INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

IT IS WELL ESTABLISHED that contraction in heart is initiated by an increase in free intracellular Ca²⁺. This increase in activator Ca²⁺ is derived from two sources, Ca²⁺ influx and Ca²⁺ release from intracellular stores. It is generally believed that the main routes for Ca²⁺ entry are voltage-gated Ca²⁺ channels and reverse-mode Na⁺/Ca²⁺ exchange, whereas the main source for Ca²⁺ release is the sarcoplasmic reticulum (SR) (3, 15, 19). Fabiato (11) has shown that a rapid rise in intracellular free Ca²⁺ can cause Ca²⁺ release from cardiac SR. This phenomenon has been called Ca²⁺-induced Ca²⁺ release (CICR) (11).

Identification of the trigger or triggers for SR Ca²⁺ release continues to be a subject of intense research. Evidence has been presented for at least three mechanisms that trigger CICR. These are 1) Ca²⁺ influx via L-type Ca²⁺ channels (4, 7, 22), 2) elevation of intracellular Ca²⁺ via reverse Na⁺/Ca²⁺ exchange in response to a rapid rise in intracellular Na⁺ during the upstroke of the action potential (15, 19), and 3) elevation of intracellular Ca²⁺ via reverse Na⁺/Ca²⁺ exchange in response to depolarization, although this last mechanism may depend on intracellular Na⁺ concentration and temperature (6, 16, 17, 29, 30).

Recently we reported a study of excitation-contraction (EC) coupling in guinea pig ventricular myocytes at 37°C that used high-resistance microelectrodes to prevent intracellular dialysis (13). Under these conditions, the threshold for activation of contractions is much more negative than the threshold for activation of L-type Ca²⁺ current (I_Ca-L). In addition, the magnitude of contraction is not proportional to the magnitude of inward current, and contractions remain maximal at membrane potentials near or beyond the reversal potential for I_Ca-L (13). In that study we were able to separate a new component of contraction that was abolished by a low concentration of ryanodine, but not by L-type Ca²⁺ channel blockers or Na⁺ channel blockers, from contraction initiated by I_Ca-L (13). The new component of contraction exhibits a sigmoidal contraction-voltage relationship in contrast to the bell-shaped relationship observed for contractions initiated by I_Ca-L. It also is unlikely that this component of contraction is initiated by reverse Na⁺/Ca²⁺ exchange, because its voltage dependence is not affected by large changes in concentrations of extracellular Na⁺ or Ca²⁺ (12, 32). Because activation of this new component of contraction clearly is dependent on membrane potential but is not proportional to macroscopic transmembrane current, we have called this component a voltage-sensitive release mechanism (VSRM) for SR Ca²⁺ (13). The present study examines the contribution of this voltage-activated Ca²⁺ release mechanism to cardiac EC coupling.

Many previous studies have reported bell-shaped contraction-voltage relationships in which the amplitudes of contractions or Ca²⁺ transients were proportional to the magnitude of I_Ca-L (2, 3, 4, 9, 10, 20). However, many of these earlier studies of cardiac EC coupling in isolated myocytes have been conducted with conditions different from those in our study, including use of holding potentials near 40 mV (2, 3, 9, 10, 20). The absence of the VSRM in those studies might be explained if the VSRM has steady-state inactivation properties and is inactivated at 40 mV. Therefore, one of the goals of this study was to determine whether the VSRM exhibits steady-state inactivation and to determine the voltage range over which inactivation occurs.

Contractions initiated by the VSRM are abolished by 30 nM ryanodine, an agent that disrupts EC coupling at the level of SR release of Ca²⁺ (13). Thus SR Ca²⁺ stores are likely essential for initiation of contraction by the VSRM. Repetitive activation of I_Ca-L serves to load or maintain SR stores of Ca²⁺ (3), and altering the voltage of repetitive depolarizations can be used to manipulate SR Ca²⁺ load and the amount of releasable Ca²⁺ (14). Therefore, a second goal of the present study was to determine and compare the effects of protocols designed to alter SR Ca²⁺ load on the components of contraction initiated by the VSRM or by I_Ca-L. SR Ca²⁺ loading and recovery of SR Ca²⁺ release also are believed to be important components in restitution of contractility, which plays a major role in adjusting the magnitude of contraction in response to changes in the interval between contractions (3, 5). Therefore, the final goal of this study was to evaluate the role of the VSRM in restitution.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Cell isolation. All experiments were performed in accordance with the guidelines published by the Canadian Council on Animal Care, and this investigation was approved by the Dalhousie University Committee on Animal Care. Most experiments were conducted on isolated guinea pig ventricular myocytes. Male guinea pigs (350-400 g, Charles River) were injected with heparin (3.3 IU/g) 30 min before anesthesia with pentobarbital sodium (80 mg/kg). The chest was opened, and the heart was rapidly cannulated in situ and immediately perfused retrogradely through the aorta (10-12 ml/min) with oxygenated (100% O₂, 36°C) Ca²⁺-free solution of the following composition (in mM): 120 NaCl, 3.8 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 10 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), and 11 glucose (pH 7.4 with NaOH). The heart was then removed from the chest during perfusion with Ca²⁺-free solution for 7-8 min. Collagenase A (35-40 mg, Boehringer Mannheim) and protease (4.8-6 mg, Sigma type XIV; Sigma, St. Louis, MO) were then included in 65 ml of this Ca²⁺-free solution, and the heart was perfused for an additional 5-8 min. After enzymatic dissociation, the ventricles were minced and washed in a substrate-enriched solution of the following composition (in mM): 80 KOH, 50 glutamic acid, 30 KCl, 30 KH₂PO₄, 20 taurine, 10 HEPES, 10 glucose, 3 MgSO₄, and 0.5 ethylene glycol-bis(-aminoethyl ether)-N,N,N',N'-tetraacetic acid (pH 7.4 with KOH). In some experiments, rat ventricular myocytes were utilized. The isolation procedure was similar to that for guinea pig myocytes except that dissociation was accomplished with collagenase (20 mg, Worthington type 2) and trypsin (2 mg, Sigma). After 1-2 h of incubation at room temperature, myocytes were placed in a modified culture dish (approximate volume = 0.75 ml) in an open-perfusion microincubator (model PDMI-2, Medical Systems) on the stage of an inverted microscope. Cells were allowed to adhere to the bottom of the chamber for 15-20 min and were then superfused at 37°C with the HEPES-buffered solution described above, supplemented with 2.0 mM Ca²⁺. After 10-15 min, cells were superfused with an oxygenated (100% O₂, 37°C) solution of the following composition (in mM): 100 choline chloride, 45 NaCl, 10 glucose, 10 HEPES, 4 KCl, 2.0 CaCl₂, 1.0 MgCl₂, and 200 µM lidocaine (pH 7.4 with NaOH). Solutions were pumped through the chamber from a buffer reservoir at a rate of 3 ml/min and were changed by switching the inlet to the pump between buffer reservoirs. The time required to replace the bath solution was determined by measuring the membrane potential response to a step change in K⁺ concentration. Transit time to the bath was ~1 min and changeover time was ~90 s.

