445

Original Contributions Direct Measurement of L-Type Ca2' Window Current in Heart Cells Yuji Hirano, Adriana Moscucci, and Craig T. January The activation and inactivation relations of several ion channel currents overlap, suggesting the existence of a steady-state or "window" current. We studied L-type Ca21 channel window current in single cardiac Purkinje cells using a voltage-clamp protocol by which channels were first inactivated nearly completely during a long-duration depolarizing step, and then the recovery of Ca21 current was observed during repolarizing steps into the L-type Ca21 window voltage range. With these conditions, a small-amplitude inward Ca21 current gradually developed after repolarization to voltages within the window but not after steps to voltages positive or negative to it. Window current was suppressed by Cd21 (50 uM), nifedipine (1 ,uM), and nicardipine (1 ,M), and it was augmented by isoproterenol (5 ,M) and Bay K 8644 (1 uM). At voltages at which window current developed, L-type Ca21 channels also recovered to a closed state from which they could be reopened by an additional depolarizing step. At voltages positive to the window range, channel recovery to a closed state(s) was absent, whereas at voltages negative to the window range, channel recovery to a closed state(s) increased, as expected from the "steady-state" inactivation relation. Our results provide direct measurement of L-type Ca21 window current and distinguish it from other processes, such as slow inactivation. Our findings support the postulate that within a window there occur channel transitions from inactivated to closed states, and these channels (re)open, and this process may occur repetitively. Some physiological and pathophysiological roles for L-type Ca21 window current are discussed. (Circulation Research 1992;70:445-455) KEY WORDS * Ca21 current * activation * inactivation * window current * Purkinje cells heart

O ne mechanism for Ca2' entry into cells is through voltage-gated Ca2' channels. These channels usually respond to an appropriate change in the membrane potential by opening and then gradually inactivating, resulting in a current in intact cells that rises to a peak value from which it then decays. For the L-type Ca2' channel current, measurements of the steady-state inactivation and activation relations that govern channel behavior have revealed a voltage range where they overlap. This region of overlap has been interpreted as indirect evidence supporting the presence of a Ca2' "window" current that exists near action potential plateau voltages.1-6Within this window, it is postulated that channel transitions may occur from inactivated to closed states (governed by the inactivation relation) and that channels may (re)open (governed by the activation relation) before inactivating again; this cycle may be repetitive. Previous attempts to measure window current have involved the use of prolonged depolarizing steps to voltages near the action potential plateau to search for From the Cardiac Electrophysiology Laboratories, Department of Medicine (Cardiology), University of Chicago (Ill.). Previously published in abstract form as a preliminary report (Biophys J 1989;55:287a). Supported by National Heart, Lung, and Blood Institute grants HL-20592 and HL-38927. Address for correspondence: Craig T. January, MD, PhD, Department of Medicine (Cardiology), Box 322, University of Chicago, 5841 South Maryland Avenue, Chicago, IL 60637. Received June 18, 1990; accepted October 15, 1991.

a persisting inward current after the decay of the initial peak current. In Purkinje fibers, Kass and Scheuer7 showed an inward Ca2' current that gradually inactivated to a nearly steady value by the end of 10-second depolarizing steps. Cohen and Lederer4 showed that L-type Ca2' current was present at the end of 200-msec depolarizing steps in isolated rat ventricular cells, and Hirano and coworkers5 showed persisting inward L-type Ca2' current at the end of 1-second depolarizing steps in canine Purkinje cells. Recordings of single L-type Ca21 channels, obtained using similar voltage protocols, also have been interpreted to support the presence of window current.8 9 However, these and other studies have shown that channel behavior is complex, including occasional long latencies before channels open, bursts of channel activity between closed and open states, and extended periods of channel activity before channels inactivate. Moreover, these types of channel activity are most pronounced near the threshold voltage for activating channels, which is also the voltage range postulated for window current. As a consequence, the interpretation of the kinetic mechanism(s) underlying the inward current persisting with prolonged depolarizing steps is uncertain. The purpose of the present experiments was to study L-type Ca21 channel window current. To avoid complications that might arise from late-activating, long-bursting, and delayed-channel inactivation, we used protocols that differ from the previous approaches. We first applied a long-duration, large-amplitude voltage step to rapidly activate and then inactivate nearly all L-type

446

Circulation Research Vol 70, No 3 March 1992

Ca2' channel current. We then studied the ability of L-type Ca2' channel current to recover from inactivation on repolarizing into its window voltage range. Thus, our results provide a direct measurement of a functional L-type Ca2' window current. Materials and Methods Preparation Single canine cardiac Purkinje cells were enzymatically isolated as described previously.10 Cells were maintained in Tyrode's solution and were used the same day they were isolated. Small aliquots of cells were transferred, as needed, to a Lucite chamber constructed on a glass coverslip and mounted on the stage of an inverted microscope (Diaphot, Nikon). The temperature in the chamber was thermostatically controlled at 30±1C during all experiments.

