Oxygen-coupled Redox Regulation of the Skeletal Muscle Ryanodine Receptor/Ca2+ Release Channel (RyR1)

Background: Within the skeletal muscle Ca2+ release channel RyR1, S-oxidation and S-nitrosylation of allosterically linked Cys residues are coupled to oxygen tension (pO2). Results: Mass spectrometry identifies multiple Cys residues as oxidized at high versus low pO2. Conclusion: Endogenous H2O2 catalyzes pO2-coupled disulfide formation within an allosteric Cys network that gates S-nitrosylation. Significance: Dynamic disulfide formation subserves physiological redox regulation of RyR1. In mammalian skeletal muscle, Ca2+ release from the sarcoplasmic reticulum (SR) through the ryanodine receptor/Ca2+-release channel RyR1 can be enhanced by S-oxidation or S-nitrosylation of separate Cys residues, which are allosterically linked. S-Oxidation of RyR1 is coupled to muscle oxygen tension (pO2) through O2-dependent production of hydrogen peroxide by SR-resident NADPH oxidase 4. In isolated SR (SR vesicles), an average of six to eight Cys thiols/RyR1 monomer are reversibly oxidized at high (21% O2) versus low pO2 (1% O2), but their identity among the 100 Cys residues/RyR1 monomer is unknown. Here we use isotope-coded affinity tag labeling and mass spectrometry (yielding 93% coverage of RyR1 Cys residues) to identify 13 Cys residues subject to pO2-coupled S-oxidation in SR vesicles. Eight additional Cys residues are oxidized at high versus low pO2 only when NADPH levels are supplemented to enhance NADPH oxidase 4 activity. pO2-sensitive Cys residues were largely non-overlapping with those identified previously as hyperreactive by administration of exogenous reagents (three of 21) or as S-nitrosylated. Cys residues subject to pO2-coupled oxidation are distributed widely within the cytoplasmic domain of RyR1 in multiple functional domains implicated in RyR1 activity-regulating interactions with the L-type Ca2+ channel (dihydropyridine receptor) and FK506-binding protein 12 as well as in “hot spot” regions containing sites of mutation implicated in malignant hyperthermia and central core disease. pO2-coupled disulfide formation was identified, whereas neither S-glutathionylated nor sulfenamide-modified Cys residues were observed. Thus, physiological redox regulation of RyR1 by endogenously generated hydrogen peroxide is exerted through dynamic disulfide formation involving multiple Cys residues.

In mammalian skeletal muscle, Ca 2؉ release from the sarcoplasmic reticulum (SR) through the ryanodine receptor/Ca 2؉release channel RyR1 can be enhanced by S-oxidation or S-nitrosylation of separate Cys residues, which are allosterically linked. S-Oxidation of RyR1 is coupled to muscle oxygen tension (pO 2 ) through O 2 -dependent production of hydrogen peroxide by SR-resident NADPH oxidase 4. In isolated SR (SR vesicles), an average of six to eight Cys thiols/RyR1 monomer are reversibly oxidized at high (21% O 2 ) versus low pO 2 (1% O 2 ), but their identity among the 100 Cys residues/RyR1 monomer is unknown. Here we use isotope-coded affinity tag labeling and mass spectrometry (yielding 93% coverage of RyR1 Cys residues) to identify 13 Cys residues subject to pO 2 -coupled S-oxidation in SR vesicles. Eight additional Cys residues are oxidized at high versus low pO 2 only when NADPH levels are supplemented to enhance NADPH oxidase 4 activity. pO 2 -sensitive Cys residues were largely non-overlapping with those identified previously as hyperreactive by administration of exogenous reagents (three of 21) or as S-nitrosylated. Cys residues subject to pO 2 -coupled oxidation are distributed widely within the cytoplasmic domain of RyR1 in multiple functional domains implicated in RyR1 activity-regulating interactions with the L-type Ca 2؉ channel (dihydropyridine receptor) and FK506-binding protein 12 as well as in "hot spot" regions containing sites of mutation implicated in malignant hyperthermia and central core disease. pO 2coupled disulfide formation was identified, whereas neither S-glutathionylated nor sulfenamide-modified Cys residues were observed. Thus, physiological redox regulation of RyR1 by endogenously generated hydrogen peroxide is exerted through dynamic disulfide formation involving multiple Cys residues.
Redox-based post-translational regulation of protein function is exerted principally through modification of Cys thiol side chains. S-Nitrosylation provides the best characterized example with a large number of endogenous substrates identified in the context of cellular signal transduction along myriad pathways (1,2). Accumulating evidence also points to a potentially broad regulatory influence of S-oxidation, catalyzed by reactive oxygen species (effectively, hydrogen peroxide (H 2 O 2 )), which could in principle be exerted through a number of independent or coupled oxidative modifications: the reversible and dynamic formation of disulfide bonds, sulfenic (and possibly sulfinic) acid, and sulfenamide as well as S-glutathionylation and S-sulfhydration (2)(3)(4). A broad role for physiological S-oxidation would be consistent with the ubiquitous expression across multicellular organismal phylogeny and cell types of the NADPH oxidases (Noxs), 3 which function to generate reactive oxygen species (5). However, although there have been significant recent advances (6,7), there are relatively few examples in which mediation or modulation of physiological signal transduction by S-oxidation has been characterized fully, that is with respect to the (enzymatic) source of endogenous H 2 O 2 and the generative stimulus, the protein targets subject to S-oxidation and the specific Cys residues modified within those targets, the chemical nature of the modification, the effect of modification on protein function, and the consequences of altered protein function in the cellular milieu.
