Journal of Physiology (1991), 434, pp.

57-83 57
With 13 figures
Printed in Great Britain

VOLTAGE CLAMP MEASUREMENTS OF THE HYPERPOLARIZATION-

ACTIVATED INWARD CURRENT If IN SINGLE CELLS FROM RABBIT
SINO-ATRIAL NODE
BY A. C. G. VAN GINNEKEN* AND W. GILESt
From the Departments of Medical Physiology and Medicine, University of Calgary,
Health Sciences Centres, 3330 Hospital Drive N. W., Calgary, Alberta,
Canada T2N 4N1
(Received 22 November 1989)

SUMMARY

1. The kinetics and ion transfer characteristics of the hyperpolarization-activated

inward current, If, have been studied in single cells obtained by enzymatic dispersion
from the rabbit sino-atrial (S-A) node. These experiments were done to assess the role
of If in the generation of the pacemaker depolarization in the S-A node.
2. The activation and the deactivation of If in these single cells are accompanied
by significant conductance increases and decreases respectively, confirming earlier
findings from multicellular man-made strips of rabbit S-A node, and from
mammalian Purkinje fibres.
3. The steady-state activation of If lies between -40 and - 120 mV, and its
voltage dependence can be described by a Boltzmann relation with the half-
activation point at approximately -70 mV.
4. The delay or sigmoidicity in both the onset of If and the deactivation of the tail
currents can be accounted for semi-quantitatively by using a second-order
Hodgkin-Huxley kinetic scheme.
5. The reversal potential for If is -24+ 2 mV (mean+ S.E.M., n = 6). It does not
change significantly as a function of the amount of If which is activated, indicating
that ion accumulation or depletion phenomena are not important variables
controlling the time course of I,, or its selectivity.
6. The fully-activated current-voltage relationship for I, is approximately linear
with a slope conductance of 12-0+0 88 nS per cell (mean+s.E.M., n = 6).
7. A simple mathematical model based on the measured values of maximum
conductance, reversal potential, and kinetics of If has been developed to simulate the
size and time cburse of If during typical spontaneous pacemaker activity in rabbit
sino-atrial node Sells. The calculations show that If can change significantly during
pacing and suggest that this current change is, at least in part, responsible for the
pacemaker depolarization.
* Permanent address: Department of Medical Physiology, University of Amsterdam, Academic
Medical Centre, AMC-MO1 Meilbergdreef 15, 1105 AZ Amsterdam, The Netherlands.
t To whom all correspondence and reprint requests should be sent.
MS 8091
58 A. C. G. VAN GINNEKEN AND W. GILES
8. Addition of CsCl (1 mM) provided further evidence for a functional role of If in
tvhe generation of the pacemaker depolarization. Caesium blocked If, slowed the
beating rate, and produced a small hyperpolarization.
9. These results indicate that If can contribute a significant net inward current
during the diastolic depolarization, and therefore may be a physiologically important
determinant of heart rate.
INTRODUCTION
An understanding of the ionic mechanism of the pacemaker potential in
mammalian cardiac muscle is of fundamental importance; nevertheless, at present it
remains incomplete and somewhat controversial (Hartzell, 1988; DiFrancesco &
Noble, 1989; Irisawa, Brown & Giles, 1991). For approximately 15 years the work of
Noble & Tsien, 1968) was taken to indicate that the decline of K+ current, iK2, was
the major time- and voltage-dependent current change underlying the pacemaker
potential. However, the interpretation of these voltage clamp results has been
changed: according to DiFrancesco (1981 a, b, 1985), the pacemaker depolarization
in cardiac Purkinje fibres is not caused by a deactivation of an outward (K+) current,
but by activation of an inward current which has been named I,. This current is
carried mainly by potassium and sodium ions with a smaller contribution by Cl-
(Seyama, 1979; van Bogaert & Carmeliet, 1985). Yanagihara & Irisawa (1980). and
Brown & DiFrancesco 1980) first identified this current in man-made 'strips' of
rabbit sinus node tissue. This type of hyperpolarization-activated inward current
has now been recorded in single isolated cells from the sino-atrial (S-A) and atrio-
ventricular (A-V) nodes (for review see DiFrancesco, 1987; Irisawa, 1987, Irisawa et
al. 1991).
Although changes in I, play an important role in the development of the
pacemaker potential (-85 to -60 mV) in cardiac Purkinje fibres, it is still not
certain whether If has a similar function in the diastolic depolarization (approxi-
mately -65 to -45 mV) of A-V or S-A node tissue (cf. Brown, 1982; Brown,
Kimura & Noble, 1982; Irisawa et al. 1991). Experimental studies designed to
determine the role of If during diastolic depolarization have yielded conflicting
results (reviewed by DiFrancesco, 1985; Irisawa et al. 1991). Moreover, most
mathematical models of electrophysiological activity in the S-A node or sinus
venosus suggest that the changes in If during diastolic depolarization may be rather
small (Yanagihara, Noma & Irisawa, 1980; Noble & Noble, 1984; DiFrancesco &
Noble, 1985; Doerr & Trautwein, 1990; Rasmusson, Clark, Giles, Shibata &
Campbell, 1990), compared with the time- and voltage-dependent current changes
during diastole due to (i) activation of ICa (Irisawa, 1987; Hagiwara, Irisawa &
Kameyama, 1988) and (ii) deactivation of IK (cf. Nakayama, Kurachi, Noma &
Irisawa, 1984; Shibata & Giles, 1985; Shibasaki, 1987; Irisawa et al. 1991). How-
ever, a number of recent studies have shown that If in cells from rabbit S-A node
can be reduced by pharmacological agents (van Ginneken, Bouman, Jongsma,
Duivenvoorden, Opthof & Giles, 1987), by Cs+ (van Ginneken & Giles, 1985; Denyer
& Brown, 1990 b), by low doses of acetylcholine (DiFrancesco & Tromba, 1987, 1988;
DiFrancesco, Ducouret & Robinson, 1989) or by adenosine (Belardinelli, Giles &
West, 1988) and these changes in If produce marked slowing of pacing. In addition,
 SINO-ATRIAL NODE PACEMAKING 59
Chang, Gao, Tromba, Cohen & DiFrancesco (1990) have shown that fl-actions of
noradrenaline change If and partly for this reason can increase the slope of the
pacemaker potential.
At present, a number of very important features of I, still have not been studied
experimentally. For example, it is not even known whether If is or is not present
consistently. Some reports indicate that If can be recorded in only approximately
50 % of viable S-A or A-V node preparations (cf. Nakayama et al. 1984; DiFrancesco,
Ferroni, Mazzanit & Tromba, 1986) while in other studies it is consistently present
in healthy spontaneously pacing cells (van Ginneken & Giles, 1985; Nathan, 1986;
Denyer & Brown, 1990 a). The role of If as a universal pacemaker current is also in
doubt, since it cannot always be identified in rabbit Purkinje tissue (Carmeliet,
Coetzee & Mubagwa, 1982) or bull-frog sinus venosus (Brown, Giles & Noble, 1977;
Giles & Shibata, 1985; but see Bois & Lenfant, 1990), both of which exhibit
pacemaker activity. In addition, many of the existing voltage clamp studies aimed
at defining the role of I, in S-A node pacemaker activity have failed to provide
consistent evidence that this current has the appropriate size or kinetics to generate
pacing. For example, even when If is present, sometimes it is activated in a voltage
range too negative (hyperpolarized) to play a functional role in primary pacemaker
tissue (Noma, Morad & Irisawa, 1983). Recently, Hagiwara & Irisawa (1989) have
shown that the size and kinetics of If are strongly modulated by intracellular
calcium. Increases in [Ca2+]i can result in large changes of If in the pacemaker range
of potentials. It is possible, therefore, that some of the observed variability in If is
due to the metabolic state of the isolated cells and/or the details of the Ca2+ buffers
or dialysis techniques used during the suction microelectrode recording.
A number of different procedures for enzymatic isolation of cells from rabbit S-A
node are now available (cf. Irisawa et al. 1991). In this study, one of these procedures
has been combined with whole-cell suction microelectrode voltage clamp techniques
(Hamill, Marty, Neher, Sakmann & Sigworth, 1981; Hume & Giles, 1983), to obtain
quantitative measurements of transmembrane ionic currents in single S-A node cells.
The size(s) and kinetics of these ionic currents can then be related quantitatively to
values of cell capacitance, and can yield quite accurate estimates of the rates at
which membrane potential will change during the activation and deactivation of
transmembrane ionic currents.
The main goals of the experiments reported here were to obtain viable,
spontaneously active isolated cells from the rabbit S-A node, and to make accurate
measurements of the size and kinetics of If in the range of potentials corresponding
to the normal pacemaker potential. A preliminary report of some of this work has
appeared (van Ginneken & Giles, 1985; van Ginneken et al. 1987).

