Cytosolic Ca²⁺-dependent Ca²⁺ release activity primarily determines the ER Ca²⁺ level in cells expressing the CPVT-linked mutant RYR2

The type 2 ryanodine receptor (RYR2) is a Ca²⁺ release channel in the SR/ER that has an indispensable role in excitation–contraction coupling in the heart. In cardiac myocytes, RYR2 is activated by Ca²⁺ influx through L-type Ca²⁺ channels during action potential with a Ca²⁺-induced Ca²⁺ release (CICR) mechanism to release more Ca²⁺, which causes muscle contraction (Fabiato, 1983; Bers, 2002). In contrast, spontaneous Ca²⁺ release, such as Ca²⁺ waves, which underlies delayed afterdepolarization of the plasma membrane via a Na⁺–Ca²⁺ exchange reaction, often results in triggered activity (Tsien et al., 1979; Lakatta, 1992). Spontaneous Ca²⁺ release can be seen when cardiomyocytes are Ca²⁺-overloaded, even in healthy hearts, but is more likely to occur in the myocardium of heart failure patients or carriers of RYR2 mutations associated with sudden cardiac death syndrome.

RYR2 mutations have been linked to several types of arrhythmogenic diseases, such as catecholaminergic polymorphic ventricular tachycardia (CPVT), left ventricular noncompaction, and idiopathic ventricular fibrillation (Priori et al., 2001; Tester et al., 2004; Medeiros-Domingo et al., 2009; Priori and Chen, 2011; Kawamura et al., 2013; Fujii et al., 2017; Uehara et al., 2017; Nozaki et al., 2020; Hirose et al., 2022). Among the nearly 300 arrhythmogenic mutations reported to date, CPVT is the most common RYR2-related arrhythmogenic disorder and is induced in response to sympathetic nerve activation without structural abnormalities of the heart (Priori et al., 2001). CPVT-linked RYR2 mutations previously characterized were associated with gain-of-function phenotypes, which are prone to induce spontaneous Ca²⁺ release from the SR.

Two Ca²⁺-dependent regulatory mechanisms of RYR2—the regulation from the luminal side of the SR/ER (luminal control) and that from the cytoplasmic side (cytoplasmic control)—may be involved in the spontaneous Ca²⁺ release. Many reports have described that SR/ER luminal Ca²⁺ plays a role in controlling Ca²⁺ release through RYR2 in cardiac cells (Bassani et al., 1995; Lukyanenko et al., 1996; Sitsapesan and Williams, 1997). In addition, effects of ER luminal Ca²⁺ on RYR2 channels have been extensively investigated under artificial noncardiac conditions using a HEK293 expression system. Single-channel analyses have indicated that luminal Ca²⁺ ([Ca²⁺]_ER) activates RYR2 at ∼10⁻³ M in the presence of calsequestrin 2 and at 10⁻² M in its absence (Qin et al., 2008). In contrast, the role of cytosolic Ca²⁺ is known as the CICR mechanism (Endo, 1977; Fabiato, 1983; Murayama and Kurebayashi, 2011; Guo et al., 2012; Rios, 2018). The cytoplasmic Ca²⁺ ([Ca²⁺]_cyt) activates RYR2 at Ca²⁺ low concentrations, below ∼10⁻⁴ M, but suppresses it at higher Ca²⁺ concentrations (above ∼10⁻³ M), resulting in a bell-shaped [Ca²⁺]_cyt dependence (Murayama and Kurebayashi, 2011; Rios, 2018). Therefore, a small local Ca²⁺ leak from the ER, which is more likely to occur at high [Ca²⁺]_ER, may activate RYR2 from the cytoplasmic side to trigger massive Ca²⁺ release by the positive feedback nature of CICR.

A question of whether CPVT mutations primarily affect the cytoplasmic or luminal regulation of RYR2 is currently under debate. Chen and colleagues (Jiang et al., 2004; Jiang et al., 2005) indicated that the spontaneous Ca²⁺ release, called store-overload–induced Ca²⁺ release (SOICR), in RYR2-expressing HEK293 cells occurs when [Ca²⁺]_ER reaches a certain critical threshold level, and they found that the CPVT-linked RYR2 cells showed a lower threshold [Ca²⁺]_ER for spontaneous Ca²⁺ release compared with WT RYR2 (Jones et al., 2008). They reported that some of the CPVT-linked RYR2 channels (e.g., R2474S and R4497C) increased luminal Ca²⁺ sensitivity, with no significant effects on [Ca²⁺]_cyt dependence (Jiang et al., 2005), suggesting that the lower threshold [Ca²⁺]_ER with these CPVT mutations is due to sensitization to luminal Ca²⁺ but not to cytosolic Ca²⁺. In contrast, Wehrens et al. (2003) reported that the same mutation, R2474S, increased [Ca²⁺]_cyt sensitivity via the phosphorylation of RYR2 by protein kinase A due to the dissociation of FKBP12.6. We also found that R2474S showed higher [Ca²⁺]_cyt-dependent activity at resting [Ca²⁺]_cyt (Uehara et al., 2017). The reason for this discrepancy may be better understood through a more quantitative analysis.

The HEK293 cell expression system allows for both functional and biochemical analyses of exogenously expressed RYRs (Jiang et al., 2002; Jiang et al., 2004; Murayama et al., 2015). In HEK293 cells, regulatory proteins specific to the myocardium, such as calsequestrin and FKBP12.6, are absent; thus, the use of these cells enables evaluation of the direct effect of mutations on RYR2 activity. Previously we have quantitatively evaluated the Ca²⁺ release activity of RYR1 using the HEK293 expression system and showed that [Ca²⁺]_ER signals correlated well with the [Ca²⁺]_cyt-dependent [³H]ryanodine binding activity of RYR1 (Murayama et al., 2015; Murayama et al., 2016). In this study, we investigated the correlation between [Ca²⁺]_cyt-dependent activity and [Ca²⁺]_ER in HEK293 cells using 10 CPVT and 6 artificial RYR2 mutations. In addition, we further validated this correlation using a mathematical model. Our results suggest that the changes in threshold [Ca²⁺]_ER for spontaneous Ca²⁺ release by CPVT mutation are explained by the [Ca²⁺]_cyt-dependent activity of RYR2 without considering a change in the [Ca²⁺]_ER sensitivity.

Full-length mouse RYR2 cDNA was constructed from cDNA fragments that were PCR-amplified from mouse ventricles and then cloned into a tetracycline-induced expression vector (pcDNA5/FRT/TO; Life Technologies; Fujii et al., 2017; Uehara et al., 2017). Mutations corresponding to V2321M, R2474S, D3638A, Q4201R, K4932R, R4497C, K4751Q, H4762P, K4805R, and I4867M were introduced by inverse PCR and confirmed by DNA sequencing (Table 1). The expression vector was cotransfected with pOG44 into Flp-In T-REx HEK293 cells (Life Technologies), according to the manufacturer’s instructions. Clones with suitable doxycycline-induced expression of RYR2 expression were selected and used in the experiments.

