Activation of Skeletal Ryanodine Receptors by Two Novel Scorpion Toxins from Buthotus judaicus*

Buthotus judaicus toxin 1 (BjTx-1) and toxin 2 (BjTx-2), two novel peptide activators of ryanodine receptors (RyR), were purified from the venom of the scorpion B. judaicus. Their amino acid sequences differ only in 1 residue out of 28 (residue 16 corresponds to Lys in BjTx-1 and Ile in BjTx-2). Despite a slight difference in EC50, both toxins increased binding of [3H]ryanodine to skeletal sarcoplasmic reticulum at micromolar concentrations but had no effect on cardiac or liver microsomes. Their activating effect was Ca2+-dependent and was synergized by caffeine. B. judaicus toxins also increased binding of [3H]ryanodine to the purified RyR1, suggesting that a direct protein-protein interaction mediates the effect of the peptides. BjTx-1 and BjTx-2 induced Ca2+ release from Ca2+-loaded sarcoplasmic reticulum vesicles in a dose-dependent manner and induced the appearance of long lived subconductance states in skeletal RyRs reconstituted into lipid bilayers. Three-dimensional structural modeling reveals that a cluster of positively charged residues (Lys11 to Lys16) is a prominent structural motif of both toxins. A similar structural motif is believed to be important for activation of RyRs by imperatoxin A (IpTxa), another RyR-activating peptide (Gurrola, G. B., Arevalo, C., Sreekumar, R., Lokuta, A. J., Walker, J. W., and Valdivia, H. H. (1999) J. Biol. Chem. 274, 7879-7886). Thus, it is likely that B. judaicus toxins and imperatoxin A bind to RyRs by means of electrostatic interactions that lead to massive conformational changes in the channel protein. The different affinity and structural diversity of this family of scorpion peptides makes them excellent peptide probes to identify RyR domains that trigger the channel to open.

contraction coupling, the process that links an electrical stimulus (depolarization) to Ca 2ϩ release and cell contraction (1)(2)(3). The amino acid sequence of the three RyR isoforms has been completely elucidated, but due in part to their tremendous molecular size (ϳ5,000 amino acids/monomer), the exact structural motif(s) of RyR involved in excitation-contraction coupling, the ATP and Mg 2ϩ regulation sites, as well as the Ca 2ϩ activation and inactivation domains still remain uncertain and controversial (4 -6). The limited availability of antibodies and other ligands against specific functional sites has also hampered the elucidation of the channel's most important structural domains.
To date, only a few molecular probes of RyRs have been well characterized. Ryanodine, a plant alkaloid, has been used for identification and characterization of RyRs because it preferentially binds to the open state of RyRs, thus serving as an index of channel activity (7). Because venoms from spiders, snakes, and scorpions are rich in peptide toxins, they have been extensively exploited as a source of peptide probes specifically targeted to various voltage-and ligand-gated channels (8 -10). From the venom of the African scorpion Pandinus imperator, two components (IpTx a and IpTx i ) have been shown to activate and inhibit RyRs, respectively (11). IpTx a , a 3.7-kDa basic peptide (12), increases [ 3 H]ryanodine binding to skeletal RyR, but not cardiac RyR, although it induces the appearance of long lived subconductance states in both isoforms (13). Interestingly, IpTx a shares strikingly similar structural and functional properties with peptide A (14), a peptide fragment from the II-III loop of the skeletal type dihydropyridine receptor ␣ 1 subunit, which also triggers Ca 2ϩ release from the SR (15). By mutating some amino acids in the cardiac and skeletal II-III loop peptide, we found that a cluster of positively charged amino acid residues on the surface of the peptide A molecule is important for activation of skeletal RyRs (16). Thus, identification of the common structural components for activation of RyR, as well as their binding sites in RyRs, will help us to better understand the mechanisms that control gating of the RyR.
Other peptide probes for the RyR have been isolated, but they are not as well characterized as the above mentioned peptides. A peptide fraction of ϳ4000 -7000 Da molecular mass from the venom of the African scorpion Buthotus hottentota increases [ 3 H]ryanodine binding and induces the appearance of subconductance states in skeletal RyRs (17). Likewise, Morrissette et al. (10) isolated ryanotoxin, an ϳ11.4-kDa peptide fraction, from Buthotus judaicus scorpion venom, which similarly induces the appearance of subconductance states in skel-* This work was supported by National Institutes of Health Grant RO1HL55438 and by a grant-in-aid from the American Heart Association (to H. H. V.). 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.
