Termination of Ca²⁺ release by a local inactivation of ryanodine receptors in cardiac myocytes

Abstract

In heart, a robust regulatory mechanism is required to counteract the regenerative Ca²⁺-induced Ca²⁺ release from the sarcoplasmic reticulum. Several mechanisms, including inactivation, adaptation, and stochastic closing of ryanodine receptors (RyRs) have been proposed, but no conclusive evidence has yet been provided. We probed the termination process of Ca²⁺ release by using a technique of imaging local Ca²⁺ release, or “Ca²⁺ spikes”, at subcellular sites; and we tracked the kinetics of Ca²⁺ release triggered by L-type Ca²⁺ channels. At 0 mV, Ca²⁺ release occurred and terminated within 40 ms after the onset of clamp pulses (0 mV). Increasing the open-duration and promoting the reopenings of Ca²⁺ channels with the Ca²⁺ channel agonist, FPL64176, did not prolong or trigger secondary Ca²⁺ spikes, even though two-thirds of the sarcoplasmic reticulum Ca²⁺ remained available for release. Latency of Ca²⁺ spikes coincided with the first openings but not with the reopenings of L-type Ca²⁺ channels. After an initial maximal release, even a multi-fold increase in unitary Ca²⁺ current induced by a hyperpolarization to −120 mV failed to trigger additional release, indicating absolute refractoriness of RyRs. When the release was submaximal (e.g., at +30 mV), tail currents did activate additional Ca²⁺ spikes; confocal images revealed that they originated from RyRs unfired during depolarization. These results indicate that Ca²⁺ release is terminated primarily by a highly localized, use-dependent inactivation of RyRs but not by the stochastic closing or adaptation of RyRs in intact ventricular myocytes.

In cardiac myocytes, Ca²⁺ release from sarcoplasmic reticulum (SR) is activated by the Ca²⁺-induced Ca²⁺ release (CICR) mechanism (1–3). Recent evidence suggests that sarcolemmal L-type Ca²⁺ channels are closely associated with a cluster of sarcoplasmic Ca²⁺ release channels, called ryanodine receptors (RyRs), in the diadic junctions forming discrete release units (4–10). According to the “cluster bomb” model (11), Ca²⁺ influx via Ca²⁺ channels serves as the local Ca²⁺ signal to activate the coupled RyR(s), causing further increase in local [Ca²⁺] and cross-activation of other RyRs within the release unit. These local release events have been visualized directly as “Ca²⁺ sparks” (12), and the recruitment of these events, as a function of Ca²⁺ channel activation and amplitude of unitary Ca²⁺ current (i_Ca), underlies the whole-cell Ca²⁺ transient induced by Ca²⁺ current (I_Ca) (13–15).

Despite improved understanding of the activation process, how Ca²⁺ release is terminated remains unclear. CICR, with its inherent positive feedback, is expected to operate in an “all-or-none” fashion. Regenerative activation of multiple RyRs within the release units should result in long-lasting Ca²⁺ sparks and unstable global Ca²⁺ oscillations. However, Ca²⁺ sparks of usually short duration (half-time of decay ≈22 ms) (12–14), and Ca²⁺ transients of graded amplitude and robust global stability occur in intact myocytes (16, 17), indicate that a regulatory mechanism(s) must exist to interrupt regenerative Ca²⁺ release. Several mechanisms have been proposed for the termination of Ca²⁺ release. (i) Ca²⁺-induced inactivation of Ca²⁺ release (2, 18): Binding of released Ca²⁺ to an inactivation site of RyRs shifts the channels to an inactivated state and shuts off Ca²⁺ release. This was originally put forward by Fabiato (2), based on studies in skinned myocytes, as the mechanism to counteract CICR. However, this process has not been demonstrated in intact myocytes (19). (ii) Adaptation of RyRs (20–22): Spontaneous decline in open probability of RyR channels in lipid bilayers after activation by a step increase in [Ca²⁺]; contrary to inactivation, the adapted channels can be reactivated by subsequent steps to higher [Ca²⁺]. (iii) Stochastic attrition (11): Simultaneous stochastic closing of RyRs in an active release unit of a few RyRs, reducing local [Ca²⁺] to a sub-threshold level and thereby extinguishing Ca²⁺ release. In addition, depletion or reduction of SR Ca²⁺ also may terminate Ca²⁺ release due to the lack of releasable Ca²⁺ or reduction in the gain of CICR (23, 24). To date, no conclusive supporting evidence has been provided for these putative mechanisms.