HEK293 cells grown on a glass-bottomed dish were treated with doxycycline 26–28 h before measurement to induce the expression of RYR2, unless otherwise noted. Single-cell Ca²⁺ imaging was carried out in HEK293 cells expressing WT or mutant RYR2 in Krebs solution (140 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 11 mM glucose, and 5 mM HEPES, pH 7.4; Murayama et al., 2015; Uehara et al., 2017). All measurements were performed at 26°C by perfusing solutions using an in-line solution heater/cooler (Warner Instruments).

For cytosolic Ca²⁺ measurements, cells were loaded with 4 µM fluo-4 AM in culture medium for 30 min at 37°C and then incubated with Krebs solution. Fluo-4 was excited at 488 nm through a 20× objective lens, and emitting light at 525 nm was captured with an electron multiplying–charge-coupled device camera at 700-ms intervals (Model 8509; Hamamatsu Photonics). The fluo-4 fluorescence signal was measured in normal Krebs solution for 5 min and then in 10 mM caffeine containing Krebs solution for 1.5 min, and at the end of each experiment the maximal fluorescence intensity (F_max) of fluo-4 was determined with a 20Ca-ionomycin solution (140 mM NaCl, 5 mM KCl, 20 mM CaCl₂, 1 mM MgCl₂, 11 mM glucose, 20 µM ionomycin, and 5 mM HEPES, pH 7.4). The fluorescence signal (F) in individual cells in Krebs solution was determined using region of interest (ROI) analysis, and cell-free background fluorescence was subtracted from F and normalized by the value F_max. The [Ca²⁺]_cyt was calculated using the parameters of effective K_D = 2.2 μM and n = 1 in situ (Harkins et al., 1993; Nelson et al., 2014). Since WT and most of mutant RYR2 mutant-expressing cells showed Ca²⁺ oscillations, the peak and resting fluorescence and Ca²⁺ oscillation frequency in normal Krebs solution were determined. For oscillation analysis, a series of cytosolic Ca²⁺ increase and decrease with ΔF/F_max > 0.05 and a duration of 0.1–1 min was regarded as one Ca²⁺ oscillation. The oscillation frequency was determined by counting the number of oscillations during 5 min in normal Krebs solution in individual cells (n/min). Usually, during one measurement in a single cell, the peak amplitudes of the oscillations were very similar. To obtain the peak [Ca²⁺]_cyt signal in individual cells, the maximum F/F₀ during 5 min in Krebs solution was determined and subtracted from the fluctuation factor, 0.01 F/F₀. To obtain the resting [Ca²⁺]_cyt signal, the minimum F/F₀ during the 5 min was determined and added with a fluctuation factor of 0.005 F/F₀.

For [Ca²⁺]_ER measurements, [Ca²⁺]_ER and [Ca²⁺]_cyt signals were simultaneously monitored using genetically encoded Ca²⁺ indicators R-CEPIA1er (Suzuki et al., 2014; Murayama et al., 2015; Uehara et al., 2017) and G-GECO1.1 (Zhao et al., 2011), respectively. Cells were transfected with G-GECO1.1 and R-CEPIA1er cDNA 26–28 h before measurements. Doxycycline was added to the medium at the same time as transfection in most experiments, or after transfection in experiments using cells with reduced RYR2 expression. Cytosolic and ER Ca²⁺ signals were determined in normal Krebs solution for 5 min and then in 10 mM caffeine containing Krebs solution for 1.5 min. After caffeine treatment, the cells were perfused with the following solutions: 0Ca-Krebs solution (140 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 11 mM glucose, and 5 mM HEPES, pH 7.4), BAPTA-0Ca-ionomycin solution (140 mM NaCl, 5 mM KCl, 5 mM 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid [BAPTA], 1 mM MgCl₂, 11 mM glucose, 20 μM ionomycin, 20 μM cyclopiazonic acid, and 5 mM HEPES, pH 7.4), 0Ca-Krebs solution, and then 20Ca-ionomycin solution. The F_min and F_max values were determined using the BAPTA-0Ca-ionomycin solution and 20Ca-ionomycin solution, respectively. Threshold and nadir [Ca²⁺]_ER levels were determined using the same protocol used for peak and resting [Ca²⁺]_cyt determination. Because G-GECO1.1 signals have a high Hill coefficient (n = 3.38; Suzuki et al., 2014), which results in a difficult [Ca²⁺]_cyt calculation, only the R-CEPIA1er signal was used for the calculation of [Ca²⁺]_ER. [Ca²⁺]_ER was calculated using the parameters determined by in situ titration (K_D = 565 μM, n = 1.7; Suzuki et al., 2014).

Ca²⁺-dependent [³H]ryanodine binding was performed as previously described (Fujii et al., 2017; Nozaki et al., 2020). Briefly, microsomes isolated from HEK293 cells were incubated for 1 h at 25°C with 5 nM [³H]ryanodine in a medium containing 0.17 M NaCl, 20 mM 3-(N-morpholino)-2-hydroxypropanesulfonic acid, pH 7.0, 2 mM dithiothreitol, 1 mM AMP, 1 mM MgCl₂, and various concentrations of free Ca²⁺ buffered with 10 mM EGTA. Free Ca²⁺ concentrations were calculated using the WEBMAXC STANDARD (https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/webmaxc/webmaxcS.htm; Bers et al., 2010). The protein-bound [³H]ryanodine was separated by filtering through polyethyleneimine-treated GF/B filters using Micro 96 Cell Harvester (Skatron Instruments). Nonspecific binding was determined in the presence of 20 µM unlabeled ryanodine. The [³H]ryanodine binding data (B) were normalized to the maximum number of functional channels (B_max), which was separately determined by Scatchard plot analysis using various concentrations (3–20 nM) of [³H]ryanodine in a high-salt medium containing 1 M NaCl. The resultant B/B_max represents the average activity of each mutant.

For estimation of ryanodine binding at resting Ca²⁺, A at pCa 7.0 (A_rest) for each mutant was calculated by Eqs. 1, 2, and 3 using the determined parameters (K_ACa, K_ICa, and A_max; Table 2), in which free [Ca²⁺] was set at 100 nM.

Microsomal proteins were separated by SDS-PAGE and transferred onto a polyvinylidene difluoride membrane. Western blotting was performed using antibodies against RYR2 (Chugun et al., 2007) and calnexin (C4731; Sigma-Aldrich).

