Stabilization of the Skeletal Muscle Ryanodine Receptor Ion Channel-FKBP12 Complex by the 1,4-Benzothiazepine Derivative S107

Activation of the skeletal muscle ryanodine receptor (RyR1) complex results in the rapid release of Ca2+ from the sarcoplasmic reticulum and muscle contraction. Dissociation of the small FK506 binding protein 12 subunit (FKBP12) increases RyR1 activity and impairs muscle function. The 1,4-benzothiazepine derivative JTV519, and the more specific derivative S107 (2,3,4,5,-tetrahydro-7-methoxy-4-methyl-1,4-benzothiazepine), are thought to improve skeletal muscle function by stabilizing the RyR1-FKBP12 complex. Here, we report a high degree of nonspecific and specific low affinity [3H]S107 binding to SR vesicles. SR vesicles enriched in RyR1 bound ∼48 [3H]S107 per RyR1 tetramer with EC50 ∼52 µM and Hillslope ∼2. The effects of S107 and FKBP12 on RyR1 were examined under conditions that altered the redox state of RyR1. S107 increased FKBP12 binding to RyR1 in SR vesicles in the presence of reduced glutathione and the NO-donor NOC12, with no effect in the presence of oxidized glutathione. Addition of 0.15 µM FKBP12 to SR vesicles prevented FKBP12 dissociation; however, in the presence of oxidized glutathione and NOC12, FKBP12 dissociation was observed in skeletal muscle homogenates that contained 0.43 µM myoplasmic FKBP12 and was attenuated by S107. In single channel measurements with FKBP12-depleted RyR1s, in the absence and presence of NOC12, S107 augmented the FKBP12-mediated decrease in channel activity. The data suggest that S107 can reverse the harmful effects of redox active species on SR Ca2+ release in skeletal muscle by binding to RyR1 low affinity sites.

FKBP12 associates predominantly with the skeletal muscle isoform to regulate RyR1 function. FKBP12.6 binds with higher affinity to the cardiac muscle isoform RyR2 to modulate cardiac muscle SR Ca 2+ release [5]. FKBPs belong to a family of immunophilins that exhibit cis/trans isomerase activity. Pharmacological removal of FKBPs using rapamycin or FK506, causes the dissociation of FKBP12 and FKBP12.6 from the RyR macromolecular complexes, uncouples RyR ion channels from their neighbors and activates Ca 2+ release from SR [6][7][8][9]. In lipid bilayers, rapamycin and FK506 increase channel activity and lead to the formation of channel openings with reduced conductance which are referred to as substates [6,7]. Two studies reported that elimination of FKBP12 binding to RyR1 decreased voltage-gated SR Ca 2+ release in myotubes [10] and fast skeletal muscle fibers [11].
RyR1 and RyR2 channels are subject to post-translational modifications in skeletal and cardiac muscle. In failing hearts, PKA-mediated hyperphosphorylation removed FKBP12.6 from RyR2 and increased channel activity [12]. In a mouse skeletal muscle model of heart failure and in patients with heart disease, exercise was linked to hyperphosphorylation and depletion of FKBP12 from RyR1 that results in increased channel activity and decreased exercise capacity [13,14]. These findings suggested that hyperphosphorylation and dissociation of the small FKBP subunits (also referred to as calstabins) from RyRs result in leaky SR Ca 2+ channels and impaired muscle function [15]. However, other laboratories failed to support this proposal [16][17][18][19]. In addition to PKA-mediated phosphorylation, mechanisms implicated in the generation of leaky Ca 2+ release channels include oxidation and Snitrosylation, and dissociation of FKBPs from RyR1 and RyR2. In dystrophic muscle, increased S-nitrosylation of RyR1 [20] and RyR2 [21] resulted in partial dissociation of FKBPs, and the formation of leaky Ca 2+ channels. Gonzalez et al. [22] reported that elimination of neuronal nitric oxide synthase (nNOS), which is closely associated with RyR2 in cardiac muscle, increased diastolic Ca 2+ levels. nNOS elimination was associated with decreased Snitrosylation, increased oxidation of RyR2, leakage of SR Ca 2+ , and arrhythmogenesis in cardiomyocytes. FKBP12.6 binding to RyR2 and RyR2 phosphorylation were not altered in homozygous nNOS knockout mice.
In the present study, [ 3 H]S107 binding and regulation of RyR1 by FKBP12 and S107 were examined. The results indicate a low equilibrium binding affinity of S107 to multiple RyR1 sites, and provide insight in how FKBP12 and S107 regulate RyR1 under reducing and oxidizing conditions.