The protein sequence(s) reported in this paper has been submitted to the GenBank TM  ʈ Established Investigator of the American Heart Association. 1 The abbreviations used are: RyR, ryanodine receptor; SR, sarcoplasmic reticulum; IpTx a , imperatoxin A; HPLC, high performance liquid etal RyRs. Nevertheless, it is not yet known whether these peptide probes consist of a single polypeptide or several distinct polypeptides of highly similar molecular weights and HPLC elution profiles (10). Therefore, it is hard to exploit these peptides as probes to study the functional domains of RyRs. In this paper, we have been able to determine the structural components of two novel toxins from the venom of scorpion B. judaicus. We report purification and complete amino acid sequences of these toxins, which are able to activate RyRs, and we compare their structure and function with IpTx a .

Materials-[ 3 H]Ryanodine was purchased from PerkinElmer Life
Sciences. Bovine brain phosphatidylethanolamine and phosphatidylserine were obtained from Avanti Polar Lipids (Birmingham, AL).
Purification of B. judaicus Toxins-100 mg of lyophilized B. judaicus venom (Alomone Laboratories Ltd., Jerusalem, Israel) was suspended in 2 ml of water and centrifuged at 14,000 rpm for 10 min to separate the soluble fraction and the mucus. The mucus was re-extracted two more times. All the supernatants were pooled and adjusted to 5 mg of protein/ml. 150 l of sample was injected into a C 18 reverse-phase HPLC column (System Gold, Beckman Instruments, San Ramon, CA). The peptide fractions were eluted with 0 -67.5% acetonitrile at a flow rate of 1 ml/min. The eluted fractions were monitored at 214 nm. Each fractionated peak was collected and lyophilized. With the same amount of peptides (w/v) in each sample as determined by A 214 , the activity of HPLC peak fractions was determined by [ 3 H]ryanodine binding.
Amino Acid Analysis and Sequencing of B. judaicus Toxins-Amino acid analysis of B. judaicus toxins was performed on samples hydrolyzed in 6 M HCl with 0.5% phenol at 110°C in evacuated, sealed tubes as described (18). Reduction of B. judaicus toxins with dithiothreitol, and alkylation with iodoacetic acid was performed as described (18). The sequences of the intact native and reduced/carboxymethylated B. judaicus toxins were determined using a model 6400/6600 automatic liquid-phase protein sequencer (Milligen/BioSearch Prosequencer) using standard Edman degradation programs and CD Immobilon membrane. To confirm the carboxyl-terminal sequence, 20 g of B. judaicus toxin was hydrolyzed with Staphylococcus aureus V8 in 100 mM ammonium bicarbonate (pH 7.8). The peptide fragments were purified and sequenced as described for native B. judaicus toxins.
Computer Modeling-To generate the three-dimensional structure, we used Insight II Discover software from Biosym Technologies (San Diego, CA) running on a Silicon Graphics Octane RS10000 Workstation. Peptide P01 (19), the three-dimensional structure of which has been determined by NMR, is a natural peptide from the North African scorpion Androctonus mauretanicus mauretanicus. The primary structure of peptide P01 is quite similar to the B. judaicus toxins we purified, thus it was used as the structural model for building. The threedimensional structure of peptide P01 was obtained from the Entrez of National Center for Biotechnology Information (NCBI) web site (www. ncbi.nlm.nih.gov/Entrez/). The electrostatic potential surface was analyzed using the Discover force field. Five hundred steps of minimization were performed using the steepest descent algorithm method until a root mean square deviation of 0.001 was obtained.
SR Microsome Preparation and Purification of RyR1-Porcine skeletal, cardiac, and liver SR-enriched microsomes were isolated as described (20). Microsomes from the last centrifugation were suspended in 0.3 M sucrose, 0.1 M KCl and 5 mM Na-PIPES (pH 7.2). RyR1 was purified from microsomes using continuous 5-20% sucrose density gradient centrifugation, as described (21).
