INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

CONTRACTION IN HEART is initiated by release of Ca²⁺ stores from the sarcoplasmic reticulum (SR). Two fundamentally different processes that trigger release of SR Ca²⁺ have been proposed. Ca²⁺ release can be initiated in response to influx of trigger Ca²⁺. This process is known as Ca²⁺-induced Ca²⁺ release (CICR; see Ref. 9) and is thought mainly to be linked to Ca²⁺ influx through L-type Ca²⁺ channels in the sarcolemma (1, 3, 6, 7, 23, 26), although Ca²⁺ entry via Na⁺/Ca²⁺ exchange (Na/Ca_EX) may contribute under some conditions (19, 21, 22, 27, 39). Contractions and Ca²⁺ transients initiated by this mechanism typically are proportional to the magnitude of L-type Ca²⁺ current (I_Ca-L; see Refs. 2 and 3). Our recent studies have provided evidence for a second mechanism, a voltage-sensitive release mechanism (VSRM), that links release of SR Ca²⁺ to depolarization of the sarcolemma (11, 13, 15, 16). The VSRM continues to operate when measurable influx of Ca²⁺ has been eliminated, and contractions and Ca²⁺ transients initiated by the VSRM are not proportional to the magnitude of I_Ca-L (11, 13, 15, 16). The electrophysiological characteristics of CICR and the VSRM predict that both mechanisms would be triggered by the cardiac action potential and therefore likely contribute to initiation of contraction in heart.

Relaxation of contraction takes place when free intracellular Ca²⁺ levels decrease. Ca²⁺ levels are decreased primarily by uptake of Ca²⁺ into the SR but also by extrusion of Ca²⁺ by sarcolemmal Na/Ca_EX, and to a lesser extent by a sarcolemmal Ca²⁺ ATPase (2). For Ca²⁺ levels to decrease, release also must terminate. It has been proposed that release of SR Ca²⁺ is terminated by inactivation of SR Ca²⁺ release channels (10, 18, 34, 37, 38). Termination of release proceeds in a time-dependent manner and is independent of the initial influx of trigger Ca²⁺ through the sarcolemma (18, 34, 37, 38). Thus relaxation would follow a time course determined by termination of release in combination with removal of Ca²⁺ from the cytosol by the SR Ca²⁺ ATPase, the Na⁺/Ca²⁺ exchanger, and the sarcolemmal ATPase. However, recently, we have observed that contractions and Ca²⁺ transients do not decline completely when depolarization of the sarcolemma is maintained in voltage-clamped cardiac myocytes (12). Sustained responses were observed under conditions that allow activation of both the VSRM and CICR. This observation led us to hypothesize that membrane potential may play a role in termination, as well as initiation of SR Ca²⁺ release. The objectives of this study were to explore the role of membrane potential in regulation of contraction and relaxation in cardiac ventricular cells and to determine whether sustained responses are caused by CICR or by the VSRM.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Studies were conducted within the guidelines published by the Canadian Council on Animal Care and were approved by the Dalhousie University Committee on Animal Care. Guinea pig ventricular myocytes were dissociated enzymatically, as described earlier (13, 16). Cells were superfused (3 ml/min) at 37°C with solution containing (in mM) 50 NaCl, 100 choline chloride, 2 CaCl₂, 4 KCl, 1 MgCl₂, 10 glucose, 10 HEPES (pH 7.4 with NaOH; gassed with 100% O₂), and 250 µM lidocaine and/or 50 µM TTX to block Na⁺ current. In some experiments, rapid solution changes at 37°C were made within 300 ms with a computer-triggered device (15).

In most experiments, discontinuous single-electrode voltage-clamp recordings were made with Axoclamp 2A electronics and high-resistance microelectrodes (18-24 M filled with 2.7 M KCl). In some experiments, discontinuous single-electrode voltage-clamp recordings were made with Axoclamp 2A electronics and 1-3 M patch pipettes that contained (in mM) 60 KCl, 70 potassium aspartate, 0.05 8-bromo-cAMP, 4 MgATP, 1 MgCl₂, 2.5 KH₂PO₄, 0.12 CaCl₂, 0.5 EGTA, and 10 HEPES, pH 7.2 with KOH, as described previously (13). The free Ca²⁺ concentration in the pipette solution was calculated to be 46 nM (Ecal for windows version 1.1; Biosoft, 1996). Liquid junction potentials were compensated before data acquisition.

