Sarcoplasmic Reticulum Ca²⁺ Cycling Protein Phosphorylation in a Physiologic Ca²⁺ Milieu Unleashes a High-Power, Rhythmic Ca²⁺ Clock in Ventricular Myocytes: Relevance to Arrhythmias and Bio-Pacemaker Design

Abstract

Basal phosphorylation of sarcoplasmic reticulum (SR) Ca²⁺ proteins is high in sinoatrial nodal cells (SANC), which generate partially synchronized, spontaneous, rhythmic, diastolic local Ca²⁺ releases (LCRs), but low in ventricular myocytes (VM), which exhibit rare diastolic, stochastic SR-generated Ca²⁺ sparks. We tested the hypothesis that in a physiologic Ca²⁺ milieu, and independent of increased Ca²⁺ influx, an increase in basal phosphorylation of SR Ca²⁺ cycling proteins will convert stochastic Ca²⁺ sparks into periodic, high-power Ca²⁺ signals of the type that drives SANC normal automaticity. We measured phosphorylation of SR-associated proteins, phospholamban (PLB) and ryanodine receptors (RyR), and spontaneous local Ca²⁺ release characteristics (LCR) in permeabilized single, rabbit VM in physiologic [Ca²⁺], prior to and during inhibition of protein phosphatase (PP) and phosphodiesterase (PDE), or addition of exogenous cAMP, or in the presence of an antibody (2D12), that specifically inhibits binding of the PLB to SERCA-2. In the absence of the aforementioned perturbations, VM could only generate stochastic local Ca²⁺ releases of low power and low amplitude, as assessed by confocal Ca²⁺ imaging and spectral analysis. When the kinetics of Ca²⁺ pumping into the SR were increased by an increase in PLB phosphorylation (via PDE and PP inhibition or addition of cAMP) or by 2D12, self-organized, “clock-like” local Ca²⁺ releases, partially synchronized in space and time (Ca²⁺ wavelets), emerged, and the ensemble of these rhythmic local Ca²⁺ wavelets generated a periodic high-amplitude Ca²⁺ signal. Thus, a Ca²⁺ clock is not specific to pacemaker cells, but can also be unleashed in VM when SR Ca²⁺ cycling increases and spontaneous local Ca²⁺ release becomes partially synchronized. This unleashed Ca²⁺ clock that emerges in a physiological Ca²⁺ milieu in VM has two faces, however: it can provoke ventricular arrhythmias; or if harnessed, can be an important feature of novel bio-pacemaker designs.

1. Introduction

Spontaneous, rare, stochastic local diastolic Ca²⁺ releases (“Ca²⁺ sparks”) [1] that occur in basal-state cardiac ventricular myocytes (VM) provide an important SR Ca²⁺ leak pathway [2]. β-adrenergic receptor stimulation (β-AR_S) of VM organizes those local diastolic Ca²⁺ releases into partially synchronized spontaneous, periodic diastolic Ca²⁺ signals (Ca²⁺ waves) that, unlike “Ca²⁺ sparks”, can be of sufficient amplitude to generate abnormal spontaneous diastolic after-depolarizations that can initiate spontaneous abnormal action potentials (AP_S) [3]. During β-AR_S, two distinct, but related, phosphorylation-dependent events occur: (i) an increase in Ca²⁺ influx into the cell, and (ii) increased Ca²⁺ pumping rate into and release from SR. An increase in intracellular Ca²⁺, due to an increase in Ca²⁺ influx effected by a β-AR_S-induced increase in phosphorylation of L-type Ca²⁺ channel subunits is thought to be the major mechanism involved in organization of local, stochastic Ca²⁺ signals into spontaneous, roughly periodic Ca²⁺ waves [4]. One viewpoint, however, is that, although β-AR_S initially increases Ca²⁺ influx, the steady-state cell Ca²⁺ load during β-AR_S does not increase (vs. that in the basal state), because Ca²⁺ efflux from the cell increases to match influx [5]. Sarcoplasmic reticulum (SR) Ca²⁺ cycling proteins, e.g. phospholamban (PLB) and ryanodine receptors (RyRs) also become phosphorylated during β-AR_S, and an increase in the phosphorylation state is associated with enhanced Ca²⁺ pumping into SR, and to changes in spontaneous activation of RyRs. A role for enhanced SR Ca²⁺ cycling in the organization of partially synchronized, roughly periodic spontaneous diastolic SR Ca²⁺ releases in VM, in the absence of Ca²⁺ overload, however, has not been directly demonstrated.

A clue that increased SR Ca²⁺ cycling in the absence of Ca²⁺ overload can indeed generate roughly periodic spontaneous local Ca²⁺ releases (referred as “LCRs”), however, has emerged from recent studies in sinoatrial nodal pacemaker cells (SANC), in which basal levels of phosphorylation of Ca²⁺ cycling proteins are well above those in basal VM in a physiologic Ca²⁺ milieu [6]. These studies in SANC, in which the surface membrane had been permeabilized, clearly demonstrated that an enhanced rate of SR Ca²⁺ cycling effected by increased basal phosphorylation of SR Ca²⁺ cycling proteins enables inherently stochastic, sub-sarcolemmal LCRs via RyRs to become organized into roughly periodic Ca²⁺ signals (Ca²⁺ wavelets), even when the ambient steady [Ca²⁺] is buffered constantly at physiologic levels [6, 7]. LCRs are Ca²⁺ wavelets, i.e. larger and more organized than Ca²⁺ sparks, but, unlike Ca²⁺ waves, propagate only locally for relatively short distances (3 to 7 μm). Since SR generated LCRs are roughly periodic, the SR in SANC has been dubbed a “Ca²⁺ clock” [8].

