Ionic currents during sustained pacemaker activity in rabbit sino-atrial myocytes



  • 1The contribution of various ionic currents to diastolic depolarization (DD) in rabbit sino-atrial myocytes was evaluated by the action potential clamp technique. Individual currents were identified, during sustained pacemaking activity reproduced under voltage clamp conditions, according to their sensitivity to supecific channel blockers.
  • 2The current sensitive to dihydropyridines (DHPs), blockers of L-type Ca2+ current (ICa,L), was small and outward during most of DD. Diastolic DHP-sensitive current was affected by changes in the driving force for K+, but it was insensitive to E-4031, which blocks the current termed IK,r; it was abolished by cell dialysis with a Ca2+ chelator.
  • 3The current sensitive to 2 mm Cs+ (ICs), a blocker of hyperpolarization-activated current (If), was inward during the whole DD and it was substantially larger than the net inward current flowing during this phase. However, diastolic IK,r, identified in the same cells as the current sensitive to the blocker E-4031, exceeded ICs 2-fold.
  • 4These findings suggest that: (a) Ca2+ influx during the pacemaker cycle increases a K+ conductance, thus inverting the direction of the net current generated by L-type Ca2+ channel activity during DD; (b) the magnitude of If would be adequate to account fully for DD; however, the coexistence of a larger IK,r suggests that other channels besides If contribute inward current during this phase.

The contribution of individual ionic conductances to sino atrial pacemaking is a matter of controversy (DiFrancesco, 1993; Irisawa, Brown & Giles, 1993). Various ionic currents, identified by voltage clamp experiments, have kinetics and voltage dependency compatible with their role in contributing to diastolic depolarization in sino-atrial myocytes. However, from the data available, the actual size and time course of each current during the pacemaking cycle can only be inferred with a remarkable degree of uncertainty. Complex summation of time-dependent currents during the cycle and changes in the intracellular ionic concentrations occurring during sustained repetitive activity may contribute to such uncertainty. This problem has been previously addressed by studying the effect on membrane potential of pharmacological blockade of individual currents (Noma, Kotake & Irisawa, 1980; Noma, Morad & Irisawa, 1983; Hagiwara, Irisawa & Kameyama, 1988; Denyer & Brown, 1990a; Boyett, Kodama, Honjo, Arai & Suzuki, 1992). Although providing useful information, this approach suffers from a major shortcoming: the changes in membrane potential induced by blockade of a single current may, in turn, affect other currents. The overlap of primary and secondary effects of blockade may thus confuse the interpretation of results.

This study exploits the action potential clamp technique (Doerr, Denger & Trautwein, 1989) (also referred to as ‘dynamic voltage clamp’) to assess the time course of individual current components during sustained sino-atrial pacemaking activity. The drug-induced changes in membrane current are measured while clamping membrane potential with action potential waveforms previously recorded from the same cell. Although it still relies on current identification by pharmacological means, such an approach overcomes the shortcoming peculiarities of membrane potential studies (see above) and may provide direct information on the time course of individual currents during the pacemaking cycle. Channel-blocking agents have been selected to obtain information on the time course of the DHP-sensitive Ca2+ currents, the delayed rectifier current (IK), and the hyperpolarization-activated current (If).


Cell isolation

Myocytes from the rabbit sino-atrial node region were isolated with the procedure described previously by DiFrancesco et al. (DiFrancesco, Ferroni, Mazzanti & Tromba, 1986). Briefly, white female rabbits (weight ranging from 0.8 to 1 kg) were anaesthetized by exposure to cotton wool soaked in tribromoethanol solution (200 mg tribromoethanol dissolved in 10 ml of ether) and killed by cervical dislocation and exsanguination. Hearts were quickly removed and placed in normal Tyrode solution (containing (mm): NaCl, 140; KCl, 5.4; CaCl2, 1.8; MgCl2, 1; d-glucose, 5.5; Hepes-NaOH, 5; adjusted to pH 7.4) where ventricles were removed and the sino-atrial node area exposed. The region enclosed laterally by the crista terminalis and the interatrial septum, and caudorostrally by the orifices of the inferior and superior venae cavae, was isolated by removing surrounding atrial tissue, and cut into a series of 1 mm wide strips perpendicularly to the crista terminals border. These were rinsed several times with a nominally Ca2+-free solution (mm: NaCl, 140; KCl, 5.4; MgCl2, 0.5; KH2PO4, 1.2; d-glucose, 5.5; taurine, 50; Hepes-NaOH, 5; adjusted to pH 6.9) and transferred to an enzyme solution containing (in the same nominally Ca2+-free solution) 224 U ml−1 collagenase Type I (tryptic activity, 0.36 Umg−1), 1.9 U ml−1 elastase (Type IV), 0.6 U ml−1 protease Type XIV, 1 mg ml−1 bovine serum albumin (BSA) and 200 μm CaCl2. The tissue was triturated in the enzyme solution for 15–25 min at 37 °C, until the pieces became soft and filamentous. The tissue was next rinsed with a Ca2+-free, potassium glutamate-based salt solution (mm: KCl, 20; glutamic acid, 70; d-hydroxybutyric acid (Na+ salt), 10; KH2PO4, 10; Hepes-KOH, 10; KOH, 80; taurine, 10; BSA 1 mg ml−1; adjusted to pH 7.4), then triturated for 15 min at 37 °C in this solution. The cell susupension was filtered through a nylon mesh and the calcium concentration gradually raised to a final concentration of 1.3 mm. The final storage solution contained (mm): NaCl, 100; KCl, 35; CaCl2, 1.3; MgCl2, 0.7; BSA, 1 mg ml−1; pH7.4; gentamicin, 18 μg ml−1. The cells were kept in this solution for up to 8 h (temperature, 4 °C), until used.