Unloaded cell shortening was measured with a video edge detector (11, 13, 16). Cell fluorescence was measured with a Photon Technology International DeltaRAM system (Brunswick, NJ). Cells were loaded with 1.0 µM fura 2-AM (Molecular Probes) for 20 min at room temperature. After loading, extracellular dye was eliminated by superfusion of the cells with physiological buffer solution for 20 min. The emission field was limited to the size of a single myocyte with an adjustable window. Cells were excited at 340 nm, and fluorescence emitted by the cell was recorded at 510 nm. Background fluorescence was not subtracted. Changes in intracellular Ca²⁺ were expressed as the ratio of peak fluorescence transient (F) over baseline fluorescence (F₀). Single wavelength excitation was used, since it allowed Ca²⁺ transients to be displayed and recorded with pCLAMP acquisition software along with current and voltage records during the experiments. Recording times with patch pipettes were kept as short as possible to minimize washout of fura 2. Significant washout of fura 2 was not detected in experiments with high-resistance microelectrodes.

Data were acquired and analyzed with pCLAMP 6.01. Recordings were digitized at sample rates up to 50 kHz. Voltage-clamp test steps were preceded by 10 conditioning pulses from a holding potential of 80 to 0 mV, to provide a consistent activation history, followed by repolarization to a postconditioning potential. Further details of specific voltage-clamp protocols are provided in the relevant sections in RESULTS.

Data are presented as means ± SE. Differences between means were assessed with a Student's t-test (P < 0.05 was considered significant). Lidocaine, tetracaine, choline chloride, nickel chloride, and cadmium chloride were purchased from Sigma Chemical. TTX was purchased from Alomone Laboratories, and ryanodine was from Calbiochem. All drugs were dissolved in deionized water.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In initial experiments, contractions were elicited with a voltage-clamp protocol (Fig. 1A) that activates the VSRM and CICR separately (11, 13, 16). This protocol elicits two phasic contractions, the first caused by the VSRM and the second initiated by CICR coupled to I_Ca-L. It is evident in Fig. 1B that phasic contractions initiated by both mechanisms were followed by sustained contractions that lasted for the duration of the depolarizing steps. To determine whether sustained contractions were caused by I_Ca-L, we blocked I_Ca-L with rapid application of 100 µM Cd²⁺. Cd²⁺ inhibited the CICR contraction coupled to I_Ca-L but not the sustained contraction (Fig. 1C). The phasic contraction initiated by the VSRM also remained. Furthermore, the sustained component began with the step that activated the VSRM, rather than the step that initiated I_Ca-L. These observations indicate that it is very unlikely that the sustained contraction was initiated by influx of Ca²⁺ via I_Ca-L.

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Depolarization of cardiac myocytes initiates voltage-dependent sustained contractions that are not blocked by Cd2+. A: voltage-clamp protocol used to separate voltage-sensitive release mechanism (VSRM) and Ca2+-induced Ca2+ release (CICR) contractions. B: representative VSRM and CICR contractions (top) and currents (bottom) initiated by steps to 40 and 0 mV. ICa-L, L-type Ca2+ current. C: rapid switch to 100 µM Cd2+ 3 s before activation steps inhibited ICa-L and the CICR contraction but left the VSRM contraction followed by a sustained contraction that lasted until repolarization. Experiments were conducted with high-resistance microelectrodes.

Persistence of the sustained contraction together with the phasic VSRM contraction during inhibition of I_Ca-L suggests that the sustained contraction also may be induced by the VSRM. It was therefore important to determine whether the magnitude of sustained contractions varied with membrane potential and whether the voltage dependence matched that of the VSRM. To investigate this, we used single voltage-clamp steps to different membrane potentials (Fig. 2A). The steps initiated phasic contractions followed by sustained contractions, both of which became greater with progressively stronger depolarization (Fig. 2B). Mean contraction-voltage relations for phasic and sustained components of contraction are shown in Fig. 2C. Figure 2C shows that the maximum amplitude of the sustained component was less than that of phasic contractions. However, the contraction-voltage relation for sustained contractions appeared to be negative relative to that for phasic contractions. To more easily visualize this, contraction-voltage relations for phasic and sustained contractions were normalized to maximum contraction and plotted in Fig. 2D. The normalized contraction-voltage relations were fitted by Boltzmann functions of the following form: y = (a  b)/{1 + exp[(V  V_h)/k]} + b, where a and b are the maximum and minimum contractions, V is the test potential, V_h is the half-maximal voltage, and k is the slope factor. V_h for the sustained component was 47.4 ± 1.5 mV (n = 11) and was significantly more negative than V_h for the phasic component (38.7 ± 1.3 mV, n = 11; P < 0.05).