METHODS
Cell isolation procedures and solutions
New Zealand rabbits of either sex, weighing 2-3 kg, were anaesthetized by i.v. injection of
sodium pentobarbitone (40 mg/kg body weight). The heart was quickly excised, mounted on a
modified Langendorff perfusion system (Giles & van Ginneken, 1985) and perfused with a Tyrode
solution containing (in mM): NaCl, 121; KCI, 5; sodium acetate, 2-8; NaHCO3, 21-8; NaH2PO4, 1-0;
MgCl2, 1; CaCl2, 2-2; glucose, 5.5. The solution was pre-heated to 37 °C and equilibrated with a
60 A. C. G. VAN GINNEKEN AND W. GILES
mixture of 95% O2and 5% C02, resulting in a pH of 7-4. The Langendorff perfusion was
maintained for 5-7 min, allowing the blood to be cleared from the heart. During this time adhering
visceral or connective tissue was removed, and then heart was perfused with 'low-Ca2+ Tyrode
solution' (approx. 15 /tM-Ca2+) for 10 min. This was followed by perfusion (20 min) with 'low-Ca2+
Tyrode solution' to which 0-1 mg/ml (15-2 U/ml) collagenase (Sigma, Type 1) was added. The right
atrium was opened with an incision along the sulcus terminalis into the vena cava. A second
incision into the superior vena cava was made as near as possible to the septum. The atrial
appendage was pinned back and the sinus node region was located visually (cf. Bleeker, MacKaay,
Masson-Pevet, Bouman & Becker, 1980; Doerr, Denger & Trautwein, 1989; Roberts, Slocum &
Riley, 1989). A strip of tissue, about 3 mm wide and 7 mm long, was removed from the intercaval
region very near and parallel to the crista terminalis.
This excised S-A node tissue was placed in a small vial which contained 1 ml of a 'dispersion'
solution, which was similar to that used by Masson-Pevet, Jongsma, Bleeker, Tsjernina, van
Ginneken, Treijtel & Bouman (1982). This solution contained (in mM): NaCl, 137-0; KCl, 5-4;
Na2HPO4, 0-35, KH2P04, 0-3; glucose, 5-5; CaCl2, 0-2; MgCl2, 0-6, HEPES, 20-0. Before use it was
equilibrated with 100% 02, and its pH was adjusted to 7-4 with NaOH; 450 U/ml collagenase
(Type 1, Sigma) was then added, and the solution was filtered (0-2 gm Micropore filter). The vial
containing the S-A node tissue was placed in a temperature-controlled bath (37 00) on top of a
magnetic stirrer. A small Teflon-coated bar (approx. 8 mm length) was used to stir the tissue gently
(60 r.p.m.). After 20-30 min the nodal tissue dissociated into either single cells or small clusters of
cells. The cell suspension was then placed in a microcentrifuge and spun at approx. 10 g for 5 min,
after which the supernatant was discarded and the pellet was re-suspended in modified Tyrode
solution, which contained 0 5 mM-Ca2+.
A few drops of the resulting cell suspension were placed in a recording chamber (volume 0-2 ml)
on the stage of an inverted microscope (Nikon Diaphot). The cells were allowed to settle on the
bottom, and then superfusion was started (0-3-1-03 ml/min) with a HEPES-buffered Tyrode
solution containing (in mM): NaCl, 121, KCl, 50, sodium acetate, 2-8; MgCl2, 1P0; CaCl2, 2-2;
glucose, 25; HEPES, 10. The pH of the solution was adjusted to 7-4 by adding NaOH, and it was
equilibrated with 100% 02. In some experiments the calcium current, Ic., was blocked either with
0-5-1P0 mM-CdCl2 or by replacing 2 mM-CaCl2 by equimolar MgCl2.
All experiments were performed at a fixed temperature (± 1 0C) between 30 and 33 0C.
Temperature was maintained using a heating plate in close contact with the bottom of the
recording chamber. This device was made using a 0-2 mm cover-slip covered (sputtered) with a thin
layer of gold. The gold layer provided sufficient resistance (approx. 6 fQ) that it could be heated by
means of a DC-stabilized current, while at the same time allowing the bottom of the recording
chamber to remain translucent (approx. 50 % light transmittance). Before each experiment current
was adjusted until the desired temperature was obtained and the temperature was monitored
continuously with a thermistor probe.
Electrophysiological measurements
Intracellular recordings were made using a modified suction microelectrode technique.
Microelectrodes were pulled from 1 mm o.d. Pyrex glass having either a star-shaped (Radnotti
Glass Technology, Monrovia, CA, USA) or square-shaped (Glass Company of America,
Bargaintown, NJ, USA) lumen using a Kopf microelectrode puller from which the magnetic core
was removed. This modification resulted in a constant pulling force of about 270 g. A heating coil,
made from 0-5 mm nickel-chrome wire (i.d. of the coil 2-1 mm, 3-5 windings), replaced the original
heating coil. A heating current of 6 A was used.
Microelectrode shanks were drawn in a single pull and were used immediately after being back-
filled, without being fire-polished. The microelectrodes were filled with either 150 mM-KCl or
150 mM-potassium gluconate with 10 mM-KCl. Ten millimolar HEPES was added to both of these
'filling' solutions and pH was adjusted to 7-4 with KOH. All filling solutions were filtered before
use (0-2 #um pore size). The microelectrode DC resistance varied between 5 and 20 MCI. These filling
solutions were used interchangeably; no consistent differences could be observed in either
spontaneous activity or in the ionic currents.
The microelectrodes were mounted in an air-tight Plexiglass holder having a suction port on the
side, and a chlorided silver wire was used to make contact with the electrode solution. Positive
pressure was applied to the electrode lumen, until the electrode was placed very near a selected cell.
 SINO-ATRIAL NODE PACEMAKING 61
When the tip touched the surface of the cell, negative pressure was applied and the formation of
a 'giga-ohm seal' was monitored by the voltage response to 10 pA pulses. After a giga-ohm seal was
obtained, an extra suction pulse usually caused the patch membrane to rupture, as signified by a
sudden change in the measured potential, and by the appearance of spontaneous pacemaker
activity.
Only cells which exhibited regular, spontaneous pacemaker activity were used in these voltage
clamp experiments. In these preparations, although the size and kinetics of the hyperpolarization-
activated current, I, were somewhat variable (± 30 %). I, was always present and did not decrease
significantly during a typical experiment (20-30 min).
Electronics
A voltage clamp circuit similar to that first described by Hamill et al. (1981) was used (see Giles
& van Ginneken, 1985). Command potentials for voltage clamping were obtained from a custom-
made programmable stimulator (Hans van der Salm, University of Utrecht). In the current clamp
mode signals from this stimulator was used as stimuli.
Data handling
Membrane potential and current records as well as trigger pulses were recorded on a FM tape-
recorder (Gould 6500) at a bandwidth of DC to 5 kHz. After the experiment, the data were
digitized using an A-D board (DT2801A) in a Compaq microcomputer (Robinson & Giles, 1986)
and were either stored on floppy discs or sent to a VAX 11-750 minicomputer (Digital Equipment
Corporation). Multiple exponential curve fitting was done using the algorithms developed by
Provencher (1976). In some cases current traces were averaged before further analysis. Raw data
and/or analysed results were displayed on a digital plotter (Hewlett-Packard Ltd, 7470A).