The ryanodine receptor/Ca 2ϩ release channel (RyR), which serves as the essential source of Ca 2ϩ release from the sarcoplasmic reticulum (SR) that mediates excitation-contraction coupling in skeletal and cardiac striated muscle, has emerged as a paradigmatic example of redox regulation of protein function through Cys-directed post-translational modification (8 -16). RyRs, which function as tetramers, are the largest channel pro-* This work was supported, in whole or in part, by National Institutes of Health teins described. Monomers of RyR1, the predominant form of RyR in skeletal muscle, comprise about 5000 amino acids (ϳ565 kDa). One hundred of those residues are Cys, and half of those Cys residues on average are in the reduced form (sulfhydryl bearing a free thiol) under basal conditions (8,16). Studies using application in vitro of exogenous redox agents (reduced glutathione, oxidized glutathione, S-nitrosoglutathione, nitric oxide (NO), and NO donors (NOC and NOR)) have reported modification of multiple Cys residues within RyR1 by S-glutathionylation, S-nitrosylation, and unspecified S-oxidation (8 -11, 13, 14, 17, 18). It has also been reported that tetanic stimulation of skeletal muscle elicits Nox-dependent production of H 2 O 2 (19) and that RyR1 can be modified by S-glutathionylation via Nox2 localized to transverse tubule membranes (20). Increasing evidence thus supports the possibility that Nox-derived reactive oxygen species may play a physiological role in skeletal muscle excitation-contraction coupling.
Physiological S-oxidation of RyR1 is coupled to endogenous S-nitrosylation of the channel, providing a paradigmatic example of allosteric cross-talk by redox-based modifications. More specifically, under physiological conditions in situ, it has been shown that RyR1 is modified by S-nitrosylation that reflects the production of NO by neuronal nitric-oxide synthase (8,10,12) and by S-oxidation that is dependent upon the production of H 2 O 2 by Nox4 (16). Physiological S-oxidation is governed by a redox cycle intrinsic to the SR that is coupled to muscle pO 2 : in isolated SR (SR vesicles) at relatively high pO 2 (21% O 2 ), an average of about six to eight free Cys thiols are oxidized consequent upon H 2 O 2 production by Nox4, and the transition to relatively low pO 2 (1% O 2 ) is accompanied by reduction that reflects suppressed H 2 O 2 production and the operation of an as yet unspecified reductive mechanism (8,16). Oxygen-coupled S-oxidation activates RyR1, enhances Ca 2ϩ release, and gates S-nitrosylation, which is blocked at high versus low pO 2 . S-Nitrosylation at low pO 2 , which also activates RyR1, targets a single Cys residue (Cys 3635 ) that is not a member of the set of Cys residues oxidized at high versus low pO 2 (10,13). Thus, S-oxidation allosterically regulates RyR1 activity, and S-oxidation and S-nitrosylation operate together to modulate Ca 2ϩ release through RyR1 over a range of pO 2 that extends from the low pO 2 that characterizes working muscle to oxidative stress.
The analysis in skeletal muscle SR provides an example of physiological S-oxidation in which the source of H 2 O 2 (Nox4) and the generative stimulus (pO 2 ), the specific protein target (RyR1), and the effects of modification on protein function (activation of RyR1 and enhanced Ca 2ϩ release from the SR) have been described. However, although it has been shown that only a relatively small subset of Cys residues on average are subject to pO 2 -coupled S-oxidation, neither the identity of the targeted Cys residues nor the nature of the oxidative modification(s) are known. We developed a mass spectrometry-based scheme for analysis of regulated protein S-oxidation that we applied to identify the sites of physiological S-oxidation within RyR1 and the form of modification. Here we identify 13 Cys residues subject to oxidation at high versus low pO 2 in isolated SR and eight additional Cys residues subject to oxidation at high versus low pO 2 but only when NADPH levels are supplemented to enhance Nox4 activity. pO 2 -coupled S-oxidation targets Cys residues that are distributed widely within the cytoplasmic domain of RyR1 in multiple functional domains. Furthermore, we show that disulfide formation is the likely form of S-oxidation.

