Ruthenium Red Modifies the Cardiac and Skeletal Muscle Ca2+ Release Channels (Ryanodine Receptors) by Multiple Mechanisms*

The effects of ruthenium red (RR) on the skeletal and cardiac muscle ryanodine receptors (RyRs) were studied in vesicle-Ca2+ flux, [3H]ryanodine binding, and single channel measurements. In vesicle-Ca2+flux measurements, RR was more effective in inhibiting RyRs at 0.2 μm than 20 μm free Ca2+. [3H]Ryanodine binding measurements suggested noncompetitive interactions between RR inhibition and Ca2+regulatory sites of RyRs. In symmetric 0.25 m KCl with 10–20 μm cytosolic Ca2+, cytosolic RR decreased single channel activities at positive and negative holding potentials. In close to fully activated skeletal (20 μmCa2+ + 2 mm ATP) and cardiac (200 μm Ca2+) RyRs, cytosolic RR induced a predominant subconductance at a positive but not negative holding potential. Lumenal RR induced a major subconductance in cardiac RyR at negative but not positive holding potentials and several subconductances in skeletal RyR. The RR-related subconductances of cardiac RyR showed a nonlinear voltage dependence, and more than one RR molecule appeared to be involved in their formation. Cytosolic and lumenal RR also induced subconductances in Ca2+-conducting skeletal and cardiac RyRs recorded at 0 mV holding potential. These results suggest that RR inhibits RyRs and induces subconductances by binding to cytosolic and lumenal sites of skeletal and cardiac RyRs.

The release and sequestration of Ca 2ϩ ions by the sarcoplasmic reticulum (SR), 1 an intracellular membrane compartment, is essential to the process of cardiac and skeletal muscle contraction and relaxation. The rapid release of Ca 2ϩ is mediated by Ca 2ϩ release channels, also known as ryanodine receptors (RyRs), because they bind the plant alkaloid ryanodine with high affinity and specificity (1)(2)(3)(4)(5)(6). Skeletal and cardiac muscles express two major isoforms of the RyR, RyR1 and RyR2, respectively. In striated muscles, RyRs are concentrated in the junctional SR membrane near transverse tubular, voltage-sensitive L-type Ca 2ϩ channels (dihydropyridine receptors). A muscle action potential initiates dihydropyridine receptor con-formational changes that activate the RyRs via a direct physical interaction in skeletal muscle or mediate the influx of Ca 2ϩ in cardiac muscle, leading to the release of Ca 2ϩ from the SR and subsequent muscle contraction. Both RyRs have been isolated as 30 S protein complexes composed of four 560-kDa (RyR polypeptide) and four 12-kDa (FK506-binding protein) subunits (1)(2)(3)(4)(5). The two channel activities are affected by endogenous and exogenous effectors, such as Ca 2ϩ , Mg 2ϩ , ATP, caffeine, ryanodine, and ruthenium red.
Ruthenium red (RR) is one of the most potent inhibitors of SR Ca 2ϩ release (2). RR also inhibits mitochondrial Ca 2ϩ uptake. However, this activity was due to a contaminant (7) that may be related to an oxygen-bridged R360 complex that, at concentrations as high as 10 M, was without effect on SR Ca 2ϩ uptake or release (8). RR is a polycationic dye with a linear structure consisting of three ruthenium atoms with a net valence of 6. In SR vesicles, RR increased the rate of Ca 2ϩ uptake and decreased the rate of Ca 2ϩ release at concentrations ranging from 1 nM to 20 M (9 -18). In muscle fibers, RR concentrations greater than 20 M are required to inhibit SR Ca 2ϩ release because of RR binding to myoplasmic proteins (19,20). In intact single frog twitch muscle fibers, the estimated free RR concentration for half-block of SR Ca 2ϩ release was 2.4 M, in good agreement with the range reported for SR vesicle preparations (19). In [ 3 H]ryanodine binding measurements, RR decreased the B max value and increased the K D value, with the latter effect being due to a slower association rate (16). In single channel measurements, micromolar concentrations of RR decreased the channel open probability of the skeletal (21) and cardiac (22) RyRs by producing prolonged channel closings. A different effect was observed when ryanodine was used to lock the channels into a permanently open, Ca 2ϩ -insensitive subconductance state. RR blocked single ryanodine-modified skeletal RyRs by binding in a voltage-dependent manner to multiple sites located in the conductance pore of the channel (23).
The present study was undertaken to clarify the effects of RR on ryanodine-unmodified skeletal and cardiac RyRs. The effects of RR on RyRs were investigated in SR Ca 2ϩ uptake, [ 3 H]ryanodine binding, and single channel measurements using SR vesicles and purified cardiac and skeletal muscle RyRs. The results show that RR modifies the gating and conductance of the RyRs by multiple mechanisms, depending on channel activity, membrane potential, and sidedness of RR addition.

