Imperatoxin A Induces Subconductance States in Ca²⁺ Release Channels (Ryanodine Receptors) of Cardiac and Skeletal Muscle

Ca²⁺ release channels (CRCs),¹ also known as ryanodine receptors (RyRs), mediate the release of Ca²⁺ from an intracellular membrane compartment, the sarcoplasmic reticulum (SR), into the cytoplasm in striated muscle cells. The CRC has been purified as a 30 S protein complex and shown to be composed of four large RyR polypeptides of ∼5,000 amino acid residues and four immunophilins (FK506 binding protein) of ∼100 amino acid residues each. Three tissue-specific isoforms of RyR have been identified. The conductances and pharmacological properties of the three isoforms are, to a large extent, similar (for reviews, see Meissner, 1994; Sutko and Airey, 1996).

Traditionally, venom from many poisonous snakes, lizards, and scorpions has been a rich source of peptide toxins that target various voltage- and ligand-gated channels including the CRCs (Furukawa et al., 1994; Morrissette et al., 1995, 1996). Recently, from the venom of the African scorpion Pandinus imperator, two factors, imperatoxin activator (IpTx_a) and imperatoxin inhibitor (IpTx_i), were isolated that selectively activated and inhibited, respectively, the CRCs (Valdivia et al., 1992; El-Hayek et al., 1995). IpTx_a is a 33 amino acid peptide that activated the skeletal CRC in [³H]ryanodine binding and single channel studies. It did not exert a significant effect on cardiac CRC (Valdivia et al., 1992; El-Hayek et al., 1995; Zamudio et al., 1997a). IpTx_i has been shown to be a heterodimeric protein (subunits of 108 and 27 amino acid residues covalently linked by a disulfide bond) with lipolytic action, and it inhibited both skeletal and cardiac CRCs by generating a lipid product (Valdivia et al., 1992; Zamudio et al., 1997b).

Here, we show that IpTx_a induces voltage- and concentration-dependent subconductance states in both skeletal and cardiac CRCs when added to the cytosolic side. Analysis of voltage and concentration dependence and stochastic analysis of the events suggests that the induction of subconductance states corresponds to reversible, voltage-dependent binding and unbinding of IpTx_a at a single, cytosolically accessible site located in the voltage drop across the channel. IpTx_a also induced a subconductance state in the ryanodine-modified skeletal muscle CRC and affected [³H]ryanodine binding to skeletal and cardiac SR vesicles depending on the level of channel activation. These results provide evidence for the binding of IpTx_a to a channel site different from that of ryanodine.

Phospholipids were obtained from Avanti Polar Lipids, Inc. (Alabaster, AL). All other chemicals were of analytical grade.

SR vesicle fractions enriched in [³H]ryanodine binding and Ca²⁺ release channel activities were prepared in the presence of protease inhibitors from rabbit skeletal and canine cardiac muscle as described (Meissner, 1984; Meissner and Henderson, 1987). The CHAPS (3-[(3-cholamido-propyl)dimethylammonio]-1-propanesulfonate)-solubilized skeletal or cardiac muscle 30 S Ca²⁺ release channel complex was purified and reconstituted into proteoliposomes by removal of CHAPS by dialysis (Lee et al., 1994).

IpTx_a was purified from P. imperator scorpion venom in three chromatographic steps as described (Valdivia et al., 1992; El-Hayek et al., 1995; Zamudio et al., 1997a). Synthetic IpTx_a was prepared as described (Zamudio et al., 1997a).

