Peptide probe of ryanodine receptor function. Imperatoxin A, a peptide from the venom of the scorpion Pandinus imperator, selectively activates skeletal-type ryanodine receptor isoforms.

We have used [3H]ryanodine binding experiments and single channel recordings to provide convergent descriptions of the effect of imperatoxin A (IpTxa), a ∼5-kDa peptide from the venom of the scorpion Pandinus imperator (Valdivia, H. H., Kirby, M. S., Lederer, W. J., and Coronado, R.(1992) Proc. Ntl. Acad. Sc. U.S.A. 89, 12185-12189) on Ca2+ release channels/ryanodine receptors (RyR) of sarcoplasmic reticulum (SR). At nanomolar concentrations, IpTxa increased the binding of [3H]ryanodine to skeletal SR and, to a lesser extent, to cerebellum microsomes. The activating effect of IpTxa on skeletal SR was Ca2+-dependent, synergized by caffeine, and independent of other modulators of RyRs. However, IpTxa had negligible effects on tissues where the expression of skeletal-type RyR isoforms (RyR1) is small or altogether absent, i.e. cardiac, cerebrum, and liver microsomes. Thus, IpTxa may be used as a ligand capable of discriminating between RyR isoforms with nanomolar affinity. IpTxa increased the open probability (Po) of rabbit skeletal muscle RyRs by increasing the frequency of open events and decreasing the duration of the closed lifetimes. This activating effect was dose-dependent (ED50 = 10 nM), had a fast onset, and was fully reversible. Purified RyR from solubilized skeletal SR displayed high affinity for [3H]ryanodine with a KD of 6.1 nM and Bmax of ∼30 pmol/mg of protein. IpTxa increased [3H]ryanodine binding noncompetitively by increasing Bmax to ∼60 pmol/mg of protein. These results suggested the presence of an IpTxa-binding site on the RyR or a closely associated regulatory protein. This site appears to be distinct from the caffeine- and adenine nucleotide-regulatory sites. IpTxa may prove a useful tool to identify regulatory domains critical for channel gating and to dissect the contribution of skeletal-type RyRs to intracellular Ca2+ waveforms generated by stimulation of different RyR isoforms.

In cardiac and skeletal muscle, the calcium release channel/ ryanodine receptor (RyR) 1 constitutes the major pathway for Ca 2ϩ release from the sarcoplasmic reticulum (SR) during excitation-contraction coupling (1). RyRs are also found in neurons (2,3), exocrine cells (4), smooth muscle cells (5,6), epithelial cells (7), lymphocytes (8), and sea urchin eggs (9). In all of these cells, RyRs play a central role in the regulation of the intracellular free Ca 2ϩ concentration, whose elevation triggers a cascade of events that culminates in muscle contraction, hormone secretion, lymphocyte activation, egg fertilization, etc. To gain the functional flexibility necessary to respond to different triggering signals, at least three tissue-specific isoforms of RyR are expressed (10). In mammals, RyR1 is expressed predominantly in fast-and slow-twitch skeletal muscle, while RyR2 is expressed predominantly in cardiac muscle. RyR3 appears to be localized to brain, smooth muscle, and epithelial cells, although low levels of expression of RyR1 and RyR2 are also found in some of these tissues (11,12). Several structural and functional characteristics confer to RyRs a distinctive earmark. They are homotetramers of large molecular size ϳ(2 million Da (13)); they form Ca 2ϩ -gated Ca 2ϩ -selective channels of large conductance (14), and they are distinctively affected by the plant alkaloid ryanodine (15).
The effects of ryanodine on single RyRs are complex and highly dependent on the concentration of the alkaloid. At low concentrations (5-50 nM), ryanodine increases the mean open time of the channel without modifying its unitary conductance (16,17). At intermediate concentrations (50 nM to 10 M), ryanodine "locks" the channel in an open subconductance state that corresponds to ϳ50% of the full conductance level. Higher concentrations (Ͼ100 M) of the alkaloid fully and irreversibly close the channel (18,19). Thus, ryanodine may act as an agonist and a blocker of the RyR.
In addition to its dual effect, ryanodine displays a slow association rate of binding (19,20) that renders the onset of activation incompatible with the time frame of electrophysiological experiments. To accelerate the onset of effect, micromolar, instead of nanomolar levels of ryanodine are used. During the initial phase of incubation, ryanodine may open RyRs and deplete the SR Ca 2ϩ content. At equilibrium, higher ryanodine concentrations may close RyRs. Therefore, although the effect of depleting the SR Ca 2ϩ content or blocking its release may be macroscopically the same (paralysis), the dual action of ryanodine makes it difficult to assess the availability of a releasable Ca 2ϩ pool at any experimental point. In addition, the recovery of SR function after treatment with ryanodine is also difficult to assess because the slow dissociation rate of ryanodine (20) makes the pharmacological effect essentially irreversible.