Solutions In most experiments, the chamber was initially superfused with a solution containing (mM) NaCl 142, KCl 5.4, CaCI2 1.8, MgCI2 0.5, NaH2PO4 0.33, glucose 5.5, and HEPES 5 (pH adjusted to 7.4 with NaOH). After a gigaohm seal was obtained, external solution was changed to a Na+- and K'-free solution containing (mM) tetraethylammonium chloride (TEA-Cl) 140, CaCI2 5, MgCl2 2, glucose 10, HEPES 10 (pH adjusted to 7.4 with TEA-OH), and 0.01-0.05 tetrodotoxin (Sigma Chemical Co., St. Louis, Mo.). Seals also were occasionally obtained directly in the Na+- and K'-free solution. All experiments were performed in the Na+- and K+free solutions. Patch pipettes contained (mM) CsCI 120, TEA-Cl 10, Cs4-BAPTA 5, MgATP 5, and HEPES 10 (pH adjusted to 7.3 with CsOH). In a few experiments, BAPTA was replaced with 5 or 10 mM EGTA without obvious difference. Patch electrode resistances were 0.5-2.5 Mfl. Exchange between the pipette solution and cytosol was complete within 5-10 minutes from obtaining a seal and rupture of the patch, as monitored by the disappearance of transient and rectifying outward currents. Under these experimental conditions, current through Ca21 channels could be studied with little contamination by other membrane currents.5 Drug and ionic interventions were made using prepared stock solutions that were added to the bath solution as required. Bay K 8644 (Miles Pharmaceuticals, West Haven, Conn.), nifedipine (Pfizer Inc., New York), and nicardipine (Syntex Laboratories, Inc., Palo Alto, Calif.) were dissolved in absolute ethanol to form 10 mM stock solutions. Isoproterenol (Sigma) was dissolved in 0.01N HCI to form a 1 mM stock solution. CdCl2 was dissolved in aqueous solution to form a 100 mM stock. In preliminary experiments, Ba21 was used as the charge carrier instead of Ca2+. However, Ba2` markedly slows the decay of Ca21 current,5,11-13 requiring the use of very long-duration voltage steps to inactivate it. Because our protocols were designed to facilitate inactivation of Ca2+ channels, we performed all experiments with Ca2' as the charge carrier.

Electrical Recordings and Voltage Protocols Recordings were made using the whole-cell conformation of the patch-clamp technique,'4 as outlined previously (see Reference 5 for details). Membrane current was recorded using a patch-clamp amplifier (model 8900, Dagan Corp., Minneapolis, Minn.; or Axoclamp-2A, Axon Instruments). Current was sampled at 1-10 kHz by a 12-bit A/D converter, controlled by a computer (IBM-PC compatible), and stored on hard disk for subsequent analysis. Pipette-cell membrane seal resistances, measured by a small-amplitude square-voltage step, usually were 5-50 GQ&. Tracings were filtered using a five-point running average, Gaussian weighted, digital technique or a computed analog technique with a cutoff at 0.2 or 1 kHz. We used the terminology of Nowycky et a115 (see also Reference 5) to describe the cardiac Ca2' currents as long lasting (L type) and transient (T type). Although cardiac Purkinje cells contain both T- and L-type Ca21 currents, T-type Ca21 current inactivates rapidly and completely16 and can be separated from L-type Ca21 current by its more negative voltage dependency of

activation and inactivation.56,16 In the experiments we Ca21 current was elicited from holding potentials (Vh) at which T-type Ca21 current was inactivated and at which the voltage-clamp step duration was adequate for complete inactivation of T-type current, thus leaving only L-type Ca2+ current. Voltageclamp protocols were applied at a frequency of 0.1 Hz or less. The specific protocols used are described in greater report, L-type

detail in "Results."

Data Analysis Most experimental data are shown as raw currents, without capacity or leak-current subtraction. In a few experiments, for reasons of clarity, the capacitive transients were removed off-line. For calculation of the activation relation, the leak-current component was obtained from hyperpolarizing voltage steps from Vh and fitted using linear regression. The extrapolated leak current was then subtracted from the current-voltage (I-V) plot before estimation of the reversal potential (see also Reference 5). In some experiments, window current tracings were curve fit as a sum of exponentials by a Fourier method.17 Results Activation and Inactivation Relations Experimentally obtained activation and inactivation relations for the L-type Ca21 channel current are shown in Figure 1. Panel A shows the protocol used to determine the inactivation relation. L-type Ca2' current was elicited by 150-msec depolarizing steps to +20 mV from various conditioning potentials (Vc). The resulting current tracings are shown superimposed; the availability of the current decreased as Vc became more positive. In most experiments, Vc was of 1-second duration. Increasing Vc up to 30 seconds did not abolish the overlap of the inactivation and activation curves, suggesting that our experimental conditions were close to a steady state. Panel B shows the peak I-V plot for the same

cell. L-type Ca2+ currents were elicited

by

10 mV

amplitude steps to different voltages, from a Vh Of-40

Hirano et al L-Type Ca' Window Current

447

B

A

20

80 FIGURE 1. L-type Ca' channel current activam tion and inactivation relations. Panel A: Inactivation protocol and currents. Depolarizing steps to +20 mV were applied for 150 msec from various conditioning voltages (V,). V, steps were of 1-second duration and in 5 mVincrements from -40 to +10 mV L-type Ca2+ current availability decreased as Vc became more positive, as shown by the inset current tracings. Panel B: Peak currentvoltage plot for L -type Ca2+ current. Steps were made to various voltages for 200 msec from a holding potential of -40 mV The peak current amplitudes (corrected for linear leak current) are plotted. The extrapolated reversal potential was +65 mV Superimposed current tracings, obtained by depolarizing steps, are shown in the inset. Panel C: Boltzmann fits through normalized peak currents are shown for steady-state inactivation (circles; half-maximal voltage, -22.8 mV; slope, 4.80) and activation (squares; half-maximal voltage, -0.5 mV; slope, -4.61). The window voltage range is predicted by the overlap of the curves between about -30 and 0 mI Experimental conditions: cell 11149, 5 mM EGTA.