Single channel recordings of purified rabbit skeletal muscle or canine cardiac muscle CRCs incorporated into planar lipid bilayers were carried out as described (Tripathy et al., 1995; Xu et al., 1996). Preliminary experiments suggested that the rate of IpTx_a-induced subconductance state formation depends on channel open probability (P_o) (see Fig. 4 for a quantitative description). Therefore, experiments were conducted under “P_o-clamp” conditions with P_o values kept in the range of 0.75–1.0. For the skeletal CRC, in most cases 1–2 mM ATP was added to the cis bilayer chamber that also contained ∼5 μM free Ca²⁺ to bring the P_o in the required range. For the cardiac CRC, where no determination of the kinetic constants was attempted, one set of experiments was conducted using cytosolic contaminant Ca²⁺ (∼4 μM) (P_o = 0.3–0.5). In a second set, conducted under optimally activating conditions (10 μM cytosolic Ca²⁺), the P_o was in the range of 0.8–1. Channel activities were recorded and analyzed using commercially available instruments and a software package (Axopatch 1D, Digidata 1200, and “pClamp 6.0.3;” Axon Instruments, Burlingame, CA). Recordings were filtered at 2 or 4 kHz through an eight-pole low pass Bessel filter (Frequency Devices Inc., Haverhill, MA) and digitized at 10 or 20 kHz. The current values for the full and subconductance states at different holding potentials were obtained by constructing all point amplitude histograms and subsequent Gaussian fits of data from at least 2 min of channel recording. The probability of substate occurrence (P_substate) was calculated as time spent in substate divided by the total recording time. Another parameter CRC_substate/CRC_{full state} was calculated as time spent in substate divided by the time in full conductance state. The durations of the bursts (when the channel was in normal gating mode between the closed and full conductance levels) and subconductance states were obtained by manual positioning of the cursors.

[³H]Ryanodine binding experiments were carried out with rabbit skeletal and canine SR vesicles (protein: 150–250 μg/ml) in 0.25 M KCl, 20 mM imidazole, pH 7.0, 0.2 mM Pefabloc, 20 μM leupeptin, 0.1 mM EGTA, and [CaCl₂] necessary to set the free [Ca²⁺] in the range of 0.1 μM to 10 mM. After incubation for 24 h at 24°C, bound radioactivity was determined by a filter assay as described (Tripathy et al., 1995). Nonspecific binding was determined with a 1,000-fold excess of unlabeled ryanodine.

Results are given as means ± SD with the number of experiments in parentheses. The SD is included within the figure symbol or indicated by error bars if it is larger. Significance of differences was analyzed with Student's paired t test. Differences were considered to be significant when P < 0.05.

The skeletal and cardiac CRCs have been shown to conduct monovalent ions more efficiently than Ca²⁺ and to be essentially impermeant to Cl⁻. In 250 mM KCl solution, the single channel conductances of both isoforms are ∼770 pS (Meissner, 1994; and this study). When 15 nM native IpTx_a was added to the cytosolic side (cis bilayer chamber) of a single skeletal CRC, the channel frequently entered a subconductance state (Fig. 1,A, left and right). In contrast to the linear current–voltage relationship of the unmodified CRC, the IpTx_a-modified channel showed a slightly rectifying behavior (Fig. 1 B). The chord conductance at −40 mV holding potential was 344 ± 37 pS (∼43% of the full conductance state) and that at +40 mV was 221 ± 17 pS (∼28% of the full conductance state) (n = 8). Addition of up to 100 nM IpTx_a to the lumenal side (trans bilayer chamber) had no effect (data not shown). This suggested that IpTx_a induced subconductance states by interacting with the channel at a site accessible from the cytosolic side of the skeletal CRC. Occasionally, the CRC entered into conductance levels smaller than the main subconductance state. However, because of their very infrequent occurrences, these were not analyzed.

Fig. 2 shows similar experiments carried out with the cardiac CRC. Cytosolic addition of 15 nM IpTx_a induced a substate (Fig. 2,A) with a magnitude and rectifying current–voltage relationship (Fig. 2 B) very similar to that of skeletal CRC. As was the case with skeletal CRC, addition of IpTx_a to the lumenal side of cardiac CRC was without effect.

With symmetrical 250 mM KCl and lumenal 25 mM Ca²⁺, the current carrier through the CRC is mostly Ca²⁺ at a holding potential of 0 mV. Under such recording conditions, IpTx_a-induced subconductance states were observed for both the skeletal and cardiac muscle CRCs (data not shown). The substate conductances were ∼30% of the control full conductance values in both cases. Thus, induction of subconductance states in the CRCs by IpTx_a does not depend upon the current carrier species. From hereon, we shall describe experiments carried out on the skeletal CRC. Channel conductances and IpTx_a binding constants (see below) of the cardiac CRC are summarized and compared with the skeletal CRC in Table I.

The appearance of IpTx_a-induced substates was highly voltage dependent. The probability of substate occurrence (P_substate) was ∼0.05 at a holding potential of −60 mV. As the holding potential was made less negative, P_substate increased and reached a value close to 1 at positive holding potentials (40–60 mV) (data not shown).