Despite these drawbacks, ryanodine has been an invaluable tool in defining the pharmacological profile of RyRs. The binding of [ 3 H]ryanodine is enhanced by activators of Ca 2ϩ release (M Ca 2ϩ , ATP, caffeine) and decreased by inhibitors of Ca 2ϩ release (Mg 2ϩ , ruthenium red), suggesting that the alkaloid binds to a conformationally-sensitive domain on the RyR protein (20,21). Therefore, [ 3 H]ryanodine has been used as a probe of the functional state of the channel. This approach has contributed to the isolation of the RyR itself (22)(23)(24) and to the identification of novel ligands (25)(26)(27), endogenous modulators (28,29), and regulatory proteins (30,31). Still, the detailed mechanism by which RyRs generate primary or amplified secondary Ca 2ϩ signals remains elusive, especially in cells where the permeability of intracellular Ca 2ϩ stores is also controlled by another Ca 2ϩ channel, the inositol triphosphate receptor (IP 3 R), or in cells where several RyR isoforms act simultaneously. Clearly, the elucidation of the mechanism of Ca 2ϩ release from intracellular stores will depend critically on the specificity of ligands that selectively alter a single intracellular Ca 2ϩ channel type.
From the venom of the African scorpion Pandinus imperator we isolated imperatoxin A (IpTx a ), a ϳ5-kDa peptide that specifically and with high affinity increased [ 3 H]ryanodine binding and enhanced the activity of single RyRs reconstituted in lipid bilayers (27). At concentrations well above the half-maximal effective concentration (ED 50 ) exhibited for RyRs, IpTx a did not affect other Ca 2ϩ channels or ion transporters of muscle and brain. The binding of a radiolabeled derivative of IpTx a to skeletal SR was displaced by ruthenium red with a K d of ϳ1 M, 2 a concentration similar to that required for the displacement of [ 3 H]ryanodine from purified receptors when assays are performed under high salt concentration (23). All these effects could be seen only on skeletal RyRs, suggesting that IpTx a preferentially affects the RyR1 isoform.
In this study, we carried out an in-depth analysis of the action of IpTx a on rabbit skeletal RyRs. We show that IpTx a binds directly to RyRs, or a closely associated regulatory protein, to produce a change in channel kinetics that results in an increased probability of the channel being open. By increasing the RyR sensitivity to Ca 2ϩ , IpTx a relieves the inhibition of RyRs caused by physiological concentrations of Mg 2ϩ . Furthermore, the effect of IpTx a is specific, rapid, and reversible, thereby overcoming some of the disadvantages of ryanodine. IpTx a thus provides the field with alternatives to ryanodine and may contribute unique information about modulatory domains of RyRs.

Materials-[ 3 H]Ryanodine
(60 -80 Ci/mmol) was from DuPont NEN. Brain phosphatidylserine and phosphatidylethanolamine were from Avanti Polar Lipids. Caffeine, AMP-PCP, doxorubicin, and CHAPS were from Sigma. All other reagents were high purity grade.
Purification of IpTx a -IpTx a was purified from P. imperator scorpion venom in three chromatographic steps as described previously (27). Crude venom was extracted from CO 2 -anesthetized scorpions, recovered with deionized water, and lyophilized. 100-mg batches of crude venom were dissolved in 2-3 ml of 20 mM NaOAc (pH 4.7) and applied onto a column (1.5 ϫ 125 cm) of Sephadex G-50 fine. Fractions were eluted with 20 mM NaOAc (pH 4.7) at a flow rate of 10 ml/h. Fraction III containing IpTx a was applied to a column (1 ϫ 25 cm) of carboxymethyl cellulose 32 (Pharmacia) equilibrated with 20 mM NaOAc (pH 4.7). Peptides were eluted at a flow rate of 12 ml/h with a linear gradient of 250 ml of 20 mM NaOAc (pH 4.7) and 250 ml of the same buffer containing 0.5 M NaCl. IpTx a eluted as a symmetric peak when the NaCl concentration at the top of the column reached 340 mM. IpTx a was dialyzed against deionized water in Spectrapor 3M dialysis membrane (Spectrum Medical Industries), concentrated by vacuum centrifugation, and injected into a C 18 reverse-phase high performance liquid chromatography column (Bondapac, Water Associates). IpTx a was eluted with a linear gradient of 5-80% acetonitrile in 0.1% trifluoroacetic acid run at 1 ml/min for 60 min. The purity and identity of IpTx a was confirmed by amino acid sequence analysis, as described for Na ϩ channel-blocking peptides (32). IpTx a was quantified by absorbance at 280 nm (A 280 nm ) using an extinction coefficient (⑀) ϭ 22,727. To avoid repetitive freezethaw cycles, IpTx a was stored frozen at Ϫ80°C in small (1-nmol) aliquots.
Preparation of SR and Cerebrum, Cerebellum, and Liver Microsomes-Heavy SR was prepared from rabbit white back and leg muscle, and separately from pig ventricle, using the procedure of Meissner (33). SR was stored frozen at Ϫ80°C in 0.32 M sucrose, 0.1 M KCl, and 5 mM sodium Pipes (pH 7.0). Cerebrum and cerebellum microsomes were prepared from rat. Three pentobarbital-anesthetized Sprague-Dawley rats were decapitated and their cerebrum and cerebellum removed and homogenized with five strokes of a motor-driven Teflon-glass homogenizer in 10 volumes of ice-cold 0.32 M sucrose, 0.9% NaCl, 5 mM Na-HEPES (pH 6.8). The resulting homogenate was sedimented at 1,000 ϫ g for 15 min, and the supernatant was used in binding experiments. Liver microsomes were prepared from adult Yorkshire pigs by a procedure similar to that used for brain microsomes, with the exception that the homogenization was carried out in a Waring blender for 1 min at high speed. Homogenizations of all tissues were carried out in the presence of the protease inhibitors pepstatin (1 M), leupeptin (1 M), and phenylmethylsulfonyl fluoride (100 M).