Materials-[ 3 H]Ryanodine was purchased from NEN Life Science
Products. Unlabeled ryanodine was obtained from Calbiochem (San Diego, CA), ruthenium red was from Fluka (Ronkonkoma, NY), and phospholipids were from Avanti Polar Lipids (Alabaster, AL). Ryanodine and ruthenium red were prepared as concentrated stock solutions in 0.25 M KCl, 20 mM KHepes, pH 7.4, before their use. All other chemicals were of analytical grade.
Preparation of SR Vesicles and Purification of RyRs-"Heavy" rabbit skeletal and canine cardiac muscle SR membrane fractions enriched in [ 3 H]ryanodine binding and Ca 2ϩ release channel activities were prepared in the presence of protease inhibitors (100 nM aprotinin, 1 M leupeptin, 1 M pepstatin, 1 mM benzamidine, 0.2 mM phenylmethylsulfonyl fluoride) as described (14,24). The 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS)-solubilized 30 S RyR complexes were isolated by rate density gradient centrifugation and reconstituted into proteoliposomes by removal of CHAPS by dialysis (25). 45 Ca 2ϩ Uptake Measurements-ATP-dependent 45 Ca 2ϩ uptake by SR vesicles was determined using a filtration method. Samples were incubated for 1 h at 24°C in 0.25 M KCl, 20 mM KHepes, pH 7.4, solutions containing 50 M Ca 2ϩ , 0.2 mM Pefabloc, and 20 M leupeptin in the absence of ryanodine or the presence of 300 M ryanodine to close the SR Ca 2ϩ release channel (26 -29). 45 where B is [ 3 H]ryanodine binding at a given [RR], and B o is the binding maximum in the absence of RR.
Single Channel Measurements and Analyses-Single channel measurements were performed by fusing proteoliposomes containing the purified RyRs with Mueller-Rudin type bilayers containing phosphatidylethanolamine, phosphatidylserine, and phosphatidylcholine in the ratio 5:3:2 (25 mg of total phospholipid per ml of n-decane) (30). The side of the bilayer to which the proteoliposomes were added was designated as the cis side. The trans side was defined as ground. Single channels were recorded in symmetric KCl buffer solutions (0.25 M KCl, 10 -20 mM KHepes, pH 7.4) with additions as indicated in the text. Unless otherwise indicated, electrical signals were filtered at 2-4 kHz, digitized at 10 -20 kHz, and analyzed as described (30). Data acquisition and analysis were performed using a commercially available software package (pClamp 6.0.3, Axon Instruments, Burlingame, CA) and an IBM-compatible computer (Pentium processor) with 12-bit analog/digital-digital/ analog converter (Digidata 1200, Axon Instruments).
In the absence of RR-related subconductance states, channel open probability in the absence of a substate (P o ) was obtained by setting the threshold level at 50% of the current amplitude between the closed (c) and open (o) states. P o values in multichannel recordings were calculated according to the equation P o ϭ ⌺ iP i /N, where N is the total number of channels, and P i is the i channel open probability. In some conditions, RR formed a predominant reduced conductance level (see Fig. 9A, right panel, bottom trace, s). RR formed additional subconductance states, but with some exceptions (see Figs. 6D and 7B), these occurred infrequently and were not further analyzed. The concentration and voltage dependence of the RR-related subconductances were obtained by determining the open probability of the full conductance events (P full ), RR-related subconductance events (P sub ), and the sum of P full and P sub (P tot ) (31). P full was obtained by setting the threshold level at 50% current amplitude between subconductance (s) and open (o) current levels (see Fig. 9A, right panel, bottom trace). This measurement eliminated inclusion of substates in P full . P tot was obtained by setting the threshold level at 50% of the current between the close (c) and subconductance (s) current levels (see Fig. 9A, right panel, bottom trace). The open probability of the subconductance state, P sub , was obtained by the equation P sub ϭ P tot Ϫ P full .
In the absence of substates, the Hill inhibition constant (K i ) and inhibition coefficient (n i ) of P o were determined by the following equation, where P o,max is P o in the absence of RR.
In the presence of substates, the Hill association constant (K a ) and coefficient (n a ) of P sub state were determined by the following equation.