[ 3 H]Ryanodine Binding Assay-[ 3 H]Ryanodine binding to pig skeletal SR and other tissue homogenates was carried out as described previously (22). The standard incubation medium contained 0.2 M KCl, 10 M CaCl 2 , 40 mM Na-HEPES (pH 7.2), 7 nM [ 3 H]ryanodine, 40 -50 g of pig skeletal or cardiac SR vesicles, and different concentrations of modulators as stated in each specific experiment. Samples (0.1 ml) were run in duplicate at 37°C for 90 min. For binding assays with different calcium concentrations, 10 M CaCl 2 in the standard assay was replaced by 1 mM EGTA and CaCl 2 necessary to set free calcium from 1 nM to 10 mM. Unless otherwise indicated, data represent the mean Ϯ S.E. with n Ն 2. Mathematical fitting of data was accomplished with the computer program Origin (version 7.0, Microcal Inc., Northampton, MA).
Planar Lipid Bilayer Technique-Pig skeletal RyRs were reconstituted into Muller-Rudin planar lipid bilayers as described previously (22). Single channel data were collected at steady voltages (ϩ35 mV, cis chamber grounded) for 2 min in symmetrical 300 mM cesium methanesulfonate and 10 mM Na-HEPES (pH 7.2). The recording solution contained ϳ5 M free Ca 2ϩ , as assessed by a calibration curve, which was sufficient to activate RyRs. Different concentrations of B. judaicus toxins were added to the cis chamber, which corresponded to the cytosolic side of the channel (22). Signals were digitized at 4 kHz and analyzed after filtering with a low-pass 8-pole Bessel filter at a sampling frequency of 1.5 kHz. Data acquisition and analysis were performed with Axon Instruments software and hardware (pClamp version 8.03, Digidata 200 AD/DA interface). The full and subconducting current values were obtained from Gaussian fits to the amplitude histograms, as described previously (23).
Spectrophotometric Ca 2ϩ Release Assay-Skeletal SR vesicles (ϳ50 g of protein) were mixed in a 1-ml solution containing 95 mM KCl, 7.5 mM sodium pyrophosphate, 250 M antipyrylazo III, 1.5 mM MgATP, 25 g/mg creatine phosphokinase, 5 mM phosphocreatine and 20 mM K-MOPS, (pH 7.0). The solution was placed in a cuvette, incubated at 37°C, and allowed to equilibrate for 2 min. Changes in free Ca 2ϩ were monitored by changes in the absorbance of antipyrylazo at 710 nm and subtraction of the absorbance at 790 nm, at 1-s intervals, using a diode array spectrophotometer (Hewlett Packard Model 8452A). Initially, vesicles were actively loaded with CaCl 2 until levels were close to the filling capacity (ϳ2 mol of total Ca 2ϩ /mg of protein, usually 10 l of 1 mM CaCl 2 added 4 times). Thapsigargin (1 M) was then added to block Ca 2ϩ uptake. The absorbance signals were calibrated by addition of a known amount of Ca 2ϩ to the complete transport mixture in the presence of the Ca 2ϩ ionophore A23187 to prevent Ca 2ϩ accumulation.

Purification of B. judaicus Toxin Activators-[ 3 H]
Ryanodine binds with high affinity (K d ϳ7 nM) to the open conformational state of the RyR. Thus, we used the standard [ 3 H]ryanodine binding assay to characterize whole B. judaicus venom and to screen the peaks separated by HPLC fractionation (as described below). Whole B. judaicus venom stimulated [ 3 H]ryanodine binding in a dose-dependent manner, and the effective concentration (EC 50 ) for the whole venom was 400 g/ml (Fig. 1C). To purify the active component, soluble proteins from whole venom were loaded onto a C 18 reverse-phase column. Two peaks, eluted at 32.8 min and 36.4 min (Fig. 1A, arrows), effectively increased [ 3 H]ryanodine binding, although most other peaks had no effect. (Two other peaks, one eluted at 6 min (not shown) and another eluted at 44 min, increased [ 3 H]ryanodine binding marginally.) Thus, the two peaks at 32.8 and 36.4 min were collected and subjected to another analytical HPLC separation to verify their purity. Both fractions appeared as single, homogenous, and symmetrical peaks (Fig. 1B), and their purity was greater than 95%. The molecular weights, determined by mass spectrometry analysis, were 2888.95 Ϯ 0.12 for peptide 1 and 2874.46 Ϯ 0.37 for peptide 2. Interestingly, the molecular mass difference between peptide 1 and peptide 2 is only 15 Da, suggesting that these two peptides may have similar primary structures. Both peptides stimulated [ 3 H]ryanodine binding in a dose-dependent manner (Fig. 1C). The EC 50 for peptide 1 and peptide 2 was 8 and 20 M, respectively. Because both peptides activating RyRs were purified from B. judaicus, we named them BjTx-1 and BjTx-2, respectively.