EXPERIMENTAL PROCEDURES
Preparation of SR Vesicles and Purification of RyR1-All preparations here and below utilized hind limb muscle from rabbit. SR vesicles were prepared essentially as described (21). Briefly, muscle was homogenized in buffer containing 20 mM Hepes, pH 7.4, 2 mM EDTA, 0.2 mM EGTA, 0.3 M sucrose, and protease inhibitors (100 nM aprotinin, 20 M leupeptin, 1 M pepstatin, 0.2 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine). Homogenates were centrifuged at 9,200 ϫ g for 20 min, and the resultant supernatant was centrifuged at 100,000 ϫ g for 1 h. The resultant pellet (membrane fraction) was resuspended and fractionated on a continuous 20 -45% (w/v) sucrose gradient by centrifugation at 100,000 ϫ g for 14 h. Fractions containing SR vesicles were eluted, collected by centrifugation at 120,000 ϫ g, resuspended, aliquoted, and stored in liquid nitrogen. RyR1 was purified from SR vesicles solubilized with CHAPS by sucrose density gradient centrifugation as described (22). Protein concentrations were determined with a bicinchoninic acid-based assay.
Quantification of Total RyR1 Free Thiols (Sulfhydryls)-The free thiol content of RyR1 was quantified by monobromobimane (MBB) fluorescence. As described (8,16), SR vesicles held at 21% or 1% O 2 (glove box) were exposed to MBB in the presence of 10 M Ca 2ϩ prior to isolation of RyR1 as above and quantification of fluorescence. When used, NADPH (1 mM) was added 30 min before MBB.
Isotope-coded Affinity Tag (ICAT) Labeling-SR vesicles were incubated in 50 mM phosphate buffer containing 10 M CaCl 2 , pH 7.5 at 1 or 21% O 2 for 30 min. When used, 10 units/ml polyethylene glycol (PEG)-coupled catalase, 20 M diphenyleneiodonium (DPI), or 1 mM NADPH was added at the beginning of the incubation. Samples were then incubated with 10ϫ excess (over calculated Cys content) ICAT reagent for 4 h in the presence of 2% SDS. Light ICAT ( 12 C) and heavy ICAT ( 13 C) were used to label samples incubated at low pO 2 (1% O 2 ) or high pO 2 (21% O 2 ), respectively. Samples (1 and 21% O 2 ) were then mixed 1:1, and excess ICAT was removed by size exclusion filtration through a P6 gel column. Proteins were separated by reducing SDS-PAGE, and the band containing RyR1 was excised and incubated with DTT followed by iodoacetamide (IA) to alkylate previously oxidized Cys residues (23). In-gel digestion by trypsin or chymotrypsin was carried out overnight in 100 mM ammonium bicarbonate at 37°C. ICATlabeled peptides were purified on avidin affinity columns, and the biotin moiety was cleaved (AB Sciex). Both ICAT-labeled peptides (avidin-bound) and unlabeled peptides (flowthrough) were collected, desalted with a C 18 UltraMicro Tip Column, and resuspended in 0.1% (v/v) formic acid for analysis by liquid chromatography-coupled tandem mass spectrometry (LC-MS/MS). The ICAT-based analytic scheme is illustrated in Fig. 1A.
Dimedone Labeling-SR vesicles were incubated with 10 mM dimedone in 50 mM phosphate buffer containing 10 M CaCl 2 , pH 7.5 at room temperature for 1 h at 21% O 2 . Proteins were separated by reducing SDS-PAGE, and the band containing RyR1 was excised and incubated with DTT followed by IA. Following in-gel digestion by trypsin or chymotrypsin, eluted peptides were desalted and analyzed by LC-MS/MS. Alternatively, muscle homogenates were incubated with 10 mM dimedone at 21% O 2 for 1 h prior to isolation of SR vesicles and analysis as above.
Preparation of RyR1 for Native Identification of Disulfide and Other Oxidative Modifications-SR vesicles were incubated with 50 mM phosphate buffer containing 10 M CaCl 2 , pH 7.5 at 1 or 21% O 2 for 20 min. Samples were then held at 1 or 21% O 2 and incubated with a 50-fold excess of IA or N-ethylmaleimide for 20 min at room temperature in the dark followed by the addition of 2% (w/v) SDS and incubation for an additional hour. Proteins were separated by non-reducing SDS-PAGE followed by excision of the band containing RyR1, in-gel digestion with trypsin or chymotrypsin, and analysis by LC-MS/MS.
