Interaction between Ryanodine and Neomycin Binding Sites on Ca 2 (cid:49) Release Channel from Skeletal Muscle Sarcoplasmic Reticulum*

Neomycin is a potent inhibitor of skeletal muscle sar- coplasmic reticulum (SR) calcium release. To elucidate the mechanism of inhibition, the effects of neomycin on the binding of [ 3 H]ryanodine to the Ca 2 (cid:49) release channel and on its channel activity when reconstituted into pla- nar lipid bilayer were examined. Equilibrium binding of [ 3 H]ryanodine was partially inhibited by neomycin. Inhibition was incomplete at high neomycin concentra- tions, indicating noncompetitive inhibition rather than direct competitive inhibition. Neomycin and [ 3 H]ryano-dine can bind to the channel simultaneously and, if [ 3 H]ryanodine is bound first, the addition of neomycin will slow the dissociation of [ 3 H]ryanodine from the high affinity site. Neomycin also slows the association of [ 3 H]ryanodine with the high affinity binding site. The neomycin binding site, therefore, appears to be distinct from the ryanodine binding site. Dissociation of [ 3 H]ry-anodine from trypsin-treated membranes or from a sol- ubilized 14 S complex is also slowed by neomycin. This complex is composed of polypeptides derived from the carboxyl terminus of the Ca 2 (cid:49) release channel after Arg-4475 the incubated Unlabeled ryanodine (cid:109) (cid:109) g trypsin incubated for 1 at 37 The proteolysis stopped by the addition of 1 mg of soybean trypsin inhibitor. Membranes solubilized in 2% 30 at 4 and centrifuged 83,000 (cid:51) g for 30 to remove insoluble material. The supernatant was layered onto 34-ml 5–20% sucrose gradient containing 0.4% CHAPS, M NaCl, M MOPS, and centrifuged at 110,000 g for 18 50 fractions were collected, and the fractions containing the peak [ 3 H]ry- anodine were pooled, diluted to 50 m M NaCl, and applied to a 25-ml DEAE-trisacryl column and eluted with 100 m M NaCl, 0.4% CHAPS, 20 m M MOPS, pH 7.4. Following a second sucrose gradient, the peak fractions were pooled and used for binding and bilayer studies. The 14 S complex was then incorporated into liposomes formed by sonication with 5 mg/ml phosphatidylcholine. The channel activity of this 14 S [ 3 H]ryanodine-labeled Ca 2 (cid:49) release channel proteolytic frag-ment the bilayer techniques described and the known values for K d1 and L , the binding constant for [ 3 H]ry- anodine in the presence of inhibitor, K d2 , can be derived. These equations will also yield the values for K i1 and K i2 , the dissociation constants of the inhibitor in the absence and presence of [ 3 H]ryanodine, respectively.

The Ca 2ϩ release channel in the terminal cisternae of skeletal muscle allows the movement of Ca 2ϩ from the lumen of the sarcoplasmic reticulum (SR) 1 to the cytoplasm in response to a signal from the transverse tubule (2). The protein that forms this channel can be activated by the binding of the plant alka-loid ryanodine at a high affinity site (3)(4)(5). The apparent affinity of [ 3 H]ryanodine for binding to the Ca 2ϩ release channel is dependent upon the functional state of the channel, and changes in binding of this ligand can be used to analyze and monitor the effects of modulators of Ca 2ϩ release channel function (6,7). High affinity ryanodine binding sites are located in the protein between Arg-4475 and the carboxyl terminus (1).
