Sarcoplasmic reticulum Ca²⁺ pumping kinetics regulates timing of local Ca²⁺ releases and spontaneous beating rate of rabbit sinoatrial node pacemaker cells

Abstract

Rationale

Sinoatrial node cells (SANC) generate local, subsarcolemmal Ca²⁺ releases (LCRs) from sarcoplasmic reticulum (SR) during late diastolic depolarization (DD). LCRs activate an inward Na⁺-Ca²⁺ exchange current (I_NCX) which accelerates DD rate, prompting the next action potential (AP). The LCR period, i.e., a delay between AP-induced Ca²⁺ transient and LCR appearance, defines the time of late DD I_NCX activation. Mechanisms that control the LCR period, however, are still unidentified.

Objective

To determine dependence of the LCR period on SR Ca²⁺ refilling kinetics and establish links between regulation of SR Ca²⁺ replenishment, LCR period and spontaneous cycle length.

Methods and Results

Spontaneous APs and SR luminal or cytosolic Ca²⁺ were recorded using perforated patch and confocal microscopy, respectively. Time to 90% replenishment of SR Ca²⁺ following AP-induced Ca²⁺ transient was highly correlated with the time to 90% decay of cytosolic Ca²⁺ transient (T-90_C). Local SR Ca²⁺ depletions mirror their cytosolic counterparts, LCRs, and occur following SR Ca²⁺ refilling. Inhibition of SR Ca²⁺ pump by cyclopiazonic acid (CPA) dose-dependently suppressed spontaneous SANC firing up to ~50%. CPA and graded changes in phospholamban phosphorylation produced by β-AR stimulation, phosphodiesterase or PKA inhibition shifted T-90_C and proportionally shifted the LCR period and spontaneous cycle length (R²=0.98).

Conclusions

The LCR period, a critical determinant of the spontaneous SANC cycle length, is defined by the rate of SR Ca²⁺ replenishment, which is critically dependent on SR pumping rate, Ca²⁺ available for pumping, supplied by L-type Ca²⁺ channel, and RyR Ca²⁺ release flux each of which is modulated by cAMP-mediated PKA-dependent phosphorylation.

Introduction

The sinoatrial (SA) node is the primary physiological pacemaker of the heart. Sinoatrial node cells (SANC) are able to generate spontaneous action potentials due to the gradual, spontaneous depolarization of the membrane, i.e. diastolic depolarization (DD)¹. Multiple mechanisms are involved in the generation of DD, including numerous ionic currents¹, and the most recently discovered mechanism, local subsarcolemmal Ca²⁺ releases (LCRs) from ryanodine receptors (RyR)^2–4. Similar to ventricular myocytes, SANC cycles Ca²⁺ via sarcoplasmic reticulum (SR) equipped with SR Ca²⁺ pumps (SERCA) and release channels, RyR, and extrudes Ca²⁺ from the cell via Na⁺-Ca²⁺ exchanger^2,5,6. LCRs appear during late DD, prior to the AP upstroke^2–4 and do not require a change in the membrane potential, but occur spontaneously: they persist during acute switch to the voltage clamp and are present in permeabilized SANC⁷.

During each spontaneous cycle, Ca²⁺ influx through L-type Ca²⁺ channels, triggered by the AP upstroke, produces a global Ca²⁺ transient, causing a global SR Ca²⁺ depletion and RyR inactivation. When the SR Ca²⁺ content is replenished by SERCA, which constantly pumps Ca²⁺ back into the SR, and RyRs recover from inactivation, LCRs begin to appear. The restitution time, i.e. the time from the AP-triggered global Ca²⁺ transient to the onset of LCR during DD is the LCR period. LCR occurrence activates an inward Na⁺-Ca²⁺ exchange current (I_NCX), which produces an exponential rise of the late DD, and is a determinant of the time at which the next rapid AP upstroke will occur^7–9. The LCR period, therefore, is a regulator of the spontaneous SANC beating rate^7–9. While it has been assumed that the LCR period depends, at least in part, upon the rate at which SR is refilled with Ca²⁺, an intrinsic link between the former and latter has never been demonstrated. The first goal of the present study is to test the hypothesis that SR Ca²⁺ refilling is a crucial determinant of both the LCR period and the spontaneous SANC beating rate.

The velocity of Ca²⁺ pumping into SR by SERCA is modulated by phospholamban (PLB) phosphorylation at the protein kinase A (PKA)-dependent Ser¹⁶ site¹⁰. While a prior study has noted a close correlation between graded changes in PLB phosphorylation and the LCR period¹¹, the mechanisms accountable for this close relationship have never been revealed. The second goal of the present study is to define mechanisms accountable for the close link between PLB phosphorylation and the LCR period in SANC.

