Characterization of recombinant rabbit cardiac and skeletal muscle Ca2+ release channels (ryanodine receptors) with a novel [3H]ryanodine binding assay.

A rapid assay for high affinity [3H]ryanodine binding to 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid (CHAPS)-solubilized recombinant or native Ca2+ release channel proteins (ryanodine receptor, RyR) was devised. The key to preservation of high affinity [3H]ryanodine binding sites in the presence of increasing concentrations of CHAPS was the addition of phosphatidylcholine. This assay was used to characterize the equilibrium and kinetic properties of [3H]ryanodine binding to recombinant skeletal (RyR1) and cardiac (RyR2) Ca2+ release channels and the effects on binding of physiological modulators including ATP, Ca2+, and Mg2+. Both RyR1 and RyR2 had a single high affinity ryanodine binding site and low affinity sites, but [3H]ryanodine binding to recombinant RyR2 was not sensitive to ATP activation or Ca2+ inactivation and was less sensitive to Mg2+ inhibition. The [3H]ryanodine binding assay was used to estimate the expression level of recombinant RyR2 and RyR1, and to show that RyR2 can be expressed at very high levels in HEK-293 cells. Analysis of the properties of recombinant RyR2 and RyR1 by measurement of intracellular Fura-2 fluorescence revealed that the different properties of RyR2 and RyR1 are retained in the recombinant expressed proteins.

Ca 2ϩ release channels (ryanodine receptors or RyR) 1 from sarcoplasmic reticulum are formed as homotetramers of 565,000-Da subunits (1,2). The activities of native Ca 2ϩ release channels from skeletal (RyR1) and cardiac muscle (RyR2) are modulated by a variety of physiologically relevant agents and by a series of pharmacological compounds (for details, see Refs. [3][4][5]. Ryanodine binds to Ca 2ϩ release channels with high affinity and was used in the original identification of Ca 2ϩ release channel proteins (6). At concentrations below 10 M, ryanodine locks the channel into a subconductance state and at concentrations above 10 M it blocks conductance. Ryanodine is often used as a probe of conformation to indicate the open state of the channel.
A major goal in studies of ryanodine receptors is to understand structure/function relationships through expression and analysis of mutant forms of both skeletal muscle (RyR1) and cardiac (RyR2) isoforms. In order to carry out such studies, a series of useful assays for ryanodine receptors expressed in homologous cell culture have been developed, including planar bilayer assays (7,8), Ca 2ϩ photometry (9), and Ca 2ϩ imaging (10). In this paper, we describe modifications to the [ 3 H]ryanodine binding assay (11) which makes it useful for microscale analyses of expressed ryanodine receptors. We also describe the expression of a rabbit cardiac ryanodine receptor (RyR2) clone in HEK-293 cells and we characterize some of its properties through use of the [ 3 H]ryanodine binding assay and through Ca 2ϩ photometry.
Cell Culture and DNA Transfection-HEK cells were maintained and DNA transfection was carried out by the calcium phosphate precipitation method described previously (8,15).
Immunocytochemical Staining-HEK-293 cells transfected with RyR1 cDNA or RyR2 cDNA were stained with mAb 34C and mAb C3-33 to detect recombinant RyR2 or RyR1 protein in the cells according to the procedures described previously (7). Alkaline phosphatase-conjugated anti-mouse IgG, 5-bromo-4-chloro-3-indolyl phosphate, and p-nitrotetrazolium chloride blue were used to develop color.
Solubilization of Transfected HEK-293 Cells-Transfected HEK-293 cells grown in 100-mm Petri dishes were washed twice with 5 ml of PBS (137 mM NaCl, 2.7 mM KCl, 8 mM Na 2 HPO 4 , 1.5 mM KH 2 PO 4 , pH 7.4) and harvested in PBS containing 5 mM EDTA. Cells were collected by centrifugation at 4000 rpm for 10 min in a Sorvall SS-34 rotor. Cell pellets were washed with PBS and centrifuged again. Unless otherwise indicated, the cell pellet from a single 100-mm dish was then solubilized in 1 ml of buffer containing 150 mM NaCl, 50 mM TrisCl, pH 8.0, 1% CHAPS, 5 mg/ml PC, and a mix of protease inhibitors (0.1 mM AEBSF, 1 mM benzamidine, 1 mg/ml leupeptin, 1 mg/ml pepstatin A, 1 mg/ml aprotinin, and 1 mg/ml E64) on ice for 45 min. Solubilized proteins were obtained by removing the debris through centrifugation at 4000 rpm, at 4°C for 10 min.

