Catecholaminergic polymorphic ventricular tachycardia is caused by mutation-linked defective conformational regulation of the ryanodine receptor

Abstract

Rationale

Catecholaminergic polymorphic ventricular tachycardia (CPVT) is caused by a single point mutation in a well-defined region of the cardiac type-2 ryanodine receptor (RyR2). However, the underlying mechanism by which a single mutation in such a large molecule produces drastic effects on channel function remains unresolved.

Objective

Using a knock-in (KI) mouse model with a human CPVT-associated RyR2 mutation (R2474S), we investigated the molecular mechanism by which CPVT is induced by a single point mutation within the RyR2.

Methods and Results

The R2474S/+ KI mice showed no apparent structural or histological abnormalities in the heart, but they showed clear indications of other abnormalities. Bidirectional or polymorphic VT was induced after exercise on a treadmill. The interaction between the N-terminal (aa 1–600) and central (aa 2000–2500) domains of the RyR2 (an intrinsic mechanism to close Ca²⁺ channels) was weakened (domain unzipping). Upon protein kinase A (PKA)-mediated phosphorylation of the RyR2, this domain unzipping further increased, resulting in a significant increase in the frequency of spontaneous Ca²⁺ transients. cAMP-induced aberrant Ca²⁺ release events (Ca²⁺ sparks/waves) occurred at much lower sarcoplasmic reticulum (SR) Ca²⁺ content as compared to the wild-type (WT). Addition of a domain-unzipping peptide, DPc10 (aa 2460–2495), to the WT reproduced the aforementioned abnormalities that are characteristic of the R2474S/+ KI mice. Addition of DPc10 to the (cAMP-treated) KI cardiomyocytes produced no further effect.

Conclusions

A single point mutation within the RyR2 sensitizes the channel to agonists and reduces the threshold of luminal [Ca²⁺] for activation, primarily mediated by defective inter-domain interaction within the RyR2.

Introduction

To date, more than 70 cardiac ryanodine receptor (RyR2) missense mutations have been identified that are linked with 2 inherited forms of sudden cardiac death—catecholaminergic polymorphic ventricular tachycardia (CPVT)¹ and arrhythmogenic right ventricular cardiomyopathy (ARVC) type 2.¹ These mutations cluster in 3 well-defined regions of the RyR2 that correspond to malignant hyperthermia (MH) or the central core disease (CCD) mutable regions, designated as the N-terminal domain (aa 1–600), central domain (aa 2000–2500), and the C-terminal transmembrane channel domain of the skeletal muscle-type ryanodine receptor (RyR1).¹ This suggests that the RyR2 shares a common domain-mediated channel regulation mechanism with RyR1. Mutations at different positions in each of these domains result in the nearly identical phenotype of channel dysfunctions such as hyperactivation of the Ca²⁺ channel and hypersensitization to agonists. To account for these phenomena, Ikemoto et al^{2, 3} proposed the so-called “domain switch hypothesis” and stated that in the resting or non-activated state, the N-terminal domain and the central domain make close contact at several sub-domains (domain zipping). Then, on physiological or pharmacological stimulation these critical inter-domain contacts are weakened, resulting in the loss of conformational constraints (domain unzipping), thus lowering the energy barrier for Ca²⁺ channel opening. Consistent with this hypothesis, single particle analysis of the three-dimensional structure of the RyR2 molecule revealed that the N-terminal and central domains (located in domains 5 and 6 of the so-called clamp region, respectively) are in a close apposition to each other. ^{4, 5}

Recent reports deal with 3 types of knock-in (KI) mice with human CPVT/ARVC-associated RyR2 mutations: R4496C,⁶ R176Q,⁷ and R2474S.⁸ Injection of caffeine plus epinephrine or exercise induces ventricular tachycardia in these mice, indicating that these point mutations can cause lethal arrhythmias. However, the underlying mechanism by which a single mutation causes lethal arrhythmia remains unresolved. We recently reported that in failing hearts, defective inter-domain interaction within the RyR2 (aberrant unzipping of the N-terminal/central domain pair and channel activation in an otherwise resting state) causes diastolic Ca²⁺ leakage and contractile dysfunction.⁹ As shown in our previous report,⁹ pathological conditions (diastolic Ca²⁺ leakage and contractile dysfunction) are reproduced in the otherwise normal system by adding DPc10, a central domain peptide (Gly²⁴⁶⁰–Pro²⁴⁹⁵) of the RyR2 that interferes with the interaction between the N-terminal and central domains of the RyR2 and causes defective domain unzipping. George et al¹⁰ showed that functional coupling between the cytoplasmic and transmembrane domains of the RyR2 is mediated by the 3722–4610 residue region, called the I-domain and that sudden cardiac death (SCD)-linked mutations occurring in the I-domain (N4104K and R4496C) caused channel instability and Ca²⁺ release dysfunction owing to defective inter-domain interactions between the I-domain and the channel domain. These findings suggest that the weakened inter-domain interaction at least in these 2 regions of the RyR2 (N-terminal domain/central domain and I-domain/channel domain) is the key mechanism underlying the pathogenesis of CPVT/ARVC and heart failure.