The concentration dependence of IpTx_a effects on skeletal CRC was investigated at a holding potential of −20 mV by adding increasing concentrations of native IpTx_a (from 6 to 100 nM) to the cis bilayer chamber. As shown in Fig. 3, P_substate reached a value of 0.8 and greater as IpTx_a concentration was raised to 100 nM. A Lineweaver-Burk plot (Fig. 3, inset) indicated that the data could be described by Michaelis-Menten-type kinetics with a P_{substate, max} close to 1 and a K_m of 20 nM. Under the same recording conditions, a lower K_m value (∼2 nM) was observed when experiments were carried out on skeletal CRC using synthetic IpTx_a (data not shown).

The rate of IpTx_a-induced substate formation was dependent on channel open probability. Fig. 4 examines this question quantitatively by displaying the rates of subconductance state appearance and disappearance (see next section for the determination of IpTx_a association and dissociation rates) as a function of P_o. The subconductance states were rarely observed when the channel was closed, and the rate of their appearance increased linearly with P_o. In contrast, IpTx_a dissociation rate from the CRC was independent of P_o.

The concentration, voltage, and P_o dependence of IpTx_a action suggested that IpTx_a binds to the open channel at a single site located in the voltage drop across the channel. Following published procedures (Strecker and Jackson, 1989; Moczydlowski, 1992; Tinker et al., 1992a), we modeled the interaction of IpTx_a with CRC according to Scheme I.

In Scheme I, V refers to voltage across the membrane, v_on (V) and k_on (V) are the observed microscopic voltage-dependent association rate and rate constant of IpTx_a binding to the channel, respectively. The velocity of IpTx_a dissociation from the subconductance state, v_off (V) is a unimolecular microscopic rate constant, k_off, and is also voltage dependent. Assuming a Boltzmann distribution between the full and subconductance states, we can write CRC_substate/CRC_{full state} = exp [(zδFV − G₀)/ RT], where CRC_substate/CRC_{full state} refers to time spent in the subconductance state divided by the time spent in the full conductance state of the CRC.

The explicit dependence of k_on and k_off on voltage can be expressed as follows: k_on(V) = k_on(0)exp(zδ₁FV/ RT), and k_off(V) = k_off(0)exp(−zδ₂FV/RT), where z is the valence of the blocking particle, δ (= δ₁ + δ₂) is the fractional electrical distance of the binding site from the reference state at the cytosolic side, δ₁ is the distance between the transition state and the reference state, and R, T, and F have their usual meanings.

Channel activities of skeletal CRC were recorded in the presence of two different concentrations (15 and 100 nM) of native and one concentration (15 nM) of synthetic IpTx_a at various holding potentials. Ln (CRC_substate/ CRC_{full state}) values from one experiment in each condition are plotted against holding potential in Fig. 5. As predicted, the data followed Boltzmann distribution. The mean zδ values at 15 and 100 nM native IpTx_a were 1.44 ± 0.28 (n = 4) and 1.42 ± 0.28 (n = 3), respectively. The mean zδ from two experiments at 15 nM synthetic IpTx_a was 1.66. When all the data were pooled, the mean zδ was 1.49 ± 0.25 (n = 9) (Table I).

The voltage dependence of k_on and k_off was investigated for the skeletal CRC at 100 nM native and 15 nM synthetic IpTx_a at holding potentials varying from −60 to −20 mV. Dwell times corresponding to the substate and the burst mode were calculated from single channel records containing 30–150 events for each state. Following the predictions of Scheme I, the distribution of dwell times in each state followed a single exponential (Fig. 6).

Plots of the natural logarithms of k_on and k_off (obtained from exponential fits of dwell-time data) from four experiments against the holding potential followed Boltzmann distribution (not shown). The mean zδ₁ and zδ₂ from four experiments (two at 100 nM native and two at 15 nM synthetic IpTx_a) were 0.48 ± 0.04 and 1.06 ± 0.21, respectively, with a mean zδ of 1.54 ± 0.21. This latter value is almost identical to the mean zδ value of 1.49 ± 0.25, which was obtained from the plots of the natural logarithms of CRC_substate/CRC_{full state} vs. holding potential (Fig. 5 and Table I). The data further show that the unbinding rate is twice as voltage dependent as the binding rate. The mean k_on increased e-fold per +53.3 (±5.6 mV, n = 4) of applied holding potential. The mean k_off and mean K_d decreased e-fold per +24.7 (±4.9 mV, n = 4) and +17.4 (±2.3 mV, n = 4) of applied holding potential, respectively.

