To clarify whether activity of the ryanodine receptor type 2 (RyR2) is reduced in the sarcoplasmic reticulum (SR) of cardiac muscle, as is the case with the ryanodine receptor type 1 (RyR1), Ca2+-dependent [3H]ryanodine binding, a biochemical measure of Ca2+-induced Ca2+ release (CICR), was determined using SR vesicle fractions isolated from rabbit and rat cardiac muscles. In the absence of an adenine nucleotide or caffeine, the rat SR showed a complicated Ca2+ dependence, instead of the well-documented biphasic dependence of the rabbit SR. In the rat SR, [3H]ryanodine binding initially increased as [Ca2+] increased, with a plateau in the range of 10–100 μM Ca2+, and thereafter further increased to an apparent peak around 1 mM Ca2+, followed by a decrease. In the presence of these modulators, this complicated dependence prevailed, irrespective of the source. Addition of 0.3–1 mM Mg2+ unexpectedly increased the binding two- to threefold and enhanced the affinity for [3H]ryanodine at 10–100 μM Ca2+, resulting in the well-known biphasic dependence. In other words, the partial suppression of RyR2 is relieved by Mg2+. Ca2+ could be a substitute for Mg2+. Mg2+ also amplifies the responses of RyR2 to inhibitory and stimulatory modulators. This stimulating effect of Mg2+ on RyR2 is entirely new, and is referred to as the third effect, in addition to the well-known dual inhibitory effects. This effect is critical to describe the role of RyR2 in excitation-contraction coupling of cardiac muscle, in view of the intracellular Mg2+ concentration.
- [3H]ryanodine binding
- stimulation by physiological Mg2+, excitation-contraction coupling in the heart
the ryanodine receptor (RyR) is a Ca2+-release channel in the sarcoplasmic reticulum (SR). It is composed primarily of a homotetramer of monomeric polypeptides of about 5,000 amino acid residues, with a MW of about 560 kDa (5, 17). There are three genetically distinct isoforms: RyR1–3. Whereas RyR1 and RyR3 occur mainly in skeletal muscles, RyR2 is expressed primarily in cardiac muscle (27, 35, 36). All three isoforms can be activated by Ca2+, giving rise to Ca2+-induced Ca2+ release (CICR) (3, 6, 7). Only RyR1 can also be activated through a conformational change of the dihydropyridine receptor, the voltage sensor, upon depolarization of the T-tubule membrane in skeletal muscle (33).
CICR plays a pivotal role in excitation-contraction coupling in cardiac muscle and is modulated by many endogenous and exogenous agents, including Ca2+, Mg2+, adenine nucleotides, SH-modifying agents, pH, calmodulin (CaM), and FK-506-binding protein (FKBP) 12/12.6, as well as many pharmacological reagents, including caffeine, procaine and halothane, and RyR domain peptides, which may affect interdomain interactions (8, 10, 18, 34, 37, 38). Ca2+ and Mg2+ are of primary importance among these factors. Ca2+ stimulates CICR at low concentrations (μM order), whereas it inhibits CICR at 0.1 mM or higher, showing biphasic Ca2+ dependence. This dependence is explained by the integration of the two independent effects on the activating (A-site) and inactivating (I-site) Ca2+ sites in the RyR molecules (13, 19, 21). In cardiac muscle, however, inhibition by a high Ca2+ concentration was not noticeable up to 0.1 mM, in contrast to the marked inhibition with RyR1 in skeletal muscle (5, 8, 17, 34, 37, 38). Therefore, the affinity for the divalent cation of the I-site of RyR2 was believed to be very low or negligible (5, 8, 17, 34, 37, 38). Mg2+ exerts an effect antagonistic to Ca2+ on the A-site and synergistic with Ca2+ on the I-site, inhibiting CICR (13, 19, 21). Therefore, Mg2+ shifts the pCa-CICR activity relationship to a higher Ca2+ concentration range (the first effect) and reduces the peak value (the second effect) (5, 6, 7, 17, 27, 35, 36, 37). An adenine nucleotide such as ATP or β,γ-methylene adenosine triphosphate (AMPPCP), a nonhydrolyzable ATP analog, stimulates CICR without a change in the Ca2+ dependence (21). The stimulation is as great as 10,000-fold (6, 27) in skeletal muscle, whereas it is not as marked (at most several-fold) in cardiac muscle (8, 18, 34, 37). For determination of [3H]ryanodine binding, a biochemical measure of CICR, in a medium simulating the cytoplasmic milieu, the addition of an adenine nucleotide is indispensable with skeletal muscle SR, whereas it is not always required with cardiac SR (5, 17, 27, 34, 35, 36). Recently, RyR1 in the SR was found to be suppressed selectively in Ca2+-dependent [3H]ryanodine binding (to about one-seventh), compared with RyR3 over the entire range of Ca2+ concentrations, without a change in its Ca2+ sensitivity, although purified RyR1 is very similar in activity to purified RyR3 (24, 30). A reduction in the affinity for the ligand, [3H]ryanodine, accompanied this suppression (24). The interdomain interaction between region 1 (the NH2-terminal domain) and region 2 (the central domain) of RyR1 was found to be the main mechanism underlying the suppression (22). This interdomain interaction has also attracted the keen interest of investigators in view of its probable role in malignant hyperthermia (10).
