Disulfide Linkage of Biotin Identifies a 106-kDa Ca2+ Release Channel in Sarcoplasmic Reticulum*

Reactive disulfide reagents (RDSs) with a biotin moiety have been synthesized and found to cause Ca2+ release from sarcoplasmic reticulum (SR) vesicles. The RDSs oxidize SH sites on SR proteins via a thiol-disulfide exchange, with the formation of mixed disul- fide bonds between SR proteins and biotin. Biotinylated RDSs identified a 106-kDa protein which was purified by biotin-avidin chromatography. Disulfide reducing agents, like dithiothreitol, reverse the effect of RDSs and thus promoted active re-uptake of Ca2+ and dissociated biotin from the labeled protein indicat-ing that biotin was covalently linked to the 106-kDa protein via a disulfide bond. Several lines of evidence indicate that this protein is not Ca2+,Mg2+-ATPase and is not a proteolytic fragment or a subunit of the 400- kDa Ca2+-ryanodine receptor complex (RRC). Monoclonal antibodies against the ATPase did not cross- react with the 106-kDa protein, and polyclonal antibodies against the 106-kDa did not cross-react with either the ATPase or the 400-kDa RRC. RDSs did not label the 400-kDa RRC with biotin. Linear sucrose Diagnostics after we had initially synthesized the product. Several experiments with the com-mercial product (99% pure) produced qualitatively similar results. However, PDP-biotin hydrazide was the superior cross-linking reagent for SH sites.

costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Recipient of a Muscular Dystrophy Association Postdoctoral Fellowship.
ll Established Investigator 84130 of the American Heart Association.
§ Supported by Physician Investigator Training Grant 5T32-DK07458 from the National Institutes of Health. **Recipient of Research Career Development Award 5 KO4 NS00909 from the National Institutes of Health. To whom correspondence should be addressed. to the 400-kDa "feet" proteins.
In the previous article, we have described a class of "reactive" disulfide (RDS)' compounds (i.e. dithiopyridines) that cause Ca2+ release from sarcoplasmic reticulum (SR) vesicles (1). The RDSs caused release by oxidizing critical sulfhydryl groups on SR proteins through a thiol-disulfide exchange reaction and the formation of mixed disulfides between the SR protein(s) and the RDS compounds. The oxidation reaction opened a Ca2+ channel pathway which was reversed by reducing the mixed disulfide bond with GSH or DTT, resulting in active Ca2+ re-uptake by SR Ca2+ pumps (1). Among the dithiopyridines that were tested, SPDP (N-succinimidyl 3-(2-pyridyldithio)propionate) is a heterobifunctional reagent for the thiolation and production of intermolecular conjugates. It provides an elegant method to covalently link proteins with an easily identifiable probe, facilitating isolation and purification of proteins involved in sulfhydryl-activated Ca2+ release.
In the present study, two methods were used to synthesize SPDP-biotin conjugates: PDP-biocytin and PDP-biotin hydrazide. Both were effective at low concentrations (10-20 p M ) in causing SR Ca2+ release, with characteristics similar to those described for RDSs (1). The RDS-biotin conjugates labeled an SR protein which was identified by biotin-avidin peroxidase reaction. The biotinylated protein isolated and purified by biotin-avidin chromatography had an apparent molecular mass of 106,000 daltons, did not cross-react with monoclonal antibodies to the Ca",M$'-ATPase, and comprised about 0.3% of total SR protein. Immunological evidence indicated that the 106-kDa is neither a fragment nor a subunit of the 400-kDa Ca2' ryanodine receptor complex (RRC) (2-6). Incorporation of purified 106-kDa protein in lipid bilayers revealed a cationic channel with a large Na+ conductance and exhibited three additional subconductance states. Like the RCC, the sulfhydryl-activated 106-kDa channel was activated by micromolar [Ca"] or millimolar [ATP] and inhibited by micromolar ruthenium red or millimolar [M$+Ifr,,. Preliminary reports of these studies have been presented (7,8). Preparation of SR Vesicles-Heavy fraction of SR vesicles were prepared from rabbit white skeletal muscle as described previously (9). After the final centrifugation step, the vesicles were suspended at 10-20 mg/ml in 0.9 M sucrose, 10 mM HEPES, at pH 7.0, and stored in liquid nitrogen until used. Junctional SR vesicles with enriched 400-kDa RRC were prepared in the presence of DIFP (1 mM) and EGTA (2 mM) as previously described (10). Protein determinations were made by the method of Lowry et al. (11).
Measurement of Ca2+ Transport-Ca" fluxes across SR vesicles were measured through the differential absorption changes of the metallochromic indicator antipyrylazo I11 (AP 111), at 720-790 nm with a time-sharing dual-wavelength spectrophotometer. The use of AP 111 to measure extravesicular free [Ca'+] in suspensions of SR and the standard controls showing its specificity for Ca2+ have been demonstrated for similar types of experiments (9,12). Uptake and release of Ca2+ was measured with SR (0.5 mg/ml) suspended in a medium containing (in mM) 100 KC1, 20 Tris-HC1, 0.5 MgCl', 0.1 mM AP 111, pH 6.80, a t 23 "C. Two aliquots of Ca2+ (25 p~) were added to calibrate the dye's response, then the addition of 50 or 100 p~ ATP along with an ATP-regenerating system consisting of creatine phosphate (5 mM) and creatine kinase (5 units). Upon completion of uptake sulfhydryl oxidizing agents were added to induce Ca2+ efflux. After the phase of efflux, the ionophore A 23187 was added to release any residual Ca'+ and determine the maximum Ca2+ that could be released by the vesicles. Control experiments ensured that the reagents used to induce release did not interfere with the measurements of free Ca2+ with AP 111, nor did they directly react with the Ca'+-ATPase to alter uptake.
Thiopyridone Production-The production of thiopyridone associated with Ca'+ release from heavy or junctional SR elicited by a RDS (i.e. SPDP-biotin conjugates) was measured in parallel experiments through the differential absorption changes a t 340-310 nm, as previously described (1).
