INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

MECHANICAL RESTITUTION is the time-dependent recovery of the ability of the heart to contract following the previous beat. That is, a premature stimulus or extrasystole introduced soon after the relaxation of a steady-state beat produces an attenuated twitch, the amplitude of which depends on the duration of the extrasystolic interval (ESI) (36). As the interval between the relaxation of a steady-state beat and the extrasystole is increased, there is an exponential increase in the amplitude of the extrasystolic beat. In fact, as the interval is prolonged beyond that of the basic cycle length, the twitch amplitude continues to increase and becomes greater than that of the steady-state beat before reaching a plateau (36). In 1975, Bass (3) termed this monoexponential recovery of force or pressure development phase II of mechanical restitution. He noted that when extrasystoles are interpolated very early and before relaxation of the steady-state beat, there is a period of enhanced relative force development (phase I) that is accompanied by action potential prolongation (4). The phase I enhancement of force development declines with increasing extrasystolic intervals, reaching a minimum value or reversal point (3). This reversal point can be taken as the boundary between phase I and phase II, but there undoubtedly is functional overlap, with the process(es) responsible for phase II possibly beginning near the extrapolated value for T₀.

Mechanical restitution and other aspects of the myocardial strength-interval relation (1, 24, 33) may reflect 1) changes in the gating mode of the L-type Ca²⁺ channels (dihydropyridine receptors, DHPRs) (29, 33); 2) changes in the extent to which Ca²⁺ release channels (ryanodine receptors, RyRs) of the sarcoplasmic reticulum (SR) recover the ability to release Ca²⁺ (34); 3) alterations in the amount of Ca²⁺ contained in the SR (6); and 4) cross-talk between the RyRs and the DHPRs wherein Ca²⁺ influx via the L-type Ca²⁺ channels triggers SR Ca²⁺ release and Ca²⁺ released from the SR inactivates the DHPRs (6). Each of these processes appears to be highly regulated and offers a potential target for therapeutic modalities designed to improve electromechanical function in diseased hearts.

The present study examined mechanical restitution in perfused rat left ventricles subjected to a variety of pharmacological perturbations. The biphasic nature of the mechanical restitution curves was abrogated by Ca²⁺ channel agonists, whereas phase II was eliminated by thapsigargin plus ryanodine. The results suggest that phase I of electromechanical restitution is dominated by changes in L-type Ca²⁺ currents, whereas phase II reflects recovery of the ability of the SR to release Ca²⁺. We also observed that phase II of mechanical restitution is markedly accelerated by agents that are thought to increase the Ca²⁺ content of the SR.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Isolated Langendorff Perfused Hearts

After a 10- to 15-min equilibration period, the atria were removed, and the atrioventricular node was crushed. This slowed the intrinsic heart rate to about 120 beats/min. Hearts were immersed in a 10-ml beaker bath containing outflow perfusate to prevent drying and to facilitate temperature control. Hearts were paced at a basic cycle length of either 333 or 500 ms using platinum electrodes, one of which was connected to the metal cannula and the other immersed in the beaker bath. Voltage was set at 5 V. A water-filled latex balloon connected via polyethylene tubing to a pressure transducer (Viggo-Spectramed) was inserted into the left ventricle, and the end-diastolic pressure was adjusted to 8 mmHg. Steady-state left ventricular developed pressure (LVDP) was typically between 120 and 150 mmHg. An IBM-XT fitted with a CIO-DACO2 card (Computerboards; Mansfield, OH) and run using a LabView virtual instrument acted as the stimulator. Data were acquired at a rate of 200 points/s.

Action Potential Measurements

The core of the optical mapping system is a 16 × 16 photodiode array (PDA) detector that combines a 16 × 16 element PDA with 256 current-to-voltage converters (feedback resistor of 10 m, resulting in a gain of 10⁷ V/A) in a single compact (136 × 136 × 154) enclosure. The PDA records from a large enough surface area of the preparation to receive enough photons per unit time for an accurate representation of an optical action potential. Subtraction of background fluorescence is done on a per-channel basis. The photocurrent produced by each photodiode is first stage amplified and converted to voltage by its own low-noise operational amplifier. Increasing the first stage amplification has been reported to improve signal-to-noise ratios in optical signals (14). The outputs of the first stage amplifiers are connected to 256-s stage amplifiers, which are reset immediately before data acquisition, to remove the DC offset of the optical signals caused by background fluorescence. Signals are filtered by Bessel filters and fed to a multiplexer and an analog-todigital converter board (Microstar Laboratories).