It remains to be clarified whether suppression occurs with RyR2, whose CICR plays a central role in excitation-contraction coupling in cardiac muscle (3). This issue is also critical with respect to elucidation of the etiology of sudden death ascribed to the mutated RyR2, in which the mutation takes place in regions comparable to the sites associated with malignant hyperthermia (10). During the course of our investigation, we found that Mg2+ exerted a third effect of stimulating [3H]ryanodine binding in an intermediate range of Ca2+, roughly 10 to 100 μM, in addition to the antagonistic effect on the A-site (the first effect) and the synergistic effect on the I-site (the second effect) (13, 19, 21). In other words, Mg2+ relieves the partial suppression that becomes evident in the state of partial activation, usually in a limited range of Ca2+. The extent of this suppression was found to differ between rabbit and rat ventricular muscles.
MATERIALS AND METHODS
Rabbit RyR2 cDNA was described previously (26). The expression vectors pcDNA5 FRT TO and pOG44, Flp-In T-REx-293 cells, Lipofectamine 2000 and the other reagents required for cell culture were purchased from Invitrogen (Carlsbad, CA). All other chemicals used were of reagent grade.
Cell culture and generation of an RyR2-inducible stable transfectant.
Flp-In T-REx-293 cells were cultured in a high-glucose Dulbecco's modified Eagle's medium containing 10% fetal bovine serum at 37°C in a humidified CO2 incubator. The day before transfection, 5 × 105 cells were seeded onto a 12-well cell culture plate. The cells were cotransfected with two expression vectors, pOG44 and pcDNA5 FRT TO/RyR2, in a 3:1 ratio by lipofection. After 48 h, the cells were replated onto a 100-mm tissue culture dish and exposed to a selective medium containing 100 μg/ml hygromycin B. Four independent clones randomly selected among 14 surviving clones were cultured until a sufficient number of cells were grown. For expression, we usually used a 60–80% confluent cell layer, induced the protein with 1 μg/ml tetracycline, and harvested the cells after 18–20 h. Expression of RyR2 was detected in all clones by Western blot analysis.
Preparation of the SR vesicle fraction.
All protocols in this study were approved by the Institutional Animal Care and Use Committee at the Juntendo University School of Medicine and were in accordance with the Guide for Care and Use of Laboratory Animals in the Institute for Laboratory Animal Research. The SR vesicle fraction was prepared as described previously, with some modifications (4). Briefly, animals (rats or rabbits) were anesthetized by intraperitoneal injection (50 mg/kg) of pentobarbital (50 mg/ml, Dainippon Seiyaku, Tokyo, Japan). Their hearts were quickly removed and perfused with 0.9% NaCl to remove the blood. In some cases, the rabbit back muscles were also used to prepare skeletal muscle SR vesicles. Both left and right ventricular muscles of animal hearts were chopped into pieces and homogenized in three volumes of distilled water containing 6.7 mM NaOH and 1 μg/ml protease inhibitor cocktail (P8340; Sigma, St. Louis, MO) in a blender-type homogenizer (AM-3; Nihonseiki Seisakusho, Tokyo, Japan), using three 15-s bursts at the maximum speed. The homogenate was centrifuged at a maximum force of 9,000 g for 20 min. The resulting supernatant was filtered with a Whatman no. 41 filter paper (Whatman, Brentford, UK), and the pellets were obtained by centrifugation at a maximum force of 27,000 g for 70 min. The pellets were resuspended in a buffer containing 50 mM KCl, 10 mM 3-(N-morpholino)-2-hydroxypropanesulfonic acid (MOPSO)-KOH, pH 6.8 and 1 μg/ml protease inhibitor cocktail and repelleted. The microsomes were resuspended in the buffer supplemented with 0.18 M sucrose, quick-frozen in liquid nitrogen, and stored at −80°C until use.
Preparation of the crude endoplasmic reticulum (ER) fraction from HEK-293 cells expressing RyR2.
After being washed once with PBS containing 0.1 mM CaCl2, cells were detached from culture dishes by pipetting and collected by centrifugation at 1,400 rpm for 5 min in a TOMY (Tokyo) TS-7 rotor. Cell pellets were resuspended in PBS containing 0.1 mM CaCl2, 0.5 mM diisopropyl fluorophosphate and 1/200 vol of protease inhibitor cocktail (P8340; Sigma) and homogenized by sonication. The homogenate was centrifuged at 1,400 rpm for 5 min (TS-7 rotor, TOMY), the supernatant was placed carefully onto a solution containing an equal volume of 0.5 M sucrose plus buffer A [50 mM KCl, 20 mM 3-(N-morpholino)propanesulfonic acid (MOPS)-KOH (pH 7.0) and 1/200 vol protease inhibitor cocktail], and recentrifuged at 25,000 rpm for 20 min in a 55T rotor in a Hitachi ultracentrifuge (Hitachi-Seisakusho, Ibaraki, Japan). The pellet was resuspended in 0.3 M sucrose plus buffer A, rapidly frozen in liquid nitrogen and then stored at −80°C until use. Typically, ∼10 mg of protein of the crude ER fraction was obtained from 18 culture dishes (100 mm in diameter).
