Multiple actions of imperatoxin A on ryanodine receptors: interactions with the II-III loop "A" fragment.

Imperatoxin A is a high affinity activator of ryanodine receptors. The toxin contains a positively charged surface structure similar to that of the A fragment of skeletal dihydropyridine receptors (peptide A), suggesting that the toxin and peptide could bind to a common site on the ryanodine receptor. However, the question of a common binding site has not been resolved, and the concentration dependence of the actions of the toxin has not been fully explored. We characterize two novel high affinity actions of the toxin on the transient gating of cardiac and skeletal channels, in addition to the well documented lower affinity induction of prolonged substates. Transient activity was (a) enhanced with 0.2-10 nm toxin and (b) depressed by >50 nm toxin. The toxin at >/=1 nm enhanced Ca(2+) release from SR in a manner consistent with two independent activation processes. The effects of the toxin on transient activity, as well as the toxin-induced substate, were independent of cytoplasmic Ca(2+) or Mg(2+) concentrations or the presence of adenine nucleotide and were seen in diisothiocyanostilbene-2',2'-disulfonic acid-modified channels. Peptide A activated skeletal and cardiac channels with 100 nm cytoplasmic Ca(2+) and competed with Imperatoxin A in the high affinity enhancement of transient channel activity and Ca(2+) release from SR. In contrast to transient activity, prolonged substate openings induced by the toxin were not altered in the presence of peptide A. The results suggest that Imperatoxin A has three independent actions on ryanodine receptor channels and competes with peptide A for at least one action.

Excitation-contraction (EC) 1 coupling is the process that facilitates Ca 2ϩ release from the sarcoplasmic reticulum (SR) of muscle fibers following depolarization of the surface/transverse (t-) tubule membrane. A protein-protein interaction between the dihydropyridine receptor (DHPR) and ryanodine receptor (RyR) underlies EC coupling in skeletal muscle. The DHPR L-type Ca 2ϩ channel in the t-tubule membrane detects surface depolarization and transmits a signal to the RyR channel in the SR via an interaction between the cytoplasmic domains of the two proteins. The interacting region of the DHPR is located between the second and third transmembrane repeats in the ␣1 subunit (II-III loop) (1). The interacting regions of the RyR are less clearly defined but are likely to involve residues 1076 -1112 (2) and residues 1837-2168 (3).
The ability of peptide A to activate RyRs is highly correlated with its capacity to adopt an ␣-helical structure (15,16,18) and with the orientation of positively charged residues along one surface of the molecule (18). Curiously, two scorpion toxins, Imperatoxin A and Maurocalcine (having 82% sequence identity), have structural features in common with peptide A. Although the intrinsically disulfide-stabilized ␤-sheet structure of the toxins is vastly different from the ␣-helical structure of the A peptide, the toxins and peptide A share a similar surface orientation of positively charged residues (18,20). Because of this structural similarity, several studies have examined the possibility that the scorpion toxins and peptide A bind to the same site on RyR1. Two studies have concluded that they bind to the same, or overlapping sites (13,18), whereas a third study concluded that Maurocalcine/Imperatoxin A and peptide A bind to independent sites (19). These different conclusions could have arisen if more than one binding site exists for either the toxins and for peptide A, and not all sites are common to the two compounds. At least two binding sites for peptide A have been defined, one within the channel pore, which leads to voltage-dependent channel block (12), and a site (or sites) on the cytoplasmic domain of the channel, which support its voltage-independent actions (10,12). Here we investigate the possibility that Imperatoxin A also has multiple actions on RyR activity and examine interactions of between the toxin and peptide A in modifying RyR channel gating.
We find that there are a least three separate effects of Imperatoxin A on cardiac and skeletal RyR channels that can be distinguished by their affinity and reversibility and by their ability to compete with peptide A in its native and a modified form. The results show that peptide A competes with Imperatoxin A for the high affinity activation of transient channel openings. On the other hand, peptide A does not prevent the characteristic toxin-induced prolonged substate activity. The results suggest that there is at least one common or overlapping binding site for Imperatoxin A and peptide A as well as independent binding sites. These observations raise the questions of which of the Imperatoxin A binding sites has been identified on the RyR (21) and whether this site is the site that also interacts with peptide A.