RESULTS
Morphology of different cell types from S-A node
In a typical suspension several different cell types were observed (Fig. 1). Only a
small number of these cells exhibited regular spontaneous beating. Relatively large
round cells, frequently observed in the cell suspension (Type II in Fig. 1), did not
exhibit spontaneous electrical activity. Since the S-A region is quite densely
innervated (Roberts et al. 1989), these round (diameter 20 ,um) cells could be nerve
or endothelial cells. We were not able to impale the very small cells (diameter 5-8 jim)
which were previously identified (Masson-Pevet, Gros & Besselsen, 1980; Masson-
Pevet et al. 1982) as the primary pacemaker cells (type 1, Fig. 1). These cells seemed
weak structurally and were usually 'sucked up' into the electrodes before a gigaohm
seal was formed.
In this series of experiments, we have considered only the spontaneously active
type III and IV cells (Fig. 1) to represent the S-A node pacemaker cells and have
used these cell types to study If. The average dimensions and the input capacitance
values for these cells are given in Table 1. Cell capacitance was measured from the
initial slope of the transmembrane voltage change in response to small applied
currents. Thus, the values for the membrane capacitance per unit area (3 9 ,sF/cm2)
are only an approximation, since these cells have an irregular shape, and infoldings
and/or caveolae are abundant on the sarcolemma (Masson-Pevet et al. 1980). The
average surface area was calculated assuming that the cells were cylindrical
structures with closed ends.
62 A. C. G. VAN GINNEKEN AND W. GILES

50 im

I 11 III IV
Fig. 1. Phase-contrast micrograph of the different cell types observed in cell suspensions
following enzymatic dispersion of the rabbit S-A node. Type I cells are round
(approximately 5 ,um in diameter) with many finger-like protrusions. They resemble the
primary pacemaker cells described by Masson-Pevet et al. (1982). Type II are spherical,
and do not exhibit either visible contractions or spontaneous electrophysiological activity
(see text for further explanation). Type III and IV cells consistently show both regular
contractions and spontaneous electrical activity. These cells were used in these
experiments.

30 mV '\1\

100 Ms

Fig. 2. Spontaneous electrical activity in an isolated rabbit S-A node cell, recorded using
a suction microelectrode technique. The beating rate, action potential duration and
diastolic potential are comparable to those recorded from the intact S-A node. The
horizontal line represents 0 mV. For further explanation see text.

TABLE 1. Dimensions of isolated cells from rabbit sino-atrial node

Mean S.E.M. n
Length, I (/tm) 32-16 2-2 15
Diameter, d (,um) 11-77 0-77 15
Volume* (10-12 1) 3-13 0-35 15
Surface areat (10-5 cm2) 1-38 010 15
Input capacitance (pF) 55-1 4-7 18
*
(Hl d2(Id/3))/4.
t Assumes a right cylindrical shape.