Synthesis of SPDP-Biotin Conjugates-Two SPDP-biotin conjugates were synthesized by the scheme shown in Fig. 1. Equimolar concentrations (10 mM) of SPDP and biotin hydrazide or biocytin were mixed in dimethyl sulfoxide and allowed to react at room temperature for 4 h. Coualent Labeling of S R Proteins with Biotin-PDP-biotin hydrazide and PDP-biocytin were used to covalently link biotin to SR proteins. SR vesicles (2 mg/ml) were incubated with an excess of an SPDP-biotin conjugate (100 p~) for 5 min in a medium consisting of (in mM) 100 NaC1, 20 Tris-HC1, 1.0 MgC12, pH 5.0, at room temper- ature. The suspension WAS diluted by 25-fold with ice-cold medium, at pH 7.0, lacking the SPDP-biotin conjugate. The vesicles were then washed by centrifugation (20,000 rpm X 60 min, in a Sorvall SS-34 pended in the original volume of 100 mM NaCl, 20 mM Tris-HC1, 1.0 rotor) to remove unreacted sulfhydryl reagent. The pellet was resus-mM MgCL, at pH 7, and washed for a second time (20,000 rpm X 60 min, in an SS-34 rotor). The labeling reaction of SR proteins with biotin was carried out in weakly acidic media, pH 5, to increase the reaction rate of thiol modification by thiol-disulfide exchange (13). However, labeling in weakly acidic media was not critical since biotinylation of SR proteins was not significantly reduced in more alkaline media (pH 7, 8, and 9, data not shown). The strategy for labeling SR proteins was based on the time course for SR Ca2+ release induced by RDS reagents followed by the interruption of the labeling reaction to minimize nonspecific biotinylation of SR proteins.
Identification of Biotinylated S R Proteins-Biotinylated SR vesicles (2 mg/ml) were suspended in SDS sample buffer containing 40% glycerol, 4% SDS, 0.004% bromphenol blue, 0.2 M Tris-HC1, at pH 6.8, and boiled for 2 min. Note that the SDS sample buffer for electrophoresis did not contain the standard sulfhydryl reducing agents. Samples (150-200 pg protein/lane for Western blots and 15 pg/lane for silver staining) were electrophoresed on 5-15% linear gradients of polyacrylamide slab gels in buffer containing: 0.1% SDS, 0.2 M glycine, 20 mM Tris-HC1, pH 6.8, a t 6.5 mA for 16 h (14). The gels were either stained with silver (15) or the proteins were electrophoretically transferred from polyacrylamide gel to nitrocellulose sheet for 45 min a t 24 V using the method of Towbin et al. (16). Nitrocellulose sheets were incubated with "blocking" buffer (10% horse serum, 1% bovine serum albumin in 50 mM Tris-HC1, 150 mM NaCl, pH 7.2 (Tris-buffered saline)) to prevent nonspecific binding of avidin. The sheets were then incubated with blocking buffer containing avidin linked to horseradish peroxidase diluted 1:lOOO in above blocking buffer. Western blots were washed with Tris-buffered saline, and bound peroxidase was visualized by staining with 4-chloro-1-naphthol and hydrogen peroxide. An identical gel was stained with silver to detect the total SR protein. Control experiments were carried out in the presence of the reducing agent DTT (5 mM) to demonstrate that the biotin label was covalently linked to SR proteins by a disulfide bond.
Isolation of Biotin-labeled SR Proteins by Biotin-auidin Chromatography-Biotinylated SR vesicles were solubilized in 1 M NaC1, 25 mM NaPIPES (or Tris maleate), 1.6% CHAPS, pH 7.1, a t 4 "C, for 2 h with continuous shaking. CHAPS (at 1% in 1 M NaC1) was found to efficiently solubilize the biotin-tagged SR proteins. The detergentsolubilized proteins were centrifuged (100,000 X g for 60 min) and the supernatant incubated with avidin-Affi-Gel-10 for 2 h at 4 "C. Avidin-Affi-Gel-10 beads were prepared by incubating avidin with affi-Gel-10 overnight at 4 "C in 100 mM MOPS, pH 7.5. The avidinconjugated bead were packed in a small column. The non-biotinylated proteins which did not bind to the column were washed with the same CHAPS-containing buffer used to solubilize SR proteins. Biotinylated proteins bound to the avidin-beads were selectively eluted with buffer containing sulfhydryl-reducing agent DTT (5 mM). The proteins eluted with DTT were no longer tagged with biotin and were identified on silver gels. Alternatively, all SR proteins bound to avidin linked to Affi-Gel-10 beads were extracted by boiling the beads in SDS sample buffer. In this case, SR proteins retained the covalently linked biotin and thus could be detected by avidin-horseradish peroxidase after Western blotting. Both procedures purified a single protein of 106-kDa apparent molecular mass. Scheme I summarizes the purification protocol of SR proteins containing the critical sulfhydryl site.
Production of Polyclonal Antibodies against the 106 kDa-Polyclonal antibodies were raised against the biotin-avidin-purified 106-kDa protein was previously described (17). Briefly, one New Zealand rabbit was immunized (intramuscular) with 20 pg (106 kDa) emulsified in Freund's complete adjuvant. After 3 weeks, immunization was repeated with antigen in Freund's incomplete adjuvant. Additional boosts were given intravenously in saline without adjuvant, four times at the interval of 3 days. In a separate experiment, the 106-kDa antigen was conjugated to keyhole limpet hemocyanin using glutaraldehyde to increase its immunogenicity (17). The same immunization protocol was used as with unconjugated protein. One week after the last boost, serum was collected and tested for the presence of antibodies. For immunoblot analysis total SR proteins were fractionated by SDS-PAGE and transferred to nitrocellulose. These Western blots were incubated either with (i) immune serum containing 106- kDa antibody (1:250 dilution in blocking buffer), (ii) preimmune serum (1:250 dilution in blocking buffer), or (iii) monoclonal antibody directed against Ca2+,Mg2'-ATPase (1:lOOO dilution in blocking buffer) for 4, 4, and 1 h, respectively. Western blots were washed for 1 h with Tris-buffered saline, then incubated with goat anti-rabbit IgG-HRP for anti-106 kDa or goat anti-mouse IgG-horseradish peroxidase, for Ca2',Mg2'-ATPase antibody. Both horseradish peroxidase-conjugated antibodies were used a t dilutions of 1:lOOO. After six washes with Tris-buffered saline, bound peroxidase-linked antibodies were visually detected after reacting with 4-chloro-1-naphthol and hydrogen peroxide.