Primary Structures-The complete amino acid sequences of BjTx-1 and BjTx-2 were determined by direct automated sequencing (Fig. 2). Both peptides are composed of 28 amino acid residues with a difference of only 1 amino acid at position 16 (i.e. a Lys in BjTx-1 but an Ile in BjTx-2). The molecular difference between Lys and Ile is 15 Da, and this precisely matches the difference in molecular mass between the two toxins as determined by mass spectrometry. This difference also explains the observed HPLC elution profile. Because Ile is a hydrophobic amino acid but Lys is electrically charged, BjTx-1 is more hydrophilic than BjTx-2 and thus elutes earlier in the reverse-phase column. The calculated molecular weights for the linear sequences are 2895 and 2880, respectively, which are 6 greater than the molecular masses determined by mass spectrometry. This suggests that both toxins may contain three pairs of disulfide bridges, which is a characteristic of most short scorpion toxins (forming a disulfide bridge removes two hydrogen atoms). However, unlike most small scorpion toxins, which are usually basic in nature, both B. judaicus toxins contain four negatively charged amino acid residues (Asp or Glu), but only two or three positively charged amino acids (Lys), making them acidic in general (calculated pI values for BjTx-1, 5.48; BjTx-2, 4.75).
To compare the similarity of the two isolated toxins with other peptides, we searched GenBank TM of the National Center for Biotechnology Information (NCBI), using the program Blast (www.ncbi.nlm.nih.gov/BLAST/). B. judaicus toxins are highly homologous with scorpion peptide Lqhpep 2, from Leiurus quinquestriatus hebraeus (24), and PepII, from Buthus sindicus (25). They also share some similarity with peptide P01 from the scorpion A. mauretanicus mauretanicus, the three-dimensional structure of which has been determined by NMR (19). However, although the complete amino acid sequences of those three peptides have been determined, their biological function remains unknown. No significant similarity in primary structure is observed between both B. judaicus toxins and IpTx a , except that all three peptides contain 6 cysteines, enabling them to form three pairs of disulfide bridges (Fig. 2).
Molecular Modeling of B. judaicus Toxins-B. judaicus toxins share some sequence similarities with scorpion peptide P01 and leiuropeptide II, the solution structures of which have been determined by NMR spectroscopy (19,24). The backbone fold for all of these structurally identified peptides (listed in Fig. 2) is virtually identical, albeit there are some differences in side chains and surface charges. Considering this information, we created a molecular model of BjTx-1 based on the three-dimensional structures of peptide P01 and leiuropeptide II. The molecular model was then subjected to energy minimization.
As shown in Fig. 3, the backbone model of BjTx-1 includes an ␣-helix connected by a tight turn to a ␤-sheet. The ␣-helix extends from residues Cys 3 to Gln 14 , with a Pro residue located at position 7, which may act as an ␣-helix breaker (19). The antiparallel ␤-sheet is composed of two strands. The first strand extends from Ala 15 to Asp 20 and the second from Gly 23 to Val 28 . The primary structures of both toxins are ordered as -Cys 1 -Cys 2 -Xaa-Xaa-Xaa-Cys 3 -Cys 4 -Cys 5 Xaa-Cys 6 -, a well known cysteine motif common to most scorpion toxins, including all the peptides we previously mentioned, and also charybdotoxin (26), iberiotoxin (27), margatoxin (28), and kaliotoxin (28). The ␣-helix and ␤-sheet are connected by three disulfide bridges, with the pattern Cys 1 -S-S-Cys 4 , Cys 2 -S-S-Cys 5 , and Cys 3 -S-S-Cys 6 (19).