[3H]Ryanodine binding was determined according to the method described by Murayama et al. (21). Briefly, the reaction was started at 25°C by the addition of [3H]ryanodine (specific activity 56 Ci/mmol; Perkin Elmer, Boston, MA) to a final concentration of 4.5 or 8.95 nM in an incubation mixture (total volume of 200 μl) containing 170 mM KCl, 20 mM MOPS-KOH (pH 7.2), CaCl2 [sufficient to get a free Ca2+ concentration of less than 0.1 mM in the presence of 1 mM ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA)], SR vesicles (final concentration, 0.5 mg protein/ml), and other ingredients if necessary. Ca2+ concentrations of 0.1 mM and higher were made by the addition of CaCl2 alone, instead of the Ca2+-EGTA buffer. In experiments with Mg2+, MgCl2 was added to the medium in a manner such that the SR vesicles were exposed to Mg2+ prior to Ca2+, and KCl was reduced to keep the ionic strength constant (21). The reaction was terminated after an 8-h incubation at 25°C by the addition of 3 ml ice-cold distilled water. Bound [3H]ryanodine was collected by filtration with a 5% polyethyleneimine-treated Whatman GF/B glass fiber filter under vacuum, followed by two more washes with 3 ml ice-cold water. The radioactivity of the bound [3H]ryanodine activity trapped on the filter was determined in a liquid scintillation counter (LS 6500, Beckman, Fullerton, CA). Specific [3H]ryanodine binding was obtained after correction for nonspecific binding in the presence of 10 μM unlabeled ryanodine. Free Ca2+ and Mg2+ concentrations were calculated using binding constants for divalent cations of EGTA (9) and AMPPCP (28).
SDS-PAGE and Western blot analysis.
To detect RyR2, SDS-PAGE was carried out on a 2–12% polyacrylamide gradient slab gel using the discontinuous buffer system of Laemmli (12). Molecular mass markers used were myosin heavy chain (205 kDa), β-galactosidase (116 kDa), phosphorylase b (97.4 kDa), bovine serum albumin (66 kDa), ovalbumin (45 kDa), and carbonic anhydrase (29 kDa). The electrophoresed proteins were transferred onto a PVDF membrane, and the membrane was probed with affinity-purified antibodies (1:1,000 dilutions) against NH2-terminal or COOH-terminal RyR2 fragments (4). After washing, the membrane was incubated with secondary antibodies conjugated to horseradish peroxidase (1:3,000 dilutions, Bio-Rad, Hertfordshire, UK). Positive bands were detected by chemiluminescence (SuperSignal West Pico Substrate, Pierce, Rockford, IL) using Fujifilm RX-U (Tokyo). To detect CaM contaminating SR, SDS-PAGE and Western blot analysis were performed in a similar way and probed with a monoclonal antibody, IM7 (11), which was kindly provided by Professor Michio Yazawa, Hokkaido University.
Data are expressed as the means ± SE (number of determinations).
In the case of duplicate determinations, half the range of the variance was within the size of a symbol in most cases and was omitted from the plots for clarity. Values of P < 0.05 were considered significant by Student's t-test. Curve-fitting was made by using SigmaPlot software version 8.01 (Systat Software, Point Richmond, CA), where R2 values were 0.99–0.90 in most cases.
The effects of Ca2+ and Mg2+ on [3H]ryanodine binding were analyzed according to the method of Murayama et al. (21). Briefly, the biphasic Ca2+ dependence can be explained by Ca2+ binding to the A-site and the I-site of RyR, respectively. Mg2+ serves as a competitive antagonist at the A-site and an agonist at the I-site. Thus, [3H]ryanodine binding, B, can be expressed by the following equation: (1) where Bp = the value B under the condition of fA = 1 and fI = 0; KA,Ca, KA,Mg,, KI,Ca, and KI,Mg are dissociation constants for Ca2+ and Mg2+ of the A- and I-sites of RyR, respectively, and n is the Hill coefficient for the corresponding site. Ca and Mg represent the concentrations of free divalent cations in the medium. In Fig. 1A, the parameters for Ca2+ of the A- and I-sites can be evaluated with a curve fit, using SigmaPlot software for nonlinear regression. In Fig. 1B, the parameters for Mg2+ of the A- and I-site can be evaluated in the presence of 3.2 μM and 0.1 or 1 mM Ca2+, respectively, in similar ways as described above. In Table 1, the dissociation constants for divalent cations at the A- and I-sites are summarized in the format of average ± SE (number of determinations). With rat cardiac SR, the apparent dissociation constants for Ca2+ are summarized in the presence of 0.3 mM Mg2+, where a biphasic Ca2+ dependence was obtained, as shown in Fig. 2.