Neomycin, a polycationic, aminoglycoside antibiotic (8), inhibits Ca 2ϩ release and blocks [ 3 H]ryanodine binding to the SR membranes (9 -13). Wyskovsky et al. (13) reported that neomycin only blocks the fast component of the release while ruthenium red completely blocks the Ca 2ϩ efflux from SR vesicles, suggesting that inhibition by neomycin has a mechanism different from ruthenium red. The exact mechanism by which neomycin inhibits channel activity is unclear. Based on the effects of neomycin on the binding of [ 3 H]ryanodine to SR membranes and on 45 Ca 2ϩ fluxes, Mack et al. (10) concluded that neomycin was competitive with ryanodine for high affinity binding sites. Furthermore, they suggested that neomycin at high concentrations slows the dissociation of [ 3 H]ryanodine from the high affinity site by an allosteric mechanism. In the present work, we examine the effects of neomycin on [ 3 H]ryanodine binding and on the behavior of the Ca 2ϩ release channel incorporated into planar lipid bilayers to demonstrate that ryanodine and neomycin bind noncompetitively to the channel. 3 H]Ryanodine (61.5 Ci/mmol) was purchased from DuPont NEN. Ryanodine was obtained from Calbiochem. Neomycin sulfate was obtained from Sigma. Phosphatidylethanolamine (bovine heart) and phosphatidylserine (bovine brain) were obtained from Avanti Polar Lipids, Inc.

Materials-[
Sarcoplasmic Reticulum (SR) Membrane Preparation-SR membranes were prepared from rabbit back and hindleg skeletal muscle and were purified by using sucrose gradient centrifugation as described elsewhere (14,15). Protein was estimated by the method of Lowry et al. (16), using BSA as standard. H]ryanodine is indicated in figure legends. SR membranes were then proteolyzed for 1 h at 37°C by trypsin with a trypsin/protein ratio of 1:500. Proteolysis was stopped with soybean trypsin inhibitor at a trypsin/ soybean trypsin inhibitor ratio of 1:10. Trypsinized SR membranes were pelleted by centrifugation in a Beckman Airfuge for 5 min at 30 p.s.i., then solubilized by resuspension in 2% digitonin, 0.3 M KCl, 100 M Ca 2ϩ , 20 mM MOPS, pH 7.4, followed by incubation for 30 min at 4°C. The solubilized membrane proteins were partially purified by sedimentation through a 5-20% linear sucrose gradient in Beckman SW28 rotor for 18 h at 24,000 rpm. Fractions (20 drops) were collected from the bottom of the sucrose gradient and assayed for bound [ 3 H]ryanodine as described (1,18).

Trypsin Digestion SR Membranes and Purification of Solubilized
Purification of the 14 S Complex of Proteolyzed SR Ca 2ϩ Release Channel-50 mg of sarcoplasmic reticulum membranes were pelleted at * This work is supported by grants from the Muscular Dystrophy Association and National Institutes of Health Grant AR41802 (to S. L. H.). The 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.
‡ Supported by Research Career Development Award Grant NS01618.
186,000 ϫ g for 30 min to remove residual protease inhibitors and then the pellets were resuspended in 5 ml of 150 mM NaCl, 100 M CaCl 2 , 50 mM MOPS, pH 7.4. To trace label the receptor to follow its purification, [ 3 H]ryanodine was added to a final concentration of 12 nM, and the membranes were incubated for 1 h at 37°C. Unlabeled ryanodine was then added to 1 M. 100 g of trypsin was added, and the membranes were incubated for 1 h at 37°C. The proteolysis was stopped by the addition of 1 mg of soybean trypsin inhibitor. Membranes were solubilized in 2% CHAPS for 30 min at 4°C and centrifuged at 83,000 ϫ g for 30 min to remove insoluble material. The supernatant was layered onto a 34-ml 5-20% sucrose gradient containing 0.4% CHAPS, 0.3 M NaCl, 50 mM MOPS, pH 7.4, and centrifuged at 110,000 ϫ g for 18 h. 50 drop fractions were collected, and the fractions containing the peak [ 3 H]ryanodine counts were pooled, diluted to 50 mM NaCl, and applied to a 25-ml DEAE-trisacryl column and eluted with 100 mM NaCl, 0.4% CHAPS, 20 mM MOPS, pH 7.4. Following a second sucrose gradient, the peak fractions were pooled and used for binding and bilayer studies. The 14 S complex was then incorporated into liposomes formed by sonication with 5 mg/ml phosphatidylcholine. The channel activity of this 14 S [ 3 H]ryanodine-labeled Ca 2ϩ release channel proteolytic fragment was determined by the bilayer techniques described below.