The cell population isolated by this procedure was mostly composed of spindle-shaped myocytes, of variable size, that, after stabilization at 36 °C, had spontaneous electrical and mechanical activity.

Experimental solutions

Isokted cells were allowed to settle to the bottom of a 30 mm diameter polylysine-coated Petri dish, which contained a plastic ring to reduce total volume to about 1 ml. The dish was placed on the stage of an inverted microscope and continuously superfused at a rate of 2 ml min−1 with normal Tyrode solution, containing (mm): NaCl, 140; KCl, 5.4; CaCl2, 1.8; MgCl2, 1; d-glucose, 5.5; Hepes-NaOH, 5; adjusted to pH 7.4. A thermostatic pipette, with multiple superfusion lines, was positioned over the individual cell being studied to allow solution changes within 1 s. The temperature of the superfusing solution was monitored at the tip of the pipette by a fast response digital thermometer (BAT-12, Physitemp, NJ, USA). Mow rate through the various lines was adjusted to maintain the temperature constant at 35 ± 0.1 °C. Except when specifically mentioned, membrane potential and currents were measured (Axopatch 200A, Axon Instruments) in the whole-cell configuration using the perforated-patch technique (Horn & Marty, 1988) implemented with amphotericin. In ruptured-patch experiments the pipette solution contained (mm): NaCl, 10; potassium aspartate, 120; KCl, 10; MgCl2, 2; ATP-Na+ salt, 2; creatine phosuphate, 5; GTP, 0.1; EGTA KOH, 10; CaCl2, 4.0 (calculated free Ca2+, 10−7m); Hepes-KOH, 10; pH 7.2). In perforated-patch experiments, ATP, creatine phosuphate and GTP were omitted, and a saturating amphotericin (Sigma) concentration (260 μm) was added to the pipette solution. In some experiments (see Results) extracellular K+ concentration was increased to 19.5 mm; the osmolarity of the solution was maintained constant by removing equimolar Na+.

Collagenase was obtained from Worthington Biochemical; elastase, protease and nifedipine from Sigma; nisoldipine and E-4031 were generous gifts from Bayer pharmaceuticals and Sanofi Recherche, respectively.

Data acquisition

In perforated-patch experiments, measurements were performed after stabilization of the series resistance (RS) at a value < 20 MΩ. Membrane capacity (Cm was estimated from the capacity compensation current required, after compensation of RS, to eliminate capacity transients during small hyperpokrizing voltage clamp pulses.

The action potential clamp technique is illustrated in Fig. 1. Transmembrane potential was recorded during stable spontaneous pacemaking activity (current clamp conditions) in a single cell during Tyrode superfusion. A membrane potential waveform, corresponding to a single pacemaking cycle, was digitized (2 kHz, 12 bit resolution) and stored in computer memory. After switching to voltage clamp mode, the acquired waveform was then used as the command signal to drive membrane potential in the same cell, thus allowing measurement of total membrane current during its pacemaking cycle. Sustained pacemaking activity was simulated by repeatedly applying the same waveform at the cycle length present during spontaneous activity. Computer processing time introduced a < 3 ms lag between subsequent cycles, during which membrane potential was held at the last value measured in the cycle. To prevent distortion of the diastolic depolarization phase, the artifact generated by the lag was made to occur during action potential upstroke. After a short stabilization period under action potential clamp conditions, total membrane current recorded during Tyrode superfusion (control) settled to a value close to zero, except for a short glitch corresponding to the intercycle kg. The current recorded during subsequent exposure to a channel blocker (compensation current), provided a mirror image of the contribution of the blocked component to the pacemaking cycle. Membrane current was filtered at 1 kHz and stored, along with membrane potential, in a PCM videorecording system. Recordings were considered acceptable only if: (1) the current recorded in control conditions was negligible, indicating that reproduction of the spontaneous activity was satisfactory; (2) wash-out of the blocker, performed after each exposure, was followed by reversal of the compensation current. Complete reversal could be obtained with all agents except E-4031, whose effects could be reversed only by approximately 70% (Verheijck, van Ginneken, Bourier & Bouman, 1995). Off-line digital analysis was performed after replaying the data into a 12-bit analog-to-digital conversion board. In standard voltage clamp experiments, square wave pulses were applied to the cell at 10s intervals from a holding potential of −40 mV; the resulting currents were filtered (5 kHz), digitized (5 kHz) and stored in computer memory for subsequent automatic analysis.

Figure 1.