[3H]Ryanodine binding to SR vesicles from rabbit and rat ventricles.
Rabbit SR vesicles showed the well-known biphasic Ca2+ dependence of [3H]ryanodine binding, as shown in Fig. 1A. The curve was represented by the following parameters, which were obtained by nonlinear regression analysis: Bp = 0.35 pmol/mg protein, KA,Ca = 1.86 μM, nA,Ca = 1.62, KI,Ca = 2.29 mM, and nI,Ca = 1.67, where Bp represents the peak amount of [3H]ryanodine binding, KA,Ca and KI,Ca, the dissociation constants for Ca2+ at the A- and I-sites, respectively, and nA,Ca and nI,Ca, respectively, represent their Hill coefficients. The range of Ca2+ concentrations (10–300 μM) with which peak [3H]ryanodine binding was obtained was broader than that with skeletal muscle SR vesicles. Mg2+ monotonically decreased Ca2+-dependent [3H]ryanodine binding in the presence of 3.2 μM, 0.1 mM, or 1 mM Ca2+, which could be described by the following parameters: KA,Mg = 0.13 mM, nA,Mg = 0.88, KI,Mg = 2.18 mM, and nI,Mg = 0.86, which gave rise to the curves shown in Fig. 1B. Similar determinations with rabbit ventricular SR were repeated, and their averaged affinities for divalent cations of the A- and I-sites are summarized in Table 1. The results obtained with rabbit skeletal muscle SR are also presented for comparison.
In contrast, rat ventricular SR vesicles showed a peculiar Ca2+ dependence in [3H]ryanodine binding: it increased with increases in Ca2+ concentration up to around 10 μM, where a plateau was formed up to 100 μM. Thereafter, it increased further, and a higher peak was attained at about 1 mM Ca2+, followed by a decrease at higher Ca2+ concentrations (Fig. 1C). [3H]ryanodine binding in the presence of 3.2 μM and 1 mM Ca2+ was monotonically decreased by Mg2+, as shown in Fig. 1D. In the presence of 0.1 mM Ca2+, however, it surprisingly increased in the presence of low Mg2+ concentrations, and the decrease was not obvious until Mg2+ concentrations were 3 mM or higher. This Mg2+ dependence of rat SR is in marked contrast to rabbit SR, where the monotonic decrease was observed even in the presence of 0.1 mM Ca2+.
The third effect of Mg2+.
As shown in Fig. 1D, [3H]ryanodine binding to rat ventricular SR in the presence of 0.1 mM Ca2+ was enhanced by 0.3 mM Mg2+. To get a deeper understanding of the finding, we determined the pCa-[3H]ryanodine-binding relationship in the presence of 0.3 mM Mg2+ (Fig. 2). The relationship became biphasic, which was familiar to us, giving an apparent KA,Ca of 6.46 μM and KI,Ca of 1.66 mM, as shown in Fig. 2. In the presence of Ca2+ concentrations less than 10 μM, Mg2+ shifted the curve rightward, whereas in the presence of Ca2+ concentrations higher than 1 mM, the shift was leftward. With further increases in Mg2+ concentrations, the stimulating curve was shifted more to the right along with the Ca2+ concentrations up to 0.1 mM, and the decreasing curve was shifted more to the left with decreased peak values, although the shift was minor in magnitude (data not shown). These effects were consistent with the predictions of a competitive antagonist effect for Ca2+ at the A-site and a synergistic effect with Ca2+ at the I-site. In the presence of 0.1 mM Mg2+, however, the pCa-binding relationship remained multiphasic, as was the case without Mg2+, although the binding was slightly increased.
Surprisingly, Fig. 2 shows that Mg2+ enhanced [3H]ryanodine binding in the plateau range of 10–100 μM Ca2+. This enhancing effect of Mg2+ is entirely new, and is referred to as the third effect of Mg2+ (in addition to its dual inhibitory effects).
It should be explained that rabbit RyR2 expressed in HEK-293 cells showed a multiphasic Ca2+ dependence in the presence of Ca2+ alone (Fig. 3). The presence of Mg2+ caused biphasic Ca2+ dependence with enhancement of [3H]ryanodine binding in the range of 10–300 μM Ca2+ (Fig. 3A). The dose-dependent effect of Mg2+ in the presence of 0.1 mM Ca2+ was upwardly convex, as is the case with rat cardiac SR: an increase with increases in Mg2+ followed by a decrease (Fig. 3B, compare with Fig. 1D). Although they presented similar results, Meissner and his colleagues (2, 40) made no comment on this matter. Rabbit RyR2, therefore, intrinsically has the potential to show multiphasic Ca2+ dependence.
Behavior of ventricular SRs in the presence of adenine nucleotides and caffeine.