First, we performed simulations of [Ca²⁺]_cyt-dependent [³H]ryanodine binding with and without luminal regulation, assuming that [Ca²⁺]_ER was the same as [Ca²⁺]_cyt, corresponding to the experimental conditions. We confirmed that the addition of E and K_ER,Ca had little effect on the fit curve of the [³H]ryanodine binding assay. Subsequently, we verified the manner in which RYR2 activity would change if the luminal Ca²⁺ concentration was maintained at high levels.

Data are presented as the mean ± SD. Statistical comparisons were performed using Prism 9 (GraphPad Software). One-way ANOVA (Figs. 1, 2, 3, 4, and 5) or two-way ANOVA (Fig. 7 D) with multiple comparisons was used to compare multiple groups. Statistical significance was determined as P < 0.05.

Fig. S1 shows calculated [Ca²⁺]_cyt (in μM) converted from fluo-4 [Ca²⁺]_cyt signals (as F/F_max) in Fig. 2 (A and B) and average [Ca²⁺]_cyt signals of fluo-4, which are the basis of Fig. 2 (C and D). Fig. S2 shows calculated [Ca²⁺]_ER (in mM) converted from R-CEPIA1er [Ca²⁺]_ER signals in Fig. 3 A, average [Ca²⁺]_ER signals, which are the basis of Fig. 3 (B and C) and Fig. 5 (D and E), and correlation between nadir [Ca²⁺]_ER and threshold [Ca²⁺]_ER. Fig. S3 shows effects of mutations at the putative ER Ca²⁺ sensing site, E4872, on Ca²⁺ signaling and [³H]ryanodine binding of RYR2.

Stable tetracycline-inducible HEK293 cell lines expressing WT and mutant RYR2s (10 CPVT and 6 artificial mutants) were generated. The CPVT mutations are in the N-terminal helical domain (V2321M, R2474S), central domain (D3638A, Q4201R), divergent region 1 (K4392R), and transmembrane domain (R4497C, K4751Q, H4762P, K4805R, and I4867M; Fig. 1 A and Table 1). Furthermore, artificial mutations were made by single alanine substitutions at the S4–S5 linker region, in which alanine substitutions differentially affect the [Ca²⁺]_cyt-dependent Ca²⁺ release activity depending on the position of the α helix in RYR1 (Murayama et al., 2011). Fig. 1 B shows representative Western blots of WT and mutant RYR2s 26 h after induction. There was no significant difference in average expression levels between WT and mutant RYR2 proteins, except for K4751Q, K4805R, and L4757A (Fig. 1 C), the expression levels of which were lower than those of the other mutants. Low expression levels may correlate with high Ca²⁺ release activity, as described later.

Single-cell Ca²⁺ imaging in HEK293 cells was performed 26–28 h after induction with doxycycline. Fig. 2 A shows representative cytoplasmic Ca²⁺ measurements determined using fluo-4. HEK293 cells expressing WT-RYR2 showed spontaneous Ca²⁺ oscillations, as reported previously (Fig. 2 A, left; Jiang et al., 2004; Fujii et al., 2017; Uehara et al., 2017). The application of 10 mM caffeine induced a transient Ca²⁺ release with an amplitude similar to that of spontaneous oscillations. Cells expressing R4497C- and R2474S-RYR2 showed Ca²⁺ oscillations with higher frequency and smaller amplitude compared with those of WT cells (Fig. 2 A [middle two], B, and C). Most of the H4762P cells showed no evident oscillations, and the application of caffeine caused only a small Ca²⁺ transient, as shown in Fig. 2 A, right. The oscillation frequencies, including no oscillation (0/min), in individual WT and mutant RYR2 cells are plotted in Fig. 2 B, with the median (red line). More than 70% of WT and CPVT mutant cells showed Ca²⁺ oscillations, with the exceptions of H4762P and K4805R. The average oscillation frequency was significantly higher in the CPVT mutants than in the WT, except for K4392R, H4762P, and K4805R (Fig. 2 B). In H4762P and K4805R cells, the median of the frequency was 0, although 10% of cells exhibited Ca²⁺ oscillations with higher frequency than those in WT cells (Fig. 2 B). The oscillation frequency in K4392R was not significantly different from that of WT. The average peak Ca²⁺ signals and calculated Ca²⁺ concentrations ([Ca²⁺]_cyt) are shown in Figs. S1 B and 2 C, respectively. Peak [Ca²⁺]_cyt was significantly smaller in CPVT than in WT, with the exception of K4392R. The average resting Ca²⁺ signals and calculated [Ca²⁺]_cyt were significantly higher in K4751Q, H4762P, and K4805R cells (Figs. S1 B and 2 D). A significant negative correlation (P < 0.0001) was observed between the average oscillation frequency and peak amplitude (Fig. 2 E); that is, a higher frequency was associated with a smaller Ca²⁺ transient.

Next, we measured [Ca²⁺]_ER levels using R-CEPIA1er, a genetically encoded ER Ca²⁺ sensor protein (Suzuki et al., 2014; Murayama et al., 2015; Murayama et al., 2016; Fujii et al., 2017; Uehara et al., 2017). Fig. 3 A shows representative [Ca²⁺]_cyt and [Ca²⁺]_ER signals from G-GECO1.1 and R-CEPIA1er, respectively, and Fig. S2 A shows the corresponding calculated [Ca²⁺]_ER. HEK293 cells expressing WT RYR2 (Fig. 3 A, left) showed a periodic decrease in R-CEPIA1er signals in normal Krebs solution, which reflected Ca²⁺ release from the ER. Before each Ca²⁺ release, the R-CEPIA1er signal reached a maximal level (threshold), rapidly decreased to reach a minimal level (nadir), and then gradually increased again toward the threshold level. Both the threshold and nadir levels were reduced in R4497C and R2474S to varying degrees, while H4762P showed no oscillation with a further decrease in [Ca²⁺]_ER (Fig. 3 A). The average ER Ca²⁺ signals and calculated [Ca²⁺]_ER are shown in Fig. S2 B and Fig. 3, B and C, respectively. All CPVT cells showed significantly decreased threshold and nadir [Ca²⁺]_ER, except for K4392R. A good correlation was observed between the nadir and the threshold [Ca²⁺]_ER (Fig. S2 D).

The properties of cytosolic Ca²⁺-dependent channel activity were evaluated with the [³H]ryanodine binding assay (Fig. 4). Both WT and mutant RYR2 channels exhibited bell-shaped biphasic Ca²⁺ dependence (Fig. 4 A). Most CPVT mutations except K4392R showed higher [³H]ryanodine binding than WT, especially at 10⁻⁵ M or lower Ca²⁺ concentrations (Fig. 4 A), suggesting higher channel activity of CPVT mutants at physiological [Ca²⁺]_cyt.

