INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DURING THE CARDIAC ACTION POTENTIAL there are complex relationships between the membrane potential, the conductances and fluxes of specific ions, and the intracellular concentration changes and compartmentalization of specific ions, particularly intracellular calcium. The L-type calcium current (I_Ca) is generally regarded as an early plateau current that, along with possible contributions from the Na⁺/Ca²⁺ exchanger (9, 10), produces calcium entry that triggers a large release of calcium from the sarcoplasmic reticulum in response to the generation of an action potential. Recent work (13, 15, 17) has shown that I_Ca has a more complex role during action potential conduction, which is discontinuous because of low coupling conductance between cardiac cells. Specifically, I_Ca becomes important as an inward current required to sustain propagation when conduction delays of greater than a few milliseconds are produced. Under these circumstances conduction is inhibited by blockers of I_Ca and facilitated by agents that increase I_Ca (5, 17). In addition, voltage clamping cells with action potential waveforms recorded during discontinuous conduction (8) has shown that discontinuous conduction may also increase the magnitude of I_Ca compared with its value in an isolated cell. A possible contribution to the increase in the calcium current during discontinuous conduction is that the partial repolarization following the peak of the action potential (which is associated with discontinuous conduction) effectively pulls the membrane potential away from the equilibrium position for I_Ca and thus increases I_Ca. In addition, the conductance of the L-type calcium channel is voltage dependent, and because discontinuous conduction alters the membrane potential, the conductance of the calcium channel during the action potential will also be altered, perhaps contributing to the increase in I_Ca.

The positive linkage between calcium entry and the subsequent release of calcium from the sarcoplasmic reticulum suggests that the process of discontinuous conduction may actually alter the time course and magnitude of the intracellular calcium concentration ([Ca²⁺]_i ) during an action potential. This possibility is particularly interesting because alterations in [Ca²⁺]_i are important for the regulation of the Na⁺/Ca²⁺ exchange current as well as contributing to the inactivation process of I_Ca. In addition, if these alterations in the waveform of [Ca²⁺]_i were cumulative and persisted into the diastolic period, there might be consequences for the gap junctional conductance because these channels have themselves been shown to be sensitive to [Ca²⁺]_i (20).

Measurements of the waveform and amplitude of [Ca²⁺]_i during a normal action potential with a variety of calcium-sensing dyes and light-sensing techniques have been made. Beuckelmann and Wier (1) recorded action potentials from isolated guinea pig ventricular cells at room temperature while monitoring [Ca²⁺]_i with the fura 2 dye and obtained resting values for [Ca²⁺]_i of 120 nM and a peak value of 1,100 nM. Yao et al. (23) used the fluo 3-AM calcium indicator on rabbit ventricular cells at 32°C and obtained a resting value for [Ca²⁺]_i of 239 nM and a peak value of 700 nM. The increased speed and availability of confocal laser scanning fluorescence microscopes has facilitated the measurement of [Ca²⁺]_i from cardiac cells, particularly for nonstimulated cells at room temperature from which the phenomenon of "calcium sparks" can be examined (2, 14, 18, 21).

To test the hypothesis that the process of action potential conduction, especially when discontinuous, alters the amplitude and waveform of [Ca²⁺]_i in cardiac cells, we have combined the technique of monitoring [Ca²⁺]_i via the intracellular fluo 3 indicator for an isolated cardiac cell with the simultaneous use of our previously developed "coupling clamp" technique (22) to allow action potential propagation from the real cell to a real-time simulation of a model cardiac ventricular cell (11). With this technique we can vary the coupling conductance, and thus change the degree to which conduction is discontinuous, and also study the action potential and the [Ca²⁺]_i waveform in the same cell when uncoupled from the model cell.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell isolation, electrodes, and solutions. The enzymatic procedure for single-cell isolation of ventricular cells was similar to that of Yazawa et al. (24) as described in our previous work (7). Hearts were removed from guinea pigs weighing 300-600 g that were anesthetized (intraperitoneally) using 50 mg/kg Nembutal injected with 500 units of heparin. The heart was cannulated for Langendorff perfusion and perfused for 3-5 min at a rate of 6-10 ml/min with normal Tyrode solution. After the blood was washed out from the coronary arteries, the heart was perfused with nominally Ca²⁺-free Tyrode solution for 5-6 min. The heart was then perfused with the nominally Ca²⁺-free Tyrode solution containing collagenase (55 mg/100 ml, type I; Sigma, St. Louis, MO) and protease (20 mg/100 ml, type XXIV; Sigma) for 4-6 min. The enzymes were then washed out from the heart with a high-K⁺, low-Cl storage solution for 5 min. After perfusion of the high-K⁺ storage solution, the right ventricle and the ventricular septum were cut into pieces and gently triturated in the high-K⁺ storage solution, filtered through a nylon mesh, and stored at 4°C. The isolated cells were transferred to an experimental chamber (Medical Systems), allowed to settle on the thin glass bottom, and continuously superfused with normal Tyrode solution at 2 ml/min at 36-37°C. Only quiescent cells with preservation of their rod-shaped appearance were studied. The pipettes were pulled from borosilicate glass and, after fire polishing, had resistances of 3-5 M when filled with the internal solution. High-resistance seals were formed with the cell membrane using light suction, and the membrane was disrupted by applying a transient suction. The junctional potential was corrected by zeroing the potential before touching the surface of the cell with the pipette tip. Recordings of membrane potential were made with an Axoclamp 2A amplifier (Axon Instruments, Foster City, CA) in the current-clamp mode.