Fractionation of SR Proteins Using Linear Sucrose Gradients-SR proteins (6 mg/ml) were solubilized with CHAPS as described above, and the supernatant after 100,000 X g for 60 min was loaded (0.3 ml) on 3-15% linear sucrose gradient in buffer containing (12 ml/tube): 1 M NaC1, 40 mM Tris-maleate, 0.9% CHAPS, with or without phospholipids (PC/PE/PS = 16:3:1 a t a concentration of 4 mg/ml), and centrifuged a t 30,000 rpm for 17 h at 4 "C in a Beckman SW 41 rotor as previously described (18). In other experiments, CHAPS solubilized junctional SR proteins were loaded on 5-20% linear sucrose gradients with buffer containing 1 M NaC1, 25 mM NaPIPES, 150 p~ CaCl,, 10 p~ EGTA, 1 mM DIFP, 2 mM ATP, pH 7.1, and were centrifuged in an SW 41 rotor a t 30,000 rpm for 17 h a t 4 "C as later modified by Lai et al. (19). After centrifugation, the gradients were separated into 16-17 fractions to analyze their protein distribution on 5-15% SDS-polyacrylamide gels by Coomassie Blue or silver staining.
Identification of Ca2+,Mg2+-ATPase with a Monoclonal Antibody-Biotin-avidin chromatography identified a single protein with an apparent molecular mass -106 kDa, as determined by SDS-PAGE. This protein represented approximately 0.3% of total SR protein, a polypeptide with an apparent molecular mass of -106 kDa as determined by SDS-PAGE. Since the bulk of SR proteins consists of Ca'+ pump proteins (-100 kDa), a low concentration of a protein with approximately the same molecular mass could be readily misconstrued as Ca'+,Mg*'-ATPase contamination. A monoclonal antibody to the Ca" ATPase, the generous gift of Dr. Kevin Campbell (University of Iowa) was used to distinguish the two proteins. Protein samples to be tested were first run on SDS-polyacrylamide slab gels and analyzed by Western blotting.
Reconstitution of 106-kDa Protein in Lipid Bilayer-Purified protein at 0.1-0.4 pg/ml was added to both sides of a Mueller-Rudin type bilayer (phosphatidylethanolamine/phosphatidylserine/phosphatidylcholine = 5:3:2; total lipid concentration = 50 mg/ml in decane) formed across an 80 or 150 pm hole drilled in a polystyrene cup, manufactured at the machine shop of the Department of Physiology (University of Pittsburgh). The maximum CHAPS concentration used in these measurements (64 wg/ml) was far less than the concentration that affected bilayer conductance in the absence of added protein. An Axopatch 1C (Axon Instruments) amplifier with a CV-3B headstage was used to amplify picoampere current fluctuations. Data were digitized (Instrutech Corp. model VR-10) stored on a video recorder and subsequently analyzed for channel activity. Analog data output from the Instrutech was digitized with an analog to digital converted (Labmaster TM-40, Scientific Solutions, Solon, OH) and analyzed using pCLAMP (a software package from Axon Instruments, Burlingame, CA).
Materials-All reagents were of analytical grade. CHAPS was purchased from Boehringer Mannheim. PIPES, HEPES, Tris, ATP, creatine phosphate, creatine kinase DTT, GSH, and EGTA, SDS-PAGE molecular weight standards, biocytin, and avidin were purchased from Sigma. AP I11 was obtained from ICN Pharmaceuticals (Plainview, NY). The ionophore A 23187 was from Behring Diagnostics; SPDP and biotin hydrazide were from Pierce. Horseradish peroxidase-avidin D was purchased from Vector Laboratories (Burlingame, CA). Affi-Gel-10 was obtained from Bio-Rad goat antimouse and anti-rabbit IgG-linked horseradish peroxidase was purchased from Organon Teknika Cappal (Malvern, PA). Lipids were purchased from Avanti Polar Lipids, Inc. (Birmingham, AL). PDPbiocytin became available from Behring Diagnostics after we had initially synthesized the product. Several experiments with the commercial product (99% pure) produced qualitatively similar results. However, PDP-biotin hydrazide was the superior cross-linking reagent for S H sites.