VC

60ms

250 pA

C

CRl-

az

L.)

c:

0

0)

Voltage (mV)

mV, and the peak currents were plotted. Superimposed currents are shown in the inset. The activation relation was calculated using the peak currents according to Isenberg and Klockner18 (see also Reference 5):

raw

'Ca

gCax(Vm-Erev) where d is the normalized conductance at each test potential (Vm), 'Ca is the peak current at each Vm, gcj is the maximal conductance, and Erev is the apparent reversal potential obtained by extrapolating the peak I-V plot through the zero-current axis. The resulting normalized inactivation (circles) and activation (squares) relations, shown with Boltzmann fits to the data points, are plotted in Figure 1C. The two curves overlap with finite conductances between approximately -30 and 0 mV, defining the L-type Ca' window current voltage range. These results are in agreement with previous findings.5 This approach, although extensively used, does not provide direct measurement of window current, and it assumes Hodgkin-Huxley formalism.

Recovery of Window Current During Repolarizing Steps If a window exists between the activation and inactivation relations, it should be possible to demonstrate the recovery of inward current by repolarizing to voltages within the window range, after first inactivating L-type Ca> channels. To test for recovery of L-type Ca> window current, we used the three-step voltage-

clamp protocol shown in Figure 2A. An initial largeamplitude step of 0.5-2-second duration was applied to activate and inactivate L-type Ca2+ current. This was followed by a repolarizing step (V,, 1-1.5 -second duration) to voltages within the predicted window voltage range. A second depolarizing step (200-msec duration) was then applied to activate current from channels that had recovered to a closed state during Vt. Figure 2B shows raw current tracings obtained with this protocol (same cell as in Figure 1). From a Vh of -80 mV, a voltage step to +20 mV (1.5-second duration) elicited a large-amplitude inward Ca>2 current (peak currents off scale, superimposed tracings). The peak current rapidly decayed to a nearly steady value by the end of this initial inactivating step. Repolarizing steps were then applied for 1.5 seconds to voltages within the predicted window voltage range. For a V, to 0 mV (positive to the window) or to -40 mV (negative to the window), the resulting current, after the decay of the capacitance transient, was nearly constant. In contrast, for a V, to -10, -20, or -30 mV (within the window) a small, time-dependent current developed in an inward direction, consistent with recovery of L-type Ca>2 window current. This pattern of recovery of inward Ca>2 current at voltages within the window, but not at voltages positive or negative to it, was observed in 30 Purkinje cells. The second depolarizing step to +20 mV activated channels that recovered to a closed state during V,. When V, was to 0 mV, no Ca2+ current transient was elicited by the second depolarizing step, suggesting that

448

Circulation Research Vol 70, No 3 March 1992

A

VT

20

mVF -80

B

AO

pAT15

C

0-A

mV 1

-40

\|

'

10 0

-25 FIGURE 2. Recovery of L-type Ca21 window current. Panel A: The voltage protocol. From a holdingpotential of -80 mV a step was made to +20 mVfor 1.5 seconds to activate and inactivate nearly completely Ca2' current. Repolarizing steps (VT; 1.5 seconds) were then made to voltages within the overlapping area of the activation and inactivation curves. Panel B: Current tracings observed in response to the voltage protocol shown in panel A. Peak current during the first depolarizing step is off scale. Currents resulting from VT to -40 (most negative current tracing), -30, -20, -10, and 0 mV (most positive current tracing) are shown. After VT, a second depolarizing step to +20 mVfor 200 msec was applied to activate L-type Ca2+ current from channels in a closed state. From a VTof 0 mV, no Ca2+ current was obtained (most positive tracing). From the VT of -10 mV, a small-amplitude inward Ca21 current was present, suggesting the recovery of a small number of channels to a closed state. From more negative VTs, the current amplitude increased, consistent with the steady-state inactivation relation. Panel C: The currentvoltage plot for the L -type Ca2+ window current. A peak inward Ca2+ current of -19 pA was obtained at a VT of -15 mV See text for additional discussion. Experimental conditions: cell 11149, 5 mM EGTA.

Ca2' channels were inactivated. In contrast, after the Vt to -10 mV, the second depolarizing step elicited a small-amplitude Ca2+ current transient, suggesting recovery of a small number of channels to a closed state during V, from which they could be voltage activated. The amplitude of the Ca2+ current was increased further after a V, to -20 mV. After more negative V,s, the peak current amplitude progressively increased, as expected for the voltage-dependent recovery of channels from inactivated to closed states. Figure 2C shows the I-V plot for the L-type Ca' window current of the same cell. Window current was measured during each Vt, as the difference between the current present just after the decay of the capacitance transient and the maximum inward current that subsequently developed. The peak window current was obtained by a V, to -15 mV. The I-V plot also shows a very small-amplitude outward current recorded at more positive voltages. This small-amplitude outward current was not observed in all cells (see Figures 3 and 4); however, when present, its effect may have been to interfere with the measurement of window current amplitude. We did not study the charge carrier for this small-amplitude current, present mainly at positive voltages, although one possibility is that it may have been carried by Cl.19,20 Effects of Bay K 8644 and Isoproterenol on Window Current Bay K 8644 is a 1,4-dihydropyridine known to cause voltage-dependent increase of L-type Ca2+ current amplitude, as well as negative shift of its activation and inactivation relations.45,21-24 At single-channel levels, Bay K 8644 promotes opening for long periods of time.25,26 Figure 3 shows the effect of Bay K 8644 (1 gM) on window current; the voltage protocol used is shown in panel A. Superimposed currents, elicited during voltage clamp back to different V,s, are shown in panel B for control conditions and after the addition of Bay K 8644. During control conditions, an inward developing Ca2+ current was present at voltages within the window voltage range, as shown previously. In the presence of Bay K 8644, the amplitude of the Ca2+ window current was increased. This finding was confirmed in 10 different Purkinje cells. Figure 3C shows the I-V plot for the window current before and after addition of Bay K 8644. During control conditions, the maximum inward current of -10 pA was recorded at a V, of -10 mV, whereas with Bay K 8644 the I-V plot was shifted negatively, and the peak inward current of -24 pA was recorded at a V, of -20 mV. The interpretation of the effect of Bay K 8644 on the window current is potentially complicated by its frequency- and voltage-dependent properties. Bay K 8644 has been shown to have potent agonist properties when depolarizing voltage steps are applied from negative holding potentials, such as those we used. However, in our protocol V, was preceded by an initial, large-amplitude depolarizing step; this initial step is likely to have decreased the agonist effect of Bay K 8644 (i.e., "relative block"23'24), but it did not reverse it. Isoproterenol has been shown to increase peak L-type Ca21 current amplitude in this preparation5 by increasing both the probability and duration of channel openings.27-30 We examined its effect on the window