[ 3 H]Ryanodine Binding Assay-[ 3 H]Ryanodine binding to rabbit skeletal SR and other tissue homogenates described in the text was carried out as described previously (26,27). Briefly, the standard incubation medium contained 0.2 M KCl, 10 mM Na-HEPES (pH 7.2), 1 mM EGTA, and CaCl 2 necessary to set free Ca 2ϩ in the range of 1 nM to 100 M. Ca 2ϩ -EGTA ratios were calculated using the stability constants of Fabiato (34). [ 3 H]Ryanodine was diluted directly into the incubation medium to achieve a final concentration of 1-40 nM. Protein concentration was in the range of 0.4 -0.6 mg/ml and was determined by the Bradford method. Unless otherwise indicated, incubations lasted 90 min at 36°C. Samples (0.1 ml) were always run in duplicate, filtered onto glass fiber filters (Whatman GF/B or GF/C), and washed twice with 5 ml of cold water using a Brandel M-24R cell harvester (Gaithersburg, MD). The filters were placed in scintillation vials, 5 ml of liquid scintillation mixture added, and the retained radioactivity measured in a Beckman LS-5000 TD ␤-counter. The specific binding was defined as the difference between the binding in the absence (total binding) and presence (nonspecific binding) of 10 M unlabeled ryanodine. Under these conditions, the average value of nonspecific binding amounted to 10 -20% of the total binding of 7 nM [ 3 H]ryanodine. Equilibrium binding data from saturation curves were fitted to a one-site model, and the dissociation constant (K D ) and maximal receptor density (B max ) were determined by nonlinear regression analysis using the computer program Origin (Microcal Inc., Northampton, MA).
Solubilization and Purification of Skeletal RyR-RyR was purified from CHAPS-solubilized skeletal SR on sucrose density gradients. SR vesicles were resuspended in 0.5% CHAPS, 1 M NaCl, 40 mM Tris maleate (pH 7.2), plus protease inhibitors described above. The homogenate was incubated 60 min at 0°C and spun at 38,000 rpm in a Beckman 45 Ti rotor for 30 min. Solubilization of SR under these conditions was Ͼ80%. 2-ml aliquots of supernatant were layered on top of a 32-ml 5-20% (w/v) linear sucrose gradient containing 0.3 M NaCl, 40 mM Tris maleate (pH 7.2), 100 M CaCl 2 , and 0.1% CHAPS. Gradients were centrifuged at 24,000 rpm in a SW-28 rotor for 15 h at 4°C. Gradient fractions were monitored for protein content, SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis), and [ 3 H]ryanodine binding activity.
Planar Bilayer Technique-A phospholipid bilayer of phosphatidylethanolamine:phosphatidylserine (1:1 dissolved in n-decane to 25 mg/ml) was "painted" with a Teflon rod across an aperture of ϳ300-m diameter in a delrin cup. The cis chamber was the voltage control side connected to the head stage of a 200A Axopatch amplifier, while the trans side was held at virtual ground. The capacitance of the bilayer was constantly monitored with a Ϯ5 mV/10 ms square pulse. Incorporation of RyRs more often occurred in bilayers of 200 -400 pico-Faraday capacitance. The cis (1.0 ml) and trans (3.0 ml) chambers were initially filled with 50 mM cesium methanesulfonate and 10 mM Tris-HEPES (pH 7.2). After bilayer formation, an asymmetric cesium methanesulfonate gradient (300 mM cis/50 mM trans) was established. A calcium: EGTA admixture was then added to the cis chamber from a 100-fold stock to a final concentration 0.994:1.0 mM, respectively. The calculated free [Ca 2ϩ ] was 10 M, which was verified with a Ca 2ϩ -sensitive electrode using the Molecular Probes (Eugene, OR) Ca 2ϩ calibration kit (range 10 nM to 50 M). The SR vesicles were added to the cis chamber, which corresponded to the cytoplasmic side of the SR, while the trans side corresponded to the luminal side. After visualization of channels, 2 H. Valdivia, B. Block, and O. Fuentes, unpublished data.
Cs ϩ in the trans chamber was raised to 300 mM to dissipate the chemical gradient. This maneuver greatly improved bilayer stability and prevented further vesicle insertion, thus allowing us to manipulate the same channel(s) over a long time period (20 -40 min). For each condition, single channel data were collected at steady voltages (ϩ30 and Ϫ30 mV) for 2-4 min. Channel activity was recorded with a 16-bit VCR-based acquisition and storage system at 10 kHz sampling rate. Signals were analyzed after filtering with an 8-pole Bessel filter at a sampling frequency of 1.5-2 kHz. Data acquisition and analysis were done with Axon Instruments (Burlingame, CA) software and hardware (pClamp v6.0, Digidata 1200 AD/DA interface).