Action potential clamp technique

(See Methods for details.) The waveform corresponding to a complete pacemaker cycle is recorded under current clamp conditions. When the waveform is reapplied in voltage clamp conditions during Tyrode superfusion (a, Control), the current approaches zero. Superfusion with a channel blocker (e.g. nifedipine (Nif)) induces a compensation current b, that reverses upon wash-out (a, Return). The blocker-sensitive current (e.g. INif) is obtained by digital subtraction of compensation currents from ‘control’ (or ‘return’) currents. In this and in the following figures, membrane potential and blocker-sensitive current traces are aligned on the time axis.


Throughout text and figures, data from action potential clamp experiments are presented in terms of ‘blocker-sensitive current’ (e.g. nifedipine-sensitive current). The latter was obtained by subtraction of compensation current traces from control traces (Fig. 1, bottom trace). Current traces from three to five cycles at steady state in each condition were averaged before subtraction.

The purpose of action potential clamp experiments in this work was to explore the time course of currents during the pacemaker cycle; thus, observations are difficult to present in quantitative terms. Nonetheless, quantitative measurements were needed to test for the reproducibility in the direction and magnitude of the currents recorded. To comply with such a requirement, we measured the ‘average current’ present in each phase of the cycle (e.g. average systolic current). This was done by integrating the current trace over the appropriate interval and dividing the result by the integration interval. Current density was estimated by dividing the average current by Cm. To avoid distortion by the lag artifact (see Methods) systolic currents were measured starting from the artifact end to the time of maximum diastolic potential (Emax). Diastole was denned as the time interval between Emax and the take-off of the following action potential (Eto), identified as the inflection of diastolic depolarization (phase 4). In all figures, membrane potential and current traces are aligned and the diastolic interval is delimited by arrows. The phase of diastolic depolarization at which each current began to contribute inward charge movement was identified by measuring the earliest diastolic potential at which the current was inward (Ec). The difference between Eo and EC was then reported as a percentage of total diastolic depolarization (EtoEmax).

Statistical analysis

Means were compared by Student's t test for paired or unpaired observations as appropriate. ANOVA for repeated measurements was performed whenever more than two means were compared. A probability level (P) < 0.05 was used to define significance throughout the study (n.s., not significant). Significance of correlation was tested by χ2 analysis. In the text and figures, values are presented as means ± standard error of the mean (s.e.m.).


Action potential parameters

All the cells studied (n= 24) had regular spontaneous pacemaking activity at a mean cycle length of 389 ± 19.7 ms and had a mean Cm of 39.9 ± 4.1 pF. Action potential parameters were similar among subgroups of cells used for measurement of specific currents; mean values for these parameters in perforated-patch experiments are reported in Table 1. Under buffering of intracellular Ca2+ (ruptured patch, 10 mm intracellular EGTA), spontaneous activity became unstable and ceased in the majority of cells several minutes after achieving the whole-cell configuration.

Table 1. Action potential parameters
 Amplitude(mV)Phase 0 slope(V s−1) E max(mV) E to(mV)Phase 4 slope (V s−1)
  1. E max, maximum diastolic potential; Eto, take-off potential.

s.e.m. 2.561.671.771.430.0068
n 2424242424

Due to such a phenomenon, possibly reflecting Ca2+ dependency of currents essential to pacemaking (see Discussion), in ruptured-patch experiments action potential waveforms were recorded shortly after attaining the whole-cell configuration, before the effects of cell dialysis could take place. Nonetheless, due to the potentially different recording conditions, action potential parameters of ruptured-patch experiments (n= S) were analysed separately (cycle length, 422 ± 51.8ms; amplitude, 92.82 ± 1.6mV; phase 0 slope, 22.18 ± 3.38 Vs−1; Emax, -58±2.86mV; take-off potential, -35.98 ± 2.91 mV; phase 4 slope, 0.092 ± 0.018 V s−1). The difference between action potential parameters recorded under ruptured- and perforated-patch conditions, respectively, did not reach statistical significance.

DHP-sensitive currents

Nifedipine-sensitive current. The current sensitive to 5 μm nifedipine (INif), a concentration adequate to block ICa,L almost completely (data not shown), was analysed in twelve cells. Figure 2 shows a pacing cycle (Fig. 2A) and the INif recorded during its course (Fig. 2B). Systolic INif displayed an initial inward peak, followed by a large inward ‘plateau’ component persisting through most of the action potential. During final repolarization INif reversed direction and was present as a small outward current during most of diastolic depolarization, to become inward again at the foot of the following action potential. A plot of INifvs. membrane potential (Fig. 2C), limited to the diastolic interval, shows that Ec corresponded to the final part of phase 4 depolarization. Figure 2D shows that upon nifedipine superfusion and wash-out, respectively, outward diastolic INif built up and decayed with slower kinetics than inward systolic INif.

Figure 2.