Measurement of [ 3 H]Ryanodine Equilibrium
Binding-Aliquots of 25 l of solubilized proteins were diluted into binding buffer B (a total of 0.25 ml) containing various concentrations of [ 3 H]ryanodine, 0.5 M KCl, 1 mM ATP, 20 M free Ca 2ϩ , 0.2 mM EGTA, 50 mM Hepes-Tris, pH 7.1, and the protease inhibitor mix. Nonspecific binding was determined using a 1000-fold excess of unlabeled ryanodine. After 2 h at 37°C, aliquots of the samples were diluted with 1 ml of ice-cold washing buffer (25 mM Hepes, pH 7.1, 0.25 M KCl) and placed on Whatman GF/B membrane filters pre-soaked with 1% polyethyleneimine in washing buffer. Filters were washed three times with 6 ml of washing buffer. The radioactivity remaining on the filters was determined by liquid scintillation counting to obtain bound [ 3 H]ryanodine. All binding assays were carried out in duplicate.
To assess the effects of modulators on high affinity ryanodine binding, cell lysates (25 l) were incubated with 2.5 nM [ 3 H]ryanodine in the presence or absence of each modulator or in the presence or absence of different concentrations of Ca 2ϩ or Mg 2ϩ in a final volume of 0.25 ml at 37°C for 2 h.
Free Ca 2ϩ was calculated using the apparent binding constants described by Fabiato (18).

Measurement of Association and Dissociation Kinetics for [ 3 H]
Ryanodine Binding-The rate of association of [ 3 H]ryanodine binding was measured by terminating the reaction by rapid filtration at times ranging from 30 s to 300 min after addition of solubilized proteins into binding buffer B with 10 nM [ 3 H]ryanodine and incubation at 37°C. Nonspecific binding was measured in the presence of 10 M unlabeled ryanodine. Dissociation of [ 3 H]ryanodine from the equilibrium complex was determined by equilibrating 10 nM [ 3 H]ryanodine with solubilized proteins in buffer B for 2 h at 37°C, followed by the addition of aliquots of the incubation mixture into a 50-fold excess of buffer B, without added ATP and Ca 2ϩ or with 40 M unlabeled ryanodine. Determination of specific binding was made at times ranging from 1 to 300 min following dilution. Nonspecific binding was measured in the presence of 10 M unlabeled ryanodine.
Fluorescence Measurements-Fluorescence measurement was described previously (9). For the generation of EC 50 values in caffeine response curves, incremental concentrations of caffeine were added gently with a Pipetman directly on top of the selected cells. Between the peak amplitudes of caffeine response, washing was carried out over 1-2 min, beginning shortly after the plateau was reached. To obtain peak amplitude (ratio (340/380 nm)) and mean half-rise time (t 1/2 ) in the 20 mM-caffeine responses, the solution changing system was used in order to minimize inaccuracy, especially for the t 1/2 measurements. Values for peak amplitudes, in arbitrary 340/380 nm ratio units, were calculated by measuring the peak ratio minus the basal ratio, while t 1/2 values (s) were calculated from measurements of the time required from the onset of the response to the half point of the response for 20 mM caffeineinduced transients. Fluorescence emission at 510 nm from about 30 cells in a specific region selected with an adjustable aperture was measured with dual excitation wavelengths of 340 and 380 nm, alternating at 50 Hz. The emission ratio at 340 and 380 nm was calculated using PTI Felix software and presented as the index of [Ca 2ϩ ] i . No systematic attempt was made to convert ratio to apparent [Ca 2ϩ ] i .
Data Analysis-Scatchard analysis was used to determine the dissociation constant (K d ) and maximal binding capacity (B max ) from equilibrium binding data. The pseudo-first-order association rate constant (k 1 ) was calculated as described previously (19 -21), based on the observed association rate constant (k obs ), which was calculated from the slope of the first-order plot of the data obtained in the association kinetics experiments, in which: SB o is the specific binding at equilibrium and SB is the specific binding at time t. The dissociation rate constant (k Ϫ1 ) was measured from the slope of the first-order plot of the data in the dissociation kinetics experiments as described previously (19 -21), in which: SB is the specific binding at time t, and SB 0 is the specific binding at time 0. All data were analyzed using Microcal Origin software (Microcal Software Ltd., Northampton, MA). Data are expressed as mean Ϯ S.E. A paired Student's t test was used for statistical comparison of mean values between paired groups. An unpaired Student's t test was used for evaluation of the mean values between groups. A value of p Ͻ 0.05 was considered to be statistically significant. EC 50 or IC 50 values were obtained by fitting the curves with an equation for logistic dose response using Microcal Origin software.