Assuming a Boltzman distribution between the P full and P sub states, their voltage dependence was described by the following equation, where Z tot is the effective gating charge of the reaction of RR with the RyR. The other terms have their usual meanings. Determination of Free Ca 2ϩ Concentrations-Free Ca 2ϩ concentrations ը1 M were determined using a Ca 2ϩ selective electrode (Nico Scientific, Philadelphia, PA). Free Ca 2ϩ concentrations of Ͻ1 M were obtained by including in the solutions the appropriate amounts of Ca 2ϩ and EGTA as determined using the stability constants and the mixed solution program published by Schoenmakers et al. (32). 45 1A) and 20 M (Fig. 1B) free 45 Ca 2ϩ . The SR Ca 2ϩ pump is present in these vesicles, but only 70 -90% have RyR1 (24). To prevent release of sequestered 45 Ca 2ϩ , RyR1 was fully closed prior to 45 Ca 2ϩ uptake by incubation for 1 h with 300 M ryanodine (26 -29). In the absence of RR, pretreatment with ryanodine increased the levels of 45 Ca 2ϩ sequestered by the vesicles 3-fold in uptake medium containing 0.2 or 20 M free Ca 2ϩ . In ryanodine-untreated vesicles at 0.2 M Ca 2ϩ , 1 M RR doubled the amount of 45 Ca 2ϩ taken up by the vesicles. 10 M RR was nearly as effective as pretreatment with ryanodine in increasing 45 Ca 2ϩ uptake. By contrast, at 20 M Ca 2ϩ , 50 M RR was required to achieve Ca 2ϩ uptake levels approaching those observed in ryanodine-treated vesicles. RR did not increase 45 Ca 2ϩ uptake by vesicles pretreated with ryanodine. For ryanodine-treated vesicles, a small decrease in the 45 Ca 2ϩ uptake rate was observed in the presence of 50 M RR, in agreement with the suggestion that high levels of RR inhibit the SR Ca 2ϩ pump (33).

Effects of RR on
The effects of RR on Ca 2ϩ uptake levels of ryanodine-treated and -untreated skeletal and cardiac SR vesicles were compared. 45 Ca 2ϩ uptake levels were determined after an uptake period of 2 min. Cardiac SR contained a smaller proportion of vesicles responding to treatment with ryanodine ( Fig. 2), in accordance with a 4 -5-fold lower B max value of [ 3 H]ryanodine binding for cardiac than skeletal SR vesicles (see Fig. 5). Ryanodine significantly increased 45 Ca 2ϩ uptake or cardiac SR vesicles at 20 M Ca 2ϩ (p Ͻ 0.05; n ϭ 4) but not at 0.2 M Ca 2ϩ (p ϭ 0.1). In ryanodine-untreated SR vesicles, RR was more effective at 0.2 M than at 20 M Ca 2ϩ in raising 45 Ca 2ϩ uptake to levels in the ryanodine-treated vesicles. In agreement with previous vesicle ion flux measurements (9 -18), the results indicate that the skeletal and cardiac muscle RyRs are inhibited by RR. The data in Fig. 2 further show that the extent of inhibition depends on the Ca 2ϩ concentration of the uptake medium. Two possible reasons for this Ca 2ϩ dependence are that SR Ca 2ϩ pump activity is dependent on Ca 2ϩ concentration and/or that Ca 2ϩ ions compete with RR for the Ca 2ϩ regulatory sites on the RyRs.
Effects of RR on [ 3 H]Ryanodine Binding to Skeletal and Cardiac Muscle SR Vesicles-The highly specific plant alkaloid ryanodine was used to obtain information on the mechanism of RR inhibition of RyR1 and RyR2 independent of SR Ca 2ϩ pump activity. [ 3 H]Ryanodine binding is widely used as a probe of channel activity because of its preferential binding to open RyR ion channel states (1-5). Fig. 3 compares the Ca 2ϩ dependence of [ 3 H]ryanodine binding to skeletal and cardiac muscle SR vesicles in 0.25 M KCl solutions containing varying concentrations of RR. In the absence of RR, two bell-shaped activation/ inactivation curves were obtained. In agreement with previous reports (1)(2)(3)(4)(5), the binding data indicate that RyR1 and RyR2 are activated by Ca 2ϩ binding to cytosolic high affinity receptor sites and inhibited by Ca 2ϩ binding to low affinity sites, with higher Ca 2ϩ concentrations required to inhibit RyR2. Higher RR concentrations were required to inhibit [ 3 H]ryanodine binding to RyR2 than RyR1 (Fig. 3 RR Inhibits and Induces Subconductances in Single RyR1s-The kinetics of RR inhibition of the RyRs were further examined in single channel measurements. Proteoliposomes containing purified RyR were fused with planar lipid bilayers. A strong dependence of single channel activities on cis-Ca 2ϩ concentration indicated that the large cytosolic region of the channels faced the cis (cytosolic) chamber in a majority (Ͼ98%) of the recordings (34,35). Channels that could not be activated by 1-10 M Ca 2ϩ were discarded. The majority of channels were recorded in symmetric 0.25 M KCl solutions rather than with a lumenal Ca 2ϩ solution to eliminate large Ca 2ϩ gradients near the cytosolic channel pore sites (34). With K ϩ as the current carrier, single channel conductance was ϳ790 pS (30).  (Table II) compared with 90 nM (Table I) suggests that RR was more effective in inhibiting the skeletal channel in the single channel than [ 3 H]ryanodine binding experiments. Average Hill coefficients of 1.3-1.6 (Table II) suggest that RR inhibited single RyRs through a weak cooperative interaction.