Materials and Methods

An extensive description of ‘Materials and Methods’ for SA node cell isolation, and electrophysiological recordings, confocal imaging of SR Ca²⁺ depletions, cytosolic Ca²⁺ transients and LCRs^2,7,11–13, cell permeabilization^7,9, and Western blotting^9,11 is provided in the online data supplement at http://circres.ahajournals.org.

Statistical analysis

Data are presented as mean±SEM. The statistical significance of effects was evaluated by Student’s t-test or analysis of variance (ANOVA) where appropriate. A value of P<0.05 was considered statistically significant.

Results

To determine the role of SERCA pumping in spontaneous SANC firing at 37°C we employed a specific and reversible SERCA inhibitor, cyclopiazonic acid (CPA)¹⁴. CPA decreased the spontaneous SANC beating rate in a dose-dependent manner (EC₅₀, 1.2 μmol/L) to a maximal suppression of 53.9 ± 3.8% (Fig. 1A, B); all effects were reversed upon CPA washout. The suppression in the beating rate was due to a marked decrease in the DD rate (Fig. 1A) from 59±3 to 21±3 mV/sec (n=4, P<0.01), supporting the idea that SR refilling plays an essential role in the control of the basal pacemaker function.

Figure 1. Inhibition of SERCA by cyclopiazonic acid (CPA) suppresses LCRs and SANC beating rate.

A, APs recorded in a representative rabbit SANC before and during superfusion with 3 μmol/L CPA. B, The relative dose-dependent decrease in SANC firing rate in response to different CPA concentrations. C, Confocal line-scan images of a representative SANC prior to and following exposure to 3μmol/L CPA, LCRs are indicated by arrowheads. The inset below the top panel illustrates how the LCR period is defined, i.e., as the time from the prior AP-induced Ca²⁺ transient to the onset of LCR. D, In 6 SANC CPA (3μmol/L) markedly decreases number of LCRs per spontaneous cycle as well as LCR’s size, measured as full width at half maximum (FWHM). * P < 0.05.

To clarify specific mechanisms involved in the CPA-induced suppression of DD and spontaneous firing of intact SANC, we studied how CPA affects Ca²⁺ cycling, specifically, subsarcolemmal LCRs. Representative images and average data in figure 1C–D show that at 4 minutes of superfusion with 3 μmol/L CPA, there was a marked decrease in LCR size as well as a reduction of LCR number during each spontaneous cycle (Fig 1C–D). CPA also markedly increased the LCR period, and the increase in the LCR period was highly correlated with a prolongation of the spontaneous cycle length (Online Figure I D), suggesting that changes in LCR characteristics could be a major mechanism of CPA-induced decrease in spontaneous SANC firing.

To define direct effects of CPA on LCRs in the absence of regularly occurring AP-induced Ca²⁺ transients, SANC were permeabilized with saponin and bathed at 100 nmol/L cytosolic free [Ca²⁺]. Similar to its effect in intact SANC, a 4-minute superfusion with CPA reduced the LCR frequency and size in permeabilized SANC (Fig. 2A, B); this was due, at least in part, to a substantial reduction in the SR Ca²⁺ content, assessed by a rapid spritz of 20 mmol/L caffeine directly onto the skinned SANC (Fig. 2C, D).

Figure 2. CPA decreases the number and size of LCRs and the SR Ca²⁺ content in permeabilized rabbit SANC.

A, Confocal line scan images of a representative saponin-permeabilized SANC bathed in 100 nmol/L free [Ca²⁺] prior to and following exposure to 3 μmol/L CPA. B, (left) The average frequency of LCRs (normalized per 1s and 100 μm); (right) LCR size, measured as full width at half maximum (FWHM) in ‘skinned’ SANC in control conditions (4 SANC; 160 LCR) and after 3 min superfusion with 3 μmol/L CPA (4 SANC; 46 LCR). C, The effect of a rapid application of caffeine to a representative permeabilized SANC in the absence (top) and presence (bottom) of 3 μmol/L CPA. D, average effect of CPA on the initial rapid component (arrow) of the caffeine-induced SR Ca²⁺ release, indexed by F/F₀; n=9 SANC in a control group; n=7 SANC following 3 min superfusion with 3 μmol/L CPA. * P< 0.05.