The Corey-Pauling-Koltun model of BjTx-1 (Fig. 3, lower model) indicates the spatial orientation of all atoms of the amino acid residues forming the carbon backbone. Three pairs of disulfide bridges formed by 6 cysteine residues (yellow) remain buried inside the molecule and function to pull it together, whereas other amino acid residues are surface-exposed. Based on the calculated electrical potentials, there appear to be two potential electrostatic zones. One of them is positive, localized in the region around Lys 11 and Lys 13 in the ␣-helix and extending to Lys 16 at the beginning of the ␤-sheet (blue residues, top of the molecule). Another charged surface area is a negative zone, including Glu 4 , Glu 5 (dark red) and Asp 20 , Asp 21 (magenta). Although far away in the primary sequence, these residues are adjacent to each other in the spatial arrangement because of the disulfide bridges.
The backbone structure of BjTx-1 is virtually identical to that of peptide P01 (Fig. 3, middle column, upper model). However, the positively charged zone is absent in peptide P01, because there is only one Lys present on the top of the molecule (middle column, bottom model). On the contrary, the backbone structure of IpTx a (third column, upper model) and its threedimensional structure, determined by NMR (29), have no similarity with BjTx-1. However, IpTx a also contains a positively charged zone, which is even more positive than that of BjTx-1, because five positively charged Lys (blue) and Arg (dark blue) residues exist in a cluster on the top of the molecule (third column, bottom model).  (Fig. 4).
B. judaicus Toxins Interact with RyR1 Directly-To determine whether BjTx-1 and BjTx-2 had a direct interaction with  (Fig. 5B), suggesting that both toxins act on RyR1 based on a direct protein-protein interaction mechanism. This property is analogous to that of IpTx a , which also activates RyR1 by a direct interaction (14,22) (Fig. 5B).

Effects of BjTx-1 and BjTx-2 on the Ca 2ϩ Dependence of [ 3 H]Ryanodine
Binding-The Ca 2ϩ dependence of [ 3 H]ryanodine binding to RyR1 has been well established. In the range of pCa 9 to pCa 5, Ca 2ϩ has an activating effect and increases [ 3 H]ryanodine binding to the RyR, whereas in the range of pCa 4 to pCa 2, Ca 2ϩ has an inactivating effect. Fig. 6A (black line) shows the dual effect of Ca 2ϩ on RyR activity, which gives rise to the characteristic bell-shaped curve (30,31). In the presence of BjTx-1 (filled circles) or BjTx-2 (filled triangles), the binding curves maintained the same bell shape; however, the absolute [ 3 H]ryanodine binding values were increased above pCa 7. Both B. judaicus toxin curves maintained the same basal threshold as control, peaked between pCa 5 and pCa 4, and almost fully inhibited RyR activity at pCa 3. For further comparison, we normalized the data points with maximal binding at pCa 5 (Fig. 6B) and found no significant difference in Ca 2ϩ sensitivity between the curves with or without the presence of B. judaicus toxins.
Combined Effects of Caffeine with B. judaicus Toxins-We tested the effect of caffeine, a classical agonist of RyRs, in the presence or absence of BjTx-1 and BjTx-2 using [ 3 H]ryanodine binding assays. As shown in Fig. 7, millimolar caffeine increased [ 3 H]ryanodine binding to RyRs, with the greatest effect observed above 10 mM. The presence of 8 M BjTx-1 or 20 M BjTx-2 (EC 50 values for the toxins, respectively) increased [ 3 H]ryanodine binding to caffeine-activated RyRs in a synergistic manner. The curves displayed a steeper slope when [caffeine] exceeded 1 mM. These results suggest that BjTx-1 and BjTx-2 do not share the same binding site with caffeine; instead, the toxin-binding site appears to display cooperative interaction with the caffeine-binding site so that occupation of one site promotes binding of ligand to the other.
Functional Activation of B. judaicus Toxins on Calcium Flux Assay-To determine whether the increase of [ 3 H]ryanodine binding to skeletal SR vesicles by B. judaicus toxins had any functional correlation, we performed calcium flux assays. Antipyrylazo III was used to detect the changes of [Ca 2ϩ ] outside of the SR vesicles. Fig. 8A shows that BjTx-1 induced Ca 2ϩ release in a dose-dependent manner. In the absence of BjTx-1, the curve was flat, indicating that most of the releasable Ca 2ϩ was stored in the SR until A23187, a Ca 2ϩ ionophore, was added (Fig. 8A, asterisk). Similar effects were also observed for BjTx-2 (Fig. 8B).