In SANC with intact sarcolemma, spontaneous, periodic Ca²⁺ wavelets during diastole generated by the SR Ca²⁺ clock are of sufficient amplitude to effect local membrane depolarization (via activation of Na⁺/Ca²⁺ exchanger) that are critically linked to the occurrence of spontaneous, rhythmic APs, i.e. normal automaticity of the cardiac impulse [8]. Since high basal levels of Ca²⁺ cycling protein phosphorylation in SANC organize stochastic Ca²⁺ releases into local wavelet-like rhythmic LCRs, i.e. Ca²⁺ clock, we hypothesized that suppression of basal SR Ca²⁺ cycling in VM linked to a suppression of basal phosphorylation of SR Ca²⁺ cycling, prevents the emergence of periodic, organized LCRs (i.e. prevents the emergence of a Ca²⁺ clock in VM); instead only stochastic, low-amplitude “Ca²⁺ sparks” occur. Specifically, we hypothesized that even in a physiologic Ca²⁺ milieu, when the basal SR Ca²⁺ cycling rate increases, e.g. either in response to an increase in SR Ca²⁺ protein phosphorylation in VM when protein phosphatase (PP) and phosphodiesterase (PDE) activities are inhibited, or when PLB- SERCA interaction is inhibited by a specific monoclonal antibody, spontaneous stochastic sparks will self-organize into synchronized, periodic LCRs, i.e., a “Ca²⁺ clock” will emerge in VM.

2. Methods

Spontaneous local Ca²⁺ release characteristics (LCR), the phosphorylation status of SR-associated proteins, PLB and RyRs in permeabilized rabbit VM bathed in 100 nM free [Ca²⁺], and cytosolic Ca²⁺ signal in electrically stimulated rabbit VM with intact sarcolemma were measured. Shortly, intact VM were permeabilized with 0.01% saponin. After washing out saponin, solution was exchanged to the recording solution that contained 0.03 mM fluo-4 pentapotassium salt, 0.114 mM CaCl₂ (free [Ca²⁺] ~ 100 nM), 100 mM C₄H₆NO₄K (DL-aspartic acid potassium salt), 25 mM KCl, 10 mM NaCl, 3 mM MgATP, 0.81 mM MgCl₂, 20 mM Hepes, 0.5 mM EGTA, 10 mM phosphocreatine, and creatine phosphokinase (5 U/ml), pH 7.2 [6]. The cytosolic free Ca²⁺ at given total Ca²⁺, Mg²⁺, ATP, and EGTA concentrations was calculated using a computer program (WinMAXC 2.50, Stanford University). A detailed description of all methods is available in the Online Data Supplementary.

Data were reported as mean ± SEM. A Student’s t test, or, when appropriate, one-way ANOVA, was applied to determine statistical significance of the differences. A P value < 0.05 was considered statistically significant.

3. Results

3.1. Phosphorylation of sarcoplasmic reticulum Ca²⁺ cycling proteins, PLB and RyRs increases in permeabilized VM when PP and PDE activities are inhibited

Inhibition of protein phosphatase (PP) by Calyculin A (CyA, 0.5 μM) or by CyA plus a broad spectrum PDE inhibitor IBMX (20 μM) markedly increased PLB phosphorylation at a protein kinase A (PKA)-specific Ser¹⁶ site, detected by Western blots (Fig. 1) and RyR phosphorylation at PKA-dependent Ser²⁸⁰⁹ site, detected by duo-immunolabeling (Fig. 2).

Fig. 1.

Enhancement of PLB phosphorylation at a protein kinase A (PKA)-specific Ser¹⁶ site detected by Western blots in response to PP and PP + PDE inhibition in permeabilized VM. (A) Representative Western blots. (B) Average data of phosphorylated PLB normalized to total PLB in response CyA (0.5 μM) or CyA + IBMX (20 μM) (n= 3 blots). *P < 0.05.

Fig. 2.

Enhancement of Ensemble RyR2 phosphorylation at Ser²⁸⁰⁹ detected by phospho-imaging of permeabilized VM in response to PDE or PP inhibition or to PP + PDE inhibition. (A) Average phosphorylation of RyR at Ser²⁸⁰⁹ by RyR duo-immunolabeling, in permeabilized VM in control (n=36) and in response to IBMX (20 μM, n=31), CyA (0.5 μM, n=32) or CyA + IBMX (20 μM, n=32). The primary antibody was omitted, and only the secondary antibodies were applied to the negative control (NC, n=27). The phosphorylation level was indexed by the average fluorescence density of phosphorylated RyR at Ser²⁸⁰⁹ normalized by the total RyR fluorescence density of a given cell; ***P < 0.001 vs. Control; ^#P< 0.05 vs. CyA; ^&&&P<0.001 vs. IBMX via one-way ANOVA. (B) Representative confocal images of permeabilized VM immunolabeled for both total RyR (red) and phosphorylated RyR at Ser²⁸⁰⁹ (green) in control, in response to 2 min incubation with CyA (0.5 μM) + IBMX (20 μM) and negative control.