We have previously shown that [Ca²⁺]_cyt-dependent Ca²⁺ release activity of RYR2 can be described by Eqs. 1, 2, and 3 with three parameters: the gain (A_max) and dissociation constants for activating Ca²⁺ (K_ACa) and inactivating Ca²⁺ (K_ICa) with fixed Hill coefficients (n_A = 2.0, n_I = 1.0; Fujii et al., 2017; Nozaki et al., 2020; Itoh et al., 2021; Hirose et al., 2022; Fig. 4 B). The determined parameters for the mutant channels are listed in Table 2 and shown in Fig. 4, C–E, as relative values. Six CPVT mutants, V2321M, R2474S, Q4201R, R4497C, K4751Q, and K4805R, had significantly increased A_max, whereas four variants, D3638A, K4392R, H4762P, and I4867M, showed similar or slightly smaller A_max than WT. All CPVT mutants, except for K4392R, showed significant increases in 1/K_ACa. Among these mutants, R2474S, K4751Q, H4762P, and K4805R showed marked increases in 1/K_ACa (twofold or more). K_ICa was increased in most mutants, with the exceptions of D3638A, K4392R, and I4867M.

To gain insight into Ca²⁺ homeostasis in HEK293 cells expressing WT and mutant RYR2s, it is necessary to know [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting Ca²⁺ (A_rest). However, direct measurements by [³H]ryanodine binding are difficult due to the small values at resting Ca²⁺ (Fig. 4 A). We instead calculated A_rest, [Ca²⁺]_cyt-dependent Ca²⁺ release activity at 100 nM [Ca²⁺]_cyt, by substituting the obtained three parameters (Table 2) into Eqs. 1, 2, and 3, and expressed it as a value relative to WT (Fig. 4 F). This procedure has been shown to effectively evaluate [Ca²⁺]_cyt-dependent Ca²⁺ release activities of RYR1 (Murayama et al., 2015; Murayama et al., 2016). All mutations, except K4392R, enhanced A_rest by a factor of ≥2. H4762P resulted in the greatest A_rest, followed by K4805R, K4751Q, and R2474S in that order. The threshold [Ca²⁺]_ER for Ca²⁺ release was plotted with A_rest on a log scale (Fig. 4 G). There was a significant inverse correlation between them, suggesting that the threshold [Ca²⁺]_ER is dependent on the [Ca²⁺]_cyt-dependent Ca²⁺ release activity of RYR2. A significant correlation was also found between oscillation frequency and A_rest, even though the two mutants with the greatest A_rest, H4762P and K4805R, rarely showed Ca²⁺ oscillation (Fig. 4 H).

The above results suggest that Ca²⁺ oscillation frequency and [Ca²⁺]_ER correlate well with the [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting [Ca²⁺]_cyt. We next examined whether this is also the case for artificial mutations associated with various [Ca²⁺]_cyt-dependent Ca²⁺ release activities. We previously reported that single alanine substitutions at the S4–S5 linker region in RYR1 differentially affect the three parameters of [Ca²⁺]_cyt-dependent Ca²⁺ release activity, depending on the position of the α helix (Murayama et al., 2011). On the basis of a high degree of sequence homology between RYR1 and RYR2 at the S4–S5 linker region, the homologous mutations were made in RYR2, and their effects on [Ca²⁺]_ER were examined (Fig. 5). The [³H]ryanodine binding assay indicated that the S4–S5 linker region mutations had different degrees of influence on the [Ca²⁺]_cyt-dependent Ca²⁺ release activity, A_rest, depending on the mutation site (Fig. 5, A and B). A_rest was greatly enhanced by L4756A (82-fold) and T4754I (9-fold) and moderately enhanced by I4755A and S4757A (2-fold), but significantly suppressed by T4754A (0.6-fold) and S2758A (0.3-fold; Fig. 5 B). These changes due to individual mutations were similar to those observed at the corresponding sites in RYR1 (Murayama et al., 2011).

We then determined [Ca²⁺]_ER in HEK293 cells. L4756A showed almost complete depletion of [Ca²⁺]_ER and no Ca²⁺ oscillations (Fig. 5 C middle, Fig. 5, D–F). T4754I, I4755A, and S4757A also significantly reduced the threshold and nadir of [Ca²⁺]_ER. In contrast, T4754A and S4758A showed significantly higher [Ca²⁺]_ER than WT (Fig. 5, C right, D, and E; and Fig. S2 B). S4757A showed more frequent Ca²⁺ oscillations, whereas S4758A exhibited less frequent oscillations than WT (Fig. 5 F). In T4754I, high-frequency Ca²⁺ oscillations were observed in 35% of cells (52/150 cells), with a median of 0 due to the high proportion of cells without clear oscillation.

Taking the data of CPVT and S4–S5 mutants together, the relationship between threshold [Ca²⁺]_ER and A_rest was plotted (Fig. 5 G). The best-fit semilog lines of CPVT (blue line) and S4–S5 mutants (pink line) were very close to each other. In addition, a similar relationship was observed with CPVT and S4–S5 mutants between Ca²⁺ oscillation frequency and A_rest (Fig. 5 H); that is, in cells with low A_rest, Ca²⁺ oscillations rarely occur, whereas cells with higher RYR2 activity show a higher frequency of Ca²⁺ oscillations. However, in cells expressing mutant RYR2 with extremely high activity (L4756A), no Ca²⁺ oscillation occurs mainly due to the depletion of ER Ca²⁺. In H4762P, K4805R, and T4754I cells, which have high activity but weaker than L4756A, only a small fraction of cells showed obvious oscillations. Importantly our findings indicate that the threshold or steady-state [Ca²⁺]_ER is highly dependent on A_rest (Fig. 5 G), the [Ca²⁺]_cyt-dependent Ca²⁺ release activity of RYR2 at resting [Ca²⁺]_cyt, regardless of CPVT mutation or artificial mutation.

We previously showed that the threshold [Ca²⁺]_ER for Ca²⁺ release decreased with time after the induction of RYR2 expression in HEK293 cells (Uehara et al., 2017). This implies that the threshold [Ca²⁺]_ER is not fixed but is affected by the RYR2 expression levels. We examined the relationship between the threshold [Ca²⁺]_ER and RYR2 expression using WT and R2474S. The protein expression of WT and R2474S increased with time after induction (Fig. 6 A) and reached a quasi-steady state at ∼24 h (Fig. 6 B). There was no significant difference in expression levels between WT and R2474S. The threshold [Ca²⁺]_ER in both WT and R2474S cells decreased with time, but that of R2474S cells decreased faster than that of WT (Fig. 6, C and D). Consequently, the threshold [Ca²⁺]_ER in R2474S cells was lower than that in WT cells at the same expression level (Fig. 6 E). Based on these findings, we hypothesized that the net Ca²⁺ release rates in situ in RYR2-expressing cells are proportional to the product of the number of RYR2 channels and their own A_rest. There was a clear correlation between the threshold [Ca²⁺]_ER and the net Ca²⁺ release rate, regardless of the type of mutant (Fig. 6 F). These results support the hypothesis that the threshold [Ca²⁺]_ER is determined by both the [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting [Ca²⁺]_cyt and the density of RYR2 channels on the ER membrane.