To elucidate the mechanistic nature of Ca²⁺ release termination, we applied a confocal-imaging technique (25), which enables both spatial and temporal resolution of local SR Ca²⁺ release fluxes, or “Ca²⁺ spikes”, at the t-tubular-SR junctional regions during strong depolarizations, to track the kinetics of local Ca²⁺ release; and we made use of the Ca²⁺ channel agonist FPL64176 (FPL) and specific voltage-clamp protocols to manipulate the properties of single Ca²⁺ channel currents for triggering Ca²⁺ release. We tested specifically (i) whether increasing open-duration and enhancing reopenings of Ca²⁺ channels prolongs Ca²⁺ release and triggers secondary Ca²⁺ release, respectively, (ii) whether the SR Ca²⁺ is depleted after a maximal Ca²⁺ release induced by I_Ca, and (iii) whether RyRs can be reactivated by a stronger Ca²⁺ stimulus immediately subsequent to a prior activation. Our results provide direct evidence indicating that Ca²⁺ release is terminated mainly by a use-dependent inactivation of RyRs, whereas stochastic closing or adaptation of RyRs, or depletion of SR Ca²⁺ is not the primary cause of release termination in intact ventricular myocytes.

MATERIALS AND METHODS

Single Channel Recordings.

Ventricular myocytes were enzymatically isolated from adult male Wistar rats (200–250 g). Cell-attached, patch-clamp recordings of L-type Ca²⁺ channels were obtained by using Sylgard (Dow Corning Midland, MI)-coated, thick-walled borosilicate glass pipettes (5–7 MΩ). Pipettes were filled with solution contained (in mM): 0.01 FPL, 10 CaCl₂, 130 tetraethylammonium chloride (TEA-Cl), and 10 Hepes (pH 7.4). Cells were bathed in high [K] solution, contained (in mM): 110 potassium aspartate, 30 KCl, 1 MgCl₂, 5 Hepes, 5 EGTA, and 3 Na₂ATP, pH 7.4, (free [Ca²⁺] = 100 nM) to approximately zero the membrane potential and enabled estimation of transpatch potentials. Unitary currents were recorded by using a cooled capacitor-feedback headstage (CV203B) and Axopatch 200B amplifier (Axon Instruments, Foster City, CA), low pass-filtered at 1 kHz, and digitized at 10 kHz. Data were collected and analyzed by using the pCLAMP software (Axon Instruments).

Simultaneous Measurement of I_Ca and SR Ca²⁺ Release Fluxes.

Myocytes were whole-cell voltage clamped with patch pipettes with tip resistance of 1.5–2.5 MΩ, and superfused with Tyrode’s solution containing (in mM): 137 NaCl, 2 CaCl₂, 5.4 KCl, 1 MgCl₂, 10 glucose, and 10 Hepes at pH 7.4, with 20 μM tetrodotoxin to block the sodium current. Membrane currents were measured with an Axopatch 200B patch-clamp amplifier. SR Ca²⁺ release fluxes were detected simultaneously with a novel laser confocal-imaging technique (25), by using the low affinity Ca²⁺ sensitive dye, Oregon Green 488 BAPTA-5N (OG-5N, Molecular Probes) in conjunction with high [EGTA], to minimize the resident time of free-released Ca²⁺ in the cytoplasm and to optimize the detection of localized high [Ca²⁺] in the release sites. The pipette solution contained (in mM): 105 CsCl, 10 NaCl, 5 MgATP, 10 Hepes, 20 TEA-Cl, 4 EGTA, 2 CaCl₂, and 1 OG-5N at pH 7.2, to eliminate K⁺ currents and buffer-free [Ca²⁺] at 150 nM for adequate Ca²⁺ loading of SR. EGTA (4 mM) has no significant effect on the local Ca²⁺ signaling between L-type Ca²⁺ channels and RyRs (9). Confocal images were acquired by using a Zeiss LSM-410 inverted confocal microscope with a Zeiss Plan-Neofluor 40x oil immersion objective (NA = 1.3), and the confocal pinhole was set to render spatial resolutions of 0.4 μm in the x–y axis and 0.9 μm in the z-axis. OG-5N was excited by the 488 nm line of an argon laser, and fluorescence was measured at >515 nm. Images were taken in the line-scan mode, with 512 pixels/line (0.104 μm/pixel) scanned at 2.09 ms intervals, and processed by using idl software (Research System, Boulder, CO). Conventional whole-cell Ca²⁺ transients were performed in some myocytes with methods described previously (9). All external solutions bathing the myocytes were exchanged rapidly (≈200 ms) with a concentration-clamp system to avoid changes in SR loading, and caffeine was rapidly applied by using a pico-spritzer. All experiments were performed at room temperature (20–22°C).