Hirano et al L-Type Ca' Window Current

A

window current during control conditions and after the addition of isoproterenol. The peak current amplitude was increased, as expected for an L-type Ca> channel current. Similar findings were obtained in two additional cells. Figure 4A also illustrates an observation recorded in a few cells exposed to interventions that increased L-type Ca> current. As shown for the V, to -10 mV, the window current reached a peak inward value within about 250 msec, and after this its amplitude declined. This biphasic pattern was observed only in cells that were stimulated with Bay K 8644 or isoproterenol. It was evident at more positive window voltages, and it could be reproduced with repetitive steps to the same voltage. A similar decline in window current from its peak value was never observed during control conditions.

VT

B

0

mV

-l-

BayKBf44

Control

-10 mV BayK8R44

Control

-20 mV

I

-

|

-30 mV

Control

W

ms20 pA 250 ms

C

10

-50

-J

mV 0

S\

S 0

*

O

Control

BayK864

pA L-30 Ca' winidow current.

FIGURE 3. Effect of Bay K 8644 on Panel A: The voltage protocol is similar to that shown previously. Panel B: Pairs ofsuperimposed current tracings, at the indicated repolarizing step (VT) show control currents and currents after the addition of Bay K 8644 (1 gM). Addition of Bay K 8644 increased the amplitude of the Ca> current. Panel C: The current -voltage plot for the window current. Bay K 8644 increased the current amplitude and shifted its peak negatively. Experimental conditions: cell 9158a, 5 mM BAPTA. current, and the results from one experiment

are shown in Figure 4. The voltage-clamp protocol previously described is shown above the current tracings. For control conditions, after the initial current inactivating step, the time-dependent recovery of window current was maximal at a V, to -20 mV (Figure 4A). Five minutes after the addition of isoproterenol (5 iiM), the protocol was repeated. The most prominent effect of isoproterenol was to increase the amplitude of the inward current that recovered within the window volt-

age range.

449

Figure 4B shows the I-V plot for the peak

Block of Window Current by Cd', Nifedipine, and Nicardipine Previous studies have shown that low concentrations of Cd>2 suppress L-type Ca2+ current.515 Figure 5 shows results from one Purkinje cell in which the effect of Cd` was studied. The voltage-clamp protocol previously described is shown above the current tracings in panel A. For control conditions, the initial step to +20 mV elicited a large-amplitude inward Ca2+ current (peak current off scale) that had decayed nearly completely by the end of the step. Repolarizing steps were then applied to different V1s within the predicted window voltage range and elicited the time-dependent inward window current, as shown previously. The second depolarizing step elicited an L-type Ca2+ current whose amplitude was consistent with the voltage-dependent recovery of these channels to a closed state from which they could be voltage activated. Five minutes after application of Cd2 (50 ,uM), the Ca2+ current amplitude elicited during the initial step to +20 mV from a Vh of -80 mV was markedly reduced. Only a smallamplitude, rapidly inactivating current component remained, as would be expected for a T-type Ca' current. More importantly, at all V,s the time-dependent increase in window current was abolished. This is also shown by the I-V plot in Figure SB. Finally, Cd` markedly suppressed the recovery of Ca2+ current that was elicited by the second depolarizing step to +20 mV. Similar findings were obtained in three additional Purkinje cells exposed to Cd>. In other experiments, we also examined the effects of the 1,4-dihydropyridines nifedipine (two cells) and nicardipine (one cell), which are known to suppress L-type Ca2+ current at the voltages we used.531 Both compounds, at 1 ,uM concentration, completely blocked the recovery of window current (data not shown). Time Course of Recovery of Window Current The recovery of Ca2+ channels from inactivated to closed states at voltages within the window was monitored using a conventional two-pulse protocol, and these results were compared with the recovery of window current. The voltage-clamp protocol we used is shown in Figure 6A, and it resembles the previously described window current protocol with the exception that the duration of Vt was varied before applying the second depolarizing step. The initial depolarizing step

d

450

Circulation Research Vol 70, No 3 March 1992

A

20 mv -W

1

VY F

Control h~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

ISP

-20 mV

-30 mV

-50 mV

SPjAH 1

-10 mV

4 1X

1

J20 pA

1 sec

B

FIGURE 4. 1~~~~~~~~~~

10

-50

0

Effect of isoproterenol (ISP) on Ca2' window current. Panel A: The voltage-clamp protocol is shown above current recordings obtained during control conditions and after the addition of ISP (5 ,ugM). After the initial voltage step to +20 mV (peak current off scale), repolarization steps (VTs) were made to the indicated voltages. ISP increased the amplitude of window current, and at more positive VTs the current reached a peak value from which it decayed. Panel B: The currentvoltage plot for the window current. ISP markedly increased peak Ca2+ window current amplitude. Experimental conditions: cell 8098, 5 mM BAPTA.