Coupling an isolated ventricular cell to a model cell. We developed an electrical circuit that has the ability to provide a variable effective coupling conductance between two isolated heart cells that are not actually in direct contact with each other (6). We extended this method to include the ability to couple a real cell to a real-time solution of a cell model (22) or to a linear strand of cell models (19). In the present study we further extended this technique to incorporate the ability to measure the intracellular calcium transients of the real cell, using a high-speed confocal laser scanning microscope, while simultaneously coupling the real cell to a model cell. Figure 1A shows how a real, isolated ventricular cell (real cell 1) is coupled to a model cell (model cell 2) by an ohmic coupling conductance (G_C). The resulting coupling current (I_C) flows between the two cells and is given by I_C = G_C · (V₁  V₂), where V₁ is the membrane potential of the real cell and V₂ is the membrane potential of the model cell. Figure 1B shows how the technique is realized by recording from a real cell in the current-clamp mode. The model cell is implemented as a real-time simulation of the Luo and Rudy (LR) (11, 12) guinea pig ventricular action potential model, as modified by Zeng et al. (25). The LR model includes sarcolemmal ionic channel currents and pump currents as well as a representation of calcium ion concentration with cytoplasmic buffers and the release and uptake of calcium by the sarcoplasmic reticulum. For each time step (t), the computer samples the voltage present in the real cell via an analog-to-digital converter. The value of coupling current is calculated for this time step, and a voltage proportional to this current is sent to the analog circuit via a digital-to-analog converter and then to the real cell via the voltage-to-current converter. The computer then updates the membrane potential of the model by integrating for one time step with the same coupling current as an additional ionic membrane current. The process is repeated at successive time steps of 60 µs. Custom-written model clamp software was compiled as a DOS real-mode application using Active Ada (version 5.3; Alsys, Boston, MA) and run on a 300-MHz Pentium Pro processor computer equipped with a fast 12-bit data acquisition board (model DigiData 1200, Axon Instruments).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Diagram of experimental setup for coupling a real, isolated ventricular cell to a real-time solution for a model cell using "coupling-clamp" technique (see METHODS). A: coupling of a real cell to a model cell by an ohmic coupling conductance GC. IC, coupling current. B: schematic showing how the coupling clamp circuit is realized along with simultaneous recording of intracellular calcium using a confocal scanning laser microscope (CLSM). V1, membrane potential of real cell 1; V2, membrane potential of model cell 2; Z1, size factor; t, time; t, time step; V to I, voltage-to-current converter; A/D, analog-to-digital converter; D/A, digital-to-analog converter.

Normalizing cell size. Because the LR model cell has a standard size and real ventricular cells do not, it is important to define some method of standardizing the "size" of the real cell. In our coupling model we included the ability to change the effective cell size of the real cell by simply scaling the coupling current that is being injected into the real ventricular cell (real cell 1) by a size factor Z₁. This produces an effective increase in the size (as represented by an increase in the current threshold and a decrease in the input resistance) of real cell 1 by a factor of 1/Z₁. In our experiments we have used a size factor Z₁ of 2.6/I_th, where I_th (in nA) is the current threshold of the real cell with current pulses of 2 ms in duration. The value of 2.6 nA was chosen as the value required for a similar stimulation of the standard-size LR model cell. Thus, using the above size factor, the real cell and the LR model cell were of equal size in our experiments, both exhibiting an effective current threshold of 2.6 nA with current pulses of 2 ms in duration. The actual value of stimulus current threshold (before normalization of the cell size) for the five cells studied was 2.5 ± 0.2 nA (mean ± SE).