Next, we tested the predictions pertaining to the IpTx_a concentration dependence of v_on and v_off (Fig. 7). Skeletal CRC channel activities were recorded at a holding potential of −20 mV at five different native IpTx_a concentrations varying from 6 to 100 nM. The regression lines drawn through the mean v_on and v_off data points have coefficients of determination of 0.99 and 0.02, respectively, thus supporting the prediction that v_on is linearly dependent on IpTx_a concentration and v_off is independent of it. From the data, a k_on (v_on/[IpTx_a]) value of 1.7 × 10⁷ M⁻¹ s⁻¹ and a k_off value of 0.31 s⁻¹ were obtained, which gave a K_d (k_off/k_on) of 18.2 nM. This K_d value is very close to the K_m value of 20 nM that was obtained from the Michaelis-Menten analysis of P_substate vs. native [IpTx_a] curve under identical recording conditions (−20 mV holding potential and 250 mM KCl medium) (Fig. 3 and Table I).

Previous studies had suggested that IpTx_a preferentially binds to and activates the skeletal CRC and that cardiac CRC is a poor target of IpTx_a (Valdivia et al., 1992; El-Hayek et al., 1995; Zamudio et al., 1997a). Since we observed that IpTx_a induces subconductance states in both the skeletal and cardiac CRCs, we reexamined the effects of IpTx_a in [³H]ryanodine binding experiments using skeletal and cardiac SR vesicles. Fig. 8 shows the Ca²⁺ activation profiles of CRCs as measured by [³H]ryanodine binding to skeletal and cardiac SR vesicles in the absence and presence of 30 nM synthetic IpTx_a. As observed previously (Valdivia et al., 1992; El-Hayek et al., 1995; Zamudio et al., 1997a), IpTx_a increased [³H]ryanodine binding to skeletal SR vesicles. Activation occurred at all tested Ca²⁺ concentrations with the possible exception of 10 mM, where [³H]ryanodine binding levels were close to background. In contrast to its effect on skeletal SR vesicles, IpTx_a decreased [³H]ryanodine binding to the cardiac SR vesicles up to twofold at a Ca²⁺ concentration of 10 μM–1 mM. At low (<1 μM) and high (>1 mM) Ca²⁺ concentration, it was without a noticeable effect.

[³H]ryanodine binding measurements have been and are extensively used as indicators of the number of channels and their activities (Meissner, 1994). We did not observe an increase in P_o of the skeletal CRC nor a decrease in P_o of cardiac CRC in bilayer experiments after addition of IpTx_a (data not shown). Therefore, we were surprised to find that IpTx_a affected the [³H]ryanodine binding to skeletal and cardiac CRCs so differently. To gain further insights, we performed Scatchard analysis of [³H]ryanodine binding to both CRC isoforms in the presence and absence of IpTx_a (data not shown). IpTx_a (30 nM, synthetic) decreased the K_d of [³H]ryanodine binding to skeletal SR vesicles approximately threefold (Table II). A small increase in maximum binding (B_max) (from 13.5 to 17.5 pmol/mg) was also observed after IpTx_a treatment. In contrast, 30 nM IpTx_a increased the K_d of [³H]ryanodine binding to cardiac SR ∼3.5-fold without an appreciable effect on the B_max value. The above experiments were carried out at 100 μM free Ca²⁺. The cardiac CRC is nearly fully active at this [Ca²⁺] (Xu et al., 1996), whereas in the absence of other activating ligands, 100 μM Ca²⁺ cannot fully activate the skeletal CRC. To test if the different effects of IpTx_a on skeletal and cardiac SR vesicles were due to the different activation levels of CRCs, we performed Scatchard analysis of [³H]ryanodine binding using maximally activating conditions (100 μM Ca²⁺ and 5 mM AMP-PCP) for both CRC isoforms. Under these conditions, IpTx_a affected the [³H]ryanodine binding to the skeletal and cardiac SR vesicles in a similar manner. The K_d of [³H]ryanodine binding to the skeletal SR vesicles increased ∼1.3-fold and that for the cardiac SR vesicles ∼4.5-fold without a change in the B_max values in both cases (Table II).