The significant correlation between threshold [Ca²⁺]_ER and A_rest suggests that threshold [Ca²⁺]_ER may be mainly determined by [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting [Ca²⁺]_cyt as expressed in Eqs. 1, 2, and 3. We attempted to test the plausibility of this hypothesis by using a mathematical model. Before performing simulations of Ca²⁺ dynamics, we initially considered introducing the luminal Ca²⁺ regulation into Eqs. 1, 2, and 3 because RYR2 has been reported to be enhanced at high luminal Ca²⁺ (Bassani et al., 1995; Lukyanenko et al., 1996; Sitsapesan and Williams, 1997; Jiang et al., 2005; Qin et al., 2008). To accomplish this, we examined the possibility of including the luminal Ca²⁺ regulation in A_max or K_ACa. When we incorporated it into A_max, however, the maximal [Ca²⁺]_cyt-dependent activity became very small at low [Ca²⁺]_ER, which could not explain experimental data at all. Therefore, we included luminal regulation in K_ACa, that is, K_ACa is reduced to K_ACa,L, as described by Eqs. 4, 5, and 6, and examined whether the actual experimental data such as [³H]ryanodine binding and bilayer single-channel activity could be reasonably reproduced. K_ER,Ca was set as 3 mM considering that the activation of WT RYR2 by luminal Ca²⁺ occurs at ∼3 mM or more, and E was arbitrarily defined as 50 considering the activating ratio at high luminal Ca²⁺.

Fig. 7 shows the simulated static effects of luminal Ca²⁺ on putative [Ca²⁺]_cyt-dependent [³H]ryanodine binding and bilayer single-channel experiments. Under experimental conditions for [³H]ryanodine binding, we assumed that [Ca²⁺]_ER was equilibrated with [Ca²⁺]_cyt (i.e., [Ca²⁺]_ER = [Ca²⁺]_cyt), because [³H]ryanodine binding did not change at all when the ER membrane was made permeable to Ca²⁺ by Ca²⁺ ionophore treatment (Fig. 7 A). Fig. 7 B shows the calculated [Ca²⁺]_cyt-dependent RYR2 channel activity for WT and R2474S without (solid lines) and with (dotted lines) luminal regulation. Although the addition of luminal control makes the slope of Ca²⁺ activation slightly steeper, there were only minor differences between the two curves for both RYR2 WT and R2474S. This indicates that we can use the three parameters determined with [³H]ryanodine binding, even with the luminal regulation.

Fig. 7 C shows calculated [Ca²⁺]_ER-dependent RYR2 activity at constant [Ca²⁺]_cyt (100 nM). The [Ca²⁺]_ER-dependent RYR2 activity increased with [Ca²⁺]_ER by the incorporation of luminal regulation as expected, and the augmentation was more prominent in R2474S than in WT. Simulated [Ca²⁺]_cyt-dependent RYR2 activity in the presence of 1 mM [Ca²⁺]_ER is shown in Fig. 7 D. [Ca²⁺]_cyt dependence for activating Ca²⁺ shifted leftward by incorporation of luminal regulation. These simulations suggest that apparent potentiation of RYR2 activity by luminal Ca²⁺ in CPVT mutant RYR2s (Fig. 7 C) is explained by increased [Ca²⁺]_cyt-dependent activity without considering a change in parameters for ER Ca²⁺ regulation, E, and K_ER,Ca. It should be noted that [Ca²⁺]_cyt-dependent activity at pCa 7.0 with and without luminal regulation—A_rest,L and A_rest, respectively—are nearly proportional as shown in Fig 7 E, indicating that the relative [Ca²⁺]_cyt-dependent RYR2 activity at resting [Ca²⁺]_cyt (Figs. 4 F and 5 B) does not change even in the presence of luminal regulation.

Because Eqs. 4, 5, and 6 were able to reasonably reproduce the static effects of luminal Ca²⁺, we then attempted to mathematically describe calcium dynamics. Cells were assumed to comprise cytosolic and ER compartments (Fig. 8 A), where the transport of Ca²⁺ between the cytosol and ER involves Ca²⁺ release through RYR2, saturable ER Ca²⁺ uptake by SERCA, and a passive leakage pathway that is intrinsic to HEK293 cells. Cellular Ca²⁺ influx via the plasma membrane is assumed to obey zero-order kinetics provided that the extracellular bulk concentration is constant and Ca²⁺ efflux is driven by a presumptive Ca²⁺ pump. In addition, we assumed that SOCE operates when ER Ca²⁺ is depleted (Rios, 2010, 2013). In actual cells, it is necessary to consider the effect of diffusion; however, for the sake of simplicity, we assumed a rapid equilibrium of mobilized Ca²⁺. First, we tested if WT Ca²⁺ oscillations were reproduced by the model. For each mutant, the A_max, K_ACa, and K_ICa were the experimental values when n_A and n_I were regarded as 2 and 1, respectively (Table 2). The V_ratio, C_cyt, K_cyt, C_ER, and K_ER values are listed in Table 3. The E and K_ER,Ca were set at 50 and 3 mM, respectively. The ε, k_leak, V_max,ER, K_m,ER, v_in, v_in,SOC, K_SOC, V_max,out, and K_m,out were set at 1,200, 0.035/s, 0.24 mM/s, 0.0003 mM, 0.001 mM/s, 0.002 mM/s, 0.170 mM, 0.01 mM/s, and 0.0004 mM, respectively (Fig. 8 A, right), to satisfy the following three conditions: (1) the Ca²⁺ concentration in ER at a steady state was 0.6–0.8 mM in mock HEK293 cells; (2) the maximum Ca²⁺ concentration in ER was 0.3–0.5 mM with a periodic time of ∼60 s for WT RYR2 cells; and (3) Ca²⁺ maximum concentration in ER was 0.1–0.2 mM in R2474S mutant.