SPDP and SPDP-Biotin Conjugates Induce SR Ca2+
Release uiu an SH Oxidation Reaction-To test the effect of SPDP and its biotin conjugates on SR Ca2+ release, the vesicles were first loaded with Ca2+ using an ATP-regenerating system. SR vesicles (0.5 mg/ml) were suspended in 2 ml of 100 mM KC1, 20 mM Tris-HC1, 0.5 mM MgC12, 0.1 mM AP 111, pH 6.8, at room temperature. Aliquots of Ca2+ (25 p M ) were sequentially added to calibrate the differential absorption changes of AP 111, then Mg-ATP (50 p~) along with the ATP-regenerating system consisting of creatine phosphate (5 mM) and creatine kinase (5 units). As shown in Fig. 2, after the phase of Ca2+ uptake (not shown), various concentrations of sulfhydryl reagents were added to cause Ca'+ release. After the phase of Ca2+ release, DTT was added to reduce the disulfide bond and induce active re-uptake of Ca2+ (Fig. 2 , A and B) then the ionophore A 23187 was added to determine the maximum Ca2+ that could be released by the vesicles. The newly synthesized SPDP-biotin conjugates were more potent at releasing Ca2+ than the parent compound SPDP, and PDP-biotin hydrazide was more effective than PDP-biocytin (Fig. 2). Control experiments indicated that additions of DTT fully reversed the effects of the biotinylated sulfhydryl reagents and when added alone, the two biotin compounds (i.e. biotin hydrazide and biocytin) did not cause Ca2+ release nor did they alter active Ca2+ uptake at the concentration used in these experiments. In principle, the lower the concentration of sulfhydryl reagent that causes rapid Ca2+ release from the SR, the more likely it is to oxidize specifically the critical SH sites involved in this Ca2+ release pathway, with fewer interactions with other accessible SH groups that are not involved in Ca2+ release. An enhanced selectivity of reactive RDS could be demonstrated by preincubating SR vesicles with other sulfhydryl reagents such as iodoacetic acid or iodoacetamide (IAM) which were found to have no significant effect on either Ca2+ uptake or release, at up to 2 mM concentrations. The presence of IAM, which is known to alkylate S H sites not involved in Ca2+ release, significantly enhanced the potency of PDP-biocytin at causing Ca2+ release, from 20 p~ in the absence of any other SH reagents to 5 p~ in the presence  determined how many of these SH sites are directly associated with Ca2+-release channel proteins and the number of oxidized SH/Ca2+ release channel is not known. Because PDP-biotin hydrazide stimulated SR Ca2+ release at faster rates and lower concentrations than PDP-biocytin, the experiments described below were focused on labeling SR proteins with PDP-biotin hydrazide.
PDP-Biotin Hydrazide Labels a Single SR Protein -106 kDa-Six different SR preparations were biotinylated with PDP-biotin hydrazide as described under "Experimental Procedures." Protein samples were run on an SDS-gel (200 pg/ lane), transferred to a nitrocelluse sheet, incubated in avidinhorseradish peroxidase, and developed to visualize the distribution of biotinylated proteins. In Fig. 5A, lanes 1-6 show that the sulfhydryl cross-linking reagent consistently identified a single SR protein which migrated with an apparent molecular mass of 106. The lane labeled DTT shows that pretreatment of the SR protein with DTT (5 mM) removed the biotin label from the 106-kDa protein. This supports the hypothesis that biotin was cross-linked to the 106-kDa protein via a disulfide bond. Fig. 5B shows a silver-stained gel of whole SR proteins from four of these preparations. The gel shows an abundance of Ca2+,Mg2+-ATPase, and high molecular weight Ca2+-ryanodine receptor complex, mass -400 kDa (top band on lanes 1-4). The data show that PDP-biotin hydrazide (Fig. 5A) does not label the 400-kDa ryanodine receptor complex which is abundant in this preparation. The labeled 106-kDa band comigrates with the Ca*+,M$+-ATPase but is not as broad as the ATPase band in the accompanying silver gels (Fig. 5B). The addition or deletion of the protease inhibitor DIFP (1 mM) did not alter the results shown in Fig.  5.
Isolation of Biotinylated Proteins-Biotinylated proteins were isolated by biotin-avidin affinity chromatography as described in Scheme 1. Biotinylated SR vesicles were solubilized in CHAPS, centrifuged, and the supernatant containing all soluble SR proteins was incubated with avidin Affi-Gel-10 beads to bind biotinylated proteins selectively to the beads. The avidin beads were washed with buffer to elute nonbiotinylated proteins, then washed with DTT to dissociate the proteins linked to biotin via a disulfide bond and to collect the SR proteins containing the reactive sulfhydryl. Fig. 6A shows a typical result of avidin affinity chromatography. The DTT-eluted fraction ( B ) contained a prominent band at 106 kDa which appeared identical to the biotinylated 106-kDa band seen in Fig. 5A. Since DTT elution cleaved the disulfide bond linking biotin to protein, this protein was no longer biotinylated but was detected on silver gels (Fig. 6A). To verify that the 106-kDa protein was retained on the avidin beads because it was biotinylated rather than because of a nonspecific interaction with the beads, proteins bound to the beads were eluted with boiling SDS sample buffer which denatures avidin on the column and releases SR proteins still attached to biotin. Fig. 6B shows the outcome of extracting SR proteins from avidin beads by boiling SDS instead of DTT and detecting biotin-tagged proteins by Western transfers. In lanes B (Fig. 6, A and B), the 106-kDa protein is clearly the most prominent band with no visible protein bands at high molecular mass, i.e. 400 kDa. The extraction of proteins with boiling SDS (Fig. 6B, lane B ) instead of DTT also yielded a doublet of minor proteins migrating between 69 and 92 kDa which could be due to a partial breakdown of the 106 kDa shown in Fig. 6A, lane B. The results show that biotin-avidin chromatography can be used to isolate specifically a highly pure form of 106-kDa protein (Fig. 6A, lane B ) and that high molecular mass proteins (i.e. 400-450-kDa ryanodine receptors) are not biotinylated nor purified by biotin-avidin chro-

45.
2s -  DTT ( l a n e B ) . B, Western blot developed to detect biotinylated proteins. Lane W , whole biotinylated S R proteins, only 106-kDa protein is detected even though the lane is heavily loaded as shown on silver stained gels (see A and W ) . Lane B, after washing the avidin beads with buffer, they were boiled in SDS sample buffer to extract all proteins non-covalently bound to the column. Lane S, the run-through of biotinylated proteins showing that the avidin beads selectively bound the biotin- matography. Protein assays carried out after extensive removal of DTT and CHAPS from the samples in dialysis indicated that 200-350 pg of 106-kDa protein was recovered/ 100 mg of total SR, typically the recovery of 106 kDa was about 0.3% of total SR protein. This ratio of 106 kDa/mg SR was most likely underestimated since (a) the labeling phase of SR protein channels with biotin was kept brief to avoid biotinylation of nonspecific proteins, (b) the elution of biotinylated proteins from the avidin column is not 100% efficient, and (c) the dialysis of protein samples extracted from the biotin-avidin column was necessary for accurate protein assays but the procedure resulted in a substantial loss of proteins.