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Phasic and sustained components of contraction are initiated by depolarization of cardiac myocytes. A: voltage-clamp steps used to examine voltage dependence of sustained contractions in experiments with high-resistance electrodes. B: contractions were composed of phasic () and sustained () components, both of which increased with depolarization. C: mean contraction-voltage curves for phasic and sustained components of contraction. D: normalized contraction-voltage relations for phasic and sustained contractions were fitted by Boltzmann functions. Half-maximal voltage (Vh) for the sustained component was 47.4 ± 1.5 mV (n = 11 cells) and was significantly different from the Vh for the phasic component (38.7 ± 1.3 mV, n = 11; P < 0.05). The phasic component likely reflects activation of phasic contractions by the VSRM in the more negative range plus activation of CICR by ICa-L as the steps became positive to 30 mV. Thus the Boltzmann relation for the phasic contraction would represent a combination of these two mechanisms of excitation-contraction coupling. E: voltage-clamp protocol used to determine the effects of ryanodine on phasic and sustained components of contraction. F: representative recordings of contraction recorded before and after exposure of a myocyte to 1 µM ryanodine (Rya). Ryanodine abolished the initial phasic contraction plus the sustained component. Con, control.

The occurrence of sustained contractions suggests that release of Ca²⁺ from the SR also may be sustained. To determine whether the sustained component of contraction arises from SR release of Ca²⁺, we superfused myocytes with 1 µM ryanodine, an agent that disrupts SR function (2). Sequential activation steps from 65 to 40 and 0 mV (Fig. 2E) elicited both phasic and tonic contractions before exposure to ryanodine (Fig. 2F). After 15 min exposure to ryanodine, both the phasic VSRM contraction and the sustained contraction were abolished (Fig. 2F). The contraction initiated by I_Ca-L also was attenuated, although a small contraction still remained. Similar results were observed in 12 of 12 myocytes. These observations suggest that the sustained component of contraction is likely caused by SR Ca²⁺ release.

To determine whether sustained elevation of intracellular Ca²⁺ accompanies maintained depolarization, additional experiments were conducted in cells loaded with the Ca²⁺-sensitive dye fura 2. Long (1-s) voltage steps to membrane potentials between 70 and +20 mV initiated Ca²⁺ transients with prominent voltage-dependent sustained components (Fig. 3A). These results confirmed that the sustained component of contraction was initiated by sustained release of SR Ca²⁺. The mean amplitudes of the sustained transients (F/F₀) were plotted as a function of test step voltage and were fitted with a Boltzmann function (Fig. 3B). The mean amplitudes of sustained transients showed a sigmoidal voltage dependence that approached a plateau near 20 mV (Fig. 3B). The voltage dependence of sustained Ca²⁺ transients (Fig. 3B) was virtually identical to that shown in Fig. 2D for sustained contractions. Furthermore, V_h for sustained transients was essentially the same as V_h observed for sustained contractions (Fig. 2D).

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Sustained Ca2+ transients exhibit voltage-dependent activation. A: progressively greater step depolarizations (top) elicited sustained Ca2+ transients (middle) in cells loaded with fura 2. Current records are shown at bottom. Experiments were conducted with high-resistance microelectrodes. f, Change in fluorescence. B: amplitudes of the sustained transients [peak fluorescence transient (F)/baseline fluorescence (F0)] showed a sigmoidal dependence on membrane potential, which was fit with a Boltzmann function (n = 11).

Figure 3A also shows that sustained transients declined only when myocytes were repolarized. To determine whether this decrease in the transient also was graded by membrane potential, we modified the voltage-clamp protocol so that an initial activation step to 0 mV was followed by repolarization to different potentials between +20 and 90 mV (Fig. 4A). With this protocol, the initial activation step always initiated a similar phasic transient. However, this was followed by sustained transients with amplitudes that varied with repolarization to different potentials (Fig. 4A). The amplitudes of the sustained transients were plotted as a function of the voltage of the repolarization step and were fitted with a Boltzmann function (Fig. 4B). The voltage dependence of the sustained transients observed with repolarizing steps was sigmoidal (Fig. 4B) and was virtually identical to that observed with activation steps (Fig. 3B). This observation suggests that sustained transients show deactivation with the same voltage dependence as activation.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Sustained Ca2+ transients exhibit voltage-dependent deactivation. A: when repolarizing steps were made after a fixed activation step to 0 mV (top), voltage-dependent sustained Ca2+ transients exhibited deactivation (middle). B: voltage dependence of sustained transients recorded with repolarizing steps (n = 11) was virtually identical to that of transients elicited with depolarizing steps (not significantly different). Experiments were conducted with high-resistance microelectrodes.