MATERIALS AND METHODS
Peptides-Peptides A and A1(D-R18) were synthesized as in Green et al. (18). Imperatoxin A was synthesized by Auspep Australia and folded using procedures outlined by Fajloun et al. (22) Vesicle Preparation-Preparation of SR vesicles, Ca 2ϩ release from SR, and single channel techniques have been described previously. Heavy skeletal SR vesicles were prepared from rabbit back and leg muscle (11,12), whereas cardiac SR vesicles were prepared from sheep heart (23).
Single Channel Techniques-Bilayers of phosphatidylethanolamine, phosphatidylserine, and phosphatidylcholine (5:3:2 w/w) (Avanti Polar Lipids, Alabaster, AL) were formed across an aperture of ϳ200-m diameter in the wall of a 1.0-ml Delrin cup (Cadillac Plastics, Australia). Terminal cisternae vesicles (10 g/ml) were added to the cis chamber. The cytoplasmic side of channels incorporated into the bilayer faced the cis solution. Bilayer potential was controlled, and single channel currents were recorded, using an Axopatch 200A amplifier (Axon Instruments, Foster City, CA). Bilayer potential is expressed as V cis Ϫ V trans (V cytoplasm Ϫ V lumen ).
Bilayers were formed and vesicles incorporated using cis solutions containing (in millimolar): 230 CsCH 3 O 3 S/20 CsCl/5.0 CaCl 2 /10 TES/ 500 mM mannitol (pH 7.4) with CsOH and a trans solution containing (in millimolar) 30 CsMS/20 CsCl/1 CaCl 2 /10 TES (pH 7.4). Following incorporation, (a) the cis solution was replaced with an identical solution, except that the [Ca 2ϩ ] was between 0.1 and 100 M and (b) 200 mM CsCH 3 O 3 S was added to the trans chamber for symmetry. Drugs were added to the cis chamber and removed by perfusion with 10 ml of cis solution.
Analysis of Channel Activity-Channel activity was analyzed over one to two 30-s periods of continuous activity at ϩ40 mV and then at Ϫ40 mV. Slow fluctuations in the baseline were corrected using an in-house baseline correction program (written by Dr. D. R. Laver). Channel activity was measured either as "mean current" (the average of all data points in a record) or as open probability (P o ), using a threshold analysis with the program Channel 2, (developed by P. W. Gage and M. Smith, John Curtin School of Medical Research). Measurements of mean current, performed on records from experiments containing 1-4 channels, included all channel activity from the smallest subconductance level to maximum openings. On the other hand, open probability (P o ), mean open time (T o ), and mean closed time (T c ) measurements are restricted to records in which the opening of a single channel only could be detected and exclude openings that fall within the baseline noise. In this case threshold levels for channel opening and closing were set to exclude baseline noise (a) at ϳ20% of the maximum single channel conductance when examining transient openings in the absence of prolonged substates or (b) at 50% of the maximum conductance or 50% of the substate level to measure P o of maximum or substate conductance respectively, when substates were present.
Channel activity is expressed as relative P o to include data in which activity varied from ϳ0.0001 to ϳ0.1 and data from bilayers containing more than one channel. Relative DIDS Modification-Channels modified by the disulfinic stilbene derivative, diisothiocyanostilbene-2Ј,2Ј-disulfonic acid (DIDS), were used in some cases with a cytoplasmic [Ca 2ϩ ] of 100 nM to enhance channel activity under control conditions. DIDS modification does not alter the regulation of RyRs by Mg 2ϩ , ryanodine, or ruthenium red (24), although it can interact with other properties of RyRs (25)(26)(27). Channels were exposed to 100 or 300 M DIDS in the cis chamber for 4 -6 min, and then DIDS was removed by perfusion. Activity increased in the presence of DIDS and then fell with removal of the reversible component of activation. However, activity remained higher than before exposure to DIDS because of covalent bonds formed between isothiocyanate groups and NH 2 , OH, and aromatic groups on a variety of amino acid residues (28).