Immunological Analysis Indicates That the 106-kDa Protein Is Different from the Ca2+,M&+-ATPase and the 400-kDa Feet Proteins-The 106-kDa protein containing the reactive sulfhydryl was isolated by biotin-avidin chromatography as in Fig. 6A (lane B), then used as an antigen to immunize two rabbits and raise polyclonal antibodies. One rabbit received 106 kDa; the other received 106 kDa conjugated with keyhole limpet hemocyanin in an attempt to enhance its immunogenicity. Immune serum from both rabbits contained anti-106 kDa antibodies which cross-reacted with biotin-avidin-purified 106 kDa whereas serum from control (nonimmunized) rabbits did not. To examine specificity of the antisera, sera from both rabbits were tested by Western blot analysis of whole SR protein. As shown in Fig. 7A, two heavy SR vesicles preparations from different rabbits (lanes 1 and 2) were run on Western blots and probed with the anti-106 kDa anti-sera. Only one moon-shaped band of protein cross-reacted; no cross-reaction with high molecular weight proteins or the bulk pg/lane. The anti-106 kDa anti-sera cross-reacted with biotinpurified proteins which on Western blots had an apparent molecular mass of 106 kDa. However, in Fig. 7A, the antisera cross-react with a band below 92 kDa. The migration of biotin-avidin-purifed 106-kDa protein was always between 100 and 110 kDa on Western blots as well as silver gels. But in the presence of high contents of Ca2+,Mg2"ATPase, the 106 kDa (detected by anti-106 antibody on Western blots, or avidin-horseradish peroxidase) migrated either above or just below the ATPase depending on gel conditions ( i e . pH or level of oxidation of protein samples). Fig. 7B shows a Western blot transfer (200 pg/lane) of the same two SR preparations (lanes 1 and 2, respectively as in Fig. 7A) probed with a monoclonal antibody against Ca2+,Mg2"ATPase. The anti-ATPase monoclonal binds to a major bulky band of protein at -80 kDa and also detects ATPase migrating at high molecular masses. The appearance of high molecular weight cross-reactive material detected by the anti-ATPase is caused by the high concentrations of proteins placed on these gels (20). The latter gels were purposely loaded with high concentrations of SR proteins to compare the immunoreactivity of anti-106 and anti-ATPase, under identical conditions. The Western blot shown in Fig. 7C demonstrates that the ATPase band appears at -100 kDa, and the smearing artifact disappears as the content of SR proteins loaded in the gels is systematically reduced from 100, 10, 5, 2, 1.5, 1.0, to 0.5 pg/ lane (Fig. 7C, lanes 1-7, respectively). The content of ATPase should be about 60% of total SR, or 0.3 pg in lane 7 where it was readily detected (immunoreactions on Western blots can detect as little as 0.1 pg of protein/band (21)).
To reach valid conclusions from such immunoaffinity analysis, it was important ot verify that all SR proteins were efficiently transferred on the Western blots. In Fig. 8 (left), the same preparations of SR proteins used in Fig. 7 were fractionated by SDS-PAGE, transferred to nitrocellulose sheets and stained with Amido black. Even with relatively insensitive Amido black stain, the 400-kDa protein was readily detected when the total protein load was as low as 50 pg, (Fig. 8, left lane 3). The same SR proteins analyzed by silverstained SDS-PAGE show the relatively high content of 400-kDa protein, when 10, 7.5, or 5 pg proteins were loaded in lanes 1, , and 3, respectively (Fig. 8, right).

Incorporation of 106-kDa Proteins into Planar Bilayers-
The sulfhydryl activated 106-kDa protein was isolated by biotin-avidin chromatography as shown in Fig. 6A and incorporated into a bilayer lipid membrane. In the presence of symmetrical 0.5 M NaCl on the ckltrans sides of the membrane, large current fluctuations were measured as a function of transmembrane potential. Channel activity was strongly activated by ATP (1 mM) and inhibited by ruthenium red (Fig. 9, A-C). Little activity was observed until ATP was added to activate the channel (in eight bilayers). The frequency distribution of current fluctuations indicated the presence of a single high conductance channel with a half-subconductance state (Fig. 9D). Analysis of current to voltage relationship of 15 bilayer experiments indicated the presence of cationic channels with a Na+ conductance, gNa+ = 375 & 15 In order to determine the cation versus anion selectivity of this channel a 5:l KC1 gradient was formed across the bilayer, and the single channel current fluctuations were measured as a function of applied voltage (Fig. 10). At the KC1 concentrations used in this experiment, the ratio of the activities of K+ and C1-across the membrane was equal to 3.6. The theoretical Nernst revesal potential for a channel perfectly selective for PS.

FIG. 8. Efficiency of SR protein transfers on nitrocellulose.
Left, total SR proteins were separated using SDS-polyacrylamide gel electrophoresis, electrophoretically transferred to a nitrocellulose sheet which was stained with Amido black to compare the relative efficiency of transfer of low and high molecular weight proteins. Under these conditions for Western transfer, large proteins (including 400 kDa) are efficiently transferred from gel to Western blot. Lanes [1][2][3] contained 200, 100, and 50 pg of total SR proteins, respectively. Right, silver staining of total SR proteins; lanes 1-3 represent 10,7.5, and 5 pg of total SR proteins.  K+ over C1would then = -32.5 mV. The measured reversal potential was -20 f 1 mV corresponding to a permeability ratio, PK+/PCLof 5. Under these experimental conditions, the channel's ionic conductance was 69 f 1 pS. Again, addition of 10 p~ ruthenium red inhibited channel activity, completely.
The Ca2+ conductance of the 106-kDa channel was measured in asymmetrical Ca2+ solutions as previously described by Lai et al. (19). The 106-kDa protein was incorporated in bilayers with asymmetrical Ca2' solutions on both sides of the chamber: trans, 53 mM Ca(OH)2/250 mM HEPES, pH 7.4; cis, 125 mM Tris, 250 mM HEPES, 20 pM CaC12, 20 p M EGTA, pH 7.4. As shown in Fig. 11, current fluctuations were activated by adding to the cis side either 2 mM ATP (Fig. 11,  traces a and b) or -80 p~ free [Ca"] (Fig. 11, traces c and d ) .