RR was similarly effective in inhibiting single RyR1s recorded at 200 M rather than 20 M free cytosolic Ca 2ϩ (Fig. 6B and Table II). Again, no subconductances were evident before or after the addition of RR. A single skeletal muscle channel recorded in the presence of 20 M free cytosolic Ca 2ϩ and 2 mM  Table I.

TABLE I Effects of RR on [ 3 H]ryanodine binding to RyR1 and RyR2
Hill inhibition constants (K i ) and coefficients (n i ) were obtained as indicated in the legend to Fig. 4. Values are the mean Ϯ S.E. of the number of experiments shown in parentheses.

RyR1
RyR2 ATP was nearly fully activated in the absence of RR (Fig. 6C). The addition of 100 nM cytosolic RR resulted in several subconductances at ϩ40 mV but not Ϫ40 mV. The predominant subconductance corresponded to 45% of the full conductance at ϩ40 mV and was interrupted by frequent transitions to a closed state. The presence of lumenal RR induced three predominant subconductances, corresponding to 85, 55, and 25% of the full conductance (Fig. 6D). The three subconductances were observed at both Ϫ40 mV and ϩ40 mV, with the frequency of substate formation being higher at the negative holding potential (n ϭ 3). The data shown in Fig. 6, C and D, indicate that formation of subconductances depended on membrane potential. Cytosolic RR-related subconductances were observed at positive membrane potentials that favored cytosolic to lumenal cation fluxes.
In the presence of lumenal RR, subconductances were prefer- entially formed at negative membrane potentials that favored lumenal to cytosolic cation fluxes. The results raised the possibility that highly positively charged RR molecules enter the conductance pathway of RyR1, resulting in reduction of single channel currents.
The SR membrane is highly permeable to monovalent cations and anions, which suggests that the membrane potential across the SR in resting muscle is likely close to 0 mV (36). The effects of cytosolic and lumenal RR on single RyR1s were therefore also determined at 0 mV holding potential in a symmetric 0.25 M KCl solution with 10 mM lumenal Ca 2ϩ , with Ca 2ϩ being the conducting ion. In this condition, a Ca 2ϩ current could be measured, ranging from 2 pA (Fig. 7C) to 3.1 pA (Fig. 7, A and  B). In Fig. 7, A and B, single RyR1 channels were initially recorded under conditions similar to those in Fig. 6C, i.e. in the presence of 20 M free cytosolic Ca 2ϩ and 2 mM ATP. Both channels were nearly fully activated in the absence of RR (top traces). Addition of 250 nM cytosolic RR induced channel closings and a subconductance (P sub ϳ 0.1) corresponding to ϳ50% of the full conductance (Fig. 7A). In Fig. 7B, the presence of 300 nM lumenal RR induced, similar to the result shown in Fig. 6D, multiple subconductances corresponding to ϳ25, 60, and 75% of the full conductance. In Fig. 7C, a single RyR1 channel was nearly fully activated using conditions closely matching those in Figs. 1 and 4, i.e. 20 M free cytosolic Ca 2ϩ and 5 mM MgATP. The addition of 60 nM cytosolic RR induced an infrequent subconductance (P sub ϳ 0.02) corresponding to ϳ25% of the full conductance. RR concentrations greater than those shown in Fig. 7 resulted in complete channel closings in a majority of the single channel recordings. Taken together, the data shown in Fig. 7 show that RR induces subconductances in Ca 2ϩ -conducting RyR1 recorded at a membrane potential considered to be physiologically relevant.
RR Inhibits and Induces Subconductances in Single RyR2s-A more detailed kinetic analysis of the effects of RR on single RyR ion channels was carried out using RyR2. Six current traces of a single K ϩ conducting cardiac release channel are shown in Fig. 8. In the top traces (control), the channel was recorded in the presence of 10 M cytosolic Ca 2ϩ and 3-4 M lumenal Ca 2ϩ at holding potentials of Ϫ30 mV (left panel) and ϩ30 mV (right panel). The addition of 30 and 60 nM cytosolic RR decreased channel activity at both holding potentials by greater than 50 and 90%, respectively. As observed for RyR1 (Fig. 6A), RR formed long channel closings without inducing subconductances. Hill inhibition constants of 46 Ϯ 1 nM at ϩ30 mV and 43 Ϯ 9 nM at Ϫ30 mV (n ϭ 6) indicated a voltageindependent inhibition of channel activity (Table II)  was to increase the duration of the closed events.