Electrophysiology
Figure 2 shows typical spontaneous activity of a single cell from the rabbit S-A
node, recorded using a single suction microelectrode. The features of this record are
very similar to recordings from the multicellular sinus node preparations (e.g.
Shibata, Giles & Pollack, 1985). Electrophysiological parameters characterizing the
 SINO-ATRIAL NODE PACEMAKING 63
spontaneous activity of these cells are summarized in Table 2. The relatively large
amplitude of the action potentials may suggest that these cells originate from the
transitional region (Bleeker et al. 1980). Alternatively, recording from these cells with
a technique which provides a very high seal resistance may be the explanation for the
larger-than-expected size of the maximum diastolic potential and action potential
(see Discussion).
Voltage clamp data: general features
In these voltage clamp experiments the holding potential (HP) was usually set at
-30 mV, the potential at which the net holding current was very near zero. Figure
3A shows superimposed current traces elicited by voltage clamp steps of increasing
amplitude with depolarizing (top) and hyperpolarizing (bottom) polarities.
In response to depolarizing steps, a TTX (10-5 M)-resistant inward current was
elicited. It was blocked by 0-5-0{1 mM-CdCl2, and was abolished when [Ca21]0 was
reduced from 2-2 to 0-2 mm (by replacing CaCl2 with equimolar MgCl2). These features
of this current are very similar to those of the Ca2+ current, ICa' in other cardiac
tissues. The I-V relationship for ICa plotted as maximum inward or minimum
outward current (Fig. 3B, 0) reaches its maximum near 0 mV. Since the major aim
of this work was to study If, no attempt was made to identify and characterize T
and/or L types of Ca2+ currents (cf. Hagiwara & Irisawa, 1989). During relatively
long depolarizations the inactivation of 'Ca is immediately followed by activation of
an outward current similar to the time- and voltage-dependent delayed rectifier K+
current, IK' previously described in single cells from rabbit S-A node (Nakayama et
al. 1984; Shibasaki, 1987). Activation of 'K results in a net outward current at the
end of the relatively long depolarizing steps (Fig. 3B, 0).
When hyperpolarizing voltage steps were applied, the size of the current change,
measured immediately after the surge of the capacitative current is small, confirming
that the inward rectifying current, IK1' was either very small or absent in these cells
(Noma, Nakayama, Kurachi & Irisawa, 1984; Fig. 2B, 0). During relatively long
hyperpolarizations, a time- and voltage-dependent current was activated (Fig. 3,
*). This current resembles the If current which is present in Purkinje fibres, in man-
made strips of S-A node tissue, as well as in single cells from rabbit A-V node and 5-
A node (cf. Irisawa et al. 1991).
Is If a 'gated' transmembrane ionic current?
Our main goal was to obtain quantitative measurements of the magnitude and
kinetics of If in isolated cells which exhibited spontaneous pacemaker activity.
Initially, it was important to determine whether If was present consistently and to
demonstrate that the time- and voltage-dependent changes in current which
characterize If are not altered by artifacts arising for example from time-dependent
changes due to accumulation or depletion of extracellular or intracellular ions. This
may seem unlikely since all of the experiments were carried out using single cells.
However, even in single cells this possibility cannot a priori be ruled out. If there is
a substantial Na+ influx into these single cells when If is activated, intracellular Na+
could rise significantly. Furthermore, caveolae are present in the sarcolemma of these
isolated cells (Masson-Pevet et al. 1980). Ion fluxes into these caveolae may produce
64 A. C. G. VAN GINNEKEN AND W. GILES

400 pA

s100
mS

60 mV

Fig. 3. Ionic currents in the rabbit S-A node cell. A, superimposed current traces recorded
during depolarizing steps to -20, 0 and + 20 mV (upper traces) and hyperpolarizing steps
to -60, -70, -90 and -100 mV (lower traces). During depolarizing steps from a holding
potential (HP) of -40 mV a TTX-resistant inward current was recorded. This current
was immediately followed by development of an outward current, presumably carried by
K+. In response to small hyperpolarizing steps (bottom four traces) relatively small
currents were measured at short times. During longer hyperpolarizations a time- and
voltage-dependent inward current was activated. B, I-V relations of these currents.
Current elicited by depolarizing steps was measured at its peak inward value; currents
elicited by hyperpolarizing steps were measured immediately after the decay of the
capacity current (0). Current measured at 300 ms after the onset of the voltage steps was
also measured and plotted as closed symbols (@). TTX (10-6 M) was added.

TABLE 2. Parameters of the action potential in isolated S-A node cells

Mean S.E.M. n
Action potential duration (ms) 72-7 3-3 20
Interbeat interval (ms) 335-4 15.0 20
Maximum diastolic potential (mV) -57.9 1*8 20
Action potential amplitude (mV) 89-8 2-1 20
Diastolic depolarization rate (mV s-1) 77.3 4.4 20
Upstroke velocity (V s-') 12-8 1-65 19
 SINO-ATRIAL NODE PACEMAKING 65

. .,.. .. ;;

300 pA
N.
\. ~~1000 ms

Fig. 4. Measurement of conductance changes during activation and deactivation of If by

application of short (20 ms, 10 mV) depolarizing potential steps, superimposed on a clamp
step from HP -40 to -90 mV which activated If. The upper panel shows the pulse
protocol. The lower panel shows the current traces. Note that conductance (measured by
the stepwise change in current during the small voltage steps) increases severalfold during
activation of If and that when If decays the conductance decreases again to its original
value.

-10 mV
A -20_,.
8\-30

-10
B --20\

500 pA
100 ms
Fig. 5. Test for changes in reversal potential of If due to accumulation-depletion
phenomena. A, If was activated by a hyperpolarization to - 130 mV for 35 ms; thereafter
the potential was clamped at - 10, -20 and -30 mV. Note that the current tails reverse
at approximately -20 mV. B, similar voltage clamp protocol, but with a Pj duration of
400 ms. Although If is much larger the reversal of the current tails remain near -20 mV,
indicating that the If current changes are not contaminated by accumulation-depletion.

enough accumulation or depletion to change concentration gradients, as well as the

reversal potentials and I-V curves of transmembrane ionic currents (e.g. K+). This
may lead to distortion of the amplitude and time courses of these currents; or in
extreme cases, to the 'appearance' of relatively slow time-dependent current
changes. Nakayama et al. (1984) and DiFrancesco et al. (1986) have drawn attention
to the considerable variability in the voltage dependence of If. It is possible that this
could have been due to accumulation-depletion phenomena.
Figure 4 shows an experiment which tests whether the time- and voltage-
dependent changes in If are caused by measurable, reproducible conductance
changes. A hyperpolarization from the holding potential (-30 mV) to -90 mV was
PHY 434
66 A. C G. VAN GINNEKEN ANVD IT' GILES
applied in order to activate If. The changes in current elicited by these small voltage
displacements reflect the relative conductance of the cell. Results from these
experiments showed consistently that during the onset of the slow inward current,
the membrane conductance increases, and also that when the potential returned to
the holding potential the conductance declines back to its resting level as the current
decreases.

