Research ArticleCell Biology

Ca2+-Dependent Phosphorylation of Ca2+ Cycling Proteins Generates Robust Rhythmic Local Ca2+ Releases in Cardiac Pacemaker Cells

See allHide authors and affiliations

Science Signaling  29 Jan 2013:
Vol. 6, Issue 260, pp. ra6
DOI: 10.1126/scisignal.2003391

Abstract

The spontaneous beating of the heart is governed by spontaneous firing of sinoatrial node cells, which generate action potentials due to spontaneous depolarization of the membrane potential, or diastolic depolarization. The spontaneous diastolic depolarization rate is determined by spontaneous local submembrane Ca2+ releases through ryanodine receptors (RyRs). We sought to identify specific mechanisms of intrinsic Ca2+ cycling by which sinoatrial node cells, but not ventricular myocytes, generate robust, rhythmic local Ca2+ releases. At similar physiological intracellular Ca2+ concentrations, local Ca2+ releases were large and rhythmic in permeabilized sinoatrial node cells but small and random in permeabilized ventricular myocytes. Furthermore, sinoatrial node cells spontaneously released more Ca2+ from the sarcoplasmic reticulum than did ventricular myocytes, despite comparable sarcoplasmic reticulum Ca2+ content in both cell types. This ability of sinoatrial node cells to generate larger and rhythmic local Ca2+ releases was associated with increased abundance of sarcoplasmic reticulum Ca2+-ATPase (SERCA), reduced abundance of the SERCA inhibitor phospholamban, and increased Ca2+-dependent phosphorylation of phospholamban and RyR. The increased phosphorylation of RyR in sinoatrial node cells may facilitate Ca2+ release from the sarcoplasmic reticulum, whereas Ca2+-dependent increase in phosphorylation of phospholamban relieves its inhibition of SERCA, augmenting the pumping rate of Ca2+ required to support robust, rhythmic local Ca2+ releases. The differences in Ca2+ cycling between sinoatrial node cells and ventricular myocytes provide insights into the regulation of intracellular Ca2+ cycling that drives the automaticity of sinoatrial node cells.

Introduction

The heart constantly supplies blood to the body by contracting 90,000 beats per day, and more than 2 billion times during a human life span. The heartbeat is initiated in the sinoatrial node, the primary physiological pacemaker of the heart in which cells constantly and spontaneously generate action potentials that govern the rate of contraction of the entire heart. In humans, Ca2+ release through ryanodine receptors (RyRs), which are sarcoplasmic reticulum (SR) Ca2+ release channels, has a key role in cardiac pacemaking. Mutations of RyRs are associated with a deterioration of human sinoatrial node function, atrial fibrillation, and atrial standstill (1). The elucidation of the features of Ca2+ cycling in cardiac pacemaker cells not only enhances our understanding of how cells of the sinoatrial node function but also provides new insights into therapies for various types of sinus node dysfunctions (2).

Spontaneous activity of sinoatrial node cells, which distinguishes them from ventricular myocytes, originates from spontaneous diastolic depolarization of the cell surface membrane, which gradually grows in amplitude during diastole, driving the membrane potential to the threshold to fire a spontaneous action potential (3). Sarcolemmal ion channel currents are critical in generating diastolic depolarization, including a hyperpolarization-activated current, If, a delayed rectifier potassium current, IK, and L- and T-type Ca2+ currents (3).

Spontaneous subsarcolemmal local Ca2+ releases through RyRs are also critical for the generation of diastolic depolarization (4). Local Ca2+ releases in sinoatrial node cells occur during diastolic depolarization, preceding the next action potential upstroke, and activate an inward Na+-Ca2+ exchange current, which exponentially increases terminal diastolic depolarization (4, 5). The local Ca2+ release period, or the time elapsing from the prior action potential triggered SR Ca2+ transient to local Ca2+ release occurrence, regulates the spontaneous beating rate of sinoatrial node cells by delineating the timing of Na+-Ca2+ exchange activation and thus prompting the next action potential (6). Rhythmic local Ca2+ releases in sinoatrial node cells reflect rhythmic Ca2+ cycling by the SR, which has been dubbed the “Ca2+ clock.” Specifically, action potential–induced Ca2+ depletion of the SR and RyR inactivation are followed by refilling of the SR Ca2+ load and recovery of RyR activation. The local Ca2+ release period critically depends on both the recovery of RyRs from inactivation and the rate of Ca2+ replenishment of the SR, which is controlled by the SR Ca2+-ATPase (SERCA) (7). Spontaneous Ca2+ release events originating from RyRs called “Ca2+ sparks” have initially been discovered in ventricular myocytes as discrete regions of increased fluorescence produced by bursts of Ca2+ (8). In contrast to spontaneous local Ca2+ releases in sinoatrial node cells, Ca2+ sparks in ventricular myocytes are relatively small, occur stochastically, and produce small local depletions of SR Ca2+, which is thought to be rapidly replenished through diffusion from neighboring, extensively interconnected regions in the SR, rather than by reuptake of Ca2+ by the SR (9, 10). Specific mechanisms that allow sinoatrial node cells, but not ventricular myocytes, to generate large and rhythmic local Ca2+ releases through RyRs in the basal state remain unclear. The cardiac SR oscillates Ca2+ through the same proteins in rabbit sinoatrial node cells and ventricular myocytes, SERCA and RyRs (11). The kinetics of Ca2+ pumping in the SR is primarily defined by the activity of SERCA. Mice with a cardiomyocyte-specific disruption of the SERCA gene and ~50% reduced abundance of SERCA show a twofold reduction in Ca2+ wave generation in ventricular myocytes compared with control mice (12). The activity of SERCA is inhibited by phospholamban (PLB), which in the dephosphorylated state binds to SERCA (13). When PLB is phosphorylated by protein kinase A (PKA) or Ca2+/calmodulin-dependent kinase II (CaMKII), it dissociates from SERCA, relieving its inhibition and thus enhancing the Ca2+ pumping rate of SERCA into the SR (13). Treating permeabilized ventricular myocytes with an antibody that recognizes PLB and disrupts its interaction with SERCA disrupts PLB-dependent suppression of SERCA activity (14) and increases the number and size of Ca2+ sparks (15).