Determination of Ca 2ϩ -ATPase Activity-Ca 2ϩ -ATPase activity of CHAPS-solubilized skeletal SR fractions was measured with malachite green by the colorimetric method of Lanzetta et al. (50), which quantitatively determines nmol of inorganic phosphate (P i ) generated by hydrolysis of ATP. The Ca 2ϩ -independent generation of P i was less than 10% of the total P i and has been substracted.  (20,26) to ensure a relatively linear relationship between the number of occupied receptors and the concentration of free ligand. Under these conditions, we could clearly distinguish between the effect of an activator from that of an inhibitor. Fig. 1A shows that IpTx a increased the binding of [ 3 H]ryanodine to skeletal SR. Binding increased to ϳ300 and ϳ500% of control at an IpTx a concentration of 12 and 100 nM, respectively. The latter concentration is 10-fold higher than the apparent K D of the IpTx a ⅐RyR complex (ϳ10 nM) (Fig. 5,Ref. 27). In contrast, IpTx a had negligible effects on the binding of [ 3 H]ryanodine to cardiac SR and cerebrum or liver microsomes. The common feature among the latter tissues is the minimal or altogether absent expression of RyR1 (skeletal-type) (11,35). However, RyR1 is expressed in cerebellar Purkinje cells (12) and using cerebellum microsomes, IpTx a produced a small but significant activation. Binding increased to 128 and 152% at 12 and 100 nM IpTx a , respectively. Since the effectiveness of IpTx a to activate binding correlated directly with the amount of RyR1 expression, the results above suggest that IpTx a acts preferentially on skeletal-type RyR isoforms. Fig. 1B shows the time course of the effect of IpTx a (50 nM) on rabbit skeletal RyRs. The stimulation of [ 3 H]ryanodine binding by IpTx a was evident after short periods of incubation and approached steady-state at ϳ100 min. k a , the association rate constants, were 0.0005 Ϯ 0.0001 (control) and 0.0007 Ϯ 0.00015 nM Ϫ1 min Ϫ1 (ϩIpTx a ). At equilibrium, the number of receptors occupied by [ 3 H]ryanodine was 1.13 (control) and 3.18 pmol/mg (ϩIpTx a ). Thus, although IpTx a did not accelerate significantly the rate of binding, it did increase the number of receptor sites available for binding. The intrinsically slow association kinetics of [ 3 H]ryanodine precluded the accurate estimation of IpTx a effects at early times, but the substantial increase of binding detected at 3 min suggested that IpTx a had a fast onset of effect.

Selective Activation of Skeletal RyRs by IpTx
Effect of IpTx a on Purified RyRs-To determine if IpTx a increased [ 3 H]ryanodine binding through a direct interaction with the RyR or a closely associated regulatory protein, we purified RyRs from skeletal SR vesicles and tested the effect of IpTx a in binding experiments. Using a modification of the sucrose gradient method of Lai et al. (24), we identified the receptor as a ϳ400-kDa band ( Fig. 2A, lane 2) highly enriched in fractions containing 14 -17% (w/v) sucrose (Fig. 2B). Since this method is likely to preserve the tight association between RyR1 and FKBP12, a 12-kDa protein that binds the immunosuppressant drug FK506 and regulates RyR activity (31), we attribute our failure to detect the latter to the fact that proteins smaller than Ϸ20 kDa migrated with the tracking dye in our SDS-PAGE. Other putative regulatory proteins of RyR1 such as triadin (ϳ95 kDa) and calsequestrin (ϳ42 kDa) were not detected in our gel. Only a ϳ106-kDa band (labeled with asterisk), positively identified as the Ca 2ϩ -ATPase of SR by a functional assay (see figure legend), migrated with the RyR and was apparent in the gel. This band was concentrated in lighter sucrose fractions (not shown) but neither [ 3 H]ryanodine binding nor IpTx a effects were found there, indicating that it did not contribute significantly to the interaction of IpTx a with the RyR. Fig. 2C (open circles) shows that the purified receptor displayed high affinity for [ 3 H]ryanodine with K D ϭ 6.1 nM and B max ϭ ϳ30 pmol/mg. IpTx a (10 nM, filled circles) increased binding at all [ 3 H]ryanodine concentrations. The main parameter affected by IpTx a was B max , which increased to ϳ60 pmol/ mg. K D remained essentially unchanged (5.7 nM). Therefore, the interaction of IpTx a with the purified receptor was noncompetitive in nature. Altogether, these results strongly suggested that IpTx a increased [ 3 H]ryanodine binding by a direct association with RyRs or the tightly associated regulatory protein FKBP12.
where B is the specific binding at time t, B eq is the amount of binding at equilibrium, and t1 ⁄2 is the half-time to reach equilibrium.
Single-channel Effects of IpTx a -To gain insight on the manner in which IpTx a alters RyR activity, we reconstituted rabbit skeletal RyRs in planar lipid bilayers in the absence of ryanodine as described (26,27). Solutions in both chambers bathing the planar bilayer were 300 mM Cs ϩ methanesulfonate and 10 mM Tris-HEPES pH 7.2. Free [Ca 2ϩ ] in the cis (cytosolic) side of the channel was 10 M. Cs ϩ , instead of Ca 2ϩ , was chosen as the charge carrier to precisely control [Ca 2ϩ ] around the channel, to increase the channel conductance (g Cs ϩ/g Ca 2ϩ ϭ 2; (14)) and to avoid interference from K ϩ channels present in the SR membrane. Cl Ϫ channels were blocked by replacing chloride with the impermeant anion methanesulfonate. Fig. 3 shows traces from continuous recordings at Ϫ30 mV holding potential in the absence ("control") and the presence of 10 nM IpTx a ("ϩIpTx a "). Channel activity was monitored for over 100 s in each condition and the mean open probability (P o ) obtained from amplitude histograms was 0.18 Ϯ 0.05 in control and 0.48 Ϯ 0.10 following addition of IpTx a , i.e. a 2.6-fold increment. The most significant kinetic effects of IpTx a were an increase in the bursting frequency and a decrease in the mean closed time (see below). No significant change in unitary conductance or mean open time was detected. As indicated by the amplitude histograms (Fig.  3, right panels), IpTx a produced an increase in the frequency of open events which significantly contributed to an increase in P o .