Next, we determined how the increase in the LCR period produced by CPA in intact SANC is related to CPA-induced suppression of SERCA function. In rabbit ventricular myocytes the rate of [Ca²⁺]_i decline during AP-induced Ca²⁺ transient is highly dependent on Ca²⁺ pumping into SR^15,16, and either the time to 90% decay of Ca²⁺ transient (T-90_C) or monoexponential [Ca²⁺]_i decline constant (τ) are convenient measures to characterize the kinetics of the SR Ca²⁺ pumping and SR Ca²⁺ refilling^15,16. Figure 3A illustrates CPA-induced time-dependent prolongation of the decay of AP-induced Ca²⁺ transient, as reflected in the increase of either the monoexponential decline constant, τ, or time to 90% decay of the cytosolic Ca²⁺ transient, T-90_c (fig. 3A,B). Note, that changes in T-90_c were paralleled by changes in τ, suggesting that either parameter faithfully indexes the Ca²⁺ transient decay. The CPA-induced time-dependent increase in the relaxation time of the Ca²⁺ transient was accompanied by a concomitant increase in the LCR period. There was a strong correlation between the time-dependent prolongation in SR Ca²⁺ refilling and prolongation of the LCR period, suggesting that Ca²⁺ refilling of SR following AP-induced Ca²⁺ release is a key determinant of the LCR period (fig. 3B). The time-dependent CPA-induced increase in LCR period predicted the time-dependent increase in the spontaneous cycle length (R²=0.99), and both parameters recovered upon washout (Fig 3C).

Figure 3. Effect of CPA to increase the LCR period is linked to a prolongation of the decay of AP-induced Ca²⁺ transient.

A, Confocal linescan images and Ca²⁺ waveforms of a representative spontaneously beating SANC before, during and after superfusion with 3 μmol/L CPA. In this SANC LCRs were inhibited after 5-minute CPA superfusion. B, Relationship between average changes in the monoexponential decline constant, τ (y=0.7x−45.5msec, R²=0.96), and time to 90% decay of the AP-induced Ca²⁺ transient, T-90_C (y=0.7x+10.9msec, R²=0.99) and average changes in the LCR period in a representative SANC (panel A) before, during and after superfusion with 3 μmol/L CPA. C, The average CPA-induced increase of the LCR period in B is linked to the average increase in the spontaneous cycle length (y=1.0x +53.2msec, R²=0.99).

On average, after a ~4 minute superfusion with 3 μmol/L CPA (n=5) there was ~51% slowing of the decay of the AP-induced Ca²⁺ transient, T-90_c (from 289.8± 29.2 to 428.7±44.8 msec, P<0.01); ~30% decrease in its amplitude (from 0.46±0.08 to 0.32±0.06 ΔF/F₀, P<0.05), accompanied by ~53% increase in the LCR period (from 361.2± 38.8 to 532.3±47.5 msec, P<0.02).

To confirm that Ca²⁺ refilling of SR is, indeed, a determinant of the occurrence of subsequent LCR, luminal SR Ca²⁺ in SANC was visualized using Fluo-5N, a low-affinity Ca²⁺ indicator¹². The SR refilling times after AP-induced Ca²⁺ transient were indexed by 90% refilling of SR Ca²⁺ store (T-90_SR) (Fig 4A). LCR-created Local SR Ca²⁺ Depletions (LCD) were detected as spatially restricted “darkenings”, clearly seen after the Ca²⁺ image was normalized (F/F₀) and processed using customized software¹², which permitted detection of LCDs as apparent dips among background noise (Fig 4A). In order to confirm that the relaxation time of global AP-induced Ca²⁺ transient at 90%, T-90_c, reflects the SR refilling rate, T-90_SR, and that the LCR period is similar to the period of LCD, we compared histograms of T-90_SR and T-90_c (Fig 4C), as well as histograms of LCD periods and LCR periods (Fig 4D). That histograms of T-90_SR and T-90_c overlap, indicates that relaxation time of global AP-induced Ca²⁺ transient, T-90_c, reflects the SR refilling time, T-90_SR. Histograms of LCD periods and LCR periods also overlap (Fig. 4D) indicating that the LCR period reflects the LCD period.

Figure 4. Refilling of SR Ca²⁺ after AP-induced SR Ca²⁺ depletion in SANC is faithfully reported by the decay of cytosolic Ca²⁺.