0

*Control ,sP

O

0

pA

from -80 to +20 mV elicited a large-amplitude Ca2' current that decayed to a nearly steady value, as shown in the superimposed current tracings in Figure 6B. The currents obtained by varying the duration of V, (-20 mV, within the window voltage range) are also shown superimposed. The current obtained during the longest V, (1.6 seconds) has been amplified and displayed in the lower tracing of Figure 6B and shows the time-dependent development of window current seen previously. When the second depolarizing step was applied, an inward L-type Ca2' current was elicited as channels were activated from a closed state. We then varied the duration of V, to study the time course of recovery of channels to a closed state. As shown in the current tracings in Figure 6B, less current was elicited by shorter duration V,s, and virtually no current was elicited at 50-msec V,. This finding is consistent with the nearly complete inactivation of L-type Ca2' channels by the end of the initial depolarizing step. As the duration of V, was increased, the peak L-type Ca2' current amplitude elicited by the second depolarizing step gradually increased to reach a steady value (see also -20 mV plot in Figure 6C), emphasizing that the process of recovery from inactivation of L-type Ca2' channels to a closed state is relatively slow at voltages near the window (e.g., action potential plateau, see also Reference 18). These results also indicate that, after channel inactivation, during a repolarizing step to voltages at which Ca2' channels were able to (re)open within their window, channels also recovered to a closed state(s) from which they then could be voltage (re)activated.

Figure 6C shows plots of the time course of recovery from inactivation of peak L-type Ca2' current measured with the second step of the two-pulse protocol shown in panel A. Recovery of channels from an inactivated to a closed state(s), and their activation, was observed at voltages within the window range (V,s of -10 and -20 mV). For a V, positive to the window range (i.e., 0 mV), channels remained nearly completely inactivated. At Vts negative to the window range (i.e., -50 mV) current recovery was enhanced, as expected from the voltagedependent behavior of the L-type Ca21 channels. Exponential fits to the recovery process for window current produced inconsistent results, with the fitting routine returning both single and multiexponential values. This may reflect complex underlying kinetic processes for the window current, as well as the noise levels associated with the measurement of this small-amplitude current. Clearly, the process was slow, usually requiring several hundred milliseconds to reach a maximum value. Although the recovery process for window current is likely to be sensitive to the recovery voltage, the limited voltage range over which it can be elicited and the inconsistency of the fits did not permit reliable quantification of the voltage dependence of the recovery time course.

Depolarizing and Repolarizing Voltage Steps Into the Window Range The voltage-clamp protocol we used first inactivated nearly completely L-type Ca2' channel current, and we then studied its ability to recover as channels reopened

Hirano et al L-Type Ca2' Window Current

451

A mV

Control

L1

1

-20 mV

FIGURE 5. Effect of Cd2+ on Ca2' window cur30 pA rent. Panel A: The voltage-clamp protocol is shown 500 ms above current recordings obtained during control conditions and after the addition of Cd2+ (50 ,uM). On-scale peak currents are shown, except for those elicited by the initial depolarizing step during control conditions. A window current developed during control conditions after a repolarizing step (VT) to -10 and 0 mV (dashed line indicates current level at the onset of each VT). The addition of Cd2+ suppressed the window current as well as the L -type Ca2+ current transients activated during both depolarizing steps. Panel B: The current-voltage plot for

-10 mV

Cd2+

B

the window current shows that Cd2+ completely suppressed the Ca2+ inward current. Experimental conditions: cell 12208, 5 mM BAPTA.

* Control O Cda

-15

after repolarization to voltages within the window. An alternative approach to measure window current has been to depolarize into the window voltage range for a duration sufficient to outlast the complex kinetics of channels activating from a closed state (e.g., channel latency before opening, bursting between closed and open states, and long periods of activity before inactivating). For a voltage-gated ion channel it should be possible to record window current independent of whether the window is approached using repolarizing or depolarizing steps. Figure 7 shows results of an experiment that tested this hypothesis. From a Vh of -80 mV, a depolarizing step was made to +20 mV for 1 second to activate Ca2' current that then decayed to a nearly steady value (tracing a, peak L- and T-type currents off scale). A Vt was then applied to -20 mV (within the L-type Ca2+ current window) for 5 seconds, and an inward current gradually developed and reached a steady value. The protocol was repeated, except that the initial depolarizing step was to -40 mV for 1 second, to activate and inactivate T-type Ca2+ current (tracing b). With the second step to -20 mV (within the L-type Ca2+ window) for 5 seconds, L-type Ca2+ current was activated. From its peak value, the current decayed to a steady level over a few seconds. As shown in Figure 7, the steady-state current amplitude reached by either method was the same.

Discussion The most significant finding of these experiments is the identification, by direct measurement, of a smallamplitude inward current with properties predicted for an L-type Ca2+ window current. This current could be elicited over a finite range of voltages, which corresponded to the region of overlap of the activation and inactivation relations. The development of window current also was time dependent, and it usually reached a maximum value within several hundred milliseconds with our experimental conditions. Our results indicate that the window current was carried by Ca2' through L-type Ca21 channels: 1) under the experimental conditions we used, the major permeable ion for the inward current was Ca2+; 2) the development of window current occurred at voltages expected for L-type Ca2' current and positive to voltages for T-type Ca2+ current; 3) the window current amplitude was enhanced by isoproterenol and Bay K 8644; and 4) the window current was suppressed by Cd2+, nifedipine, and nicardipine. This small-amplitude current is not likely to represent the deactivation of an outward current through K' channels (e.g., IK) activated by the initial depolarizing step for several reasons. 1) The solutions (bath and pipette) were K+ free. To further suppress K+ current, TEA+ was present in large concentrations in the bath,