Ca 2ϩ Release from SR-Extravesicular Ca 2ϩ was monitored at 710 nm using a Cary 3 spectrophotometer (12). The cuvette solution was stirred continuously, and temperature was controlled at 25°C. Skeletal SR (100 g of protein) was added to the cuvette solution (final volume of 2 ml), containing (in millimolar): 100 KH 2 PO 4 (pH 7); 4 MgCl 2 ; 1 Na 2 ATP; 0.5 antipyrylazo III. Ca 2ϩ ,Mg 2ϩ -ATPase activity was suppressed with thapsigargin (200 nM (29)). The same solutions were used with cardiac SR except that an ATP-regenerating system (phospho(enol)pyruvate (5 mM) and pyruvate kinase (25 g/ml)) was added, and Ca 2ϩ -induced Ca 2ϩ release was triggered by addition of 20 M Ca 2ϩ to the cuvette solution.
When toxin was examined alone, it was added 2 min after thapsigargin. When peptide was added before toxin, either peptide or an equivalent volume of water (vehicle) was added 2 min after thapsigargin, and then toxin or an equivalent volume of water (vehicle) was added after a further 2 min. Release rates in each experiment were measured 10 -20 s after toxin addition. Statistics-Average data are given as mean Ϯ S.E. The significance of the difference between control and test values was tested a using either (a) a Student's t test, either one or two sided and either for independent or paired data, as appropriate or (b) using the non-parametric "sign" test (30). Differences were considered to be significant when p Յ 0.05.

RESULTS
RyR activity demonstrated two distinct gating modes, a "transient" mode and a maintained "substate" mode. The transient mode comprised all control channel activity and was observed in toxin-modified channels. In this mode, the duration of channel openings varied from ϳ0.5 to ϳ1000 ms with brief submaximal openings and longer openings to the maximum conductance. The maintained substate mode was seen only after exposure to Imperatoxin A and was characterized by openings lasting from ϳ1 to 20 s. The number of transient openings increased with toxin concentrations between 200 pM and 20 nM, then declined with Ͼ20 nM toxin. Prolonged substate openings appeared with Ͼ100 nM toxin.

Imperatoxin A Alters the Probability of Transient RyR Channel Openings
Increased Transient Activity-Imperatoxin A applied at picomolar concentrations to the cis (cytoplasm) side of RyR channels caused an increase in the frequency of transient openings. The increase occurred rapidly, within the 30-s period of toxin addition and stirring. Prolonged substate openings were not induced with these low toxin concentrations. The activating effect on transient openings was only slowly reversible when the toxin was perfused from the cis chamber ( Fig. 1, A-C).
The changes in channel gating associated with increased activity were measured in a subset of data from skeletal and cardiac RyR channels (Table I). The increased activity of skeletal RyRs with 1 and 10 nM toxin was due to an increase in channel open time, and a reduction in closed time. The increase in open time was also seen in cardiac channels, but in contrast to skeletal channels, there was no significant change in closed durations at lower toxin concentrations.
Inhibition of Transient Activity-Transient channel openings decreased when the toxin was increased to between 50 and 500 nM (see Fig. 1C). Some channels were activated for several minutes after exposure to higher [toxin] and were then inhibited ( Fig. 2A). Activity was rapidly restored to an activated (greater than control) level upon removal of the toxin, indicating raid reversibility of the inhibitory action.
The inhibition of transient activity was independent of prolonged substate openings. Although all channels exhibiting prolonged substates also showed fewer transient openings, the number of transient openings was often substantially reduced in the absence of substate activity (100 nM toxin ( Fig. 2A)). Inhibition reduced P o to less than control when the activity was initially high (in channels with activating cis Ca 2ϩ and/or ATP ( Fig. 2A)). When activity was low (in the presence of 100 -300 nM Ca 2ϩ and/or MgATP), P o fell to less than the toxin-activated level, but often remained higher than control (e.g. the skeletal channels in Table I).