Data Analysis.

Single Ca²⁺ channel records were leak and capacitive current eliminated by subtracting the original traces with blank sweeps. Open events were idealized by half-height criteria, and single channel patches were verified by the absence of stacked openings in entire data sets (>900 sweeps). Amplitudes of i_Ca were estimated from the Gaussian distributions of single channel currents, and open-time distribution was fitted with a bi-exponential probability distribution function. Latency distributions and averaged currents of the first openings and reopenings of Ca²⁺ channels were constructed from idealized events. Ca²⁺ release-induced inactivation of I_Ca was quantified as the fraction of peak I_Ca at 25 ms of clamp pulses (I_{25 ms}/I_peak) and compared before and after complete depletion of SR Ca²⁺ with 10 mM caffeine. SR Ca²⁺ release fluxes were determined from line-scan confocal images as described by Song et al. (25). Briefly, spatially averaged OG-5N fluorescent signals from confocal images were normalized with basal fluorescence and expressed as F/F₀. The change in the OG-5N signal (ΔF/F₀), in the presence of high EGTA concentration, is the sum of two components, a prominent spike component directly related to SR Ca²⁺ release fluxes (f_r), and a small pedestal component (f_s) representing the weighted running integral of the release fluxes, thus ΔF/F₀ = f_r + f_s, where f_s = α∫f_rdt. f_s and α were determined experimentally by fitting f_s to the pedestal level of ΔF/F₀ after repolarization, at which Ca²⁺ release was expected to be zero. f_r was then generated by subtracting f_s from the ΔF/F₀ trace. Numerical simulation by using realistic buffer kinetics and concentrations showed that Ca²⁺ spikes reproduced well the waveform of Ca²⁺ fluxes, with an “on” and “off” response time of <1 ms, and the amplitude of the spike was linearly related to Ca²⁺ release flux over a wide range (25). All data were expressed as mean ± SEM and were compared by using paired t tests. P values <0.05 were considered statistically significant.

RESULTS AND DISCUSSIONS

Stochastic Closing of RyRs and Termination of SR Ca²⁺ Release.

Previous numerical analysis of whole-cell Ca²⁺ transients (26, 27) and our recent direct measurement of SR Ca²⁺ release fluxes (25) revealed that Ca²⁺ release occurs and terminates shortly after the onset of a depolarizing pulses. Because L-type Ca²⁺ channel openings are brief (28–30), and reopenings of Ca²⁺ channels are inhibited to a large extent by the inactivation induced by Ca²⁺ release from SR (8–10) and Ca²⁺ influx via Ca²⁺ channels (29, 30), the transient nature of SR Ca²⁺ release may simply reflect the random stochastic closing of RyRs in the absence of a sustained trigger (11). In this scenario, an increase in the open duration of Ca²⁺ channels should provide a more prolonged stimulus to sustain the regenerative activation of RyRs; and reopenings of Ca²⁺ channels after release termination should provide new stimuli to reactivate the extinguished RyRs. To test this paradigm, the Ca²⁺ channel agonist FPL64176 (FPL) (31, 32) was used to prolong the open duration and enhance the reopenings of Ca²⁺ channels. Fig. 1 shows single Ca²⁺ channel currents recorded under cell-attached mode with Ca²⁺ (10 mM) as the charge carrier in the presence of FPL (10 μM). After depolarization to 0 mV, the probability of an active sweep was 0.60 ± 0.05, and open probability (P_o) of the active sweeps was 0.28 ± 0.04 (n = 3). The high P_o was mainly due to a prolonged open lifetime (mean open time = 15.9 ± 2.9 ms, n = 3 patches) (Fig. 1 A and B), which was ≈2 orders of magnitude longer than those recorded without FPL (mean open-time = 0.27 ms) (28). Latency distributions show that the first openings dominated the first 10 ms, and reopenings of Ca²⁺ channels constituted virtually all of the events after 30 ms of the clamp pulse, giving rise to the prominent maintained phase of I_Ca (Fig. 1 C and D). Additionally, FPL abolished the inactivation of Ca²⁺ channel induced by Ca²⁺ release. Under control conditions, whole-cell I_Ca elicited at 0 mV inactivated rapidly (I_{25 ms}/I_peak = 0.13 ± 0.01, n = 13); the rate of inactivation was significantly reduced when SR Ca²⁺ release was abolished by 10 mM caffeine (I_{25 ms}/I_peak = 0.28 ± 0.02, P < 0.05). In the presence of FPL, inactivation of I_Ca was prolonged significantly (I_{25 ms}/I_peak = 0.71 ± 0.03, n = 13); inhibition of Ca²⁺ release with caffeine had minimal effect on the maintained phase and the rate of inactivation of I_Ca (I_{25 ms}/I_peak = 0.73 ± 0.03, n = 13) (Fig. 1 E and F). Reopenings of Ca²⁺ channels, thus, were not prevented by Ca²⁺ release triggered by their own first openings in the presence of FPL and provided multiple stimuli to the coupled RyRs during a depolarizing pulse.