The current to voltage relationship analyzed from three bilayer experiments indicated the presence of Ca2+ channels with a maximum conductance, gca2+ = 107.7 f 12 pS (Fig. 12). In view of reports identifying the 400 kDa feet proteins as the Ca" release channel in SR and the similarity in the channel characteristics of the 400-and 106-kDa proteins, the procedure used to purify 400-kDa feet proteins was re-examined to determine if the 106-kDa copurified with 400-kDa proteins.
Fractionation of SR Proteins by Linear Sucrose Gradients-SR vesicles were solubilized in CHAPS and fractionated on linear sucrose gradients. Fig. 13 shows an analysis of the various sucrose fractions from 3 to 15% (lanes 1-17, respectively) using SDS-PAGE and staining with silver. As previously described (6,17), the 400-kDa ryanodine receptor tends to migrate to high sucrose densities (lanes 13-15). However, in the heavily loaded silver stained gels, the same lanes contained a second high molecular mass band (-350 kDa), a band at about 150 kDa, and a dark band of protein in the range of 97-116 kDa. Lanes 16 and 17 are devoid of 400-kDa bands and contain a single protein band in the 97-116-kDa molecular mass range. A key question is whether the protein bands (-100 kDa in lanes 13-15) comigrating with the 400 kDa consist of sulfhydryl-activated 106-kDa Ca2+ channel  Transmembrane Potential (mV) FIG. 12. Ca2+ conductance of 106-kDa Ca2+ release channel protein incorporated in planar bilayer. The 106-kDa channel protein was incorporated in a planar bilayer in asymmetrical Ca2+ solutions as described for Fig. 1. The current-to-voltage relationship was plotted from single channel current recorded at various holding potentials ranging from -40 to +50 mV. A linear regression algorithm was used to obtain the best fit through from -40 to +50 mV. A linear regression algorithm was used to obtain the best fit through these data points. The most probable open state of the channel had a conductance gcaz+ = 107.7 2 12 pS. The current fluctuations also revealed three subconductances states. The next most probable openstate of the channel had a conductance, gaz+ = 45 * 10 pS. Reversal Dotential was eaual to + 20 mV, such that gC.z+/PTd. was approximately equal to 4. linear sucrose gradient and separated in 17 fractions of equal volume but of increasing sucrose density. Each fraction was analyzed on SDS gels followed by silver staining as shown in lanes 1-1 7, respectively. The RRC appears primarily on lanes 13-15 and 106-kDa protein appears in lanes 13-17. Lanes 1-6 received 7 pl from fractions 1-6 of the linear sucrose gradient and lanes 7-17 received 45 pl from fractions 7-17 of the linear gradient. protein or Ca2+,Mg2+-ATPase. The issue was addressed in the following way. First, samples from lanes 3, 7, 14, and 17 were run on SDS gels, transferred by Western blot techniques to nitrocellulose sheets, and cross-reacted with monoclonal antibody raised against Ca2',Mg2"ATPase. As shown in Fig. 14,  lanes 3 and 7, are rich in ATPase, but the low molecular mass bands (-100 kDa) in lane 14 and 17 did not significantly cross-react with anti-ATPase monoclonal antibody. The concentrations of protein layered in lanes 3, 7, 14, and 17 were 19,8, 1.25, and 0.16 pg of protein, respectively. Moreover, the anti-ATPase monoclonal can readily detect as little as 0.1 pg of ATPase protein on Western blots which strongly implies that the low molecular mass bands in lanes 14 and 17 are primarily non-ATPase proteins. In spot tests on nitrocellulose sheets, both anti-106-kDa polyclonals were found to crossreact with proteins from lanes 14 and 17 which indicated the presence of 106-kDa channel protein comigrating with the feet proteins in lane 14. Protein in lane 17 (0.15 pg/2 ml) which did not contain 400-kDa protein was added to the cis side of a planar bilayer and upon incorporation exhibited Na' current fluctuations (Fig. 15) and a Na' conductance of about 400 pS, under the conditions described for Fig. 9. The analysis of SR proteins by linear sucrose gradients (as in Figs. 13 and 14) were reproduced with eight separate SR preparations. In each case, the vesicles were prepared with careful attention to maintaining low temperatures (4 C) and in the presence of the protease inhibitor DIFP (1 mM), which was replenished every 4 h to ensure its potency. The latter precautions protected the 400 kDa from proteolytic breakdown and did not alter the distributions of 106-kDa protein shown in Figs. 13  and 14.
On the other hand, other studies obtained highly purified 400 kDa feet proteins using the same methods but with junctional SR vesicles that are enriched with feet proteins (16,17). The experiments of Lai et al. (19) were reproduced and analyzed in SDS-PAGE stained with Coomassie Blue. As shown in Fig. 16, top, a single band of seemingly pure 400-kDa protein can be detected with no other proteins on the same lanes (11-14). It is important to note that this linear sucrose gradient was run with junctional SR, in the presence of ATP (2 mM), DIFP (1 mM), and free Ca2+ (150 pM CaCh SRproteins (1.25 mg/ml) were solubilized in 1.6% CHAPS, 1 M NaCI, 25 m M NaPIPES, 1 mM DIFP, pH 7.1, for 2 h at 20°C. The 100,000 X g supernatant was loaded onto 5-20% linear sucrose gradients with a total volume of 12 ml (see "Experimental Procedures") and centrifuged for 17 h at 30,000 X rpm its SW 41 rotor. At the end of the run 16 fractions (700 pl/each) were collected from each gradient. Fractions were analyzed using SDS-polyacrylamide gel electrophoresis and stained with Coomassie Blue ( A ) or silver staining ( E ) . For Coomassie Blue-stained gel, all the lanes contained 45 pl from each fraction of the linear gradient. For silver staining, lanes 1-6 contained 7 pl each and lanes 7-16 contained 45 pl of their respective fractions from the linear sucrose gradient. + 100 p~ EGTA) to protect feet proteins from proteolytic degradation, maximize the amount of purified RRC and its binding activity to ryanodine. However, when the same experiment is analyzed by silver staining, the same lanes (11- 14) are found to contain high molecular mass "doublets" and significant protein bands at about 150 and 100 kDa (Fig. 16,  bottom). The latter result was reproduced in several different ways with linear sucrose gradient (5-20%) fractionation of junctional SR vesicles with 1 mM DIFP (six times) or 0.1 pg/ ml of leupeptin, with (twice) or without (twice) lipids in the sucrose gradient. The appearance of protein bands other than a 400-kDa singlet was due to the higher amounts of protein loaded on the gels and the higher sensitivity of silver staining techniques. The isolation of the feet protein as shown in Fig.  16, top lanes 11-13, would otherwise "appear" to yield a superbly pure sample.