Experimental methods. Discontinuous single-electrode voltage-clamp recordings (sample rate 10-14 kHz) were made with an Axoclamp 2A amplifier (Axon Instruments, Foster City, CA). Recordings were made with high-resistance microelectrodes (18-25 M, filled with 2.7 M KCl) to reduce cell dialysis. A 2.7 M KCl-agar bridge was used as a bath ground to minimize liquid junction potential changes. Voltage-clamp protocols were generated with pCLAMP software (Axon Instruments); pCLAMP software also was used to acquire and analyze data on computer. Recordings were only made from rod-shaped myocytes with clear, well-organized striations and with resting potentials more negative than 85 mV. In all experiments, both current and transmembrane voltage were recorded. During discontinuous single-electrode voltage clamp, we continuously monitored the output of the switching circuit to ensure that adequate settling time for accurate voltage measurement was maintained.

Inward and delayed rectifier potassium currents were not blocked in this study because the effects of blocking agents on the VSRM are not known. Therefore, changes in steady currents at the ends of the activation steps represent steady-state current-voltage (I-V) relations. After trains of conditioning pulses, the membrane potential was usually repolarized to a postconditioning potential (v_PC) more positive than the holding potential. Thus, background currents during the v_PC period also reflect steady-state I-V relations.

I_Ca-L was measured as the difference between the peak inward current and a reference point at the end of the voltage step (normally 250 ms). Figure 1 demonstrates that current measured in this way is abolished by 2 µM verapamil, an established L-type Ca²⁺ channel blocker. The top trace in Fig. 1A shows the membrane potential response to a voltage-clamp step from 65 to 5 mV, measured by the current-passing electrode. In a previous study, we demonstrated that the voltage measured by the current-passing electrode is an accurate measurement of the membrane potential by monitoring membrane potential with a second independent electrode (13). The next two traces show the corresponding current records before and after exposure of the cell to 2 µM verapamil. The inward current deflection following the capacitive transient was completely blocked by verapamil. The bottom trace is a difference trace derived by subtracting the current trace in the presence of verapamil from the control trace. Figure 1B shows I-V relationships derived from voltage steps to different potentials before and after exposure to verapamil. Figure 1C shows the I-V relationship derived from the difference currents. The results indicate that verapamil completely blocked the current identified as I_Ca-L and that this current was identical to the difference in current before and after drug treatment.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Verapamil inhibits currents identified as L-type Ca2+ currents (ICa-L). Membrane potential response to a voltage-clamp step from 65 to 5 mV was measured by the current-passing electrode (top trace in A). Next 2 traces show corresponding current records before and after exposure of the cell to 2 µM verapamil. Bottom trace shows difference between top 2 traces. B: current-voltage (I-V) relationships when current was measured as difference between peak inward deflection and a point at the end of the voltage step, before and after exposure to verapamil. C: I-V relationship derived from the difference currents. Results demonstrate that the verapamil-sensitive current was identical for both methods of measurement.

Cell images were monitored with a closed-circuit television camera with interlace defeat and partial scan capability (model 1-GP-CD60, Panasonic) and were displayed on a video monitor (model VM-1220C, Hitachi Densi). Unloaded cell shortening was sampled at 120 Hz with a video edge detector (Crescent Electronics, Sandy, UT) coupled to the television camera. Details of specific voltage-clamp protocols are provided in the appropriate sections in RESULTS. In most experiments, voltage-clamp protocols were repeated two to three times and the data were averaged. Current, voltage, and contractions were digitized with a Labmaster A/D interface at 125 kHz (TL1-125, Axon Instruments) and stored on hard disk for subsequent analysis.

Data analyses. Ionic current, voltage, and contraction were measured with pCLAMP analysis software. Significance of differences between population means was tested with a Student's t-test with a Bonferroni correction for multiple comparisons. I-V relationships, contraction-voltage relationships, and time courses were analyzed with a two-way repeated-measures analysis of variance. Post hoc comparisons were made with a Bonferroni test. All statistical analyses were performed with Sigma Stat (Jandel, version 1.02) or with SAS (SAS Institute). Nonlinear curve-fitting procedures were conducted with SigmaPlot (Jandel, version 2.0). Data are presented as means ± SE. The value of n represents the number of myocytes sampled; no more than two replicates (myocytes) were collected from the same heart.

Sources of drugs and chemicals. Lidocaine was purchased from Sigma, and ryanodine was purchased from Calbiochem (San Deigo, CA). All drugs were dissolved in distilled water, except nifedipine stock, which was prepared in ethanol. Choline chloride was purchased from Fisher Scientific (Fairlawn, NJ).

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Effects of v_PC on contraction-voltage and I-V relations. Figure 2 (A and B) shows representative original recordings of membrane currents and contractions determined in the presence of 200 µM lidocaine to inhibit inward Na⁺ current. A schematic of the voltage-clamp protocol is illustrated at the top of each panel. The complete voltage-clamp protocol was repeated every 7 s. Each test step was preceded by a train of 10 200-ms conditioning steps to 0 mV to activate I_Ca-L and maintain SR Ca²⁺ loading. With this conditioning protocol, signs of Ca²⁺ overload were not observed. In Fig. 2A, the test step was separated from the last conditioning step by a 300-ms v_PC of 40 mV. When the v_PC was 40 mV, a test step to 10 mV initiated both inward current and a large contraction, whereas a test step to +80 mV activated no inward current and only a very small contraction. Traces in Fig. 2B were recorded with a similar voltage-clamp protocol, except that the v_PC was 70 mV. Here a test step to 40 mV activated a small inward current and a large contraction, and a test step to 10 mV activated a larger inward current and contraction (Fig. 2B). However, in contrast to the results shown in Fig. 2A, a depolarizing step to +80 mV initiated a large contraction but did not activate inward current when the v_PC was 70 mV.