Figure 1.

FPL increases open probability, prolongs open duration, and prevents Ca²⁺ release-induced inactivation of Ca²⁺ channel. (A) Representative traces of unitary Ca²⁺ current recorded from a single channel patch at 0 mV, using 10 mM Ca²⁺ as the charge carrier in the presence of 10 μM FPL. (A Bottom) the ensemble averaged current of 925 sweeps. Notice the prolonged openings and reopenings during the pulse. (B) Open-time histogram of 1,907 opening events, bin width = 0.4 ms. The smooth line represents the best-fitted probability distribution function and τs denote the time constants. (C) Latency distribution of the first openings and reopenings of Ca²⁺ channel, bin width = 1 ms. Number of first openings is 637 and reopenings is 1,270. (D) Currents of the first opening (I_1st) and reopenings (I_R) of Ca²⁺ channels, reconstructed from idealized events. (A to D) Generated from the same patch. (E) Whole-cell I_Ca and Ca²⁺ transients in the presence of 3 μM FPL, before and during superfusion of 10 mM caffeine. (F) Inactivation of whole-cell I_Ca, quantified as I_{25 ms}/I_peak, before and during superfusion of 10 mM caffeine, in control and FPL-treated (n = 13) myocytes. ∗, Significant difference from control (P < 0.05).

Confocal images of SR Ca²⁺ release fluxes showed that depolarizations to 0 mV activated spatially discrete, localized Ca²⁺ spikes, which occurred and terminated within 40 ms after the onset of clamp pulse (Fig. 2A). They usually occurred only once at each release site (≈1.8 μm apart), without reactivation during the later part of depolarization. Estimations based on anatomical data (5, 33, 34) suggest that a resolvable volume (0.144 μm³) of confocal images encompasses multiple diadic junctions; hence, a Ca²⁺ spikes represents the ensemble Ca²⁺ release fluxes originated from multiple release units within the same site. Rapid application of FPL (10 μM, 10 s before a clamp-pulse) caused an immediate enhancement of I_Ca (I_peak = 0.92 ± 0.11 nA in control and 3.64 ± 0.34 nA in FPL, n = 10, P < 0.05), a slowing of its inactivation, and an increase in the spatially averaged release transient (ΔF/F₀ = 0.18 ± 0.02 in control, and 0.23 ± 0.02 in FPL, n = 10, P < 0.05). Surprisingly, the time course of Ca²⁺ release was unaltered (time to peak = 17.2 ± 1.2 ms in control and 15.2 ± 0.9 ms in FPL; time to 75% relaxation = 35.9 ± 2.1 ms in control and 34.2 ± 2.6 ms in FPL, n = 10); and no major secondary Ca²⁺ release was observed despite the presence of a prominent maintained I_Ca. The disparity between the kinetics of I_Ca and Ca²⁺ release is illustrated by superimposing the peak normalized I_Ca and spatially averaged Ca²⁺ release transients (Fig. 2A Right Bottom). These results were confirmed by using the first derivative of “conventional” Indo-1 Ca²⁺ transient (d[Ca²⁺]/dt), as an empirical indicator of Ca²⁺ release (data not shown). Latency analysis shows that the occurrence of Ca²⁺ spikes, mostly within 2–10 ms of depolarization, coincided with the first latency of Ca²⁺ channels, but was completely dissociated from Ca²⁺ channel reopenings (Fig. 2B). These results argue against the stochastic attrition (11) as the primary mechanism for terminating Ca²⁺ release, because it predicts an increase in open duration of L-type Ca²⁺ channel would prolong Ca²⁺ release, and multiple Ca²⁺ channel reopenings would simply give rise to multiple release events. Moreover, the finding that Ca²⁺ release at 0 mV was activated exclusively by the first openings of Ca²⁺ channels (35) is consistent with the previous finding that Ca²⁺ release is gated by the initial phase of I_Ca (7, 14, 26, 27, 31, 36), but is in contrast to the observation that the latency distribution of Ca²⁺ sparks in the presence of Ca²⁺ channel blocker resembles the time course of I_Ca (13, 15). In the latter case, however, only a few Ca²⁺ release units were triggered at the onset, hence leaving plenty of unfired RyRs for activation in the later part of depolarizing pulses.