So far, rat and rabbit ventricular SRs appeared to differ with respect to whether they exhibit the third effect of Mg2+. In the next step, we determined the effects of adenine nucleotides and caffeine, which are well-known stimulators of CICR. Figure 4 shows the effect of AMP on rabbit ventricular SR. AMP has very low affinities for Ca2+ and Mg2+, unlike ATP, ADP, or AMPPCP, and it causes only minor changes in the concentrations of the free divalent cations (28). The presence of 3 mM AMP converted the biphasic Ca2+-dependent curve for rabbit cardiac SR into a multiphasic relationship that was similar to that for rat SR. Notably, the stimulatory effect was most marked around 1 mM Ca2+, but less around 10–100 μM Ca2+. Supplementation with 0.3 mM Mg2+ restored the biphasic relationship, with an enhanced peak value around 0.1 mM Ca2+. Ca2+ at 1 mM gave similar [3H]ryanodine binding in the absence and presence of Mg2+. This stimulation is also the case with other adenine nucleotides, e.g., ATP and AMPPCP (data not shown). Caffeine at 10 mM, similarly, caused multiphasic Ca2+ dependence in addition to its familiar effects [20-fold sensitization to a low Ca2+ concentration and increase in the peak binding (27, 29)] (Fig. 5A). Enhancement of [3H]ryanodine binding was marked at concentrations lower than 1 μM Ca2+ and around 1 mM Ca2+, as observed with AMP, and the plateau at the intermediate level became markedly broader in range because of the Ca2+-sensitization effect. The plateau level in the [3H]ryanodine binding to rabbit SR in the presence of caffeine gradually increased with the increase in Ca2+ concentrations in the range of pCa 6–4. Mg2+ restored the biphasic Ca2+ dependence, together with an enhanced and widely expanded peak level (Fig. 5A). The sensitization to a low Ca2+ concentration, the effect characteristic of caffeine, remained unchanged (Fig. 5A). The stimulatory effect of Mg2+ also was observed in a dose-dependent manner in the presence of caffeine, in contrast to the monotonic decrease by Mg2+ in its absence (Fig. 5B).
With rat ventricular SR, multiphasic Ca2+ dependence was also similar to that depicted in Figs. 4 and 5 in the presence of adenine nucleotides or caffeine (data not shown). The plateau level of [3H]ryanodine binding at the intermediate concentrations of 10–100 μM Ca2+ (or 1–100 μM Ca2+ in the presence of 10 mM caffeine), however, was negligibly increased by adenine nucleotides or caffeine in the absence of Mg2+, the peak around 1 mM Ca2+ being more conspicuous. The effect of caffeine on rat RyR2 would be misunderstood as Ca2+ sensitization alone if the survey was not made above 0.1 mM Ca2+(3, 5, 7, 8, 17; but see also 6, 29). The presence of Mg2+ also converted the multiphasic relationship to a biphasic one, and the stimulating effect of Mg2+ at intermediate Ca2+ concentrations was similar to those shown in Figs. 4 and 5.
Effects of CICR modulators in the presence of 0.3 mM Mg2+.
So far, the activity of RyR2 at 0.1 mM Ca2+ appeared as if it would freeze in the absence of Mg2+ and be relieved by 0.3 mM Mg2+. Figure 6 shows the dose-dependent effect of caffeine on [3H]ryanodine binding to rat ventricular SR in the presence of 0.1 mM Ca2+. In the absence of Mg2+, caffeine did not appreciably increase this binding. Addition of Mg2+ increased the binding up to twofold. Caffeine showed a dose-dependent enhancement in the presence of 0.3 mM Mg2+, and the maximum effect was attained at 3–10 mM of caffeine. A higher concentration of caffeine was somewhat inhibitory. The summary of this stimulatory effect of 10 mM caffeine at 0.1 mM Ca2+ with and without 0.3 mM Mg2+ is shown with the number of determinations in Fig. 7. Figure 7 also shows similar results with various other kinds of stimulators and inhibitors. They were examined at their optimum concentrations, which were found to be much lower with RyR2 compared with RyR1 of skeletal muscle. The amounts of bound [3H]ryanodine changed little in the absence of Mg2+, whereas they were greatly affected in the presence of Mg2+, as shown in Fig. 7 (see also Fig. 6). In other words, Mg2+ makes RyR2 more reactive to various agents. The enhancing effect of AMPPCP in the absence of Mg2+ was greater than that of caffeine, as is true with RyR1. In the presence of Mg2+, it was more remarkable. Whereas the stimulatory effect of 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), which was used to remove the selective suppression of RyR1, was maximum at 2% on mammalian skeletal muscles (22), the maximum effect was obtained at 0.1%-0.2% with rat ventricular SR. CHAPS at 1% was rather inhibitory to rat or rabbit RyR2 (data not shown). It should be pointed out that the stimulatory effect of CHAPS was also dependent on the presence of Mg2+. FK-506, which deprives RyR of FKBP 12/12.6, is reported to increase the activity of RyR. FK-506, however, did not affect [3H]ryanodine binding in the absence or presence of Mg2+, excluding the possibility of the involvement of FKBP 12/12.6 in the third effect of Mg2+ (Fig. 7). FK-506 treatment did not affect the responses of RyR2 to caffeine or AMPPCP, either (data not shown).