View larger version (28K): [in this window] [in a new window]   Fig. 2.   Effects of varying postconditioning potential (vPC) on contraction-voltage and I-V relationships. A: representative original recordings of transmembrane current and corresponding contractions elicited by voltage steps to +80 and 10 mV from a vPC of 40 mV. Schematic of voltage-clamp protocol is shown at top of A. Each test step was preceded by a train of 10 conditioning pulses from 80 to 0 mV to maintain sarcoplasmic reticulum (SR) Ca2+ stores. The step to +80 mV elicited no inward current and only a very small contraction. However, the step to 10 mV elicited both inward current and a large contraction. B: representative original recordings of currents and contractions elicited by voltage steps to +80, 10, and 40 mV from a vPC of 70 mV. Voltage-clamp protocol is illustrated at top of B. With this vPC, the test step to +80 mV still elicited no inward current but now elicited a large contraction. Test steps to 10 and 40 mV elicited both contractions and inward current. C: mean contraction-voltage relationships determined from vPC values of 40 and 70 mV. When vPC was 40 mV, the contraction-voltage relationship was bell shaped. In contrast, when vPC was 70 mV the contraction-voltage relationship was sigmoidal. D: mean I-V relationships recorded simultaneously with the contractions shown in C. Contractions were inversely proportional to the magnitude of inward current when the vPC was 40 mV, whereas contractions were not proportional to the magnitude of inward current when vPC was 70 mV. * Significantly different than corresponding points when vPC was 70 mV (P < 0.05).

The effect of v_PC on contractions and inward currents is clearly seen when contraction-voltage and I-V relationships are plotted as shown in Fig. 2, C and D, respectively. These curves represent mean values ± SE for 12 cells studied with a v_PC of 70 mV and 7 cells studied with a v_PC of 40 mV. When the v_PC was 40 mV, a bell-shaped contraction-voltage relation was observed (Fig. 2C). The magnitude of contractions increased to a peak at 0 mV and then declined at more positive potentials. These changes in the magnitudes of contractions were proportional to changes in Ca²⁺ current, as shown in Fig. 2D. Figure 2C also shows contraction-voltage relationships determined by voltage steps from a v_PC of 70 mV. The threshold for activation of these contractions was about 60 mV, and contractions increased to a plateau at approximately 20 mV (Fig. 2C). Contractions remained large at very positive potentials, even though peak inward current decreased (Fig. 2D). We previously suggested that the additional component of contraction activated when the v_PC is more negative than 40 mV may represent activation of a VSRM (13).

Steady-state inactivation properties of the VSRM. Our observation that the VSRM component of contraction was inhibited when test steps were made from a v_PC of 40 mV suggests that the VSRM might show voltage-dependent inactivation. To determine whether the VSRM exhibits the property of steady-state voltage inactivation, we examined the effects of systematically changing the v_PC on the magnitude of contractions activated by voltage steps to 40 mV. We have shown previously that contractions initiated by an activation step to 40 mV are entirely attributable to the VSRM (13). The schematic in Fig. 3A depicts the voltage-clamp protocol used in these studies. Each test step was preceded by 10 conditioning steps to 0 mV followed by repolarization to the v_PC for 700 ms. The v_PC was changed in 5-mV steps, from 35 to 70 mV, with each repetition of the voltage-clamp protocol. Each 700-ms v_PC was followed by a very brief (10 ms) step to 65 mV. The complete voltage-clamp protocol was repeated every 9 s. Representative original records of contraction and current also are shown in Fig. 3B. When the v_PC was 35 mV, no contraction was observed with the step to 40 mV. However, when the v_PC was changed to 45 mV, a small contraction was elicited by the activation step. Successively larger contractions were observed with v_PC values of 55 and 65 mV (Fig. 3B). The activation step elicited very little inward current with any of the v_PC values tested. Figure 3C shows a plot of mean magnitudes of peak contraction (±SE) as a function of v_PC for eight myocytes. These data were fitted with a Boltzmann equation of the following form: y = (a  b)/{1 + exp[(v_PC  v_h)/k]} + b, where a is the maximum contraction, b is the minimum contraction, v_h is the half-inactivation voltage, and k is the slope factor. Mean data also were normalized to the maximum contraction and plotted as a function of v_PC in Fig. 3D. The line in Fig. 3D represents a Boltzmann function fitted to the normalized data using the following equation: y = 1/{1 + exp[(v_PC  v_h)/k]}. v_h and k were calculated for each myocyte (n = 8). Mean v_h was 47.6 ± 1.0 mV; mean k was 4.37 ± 0.65 mV. These observations clearly show that the contractions initiated by the activation step to 40 mV exhibit the property of steady-state inactivation.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Contractions initiated by voltage-sensitive release mechanism (VSRM) exhibit steady-state inactivation. Voltage-clamp protocol is shown by schematic in A. Holding potential was 80 mV, and test steps were preceded by 10 conditioning steps to 0 mV. The last conditioning potential was followed by repolarization to a vPC for 700 ms. With each repetition of the protocol, vPC was changed in 5-mV steps. After the vPC, the membrane potential was stepped to 65 mV for 10 ms and then to 40 mV to assess availability of currents and contraction. B: traces recorded with representative vPC values. Amplitudes of contractions increased progressively as vPC was made more negative. A small inward current was observed with the more negative vPC values. C: mean contraction amplitudes as a function of vPC for data from 8 myocytes. Line represents a Boltzmann function fitted to the mean data. D: normalized steady-state inactivation relationship for the same data. Half-inactivation voltage (vh) was 47.6 ± 1.0 mV, and slope factor (k) was 4.37 ± 0.65 mV.

Separation of contractions induced by the VSRM and I_Ca-L by voltage. The observation that contractions initiated by the VSRM exhibit the property of steady-state inactivation suggests that the activation and inactivation properties of contraction can be used to separate contractions initiated by the VSRM from those initiated by I_Ca-L. Indeed, we previously have shown that contractions initiated by the VSRM can be separated from contractions initiated by I_Ca-L by sequential steps to 40 and 0 mV (13). Figure 4 shows records in which the VSRM was activated by a 250-ms step to 40 mV from a v_PC of 65 mV, and I_Ca-L was activated by the second step to 0 mV. Figure 4A shows contractions and currents recorded under control conditions. The step to 40 mV activated a large contraction and little if any inward current. The second step to 0 mV activated an additional contraction and I_Ca-L. Figure 4C shows the effects of exposing the same cell to 2.5 µM nifedipine. Nifedipine selectively blocked I_Ca-L and the corresponding contraction. The VSRM contraction was only slightly reduced in amplitude. Mean results for five myocytes exposed to nifedipine are presented in Fig. 5. Nifedipine significantly inhibited I_Ca-L and the contraction elicited by the step to 0 mV (Fig. 5, D and B, respectively) but had no significant effect on the current or contraction elicited by the step to 40 mV (Fig. 5, C and A, respectively). Similar results were observed in an additional five cells exposed to verapamil (2 µM) (not illustrated).

