Activation of Ryanodine Receptors by Imperatoxin A and a Peptide Segment of the II-III Loop of the Dihydropyridine Receptor*

Excitation-contraction coupling in skeletal muscle is believed to be triggered by direct protein-protein interactions between the sarcolemmal dihydropyridine-sensitive Ca2+ channel and the Ca2+ release channel/ryanodine receptor (RyR) of sarcoplasmic reticulum. A 138-amino acid cytoplasmic loop between repeats II and III of the α1 subunit of the skeletal dihydropyridine receptor (the II-III loop) interacts with a region of the RyR to elicit Ca2+ release. In addition, small segments (10–20 amino acid residues) of the II-III loop retain the capacity to activate Ca2+ release. Imperatoxin A, a 33-amino acid peptide from the scorpion Pandinus imperator, binds directly to the RyR and displays structural and functional homology with an activating segment of the II-III loop (Glu666-Leu690). Mutations in a structural motif composed of a cluster of basic amino acids followed by Ser or Thr dramatically reduce or completely abolish the capacity of the peptides to activate RyRs. Thus, the Imperatoxin A-RyR interaction mimics critical molecular characteristics of the II-III loop-RyR interaction and may be a useful tool to elucidate the molecular mechanism that couples membrane depolarization to sarcoplasmic reticulum Ca2+ release in vivo.

Compelling evidence indicates that the skeletal DHPR subtype is indispensable to elicit a Ca 2ϩ -independent (skeletaltype) contraction (6,7) and that the 138-amino acid cytoplasmic loop between repeats II and III of the ␣ 1 subunit participates in this process (6,8). In experiments with isolated peptides, the II-III loop activates purified RyRs (9), and a small fragment of the II-III loop (Thr 671 -Leu 690 ) induces Ca 2ϩ release from SR vesicles (10). In dysgenic myotubes, skeletal-type E-C coupling is partially restored by a chimeric DHPR that is entirely cardiac except for a short segment of skeletal II-III loop (Phe 725 -Pro 742 ) (8). Although apparently contradictory in the identity of the activating region, these results suggest that specific domains of the II-III loop directly interact with the RyR to change its conformational state and produce Ca 2ϩ release. Therefore, in skeletal muscle, the II-III loop stands as the strongest candidate among regions of the DHPR to bind to RyRs. However, the precise amino acid residues of the II-III loop that trigger Ca 2ϩ release remain unknown. Furthermore, other DHPR segments (11) or subunits (12) have not been discarded as points of contact.
We have previously shown that Imperatoxin A (IpTx a ), a 33-amino acid peptide from the scorpion Pandinus imperator, is a high-affinity activator of RyRs (13,14). The biological significance of IpTx a is unknown, because the apparent target for this membrane-impermeable peptide is located intracellularly. Because some peptide toxins activate intracellular signaling pathways by mimicking surface receptors (15,16), we tested the hypothesis that IpTx a activates RyRs by mimicking a domain of the DHPR that is critical to trigger Ca 2ϩ release. We found that IpTx a and a synthetic peptide with an amino acid sequence corresponding to a segment of the II-III loop (Glu 666 -Leu 690 ) (10) activate RyRs in a similar manner and appear to compete for a common binding site on the channel protein. Both peptides bind to RyRs via a structural domain consisting of a cluster of basic amino acids (Arg 681 -Lys 685 of the II-III loop and Lys 19 -Arg 24 of IpTx a ) followed by a hydroxylated amino acid (Ser 687 of the II-III loop and Thr 27 of IpTx a ). Thus, IpTx a presents an interesting case of toxin mimicry of effector proteins that may be used to identify regions of the RyR that trigger Ca 2ϩ release. If the peptide segment emulated by IpTx a is an actual participant in the DHPR/RyR interaction, IpTx a may also be exploited to identify regions of the RyR involved in E-C coupling. 3 H]Ryanodine (60 -80 Ci/mmol) was from NEN Life Science Products, agelenin and Tx2-9 were from The Peptide Institute, Inc. (Osaka, Japan), bovine brain phosphatidylethanolamine and phosphatidylserine were from Avanti Polar Lipids (Birmingham, AL), and Fmoc-amino acids were from Applied Biosystems. Polyclonal skeletal RyR antibody was from Upstate Biotechnology. Peroxidase-conjugated secondary antibody and -conotoxin were from Calbiochem. The chemiluminescence detection kit was from Boehringer Mannheim. Pre-cast linear gradient polyacrylamide gels were from Bio-Rad. All other re-* This work was supported by National Institutes of Health Grants HL55438 and PO1 HL47053 (to H. H. V. and J. W. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