A, Confocal linescan image of SR luminal Ca²⁺ (top) and corresponding Ca²⁺ waveform (bottom) in a representative spontaneously beating SANC visualized using low affinity Ca²⁺ indicator Fluo-5N. The bottom panel illustrates how the LCD period and AP-induced global Ca²⁺ depletions, T-90_SR, are defined. B, Confocal linescan image (top) and Ca²⁺ waveform (bottom) illustrate changes in cytosolic Ca²⁺ in a representative SANC. The inset below the bottom panel illustrates how the LCR period and decay of AP-induced Ca²⁺ transient, T-90_C, are defined. C, Histograms of the SR Ca²⁺ refilling times, T-90_SR (14 SANC), and decay of Ca²⁺ transient, T-90_C (9 SANC), overlap, indicating that T-90_C truthfully reproduce T-90_SR. D, Histograms LCDs periods (82 LCDs from 28 SANC) and LCR periods (92 LCRs from 9 SANC) overlap, indicating that LCR period truthfully reproduce LCD periods.

In ventricular myocytes, β-AR stimulation increases cAMP-mediated, PKA-dependent PLB phosphorylation, leading to relief of inhibition of SERCA and to a decrease in the SR refilling time^10,16. Prior studies have demonstrated that spontaneous beating of rabbit SANC is critically dependent upon cAMP-mediated, PKA-dependent phosphorylation, particularly, phospholamban (PLB) phosphorylation at PKA-dependent Ser 16 site^9,11. However, the relationship between changes in PLB phosphorylation during β-adrenergic (β-AR) stimulation in SANC, changes in SR refilling time and LCD period were not directly measured. Figure 5 shows that β-AR stimulation with isoproterenol (ISO, 0.1 μmol/L) decreases SR refilling times and shifts the histogram of T-90_SR to the left (Fig. 5C). The decrease in the SR refilling time is linked to the β-AR stimulation-induced increase in PLB phosphorylation (Fig. 8A) and consequent acceleration of the SERCA Ca²⁺ pumping rate. The decrease in T-90_SR is accompanied by a decrease in LCD period and a shift of LCD period histogram to the left (Fig. 5B vs 5C). The decrease in the LCD period is accompanied by a decrease in the spontaneous cycle length (Fig. 5D), which confirms that the refilling time is indeed the key factor that controls the LCD period and spontaneous SANC beating rate.

Figure 5. β-AR stimulation induced reduction in the spontaneous AP cycle length is linked to acceleration of SR Ca²⁺ refilling.

A, Confocal linescan image of SR luminal Ca²⁺ (middle) and SR luminal Ca²⁺ waveform from the whole image (top) and from the area inside a dashed line (bottom) during β-AR stimulation visualized using low affinity Ca²⁺ indicator Fluo-5N in a representative spontaneously beating SANC. B, C, Histograms of the SR Ca²⁺ refilling times, T-90_SR, and LCD periods (82 LCD) in subset of control SANC and in another subset of SANC, after superfusion with 0.1 μmol/L ISO (62 LCDs), respectively. D, The reduction in the spontaneous cycle length during β-AR stimulation is linked to the reduction in the LCD period.

Figure 8. Graded changes in PLB phosphorylation are paralleled by proportional changes in SR Ca²⁺ refilling time, T-90_C, and LCR period, which are highly correlated with concurrent changes in the spontaneous cycle length.

A, Representative western blots of PLB phosphorylated at serine¹⁶ site and total PLB in rabbit SANC in the basal state and following milrinone (50μmol/L), IBMX (100μmol/L), β-AR stimulation (0.1 μmol/L (ISO1) or 1 μmol/L (ISO2) isoproterenol), and PKI (10 μmol/L). B, Graded changes in PLB phosphorylation by β-AR stimulation (0.1 μmol/L ISO, n=8), a broad-spectrum PDE inhibitor (100 μmol/L IBMX, n=8), specific PDE-3 inhibitor (50 μmol/L milrinone, n=8), or by specific PKA inhibitor peptide (10 μmol/L PKI, n=8) are linked to inverse changes in T-90_C. C, Changes in T-90_C produced by either changes in PLB phosphorylation or by CPA are paralleled by concomitant changes in the LCR period. D, Changes in the spontaneous cycle length are tightly linked to the concurrent changes in the LCR period.