Circulation Research Vol 70, No 3 March 1992

452

A 20

vt

-80

AT

B

AT

so"0 2oo 400

oo

ioo ms

1I1O I

p1 A C

0 cO

T~

t_ T

AT (ms) 800

1600 O mY -10 mV

v

t-20 mv

FIGURE 6. Time course of recovery of L-type Ca> window current. Panel A: The voltage-clamp protocol is similar to that described previously, except that the duration (AT) of the repolarization step (VT) was varied. Panel B: Inward Ca> currents obtained by a VT to -20 mV are shown (solid line indicates current level at the onset of the VT). The development of Ca2 window current during the longest duration VT (1.6 seconds) has been amplified (lower current tracing) to show its time dependency. The time course of recovery of L -type Ca>2 channels to a closed state was studied by varying AT of the VTs before applying the second depolarizing step to +20 mV Recovery of Ca2+ current was time dependent, and at very short AT there was little recovery. Panel C: Time-dependent recovery ofpeak L-type Ca2+ current at four different VTs. See text for discussion. Experimental conditions: cell 8188, 5 mM BAPTA.

O -250-

-W0 mV -500-

and Cs+ was present in both the bath and the pipette. 2) The window current we measured did not develop after repolarizing steps to voltages of -40 mV (negative to the window) or to voltages of 0 mV (positive to the

2020r -40 -80

a b

a,b

window). The decay of an IK tail current would not be expected to show such a selective voltage dependence. Also, the absolute value of window current (shown in Figure 4) became more inward in the presence of -20 mV

ap (bbJ v >~ ~ ~ ~pA~ 5

FIGURE 7. Ca2+ currents obtained in response to 5-second repolarizing and depolarizing steps into the window voltage range. Both protocols result in the same steady-state current amplitude at the end of the pulse. Voltage protocols are shown above the current tracings. Letters identify corresponding voltage steps and resulting currents. Experimental conditions: cell 9138, 5 mM BAPTA.

Hirano et al L-Type Ca' Window Current

isoproterenol, opposite to that expected if this were enhanced IK tail current. 3) In some Purkinje cells in which window current was present, we applied a protocol intended to elicit IK tail currents (steps up to + 100 mV for 800 msec before repolarizing to -40 mV; see Reference 32), without ever observing such current tails (data not shown). 4) In several experiments, the window current was not blocked by the addition of the lK blocker 4-aminopyridine (2 mM) to the external solution. Because our solutions were K' and Na+ free, the window current also is unlikely to arise from electrogenic transport (Na+-K+ pump and Na+-Ca2+ exchange). The external solution also contained tetrodotoxin, making current through Na+ channels an unlikely possibility. Finally, the window current is not likely to have been carried through Cl- channels, since Cl- current shows little rectification under conditions of a symmetrical Clgradient, it is time independent, and it is negligible in the absence of cAMP-dependent kinase activation and the presence of high concentrations of intracellular Ca2+ buffer (e.g., millimolar EGTA).'9,20 Our voltage protocol provided a method for distinguishing a membrane current arising from channels (re)opening within a window from a membrane current resulting from the complex repetitive gating of channels from a closed state near their threshold voltage. The initial voltage step of our protocol was designed to activate Ca2+ channels from a closed state(s) and then inactivate them nearly completely over the duration of this step. On repolarization into the window voltage range, channel recovery began from an inactivated state and not from a closed state, as occurs when window current is studied using a depolarizing step. The ability of single L-type Ca2+ channels that have been inactivated to (re)open on repolarization to voltages within the window range has also been shown recently.13

Channel (Re)Openings for Window Current Are From Closed States Do Ca2+ channels first recover to a closed state from which they then (re)open, or can channels (re)open directly from an inactivated state? Our data suggest that the former possibility occurs. 1) At window voltages at which channels (re)opened, additional channels were shown to recover to a closed state from which they could be activated by the second depolarizing step of our protocol. 2) At voltages positive to the window range, the recovery of channels to a closed state also did not occur, as shown by the absence of current elicited by the second depolarizing step. 3) The time course of the development of window current required several hundred milliseconds for the current to reach a maximum. Previous reports suggest that the recovery of channels to a closed state has a similar time course (see Figure 6).18 Our data do not exclude the possibility of channels (re)opening directly from an inactivated state or from a long-lived closed state. However, under our experimental conditions, the recovery of L-type Ca>2 window current was always associated with recovery of additional channels to a closed state. Thus, our findings support the presence of a voltage range where L-type Ca>2 window current can be directly measured, as channels recycle from inactivated to closed states, and then reopen transiently (see also Reference 13).

453

The mechanism of the apparent decay of window current from its peak (see Figure 4) remains uncertain. One possibility is the development of an overlapping, large anionic (i.e., Cl-) current. Intriguing alternative explanations are that increased cytosolic free Ca> may modulate regulatory processes (e.g., Ca'+-dependent inactivation and phosphorylation) or that Ca2' accumulates in a restricted intracellular space, leading to alteration of the transmembrane ion gradient.3334 In such circumstances, it would be possible for a current through voltage-gated ion channels operating within their window to display time-dependent behavior, thus modulating the window current.