Ethics Statement
This study was carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the University of North Carolina at Chapel Hill Institutional Animal Care and Use Committee (10-056).

Preparation of Skeletal Muscle Homogenates and Membrane Fractions
Whole muscle homogenates and SR membrane fractions were obtained from leg and back muscle of 2-3 month old rabbits [31]. Rabbit muscle was homogenized at 4uC using a Waring Blender in 6.5 volumes of 0.1 M NaCl, 2 mM EDTA 0.2 mM ethylene  [32]. Aliquots were quick frozen and stored at 280uC.

Preparation of Supernatant Fraction
Rabbit skeletal muscle was homogenized at 4uC in 3 volumes of 0.1 M NaCl, 20 mM imidazole, pH 7, and protease inhibitors using a Tekmar Tissumizer for 367 s at 13,500 rpm. A supernatant fraction was prepared by centrifuging homogenates for 45 min at 35,000 rpm in a Beckman Ti50 rotor.

FKBP12 Preparation
Recombinant rabbit FKBP12 engineered with an aminoterminal GST tag and TEV protease cleavage site was purified from E. coli using GST and Q-Sepharose chromatography. Following buffer exchange into phosphate buffered saline and cleavage of the GST affinity tag by the addition of histidine-tagged TEV protease, the cleaved FKBP12 protein was passed through nickel and GST-Sepharose columns to remove the histidinetagged TEV protease and cleaved GST tag.
[ 3 H]S107 Binding [ 3 H]S107 binding to SR vesicles was determined at 24uC using a centrifugation assay. Unless otherwise indicated, SR vesicles were incubated with 0.1 mM [ 3 H]S107 and 0.1 to 100 mM unlabelled S107 in 0.25 M KCl, 20 mM imidazole, pH 7.0 containing 50 mM free Ca 2+ and protease inhibitors, followed by centrifugation in a Beckman Airfuge for 30 min at 90,0006g. Nonspecific binding was determined by measuring [ 3 H]S107 binding to SR vesicles incubated for 10 min at 95uC. Bound [ 3 H]S107 was determined by scintillation counting after solubilisation of pellets in 50 mM Tris-HCl, pH 8.5 containing 2% SDS. Specific binding parameters were calculated using the four parameter logistic equation (SigmaPlot11).
where Y is specifically bound S107 per mg protein, min and max are the minimal and maximal S107 binding values per mg protein, respectively, x is the free S107 concentration, EC50 is the concentration of S107 that results in half-maximal binding, and the Hillslope represents an apparent n H . Representative immunoblot of SR vesicles not treated and treated with FK506. SR vesicles were incubated with 10 mM FK506 as described in Materials and Methods, followed by centrifugation to remove FK506 and dissociated FKBP12. (B and C) FKBP12 dissociation from SR vesicles in the presence of GSH and GSSG. Immunoblots of SR vesicles not treated with FK506 were incubated for 1 and 20 h at 24uC in 0.25 M KCl, 20 mM imidazole, pH 7.0, 50 mM free Ca 2+ , protease inhibitors, and the indicated concentrations and ratios (5 mM total glutathione) of GSH and GSSG in the absence and presence of 44 mM S107. Free FKBP12 was removed by centrifugation. Data were normalized to SR vesicles not incubated (gray bar, 0 min) and are the mean 6 SEM of 4-5 experiments. *p,0.05 compared to SR vesicles at 0 min not treated with S107. doi:10.1371/journal.pone.0054208.g002

Distribution Coefficient of [ 3 H]S107
The differential solubility of 0.1 mM [ 3 H]S107 between an aqueous and hydrophobic phase was measured using equal volumes 0.25 M KCl, 50 mM Ca 2+ , 20 mM imidazole, pH 7.0 solution and vegetable oil. The two phases were mixed vigorously and separated by centrifugation. Measurement of 3 H radioactivity in the two phases by liquid scintillation counting yielded a distribution coefficient of 0.6160.05 (n = 2) between the hydrophobic and aqueous phases.