Current sensitive to 5 μm nifedipine (INif)

A, membrane potential. B, INif recorded during the pacemaker cycle shown in A, C, dynamic I-V relation obtained by plotting the current shown in B vs. the potential shown in A during the diastolic depolarization phase (Ec, membrane potential at which the current becomes inward during the diastole). D, systolic (filled squares) and diastolic (filled circles) currents measured every cycle during nifedipine superfusion (marked by the bar) and after wash-out; the time at which the currents were measured during each cycle is shown in B by a filled square (systolic) and a filled circle (diastolic). In all panels of this and the following figures, arrows delimit, the diastolic interval.

Mean systolic INif density was –2.09 ± 0.24 pA pF−1 (n= 12). Mean diastolic INif density was 0.105 ± 0.030 pApF−1 (n= 12). The membrane potential at which INif became inward (Ec) was –40.7 ± 1.8 mV (n= 12), which represented 74.1 ± 6.4% of the diastolic depolarization.

Outward diastolic INif might result from direct blockade of outward currents by the rather high concentration of nifedipine required to inhibit ICaL. To test this hypothesis we repeated the experiments using nisoldipine, a highly specific dihydropyridine Ca2+ channel blocker, active at submicromolar concentrations.

Nisoldipine-sensitive current. The current sensitive to 0.2 μm nisoldipine (INis), a concentration adequate to largely inhibit ICa,L (Daleau & Turgeon, 1994), was studied in ten cells.

The time course of INis, recorded in perforated-patch experiments (n= 5; Fig. 3A and B), was similar to the one described for INif (see Fig. 2). Indeed, INis was outward during diastolic depolarization, becoming inward only in the late portion of this phase, in all the cells tested. INif was not significantly different from INif, during the action potential plateau (–1.30 ± 0.25pA pF−1; n.s. vs. INif) as well as during the diastolic depolarization (0.21 ± 0.02 pA pF−1; n.s. vs. INif). The membrane potential at which INif became inward (Ec) was –34.8 ± 3.6 mV, which represented 90.9 ± 3.3% of total diastolic depolarization (n.s. vs. INif). Thus, during diastolic depolarization, DHP-sensitive current was unlikely to result from drug-specific ancillary properties.

Figure 3.

Current sensitive to 0.2 μM. nisoldipine (INis) recorded in perforated- (A and B) and ruptured- (C and D) patch experiments

A and C, membrane potential; B and D, INis recorded during the pacemaker cycle shown in the upper panels. Bottom panels: comparisons between perforated- and ruptured-patch experiments; left, mean INis density during diastolic depolarization (INis,DD); right, ratio between total diastolic depolarization span (▵EDD=Eto-Emax) and Ec (▵EC=EC - E (*P < 0.05).

Nature of the diastolic DHP-sensitive current. We considered the hypothesis that outward diastolic DHP-sensitive current might represent the sensitivity of an outward current to changes in intracellular Ca2+ levels. This hypothesis was tested by measuring INis while buffering intracellular Ca2+ fluctuations (ruptured-patch configuration, 10 mm EGTA in pipette solution; n= 5) (Fig. 3C and D).

Due to unstable spontaneous activity under such conditions (see above), action potential waveforms were recorded before significant cell dialysis occurred. Total membrane current, albeit close to null immediately after switching to the voltage clamp mode (see Methods), slightly changed over the following minutes, as cell dialysis progressed; thus, a new steady state had to be attained before INis could be measured. In the presence of intracellular Ca2+ buffering, diastolic INis became very small and reversed to inward direction earlier during diastolic depolarization (Ec at -51.6 ± 5.1 mV; P < 0.05vs. perforated-patch experiments), thus resulting in an inward average diastolic current (-0.155 ± 0.06 pA pF−1; P < 0.05vs. perforated-patch experiments). Mean values of diastolic INis density and Ec in perforated- vs. ruptured-patch experiments are compared in the bar graphs at bottom of Fig. 3. Mean systolic INis was larger (more inward) in ruptured- than in perforated-patch experiments (-2.257 ± 0.227 vs. -1.304 ± 0.247 pA pF−1; P < 0.05). This is likely to reflect inhibition, by intracellular Ca2+ buffering, of the outward Ca2+-dependent conductance also observed during diastole; removal of Ca2+-dependent inactivation of ICa,L (Lee, Marban & Tsien, 1985) might also partly account for the phenomenon.

The small magnitude of the DHP-sensitive diastolic current and its close dependency on the shape of the potential waveform prevented an analysis of its ionic nature by means of traditional voltage clamp protocols. Nonetheless, indicative information was obtained from action potential clamp experiments in which INif was recorded at two extracellular K+ concentrations (5.4 and 19.5 mm), producing estimated K+ equilibrium potentials negative to the diastolic depolarization range (-84.5 mV) and comprised by it (-50.4 mV), respectively. As shown in Fig. 4, switching to higher extracellular K+ concentration reversibly shifted diastolic INif in the negative direction, resulting in a change of mean diastolic current from 0.083 ± 0.039 pA pF−1 to -0.206 ± 0.063 pA pF−1 (n= 1; P < 0.05).

Figure 4.