Rapid and Reversible Effect of IpTx a -The effect of IpTx a on skeletal RyRs was rapid and reversible. Fig. 4 shows traces from continuous recordings at Ϫ30 mV holding potential prior to and after the cis addition of 10 nM IpTx a , and after extensively washing the cis chamber with a peptide-free solution. For the channel depicted in Fig. 4, activity was monitored for over 100 s in each condition and had a P o ϭ 0.14 in control that increased to 0.36 following addition of IpTx a . After washing, P o dropped again to control levels (0.12) and remained at this level for the duration of the experiment. The complete diary of channel activity under each condition is presented in the right panel of Fig. 4. The length of each bar represents P o during consecutive sweeps. Each sweep had a duration of 6 s. Empty spaces between clusters of bars represent gaps in recording due to stirring or perfusion, but the total elapsed time is indicated in the x axis. The arrow at time 0 marks the addition of IpTx a . As observed, the increased frequency of openings following addition of the peptide gave rise to bursts of activity that resulted in sweeps of high P o . The appearance of a cluster of sweeps of P o consistently higher than those in control was taken as an indication that the effect of IpTx a was fully established. Using this criterion, the lag time for the IpTx a effect on this particular channel was ϳ40 s. Indeed, the average lag time for 3 independent observations was Յ1 min. Since these clusters of high activity disappeared after washing IpTx a (segment labeled "wash"), we took that as an indication that the effect was reversible. Thus, unlike ryanodine, IpTx a displays both a rapid onset and a reversible effect.
Dose-response Relation of IpTx a Effect and Kinetic Alterations Resulting from IpTx a - Fig. 5 shows representative current records and close time histograms of a single RyR recorded in the presence of 0, 3, 10, 30, and 100 nM IpTx a . The traces show that channel activity was increased by raising the cytosolic concentration of IpTx a in a dose-dependent manner. In the absence of IpTx a , open events could be fitted with two exponen-tials, open1 ϭ 0.64 ms and open2 ϭ 4.6 ms. At the maximal IpTx a concentration tested, the open lifetimes had values open1 ϭ 0.71 and open2 ϭ 5.8 ms (data not shown). Thus, IpTx a increased the open lifetimes by a margin per se insufficient to explain the noticeable increase in P o . However, a noticeable effect of IpTx a could be found in the analysis of closing events. In the presence of IpTx a , long closings normally found in control were substituted by numerous brief closings. A quantitative description of this effect is presented in the closed time histograms (Fig. 5B). In control, the closed time histogram could be fitted by a triple exponential, closed1 ϭ 0.8 ms, closed2 ϭ 1.7 ms, and closed3 ϭ 7.5 ms. The relative proportion of each of these components was 62, 27, and 11%, respectively. In the presence of 100 nM IpTx a (Fig. 5B, ϩIpTx a ) the histogram was biexponential with close1 ϭ 0.7 ms and closed2 ϭ 1.6 ms. However, the proportion of closed1 was 95%. Thus, IpTx a totally eliminated close3 and substantially decreased close2 . The overall result was an increase in P o due to an increase in the number of open events per unit time.

FIG. 4. Onset and reversibility of IpTx a effect.
A, representative single channel recordings of a skeletal RyR activated by 10 M Ca 2ϩ before (control) and following addition of 10 nM IpTx a to the cis chamber. Traces labeled "wash" were taken from the same channel after extensive perfusion of the cis chamber with a peptide-free solution. B, 6 min of continuous records spanning the whole experiment were divided into intervals of 6 s; P o in each interval is plotted as a bar of length 0 -1. Empty spaces are gaps in recording due to addition of IpTx a or perfusion. Average P o for the channel shown was 0.14 (control), 0.36 (ϩ IpTx a ), and 0.12 (washing).
creased P o consistently and in a dose-response manner. However, P o did not reach ϳ1.0 even at saturating concentrations of the peptide. Data were fitted with the equation:  (36,37). The EC 50 for the activation of [ 3 H]ryanodine binding by Ca 2ϩ (ascending limb of the curve) was 3 M, while that for the inhibition was 250 M. In the presence of IpTx a (filled circles) the binding curve also had a bell-shape but was dramatically augmented in absolute values. The augmentation of [ 3 H]ryanodine binding produced by IpTx a increased with [Ca 2ϩ ]. At pCa ϭ 7, 6, and 5, the net IpTx a -stimulated binding (ϩIpTx a Ϫ control) was 0.4 Ϯ 0.06, 1.9 Ϯ 0.2, and 2.9 Ϯ 0.31 pmol/mg (mean Ϯ S.E., n ϭ 4). Furthermore, the threshold for activation was shifted by IpTx a to pCa 8 and the optimal binding to pCa 5. This resulted in a shift of the ascending limb of the curve toward lower [Ca 2ϩ ], with a midpoint ϭ 0.8 M. On the other hand, the midpoint for the descending limb of the curve remained unchanged (280 M). These results suggest that IpTx a exerts its stimulatory effect by sensitizing RyRs to Ca 2ϩ .