[ 3 H]Ryanodine and 125 I-IpTx a Binding Assay-[ 3 H]Ryanodine (7 nM) was incubated for 90 min at 36°C with 40 -50 g of rabbit skeletal SR vesicles in medium containing 0.2 M KCl, 10 M CaCl 2 , and 10 mM Na-Hepes (pH 7.2) in the absence and presence of peptides. Free ligand and bound ligand were separated by rapid filtration on Whatman GF/B glass fiber filters, as described previously (13,14). Native IpTx a (10-g batches) was purified according to established procedures (13,14) and iodinated to a specific activity of 60 -80 Ci/mmol with the Bolton-Hunter© method following the specifications of the manufacturer (New England Nuclear). The binding of 125 I-IpTx a to skeletal SR and Chinese hamster ovary (CHO) cell homogenates was performed under conditions identical to those described for [ 3 H]ryanodine, except that the protein concentration was 0.1-0.2 mg/ml in the case of CHO cells. B max and K D of the 125 I-IpTx a -receptor complex were obtained by fitting data points with the following equation: B ϭ B max ϫ 125 I-IpTx a /(K D ϩ 125 I-IpTx a ), where B is the specific binding of 125 I-IpTx a .
Transfection of CHO Cells with RyR-CHO cells were transfected by lipofection with plasmid pRRS11, the rabbit skeletal muscle RyR (RyR1), as described previously (17). Expression of the RyR was confirmed by immunoblot analysis using monoclonal antibodies against the skeletal RyR and by [ 3 H]ryanodine binding. Control and transfected cells were homogenized in 500 mM sucrose, 1 mM EGTA, and 10 mM Hepes-Tris (pH 7.4) and spun at 44,000 ϫ g for 30 min (17). The pellet was recovered and used for [ 3 H]ryanodine binding experiments.
Synthesis of Peptides-Linear analogs of IpTx a and the II-III loop were synthesized by the solid-phase methodology with Fmoc amino acids in an automated peptide synthesizer and subjected to the same cyclization and HPLC purification method as described previously (14). Analytical HPLC, amino acid analysis, and mass spectrometry confirmed the structure and the purity of the synthetic peptides. Photoactivatable IpTx a was prepared by inserting p-benzoyl-phenylalanine, a photoactivatable cross-linker (18), in place of Leu 7 during the synthesis of IpTx a . Photoactivatable IpTx a was subjected to the same cyclization and HPLC purification method as described for synthetic IpTx a (14).
Planar Bilayer Recording of RyRs-Recording of single RyR in lipid bilayers was performed as described previously (13,19). Single channel data were collected at steady voltages (ϩ30 mV) for 2-5 min in symmetrical 300 mM cesium methanesulfonate, 10 M CaCl 2 , and 10 mM Na-Hepes (pH 7.2). IpTx a and the II-III loop peptide were added to the cis chamber, which corresponded to the cytosolic side of the channel (13,19). The addition of the peptides to the trans (luminal) side of the channel was without effect. In some experiments, we added 10 mM CaCl 2 to the trans solution. At 0 mV, Ca 2ϩ was the only charge carrier in these experiments, and both peptides were effective in inducing a subconductance state of about one-fourth of the full conductance level. However, the low signal:noise ratio obtained under these conditions made the analysis of the kinetic effect difficult. For the experiments presented here, we omitted Ca 2ϩ in the trans solution. Signals were filtered with an 8-pole low pass Bessel filter at 2 kHz and digitized at 5 kHz. Data acquisition and analysis were done with Axon Instruments software and hardware (pClamp v6.0.2, Digidata 200 AD/DA interface), as described previously (13,19). The current values for the full and subconductance states were obtained from Gaussian fits to the all point amplitude histograms, as described previously (19).