Effect of increasing extracellular K+ concentration ([K−1]0) on nifedipine-sensitive current (INif) during diastolic depolarization

A, I Nif recorded, during diastolic depolarization, in the presence of [K+]0 of 5.4 mm (Control, Return) and 19.5 mm (high [K+]0). B, dynamic I–V relation obtained by plotting the current shown in A vs. membrane potential (not shown). C, mean values of the average diastolic INif density (INif,DD), measured with [K+]0 of 5.4 mm and 19.5 mm in 7 cells (*P < 0.05).

Sensitivity to changes in the K+ gradient was consistent with contribution of a K+ conductance to the DHP-sensitive diastolic current. Since the E-4031 -sensitive component (IK,r) should represent most of the delayed rectifier current in sino-atrial myocytes (Verheijck et al. 1995), experiments were performed to test whether outward DHP-sensitive diastolic current might result from Ca2+-induced enhancement of IK,r. To this end, INif was measured after blocking IK,r by 1 μm E-4031 (Fig. 5). However, the time course and magnitude of INif, particularly of its diastolic component (0.15 ± 0.05 pA pF−1; n= 4; n.s. vs. INif in control), were unchanged under these conditions. Thus, the outward direction of INif during the diastolic depolarization could not be accounted for by Ca2+-dependent enhancement

Figure 5.

Current sensitive to 5 μM. nifedipine (INif) recorded in the presence of 1 μM. E-4031

A, membrane potential. B, INif recorded during the pacemaker cycle shown in A. C, dynamic I–V relation obtained by plotting the current shown in B vs. the potential shown in A during the diastolic depolarization phase. D, comparison between mean INif density during diastolic depolarization (INif,DD) measured in the absence (Control) and presence of E-4031. The current sensitive to 1 μm E-4031, recorded in this cell before applying nifedipine, is shown in the inset of B.

Cs+-sensitive current

The current sensitive to 2 mm Cs+ (ICs) was studied in six cells (Fig. 6). Systolic ICs, was consistently observed as an outward current, peaking during the repolarization phase (Fig. 6A and B). Mean systolic ICs density was 0.42 ± 0.12 pA pF−1. Although the time course and magnitude of systolic ICs, were somewhat variable among cells, its direction was outward at membrane potentials positive to -52.2 ± 1.9 mV, thus substantially lower than the expected reversal potential of If (around −18 mV) (DiFrancesco et al. 1986; Zaza, Rocchetti & DiFrancesco, 1996). To test whether outward systolic ICs, was due to a partial blockade of IK, the effects on IK of 2 mm Cs+ and of 1 μm E-4031 were compared in standard whole-cell voltage clamp experiments (Fig. 7). IK was activated by 1 s depolarizing steps from a holding potential of −40 mV, and measured as the amplitude of the tail current upon returning to the holding potential. Contamination by Ca2+ currents was minimized by adding 0.2 μm nisoldipine to all solutions. As can be better appreciated from Fig. 7B, IK was slightly, but reversibly reduced by Cs+. In seventeen cells, IK was 229 ± 34.4 pA in control, was reduced to 205 ± 33.9 pA (-16 ± 9 ± 3.3%; P < 0.05) and returned to 229 ± 34.4 pA upon wash-out. In thirteen of these cells, Cs+ and E-4031 could be tested in sequence. In these cells Cs+ blocked 14.4 ± 3.8% of control IK tail magnitude (P < 0.05) compared with a 48.9 ± 7.9% blockade (Fig. 7C) induced by E-4031; the proportion of tail current sensitive to E-4031 (IE-4031) was rather variable among different cells (range, 13–100%) and inversely correlated to control tail amplitude (correlation coefficient (r) = 0.78; P < 0.05). The ratio between Cs+ and E-4031 effects was 0.29 ± 0.05. This can be compared with the ratio between ICs (Fig. 7D) and IE-4031 (Fig. 7E) in four cells in which both currents could be measured sequentially in action potential clamp conditions. The ratio between systolic ICs, measured at its peak value, and IE-4031, measured at the same instant of ICs, peak (close but not always identical to IE-4031 peak time), was 0.39 ± 0.07. The ratio between Cs+ and E-4031 effects in standard voltage clamp and action potential clamp experiments is compared in Fig. 7F.

Figure 6.

Current sensitive to 2 mm Cs+ (ICs)

A, membrane potential. B, ICs recorded during the pacemaker cycle shown in A. C, dynamic I–V relation obtained by plotting the current shown in B vs. the potential shown in A during the diastolic depolarization phase. Both B and C also show the net membrane current (Inet, represented by *), obtained by multiplying the derivative of the trace shown in A (membrane potential) by -Cm.

Figure 7.

Effect of 2 mm Cs+ and 1 μm E-4031 on the delayed rectifier current (IK) in standard voltage clamp and action potential clamp experiments

A, effect of 2 mm Cs+ and 1 μm E-4031 on the current recorded when stepping to 0 mV from a holding potential of −40 mV in the presence of 5 μm nifedipine; tail current amplitude was measured as the difference between peak current upon return to holding potential and holding current. B, tail current amplitude sampled through the experiment; the times of Cs+ and E-4031 superfusion are marked by the bars. C, mean changes in tail current amplitude induced by Cs+ and E-4031 in 13 cells (Con, control; Ret, return). *P < 0.05vs. control. D and E, ICs and IE-4031 recorded from a single cell under action potential clamp conditions; the arrows mark the time of peak ICs, in both panels (traces aligned on the time axis). F, the ratio between Cs+ and E-4031 effects (expressed as ICs, and IE-4031) is compared between standard voltage clamp (IK) and action potential clamp (Isyst) experiments.