Fluorescence measurements with confocal laser scanning microscopy. Intracellular calcium was monitored using the calcium fluorescence indicator fluo 3 (pentapotassium salt, Molecular Probes), which was added to the pipette solution (50-75 µM) and dialyzed into the cell. The fluorescence imaging was performed with a confocal laser scanning microscope (OZ, Noran Instruments) coupled to an inverted microscope (IX-70, Olympus) equipped with a ×60 water-immersion objective (NA = 1.2, UPlanApo, Olympus). Fluo 3 fluorescence was excited with the 488-nm line of a krypton-argon laser. Emitted fluorescence was measured at wavelengths >500 nm.

Experimental protocol. To align the fluorescence measurements with the membrane potential recordings, it was necessary to synchronize the acquisition of the confocal images with the acquisition of the membrane potential of the real cell and the corresponding computed potential of the model cell. For each experiment, the real cell was stimulated at 1 Hz using a MECA programmable stimulator applied to the external step command of the Axoclamp 2A amplifier. The PC was triggered to begin the protocol at the start of a train of 10 beats. The stimulator was programmed to send a transistor-transistor logic (TTL) pulse to the SGI computer 130 ms before the eighth beat of the train, which signaled the confocal system to begin acquisition of the fluorescence images. At 70 ms before the eighth beat of the train, another TTL pulse was sent to a light-emitting diode placed in the path of the photomultiplier tube of the confocal system, resulting in saturated images during the diode flash. Knowing that the diode flash began 70 ms before the stimulation of the eighth beat of the train enabled us to align the fluorescence images with their corresponding membrane potential recordings.

Statistical analysis. Statistical analysis was performed using SigmaStat for Windows (Jandel Scientific). Statistical significance was determined by Student's t-test for paired data. P values <0.05 were regarded as significant. Data are presented as means ± SE in the text.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In previous work we showed that discontinuous conduction between two coupled guinea pig ventricular cells results in a larger I_Ca in the leader cell than in the follower cell (8). We hypothesized that the increase in I_Ca would result in an increased intracellular calcium transient. Both uncoupled and coupled membrane potential records are shown in Fig. 2A. Note that the time axis was adjusted to align the maximum upstroke velocity of the membrane potentials of the real cells. Traces a and c in Fig. 2A are the membrane potentials of the real cell when G_C was set to 0 nS (control); trace b shows the membrane potential for the real cell (solid line) when G_C was set to 8 nS and the corresponding membrane potential of the LR model cell (dashed line). There was a 10.7-ms delay between the upstroke of the real cell and the upstroke of the model cell. The low value of G_C limited the amount of current that passed to the model cell; thus the model cell was slowly brought to threshold, resulting in a long delay between the activation of the two cells. We chose to study discontinuous conduction by setting G_C to 8 nS for several reasons. First, 8 nS is slightly larger than the critical value of G_C (7.0 ± 0.2 nS, n = 8) required for propagation between a real guinea pig ventricular cell and an LR model cell as determined in our previous work (22). Second, G_C at 8 nS was large enough to consistently provide sufficient current for propagation to occur between the two cells.

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Membrane potential recordings and corresponding calcium transients from real ventricular cells under control and discontinuous conduction conditions. A: recordings from a guinea pig ventricular cell (solid line) and a simultaneous real-time computation of Luo and Rudy (LR) ventricular cell model (dashed line). Recordings represent 8th beat of a train of 10 action potentials, where the real cell is repetitively stimulated at 1 Hz with a 2-ms current pulse. Traces a and c: membrane potential of real cell when uncoupled from model cell (GC = 0 nS). Trace b: membrane potential of real cell when GC = 8 nS, with associated membrane potential of model cell. B: corresponding calcium transients recorded from real cell and plotted as fluorescence ratio F/Fo, where F is fluorescence at a given time and Fo is average fluorescence measured 40 ms before cell was stimulated.