Mass Spectrometric Analysis-LC-MS/MS was performed using an LTQ Orbitrap XL linear ion trap mass spectrometer (Thermo Fisher Scientific). Reverse-phase high performance liquid chromatography (HPLC) was conducted with an Ultimate 3000 HPLC system (Dionex) equipped with a Dionex C 18 Acclaim PepMap 100 column (0.075 ϫ 150 mm) in a linear gradient of acetonitrile from 2 to 60% (v/v) for a period of 60 min at a flow rate of 0.3 l/min. Spectra were acquired in the positive ionization mode by data-dependent methods consisting of a full MS scan in high mass accuracy Fourier transform MS mode at 60,000 resolution and MS/MS on the six most abundant precursor ions in collision-induced dissociation mode with a normalized collision energy of 35%. A dynamic exclusion function was applied with repeat count of 2, repeat duration of 45 s, exclusion duration of 15 s, and exclusion size list of 350. Peptides were identified with Mascot Daemon (version 2.3.0, Matrix Science), and the data were searched against the RyR1 primary sequence. The mass tolerance was set at 10 ppm for precursor ions and 0.8 Da for product ions. Carbamidomethylation or N-ethylmaleimide labeling of Cys residues, dimedone labeling of Cys residues and sulfenic acids (when used), and oxidation of Met residues were set as variable modifications. Other possible modifications of Cys, including sulfenation, sulfination, sulfonation, glutathionylation, and sulfenamide formation, were also tested. The identification of disulfide bonds was performed using the on-line version of MassMatrix software (24). The modification(s), including disulfide bond formation, suggested by the search engine were verified by manual examination of each tandem mass spectrum and compared with corresponding unmodified peptides. For ICAT labeling analysis, both light and heavy ICAT-labeled peptides were first identified using Mascot Daemon and then confirmed by manual examination, especially in the case of peptides containing multiple Cys residues and mixed labeling with IA and ICAT.
MS Data Analysis-Quantification of ICAT labeling was conducted using Mascot Distiller (Matrix Science). All reported peptides labeled with ICAT were identified with standard error Ͻ0.1 and correlation coefficient Ͼ0.95. The two peaks from heavy and light ICAT labeling were identified as separated by 9 Da. The ratio of light ICAT (labeled at 1% oxygen) to heavy ICAT (labeled at 21% oxygen) was calculated and averaged on the basis of peptide peak heights for each individual scan. The ratios for the same peptide but with different charge and methionine oxidation state were then averaged manually. For peptides with multiple Cys residues and mixed labeling by IA and ICAT, the ratio was manually calculated from the peptide peak heights extracted from the selected ion chromatogram. Standard errors were derived from replicated biological samples, different charge and methionine oxidation states, or multiple scanning. Quantification of disulfide was conducted manually. Two native reference peptides, A (LSLPVQFHQHFR) and B (GDGGEGEGEVQ-FLR), that appeared unmodified in each of several repeated experiments were chosen (25). Peaks from the same disulfide-containing peptide with different methionine oxidation states were added together. After normalization with respect to the two reference peptides at the different pO 2 values, relative disulfide abundance at high versus low pO 2 was calculated as an average of three biological replicates.

Identification of Cys Residues within RyR1 Subject to pO 2 -dependent Oxidative Modification-RyR1
is an unusually large protein (ϳ5,000 amino acids) that contains 100 Cys residues. We developed an approach that allowed us to assess by mass spectrometric analysis the redox state of 93 of 100 Cys residues. In this approach (Fig. 1A), an RyR1-enriched subcellular fraction (SR vesicles) is solubilized in SDS under non-reducing conditions, and free thiols are labeled with ICAT reagent. The efficacy of ICAT labeling under the conditions used is demonstrated by the finding that incubation with ICAT reagent at either low or high pO 2 eliminates subsequent labeling by the thiol-specific agent MBB (Fig. 1B). RyR1 is then enriched by SDS-PAGE under reducing conditions followed by in-gel alkylation with IA of thiols freed by reduction and in-gel digestion with either trypsin or chymotrypsin. Digests are subjected to avidin affinity purification, and both bound and unbound peptides are analyzed by LC-MS/MS. The dual enrichment provided by selective ICAT labeling of Cys-containing peptides and gel purification resulted in coverage of 93% of RyR1 Cys residues when the results of trypsin and chymotrypsin digestion were combined.
After labeling SR vesicles held at low versus high pO 2 (1% O 2 versus 21% O 2 ) with light ( 12 C) versus heavy ( 13 C) ICAT reagent, respectively, followed by reduction and labeling with IA, RyR1 Cys residues fell into one of four principal groups. Seven Cys residues were not identified in our analysis (Cys 146 , Cys 1489 , Cys 1491 , Cys 2651 , Cys 4657 , Cys 4876 , and Cys 4882 ) (including additional experiments using digestion with Asp-N), although the predicted masses of the peptides containing those Cys residues were within the detectable limit (ϳ800 -3,000 Da). Fifteen Cys residues were labeled with IA only and were therefore identified as fully oxidized (although reducible by DTT) at both low and high pO 2 (Fig. 2), consistent with their participation in "structural" disulfides. Remarkably, 78 of 93 identified Cys residues were detected with ICAT labeling and were thus present in some proportion in reduced form. Twenty-two of those were labeled with ICAT reagent but not IA (Fig. 2) and displayed a light:heavy ratio of ϳ1:1, consistent with their presence in fully reduced form at both low and high pO 2 (Table 1). An additional two Cys residues detected with ICAT only (Cys 566 and Cys 762 ) were pO 2 -sensitive (increased light:heavy ratio) ( Table 2), although no IA-labeled form was detected presumably because the IA-labeled peptide was of low abundance (consistent with the existence of those Cys residues in predominantly reduced form at low pO 2 ).