As shown in Fig. 6B, approximately at the time of Emax, ICs, reversed to become an inward current that reached a maximum in the last third of the diastolic depolarization. Diastolic ICs was consistently observed as an inward current; its mean density was -0.24 ± 0.05 pA pF−1 (equivalent to 33.4 ± 7.8 mC F−1 of normalized charge; n= 6), and the threshold of inward ICs (EC) almost coincided with Emax (-52.2 ± 1.9 mV representing 3.8 ± 2.5% of diastolic excursion). ICs can be compared with the net membrane current (Inet) required to support diastolic depolarization (marked by an asterisk in Fig. 6B and C) in the same cells, calculated as follows:

display math(1)

where Cm is membrane capacity and dV/dt is the rate of membrane potential change. Diastolic ICs, exceeded Inet in all the cells tested; the mean value of the ratio ICs/Inet was 3.69 ± 1.12(n= 6).

It has been recently reported that IK,r, recorded as the E-4031-sensitive current, persists to a significant extent during the whole diastolic interval (Ono & Ito, 1995). The presence of extra outward diastolic current might account for the observation that If alone exceeded the net inward current flowing during diastolic depolarization. However, previous experiments were performed in ruptured-patch conditions, preventing a safe comparison with the other currents recorded in the present experiments. To allow such a comparison, the current sensitive to 1 μm E-4031 (IE-4031) was measured from seven cells in perforated-patch conditions. IE-4031 was outward during the whole pacemaking cycle, with a time course similar to the one previously described (Ono & Ito, 1995) (for an example see inset of Fig. 5). Peak systolic IE-4031 density was 2.8 ± 0.45 pA pF−1. IE-4031 was larger than the net current required to support repolarization as calculated by eqn (1): the ratio of IE-4031/Inet peak values was 2.08 ± 0.40. Although peak IE-4031 occurred slightly later than the maximum repolarization rate (98.6 ± 7.48 ms vs. 89.5 ± 8.73 ms from upstroke; P < 0.05), the two times were highly correlated (r= 0.97; P < 0.05). However, in the same cells, peak magnitudes of IE-4031 and of the net estimated current (Inet, eqn (1)) were not correlated (r= 0.07; n.s.). IE-4031 appeared as an outward current during the whole diastolic depolarization, with a mean density of 0.41 ± 0.08 pA pF−1. Although IE-4031 magnitude progressively decayed during diastole, a substantial proportion of it persisted at the end of each cycle; thus, as previously reported, IE-4031 diastolic deactivation was apparently incomplete.


Properties of sino-atrial myocytes

Even if morphologically similar, pacemaking cells from different regions within the sinus node area may have distinct functional properties, usually catergorized either as peculiar to the central nodal area (earliest activation site) or to ‘transitional’ elements. Morphological and electro-physiological properties of automatic cells in our preparation do not allow a clear-cut separation between different cell types. Thus, although only spindle-shaped, spontaneously active cells were studied, no effort was made to discriminate potentially dominant pacemakers from transitional elements. As estimated from Cm, cell size was in the high range, closer to the one reported for transitional cells (Denyer & Brown, 1990b). The rate of action potential upstroke (16 V s−1) was also higher than the one described in dominant pacemaker cells. However, maximum diastolic potentials, around −50 mV, would be compatible with those recorded in the central nodal area (Denyer & Brown, 1990b; Bleeker, MacKaay, Masson-Pevet, Bouman & Becker, 1980; Kodama & Boyett, 1985). Thus, it is difficult to attribute the findings of this study to either cell type. On the other hand, most of the information on action potential parameters comes from impaled multicellular preparations or ruptured-patch recordings in isolated myocytes. Due to potential differences between such techniques and the perforated-patch method, the comparison might be unwarranted.

Intracellular Ca2+ buffering, while making diastolic DHP-sensitive current more inward, inhibited pacemaking. Such apparently contrasting findings might result from Ca2+ dependency of inward pacemaking currents other than ICa,L: possibly If (Hagiwara & Irisawa, 1989).

DHP-sensitive current

Except for the outward diastolic component, INif and INis profiles during the action potential were similar to the one predicted for ICa,L (Demir, Clark, Murphey & Giles, 1994), which represents the main component of the DHP-sensitive current. However, changes in other currents, sensitive to intracellular Ca2+ levels, may result from inhibition of Ca2+ influx and might thus contribute to DHP-sensitive current. These include the Na+–Ca2+ exchanger current and K+ or Cl currents carried by Ca2+-sensitive channels (Zygmunt & Gibbons, 1992; Nitta, Furukawa, Marumo, Sawanobori & Hiraoka, 1994).