3.2. Periodic, high-power Ca²⁺ signals emerge from stochastic Ca²⁺ sparks when phosphorylation of SR Ca²⁺ cycling proteins becomes increased in response to PP and PDE inhibition or exogenous cAMP

In a free [Ca²⁺] of 100 nM spontaneous Ca²⁺ sparks in VM are stochastic, non-periodic event of low power in the frequency domain, and of a low amplitude in the space-time domain (Control, Figs. 3A–D). When, in response to PP inhibition by CyA, PKA-dependent PLB phosphorylation is increased (Fig. 1) and the kinetics of SR Ca²⁺ cycling increase, multiple wavelet-like, rhythmic local Ca²⁺ oscillations, i.e. LCRs, emerge (CyA, Fig. 3A and B). When studied in the frequency domain by Fourier analysis, LCRs are synchronized at a dominant frequency of 2.5 Hz (Fig. 3B) and in the space-time domain of the confocal image resulted in high-amplitude individual LCRs Ca²⁺ signals (CyA, Fig. 3C) and summation of these individual Ca²⁺ signals produced a high-amplitude whole-cell (macroscopic) Ca²⁺ signal (ensemble of LCRs) (CyA, Fig. 3D). In other terms, a “Ca²⁺ clock” emerges in VM in a physiologic Ca²⁺ milieu. In the presence of CyA the addition of IBMX, a broad spectrum PDE inhibitor that increases cAMP, and leads to an increase in PKA-dependent phosphorylation [9] (Figs. 1 and 2), further increases the power of the partially synchronized Ca²⁺ signal in the frequency domain (CyA+IBMX, Fig. 3B) and this enhanced synchronization not only further amplified the space-time domain Ca²⁺ signal of individual LCR’s (CyA+IBMX, Fig. 3C), but also amplified the Ca²⁺ signal of the LCR ensemble by 8-fold over control (CyA+IBMX, Fig. 3D). On average, LCR periodicity in the frequency domain was observed in 77% and 86% of cells in response to inhibition of PP or PP plus PDE, respectively (Fig. S1A), and the dominant LCR period averaged 3.2 ±0.2 Hz (Fig. S1B).

Fig. 3.

Inhibition of PP and PP + PDE in permeabilized VM organizes stochastic sparks into synchronized, rhythmic, high power, spontaneous local Ca²⁺ releases (LCR’s/wavelets). (A) Representative confocal line-scan images of a permeabilized VM bathed in 100 nM free [Ca²⁺] in control and after incubation with CyA (0.5 μM) and CyA + IBMX (20 μM). (B) Respective fast Fourier transforms (FFT) of the rhythmic of Ca²⁺ oscillations recorded in A. (C, D) Average amplitudes of Ca²⁺ signals of individual LCR’s and of the LCR ensemble in control and in response to CyA and CyA + IBMX. *P < 0.05, ***P < 0.001 and ^##P < 0.01 vs. CyA, n = 4 – 7 cells for each data point.

We employed PKI, a specific peptide inhibitor of PKA activity, to ascertain a specific role for PKA-dependent phosphorylation in the emergence of rhythmic spontaneous local Ca²⁺ releases in the presence of CyA+PKI reversibly abolished LCR periodicity in the frequency domain (CyA+PKI, Figs. 4A, B and Figs. S1A, B) and reduced Ca²⁺ signal amplitude in the space-time domain (CyA+PKI, Figs. 4A–D). Thus, when phosphatase activity was inhibited by CyA, PKA-dependent protein phosphorylation was required for synchronization of stochastic Ca²⁺ sparks into partially synchronized, high-amplitude LCRs.

Fig. 4.

PKI, a specific PKA inhibitor peptide, reversibly abolished LCR (wavelets) periodicity and reduced the amplitude of Ca²⁺ signals affected by PP inhibition with CyA. (A) Representative confocal line-scan images in control and during incubation with CyA (0.5 μM) or CyA + PKI (15 μM), and after washing out PKI, with CyA (0.5 μM) still present. (B) Respective Fast Fourier Transforms (FFT) of the LCR’s recorded in A. (C, D) Average amplitudes of Ca²⁺ signals of individual LCR’s and of the LCR ensemble in control, in response to CyA, to CyA + PKI, and following wash out of PKI with CyA. *P < 0.05, **P < 0.01, ***P < 0.001vs. Control and ^###P < 0.001 vs. CyA+PKI, n = 4 – 6 cells for each data point.