The distribution of RyRs, which are Ca2+ release channels of the SR, is similar in sinoatrial node cells and ventricular myocytes, with RyRs located both beneath sarcolemma and in the interior of the cell (11, 16, 17). Within sinoatrial node cells, RyRs are localized in transverse bands spaced ~2 μm apart (17), an interval similar to the sarcomere spacing in ventricular myocytes. The highest density of RyR is found beneath the sarcolemma of sinoatrial node cells, which exceeds that in ventricular myocytes by almost threefold (11).

Considering the different sets of surface membrane ion channels in sinoatrial node cells and ventricular myocytes and the regulation of many of these channels by Ca2+ (3), intracellular Ca2+ cycling cannot be explicitly compared in intact sinoatrial node cells and ventricular myocytes. Cell surface membrane permeabilization has been used to study intracellular Ca2+ cycling in cardiac cells and has revealed the mechanism of Ca2+-induced release of Ca2+ (CICR) from the SR in cardiac cells (18). When the cell surface membrane is permeabilized to remove membrane currents, Ca2+ cycling by the SR becomes “free running” and is controlled mostly by the concentration of free cytosolic Ca2+ and the kinetics of Ca2+ pumping into and release from the SR.

We assessed Ca2+ cycling in permeabilized sinoatrial node cells and ventricular myocytes under controlled conditions and similar free cytosolic Ca2+ concentrations to determine the differences in subsarcolemmal Ca2+ releases in sinoatrial node cells and ventricular myocytes at a steady concentration of cytosolic Ca2+, how variations of cytosolic Ca2+ modulate spontaneous Ca2+ releases in the two cell types, and the Ca2+ cycling mechanisms that differ in pacemaker cells and ventricular myocytes.

Our results showed that at the same physiological cytosolic Ca2+ concentration, sinoatrial node cells cycled Ca2+ beneath the sarcolemma more efficiently than ventricular myocytes. Specifically, sinoatrial node cells could sustain larger and more rhythmic spontaneous SR Ca2+ releases than ventricular myocytes at similar amounts of Ca2+ in the SR, which correlated with increased abundance of SERCA and reduced abundance of PLB. Moreover, in response to increased concentrations of cytosolic Ca2+, phosphorylation of PLB and RyR by PKA and CaMKII was increased to a greater extent in sinoatrial node cells than in ventricular myocytes. We speculate that the resulting phosphorylation-dependent modulation of Ca2+ pumping in and release from the SR in sinoatrial node cells permitted robust, controlled, spontaneous Ca2+ releases by RyRs that were linked to the pacemaker nature of sinoatrial node cells.

Results

Spontaneous local Ca2+ releases were more robust and periodic in sinoatrial node cells than in ventricular myocytes

Representative subsarcolemmal line-scan images demonstrated that at the same free cytosolic Ca2+ concentration, spontaneous local Ca2+ releases in permeabilized sinoatrial node cells were comparatively large clustered events, whereas spontaneous Ca2+ releases in ventricular myocytes (Ca2+ sparks) were significantly smaller (Fig. 1, A to D, and fig. S1A). Because the number of Ca2+ releases within a given time and space does not adequately reflect the total amount of Ca2+ released from the SR by either cell type, we estimated the total Ca2+ signal mass released by sinoatrial node cells or ventricular myocytes by integrating signal masses of all spontaneous Ca2+ releases within a given time and space of the line-scan image (6). Sinoatrial node cells and ventricular myocytes released comparable total Ca2+ signal masses at relatively low cytosolic Ca2+ concentrations (50 to 100 nM). When the cytosolic Ca2+ concentration was increased above 150 nM, the total Ca2+ signal mass released by sinoatrial node cells was about two- to threefold larger than that released by ventricular myocytes (Fig. 1E). Because diastolic free Ca2+ concentrations are between ~160 and 250 nM in intact, spontaneously beating rabbit sinoatrial node cells and paced ventricular myocytes (6, 16, 19), these concentrations of free cytosolic Ca2+ in permeabilized cells are physiologically relevant.

Fig. 1

Regulation of spontaneous Ca2+ releases by cytosolic Ca2+ differed in permeabilized sinoatrial node cells (SANC) and ventricular myocytes (VM). (A) Representative images and Ca2+ waveforms from bands (indicated by arrows) of SANC exposed to different concentrations of cytosolic free Ca2+. (B) FFT of Ca2+ waveforms in (A). (C) Representative images and Ca2+ waveforms of VM bathed at different cytosolic Ca2+ concentrations. (D) FFT of the Ca2+ waveforms from bands indicated by the color-matched arrows in (C). (E) Comparison of total Ca2+ signal mass released by either SANC (n = 9 to 54 cells for each data point) or VM (n = 5 to 12 cells for each data point) at different cytosolic Ca2+ concentrations. (F) Relative number of SANC and VM (% of total) generating periodic spontaneous Ca2+ releases. The number of cells with periodic releases is shown as a fraction of the total number of cells studied at each cytosolic Ca2+ concentration. Logistic regression analysis demonstrated a significant difference between the curves for SANC and VM (P < 0.002). The average FFT frequency of spontaneous Ca2+ releases in SANC or VM was ~3 Hz at different free cytosolic Ca2+ concentrations. *P < 0.05.

Subsarcolemmal local Ca2+ releases in sinoatrial node cells were also rhythmic (Fig. 1A), whereas sparks in ventricular myocytes occurred randomly (Fig. 1C). To study the periodicity of spontaneous Ca2+ releases in sinoatrial node cells or ventricular myocytes, we used fast Fourier transform (FFT) (20), which transforms signals in the time domains (Fig. 1, A and C) into their frequency domain representations (Fig. 1, B and D). For comparable analysis, Ca2+ waveforms were calculated from consecutive 2- to 3-μm strips of line-scan image, which was the average spark size. A single peak in the frequency domain in sinoatrial node cells (Fig. 1B) indicated that spontaneous Ca2+ releases appeared repeatedly after about the same time interval and thus were periodic. In contrast, multiple uniformly distributed peaks in the frequency domain in ventricular myocytes (Fig. 1D) indicated a lack of periodicity in spontaneous Ca2+ releases. Sinoatrial node cells, but not ventricular myocytes, generated periodic spontaneous Ca2+ releases at relatively low cytosolic Ca2+ concentrations (50 to 150 nM) (Fig. 1, A and F). Increasing the cytosolic Ca2+ concentration resulted in increases in the Ca2+ release size (fig. S1A) and the number of cells with periodic Ca2+ releases in both cell types (Fig. 1F and fig. S2). Ca2+ releases in ventricular myocytes became periodic when their size reached ~4 μm (fig. S2), and the size of spontaneous Ca2+ releases correlated with the number of cells generating periodic Ca2+ releases (fig. S2).