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Contractions initiated by the VSRM and ICa-L can be separated by membrane potential. Voltage-clamp protocol is illustrated at top. After 10 conditioning steps to 0 mV, VSRM was activated by a 250-ms step to 40 mV from a vPC of 65 mV, and ICa-L was activated by a second step to 0 mV. Under control conditions, the step to 40 mV activated a large contraction and little if any inward current, whereas the second step to 0 mV activated an additional contraction and ICa-L (A). C: 2.5 µM nifedipine selectively blocked ICa-L and the corresponding contraction, with little effect on the VSRM contraction. To maintain SR Ca2+ stores in the presence of Ca2+ channel blockade, conditioning pulses to positive potentials (e.g., +80 mV) were used to cause Ca2+ entry via Na+/Ca2+ exchange. B: control recordings of currents and contractions initiated by sequential steps to 40 and 0 mV in a separate cell. Exposure of this cell to 30 nM ryanodine inhibited VSRM contraction but had little effect on current and contraction initiated by the step to 0 mV (D).

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Mean effects of nifedipine, ryanodine, and conditioning pulses (CP) to 40 mV on contractions and currents initiated by sequential activation steps to 40 and 0 mV. Shown are histograms of mean data collected with the same voltage-clamp protocol as the examples shown in Fig. 4. Control data were collected when the conditioning pulses were to 0 mV. Nifedipine (2.5 µM) had no significant effect on contractions or currents initiated by voltage steps from 65 to 40 mV (A and C). However, nifedipine significantly inhibited both contractions and currents initiated by steps to 0 mV (B and D). In contrast, either ryanodine (30 nM) or conditioning pulses to 40 mV caused significant decreases in the magnitudes of contractions initiated by the VSRM (A). Magnitudes of contractions initiated by ICa-L were only moderately reduced by ryanodine and were not affected by conditioning pulses to 40 mV (B). Ryanodine had no effect on peak inward current recorded in response to steps to 40 or 0 mV (C and D, respectively). However, conditioning pulses to 40 mV significantly increased Ca2+ current associated with the step to 0 mV (D). n = 25 cells for control, 5 cells for nifedipine, 20 cells for ryanodine, and 14 cells for conditioning pulses to 40 mV. * Significantly different from control.

In contrast to Ca²⁺ channel blockers, ryanodine (30 nM) strongly inhibited the VSRM contraction, rather than the I_Ca-L contraction. Figure 4B shows control recordings of currents and contractions initiated by sequential steps to 40 and 0 mV. Fig. 4D shows that exposure of the same cell to ryanodine abolished the contraction initiated by the step to 40 mV but had little effect on current and contraction initiated by the step to 0 mV. Mean data for ryanodine also are presented in Fig. 5. Ryanodine virtually abolished contractions initiated by the VSRM (from 2.2 ± 0.2 to 0.10 ± 0.03, P < 0.001). In contrast, contractions initiated by the second step to 0 mV were only moderately decreased by exposure to ryanodine (from 1.64 ± 0.2 to 1.06 ± 0.14, P < 0.05). Treatment with ryanodine did not affect inward currents associated with steps to either 40 or 0 mV (Fig. 5, C and D).

These observations indicate that sequential steps to 40 and 0 mV separate two types of contractions with clearly different pharmacological sensitivities. These observations also suggest that VSRM contractions should be very sensitive to manipulations designed to alter SR Ca²⁺ load, whereas contractions initiated by I_Ca-L should be much less sensitive to these manipulations.

Effects of conditioning pulses designed to alter SR load on contractions initiated by L-current and VSRM. To increase SR loading, we used trains of conditioning pulses to 0 mV to repetitively activate I_Ca-L. Figure 6A shows representative recordings of voltages and contractions with conditioning pulses to 0 mV, which is close to the peak of the I-V relation for I_Ca-L. After the first contraction, which is a rest contraction, a positive staircase was observed for the next nine conditioning pulses. In this example, the train of conditioning pulses was followed by repolarization to a v_PC of 70 mV without a test step. To reduce SR load, trains of conditioning pulses to 40 mV were used to activate release of SR Ca²⁺ with minimal activation of I_Ca-L (Fig. 6B). With this protocol, the rest contraction had the same magnitude as the corresponding rest contraction in Fig. 6A, but there was no positive staircase. Clearly, conditioning pulses to 0 or 40 mV had very different effects on magnitude of contraction during the conditioning trains.

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Effects of changing conditioning-pulse voltage from 0 to 40 mV on contraction amplitude. A: representative original recordings of voltage and contractions during trains of conditioning pulses to 0 mV, followed by return to a vPC of 70 mV. After the first contraction of the train, contractions exhibited a positive staircase. B: similar sequence recorded when the voltage of conditioning pulses was 40 mV. Contractions during the trains were greatly decreased in amplitude.

Next we determined the effects of conditioning pulses to 0 and 40 mV on contractions initiated by sequential test steps to 40 and 0 mV. Figure 7A was recorded when the activation steps were preceded by a series of 10 conditioning pulses to 0 mV. With this protocol the step to 40 mV activated a large contraction and little if any inward current. The second step to 0 mV activated an additional contraction and I_Ca-L. Figure 7B was recorded after 10 conditioning pulses to 40 mV. With this protocol the step to 40 mV no longer elicited a contraction. However, the contraction associated with activation of I_Ca-L was only slightly decreased. Currents associated with the steps to 40 and 0 mV were not affected by changing the conditioning-pulse voltage in this example. Mean data for the effects of conditioning-pulse voltage are shown in Fig. 5. Contractions initiated by the VSRM were significantly reduced in amplitude (from 2.2 ± 0.2 to 0.3 ± 0.1 µm, P < 0.001, Fig. 5A), whereas the amplitude of I_Ca-L contractions was not significantly decreased by conditioning pulses to 40 mV (from 1.6 ± 0.2 to 1.5 ± 0.3 µm, NS, Fig. 5B). The small inward current initiated by the step to 40 mV was not significantly affected (Fig. 5C); however, there was a significant increase in peak I_Ca-L when conditioning-pulse voltage was changed from 0 to 40 mV (Fig. 5D, P < 0.05). The effects of changing conditioning-pulse voltage to 40 mV on the VSRM contraction were very similar to those of ryanodine.

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Comparison of effects of conditioning pulse voltage on contractions initiated by the VSRM and ICa-L. Sequential steps to 40 and 0 mV were used to activate VSRM and ICa-L contractions, respectively. A: recordings of contractions (top) and currents (bottom) when conditioning-pulse voltage was 0 mV. Each step initiated contractions; however, inward current was only observed with the step to 0 mV. B: when voltage of the conditioning pulses was changed to 40 mV, the contraction initiated by the VSRM was abolished, but ICa-L contraction was decreased only slightly.