Comparison With Other Window Currents The concept of a steady-state current arising within the window of overlap between an ion channel's inactivation and activation relations is not new. Measurements of voltage-dependent activation and inactivation relations for Na+ current in squid axon have shown such a region.35 Subsequent computer models of the cardiac action potential have incorporated window current,36 and studies of selective ion channel block during long depolarizing steps gave results to support a steady-state Na+ window current.37 However, more recent findings, showing a time-dependent decay of the Na+ window current,38,39 the identification of late activating and inactivating Na+ channel current,40A41 and multiple channel openings,4243 have resulted in uncertainty regarding the existence of a Na+ window current. A window current may also exist for T-type Ca>2 current, as suggested by the overlap of its inactivation and activation relations.5,644

Physiological and Pathophysiological Roles of L-Type Ca>2 Window Current Ca>2 current has long been known to contribute to the regulation of the action potential plateau and its duration. Our findings suggest at least two mechanisms for the participation of L-type Ca2+ channels. First, action potential depolarization initiates channel openings from a closed state. These channels may open with latency, burst between open and closed states, and undergo long periods of activity before inactivating. At physiological temperatures and conditions, the action potential overshoot reaches positive voltages, which will favor more rapid channel activation and inactivation. A second mechanism for Ca2+ channel participation in the action potential is through channels that recover from inactivated to closed states and then (re)open within their window voltage range. Because this recovery process is kinetically slow, lengthening of the time the action potential plateau remains within the window voltage range would increase the contribution of Ca> channels (re)opening within the window. Therefore, during an action potential the L-type Ca2+ current arising from channels undergoing their initial activation from a closed state will decline, whereas during the plateau of the action potential channels will be able to (re)open from an inactivated state(s). The onset of phase 3 repolarization may be facilitated, in part, by a decline of L-type Ca>2 window current. Ca2+ entry through L-type channels (re)opening within their window may contribute to loading of cells

454

Circulation Research Vol 70, No 3 March 1992

with Ca2`. This provides a mechanism for window current participation in the regulation of subsequent contractions.445 Also, in other tissues that undergo depolarization, such as smooth muscle and secretory cells, the L-type Ca2' window current could contribute to the regulation of Ca2' homeostasis and other Ca2'dependent signaling events.46-48 Furthermore, because of the steepness of the activation and inactivation relations, small changes in membrane potential could exert steeply graded changes in transmembrane Ca' entry and the cellular processes coupled to this. The L-type Ca' window current has also been proposed to be an arrhythmogenic mechanism. January and Riddle24 postulated that early afterdepolarizations result from the time- and voltage-dependent (re)opening of L-type Ca' channels within their window voltage range during the action potential plateau and that this recovery of inward current shifts the balance of membrane currents toward depolarization. When depolarization was initiated, additional L-type Ca' channels could then be recruited to open from a closed state(s), thereby augmenting the depolarizing Ca> current. In this model, the presence of L-type Ca> window current was an essential component to the induction of early afterdepolarizations.

Acknowledgments The authors gratefully acknowledge the assistance of Dr. Christopher A. Sanders in the performance of some experiments and the assistance of Ms. Gloria Johnson and Mr. Peter Cunningham in the preparation of this manuscript.

References 1. Reuter H, Scholz H: A study of the ion selectivity and the kinetic properties of the calcium dependent slow inward current in mammalian cardiac muscle. J Physiol (Lond) 1977;264:17-47 2. Brown HF, Kimura J, Noble D, Noble S, Taupignon AI: The slow inward current, is,, in the rabbit sino-atrial node investigated by voltage clamp and computer simulation. Proc Royal Soc Lond [Biol] 1984;222:305-328 3. Josephson IR, Sanchez-Chapula J, Brown AM: A comparison of calcium currents in rat and guinea pig single ventricular cells. Circ Res 1984;54:144-156 4. Cohen NM, Lederer WJ: Calcium current in isolated neonatal rat ventricular myocytes. J Physiol (Lond) 1987;391:169-191 5. Hirano Y, Fozzard HA, January CT: Characteristics of L- and T-type Ca2' currents in canine cardiac Purkinje cells. Am J Physiol 1989;256:H1478-H1492 6. Tseng G-N, Boyden PB: Multiple types of Ca2+ currents in single canine Purkinje cells. Circ Res 1989;65:1735-1750 7. Kass RS, Scheuer T: Slow inactivation of calcium channels in the cardiac Purkinje fiber. J Mol Cell Cardiol 1982;14:615-618 8. Cavalie A, Ochi R, Pelzer D, Trautwein W: Elementary currents through Ca2' channels in guinea pig myocytes. Pflugers Arch 1983;398:284-297 9. McDonald TF, Cavalie A, Trautwein W, Pelzer D: Voltagedependent properties of macroscopic and elementary calcium channel currents in guinea pig ventricular myocytes. Pflugers Arch 1986;406:437-448 10. Sheets MF, January CT, Fozzard HA: Isolation and characterization of single canine Purkinje cells. Circ Res 1983;53:544-548 11. Kass RS, Sanguinetti MC: Inactivation of calcium channel current in the calf cardiac Purkinje fiber: Evidence for voltage- and calcium-mediated mechanisms. J Gen Physiol 1984;84:705-726 12. Lee KS, Marban E, Tsien RW: Inactivation of calcium channels in mammalian heart cells: Joint dependence on membrane potential and intracellular calcium. J Physiol (Lond) 1985;364:395-411 13. Shorofsky SR, January CT: L- and T-type Ca2+ channels in canine cardiac Purkinje cells: Single-channel demonstration of L-type Ca2' window current. Circ Res 1992;70:456-464 14. Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ: Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflugers Arch 1981;391:85-100