Amplified Ca2+ release from the SR in sinoatrial node cells was not attributable to increased Ca2+ content in the SR

We compared the Ca2+ load of the SR in permeabilized sinoatrial node cells and ventricular myocytes by applying a rapid pulse of caffeine, which binds to RyRs and causes them to open, thereby emptying Ca2+ stores of the SR (16). The caffeine-releasable Ca2+ content in the SR was the same in both cell types over a broad range of physiologically relevant free cytosolic Ca2+ concentrations (50 to 250 nM) (Fig. 2, A to C). Thus, sinoatrial node cells could sustain larger, rhythmic spontaneous Ca2+ releases through RyRs at similar Ca2+ loads in the SR as ventricular myocytes. When cytosolic Ca2+ was further increased to 300 nM, spontaneous local Ca2+ releases in sinoatrial node cells disappeared (fig. S3A) as SR Ca2+ load decreased (Fig. 2C). In contrast, increasing the cytosolic Ca2+ concentration to 300 nM or greater increased the Ca2+ content of the SR in ventricular myocytes (Fig. 2C), leading to the appearance of rhythmic Ca2+ wavelets (figs. S2 and S3).

Fig. 2

The Ca2+ content in the SR remained comparable in permeabilized sinoatrial node cells (SANC) and ventricular myocytes (VM) at physiological cytosolic Ca2+ concentrations. (A) Caffeine-induced Ca2+ transients in representative SANC. (B) VM at different cytosolic Ca2+ concentrations. The initial rapid component of the caffeine-induced Ca2+ transient (indicated by arrows) was analyzed. (C) Average total SR Ca2+ content in SANC (n = 9 to 31 cells for each data point) and VM (n = 6 to 11 cells for each data point). *P < 0.05.

Increased Ca2+ releases through RyRs in sinoatrial node cells in the absence of detectable depletion of Ca2+ in the SR suggested that Ca2+ pumping into the SR through SERCA was increased. Western blots of SERCA in tissue homogenates demonstrated that the abundance of SERCA in rabbit sinoatrial node and atrium exceeded that in ventricle by ~1.5- and ~2-fold, respectively (Fig. 3A). The Western blot analysis was repeated in isolated cells to exclude the possibility of contamination from other cell types and confirmed that the abundance of SERCA in sinoatrial node cells surpassed that in ventricular myocytes by ~1.5-fold (Fig. 3B). In addition, the abundance of PLB in sinoatrial node cells was ~2-fold less than that in ventricular myocytes (Fig. 3C), suggesting that inhibition of SERCA by PLB could be lower in sinoatrial node cells than in ventricular myocytes.

Fig. 3

The SR Ca2+ pump SERCA was more abundant and the SERCA inhibitor PLB was less abundant in sinoatrial node cells (SANC) than in ventricular myocytes (VM). (A) Top, representative Western blots of SERCA in right atrium, sinoatrial node (SAN), and ventricular tissues. Bottom, average data (n = 3 blots) were normalized and shown as percent of the amount of SERCA protein in ventricular tissue. (B) Representative Western blots of SERCA in SANC and VM. Bottom, average data (n = 5 blots) are shown as percent of the amount of SERCA protein in VM. (C) Top, representative Western blot of total PLB in SANC and VM. Bottom, average data (n = 6 blots) are shown as percent of the amount of PLB in VM. *P < 0.05.

Increased cytosolic Ca2+ concentration resulted in increased phosphorylation of PLB in cardiac pacemaker cells

Phosphorylation of PLB disengages it from SERCA, thereby relieving the inhibition of SERCA and increasing its Ca2+ pumping rate (13, 16). We assessed the phosphorylation status of PLB at Ser16, a PKA targeted site. Increased cytosolic Ca2+ concentrations might stimulate Ca2+-activated adenylyl cyclases in sinoatrial node cells (21, 22) and increase cyclic adenosine 3′,5′-monophosphate (cAMP) concentrations and cAMP-mediated PKA-dependent phosphorylation (23). Indeed, an increase in the cytosolic Ca2+ concentration resulted in increases in both the cAMP concentration (Fig. 4A) and the phosphorylation of Ser16 in PLB (Fig. 4B) in permeabilized sinoatrial node cells but did not alter the phosphorylation of Ser16 in PLB in ventricular myocytes (Fig. 4C).

Fig. 4

Ca2+ dependence of PKA-mediated phosphorylation of PLB at Ser16 differed in permeabilized sinoatrial node cells (SANC) and ventricular myocytes (VM). (A) An increase in the cytosolic Ca2+ concentration was associated with increased cAMP concentration in permeabilized SANC (n = 8 rabbits). (B and C) Top, representative Western blots of phosphorylated and total PLB in SANC and VM exposed to different cytosolic Ca2+ concentrations. Bottom, quantification of PLB phosphorylation in SANC (n = 5 blots) and VM (n = 4 blots). *P < 0.05.

Ca2+ also activates CaMKII (24, 25), which phosphorylates PLB at Thr17 (13, 16) and RyR (26, 27). An increase in cytosolic Ca2+ concentration was associated with a significant increase in the phosphorylation of PLB at Thr17 in sinoatrial node cells (fig. S4), but not in ventricular myocytes (figs. S4 and S5). The CaMKII inhibitor KN-93, but not its inactive analog KN-92, suppressed phosphorylation of PLB at Thr17 in sinoatrial node cells (fig. S4C).

Ca2+-dependent phosphorylation of RyRs was associated with amplified Ca2+ release from RyRs

Phosphorylation of RyR2 is a key mechanism to alter RyR2 activity and Ca2+ release from the SR, although the exact effects of phosphorylation on RyR function are unclear (16, 2629). Cardiac RyRs in ventricular myocytes are phosphorylated by both PKA and CaMKII at Ser2808 (in mice) or Ser2809 (in rabbit) (27, 30), which increases the activation of RyRs by cytosolic Ca2+ and enhances RyR Ca2+ release (16, 26, 28). Mice with a form of RyR2 that cannot be phosphorylated by PKA at Ser2808 have reduced basal heart beating rates and blunted chronotropic responses to β-adrenergic receptor stimulation (31). In addition, in these mice, the ability of RyRs to respond to a higher Ca2+ load in the SR is reduced, suggesting a link between phosphorylation of RyR2 at Ser2808 and sensing of luminal Ca2+ concentrations in the SR (32).