Although the domain peptide approach provided important information regarding the underlying mechanism of the RyR2 abnormalities during heart failure and lethal arrhythmia, further in vivo studies with the mutation-linked disease model are required for a straightforward test of the inter-domain interaction hypothesis. In the present study, using the KI mouse model with a human CPVT-associated RyR2 mutation, R2474S, we investigated the molecular mechanism by which CPVT is induced by a single point mutation within the RyR2. The data presented here suggested that the introduced mutation, in fact, causes defective inter-domain interaction in the RyR2, reduces the threshold of luminal Ca²⁺-dependance for channel activation, sensitizes RyR2 to protein kinase A (PKA)-dependent phosphorylation, and in turn leads to CPVT.

Methods

For a detailed description, see the expanded Materials and Methods sections and Online Figures I and II in the online data supplement.

Animals

This study conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). The care of the animals and the protocols used were in accordance with guidelines laid down by the Animal Ethics Committee of Yamaguchi University School of Medicine.

Statistics

Paired or unpaired t-tests were used for statistical comparisons of data obtained during the 2 different situations, while ANOVA with a post hoc Scheffe’s test was used for statistical comparison of concentration-dependent data. All data are expressed as mean ± standard error (SE). A p value less than 0.05 was considered statistically significant.

Results

There was no appreciable change in structural or functional characteristics of R2474S/+ KI mice during the resting (non-activated) state

In the absence of activation, there was no appreciable difference in the structural or functional features of the hearts between WT and KI mice. Thus, the cross-sectional view showed the identical features (Figure 1A). Echocardiography revealed no functional difference between WT and KI mice (Figure 1B). There was no appreciable change in the KI mice regarding the expression or phosphorylation levels of any of the sarcoplasmic reticulum (SR) proteins examined (Online Figure III).

Figure 1.

Structural and functional characterization of hearts from WT and R2474S/+ KI mice. A, Representative images of Hematoxylin-Eosin-stained hearts from WT and KI mice. B, Summarized data of body weight (BW), heart rate (HR), left ventricular end-diastolic diameter (LVEDD), left ventricular end-systolic diameter (LVESD), intra-ventricular septum diastolic thickness (IVSD), left ventricular posterior wall diastolic thickness (LVPWD), left ventricular fractional shortening ((LVEDD−LVESD)/LVEDD×100). N: the number of mice examined.

Exercise or drug (epinephrine and caffeine) administration induced VT in R2474S/+ KI mice

In the resting conscious condition, we frequently observed polymorphic ventricular premature contractions in KI mice in response to even very weak stimuli, like light or sound, but not in WT mice (data not shown). Injection of caffeine plus epinephrine (i.p.) or exercise on a treadmill induced bidirectional or polymorphic VT in KI mice, but not in WT mice (Figure 2A). The duration of VT in most KI mice was less than 30 sec (Figure 2B).

Figure 2.

Effect of epinephrine plus caffeine or exercise on ventricular arrhythmia. A, Representative ECG recordings in WT and R2474S/+ KI mice. In all KI mice examined, bidirectional or polymorphic ventricular tachycardia (VT) was induced by epinephrine (2 mg/kg of body weight (BW) i.p.) and caffeine (120 mg/kg of body weight i.p.) (n=10) (top) or exercise with treadmill (n=6) (bottom). In WT mice, neither epinephrine plus caffeine (n=8) nor exercise (n=6) produced VT. B, Durations of VT and numbers of VT episodes observed during the 5 minutes period in WT and KI mice. VT was induced by only epinephrine (2 mg/kg of body weight i.p.) without caffeine. N: the number of mice examined.

Seizures were not observed in R2474S/+ KI mice

To determine whether seizures occurred in KI mice, as previously reported,⁸ we monitored the behavior of mice by video recording for 1 week and assessed their susceptibility to seizures by pharmacologic induction (see expanded Materials and Methods in the online data supplement). Thus far, we have not observed spontaneous seizures in KI or WT mice. Furthermore, a pharmacological provocation test with 4-aminopyridine and caffeine showed no difference in the latency to the development of generalized tonic-clonic seizures (Online Figure IV).

Relaxation phase of cell shortening and Ca²⁺ transient was prolonged in the isoproterenol-activated R2474S/+ KI mice

There was no statistically significant difference in the contour or kinetic parameters of Ca²⁺ transient and cell shortening at baseline between the WT and KI mice (Figure 3A, Online Table I). In response to isoproterenol (ISO), however, the time from the peak to 80% decline in cell shortening or Ca²⁺ transient was prolonged in KI cardiomyocytes, suggesting a delay in the inactivation of Ca²⁺ release and/or spontaneous Ca²⁺ release events. Moreover, the SR Ca²⁺ content, determined by caffeine application, was significantly lower in KI cardiomyocytes than in WT cardiomyocytes, both before and after the addition of isoproterenol (Figure 3B). SERCA2 activity could contribute to the slowed decay kinetics of the Ca²⁺ transient during heart failure. Thus, we next measured the SERCA2-mediated Ca²⁺ uptake by monitoring the time-dependent change in the intra-SR [Ca²⁺] (Online Figure V). As shown, there was no appreciable change in the time course of SR Ca²⁺ uptake.