The decline in transient activity was caused by an increase in closed time (Table I). Curiously, the open times in skeletal RyRs continued to increase as toxin concentration increased (Table I), perhaps indicating that the activating effect of the toxin increased with [toxin], but was overwhelmed by an independent inhibitory effect. Persistent inhibition was also seen following perfusion of inhibiting [toxin] from the cis chamber, when removal of inhibition revealed a strong slowly reversible activation (e.g. Fig. 2A). On average, cardiac RyRs showed a significant decline in open times and increase in closed dura-  (Table I). Inhibition was seen in all cardiac (n ϭ 13) and skeletal (n ϭ 9) channels exposed to higher toxin concentrations and was independent of bilayer potential or cis [Ca 2ϩ ] (n ϭ 9 for Յ300 nM Ca 2ϩ or n ϭ 13 for 10 M Ca 2ϩ ), and occurred without ATP or Mg 2ϩ (n ϭ 8) or with 2 mM ATP (n ϭ 14) and 2 mM Mg 2ϩ (n ϭ 4), as well as in DIDS-modified channels (n ϭ 5) (Fig. 2B).
Toxin-activated channels were further activated by Ca 2ϩ , ATP, or AMP-PNP, whereas inhibited channels could not be activated to the same extent by these ligands (Fig. 2C). In this experiment, with 100 nM cis Ca 2ϩ , cardiac RyR channels were either activated or inhibited by 200 nM toxin. The toxin-acti-    Three Discrete Actions of Imperatoxin A on RyRs vated channels were further activated when cis Ca 2ϩ was increased to 100 M and activated again by either 2 mM ATP or AMP-PNP. In contrast, although activity increased in the toxin-inhibited channels, the increase (relative to control) was significantly less than in the activated channels.
Peptide A Competes with Imperatoxin A for Activation of Transient Opening-It has been suggested that peptide A and Imperatoxin A compete for a single site on the RyR. Cardiac channels were activated by 1 nM toxin and by subsequent additions of peptide A up to 500 nM (Fig. 3A). Activity declined when the peptide was increased to 1 M. However, the plateau of activation with peptide plus toxin was no greater than that with peptide alone, suggesting that there was no summation of the effects of the two compounds and supporting the concept of the same or overlapping binding sites. Similar results were obtained when channels were exposed to 10 nM toxin, although there was a significant reduction in activity when the [peptide] reached 500 nM. In a similar experiment (Fig. 3B), channels were first exposed to 50 nM peptide A and then to increasing concentrations of toxin. Activity tended to increase with peptide plus toxin up to 600 pM, but activation was significantly less than expected from the summation of two independent processes (Fig. 3B). Channel activity declined when higher concentrations of toxin were added with peptide. The decline in activity with higher concentrations of toxin plus peptide A (or modified peptide A (A1-R18D) (18)) was particularly apparent in skeletal RyR channels (Table II and Fig. 3, C and D). Transient openings declined with 1 nM toxin in the presence of 1 M peptide A (Fig. 3), even though there was no inhibition of RyRs by either 1 nM toxin or 1 M peptide A alone.

Prolonged Substate Openings Induced by Imperatoxin A
Imperatoxin A induces prolonged channel openings to substate levels in cardiac and skeletal RyRs (31). In the present experiments, toxin concentrations of 1-10 M were required to consistently observe substate activity within 2-10 min of exposure to the toxin, although substate activity could occasionally be seen after prolonged exposure to 10 -100 nM toxin. Channels demonstrated abrupt transitions between the substate and transient gating modes (Fig. 4A). In the substate mode, the channel was almost continually open with P o close to 1.0 and average open times of ϳ1-20 s. Substate gating was interrupted by bursts of transient activity with frequent channel openings from the closed to the maximum conductance levels (Fig. 4A).