Figure 2.

Effect of FPL on SR Ca²⁺ release. (A) Ca²⁺ release immediately before (Left column) and during the first depolarizing pulse (Right column) after application of 10 μM FPL. From Top to Bottom: voltage protocols, confocal line-scan images, normalized spatially averaged OG-5N signals (F/F₀), I_Ca, and superimposed peak normalized I_Ca (black traces) and SR Ca²⁺ release function (f_r, red traces). Vertical and horizontal axes of line-scan images are axes of space and time, respectively. Smooth lines in the F/F₀ panels represent f_s and red lines in the Bottom panels represent f_r generated by subtracting f_s from the ΔF/F₀ traces. (B) Latency distribution of Ca²⁺ spikes (n = 376) elicited in 10 myocytes during the first and second pulses after application of FPL (10 μM). The red and blue lines are the scaled latency distributions of the first openings and reopenings, respectively, of Ca²⁺ channel of Fig. 1C.

Recovery of Ca²⁺ release also was examined by applying a second depolarizing pulse at different intervals (50–1200 ms) after a maximal initial release at 0 mV. An apparent absolute refractory period of ≈150 ms was observed, followed by a second phase of recovery of Ca²⁺ release with a half-time of ≈500 ms (data not shown), similar to the previous observation on the interactions of evoked Ca²⁺ release with Ca²⁺ waves (38). The refractory period after a maximal Ca²⁺ release was, hence, significantly longer than that following a spontaneous spark (≈30 ms) (39) generated by only a single/few RyRs.

SR Ca²⁺ Depletion and Termination of Ca²⁺ Release.

The refractoriness of SR after a maximal Ca²⁺ release could be due to global or local depletion of SR Ca²⁺. To explore this possibility, high concentration of caffeine was applied to myocytes, which was exposed to FPL for 10 s, via a pico-spritzer at 40 ms after depolarization to cause complete release of Ca²⁺ from the SR. The rapid injection of caffeine caused a large release of Ca²⁺ during the otherwise silent later period of the clamp pulse (Fig. 3A), with Ca²⁺ spikes occurring at sites activated by I_Ca before the caffeine application. The amount of Ca²⁺ released by I_Ca compared with caffeine, quantified by integrating their respective release transients, had a ratio of 1:2.2 (Fig. 3B). Assuming the total releasable SR Ca²⁺ equals the sum of Ca²⁺ released by I_Ca and caffeine, the fractional release of Ca²⁺ induced by I_Ca was 33.1 ± 3.4% (n = 6), consistent with previous estimations (24, 40). This indicates that SR Ca²⁺ was not depleted after the initial release, and a substantial amount of Ca²⁺ was immediately available for release when RyRs were allowed to reopen. Moreover, the reduction of SR Ca²⁺ by one-third was unlikely the primary mechanism for the refractoriness of SR because I_Ca is able to trigger Ca²⁺ release after a similar reduction of Ca²⁺ loading (40), and Ca²⁺ release can be elicited within 2–3 depolarizing pulses immediately after complete depletion of SR by caffeine(9, 17, 40), suggesting that SR is capable of releasing Ca²⁺ with an even lower Ca²⁺ content (less than two-thirds of normal load). Moreover, spontaneous Ca²⁺ sparks were observed after 64% reduction in SR Ca²⁺ loading (41). However, the data do not exclude the possibility that the partial SR Ca²⁺ depletion may play a contributing role in terminating Ca²⁺ release by altering RyR activities (42, 43), and reducing the gain of CICR (24).