The calcium transients that correspond to the membrane potentials of Fig. 2A are plotted in Fig. 2B. As described in METHODS, the calcium transients are plotted as F/F_o. The calcium transient (Fig. 2B, trace a) corresponds to the first control, for which G_C was set to 0 nS. The calcium transient began to rise shortly after the action potential was initiated (Fig. 2A, trace a) and reached its peak value within 40 ms. The transient was maintained throughout the plateau of the action potential and began to decrease as the action potential began to repolarize, but the decrease in the calcium transient was much slower than the repolarization of the action potential. After repolarization of the action potential, the calcium transient had not yet completely returned to its diastolic value for the time interval plotted in Fig. 2B.

When the real cell was then coupled to a model cell with a G_C of 8 nS, the calcium transient was altered, as shown in Fig. 2B, trace b. The time course of the upstroke of the calcium transient was faster in the case of a propagating action potential compared with the uncoupled system. In additional, the calcium transient reached a higher peak within 20 ms of the action potential initiation. The overshoot of the calcium transient was short-lived and within 45 ms had returned to the value of the peak of the control calcium transient. Thus, although discontinuously coupling two cells did cause an increase in the calcium transient, that increase was not maintained throughout the action potential. In fact, after ~50 ms, the control and the coupled calcium transient had a very similar time course. After the cells were coupled, we again did a control experiment (G_C = 0 nS), and the results are shown in Fig. 2, A and B, trace c. The calcium transient for the second control is similar in magnitude and time course to that for the control before coupling, demonstrating that the effects of discontinuous conduction were completely reversible.

Because coupling two cells at 8 nS caused a brief change in the calcium transient, we replotted the first 35 ms of Fig. 2 in Fig. 3, omitting the second control (trace c) for clarity. As shown in Fig. 2, when the real cell was coupled to a model cell, the corresponding calcium transient rose faster and obtained a higher value. For this experiment, the images of the calcium transient were taken with a time step of 2.08 ms. For both the uncoupled and the coupled case, the calcium transient began within 4 ms of the action potential upstroke. When coupled, the calcium transient had a much faster upstroke, which can be seen by comparing Fig. 3B, traces a and b. The increase in the rate of the rising phase of the calcium transient was caused by the rapid repolarization of the membrane potential of the real cell when it was coupled to a model cell by 8 nS. As shown in Fig. 3A, trace b, after the action potential was initiated in the real cell, there was a delay of 10.7 ms before the model cell was activated. During this time the membrane potential of the real cell was pulled down by the load of the model cell, resulting in a rapid repolarization of 25.6 mV. The rapid repolarization increased the calcium current, thus bringing more calcium into the cell, resulting in a faster, larger calcium transient. For five cells studied, the peak calcium transient for control was 2.09 ± 0.24 and increased to 2.43 ± 0.28 (P < 0.01). This was an average increase of 16%.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Data in Fig. 2 plotted on an expanded time scale. Trace c was deleted for simplicity. A: trace a represents membrane potential of real cell when GC = 0 nS, whereas trace b represents membrane potential of real cell when GC = 8 nS. B: rising phase of calcium transient corresponding to membrane potential recordings in A.

The images that correspond to the calcium transients in Fig. 3B are shown in Fig. 4. In this instance, the long axis of the cell was nearly perpendicular to the long axis of the image window; thus a small portion of the cell was imaged. For an image time step of 2.08 ms, the image window is ~53 × 8 µm in size. In Fig. 4, the image has been cropped so that only the portion of the image containing the cell (30 × 8 µm) is shown. The images were color coded so that blue and green indicated low levels of calcium, and as calcium increased, the color changed from yellow to red. In Fig. 4, the faster time course of the calcium transient for a coupled cell is easily noted. At time 0, the images for both the uncoupled and coupled records are similar. As time increased to 8.32 ms, it is clear that the fluorescence was greater when discontinuous conduction occurred than when the cell was uncoupled.

View larger version (80K): [in this window] [in a new window]   Fig. 4.   Confocal images corresponding to calcium transients shown in Fig. 3B. Images at left are from the uncoupled cell, whereas images at right are from the cell coupled with a GC of 8 nS. Images were acquired with a time step of 2.08 ms and have been cropped so that only that portion of image containing the cell is shown. Images are color coded so that blue and green indicate low levels of calcium; as calcium level increases, color changes from yellow to red. Image labeled "background" was recorded before cell was loaded with fluo 3, and average pixel value of this image was subtracted from fluorescence measurements before fluorescence ratio was calculated.