RESULTS
Transient Expression of RyR1 and RyR2 cDNAs-As we were interested in analyzing [ 3 H]ryanodine binding to recombinant RyR1 and RyR2, full-length RyR1 and RyR2 cDNAs were subcloned into pcDNA3.1(Ϫ), and transfected into HEK 293 cells (8). The characteristics of recombinant RyR1 have been described (8), but the characteristics of recombinant RyR2 have not been so clearly defined. Immunostaining of transfected cells using mAb 34C and mAb C3-33 for RyR1 and RyR2 respectively was used to demonstrate the expression of both proteins and their absence in pcDNA-transfected cells ( Fig. 1). High transfection efficiency (up to 60%) could be observed around the edges of clusters of cells (9). There was no apparent difference in transfection efficiency between RyR1 and RyR2 cDNA-transfected cells (Fig. 1, A-C).
We also used these two monoclonal antibodies to detect the presence of the expressed proteins in the lysates of transfected HEK-293 cells, using heavy cardiac and skeletal muscle sarcoplasmic reticulum preparations as controls. Antibody 34C detected both RyR1 and RyR2 proteins in transfected cells (Fig.  1D). Recombinant RyR2, like native RyR2, had a slightly higher mobility than recombinant and native RyR1 in SDS-PAGE. This result is consistent with previous reports for native RyR proteins (22-24), but does not agree with the deduced molecular weights (14). Monoclonal antibody C3-33 detected only native and recombinant RyR2 (Fig. 1E). Western blotting of non-transfected cells did not detect any trace of a protein with a similar size with either of these two monoclonal antibodies ( Fig. 1, D and E), demonstrating that HEK-293 cells do not contain detectable amount of endogenous RyR proteins.
Caffeine-induced Ca 2ϩ Release from HEK-293 Cells Transfected with RyR1 or RyR2 cDNA Is Inhibited by Ryanodine and Thapsigargin-We measured Ca 2ϩ release in HEK-293 cells in which RyR2 was expressed, using Fura-2 fluorescence as an assay (9). Caffeine and ryanodine are the standard pharmacological tools for demonstrating the activation and inactivation of the ryanodine receptor (25). Fig. 2A shows that 20 mM caffeine caused a rapid Ca 2ϩ release in HEK-293 cells transfected with RyR2 cDNA. This response could be obtained repeatedly, provided that previously applied caffeine was washed out. Subsequent addition of 50 M ryanodine did not cause significant intracellular Ca 2ϩ release in three of four experiments over a period of 2-3 min. The first application of 20 mM caffeine after the addition of ryanodine induced a release phase that was similar to the previous one, indicating that ryanodine did not have any direct effect on caffeine-induced Ca 2ϩ release. Further caffeine challenges greatly diminished and eventually could not stimulate any Ca 2ϩ release from the internal store, even in the absence of ryanodine in the extracellular solution ( Fig. 2A). Similar results were also obtained in HEK-293 cells transfected with RyR1 cDNA for caffeine and ryanodine (data not shown). In untransfected cells, 20 mM caffeine and 50 M ryanodine did not cause any significant Ca 2ϩ release (data not shown). The results illustrate that recombinant RyR2 is sensitive to these two agents and that ryanodine binds only to open-state channels, induced by caffeine. Ryanodine then blocks those channels that were responsive to caffeine.
Thapsigargin is a potent and specific inhibitor of sarco(endo)plasmic reticulum Ca 2ϩ -ATPases. Ca 2ϩ is probably released from intracellular stores, through endogenous leaks that become apparent following Ca 2ϩ pump inhibition (26,27). In untransfected HEK-293 cells, 1 M thapsigargin caused a slow phase of Ca 2ϩ release (n ϭ 3, data not shown). In HEK-293 cells transfected with RyR2 cDNA, in which caffeine-induced Ca 2ϩ release was evident, application of 1 M thapsigargin induced Ca 2ϩ release similar to that observed in untransfected cells, and abolished caffeine-induced Ca 2ϩ release ( Fig. 2B), demonstrating that caffeine-induced Ca 2ϩ release was from thapsigargin-sensitive Ca 2ϩ stores.