There is little information regarding phosphorylation of RyR at Ser2809 in sinoatrial node cells. Immunofluorescence analysis indicated that total and phosphorylated RyRs were preferentially distributed beneath the sarcolemma of sinoatrial node cells, in the area where local Ca2+ releases were recorded (Fig. 5), whereas ventricular myocytes showed a more uniform distribution of total and phosphorylated RyR (fig. S6). Raising the concentration of cytosolic Ca2+ within the physiological range increased phosphorylation of RyR at Ser2809 in permeabilized sinoatrial node cells but not in ventricular myocytes (Fig. 5E and fig. S6). Phosphorylation of RyR at Ser2809 was reduced in sinoatrial node cells by the PKA inhibitor peptide PKI or the CaMKII inhibitor KN-93, suggesting that this site was phosphorylated by both PKA and CaMKII in these cells. This increased phosphorylation of both RyR and PLB at physiological concentrations of cytosolic Ca2+ may promote robust basal Ca2+ cycling in sinoatrial node cells.

Fig. 5

Phosphorylation of RyR in permeabilized sinoatrial node cells (SANC) was modulated by the cytosolic Ca2+ concentration. (A and B) Representative confocal images of SANC immunolabeled for total RyR (red) and RyR phosphorylated at Ser2809 (green) at the indicated cytosolic Ca2+ concentration. (C and D) A graph of the pixel-by-pixel fluorescence intensities of total and phosphorylated RyR labeling along an arbitrary line, indicated by a white line in (A) and (B) at the indicated cytosolic Ca2+ concentration. The dashed lines in (C) and (D) show the average pixel intensity (over the entire cell) for total RyR (red) and phosphorylated RyR (green). (E) Relative changes in the phosphorylation of RyR at Ser2809 normalized to total RyR in SANC exposed to different cytosolic Ca2+ concentrations (n = 49 to 80 cells for each data point) and VM (n = 80 to 81 cells for each data point). *P < 0.05.

Both PKA- and CaMKII-dependent phosphorylation events were required for the generation of robust periodic local Ca2+ releases in sinoatrial node cells

Suppression of PKA-dependent phosphorylation by PKI resulted in a decrease in the total Ca2+ signal mass released by sinoatrial node cells (Fig. 6A) as well as a decrease in the size (fig. S7A) and periodicity (Fig. 6B) of local Ca2+ releases. However, the Ca2+ content of the SR was unchanged when the cytosolic Ca2+ concentration was below 150 nM (Fig. 6C).

Fig. 6

Inhibition of PKA-dependent protein phosphorylation suppressed spontaneous, periodic, local Ca2+ releases in permeabilized sinoatrial node cells (SANC). (A) Left, confocal line-scan images of a representative SANC bathed in 200 nM cytosolic Ca2+ before (top) and after (bottom) superfusion with PKI. Right, total Ca2+ signal mass released by SANC before and after superfusion with PKI (n = 5 to 12 cells for each data point). (B) Left, FFT of Ca2+ oscillations (from bands indicated by arrows) in (A), before and after PKI. Right, relative number of SANC that generated periodic spontaneous Ca2+ releases in control and after PKI; logistic regression analysis demonstrated a significant difference between curves (P < 0.0001). (C) Left, representative images and caffeine-induced Ca2+ transients in SANC under control conditions and after treatment with PKI. Right, average data of the total Ca2+ content in the SR before and after PKI treatment. n = 3 to 9 cells for each data point. *P < 0.05.

Similar to suppression of PKA, suppression of CaMKII-dependent protein phosphorylation by autocamtide-2 related inhibitory peptide (AIP) in sinoatrial node cells was associated with reduced total Ca2+ signal mass (Fig. 7A) and decreased size (fig. S7D) and periodicity (Fig. 7B) of local Ca2+ releases. Similar to PKI, AIP did not change the Ca2+ content of the SR when the cytosolic Ca2+ concentration was below 150 nM (Fig. 7C). This would be possible only if decreased Ca2+ release through RyRs produced by inhibition of either kinase was balanced by reduced Ca2+ pumping into SR. Concurrent reductions in the phosphorylation of PLB (23) (fig. S4C) and phosphorylation of RyR (Fig. 5E) produced by inhibition of either PKA or CaMKII might account for a constant Ca2+ content in the SR. At a cytosolic Ca2+ concentration of 200 nM, the Ca2+ content of the SR was reduced by inhibition of either kinase (Figs. 6C and 7C), suggesting that less Ca2+ was pumped back into the SR by SERCA.

Fig. 7

Inhibition of CaMKII suppressed spontaneous, periodic, local Ca2+ releases in permeabilized sinoatrial node cells (SANC). (A) Left, confocal line-scan images of a representative SANC bathed in 200 nM cytosolic Ca2+ before (top) and after (bottom) superfusion with AIP. Right, AIP treatment resulted in decreased total Ca2+ signal mass released by SANC. n = 4 to 8 cells for each data point. (B) Left, FFT of Ca2+ oscillations (from bands indicated by arrows) in (A), before and after AIP treatment. Right, relative number of SANC that generated periodic spontaneous Ca2+ releases under control conditions and after AIP treatment. Logistic regression analysis demonstrated a significant difference between the two curves (P < 0.002). (C) Left, representative images and caffeine-induced Ca2+ transients in control conditions and after treatment with AIP. Right, average data of the total SR Ca2+ content before and after AIP treatment (n = 3 to 13 cells for each data point). *P < 0.05.

Discussion

Spontaneous beating of sinoatrial nodal pacemaker cells is produced by spontaneous diastolic depolarization, and an ensemble of ion channels and membrane transporters creates a “membrane clock” that generates diastolic depolarization and spontaneous action potentials (3, 33). Spontaneous subsarcolemmal local Ca2+ releases through RyR occur during diastolic depolarization and are envisioned to be a Ca2+ clock (34, 35). In spontaneously firing sinoatrial node cells, the membrane clock and Ca2+ clock do not operate in isolation but are mutually entrained through multiple interactions to form a stable, coupled system that drives the automaticity of cardiac pacemaker cells (3335).