It is unlikely that I_Ca-L contributes to sustained transients initiated with the protocol shown in Fig. 4A, as I_Ca-L should inactivate during the initial depolarizing step and should remain inactivated during maintained depolarizations. To test this, we interpolated an activation step to 0 mV at the end of the repolarization steps as shown in Fig. 5A. The activation step to 0 mV was preceded by a 3-ms return to 70 mV to provide a constant activation step. The magnitude of I_Ca-L elicited by the test step varied in response to changes in preceding potential (Fig. 5A, bottom). Peak inward current was plotted as a function of the preceding conditioning potential in Fig. 5B. A Boltzmann function fitted to these mean data had a V_h of 33.6 mV. The relation between magnitude of peak inward current and the preceding potential demonstrated that I_Ca-L was inactivated at potentials that elicited maximal sustained transients and was only fully available at potentials where sustained transients were minimal (Fig. 5B). Thus the curve describing the voltage dependence of availability of I_Ca-L (Fig. 5B) was opposite to the curve exhibited by sustained transients (Figs. 3B and 4B), which further indicates that a role for I_Ca-L in sustained transients is highly unlikely.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Sustained Ca2+ transients are maximal at potentials that inactivate ICa-L. A: to assess the availability of ICa-L during sustained Ca2+ transients, an activation step to 0 mV was introduced at the end of the deactivation protocol (middle). Sustained transients (top) were maximal at membrane potentials that abolished ICa-L (shown enlarged, bottom). B: mean steady-state inactivation curve for ICa-L (Vh = 33.6 mV, n = 5) showed that ICa-L was inactivated at membrane potentials where the sustained transient was maximal. Experiments were conducted with high-resistance microelectrodes.

To evaluate a possible role of Na/Ca_EX in sustained transients, we used several approaches. First, we determined the voltage dependence of deactivation in cells superfused with extracellular solution that contained 50 or 100 mM extracellular Na⁺. If sustained transients are caused by Na/Ca_EX, the voltage dependence of the transient-voltage relation should shift in response to a twofold change in extracellular Na⁺ concentration (4, 8, 17). Figure 6A shows plots of mean amplitudes of sustained Ca²⁺ transients recorded with the deactivation protocol shown in Fig. 4A. Curves recorded in 50 or 100 mM extracellular Na⁺ were superimposable. V_h for deactivation in 100 mM Na⁺ was 43.0 mV (n = 4) and was not significantly different from V_h of 44.8 mV determined in 50 mM Na⁺ (n = 11). The absence of an effect of doubling of Na⁺ concentration on deactivation indicates that it is unlikely that Na/Ca_EX contributed to these transients.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Sustained Ca2+ transients generated in the present study are not caused by Na+/Ca2+ exchange (Na/CaEX). A: 2-fold changes in extracellular Na+ (Na+ replaced with choline chloride) had no effect on the voltage dependence of deactivation for sustained Ca2+ transients, determined with the protocol shown in Fig. 4A, top. Experiments were conducted with high-resistance microelectrodes. Vh for deactivation in 100 mM Na+ was 43.0 mV (n = 4) and in 50 mM Na+ was 44.8 mV (n = 11; not significantly different). B: sustained Ca2+ transients also could be elicited in myocytes dialyzed with patch pipette solution containing 0 mM Na+. Activation/deactivation curves (n = 7 cells) were the same as those determined in undialyzed cells (Figs. 3B, 4B, and 6A). Experiments were conducted with 1- to 3-M patch pipettes.