Figure 3.

Caffeine-induced Ca²⁺ release during the maintained phase of I_Ca. (A Top to Bottom) voltage protocol, confocal line-scan image, spatially averaged OG-5N fluorescence signal, and the simultaneously measured I_Ca. FPL (10 μM) was applied 10 s before depolarization, and caffeine was rapidly injected onto the cell 40 ms after the onset of depolarization. The smooth line in the Middle panel represents f_s. (B) Amount of SR Ca²⁺ released by I_Ca and caffeine quantified by integrating their respective release fluxes (f_r) transients (n = 6).

Inactivation vs. Adaptation of RyRs in Terminating SR Ca²⁺ Release.

The refractoriness of SR to reopenings of L-type Ca²⁺ channels could be due to RyR adaptation or inactivation. To distinguish between these two possibilities, we devised a voltage-clamp protocol to test whether the once fired RyRs could be reactivated by a stronger Ca²⁺ stimulus, as would be predicted by the adaptation, but not by the inactivation hypothesis. Depolarizing pulses to either −30, 0, +30, or 60 mV were applied for 50 ms, followed by a hyperpolarization to −120 mV. Single channel recordings in the presence of FPL showed that i_Ca at −30, 0, and +30 mV were −0.42, −0.21, and −0.06 pA, respectively. Hyperpolarizing steps to −120 mV caused an instantaneous jump of i_Ca to −1.00 pA (Fig. 4A) due to an increase in the electrochemical gradient for Ca²⁺ influx. According to the adaptation hypothesis, the multi-fold increase in i_Ca during the hyperpolarizing steps should provide a local [Ca²⁺] sufficient to reactivate the adapted RyRs (20), even though these RyRs no longer respond to the smaller i_Ca of Ca²⁺ channel reopenings during depolarization. However, when this protocol was applied to intact myocytes, hyperpolarizing steps subsequent to the maximal Ca²⁺ releases elicited by depolarizations to −30 and 0 mV failed to trigger, or only activated a minimal Ca²⁺ release, despite the activation of a larger tail than initial I_Ca (Fig. 4 B and C). This result argues against the adaptation of RyRs but supports the hypothesis of strong inactivation of RyRs after their initial activation.

Figure 4.

Single channel current and SR Ca²⁺ release induced by depolarization and subsequent hyperpolarization. (A) Representative traces of single Ca²⁺ channel current recorded in the presence of 10 μM FPL, during depolarizing pulses (50 ms) to −30, 0, and +30 mV followed by hyperpolarizing steps to −120 mV. C, O, and T, mark the i_Ca when the channel was closed, open, and open during hyperpolarization (tail opening), respectively. Notice the multi-fold step increase in i_Ca upon hyperpolarization. (B) Representative line-scan images, spatially averaged OG-5N fluorescence signals and I_Ca simultaneously recorded in a myocyte. (C) Peak Ca²⁺ release fluxes elicited by a depolarizing pulse (to −30, 0, +30, or +60 mV, black bars) followed by a hyperpolarizing step (gray bars). (D) Cumulative Ca²⁺ release elicited by both depolarizing and hyperpolarizing pulses quantified by integrating f_r. Bar graphs C and D are the averaged data from five cells.