Each of the remaining 54 Cys residues were labeled with IA or ICAT reagent and identified as members of a population that contained both reduced and oxidized Cys residues at both low and high pO 2 . The observed light ICAT labeling versus heavy ICAT labeling (L:H) ratios ranged from 0.75 to 1.47 with the large majority distributed from 0.94 to 1.21 (Fig. 3A). Greater relative abundance of light ICAT labeling versus heavy ICAT labeling identified Cys residues subject to oxidation at high versus low pO 2 . Based upon an arbitrary threshold L:H of 1.24 (which in terms of tetrameric RyR1 would represent at the extremes a population of ϳ25% of RyR where all subunits are modified or 100% of RyR where a single monomer is modified), the redox states of 13 Cys residues (L:H of 1.24 -1.47) were coupled to pO 2 ( Fig. 4 and Table 2). Importantly, as illustrated in Table  2, in every case examined, the increase in L:H ratio at high versus low pO 2 was eliminated when SR vesicles were exposed to 20% O 2 in the presence of either PEG-coupled catalase to remove H 2 O 2 or the Nox inhibitor DPI, confirming the essential role of endogenous H 2 O 2 production by SR-resident Nox4 (16).
It is of note that the two Cys residues exhibiting anomalous behavior, Cys 3193 and Cys 3635 , include the identified site of endogenous S-nitrosylation (Cys 3635 ) ( Fig. 3A and Table 1). Cys 3635 is pO 2 -insensitive (L:H ϳ 1 under control conditions), consistent with the previous finding that Cys 3635 is not subject to pO 2 -coupled S-oxidation (13). However, the L:H ratio drops to ϳ0.78 when H 2 O 2 is removed at high pO 2 (Table 1). One possible explanation for this finding is that, inasmuch as some proportion of Cys 3635 is present in isolated SR vesicles in the S-nitrosylated form (not shown), enhanced denitrosylation (26) at high pO 2 in the absence of H 2 O 2 (as in the hypothetical case of a denitrosylase that is inhibited by reactive oxygen species) would increase ICAT labeling at high versus low pO 2 and thereby decrease the L:H ratio.
We reported previously that both generation of H 2 O 2 and RyR1 activity in SR vesicles were enhanced at 21 versus 1% oxygen and that these effects were associated with pO 2 -coupled oxidation of a set of thiols within RyR1 (16). In addition, we reported that supplementation with NADPH (to enhance Nox4 activity) enhanced H 2 O 2 production (16). Here we confirmed directly that the loss of free thiols within RyR1 in SR vesicles exposed to high versus low pO 2 was enhanced by addition of NADPH (Table 3). When SR vesicles were incubated with NADPH (1 mM), the L:H ratios were also enhanced for 6 of 11 tested Cys residues of the 13 Cys residues that were oxidized at high versus low pO 2 ( Fig. 4 and Table 2). Furthermore, addition of NADPH resulted in an increase in L:H ratio from ϳ1:1 to Ͼ1.24:1 in a population of eight additional Cys residues ( Fig.  3B; Table 4). Thus, 21 Cys residues within RyR1 are subject to pO 2 -coupled redox regulation.
Identification of the Form of pO 2 -coupled S-Oxidation-To investigate the form of pO 2 -coupled S-oxidation, we incubated SR at low or high pO 2 , blocked free thiols with IA or N-ethylmaleimide, and then purified RyR1 by non-reducing SDS-PAGE prior to trypsin or chymotrypsin digest and analysis by LC-MS/MS. We did not detect S-glutathionylation at either low or high pO 2 . Western blot analysis of intact RyR1 using anti-glutathione antibodies also failed to detect S-glutathionylation, although S-glutathionylation of RyR1 was detected by Western blot when SR vesicles were incubated with oxidized glutathione as a positive control (data not shown). We also did Cys residues within RyR1, SR vesicles maintained at 21% O 2 or 1% O 2 are solubilized in SDS, and free thiols are labeled with heavy or light ICAT, respectively. Samples are then combined, and RyR1 is isolated by SDS-PAGE under reducing conditions. The band containing RyR1 is excised, and free thiols are alkylated with iodoacetamide followed by in-gel digestion with either trypsin or chymotrypsin. Eluted peptides are passed over an ICAT affinity column, and both bound peptides and the flow-through are analyzed by LC-MS/MS. Thus, for a given peptide, an increase in the light:heavy ratio at 1% O 2 versus 21% O 2 indicates Cys residues that are oxidized at high versus low pO 2 . B, ICAT fully labels RyR1 free Cys thiols. SR vesicles maintained at 1% O 2 (low pO 2 ) or 21% O 2 (high pO 2 ) were solubilized in SDS, and free thiols were labeled with the fluorescent reporter MBB. Prior incubation with ICAT reagent eliminates MBB fluorescence, demonstrating that free thiols are quantitatively labeled by ICAT reagent. Note that MBB fluorescence is decreased at high versus low pO 2 , indicating a loss of free thiols to S-oxidation (8,16). Error bars represent S.E.  