Purification of SR Vesicles and Determination of Ca 2ϩ Release-SR vesicles were purified from rabbit white fast skeletal muscle, as described previously (13,19). Maximal [ 3 H]ryanodine binding site density was typically 3-5 pmol/mg protein. Ca 2ϩ release from SR vesicles was measured by the method of Palade (20), with slight modifications. Briefly, SR vesicles (60 g of protein in 20 l of reaction medium) were placed in a cuvette containing 980 l of 95 mM KCl, 20 mM K-MOPS (pH 7.0), 7.5 mM sodium pyrophosphate, 250 M Antipyrylazo III, 1.5 mM MgATP, 25 g of creatine phosphokinase, and 5 mM phosphocreatine. The mixture was allowed to equilibrate for 3 min at 37°C under constant stirring. Free Ca 2ϩ was monitored by measuring A 710 -790 nm using a diode array spectrophotometer (Hewlett-Packard Model 8452A). Vesicles were actively filled by three to five consecutive additions of 10 nmol of CaCl 2 before the addition of IpTx a or the II-III loop peptide. The total amount of Ca 2ϩ loaded was quantified by the addition of 5 M of the Ca 2ϩ ionophore A23187 at the end of each experiment.
Cross-Linking of Photoactivatable IpTx a , SDS-Polyacrylamide Gel Electrophoresis, and Western Blot Analysis of RyR-SR microsomes (0.4 mg/ml) were incubated with 30 nM photoactivatable IpTx a (see above) in buffer containing 10 M free Ca 2ϩ , 200 mM KCl, and 10 mM Na-Hepes (pH 7.2) in the absence and the presence of 50 M IpTx a . After 60 min at 36°C, 1-ml aliquots were spread over 1-cm-diameter plastic wells and irradiated at short range with ultraviolet light (360 nm) for 30 min.
Samples were washed twice with incubation buffer by centrifugation in a table-top minifuge at 12,000 rpm. Pellets were then resuspended in Laemmli buffer (0.25 M Tris, pH 6.8, 0.4 M dithiothreitol, 8% SDS, 40% glycerol, and 0.04% bromphenol blue) and subjected to SDS-polyacrylamide gel electrophoresis on two identical linear gradient acrylamide gels (4 -12%). Proteins contained in one gel were stained with Coomassie Blue, whereas proteins in the other gel were transferred to nitrocellulose membranes for Western blot analysis. Blots were probed first with a rabbit polyclonal RyR1 antibody (dilution, 1:3,000) and then with an anti-rabbit peroxidase-conjugated secondary antibody. Dried gels were then exposed to x-ray film for 2 days.

Selective Activation of [ 3 H]Ryanodine
Binding by IpTx a among Ca 2ϩ Channel Toxins-The amino acid sequence of IpTx a (14) exhibits no significant homology with the well-characterized Na ϩ and K ϩ channel scorpion toxins (data not shown). However, IpTx a does share 45% and 42% sequence identity with agelenin (21) and Tx2-9 (22), respectively, two spider toxins that block presynaptic (P-type) Ca 2ϩ channels (Fig. 1A). The Cys residues, which stabilize the three-dimensional structure by forming disulfide bridges (16), are similarly arranged in these three peptides (gray boxes). Indeed, they may be used as a frame to align the amino acid sequence of -conotoxin MVIIC, a snail peptide that blocks P-type Ca 2ϩ channels (23), and to reveal regions of homology (open boxes). Fig. 1B shows that, despite the demonstrated structural kinship among these peptides, only IpTx a is capable of enhancing [ 3 H]ryanodine binding. ED 50 , the concentration of IpTx a required to produce a half-maximal effect (6.4 Ϯ 3.1 nM, mean Ϯ S.D.; n ϭ 18), is only slightly higher than that exhibited by [ 3 H]ryanodine among ligands of RyRs (24). This selective and high-affinity effect suggests that IpTx a possesses a unique structural motif that activates RyRs, which is not present even in structurally related peptides.