To calculate the fluorescence intensity and plot it as F/F_o, we drew a region of interest around the cell and averaged the pixel values within that region. For this particular cell, the region of interest was 151 µm², which corresponds to 2,616 square pixels. Because we loaded the fluorescent dye via the pipette, we were able to capture an image of the cell without any dye, recording the autofluorescence of the cell before touching the cell with the pipette. An example of the background image for this particular cell is shown in Fig. 4. This image had a much lower signal than either of the baseline images for the uncoupled or coupled records (time = 0.0 ms). The average pixel value for the region of interest for the background image was 24.7. (The confocal laser scanning microscope "mapped" the fluorescence counts at each pixel to a value from 0 to 255 that represents F as a unitless arbitrary measurement of the intensity of light.) The background fluorescence was subtracted from the fluorescence measurements before the fluorescence ratio was calculated.

To determine that the individual pixels of the images were not being saturated at the peak of the calcium transient, in Fig. 5 we plotted histograms of the region of interest drawn within the cell boundaries. Figure 5, A and C, corresponds to the images shown at time = 0.0 ms in Fig. 4. The mean pixel value in the histogram for the control (Fig. 5A) was 48.9, whereas the mean pixel value for the histogram for the coupled record (Fig. 5C) was 50.3; the closeness of these values indicates a stable baseline value. The maximum pixel value allowed is 255, and for Fig. 5, A and C, no pixels reached that value. Figure 5, B and D, corresponds to the images shown in Fig. 4 at time = 18.72 ms. The mean pixel value in the histogram for the uncoupled record (Fig. 5B) was 80.9, whereas the mean pixel value for the coupled record was 100.3. The image for the control had no pixel values equal to the maximum value (255), whereas the image for the coupled calcium transient had a few pixels that were saturated, but the bulk of the pixels were below saturation. If many pixels had reached the saturation level, we would have been underestimating the maximum level of the calcium transient. These histograms indicate that we were utilizing the full range of the system to measure the calcium transients.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Histograms of regions of interest drawn within cell boundaries for several images from Fig. 4. A and C: histograms corresponding to coupled and uncoupled images at time = 0.0 ms in Fig. 4 and labeled (arrow) with mean pixel value. B and D: histograms corresponding to coupled and uncoupled images at time = 18.72 ms in Fig. 4. Histograms in A-C show no pixel values equal to maximum pixel value (255), whereas histogram in D shows only a few pixels at that value; thus the bulk of the pixels for even the largest fluorescence values were below saturation.

We demonstrated in Figs. 2 and 3 that the calcium transient briefly increased when the action potential of a real ventricular cell propagated to an adjacent model cell when the two cells were poorly coupled. In Fig. 6 we show what happened to the calcium transient when G_C was increased so that the two cells were better coupled. In Fig. 6A, for a different cell from that used for obtaining data in Figs. 2 and 3, we show the membrane potential for the real cell (solid lines) and the model cell (dashed lines) when coupling was set to either 0.0, 8.0, or 20.0 nS (Fig. 6A, traces a-c). Figure 6B shows the corresponding calcium transients plotted as the fluorescence ratio. Note that the time step for the calcium transients was 4.16 ms. For the control record (Fig. 6B, trace a), where G_C = 0.0 nS, the calcium transient is similar to the one shown in Fig. 3. The fluorescence began to rise within 4 ms of the action potential and reached a peak after 35 ms. The peak value of the fluorescence ratio was 2.56. When we coupled the real cell to a model cell with a G_C of 8.0 nS, there was again a delay (11.5 ms) between the time when the real cell was activated and the time when the model cell was activated. There was also a rapid repolarization (32.0 mV) of the membrane potential of the real cell before the model cell was activated. As shown in Fig. 3, this rapid repolarization enabled more calcium to be drawn into the cell, resulting in a more rapid increase in the calcium transient (Fig. 3B, trace b) under coupled conditions. Increasing the coupling conductance between the real cell and the model cell to 20.0 nS (Fig. 6A, trace c) resulted in a decrease in the delay (3.1 ms) before activation of the model cell. Note that the rapid repolarization of the membrane potential of the real cell was 25.3 mV, a value that is slightly less than the repolarization when G_C equals 8.0 nS. The corresponding calcium transient is plotted in Fig. 6B, trace c. The calcium transient for G_C = 20.0 nS rose more rapidly than the transient for G_C = 0.0 nS, but it did not reach as high a peak as for G_C = 8.0 nS. The peak value of the fluorescence ratio was 2.98 for G_C = 20.0 nS but reached a value of 3.13 for G_C = 8.0 nS. Because of the shorter delay between activation of the real cell and the model cell, there was less time for the large calcium entry during the early repolarization and thus the calcium transient was slightly smaller. Even though a G_C of 20.0 nS is higher than the value 8.0 nS, conduction remains discontinuous and thus calcium entry was facilitated.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Results from recordings for which the degree of discontinuous conduction was varied by altering GC between real cell and model cell. A: membrane potential recordings (traces a-c) from real cell (solid lines) with corresponding computations of membrane potential of model cell (dashed lines) for various values of GC. B: calcium transients from real cell recorded simultaneously with membrane potentials shown in A. Note that time step for calcium transient recordings is 4.16 ms. Data were recorded from a different cell than that used to obtain data in Figs. 2-5.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Combining the fluorescent dye technique to measure calcium transients with our coupling clamp technique to produce discontinuous action potential propagation confirmed our hypothesis that discontinuous action potential conduction can lead to an increased calcium concentration in the leader cell. We demonstrated that coupling conductance values slightly above the critical value required for propagation cause a brief increase in the calcium transient as well as an increase in the rate of rise of the calcium transient. Similar results were found by further increasing the coupling conductance, but the effect was not as large.