Figure 3.

Cell shortening and intracellular Ca²⁺ transient in intact cardiomyocytes. A, Representative recordings of cell length and fura-2 fluorescence signal, at a pacing rate of 2 Hz, with isoproterenol (ISO) (10 nmol/L) activation (ISO) and without it (Baseline). B, Representative recording of fura-2 fluorescence signal after addition of caffeine (top figure), which is a measure of the SR Ca²⁺ content, and the summarized data of SR Ca²⁺ content (bottom). Caffeine-induced Ca²⁺ transient was measured by first applying a stimulation train at 2 Hz, and then by rapidly switching the super fusing solution to a solution containing 20 mmol/L caffeine for 5 to 6 seconds. N: the number of cells from 3–9 hearts.

RyR2 Ca²⁺ channels of R2474S/+ KI cardiomyocytes were hypersensitive to channel activation by isoproterenol, PKA phosphorylation, and luminal calcium

In KI cardiomyocytes, Ca²⁺ spark frequency (SpF) was higher than that in WT cardiomyocytes, both before and after the addition of isoproterenol (Figure 4A, Online Figure VIA for 3-dimensional images). In the presence of a higher concentration of isoproterenol (100 nmol/L), the frequency of spontaneous Ca²⁺ waves was much higher in KI cardiomyocytes than in WT cardiomyocytes and the duration of local Ca²⁺ release provoked by the transmission of the Ca²⁺ waves was markedly prolonged. (Online Figure VII). In both KI and WT cardiomyocytes, isoproterenol (10 nmol/L) increased the peak amplitude, full width at half maximum (FWHM), and full duration at half maximum (FDHM) (Online Table II). In KI cardiomyocytes, however, FDHM increased approximately twice as much as in WT cardiomyocytes (Online Table II). In response to isoproterenol (10 nmol/L), the level of PKA-dependent phosphorylation at Ser2808 of the RyR2 showed a rather modest increase that was slightly less than half of maximum phosphorylation (see Online Figure II). But in this case, the extent of the increase was nearly the same in the KI and WT cardiomyocytes (Figure 4B). These results suggest that the sensitivity of the channel to both SR content (luminal [Ca²⁺]) and PKA phosphorylation is markedly increased in KI channels.

Figure 4.

Local Ca²⁺ release events in intact cardiomyocytes and Ser2808 phosphorylation in the RyR2. A, Spontaneous Ca²⁺ sparks. Top: Recordings of line-scan images of fluo-4 AM fluorescence with or without isoproterenol (10 nmol/L) in intact WT and R2474S/+ KI cardiomyocytes. Bottom: Summarized data of Ca²⁺ spark frequency. N: the number of cells from 6–8 hearts. B, Comparison of protein kinase A (PKA)-mediated phosphorylation level at Ser2808 of the RyR2 (P-Ser2808-RyR2) in intact KI and WT cardiomyocytes. In the presence of isoproterenol (10 nmol/L), KI and WT cardiomyocytes were incubated in the lysis buffer, and centrifuged. Then, the supernatant fraction containing crude homogenate was used for the phosphorylation assay. C, Comparison of P-Ser2808-RyR2 in KI and WT cardiomyocytes. In the presence of cAMP (1 μmol/L), okadaic acid (1 μmol/L), and CaMKII inhibitor KN-93 (1 μmol/L), KI and WT cardiomyocytes were incubated in the lysis buffer, and centrifuged. Then, the supernatant containing crude homogenate was used for phosphorylation assay. N: the number of cell lysate from 3–5 heats. D, Effect of cAMP on the RyR2-FKBP12.6 association in WT and KI cardiac homogenates. (Top) Representative Western blots of the RyR2 and FKBP12.6 after immunoprecipitation (cf. online supplemental methods). (Bottom) Summarized data of the ratio of FKBP12.6 to the RyR2, which was normalized to control condition. Although FK506 dissociated FKBP12.6 from the WT RyR2, shown as positive control, cAMP (1 μmol/L) did not dissociate the RyR2-bound FKBP12.6 in WT and KI cardiac homogenates. N: the number of SR preparations from 7–14 hearts.

For further analysis of the increased sensitivity of KI channels to PKA phosphorylation and the SR Ca²⁺ content, we measured the PKA phosphorylation level at Ser2808 of RyR2 after addition of cAMP under the same conditions as the Ca²⁺ spark assay ([Ca²⁺] = 30 nmol/L, buffered by 0.5 mmol/L ethylene glycol tetraacetic acid (EGTA)). For this purpose, we added the Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) inhibitor KN-93 (1 μmol/L) to inhibit the effect of intrinsic CaMKII on the phosphorylation of the RyR2. We then measured both SpF and SR Ca²⁺ content in the absence and in the presence of 1 μmol/L cAMP in the saponin-permeabilized cardiomyocytes. There was no significant difference in the PKA phosphorylation level at Ser2808 between the KI and WT cardiomyocytes, both in the absence and the presence of cAMP (Figure 4C). Because there is marked variation in specificity among antibodies generated against Ser2808, ¹¹ we compared the phosphorylation level at Ser2808 of RyR2 by using another different antibody against Ser2808 (Badrilla, UK). There was no significant difference in the PKA phosphorylation level at Ser2808 between KI and WT cardiomyocytes (Online Figure VIII).