Physical Interaction of IpTx a with RyRs-To test whether IpTx a may be used independently as a high-affinity, specific ligand for RyRs, we radiolabeled IpTx a and conducted binding experiments in the absence of ryanodine. Fig. 2A shows that the radiolabeled derivative of IpTx a retained high affinity (K D ϭ 11 Ϯ 3 nM) and bound to skeletal SR with a maximal receptor site density (B max ) of 16.1 Ϯ 1.9 pmol/mg protein (n ϭ 3). In the same tissue, the B max for [ 3 H]ryanodine was 3.7 Ϯ 0.6 pmol/mg protein. Thus, assuming all 125 I-IpTx a binding occurs to the RyR, the 125 I-IpTx a :[ 3 H]ryanodine binding site stoichiometry is 4.3:1. Because one [ 3 H]ryanodine molecule binds with high affinity to the tetrameric RyR (24), this ratio suggests that about four IpTx a molecules bind to every RyR tetramer. In CHO cells transfected with the skeletal RyR (Fig. 2B, ϩ RyR1), the 125 I-IpTx a :[ 3 H]ryanodine binding site stoichiometry is 4.6:1 (n ϭ 2). In naïve CHO cells, there is neither 125 I-IpTx a (Fig. 2B, Untransfected) nor [ 3 H]ryanodine binding (data not shown).
To confirm that IpTx a physically interacts with the RyR monomer in skeletal SR, we prepared photoactivatable IpTx a , a synthetic derivative of IpTx a in which Leu 7 was replaced by the light-sensitive cross-linker p-benzoyl-phenylalanine (18). The photoreactive derivative retained high affinity for the RyR (K D ϭ 12 Ϯ 4 nM; n ϭ 3; data not shown). Fig. 3A shows a SDSpolyacrylamide gel electrophoresis profile of SR proteins that were radiated with ultraviolet light after incubation with 30 nM photoactivatable 125 I-IpTx a in the absence (lane 1) and the presence (lane 2) of 50 M unlabeled IpTx a . An immunoblot analysis using a skeletal RyR polyclonal antibody recognized only the high molecular weight band of SR proteins (Fig. 3B). The autoradiogram of the SDS-gel (Fig. 3C) shows clear labeling of the band corresponding to the RyR (lane 1). Other bands are also labeled, most likely from a nonspecific interaction with the toxin, because labeling persists in the presence of excess IpTx a (lane 2). Together with data from Fig. 2, these results indicate that IpTx a makes direct protein-protein interactions with the RyR with a stoichiometry of four IpTx a molecules per single RyR channel.
Functional Effect of IpTx a and a Short Segment of the II-III Loop-The functional effect of the IpTx a -RyR interaction was tested in planar lipid bilayer and Ca 2ϩ release experiments. Fig. 4A shows that 50 nM IpTx a added to the cytoplasmic (cis) side of the skeletal RyR induced the appearance of a subconductance state corresponding to ϳ25% of the full conductance as previously shown (25). Although of small amplitude, the subconductance state displayed a mean open time that was Ͼ100-fold longer than that of unmodified channels. Ion flow would therefore be expected to be greater for an IpTx a -modified channel, despite its lower conductance. Fig. 4B shows that IpTx a elicited Ca 2ϩ release from actively loaded SR vesicles in a dose-dependent manner. The effect of IpTx a was blocked by ruthenium red, consistent with Ca 2ϩ release occurring through RyRs. Strikingly similar results were observed with a 25-amino acid synthetic peptide with primary sequence (Glu 666 -Leu 690 ; Fig. 5A) overlapping that of peptide A (Thr 671 -Leu 690 ), a segment of the II-III loop that activates RyRs (10). The 25-amino acid segment of the II-III loop (henceforth termed "the II-III loop peptide"), like IpTx a , induced the appearance of a small- amplitude and long-lifetime subconductance state (Fig. 4C) and elicited Ca 2ϩ release from SR (Fig. 4D). Thus, albeit with different affinity, the two apparently unrelated peptides exhibit similar functional effects on RyRs.