Characterization of Caffeine-induced Ca 2ϩ Release from Internal Stores in HEK-293 Cells Transfected with RyR2 and RyR1 cDNA- Fig. 3 (A and B) shows representative fluorometric recordings for intracellular Ca 2ϩ transients in response to incremental applications of caffeine (0.03-30 mM) in HEK 293 cells transfected with RyR2 and RyR1 cDNA. Caffeine up to 30 mM did not cause any significant Ca 2ϩ release in pcDNAtransfected cells (not shown). Transient traces in Fig. 3 (A and  B) showed that RyR2 cDNA-transfected cells responded to lower concentrations of caffeine than RyR1 cDNA-transfected cells. Threshold values ranged from 0.052 to 0.1 mM for RyR2 and from 0.3 to 0.52 mM for RyR1, indicating that the sensitivity to caffeine in recombinant RyR2 is higher than that in recombinant RyR1. EC 50 values for RyR2 and RyR1 were calculated by measuring the peak amplitude in each caffeine application and normalized to the peak amplitude for maximal Ca 2ϩ release induced by 30 mM caffeine. The EC 50 values, generated from fitting the caffeine dose-response curves with an equation for logistic dose response, were 0.99 mM for RyR2 and 2.77 mM for RyR1 with Hill coefficients of 1.08 for RyR2 and 1.66 for RyR1 (Fig. 3C). Since 20 -30 mM caffeine can release all of the Ca 2ϩ from caffeine-sensitive stores (28), the response of RyR2 and RyR1 to caffeine was reflected in the peak amplitude Caffeine was washed out to restore the resting Ca 2ϩ level, after peak amplitude (peak of change in the ratio of fluorescence at 340/380 nm, indicating peak changes in [Ca 2ϩ ] i ) was obtained. About 60 s after resting Ca 2ϩ level was achieved, 50 M ryanodine was applied. Subsequent addition of 20 mM caffeine caused one Ca 2ϩ release similar to that before addition of ryanodine, but subsequent responses to caffeine were blocked, even in the absence of extracellular ryanodine. B, the same strategy was used to observe the effect of 1 M thapsigargin.
(ratio (340/380 nm)) and the mean half-rise time (t 1/2 ) in the 20 mM caffeine responses. These two values, calculated from independent transients, are presented in Table I. Significant differences in both peak and t 1/2 between RyR2 and RyR1 cDNAtransfected cells indicate that recombinant RyR2 and RyR1 have different kinetics of caffeine-induced Ca 2ϩ release.

Characterization of a [ 3 H]Ryanodine Binding Assay-
The detergent CHAPS is widely used in solubilization and purification of muscle sarcoplasmic reticulum Ca 2ϩ release channels (6,29). We solubilized rabbit skeletal muscle microsomes at a concentration of 0.4 mg/ml in 1% CHAPS (16 mM), which is 2-fold higher than its critical micelle concentration. Equilibrium binding experiments were carried out following 10-fold dilution with 0.1% CHAPS in the binding buffer, and results were compared with controls with no CHAPS in the binding buffer. Scatchard plot analysis showed that B max for the solubilized protein was significantly lower than that for controls (3.6 versus 7.7 pmol of ryanodine/mg of protein) (p Ͻ 0.05) with no alteration of K d (Fig. 4A and Table II). Although 0.1% CHAPS in binding buffer has been reported to stabilize [ 3 H]ryanodine binding to RyR (30), our observation of a decrease in the number of high affinity binding sites suggested that endogenous lipid molecules were stripped away from the solubilized RyR protein and replaced by CHAPS molecules (31). To avoid the formation of a preponderance of CHAPS:RyR protein micelles, we tested the effect of addition of exogenous phospholipids, since PC has been used to stabilize the cardiac ryanodine receptor during its purification from dog heart sarcoplasmic reticulum in the presence of CHAPS (22). The addition of 5 mg/ml PC in the solubilization buffer led to a higher recovery of high affinity [ 3 H]ryanodine binding sites. We observed a B max of 11.8 pmol/mg of protein with an unchanged K d (Fig. 4A and Table II), indicating that the ryanodine binding function could be preserved by increasing the phospholipid concentration to form mixed micelles (detergent:protein:phospholipid).