To test the postulated possibility that de-stabilization of the RyR2 caused by the dissociation of FK506 binding protein 12.6 (FKBP12.6) from the RyR2 is a common pathogenic mechanism underlying heart failure and lethal arrhythmia,^{12, 13} we determined the RyR2-bound FKBP12.6 by using a pull-down assay. As shown in Figure 4D, between the WT and KI mice, there was no significant difference in the RyR2-bound FKBP12.6 in the absence or the presence of cAMP (1 μmol/L).

Figure 5A shows the representative traces of Ca²⁺ sparks in the presence of various concentrations of cAMP (0.1–1 μmol/L) (see also Online Figure VIB for 3-dimensional images of Ca²⁺ sparks). The dependence of SpF on the SR Ca²⁺ content is plotted in Figure 5B. To obtain the point at lower SR Ca²⁺ content, thapsigargin was added to the cardiomyocytes. Then, both Ca²⁺ sparks and SR Ca²⁺ content were measured 3, 5, and 10 minutes after the addition of thapsigargin. SpF was higher in KI than in WT cardiomyocytes, although there was a considerable reduction in the SR Ca²⁺ content in KI cardiomyocytes. As a result, there was a considerable left-shift of the SpF versus SR Ca²⁺ content plot in the case of KI cardiomyocytes (Figure 5B). The effect of cAMP on Ca²⁺ spark characteristics are summarized in Online Table III. Compared to the WT cardiomyocytes, both the peak amplitude and FWHM decreased whereas FDHM increased in KI cardiomyocytes, suggesting a delay in the inactivation of the RyR2. These results suggest that the threshold of luminal [Ca²⁺] for channel opening decreased considerably owing to the single R2474S CPVT mutation of the RyR2. To assess the maximum capacity of Ca²⁺ loading of SR, we added tetracaine (1 mmol/L) to inhibit Ca²⁺ release and then evaluated SR Ca²⁺ content by caffeine application. After treatment with the Ca²⁺ release blocker tetracaine (arrows, Online Figure IX), the SR Ca²⁺ content increased and reached nearly the same level (approximately 4 F/F₀) in both WT and KI cardiomyocytes. This suggests that the reduced SR Ca²⁺ content in the KI cardiomyocytes was not due to reduced capacity of Ca²⁺ loading, but due to increased activation of the Ca²⁺ release flux relative to the influx.

Figure 5.

Decreased threshold of SR Ca²⁺ content to induce spontaneous Ca²⁺ sparks in saponin-permeabilized cardiomyocytes. A, Representative line-scan images of cardiomyocytes after addition of cAMP (0.1–1 μmol/L), at 30 nmol/L [Ca²⁺] buffered by 0.5 mmol/L EGTA. B, Relationship between Ca²⁺ spark frequency (SpF) and SR Ca²⁺ content. SR Ca²⁺ content was measured by addition of 10 mmol/L caffeine. To obtain the point at lower SR Ca²⁺ content, thapsigargin was added to the cardiomyocytes. Then, both Ca²⁺ sparks and SR Ca²⁺ content were measured 3, 5, and 10 minutes after the addition of thapsigargin [WT (n=6–12 hearts): (−) thapsigargin:107–441 cells for SpF and 10–84 cells for SR Ca²⁺ content; (+) thapsigargin: 4 cells for SpF and 4 cells for SR Ca²⁺ content, R2474S/+ KI (n=6–9 hearts): (−) thapsigargin: 91–361 cells for SpF and 14–46 cells for SR Ca²⁺ content; (+) thapsigargin: 4 cells for SpF and 4 cells for SR Ca²⁺ content]. C, Effect of protein kinase A (PKA) phosphorylation on Ca²⁺ spark frequency at comparable SR [Ca²⁺] content in WT and R2474S/+ KI saponin-permeabilized cardiomyocytes. For PKA phosphorylation of the RyR2, both cAMP (1 μmol/L) and okadaic acid (1 μmol/L) were added to the cardiomyocytes at 30 nmol/L [Ca²⁺] buffered by 0.5 mmol/L EGTA. Two minutes after the addition of cAMP, the cAMP was washed away to prevent further PKA phosphorylation. Then, thapsigargin (Thap) (0.3 μmol/L) was added. One minute later, line-scanned images were obtained. Arrow indicates the change in the data point produced by the addition of thapsigargin, Note that addition of cAMP increased SpF at a comparable SR Ca²⁺ content only in the KI cardiomyocytes.