Single Channel Recordings
Single channel measurements were performed at 24uC in planar lipid bilayers containing 5:3:2 ratio of phosphatidylethanolamine, phosphatidylserine, and phosphatidylcholine (25 mg of total phospholipid/ml of n-decane) [33]. SR vesicles were added to the cis cytoplasmic side of the bilayer, with the trans SR luminal side defined as ground. Measurements were made using 0.25 M Cs methanesulfonate, 10 mM CsHepes, pH 7.4 on both sides of the bilayer to minimize K + and Cl 2 channel activities in SR vesicles [34]. Data acquired using test potentials of 635 mV were sampled at 10 kHz and filtered at 2 kHz. Channel open probability (P o ) was determined from at least 2 min of recordings for each condition.

[ 3 H]Ryanodine Binding
The plant alkaloid ryanodine is widely used as a highly specific probe of RyR channel activity because it preferentially binds the open channel configuration [35]. In the present study a relatively low concentration of [ 3 H]ryanodine was used to detect changes in binding and thus RyR1 activity. SR vesicles were incubated with 3 nM [ 3 H]ryanodine at 24uC in 0.25 M KCl, 20 mM imidazole, pH 7.0 with ,7 mM free Ca 2+ and protease inhibitors. Nonspe- (A-C) SR vesicles not treated with FK506 were incubated for 5 h at 24uC with or without 0.10 mM NOC12 in the absence and presence of 44 mM S107 in 0.25 M KCl, 20 mM imidazole, pH 7.0, 7 mM free Ca 2+ and protease inhibitors. S-nitrosylation was stopped by centrifugation. Resuspended samples were separated on 8-20% (FKBP12) and 3-12% (RyR1 and Cys-SNO) gradient SDS-PAGE gels and transferred to nitrocellulose membranes to detect S-nitrosylation of RyR1, and FKBP12 and RyR1 proteins. Data are the mean 6 SEM of 4 determinations. *p,0.05 compared to control samples (B) and samples with NOC12 and S107 (C). (D and E) SR membranes were incubated with and without 44 mM S107 and 0.1 mM NOC12 at 24uC for 90 min, solubilized, and immunoprecipitated as described in Methods. Immunoblots of RyR1 and FKBP12 are shown. Data are the mean 6 SEM of 4 experiments. * p,0.05 compared to control samples and samples incubated with NOC12 and S107. doi:10.1371/journal.pone.0054208.g003 S107 Effects on RyR1-FKBP12 PLOS ONE | www.plosone.org cific binding was determined using 1000-fold excess of unlabelled ryanodine. Aliquots of samples were diluted 9-fold with ice-cold water and placed on Whatman GF/B filters saturated with 2% polyethyleneimine. Filters were washed with three 5-ml volumes of ice-cold 0.

SDS-PAGE and Immunoblot Analysis
RyR1 was detected using 3-12% gradient SDS-PAGE and transfer to nitrocellulose membranes. Experiments for detection of S-nitrosocysteines were performed under nonreducing conditions [36]. Membranes were blotted with 2% ECL Advance blocking reagent in 0.5% Tween 20, Tris buffered saline (TBS), pH 7.4 at 24uC for 2 h. Membranes were probed with primary anti-Cys-SNO polyclonal antibody and secondary peroxidase-conjugated anti-rabbit IgG antibody, and with anti-RyR1 polyclonal antibody and peroxidase-conjugated anti-mouse IgG, using the ECL detection method. FKBP12 was detected using 8-20% gradient SDS-PAGE and anti-FKBP12 monoclonal antibody.

Coimmunoprecipitation of RyR1 and FKBP12
For co-immunoprecipitation experiments, SR membranes were incubated with and without 44 mM S107 and 0.1 M NOC12 at 24uC for 90 min and solubilized at 2 mg protein/ml in 15 mM NaPipes pH 7.2 containing 0.5 M sucrose, 1 M NaCl, 2.5% Triton X-100 and protease inhibitors [16]. Following incubation for 10 min on ice, samples were diluted with equal volumes of icecold water, incubated for another 20 min on ice, and centrifuged for 10 min at 13,000 rpm. After dilution to 0.125 M NaCl and 0.3% Triton X-100 and addition of 0.9 mM Ca 2+ , the supernatant fraction was incubated with RyR1 monoclonal antibody D286 (1:10 dilution) for 20 h at 4uC. Protein G coupled to magnetic beads was added and samples were incubated for 2 h at 4uC. After washing twice with 20 mM NaPipes, pH 7.2 containing 0.125 M NaCl, 0.3% Triton X-100 and protease inhibitors, bound proteins were released with 30 ml 46 SDS buffer at 37uC for 20 min and separated on SDS gels and analyzed on immunoblots as described above.