Yasui et al. (22) showed that depolarization to +30 mV, in the presence of FPL, elicits a transient Ca²⁺ release that terminates despite continued I_Ca, yet additional Ca²⁺ release is triggered by the tail current after repolarization; this has been interpreted as the evidence for adaptation of RyRs in situ. When our myocytes were first depolarized to +30 (or +60 mV) to activate a submaximal Ca²⁺ release, the subsequent hyperpolarizing step indeed triggered a “tail” release transient (Fig. 4B Right). However, the total amount of Ca²⁺ released by the depolarizing pulse and the subsequent hyperpolarizing step was similar to the maximal Ca²⁺ release at −30 or 0 mV (Fig. 4D), indicating that the large tail i_Ca did not trigger additional Ca²⁺ release from RyRs that fired during depolarization; rather, the tail transients likely represent the activation of release units that were not opened by the small i_Ca during the submaximal initial release. Because a Ca²⁺ spike is the ensemble release fluxes from multiple release units within a junctional site, in the latter case, the tail currents should elicit larger Ca²⁺ spikes in junctional sites where only a few release units were activated by the preceding depolarizing pulse, and trigger smaller Ca²⁺ release in sites where most of the release units were fired during the initial activation. Indeed, such a complementary release pattern was apparent at individual junctional sites: large initial Ca²⁺ spikes were often followed by either no or small tail releases at the same sites (Fig. 5A, sites 2, 4, and 6), whereas small or no initial releases were usually associated with larger tail Ca²⁺ spikes (Fig. 5A, sites 1, 3, and 5). Quantitatively, the amplitude of the initial Ca²⁺ spikes, elicited in 34 different release sites, correlated negatively (r = −0.44, n = 170, P < 0.001) with the tail Ca²⁺ spikes (Fig. 5B). These results clearly indicate that the large tail Ca²⁺ spikes were not generated by the reactivation of adapted RyRs because they were originated from sites where the initial activation of RyR was absence or minimal. The tail current failed to elicit secondary Ca²⁺ spikes from sites of large initial release further supports the inactivation of once activated RyRs at subcellular release sites.

Figure 5.

Local Ca²⁺ release transients elicited at individual t-tubular-SR junctional sites in the presence of 10 μM FPL. (A) Confocal line-scan image (Upper) and Ca²⁺ spikes (Lower) elicited by a single depolarizing pulse to +30 mV for 50 ms, followed by repolarization to −60 mV. Numbers in the image indicate the spatial locations of the six individual release sites at which the Ca²⁺ spikes displayed in the Lower panel were recorded. (B) Scatter plot of amplitudes of depolarization-induced Ca²⁺ spikes vs. amplitudes of those triggered by the subsequent repolarization. The straight line is the least square linear regression of the data, r is the correlation coefficient, and n = 170.

Concluding Remarks.

Using the imaging technique to directly measure SR Ca²⁺ release fluxes, in conjunction with the unique agonist, FPL64176, to manipulate the gating properties of L-type Ca²⁺ channels, we have provided compelling evidence that SR Ca²⁺ release during excitation–contraction coupling is terminated mainly by a local inactivation of RyRs in intact myocytes, whereas stochastic attrition, depletion of SR Ca²⁺, and the adaptation of RyRs observed in lipid-layers (20, 21) do not participate or only play a contributing role in terminating Ca²⁺ release in situ. This inactivation of RyRs may depend on the high local [Ca²⁺]consequential to their own Ca²⁺ release, as suggested previously in skinned fibers (2, 18), SR vesicles (44), and in single RyRs in lipid bilayers (45–47), as well as recently in intact myocytes (48), showing that the rate of Ca²⁺ spark termination is related to the magnitude of release flux. However, the possibility that the process is obligated to the activation of RyRs per se (49) cannot be excluded. Nevertheless, the inactivation process is highly localized, and use-dependent, and may be modulated, e.g., by cyclic adenosine monophosphate (21), Ca²⁺/calmodulin-dependent protein kinases (50, 51), or FK506-binding protein (52–54). Because of this use-dependent inactivation, RyRs once activated are precluded from reactivation during a cardiac cycle; therefore, it interrupts the positive feedback of CICR, providing both global and micro-stability in cardiac myocytes. Since the mobilization of Ca²⁺ from intracellular stores through Ca²⁺ release channels is pivotal for signal transduction in many cell types, our findings as well as our experimental approaches may have important applications to studies of Ca²⁺ signaling in general.

ABBREVIATIONS

RyRs

ryanodine receptors

SR

sarcoplasmic reticulum

CICR

Ca²⁺-induced Ca²⁺ release

FPL

Ca²⁺ channel agonist, FPL64176

OG-5N

Oregon Green 488 BAPTA-5N

i_Ca

unitary Ca²⁺ current

I_Ca

whole-cell Ca²⁺ current