Procaine, which is a well-known inhibitor of CICR, decreased [3H]ryanodine binding in both the presence and absence of Mg2+. The magnitude of reduction by 1 mM procaine was more marked in the presence of Mg2+ than that in its absence (Fig. 7). The difference in [3H]ryanodine binding with and without Mg2+, however, decreased with the increase in its concentration. Procaine at 10 mM reduced [3H]ryanodine binding so that it was negligible, irrespective of Mg2+ concentration. Interesting results were obtained with CaM. In the crude SR vesicles used in these experiments, CaM was detected on Western blot analysis after SDS-PAGE. Exogenously added CaM nevertheless reduced [3H]ryanodine binding to one-third in the presence of 0.3 mM Mg2+, reaching the maximum at 0.2 μM CaM. In the absence of Mg2+, in contrast, a statistically significant increase in [3H]ryanodine binding, albeit as little as 0.06 pmol/mg protein, was observed. The difference might be practically negligible. Meissner and colleagues (2, 40) reported that the inhibitory effect of CaM could be observed exclusively in the channel activity in the lipid bilayer membrane charged with a voltage difference, but it was not detected by [3H]ryanodine binding. These observations indicate that the interaction between CaM and RyR would be dependent on their conformations, as reported by Rodney et al. (32).
In conclusion, the presence of 0.3 mM Mg2+ greatly amplified the responses of RyR2, not only to stimulatory reagents, but also to inhibitory ones, with the exception of FK-506.
RyR2 is in a more activated state in the presence of Mg2+.
In the standard assay medium containing 0.17 M KCl and 0.1 mM Ca2+, [3H]ryanodine binding to rat ventricular SR was determined in the presence and absence of 0.3 mM Mg2+ at varied concentrations of the ligand (Fig. 8). As shown in Fig. 8, the results can be explained by the following parameters: Bmax = 2.09 pmol/mg protein, Kd = 33.3 nM in the absence of Mg2+, and Bmax = 1.94 pmol/mg protein, Kd = 16.0 nM in the presence of 0.3 mM Mg2+. It should be explained that Kd would be 1–3 nM if determined in a medium containing 1 M KCl as reported (8, 18, 34, 37, 38). These results suggest that the single class of binding sites is homogeneous and that Mg2+ increased the affinity for the ligand without a change in the maximum number of these binding sites (Bmax). The increased affinity of RyR2 for the ligand indicates that RyR2 is more activated in the presence of Mg2+. The results of determinations in the medium containing 0.17 M NaCl instead of KCl (Fig. 9) may help to deepen our understanding of the underlying mechanism of the third effect of Mg2+. As shown in Fig. 5A, rabbit ventricular SR showed a multiphasic Ca2+ dependence in the presence of 10 mM caffeine in the 0.17 M KCl medium. In the 0.17 M NaCl medium, however, the Ca2+ dependence was biphasic, with a broader plateau (1–1,000 μM Ca2+), even in the presence of 10 mM caffeine. The findings that Ca2+ dependence of [3H]ryanodine binding was changed by monovalent ions were already reported with RyR1 (20, 23) and with RyR2 (15). The results in Fig. 9 fit the following parameters: in the absence of caffeine or Mg2+ (open circles), Bp= 0.42 pmol/mg protein, KA,Ca = 3.89 μM, nA,Ca = 1.46, KI,Ca = 1.91 mM, and nI,Ca = 1.25, whereas in the presence of 10 mM caffeine without Mg2+ (open diamonds), Bp = 0.64 pmol/mg protein, KA,Ca = 0.22 μM, nA,Ca = 1.50, KI,Ca = 6.31 mM, and nI,Ca = 1.69. The effects of caffeine in 0.17 M NaCl medium would be described as follows: Ca2+ sensitization as great as 18-fold, increase in Bp, i.e., increase in CICR activity, and decrease in the affinity of the I-site for Ca2+ with the increase in the Hill coefficient. These effects of caffeine are consistent with the results already reported (1, 6, 18, 27, 29). The supplement of 0.3 mM Mg2+ (closed diamonds) further increased Bp (0.92 pmol/mg protein) and restored the affinity of the I-site for Ca2+(KA,Ca = 0.40 μM, nA,Ca = 1.63, KI,Ca = 1.78 mM, nI,Ca = 0.93). The increase in Bp was due to the increase in the affinity for the ligand, as shown in Fig. 8. These findings indicate that RyR2 may be susceptible to the allosteric change in the conformation and that coexistence of 0.3–1 mM Mg2+ is necessary for RyR2 to function properly.
Characteristics of RyR2 as the CICR channel in cardiac muscle.