Biochemical Assays and Data Analysis
Free Ca 2+ concentrations were determined by adding the appropriate amounts of Ca 2+ and EGTA calculated using stability constants and a published computer program [37]. Free Ca 2+ concentrations were verified with the use of a Ca 2+ selective electrode. Unless otherwise indicated, differences between samples were analyzed by one-way Anova followed by Tukey t-test; p,0.05 was considered significant.

[ 3 H]S107 Binding to SR Vesicles
A distribution coefficient of 0.61 (see Materials and Methods) suggests that S107 can be taken up by cells and partition into the large RyR complexes. The time course of total and nonspecific [ 3 H]S107 binding was determined by incubating SR vesicles enriched in RyR1 with 0.2 mM [ 3 H]S107 for 1 to 9 h at 24uC. Specific [ 3 H]S107 binding to the vesicles reached a maximum after ,7 h (Fig. 1A). A similar time course of specific [ 3 H]S107 binding was observed when SR vesicles were incubated with 2 mM [ 3 H]S107 and 20 mM [ 3 H]S107 (Fig. S1).
Equilibrium binding of [ 3 H]S107 was determined in a homologous competition assay. SR vesicles were incubated for 7 h with 0.1 mM [ 3 H]S107 in the presence of 0.1 mM to 100 mM unlabelled S107. Total, nonspecific and specific [ 3 H]S107 binding are shown in Fig. 1B. Analysis of specific binding using a four parameter logistic equation (see Methods) showed SR vesicles bound 353646 pmol [ 3 H]S107/mg protein (n = 93, p,0.0001) with EC 50 = 51.868.8 mM S107 (p,0.0001) and Hillslope = 2.160.4 (p,0.0001). A Hillslope of 2.1 reflected low cooperativity among multiple S107 binding sites. The B max of Figure 5. Effect of S107 on the stability of FKBP12-RyR1 complex in skeletal muscle homogenates. (A and B) Skeletal muscle homogenates were incubated without (control) or with 5 mM GSH, 5 mM GSSG or 0.10 mM NOC12 in the absence or presence of 44 mM S107 for 20 h at 24uC. Unbound FKBP12 was removed by centrifugation and the amounts of RyR1 and FKBP12 were detected using anti-RyR1 and anti-FKBP12 antibodies. Homogenates incubated without glutathione and NOC12 served as control. Data are the mean 6 SEM of 8 experiments. *p,0.05 compared to control homogenates without S107. # p,0.05 compared to homogenates incubated with NOC12 in the absence of S107. doi:10.1371/journal.pone.0054208.g005 Figure 6. Effects of S107 on FKBP12 binding to RyR1. (A) Immunoblots of RyR1 and FKBP12. SR vesicles depleted of FKBP12 were incubated for 20 h at 24uC in 0.25 M KCl, 20 mM imidazole, pH 7.0, 7 mM free Ca 2+ , protease inhibitors and 10 nM FKBP12 without or with 44 mM S107 in the absence and presence of 5 mM GSH, 5 mM GSSG or 0. 10  When SR vesicles were treated with FK506 to remove FKBP12 ( Fig. 2A), a similar time course of [ 3 H]S107 binding was observed with similar amounts of [ 3 H]S107 bound to SR vesicles containing or depleted of FKBP12 (Fig. S2, Table 1). A slower time course of [ 3 H]S107 binding was observed when vesicles were incubated in the presence of free Ca 2+ that closed (,0.01 mM Ca 2+ ) compared to partially activated (50 mM Ca 2+ ) RyR1 (Fig. S2, Table 1). However, similar amounts of [ 3 H]S107 were bound to the SR vesicles after incubation of 9 h. We conclude that neither the removal of FKBP12 nor a change in RyR1 activity altered [ 3 H]S107 binding to RyR1.