DISCUSSION
The present study shows that RDSs can be used to crosslink biotin to an SR protein containing the critical sulfhydryl site involved in the opening and closing of a Ca2+ channel pathway. Two compounds were synthesized, PDP-biotin hydrazide and PDP-biocytin which trigger Ca2+ release by oxidizing a free sulfhydryl on an SR protein. Conversely, the reduction of the newly formed disulfide bond between the SR protein and the biotin resulted in active re-uptake of Ca2+. Biotin-avidin chromatography made it possible to isolate a single 106-kDa protein associated with this critical and highly reactive sulfhydryl which is involved in a pathway for Ca2+ release. The sulfhydryl properties of the Ca2+ release pathway are important because they made it possible to separate and distinguish a protein which is merely 0.3% of total SR protein from the Ca2+,Mg2"ATPase which has a similar molecular mass and comprises about 60% of total SR proteins. Monoclonal antibody to the Ca2+,Mg2"ATPase did not cross-react with the 106-kDa and biotinylation of proteins using SPDPbiotin conjugates did not label either the Ca2+,Mg2"ATPase or the 400-kDa RRC. Isolation of the 400-kDa RRC by linear sucrose gradients shows that the 106-kDa migrates in even higher sucrose densities (lanes 16- 17, Fig. 13) and thus provided an alternate method to separate the 106 kDa. Incorporation of purified (>95% pure) 106 kDa in planar bilayers revealed the existence of a Ca2+ channel with large single unit conductance (-107 zk 12 pS) which is activated by free Ca2+ and ATP (1 mM) on the cis side and inhibited by Me.
A most important issue raised by this work is the relationship between the 400-kDa ryanodine receptor complex (18,19,(22)(23)(24) and the 106-kDa sulfhydryl-activatedprotein. Several possibilities come to mind (a) the 106-kDa may be a proteolytic fragment of the 400-kDa which contains the Ca2+ release channel and perhaps some of the site(s) involved in the regulation of the channel; (b) it may be a subunit of the 400-kDa protein; (c) it may be an entirely different SR channel protein; or (d) the RRC is composed of high molecular mass "bridging" proteins associated with 106-kDa protein(s).
Despite the large and impressive body of evidence identifying the feet proteins as the Ca2+ ryanodine receptor complex and the physiological Ca2+ release channel, the present study questions the level of purification of 400-kDa proteins incorporated in planar bilayers. The purification of 400-kDa proteins by linear sucrose gradients and its identification as the Ca2+ release channel is primarily based on the correlation of radiolabeled ['Hlryanodine peak with the migration of the 400 kDa (19)(20)(21)(22). However, there are other proteins at about 100 kDa that comigrate with the 400 kDa. These "contaminating" proteins are not seen in Coomassie Blue gels (Fig. 16) or even silver gels unless heavily loaded with protein. The general consensus in the literature is that feet proteins are the high affinity ryanodine receptors which can be purified by linear sucrose density gradients. However, at least two reports suggest that the situation may be more complex. First, Pessah et al. (25) noted that proteins in the C & 3 detergentinsoluble fractions that could be solubilized with CHAPS consisted primarily of the high molecular mass ryanodine receptor and 100-kDa proteins, possibly Ca2+,Mg2+-ATPase. The 100-kDa proteins could be removed from the CHAPSsoluble protein fractions ( i e . primarily ryanodine receptor) by prior treatment with increasing concentrations of C & I . However, as the content of 100-kDa protein was systematically reduced, ['Hlryanodine binding to the high molecular mass receptor site decreased and was abolished as the content of 100-kDa proteins (presumably Ca2+, Mg2"ATPase) was removed (25). At low CI2E9 to protein ratios (0.5 mg/mg A 106-kDa Ca2+ Release Channel Protein in Skeletal SR protein), the soluble fraction contained a high density of high affinity [3H]ryanodine-binding sites (18 pmol of ryanodinel mg SR, at 10 nM free ryanodine), yet there was no detectable high molecular mass protein in the silver geh2 The conclusion was that the "ATPase" and feet proteins were somehow associated, and the interaction was necessary for ryanodine binding (25). Second, Meissner et al. (26) used sucrose density purified 400-kDa proteins (by the method shown in Fig. 16, lanes [10][11][12][13][14] as antigens to produce polyclonal antibodies and thus follow the various fragments of 400-kDa proteins after mild tryptic digestion. The polyclonals thus produced crossreacted with total heavy SR proteins and as anticipated were found to correctly label 400-kDa bands. But the same polyclonals also cross-reacted with a "faint" band of 100-kDa protein (26). The latter band was faint suggesting that it was a minor protein, i.e. not the bulky ATPase band. In any case, these results indicate that the 400-kDa immunogen was not free of other proteins.