Although Ca2+ cycling is essential for the spontaneous activity of cardiac pacemaker cells (1, 47, 17, 3336), the specific features of intracellular Ca2+ cycling remain unclear. Here, we have demonstrated that sinoatrial node cells generated more robust and rhythmic spontaneous local Ca2+ releases through RyRs at the same physiological cytosolic Ca2+ concentrations and similar SR Ca2+ content than did ventricular myocytes. Sinoatrial node cells and ventricular myocytes had different abundances of proteins involved in Ca2+ cycling by the SR [SERCA, PLB, and RyR (Figs. 3 and 5)]. Sinoatrial node cells, but not ventricular myocytes, had Ca2+-induced increases of both PKA- and CaMKII-dependent phosphorylation of these proteins.

To sustain a long-lasting, increased Ca2+ release through RyRs (Fig. 1), the SR Ca2+ stores in sinoatrial node cells must be rapidly replenished by SERCA (7, 13). We found that SERCA was more abundant in rabbit sinoatrial node cells than in ventricular myocytes by ~1.5-fold (Fig. 3B). Considering that the amount of PLB in sinoatrial node cells was ~2-fold less than in ventricular myocytes (Fig. 3C), the SERCA/PLB ratio could be at least ~3-fold larger in sinoatrial node cells than in ventricular myocytes. This suggests that there was more uninhibited (not bound to PLB) SERCA in sinoatrial node cells than in ventricular myocytes, which could result in increased Ca2+ pumping into the SR, which is required to support robust local Ca2+ releases. An increase in the cytosolic Ca2+ concentration in sinoatrial node cells was associated with increased phosphorylation of PLB at both Ser16 and Thr17 (Fig. 4 and fig. S4, respectively), which would be expected to relieve its inhibition of SERCA-mediated Ca2+ pumping. Accordingly, sinoatrial node cells maintained a sufficient SR Ca2+ load (Fig. 2), which, in the presence of an increased Ca2+ release from RyR, could be possible only with increased Ca2+ reuptake by the SR. Thus, coordinated and synchronized increases in both Ca2+ release and reuptake by the SR could sustain large and rhythmic spontaneous local Ca2+ releases through RyRs in sinoatrial node cells.

The interplay between Ca2+ release through RyRs and reuptake by SERCA is likely to be a major factor that regulates the periodicity of local Ca2+ releases. Indeed, each local Ca2+ release locally depletes SR Ca2+ content and inactivates RyR clusters that generate local Ca2+ releases (4). When SR Ca2+ is replenished by SERCA and RyRs recover from inactivation, the next spontaneous Ca2+ release could occur. Robust, rhythmic local Ca2+ releases in pacemaker cells required high basal PKA- and CaMKII-dependent protein phosphorylation because inhibition of either PKA- or CaMKII-dependent phosphorylation resulted in small (fig. S7, A and D), stochastic Ca2+ releases that resembled Ca2+ sparks in ventricular myocytes (Figs. 6 and 7). Moreover, small and stochastic Ca2+ releases in ventricular myocytes also became large and periodic upon exposure to a higher concentration of cytosolic Ca2+ (figs. S2 and S3). Thus, the size of spontaneous Ca2+ releases could be a critical factor involved in spontaneous Ca2+ release periodicity in both cell types (fig. S2). When cytosolic Ca2+ was increased above 250 nM, the Ca2+ content in the SR was increased only in ventricular myocytes (Fig. 2), which led to generation of large, rhythmic Ca2+ wavelets (fig. S3), whereas the SR Ca2+ load (Fig. 2) was decreased and local Ca2+ releases were eliminated in sinoatrial node cells (fig. S3). In permeabilized canine ventricular myocytes exposed to very high cytosolic Ca2+ concentrations, RyRs stay continuously open, which promotes a large Ca2+ leak and leaves the SR nearly empty of Ca2+, thereby suppressing spontaneous Ca2+ releases through RyRs (37).

The RyR is a multimolecular complex that includes regulatory proteins that control Ca2+ release from the SR (16). Cytosolic Ca2+ activates RyRs in cardiac cells through CICR, and the sensitivity of RyRs to cytosolic Ca2+ is also regulated by the Ca2+ content in the SR (16). However, permeabilized sinoatrial node cells discharge more Ca2+ than ventricular myocytes at a similar Ca2+ content in the SR, suggesting that other factors account for differences in spontaneous Ca2+ releases from RyRs between two cell types. Ca2+-induced phosphorylation of RyRs at physiological cytosolic Ca2+ concentrations was enhanced in sinoatrial node cells (Fig. 5), which could be related to the differences in local control mechanisms. In sinoatrial node cells, adenylyl cyclases are constitutively activated by Ca2+ (21, 22) and the basal concentration of cAMP and basal PKA-dependent phosphorylation are enhanced compared to ventricular myocytes (23). An increase in cytosolic Ca2+ concentrations in permeabilized sinoatrial node cells increased cAMP production (Fig. 4A), presumably because of increased activation of adenylyl cyclases, leading to increased PKA-dependent phosphorylation of RyR (Fig. 5E) and, likely, increased sensitivity of RyR to the concentration of luminal Ca2+ in the SR. However, it remains to be shown that Ca2+ release from RyRs in sinoatrial node cells requires these phosphorylation events. Ca2+-inhibited adenylyl cyclases 5 and 6 are the most abundant adenylyl cyclase isoforms in the ventricle (38), which might partially explain why PKA-dependent phosphorylation of both RyR (Fig. 5E and fig. S6) and PLB (Fig. 4C) was not observed in ventricular myocytes in response to increased cytosolic Ca2+ concentrations.

Physiological concentrations of cytosolic Ca2+ in sinoatrial node cells would also be expected to activate CaMKII (24, 25), and accordingly, these cells showed increased CaMKII-dependent phosphorylation of PLB (fig. S4) and RyR (Fig. 5E). Activated CaMKII is located beneath the surface membrane of sinoatrial node cells (39), in the area with highest RyR density (Fig. 5) where local Ca2+ releases are generated. In contrast, CaMKII-dependent phosphorylation of PLB (fig. S4) and RyR (Fig. 5E) in ventricular myocytes was not affected by increased cytosolic Ca2+ concentrations, which is consistent with a lack of change in CaMKII-dependent phosphorylation of RyRs in permeabilized canine ventricular myocytes in response to increased cytosolic Ca2+ (37).