In Fig. 9, we explored the cointeraction of IpTx_a and ryanodine with the skeletal CRC at the single channel level. After addition of 100 nM IpTx_a to the cis bilayer chamber, the channel current fluctuated between the full conductance and IpTx_a-induced subconductance states (Fig. 9,A, top). After further addition of 5 μM ryanodine to the cis chamber, the channel current fluctuated between two new levels (Fig. 9,A, bottom). The current–voltage relationships in presence of both IpTx_a and ryanodine are shown in Fig. 9,B. The higher subconductance state had linear current–voltage characteristics and was identified as the ryanodine-modified subconductance state of the CRC (Lai et al., 1989). The less conducting IpTx_a-induced substate of the ryanodine-modified CRC had a rectifying current–voltage characteristic. For comparison, the control and the IpTx_a-induced substate current–voltage curves (in the absence of ryanodine) from Fig. 1,B are redrawn as a dashed and dotted line, respectively, in Fig. 9 B. The new substate conductance in the presence of both IpTx_a and ryanodine was significantly smaller than the IpTx_a-induced substate (in the absence of ryanodine) at all applied holding potentials.

The present studies explored the effects of IpTx_a on purified CRCs in single channel experiments, and on SR vesicles in [³H]ryanodine binding experiments. In these studies, we provide the first evidence that IpTx_a directly interacts with both skeletal and cardiac CRCs to induce voltage- and concentration-dependent subconductance states in both isoforms by binding at a single site, located in the voltage drop across the channel. Second, we show that IpTx_a affects the [³H]ryanodine binding characteristics of the two isoforms depending on the level of channel activation.

We modeled the interaction of IpTx_a with the CRC as outlined in Scheme I. Briefly, we envisaged that when the channel opened, an IpTx_a molecule entered the conductance pathway from the cytosolic side to bind at an internal site. The model presented us with some testable predictions that were all satisfied. The distribution of dwell-times in each state (burst and subconductance) followed a single exponential. The reciprocal mean dwell-time in the burst state was linearly dependent on, and the reciprocal mean dwell-time in the subconductance state was independent of, IpTx_a concentration. A K_d value of IpTx_a binding to skeletal CRC could be calculated from the dwell-time analysis, and it was very close to that obtained from the Michaelis-Menten-type concentration-dependence curve under identical conditions.

The effective valence of the reaction leading to substate formation was determined to be ∼1.5 for the skeletal CRC (the value for cardiac CRC was ∼1.3). From the known amino acid sequence of IpTx_a (Zamudio et al., 1997a) and using the program “Isoelectric” of the GCG suite (Genetics Computer Group, Inc., Madison, WI), we calculated a net charge of 6.5 at pH 7.3. Dividing the effective valence by this number gives an electrical distance of ∼0.23 from the cytosolic side for the location of IpTx_a binding site. This is based on the assumption that all the charges of IpTx_a enter the CRC conduction pathway. The unbinding reaction of IpTx_a from the CRC has a twofold higher voltage dependence than the binding reaction. Higher voltage dependence for the unbinding rate has also been observed in case of tetrabutyl ammonium (TBA⁺)–induced substate in the cardiac CRC (Tinker et al., 1992a) and curare-induced subconductance state in the acetylcholine receptor channel (Strecker and Jackson, 1989).

IpTx_a is a relatively small peptide of 3,765 D with three pairs of cysteine residues that would stabilize the three-dimensional (3-D) conformation and help it assume a compact globular form by forming disulfide bridges (Zamudio et al., 1997a). Preliminary modeling work suggests that IpTx_a structure can be roughly approximated as a sphere of ∼2.5 nm in diameter. The selectivity filter of the skeletal CRC has a diameter of ∼0.7 nm, as choline⁺, Tris⁺, and glucose can permeate through the channel (Meissner, 1986; Smith et al., 1988), while sucrose cannot (G. Meissner, unpublished observations). Therefore, IpTx_a cannot enter the selectivity filter of the CRC. On the other hand, studies with large tetraalkyl ammonium ions (Tinker et al., 1992a) and charged local anesthetics (Tinker and Williams, 1993) have suggested that the sheep cardiac CRC has a relatively large vestibule facing the cytosolic side. Theoretical considerations have suggested that a significant proportion of the voltage drop may fall over the wide vestibule of a channel (Jordan, 1986). Our studies showed that the vestibule has to be at least 2–2.5-nm wide to accommodate IpTx_a, and at least ∼23% of the voltage falls over the vestibule. Interestingly, 3-D cryo-electron micrographic reconstruction of the skeletal muscle CRC (Radermacher et al., 1994; Serysheva et al., 1995) shows a central opening of ∼5 nm in the cytoplasmic domain (facing the transverse tubule), and thus supports the presence of a wide vestibule facing the cytoplasmic side of the CRC.

