inhibition in the diamide-pretreated membranes. Sulfhydryls involved in channel activation by diamide, therefore, appear to be different than those involved in activation by NEM. Continued incubation with NEM leads to channel inhibition. Additional activation by NEM is also seen with channels pretreated with higher concentrations (500 M to 1 mM) of diamide (data not shown).
To examine further the relationship between sulfhydryls altered by diamide and those that react with NEM, we examined the effect of pretreatment of membranes with diamide on the ability of NEM to alter [ 3 H]ryanodine binding (Fig. 7A). In these experiments the membranes were reacted first with 250 M diamide for 30 min and then with 5 mM NEM. At various times after NEM addition, aliquots were removed and the reaction stopped by the addition of DTT. The samples were then assayed for [ 3 H]ryanodine binding. As can be seen in Fig.   FIG. 4

. Formation of dimers as detected by SDS-PAGE (A) and two-dimensional gels showing dimer reduces to monomer (B). A,
Coomassie-stained 5% SDS-PAGE. Electrophoresis performed under non-reducing condition. Lanes 1-6, increasing concentrations of diamide: 0, 0.1, 0.2, 0.5, 0.7, and 1.0 mM, respectively. Band A represents the dimer (2 ϫ 565 kDa) RYR, and band a is the monomer (565 kDa) of RYR. B, second dimension Coomassie-stained 5% SDS-PAGE of sample cross-linked with 500 M diamide. Two-dimensional electrophoresis was performed under reducing condition as described under "Experimental Procedures." The top gel, which is shown horizontally, is the first dimension when electrophoresis was performed under nonreducing conditions. Band A is the dimer (2 ϫ 565 kDa) of full-length RYR, and band B is the dimer (2 ϫ 400 kDa) of calpain-digested RYR. Bands a and b are the monomers of full-length (565 kDa) and proteolyzed (400 kDa) RYR. 7A, diamide pretreatment prevented the phase 1 inhibition of [ 3 H]ryanodine binding and greatly increased phase 2 enhancement. DTT treatment reduced the cross-links formed by diamide (Fig. 8A). Therefore, the diamide-induced cross-links did not contribute to the effects on [ 3 H]ryanodine binding; instead, the oxidation appeared to be protecting sulfhydryls during NEM alkylation.
As shown in Fig. 2, lowering the NEM concentration greatly slows the onset of the phase 2 and 3 effects, allowing us to look at a reaction that is primarily phase 1. To examine the effect of diamide pretreatment on the isolated phase 1 reaction, we pretreated membranes with 250 M diamide and then examined the effect of 500 M NEM on [ 3 H]ryanodine binding (Fig.  7B). No phase 1 inhibition was seen with the diamide-pretreated membranes. These findings suggest that diamide is interacting with and protecting the phase 1 sulfhydryls. Alternatively diamide may be altering the conformation of the Ca 2ϩ release channel, such that the phase 1 sulfhydryls no longer react as rapidly. To demonstrate this protection we pretreated membranes with 250 M diamide, diluted the diamide to less than 25 M, reacted with 5 mM NEM for 30 min on ice, and then reduced with 10 mM DTT, washed, and tested the effect of a second addition of 500 M NEM on the binding of [ 3 H]ryanodine. The reaction was again stopped with 10 mM DTT prior to the binding assay. These data are shown in Fig. 7C. As can be seen in this figure, reduction after diamide and NEM treatment restored the ability of NEM to inhibit [ 3 H]ryanodine binding in a phase 1-like reaction. The initial binding and plateau binding to the membranes treated with diamide and then NEM was increased, as would be expected from membranes that have been alkylated at the phase 2 sites (see Fig. 7A).
To obtain additional evidence that diamide has not altered the ability of NEM to enhance binding in a phase 2 reaction, membranes were pretreated with diamide and with low concentrations of NEM under conditions that in the absence of diamide would have produced phase 1 inhibition. NEM was then added under conditions determined to produce phase 2 enhancement (Fig. 7D). Phase 2 enhancement is not blocked by diamide pretreatment.