-1-1

1 Pi -1

600 pA

40

200

0
-80 mV
-200

B
-400

-600

-800

L-1 000
Fig. 6. Current-voltage relationship for If in an isolated cell from the rabbit S-A node.
Panel A illustrates the protocol and raw data. One second hyperpolarizations from HP
-40 to -90 mV were used to activate If. Current tails were measured immediately after
a subsequent depolarizing step (P2) to various potentials. The initial amplitude of the
current tail was plotted versus the P2 voltage. The I-V curve in panel B shows that the
current reverses near -30 mV and that the I-V relation is approximately linear. Steady-
state activation of If was not reached in this experiment. Calcium current was blocked by
0 5 mM-CdCl2.
 SINO-A TRIAL NODE PACEMAKING 67
These findings demonstrate that the activation and decay of If are mediated by
changes in a membrane conductance. However, it is still possible that shifts in its
reversal potential are taking place during the clamp step. This possibility was
assessed in protocols such as the one shown in Fig. 5. A two-pulse procedure was used
to measure the reversal potential of If. The duration of the P1 pulse (which elicited
If) was adjusted so that different amounts of currents were activated during the
shorter (35 ms; panel A) and longer P1 pulses (390 ms; panel B). The data in panel B
show that there is no significant change in the reversal potential, even though the
amount of charge transfer during activation differs by more than 10-fold in these two
cases. It is therefore very unlikely that ion accumulation or depletion phenomena
significantly modulate the time and voltage dependence or magnitude of If.
I-V relation of r in iisolated S-A node cells
Since the inwardly rectifying K+ current IK1' is very small or absent in cells from
the S-A node (Noma et al. 1984; Irisawa, 1984), the instantaneous 1-V relation of I
can be obtained without interference from other currents. Typical raw I-V data
(panel A) and an instantaneous I-V relationship (panel B) are shown in Fig. 6. These
results confirmed that If is a slowly activating anward current. In principle, therefore
If could contribute to the inward depolarizing current which generates the
spontaneous diastolic depolarization or pacemaker potential. The I-V relationship
for If is linear over a large range of potentials: the average fully activated
conductance is 12 0 + 0-88 x 10-9 S (S.E.M., n = 6). The reversal potential for If was
-24 + 1-5 mV (S.E.M., n = 6) in normal Tyrode solution.
Time dependence and steady-state voltage dependence of If
The results in Figs 4 and 5 demonstrate that there is a substantial conductance
increase during activation of If and that no significant change in the reversal
potential for If occurs, even during relatively long (e.g. 10 s) hyperpolarizations.
Interpretation of steady-state voltage dependence of If requires knowledge of
whether or not the time- and voltage-dependent change in If can be accounted for
entirely by these conductance changes. Evidence for this was obtained from an
envelope of tails' protocol, as shown in Fig. 7.
Voltage steps to -110 mV were applied for various durations. After the return to
-60 mV, inward current tails were recorded. Figure 7B compares the time courses
of the envelope of tails with the onset of the current. This was constructed by scaling
the current tails by a constant such that the initial values of the current tails were
very similar to the amplitude of the activating current. It is apparent that the
envelope of tails matches very closely the time course of the onset of If. As a result,
a Hodgkin-Huxley mathematical formulation was used to describe the time and
voltage dependence of If.
Since If does not inactivate, measurement of the voltage dependence of its steady-
state activation is quite straightforward. In these experiments, hyperpolarizing
voltage steps to various potentials (P1) were applied for 3 s; thereafter a second step
(P2) to a fixed potential (-40 mV) was made. The current tails, recorded immediately
after return to P2, reflect the deactivation of If. The initial amplitudes of these tails
can be used to construct an activation curve for If (Fig. 8A). The relatively small
3-2
68 A. C. G. VAN GINNEKEN AND W. GILES
variability in the voltage dependence of If is demonstrated by the standard error
bars. This aspect of our results contrasts those reported in previous studies on these
single cells (Nakayama et al. 1984; DiFrancesco et al. 1986; Hagiwara & Irisawa,
1989).

A -30 mV 60
-110

200 pA
250 ms

300 pA
123 pA
100 ms

Fig. 7. An envelope-of-tails test for the I, current. A, the superimposed current onset
records were elicited by hyperpolarizations from -30 to -1 O mV. The envelope of tails
was recorded by changing the duration of the hyperpolarization, and then returning to
-60 mV. Each decay tail was fitted with a single-exponential function. As shown in panel
B the envelope of the tails is very similar to the time course of the onset of I,. B, in this
panel the tail amplitudes have been scaled by a constant factor in order to match the
amplitude of the onset of I,. After sufficient current is activated to yield a resolvable tail,
the envelope of tails appears to match reasonably well with the time course of the onset.

Note, however, that neither the activation kinetics nor the deactivation kinetics
are well-described by a single exponential function. The activation phase begins after
a 'lag' (Figs 3 and 5), giving it a sigmoid appearance. The deactivation process may
also exhibit a somewhat complex diphasic time course. In mammalian cardiac
Purkinje fibres the kinetics of I, have been analysed in some detail and are complex
(DiFrancesco, 1985). In contrast, in both multicellular preparations (Yanagihara &
Irisawa, 1980) and in single cells from rabbit S-A node (Nakayama et al. 1984) the
kinetics of If have been described as a single exponential function.
 SINO-ATRIAL NODE PACEMAKING 69
Although developing a detailed kinetic model for I, in single pacemaker cells was
not the primary purpose of this study, we considered a number of ways of accounting
for the somewhat complex kinetic behaviour of If. A second-order Hodgkin-Huxley
scheme provided satisfactory fits to most of the current onsets and decay tails. This

A P, P2

- -==-m-wVww

200 pA

500 ms

B *1*0 a
-0
E

.
z
0 Z
-160 -120 -80 -40 0
Pi (mV)
Fig. 8. Steady-state voltage dependence of I, in rabbit S-A node cells. A, upper traces show
the voltage clamp protocol. Hyperpolarizing steps to various potentials (P,) for 3 s were
applied to activate I. Thereafter the potential was returned to holding potential
(-40 mV). Lower traces, superimposed current traces, starting 2 s after onset of P,. Note
that maximum tail amplitudes are recorded at approx. - 10 mV. B, the initial amplitude
of the tail was measured at each P1 level; these values were then normalized to the
maximum tail size and plotted as an activation curve. In the activation curves shown the
data have been plotted in two different ways: uncorrected (v) or corrected for the
observed sigmoidicity (by obtaining the square root, refer to text) shown as open symbols
(0). The data points and error bars denote mean+s.E.M. (n = 7).
implies that the initial sizes of the decay tails represent the values of the activation
parameter, q, raised to the power 2. The square roots of the initial tail amplitudes
were therefore measured and plotted versus membrane potential (Vm, Fig. 8, 0) to
obtain the steady-state activation curve. These data were well-fitted by a Boltzmann
distribution described by:

q(v, t) --
I + exp [(V,,, + 64-0)/ 13-5]
(1)
70 A. C G. VAN GNINNEKEN AND IY GILES
For comparison with previously published data, both the original (raw) data points
(a), and the 'corrected' activation data assuming a second-order process (0), are
shown in Fig. 8. The activation curve obtained using the second-order kinetic scheme
lies some 10 mV positive to the raw data. Obviously, the choice of the kinetic scheme
for If can be very important in deducing its functional role (see Discussion).