Functional Activation of B. judaicus Toxins on Single Channels-To further investigate the functional properties of the B. judaicus toxins, we reconstituted skeletal RyRs into planar lipid bilayers to determine their effects on single channel activity. Before addition of peptide, the RyR1 channel remained mostly in a closed state, with only a few intermittent open events (Fig. 9A). The unitary conductance of the control channel corresponded to 640 pS. Current amplitude histograms representing 2 min of recording showed a symmetric peak centered at 5 pA (closed), and a small peak at 25 pA (open) (Fig. 9E). Addition of 1 M BjTx-1 induced a small but significant increase in open events (Fig. 9, B and F). However, when 5 M (Fig. 9C) or 10 M (Fig. 9D) BjTx-1 was added to the cis side, it induced subconductance states measuring 160 pS, which corresponded to ϳ25% of the full conductance openings. As reflected in the amplitude histograms (Fig. 9, G and H), the time the channel spent in a subconducting state was dose-dependent.
A similar effect was observed with BjTx-2 (Fig. 10). Addition of BjTx-2 into the cis side of the channel also induced subconductance states in a dose-dependent manner, as de- picted in the single channel recordings (Fig. 10, B-D) and the amplitude histograms (Fig. 10, F-H). These results suggest that both B. judaicus toxins activate the RyR1 by inducing subconductance states. Despite their small amplitude, these toxin-induced subconductance states displayed a long mean open time that is Ͼ 100 times longer than that of the unmodified channels; therefore, total ion flux would be greater for the toxin-modified channels.

DISCUSSION
In this study we reported identification of two novel activators of RyR1 from the scorpion B. judaicus. These two toxins show no similarities in primary structure with any known peptide probes of RyR1. Both toxins increased binding of [ 3 H]ryanodine to skeletal SR at micromolar concentrations but had no effect on cardiac or liver microsomes. Their activating effect was Ca 2ϩ -dependent and was synergized by caffeine. B. judaicus toxins also increased binding of [ 3 H]ryanodine to the purified RyR1, suggesting that a direct protein-protein interaction mediates the effect of the peptides. BjTx-1 and BjTx-2 induced Ca 2ϩ release from Ca 2ϩ -loaded SR vesicles in a dose-dependent manner and induced the appearance of long lived subconductance states in skeletal RyRs reconstituted into lipid bilayers.
Morrissette et al. (10) purified ryanotoxin, a peptide fraction from the scorpion B. judaicus, the same venom used here to purify BjTx-1 and BjTx-2. In the same manner as BjTx-1 and BjTx-2, ryanotoxin increased [ 3 H]ryanodine binding to skeletal SR vesicles and induced subconductance states measuring 163 Ϯ 12 pS in RyRs reconstituted into lipid bilayers. However, this does not mean that the two toxins we isolated here are ryanotoxin; in fact, we have evidence that shows they are separate toxins. First, using the same C 18 reverse-phase column, ryanotoxin was eluted by 67.5% ACN (10), whereas BjTx-1 and BjTx-2 were eluted by 31 and 38% ACN, respectively. The shorter retention times reflect that the toxins we isolated have different hydrophobicity values than that of ryanotoxin. Another obvious difference between B. judaicus toxins and ryanotoxin is their molecular weight. The molecular masses of BjTx-1 and BjTx-2 are 2889 and 2874 Da, respectively, much less than that of ryanotoxin, which is 11.4 kDa as presented by Morrisette et al. (10). Actually, based on our protein elution profile, it is possible that the ryanotoxin corresponds to the peak eluted at 44 min, which also showed an activation effect in [ 3 H]ryanodine bindings. However, further studies must be done to confirm these data. Interestingly, Morrissette et al. (10) did not detect the B. judaicus toxins that we isolated probably because they used a wavelength of 254 nm to detect peptides, which can resolve only 6 peaks from B. judaicus venom. Conversely, we used 214 nm to monitor the peptides, which separates the venom into more than 30 peaks.