A second approach to evaluate the role of Na/Ca_EX in sustained transients was to determine whether sustained transients were present in myocytes dialyzed with 0 mM Na⁺, which has been shown to prevent contractions initiated by reverse Na/Ca_EX (20, 27, 39). Figure 6B, inset, shows that a sustained Ca²⁺ transient was still present in a representative experiment conducted with patch pipettes that contained 0 mM Na⁺. Figure 6B also shows a plot of mean amplitudes of sustained transients, recorded from cells dialyzed with 0 mM Na⁺, as a function of repolarization potential. The protocol used was identical to that shown in Fig. 4A. The deactivation curve (Fig. 6B) exhibited a value of V_h very similar to that determined in undialyzed cells (Figs. 3B, 4B, and 6A). Thus experimental manipulation of intracellular Na⁺ also indicates it is very unlikely that Na/Ca_EX generated these sustained transients.

A third approach to evaluate a possible role of Na/Ca_EX in sustained responses was to compare effects of agents known to inhibit either Na/Ca_EX or the VSRM. Effects of Ni²⁺ and tetracaine were examined on sustained contractions initiated by the voltage-clamp protocol shown in Fig. 7A, inset. An initial test step from 65 to 0 mV was followed by a 7-s maintained repolarization step to 20 mV. This resulted in a rapid phasic contraction followed by a sustained contraction (Fig. 7A). Test agents were applied with the rapid solution changer during the step to 20 mV. Application of 2 mM Ni²⁺, an inhibitor of Na/Ca_EX (17), had little effect on sustained contractions (Fig. 7A). In contrast, application of 200 µM tetracaine (Fig. 7B), which does not block Na/Ca_EX (28, 29) but which inhibits the VSRM (24), caused a prompt but reversible inhibition of the sustained contraction. These results in combination demonstrate that sustained responses are not caused by Na/Ca_EX but are likely caused by the VSRM.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Sustained contractions generated in the present study are inhibited by tetracaine but not Ni2+. A: voltage-clamp protocol (inset) activated both a phasic contraction and a sustained contraction. A rapid switch to solution that contained 2 mM Ni2+, which inhibits Na/CaEX, did not inhibit the sustained contraction when applied during prolonged depolarizations. In this example, the trace corresponding to Ni2+ application is slightly below the control trace and rose slightly during Ni2+ application. However, Ni2+ had no systematic effect on sustained contractions in 20 of 20 cells examined. B: rapid switch to solution that contained 200 µM tetracaine, which blocks the VSRM but not Na/CaEX, reversibly inhibited sustained contractions. Similar effects of tetracaine were observed in 20 of 20 cells. Experiments were conducted with high-resistance microelectrodes.

One new characteristic not previously reported for the VSRM is the ability to maintain continuous release of Ca²⁺. Our previous studies showed that phasic VSRM contractions are subject to steady-state inactivation (13, 16). The existence of sustained contractions and Ca²⁺ transients, however, suggests that there is a noninactivating component of the VSRM. We therefore compared the role of inactivation in determining the magnitude of sustained and phasic VSRM contractions with the protocol shown in Fig. 8A, inset. Here we used an activation step to 35 mV to elicit VSRM contractions selectively (13, 16). To evaluate inactivation, the test step was preceded by conditioning steps to potentials between 30 and 105 mV. The test steps elicited phasic contractions (b) superimposed upon sustained contractions (c) (Fig. 8A). The amplitudes of the phasic contractions varied with preceding conditioning potential, whereas the amplitude of the sustained component was constant (Fig. 8A). The mean amplitudes of phasic contractions (b-c) are plotted as a function of preceding conditioning potential in Fig. 8B. This plot shows that phasic contractions exhibited steady-state inactivation and approached a minimum with conditioning potentials near 30 mV (Fig. 8B). The representative recordings shown in Fig. 8A show that cell length preceding phasic contractions (a) also changed in response to conditioning potential. These changes in cell length (a) measured with respect to a constant reference at d were plotted as a function of conditioning potential in Fig. 8C. The contraction-voltage relation for cell length followed the same activation-deactivation relation as that described for Ca²⁺ transients (Fig. 4B).

View larger version (22K): [in this window] [in a new window]   Fig. 8.   VSRM contractions are composed of a phasic, inactivating component and a sustained, noninactivating component. A: inactivation was examined with a protocol in which a voltage-clamp step to 35 mV, to selectively activate the VSRM, was preceded by conditioning steps to different membrane potentials (inset). Phasic VSRM contractions (b) decreased in amplitude as the conditioning potential was made more positive, whereas the sustained component (c) was not affected. B: phasic component of the VSRM (b-c) exhibited steady-state inactivation, with Vh near 47 mV. C: sustained contraction (a), measured with respect to a constant reference (d), varied with conditioning potential as expected for activation-deactivation. D: total developed contraction (b-a) was dependent on the preceding conditioning potential, as shown by the steady-state inactivation curve. Steady-state inactivation curve reflects opposite changes in the amplitudes of phasic (b) and sustained (a) components of the VSRM. A-D show mean data from 11 cells. Experiments were conducted with high-resistance microelectrodes.