- 20

-15

10 I

V
-160 -120 -80 -40 0
Membrane potential (mV)
Fig. 9. Kinetics of activation and deactivation of If. The time courses of onsets and current
tails were fitted by a function consisting of an exponential raised to the power 2. The rate
constant of the slowest and largest component of the onset was plotted as closed symbols
(@); the rate constant of decay of the tails multiplied by 2 was plotted as open symbols
(0). These results are the means of four to seven experiments. Bars denote+ S.E.M. The
curve was fitted by the sum of a positive and negative exponential functions reflecting the
forward and backward rate constants af and /f respectively (dashed lines, see text).

Time dependence of If
Figures 3 and 4 shows that there is a 'lag' or delay in the onset of If. As indicated
previously, to account for this we have used the previously described modified
Hodgkin-Huxley (1952) scheme with a voltage-dependent activation parameter
raised to a power of 2 to model this current change (cf. Bezanilla, 1985; DiFrancesco,
1985). The onset of the current can be fitted using the expression:
If (t) = [1 - exp (-t/Tr)]2. (2)
where If is the hyperpolarization-activated current, and Tq is the time constant of its
gating variable, q. In a small number of cases, the current tails near the foot of the
activation curve could be adequately fitted using a single-exponential function.
However, when the preceding hyperpolarization was relatively large, there was a
distinct delay before they began to deactivate (see Fig. 7 and inset to Fig. 8).
 SINO-A TRIAL NODE PACEMAKING 71
It is possible that the more complex kinetic scheme proposed by DiFrancesco
(1985) could provide a better fit; none the less we chose to model the current with
simple second-order kinetics, since moderate and short hyperpolarizations (as occur
during normal spontaneous activity) do not result in this diphasic time course or a

200 ms

Fig. 10. Estimation of the size and time course of If during spontaneous pacing in rabbit
S-A node. I was calculated as a function of time from values for maximum conductance,
reversal potential and the rate constants which were obtained from our experimental
data. Pre-recorded action potentials were used as command potentials in a simulated
voltage clamp experiment. A, the action potential and pacemaker potential. B,
superimposed current traces. The noisy trace reflects net ionic current: -Cm dV/dt. The
value of dV/dt was obtained from the action potential; Cm equals the average input
capacitance of the cells, 55-1 pF. The second trace (thick continuous line) is the calculated
If during the potential displacements which correspond to the action potential and
pacemaker potential (A). See text for further description.

measurable 'delay' in deactivation kinetics of the current tail. Adopting this

'second-order' approach, the rate of decay of tails is described by:
If(t), tail = exp (- 2t/rq). (3)
Thus, the time constant with which the activation variable (q) changes during the
tails equals half the time constant measured directly from the tail currents. In Fig.
9 the rate constants of activation are plotted versus membrane potential after the
delay in onset was accounted for (0). These data points were then fitted by the sum
72 A. C. G. VAN GINNEKEN AND W. GILES
of negative and positive exponential functions, representing the voltage-dependent
rate constants aCf and /f respectively. According to the Hodgkin and Huxley
formalism:
Tf= 1
+ (4)

since, in the Hodgkin-Huxley formalism, the steady-state values of the activation

variable are described by:
f= 1 (5)

The ratio of /Jf/af was determined from the best-fitting Boltzmann equation (eqn (1))
and this ratio was used to restrict the number of possible combinations of af and /f3.
The best-fitting combination of af and flf was found to be:
af = exp [- 00220741( Vm + 386 9)] ms-1, (6)
and Af = expL-0-052(Vm-7308)]ms-'. (7)
Estimation of the changes in If during the pacemaker potential
Our data describing the ion transfer and kinetic characteristics of If can provide
useful information about the role of If in the development of the pacemaker
potential. By combining calculations and experimental measurements of the size and
time course of If, estimates concerning the size and direction of this current change
during the pacemaker potential can be obtained. Assuming that the cell is uniformly
polarized, the following relationship is applicable:
Ii = -CmdV/dt. (8)
where Ii is the net ionic current. The values for dV/dt can be obtained directly from
a typical pacemaker potential recording and the membrane capacitance (Cm, 55 pF)
has been measured in the single cells (Table 1). Panel A of Fig. 10 shows typical
spontaneous activity recorded from the type III and IV cells. The first derivative of
the action potential is shown in panel B. The continuous line in panel B was
calculated from the experimental information concerning the ion transfer and kinetic
behaviour using the following relationship:
If = q2qf(E -Ef), (9)
where q is the activation variable, 9f is the fully activated conductance, E is
membrane potential, and Ef is the reversal potential for If.
These findings suggest that the changes in If which we have identified could
produce a significant inward current in the range of potentials in which the
pacemaker depolarization occurs in these cells. However, this calculation is not
completely conclusive since: (1) some variability in the size and time course of If as
well as the maximum diastolic potential and the spontaneous firing rate is observed;
and in addition, (2) the deactivation of IK and the activation of ICa are also
importantly involved in controlling the pacemaker potential (cf. Irisawa et al. 1991).
We therefore attempted to obtain additional, independent evidence concerning
 SINO-ATRIAL NODE PACEMAKING 73
whether changes in If are functionally important in the development of the
pacemaker potential.
In this second experimental approach, we blocked If and recorded the resulting
changes in the pattern of spontaneous activity. If (i) If can be blocked selectively,

Control

Fig. 11. Effects of 1 mM-CsCl on spontaneous activity and membrane currents in rabbit
S-A node. A, control spontaneous activity (top) recorded from a single S-A node cell
superimposed on voltage steps applied during a voltage clamp experiment on the same
cell. The holding potential was set to the zero net current, and hyperpolarizing clamp
steps were applied in 10 mV increments. Bottom, transmembrane currents showing the
presence of If at hyperpolarized potentials. B, analogous spontaneous activity and ionic
currents in the presence of 1 mM-CsCl. Note that the rate of spontaneous discharge
decreased after Cs2" treatment and that If was almost completely blocked. The same
voltage clamp protocol as in the control was used.

and (ii) If does change as suggested by the data in Fig. 10, then after If is blocked there
should be a hyperpolarization of the maximum diastolic potential and a significant
slowing of heart rate.
Figure 11 shows an experiment in which the spontaneous pacemaker activity was
recorded, and then voltage clamp measurements were made on the same cell, in the
absence and presence of 1 mM-CsCl, to block If. Panel A (upper traces) shows the
control spontaneous activity superimposed on the voltage records used during
the subsequent voltage clamp protocols. Changes in If during these hyperpolarizing
voltage steps are shown in the lower traces. Note that changes in If could be recorded
74 A. C. G. VAN GINNEKEN AND W4' GILES
in the range of potentials corresponding to the pacemaker depolarization. In Fig. 12,
*the upper traces show spontaneous activity measured in the same cell after
application of 1 mM-CsCl. The spontaneous activity slowed significantly, and the
current records (lower traces) showed that 1 mM-CsCl strongly inhibited If. In other