Because of high Fluo-5N bleaching, it is not possible to measure refilling times, T-90_SR, and LCD periods in the same SANC before and after β-AR stimulation. But, since the relationship of T-90_C and LCR period is similar to that of T-90_SR and LCD period (Fig. 4), we can use T-90_C and LCR period readouts to index effects of β-AR stimulation in the same SANC. Online data supplement Figure II shows that β-AR stimulation decreases both T-90_C (from 222.4±10.7 to 183.0±8.6 msec, n=6) and LCR period (from 364.1±16.2 to 288.6±7.9 msec), shifting their histograms to the left. The decrease in the LCR period is accompanied by the decrease in the spontaneous cycle length (Online figure II). β-AR stimulation with ISO also increases the average number of LCRs per each spontaneous cycle (from 1.00±0.08 to 1.50±0.08; P<0.01); LCR’s size (from 6.74±1.08 to 7.59±0.91 μm) and LCR’s amplitude (from 0.93±0.09 to 1.11±0.08 ΔF/F₀, P<0.02).

Like β-AR stimulation, PDE inhibition markedly increases the level of cAMP and consequentially increases cAMP-mediated PKA-dependent phosphorylation in conjunction with its effect to accelerate spontaneous firing of rabbit SANC¹¹. Similar to β-AR stimulation, an increase in PKA-dependent PLB phosphorylation by PDE inhibition (Fig. 8A) would be expected to relieve SERCA from PLB inhibition, accelerate pumping Ca²⁺ into SR and its refilling decreasing the time of the decay of AP-induced Ca²⁺ transient, T-90_C. Indeed, suppression of PDE activity by either a broad-spectrum PDE inhibitor, IBMX, or PDE3 inhibitor, milrinone, markedly decreases T-90_C and concomitantly decreases the LCR period, shifting their histograms to the left (for IBMX, Fig. 6; for milrinone, Online Figure III, respectively). The reduction in the LCR period produced by either IBMX or milrinone is paralleled by a reduction in the spontaneous cycle length (Fig. 6D and Online Figure III D, respectively). Both IBMX and milrinone markedly increase the number of LCRs per cycle, LCR size and amplitude (Online Figure IV).

Figure 6. Suppression of phosphodiesterase activity with IBMX increases SR Ca²⁺ refilling rate and reduces T-90_C and LCR period.

A, Confocal line-scan images of a representative SANC prior to and during exposure to 100 μmol/L IBMX, LCRs are indicated by arrowheads. B, Histograms of the decay of AP-induced Ca²⁺ transient, T-90_C, before and during superfusion with IBMX. C, Histograms of the LCR period in control (46 LCRs from 5 SANC) and after superfusion with IBMX (124 LCRs from 5 SANC). D, The IBMX-induced decrease in the spontaneous cycle length is linked to the decrease in the LCR period.

The suppression of PKA-dependent phosphorylation in SANC by a specific PKA inhibitor peptide, PKI, is likely mediated by a reduction in phosphorylation of multiple proteins that regulate SANC Ca²⁺ balance, including PLB (Fig. 8A), RyR, L-type Ca²⁺ channels and probably others⁹. In SANC, PKI markedly prolongs a decay of AP-induced Ca²⁺ transient, T-90_C, shifting its histogram and that of LCR periods, to longer times (Fig. 7). This increase in the LCR period is accompanied by an increase in the spontaneous SANC cycle length (Fig. 7D). PKI also substantially decreases the number of LCR’s per each spontaneous cycle (from 1.24±0.18 to 0.72±0.06; n=4, P<0.05), LCR amplitude (from 1.06±0.04 to 0.84±0.04 ΔF/F₀; n=5, P<0.05) and size (from 7.05±0.35 to 4.95±0.36 μm, n=4, P<0.05).

Figure 7. Suppression of PKA-dependent phosphorylation with PKI increases SR Ca²⁺ refilling time and prolongs the LCR period.

A, Confocal line-scan images of a representative SANC prior to and following exposure to 5 μmol/L PKI, LCRs are indicated by arrowheads. B, Histograms of the decay of the Ca²⁺ transient, T-90_C, before and during superfusion with PKI. C, Histograms of the LCR period in control (n=5 cells, 52 LCRs) and during superfusion with PKI (n=5 cells, 18 LCRs). D, The PKI-induced increase in the spontaneous cycle length is linked to the increase in the LCR period.