The mathematical model successfully reproduced the periodic cytoplasmic Ca²⁺ increase and [Ca²⁺]_ER decrease in WT cells. This was similar to the [Ca²⁺]_cyt and [Ca²⁺]_ER observed in the experiments (Fig. 8 B, left). Then, Ca²⁺ oscillations for mutant cells were calculated using the same constants and Ca²⁺ release activity with the three parameters for individual mutants. The model also effectively reproduced a reduced threshold [Ca²⁺]_ER for Ca²⁺ release and an increased oscillation frequency in R4497C and R2474S (Fig. 8 B, middle) and completely diminished Ca²⁺ oscillation with a further decrease in [Ca²⁺]_ER in H4762P cells (Fig. 8 B, right). We then calculated the oscillation frequency, [Ca²⁺]_cyt, and [Ca²⁺]_ER at various A_rest values, by changing the A_max or K_ACa values, and plotted their values in Fig. 8, C–E. Good correlations of simulated oscillation frequency versus A_rest (Fig. 8 C), peak and resting [Ca²⁺]_cyt versus A_rest (Fig. 8 D), and simulated threshold [Ca²⁺]_ER versus A_rest (Fig. 8 E) were reproduced in the A_rest range of 0.3–10. Furthermore, the lack of oscillation in cells with very large A_rest, (>10), that is, H4762P and L4756A, was also reproduced (Fig. 8, C–E). These results strongly support the idea that [Ca²⁺]_ER is largely determined by the [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting [Ca²⁺]_cyt.

Chen et al. (2014) have proposed that a glutamate residue (E4872) in the helix bundle crossing the proposed gate of the RYR2 is important for luminal Ca²⁺ sensing, in which mutation in the residue (E4872A or E4872Q) abolished SOICR but not cytosolic RYR2 Ca²⁺ activation. To test this hypothesis, we carried out [Ca²⁺]_ER measurements of these mutants in HEK293 cells. Introduction of E4872A or E4872Q mutation in WT, R2474S, and R4497C increased average [Ca²⁺]_ER signals (Fig. S3, A and B) and no or substantially reduced Ca²⁺ oscillations (Fig. S3, A and C). This phenomenon was consistent with those reported previously (Chen et al., 2014).

We then determined [Ca²⁺]_cyt-dependent [³H]ryanodine binding of these mutations. Introduction of E4872A mutation in WT (E4872A), R2474S (R2474S/E4872A), and R4497C (R4497C/E4872A) caused no measurable [³H]ryanodine binding (Fig. S3 D). The introduction of E4872Q mutation also abolished [³H]ryanodine binding in WT (E4872Q), but measurable binding was observed with double mutants, R2474S/E4872Q and E4497C/E4872Q, in which the former exhibited higher activity than the latter (Fig. S3 D, right). This order was the same as that of mutants without E4872Q mutation (Fig. 4, A and C). These results indicate that the primary effects of the E4872A and E4872Q mutations on Ca²⁺-dependent [³H]ryanodine binding was a reduction in A_max, and thereby a considerable reduction in [Ca²⁺]_cyt-dependent RYR2 activity.

The results in Figs. 2, 3, and 4 successfully provided the [Ca²⁺]_cyt-dependent Ca²⁺ release activities of individual CPVT mutants examined in this study. To address their pathological relevance, we plotted the age of onset of arrhythmia or SCD against A_rest (Fig. 9). Mutations with high A_rest, R2474S, K4751Q, and K4805R, caused disease at a young age (<8 yr), and only one or two (twin) individuals were affected in the family, suggesting very low penetrance. In contrast, two of the mild mutations, Q4201R and R4497C, were detected in multiple individuals of various ages of onset from young to old age, indicating relatively high penetrance (Laitinen et al., 2001; Priori et al., 2001). Only a single case has been reported for each of the other mutations with moderate [Ca²⁺]_cyt-dependent Ca²⁺ release activity, V2321M, D3638A, and I4867M. There was a significant difference in the ages of onset between patients with moderate and severe RYR2 mutations. The H4762P proband carries another mutation, G4662S, and her three family members with the H4762P mutation have been reported to be asymptomatic. Notably, a patient with K4392R did not show any difference from the WT. The mother of this patient had the same mutation and was asymptomatic (Arakawa et al., 2015). Recently, K4392R (rs753733164) was found to be a benign variant that does not exert arrhythmogenic effects in the majority of its carriers.

CPVT mutations in RYR2 are prone to cause abnormal spontaneous Ca²⁺ release, which leads to lethal arrhythmia. The HEK293 expression system is a good tool for studying the properties of mutant RYR2 and their effects on Ca²⁺ homeostasis. To explore the mechanisms controlling spontaneous Ca²⁺ release and threshold [Ca²⁺]_ER level for the release, we measured [Ca²⁺]_cyt and [Ca²⁺]_ER in cells expressing WT and mutant RYR2 and determined [Ca²⁺]_cyt-dependent Ca²⁺ release activity with a [³H]ryanodine binding assay. Our quantitative analysis indicates that all the CPVT mutations are associated with enhanced [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting [Ca²⁺]_cyt and enable ranking of the severity of these mutations. The model simulation reasonably explains the experimental results. Our results indicate that the [Ca²⁺]_cyt-dependent Ca²⁺ release activity of RYR2 is critically involved in determining the threshold [Ca²⁺]_ER for spontaneous Ca²⁺ release.

Chen and colleagues (Jiang et al., 2004; Jiang et al., 2005) have reported that CPVT mutations reduce threshold [Ca²⁺]_ER for spontaneous Ca²⁺ release. Based on the findings that many CPVT mutations increase [Ca²⁺]_ER-dependent activity without altering [Ca²⁺]_cyt-dependent RYR2 activity, they suggested that these mutations increase luminal Ca²⁺ sensitivity. In this study, we confirmed the decrease in threshold [Ca²⁺]_ER for spontaneous Ca²⁺ release by CPVT mutations. However, we show that all CPVT mutations have increased [Ca²⁺]_cyt-dependent activity, which is in marked contrast to their finding.

A significant difference between this and the previous studies is the assessment of Ca²⁺-dependent [³H]ryanodine binding activity of the mutant RYR2. For example, R2474S and R4497C mutations did not significantly affect [Ca²⁺]_cyt-dependent [³H]ryanodine binding in the studies by Jiang et al. (2004, 2005). In our studies, however, the two mutants exhibited increased [³H]ryanodine binding (Fig. 4).

There are two possible reasons for this difference. One is the data presentation of [³H]ryanodine binding. We expressed the binding activity as B/B_max, in which the determined activity was normalized by the number of total RYR2 molecules in the preparation (Murayama and Kurebayashi, 2011; Murayama et al., 2015; Fujii et al., 2017; Uehara et al., 2017; see Materials and methods). Using this method, we showed that most CPVT mutants, including R2474S and R4497C, significantly increased the A_max (Fig. 4). In contrast, in some studies, the binding activity was normalized by the peak values at the optimal Ca²⁺ (pCa ∼4; Jiang et al., 2004; Jiang et al., 2005; Xiao et al., 2016). This is useful for evaluating the [Ca²⁺]_cyt dependence but masks the difference in A_max values.