Effects of S107 on Stability of FKBP12-RyR1 Complex in the Absence of Exogenous FKBP12
RyR1s are redox-sensitive Ca 2+ channels whose activity is modulated by reduced (GSH) and oxidized (GSSG) glutathione [38][39][40]. To determine the effect of S107 on the stability of the FKBP12-RyR1 complex under reducing and oxidizing conditions, SR vesicles were incubated in the presence of 5 mM GSH or 5 mM GSSG for 1 h and 20 h (Figs. 2B and C). After 1 h in the presence of GSH, 44 mM S107 blocked the release of 10-15% of vesicle-associated FKBP12. Similar amounts (20%) of FKBP12 They were corrected for amounts of FKBP12 associated with FK506 treated SR vesicles kept on ice and normalized to SR vesicles not treated with FK506 (gray bar). *p,0.05 compared to SR vesicles not treated with FK506 and not incubated with FKBP12 and S107. #p,0.05 compared to FKBP12-depleted SR vesicles incubated for the same time (10 min or 2 h) in the presence of FKBP12 but absence of S107. doi:10.1371/journal.pone.0054208.g007 S107 Effects on RyR1-FKBP12 PLOS ONE | www.plosone.org were released after 20 h in the presence and absence of S107. In the presence of 5 mM GSSG, significantly larger amounts of FKBP12 were released from the vesicles (35% and 95% after incubation for 1 h and 20 h, respectively). Addition of 44 mM S107 did not alter FKBP12 release from RyR1 in the presence of 5 mM GSSG (Figs. 2B and C).
The activity of RyR1 depends on the GSH/GSSG ratio [38,39]. Incubation of SR vesicles for 20 h with 4.5 mM GSH and 0.5 mM GSSG (9:1 GSH/GSSG) released amounts of FKBP12 similar to vesicles incubated with 5 mM GSH in the presence and absence of 44 mM S107 (Fig. 2C). About half of FKBP12 content was released from SR vesicles after incubation for 20 h at GSH/ GSSG ratio of 4:1. Taken together, Fig. 2C shows that a decrease in the GSH/GSSG ratio decreased the stability of the FKPB12-RyR1 complex.
NO and the NO donor NOC12 activate RyR1 by Snitrosylation of a single cysteine residue (Cys3635) [36,41]. Figs. 3A-C show that treatment of SR vesicles with 0.10 mM NOC12 significantly increased RyR1 S-nitrosylation and decreased FKBP12 association with SR vesicles. S107 increased the amount of FKBP12 bound to NOC12-treated vesicles (Fig. 3C) without altering Cys-NO content (Fig. 3B). This suggested that S107 increased the binding affinity of FKBP12 to the Snitrosylated RyR1.
NOC12-and S107-treated vesicles were solubilized, and RyR1 and FKBP12 co-immunoprecipitated using an antibody against RyR1. Immunoblots indicate that the presence of S107 minimized the dissociation of FKBP12 from the RyR1-FKBP12 complex in SR membranes treated with NOC12 (Figs. 3D and E). Single channel data were obtained as described in A. Data are the mean 6 SEM of 4-12 single channel recordings. * p,0.05 compared to RyR1 not treated with FK506 in the absence of S107 and FKBP12. # p,0.05 compared to FK506-treated RyR1 incubated in the absence of S107 and FKBP12. % p,0.05 compared to FK506treated RyR1 incubated with FKBP12 in absence of S107. p values were determined by Student's t-test. doi:10.1371/journal.pone.0054208.g008 S107 Effects on RyR1-FKBP12 PLOS ONE | www.plosone.org The effects of NOC12 and S107 on RyR1 activity were further investigated in a ligand binding assay using the RyR1-specific probe ryanodine. Preferential binding [ 3 H]ryanodine to the open RyR1 provides an indirect measure of RyR1 activity [35]. NOC12 increased [ 3 H]ryanodine binding, suggesting an increase in RyR1 activity at 0.03 to 0.30 mM (Fig. 4A). Activation by NOC12 was partially inhibited by 44 mM S107. More than 10 mM S107 was required to observe a significant decrease in NOC12-induced increase in RyR1 activity using the [ 3 H]ryanodine binding assay (Fig. 4B). S107 reduced [ 3 H]ryanodine binding to NOC12-treated SR vesicles containing FKBP12 but not depleted of FKBP12 (Fig. 4C). Taken together, the data suggest that S107 decreased NOC12-mediated RyR1 activation by attenuating the dissociation of FKBP12 from RyR1.
Effect of S107 on the Stability of the FKBP12-RyR1 Complex in Skeletal Muscle Homogenates SDS-PAGE and immunoblot analysis indicated that the supernatant fraction of rabbit skeletal muscle homogenates contains 0.4360.05 mM FKBP12 (n = 5) (data not shown). This suggests, considering the 3 volume homogenization buffer, that myoplasmic FKBP12 in rabbit skeletal muscle is ,1.7 mM, which is in reasonable agreement with a myoplasmic FKBP12 concentration of 3 mM reported previously [42]. To determine the effects of exogenously added and myoplasmic FKBP12 on the stability of the FKBP12-RyR1 complex under reducing and oxidizing conditions, SR vesicles (in presence of 0.15 mM exogenous FKBP12) or homogenates were incubated without or with 5 mM GSH, 5 mM GSSG or 0.1 mM NOC12 for 20 h.
FKBP12 did not dissociate from SR vesicles incubated with 0.15 mM FKBP12 in the presence of 5 mM GSH, 5 mM GSSG or 0.1 mM NOC12 (Fig. S3). FKBP12 was also not released from the particulate matter of homogenates incubated for 20 h with GSH (Fig. 5). However, in the presence of GSSG and NOC12, ,35% and 45% FKBP12 dissociation, respectively, was observed in homogenates with 0.43 mM myoplasmic FKBP12. Addition of S107 increased the amount of FKBP12 associated with particulate matter of homogenates, which suggests that S107 minimized GSSG-and NO-mediated dissociation of FKBP12 from RyR1.