A Control Cs+

0
C
-200
-400
-600
-800
Control -1000
-1200
-1400 pA
Fig. 12. Detailed analysis of the effect of 1 mM-CsCl on spontaneous activity and
membrane current(s) in rabbit S-A node cells. A, current clamp experiments showing the
effect on maximum diastolic potential and heart rate of an applied hyperpolarizing
current in control condition (left) and after 1 mM-Cs' was applied (right), compared to the
spontaneous activity without this hyperpolarizing current. Note that the effect of the
hyperpolarizing current is much larger when Cs' is present. B, current traces recorded
during hyperpolarizing clamp steps illustrated CsCl block of If. Panel C shows currents
measured at the end of 3 s hyperpolarizing steps plotted versus membrane potential, in
control condition (0) and in the presence of 1 mM-CsCl (@). Note that the effect of Cs+
is larger at more negative potentials, indicating that the block of I, by Cs+ is voltage
dependent.

experiments of this kind, the addition of 1 mM-CsCl did not always result in this
slowing of spontaneous activity. However, in these cases, when a current was applied
to hyperpolarize the maximum diastolic potential by approximately 10 mV, a
 SkINO-ATRIAL NOVODE PACEMAKING 75
significant slowing of the spontaneous rate was observed, and subsequent application
of 1 mM-CsCl completely abolished the spontaneous activity (see Discussion).
Figure 12 shows results from this type of CsCl experiment. In panel A spontaneous
activity was recorded from a Type III cell, and then the current clamp was used to
hyperpolarize the cell by approximately 10 mV. This resulted in a substantial
slowing of the pacing rate, and when CsCl (1 mm) was applied, the spontaneous
activity was abolished completely. The voltage clamp measurements from this
experiment showed that CsCl (1 mM) inhibited If (Fig. 12B and C). The
current-voltage relationships in Fig. 12 C confirm that the Cs'-induced inhibition of
If is strongly voltage dependent (cf. DiFrancesco, 1985).

DISCUSSION
Single cell preparation
Our enzymatic dispersion procedure consistently yields viable, spontaneously
active single cells from the pacemaker region of the rabbit S-A node. In agreement
with most previous anatomical and electrophysiological work, the cells which are
obtained are not uniform in size, shape or appearance (cf. Denyer & Brown, 1990a).
Our original objective was to perform electrophysiological and voltage clamp
experiments only on cells from the primary region of the anatomical S-A node.
However, the very small spherical cells which originate from this area (Mason-Pe'vet
et al. 1982) could not be impaled using our suction microelectrode methodology. For
that reason we studied the larger cells (types III and IV shown in Fig. 1). These cells
have the complement of transmembrane ionic currents which one would expect to
find in cardiac pacemaker tissue:ICa, a transient inward calcium current; I., a time-
and voltage-dependent delayed rectifier potassium current; and If, a time- and
voltage-dependent inward current which is activated upon hyperpolarization. The
ability to record normal action potentials and pacemaker potentials and the
consistent presence of each of these currents indicate that these cells are normal and
healthy. Moreover, the absence of both IK, the inwardly rectifying K+ current, and
It, the transient outward current, are consistent with most previous work on rabbit
S-A node (cf. Irisawa et al. 1991).
A number of other observations also suggest that these cells are not damaged
significantly. (1) Bleeker et al. (1980) have pointed out that pacemaker cells from the
rabbit S-A node are rather 'empty'; i.e. intracellular structures such as well-
organized myofibrils and/or cytoskeletal elements could be identified only in S-A
node pacemaker cells. Hence, the tendency for our cells to 're-organize' into spherical
shapes when not supported by an extracellular matrix may be due to a fundamental
property of their microanatomy. (2) In the solutions containing S-A node cells, often
there are single cells from the crista terminalis of the rabbit heart (cf. Giles & van
Ginneken, 1985). These crista terminalis cells do maintain their original rod-shaped
structure, indicating that the dissociation procedure per se does not cause all viable
cells to round-up and/or exhibit a calcium paradox phenomenon. (3) The beating
rate and the action potential characteristics of the isolated cells were comparable to
those recorded using conventional microelectrode techniques in the rabbit S-A node
(e.g. Shibata et al. 1985). Nevertheless, since we failed to record from the small
76 A. C. G. VAN GINNEKEN AND W. GILES
spherical cells (which are thought to be the primary pacemaker cells) we cannot be
certain whether the spontaneous electrical activity of the cells used in this study does
or does not represent latent or secondary, as opposed to primary, pacemaker activity
(Bleeker et al. 1980; Kreitner, 1985).
It is of interest to compare morphology of these pacemaker cells with those
obtained using previously reported techniques for isolation of single cells from either
the rabbit S-A node or A-V node (Taniguchi, Kokobun, Noma & Irisawa, 1982;
Nakayama et al. 1984; DiFrancesco et al. 1986). As shown in Table 1, the average
dimensions of the cells (types III and IV) used here were: length, 32 /tm; diameter,
11 m. In contrast the isolated cells used by Nakayama et al. were spheroids of the
following dimensions: (S-A node cells, 24 x 28 ,um; A-V node cells, 22 x 31 ,um) and
those studied by Denyer & Brown (1990 a) were approximately 95 ,um in length and
7 ,im in width. Our isolated cells, like those of Denyer & Brown (1990a), had an
extremely high input resistance (1-2 Q) and a relatively small total cell capacitance
(50-70 pF). These features, plus the presence of only relatively small and slow
transmembrane ionic currents, strongly suggest that space clamp problems will be
minimal in these experiments. The apparently very high specific capacitance of
these cells (approximately 4 jF/cm2) may arise from the extremely convoluted
sarcolemmal contour. Electron micrographs show that many caveolae are present
(Masson-Pevet et al. 1980). As indicated in the Results, these caveolae may act as
significant barriers to ion diffusion even in populations of isolated single cells (see
Scheid & Fay, 1980).
It is important to recall that Kreitner (1985) has reported that the subsidiary
pacemaker cells (plateau fibres or fl-cells in her terminology) exhibited action
potentials which were somewhat sensitive to tetrodotoxin (TTX). The majority of
the S-A node cells used by Denyer & Brown (1990 a) also exhibited some TTX
sensitivity. In contrast, the spontaneous activity of our isolated cells was not
inhibited by TTX (10-6 M).
The presence of the hyperpolarization-activated inward current: If
An important difference between our results and some of those from previous
studies on single S-A node cells is apparent in a comparison of the ability to
consistently record the current If. Nakayama et al. (1984) observed I, in only
approximately one-half of their cells. In contrast, in our experiments, If was present
in every one of the type III or type IV cells which was successfully impaled. Nathan
(1986) and Denyer & Brown (1990 a) have also consistently recorded If in cells from
rabbit S-A node. It is possible that in the study of Nakayama et al., cells from the
crista terminalis were also included. Previous work (Giles & van Ginneken, 1985;
Nathan, 1986) shows that while a variety of cell types from the S-A node consistently
exhibit If, those from the crista terminalis do not. This may also explain why
DiFrancesco et al. (1986) failed to find If in the 'larger' type of the cells which they
isolated from rabbit heart.
The ability to record I, consistently in isolated pacemaker cells is an essential pre-
requisite for quantitative studies of this current aimed at determining its functional
role. Until recently (cf. DiFrancesco, 1987; Irisawa et al. 1990), in spite of a
substantial amount of previous experimental investigation and computer modelling
(cf. DiFrancesco, 1985; DiFrancesco & Noble, 1989), the evidence that If was an
 SINO-ATRIAL NODE PACEMAKING 77
important pacemaker current remained indirect and in many cases was only
circumstantial, since it was based upon correlations. The absence of If in some
'pacemaker cells' from rabbit S-A or A-V node (Nakayama et al. 1984), in bull-frog
sinus venosus cells (Giles & Shibata, 1985; Shibata & Giles, 1985; but see Bois &
Lenfant, 1988) and in spontaneously active rabbit Purkinje fibres (Carmeliet et al.
1982) rules out the suggestion that I, is the only pacemaker current in heart. In
addition, the ability to record If-like current changes in multicellular trabeculae but
not in single cells in bull-frog sinus venosus (compare Brown et al. 1977 with Giles &
Shibata, 1985) and human atrium (compare van Bogaert & Carmeliet, 1985 with
Shibata, Drury, Refsum, Aldrete & Giles, 1989) could be taken as evidence that If is
not generated solely by a conductance change, but instead is 'contaminated' by
accumulation-depletion of Na+ inside the cell, and/or K+ in the extracellular space
(Maylie, Morad & Weiss, 1981; Maylie & Morad, 1984). However, our results
demonstrate that If behaves in a way expected of a time- and voltage-dependent
conductance, as opposed to arising from a shift in driving force due to ion
accumulation or depletion. We have shown that: (1) the activation and deactivation
of the f-current are accompanied by increases and decreases of membrane
conductance respectively (Fig. 4); (2) the envelope of the tail currents also closely
matches the time course of the onset of the current (see Fig. 8), an important
criterion in determining whether an ionic current change is likely to obey
Hodgkin-Huxley kinetics; (3) the voltage dependence of the steady-state activation
is well-described by a Boltzmann distribution (see Fig. 9); (4) the reversal potential
does not change significantly even in response to very large and long If currents (Fig.
5). The strongest evidence that If is a conventional, channel-mediated time- and
voltage-dependent ionic current comes from the work of DiFrancesco (1986) which
identified the single-channel events that generate this current.
The absence of a shift in the reversal potential is rather surprising when one
considers the relatively large size of this current with respect to the rather small cell
volumes. Using the average fully activated conductance of a cell (12 nS) and the
average value of the reversal potential (-24 mV), we calculated that when the
current is fully activated at - 120 mV, about 1-2 x 10-14 moles of positive charge
enters the cell per second. Assuming that at this voltage most of this charge is carried
by Na+, and that Na+ ions rapidly and uniformly distribute throughout the average
cell volume (3-13 x 1012 1), an increase in [Na+]i concentration of about 3-9 mm per
second is predicted. Recently, Na+-sensitive microelectrode measurements have been
made in sheep Purkinje fibre tissue during activation of If. These data provide direct
evidence for a substantial Na+ influx and change in intracellular Na+ during
activation of If (Glitsch, Pusch & Verdonck, 1986). However, our measurements
show that these changes in Na+ may result in only very small changes in the reversal
potential for If. Assuming that the If channels are equally permeable to Na+ and K+
the shift in reversal potential would be expected to be only 0-6 mV. This small shift
could not be detected by the electrophysiological methods used in this study (Fig. 5).
However, it is important to note that this change in [Na+]i (- 4-8 mm) could be
significant for other reasons. For example, it could produce a significant stimulation
of the electrogenic Na+-K+ pump (cf. Gadsby, 1984; Nakao & Gadsby, 1989).
78 A. C. G. VAN GINNEKEN AND l'V GILES