To demonstrate that the phase 1 reaction with NEM protects the sulfhydryls involved in the diamide cross-link, we assessed the ability of NEM in the different phases to block dimer formation. The steps in this experiment were: 1) incubation of membranes with 5 mM NEM for different periods of time, 2) reduction with DTT, 3) removal of DTT, and 4) cross-linking with 250 M diamide. NEM at the earliest incubation times blocked dimer formation (Fig. 8B). If dimer formation is truly protecting sulfhydryls from the phase 1 reaction, it should be possible to reduce diamide-formed disulfides after the phase 2 NEM reaction and then reform dimers with diamide. The experiment involved the following steps: 1) incubation of membranes with 250 M diamide, 2) treatment with NEM for different periods of time, 3) reduction with DTT, 4) removal of DTT, and 5) cross-linking with diamide. Dimers could reform if, prior to NEM and subsequent DTT treatment, the sulfhydryls were protected by diamide cross-linking (Fig. 8C).
The phase 1 effects on [ 3 H]ryanodine binding correlated well with the rate of incorporation of [ 14 C]NEM into the 565-kDa band of the RYR1. This is demonstrated in Fig. 9, where membranes with and without diamide pretreatment were incubated with 1 mM [ 14 C]NEM for increasing periods of time. The 565-kDa band from a Coomassie Brilliant Blue-stained gel was excised, digested as described under "Experimental Procedures," and the radioactivity in the bands was quantitated by liquid scintillation counting. The data were fit as the sum of two exponentials, with the rapid labeling component having a k obs of 0.26 min Ϫ1 and the slower component having a k obs of 0.067 min Ϫ1 . From the fits it was determined that diamide pretreatment reduced the fast phase by 81% and the slow phase 13%. In these experiments the incorporation of

DISCUSSION
Redox modulation of excitation-contraction coupling is important physiologically. In the native state, intact skeletal muscle fibers produce detectable levels of both reactive oxygen species (12) and NO derivatives (13,14). Oxidant depletion alters contractile function. In unfatigued muscle, selective scavenging of reactive oxygen species depresses force generation (12); inhibition of NO synthesis has the opposite effect, increasing force (13). Strenuous contractile activity accelerates the rate at which myocytes produce free radicals, and other reactive oxidants, e.g. superoxide anion radicals (16), hydroxyl radicals (17), and NO derivatives (14). Oxidants accumulate in the active muscle and contribute directly to the loss of contractile function that occurs in muscular fatigue (15). Oxidative effects on both unfatigued and fatigued muscle are consistent with an increase in cytoplasmic Ca 2ϩ concentrations due to activation of the sarcoplasmic reticulum Ca 2ϩ release channel (21). An important role of sulfhydryl groups in the modulation of the activity of the skeletal muscle Ca 2ϩ release channel has been demonstrated by several laboratories (3-9, 27, 28). Both oxidizing compounds (3)(4)(5)(6)(7)(8)(9) and reagents that modify free sulfhydryls (27,28) alter the activity of Ca 2ϩ release channel.
To explore these phenomena we choose NEM, a reagent that alkylates cysteine residues, and diamide, a sulfhydryl oxidizing agent that cross-links near neighbor cysteine residues. NEM shows a three-step effect on the activity of the Ca 2ϩ release channel; first it inhibits, then it activates, and finally it again inhibits the channel reconstituted into planar lipid bilayers. In general, agents that activate the Ca 2ϩ release channel increase the apparent affinity of the protein for [ 3 H]ryanodine, while those that inhibit channel activity decrease the apparent affinity (36). Consistent with this, NEM produces a similar threephase effect on [ 3 H]ryanodine binding: phase 1 inhibition, phase 2 enhancement, and phase 3 inhibition. As shown by the effects on [ 3 H]ryanodine binding and channel activity, and by direct [ 3 H]NEM labeling, sulfhydryls on the Ca 2ϩ release channel are reacting sequentially with NEM to alter the structure and function of the protein.