The evidence gathered in this study strongly suggests that the 106-kDa protein is not a proteolytic fragment and is not a subunit of the 400 kDa. The 106 kDa was isolated by the following two different procedures: ( a ) fractionation of SR proteins by linear sucrose gradients, using junctional SR prepared in the presence of EGTA (1 mM) and protease inhibitor (DIFP = 1 mM) to protect the 400-kDa RRC from proteolysis. The presence or absence of the latter agent did not cause a measurable change in the migration of 106 kDa. ( b ) Biotin-avidin chromatography in the presence of DIFP resulted in labeling and isolation of the 106 kDa but did not interact with the 400-kDa RRC. The implication of the latter experiment is that the sulfhydryl site involved in cross-linking biotin to the protein is located on the 106 but not the 400-kDa protein.
Alternatively, the possibility remained that the 106 kDa was a proteolytic fragment of the 400-kDa protein and the critical SH site is part of the 400-kDa protein but is sterically protected from oxidation by PDP-biotin hydrazide or PDPbiocytin, until proteolysis makes the site accessible. TO address the latter possibility, polyclonal antibodies were obtained against 106 kDa isolated by biotin-avidin chromatography by immunizing two rabbits. The antibodies from both rabbits did not cross-react with either the 400-kDa RRC nor the Ca2+,Mg2+-ATPase (Fig. 7A). These results indicate that (i) the 106 kDa is neither a fragment nor a subunit of the 400-kDa protein and that (ii) the original 106-kDa antigen used to raise these polyclonal antibodies consisted of a highly pure protein free of ATPase and 400-kDa feet proteins.
The purification and incorporation of the 106-kDa protein (isolated by biotin-avidin techniques) in planar bilayers indicate that it is a cationic channel with a large Na+ conductance of 375 f 15 pS (mean value +. S.E. from 15 bilayer measurements). The channel is activated by ATP and is inhibited by ruthenium red. Moreover, similar results have been obtained from 106-kDa proteins (Fig. 15) isolated by linear sucrose gradients (Fig. 13, lanes 16 and 17). With Ca2+ solutions in the cisltrans sides of the bilayer chamber, the biotin-avidin purified 106-kDa protein had a Ca2+ conductance gca2+ = 107.7 * 12 pS; it was activated by adding on the cis side 80 PM [CaZ+]free, ATP (1 mM) (Fig. 11), Ag+ (10 W ) or 2,2'-dithiodipyridine (20 PM) (not shown). The available data indicates that the 106-kDa and 400-kDa channels have (within experimental error) similar conductance values for Na+, K+, Ca2+, and C1-ions and respond in a similar fashion to modulators of Ca2+ release (27). Further bilayer studies are needed with careful perfusion of cis and trans sides of the I. N. Pessah, personal communication.
bilayers to fully characterize single channel properties and regulation by other agents that alter SR Ca2+ release.
The molecular masses of 106-kDa Ca2+ release channels and the Ca2+, Mg2"ATPase' pumps are almost identical. When biotinylated, the 106-kDa protein appears as a sharp band just above the ATPase and on gels heavily loaded with whole SR proteins, it migrates either above the ATPase or as a crescent moon just below the ATPase. Monoclonal antibody raised against Ca2+,M$+-ATPase did not cross-react with 106-kDa protein and polyclonal antibodies raised against 106-kDa protein did not cross-react with either the ATPase or the 400-kDa ryanodine-binding protein. The latter result was obtained with 106-kDa protein isolated by either biotin-avidin or linear sucrose gradients, as in Fig. 13 and indicates that Ca2+ pump and the 106-kDa protein are two different proteins. Because of the proximity in the apparent molecular mass of these two proteins and the overwhelming concentration of ATPase in SR preparations, the 106-kDa release channel could not be previously distinguished from the ATPase without the present biotin labeling technique and the production of polyclonal antibodies against the 106 kDa. Moreover, the presence of sulfhydryl-activated 106-kDa channel protein detected in the same sucrose density as the 400-kDa RRC could be readily misconstrued as ATPase contamination. Studies that depend on highly purified fractions of Ca2+, M P -A T Pase may require more stringent procedures to remove non-ATPase -100-kDa proteins from their preparations. For instance, Gould et al. (28) reconstituted presumably purified 100-kDa Ca2+ pumps in liposomes and reported that Ag+ ions induced Ca2+ release by acting at the Ca2+,Mg2+-ATPase. From such experiments, Gould et al. (29) further argued that both processes of uptake and release from SR vesicles could be solely attributed to the properties of the Caz+, M$+activated ATPase. However, the present identification of the 106-kDa Ca2+ release channel points out that such an interpretation cannot be justified without more sophisticated purification of the ATPase.
It is important to note that feet proteins have been purified and reconstituted in planar bilayers by numerous methods: (a) sucrose density gradients (18,19), ( b ) sequential column chromatography on heparin and hydroxylapatite columns (30,31), and ( c ) immunoaffinity chromatography (23). In addition, the recent cloning of the cDNA coding for the 5037 amino acids comprising the ryanodine receptor complex and its expression resulting in ryanodine binding activity (32), all comprise overwhelming evidence in favor of the feet proteins as the physiological site for SR Ca2+ release.
Nevertheless, the present data on 106-kDa Ca2+-release channel proteins indicate that SR Ca2+ release involves more than one pathway and suggests that linear sucrose gradients may not always separate feet proteins from the sulfhydrylactivated 106-kDa channels. It does not address the relationship between the 106-and 400-kDa proteins when feet proteins are purified by the other methods mentioned above. Certainly the understanding of the relationship between these two channels is far from complete and important questions remain regarding the role of the 106-kDa Ca2+ release channel protein, such as: ( a ) could the 106-kDa protein still be a fragment of the feet proteins ( i e . the Ca2+ channel region of the feet protein cloned by Takeshima et al. (32)) despite the lack of cross-reactivity between anti-106 antibodies and feet proteins? ( b ) Could there be more than one Ca2+ release channel in SR? (c) Since reactive disulfides can cause some release of Ca2+ from light SR vesicles, are 106-kDa proteins distributed in the longitudinal as well as terminal cisternae of the SR network?