The spectroscopic properties of styryl dyes (RH-421, di-4-ANEPPS, and di-8-ANEPPS) have been shown to change linearly with membrane potential changes in the normal physiological range of transmembrane voltages in the axon (31) and heart (32). The precise depth of optical recordings in preparations stained by perfusion has been subject to debate. Model calculations by Salama (32), based on the depth of field of the optical system, predicted a depth of 144 µm. Measurements by Knisley et al. (23) in a tapering wedge of tissue showed that the intensity of optical action potentials ceased to increase if the thickness of the tissue was >300 µM. From the measurements of the absorption coefficient of myocardium for the excitation and emission spectrum, Girouard et al. (15) predicted that 95% of the signal originates from a tissue depth of 500 µM or less.

A major limitation of current techniques of potential imaging is the lack of absolute calibration. Unlike many ratio-metric fluorescent probes for calcium imaging, voltage-sensitive dyes can provide only relative information about changes in transmembrane voltage. Although the changes in the absolute amount of fluorescence excited at one wavelength linearly depends on transmembrane voltage of the viewed cells, accurate calibration has so far been impossible because the number of cells contributing to the signal is unknown.

To measure action potential (AP) amplitudes and areas at 50% repolarization, we chose to average fluorescence signals that originated from the surface of the central part of the left ventricle, where motion artifacts are minimal. For each ESI, we averaged signals coming from at least four adjacent photodiodes facing the central region of the epicardium. We measured AP amplitudes, and we calculated by integration the area under the AP from the origin of the upstroke up to the point corresponding to 50% repolarization. Because of the lack of absolute calibration, both areas under AP and AP amplitudes are normalized to the F₀ areas or amplitudes.

Most studies of electrical activity and mechanical restitution were done separately because the presence of an intraventricular balloon increased movement artifacts in the fluorescence traces. Agents such as butane dione monoxime (2) or cytochalasin D (38) that directly interfere with force development were not used in any of these studies. Instead, only those fluorescence traces showing a clear absence of movement artifacts, usually from the central region of the preparation, were analyzed.

Experimental Design

Pacing protocols. After a priming period of 100 beats at 3 Hz, test beats (extrasystoles) were introduced at varying time intervals, from as early as 40 ms to as late as 600 ms. Beyond 600 ms the hearts generally exhibited spontaneous activity. In experiments where high concentrations of thapsigargin and ryanodine were employed to disable the SR, the pacing rate was reduced to 2 Hz to allow for more complete relaxation of the steady-state beats. All extrasystoles were followed by an immediate return to the basic cycle length. That is, when the steady-state pacing frequency was 3 Hz, there was a standard 333-ms interval between the extrasystolic and the first postextrasystolic stimulus. Likewise, when the steady-state pacing rate was 2 Hz, there was a standard 500-ms interval between the extrasystolic and first postextrasystolic stimulus. Each pacing protocol was repeated two to three times for each heart, and the F₁/F₀ data were averaged to obtain a single set of values for each heart. The average standard error for triplicate F₁/F₀ values was ~2% of the F₁/F₀ ratio.

Reagents. BAY K 8644 and ryanodine were purchased from Calbiochem (LaJolla, CA). KB-R7943 (KBR) was the generous gift of Kanebo Pharmaceutical Laboratories (Osaka, Japan). All other reagents were purchased from Sigma (St. Louis, MO).

Data Analysis

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Example of a fused action potential (A) and twitch (B) for an electrical stimulus delivered 90 ms after a steady-state (SS) beat at 3 Hz. Also shown are resolved waveforms resulting from the subtraction of the previous SS beat. Solid line in B is the original recording. Dash-dotted line shows two SS contractions followed by an extension of the end-diastolic pressure offset by 333 ms and superimposed on the original pressure trace. Dotted line gives the results obtained by subtracting the dash-dotted waveform from the original pressure trace. ES BEAT, resolved left ventricular developed pressure (LVP) for the extrasystole; PEP, postextrasystolic potentiated beat.

Phase II of mechanical restitution was curve fit using Origin software and the equation Y = Y_max · {1  exp[(X  T₀)/B]}, where X is the ESI, T₀ is the X-axis intercept, and B = /ln 2, where  is the time constant for mechanical restitution. The values for the reversal point (nadir) and the data point immediately following it (e.g., ESIs of 230 and 250 ms in Fig. 2) were not used in the curve fitting because of the apparent overlap with phase I. 