Diamide activates the channel reconstituted into planar lipid bilayers and enhances [ 3 H]ryanodine binding by increasing apparent affinity. The observation that either an oxidizing or an alkylating reagent can activate the channel raised the question of whether the activation was due to the loss of a free sulfhydryl. If this were true, we would expect the activation by the two reagents to involve the same cysteine residues. This is not the situation for diamide and NEM; instead, to activate the channel, NEM and diamide appear to react with different sulfhydryls. This conclusion is based on a number of observations: 1) NEM can further stimulate the activity of the diamideactivated single channel, 2) diamide blocks the phase 1 inhibitory effects of NEM on [ 3 H]ryanodine binding, 3) phase 1 alkylation by NEM blocks dimer formation by diamide, and 4) diamide pretreatment enhances the activating effects of NEM on [ 3 H]ryanodine binding. When phase 1 sulfhydryls are protected by diamide cross-linking, NEM alkylation markedly enhances [ 3 H]ryanodine binding. Our data are most consistent with a model in which the activation by diamide is due to cross-linking of phase 1 sulfhydryls while the NEM activation is associated with alkylation of phase 2 sulfhydryls. Without actually identifying the cysteine residues involved in phase 1 alkylation and those involved in diamide cross-linking, it is impossible to totally eliminate the possibility that diamide is having a long distance effect on phase 1 sulfhydryls. However, extensive reaction with NEM after diamide cross-linking fails FIG. 7. Phase 1 inhibition by NEM is not seen with membranes pretreated with diamide. SR membranes (0.56 mg) were treated with 250 M diamide for 10 min at room temperature. These membranes were then treated with NEM on ice for the indicated times. At these times 20-g aliquots were removed and the reaction stopped with 10 mM DTT. A, effect of diamide pretreatment on the subsequent ability of 5 mM NEM to alter [ 3 H]ryanodine binding. f, diamide-pretreated membranes; q, control membranes. B, effect of diamide pretreatment on the subsequent ability of 500 M NEM to alter [ 3 H]ryanodine binding. f, diamide-pretreated membranes; q, control membranes not treated with diamide. C, prior to data collection SR membranes were pretreated in five steps: 1) incubated with 250 M diamide for 30 min on ice, 2) alkylated with 500 M NEM for 20 min, 3) reduced with 5 mM DTT for 30 min at room temperature, 4) washed twice by pelleting in a Beckman Airfuge, and 5) resuspended in binding buffer. Then at time 0, membranes were re-exposed to 500 M NEM for the indicated periods of time. The reaction was stopped with 5 mM DTT, and the membranes were assayed for [ 3 H]ryanodine binding as described previously. f, membranes pretreated with diamide, then NEM; q, control membranes with no pretreatment; E, membranes prealkylated with NEM only. D, effect of 400 M NEM addition followed by 4 mM NEM addition on [ 3 H]ryanodine binding to diamide-pretreated membranes. E, diamide-pretreated membranes; q, control membranes not treated with diamide.
to prevent the recovery of the phase 1 effects of NEM after reduction of the cross-linked sulfhydryls, a finding that strongly supports our model of phase 1 sulfhydryls being at subunit-subunit contact domains. The effects of oxidation, reduction, and alkylation on the binding of [ 3 H]ryanodine and on channel activity are summarized in Table I. Intersubunit crosslinks are likely to involve different cysteines on the two adjacent subunits. We have not yet determined whether both of these sulfhydryls are alkylated by NEM in the phase 1 reaction. The second sulfhydryl may be alkylated at a later stage, or it may not react with NEM. A completely different sulfhydryl could be alkylated in phase 2. The sequentially reacting cysteine residues could be at different locations in the primary sequence of each subunit or could be the same residues on different subunits, reacting at different rates as a result of conformational changes in the protein. These issues remain to be resolved. Diamide-induced cross-links are detected in intact membranes, in detergent-solubilized membranes, and in purified ryanodine-binding proteins, 3 suggesting that the cross-links are occurring between subunits of a tetramer. Extended incubations with diamide lead to the formation of higher molecular weight oligomers, which could be trimers or tetramers. Formation of higher oligomers suggests either that the cross-link involves different sulfhydryls on adjacent subunits or that there is more than one type of cross-link formed. The partners in the formation of the