Because CICR plays a pivotal role in excitation-contraction coupling in cardiac muscle, it should be helpful to discuss the characteristics of RyR2 that were uncovered by this study, focusing on the dual inhibitory effect of Mg2+ on RyR2. It is widely expected that RyR2 should be more sensitive to Ca2+ than RyR1. Many comparisons between cardiac and skeletal muscle SRs, however, were performed with preparations of different animals, and very few reports to date compare RyR1 and RyR2 from the same animal (3, 5, 7, 8, 17, 34, 37). Table 1 summarizes the results of rabbit RyR1 and RyR2. KA,Ca of RyR2 was greater than that of RyR1 (P = 0.02), in contrast to what was anticipated. In addition to these findings, the effects of Mg2+ also must be considered with regard to physiological relevance. The common characteristics between the two isoforms are that the A-sites show much greater preference for Ca2+ than for Mg2+, whereas the I-sites show no significant difference in affinity between the two divalent cations. RyR1 and RyR2, however, are distinctly affected by Mg2+. KA,Mg of RyR2 was about 1/4 of that of RyR1. Furthermore, the I-site of RyR2 exhibits very low affinities for divalent cations with Kd values of about 2–3 mM, whereas the counterpart of RyR1 has very high affinities of 0.15–0.2 mM. Mg2+ is more effective on RyR2 than on RyR1 as a competitive inhibitor of Ca2+ at the A-site (P = 0.012), whereas it is more potent at the I-site of RyR1 than that of RyR2. This means that RyR2 is largely reduced in Ca2+ sensitivity in the presence of Mg2+, whereas in RyR1, it is the peak value that is largely reduced. Similar findings were reported by Pessah et al. (31). These suggest that the cardiac RyR2 would be less sensitive to Ca2+ than the skeletal muscle RyR1.
Table 1 also summarizes the results of multiple similar determinations with rat ventricular SR in the presence of 0.3 mM Mg2+. Because Fig. 1, B and D, showed similar effects of Mg2+ on [3H]ryanodine binding in the presence of 3.2 μM and 1 mM Ca2+, the inhibitory effects of Mg2+ on rat ventricular SR may be similar to those for rabbit ventricular SR. The results shown in Table 1 indeed indicate that affinities of the A-site and the I-site for Ca2+ and Mg2+ would be similar between rabbit and rat ventricular SRs. Table 1 also indicates that the properties of frog skeletal muscles may differ from those of rabbit skeletal muscles.
The third, stimulatory effect of Mg2+ on RyR2.
Rabbit RyR2 in the ventricular SR shows biphasic Ca2+ dependence in the presence of Ca2+ alone. The effect of Mg2+ on rabbit RyR2 is due to competitive antagonism to Ca2+ at the A-site and synergistic action with Ca2+ at the I-site (Fig. 1, A and B, and Table 1) (1, 13, 19, 21). Rat RyR2 in the ventricular SR, in contrast, showed a peculiar multiphasic Ca2+ dependence, with a plateau of an intermediate level of bound [3H]ryanodine in the range of 10–100 μM Ca2+ (Fig. 1, C and D). The presence of Mg2+ at 0.3 mM or higher converted the complexed Ca2+ dependence to a biphasic relationship, with enhanced [3H]ryanodine binding in the intermediate range of Ca2+ (Fig. 2). This biphasic Ca2+ dependence in the presence of Mg2+ can be explained by affinities for Ca2+ and Mg2+ of the A-site and I-site, which are similar to those of rabbit ventricular SR (Table 1). In the presence of an adenine nucleotide or caffeine, the multiphasic Ca2+ dependence with the intermediate level of the plateau was observed, irrespective of rabbit or rat RyR2. Mg2+ was found to exhibit the third effect on RyR2, i.e., enhanced binding in the intermediate Ca2+ concentrations in addition to the actions on the A- and I-sites (Figs. 2–5). The third, stimulatory effect of Mg2+ on RyR2 is dose-dependent, and its minimum concentration was about 0.3 mM. This prevents us from observing the third effect, if any, with RyR1 or RyR3, which has a KI,Mg of around 0.2 mM (Table 1). Only RyR2, which has a large KI,Mg of 2–3 mM, can be partly immune to the inhibitory effect of the I-site (Table 1). Furthermore, the third effect of Mg2+ is likely to be independent of the dual inhibitory effects and additive to them.
In the SR membrane of skeletal muscles, on the other hand, the activity of RyR1 is selectively suppressed to about one-seventh over the whole range of Ca2+ concentrations compared with RyR3, as evidenced by a reduced affinity for [3H]ryanodine to a comparable extent without a change in Ca2+ dependence (22, 24, 30). This means that the attenuating coefficient (1/7–1/8) for RyR1 is constant over the whole range of Ca2+ concentrations, without a change in the affinities for the divalent cations of the A- or I-site. The interdomain interaction between region 1 and region 2 in RyR1 is suggested to be involved primarily in the selective suppression (22). RyR2, in contrast to RyR1, appears to freeze to the state of an intermediate activity in the range of 10–100 μM Ca2+. In this sense, RyR2 is subjected to the partial suppression, which is removed by Mg2+. Because CHAPS stimulated RyR2 (Fig. 7), an interdomain interaction such as within RyR1 could be considered, although it is not identical.