S107 Increases FKBP12 Binding to RyR1
To verify that S107 increases FKBP12 binding to RyR1, FKBP12-depleted SR vesicles were incubated for 20 h with 10 nM FKBP12 in the absence and presence of 44 mM S107 (Fig. 6). In absence of S107, highest levels of FKBP12 binding were seen with SR vesicles incubated in presence of 5 mM GSH, intermediate levels with 0.1 mM NOC12, and lowest levels to SR vesicles with 5 mM GSSG. FKBP12 binding to FKBP12-depleted vesicles was increased by S107 in presence of GSH and NOC12 but not GSSG.
To determine how S107 affects the association rate of FKBP12 binding to RyR1, FKBP12-depleted SR vesicles were incubated for 10 min, 2 h and 20 h with 1 mM FKBP12 in the presence of 5 mM GSH with and without 44 mM S107, followed by immunoblot analysis (Fig. 7). The amounts of bound FKBP12 were compared with SR vesicles not treated with FK506. After 10 min, significantly greater amounts of FKBP12 were bound to vesicles in the presence than the absence of 44 mM S107 (4967 and 3665%, respectively, of FKBP12 in SR vesicles not treated with FK506). Incubation for 2 h significantly increased the level of FKBP12 from 6864% in the absence of S107 to 8664% in presence of S107. At 20 h, FKBP12 binding was similar to untreated SR vesicles in the presence and absence of S107. That S107 increased the rate of FKBP12 binding to FKBP12-depleted SR vesicles was confirmed using the [ 3 H]ryanodine binding assay ( Spplemental Information, Fig. S4).
Effects of S107 and FKBP12 on RyR1 Single Channel Activity Using the Planar Lipid Bilayer Method SR vesicles not treated with or exposed to FK506 to remove FKBP12 ( Fig. 2A) were fused with a lipid bilayer. Before addition to the cis bilayer chamber, vesicles were incubated for 30 min without and with 25 mM S107 and 5 mM FKBP12, conditions similar to those in Fig. 7. A shorter preincubation period and lower FKBP12 concentration resulted in smaller differences in channel activity without and with S107, whereas longer incubations to achieve equilibrium equalized binding in the absence and presence of S107 (data not shown).
FKBP12-depleted RyR1 channels recorded at 2 mM cis cytoplasmic Ca 2+ exhibited a 4-fold greater single channel open probability (P o ) than RyR1 channels without FK506 (Fig. 8A,  traces 1 and 2, Fig. 8B). Removal of FKBP12 from the RyR1 complex is reported to result in the formation of channel openings with reduced conductance which are referred to as substates [6]. However, current traces and current histograms of Fig. 8A (traces 1 and 2) show that removal of FKBP12 only minimally increased the appearance of substates. A two-fold significant decrease in P o was observed for FKBP12-depleted RyR1 channels preincubated with FKBP12 (Fig. 8A, traces 2 and 4, Fig. 8B). A further 2-fold significant decrease in P o was observed for vesicles preincubated with 5 mM FKBP12 in the presence of 25 mM S107 (Fig. 8A,  traces 4 and 5, Fig. 8B). The results suggest that S107 increased binding of FKBP12 to RyR1. In agreement with the [ 3 H]ryanodine binding experiments of Fig. 4C, preincubation with 25 mM S107 alone had no effect (Fig. 8A, traces 2 and 3, Fig. 8B). In controls, preincubation with 25 mM S107, 5 mM FKBP12, or the combined preincubation with 25 mM S107 and 5 mM FKBP12 had no effect on channel activities not exposed to FK506 (Fig. 8B). Reduction of cytoplasmic Ca 2+ from 2 mM to 0.1 mM decreased P o to a similar low value for vesicles treated and not treated with FK506 (data not shown).
Single channel measurements were also performed with SR vesicles preincubated with 0.1 mM NOC12 (Fig. S5). As shown in Fig. 8, S107 lowered P o of FKBP12-depleted RyR1s in the presence but not the absence of FKBP12.