[ 3 H]Ryanodine Binding to Recombinant Channels-In the situation where RyR1 cDNA-transfected HEK-293 cells were less than 40% confluent, solubilizing the cells with 1% CHAPS alone resulted in no significant [ 3 H]ryanodine binding (Fig.  4B). However, addition of exogenous PC (5 mg/ml) in the solubilizing buffer in the same batch of cells led to preservation of [ 3 H]ryanodine binding (Fig. 4B). We also observed that treat-  ment of transfected cells that were ϳ80 -90% confluent with 1% CHAPS alone did not abolish [ 3 H]ryanodine binding (see below), suggesting that increased endogenous lipid might help to preserve [ 3 H]ryanodine binding. In five independent experiments (of which two were with ϳ40% confluent cells and three were with ϳ80% confluent cells), an average B max of 0.18 pmol/mg of protein with a K d of 2.2 nM (Table II) was observed when cells were solubilized with 1% CHAPS and 5 mg/ml PC. The K d value was similar to that of native RyR1.
The effect of CHAPS and PC on the extraction of protein and on [ 3 H]ryanodine binding was determined by assaying different concentrations of CHAPS in solubilizing and binding buffers. We increased the concentration of CHAPS in the presence of PC, but maintained a PC:CHAPS ratio of 0.5:1 except for 2% CHAPS, in which the PC concentration was maintained at 5 mg/ml. The presence of PC in the solubilizing buffer increased both [ 3 H]ryanodine binding and protein concentration in the lysate from ϳ80% confluent cells (Fig. 5, A and B). At 1% CHAPS in the presence of 5 mg/ml PC, solubilization of protein began to plateau. When the same batch of cells was solubilized with CHAPS alone, protein solubilization was increased, but [ 3 H]ryanodine binding decreased (Fig. 5, A and B). Increasing concentrations of CHAPS in the binding buffer also inhibited [ 3 H]ryanodine binding (Fig. 5C) (Table II) and cardiac muscle (2.1 Ϯ 0.3 nM, n ϭ 4) using sarcoplasmic reticulum microsomes, under identical conditions. B max was 0.18 Ϯ 0.02 and 0.76 Ϯ 0.07 pmol/mg of protein for recombinant RyR1 and RyR2, respectively. There was no specific binding in lysates isolated from pcDNA-transfected HEK cells.
Association and Dissociation Kinetic Studies of High Affinity Binding Sites-The association kinetics for high affinity binding were measured for recombinant RyR1 and RyR2, with 10 nM [ 3 H]ryanodine. Ryanodine binding reached saturation within 60 min in both RyR1 and RyR2. The association rate constant (k 1 ), calculated from the slope (k obs ) of the pseudo-first-order plot was similar for recombinant RyR1 and RyR2 ( Fig. 6A and Table III) using the equation listed in the legend to Table III. The rate of dissociation of [ 3 H]ryanodine from the high affinity binding sites in recombinant RyR1 and RyR2 was also examined by 50-fold dilution into binding buffer B without added ATP and Ca 2ϩ so that rebinding of [ 3 H]ryanodine was prevented. The first-order plots were linear for both RyR1 and RyR2 (Fig. 6B), indicating single high affinity ryanodine binding sites, which do not have cooperativity. The dissociation rate constant, k Ϫ1 , was similar for recombinant RyR1 (0.01056 min Ϫ1 ) and RyR2 (0.01106 min Ϫ1 ) ( Table III). The dissociation constant (K d ) for ryanodine binding to high affinity sites calculated from the kinetic rate constants for association and dissociation (K d ϭk Ϫ1 /k 1 ) was virtually the same for RyR1 (1.8 nM) and RyR2 (1.8 nM) and was similar to that obtained from Scatchard analysis (Table III).