Comparison of the effects of conditioning pulses to 40 mV with effects of ryanodine on contraction-voltage and I-V relationships. The results of the preceding experiments (Fig. 5) show that the VSRM is affected much more than I_Ca-L-induced contractions by conditioning steps to 40 mV or by exposure to ryanodine. Thus one would predict that changing conditioning-pulse voltage might also affect the contribution of the VSRM to contractions initiated when both mechanisms are available. Therefore, we determined the effects of conditioning-pulse amplitude on contraction-voltage relationships determined with voltage steps to a wide range of potentials. First, we determined the effects of conditioning-pulse voltage on I-V and contraction-voltage relationships when the VSRM was inactivated by a v_PC of 40 mV (Fig. 8, A and B). I-V and contraction-voltage relationships determined from a v_PC of 40 mV were both bell shaped. When conditioning-pulse voltage was changed from 0 to 40 mV, I-V relationships were not affected; however, the magnitudes of the contractions were significantly decreased at all voltages (P < 0.05). Figure 8C shows the difference between contraction-voltage relationships determined with the two conditioning-pulse voltages. The component inhibited by changing conditioning-pulse voltage also was bell shaped.

View larger version (33K): [in this window] [in a new window]   Fig. 8.   Effect of conditioning-pulse voltage on contraction-voltage and I-V relationships determined from vPC values of 40 and 70 mV. Changing the voltage of the conditioning pulses had no effect on I-V relationships determined with vPC values of either 40 or 70 mV (A and D). However, the bell-shaped contraction-voltage relationship elicited by steps from a vPC of 40 mV was significantly decreased in amplitude (P < 0.05) when conditioning-pulse voltage was changed from 0 to 40 mV (B). The difference contraction-voltage relationship illustrated in C shows component of contraction inhibited by changing conditioning-pulse voltage. When vPC was 70 mV, the magnitude of contraction also significantly decreased (P < 0.05) and the contraction-voltage relationship changed from sigmoidal to bell shaped when the voltage of the conditioning pulses was changed from 0 to 40 mV (E). The component of contraction inhibited by changing conditioning-pulse voltage was sigmoidal, as shown in F. n = 12 cells for vPC 70 mV and 7 cells for vPC 40. * Significantly different from conditioning pulses to 40 mV, P < 0.05.

We also examined the effects of conditioning pulses on I-V and contraction-voltage relationships determined with activation steps from a v_PC of 70 mV (Fig. 8, D and E). Under these conditions both I_Ca-L and the VSRM were available to initiate contraction. The I-V relationship was again bell shaped and was not affected by changing conditioning-pulse voltage (Fig. 8D). When conditioning pulses were to 0 mV, the contraction-voltage relationship was sigmoidal and contractions remained large even at very positive potentials (Fig. 8E). However, when conditioning pulses to 40 mV were used, the magnitudes of contractions were significantly decreased (P < 0.05). Furthermore, the contraction-voltage relationship became bell shaped with a peak at 0 mV. Figure 8F shows the difference between the contraction-voltage relationships determined with the two conditioning-pulse voltages. Here the difference was sigmoidal.

We also determined the effects of 30 nM ryanodine on contraction-voltage and I-V relations initiated by activation steps from v_PC values of 40 and 70 mV. In these experiments, all conditioning pulses were to 0 mV (Fig. 9). Ryanodine significantly reduced (P < 0.05) inward current determined with either v_PC but did not shift the voltage dependence (Fig. 9, A and D). Figure 9B shows that ryanodine significantly decreased (P < 0.05) the amplitudes of contractions initiated from a v_PC of 40 mV. The contraction-voltage relationship in the presence of ryanodine remained bell shaped. Figure 9C shows the difference between the contraction-voltage relationships determined in the absence and presence of ryanodine. When contraction-voltage relationships were determined from a v_PC of 70 mV (Fig. 9E), ryanodine changed the shape of the contraction-voltage relationship from sigmoidal to bell shaped. Ryanodine also significantly decreased (P < 0.05) the amplitudes of contractions initiated at virtually all test voltages (Fig. 9E). The component of contraction inhibited by ryanodine showed a sigmoidal voltage dependence (Fig. 9F). These effects of ryanodine on contraction-voltage relationships were very similar to those of changing conditioning-pulse voltage from 0 to 40 mV.

View larger version (33K): [in this window] [in a new window]   Fig. 9.   Effect of ryanodine on contraction-voltage and I-V relationships determined from vPC values of 40 and 70 mV. Ryanodine significantly inhibited ICa-L elicited with voltage steps from vPC of either 40 or 70 mV (A and D). When the vPC was 40 mV, the contraction-voltage relationship was bell shaped. Amplitudes of contractions decreased significantly in response to 30 nM ryanodine (B, P < 0.05). C: ryanodine inhibited a bell-shaped component of contraction. When vPC was 70 mV, the control contraction-voltage relationship was sigmoidal (E). Ryanodine significantly decreased the magnitude of contraction and made the contraction-voltage relationship bell shaped (E, P < 0.05). F: component of contraction inhibited by ryanodine was sigmoidal when vPC was 70 mV. n = 12 cells for vPC 70 mV and 7 cells for vPC 40. * Significantly different from ryanodine, P < 0.05.

The effects of thapsigargin on contraction-voltage and I-V relationships determined from a v_PC of 70 mV in rat ventricular myocytes. An alternate method of evaluating the contribution of SR Ca²⁺ release to contraction is to inhibit SR Ca²⁺ uptake with thapsigargin, an agent that blocks the SR Ca²⁺-adenosinetriphosphatase. Figure 10, A and B, shows representative recordings from a rat ventricular myocyte before and after exposure to 0.2 µM thapsigargin. Contractions and currents initiated by sequential steps to 40 and 0 mV from a v_PC of 65 mV were very similar to those observed with guinea pig ventricular myocytes. Thapsigargin strongly inhibited the contraction initiated by the VSRM but only partially inhibited the contraction accompanying I_Ca-L. Figure 10, C and D, shows mean contraction-voltage and I-V relationships recorded with voltage steps from a v_PC of 70 mV (n = 3). Thapsigargin had no effect on I_Ca-L, but significantly reduced the amplitudes of contractions (P < 0.05). In addition, thapsigargin, like ryanodine or conditioning pulses to 40 mV, caused the contraction-voltage relationship to become bell shaped.