15. Nowycky MC, Fox AP, Tsien RW: Three types of neuronal calcium channel with different calcium agonist sensitivity. Nature 1985;316:440-442 16. Hirano Y, Fozzard HA, January CT: Inactivation properties of T-type calcium current in canine cardiac Purkinje cells. Biophys J 1989;56:1007-1016 17. Provencher SW: A Fourier method for the analysis of exponential decay curves. Biophys J 1976;16:27-41 18. Isenberg G, Klockner U: Calcium currents of isolated bovine ventricular myocytes are fast and of large amplitude. Pflugers Arch 1982;395:30-41 19. Bahinski A, Nairn AC, Greengard P, Gadsby DC: Chloride conductance regulated by cyclic AMP-dependent protein kinase in cardiac myocytes. Nature 1989;340:718-721 20. Zygmunt AC, Gibbons WR: Calcium-activated chloride current in rabbit ventricular myocytes. Circ Res 1991;68:424-437 21. Sanguinetti MC, Kass RS: Regulation of cardiac calcium channel current and contractile activity by the dihydropyridine Bay K 8644 is voltage dependent. J Mol Cell Cardiol 1984;16:667-670 22. Thomas G, Chung M, Cohen CJ: A dihydropyridine (Bay K 8644) that enhances calcium currents in guinea pig and calf myocardial cells: A new type of positive inotropic agent. Circ Res 1985;56:87-96 23. Sanguinetti MC, Krafte DS, Kass RS: Voltage-dependent modulation of Ca channel current in heart cells by Bay K 8644. J Gen Physiol 1986;38:369-392 24. January CT, Riddle JM: Early afterdepolarizations: Mechanism of induction and block: A role for L-type Ca21 current. Circ Res 1989;64:977-990 25. Hess P, Tsien RW: Mechanism of ion permeation through calcium channels. Nature 1984;309:453-456 26. Kokubun S, Reuter H: Dihydropyridine derivatives prolong the open state of Ca channels in cultured cardiac cells. Proc NatlAcad Sci USA 1984;81:4824-4827 27. Yue DT, Herzig S, Marban E: ,B-Adrenergic stimulation of calcium channels occurs by potentiation of high-activity gating modes. Proc Natl Acad Sci USA 1990;87:753-757 28. Reuter H: Calcium channel modulation by neurotransmitters, enzymes and drugs. Nature 1983;301:569-574 29. Brum G, Osterrieder W, Trautwein W: ,B-Adrenergic increase in the calcium conductance of cardiac myocytes studied with the patch clamp. Pflugers Arch 1984;401:111-118 30. Tsien RW, Bean BP, Hess P, Lansman JB, Nilius B, Nowycky MC: Mechanisms of calcium channel modulation by 13-adrenergic agents and dihydropyridine calcium agonists. J Mol Cell Cardiol 1986;18:691-710 31. January CT, Hirano Y, Riddle JM, Fozzard HA: Effect of nicardipine on electrophysiological properties of cardiac Purkinje cells (abstract). Circulation 1989;80(suppl II):II- 607 32. Sanguinetti MC, Jurkiewicz NK: Two components of cardiac delayed rectifier K+ current: Differential sensitivity to block by class III antiarrhythmic agents. J Gen Physiol 1990;96:195-215 33. Yue DT, Backx PH, Imredy JP: Calcium-sensitive inactivation in the gating of single calcium channels. Science 1990;250:1735-1738 34. Lederer WJ, Niggli E, Hadley RW: Sodium-calcium exchange in excitable cells: Fuzzy space. Science 1990;248:283 35. Hodgkin AL, Huxley AF: A quantitative description of membrane current and its application to conduction and excitation in nerve. J Physiol (Lond) 1952;117:500-544 36. McAllister RE, Nobel D, Tsien RW: Reconstruction of the electrical activity of cardiac Purkinje fibres. J Physiol (Lond) 1975; 251:1-59 37. Attwell D, Cohen I, Eisner D, Ohba M, Ojeda C: The steady state TTX-sensitive ("window") sodium current in cardiac Purkinje fibres. Pflugers Arch 1979;379:137-142 38. Gintant GA, Datyner NB, Cohen IS: Slow inactivation of a tetrodotoxin-sensitive current in canine cardiac Purkinje fibers. Biophys J 1984;45:509-512 39. Carmeliet E: Slow inactivation of the sodium current in rabbit cardiac Purkinje fibres. Pflugers Arch 1987;408:18-26 40. Patlak JB, Ortiz M: Slow currents through single sodium channels of the adult rat heart. J Gen Physiol 1985;86:89-104 41. Kiyosue T, Arita M: Late sodium current and its contribution to action potential configuration in guinea pig ventricular myocytes. Circ Res 1989;64:389-397 42. Kunze DL, Lacerda AE, Wilson DL, Brown AM: Cardiac Na currents and the inactivating, reopening, and waiting properties of single cardiac Na channels. J Gen Physiol 1985;86:691-719 43. Scanley BE, Hanck DA, Chay T, Fozzard HA: Kinetic analysis of single sodium channels from canine cardiac Purkinje cells. I Gen Physiol 1990;95:411-437

Hirano et al L-Type Ca2' Window Current 44. Hagiwara N, Irisawa H, Kameyama M: Contribution of two types of calcium currents to the pace-maker potentials of rabbit sinoatrial node cells. J Physiol (Lond) 1988;395:233-253 45. Mitra R, Morad M: Two calcium channels in guinea pig ventricular myocytes. Proc Natl Acad Sci U S A 1986;83:5340-5344 46. Boyd AB Ill: Sulfonylurea receptors, ion channels, and fruit flies. Diabetes 1988;37:847- 850

455

47. Yatani A, Seidel CL, Allen J, Brown AM: Whole-cell and single-channel calcium current of isolated smooth muscle cells from saphenous vein. Circ Res 1987;60:523-533 48. Aaronson PI, Bolton TB, Lang RJ, MacKenzie I: Calcium currents in single isolated smooth muscle cells from rabbit ear artery in normal calcium and high-barium solutions. J Physiol (Lond) 1988; 405:57-75