Study limitations

To compare intrinsic Ca2+ cycling at the same physiological cytosolic Ca2+ concentration without interference from surface membrane channels, we used saponin-permeabilized sinoatrial node cells and ventricular myocytes. However, there are several caveats regarding permeabilized cells. First, small molecules in the cytoplasm may leak out, which could alter the regulation of intrinsic Ca2+ cycling. The phosphorylation response of permeabilized sinoatrial node cells appeared to be intact, suggesting that critical elements of Ca2+ cycling in these cells remained intact. Second, it would be an oversimplification to assume that Ca2+ cycling in permeabilized cardiac cells is identical to that within intact cells. Additional modulators of intrinsic Ca2+ cycling in cardiac cells include transmembrane ionic channels and membrane-bound regulatory proteins. Different sets of ionic channels and membrane-bound regulatory proteins in intact sinoatrial node cells and ventricular myocytes create additional dissimilarities between Ca2+ cycling in these cardiac cells. However, although trade-offs are required for the quantitative comparison between different cell types, we propose that our work can contribute to understanding the differences in intrinsic Ca2+ cycling in cardiac pacemaker cells and ventricular myocytes.

Potential clinical relevance of distinct characteristics of Ca2+ cycling in cardiac pacemaker cells

Human sinoatrial node dysfunction includes sinus bradycardia and sinus arrest, both of which show increased incidence with aging (2). Implantation of an electronic pacemaker is the common therapy to treat sinoatrial node dysfunction (40); however, it can cause premature death because of electronic failure or chronic infections (41). Although the creation of biological pacemakers is an attractive alternative, attempts to create one have been unsuccessful (4244). The present study showed that cardiac pacemaker cells rely on Ca2+ cycling by the SR to a greater extent than previously assumed. In contrast to current approaches that focus exclusively on ionic channels (42), a robust and functional biological pacemaker design might combine proteins involved in intracellular Ca2+ cycling with ionic channels.

Materials and Methods

Sinoatrial node cell isolation

Single sinoatrial node cells were isolated from New Zealand White rabbits (Charles River Laboratories) as previously described (6, 7). The sinoatrial node region was removed from the heart and cut into ~1.0-mm strips perpendicular to the crista terminalis. The sinoatrial node preparation was washed in Ca2+-free Tyrode solution containing 140 mM NaCl, 5.4 mM KCl, 0.5 mM MgCl2, 0.33 mM NaH2PO4, 5 mM Hepes, and 5.5 mM glucose (pH 6.9), and then incubated at 34 ± 0.5°C for 30 min in Ca2+-free Tyrode solution containing elastase type IV (0.6 mg/ml; Sigma-Aldrich Corp.), collagenase type 2 (0.6 mg/ml; Worthington), and 0.1% bovine serum albumin (BSA) (Sigma-Aldrich Corp.). Thereafter, the sinoatrial node preparation was washed in modified Kraftbruhe (KB) solution, containing 70 mM potassium glutamate, 30 mM KCl, 10 mM KH2PO4, 1 mM MgCl2, 20 mM taurine, 10 mM glucose, 0.3 mM EGTA, and 10 mM Hepes (titrated to pH 7.4 with KOH), and kept at 4°C for 1 hour in KB solution containing polyvinylpyrrolidone (50 mg/ml; PVP 40, Sigma-Aldrich Corp.). Cells were dispersed from the sinoatrial node by gentle pipetting in KB solution and stored at 4°C.

Ventricular myocyte isolation

The rabbit heart was perfused in Langendorff mode for 5 min with a nominally Ca2+-free modified Krebs solution (containing 120 mM NaCl, 5.4 mM KCl, 1.6 mM MgSO4, 1.0 mM NaH2PO4, and 20 mM NaHCO3) at 37°C, bubbled with 95% O2, 5% CO2. Perfusion continued for 3 to 4 min with protease (0.02 mg/ml, type XIV, Sigma-Aldrich Corp.) and collagenase [1 mg/ml; type B, 220 to 230 U/mg (Boehringer-Mannheim), or type 2 (Worthington)], and then for a further 10 to 15 min with 50 μM CaCl2 added. The ventricles were separated from the heart, chopped into chunks, and placed for a second digestion for 10 to 15 min in a shaker (60 to 70 rpm) at 37°C, with Tyrode solution containing 100 μM CaCl2 and collagenase (1 mg/ml). Digestion was quenched by filtering the supernatant for centrifugation at 500g and three washes with a modified Tyrode solution [137 mM NaCl, 4.9 mM KCl, 15 mM glucose, 1.2 mM MgSO4, 1.2 mM NaH2PO4, 20 mM Hepes (pH 7.4)]; Ca2+ concentrations were successively increased to 250, 500, and 1000 μM. Ventricular cells were stored at room temperature in 1 mM Ca2+ Tyrode solution.

Cell permeabilization

Intact sinoatrial node or ventricular cells were plated on laminin (20 μg/1 ml) in Petri dishes (MatTek Culterware) for 20 min to attach. Cells were permeabilized with 0.01% saponin, as previously described (6), in a solution containing 100 mM C4H6NO4K (dl-aspartic acid potassium salt), 25 mM KCl, 10 mM NaCl, 3 mM MgATP, 0.81 mM MgCl2 (~1 mM free Mg2+), 20 mM Hepes, 0.5 mM EGTA, 10 mM phosphocreatine, and creatine phosphokinase (5 U/ml), pH 7.2. After washing out saponin, permeabilization solution was changed to a continuous superfusion with an experimental solution that contained 30 μM fluo-4 pentapotassium salt (Life Technologies) and Ca2+. The cytosolic free Ca2+ at given total Ca2+, Mg2+, ATP, and EGTA concentrations was calculated using a computer program (WinMAXC 2.50, Stanford University).