Developed contraction normally is measured as the difference between peak contraction and preceding diastolic length. In the present experiment, this would correspond to the difference between b and a (Fig. 8A), which is plotted as a function of preceding conditioning potential in Fig. 8D. The resulting curve is identical to the steady-state inactivation curve previously described for the VSRM (13, 16). However, here we show that the magnitude of developed contraction initiated by the VSRM is determined by a combination of inactivation of the phasic component of the VSRM and voltage-dependent changes in diastolic length regulated by the activation/deactivation properties of the sustained component.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The objectives of this study were to determine whether sustained contractions represent a component of excitation-contraction coupling, which contributes to initiation and/or relaxation of contraction in cardiac myocytes, and to determine whether sustained responses are caused by CICR or the VSRM. Our results indicate that maintained depolarization causes sustained Ca²⁺ transients and contractions, and that relaxation of these sustained responses occurs only upon repolarization to membrane potentials near the normal resting membrane potential of ventricular myocytes. Our experiments indicate that the sustained component represents maintained SR Ca²⁺ release. Furthermore, our observations indicate that it is improbable that CICR coupled to I_Ca-L or Na/Ca_EX initiates these sustained responses. Instead, our observations indicate that sustained responses represent a component of the VSRM. The contribution of the VSRM to initiation of contraction likely involves both phasic and sustained components. Although relaxation of the phasic component of VSRM contraction proceeds independently of membrane potential, relaxation of the sustained component is controlled by membrane potential.

A major finding in this study was that cardiac myocytes exhibit a component of contraction that is maintained at potentials positive to the resting membrane potential. The voltage dependence of these contractions or transients is described by a Boltzmann function, as would be predicted for a phenomenon that is regulated by charge movement in a voltage sensor (14). The voltage dependence of sustained transients was the same regardless of the direction of membrane potential change. This suggests that sustained transients are regulated by a mechanism that exhibits activation and deactivation with virtually identical voltage dependencies. Furthermore, no evidence of inactivation was observed with sustained responses up to 7 s in duration.

In theory, several mechanisms might be responsible for generation of sustained Ca²⁺ transients and contractions; these include CICR coupled to I_Ca-L, Na⁺ current, or Na/Ca_EX , or alternatively, the VSRM. It is unlikely that I_Ca-L causes the sustained phenomena, since sustained contractions persisted in the presence of I_Ca-L blockade with Cd²⁺ or when I_Ca-L was inactivated by sustained depolarization. In fact, sustained Ca²⁺ transients were maximal at membrane potentials that strongly inactivated I_Ca-L. Also, sustained contractions and Ca²⁺ transients appeared at membrane potentials more negative to those at which I_Ca-L is significantly activated (25). The activation-deactivation curve for sustained transients was similar to the voltage dependence of the VSRM described previously (11, 13, 16).

Our results also suggest that it is unlikely that the sustained responses are initiated by mechanisms of excitation-contraction coupling linked to activation of Na⁺ channels. Ca²⁺ influx through Na⁺ channels has been reported to initiate cardiac contraction in isolated myocytes treated with isoproterenol or ouabain (31). In addition, Na⁺ influx through Na⁺ channels may initiate CICR by causing Ca²⁺ influx via reverse-mode Na/Ca_EX (19, 22 but also see 5, 33, 36). Both of these mechanisms for initiation of contraction are blocked by Na⁺ channel blockers. However, the sustained responses in the present study could not have been initiated by either of these mechanisms, since they were not affected by Na⁺ channel blockade with either lidocaine or TTX.