Structural Analogy between IpTx a and the II-III Loop Peptide-A one-to-one residue alignment between IpTx a and the II-III loop peptide does not reveal significant homology in their amino acid sequence (Fig. 5A). However, both IpTx a and the II-III loop peptide display a structural motif consisting of a cluster of basic amino acids (boxes, Lys 19 -Arg 24 and Arg 681 -Lys 685 , respectively) followed by Thr or Ser, two hydroxylated amino acids (ovals, Thr 26 and Ser 687 , respectively). In IpTx a , the cluster of basic amino acids is interrupted by Cys 21 and encompasses the sequence KCK, which is also found in agelenin and Tx2-9 (Fig. 1A). Therefore, it is likely that the KCK sequence alone does not suffice to activate RyRs and that Cys 21 stabilizes the peptide structure without intervening in proteinprotein interactions with the RyR. Hydroxylated amino acids in a position close to Thr 26 of IpTx a are also found in agelenin (Ser 28 ) and in Tx2-9 (Thr 23 and Thr 24 ), but none is preceded by a cluster of basic amino acids. Likewise, the motif RRG, which appears in IpTx a and -conotoxin MVIIC, is only followed by a hydroxylated amino acid in the former peptide. Indeed, similar to Ser 687 of the II-III loop (27), the distinctive arrangement of Thr 26 of IpTx a with the preceding residues produces a phosphorylation consensus for several protein kinases (28). As presented in Fig. 5B, the structural motif consisting of a cluster of positively charged amino acids followed by a hydroxylated residue is not found in other peptide toxins or in other regions of the DHPR, including the ␤-subunit. Thus, IpTx a and the II-III loop peptide share a specific arrangement of amino acid residues that may be responsible for their similar functional effect on RyR channels.
Competition Experiments between IpTx a and the II-III Loop Peptide-To test whether the analogous functional effects produced by IpTx a and the II-III loop peptide (Fig. 4) result from activation of the same modulatory site on the RyR, we carried out competition experiments between the two peptides. Fig. 6A shows that the II-III loop peptide incrementally decreases the capacity of IpTx a to activate RyRs. The ED 50 for II-III loop-inhibition of the IpTx a effect was 1.3 Ϯ 0.7 M (n ϭ 3), in agreement with the value calculated from direct activation of [ 3 H]ryanodine binding by the II-III loop peptide (Fig. 7B). In contrast, a scrambled II-III loop (a synthetic peptide with amino acid composition identical to the II-III loop peptide but in random sequence) was incapable of stimulating [ 3 H]ryanodine binding (data not shown) or of abolishing the IpTx a effect (Fig. 6B). Thus, the effects of the II-III loop peptide require a defined amino acid sequence and are unrelated to peptide mass or electrical charge. In other competition studies, the II-III loop peptide displaced the binding of 125 I-IpTx a to SR vesicles with an ED 50 of 36 Ϯ 4 M (Fig. 5C). This reduced affinity may be due to displacement of 125 I-IpTx a from sites of nonequivalent affinity, or it may result from positive allosteric interaction between the II-III loop peptide and 125 I-IpTx a . Nevertheless, the II-III loop-125 I-IpTx a competition appears to be specific, because the scrambled II-III loop had no significant effect at concentrations up to 300 M (Fig. 6C).