The effects of cytosolic RR on RyR2 were also evaluated at a cytosolic [Ca 2ϩ ] of 200 M. At this Ca 2ϩ concentration the cardiac channel is close to maximally activated (Fig. 9A, top  panels). Addition of 60, 150, and 220 nM cytosolic RR induced a subconductance state at ϩ30 mV but not at Ϫ30 mV. The current amplitude histogram of Fig. 9B shows that the frequency but not current amplitude of the substate depended on RR concentration. The histogram shows open channel currents of ϳ24 pA. The addition of RR resulted in a major substate with a current of 8 -9 pA at ϩ30 mV and closed channel events (at 0 pA) at Ϫ30 and ϩ30 mV. Current-voltage relationships of the full and subconductance states in the absence and presence of 150 nM RR are shown in Fig. 9C. RR did not affect the full open conductance at negative and positive holding potentials. In contrast, the subconductance current-voltage relationship was nonlinear, with the subconductance corresponding to 28 Ϯ 5 and 12 Ϯ 2% of the full conductance at ϩ30 mV and ϩ60 mV, respectively (n ϭ 5). A decrease to 60 nM RR or increase to 220 nM RR (Fig. 9A) was without effect on the current-voltage relationship of the subconductance (not shown). The fractional increase in channel open probability of the substate (P sub /P tot ) at ϩ30 mV (Fig. 9A) could be fitted by Equation 3, which describes the O Ca N O sub transitions (Fig. 9D, solid line). The average Hill constant was 140 Ϯ 20 nM (Table II). A Hill coefficient of 1.6 Ϯ 0.1 suggested that more than one RR molecule was involved in forming the substate. RR also increased the rate of full channel closings (Fig. 9A). This effect was analyzed in single channel recordings obtained at negative holding potentials because these lacked subconductances, therefore simplifying the analysis. Broken lines in Fig. 9D were obtained by fitting P o to Equation 2. The average Hill constants of the full channel closings were 2-3-fold higher than the Hill constants of substate formation and ϳ10-fold higher than the Hill inhibition constants at 10 M Ca 2ϩ (Table II).
RR formed a subconductance in 8 out of 10 recordings at 200 M Ca 2ϩ and 1 out of 2 recordings at 5 mM Ca 2ϩ . In the remaining three recordings, RR decreased channel activity without inducing a subconductance state. Time analysis of channel recordings at 200 M Ca 2ϩ suggested differences in channel gating that appeared to be directly related to the ability of channels to form subconductances. Channels that formed subconductances gated more slowly before the addition of RR than channels that did not form subconductances (mean open time, 32 Ϯ 9 ms (n ϭ 8) versus 3 ms (n ϭ 2), respectively). This result suggests that slowly gating channels are more likely to form subconductances than rapidly gating channels.
In support of this idea, we found that subconductances were formed at positive membrane potentials when RyR2 was recorded in the presence of 10 M cytosolic free Ca 2ϩ , 5 mM ATP, and 300 nM RR (not shown). This condition in the absence of RR resulted in a nearly fully activated channel with long open times. Lumenal RR induced a major subconductance in RyR2 at negative membrane potentials (Fig. 10). A single Ca 2ϩ -activated RyR2 was recorded at Ϫ30 mV and ϩ30 mV in the presence of 10 M cytosolic Ca 2ϩ and increasing concentrations of lumenal RR. In the absence of RR (Fig. 10A, top traces), the channel was nearly fully activated. Addition of 150 -500 nM lumenal RR affected the channel (Fig. 10A) in a manner reminiscent of that of cytosolic RR in the presence of 200 M Ca 2ϩ (Fig. 9A), except that lumenal RR induced a major substate at the negative (Fig. 10A, left panel) instead of positive (right panel) holding potential. Fig. 10B shows IV curves of the full and subconductance states in the presence and absence of 150 nM lumenal RR. The presence of RR did not affect the full open conductance, whereas the lumenal RR-related subconductance current showed a slightly rectifying behavior with a conductance corresponding to 49 Ϯ 1 and 40 Ϯ 1% of the full conductance at Ϫ30 and Ϫ60 mV, respectively (Fig. 10B, n ϭ 5). An increase in cytosolic Ca 2ϩ from 10 to 200 M and a change in lumenal RR concentration were without effect on the currentvoltage relationship of the subconductance (not shown).