Our observations also indicate that it is highly unlikely that CICR coupled to influx of Ca²⁺ through Na/Ca_EX initiates the sustained contractions and transients described in this study. The voltage dependence of activation-deactivation was independent of changes in concentration of extracellular Na⁺, which have been shown to shift the voltage dependence of the exchanger (4, 8, 17). Also, sustained transients with an identical voltage dependence were demonstrated in experiments in which cells were dialyzed with patch pipette solution with 0 mM Na⁺ to inhibit reverse-mode Na/Ca_EX (20, 27, 39). Furthermore, sustained contractions were not affected by 2 mM Ni²⁺, which significantly inhibits Na/Ca_EX (17). In contrast, 200 µM tetracaine, which inhibits the VSRM (24) but not Na/Ca_EX (28, 29), strongly inhibited sustained contractions.

The sustained components of contractions and Ca²⁺ transients described in this study exhibit properties that are characteristic of the VSRM. Sustained responses exhibited a sigmoidal voltage dependence that was well described by a Boltzmann function with a V_h identical to that described for the VSRM (13, 16). The sustained responses, like the VSRM, were not inhibited by agents or conditions that inhibit contractions and transients initiated by I_Ca-L, Na/Ca_EX, or Na⁺ current. Although sustained contractions and transients were insensitive to Cd²⁺ and Ni²⁺, they were strongly inhibited by tetracaine, a known blocker of the VSRM (24). These similarities indicate that the phasic and sustained components are most likely initiated by the same event. However, one striking difference between these two phases is that the phasic component exhibits inactivation, whereas the sustained component does not. This difference is reminiscent of the properties of two components of Ca²⁺ release described for skeletal muscle excitation-contraction coupling. Indeed, the VSRM transients observed in the present study bear a striking resemblance to Ca²⁺ transients recorded from skeletal muscle (30, 35). Transients in skeletal muscle exhibit an initial rapid phase, which inactivates, followed by a noninactivating sustained component (32, 35). In skeletal muscle, the sustained component is believed to represent activation of ryanodine receptors by the voltage sensor. It has been suggested that the initial phasic component is large because Ca²⁺ released by ryanodine receptors coupled to voltage sensors recruits adjacent ryanodine receptors through CICR (30). This results in a multiplier or amplification system that exhibits inactivation, possibly through a Ca²⁺-dependent mechanism (30, 32). One may speculate that similar mechanisms are responsible for the phasic and sustained components of Ca²⁺ release observed in cardiac myocytes. Clearly, additional studies will be required to test this possibility.

In previous studies, we have focused on the phasic component of the VSRM. However, evidence for the sustained component can be found in many of our figures published earlier [e.g., Fig. 10 (11); Figs. 3 and 5 (15); Fig. 10 (16)]. In other examples, the existence of the sustained component was obscured by use of less negative postconditioning potentials and recording windows that did not include sufficient baseline recording both before and after the contractions and transients. In the present study, we have used techniques to clearly show the relation of the contractions and Ca²⁺ transients to the diastolic levels before and after the events. With this approach, the sustained component is very evident.

Interestingly, sustained VSRM contractions and transients were normally very stable and did not exhibit oscillatory behavior. Our observations with ryanodine indicate that sustained VSRM contractions are generated by release of SR stores of Ca²⁺. Thus our observations suggest that it is likely that released Ca²⁺ is taken back up by the SR and rereleased. These considerations suggest that, under the conditions of our experiments, Ca²⁺ release, uptake, and rerelease must reach a stable equilibrium that does not spontaneously oscillate.

In contrast to earlier concepts of excitation-contraction coupling in heart, our results demonstrate that sustained release of Ca²⁺ can be regulated by membrane potential through the VSRM. Through voltage-dependent activation and deactivation, membrane potential adjusts release of Ca²⁺ stores and determines contraction and relaxation of the cardiac cell. Activation of contraction by the VSRM includes activation of both sustained and phasic components of Ca²⁺ release. Relaxation of the phasic component of contraction occurs even when the cell membrane remains depolarized. However, termination of sustained Ca²⁺ release requires repolarization. Thus the VSRM is a mechanism of excitation-contraction coupling that actively controls both contraction and relaxation in cardiac myocytes. Our study demonstrates that we must revise our concepts of excitation-contraction coupling in heart to include an important contribution of Ca²⁺ release linked to membrane potential, in addition to CICR coupled to Ca²⁺ influx.

Because the sustained component of the VSRM is graded by voltage, this component might delay relaxation in cardiac hypertrophy or failure where the action potential is significantly prolonged (40). Furthermore, sustained Ca²⁺ release graded by the VSRM may alter diastolic compliance, especially in disease conditions that result in reduction of membrane potential. Thus our observations have important implications for regulation of contraction in diseased and normal hearts.