Specific changes in the characteristics of individual local Ca²⁺ releases that become organized into periodic high-amplitude Ca²⁺ signals in the space-time domain are illustrated in Supplemental Figure S2. Note that the increase of space-time integral of each LCR Ca²⁺ signal results from increases in the average amplitude, width, duration and an increase in number of wavelet occurrences within the time window of observation (Fig. S2). This increase in amplitude of individual Ca²⁺ wavelets at a given locus may in part involve local recruitment of activation of neighboring RyRs which, in part may be attributed to Ca²⁺-induced Ca²⁺ release. That the number of sites at which Ca²⁺ wavelets occur triples compared to control (Fig. S2), however, indicates that in addition to Ca²⁺-induced Ca²⁺ release, RyR activation must become partially synchronized among remote sites (along a confocal line scan image), some of which were not visualized in control. To determine the cell-wide emergence of periodicity and synchronization of local Ca²⁺ signaling, we captured the local Ca²⁺ release in permeabilized VM in two-dimensional whole-cell images by Hamamatsu camera before and during PP inhibition with 0.1 μmol/L CyA (Fig. 5, Fig. S3, and movie 1-4). Supplemental Figure S3 illustrates how a phosphorylation-induced emergence of a powerful whole cell Ca²⁺ signal emerges from the ensemble of local Ca²⁺ oscillators following of CyA exposure. As exposure time in CyA increases, during which time a protein phosphorylation likely increases, LCRs evolve from rare small Ca²⁺ sparks in control (Fig. 5A and movie 1) into larger, powerful, and rhythmic Ca²⁺ oscillations over the entire cell (Figs. 5B–D and movie 2–4). In the frequency domain, partial synchronization of local spontaneous Ca²⁺ releases results in a Ca²⁺ signal as a sharp, high-amplitude peak in the power spectra (Fig. 5D).

Fig. 5.

Time dependent emergence of the LCR (wavelets) synchronization during protein phosphatase (PP) inhibition in a representative saponin-permeabilized VM, recorded in two dimensions by a high speed camera. (A-D) Whole cell 2D images and their power spectra recorded in control and during 3 minutes, 8 minutes and 12 minutes of CyA (0.1 μmol/L) exposure. Shown is a representative example from 5 cells tested. Panel A also shows the transmission light image of the permeabilized cell (middle panel).

Figure 6 illustrates, that following the addition of exogenous cAMP (3 μM), similar to PP and PDE inhibition, rhythmic “clock-like” partially synchronized LCRs in the frequency and space-time domains emerge from stochastic Ca²⁺ sparks. This occurred in 78% of cells tested (cAMP, Fig. S1A), the average size of the local Ca²⁺ release increased 1.6 fold, and the average duration increased 1.5 fold, and the number of individual releases occurring per unit time increased by 1.9 fold (Fig. S2A). The space-time Ca²⁺ signal of individual LCRs increased by 4.9 fold (Fig. 6C), and that the LCR ensemble increased by 7.7 fold (Fig. 6D).

Fig. 6.

Addition of cAMP organizes stochastic sparks into synchronized, rhythmic, high power, spontaneous local Ca²⁺ wavelets (LCR’s) in permeabilized VM. (A) Representative confocal line-scan images in control and in response to 3 μM cAMP. (B) Respective Fast Fourier Transforms (FFT), of the local Ca²⁺ oscillations recorded in A. (C, D) Average amplitude of Ca²⁺ signals of individual LCR’s and of the LCR ensembles in control and in response to cAMP. **P < 0.01, n = 9 cells.

3.3. A specific monoclonal antibody that inhibits the PLB-SERCA interaction mimics the effects of PLB and RyR phosphorylation on synchronization of local Ca²⁺ releases in VM

To establish a link between an increase in SR Ca²⁺ pumping and the emergence of rhythmic local Ca²⁺ wavelets when Ca²⁺ cycling protein phosphorylation is increased (Figs. 1, 2), we employed a specific anti-PLB monoclonal antibody, (2D12), that mimics the effect of PLB phosphorylation by inhibiting the PLB-SERCA-2 interaction, markedly increasing Ca²⁺ pumping into SR [10-12], without directly effecting the PLB phosphorylation. Indeed, similar to PDE and PP inhibition or cAMP application (Figs. 3-6), a 2-minute incubation with 2D12 (13.1μg/1ml) induced periodic, high-power rhythmic Ca²⁺ signals in the frequency domain (2D12, Figs. 7A, B) and high-amplitude ensemble local Ca²⁺ wavelets (2D12, Figs. 7C, D) in permeabilized VM. Periodicity emerged in 50% of cells tested (2D12, Fig. S1A), and the dominant frequency averaged at 2.6±0.17Hz, i.e. similar to interventions that increased SR Ca²⁺ cycling protein phosphorylation (Fig. S1B). In response to 2D12 and in response to any other perturbation employed in our study, the increase in the Ca²⁺ signal of LCR ensemble was highly correlated with the percent of cells that produced periodic LCRs (wavelets) (Fig. S5). The increased size, duration and number of local Ca²⁺ releases in response to 2D12 also resembled that of other perturbations that increase protein phosphorylation (Fig. S2). The Ca²⁺ signal of individual LCR’s in response to 2D12 increased by 4.6 fold (Fig. 7C) and that of the LCR ensemble increased by 6.4 fold over control (Fig. 7D), powerful, rhythmic Ca²⁺ signals that are strikingly similar to those induced by increased protein phosphorylation.