In the absence of IpTx a , the Ca 2ϩ dependence of [ 3 H]ryanodine binding to cardiac RyRs (Fig. 6B) also had a threshold for detection at ϳpCa 7 and a maximum at pCa 5. The EC 50 for activation of binding by Ca 2ϩ was also ϳ3 M. However, a fundamental difference with skeletal receptors was the lack of inactivation produced by 1 mM [Ca 2ϩ ]. Binding decreased only by ϳ20% with respect to maximum in cardiac RyRs while it decreased by ϳ90% in skeletal RyRs. These results are consistent with those of Chu et al. (37), who failed to detect inactiva-tion of cardiac RyRs at 1 mM [Ca 2ϩ ]. In contrast to the dramatic effect on skeletal RyRs, the effect of IpTx a on cardiac RyRs was erratic, showing no effect at pCa 8 and 7, a modest stimulation at pCa 6 (from 0.093 Ϯ 0.015 to 0.23 Ϯ 0.08 pmol/mg protein, n ϭ 3; values significantly different at p ϭ 0.05 by one-way analysis of variance), and a modest, but statistically insignificant inhibition at pCa 5 to 3. Therefore, the activation induced by IpTx a on cardiac RyRs occurred at one [Ca 2ϩ ] and constituted a small fraction of that observed on skeletal RyRs. Furthermore, in the presence of a low [Ca 2ϩ ] (100 nM), a radiolabeled derivative of IpTx a failed to bind to cardiac SR, but it did bind to skeletal SR (38). Taken together, the results indicate that the activating effect of IpTx a on cardiac RyRs is restricted to a well defined range of [Ca 2ϩ ] and is weak in absolute quantity. Given these restrictions and the fact that intracellular [Ca 2ϩ ] is a dynamic parameter of many cells that express more than one RyR isoform, we propose that cardiac RyRs are poor substrates for IpTx a . By contrast, the broad Ca 2ϩ dependence and conspicuous response of skeletal RyRs make these receptor isoforms the preferred target of IpTx a .
Interaction of IpTx a with Other Modulators of RyRs-To assess a potential interaction between IpTx a and other modulators of RyRs, we bound [ 3 H]ryanodine to skeletal RyRs in the presence of several agonists and inhibitors of Ca 2ϩ release and tested the effect of IpTx a (Fig. 7). The binding observed in the absence of modulators and IpTx a was defined as 100% and all results are presented relative to this value. In the absence of modulators (Control), IpTx a (50 nM) increased binding 360 Ϯ 40%. When added alone, doxorubicin (300 M) and AMP-PCP (3 mM) increased binding 210 Ϯ 24% and 189 Ϯ 17%, respectively. Thus, under our experimental conditions, doxorubicin and AMP-PCP were less effective than IpTx a to activate RyRs. When doxorubicin was combined with IpTx a , binding increased dramatically to 825 Ϯ 75%. However, in the case of AMP-PCP plus IpTx a , binding increased only to 470 Ϯ 38%. Since doxorubicin and caffeine share a common binding site (39), these results suggest that the IpTx a -binding site may interact allosterically with the caffeine-binding site to produce a synergistic effect; by contrast, the nucleotide-binding site appears to be independent of the IpTx a -binding site (see further below).
When added individually, ruthenium red (3 M), calmodulin (3 M), and free Mg 2ϩ (1 mM) decreased binding to 8 Ϯ 2%, 49 Ϯ 11%, and 6 Ϯ 3%, respectively. When IpTx a was added in tandem with each of these inhibitors, not only was binding restored but increased above control, to 180 Ϯ 20%, 230 Ϯ 28%, and 110 Ϯ 20%, respectively. Therefore, the interaction of IpTx a with RyRs relieves the inhibitory effect of ruthenium red, CaM, and Mg 2ϩ , and evokes a new level of RyR activity that is higher than that observed in the absence of the inhibitors.
Relation of the IpTx a -binding Site to the Caffeine-and Nucleotide-binding Sites-To discern in detail the interaction of FIG. 6. Effect of IpTx a on the Ca 2؉ dependence of [ 3 H]ryanodine binding. The binding of [ 3 H]ryanodine to rabbit skeletal (A) and porcine cardiac (B) SR was performed as described in the legend to Fig. 1, with the exception that the incubation medium was supplemented with 1 mM EGTA and several CaCl 2 concentrations to yield the indicated [free Ca 2ϩ ]. Ca 2ϩ :EGTA ratios were calculated with a computer program using the stability constants given by Fabiato (34). IpTx a (100 nM) was present throughout the incubation period. Data points are the mean (ϮS.E.) of four independent determinations. Smooth lines linking data points have no theoretical meaning.
IpTx a -, caffeine-, and nucleotide-binding sites in the receptor complex, we tested the effect of IpTx a in the absence and presence of incremental concentrations of caffeine and AMP-PCP. Caffeine activates RyRs at low (submicromolar to micromolar) [Ca 2ϩ ] (40,41), while AMP-PCP does it independent of the [Ca 2ϩ ]. Fig. 8A shows the interaction of caffeine and IpTx a . At pCa 7, binding was 0.061 pmol/mg in the absence of IpTx a and caffeine and increased to 0.52 Ϯ 0.04 pmol/mg (n ϭ 3) in the presence of 30 mM caffeine (open circles). This effect was rather modest, but is consistent with previous determinations carried out at low [Ca 2ϩ ] (41). In the presence of 50 nM IpTx a (filled circles), binding was 0.75 Ϯ 0.06 pmol/mg in the absence of caffeine and increased to 2.41 Ϯ 0.32 pmol/mg in the presence of 30 mM caffeine, i.e. a gain of Ͼ1.6 pmol/mg. Thus, the binding increment induced by the combined addition of caffeine and IpTx a was substantially larger than the sum of that induced by caffeine or IpTx a alone. This synergistic effect suggests a cooperative interaction between the IpTx a -and the caffeine-binding sites.