View larger version (24K): [in this window] [in a new window]   Fig. 10.   Currents and contractions in rat ventricular myocytes. A and B: representative recordings of currents (top) and contractions (bottom) elicited by sequential voltage steps to 40 and 0 mV (shown schematically above A). Thapsigargin (0.2 µM) almost abolished the contraction initiated by the step to 40 mV but only partially inhibited the contraction initiated by the step to 0 mV. C: mean contraction-voltage relationships determined from a vPC of 70 mV. Thapsigargin significantly reduced the amplitudes of contraction and converted the contraction-voltage relationship from sigmoidal to bell shaped. Current was unaffected by thapsigargin (D). * P < 0.05 with respect to control. E: mean steady-state inactivation curves for the VSRM contraction and ICa-L determined in 9 rat myocytes. Curves were normalized to maximum contraction or inward current. Voltage-clamp protocols were similar to that shown in Fig. 3. Specific voltage steps are given in RESULTS. vh for VSRM (53.2 ± 0.4 mV) was significantly different from that of ICa-L (25.3 ± 0.9 mV, P < 0.001). Values for k also were significantly different (4.6 ± 0.2 mV for VSRM and 6.0 ± 0.2 mV for ICa-L, P < 0.001).

In rat myocytes there is clear separation between the threshold for activation of contraction and the threshold for activation of inward current (Fig. 10, C and D), probably because of the absence of T-type Ca²⁺ current in this species (28). Because rat myocytes only have L-type Ca²⁺ current, we were able to compare the steady-state inactivation properties of the VSRM to those of I_Ca-L. The voltage-clamp protocol used in these experiments is similar to that shown in Fig. 3. The v_PC was changed in 5-mV steps with each repetition of the voltage-clamp protocol. Each 700-ms v_PC was followed by a very brief (10 ms) step to 65 mV, followed by a step to 35 mV to activate the VSRM, or a brief step to 50 mV followed by a step to 0 mV to activate I_Ca-L. Figure 10E shows mean normalized steady-state inactivation curves for the VSRM and I_Ca-L determined in nine myocytes. The VSRM had a v_h of 53.2 ± 0.4 mV and a k of 4.6 ± 0.2 mV, whereas the corresponding values for I_Ca-L were 25.3 ± 0.9 mV and 6.0 ± 0.2 mV, respectively. The v_h of the VSRM was 28 mV negative to that of I_Ca-L and was significantly different from v_h of I_Ca-L (P < 0.001). The values of k for the VSRM and I_Ca-L also were significantly different (P < 0.001). Thus, when the v_PC was 40 mV, the VSRM was completely inactivated but I_Ca-L was still fully available.

Comparison of the effects of conditioning pulses to 40 mV with effects of ryanodine on restitution of contraction. Contractions initiated by an early test stimulus following a previous activation are small, but show recovery or restitution when the test interval increases (5). We determined the time course of restitution of contraction in isolated myocytes as well as restitution of I_Ca-L. The voltage-clamp protocol used in these experiments is shown at the top of Fig. 11A. Test steps to 0 mV from a v_PC of 65 mV were used to activate both VSRM and I_Ca-L components of contraction (total contraction). The interval (t) between the last conditioning pulse and the test step was increased progressively from 6 to 246 ms, in 20-ms increments. Figure 11A shows representative original recordings of current and contraction for three selected test intervals (26, 106, and 186 ms). Contraction increased progressively as the interval was lengthened. In contrast, Ca²⁺ current increased when the test interval was lengthened from 26 to 106 ms, but remained relatively constant with further increases in test interval. Mean data for restitution of contraction are shown in Fig. 11B. The line represents a single exponential fit to the mean data. The exponential function for restitution of contraction had a time constant of 57.8 ms. Mean data for recovery of Ca²⁺ current are shown in Fig. 11C. The exponential fit to these data had a time constant of 31.1 ms. Thus the recovery of contraction followed a slower time course than the recovery of peak inward Ca²⁺ current.

View larger version (23K): [in this window] [in a new window]   Fig. 11.   Time course of restitution of contraction and restitution of inward current. A schematic of the voltage-clamp protocol is shown at top of A. Test steps to 0 mV from a vPC of 65 mV were preceded by 10 conditioning pulses to 0 mV to maintain SR Ca2+ load. The interval between the last conditioning pulse and the test step (t) was progressively lengthened from 6 to 246 ms. A: representative original records of contraction (top trace in each pair) and current (bottom trace in each pair) for 3 test intervals, 26, 106, and 186 ms. Contractions increased in magnitude with increases in test interval. However, inward current did not increase in magnitude at test intervals beyond 106 ms. Mean data for the time course of restitution of contractions are shown in B, and corresponding data for restitution of current are shown in C. Recovery of contraction followed a slower time course than recovery of current. , Time constant of a single exponential fitted to the mean data.

We next determined the effects of 30 nM ryanodine on restitution of total contraction determined with test steps to 0 mV from a v_PC of 65 mV (Fig. 12A). The voltage-clamp protocols are shown at the far right of Fig. 12. For the ryanodine experiments, the conditioning pulses were to 0 mV. Ryanodine significantly (P < 0.05) decreased the magnitude of contractions initiated by activation steps from 65 to 0 mV (total contraction), and the plateau of the curve was reached at a shorter test interval than in control (Fig. 12A).

View larger version (36K): [in this window] [in a new window]   Fig. 12.   Effects of ryanodine and conditioning-pulse voltage on restitution of contraction. Experiments were conducted as described in Fig. 11. Schematics illustrating voltage-clamp protocols are shown at right. A: mean contraction-interval relationship for contractions initiated by a test step from 65 to 0 mV in the absence and presence of ryanodine. Ryanodine decreased the magnitude of contractions at all test intervals. B: contraction-interval relations determined with a test step from 65 to 30 mV to selectively activate VSRM. Restitution of VSRM followed a similar time course to that of restitution of total contraction and was similarly sensitive to ryanodine. C: contraction-interval relationships for contractions initiated by depolarizing steps from 35 to 0 mV to selectively activate ICa-L contractions. Contractions initiated by ICa-L showed minimal restitution and were virtually insensitive to ryanodine. D: effects of changing conditioning-pulse voltage on restitution of contractions initiated by a test step from 65 to 0 mV. Conditioning pulses to 40 mV greatly reduced the magnitude of contractions at all test intervals and virtually eliminated restitution. E: conditioning pulses to 40 mV inhibited restitution of contractions initiated by the VSRM. F: contractions initiated by ICa-L were not affected by conditioning pulses to 40 mV. Curves represent mean data from 5-7 experiments. * Statistically significant with respect to corresponding controls, P < 0.05.

We then determined whether contractions initiated by the VSRM exhibited restitution. VSRM contractions were elicited by a test step to 30 mV from a v_PC of 65 mV. Figure 12B shows changes in magnitude of VSRM contractions with progressive increase in test interval. In the absence of ryanodine, restitution of VSRM contractions (Fig. 12B) followed a time course similar to that for total contractions (Fig. 12A). Ryanodine had virtually the same effect on restitution of VSRM contractions as on total contraction (Fig. 12B, P < 0.05). Thus restitution of total contraction could be accounted for almost entirely by restitution of the VSRM component of contraction. We also evaluated restitution of contractions initiated at different test intervals by test steps to 0 mV from a v_PC of 35 mV to inactivate the VSRM. With this protocol, contractions were smaller and exhibited very little restitution (Fig. 12C). In addition, with this protocol, contractions at all test intervals were insensitive to 30 nM ryanodine.