D, E, Top: Representative line-scan images of cardiomyocytes after addition of DPc10 (50 μmol/L) and/or dantrolene (1 μmol/L), at 30 nmol/L [Ca²⁺] buffered by 0.5 mmol/L EGTA. Bottom: The dependence of SpF on the SR Ca²⁺ content after addition of DPc10 in the absence or in the presence of dantrolene. Arrow indicates the shift of the data points by the addition of DPc10 in the absence or in the presence of dantrolene [WT (n=5–12 hearts): {(−) DPc10: 441 cells for SpF and 84 cells for SR Ca²⁺ content; (+) DPc10: 13 cells for SpF and 7 cells for SR Ca²⁺ content; (+)DPc10, (+)dantrolene: 47 cells for SpF and 29 cells for SR Ca²⁺ content}, KI (n=3–9 hearts): {(−) cAMP, (−) DPc10, (−) dantrolene: 91 cells for SpF and 14 cells for SR Ca²⁺ content; (+) cAMP (1 μmol/L), (+) DPc10: 15 cells for SpF and 10 cells for SR Ca²⁺ content; (+) cAMP (1 μmol/L), (+) dantrolene (1 μmol/L): 37 cells for SpF and 25 cells for SR Ca²⁺ content}].

PKA-phosphorylation modulates not only RyR2 function but also SR Ca²⁺ load-dependence of channel activation. Thus, to clarify the direct effect of PKA-phosphorylation on RyR2 function, we measured the SpF and SR Ca²⁺ content in the presence of 0.3 μmol/L thapsigargin (inhibitor of SERCA2a activity) (Figure 5C). At comparable SR Ca²⁺ content, PKA phosphorylation increased SpF in KI cardiomyocytes, but not in WT cardiomyocytes. This suggests that PKA-dependent phosphorylation sensitized diastolic SR Ca²⁺ release, but only in CPVT, and not WT RyR2 channels.

Domain unzipping peptide DPc10 mimics the phenotype of KI channels in WT cardiomyocytes

As shown previously,⁹ DPc10 (a peptide corresponding to the 2460–2495 region of the central domain) mimics the channel disorder in the CPVT mutant (R2474S) by interfering with the channel-stabilizing inter-domain interactions between the N-terminal and central domains (i.e., domain unzipping). To assess the effect of domain unzipping on the dependence of SpF on the SR Ca²⁺ content, we added DPc10 to the WT saponin-permeabilized cardiomyocytes. Successful incorporation of DPc10 was confirmed by the intracellular fluorescence signal of Alexa fluor 488-labeled DPc10 (Online Figure X). DPc10 (50 μmol/L, applied externally) induced Ca²⁺ sparks in WT cardiomyocytes (Figure 5D, top). Importantly, addition of DPc10 caused a left-shift of the SpF/SR Ca²⁺ content relationship in WT cardiomyocytes (Figure 5D, bottom), resulting in the nearly identical SpF/SR Ca²⁺ profile as that of KI cardiomyocytes (Figure 5B). Similar to cAMP-treated KI cardiomyocytes, addition of DPc10 to WT cardiomyocytes prolonged FDHM (21.32 ± 0.31 ms (n = 27 cells) to 23.82 ± 0.54 ms (n = 48 cells); p < 0.01), but decreased the peak (1.74 ± 0.03 to 1.69 ± 0.01; p = 0.068) and FWHM (2.05 ± 0.03 μm to 2.00 ± 0.02 μm; p < 0.01). Addition of DPc10 to the (cAMP-treated) KI cardiomyocytes produced no further effect (Figure 5E). To assess whether suppression of domain unzipping reversed the left-shift of the SpF/SR Ca²⁺ content relationship in KI cardiomyocytes, we added dantrolene, which has corrected aberrant domain unzipping and prevented the development of heart failure¹⁴. Interestingly, dantrolene shifted back the SpF/SR Ca²⁺ content relationship to the right, almost towards the point corresponding to that of WT cardiomyocytes (Figure 5E). Dantrolene also prevented the DPc10-induced left-shift of the SpF/SR Ca²⁺ content relationship in WT cardiomyocytes (Figure 5D). Collectively, these findings indicate that the defective inter-domain interaction within the RyR2 caused by either point mutation (R2474S) or domain unzipping peptide (DPc10) reduced the threshold of luminal [Ca²⁺] for channel activation, leading to the phenotype of CPVT (hyper-activation of the RyR2 channel). This suggests that as in the case of MH, correction of the defective inter-domain interaction by dantrolene may provide an effective method to treat CPVT.

Luminal [Ca²⁺] of SR in saponin-permeabilized WT and KI cardiomyocytes

To confirm the above notion that the level of SR Ca²⁺ load was decreased in KI cardiomyocytes, we monitored luminal [Ca²⁺] in both cardiomyocytes using fluo-5N as a luminal Ca²⁺ probe. The average level of free diastolic SR Ca²⁺ decreased in KI cardiomyocytes, and upon caffeine-induced discharge of luminal Ca²⁺, the luminal Ca²⁺ reached the same basal level in both WT and KI cardiomyocytes (Figure 6).