Fig. 7.

A monoclonal antibody against phospholamban (2D12) that inhibits SERCA2a-PLB interaction, like PLB phosphorylation, organizes stochastic sparks into synchronized, rhythmic, high power, spontaneous local Ca²⁺ wavelets (LCR’s) in permeabilized VM. (A) Representative confocal line-scan images in control and after 2 min application of 2D12 (13.1μg/1ml). (B) Respective Fast Fourier Transforms (FFT) of the local Ca²⁺ oscillations recorded in A. (C, D) Average amplitude of Ca²⁺ signals of individual LCR’s and of LCR ensembles in control and in response to 2D12. **P < 0.01, n = 6 cells.

3.4. How partially synchronized SR Ca²⁺ depletion in local microdomains in a physiologic Ca²⁺ milieu affects global SR Ca²⁺ load as assessed by caffeine in permeabilized VM

The Ca²⁺ release measured by a rapid caffeine application in permeabilized VM bathed in 100 nM [Ca²⁺] is the SR Ca²⁺ load determined by difference between the magnitudes of Ca²⁺ pumped into SR throughout the cell and Ca²⁺ lost from SR in microdomains by the ensemble of spontaneous local Ca²⁺ releases. When the magnitude of spontaneous local Ca²⁺ releases equals that of the Ca²⁺ influx into SR due to enhanced Ca²⁺ pumping throughout the cell, the amplitude of the caffeine-induced SR Ca²⁺ release will be the same as that in control, even though SR Ca²⁺ pumping had been markedly increased throughout of the cell. In this case it might appear as if perturbations that increase Ca²⁺ pumping into SR do not increase the SR Ca²⁺ load, but this would be an inaccurate conclusion, because the caffeine-induced Ca²⁺ release reports the sum SR Ca²⁺ load within local domains in which spontaneous Ca²⁺ releases had not just occurred prior to caffeine application, and in other domains in which partially synchronized spontaneous Ca²⁺ releases had just occurred.

The effect of gradations in the amplitude of spontaneous, partially synchronized local SR Ca²⁺ releases in response to a perturbation that increases SR Ca²⁺ loading on caffeine-induced Ca²⁺ release is illustrated in Figure 8. In response to a low concentration of 2D12 (1.31μg/1ml) the amplitude of the caffeine-induced Ca²⁺ release increases indicating, therefore, that concomitant increase in the local SR Ca²⁺ depletion resulting from partially synchronized local SR Ca²⁺ releases, is less (2D12, 1.31μg/1ml, Fig. 8D) than the effect of 2D12 to increase cell wide Ca²⁺ pumping into SR (2D12, 1.31μg/1ml, Fig. 8C). A 10 fold increase in 2D12 concentration (2D12, 13.1μg/1ml, Fig. 8D), that maximally increases Ca²⁺ pumping into SR (Fig. 9D in [11]) also markedly further increases spontaneous Ca²⁺ releases within the local microdomains. In this case, the net SR Ca²⁺ load detected by caffeine does not increase even though Ca²⁺ pumping into SR is markedly enhanced, indicating this increased Ca²⁺ pumped into SR is balanced by an enhanced synchronization and magnitude of local Ca²⁺ release fluxes throughout the cell (Fig. 5 and Figs. S3, S5). It is noteworthy in this regard, that further synchronization of spontaneous local Ca²⁺ releases, per se, is a mechanism that increases the summated spontaneous RyR release flux. Enhanced local spontaneous Ca²⁺ flux likely increases local RyR phosphorylation by activating CaMKII. Both an increase in RyR phosphorylation and disengagement PLB from SERCA, which increases the kinetics of Ca²⁺ pumping into SR, are mechanisms that underlie the extent of this augmentation of synchronization of local spontaneous Ca²⁺ releases.

Fig. 8.

The relationship of the magnitude of the summated Ca²⁺ signal of spontaneous partially synchronized local Ca²⁺ releases (LCR’s/wavelets) to the amplitude of the cell-wide caffeine-induced SR Ca²⁺ release. (A, B) Representative confocal images and caffeine-induced Ca²⁺ transients elicited by 20 mM caffeine in permeabilized VM in control conditions bathed in 100 nM [Ca²⁺] (left) and after 2 min treatment with 2D12 (1.31μg/1ml and 13.1μg/1ml) (right). (C, D) Average changes in the amplitudes of the caffeine-induced SR Ca²⁺ release (n=6 to 18 cells for each data point) and Ca²⁺ signal of the LCR ensembles (n= 6 cells for each data point) in permeabilized VM in control and after application of a low (1.31μg/1ml) or a high (13.1μg/1ml) concentration of 2D12. *P < 0.05, **P < 0.01.

Similar to the case of a high 2D12 concentration, the amplitude of Ca²⁺ released by caffeine did not, on average, significantly increase from control in response to cAMP or PDE/PP inhibition, indicating that the magnitude of the partially synchronized ensemble of LCRs within microdomains in permeabilized VM bathed in 100nM [Ca²⁺] equals that of the magnitude of increased Ca²⁺ pumped into SR (Fig. S4).