Lumenal RR induced a major subconductance with a lower affinity than cytosolic RR, as indicated by a Hill constant of 1700 Ϯ 700 nM for lumenal RR (Fig. 10C and Table II) as compared with 140 Ϯ 20 nM for cytosolic RR (Fig. 9D and Table  II). At ϩ30 mV in the presence of 300 and 500 nM RR, brief, poorly resolved channel closings were observed (Fig. 10A). Their occurrence suggested that RR also affected the channel at positive holding potentials, but unlike at Ϫ30 mV, the durations were too short to be fully resolved at the limited time resolution of the bilayer set-up (35). In addition, small, variable increases in full channel closings were observed at elevated lumenal RR (Fig. 10C and Table II).
Cytosolic RR was effective in inducing a subconductance at positive holding potentials (Fig. 9), whereas lumenal RR was effective at negative potentials (Fig. 10). The voltage depend-  Table II. ence of P sub /P full on holding potential was analyzed by the Boltzmann equation given under "Experimental Procedures" (Equation 4). A single RyR2 was recorded in the presence of 200 M cytosolic Ca 2ϩ and 60, 150, and 220 nM cytosolic RR. Three lines with similar slopes but different intercepts were obtained at the three RR concentrations (Fig. 11). The mean slope was 0.050 Ϯ 0.002, from which an effective gating charge (Z tot ) of 1.3 Ϯ 0.1 was obtained (Table II). P sub /P full of the lumenal RR-related substate of RyR2 obeyed Equation 4, with a Z tot value of 0.8 Ϯ 0.1 (Table II).
In Fig. 12, the effects of cytosolic and lumenal RR on single RyR2 channels were determined as in Fig. 7, i.e. at 0 mV with 10 mM lumenal Ca 2ϩ as the current carrier. Single channel conductance varied from 2 pA (Fig. 12D) to 3.4 pA (Fig. 12A). In Fig. 12, A and B, single RyR2 channels were recorded under conditions similar to those in Figs. 9 and 10, i.e. in the presence of 200 and 10 M free cytosolic Ca 2ϩ , respectively. Addition of 300 nM cytosolic (Fig. 12A) and 500 nM lumenal (Fig. 12B) RR induced subconductances (P sub Ͻ0.06) corresponding to ϳ25 and 35% of the full conductance, respectively. In Fig. 12, C and D, single RyR2 channels were recorded in the presence of 5 mM cytosolic MgATP and 20 or 200 M free Ca 2ϩ , i.e. under conditions similar to those in Figs. 2 and 4. In the presence of cytosolic RR, subconductances were observed in the 200 M Ca 2ϩ medium (Fig. 12D, P sub ϳ0.01) but rarely in the 20 M (Fig. 12C, P sub Ͻ 0.001) Ca 2ϩ medium. RR concentrations in excess of those in Fig. 12 resulted in complete channel closings. We conclude from these observations that RR induces subconductances in Ca 2ϩ -conducting RyR2 channels recorded at 0 mV.

DISCUSSION
In the present study, the effects of RR on the cardiac and skeletal muscle RyRs were investigated in vesicle Ca 2ϩ flux, [ 3 H]ryanodine binding, and single channel measurements. Scheme 1 describes the results.  Fig. 8 except that increasing RR concentrations were added to the lumenal instead of the cytosolic chamber of the bilayer apparatus. At negative holding potential a lumenal RR-related subconductance state (see left panels) became increasingly obvious as the RR concentration was increased, but the current of the subconductance did not change with RR concentrations. B, current-voltage relationships of full conductance state in the absence of RR (G) and full conductance (E) and subconductance (‚) states in the presence of 150 nM lumenal RR. Values are means Ϯ S.E. of five recordings (error bars are smaller than symbols). C, dependence of P o (G) (at ϩ30 mV) and P sub /P tot (OE) (at Ϫ30 mV) on lumenal RR concentration. Solid line was calculated according to Equation 3. Data are the means Ϯ S.E. of four experiments. Average Hill constants and coefficients are shown in Table II. It is assumed that RR binds to voltage-independent and -dependent sites of the RyRs. The Ca 2ϩ release channels are present in a closed Ca 2ϩ -free form (C) at cytosolic [Ca 2ϩ ] Ͻ0.1 M and Ca 2ϩ -activated form (O Ca(ϪATP) ) at cytosolic [Ca 2ϩ ] Ͼ0.1 M. Micromolar cytosolic Ca 2ϩ and the presence of millimolar ATP (ϮMg 2ϩ ) yields O Ca(ϮATP) . Binding of RR to cytosolic channel sites of C and O Ca(ϮATP) results in closed channel states C RRcyt and C Ca(ϮATP),RRcyt . Cytosolic and lumenal RR binding to sites of the open channels is influenced by membrane potential and results in subconductance channel states, O sub,RRcyt and O sub,RRlum , respectively. The oligomeric RyR contains cooperatively interacting Ca 2ϩ activation and RR binding sites; however, only one each is shown. Also not shown are the channel Ca 2ϩ -inactivation sites and channel closings originating from substates.