The second reason is the differences in experimental conditions. The Ca²⁺ sensitivity and maximum value of [³H]ryanodine binding vary depending on ionic strength, pH, Mg²⁺, and nucleotide concentrations (Meissner et al., 1997; Jiang et al., 2004; Chugun et al., 2007; Murayama and Kurebayashi, 2011). In addition, a difference in sample condition (microsomes or solubilized proteins) also affects the activity values (Murayama and Ogawa, 2004). These conditions affect not only the three parameters (A_max, K_ACa, and K_ICa) but also the Hill coefficients (n_ACa and n_ICa) in Eqs. 1, 2, and 3. Thus, the differences between ours and theirs in pH (7.0 versus 7.4), Mg²⁺ (1.0 versus 0 mM), and preparation (microsomes versus solubilized proteins) might contribute to the difference in experimental results.

In addition to the above, the fitting and extrapolation procedures using Eqs. 1, 2, and 3 enabled quantitative evaluation of [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting [Ca²⁺]_cyt, at which the direct determination of [³H]ryanodine binding activity is difficult. In particular, since the parameter K_ACa contributes to [Ca²⁺]_cyt-dependent Ca²⁺ release activity with a squared weight, the determination of K_ACa by a fitting procedure is important. In fact, in the study by Jiang et al. (2005), R2474S appeared to have a smaller K_ACa than the WT. Although the activities at resting [Ca²⁺]_cyt are very low compared with the maximum value, they are not zero and certainly cause Ca²⁺ release to affect basal [Ca²⁺]_ER. A similar conclusion was obtained with HEK293 cells expressing RYR1 (Murayama et al., 2015; Murayama et al., 2016; Murayama et al., 2018). Consequently, the CPVT mutations increase [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting [Ca²⁺]_cyt, although the mechanism of activation is mutation-site specific; some substantially increased sensitivity for K_ACa, whereas others predominantly increased A_max.

Previous reports have indicated that HEK293 cells expressing RYR2 with CPVT mutations show reduced threshold [Ca²⁺]_ER for spontaneous Ca²⁺ release (Jiang et al., 2004; Jiang et al., 2005; Xiao et al., 2016). Using R-CEPIA1er, an ER-targeted Ca²⁺ indicator and proteins with low Ca²⁺ affinity (K_D = 565 μM) and determining F_max and F_min values in the Ca²⁺ ionophore–containing solutions (Suzuki et al., 2014; Bovo et al., 2016; Fujii et al., 2017; Uehara et al., 2017), we confirmed these results in a more quantitative manner. There are two possible explanations for this phenomenon: [Ca²⁺]_ER is determined either by [Ca²⁺]_cyt-dependent Ca²⁺ release activity, the regulation by Ca²⁺ on the cytoplasmic side, or by the SOICR mechanism, which is regulated by Ca²⁺ on the ER luminal side. Our results indicate a good correlation between [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting [Ca²⁺]_cyt and [Ca²⁺]_ER (Fig. 5 G). More precisely, the threshold [Ca²⁺]_ER for Ca²⁺ release in individual mutants varies depending on the net permeability of Ca²⁺ through RYR2 channels: (1) with the same mutant, higher expression of RYR2 channels lowers the threshold [Ca²⁺]_ER; (2) at the same expression level, RYR2 mutants with higher [Ca²⁺]_cyt-dependent Ca²⁺ release activity presented a lower threshold [Ca²⁺]_ER; and (3) the [Ca²⁺]_ER is inversely correlated with net [Ca²⁺]_cyt-dependent Ca²⁺ release activity at rest, as a product of mutant-specific [Ca²⁺]_cyt-dependent Ca²⁺ release activity at rest and density of RYR2 molecules in the cell, on the ER membrane (Fig. 6 F). Xiao et al. (2016), also reported that many RYR2 CPVT mutations enhance cytosolic Ca²⁺ activation and decrease threshold [Ca²⁺]_ER for Ca²⁺ release. These results support the idea that [Ca²⁺]_cyt-dependent Ca²⁺ release activity, but not SOICR, mainly determines threshold [Ca²⁺]_ER.

To express both bell-shaped cytoplasmic [Ca²⁺]_cyt-dependent RYR2 activity (Murayama and Kurebayashi, 2011; Uehara et al., 2017; Nozaki et al., 2020; Hirose et al., 2022) and [Ca²⁺]_ER-dependent activations at ∼1 mM or more (Bassani et al., 1995; Lukyanenko et al., 1996; Sitsapesan and Williams, 1997; Gyorke and Gyorke, 1998; Jiang et al., 2004), we incorporated parameters for luminal Ca²⁺ activation (E and K_LCa) into Eqs. 1, 2, and 3 to obtain Eqs. 4, 5, and 6, where we assumed that K_ACa is reduced by [Ca²⁺]_ER. These equations predict that an increase in [Ca²⁺]_cyt-dependent RYR2 activity causes an apparent increase in [Ca²⁺]_ER-dependent RYR2 activity with unchanged E and K_LCa (Fig. 7 C). This can be easily understood by considering that the channel activity at resting [Ca²⁺]_cyt was originally higher in CPVT mutants than WT, even though their absolute values are apparently small, and luminal activation equally enhances the [Ca²⁺]_cyt-dependent activity to make them more visible.

To obtain further insight into the luminal Ca²⁺ regulation, we examined effects of mutations at the putative luminal Ca²⁺ sensing site, E4872A or E4872Q, which have been reported to increase SOICR threshold with unchanged cytosolic RYR2 Ca²⁺ activation (Chen et al., 2014). Contrary to our expectation, these mutations exhibited strongly suppressed [Ca²⁺]_cyt-dependent [³H]ryanodine binding, the properties of which are similar to those of loss-of-function disease mutations (Fujii et al., 2017; Hirose et al., 2022). E4872 is located in the cytoplasmic side of S6 segment and may form an intersubunit salt-bridge with R4874 (Peng et al., 2016). Thus, E4872 may not contribute to RYR2 activation by luminal Ca²⁺. Further structural and functional studies of RYR2 will be required to understand the molecular basis of luminal Ca²⁺ activation of RYR2.