Figure 8B illustrates the continuum between changes in the PLB phosphorylation produced by β-AR stimulation, PDE inhibition or PKI and changes in T-90_C. Thus, the rate at which the SR reloads with Ca²⁺ is dependent on the level of PLB phosphorylation: an increase or decrease in PLB phosphorylation is linked to, proportional inverse changes in SR refilling times (Fig. 8B). Furthermore, the effects of PLB phosphorylation on T-90_C are reflected in the changes of the LCR period (Fig. 8C), i.e. the acceleration of SR refilling produced by either β-AR stimulation or PDE inhibition decreases, while PKI increases, the LCR period. Note, that the increase in the SR refilling time and LCR period produced by direct inhibition of SERCA by CPA, which does not change PLB phosphorylation, lies along the same function that depicts the effects of maneuvers that change PLB phosphorylation (Fig. 8C). Finally, relative changes in the LCR period, effected by perturbations that alter SR pumping rate in figure 8C, are tightly linked to the relative changes in the spontaneous cycle length (R²=0.98), and the function describing this conforms to the line of identity (Fig. 8D).

Discussion

The first novel finding of the present study is that the restitution time for LCRs, the LCR period, and the spontaneous SANC cycle length, are both determined by speed at which SR is refilled with Ca²⁺. The counterpart of cytosolic LCRs, i.e. LCR-generated local SR Ca²⁺ depletions, LCDs, visualized with a low-affinity Ca²⁺ indicator Fluo-5, appear only when SR is replenished with Ca²⁺ (Fig. 4).

A crucial role of SR Ca²⁺ cycling and, particularly, SERCA function for normal spontaneous pacemaker firing (Fig. 1) is confirmed by suppression of SERCA function and SR Ca²⁺ reuptake with CPA, a specific Ca²⁺-ATPase inhibitor^14,17, which dramatically, in a time- and dose-dependent manner decreases the spontaneous beating rate of rabbit SANC by up to ~50% (Fig. 1B). The second novel finding of the present study is that the CPA-induced increase in the spontaneous cycle length (reduction in the beating rate) is due to the prolongation of the LCR period, caused by a substantial increase in the time of SR refilling with Ca²⁺ (Fig. 3, Online Figure I). The increase in the LCR period and decrease in LCR number and size produced by CPA postpones the occurrence of LCR-activated, Na⁺-Ca²⁺ current and reduces its amplitude leading to decrease in the slope of DD (Fig. 1A) and, as a result, prolongation of the spontaneous cycle length (Online Figure I).

In skinned SANC the suppression of SERCA function by CPA decreases the SR Ca²⁺ load (Fig. 2D) and markedly suppresses LCRs (Fig. 2B) indicating that SR Ca²⁺ load is a crucial determinant of LCR characteristics. Similar to intact SANC, CPA also markedly increases a time interval between spontaneous LCRs in skinned SANC (Fig. 2A). When SR Ca²⁺ pumping rate is inhibited by CPA, LCR number and size both in intact (Fig. 1D) and permeabilized SANC are markedly decreased (Fig. 2B).

The decay rate of the AP-induced transient increase in cytosolic [Ca²⁺] reflects the activities of both the SR Ca²⁺-ATPase and the Na⁺-Ca²⁺ exchanger, and their contribution is species dependent^15,16. In rabbit ventricular myocytes ~70% of Ca²⁺ released in cytosol during AP-induced Ca²⁺ transient is actively transported back into SR by SR Ca²⁺- ATPase and ~30% is extruded from the cell via Na⁺-Ca²⁺ exchanger^15,16. Thus, the dominance of SERCA over Na⁺-Ca²⁺ exchanger in rabbit ventricular myocytes is about 2 to 3-fold^15,16. In the present study in rabbit SANC, inhibition of SERCA function by CPA, increased the time of the decay of AP-induced Ca²⁺ transient, T-90_C, by ~2-fold, which is consistent with data in rabbit ventricular myocytes.^15,16

In SANC the subsystems of ionic channels and Ca²⁺ cycling work together to guarantee the beating rate stability in a given steady state.^18,19 There are multiple interactions between these two subsystems: surface membrane proteins not only control changes in membrane potential, but also directly or indirectly regulate intracellular Ca²⁺ cycling; and vice versa, intracellular Ca²⁺ cycling proteins also regulate membrane potential via Ca²⁺-modulation of surface membrane electrogenic molecules. The present results show how manipulations that specifically target the Ca²⁺ cycling subsystem, i.e. suppression of SERCA function by CPA, affect the entire coupled-system and lead to decrease in the spontaneous beating of SANC.

Though a moderate CPA-induced reduction of guinea-pig isolated SA node or SANC beating rate has been noted before⁶, neither the full range of this effect had been studied, nor the mechanisms of CPA-induced suppression of the SANC beating rate. The experimentally observed CPA-induced decrease in the SANC spontaneous beating rate via modulation of the Ca²⁺-ATPase pumping rate in the present study is faithfully simulated by a novel numerical model of SANC²⁰. Consistent with experimental results, these model simulations predict 40% suppression of the SANC beating rate when Ca²⁺-ATPase pumping rate is decreased to approximately one-third of its basal value²⁰.