In perforated-patch experiments, diastolic DHP-sensitive current was a small outward component that reversed to inward only in the proximity of the action potential threshold. This partly unexpected finding was consistently reproduced for both INif and INis. The outward component, while unchanged by concomitant IK,r blockade by E-4031, showed sensitivity to changes in K+ driving force; thus suggesting K+ as the permeant ion. Such current disappeared after rupturing the patch in the presence of a high pipette concentration of the Ca2+ buffer EGTA. Moreover, upon nifedipine superfusion the development of the outward diastolic component of INif lagged behind ICa,L blockade, as revealed by the earlier onset of inward systolic current (Fig. 2D). All these findings concur to support the hypothesis that Ca2+ induced enhancement of a K+ current, insensitive to E-4031, may contribute to charge movement during diastolic depolarization. A previous report indicating dependency on Ca2+ influx of the enhancement of sino-atrial IK by adrenergic agonists (Brown & DiFrancesco, 1980) is consistent with our findings and suggests that an E-4031-insensitive component, possibly similar to IKs of ventricular myocytes (Nitta et al. 1994), may actually contribute to sino-atrial IK. In the ruptured-patch experiments, in which the outward component was largely removed, inward DHP-sensitive current appeared at about −52 mV, a potential substantially negative to the activation threshold of ICa,L, as measured by standard voltage clamp protocols (Hagiwara et al. 1988). In a previous study Doerr et al. (Doerr et al. 1989) reported that the current sensitive to the Ca2+-channel blocker D600 was inward during diastolic depolarization and that its threshold potential was about −50 mV. Since Doerr's experiments were performed under ruptured-patch conditions (10 mm EGTA in the pipette solution) such findings are fully consistent with those of the present study.

A steady-state inward DHP-sensitive current, present at diastolic membrane potentials (between −70 and −50 mV), was recently described in rabbit sino-atrial myocytes (Guo, Ono & Noma, 1995). Inactivation of this current positive to -50mV, with slow recovery kinetics (1.36s at -80mV) (Guo et al. 1995), might explain why such a current was not detected in the present experiments: during repetitive activity, the diastolic interval (about 160 ms) might be too short to allow for the recovery of a significant number of these channels. On the other hand, even in ruptured-patch experiments, Ca2+-activated outward current might partially persist, thus masking inward DHP-sensitive components. The finding of outward DHP-sensitive current during diastolic depolarization does not conflict with the notion that DHPs may reduce sino-atrial pacemaking rate. Indeed, DHPs may reduce pacemaking rate by increasing activation threshold, rather than depressing diastolic depolarization (A. Zaza, unpublished observation).

Cs+-sensitive current

I Cs had a complex time course during the activation cycle: a substantial outward component, peaking during the repolarization phase, was followed by an inward component, progressively developing from the very beginning of diastolic depolarization. Previous reports suggested that 2 mm Cs+ almost completely blocked If, without affecting IK in sino-atrial myocytes (Denyer & Brown, 1990a). Nonetheless, the systolic component of ICs is unlikely to represent If for two reasons: (a) the current was outward at membrane potentials negative to the expected If reversal potential (DiFrancesco et al. 1986; Zaza et al. 1996); (b) since Cs+ blockade of outward If is negligible (DiFrancesco, 1982; DiFrancesco et al. 1986), outward ICs should not include If. By measuring IK in standard voltage clamp experiments, we have shown that this current was slightly but consistently reduced by 2 mm Cs+. A quantitative comparison between outward ICs and IK, measured as E-4031-sensitive current, suggests that Cs+ blockade of IK might account for the outward component of ICs.

The inward component of ICs is likely to reflect If activation, and suggests that a substantial amount of this current is present during diastolic depolarization. Since ICs may contain a small proportion of IK, which is present as an outward current during the whole diastole, ICs may actually underestimate If. Moreover, due to superimposed IK deactivation and voltage dependency of If blockade, the time course of ICs might also slightly diverge from that of If. The magnitude and time course of ICs may be compared to those predicted for If in numerical models of sino-atrial pacemaking. Various models of sino-atrial activity, reviewed by Wilders et al. (Wilders, Jongsma & vane charge movement contributed by If during a pacemaking cycle (from 5.7 to 35 mC F−1). The mean charge movement contributed by ICg during a diastolic interval was 33.23 mC F−1 in our experiments, a value reasonably close to the one (35 mC F−1) obtained in the most recent computations (Wilders et al. 1991). The time course of ICs illustrated in Fig. 6 is also comparable to the one predicted for If in the same simulations.