FIGURE 2. RyR1 Cys residues insensitive to pO 2 . Following analysis according to the scheme illustrated in Fig. 1A, 72 Cys residues were identified as pO 2 -insensitive, including a set of 37 Cys residues labeled with IA or ICAT only and therefore members of a fully oxidized or a fully reduced population, respectively. Fully oxidized Cys residues are likely to participate in structural disulfides with the exception of Cys 537 , which could be labeled with dimedone (see Table 5). Note that the analysis could not discriminate between multiple Cys residues located within a single peptide. In the four cases in which multiple Cys residues were present in a single peptide (Cys 65/66 , Cys 206/208/209 , Cys 604/607/609 , and Cys 746/747 ), one Cys was ICAT-labeled, and the remainder were IA-labeled in a pO 2 -insensitive fashion. The locations within RyR1 of fully oxidized or reduced Cys residues are indicated. Transmembrane domains; regions implicated in the interaction of RyR1 with the L-type Ca 2ϩ channel (dihydropyridine receptor (DHPR)), FK506-binding protein 12 (FKBP12), and calmodulin (CaM); leucine/ isoleucine zipper (LIZ) motifs; and the SPRY1 domain as well as hot spot regions implicated in malignant hypothermia/central core disease (MH, CCD) are delineated on the basis of the human RyR1 sequence (for a recent review, see Ref. 33). Quantitative data are provided in Table 1. Black circles, fully oxidized; red circles, fully reduced; half red-half black circles, fully oxidized or reduced.

TABLE 1 pO 2 -insensitive RyR1 Cys residues
Cys residues labeled with IA alone (fully oxidized at low and high pO 2 ) are not shown (see Fig. 2). Cys residues labeled with ICAT alone (fully reduced at high and low pO 2 ) are underlined. For all remaining pO 2 -independent Cys residues, some proportions were detected with IA and with ICAT labeling, and L:H ICAT ratios were ϳ1:1 at low or high pO 2 and at high pO 2 with NADPH supplementation. Note anomalous values for Cys 3193 and Cys 3635 (see Fig. 3 and text). Standard error was derived from replicated biological samples (n ϭ 2-5), multiple peptides with different charge and/or methionine oxidation, or multiple scanning. PEG-CAT, polyethylene glycol-coupled catalase; DPI, diphenyleneiodonium; -, data not available.  (Fig. 5). Quantification of disulfide formation with a native peptide reference method (see "Exper-imental Procedures") (25) yielded H:L ratios of 1.23 Ϯ 0.16 and 1.45 Ϯ 0.04 (n ϭ 3 experiments) for Cys 2305/2310 and Cys 2602/2611 , respectively. We did not detect disulfide formation between Cys residues located in separate peptides, but theoretical analysis of the results of trypsin/chymotrypsin digest indicates that very few composite peptides resulting from disulfide formation would possess a mass low enough to allow their detection by MS/MS.

Cys
Although we did not observe sulfenic acid at either low or high pO 2 , we considered the possibility that sulfenic acid might represent a transient intermediate en route to disulfide formation. We did not identify sulfenic acid by dimedone labeling (28) following incubation of SR vesicles at low or high pO 2 . However, if muscle homogenates were incubated with dimedone (in room air; 21% O 2 ) prior to the isolation of SR vesicles and purification of RyR1, then 13 dimedone-modified Cys residues were detected ( Table 5). Eleven of those 13 Cys residues were not among the set of pO 2 -sensitive Cys residues (listed in Tables 2 and 4). However, Cys 36 and Cys 120 were identified as both dimedone-labeled and pO 2 -sensitive, consistent with a role for sulfenic acid as a transient intermediate to higher oxidation, presumably disulfide formation.