The results presented thus far have reported how synchronization of stochastic local Ca²⁺ sparks to produce periodic Ca²⁺ wavelets (LCR’s) is affected by perturbations that increase Ca²⁺ cycling through the SR in the context of a physiologic free [Ca²⁺]. Numerous prior studies in VM have indicated that an increase in the bathing [Ca²⁺], which causes cell and SR [Ca²⁺] loading to increase also induces spontaneous Ca²⁺ waves [13, 14].

4. Discussion

The present study demonstrates, for the first time, that similar to pacemaker cells, self-organized, partial synchronization of spontaneous local Ca²⁺ releases operates in the absence of Ca²⁺ overload in the form of partially synchronized local rhythmic Ca²⁺ wavelets (LCRs) that generate a powerful ensemble signal, observed as a sharp, high-power peak in the frequency domain, and a high-amplitude ensemble Ca²⁺ signal in the space-time domain.

In our previous work by Lyashkov et al. Circ Res. 2007 [15], we have compared the distribution of RyR in rabbit SANC and VM. We showed that RyR’s within SANC are located both beneath sarcolemma and in the interior of the cell, with the highest density beneath the sarcolemma. The localization of RyR’s in transverse bands spaced ~2 mm apart within the SANC interior resemble the sarcomere spacing in VM, but does not depend on the presence of T-tubules, as those are not present in SANC [15, 16].

The ability of permeabilized SANC, but not permeabilized VM to generate periodic clock-like Ca²⁺ releases, demonstrated in our previous work, has been attributed to an experimentally detected in SANC versus VM increase in the abundance of SERCA, reduced abundance of the SERCA inhibitor protein, PLB, and to increased Ca²⁺-dependent basal phosphorylation of PLB and RyR [6].

4.1. Emergent partial synchronization and periodicity of spontaneous local Ca²⁺ releases is due, in part at least, to increased Ca²⁺ pumping into SR

In the present study we show that in contrast to pacemaker cells, the spontaneous Ca²⁺ clock in VM does not normally operate in the basal state, but must be unleashed by factors that accelerate SR Ca²⁺ cycling, e.g. by inhibiting the mechanisms that restrain Ca²⁺ and cAMP-dependent SR protein phosphorylation (such as phosphatases and phosphodiesterases) or during β-ARs. A phosphorylation-dependent facilitation of restitutions distributed across many SR functions mechanisms that govern SR Ca²⁺ cycling kinetics distributed across different processes is a general mechanism that underlies the self-organization of low-amplitude, stochastic oscillations (Ca²⁺ sparks) into cell-wide (line scan) Ca²⁺ signals of substantial amplitude when phosphorylation of SR Ca²⁺ cycling proteins increases. These functions include: 1) an increase in the rate of SR Ca²⁺ pumping into SR; 2) an increase in kinetics of mechanisms that remove RyR inactivation following a Ca²⁺ release; 3) a reduction in the threshold required for spontaneous RyR activation (see below).

Cyclic AMP-mediated, protein kinase A (PKA)-dependent phosphorylation of phospholamban (PLB), an accessory protein of the SR Ca²⁺ pump (SERCA-2) causes PLB to disengage from SERCA-2 [17], enhancing the rate of Ca²⁺ pumping into SR. Detailed studies of regulation of the kinetics of Ca²⁺ pumping by SERCA2 indicate that removal of PLB regulation converts SERCA2A to a functionally oligomeric state with increased intersubunit free energy exchange [18-20]. To demonstrate that effect of PLB phosphorylation induced by PP and PDE inhibition increases kinetics Ca²⁺ pumping into SR and that this effect is linked to partial synchronization of local Ca²⁺ releases from SR, we employed a non-phosphorylation-dependent mechanism, i.e. a specific antibody, 2D12, that inhibits PLB-SERCA interaction [11] and increases Ca²⁺ pumping into SR. Indeed, 2D12 reproduces the synchronizing effects of PP and PDE inhibition on spontaneous Ca²⁺ release. While this effect of selective inhibition of the association of PLB and SERCA2 by 2D12 initiates self-organized emergence of high amplitude, partially synchronized spontaneous local SR Ca²⁺ releases, it cannot be interpreted to indicate that an increased rate of Ca²⁺ pumping into SR is sufficient to fully account for this phenomenon. Indeed, the local high [Ca²⁺] in the vicinity of RyRs generated by the high-amplitude LCRs occurring in the contest of an increase in local SR Ca²⁺ load due to increase in Ca²⁺ pumping into SR can affect Ca²⁺-dependent RyR phosphorylation, via e.g. calmodulin kinase II (CAMKII) [21], which in VM also increases when PDE and PP are inhibited [22, 23]. Although some studies provide evidence that RyR phosphorylation by PKA in VM modulates SR luminal Ca²⁺ sensitivity [24], how RyR phosphorylation by CAMKII affects RyR gating or activation threshold, however, remains controversial [25-27].