Effect of Mutations in IpTx a and the II-III Loop Peptide-If, by analogy with other peptide toxin-ion channel associations (29,30), the high-affinity IpTx a -RyR interaction is mediated by electrostatic forces, then mutations in the binding domain of IpTx a should alter its electrostatic potential and the measured affinity constant of the IpTx a -RyR complex. Likewise, if IpTx a and the II-III loop peptide bind to the skeletal RyR via a common structural motif, then corresponding mutations should evoke parallel changes in affinity for both peptides. Fig. 7A shows that synthetic IpTx a (a synthetic peptide with an amino acid sequence identical to that of native IpTx a ; Ref. 14) activates [ 3 H]ryanodine binding to skeletal SR with potency and affinity identical to native IpTx a (5 Ϯ 3 nM; n ϭ 6). Other synthetic derivatives of IpTx a in which Thr 26 was replaced with Ala (T26A) or with the negatively charged residue Glu (T26E) increased [ 3 H]ryanodine binding with an affinity 12-and 160fold lower (ED 50 ϭ 60 Ϯ 5 and 800 Ϯ 78 nM nM, respectively). Mutations to the II-III loop peptide elicited qualitatively similar results (Fig. 7B). The replacement of Ser 687 with Ala (S687A) or with Glu (S687E) decreased the affinity of the II-III loop peptide from 0.81 Ϯ 0.2 M (control) to 4.2 Ϯ 1.1 and 22 Ϯ 4 M, respectively. Therefore, substituting Thr or Ser of IpTx a and the II-III loop peptide with nonpolar or charged amino acids has a substantial impact on the ability of both peptides to activate RyRs. However, neither mutation results in a complete loss of peptide activity.
Mutations within the cluster of basic amino acids elicit more dramatic effects. In IpTx a , replacing Arg 23 with Glu (R23E) abolished IpTx a stimulation of [ 3 H]ryanodine binding (Fig. 8A). The corresponding substitution in the II-III loop peptide, Arg 684 to Glu (R684E), also abolished its effect on [ 3 H]ryanodine binding (Fig. 8B). In contrast, replacing either Lys 8 of IpTx a or the comparable amino acid Lys 675 of the II-III loop peptide with Glu (K8E and K675E, respectively) yielded ED 50 values of 19 Ϯ 4 nM and 1.5 Ϯ 0.3 M, respectively. Thus, the dramatic effect of mutations within the cluster of basic amino acids cannot be solely attributed to a change of electrical charge because mutations distant to the cluster but producing the same electrical change have relatively minor effects. DISCUSSION The molecular mechanism by which depolarization of the skeletal T-tubule membrane induces Ca 2ϩ release from the SR remains an outstanding problem in E-C coupling, with a physical DHPR/RyR interaction being the most plausible hypothesis. Whereas the identification of the RyR region(s) involved in this interaction is in progress (31,32), there is already substantial evidence to invoke the participation of the II-III loop of the ␣ 1 subunit of the DHPR (6, 8 -10). As discussed below, there is still controversy regarding the precise structural domain(s) of the II-III loop involved in this interaction. In this study, we identified a unique structural motif involved in activation of RyRs in IpTx a and in a short segment of the II-III loop. Although the participation of this structural motif in E-C cou-pling awaits further testing, our results provide a structural framework for a mechanical model in which specific amino acids of the II-III loop are capable of interacting with RyRs and triggering Ca 2ϩ release.