Figure 6.

Spectroscopic determination of luminal [Ca²⁺] of the SR in saponin-permeabilized WT and R2474S/+ KI cardiomyocytes (see “Expanded Materials and Methods” in online data supplements for measurement of intra-SR [Ca²⁺]). (Left) Representative fluo-5N images before and after addition of caffeine (20 mmol/L). (Right) Summarized data of the luminal [Ca²⁺] normalized to control condition of WT cardiomyocytes. N: the number of cells from 3–4 hearts.

Spectroscopic evidence that the inter-domain interaction is defective in R2474S/+ KI cardiomyocytes

To investigate whether the inter-domain interaction was defective in KI mice, we used the fluorescence quench technique that permits spectroscopic monitoring of the state of the inter-domain interaction (zipped or unzipped).¹⁵ The methylcoumarin acetamido (MCA) probe that has been attached to the critical domain would be inaccessible to a bulky fluorescence quencher (QSY^®-BSA conjugate) in a zipped configuration of the interacting domains, while it would become accessible to the quencher upon unzipping. As in our previous study with DPc10,⁹ site-specific intense fluorescence labeling of the RyR2 band was achieved using DPc10 as a carrier (Figure 7A, left lane), but there was no MCA labeling when DPc10-mut was used as a carrier (middle lane). As shown in the “cold-chase” experiment (right lane), excess unlabeled DPc10 (10 mmol/L) prevented DPc10-mediated MCA labeling. DPc10 increased the slope of the Stern-Volmer plot (K_Q), a measure of the extent of unzipping between the N-terminal (aa 1–600) and central (aa 2000–2500) domains in WT SR (Figure 7B, left). Addition of cAMP (1 μmol/L) had no effect (no unzipping) in WT SR (Figure 7B, left). In contrast, KI SR showed a high K_Q value (Figure 7B, right). Addition of cAMP (1 μmol/L) further increased the K_Q, which was comparable with the value of WT SR with added DPc10 (Figure 7B, right). Addition of DPc10 on the top of cAMP (1 μmol/L) produced no further increase in K_Q. Interestingly, dantrolene (1 μmol/L) partially reversed the K_Q that had been increased by cAMP (Figure 7B, right). Dantrolene also prevented the DPc10-induced unzipping in WT SR (Figure 7B, left). These K_Q values are summarized in Figure 7C.

Figure 7.

Spectroscopic monitoring of domain-domain interaction between the N-terminal (aa 1–600) and central (aa 2000–2500) domains in the RyR2. A, Site-directed fluorescence labeling of the RyR2 with MCA. MCA fluorescence labeling took place only when DPc10-SAED was used to mediate site-specific labeling (left lane). No MCA fluorescence was seen when DPc10-mut-SAED was used (middle lane) or when an excess concentration of DPc10 (10 mmol/L) was added during the labeling (‘cold chase’, right lane). B, Left: Stern-Volmer plots of the MCA fluorescence quenching data with QSY^®7-BSA of WT mice. Right: Comparison of the Stern-Volmer plots of the MCA fluorescence quenching data of R2474S/+ KI mice with those of WT mice. Note that the K_Q increased in the SR vesicles from KI to a similar level as that of DPc10-added WT SR vesicles. The values of K_Q are means±SE of 5–9 experiments using 2–3 SR preparations from 20–30 mice heart. C, Summarized data of K_Q shown in Figures 7B.

Arrhythmogenic characteristics of the membrane potential in R2474S/+ KI cardiomyocytes

Because altered membrane potential events represent an important landmark in arrhythmogenesis,¹⁶ we measured membrane potentials of WT and KI cardiomyocytes by recording the di-8-ANEPPS fluorescence while pacing in the presence of 30 nmol/L isoproterenol (Online Figure XI). WT cardiomyocytes showed no spontaneous after potential (Figure 8A) and no spontaneous Ca²⁺ transient (Figure 8B). However, KI cardiomyocytes showed spontaneous Ca²⁺ transients and spontaneous after potentials in response to isoproterenol (30 nmol/L) when we increased the pacing rate from 1 to 5 Hz (Figures 8A and 8B). Interestingly, the spontaneous after potentials and Ca²⁺ transients disappeared in the presence of dantrolene (1 μmol/L), which corrected the defective inter-domain interaction between the N-terminal and central domains (Figures. 8A and 8B).

Figure 8.

Provocation of triggered activity and spontaneous Ca²⁺ transient in R2474S/+ KI cardiomyocytes. Representative recording of di-8-ANEPPS fluorescence (A) and Ca²⁺ transients (B) before and after addition of isoproterenol (30 nmol/L) in intact WT and KI cardiomyocytes. Summarized data are shown at the bottom of each figure. After-potentials (indicated by arrows and shown in the insets) and spontaneous Ca²⁺ transients took place after pacing (5Hz) in KI cardiomyocytes. N: the number of hearts.