The currents from the DIDS-modified channel in Fig 7). The broken line shows expected activity for added effects (A alone at 50 nM (second bar) plus the mean of 1 and 10 nM toxin from A). In A and B, note 1) maximum activation with toxin plus peptide A is no different from that with peptide A alone and 2) a decline in activity with higher concentrations of toxin plus peptide. C and D, exclusive activation of skeletal RyRs by either peptide A or Imperatoxin A: 5-s current records (left) and 30-s all-points histograms (right) with a cis [Ca 2ϩ ] of 100 nM. C, 10 nM toxin increased channel activity at ϩ40 mV and activity declined with 1 M peptide A plus toxin. Conversely, D shows an increase in activity at Ϫ40 mV with 1 M peptide A and then a decrease after adding 1 nM toxin. the similarity of the toxin's action (and substate levels) at ϩ40 and Ϫ40 mV; and 3) rapid recovery of transient openings after removal of the toxin. The substate activity remained for several minutes after removal of the toxin and restoration of transient activity. Two or three different conductance levels are apparent in the toxin-induced substate activity (Fig. 4). The highest conductance substate dominated and was a similar fraction of the maximum conductance at ϩ40 and Ϫ40 mV (Table III). The  (A1-R18D), on the relative probability of transient skeletal RyR channel activity The concentrations of toxin and peptide used were maximal for activation of transient openings. Data obtained at ϩ40 and Ϫ40 mV have been combined, because similar results were obtained at the two potentials. Each of the toxins, peptide A and A1-R18D, increase the relative open probability when added alone. However, subsequent addition of a second compound failed to increase activity further. Some of this data were presented in Green et al. (18) and have been re-evaluated for comparison with data in this report.  substate was significantly higher in the skeletal RyRs (ϳ33% maximum) than in cardiac RyRs (29% maximum).
The substate parameters were not obviously affected by the channel type (cardiac or skeletal), voltage across the bilayer, cis Ca 2ϩ or Mg 2ϩ or modification by DIDS. The average combined data in Table III shows that the substate mode occupied ϳ70% to 80% of channel activity and that, within that mode, the P o was ϳ1.0, the mean open time was ϳ2-7 s, and mean closed time was between ϳ11 and 40 ms.
The Effects of Peptide A on the Toxin-induced Substate-The interaction between peptide A and Imperatoxin A during prolonged substate activity was examined in DIDS-modified channels with a cis [Ca 2ϩ ] of 300 nM. The RyR in Fig. 5 was activated to a P o of 0.999 by 300 M DIDS. Activity fell to 0.44 when DIDS was removed ("Materials and Methods"). 1 M peptide A increased P o of the DIDS-modified channel to 0.977. Subsequent addition of 1 M toxin resulted in an immediate decline in transient openings (P o ϭ 0.15) and the slower appearance of prolonged substate activity with the usual gating transitions between substate and transient modes. Additional substate levels between 50 and 90% of maximum were apparent with toxin plus peptide in some channels (e.g. Fig. 5). These substate levels were not obvious with 1 M peptide A alone (Fig. 5, A and B) but were occasionally seen when peptide A was added to normal (not DIDS-modified) channels at concentrations of Ͼ5 M (12).
The records in Fig. 5 (A and B) reinforce several novel features of the actions of Imperatoxin A and peptide A. 1) Peptide A activates cardiac RyR channels after DIDS modification. 2) 1 M Imperatoxin A inhibits transient channel gating with minimal substate activity. 3) Imperatoxin A induces its characteristic substate in the presence of peptide A. 4) Peptide A-induced substates, between 50 and 90% s of the maximum conductance, occur in the presence of toxin-induced substate activity.
When 1 or 10 M peptide A was added to channels after 1 M toxin in three experiments, no change in Imperatoxin A-induced substate activity was observed, but additional substate activity occasionally appeared, similar to that shown in Fig. 5.
Taken together these results suggest that, as with maurocalcine (19), the effects of peptide A and the ability of Imperatoxin A to induce prolonged substate openings are independent actions, probably due to the compounds binding to separate sites on the RyR.