According to If kinetics estimated from standard voltage clamp experiments, only a very small amount of this current should be activated during sino-atrial pacemaking (DiFrancesco, 1991). The relevance of If in providing diastolic inward current has been questioned on this basis (Irisawa et al. 1993). The results of the present study strongly suggest that a substantial If activation occurs during diastolic depolarization; the resulting current may be 3-fold larger than the net current required to support automaticity. The reasons for the discrepancy between our observation and predictions from standard voltage clamp experiments are not obvious. Still, it should be considered that processes, potentially affecting If kinetics, may be influenced by the presence of repetitive activations, a factor generally disregarded in standard voltage clamp studies. For instance, If magnitude may be increased, although indirectly (Zaza, Maccaferri, Mangoni & DiFrancesco, 1991), by elevated intracellular Ca2+ activity (Hagiwara & Irisawa, 1989), a plausible consequence of fast repetitive activations. The presence of measurable inward ICs almost immediately after repolarization is apparently at odds with the slow kinetics of If activation at diastolic potentials (van Ginneken & Giles, 1991). This suggests that If may not deactivate completely during repetitive activity; this would also increase the amount of If available during diastolic depolarization. Direct comparisons between If kinetics under ruptured- vs. perforated-patch conditions are, to our best knowledge, not available; however, data from different studies performed in either of the two conditions (Zaza et al. 1996; Accili, Robinson & DiFrancesco, 1997) suggest that If activation potentials and conductances may not differ markedly. Nonetheless, it is difficult to rule out that the absence of cell dialysis may contribute to the relatively large diastolic If observed in the present study.

The presence of a sustained outward component contributed by IK,r(IE-4031) may account for the observation that If alone may exceed the net inward current present during diastolic depolarization. Diastolic IE-4031 was actually twice as large as ICs; even if precise quantitative comparisons might be hampered by a slight underestimation of If (see above), such a large difference suggests that inward currents besides If may exist during diastolic depolarization (Hagiwara, Irisawa, Kasanuki & Hosoda, 1992; Verheijck et al. 1995; but see DiFrancesco, 1991).

The proportion of total tail current blocked by E-4031 in standard voltage clamp experiments (Fig. 7A–C) averaged 48.9%. However, control tail amplitudes and their percentage reduction by E-4031 were both widely variable among cells and inversely correlated to each other. Such findings suggest that variable amounts of IK,s may be expressed in different cells, thus resulting in a large variability of the percentage of tail amplitude sensitive to E-4031. Since, due to its slow kinetics, IK,s contribution to sino-atrial node repolarization may be minor (Lei, Varghese, Gerlach, Lang & Kohl, 1997), variable expression of this current may not necessarily translate into heterogeneity of repolarization rates.

Although the times of occurrence of peak systolic IE-4031 and peak systolic net current were highly correlated, the respective current magnitudes were not. This might imply that, although fast repolarization may indeed be caused by IK,r, its rate may be significantly affected by other concomitant currents (e.g. background currents).

Limitations of the study

In action potential clamp studies the individual components of total current are dissected according to their sensitivity to a blocking agent (drug). The extent to which a drug-sensitive current reflects a single ionic current depends on drug selectivity, and the existence of other conductances modulated by the ion normally permeating the channel. Such aspects, particularly relevant for Ca2+ currents, have been addressed in the appropriate section. How closely the blocker-sensitive current reflects the magnitude and time course of the ionic current depends on the potency and voltage dependency of block. DHPs were used at concentrations almost completely suppressing ICa,L at all potentials. On the other hand, ICs, may not accurately reflect the time course of If due to voltage dependency of block by Cs+ (DiFrancesco, 1982; DiFrancesco et al. 1986). Since 1 μm E-4031 only partially blocks IK, the magnitude of this current may be underestimated by IE-4031; experiments in sino-atrial myocytes do not suggest a strong voltage dependency for IK blockade by this agent (Verheijck et al. 1995).

Due to the effects of series resistance and stray capacitance, the time courses of the membrane potential and of the command potential may diverge. Such a divergence will be maximal during fast potential changes (i.e. the action potential upstroke) and in the presence of large currents. Moreover, the intercycle lag (see Methods) produced an artifact during the upstroke phase. Thus, current measurements during this phase are unreliable and their interpretation is unwarranted. Conversely, during diastolic depolarization, membrane potential changes were slow and the currents developing following channel blockade were small (tens of pA), thus reducing the series resistance error to an acceptable level (0.1–0.2 mV per 10 pA of current).


While keeping in mind the limitations inherent to the experimental method used, several conclusions, relevant to the ionic mechanisms of sino-atrial pacemaking, can be drawn from this study. Ca2+ influx activates (or enhances) a K+ conductance, resulting in a net outward DHP-sensitive current during diastolic depolarization. Thus, the effect of Ca2+ influx on diastolic membrane potential may be more complex than the depolarizing effect expected from the inward direction of ICa,L.

The magnitude of If during diastolic depolarization exceeded previous predictions based on extrapolation of current kinetics (Irisawa et al. 1993). If may be larger than the net current required to support automaticity; thus, If-based automaticity would theoretically be possible in sino-atrial myocytes.

In agreement with previous reports (Ono & Ito, 1995), IK‘accumulation’ during repetitive activity generated a significant pseudo-instantaneous component during diastole. The balance between ICs and IE-4031 was outward, thus suggesting that other inward currents, besides If, are present during diastolic depolarization. Obviously, this does not rule out that If contribution may be necessary to cause diastolic depolarization.


The authors are grateful to Professors Arnaldo Ferroni for reading the manuscript and providing constructive criticism, and to Mr Gaspare Mostacciuolo for expert technical assistance. This work was supported by grants from the Ministero dell'Università e della Ricerca Scientifica e Tecnologica (MURST).