We considered it unlikely that IpTx_a would produce subconductance states by a surface charge mechanism (Green and Andersen, 1991). The estimated decrease in local [K⁺] using the Guoy-Chapman double-layer theory (McLaughlin, 1977) was only approximately fourfold (assuming zero net surface charge near the cytosolic face of the CRC and the IpTx_a net positive charge of 6.5 spread out over a sphere of ∼2.5-nm diameter). This decrease (from 250 to 62.5 mM) was not sufficient to cause the subconductance state as the K_d of K⁺ for the cardiac CRC has been shown to be ∼20 mM (Tinker et al., 1992b). However, if the positive charges of IpTx_a are clustered in a much smaller area (leading to a higher charge density), the local [K⁺] could be much lower, which may explain the formation of subconductance state. The above possibility is also not very likely as the ratios of subconductance to full conductance values were similar, and there was no indication of any relief of block, when the [KCl] in the bilayer chambers was increased symmetrically from 250 to 500 mM (data not shown). However, we observed an increase in the K_d of IpTx_a binding to CRC when [KCl] was raised from 250 to 500 mM (our unpublished studies). Similar effects of increasing [KCl] on K_d have been observed by other investigators studying other toxins (Anderson et al., 1988; Lucchesi and Moczydlowski, 1990).

In our experiments with the purified CRCs in bilayer experiments, substates of different amplitudes are sometimes observed even in control recordings. These substates show linear current–voltage relationships. Therefore, it is conceivable that IpTx_a preferentially binds to and stabilizes a normally occurring CRC substate. However, we think this to be an unlikely mechanism as the IpTx_a-induced substates had different conductances at positive and negative holding potentials (Figs. 1 and 2, and Table I).

How then does IpTx_a reduce channel conductance? Is it by partial occlusion of the channel pore or by a conformational change? Tetrabutyl and tetrapentyl ammonium–induced substates in the cardiac CRC have been explained in terms of partial occlusion of the conduction pathway (Tinker et al., 1992a). Convincing arguments have been put forward in favor of partial block of acetylcholine receptor channel by curare (Strecker and Jackson, 1989) and complete block of a maxi Ca²⁺-activated K⁺ channel by charybdotoxin (Anderson et al., 1988; MacKinnon and Miller, 1988). In the former case, displacing curare from the agonist binding sites of acetylcholine receptor channel had no effect on the kinetics of subconductance state occurrence. In the latter case, enhancement of external toxin dissociation by internal K⁺ and competition for the toxin-binding site by external tetraethylammonium ions gave strong support to the occlusion model. On the other hand, Zn²⁺-induced substates in canine cardiac sarcolemma sodium channel (Schild et al., 1991) and proton and deuterium ion-induced substates in a voltage-dependent Ca²⁺ channel (Prod'hom et al., 1987; Pietrobon et al., 1989) have been explained in terms of conformational changes; as in the above two cases, v_off was dependent on the blocker concentration. Furthermore, in case of dendrotoxin-I (DTX-I)-induced substate formation in a maxi Ca²⁺-activated K⁺ channel, a conformational change mechanism was also favored as the toxin converted the linear control current–voltage relationship to a rectifying one, though in this case v_on and v_off were linearly dependent on and independent of the toxin concentration (Lucchesi and Moczydlowski, 1990). In the present case, IpTx_a binding converted an ohmic CRC to a rectifying one. Furthermore, IpTx_a-modified CRCs (both skeletal and cardiac) had higher K_ds of [³H]ryanodine binding under fully activating conditions (conditions similar to bilayer recording conditions) (see below for further discussion on this aspect). These observations favor an IpTx_a-induced conformational change in both the skeletal and cardiac CRCs as the mechanism causing the subconductance state. Alternatively, resultant conformational changes by the membrane electric field may cause IpTx_a to differentially occlude the CRC at positive and negative holding potentials to produce substates of different conductances. Regardless of whether partial occlusion of the pore or a conformational change is the mechanism causing subconductance state formation, our studies strongly suggest the existence of a cytosolically accessible IpTx_a binding site in the conductance pathway of both the skeletal and cardiac CRC.