cross-linked complexes were identified by two-dimensional electrophoresis. The membrane preparation used for these studies contained a significant amount of a 400-kDa band, which has been identified previously as a proteolytic fragment of the Ca 2ϩ release channel (38,39). Experiments have been performed with membrane preparations that have very little proteolysis and membranes that were proteolyzed by endogenous calpain. The effects of NEM and diamide on the binding of [ 3 H]ryanodine was unaffected by this proteolytic event. Both the 565-kDa full-length ryanodine-binding protein subunit and the 400-kDa fragment appear to be crosslinked by diamide to form dimers, suggesting that the site of the intersubunit cross-link is within the 400-kDa fragment. Very little cross-linking, however, occurs between the 400-kDa and the 565-kDa bands. This surprising finding suggests that the proteolytic events that produce the 400-kDa fragment are The diamide was removed by pelleting the membranes in a Beckman Airfuge (4 min at 30 p.s.i.) and resuspending the membranes in binding buffer (6.2 mg/ml). The membranes were then incubated with 1 mM [ 14 C]NEM for the indicated periods of time. The reaction was stopped with the addition of 20 mM DTT. Aliquots (22 g) were then electrophoresed on 5% SDS gels, and, after Coomassie Brilliant Blue staining, the 565-kDa RYR1 band was excised, digested, and counted as described under "Experimental Procedures." All samples were electrophoresed three times. q, control; f, diamide-pretreated. Also shown in this figure (E) is the effect of a similar treatment of the same membranes at 6 mg/ml with unlabeled 1 mM NEM on the binding of [ 3 H]ryanodine. This was performed as described in Fig. 2A. nonrandom. The most frequent neighbor of a 400-kDa proteolyzed subunit within a tetramer is another 400-kDa subunit, while the most frequent neighbor of the full-length 565-kDa protein is another 565-kDa protein. The proteolytic event is presumably due to the action of endogenous calpains (38,39). One interpretation of these findings is that some of the Ca 2ϩ release channel tetramers are by some means targeted for proteolysis such that all of the subunits in the protein are proteolyzed simultaneously. Other tetramers contain all intact subunits. This would suggest that some modification of the Ca 2ϩ release channel allows it to be recognized by endogenous calpains. A second possibility is that once calpain binds to the tetramer it remains bound until all of the subunits are cleaved. We are currently investigating these intriguing possibilities.
Cross-linking of sulfhydryls by diamide activates the channel, indicating that sulfhydryls within this class may be the targets of redox modulation of the Ca 2ϩ release channel and that these redox-sensitive sulfhydryls are located in domains where subunits contact one another. There is no indication of the involvement of any other protein in the disulfide bond formation induced by diamide. This is in contrast to the results of Liu et al. (22,23), who suggest that triadin is cross-linked to the Ca 2ϩ release channel by oxidizing agents. We were unable to find evidence of such a cross-link.
In summary, we demonstrate the existence of at least three classes of functionally important sulfhydryls on the Ca 2ϩ release channel, modification of which alter channel activity and [ 3 H]ryanodine binding (Table I). Both NEM and diamide can activate the Ca 2ϩ release channel, but the reactions involve different sulfhydryls. Diamide cross-links subunits within the tetramer. This cross-linking appears to correlate with channel activation and with oxidation of a class of sulfhydryls, which, in the absence of diamide, react rapidly with NEM to inhibit the channel. Alkylation of these phase 1 sulfhydryls prevents dimer formation. The phase 1 sulfhydryls, therefore, may be located at contact domains between subunits. These studies will be useful in designing strategies to differentially label phase 1 and phase 2 sulfhydryls to map their location in the primary sequence of the Ca 2ϩ release channel. This may also enable us to define some of the parts of the protein that are in regions of subunit-subunit contact in the three-dimensional structure of the protein.

TABLE I
Model for diamide and NEM effects if phase 1 sulfhydryls are also the sulfhydryls which are cross-linked by diamide S, a cysteine residue in the Ca 2ϩ release channel (the subscripts a, b, and c are used to indicate different residues); A, alkylated, S-S ϭ disulfide bond; SH, free sulfhydryl; S b and S c are on different subunits; S?, currently unknown whether these sulfhydryls are alkylated. ND, not determined.