Discussion
This report describes the regulation of the skeletal muscle RyR1 ion channel by the 1,4 benzoderivative S107 and FKBP12 subunit. Previous studies suggested that S107 reverses the deleterious effects of in vivo RyR oxidation and S-nitrosylation on SR Ca 2+ release by reducing FKBP dissociation from cardiac muscle and skeletal muscle RyR complexes [20,28,29]. The present study provides the first direct insights into the S107 binding characteristics of RyR1. [ 3 H]S107 binding studies indicate that S107 binds with micromolar affinity to multiple RyR1 sites. Immunoblot, [ 3 H]ryanodine binding and single channel measurements showed that S107 inhibited S-nitrosylation-mediated FKBP12 dissociation and RyR1 activation, and attenuated the destabilizing effects of oxidized glutathione and S-nitrosylation in a skeletal muscle homogenate.
[ 3 H]S107 binding studies indicated a high degree of nonspecific and specific binding, which suggests that S107 may have additional functional effects by interacting with other SR proteins and possibly RyR1 associated lipids. Analysis of the [ 3 H]S107 binding data using a four parameter logistic equation suggested a low binding affinity (EC 50 ,50 mM) to RyR1. An unexpectedly high number of binding sites was observed that exceeded the concentration of RyR1 in SR vesicles. Further studies are needed to address why S107 binds slowly to RyR1, how S107 alters the regulation of RyR1 by other endogenous effectors, and the basis of the high number of S107 binding sites in RyR1. A distribution coeffficient of 0.61 suggests that S107 partitions into the large subunits of RyR1 allowing the binding to multiple sites.
Four FKBP12 molecules are bound to the homotetrameric RyRs [32]. Dissociation of FKBP12 from the RyR1 complex increased open channel probability (P o ) and induced substates [6,7,43,44]. In agreement with Barg at al. [45], we found that removal of FKBP12 increased RyR1 activity at 2 mM free Ca 2+ , but did not increase the frequency of RyR1 substates in single channel measurements. The reason for this discrepancy is unclear but may result from differences in sample preparation and recording conditions.
The RyRs are redox-sensitive channels whose redox state and activity depend on the glutathione redox potential and reactive oxygen and nitrogen species [29,38,39,[46][47][48][49][50]. There are 100 cysteine residues in the RyR1 subunit and one cysteine per FKBP12 subunit that can be potentially modified [2]. RyR1s are likely in a reduced state in vivo based on the reducing environment of thiol-reducing compounds, the most abundant being glutathione [51]. Unlike the RyR1 polypeptides, which are readily modified by redox active species, the small FKBP12 subunit was neither S-nitrosylated using NO-donor NOR-3 nor S-glutathionylated by GSH and H 2 O 2 [52]. S-nitrosylation decreased RyR1 binding affinity for FKBP12 several fold in a [ 35 S]FKBP binding exchange assay, whereas S-glutathionylation had no effect. Treatment with H 2 O 2 and diamide lowered FKBP12 binding to RyR1 [53]. In the present study with SR vesicles in the absence of exogenously added FKBP12, oxidation resulted in FKBP12 dissociation from the RyR1 complex when free thiol content was decreased from 37.561.0 thiols/RyR1 subunit in the presence of GSH to 30.862.7 thiols/RyR1 subunit in the presence of GSSG [54], and when one cysteine (Cys3635) per RyR1 subunit was Snitrosylated by the NO donor NOC12 [36,41]. The results indicate that oxidation and S-nitrosylation of RyR1 destabilize the FKBP12-RyR1 complex. S107 has a significant effect on skeletal muscle function by minimizing the release of FKPB12 from RyR1 and thereby NOmediated SR Ca 2+ leakage during exercise [30], and oxidation and S-nitrosylation in dystrophic and aged skeletal muscle [20]. Rapamycin-induced deleterious effects on mitochondrial function could be prevented by treatment of FDB fibers with 5 mM S107 [28]. Daily administration of 1.3 mg S107 (,20 mmol S107/kg) for 4 weeks improved muscle function and exercise capacity in WT but not FKBP12 2/2 mice [28]. We found that S107 increased FKBP12 binding to RyR1 incubated with the NO donor NOC12. Greater than 10 mM S107 was required to significantly inhibit NOC12-mediated activation of RyR1. This was in agreement with the [ 3 H]S107 binding studies, which showed that S107 binds with low affinity to RyR1. Removal of FKBP12 from SR vesicles did not appreciably alter [ 3 H]S107 binding, which suggests that S107 binding did not depend on FKBP12.
Previous immunoprecipitation and immunoblot experiments showed that treatment of mice with S107 stabilized binding of FKBP12 to RyR1 without attenuating RyR1 oxidation and Snitrosylation in skeletal muscle of 24-month-old mice [28]. We observed GSSG-and NO-mediated dissociation of FKBP12 when SR vesicles were incubated in the absence of FKBP12. Presence of 0.15 mM FKBP12 eliminated the dissociation of FKBP12 from SR vesicles. However, GSSG-and NO-mediated dissociation was observed in skeletal muscle homogenates that contained 0.43 mM myoplasmic FKBP12. At this time, the mechanism(s) leading to the dissociation of FKBP12 in homogenates remains unclear. One possibility we cannot rule out is that binding of FKBP12 to myoplasmic proteins decreases the free FKBP12 concentration. In this case, FKBP12 could dissociate from RyR1, and S107 could attenuate dissociation of FKBP12 by increasing FKBP12 binding affinity to RyR1. Another possibility is that unknown factors lead to the release of FKBP12 in homogenates. Further studies will address the mechanism(s) for the release of FKBP12 and enable S107 to stabilize the FKBP12-RyR1 complex in skeletal muscle.