Most of our experiments were conducted using native, purified IpTx_a. In some experiments carried out using synthetic IpTx_a, we observed that the synthetic toxin was ∼10-fold more effective in inducing substates in skeletal and cardiac CRCs than its native counterpart. The reasons for this difference are not clear. One possibility is that during purification the native toxin lost part of its biological activity.

[³H]ryanodine binding correlates reasonably well with CRC activation levels. Because of its simplicity and convenience, it is extensively used to explore the effects of many drugs/toxins on CRC. But as this study shows, direct measurements of channel activity can provide insights that might not be detected by [³H]ryanodine binding measurements. When the effects of IpTx_a were assessed at 100 μM free Ca²⁺, an increase in [³H]ryanodine binding level for the skeletal but a decrease for the cardiac SR vesicles was observed (Fig. 8), though the net effect is the formation of subconductance states in both isoforms in bilayer experiments. Furthermore, using purified skeletal CRC, we demonstrated that IpTx_a can still bind and induce substates of slightly smaller conductances in the ryanodine-modified channel (Fig. 9, A and B). Similar paradoxical results were observed in studies exploring the effects of polyamines on the CRCs. Polyamines increased [³H]ryanodine binding to skeletal SR vesicles approximately fivefold (Zarka and Shoshan-Barmatz, 1992). However, in single-channel studies, it caused a blockade of the cardiac CRC, and no increase in P_o was observed (Uehara et al., 1996). Though [³H]ryanodine binding and single channel experiments were done on different CRC isoforms, the above experiments and the present study call for caution in interpreting [³H]ryanodine binding results in the presence of other drugs and toxins.

The apparently paradoxical effects of IpTx_a on skeletal and cardiac SR vesicles in [³H]ryanodine binding measurements led us to investigate its effects in detail. At 100 μM Ca²⁺, IpTx_a caused a decrease in K_d (with a small increase in B_max) for the skeletal SR vesicles, but an increase in K_d with little or no change in B_max for the cardiac SR vesicles (Table II). The affinity of [³H]ryanodine binding to SR vesicles correlates with the CRC activation level, suggesting that ryanodine binds to an open channel (Meissner, 1994). The IpTx_a-induced long-lived substates in the skeletal CRC would thus provide a favorable environment over the rapidly gating control low channel activity (at 100 μM Ca²⁺) for ryanodine entry into the CRC interior and subsequent binding. For the cardiac CRC, which is maximally active at 100 μM Ca²⁺, the substates provided no further improvement. On the contrary a decrease was observed, probably because the IpTx_a-modified channel may have reduced affinity for ryanodine. In favor of this idea, we observed that under maximally activating conditions for both CRC isoforms (100 μM Ca²⁺ and 5 mM AMP-PCP), IpTx_a decreased the K_ds of [³H]ryanodine binding to both skeletal and cardiac SR vesicles (Table II). Furthermore, IpTx_a increased [³H]ryanodine binding to cardiac SR vesicles under minimally activating conditions (∼1 μM Ca²⁺) just as it increased [³H]ryanodine binding to skeletal SR vesicles under suboptimally activating conditions (100 μM Ca²⁺) (our unpublished studies). Thus, our data show that the IpTx_a-modified CRC (both skeletal and cardiac) has a K_d of [³H]ryanodine binding intermediate between those of the minimally and maximally activated channel.

In conclusion, the present work demonstrates a direct, high affinity interaction of IpTx_a with both the skeletal and cardiac CRCs. The use of IpTx_a should aid in the structural elucidation of the ion-conductance pathway of the skeletal and cardiac muscle CRCs. Furthermore, we show that experiments exploring effects of drugs/toxins on CRCs by [³H]ryanodine binding measurements must be interpreted with caution and be correlated with measurements of channel activity.

This work was supported by National Institutes of Health (NIH) grant HL-55438 and Grant-in-Aid from the American Heart Association (to H.H. Valdivia) and NIH grants AR-18687 and HL-27430 (to G. Meissner).

CRC

Ca²⁺ release channel

IpTx_a

imperatoxin activator

RyR

ryanodine receptor

SR

sarcoplasmic reticulum