Equilibrium at 37°C for 2 h, then proteolyzed by trypsin (1:500 trypsin:SR membranes) for 1 h at 37°C. The membranes were solubilized in 2% digitonin, and the 14 S complex was isolated using sucrose gradient centrifugation. Aliquots (440 l) of the 14 S complex were diluted into 10 ml of Buffer I in the presence and absence of 20 M neomycin. Dissociation was performed at room temperature, and aliquots (0.4 ml) of diluted sample were added to 200 l of ice-cold Buffer I containing rabbit ␥-globulin (5 mg/ml), BSA (5 mg/ml), and 10% polyethylene glycol (PEG) 8000. After 15 min of incubation at 4°C, the samples were filtered and washed with 5 ϫ 5 ml of ice-cold wash buffer containing 10% polyethylene glycol.
Bilayer Techniques-Planar bilayers consisting of 1:1 L-␣-phosphatidylethanolamine and L-␣-phosphatidylserine were formed following the Mueller-Rudin procedure across a 100-m diameter aperture in teflon cups as described previously (17). The mixture of phospholipids was dissolved in n-decane (Sigma) at a concentration of 25 mg/ml. Both chambers were filled with buffer solution (25 mM Cs 2 SO 4 , 10 mM MOPS, and 8 M CaCl 2 , pH 7.4). After bilayer formation, 5 l of SR membranes were added to the cis chamber to give a final protein concentration of 0.5 mg/ml. The other side of the bilayer was defined as trans. An osmotic gradient was formed between the cis and trans chambers by adding a concentrated salt solution (1 M Cs 2 SO 4 ) to the cis chamber. Recording solutions contained 225 mM Cs 2 SO 4 in the cis chamber and 25 mM Cs 2 SO 4 in the trans chamber. All subsequent additions were made to the cis chamber. Agar/KCl bridges were used to connect the chambers to Ag/AgCl electrodes immersed in 2 M KCl. Holding potential was ϩ40 mV. The data were filtered at 2.5 kHz and digitized at 10 kHz. Analysis was made using pCLAMP software: programs CLAMPEX, FETCHAN, and pSTAT (Axon Instruments, Inc). To facilitate incorporation and detection of channels from the 14 S complex, K ϩ was used as a current carrier, and all solutions contained KCl at the same concentrations used for Cs 2 SO 4 .
Data Analysis-[ 3 H]Ryanodine binding data were analyzed with nonlinear curve-fitting using Sigma Plot (Jandel Scientific). Nonspecific binding was subtracted prior to analysis.
Association kinetics: where B t ϭ bound ligand at time t, t ϭ time after addition of [ 3 H]ryanodine, n ϭ number of components, A i ϭ amount of ligand bound to component i, a i ϭ 1/k obs , n Յ 3. Dissociation kinetics: where B t ϭ bound ligand at time t, n ϭ number of components, Competitive interaction between neomycin and ryanodine at the high affinity binding site.
MODEL 2. Noncompetitive binding of neomycin allosterically alters the high affinity ryanodine binding site.

Inhibition of [ 3 H]ryanodine equilibrium binding by neomycin was
analyzed by nonlinear least squares fitting to the following equation: where B I is the observed binding in the presence of the inhibitor concentration I, B ϱ is the bound [ 3 H]ryanodine concentration at maximal inhibition, A is the inhibitable binding and is equal to B o Ϫ B ϱ , and B o is the bound [ 3 H]ryanodine at zero inhibitor concentration. In Model 1, for competitive binding, B ϱ equals nonspecific binding, and and L ϭ free Thus where L is the [ 3 H]ryanodine concentration and K d1 is the equilibrium dissociation constant for [ 3 H]ryanodine and B T is the total binding site concentration. Likewise Thus, from the values of B o , B ϱ , and K app (determined by curve-fitting), and the known values for K d1 and L, the binding constant for [ 3 H]ryanodine in the presence of inhibitor, K d2 , can be derived. These equations will also yield the values for K i1 and K i2 , the dissociation constants of the inhibitor in the absence and presence of [ 3 H]ryanodine, respectively.