In studies with skeletal and cardiac muscle sarcoplasmic reticulum, low affinity binding sites for ryanodine were identified, and occupation of low affinity sites decreased the dissociation rate of bound ryanodine from high affinity sites, suggesting that positively cooperative interactions occur between high and low affinity sites (20,21). To determine whether this characteristic is preserved in recombinant RyR1 and RyR2, we tested for low affinity sites by 50-fold dilution in binding buffer B containing 40 M unlabeled ryanodine, without added ATP and Ca 2ϩ . A dramatic decrease in [ 3 H]ryanodine dissociation from high affinity sites was observed in both recombinant RyR1  Table II. and RyR2 when 40 M ryanodine was present in the dissociation buffer (Fig. 6B). The k Ϫ1 values were 0.00167 min Ϫ1 and 0.00115 min Ϫ1 for recombinant RyR1 and RyR2, respectively (Table III). These results suggest that recombinant RyR1 and RyR2, like the native channels, possess both high and low affinity binding sites and that there is positive cooperativity between these binding sites.

Effects of Ca 2ϩ , AMPPCP, and Mg 2ϩ on [ 3 H]Ryanodine
Binding-The effects of the physiological modulators of muscle Ca 2ϩ release channels, Ca 2ϩ , ATP, and Mg 2ϩ , were tested on recombinant RyR1 and RyR2. The Ca 2ϩ dependence of [ 3 H]ryanodine binding and the effect of 1 mM AMPPCP on the Ca 2ϩ depend-ence of [ 3 H]ryanodine binding to recombinant RyR2 are shown in Fig. 7A and for RyR1 in Fig. 7B. At low Ca 2ϩ concentrations (pCa Ն 7) in the binding buffer without ATP, no significant binding was detected in either recombinant RyR1 or RyR2. [ 3 H]Ryanodine binding was activated by increasing Ca 2ϩ concentrations, reaching optimal binding at about pCa 5. At higher Ca 2ϩ concentration (pCa Ͻ 4), [ 3 H]ryanodine binding was decreased in RyR1. However, binding to RyR2 was not sensitive to inactivation by high Ca 2ϩ concentration up to pCa 2.0. Thus, the Ca 2ϩ dependence of [ 3 H]ryanodine binding was bell-shaped in RyR1 but not in RyR2. The EC 50 and Hill coefficient for Ca 2ϩ -activation in RyR2 and RyR1 and the IC 50 and Hill coef-  These results are all in agreement with those obtained in skeletal and cardiac muscle sarcoplasmic reticulum (3,5).
In the presence of 1 mM AMPPCP, the Ca 2ϩ activation of [ 3 H]ryanodine binding was shifted to the left in RyR1 (Fig. 7B), but was not changed in recombinant RyR2 (Fig. 7A). The EC 50 values calculated from three independent experiments are listed in Table IV. The inactivation curve was shifted to the right in RyR1 with IC 50 from pCa 2.33 to pCa 1.81 (p Ͻ 0.05) ( Fig. 7B and Table IV). These results suggest that the sensitivity to ATP was higher in recombinant RyR1 than in RyR2, consistent with those reported in muscle (34,35).
Mg 2ϩ at millimolar concentrations inhibits Ca 2ϩ release and [ 3 H]ryanodine binding in skeletal muscle sarcoplasmic reticulum and, to a lesser extent, in cardiac muscle sarcoplasmic reticulum (3,5). The inhibition of [ 3 H]ryanodine binding to recombinant RyR1 and RyR2 by increased Mg 2ϩ concentrations is shown in Fig. 8. The IC 50 value was 3.1 Ϯ 1.2 mM (n ϭ 3) for RyR1 and 8.7 Ϯ 0.7 mM (n ϭ 3) for RyR2 (p Ͻ 0.05), indicating a significant difference in the mechanism by which Mg 2ϩ affects the [ 3 H]ryanodine binding. In both proteins, Mg 2ϩ is proposed to displace Ca 2ϩ competitively from its activation site(s) (5 (6). While native RyR1 from sarcoplasmic reticulum could be dissolved in CHAPS and purified to homogeneity without lipids and without loss of ryanodine binding, the highest [ 3 H]ryanodine binding ability was obtained after the purified RyR1 was incorporated into liposomes (6).