Confocal imaging of local subsarcolemmal Ca2+ releases

Permeabilized sinoatrial node cells or ventricular myocytes loaded with fluo-4 pentapotassium salt were placed on the stage of a Zeiss LSM-410 inverted confocal microscope (Carl Zeiss Inc.). All images of local Ca2+ releases in sinoatrial node cells and sparks in ventricular myocytes were recorded in the line-scan mode, with the scan line oriented along the long axis of the cell and processed with IDL software (6.1, Research Systems). The bath temperature was maintained at 35 ± 0.5°C with a temperature controller (Cell Micro Controls). Spontaneous Ca2+ releases in permeabilized sinoatrial node cells and ventricular myocytes were analyzed as previously described (6, 23): The amplitude of each spontaneous Ca2+ release was calculated as the peak value (F) normalized to minimal fluorescence (F0), its spatial size was indexed as the full width at half maximum amplitude (FWHM), and its duration was characterized as the full duration at half maximum amplitude (FDHM). Parameters of spontaneous Ca2+ releases in sinoatrial node cells and ventricular myocytes were different; thus, the number of spontaneous Ca2+ releases did not reflect the amount of Ca2+ released by each cell type. To avoid this problem, we calculated the total Ca2+ signal mass released by each cell type (6, 43). The signal mass of individual Ca2+ releases (local Ca2+ releases in sinoatrial node cells or sparks in ventricular myocytes) was estimated as follows:M=FWHM×FDHM×1/2ΔF/F0 (where ΔF/F0=F/F01)(1)

[Ca2+]i value in permeabilized cells was calculated using pseudoratio (8):[Ca2+]i=Kd(F/F0)/(Kd/[Ca2+]f+1F/F0)(2)where Kd is 864 nM (45) and free Ca2+ at rest [Ca2+]f in the “internal” solution in permeabilized cells is known.

The spontaneous Ca2+ release amplitudes measured as F/F0 were converted to changes in Ca2+ concentration:Δ[Ca2+]i=peak[Ca2+]irest[Ca2+]f(3)where peak [Ca2+]i was calculated using Eq. 2 and rest [Ca2+]f is known.

The total Ca2+ signal mass collectively produced by either local Ca2+ releases in sinoatrial node cells or sparks in ventricular myocytes was estimated, as previously described (6, 46), by integrating signal masses of all individual Ca2+ releases produced by either cell type during a 400-ms time interval and along a 100-μm line-scan. In permeabilized sinoatrial node cells, local Ca2+ releases appear after a specific delay, demonstrating a roughly periodic behavior. In contrast, sparks in permeabilized ventricular myocytes are random at low cytosolic Ca2+ concentrations (50 to 150 nM), but when free cytosolic Ca2+ reaches 200 nM, they increase in size and become roughly periodic (figs. S2 and S3). To compare periodicity of spontaneous Ca2+ releases in sinoatrial node cells and ventricular myocytes, we used FFT (20) with Ca2+ waveforms as time domains for FFT. FFT analysis was performed using Clampfit 10.3 (Axon Instruments Inc.) with a rectangular window and the power displayed on a linear scale. To provide similar approach for two different cell types and considering that the spark size in ventricular myocytes is ~2 to 3 μm, we calculated Ca2+ waveforms from 2- to 3-μm bands. Although several spots with periodic Ca2+ releases could be present in the single sinoatrial node cell, for simplicity, a single spot was enough to consider that the cell had periodic Ca2+ releases.

The Ca2+ content in the SR in both cell types under control conditions and in the presence of drugs was determined from the peak amplitude of Ca2+ transients produced by rapid application of 20 mM caffeine on permeabilized cells. The change in intracellular [Ca2+]i on caffeine application (peak [Ca2+]i − rest [Ca2+]f before caffeine) was converted to the change in total SR Ca2+ (ΔCa2+ = [Ca2+]t − basal [Ca2+]t), as previously described (47), taking into account Ca2+ binding by EGTA and Ca2+ indicator:[Ca2+]t=[Ca2+]i+[EGTA]/(1+KEGTA/[Ca2+]i)+[fluo-4]/(1+Kfluo/[Ca2+]i)where [EGTA] = 0.5 mM, [fluo-4] = 30 μM, and Kfluo = 864 nM. For this type of experiment, images were taken with the scan line oriented perpendicular to the long axis of the cell to prevent cell shifting during caffeine application.

Western blot (SERCA)

Heart tissues (atrial, sinoatrial, or ventricular) or intact sinoatrial node cells or ventricular myocytes were frozen in liquid nitrogen and subsequently mixed with ice-cold lysis buffer [radioimmunoprecipitation assay: 50 mM tris-HCl (pH 8.0), 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM EDTA] with protease inhibitor cocktail (Sigma-Aldrich Corp.). Total protein content was determined using standard commercially available 2D protein Quant kit (GE Healthcare); accuracy of protein quantification was verified using 4 to 12% bis-tris polyacrylamide SDS–polyacrylamide gel electrophoresis (SDS-PAGE) gel (Life Technologies). Total integral of fluorescence intensity for each protein lane was calculated using standard ImageQuant5 software (GE Healthcare). Twelve micrograms of protein from either tissue or cell lysates was mixed with NuPAGE sample buffer and dithiothreitol and resolved on 4 to 12% bis-tris polyacrylamide SDS-PAGE gel, using standard Mops running buffer. Proteins were transferred onto a polyvinylidene difluoride membrane and incubated with primary anti-SERCA mouse monoclonal antibodies (ABR, 1:5000) and subsequently with goat anti-mouse polyclonal horseradish peroxidase (HRP) secondary antibodies (Dako North America Inc.). For actin, primary anti-actin goat polyclonal antibodies (Santa Cruz Biotechnology Inc., 1:3000) were used in the same gel, followed by secondary donkey anti-goat polyclonal HRP antibodies (Santa Cruz Biotechnology Inc.).

Western blot of PLB in intact or permeabilized sinoatrial node cells and ventricular myocytes

Rabbit sinoatrial node cells and ventricular myocytes were permeabilized and incubated at different Ca2+ concentrations (0 to 250 nM) for 10 min at 35 ± 0.5°C. Subsequently, intact or permeabilized cells were solubilized, and extracts were snap-frozen in liquid nitrogen. To dissociate PLB into monomers (66 kD), we boiled the samples at 95°C for 10 min. The protein load was determined using a BCA protein assay kit (Pierce Biotechnology). PLB proteins were resolved by 7.5% urea/SDS-PAGE gel and transferred (10 μg of protein per lane) onto polyvinylidene difluoride (PVDF) membranes (Amersham Pharmacia Biotech). For detection of total PLB, membranes were probed with the total PLB monoclonal antibody [immunoglobulin G1 (IgG1), 1:10,000, Badrilla]. Detection of site-specific PLB phosphorylation at PKA-dependent Ser16 site was performed using site-specific PS-16 polyclonal antibody (1:10,000, Badrilla) and HRP-conjugated anti-rabbit secondary antibody (1:15,000) (Bio-Rad Laboratories), as previously described (48). PVDF membranes were exposed to chemiluminescence (ECL, Amersham Pharmacia Biotech) reaction and quantified with a video documentation system (Bio-Rad Laboratories).