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Composite mechanical restitution data for extrasystolic beats (F1/F0, filled squares) and the first postextrasystolic beat (F2/F0, open circles). Data are means ± SE; n = 15. Also shown is a typical steady-state beat (dashed line).

Comparisons between complete restitution curves (phase I plus phase II) were done using Origin software and a resident statistical package for repeated measures one-way analysis of variance (ANOVA). When the complete curves were significantly different from each other we used a Newman-Keuls post hoc test to determine which time points were significantly different between the two curves. We also used one-way ANOVA to determine which F₁/F₀ values were significantly different from 100%. A P value <0.05 was considered to be statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Mechanical Restitution

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Restitution of action potential (AP) amplitudes (open triangles) and AP areas (filled circles) at 50% repolarization. Data are means ± SE for 5 hearts. AP area ratios (F1/F0 × 100%) were significantly different from 100% at extrasystolic intervals (ESI) from 120 to 170 ms.

Figure 4 illustrates the effects of isoproterenol and the Ca²⁺ channel agonists BAY K 8644 and FPL-64176 on mechanical restitution. Effects of these compounds on steady-state left ventricular contractile performance are summarized in Table 1. Both -adrenergic stimulation and L-type Ca²⁺ channel activation significantly accelerated phase II of mechanical restitution, thereby decreasing the F₁/F₀ at the plateau. The Ca²⁺ channel agonists also eliminated the decline in F₁/F₀ between 150 and 230 ms. Isoproterenol did not alter the biphasic nature of the restitution curve but it shifted the reversal point leftward from 230 to 150 ms. Effects of the Ca²⁺ channel and -adrenoceptor agonists on phase II of mechanical restitution were partially reversed by the Ca²⁺-calmodulin (CaM) kinase inhibitor KN-93 (1 µM, e.g., Fig. 5). KN-93 had no significant effect on steady-state contractile parameters (n = 6). KBR, a putative inhibitor of the Na⁺/Ca²⁺ exchanger (21), slightly reduced the amplitude of the early phase I ratios, but the effect failed to reach statistical significance (Fig. 6). KBR had no statistically significant effects on steady-state LVDP or time to peak pressure (n = 4), but there was a significant decline in maximal rates of pressure increase (+dP/dt_max) (2,040 ± 230 vs. 2,520 ± 160 mmHg/s) and decrease (dP/dt_max) (1,260 ± 170 vs. 1,630 ± 90 mmHg/s).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Effects of isoproterenol (1 µM, A, n = 3), FPL-64176 (150 nM, B, open squares, n = 4) and BAY K 8644 (100 nM, A, filled circles, n = 6) on mechanical restitution.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Effect of 1 µM KN-93 on mechanical restitution in hearts perfused with 150 nM FPL-64176. Open circles, +KN-93; closed squares, KN-93. The two curves are significantly different at ESIs between 230 and 290 ms. (FPL-64176, n = 5; FPL + KN-93, n = 4.)

View larger version (14K): [in this window] [in a new window]   Fig. 6.   There is no statistically significant effect of 10 µM KB-R7943, a putative inhibitor of Na+/Ca2+ exchange on mechanical restitution. Filled squares, control; open circles, +KB-R7943 (n = 5).

To assess the involvement of the SR in phase I of mechanical restitution, hearts were perfused with thapsigargin, a specific inhibitor of the SR Ca²⁺ pump (22), plus ryanodine, a compound that renders the SR leaky to Ca²⁺ (17). This reduced peak systolic left ventricular pressure (LVP) to 88 ± 22 mmHg, increased end-diastolic pressure to 48 ± 14 mmHg, decreased +dP/dt_max to 410 mmHg/s, and decreased dP/dt_max to 235 ± 62 mmHg/s (n = 3). Figure 7 is a typical pressure recording. Note the greatly enhanced relative amplitude of the extrasystole at 170 ms (basic cycle length, 500 ms) and the absence of postextrasystolic potentiation. Figure 8 summarizes the mechanical restitution data for a group of hearts perfused with thapsigargin plus ryanodine. With the SR unable to participate in excitation-contraction coupling, phase I of mechanical restitution was greatly exaggerated, whereas phase II was nearly flat. KN-93 (3 µM) had no statistically significant effect on mechanical restitution in the presence of thapsigargin plus ryanodine (Fig. 8).