Vesicle-Ca 2ϩ flux measurements suggest the effectiveness of RR in inhibiting RyRs depended on Ca 2ϩ concentration. One likely reason for the requirement of relatively high RR concentrations in these studies was that Ca 2ϩ flux measurements were carried out in the presence of the allosteric activator molecule ATP (5). This notion was confirmed by our [ 3 H]ryanodine binding experiments that showed that the presence of AMPPCP and MgAMPPCP reduced RR inhibition of RyR. In overlay studies with RyR1 fusion peptides, RR bound to several Ca 2ϩ binding peptides (37), raising the possibility that RR competes with Ca 2ϩ for the Ca 2ϩ activation sites. In support of this suggestion, RR decreased [ 3 H]ryanodine binding by competitive inhibition (18). Alternatively, AMPPCP might have induced a conformational change that lowers the affinity of RR binding to sites not directly involved in regulation by Ca 2ϩ . In support of a noncompetitive mechanism is the finding that RR decreased B max and increased K D of [ 3 H]ryanodine binding (Ref. 16 and this study), with the latter effect being due to a slower association rate (16). In single channel experiments at low micromolar [Ca 2ϩ ] in the absence of ATP, cytosolic RR decreased the channel activities to the same extent at negative and positive holding potentials, without the appearance of channel states of reduced conductance. At elevated Ca 2ϩ concentrations, cytosolic RR also induced full channel closings at both holding potentials. These were analyzed for RyR2 only at negative holding potentials, at which substates were not observed. Because RR is a polycationic molecule, a voltage-independent inhibition suggests that RR fully closed the channels by binding to cytosolic sites that are located outside the electrical field.
Several differences were observed in the interaction between RR and the RyRs. At 10 -15 M Ca 2ϩ , RR was more effective in inhibiting the cardiac and skeletal RyRs in single channel than  Table II. cardiac and skeletal muscle RyRs, respectively. For RyR1, average Hill inhibition coefficients of 1.1-1.6 were obtained, suggesting no or a weak cooperative binding of RR. Hill inhibition coefficients of 2.4 and 2.3 suggest that RR inhibited single RyR2s by a cooperative interaction, whereas Hill coefficients of ϳ1 argue against cooperative inhibition of [ 3 H]ryanodine binding to RyR2. The reasons for these differences are not clear but may be related to isoform-specific differences, the use of SR vesicles versus purified RyRs, or the absence of a membrane potential in [ 3 H]ryanodine binding but not single channel measurements. Another difference was that in the presence of AMPPCP, RR fully inhibited [ 3 H]ryanodine binding to RyR1 but not RyR2 at RR concentrations as high as 100 M. Differences in the effectiveness of RR in inhibiting RyR isoforms were also noted in 45 Ca 2ϩ uptake measurements using skeletal and cardiac SR vesicles and in single channel measurements using purified RyRs.
Ruthenium red has been extensively used as a diagnostic tool in single channel measurements of RyRs (2). In general, RR was applied to RyRs under conditions that favored full channel closings and disfavored the formation of substates. The present study defines conditions that result in the formation of cytosolic and lumenal RR-related substates in both RyR1 and RyR2. The majority of experiments were done in a symmetric KCl solution in the absence of a high lumenal Ca 2ϩ concentration because it avoided the formation of a large Ca 2ϩ gradient near the cytosolic channel pores (34), thereby simplifying analysis of the interaction of RyRs with RR. The RyR ion channels conduct Ca 2ϩ across the SR membrane in muscle at a membrane potential that is likely close to 0 mV (36). The effects of RR on single channels were therefore also determined at 0 mV with 10 mM Ca 2ϩ as the conducting ion. In both recording conditions, lumenal and cytosolic RR induced subconductance states. We found that probability of substate formation (P sub ) depended on cytosolic and lumenal ruthenium red concentration, P o , and holding potential. Cytosolic RR induced a predominant substate in K ϩ -conducting channels at positive membrane potentials that favored cytosolic to lumenal cation fluxes. Substate formation was favored by conditions that resulted in nearly full channel activation. The most straightforward explanation is that cytosolic RR binds to sites located in the conductance pathways of RyR1 and RyR2 and that channel openings with long durations were required for RR to access these sites. The interaction of RyR2 with lumenal RR differed in several respects from that with cytosolic RR. First, lumenal [RR] Ͼ1 M was required to induce full channel closings, suggesting the absence of high affinity lumenal regulatory sites for RR. Second, in contrast to cytosolic RR, lumenal RR induced a major subconductance in RyR2 at a negative holding potential. At a positive membrane potential, elevated levels of lumenal RR produced channel closings. However, these were too brief and poorly resolved at the time resolution of the lipid bilayer set-up to be further analyzed. Interaction of lumenal RR with RyR1 was more complex and induced three major substates. Substates were observed at negative and positive holding potentials, which suggested a low voltage dependence for their formation. Further studies are required to clarify the mechanism(s) of the formation of the RR-lumenal related RyR1 substates.