When the functional effect of B. judaicus toxins is compared via the Ca 2ϩ release assay versus the single channel recordings, we notice an obvious difference in dose dependence for RyR activation. In the Ca 2ϩ release assay, low nanomolar toxin induced Ca 2ϩ release from the SR, whereas in single channel recordings, micromolar toxin was necessary to induce the occurrence of subconductance states. Likewise, micromolar toxin was necessary to increase [ 3 H]ryanodine binding to SR vesicles. This discrepancy could be explained by differences in Ca 2ϩ loading on the luminal side of SR. In the Ca 2ϩ release assay, SR vesicles  Fig. 1, except that the incubation medium was supplemented with 1 mM EGTA and various CaCl 2 concentrations to yield the indicated free [Ca 2ϩ ]. The Ca 2ϩ :EGTA ratios were calculated with a computer program using the stability constants given by Fabiato (33). Data points are the mean (Ϯ S.E.) of three experiments. B, normalized data of A. Each data point was normalized to the binding at pCa 5, which is the maximal binding for each group of data. were fully loaded with Ca 2ϩ , which increased the activity of RyRs and also increased their sensitivity to the challenge of toxin. One study that supports this explanation was done by Koizumi et al. (32), who investigated the properties of quantal Ca 2ϩ release evoked by activation of RyR in PC12 cells. They found that the sensitivity of RyR to caffeine was altered by luminal Ca 2ϩ . Specifically, the threshold for RyR activation by caffeine was sensitized ϳ10-fold as the Ca 2ϩ load increased from minimal to maximal loading. In addition, the fraction of Ca 2ϩ released by low caffeine concentrations was increased (32).
Surprisingly  (22). Downward arrows indicate the addition of 10 nmols of Ca 2ϩ to the 1-ml reaction medium. Thapsigargin (1 M) was added simultaneously with the peptides (upward arrows) 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). single RyR1 channels reconstituted into planar lipid bilayers. Nevertheless, there are also functional differences between the toxins. As seen in the [ 3 H]ryanodine binding curves in Fig. 6, for example, B. judaicus toxins do not alter RyR Ca 2ϩ sensitivity. However, IpTx a shifts the threshold of activation from pCa 7 to pCa 8 and the threshold of inactivation from pCa 3 to pCa 2. Consequently, IpTx a exerts its stimulatory effect by sensitizing RyRs to Ca 2ϩ activation and decreasing the inhibitory effect at low Ca 2ϩ (22). Despite significant differences in their primary sequences, structural modeling (Fig. 3) suggests that the position of the charge distribution at the level of the three-dimensional structures of B. judaicus toxins and IpTx a are comparable, which may explain why they share functional similarity. In BjTx-1, there is a positively charged domain which consists of Lys 11 , Lys 13 , and Lys 16 . A similar positively charged domain also exists in the three-dimensional structure of IpTx a (Fig. 3). Indeed, structure-function relationship studies have demonstrated that the electrostatic effect plays a very important role in toxin-receptor interactions. For example, with IpTx a , replacement of Arg 23 , localized in a cluster of five positively charged amino acids, with the negatively charged Glu abolished its activation effect on [ 3 H]ryanodine binding. However, replacement of Lys 8 , which is far away from the positively charged domain, with Glu has a negligible effect on IpTx a activity (14). Comparable results are obtained from point mu-tations of peptide A, a peptide fragment within the II-III loop of the ␣1 subunit of dihydropyridine receptors, which is critical for Ca 2ϩ release via activation of RyRs. Mutation of Arg 684 , located in a cluster of five basic amino acids, with negatively charged Glu also abolished its function as a RyR activator (14), whereas mutation of Lys 675 -Glu, has no effect (14). Taken together, it is plausible that the positive potential zones within IpTx a and BjTx-1 are also involved in their interactions with the RyR. Sound support to this idea is the fact that BjTx-2, a natural mutant of BjTx-1 with Lys 16 replaced by Ile, displays lower affinity for interaction with RyR compared with BjTx-1, as shown in Fig. 1C.
In summary, peptide toxins are emerging as useful probes of RyR structure and function. The B. judaicus toxins that we isolated bind directly to the RyR and act rapidly and reversibly. These intrinsic properties confer upon them a unique set of attributes that make them useful tools to identify regulatory domains critical for channel gating.