RESULTS
Neomycin is a potent inhibitor of [ 3 H]ryanodine binding and of Ca 2ϩ release channel activity (10 -13). It has been proposed that neomycin blocks channel activation by ryanodine by competitively binding at the same site as ryanodine (10,11). To investigate this mechanism more closely, the effects of neomycin on [ 3 H]ryanodine binding were examined in detail. The While the data of Fig. 1, A and C, are inconsistent with simple competitive binding, the data can be analyzed in terms of a cyclic model for allosteric, noncompetitive inhibition (see "Experimental Procedures"). The data of Fig. 1A were well fit by Equation 3 as shown by the solid lines. From the values of B o , B ϱ , and K app obtained by nonlinear least squares fitting and the affinity of [ 3 H]ryanodine (14 nM) and using the equations described under "Experimental Procedures," the binding affinity for neomycin was calculated to be 300 nM and 1100 nM in the absence and presence of ryanodine, respectively (Table I). Likewise, the affinity of [ 3 H]ryanodine in the presence of high concentrations of neomycin was calculated to be 50 nM, consistent with the value of 46 nM determined by direct binding isotherm in the presence of 10 M neomycin in Fig. 1C.
To further explore the interaction between ryanodine binding and neomycin binding, the kinetics of association and dissociation in the presence of [ 3 H]ryanodine were examined. If there is no effect on association rate, then the binding cannot be competitive. Changes in the rate of association can, however, occur with either a competitive or a noncompetitive mechanism. The association of [ 3 H]ryanodine (Fig. 2) is characterized by a single component (k 1 ϭ 0.0025 min Ϫ1 M Ϫ1 ). Neomycin decreases the apparent rate of association to 0.00041, a 6.1-fold effect. The predicted dissociation rate constant from the k obs plot is 0.011 min Ϫ1 and does not appear to change significantly in the presence of neomycin.
Neomycin also slows the dissociation rate of bound [ 3 H]ryanodine from the site (Fig. 3), a finding consistent only with a noncompetitive interaction between the neomycin and the ryanodine binding sites. In the absence of neomycin, the dissociation data were fit with 3 descending exponential components (k Ϫ1 ϭ 0.013 min Ϫ1 , k Ϫ2 ϭ 0.0026 min Ϫ1 , and k Ϫ3 ϭ 0.00045 min Ϫ1 ). The relative amounts of each of these components varied only slightly among membrane preparations. In 12 prep- association kinetics (k obs plot, Fig. 2). The effect of neomycin concentration on the relative amounts of the three components is shown in Fig. 4. At low concentrations of neomycin, the fast component appears to be converted to the intermediate component, and, at higher neomycin concentrations, the intermediate is converted to the slow component.
To investigate the mechanism by which neomycin blocks the channel, we examined its effects on single channels reconstituted into planar lipid bilayers. The effect of neomycin on the Ca 2ϩ release channel and the ryanodine-modified Ca 2ϩ release channel reconstituted into planar lipid bilayers is shown in  (n ϭ 3). The channel opens only to the ryanodine-modified conductance level in the presence of both ryanodine and neomycin (Fig. 5, c-e), supporting a model in which the two ligands can bind simultaneously to the channel.
These data again support a noncompetitive interaction between neomycin and ryanodine.
Previous work from this laboratory has shown that both high and low affinity ryanodine binding sites are located in a 14 S proteolytic complex composed only of polypeptides derived from the carboxyl terminus after Arg-4475 in the primary sequence of the Ca 2ϩ release channel (1). In SR membranes (not solubi-   Fig. 3 at the indicated concentrations of neomycin in the dilution buffer. Each set of dissociation data were fit to a three-component exponential decay as described in Fig. 3 and under "Experimental Procedures." The counts/min corresponding to the fast (q), intermediate (f), and slow (å) components at each neomycin concentration are shown. lized), proteolysis with trypsin under the conditions which generate the 14 S complex does not significantly alter the rate of dissociation of [ 3 H]ryanodine (bound prior to proteolysis) nor does it alter the ability of neomycin to slow dissociation (data not shown), suggesting that the neomycin and ryanodine binding sites are still capable of interacting after trypsin treatment. However, proteolyzed SR membranes may also contain other proteolytic fragments that contribute to the neomycin effect on ryanodine dissociation.