4.2. Partial synchronization of phases of release and restitution of SR Ca²⁺ cycling within and among cell loci (microdomains) create a rhythmic clock in VM

The emergence of a macroscopic Ca²⁺ oscillator (i.e. whole-cell Ca²⁺ clock) in the present study involves the emergence of partial synchronization of local Ca²⁺ releases; i.e. Ca²⁺ oscillations having nearly the same period. This permits cooperative (and partially synchronized) operation of an ensemble of local Ca²⁺ oscillators in cell areas that are remote from each other. This type of synchronization mechanism is conceptually similar to that reported by Kort et al. in 1985 [28] and Stern et al. [29], who numerically modeled the emergence of powerful periodic macroscopic Ca²⁺ signal (whole cell signal in our case) on the basis of summation of local (microscopic) intracellular Ca²⁺ oscillators whose periods become synchronized around same value. This emergent macroscopic signal of VM in the present study is conceptually similar to the powerful Ca²⁺ signals in SANC when the phases of the intracellular Ca²⁺ clock synchronized by the last action potential prior to a voltage clamp at the maximum diastolic potential remain partially synchronized during the voltage clamps (e.g. Fig. 3 in [30]).

4.3. Relevance of synchronizing effects of phosphorylation-mediated SR Ca²⁺ cycling in permeabilized VM in the present study to electrically paced cells with intact sarcolemmal function during β-ARs

β-ARs during external pacing of intact VM engages the same mechanisms used in our experiments to augment Ca²⁺ cycling protein phosphorylation. β-AR_S (1) suppresses protein phosphatase activity, via PKA-dependent phosphorylation of the phosphatase inhibitor I-1[31], as does PP inhibition in permeabilized VM in the present study (Figs 2, 3D and Figs. S2, S3, S4); (2) increases PLB phosphorylation via increased activation of protein kinase A, by activating adenylyl cyclase resulting in an increase in cAMP production [32], as does PDE plus PP inhibition (Fig. 1) or application of cAMP to permeabilized VM in the present study (Fig. 6 and Figs. S2, S4). The increase in cAMP-mediated, PKA-dependent phosphorylation of PLB during β-AR_s in intact VM permits Ca²⁺ to be pumped into the SR at a greater rate [33], as does disengagement PLB from SERCA2 in permeabilized VM by the 2D12 antibody (Figs. 7, 8 and Fig. S2).

During βAR_s the robust Ca²⁺ “clock-like” behavior of SR Ca²⁺ cycling in VM demonstrated by the present results has two faces: when regulated normally, it not only prepares the myocardium to generate a strong organized Ca²⁺ release in response to APs occurring at a given frequency; but when it becomes dysregulated, it has the potential to trigger spontaneous abnormal APs that can initiate life-threatening arrhythmias.

4.3.1. The “upside” of β-AR_s is to optimally synchronize local RyR activation in response to an AP

During β-AR_s rhythmic APs occur at an increased rate, requiring that restitution of SR Ca²⁺ cycling processes becomes accelerated. The effect of phosphorylation of SR Ca²⁺ cycling proteins demonstrated in permeabilized cells in the present study would be expected to poise RyRs to activate at sooner times and with increased synchrony in response to subsequent single L-type channel activation, which is also enhanced by phosphorylation of its subunits during β-AR_S [34]. The net result of phosphorylation-dependent synchronization of both single L-type Ca²⁺ channels and SR Ca²⁺ cycling within and among Ca²⁺ release units of a single I_CaL-SR junction, and among I_CaL-SR junctions cell-wide in response to an AP entrains the periodicity of VM SR Ca²⁺ cycling to ensure optimal synchronized RyR activation in response to that AP [33]. In other words, the SR Ca²⁺ clock and L-type channel gating clock mechanisms become more synchronized during β-AR_S to generate an AP-induced cytosolic Ca²⁺ signal of increased amplitude that elicits a contraction of not increased amplitude, but of increased synchrony as well [33].

4.3.2. The “downside of β-AR_S to synchronize local spontaneous diastolic Ca²⁺ releases

During β-AR_S, the periods of spontaneous local Ca²⁺ releases of the SR Ca²⁺ clock are partially synchronized in part by phosphorylation of its proteins, and in part by the relatively cell-wide homogeneous SR Ca²⁺ depletion induced by the prior AP. During β-AR_S regularly occurring APs arrive at the ventricular myocardium at shorter intervals and this faster heart rate entrains the SR Ca²⁺ clock to AP firing rate, i.e., the SR Ca²⁺ clock ticks faster. When the external stimulation rate during β-AR_S is abruptly lowered, the entrained the SR clock generates partially synchronized Ca²⁺ releases spontaneously during diastole at or shortly following the time at which the next regular AP at the higher pacing rate was due. Supplemental Figure 6A illustrates the emergence of highly organized spontaneous diastolic Ca²⁺ releases when the Ca²⁺ clock ticks faster than the external pacing rate when the external pacing rate is abruptly lowered from 2Hz to 0.5 Hz. The periods of spontaneous Ca²⁺ release (measured as the time of their occurrence following the prior AP induced makes systolic Ca²⁺ transient ranged from 0.9 second up to 1.57 second (Fig. S6A). In the continued presence of β-AR_s, increasing external pacing frequency to 3Hz, with a frequency less than Ca²⁺ clocks spontaneous Ca²⁺ release overdrives the spontaneous SR Ca²⁺ clock (Fig. S6A (iii)).