DISCUSSION
It has been established that the redox state of pO 2 -sensitive RyR1 Cys residues is regulated by a cycle of dynamic S-oxidation and reduction that operates within the SR, but the identity of pO 2 -sensitive Cys residues has remained unknown (16,17). We show here that pO 2 -coupled redox regulation of RyR1 is exerted through S-oxidation of 21 Cys residues that are distributed widely within RyR1. Previous reports based upon thiol labeling have reported pO 2 -coupled oxidation of a set of approximately six to eight thiols on average (8). The analytic scheme used here based upon ICAT labeling and mass spectrometric analysis revealed that this set of thiols represents substoichiometric oxidation of a larger set. (Indeed, only two of the Cys residues we identified as pO 2 -sensitive appeared to be fully reduced at low pO 2 , and more generally, 54 of the 93 Cys residues we identified comprise mixed populations of reduced and oxidized Cys residues at both low and high pO 2 ). In the best characterized cases of redox regulation of protein function by S-oxidation, including most prominently protein-tyrosine phosphatases, modification preferentially targets a single,

-sensitive RyR1 Cys residues
For most pO 2 -dependent Cys residues, some proportions were detected with IA and with ICAT labeling, and ICAT L:H ratios Ն1:1.24 (L represents light ICAT, low pO 2 ; H represents heavy ICAT, high pO 2 ) indicate enhanced oxidation at high versus low pO 2 . In two cases (indicated by underlining), only ICAT labeling was detected. In all cases, the increase in L:H ratio at high versus low pO 2 was eliminated by inhibiting NADPH oxidase activity with DPI or scavenging H 2 O 2 with polyethylene glycol-coupled catalase (PEG-CAT) and was enhanced by supplementation with NADPH. Standard error was derived from replicated biological samples (n ϭ 2-5), multiple peptides with different charge and/or methionine oxidation, or multiple scanning. -, data not available.   Fig. 1A). The graph includes the 71 Cys residues that were identified as ICAT-labeled under control conditions (low and high pO 2 with no exogenous additives) in more than one biological replicate and that were contained in peptides that included only a single Cys (allowing unambiguous identification). Light:heavy ratios Ͼ1 indicate Cys oxidation (loss of free thiol) at high versus low pO 2 . A light:heavy cutoff value of 1.24:1 (indicated by the upper dotted line) yielded a set of 13 Cys residues. An asterisk indicates Cys 3193 and Cys 3635 , which exhibited anomalous behavior (see Table 1). B, Cys residues identified as subject to oxidation at high versus low pO 2 are listed with reference to their location within RyR1. Note that the L:H ratio was enhanced for all listed Cys residues by supplemental NADPH (see Table 2), but that the L:H ratio for Cys residues listed in the right-hand column was ϳ1:1 in the absence of supplemental NADPH; that is, pO 2 -dependent oxidation required supplemental NADPH (see Table 3).   active site Cys within a minor population of protein (27). Thus, our results provide an apparently novel example of multisite redox regulation by physiological S-oxidation. Physiological S-Oxidation of Cys Residues in RyR1-There are few analyses of oxidative modification of protein function in situ that have identified the source and nature of oxidizing equivalents, the protein targets and Cys residues modified, the nature of the modification, and the consequences of modification for protein function. Skeletal muscle RyR1, which contains 100 Cys residues, has long provided a model of Cys-based redox indicates that both Cys residues were in the reduced state prior to initial blocking by alkylation. Note that all four forms are detected in samples prepared from SR vesicles held at high or low pO 2 but that the disulfide form is more abundant (1.23-fold) in samples prepared at high pO 2 (see "Results").  regulation. In particular, it has been reported that Nox2 is localized to transverse tubule membranes in skeletal muscle and that addition of NADPH to isolated triads (containing transverse tubule as well as SR membranes) results in the glutathionylation of RyR1 (20). However, we did not identify S-glutathionylated Cys residues within RyR1 following pO 2 -coupled activation of RyR1, consistent with previous work showing that Nox2 may play a role in tetanic contraction but is not a source of pO 2 -coupled reactive oxygen species production (16). In contrast, Nox4 was identified previously as the source of pO 2 -coupled S-oxidation of RyR1 that activates RyR1 to enhance Ca 2ϩ release from the SR (16). Here, in addition to identifying pO 2sensitive Cys residues within RyR1, we identified disulfide as a stable product of S-oxidation but not other potential, reversible oxidative modifications, including S-glutathionylation, sulfenic acid, sulfinic acid, or sulfenamide (a form of modification within protein-tyrosine phosphatases (29,30)). Thus, glutathionylation and disulfide modification may represent signatures of contraction-and oxygen-regulated RyR1 activity, respectively. Previous studies using mass spectrometric analysis but relying primarily upon the administration of exogenous oxidants reported modification of multiple Cys residues within RyR1 by glutathionylation or unidentified S-oxidation as well as (by inference) disulfide formation (14). Although it is not possible to deduce the coverage of RyR1 Cys residues in previous analyses, remarkably, none of the Cys residues reported previously to be oxidatively modified are among the set of Cys residues shown here to be subject to physiological, pO 2 -coupled S-oxidation with the exception of Cys 36 , which was identified previously as a possible participant in disulfide formation catalyzed by the addition of high concentrations of H 2 O 2 (5 mM) (14). Conversely, with the exception of Cys 36 , Cys residues identified here as pO 2 -sensitive do not appear among the Cys residues reported to be oxidatively modified by exogenous agents (14), although Cys 2606/2611 , which we identified as participating in a pO 2 -sensitive disulfide, were identified as hyperreactive by selective labeling with a thiol-reactive coumarin (31). Interestingly, a recent crystal structure of the N-terminal domain of RyR2 in which Cys 36  Trans-Transmembrane membrane FIGURE 6. Schematic illustration of the distribution within RyR1 of Cys residues subject to pO 2 -coupled redox modification. Among the 100 Cys residues of rabbit RyR1, color-coding identifies Cys residues unaffected by pO 2 (n ϭ 72; black circles), oxidized at high versus low pO 2 (enhanced by exogenous NADPH) (n ϭ 13; red circles), and oxidized at high pO 2 only with addition of exogenous NADPH (n ϭ 8; green circles). Cys residues not detected in our analysis are also indicated (n ϭ 7; blue circles). Identified disulfides formed more abundantly at high versus low pO 2 are indicated by brackets (Cys 2305 -Cys 2310 and Cys 2606 -Cys 2611 ), and for reference, the previously identified site of S-nitrosylation is indicated (Cys 3635 ). The following functional domains within RyR1 are delineated: transmembrane domains; regions implicated in the interaction of RyR1 with the L-type Ca 2ϩ channel (dihydropyridine receptor (DHPR)), FK506-binding protein 12 (FKBP12), and calmodulin (CaM); leucine/isoleucine zipper (LIZ) motifs; and the SPRY1 domain as well as hot spot regions implicated in malignant hypothermia/central core disease (MH, CCD).