Analysis of the amino acid sequence of IpTx a and the effect of indicator Antipyrylazo III, as described previously (20). Arrows indicate the addition of 10 nmol of Ca 2ϩ to the 1-ml reaction medium. Thapsigargin (1 M) was added simultaneously with the peptides to block Ca 2ϩ uptake by the SR. When equilibrium was reached, the Ca 2ϩ ionophore A23187 (5 M) was added to assess SR Ca 2ϩ content (asterisks). Calibration bars, 60 s (horizontal) and 5 M Ca 2ϩ (vertical). mutations strongly suggest that the structural motif encompassing Lys 19 -Arg 24 , the cluster of basic amino acids, followed by Thr 26 , the hydroxylated amino acid, is involved in IpTx a binding to the RyR. Our results, however, do not allow us to ascertain the weight of individual amino acids within this structural domain for participation in binding. For instance, agelenin, Tx2-9, and -conotoxin, three structurally related peptides that are incapable of increasing [ 3 H]ryanodine binding (Fig. 1), bear resemblance to IpTx a in several regions, except in the proposed structural domain. However, agelenin and Tx2-9 do maintain the KCK motif, and thus the actual involvement of these residues in the binding of IpTx a remains undetermined. Either the KCK motif is independent of the binding site or, in analogy to other peptide toxins (29,30), it may be involved in IpTx a docking without participating in activation. The pair of Arg residues following the KCK motif is clearly unique to IpTx a . As shown in Fig. 8, replacing Arg 23 with Glu (R23E) totally abolished the capacity of IpTx a to activate RyRs. Because a similar substitution in a region distant to this cluster (K8E; Fig. 8) was without major functional consequences, the lack of effect of R23E was probably the result of local changes in the electrostatic potential of the toxin's binding site, rather than global conformational changes. Lastly, the involvement of Thr 26 was demonstrated by hydrophobic (T26A) and charged (T26E) substitutions (Fig. 5). Although the mutated peptides retained their ability to activate RyRs, there was a significant loss of affinity for both peptides.
Replacing equivalent amino acids in the II-III loop peptide produced results that were qualitatively similar to those obtained with IpTx a mutants (Figs. 7 and 8). These data, plus the competition experiments in which the II-III loop peptide inhibited the effect of IpTx a (Fig. 6A) and displaced the binding of 125 I-IpTx a (Fig. 6C), strongly suggest that both peptides activate RyRs by a physical interaction with the channel protein via the aforementioned structural domain. If this structural domain of the II-III loop is important for the activation of RyR in vivo, then IpTx a is a peptide mimetic of an effector protein, and its structure-activity information may have direct implications for the mechanism of E-C coupling in skeletal muscle. In RyRs reconstituted in lipid bilayers, IpTx a and the II-III loop peptide induced the appearance of a subconductance state, again supporting the notion that both peptides lead to the same conformational changes in the channel protein.
The small-amplitude, long-lifetime subconductance state produced by IpTx a and the II-III loop peptide was seen in Cs ϩ -conducting (Figs. 4 and 5) and in Ca 2ϩ -conducting RyRs (data not shown; see Ref. 25) and is clearly different from that produced by FK506 and ryanodine. FK506 and other immunosupressants that strip RyRs of the accessory protein FKBP12 produce rapidly fluctuating subconductance states that represent approximately one-half and one-fourth of the full conductance state (33,19). Ryanodine irreversibly locks the channel in a subconductance state of variable amplitude, depending on the charge carrier (34). However, the effects of IpTx a and the II-III loop peptide are also different from those produced by the whole 138-amino acid II-III loop, which activates RyRs by increasing the frequency of full openings without inducing the appearance of a subconductance state (9,27). These results argue against the II-III loop fragment studied here being an exact functional equivalent of the whole II-III loop. They suggest instead that the II-III loop-RyR interaction is more complex, perhaps encompassing several points of contact that, when acting concertedly, modify the RyR kinetics in a way that is more similar to the activation produced in vivo. In skeletal muscle cells, the DHPR-RyR interaction involves orthograde (DHPR3 RyR) as well as retrograde (RyR3 DHPR) signals (32). Thus, providing that the II-III loop peptide studied here is an actual participant in the E-C coupling mechanism, this fragment and IpTx a may be regarded as partial effectors of the DHPR-RyR interaction whose critical task is to relay the signal that initiates Ca 2ϩ release.
Using synthetic peptides corresponding to small segments of the II-III loop, El-Hayek et al. (10) found that only the aminoterminal portion (peptide A, Thr 671 -Leu 690 ) was capable of activating RyRs. More recently, the same group (35) reduced the length of the activator peptide to 10 residues (Arg 681 -Leu 690 ) that encompassed the structural domain studied here. Inter-estingly, a pentapeptide corresponding to the cluster of basic amino acids was insufficient to activate RyRs (35), and our own results indicate that a scrambled II-III loop peptide is also without effect (data not shown). Therefore, the distinctive distribution of the positively charged residues and their relation to the hydroxylated amino acid are critical determinants of the structural domain that binds and activates RyRs.