FIG. 4. The effect of neomycin on the 3 components of [ 3 H]ryanodine dissociation. Dissociation kinetics of [ 3 H]ryanodine was performed as described in
To determine whether the neomycin binding site is on the 14 S complex, we generated the 30 S (intact) and the 14 S (proteolyzed) solubilized and purified forms of the protein, each containing bound [ 3 H]ryanodine. The sucrose gradient profile is shown in Fig. 6. We then determined the effect of neomycin on the rate of dissociation of [ 3 H]ryanodine from the purified 30 S and 14 S complexes (Fig. 7, A and B). Similar to the results obtained from SR membranes, 20 M neomycin slows [ 3 H]ryanodine dissociation from both the 30 S intact Ca 2ϩ release channel (Fig. 7A) and 14 S complex (Fig. 7B). Dissociation of [ 3 H]ryanodine from the detergent-solubilized and gradient-purified ryanodine receptor is characterized by two components (k Ϫ1 ϭ 0.016 min Ϫ1 , k Ϫ2 ϭ 0.0063 min Ϫ1 ), both of which are faster than those observed in membranes (Fig. 3). Dissociation from 14 S is even faster and is best fit by 3 components (k Ϫ1 ϭ 0.026 min Ϫ1 , k Ϫ2 ϭ 0.011 min Ϫ1 , k Ϫ3 ϭ 0.0004 min Ϫ1 ).
The question of whether the 14 S complex still behaves as an ion channel was addressed by reconstitution of the 14 S complex into planar lipid bilayers. To obtain a higher purity for the 14 S complex, the fraction purified by a sucrose gradient and the DEAE-column was further purified on a second sucrose gradient. The gradient profile is shown in Fig. 8A, and the polypeptides from fractions 10 through 24 are shown on a silver-stained gel in Fig. 8B. A Western blot with an antibody to the last 9 amino acids is shown in Fig. 8C trin) (c), 37 kDa (derived from N terminus of 76 kDa and recognized by an antibody to residues 4670 -4685) (d), and 27 kDa (beginning with 4756 and containing the carboxyl terminus) (e). In this preparation, the identity of the 76-kDa, the 56-kDa, and the 27-kDa peptides were confirmed by aminoterminal sequencing. All other bands were identified in Western blots as described previously. The 14 S complex isolated with [ 3 H]ryanodine bound to the high affinity site was reconstituted into planar lipid bilayers, and the channel activity was monitored. In Fig. 9, 3 channels were incorporated, each with a conductance of 570 pS in 225 mM KCl (Fig. 9A). Neomycin caused the channel to close more frequently (Fig. 9B). Similar channels were seen in 6 different 14 S preparations and in a total of 8 trials. DISCUSSION The data presented here demonstrate that the channel-forming regions of the channel, the ryanodine binding sites, and the neomycin binding sites are all located between Arg-4475 and the carboxyl terminus of the Ca 2ϩ release channel, but that the neomycin binding site is distinct from that of ryanodine. Neomycin inhibits [ 3 H]ryanodine binding and the activity of both the intact channel and the ryanodine-modified channel in a manner consistent with neomycin binding being noncompetitive with ryanodine binding.