We also examined the effects of conditioning-pulse voltage on restitution of contraction. Figure 12D shows the effects of changing the voltage of conditioning pulses from 0 to 40 mV on total contraction. When test steps were preceded by conditioning pulses to 40 mV, restitution was inhibited (P < 0.05). Figure 12E shows the effects of changing conditioning-pulse amplitude on contractions initiated by the VSRM. The magnitudes of contractions also were decreased, although the difference was not statistically significant. Figure 12F shows that contractions initiated by steps from 35 to 0 mV were virtually unaffected by changing conditioning-pulse voltage. Thus conditioning pulses to 40 mV and treatment with ryanodine virtually eliminate restitution of contraction, and this effect was mediated primarily through actions on the VSRM.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The objectives of this study were 1) to determine whether the VSRM exhibits steady-state inactivation and to determine the voltage range over which inactivation occurs, 2) to determine and compare the effects of protocols designed to alter SR Ca²⁺ load on components of contraction initiated by the VSRM or by I_Ca-L, and 3) to evaluate the role of the VSRM in restitution. Our observations demonstrate that the VSRM does show steady-state inactivation, is exquisitely sensitive to conditioning-pulse voltage and drugs that disrupt SR function, and is an important determinant of restitution of cardiac contraction.

One of the central observations of this study was that the VSRM exhibits steady-state inactivation properties. The VSRM was fully available when activation steps were made from membrane potentials more negative than 60 mV. The v_h of the VSRM was found to be approximately 48 mV in guinea pig and 53 mV in rat. Complete inactivation occurred near 35 and 40 mV in the two species, respectively. Clearly, earlier studies of cardiac EC coupling, which utilized holding or conditioning voltages near 40 mV, would have almost completely inactivated the VSRM (2, 3, 9, 10, 20). Indeed, the bell-shaped contraction-voltage or Ca²⁺ transient-voltage relationships reported in those studies likely reflects activation of contraction by I_Ca-L in the absence of the VSRM.

The steady-state inactivation properties of the VSRM help distinguish it from CICR coupled to I_Ca-L. The steady-state inactivation curve for the VSRM was found to be significantly different statistically from that for I_Ca-L, with respect to both v_h and k values. The steady-state inactivation curves for the VSRM and I_Ca-L were separated by 28 mV at v_h. Because of this wide separation, the VSRM was completely inactivated when v_PC was 40 mV but I_Ca-L was still fully available. Because the inactivation properties of the VSRM and I_Ca-L are widely divergent, it is very unlikely that the VSRM represents CICR initiated by I_Ca-L.

The steady-state inactivation properties of the VSRM allowed us to utilize two sequential activation steps to 40 and 0 mV to separate VSRM and I_Ca-L-induced contractions within a single voltage-clamp protocol. The ability of nifedipine to selectively inhibit I_Ca-L and the I_Ca-L-induced contraction with minimal effect on the VSRM contraction confirmed this separation. This selective effect of nifedipine also indicates that initiation of the VSRM contraction cannot be attributed to a very small influx of Ca²⁺ through L-type Ca²⁺ channels on the step to 40 mV, because inhibition of the large Ca²⁺ current on the step to 0 mV resulted in strong inhibition of the I_Ca-L-induced contraction.

The steady-state inactivation properties of the VSRM also serve to distinguish it from contractions triggered by reverse Na⁺/Ca²⁺ exchange. Contractions attributed to reverse Na⁺/Ca²⁺ exchange can be elicited by activation steps from either 70 or 40 mV (30). This indicates that contractions initiated by Na⁺/Ca²⁺ exchange do not show inactivation at 40 mV. Indeed, Na⁺/Ca²⁺ exchange is not known to exhibit voltage-dependent inactivation. Furthermore, contractions attributed to reverse Na⁺/Ca²⁺ exchange demonstrate sigmoidal, N-shaped, or progressively increasing contraction-voltage relations with activation steps from 40 mV, depending on the intracellular concentration of Na⁺ (16, 23, 29, 30). The present study was conducted with high-resistance microelectrodes, which would minimize intracellular dialysis and which did not contain Na⁺. Under these conditions the contraction-voltage relationships were always bell shaped when the v_PC was 40 mV.

Steady-state inactivation is generally described as a property of voltage-gated ion channels. In skeletal muscle, which also shows voltage-dependent release of SR Ca²⁺, the voltage sensor for EC coupling is believed to be the sarcolemmal L-type Ca²⁺ channel (24). The L-type Ca²⁺ channel is believed to be linked physically to the Ca²⁺ release channel (ryanodine receptor) in junctional SR (24). Interestingly, the release mechanism in skeletal muscle exhibits voltage-dependent inactivation (8, 24). Thus demonstration that the VSRM in cardiac myocytes also exhibits steady-state inactivation suggests that the voltage sensor for the VSRM may be a voltage-gated ion channel. The identity of the voltage sensor in cardiac myocytes is unknown. It would seem unlikely that the L-channel serves as the voltage sensor because the v_h of the cardiac L-channel is near 25 mV under our conditions as well as in other studies (21). However, we cannot completely exclude this possibility because the voltage sensitivity of L-type Ca²⁺ channels might change when they are coupled to another protein such as Ca²⁺ release channels in the SR. Of the known voltage-sensitive channels in the cardiac cell membrane, only the T-type Ca²⁺ channel has a steady-state inactivation curve that closely matches that of the VSRM. The v_h of the T-type Ca²⁺ channel has been reported to be near 50 mV, with a k value of ~5 mV (27, 31). These values correspond closely to the v_h and k values determined for the VSRM in the present study. However, in the present study we also were able to demonstrate the VSRM in rat ventricular myocytes, which have been reported not to have T-type Ca²⁺ current (28). This clearly excludes CICR in response to T-type Ca²⁺ current as the trigger for the VSRM, but does not necessarily exclude a role for T-channels as voltage sensors. T-channels might serve as voltage sensors in rat, but one would have to postulate that the T-channels have lost their current carrying capacity but retained their gating properties.