TABLE 5 Dimedone labeling of Cys thiols in RyR1
Muscle homogenates were incubated with dimedone prior to preparation of SR vesicles and analysis. Dimedone labeling (red) was identified by MS/MS. Mass is given as Daltons.
which Cys 36 could form a vicinal pair with Cys 65 (32). However, we did not identify Cys 65 as pO 2 -sensitive. Generally, studies of oxidative modification of Cys residues within proteins that are involved in cellular signal transduction have relied on the in vitro application of oxidants, principally H 2 O 2 , to address the form of modification. In the much studied case of protein-tyrosine phosphatases, in vitro analysis has identified multiple possible oxidative modifications, including intramolecular disulfides and sulfenamide (27). However, there is little evidence bearing on the nature of oxidative modification of Cys residues induced by endogenously generated H 2 O 2 , and more generally, the nature of oxidative modification that may subserve physiological signal transduction remains largely unexplored (2). Our results indicate that disulfide formation serves as a principal dynamic modification of Cys residues within RyR1 that is governed by endogenous oxidative and reductive mechanisms to regulate RyR1 function allosterically.
Our finding that ϳ21 Cys residues contribute to a total of 6 -8 mol of thiol/mol of RyR monomer that are subject to pO 2coupled oxidation is in keeping with the evolving understanding that regulatory post-translational modifications of proteins are typically present in substoichiometric amounts. Subpopulations of proteins may be subject to differential regulation, and sub-stoichiometric oxidative modification of RyR1 may be viewed as providing a selective gain of function. However, we cannot determine whether a subpopulation of RyR1 is subject to privileged modification (e.g. by Nox4 that may be associated with a fraction of RyR1) or whether redox modifications are in fact evenly distributed across the entire population of RyR1 in which case oxidation of single subunits of the RyR1 tetramer may cooperatively influence the activity of other subunits to subserve pO 2 -coupled regulation of calcium release.
The Distribution of Physiologically S-Oxidized Cys residues in RyR1-The localization of identified Cys residues within RyR1 is illustrated in Fig. 6. All pO 2 -sensitive Cys residues are located in cytoplasmic portions of RyR1 and with a single exception (Cys 4663 ) within the N-terminal domain. Cys residues subject to pO 2 -dependent S-oxidation are located within multiple functional domains (for a recent review, see Ref. 33), including domains implicated in activity-regulating interactions of RyR1 with FK506-binding protein 12 (33,34) and the dihydropyridine receptor (33,35). In addition, pO 2 -sensitive Cys residues are located in domains implicated in malignant hyperthermia and central core disease (33). It is not possible to deduce which identified Cys residues play a role in pO 2 -dependent enhancement of RyR1 activity or may be subject to dysregulated S-oxidation in disease, and concerted effects involving multiple Cys residues seem likely. However, it should be noted that mutation of Cys 36 (Cys 35 in the human sequence) has been implicated directly in malignant hyperthermia and central core disease (36) and that oxidative stress associated with aging and in a model of muscular dystrophy results in the displacement of FK506-binding protein 12 from RyR1 and consequently Ca 2ϩ leakage through RyR1 that contributes to loss of muscle function (37,38). It is also of interest that Cys 3635 , which is not a target of pO 2 -coupled S-oxidation but which serves as the principal target of endogenous S-nitrosylation that is gated allosterically by RyR1 S-oxidation of alternative Cys residues (10, 13), is located distantly in the primary sequence from identified pO 2sensitive Cys residues within RyR1 (Fig. 6).
Ninety-eight of 100 RyR1 Cys residues are conserved in mammals, including all pO 2 -sensitive Cys residues identified here with the exception of Cys 305 (oxidized at high pO 2 only in the presence of NADPH). Furthermore, the majority of RyR1 Cys residues that we identified as subject to physiological redox regulation are also conserved in RyR2 and RyR3, consistent with the finding that the activity of RyR2 is enhanced at high versus low pO 2 (15). Thus, our findings suggest that S-oxidation of Cys residues within the RyRs is likely to serve as a mechanism for physiological redox regulation of RyR function across many mammalian cell types and tissues.