We simulated Ca²⁺ influx and efflux across the cytoplasmic and ER membranes in HEK293 cells using a minimal essential model instead of detailed models with many components (Fig. 8). Our simple model reasonably reproduced the Ca²⁺ oscillations and the effects of changes in [Ca²⁺]_cyt-dependent Ca²⁺ release activity on Ca²⁺ oscillation frequency, threshold [Ca²⁺]_ER level, and peak amplitude of [Ca²⁺]_cyt in WT and mutant RYR at constant luminal Ca²⁺ regulation parameters, E and K_ER,Ca. However, the simulation and actual measurements suggest that Ca²⁺ oscillations may not always be observed. In any case, resting/threshold [Ca²⁺]_ER levels will vary depending on the RYR2 activity, with or without oscillations. Ca²⁺ oscillation does not occur if the RYR2 activity is too low or too high (Fig. 3 A, H4762P; Fig. 5 C, L4756A; and Fig. 8, C–E). In addition, Ca²⁺ oscillations were rarely seen in HEK293 cells expressing WT or mutant RYR1, but [Ca²⁺]_ER drastically changes depending on RYR1 activity (Murayama et al., 2015; Murayama et al., 2016). The A_max, K_ACa, K_ICa, n_ACa, and dependence on [Ca²⁺]_ER of RYR1 are different from those of RYR2. Moreover, it should be noted that in the actual experiments, there were cell-to-cell variations in oscillation frequency, threshold and nadir [Ca²⁺]_ER, and peak [Ca²⁺]_cyt. These variations are probably due to cell-to-cell variations in Ca²⁺ influx and efflux parameters.

Ca²⁺ oscillations would appear when the [Ca²⁺]_cyt-dependent Ca²⁺ release parameters and the Ca²⁺ flux parameters were well balanced. They can occur under conditions favorable for the regenerative activation of RYR2. For example, Ca²⁺ oscillation is probably more likely to occur with higher Hill coefficients for activating [Ca²⁺]_cyt, because released Ca²⁺ will increase [Ca²⁺]_cyt to activate RYR2 to induce more Ca²⁺ release. But this is not an absolute requirement. Our understanding of the mechanism by which RYR2 causes Ca²⁺ oscillation is not yet sufficient. More systematic simulation and experimental data will help elucidate the mechanism.

Our simulation indicates that the effects of CPVT mutations on the basic phenomenon of cytoplasmic Ca²⁺ oscillations and corresponding [Ca²⁺]_ER changes can be explained only by the difference in [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting [Ca²⁺]_cyt determined by [³H]ryanodine binding measurements. However, it is possible that some specific mutants may have a larger E or smaller K_ACa,L compared with WT. In this case, these mutants will show more severe leak and greater reduction in [Ca²⁺]_ER than expected from the cytoplasmic three parameters. The model simulation would be useful to find such mutations. In addition, combination of model simulation and measurements using HEK293 expression system may be also helpful in exploring effects of the parameters of Ca²⁺ influx and efflux through plasma membrane and ER membrane, such as effects of the adrenergic β stimulation, which induces CPVT by affecting RYR2 phosphorylation, SERCA activity, and Ca²⁺ influx through plasma membrane.

Next-generation sequencing technology has revealed numerous novel RYR2 mutations linked to arrhythmias. However, it is difficult to predict their functional effects from the results of such sequencing. Heterologous expression systems with HEK293 cells have been successfully used to assess the phenotypes resulting from mutations in RYR2 that cause both loss-of-function and gain-of-function mutations (Jiang et al., 2002; Jones et al., 2008; Fujii et al., 2017; Uehara et al., 2017; Hirose et al., 2022; Nozaki et al., 2020). Here we show that the [Ca²⁺]_cyt-dependent Ca²⁺ release activity at resting [Ca²⁺]_cyt, A_rest, of mutant RYR2 correlates well with the age of onset of arrhythmia or SCD, an index of severity of the disease (Fig. 9): mutations with A_rest 2 to 3 times higher than WT tend to cause relatively mild familial arrhythmias, whereas mutations with A_rest 10 times higher than WT cause more severe arrhythmias at younger age. However, SCD may occur even with mutations with mild effects. Quantitative analysis of the [Ca²⁺]_cyt-dependent Ca²⁺ release activity will be useful not only to distinguish the phenotypes but also to assess the severity of diseases.

In this study, we investigated the properties of the homotetrameric mutant channels expressed in HEK293 cells. Because CPVT patients have heterozygous RYR2 channels, their changes in activity are expected to be smaller than those of homozygous channels. Therefore, the [Ca²⁺]_ER in cells with heterozygous channels may not be reduced as much as that in cells with homozygous channels. However, we need to determine the activities of heterotetrameric channels consisting of WT and mutant RYR2 in various combinations. The second limitation is that we do not know the effects of mutations on interactions between RYR2 and cardiomyocyte-specific proteins such as FKBP12.6, triadin, calsequestrin, PKA, and CaMKII, because we performed observations only in HEK293 cells. In particular, effects of adrenergic stimulations that cause CPVT need to be investigated more with cardiomyocytes or HEK293 cells expressing cardiac proteins. Furthermore, other genetic and environmental factors are also considered to be involved in human diseases. However, clarifying the impact of mutations on RYR2 channel activity provides an essential basis for understanding mutant RYR2 channel regulation and for designing therapeutic strategies.

Eduardo Ríos served as editor.

We thank Ikue Hiraga and the Laboratory of Radioisotope Research, Research Support Center, Juntendo University Graduate School of Medicine, for their technical assistance. We also thank Edanz for editing the English text of the draft of this manuscript.

This work was supported by JSPS KAKENHI grants 19K07105 and 22K06652 to N. Kurebayashi and 19H03404 and 22H02805 to T. Murayama; the Practical Research Project for Rare/Intractable Diseases (19ek0109202 to S. Ohno and N. Kurebayashi) from the Japan Agency for Medical Research and Development (AMED); Platform Project for Supporting Drug Discovery and Life Science Research (Basis for Supporting Innovative Drug Discovery and Life Science Research grant JP20am0101080 to T. Murayama and N. Kurebayashi); an Intramural Research Grant (2-5) for Neurological and Psychiatric Disorders from the National Center of Neurology and Psychiatry to T. Murayama; and the Vehicle Racing Commemorative Foundation to T. Murayama (grants 6114, 6237, and 6303).

The authors declare no competing financial interests.

Author contributions: N. Kurebayashi, T. Murayama, and T. Kobayashi performed the cell biological experiments; R. Ota and F. Yamashita developed the simulation program; R. Ota, F. Yamashita, and N. Kurebayashi performed the simulations and analyses; J. Suzuki, K. Kanemaru, and M. Iino provided experimental tools; N. Kurebayashi, T. Murayama, R. Ota, J. Suzuki, K. Kanemaru, T. Kobayashi, and T. Sakurai analyzed the experimental data; S. Ohno and M. Horie analyzed the clinical data; and N. Kurebayashi, T. Murayama, S. Ohno, M. Iino, and F. Yamashita wrote the manuscript. All authors discussed the results and approved the final version of the manuscript.