The intracellular Ca²⁺ transient that initiates cardiac contraction is derived from release of SR Ca²⁺ as well as influx of Ca²⁺ through the sarcolemma. The magnitude of Ca²⁺ released from the SR varies with SR load (14). Han et al. (14) demonstrated that test pulses to 0 mV activate only a small Ca²⁺ transient when preceded by conditioning pulses to 30 mV, to provide a low SR Ca²⁺ load. However, the same test pulse initiated a much larger Ca²⁺ transient, when preceded by conditioning pulses to 0 mV, to increase SR Ca²⁺ load. In the present study, contractions initiated by the VSRM and I_Ca-L were affected differentially by similar changes in conditioning-pulse voltage. VSRM contractions were greatly reduced in amplitude by conditioning-pulse protocols designed to reduce SR Ca²⁺ loading. This is compatible with the VSRM contraction, depending on release of SR Ca²⁺. In contrast, contractions initiated by activation of I_Ca-L were much less affected by conditioning-pulse voltage. These observations suggest that the I_Ca-L contraction, unlike the VSRM contraction, depends on both SR Ca²⁺ release and Ca²⁺ influx. This interpretation is also supported by the effects of ryanodine and thapsigargin, which interfere with SR function. Both agents strongly inhibited VSRM contractions but had much less effect on I_Ca-L contractions. Our observations with 30 nM ryanodine imply that, in undialyzed guinea pig cells at 37°C, part of the I_Ca-L contraction may be mediated by direct activation of myofilaments by Ca²⁺ influx. There are conflicting reports as to whether influx of Ca²⁺ via I_Ca-L is sufficient to initiate contraction directly (3). Effects of ryanodine vary widely depending on temperature, duration of exposure, concentration, stimulation interval, atrial versus ventricular tissue, and species (3). Thus in some species virtually all contraction can be eliminated by ryanodine (e.g., adult rat ventricle), whereas in others a ryanodine-resistant component has been observed (e.g., guinea pig ventricle) (3). For example, in field-stimulated guinea pig ventricular myocytes (18) and multicellular preparations (26) studied at physiological temperatures, 25-80% of contraction amplitude was retained in the presence of ryanodine and/or thapsigargin. However, these earlier studies did not separate effects of ryanodine or thapsigargin on the two components of EC coupling examined in this study.

Contraction-voltage relationships determined from a v_PC of 70 mV include both VSRM and I_Ca-L components of contraction. Both changes in conditioning-pulse voltage and ryanodine caused very large reductions in amplitudes of contractions initiated by steps from 70 mV. The component of contraction inhibited by both of these manipulations had a sigmoidal voltage dependence. The magnitude of the component that was inhibited was large even at very positive membrane potentials near the reversal potential of I_Ca-L. In contrast, the component that was inhibited was very small at positive membrane potentials when the VSRM was inactivated by a v_PC of 40 mV. Thus, when both the VSRM and I_Ca-L are available, most of the contraction elicited at potentials corresponding to the action potential peak and plateau appears to attributable to the VSRM.

Our studies also demonstrated an inhibitory effect of ryanodine on I_Ca-L that was apparent in the I-V relationships. It is unlikely that ryanodine directly inhibits the L-type Ca²⁺ channel, because experiments by others in which intracellular Ca²⁺ levels were strongly buffered showed no effect of ryanodine on I_Ca-L (1). In our experiments, in which intracellular Ca²⁺ was not buffered, inhibition of I_Ca-L by ryanodine may reflect Ca²⁺-mediated inhibition of L-type Ca²⁺ channels (21). Low concentrations of ryanodine lock the SR Ca²⁺ release channel in an open subconducting state; however, the SR can still take up Ca²⁺ (25). This Ca²⁺ can gradually leak out of the SR over a period of several hundred milliseconds (3, 26). In the protocols used to determine I-V relationships, test steps were preceded by conditioning pulses and a postconditioning period of 500 ms. A possible explanation for the effect of ryanodine on I_Ca-L is that the L-type Ca²⁺ channels may have been inhibited by Ca²⁺ leaking from the SR during the period between the last conditioning pulse and the test step. At present we have no direct evidence supporting or refuting this explanation. However, it is interesting to note that thapsigargin, which does not lock the release channel in an open subconducting state, had no effect on I_Ca-L.

The effects of ryanodine and conditioning-pulse voltage on I_Ca-L and on I_Ca-L contractions were slightly different depending on the voltage-clamp protocol used. For example, the I_Ca-L contraction was not significantly inhibited by conditioning pulses to 40 mV when sequential activation steps to 40 and 0 mV were used, but was significantly inhibited when contraction-voltage relations were determined. This difference might be attributable to differences in the activation sequence between the two protocols. With the sequential steps, the VSRM is activated immediately before the I_Ca-L contraction. Because the VSRM contraction is not accompanied by I_Ca-L in this protocol, activation of the VSRM might partially reduce SR Ca²⁺ and decrease the sensitivity of I_Ca-L contractions to changes in SR load.

The third goal of this study was to examine the role of the VSRM in recovery of contraction after a previous activation, a process termed restitution of contraction (5). In the present study we were able to demonstrate restitution of contraction in isolated myocytes under voltage-clamp conditions. Restitution was measured as the change in magnitude of test contractions initiated at different intervals after a previous activation. We found that the magnitude of the test contraction was negligible at short intervals but gradually increased with longer test intervals. Contractions initiated by depolarizing steps from 65 to 0 mV, to activate both VSRM and I_Ca-L components of contraction, showed restitution of contraction with a time constant near 60 ms. In contrast, the time constant of restitution of I_Ca-L was near 30 ms, which is similar to previous reports when one takes temperature into account (21). Thus the time course of restitution of contractions was slower than restitution of I_Ca-L. Therefore, restitution of contraction is not limited by recovery of I_Ca-L.

We found that contractions initiated by the VSRM, selectively activated by steps from 65 to 30 mV, also exhibited restitution. The magnitude of restitution of VSRM contractions was similar to that observed when both VSRM and I_Ca-L components of contraction were activated by steps from 65 to 0 mV. Thus recovery of the VSRM could account for much of restitution of contraction. This observation further demonstrates that restitution of contraction is not limited by recovery of I_Ca-L, because the VSRM contractions were elicited by voltage steps that did not activate measurable I_Ca-L. Indeed, I_Ca-L contractions initiated from a v_PC of 35 mV showed only minimal restitution and only during the first 25-50 ms after the previous activation.

Manipulations to decrease SR Ca²⁺ load greatly attenuated restitution of total contractions initiated by voltage steps from 65 to 0 mV. The same manipulations, conditioning pulses to 40 mV or exposure to ryanodine, also reduced restitution of contractions initiated by the VSRM. These observations indicate that the VSRM plays an important role in restitution of cardiac contraction and that restitution of the VSRM is very sensitive to manipulations designed to reduce SR Ca²⁺ load.

The component of EC coupling that determines the time course of restitution of contraction has yet to be identified. It is very unlikely that the time co