View larger version (32K): [in this window] [in a new window]   Fig. 7.   Effects of 3 µM thapsigargin (Thapsi) plus 200 nM ryanodine on steady-state LVP (A) and on an extrasystole triggered at 170 ms after a steady-state beat at 2 Hz (B). Note the large potentiation of the extrasystole and the absence of postextrasystolic potentiation.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Effects of 3 µM thapsigargin plus 200 nM ryanodine ± 3 µM KN-93 on mechanical restitution. Closed squares, no KN-93 (n = 6), open circles, +KN-93, (n = 5). Hearts were perfused with thapsigargin plus ryanodine for at least 10 min before the restitution experiments. The two curves were not significantly different.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present study demonstrates that the electromechanical restitution of perfused rat left ventricles consists of two distinct phases: phase I, which is completed before the full relaxation of the steady-state beats, and phase II, which occurs later. Aside from the classic 1975 papers by Bass (3, 4), which described the two phases of electromechanical restitution in cat papillary muscles, most investigations have focused on phase II, specifically on that portion of phase II that follows complete relaxation of the steady-state beat (1, 36, 40). In such studies, there typically are mathematical extrapolations to zero developed force or zero intracellular Ca²⁺ concentration ([Ca²⁺]_i) transient amplitude as the ESI is shortened; fused waveforms are not dealt with (36). Phase II appears to be a monoexponential process that results from the recovery of the ability of the RyRs of the SR to release calcium (34, 36). This may result from the recovery of the RyRs from inactivation (34) or adaptation (16). There could also be a time-dependent refilling of highly localized SR Ca²⁺ stores (8).

The characterization of phase I of mechanical restitution at first appears to be problematic in that it requires the computerized subtraction of a steady-state beat to separate out the amplitude and configuration of very premature extrasystoles. This raises questions about whether force development, APs, and Ca²⁺ transients can realistically be thought of as additive for fused responses that result from two closely spaced stimuli. The curve describing the relative strength of the first postextrasystolic beat at short ESIs, which is similar in shape to the F₁/F₀ curve, suggests that the computerized subtractions do provide realistic data. Extra calcium entering the cell in response to a premature stimulus is thought to be taken up by the SR, becoming available for release during the subsequent beat (5, 28). When the premature stimulus occurs while the RyRs are refractory, the extra contractile response of the postextrasystolic beat to this extra Ca²⁺ is roughly proportional to the amount of Ca²⁺ entering the cell during the extrasystole. Thus the time course and magnitude of changes in F₁/F₀ as a function of the ESI for both the extrasystoles and the associated first postextrasystolic beats (F₂/F₀) were similar during phase I. Moreover, rat ventricular APs, because of their brief duration, are less fused than the corresponding mechanical traces at short ESIs, and their AP areas at 50% repolarization also paralleled the computer-resolved LVDPs for early extrasystoles.

The present study demonstrated that phase I of electromechanical restitution is independent of SR Ca²⁺ release and can be observed when the SR is chemically disabled with thapsigargin plus ryanodine and is unable to accumulate or retain significant amounts of Ca²⁺. Phase II, on the other hand, appears to be almost entirely attributable to recovery of the SR Ca²⁺ release process because this phase is flat when the SR can no longer accumulate or retain Ca²⁺.

Interestingly, neither phase was significantly affected by the putative cardiac Na⁺/Ca²⁺ exchange inhibitor KBR (19, 25). We used 10 µM KBR because of a reported IC₅₀ of 1.2-2.4 µM for reverse mode Na⁺/Ca²⁺ exchange inhibition. Concentrations higher than 10 µM might have significantly affected other ion channels (19), which would have confounded interpretation of the data. Ladilov et al. (25) found 20 µM KBR to be the maximally effective dose in studies of perfused rat hearts and isolated adult rat cardiomyocytes, but significant inhibition was obtained with 10 µM KBR. Because there was a trend toward a decrease in the maximum F₁/F₀ ratios during phase I in hearts perfused with 10 µM KBR, a portion of the increase in F₁/F₀ during phase I in the absence of drug could have been due to Ca²⁺ influx via reverse mode Na⁺/Ca²⁺ exchange.