Effects of Imperatoxin A on Ca 2ϩ Release from SR
The effects of Imperatoxin A on Ca 2ϩ release were examined to see whether the high affinity action of the toxin on RyR channels could be seen in a system that was more intact and physiological than the bilayer.
Concentration Dependence of Effects of Imperatoxin A on Ca 2ϩ Release-Addition of only 1 nM Imperatoxin A significantly enhanced resting Ca 2ϩ release from thapsigargin-blocked skeletal SR vesicles (Fig. 5C). Higher concentrations of toxin resulted in the release of Ca 2ϩ at progressively higher rates, up to Ͼ5000 nmol/mg/min with 40 M toxin (Fig. 5C). The data were best fitted by the sum of two Hill functions, the first with a maximum rate of 650 nmol/mg/min, an AC 50 of 190 nM, and Hill coefficient of 0.95. The second had with a maximum of 5300 nmol/mg/min, an AC 50 of 4.4 M, and a Hill coefficient of 1.0. Single Hill functions did not fit the data, particularly at low toxin concentrations (broken line, Fig. 5C). Similar results were obtained with Ca 2ϩ -induced Ca 2ϩ release from cardiac SR (Fig. 5D). 2 These results are consistent with the single channel data indicating two activation processes, one high affinity, equivalent to the increase in the probability of transient openings, and one with lower affinity, equivalent to prolonged substate openings.
Competition between Imperatoxin A and Peptide A in Ca 2ϩ Release from SR-Exposure of skeletal vesicles to 1 M peptide A resulted in an increase in the rate of release that was similar to that induced by exposure to 1 nM Imperatoxin A (Fig. 5D). However, the rate of release was not enhanced when 1 nM toxin was added after 1 M peptide A. Thus there appeared to be no additive effect on release when peptide A was present with a low concentration of toxin. On the other hand, the rate of release of Ca 2ϩ induced by 5 M toxin tended to be greater when toxin was added to peptide-activated channels. These results are again consistent with the single channel data, which showed interdependent actions of the peptide and the high affinity effect of the toxin, but an additive effect of the peptide and the lower affinity effect of the toxin (i.e. substate openings).

DISCUSSION
Activation of transient gating of skeletal and cardiac RyR channels by picomolar concentrations of Imperatoxin A is described for the first time. Novel observations with DIDS-modified channels illustrate the robust nature of (a) high affinity activation and lower affinity inhibition of transient activity by Imperatoxin A, (b) toxin-induced substate gating, and (c) activation by peptide A. We show that peptide A competes with Imperatoxin A only for actions on transient channel gating. The toxin-induced substate persists in the presence of peptide A, with additional transient openings to conductance levels peculiar to the peptide. The results raise questions about the functional effects of binding to the site identified for Imperatoxin A (21) and whether peptide A also binds to the site. Finally the multiple actions of the toxin provide an explanation for divergent reports of the existence of a common binding site with peptide A (13,18,19). 2 Note that the rates of Ca 2ϩ release from the crude cardiac SR preparation are significantly lower than those from the heavy skeletal SR preparation, largely because of an ϳ10-fold greater contamination with longitudinal SR vesicles that do not contain Ca 2ϩ release channels.

Three Discrete Actions of Imperatoxin A on RyRs
Multiple Independent Actions of Imperatoxin A-The effects of Imperatoxin A are separated by concentration dependence, reversibility, gating, and interactions with peptide A. The increase in transient activity, with as little as 200 pM toxin, occurred with an increase in open time and decrease in closed times, and was irreversible in the short term. Activated channels were further activated by Ca 2ϩ and ATP (or AMP-PMP). Peptide A competes with the toxin for transient activation.
Toxin at Ͼ10 nM inhibited transient openings and prolonged closed times. Inhibited channels could not be fully activated by ATP or AMP-PNP. Inhibition was (a) rapidly reversed when the toxin was removed and (b) enhanced by peptide A. The third action of the toxin was the induction of prolonged substates that persisted after recovery of transient gating when the toxin was removed and persisted in the presence of peptide A. The enhancement of transient RyR activity by picomolar concentrations of Imperatoxin A has not been reported, possi-bly because previous studies have not systematically examined of the concentration dependence of the toxin, but focused on the toxin-induced substate (31) and compared it with peptide A-induced substates (13).