Discussion

Many point mutations have been found in RyR2 in patients with CPVT. Several pieces of biochemical evidence suggest that these mutations cause defective channel gating, leading to diastolic Ca²⁺ leakage.¹⁷ More direct evidence that a single point mutation in the RyR2 is the primary cause of channel dysfunctions in CPVT patients have been obtained by recent studies with a knock-in (KI) mouse model.^6–8 A single point mutation (R4496C) introduced in the RyR2 was found to cause bidirectional or polymorphic VT in the KI mice.⁶ In isolated cardiomyocytes from the R4496C KI mice, both delayed after depolarization (DAD) and triggered activity were induced on stimulation with isoproterenol. ¹⁸ A more recent study using the same model ¹⁹ showed that a dramatic increase in the Ca²⁺ sensitivity of the RyR2 channel resulted in the increased frequency of Ca²⁺ sparks and Ca²⁺ waves, which was further amplified by either isoproterenol or high pacing rates. Further studies^7,8 confirmed mutation-linked dysfunction of RyR2, namely spontaneous Ca²⁺ release events and DAD. These studies show a close inter-relationship between the single point mutation, increased Ca²⁺ sensitivity of the channel gating, and the resulting lethal arrhythmia. However, the underlying mechanism by which a single mutation in such a large molecule causes drastic effects on cardiac function has remained unclear.

Defective inter-domain interaction is the source mechanism of mutation-linked channel disorder

The most important new aspect of the present study is the finding that introduction of the R2474S CPVT mutation into the central domain of RyR2 induced a defective interaction between the central domain and the N-terminal domain, as predicted from the “domain switch hypothesis” (cf. Introduction), and this caused channel dysfunction similar to that of CPVT patients in KI mice. The three lines of evidence are consistent with this. First, DPc10 (aa 2460–2495), which contains the mutable R2474 residue and is known to interfere with normal inter-domain interaction between the N-terminal and central domains,⁹ reproduced the abnormal cellular Ca²⁺ events seen in the R2474S/+ KI mice (e.g., increased frequency of Ca²⁺ sparks) in an otherwise normal system (i.e., in cardiomyocytes isolated from WT mice). However, the addition of DPc10 to the (cAMP-treated) R2474S KI cardiomyocytes produced no further effect, suggesting that the defective inter-domain interaction (aberrant domain unzipping) had already taken place in the KI cardiomyocytes. Second, The R2474S mutation, introduced into the central domain of RyR2 of KI mice, did produce defective inter-domain interaction between the N-terminal and central domain (aberrant domain unzipping), as evidenced by the accessibility of the fluorescent probe MCA attached to the N-terminal domain to a high molecular weight fluorescence quencher QSY^®7-BSA was considerably higher in the KI RyR2 than the WT RyR2. Finally, dantrolene, which corrects aberrant domain unzipping, did suppress aberrant phenomena characteristic of CPVT KI mice, such as reduced threshold of luminal Ca²⁺ for channel activation, spontaneous Ca²⁺ sparks, and DAD.

PKA-dependent phosphorylation of the RyR2 at Ser2808 facilitates domain unzipping only in the CPVT mutant ryanodine receptor

An interesting new finding in the present study is that the threshold of luminal [Ca²⁺] for activation of Ca²⁺ sparks was much lower in R2474S/+ KI mice than in WT mice. In other words, the sensitivity of the RyR2 channel to activation by luminal [Ca²⁺] was increased in R2474S/+ KI mice. More importantly, we could reproduce this sensitized channel gating to luminal [Ca²⁺] that is characteristic of the R2474S/+ KI mice, in WT cardiomyocytes by adding DPc10 (Figure 5D). This provides further support for the notion that the aberrant channel gating in R2474S/+ KI mice is produced by defective inter-domain interaction between the N-terminal and central domains.

This study also showed that the level of PKA-dependent phosphorylation of the RyR2 at Ser2808 was virtually indistinguishable between KI and WT RyR2s (Figure 4B, C), yet PKA phosphorylation produced a much larger effect in increasing the frequency of Ca²⁺ sparks (SpF) in the KI cardiomyocytes than the WT myocytes (Figures 4A and 5A). This suggests that the CPVT mutation also sensitizes the channel to PKA phosphorylation-dependent activation. In the three-dimensional image of the RyR2, Ser2808, the site of PKA phosphorylation, has been localized in the vicinity of the boundary between the N-terminal (aa 1–600) and the central domains (aa 2000–2500).²⁰ Earlier cryo-electron microscopy single particle study of RyR1 ²¹ also showed that domain 5 (including the N-terminal domain) and domain 6 (including the central domain) at the clamp region are indeed in close apposition to each other in the resting state, whereas these domains become separated in the activated (channel-open) state (e.g., in the presence of cAMP and activating Ca²⁺). Thus, it is tempting to suggest that PKA phosphorylation at Ser2808 accelerates domain unzipping in the KI channel, where domain unzipping has already progressed because of a weakened inter-domain interaction caused by the CPVT mutation.