Phase I of mechanical restitution was markedly exaggerated by perfusion with thapsigargin plus ryanodine, which were used to limit respectively SR Ca²⁺ uptake and retention. On the other hand, phase II of mechanical restitution and postextrasystolic potentiation were abolished by the SR poisons. (Identical effects were obtained with thapsigargin alone, data not shown.) Thapsigargin is a potent and selective inhibitor of the SR Ca-ATPase (22), and it prevents Ca²⁺ uptake by the SR. Ryanodine leaves the Ca²⁺ release channels (RyRs) of the SR in an open but subconducting state (17), precluding significant retention of any SR Ca²⁺ that could have been accumulated despite the presence of a maximally effective concentration of thapsigargin (22).

Janczewski and Lakatta (20) demonstrated that thapsigargin increases [Ca²⁺]_i transients in guinea pig ventricular cardiomyocytes after SR Ca²⁺ depletion with caffeine, indicating that some Ca²⁺ entering the cell via the L-type Ca²⁺ channels is immediately sequestered by the SR and does not contribute to the global [Ca²⁺] transient. This effect of thapsigargin may have contributed to the amplitude of the steady-state beats shown in Fig. 7, which are larger than what would be predicted from data showing that SR Ca²⁺ release contributes >90% of myofibrillar activator Ca²⁺ in response to a normal rat heart AP (5) (also see Ref. 22). Our results are consistent with early studies of the effects of caffeine on cat papillary muscles (3), but thapsigargin and ryanodine are more specific inhibitors of SR function, leaving the data less open to alternative explanations.

A study of mechanical restitution in ferret papillary muscles failed to detect an overshoot in Ca²⁺-dependent aequorin luminescence for early fused beats in the presence of ryanodine (36), as might have been predicted by our data. Unfortunately, in that study, fused mechanical responses were not resolved, leaving it unclear whether such muscles exhibited an exaggerated mechanical phase I. It is also unclear whether ryanodine alone is as effective as ryanodine plus thapsigargin in eliminating the contribution of the SR to excitation-contraction coupling. The present study did not examine the effects of ryanodine alone on electromechanical restitution.

It should be emphasized that the exaggeration of phase I of mechanical restitution in the presence of thapsigargin and ryanodine was associated with a ~70% reduction in steady-state LVDP. LVDP averaged only 40 mmHg compared with ~130 mmHg for control steady-state beats. Because the peak F₁/F₀ shown in Fig. 8 is 190%, it follows that the mean peak F₁ amplitude is only ~75 mmHg, or about half that of typical control steady-state beats at 3 Hz. Nevertheless, perfusion with thapsigargin plus ryanodine allowed a near doubling of LVDP for very early extrasystoles, over and above that which could be attributed to the absence of L-type Ca²⁺ current inactivation by SR Ca²⁺ release (6) and the lack of SR sequestration of Ca²⁺ entering through the L-type Ca²⁺ channels (20). It has been estimated that L-type Ca²⁺ current can double when Ca-dependent current inactivation by Ca²⁺ released from the SR is eliminated (6). Our data suggest that a further doubling of Ca²⁺ influx is possible for very premature beats in the intact rat heart and emphasize the importance of Ca²⁺-dependent L-type Ca²⁺ current inactivation and facilitation as mechanisms for grading cardiac contractility.

As shown in the companion paper (9), electromechanical restitution kinetics are altered significantly in failing spontaneously hypertensive, heart failure (SHHF-Mccfa^cp) rat hearts. We therefore initiated a series of experiments to characterize factors that might regulate either phase I or phase II of mechanical restitution in normal rat hearts and thus possibly explain the differences in restitution kinetics accompanying heart failure. Phase II could be accelerated dramatically by positive inotropic agents such as the -adrenergic agonist isoproterenol, and by the dihydropyridine Ca²⁺ channel agonist BAY K 8644 and the nondihydropyridine Ca²⁺ channel agonist FPL-64176. All three of these compounds would be predicted to increase the amount of Ca²⁺ contained in the SR due to the increase in L-type Ca²⁺ current. Isoproterenol should have the additional effect of causing phosphorylation of phospholamban, the endogenous inhibitor of the SR Ca-ATPase (7). In vivo studies of mechanical restitution in transgenic mice with varying amounts of myocardial phospholamban (18) indicated that restitution was fastest in phospholamban knockout mice, intermediate in wild-type mice, and slowest in mice overexpressing phospholamban in the heart. These mouse data lend support to the concept that SR Ca²⁺ content is a major determinant of the rate of phase II mechanical restitution.