Conclusion
The present study shows that S107 binds to multiple RyR1 sites with low affinity. Stabilization of the RyR1-FKBP12 complex by S107 depended on the redox state of RyR1. In homogenates, GSSG-and NO-mediated dissociation of FKBP12 occurred in the presence of a relatively high myoplasmic FKBP12 concentration, which suggests that other to be identified factors may be involved in destabilization of the FKBP12-RyR1 complex in skeletal muscle. Reduction of NOC12-mediated release of FKBP12 by S107 in homogenates is of interest, because NO-mediated RyR1 S-nitrosylation triggers the release of Ca 2+ in brain [55] and skeletal muscle [20,28,56] under physiological and pathological conditions. Effects of FKBP12 and S107 on [ 3 H]ryanodine binding. SR vesicles treated and not treated with FK506 were incubated at 24uC for the indicated times in the above buffer containing the indicated concentrations of FKBP12 and S107. Data show that in the absence of added FKBP12, S107 did not alter [ 3 H]ryanodine binding to SR vesicles with (2FK506) or freed (+FK506) of FKBP12, whereas S107 significantly decreased [ 3 H]ryanodine binding to FKBP12-depleted vesicles in the presence of FKBP12 after preincubation for 10 min and 30 min. Data are the mean 6 SEM of 4-6 experiments. *p,0.05 compared to vesicles minus FK506 at 10 and 30 min, respectively. # p,0.05 compared to FKBP12-depleated vesicles plus added FKBP12 and minus S107 at 10 and 30 min, respectively. (TIF) Figure S5 Single channel measurements with NOC12treated RyR1s. (A) FK506-treated SR vesicles were incubated for 30 min at 24uC with 0.10 mM NOC12 in 0.3 M sucrose, 0.25 M KCl, 20 mM imidazole, pH 7.0 without (top trace), with 44 mMS107 (trace 2), 5 mM FKBP12 (trace 3) or 44 mM S107 plus 5 mM FKBP12 (bottom trace). Vesicles were then fused to a lipid bilayer and recorded at 2 mM cis cytoplasmic Ca 2+ and 235 mV as described in Materials and Methods. Representative single channel recordings (downward deflections from closed levels, c-) (left) and current histograms (right) are shown. (B) Single channel data were obtained as described in A. Mean channel open probabilities (Po) were greater in the presence of NOC12 (this figure) than absence of NOC12 (Fig. 8) by 11% in absence of S107 and FKBP12, by 11% in presence of S107, by 3% in presence of FKBP12, and 35% in presence of S107 and FKBP12. Data are the mean 6 SEM of 5-8 single channel recordings. * p,0.05 compared to RyR1s not incubated with S107 and FKBP12, # p,0.05 compared to RyR1s incubated with FKBP12 in absence of S107. Significance of differences of data was analyzed with Student's t-test. (TIF) Author Contributions