One limitation in the quantitative interpretation of these experiments is the uncertainty in both the correlation between F/F_o and the true value of [Ca²⁺]_i as well as uncertainty in the frequency response of the F/F_o signal as a representation of the time course of the change in [Ca²⁺]_i. The use of fluo 3 or any other fluorescence indicator that binds calcium ions depends on the rate constants for association and dissociation of calcium ions to the dye as well as the possible compartmentalization of the dye molecules. Smith et al. (16) discussed these problems in detail with regard to the use of fluorescence to record the time course and amplitude of calcium sparks. One problem is that the dissociation and association rate constants for fluo 3 both appear to be increased (but by different amounts) in a cytoplasmic environment compared with an aqueous solution (3, 4). Thus the dissociation rate constant for fluo 3 is raised from 0.4 to 1.13 µM. If we use the association and dissociation rate constants of Smith et al. (16) and assume a diastolic [Ca²⁺]_i of 200 nM, then the average F/F_o values that we measured for the uncoupled state (2.09) would predict a peak systolic [Ca²⁺]_i of 600 nM. Given the same reaction rate constants, if the calcium transient had actually occurred instantaneously, then the response of the fluo 3 signal would have occurred with a half-time to peak of 5 ms. This suggests that our measurement of the time course of [Ca²⁺]_i is not significantly affected by the slower reactions of the fluo 3 in the cytoplasmic environment, but the peak effect of the discontinuous conduction may have been reduced by the finite time response of the fluo 3 signal. Another factor that may be important in the time course of the calcium transient is the presence of intracellular buffers for calcium other than fluo 3. We used a small level of EGTA (0.1 mM) in the pipette to avoid loading the cell with calcium from the pipette solution. In three experiments performed without 0.1 mM EGTA in the pipette solution, we found that the cells had similar responses (18% increase in F/F_o with discontinuous conduction) but lower values of F/F_o (1.43 ± 0.05 for uncoupled vs. 1.69 ± 0.10 for coupled) and a shorter time before rundown.

In this study we showed that discontinuous action potential propagation increased the amplitude of the calcium transient. We also showed that this effect is clearly transient in these ventricular cells. After conduction, the voltage waveform of the leader cell and the follower cell become nearly identical and the waveforms for calcium concentration then become identical for the coupled and uncoupled conditions. In our previous work (8), we applied the voltage-clamp technique to ventricular cells with waveforms chosen from prior experiments in which we had measured the action potentials waveforms during discontinuous conduction. In these experiments, discontinuous conduction produced an increase in the peak value of I_Ca from 5.93 pA/pF to 8.88 pA/pF, an increase of 50%. The percent increase we measured in the present experiments in the peak amplitude of the calcium transient was 16%. The lack of quantitative agreement between these two percent changes may be due to the fact that most of the increase in [Ca²⁺]_i after the initiation of the action potential is thought to be the result of calcium-induced release of calcium from the sarcoplasmic reticulum rather than a simple integration of the I_Ca.