Ca 2ϩ -conducting channels recorded at 0 mV formed substates with conductances similar to those formed by the K ϩconducting channels, with a P sub Յ 0.1 at RR concentrations Յ500 nM. Higher concentrations caused complete channel closings in the majority of the experiments. Thus, the substates are likely of physiological relevance; however, the dominant mech-anism of RR interaction with the Ca 2ϩ -conducting RyRs appears to be full closure.
Ma (23) described the effects of RR on the ryanodine-modified skeletal RyR. Several cytosolic RR molecules produced an all-or-none flickery block with a Hill coefficient of 2. Lumenal RR created a different blocking effect by attenuating the single channel currents instead of forming a substate. These results suggested that different binding sites are located in the conduction pore of the ryanodine-modified skeletal muscle RyR. Ma (23) did not describe the effects of RR on ryanodine-unmodified RyR1 or RyR2.
Does RR reduce the channel conductance of the RyRs by partial occlusion of the channel pore, or are the subconductances the result of RR-dependent changes in channel protein conformation? For other ion channels, examples for both mechanisms exist (38 -42). In the cardiac RyR, large cytosolic monovalent tetraalkyl ammonium cations produced a substate with a linear current-voltage relationship, a Hill coefficient of 1, and an effective gating charge of 1.7 (31). These results were explained in terms of a partial block in which the binding of more than one tetralkyl ammonium cation in the voltage drop of RyR2 produced an electrostatic barrier for ion translocation. The binding of the reversible ryanoid 21-amino-9␣-hydroxyryanodine to a cytosolic site of the open RyR2 channel was strongly influenced by the membrane potential (43). The results suggested that the voltage dependence was due to a voltage-driven conformational change that altered the affinity of the binding site. Imperatoxin activator, a positively charged 33-amino acid peptide, induced a rectifying subconductance state by binding to a single cytosolic site of the cardiac and skeletal muscle RyRs (30). The formation of rectifying substates suggested that imperatoxin activator might have induced a conformational change in both RyRs. RR also induced rectifying subconductance states in the RyRs. The rate of formation but not the current amplitude of the substates was dependent on RR concentration. These results favor a conformational change as the mechanism for the formation of the substates. A conformational change can explain why the binding of one or more RR produces only one predominant substate.
Voltage dependence of substates suggests that RR binds to cytosolic and lumenal sites in the conductance pathway of RyR2. However, the current amplitudes of the cytosolic and lumenal RR-related substates were different, which argues against a common binding site. For the cytosolic RR binding site, an effective gating charge of 1.3 was determined. For the lumenal site, the gating charge was 0.8. Dividing these numbers by the six positive charges of RR yields an electrical distance of 0.22 from the cytosolic side and 0.13 from the lumenal side for the location of the RR binding sites in the voltage drop of the RyR2. The calculations are based on the assumption that all charges of RR enter the conduction pathway. The above numbers would be smaller if more than one RR molecule enters the conduction pathway. The RyRs consist of a large 29 ϫ 29 ϫ 12-nm cytosolic foot region with a large, centrally located vestibule and a smaller transmembrane region that extends ϳ7 nm toward the SR lumen (44,45). Therefore, the total length of the conductance pathway may be 19 nm. By comparison, RR has an estimated length of 1.2 nm with a diameter of 0.4 nm. Accordingly, it is conceivable that several RRs simultaneously enter the conduction pathway of the tetrameric channel complexes. Theoretical considerations suggest that a significant portion of the voltage drop may occur in the vestibules of a channel (46). Thus, the RR binding sites may be close to the cytosolic and lumenal entrances of the conductance pathway.
In conclusion, RR modifies RyRs by at least two different mechanisms. RR inhibits the Ca 2ϩ -activated RyRs by a voltage-independent mechanism involving the noncompetitive interaction between Ca 2ϩ regulatory and RR inhibition sites. In addition, cytosolic and lumenal RR can induce substates in RyRs. Formation of substates is influenced by membrane potential, and the rate of formation but not conductance of the substates is dependent on RR concentration. These observations argue against partial occlusion of the channel pore by RR. An alternative explanation is that the substates are the result of voltage-and RR-dependent changes in channel protein conformation. The formation of substates reduces the effectiveness of RR as an inhibitor of RyR channel activity. Therefore, caution is required when experiments are done under conditions that, as defined in this study, favor formation of these substates.