The speed at which Ca²⁺ is pumped back in the SR is modulated by phospholamban (PLB), which, in its unphosphorylated state, binds to SERCA and inhibits its function^10,21. In ventricular myocytes, phosphorylation of PLB by cAMP-mediated, PKA-dependent activation relieves SERCA inhibition, leading to acceleration of Ca²⁺ uptake by SR and an increase in the speed of cardiac muscle relaxation^10,16. Although it has been demonstrated that the positive chronotropic effect of β-AR stimulation is critically dependent on its effect to accelerate SR Ca²⁺ cycling,^{3,18,19,22–24} shifting the LCR period and LCR-activated I_NCX current to earlier times during DD,²⁴ the mechanisms responsible for these changes have never been demonstrated. The third novel finding of the present study, which directly monitors changes in the SR Ca²⁺ in SANC, is that an increase in PKA-dependent phosphorylation by β-AR stimulation, indexed by PLB phosphorylation (fig. 8A), decreases the SR refilling time, T-90_SR, and consequently shortens the counterpart of LCR period, the LCD period, producing a decrease in the spontaneous cycle length (Fig. 5C). The effects of β-AR stimulation to decrease the LCD period (Fig. 5C) are due, in part at least, to the increase in the SERCA pumping rate which is relieved by PLB phosphorylation. As noted, PKA-dependent effects on sarcolemmal molecules are also involved, because these effects regulate Ca²⁺ that is provided for the SR for pumping. The important role of PLB phosphorylation in cardiac pacemaker function has also been confirmed in genetically manipulated mice (see Online data supplement).

A comparison of the time of SR Ca²⁺ refilling, T-90_SR, and the decay time of AP-induced Ca²⁺ transient, T-90_C, demonstrates that the latter faithfully reproduce the dynamics of SR Ca ²⁺ refilling and can be used to index it (Fig. 4). This observation has also been confirmed during β-AR stimulation, i.e. changes in cytosolic Ca²⁺, T-90_C and LCR period (Online data supplement figure II) are fully consistent with the data in figure 5, which illustrates changes in the SR Ca²⁺ refilling time, T-90_SR, and LCD period. Inhibition of PKA-dependent phosphorylation by PKI decreases PLB phosphorylation (Fig. 8) and markedly increases the SR refilling time, indexed by T-90_C (Fig. 7). The decrease in PLB phosphorylation, produced by PKI, inhibits SERCA function and SR Ca²⁺ uptake, prolonging SR Ca²⁺ refilling time and increasing both the LCR period and spontaneous cycle length.

In the basal state spontaneous SANC firing is highly regulated by constitutive adenylyl cyclase activity^25,26 and constitutive phosphodiesterase activity¹¹. The concurrent activation of both cAMP production and degradation mechanisms permits rapid responses to signals that change Ca²⁺. When PDE activity is inhibited either by broad-spectrum PDE inhibitor, IBMX, or by PDE3 inhibitor, milrinone, a substantial decrease in the SR refilling time, indexed by T-90_C, accompanied by a decrease in the LCR period is observed (Fig. 6 and Online Figure III, respectively). However, PKA-dependent phosphorylation affects several major targets involved in the regulation of Ca²⁺ cycling and spontaneous SANC beating rate: L-type Ca²⁺ current^11,16,24, PLB^8,11,16 and RyR^8,27,28. Since these proteins operate in concert and are all critically involved in the regulation of Ca²⁺ cycling in SANC, it is a challenging task to estimate individual contributions of each of these players. PKA-dependent phosphorylation of RyR might modulate RyR Ca²⁺ release characteristics and, similar to ventricular myocytes^27,28, contribute to SR Ca²⁺ replenishment and Ca²⁺ cycling in SANC. An increase in cAMP-mediated PKA-dependent phosphorylation produced by β-AR stimulation or PDE inhibition markedly increases L-type Ca²⁺ current (I_Ca,L) amplitude in SANC by 75% and 45% respectively^24,11. An increase in Ca²⁺ influx through I_Ca,L increases the amount of Ca²⁺ available for pumping into SR, contributing to increase of SR Ca²⁺ load^11,16,24. As a result, LCR amplitude and spatial width are markedly increased and LCR-activated inward current via Na⁺-Ca²⁺ exchanger is substantially augmented^24,11, leading to an increase in the spontaneous SANC beating rate. The present study demonstrates a key role of SR Ca²⁺ replenishment, controlled by SERCA function and its regulation by PLB phosphorylation, to modulate the SERCA pumping rate and SR Ca²⁺ load to control the LCR period.