Immunostaining of PLB and RyR in permeabilized rabbit sinoatrial node cells and ventricular myocytes

Rabbit sinoatrial node cells and ventricular myocytes were permeabilized and incubated at different Ca2+ concentrations (0, 150, 200, or 250 nM) at 35 ± 0.5°C. In a subset of experiments, permeabilized sinoatrial node cells were treated with either PKA inhibitor (15 μM PKI) or CaMKII inhibitor (3 μM KN-93 or its inactive analog KN-92) for 20 min. The cells were fixed with 4% paraformaldehyde, treated with 1% Triton, and incubated with blocking solution (1× phosphate-buffered saline containing 2% BSA + 5% donkey serum + 0.01% NaN3 + 0.2% Triton). For assessment of CaMKII-dependent phosphorylation of Thr17, permeabilized cells were incubated with a polyclonal antibody specific for phosphorylated Thr17 (rabbit, 1:200, Badrilla) or total PLB monoclonal antibody (mouse, 1:1000, Badrilla). For RyR2 immunostaining, permeabilized cells were treated with an antibody that recognized total RyR2 (mouse, 1:1000, ABR) and an antibody that recognized RyR2 phosphorylated at Ser2809 (rabbit, 1:200, Badrilla). Secondary Cy5-conjugated anti-mouse IgG antibody (1:1000, Jackson ImmunoResearch Laboratories) was used for either total PLB or total RyR, and secondary Cy3-conjugated anti-rabbit IgG antibody (1:1000, Jackson ImmunoResearch Laboratories) was used for phosphorylated PLB or RyR immunostaining. Dual confocal images of central sections of sinoatrial node cells and ventricular myocytes were obtained with a Zeiss LSM 510 (Carl Zeiss Inc.). For either RyR or PLB, the total amount of protein or fluorophore Cy5 was detected by 633-nm excitation and 650-nm long-pass emission, and 543-nm excitation and 565- to 615-nm band-pass emission were used for the detection of phosphorylation or fluorophore Cy3. Considering that cells vary in size and total RyR or PLB protein density might vary from cell to cell (11), the extent of RyR or PLB phosphorylation was estimated as the ratio of phosphorylated to total RyR or PLB protein. Specifically, the average fluorescence density of PLB phosphorylated at Thr17 in a cell was normalized to total PLB, and the fluorescence density of RyR2 phosphorylated at Ser2809 in a cell was normalized to the total RyR2 fluorescence density (the nuclear area was excluded in all cells). Only secondary antibodies were applied to the negative control, which displayed negligible fluorescence.

cAMP measurements in permeabilized rabbit sinoatrial node cells

Permeabilized sinoatrial node cells were incubated for 5 min at 35°C in a solution containing 0 or 150 to 180 nM free cytosolic Ca2+. The buffer and sinoatrial node cells were collected, and the reaction was stopped with ice-cold ethanol. Supernatants were adjusted to 65% ethanol and used for cAMP estimation, using a LANCE cAMP 384 kit (PerkinElmer). Fluorescence intensity was measured on a Victor instrument (PerkinElmer) with the LANCE high count 615/665 protocol. Total protein was determined with a BCA assay (Pierce Biotechnology). The amount of the cAMP was calculated in picomoles per milligram of protein.

Drugs

The CaMKII inhibitor KN-93, its inactive analog KN-92, the CaMKII inhibitor peptide AIP (30 μM), and a cell-permeable PKA inhibitory peptide amide (14–22) (PKI, 15 μM) were from EMD4 Biosciences, Calbiochem.

Statistics

Data are shown as means ± SEM. Statistical significance was evaluated by t test and analysis of variance (ANOVA) followed by Bonferroni post hoc tests. Normalized data were analyzed using a one-sample t test. If heterogeneous random errors have been observed in the data set, a logarithmic transformation of the data was performed before the analysis to stabilize the variance. Logistic regression analysis was used to evaluate statistical significance between fitting curves of periodic Ca2+ releases in two different cell types and in sinoatrial node cells before and after treatment with PKI or AIP. A value of P < 0.05 was considered statistically significant.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/6/260/ra6/DC1

Fig. S1. Local Ca2+ release characteristics are Ca2+-dependent and differ in sinoatrial node cells and ventricular myocytes.

Fig. S2. The size of spontaneous local Ca2+ releases delineates their periodicity.

Fig. S3. High cytosolic Ca2+ concentrations eliminate local Ca2+ releases in sinoatrial node cells.

Fig. S4. Increased cytosolic Ca2+ concentrations increase phosphorylation of PLB at Thr17 in sinoatrial node cells, but not in ventricular myocytes.

Fig. S5. Increased cytosolic Ca2+ concentrations do not change the phosphorylation of PLB at Thr17 in ventricular myocytes.

Fig. S6. Phosphorylated RyR immunolabeling in ventricular myocytes is not Ca2+-dependent.

Fig. S7. Effects of PKI and AIP on the characteristics of local Ca2+ releases in permeabilized sinoatrial node cells.

References and Notes

Acknowledgments: We are deeply grateful to professional statisticians C. Morrell and V. Shetty for their help with statistical analysis and R. Sadler for her excellent technical support. Funding: This study was supported by the Intramural Research Program of the National Institute on Aging, NIH. Author contributions: E.G.L. and T.M.V. created the project and designed the experiments. S.S., E.G.L., and T.M.V. wrote the paper. S.S. and T.M.V. performed the analysis of data and statistical analysis. S.S., D.Y., Y.L., A.E.L., and Y.O.L. performed the experiments: S.S. performed measurements of local Ca2+ releases and SR Ca2+ contents in permeabilized sinoatrial node cells and ventricular myocytes; D.Y. performed immunostaining of PLB and RyR; Y.L. performed Western blots of PLB and analyzed the data; A.E.L. performed Western blots of SERCA and analyzed the data; and Y.O.L. performed Western blots of PLB and cAMP measurements and analyzed the data. Competing interests: The authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Signaling

Navigate This Article