Purified RyR2 showed monophasic Ca2+ dependence without inhibition by high Ca2+, independent of animal sources. Similar monophasic Ca2+ dependence was observed in 1 M KCl-containing medium with cardiac SR vesicles (8, 18, 34, 37, 38). These results might strongly suggest the involvement of some accessory proteins in multiphasic Ca2+ dependence. In these cases, however, the activation by Ca2+ was very steep, nearly of an all-or-none type. Because the Hill coefficient for the activating Ca2+ was near 2 with RyR2 in the SR, the allosteric conformation may be changed after purification. Yamaguchi et al. (40) reported that inhibition by CaM can be detected by lipid bilayer experiments only under a charged voltage difference, but not by [3H]ryanodine binding. The effect of CaM may be dependent on the conformation of RyR molecules, as reported by Rodney et al. (32). Addition of exogenous CaM to crude SR intriguingly reduced [3H]ryanodine binding to one-third in the presence of 0.3 mM Mg2+, but caused no inhibition in the absence of Mg2+ (Fig. 7). It could be hypothesized that added Mg2+ would bind to CaM to remove the inhibitory effect. The stimulatory effect of Mg2+ on RyR2, however, was observed in a limited range of Ca2+ concentrations, whereas the inhibitory effect of CaM on RyR2 was homogeneous over the whole range of Ca2+ concentrations (2, 18, 40). CaM (17), therefore, is excluded from the list of candidates for the responsible cofactor. Treatment of the SR with FK-506 did not change [3H]ryanodine binding in the presence or absence of Mg2+(Fig. 7). Washing of rat SR with 0.17 M KCl medium containing 1 mM EGTA, which is effective to deprive sorcin of RyR2 (16), did not change the multiphasic Ca2+ dependence (data not shown). These findings suggest that either FKBP (38) or sorcin (16) would play a minor role in the third effect of Mg2+. The possibility of other unidentified proteins cannot be excluded, and further investigations are required to clarify the underlying mechanism of the third effect of Mg2+.
The third effect of Mg2+ was not detectable at 0.1 mM, but was clearly observed at 0.3 mM or more. Correspondingly, the second increase of [3H]ryanodine binding in multiphasic Ca2+ dependence in the absence of Mg2+, was observed at 0.3 mM Ca2+ and higher, with the apparent peak around 1 mM Ca2+ (Figs. 1C and 2–5). These findings strongly suggest that Ca2+ and Mg2+ are of equal importance in the third effect.
The regulation will be dependent on the extent of activation of RyR2 molecules, but not on the Ca2+ concentration itself, because the plateau of [3H]ryanodine binding started at a lower Ca2+ concentration in the presence of caffeine with the unchanged plateau level. The results in Figs. 4 and 5A show that the intermediate plateau with rabbit ventricular SR was gradually increased as Ca2+ increased up to 100 μM in the absence of Mg2+, whereas it was not appreciably changed with rat ventricular SR. This gradual increase in the plateau level, with the increase in Ca2+, was also observed with rabbit RyR2 expressed in HEK-293 cells (Fig. 3). These results may suggest that rabbit RyR2 is more vulnerable to the third effect than rat RyR2 or that rabbit RyR2 may be looser in its regulation than rat RyR2. More comparative studies are required, because it is well known that species specificity is great even among mammalian hearts.
Physiological relevance of the current findings.
It has so far been claimed that no inhibition of CICR at high Ca2+ concentrations is characteristic of RyR2 (3, 5, 8, 17, 34, 37). We showed clearly the effectiveness of the I-site, although RyR2 shows a lower affinity than RyR1 and RyR3 (Table 1). One explanation for this discrepancy is the difference in experimental conditions: a high salt-medium, 1 M KCl, was used instead of 0.17 M KCl, and/or purified RyR2 was used for the experiments (8, 14, 15, 18, 34, 37). Another explanation is the limited range of Ca2+ concentrations of up to pCa 4 (14, 15). Another intriguing determination was made by Xu et al. (39) in which 5 mM Mg-AMPPCP was used. In this case, the free Mg2+ concentration inevitably amounts to about 0.8 mM, which certainly would exert the third effect.
Mg2+ makes RyR2 sensitive to well-known modulators (Fig. 7); therefore, the presence of Mg2+ is inevitable with experiments on RyR2. In the absence of Mg2+, the partial suppression of RyR2 may be obvious. In the presence of Mg2+, RyR2 is very similar among mammalian hearts by removal of the suppression. The physiological intracellular concentration of Mg2+ in cardiac muscle is assumed to be 0.5–1 mM (3, 20, 25), which is sufficient for the stimulatory effect of Mg2+ on RyR2. Mg2+, on the other hand, will move the pCa-activity curve about 0.5–0.7 pCa unit to a higher Ca2+ concentration range. The several-fold enhancement of the activity will compensate for this handicap in the presence of Mg2+ and ATP. The presence of Mg2+ at the physiological intracellular concentration is essential for RyR2 to play a key role in excitation-contraction coupling of cardiac muscle.
We thank Drs. N. Kurebayashi, T. Murayama, and T. Kashiyama of Juntendo University for suggestions and encouragement.
Present address of O. Sato: Dept. of Physiology, University of Massachusetts Medical School, 55 Lake Ave. North, Worcester, MA 01655.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society