Ca 2ϩ Release from SR-Imperatoxin A at Ն1 nM released Ca 2ϩ from SR. A dual activating effect of the toxin was suggested by the need for two Hill functions fit to the data. The fits suggested a small high affinity effect added to a large low affinity effect, corresponding, respectively, to the high affinity activation of transient gating and the lower affinity substate induction. The effects on Ca 2ϩ release were at ϳ10-fold higher concentrations than effects on channels in bilayers. The reason for this shift to higher [toxin] in Ca 2ϩ release experiments is unclear but is also seen with the DHPR II-III loop and loop peptides (4,12). Because each of the four RyR subunits are likely to contain toxin binding sites, it is curious that the Hill coefficients were around 1. This could mean that binding of one toxin molecule to the tetramer was sufficient to evoke a maximum effect. However, by analogy with Ca 2ϩ inhibition (23), it is likely that the low coefficients are due to averaging release through many channels with individual variations in responses to the toxin.
The inhibition of transient activity was not reflected in Ca 2ϩ release from SR, possibly because (a) of variations between individual channels in the toxin concentration at which inhibition was apparent, (b) activity may have remained greater than control in some channels, even though it was less than the maximum toxin-activated level, and (c) any inhibitory effect on transient activity would have been swamped by the current flowing through substate openings at higher [toxin]. Under the bilayer conditions, the P o for transient openings with low cis [Ca 2ϩ ] in the presence of inhibiting toxin concentrations was between 0.001 and 0.0001, and thus the average current flowing through one channel would be between 0.01 and 0.001 pA, assuming there was a maximum single channel current of 10 pA. In contrast, when the channel was open at the substate level at 30% of the maximum conductance for 80% of the time, the average current through the channel would be ϳ2.5 pA.
Actions of Peptide A-Peptide A also has multiple independent effects. First, it increases transient activity in skeletal RyRs with high affinity (Ն10 nM) and releases Ca 2ϩ from skeletal SR (5,6,10,(12)(13)(14)(15)(16)(17)(18). Ca 2ϩ -induced Ca 2ϩ release from cardiac SR is enhanced by the peptide and cardiac RyRs activated at sub-activating cis [Ca 2ϩ ]. 3 The second action is a voltage-independent inhibition, seen specifically with the modified A1-R18D at Ն10 nM in skeletal and cardiac RyRs with 100 M cis Ca 2ϩ . 3 The native peptide A is less active with 100 M cis Ca 2ϩ than with lower [Ca 2ϩ ] (12,17). Because peptide structure is not Ca 2ϩ -dependent, 3 this inhibition raises the possibility either that there is a separate inhibitory site or the activation site can become inhibitory. The third action is a voltagedependent block of the skeletal and cardiac RyR pore with Նϳ5 M peptide A, leading to transient substate activity, particularly when the current pulls the peptide into the pore (at ϩ40 mV) (10,12,32). The substates have conductances of Ͻ10 -90% of the maximum conductance, with lower levels dominating at positive potentials.
Interactions between Imperatoxin A and Peptide A: Common and Independent Binding Sites-The non-additive nature of activation of transient gating by Imperatoxin A and peptide A is indicative of identical or overlapping sites so that binding of either peptide or toxin precludes binding of the other compound. In contrast, the toxin-induced substate activity persisted in the presence of peptide A suggesting independent binding sites (Fig. 6).