Lu et al. (27) found that phosphorylation of Ser 687 of the skeletal II-III loop (or substitution by Ala) abolished the capacity of the whole loop to activate RyRs. Leong and MacLennan (36) used the II-III loop in an affinity column and found that replacement of Lys 677 or Lys 682 with Glu decreased the capacity of the II-III loop to interact with peptide fragments of the skeletal RyR. These results are in line with the hypothesis that the amino-terminal region of the II-III loop binds to and activates the skeletal RyR. However, expressing chimeras of cardiac and skeletal DHPR in dysgenic myotubes, Nakai et al. (8) found that Phe 725 -Pro 742 was the most important region of the skeletal DHPR to elicit a skeletal-type contraction. This region is located almost at the middle of the II-III loop, more than 30 residues downstream from peptide A. Although the results with DHPR chimeras are apparently contradictory to those obtained with isolated peptides, it is still possible that the structural domain of peptide A postulated here as being important for activation of RyRs participates in the transmission signal from DHPR to RyR. A potential scenario would be that Phe 725 -Pro 742 , which determines skeletal-type E-C coupling (8), binds to the skeletal RyR at rest to inhibit Ca 2ϩ release and that depolarization of the T-membrane rearranges the II-III loop such that peptide A interacts with and activates the RyR. In agreement with this hypothesis, El-Hayek et al. (10) found that the activating effect of peptide A was antagonized by Glu 724 -Pro 760 , a fragment of the skeletal II-III loop encompassing the amino acid sequence identified by Nakai et al. (8). However, even in this simplified scenario, outstanding issues remain unsolved. For example, skeletal-type contractions do not require entry of external Ca 2ϩ , whereas activation of RyRs by peptide A (10), IpTx a (13), or the whole II-III loop (9) requires Ca 2ϩ , at least at suboptimal levels. Again, this may reflect the fact that these peptides are only partial effectors of the transmission signal, with other segments of the DHPR conferring Ca 2ϩ independence to the DHPR-RyR interaction.
Toxin mimicry of surface receptors is not unprecedented. Mastoparan, a peptide from wasp venom, is a potent secretagogue that mimics an intracellular loop of G protein-coupled receptors (15). As previously shown (16), the capacity of foreign peptides to activate intracellular signals may yield insights into the molecular mechanisms of signal transduction in the target cells. In the case of IpTx a , the high affinity of 125 I-IpTx a demonstrates a direct protein-protein interaction with the RyR and reveals a ϳ4:1 stoichiometry of IpTx a /RyR binding sites (Fig. 2). Assuming that IpTx a is a surrogate high-affinity ligand for the II-III domain that triggers Ca 2ϩ release, these results suggest that in the intact muscle, up to four II-III loops directly gate one RyR. This would be consistent with the current structural model of E-C coupling in skeletal muscle (26), where every other foot protein (RyR) faces a tetrad of T-tubule particles representing four DHPR molecules. Another inference based on the above assumption is that the II-III loop but not the carboxyl terminus of the ␣ 1 subunit (11) or the ␤-subunit (12) is structurally endowed to trigger Ca 2ϩ release. This is because the structural domain identified as being responsible for activating RyRs (Fig. 4) is contained within the II-III loop sequence only. However, as mentioned above, our results do not rule out the possibility of multiple sites of interaction between the DHPR and the RyR because inhibitory sites of contact have been detected in other segments of the ␣ 1 subunit (11) and even within the II-III loop itself (10).
Regardless of its potential use as a peptide probe to study the molecular determinants of E-C coupling, it is important to keep in mind that IpTx a triggers conformational changes in the channel protein that ultimately lead to Ca 2ϩ release (Fig. 4). The specificity, high affinity, and reversibility of the IpTx a -RyR interaction may thus be exploited to identify amino acids of the RyR directly responsible for channel opening.