In contrast, the nucleotide-binding site appeared to be independent of the IpTx a -binding site. Fig. 8B shows the effect of incremental concentrations of AMP-PCP, a non-hydrolyzable analog of ATP, on the binding of [ 3 H]ryanodine to skeletal SR in the absence and presence of IpTx a . In these experiments, AMP-PCP was preferred over ATP to avoid activation of both the Ca 2ϩ -ATPase of SR (which may alter the [Ca 2ϩ ] of the incubation medium) and endogenous protein kinases (which may alter the activity of the RyR (42)). In the absence of IpTx a , AMP-PCP increased the binding of IpTx a Relieves the Inhibitory Effect of Mg 2ϩ -Mg 2ϩ is an important modulator of RyRs (43). When present at a physiological concentration (free ϳ1 mM), Mg 2ϩ severely depresses the RyR response to transient and steady concentrations of Ca 2ϩ (44). We therefore tested the ability of IpTx a to overcome the inhibitory effect of Mg 2ϩ . The Ca 2ϩ dependence of [ 3 H]ryanodine binding in the absence of Mg 2ϩ and IpTx a had a threshold for activation at pCa 7 and was optimal at pCa 5 ( Fig.  9, open circles; see also Fig. 6). 1 mM free Mg 2ϩ (filled circles) shifted the threshold for activation by Ca 2ϩ to pCa 6 (1 order of magnitude higher than control) and decreased the maximal binding to ϳ15% of control. Thus, Mg 2ϩ decreased the RyR response to Ca 2ϩ and shifted the Ca 2ϩ dependence of activation to higher [Ca 2ϩ ]. When Mg 2ϩ was added in tandem with 100 nM IpTx a (filled squares), the threshold for activation was restored to pCa 7, the binding curve was again bell-shaped and centered at pCa 5 to 4, and the absolute binding had values similar to control. At pCa 4 and 3, even in the presence of Mg 2ϩ , IpTx a was able to increase binding above control. Thus, IpTx a overcame the Mg 2ϩ inhibition to produce a substantial activation of RyRs. The capacity of IpTx a to relieve the Mg 2ϩ inhibition and restore Ca 2ϩ sensitivity to RyRs is reminiscent of the effect of phosphorylation. Both IpTx a and phosphorylation by CaMKII and protein kinase A increase the P o of RyRs when the concentration of Mg 2ϩ is whithin physiological levels (42). DISCUSSION We have used two different parallel indicators of channel function (i.e. [ 3 H]ryanodine binding and direct single-channel recordings), to provide complementary lines of evidence for a selective, fast and reversible activation of skeletal RyRs by IpTx a . The near identity of dose-response curves for the IpTx a effect on [ 3 H]ryanodine binding (27) and single-channel recordings (Fig. 5) lends credence to the assumption that [ 3 H]ryanodine binding is proportional to P o , since it is difficult to have a systematic error in both assays and maintain the similarity shown.
The mechanism involved in the activation of skeletal RyRs by IpTx a is qualitatively similar to that produced by caffeine (39) and the eudostomin-derivative MBDE (25). The latter compounds increase the affinity of the RyR for Ca 2ϩ ("Ca 2ϩ sensitization"), an effect that is reflected as a selective shift of the ascending limb of the Ca 2ϩ dependence of P o curve to the left. The descending limb of the curve, presumably reflecting binding of Ca 2ϩ to an inactivation site, is unaffected by these compounds. The inactivation of RyRs by high Ca 2ϩ is only observed in skeletal RyR (37), and its physiological significance has been questioned (45). At nanomolar concentrations, IpTx a increased the amplitude and displaced the midpoint of the ascending limb of the curve from 3 M to 0.8 M Ca 2ϩ . The midpoint of the inactivation curve remained unchanged (Fig.  6A). Therefore, given the resemblance between the effect of caffeine and IpTx a , it may be suggested that they both activate RyRs by binding to a common receptor site. However, two lines of evidence indicated that this is probably not the case. First, in the presence of a saturating concentration of caffeine, IpTx a produced an additional 6-fold increase in [ 3 H]ryanodine binding (Fig. 8), an effect that would not be expected if occupation of the IpTx a site by caffeine had occurred. Second, caffeine activates both the cardiac and skeletal RyR by a mechanism similar to activation of skeletal RyRs by IpTx a . If IpTx a occupied the same site as caffeine, IpTx a would also be expected to activate cardiac and skeletal RyRs alike.
Following the same argument, the nucleotide-binding site appeared to be different from and non-cooperative with the IpTx a site. This is evident in binding experiments where the combined effect of the two agonists is simply the sum of each agonist acting separately (Fig. 8). Furthermore, the nucleotide effect spans the ascending and descending limbs of the Ca 2ϩ dependence of [ 3 H]ryanodine binding curve (10), as opposed to the selective effect of IpTx a on the ascending limb of the curve (Fig. 6). In addition, single-channel experiments reveal that the kinetic parameters affected by each agonist are different. While nucleotides activate RyRs by increasing the frequency and duration of open events (46), IpTx a does it by decreasing the duration and proportion of the closed lifetimes (Fig. 5) without affecting the duration of the open lifetimes. This different mechanism of action highlights a distinctive functional output from each agonist site.