To delineate the relative contributions of the SR Ca²⁺ pump and ionic channels to the positive chronotropic effect of PDE inhibition, experimentally measured amplifications of ionic currents produced by PDE inhibition, i.e. I_Ca,L increased by 45%¹¹, I_K increased by 12%¹¹, have been introduced in our new numerical model²⁰. The model faithfully simulated the experimentally observed ~50% increase in the spontaneous SANC beating rate when Ca²⁺-ATPase pumping rate was increased by ~4-fold²⁰. However, in the absence of changes in the SR Ca²⁺-ATPase pumping rate, changes in membrane currents alone produced only a modest ~13% increase in the spontaneous SANC beating rate, supporting the interpretation of the present experimental results and confirming a pivotal role of Ca²⁺-ATPase pumping rate and SR refilling time in the modulation of the LCR period and spontaneous SANC beating rate.

In summary, present study shows, for the first time, that the SR Ca²⁺ refilling time is a key parameter in regulation of the LCR period and the spontaneous SANC cycle length. Gradations in SR Ca²⁺ refilling time via PKA-dependent phosphorylation, indexed by PLB phosphorylation at Ser¹⁶ site or direct SERCA inhibition by CPA, are closely linked to gradations in the LCR period and cycle length across the wide physiological range of the SANC beating rate.

Novelty and Significance.

What is Known?

• For over sixty years, a prevailing view has been that the physiologic timekeeping mechanism of the heart’s pacemaker resides primarily in an ensemble of surface membrane ion transport proteins.

• More recently, the view that the sarcoplasmic reticulum (SR) of sinoatrial nodal cells (SANC) functions as a “Ca²⁺ clock” by generating rhythmic local Ca²⁺ releases (LCRs) beneath the cell membrane has emerged as an important player in pacemaker timing.

• Thus, the timekeeping mechanism of the heart’s pacemaker cells is regulated by a robust, coupled-clock system involving surface membrane and intracellular oscillators.

What New Information Does the Article Contribute?

• By directly measuring local SR Ca²⁺ depletion and refilling time, the rate at which the SR refills with Ca²⁺ following the prior AP-triggered Ca²⁺ release is shown to regulate the LCR period

• The Ca²⁺ refilling time of the SR can be inferred from the decay kinetics of the cytosolic Ca²⁺ transient to predict changes in LCR period and spontaneous AP cycle length in response to β-adrenergic receptor stimulation (β-ARs).

Summary

The present study presents the first direct evidence that the LCR period (i.e., the restitution time for LCR occurrence following a prior AP, or the Ca²⁺ clock’s “ticking speed,”) is determined by the speed at which SR refills with Ca²⁺. Changes in the LCR period (e.g. in response to G-protein-coupled receptor stimulation) induce changes in the timing of LCR-activated inward Na⁺-Ca²⁺ current, leading to changes in the spontaneous depolarization of SANC, resulting in variations of the spontaneous AP cycle length. The findings of the present study should motivate further studies of how the integrated response of Ca²⁺ handling proteins, including L-type Ca²⁺ channels, ryanodine receptors, and SR Ca²⁺ pumps, contribute to the SR Ca²⁺ refilling time and, consequently, the rate of pacemaker firing. Overall the work elucidates the molecular interactions that determine the complex coupled clock systems that robustly regulate basal SANC AP firing rate, and yet are sufficiently flexible to respond to neurotransmitter input to shift the clock’s ticking speed. These findings could help in the development of a rational design of biologic pacemakers that can be translated into therapies for patients with heart rate and rhythm disorders.

List of non-standard abbreviations and acronyms

AC

adenylyl cyclase

AP

action potential

AR

adrenergic receptor

CPA

cyclopiazonic acid

DD

diastolic depolarization

IBMX

3-Isobutyl-1-methylxanthine

I_NCX

Na⁺/Ca²⁺ exchange current

ISO

isoproterenol

LCR

subsarcolemmal local Ca²⁺ releases

PDE

phosphodiesterase

RyR

ryanodine receptor

SA

sinoatrial

SANC

sinoatrial nodal cell

SERCA

sarco-/endoplasmic reticulum Ca²⁺/ATPase

SR

sarcoplasmic reticulum