A new molecular mechanism for CPVT

Although recent studies using KI mouse models have demonstrated that CPVT is caused by mutation-linked dysregulation of intracellular Ca²⁺ events and membrane potential events (DAD and triggered activity), it is still unclear how a single mutation changes the conformational state of the RyR2, leading to leaky channel. We propose a new molecular mechanism underlying CPVT (Online Figure XI). In the normal channel, domain-domain interaction between the N-terminal (aa 1–600) and central (aa 2000–2500) domains is maintained in a zipped state, and thus, the channel is stabilized, preventing Ca²⁺ leakage and DAD at a physiological range of SR Ca²⁺ contents. In the mutant channel, the stabilized inter-domain interaction is disrupted, causing aberrant domain unzipping; domain unzipping is further aggravated by the PKA-phosphorylation of Ser2808, located at the boundary between the 2 domains at the clamp region.²⁰ In turn, the threshold of luminal [Ca²⁺] for channel activation is decreased (cf. ref. 17). Together, this results in SR Ca²⁺ leakage, DAD, and lethal arrhythmia. It should be noted, however, that DAD-triggered arrhythmia can also be induced by intracellular Ca²⁺ overload, for example, the toxic arrhythmogenic effects of the cardiac glycosides.²²

In contrast to a previous report, ⁸ the RyR2-FKBP12.6 link was not disrupted in R2474S/+ KI mice. In the present study, the level of protein kinase A (PKA)-dependent phosphorylation at Ser2808 of the RyR2 showed only a modest increase in response to 1 μmol/L cAMP (slightly less than half of maximum phosphorylation; see Online Figure II). This level of phosphorylation may not have been high enough to dissociate FKBP12.6 from RyR2. Nonetheless, the fact that such a modest increase in the PKA-phosphorylation at Ser2808 induced domain unzipping and the resulting Ca²⁺ leakage in KI mice, suggests that the defective inter-domain interaction within the RyR2 is the source mechanism of CPVT and that FKBP dissociation is not a direct cause of defective channel gating in CPVT. As we previously reported, ⁹ however, it is also true that FKBP dissociation induces domain unzipping, which may account for the destabilized channel gating in heart failure. Thus, either a CPVT-type mutation or FKBP dissociation in heart failure commonly induces domain unzipping as an independent trigger, resulting in aberrant Ca²⁺ release in diseased hearts.

The reduction in SR Ca²⁺ content in heart failure may be partly due to increased NCX function and the increased SR Ca²⁺ leakage (in addition to reduced SERCA function), causing contractile dysfunction as well as arrhythmia. We previously demonstrated that dantrolene corrects defective inter-domain interactions within the RyR2 in failing hearts, inhibits spontaneous Ca²⁺ leakage, and in turn improves cardiomyocyte function in failing hearts.¹⁴ In this study, we showed that dantrolene was equally effective in the CPVT-type mutated RyR2 as in failing hearts. This indicates that channel dysfunction in CPVT and heart failure are caused by a common mechanism, that is, defective inter-domain interaction within the RyR2.

According to a recent report by Lehnart et al,⁸ R2474S KI mice, which harbors the same mutation as those used in the present study, exhibited spontaneous generalized tonic-clonic seizures. In our KI mice, however, we did not observe spontaneous tonic-clonic seizures. We have no explanation for this difference. It may be ascribable to differences in the background of the mouse model. In generating the RyR2 KI mice, we used ES cells derived from C57BL/6J mouse (no difference in the background between ES cell line mice and KI mice), but Lehnart et al. used those from 129 mice, followed by backcrossing with C57BL/6J mice.

In this, study, we evaluated the phosphorylation status only at Ser2808. However, because beta-adrenergic stimulation has been reported to activate CaMKII, ^{11, 23} a further investigation is clearly needed to assess the role of other phosphorylation sites (e.g. Ser2814 and Ser2030) on CPVT-type channel disorder.

In conclusion, a single point mutation within the RyR2 sensitizes the RyR2 channel to activation by luminal [Ca²⁺] (i.e., a decreased threshold of luminal [Ca²⁺] for channel activation), and in turn induces spontaneous Ca²⁺ sparks and DAD, leading to CPVT. More importantly, this aberrant channel opening is primarily mediated through defective inter-domain interaction between the N-terminal (aa 1–600) and central (aa 2000–2500) domains.

Non-standard Abbreviations and Acronyms

CPVT

catecholaminergic polymorphic ventricular tachycardia

RyR

ryanodine receptor

WT

wild-type

KI

knock-in

PKA

protein kinase A

MH

malignant hyperthermia

CCD

central core disease

ARVC

arrhythmogenic right ventricular cardiomyopathy

SR

sarcoplasmic reticulum

VT

ventricular tachycardia

SpF

Ca²⁺ spark frequency

FWHM

full width at half maximum

FDHM

full duration at half maximum

CaMKII

Ca²⁺/Calmodulin-dependent protein kinase II

MCA

methylcoumarin acetamido

DAD

delayed after depolarization

SERCA

SR Ca²⁺-ATPase

PLB

phospholamban

CASQ

calsequestrin

NCX

Na⁺/Ca²⁺ exchanger