It is widely recognized that spontaneous diastolic Ca²⁺ releases can generate spontaneous diastolic depolarization, which, in single myocytes, are capable of generating spontaneous abnormal APs [3]. But it is also widely recognized that in ventricular myocardium spontaneous diastolic depolarization of the surface membrane of a single VM source in which a DAD arises would dissipate into the sink of surrounding VM, and if therefore, an AP occurred within the single VM, it would not likely excite adjacent cells, due to the well-known large source-sink current mismatch [35].

Based upon the phosphorylation induced emergence of synchronized spontaneous local SR Ca²⁺ releases as demonstrated in the present study, we propose that during β-AR_S local Ca²⁺ releases not only become synchronized within a given VM, but also become partially synchronized among VM residing within local “neighborhoods” of ventricular tissue, and that partial synchronization of periods of spontaneous local diastolic Ca²⁺ releases among cells is a mechanism to explain how spontaneous diastolic Ca²⁺ oscillations within cells can generate spontaneous APs that trigger arrhythmias that arise in ventricular tissue. Supplemental Figure 6B demonstrates spontaneous Ca²⁺ oscillations occurring in the ten different VM residing together in a “neighborhood” of myocardium. Following a pause in the external stimulation during β-AR_s spontaneous Ca²⁺ releases occur with roughly, i.e. not exactly, the same period. Note, the summated spontaneous Ca²⁺ release signal among the ten cells generates a high-amplitude spontaneous Ca²⁺ signal (red trace in Fig. S6B). This phosphorylation-dependent mechanism to synchronize emergent Ca²⁺ clocks among different myocytes during β-ARs can explain how spontaneous Ca²⁺ oscillations arising within individual VM can overcome the “source-sink” safety feature of ventricular myocardium (that protects the heart under normal conditions against an eventual afterpotential of single cell or a few cell) [35, 36]. This mechanism of abnormal automaticity originating from partially synchronized spontaneous diastolic Ca²⁺ releases among VM within ventricular tissue would be likely to occur during a bradycardic pause or between regular impulses emanating from the SA node at inadequately low rates that are unable to override the intrinsic “Ca²⁺ clock”. It is important to note that this clock-like synchronization mechanism to trigger a focal tissue excitation does not directly require or involve cell-to-cell interactions, and operates without dependence on cell density. In other words, the “source-sink” safety feature of ventricular myocardium (that protects the heart under normal conditions against an eventual after-depolarization of a single cell or a few cells) to generate an AP, can be overcome, because not a single VM, but a large number of VM within the “neighborhood” are the “source” in this instance.

4.4. Relevance of Present Results to Bio-pacemaker Design

Ca²⁺ clocks within SANC are persistently activated and driven by cAMP-activated PKA- dependent phosphorylation, due to presence and activation of Ca²⁺-activated adenylyl cyclases [37]. Novel designs of genetically engineered biological pacemakers, conceptually, could feature emergent Ca²⁺ clocks within VM and among other cardiac cells that are uncoupled from the impulses generated by SA node but are sufficiently coupled to activation of the membrane clock [38]. A recent study numerically tested hundreds of thousands of different bio-pacemaker designs, concluded that a Ca²⁺ clock is required for not only robust, but also flexible pacemaker function [39]. One specific biological pacemaker design (that mimics nature’s design) [37, 40] employs genetically controlled overexpression of Ca²⁺-activated adenylyl cyclases that drives the phosphorylation-driven automaticity. Overexpression of a Ca²⁺-activated adenylyl cyclase has been shown to be sufficient, in the absence of “funny-current” activation, to produce biological pacemaking for seven days in experimental dogs [41]. Another new and interesting approach to create natural coupled-clock functionality within a bio-pacemaker is to reactivate the genetic program of the SA node within VM. A recent study, successfully demonstrated a proof principle for this approach, using tbx18-infected VM that exhibit that “Ca²⁺ clock mechanisms of automaticity” pacing the heart for up to 8 weeks [42].

Highlights.

• A Ca²⁺ clock emerges in permeabilized ventricular myocytes at normal [Ca^2+]i

• The Ca²⁺ clock has phosphorylation-dependent mechanism

• The Ca²⁺ clock is manifested by rhythmic local Ca²⁺ releases as in pacemaker cells

• The Ca²⁺ clock can be activated by PP and PDE inhibition

• The Ca²⁺ clock can be arrhythmogenic and also a candidate for biopacemaker design

Abbreviations

AP

Action potential

β-ARs

β-adrenergic receptor stimulation

Ca²⁺

Calcium

IBMX

Isobutyl-1-methylxanthine

LCR

Local Ca²⁺ releases

PDE

Phosphodiesterase

PKA

Protein kinase A

PLB

Phospholamban

PP

Protein phosphatases

RyRs

Ryanodine receptors

SANC

Sinoatrial node cells

SERCA-2

SR Ca²⁺ pump

SR

Sarcoplasmic reticulum

VM

Ventricular myocytes