Since IpTx a fails to substantially activate cardiac RyRs (as well as cerebrum and liver RyRs) (Figs. 1A and 6B, see also Ref. 27), this in all likelihood reflects the existence of a distinct modulatory site for IpTx a on RyR1. Thus, we conclude that IpTx a may be used as a selective activator of skeletal-type RyRs (RyR1). The significant stimulation obtained with cerebellum microsomes (as opposed to cerebrum microsomes) is also consistent with this notion, since there seems to be a relatively large proportion of RyR1 in cerebellar Purkinje cells. But even so, we cannot rule out that the IpTx a -binding site is altogether absent in cardiac (RyR2) or brain (RyR3) isoforms. The Ca 2ϩ dependence of [ 3 H]ryanodine binding to cardiac RyRs (Fig. 6B) was, albeit modestly, altered by IpTx a . This suggests that IpTx a may interact, directly or indirectly, with cardiac RyRs; however, the functional consequence of such interaction is almost null and occurs only over a narrow range of [Ca 2ϩ ]. By contrast, IpTx a was highly effective in activating skeletal RyRs over a wide range of [Ca 2ϩ ] (Fig. 6A). Thus, in systems where intracellular [Ca 2ϩ ] fluctuates rapidly due to the concurrent action of several RyR isoforms, as it does in non-mammalian skeletal muscle, IpTx a is likely to target skeletal-type RyRs and produce an exclusive and substantial modification of the parameters directly controlled by these isoforms.
In both cardiac and skeletal RyRs, binding was maximum when [Ca 2ϩ ] was ϳ10 -50 M. At this optimum [Ca 2ϩ ], the P o of single RyR channels was far from being 1.0 (P o ϳ0.2, Figs. [3][4][5], suggesting that Ca 2ϩ does not fully activate the channel when it is present as the only activating ligand. However, it has been established that if a fast and sustained [Ca 2ϩ ] step is applied, channel activity increases to ϳ1.0 and then spontaneously decays to a new steady-state level (47,44). The spontaneous decay of RyR activity in the presence of a sustained [Ca 2ϩ ] has been termed "adaptation" (47) by analogy to other receptors that decrease their level of response despite persistent occupancy by the agonist that initiated the response (48). In the case of RyRs, adaptation may be explained by the rapid association of Ca 2ϩ to a binding site (O-domain) that leads to channel opening, followed by the slower binding of Ca 2ϩ to another binding site (A-domain) responsible for channel closing or adaptation (49). Thus, it is conceivable that the low P o of RyRs observed under our experimental conditions is due to the absence of the transient peak of activity evoked by a fast increment of [Ca 2ϩ ]. Since [Ca 2ϩ ] remained constant during the course of our experiments, we effectively tracked the activity elicited by the binding of Ca 2ϩ to the A-domain, which is indeed proportional to the stationary Ca 2ϩ concentration (44) but saturates at a P o Ͻ 1.0 (49). Within this scheme, the IpTx a -induced increase of channel activity (Fig. 5) and [ 3 H]ryanodine binding (Fig. 6)   range is between 0.1 and 10 -100 M [Ca 2ϩ ] (49). However, this conclusion awaits further experiments involving fast application of Ca 2ϩ to measure directly the parameters of these Ca 2ϩbindings domains and the effect of IpTx a on the A-site.
The intrinsic properties of IpTx a and ryanodine confer to each ligand a unique set of attributes that may both restrict and promote their use as probes of RyRs. The activating effect of IpTx a on skeletal RyRs is fast and fully reversible (Fig. 4). In this respect, IpTx a is more advantageous to physiologists than ryanodine, which is too slow to bind and dissociate from its effector site to attain equilibrium during the course of experiments with intact cells. On the other hand, it was precisely the slow dissociation of ryanodine that facilitated the purification of the receptor (24). As a natural peptide, IpTx a is suitable for iodination, and for attachment of cross-linkers for covalent binding to the receptor protein, thus offering the prospect of receptor tagging and mapping. The preferential effect of IpTx a for the skeletal RyR may be extremely useful to dissect the role of RyR1 in systems where complex Ca 2ϩ waveforms represent the functional output of several RyR isoforms acting in concert. This very feature, however, may limit the use of IpTx a in a wide variety of cells where the levels of expression of RyR1 is low or totally absent. Finally, the results provided here do not allow us to safely ascertain if the binding of IpTx a , like that of ryanodine, is dependent on the conformational state of the receptor. Although the effect of IpTx a was potentiated by ligands that open RyRs such as Ca 2ϩ (Fig. 6) and caffeine (Fig.  8), an unequivocal demonstration that IpTx a binds to the open receptor must be provided by experiments where IpTx a is directly used as the radiolabeled ligand.
Amino acid analysis indicates that IpTx a is a positivelycharged ϳ5-kDa peptide rich in basic residues. 3 Thus, it is important to note that although IpTx a may contribute considerably to the understanding of the structure-function relationship of the RyR in isolated SR vesicles or with the purified receptor, its usefulness in intact cells may be limited by its incapacity to permeate the external membrane. In this respect, the lipophylic nature of ryanodine is clearly advantageous. However, the possibility that IpTx a cannot reach its receptor raises several questions. Is the skeletal RyR its intended target? And if it is not, how can IpTx a display such an exquisite selectivity and high affinity? Is IpTx a , like Escherichia coli enterotoxin, allowed into the cell by another molecular component of the venom? Or are there instances yet unknown where the RyR (or a structurally related protein) is expressed in the external membrane? In all cases the possibilities are intriguing.