The results with inhibition are less easily explained. Addition of peptide plus toxin at activating concentrations for the individual compounds, resulted in a decline in activity at ϩ40 and Ϫ40 mV and shift of inhibition to lower [toxin]. Peptide A alone does not cause a voltage-independent inhibition with 100 nM cis Ca 2ϩ (previous section), so that the results could suggest that peptide A binds to a site near the toxin inhibition site to increase the affinity for the toxin. However, an additive effect cannot be ruled out, because the modified A1-R18D can cause voltage-independent inhibition under appropriate conditions. Thus the shift in inhibition to lower toxin concentrations could be explained by the compounds binding to the same or different inhibitory sites. Additive effects could not be determined because inhibition by either compound reduced transient channel activity to very low levels. Overall, these data suggest that there is (a) a common transient activation site for Imperatoxin A and peptide A, (b) independent substate induction sites for the toxin and peptide, and (c) a toxin inhibition site whose relationship to the peptide A site is unclear (Fig. 6).
Similar Actions of the Toxin and Peptide on Skeletal and Cardiac RyR Channels-Imperatoxin A induces similar substate activity in skeletal and cardiac RyR channels (Table III  and Ref. 31). Similarly, both skeletal and cardiac II-III loop and II-III loop peptides can interact with the either the skeletal or cardiac RyR (33,34). 3 These results imply that a physical interaction can occur between either skeletal or cardiac RyRs and DHPRs, if the proteins were appropriately targeted (32). This is difficult to test, because DHPRs do not form tetrads and are not aligned with cardiac RyRs (35).
The Imperatoxin A-induced Substate-The toxin-induced substate was ϳ30% of the maximum conductance at both ϩ40 and Ϫ40 mV in cardiac and skeletal RyRs. This differs from reported Imperatoxin A-induced substates of 43% at Ϫ40 mV and 28% at ϩ40 mV (31), perhaps because we used Cs ϩ as the current carrier (as opposed to K ϩ ), or native rather than CHAPS-solubilized RyRs. Maurocalcine-induced substates are also voltage-independent (19).
Previous Studies of Imperatoxin A and Peptide A Binding-Findings that peptide A and Imperatoxin A do, or do not, bind to one site are consistent with multiple actions of the compounds. Gurrola et al. (13) showed competition between the toxin and peptide A in Ca 2ϩ release from SR, [ 3 H] ryanodine binding and 125 I-Imperatoxin A binding. Because these parameters do not reveal channel gating, the competition could have been for any of the functional sites and was most likely for the transient activation site. Indeed Green et al. (18) found interactions between peptide A and effects of the toxin on transient gating. Chen et al. (19) report independent peptide A and Maurocalcine-induced substates. Thus previous reports are consistent with the present finding that the toxin and peptide A compete for transient activation, but not for substate induction. In addition, the observation that peptide A inhibits the increase in [ 3 H]ryanodine binding induced by the Imperatoxin A (13), even though toxin and peptide individually activate Ca 2ϩ release, is consistent with the fact that transient channel gating is inhibited when toxin and peptide A are added together.
A binding site for Imperatoxin A has been identified on the 3 A. F. Dulhunty and M. G. Casarotto, unpublished data.

FIG. 6. A schematic representation of the various Imperatoxin A and peptide A binding sites on a RyR channel in a bilayer situation.
cytoplasmic domain of the RyR, close to the membrane (21). Because the site is likely to be a high affinity site, toxin binding to it may activate transient openings and thus also be the peptide A binding site. This site is remote from RyR domain 6, which is closest to the t-tubule membrane and the DHPR (36,37). If the A region of the II-III loop does interact with the RyR under any conditions in vivo, it is likely that the interaction is with a part of the protein that is closer to domain 6, than the toxin binding site. It is thus possible that the inhibitory action of the peptide A and the toxin are caused by binding to a site that differs from the activation site and is located closer to the T-tubule membrane (Fig. 6).
In conclusion, we find that Imperatoxin A has three independent functional effects on cardiac and skeletal RyR channels. The skeletal DHPR II-III loop-derived peptide (peptide A) competes with the toxin for its high affinity activation of transient channel openings. Peptide A modifies a lower affinity toxin-induced inhibition of transient activity and has an independent action on the lowest affinity effect of the toxin in inducing prolonged substate activity. We suggest that there are three functional binding sites for the toxin and that at least one site is shared by peptide A.