The acceleration of phase II of mechanical restitution by isoproterenol and the Ca²⁺ channel agonists was retarded slightly by the CaM kinase II inhibitor, KN-93. Thus CaM kinase-dependent phosphorylation of RyR2 (37) or an accessory protein (27) may have contributed somewhat to the acceleration of phase II of mechanical restitution. However, the pronounced increase in SR Ca²⁺ load caused by these positive inotropic agents was probably responsible for most of the acceleration of mechanical restitution (6). All three drugs induced spontaneously reversible periods of visibly disorganized contractile behavior and zero LVDP at the initiation of pacing. This behavior was suggestive of ventricular fibrillation, which is commonly associated with an SR Ca²⁺ overload (6, 26). Under conditions where -adrenoceptor stimulation does not increase the SR Ca²⁺ load, it has been reported that there is no effect of -agonists on the recovery of the ability of cardiac RyRs to release Ca²⁺ (35).

Isoproterenol had little effect on phase I of mechanical restitution, except to abbreviate it, owing for the most part to a greater overlap with the accelerated phase II and a shift of the reversal point from 230 to 150 ms. The two chemically distinct Ca²⁺ channel agonists, on the other hand, abrogated the reversal point separating phase I and phase II. This would be explained if the Ca²⁺ channel agonists elicited maximum activation of the L-type Ca²⁺ channels and if the rise in F₁/F₀ during phase I with Ca²⁺ channel agonists absent were due to transient voltage and/or Ca²⁺-dependent L-type Ca²⁺ channel facilitation coupled with the time-dependent recovery of electrical excitability. Facilitation involves an increase in the type II L-type Ca²⁺ channel gating mode that is characterized by very prolonged openings (29). Ca²⁺ channel agonists also act to increase type II openings (13). The subsequent decline in F₁/F₀ in the absence of the Ca²⁺ channel agonists could then be due to the time-dependent decay of facilitation, which would not occur with the Ca²⁺ channel agonists present. Isoproterenol also increases type II openings (29), but increases in SR Ca²⁺ uptake as a consequence of phospholamban phosphorylation (11) account for a large fraction of the resulting positive inotropic effect. As shown in Table 1, the maximally effective dose of isoproterenol (1 µM) had much larger effects on ±dP/dt_max than did the Ca²⁺ channel agonists, suggesting a more pronounced effect on the kinetics of SR Ca²⁺ accumulation and possibly a lesser effect on the L-type Ca²⁺ channels.

Surprisingly, phase I of mechanical restitution was not affected by CaM kinase inhibition with KN-93, despite good evidence that Ca²⁺ current facilitation in single cardiomyocytes results from CaM kinase activation (10, 39). It should be noted that the concentration of KN-93 used in the experiments summarized in Fig. 8 (3 µM) was threefold higher than that used for the experiments summarized in Fig. 5, where a small but statistically significant effect of KN-93 was observed. Thus we should have seen an effect of KN-93 on phase I of mechanical restitution in the presence of thapsigargin plus ryanodine if CaM kinase II were responsible for the observed increase in F₁/F₀ at short ESIs. That we did not raises the possibility that cardiac Ca²⁺ current facilitation is not a single process in the intact heart but is instead several processes that together regulate frequency-dependent changes in Ca²⁺ influx across the sarcolemma. The changes observed in the present study peaked at ~150 ms at a slightly hypothermic temperature of 32°C and relatively slow pacing frequencies for rat hearts of 2-3 Hz. Most evidence for CaM kinase-dependent Ca²⁺ channel facilitation has dealt with longer time scales, as might be more appropriate for a phosphorylation-dependent mechanism.

Both Ca²⁺-dependent facilitation and inactivation of L-type Ca²⁺ channels require the presence of calmodulin bound to an isoleucine glutamine (IQ) motif in the carboxy tail of the ₁-subunit (41). Replacement of the native isoleucine of the IQ domain with alanine removes Ca²⁺-dependent channel inactivation and unmasks a strong facilitation, whereas conversion of the isoleucine to glutamate eliminates both facilitation and inactivation (41). A peptide representing another region in the carboxy-terminal sequence of the L-type Ca²⁺ channel, the CB peptide, binds Ca²⁺ CaM and enhances Ca²⁺-dependent Ca²⁺ current (I_Ca) facilitation when injected into cardiac myocytes, again suggesting a direct effect of Ca²⁺CaM on this process (30).

Limitations of the Present Study