To demonstrate clearly the existence of a distinct binding site for neomycin, it was necessary to show that the binding was inconsistent with competitive binding to the ryanodine binding sites. Competitive inhibition can be distinguished from noncompetitive binding by analysis of equilibrium binding and by kinetic experiments. Since noncompetitive inhibition is mediated by allosteric effects, the result is a decrease in binding affinity with concomitant changes in the association rate, dissociation rate, or both. The change in affinity is saturable with inhibitor concentration as the inhibitor site becomes fully occupied. Competitive binding should reveal a change in apparent affinity that is not saturable with increasing inhibitor concentration and have no changes in the dissociation rate. Neither competitive nor noncompetitive inhibition will display a change in the number of binding sites.  The data in Fig. 1 clearly show incomplete inhibition of [ 3 H]ryanodine binding by neomycin, which is inconsistent with competitive inhibition but fully consistent with noncompetitive inhibition. Thus, the effect of neomycin is to change the affinity of ryanodine about 4-fold while ryanodine has a reciprocal 4-fold effect on the affinity of neomycin. The affinity of neomycin for the ryanodine receptor is about 300 nM. The kinetic data further support noncompetitive inhibition of [ 3 H]ryanodine binding by neomycin. At a high concentration of 100 M, neomycin decreases the association rate of [ 3 H]ryanodine dramatically. (If neomycin was competitive with ryanodine, the association rate should decrease proportionally with the predicted occupancy of neomycin. With a K d ϭ 300 nm the association rate should have decreased about 300-fold but actually changed only 6.1-fold.) More definitive is the analysis of the dissociation rates. A competitive inhibitor should not affect the dissociation rate under conditions where there is no rebinding of ligand. However, neomycin clearly inhibits dissociation of [ 3 H]ryanodine and alters the proportion of the three distinct rates of dissociation toward the slower components. Slowing of both association and dissociation rates is consistent with allosteric effects. Since the effect on the association rate constant is stronger, the net result is a decrease in ryanodine binding affinity at equilibrium. Inhibition of [ 3 H]ryanodine binding by neomycin was interpreted by Mack et al. (10) as competitive. However, their data are not discrepant with the data presented here and are fully consistent with an allosteric model for noncompetitive inhibition. They also display data demonstrating slower dissociation of [ 3 H]ryanodine in the presence of high concentrations of neomycin, a result incompatible with a simple competitive mechanism.
Further support for a model wherein neomycin binds a site distinct from the ryanodine binding site is obtained from functional assays of channel activity in planar lipid bilayers. Ryanodine alone promotes long open states with a lower conductance than seen in the absence of ryanodine. The further addition of neomycin produces frequent fast closings: the mean open time is decreased. However, the channel opens to the conductance level seen in the presence of ryanodine, never to the higher conductance level characteristic of the unbound channel. If neomycin acted by competitive displacement of ryanodine, the channel would be expected to occasionally reopen to the unmodified level and this is not observed. Ryanodine, thus, appears to remain bound in the presence of neomycin inhibition of channel activity. The effect of neomycin on the affinity of [ 3 H]ryanodine does not directly account for inhibition of the ryanodine-modified channel. The mechanism of neomycin inhibition may be through stabilization of a closed conformation or by direct channel block.
Neomycin slows the dissociation of [ 3 H]ryanodine from the purified Ca 2ϩ release channel (30 S) and a 14 S complex which we have previously shown to be composed of peptides derived from the carboxyl terminus after Arg-4475 (1). The ryanodinemodified 14 S complex purified after trypsin digestion forms a channel in the bilayer, and this activity is inhibited by neomycin. This is consistent with the slowing of the dissociation of [ 3 H]ryanodine from its binding site on the 14 S complex by neomycin. It is extremely difficult to eliminate the possibility that a minor contaminant of this preparation is forming the ion channels. However, these data taken together with the binding data support a model in which the channel-forming portion of the protein is localized in a complex of a 76-kDa peptide fragment which is the part of the protein between amino acid 4476 and the carboxyl terminus of Ca 2ϩ release channel. This same region contains both high and low affinity ryanodine binding sites (1) and part or all of the putative transmembrane domains of the Ca 2ϩ release channel.
In summary, neomycin inhibits Ca 2ϩ release and noncompetitively inhibits [ 3 H]ryanodine binding sites by an allosteric mechanism. The neomycin binding sites as well as the high and low affinity ryanodine binding sites are located in a peptide region encompassing the amino acid sequence from Arg-4475 to the carboxyl terminus.