Our modified assay has proven to be procedurally simple and  to be highly reliable for the investigation of the properties of [ 3 H]ryanodine binding to recombinant channels and the effects of modulators on [ 3 H]ryanodine binding. One of the important applications of the assay has been to show that expression levels for functional expressed channels can be readily calculated, since the bulk of the expressed recombinant proteins are extracted for assay. After solubilization with 1% or 2% CHAPS in the presence of PC, the debris of transfected cells with RyR cDNAs had less than 5% of the [ 3 H]ryanodine binding activity in solubilized proteins. Assuming that the expressed RyR proteins are all assembled into tetrameric complexes, that one tetramer has one high affinity binding site, and that a tetrameric receptor binds about 400 pmol of ryanodine/mg of protein (6, 7), we calculated that our solubilized recombinant protein in whole cell lysates is about 0.045% (0.18/400) of total cellular proteins for recombinant RyR1 (B max ϭ 0.18 pmol/mg) and about 0.2% (0.76/400) for recombinant RyR2 (B max ϭ 0.76 pmol/mg). By contrast, the expression of RyR1 was only 0.05% in isolated microsomes from transfected COS-1 cells (7). Thus, the expression of RyR protein in HEK-293 cells is indeed higher than in COS-1 cells (8). The high expression of RyR2 in HEK-293 cells could be confirmed directly by staining of gels with Coomassie Blue and silver, when compared with pcDNAtransfected cells. The corresponding band for RyR1 could not be distinguished, probably because it shared a similar molecular weight with endogenous proteins (data not shown). The transfection efficiency was similar on the basis of immunochemical staining (Fig. 1), suggesting that RyR2 is truly expressed to a higher level than RyR1 in similarly transfected cells.
RyR1 has been expressed in COS-1, CHO, and HEK-293 cells, and its functional properties have been characterized (7,8,36,37). Although RyR2 has also been expressed in both CHO cells and Xenopus oocytes (38,39), its properties have not been well characterized. Using stably transfected CHO cells with RyR2 cDNA, Imagawa et al. (39) observed a B max of 19.8 fmol/mg of cellular protein in [ 3 H]ryanodine binding, reflecting the limitation of the expression system and perhaps the [ 3 H]ryanodine binding assay. With our transient expression system, expression levels were 40-fold higher for RyR2. This has enabled us perform more detailed analyses of the equilibrium and kinetic characteristics of the recombinant channel, the Ca 2ϩ dependence of activation and inactivation, and the responses to physiological modulators with the [ 3 H]ryanodine binding assay. We have demonstrated that [ 3 H]ryanodine binding to recombinant RyR2 and RyR1 has similar equilibrium and kinetic properties, but different characteristics in Ca 2ϩ dependence and ligand responses to ATP and Mg 2ϩ . This may reflect the likelihood that binding sites for ATP and Ca 2ϩ in RyR2 and RyR1 molecules are not well conserved. One approach to the understanding of the molecular basis for the different properties of RyR1 and RyR2 will be to generate mutations or chimeras surrounding potential functional domains. The [ 3 H]ryanodine binding assay should be useful in assaying the functional consequences of those mutations and chimeras that affect channel opening and ryanodine binding.
Caffeine-induced Ca 2ϩ Release-Caffeine, a potent activator of Ca 2ϩ release in native skeletal and cardiac muscle preparations, released Ca 2ϩ from RyR1 and RyR2 cDNA-transfected HEK-293 cells in a dose-dependent manner. However, recombinant RyR2 was activated by lower concentrations of caffeine than RyR1 in terms of both threshold and EC 50 values, indicating that recombinant RyR2 is more sensitive to caffeine than RyR1. This is in agreement with results obtained with skeletal and cardiac sarcoplasmic reticulum (35,40).
The time course of Ca 2ϩ release for recombinant RyR2 was faster than that for recombinant RyR1, and the peak amplitude of the Ca 2ϩ transient was higher for recombinant RyR2. This does not necessarily mean that the same dose of caffeine releases more Ca 2ϩ in RyR2 cDNA-transfected cells, since high concentrations of caffeine can release Ca 2ϩ completely from both skeletal and cardiac muscle sarcoplasmic reticulum (28). Thus, the different caffeine response for recombinant RyR1 and RyR2 might reflect different properties of these two channels. We do not know, however, to what extent the differences in expression level between RyR1 and RyR2 cDNA-transfected cells will contribute to the differences in time course. The Hill coefficient for caffeine-induced Ca 2ϩ release in recombinant RyR2 was close to 1, implying that a single binding site exists for caffeine, while the value in recombinant RyR1 was greater than unity, indicating a cooperative action. The caffeine binding site has been proposed to be located in the first 4000 amino acids (37), so the precise site of molecular interaction of caffeine with RyR is still not known. These differences in caffeine activation might be exploited to pinpoint the site of caffeine activation in future studies of chimeric RyR molecules.