Voltage dependence and ion selectivity of If

We did not attempt to study in detail the selectivity of If. The observed reversal
potential, approximately -25 mV. is very similar to previous findings in multi-
cellular. Purkinje fibres and man-made sino-atrial node strips (cf. DiFrancesco, 1985)
as well as in enzymatically isolated single cells (DiFrancesco et al. 1986). This reversal
potential value indicates that If is carried by both sodium and potassium ions and
that the permeability for these two ions may be approximately equal. Recently.
Glitsch et al. (1986) have evaluated this using Na'-selective electrodes in voltage-
clamped sheep Purkinje fibres. Their data show that the permeability for K+ is about
10-20 times that for Na+. At present, there is very little detailed information
regarding the selectivity of If for anions. A significant chloride contribution to If has
been reported in Purkinje fibres and human atrium (van Bogaert & Carmeliet, 1985)
as well as in rabbit S-A node (Seyama, 1979; Yanagihara & Irisawa, 1980).
To determine the functional role of If in cardiac pacemaker tissue it is essential to
obtain quantitative descriptions of its kinetics in the ranges of potential
corresponding to the pacemaker depolarization and the action potential. As noted in
the Results, both the onset of this current and the decay tails have somewhat sigmoid
time courses. DiFrancesco (1985) has chosen a complex kinetic scheme involving
three closed and five open states to describe his data. In the absence of more
extensive and quantitative data which require this level of complexity, we have
somewhat arbitrarily chosen the most simple model which provided an acceptable fit
to the data. This second-order model is adequate in the ranges of potentials of most
interest; i.e. when only small hyperpolarizations are used (-60 to -90 mV").
However, the sigmoidicity observed during the tail currents may not arise entirely
from voltage-dependent gating phenomena. Instead it could be due to an ion-ion
interaction phenomenon; e.g. an intracellular ion could block the channel after a
large hyperpolarization. This type of effect has been observed when for example Ba2"
blocks the delayed rectifier K+ current in squid axon (Eaton & Brodwick. 1980; cf.
Armstrong, 1990). Alternatively, the complex kinetics may arise from the recently
described dependence of If on intracellular Ca2+ levels (Hagiwara & Irisawa, 1989).

The functional role of If in sino-atrial node pacemaker depolarization

The steady-state voltage dependence of If in these single cardiac pacemaker cells
(Fig. 9), shows that If is activated significantly at voltages within the pacemaker
range of potentials. These data are comparable to the data reported by Yanagihara
& Irisawa (1980) from man-made strips from the rabbit S-A node. those in cardiac
Purkinje strands (DiFrancesco, 1981 b) and the results of DiFrancesco et al. (1986)
from isolated cells of the rabbit S-A node. However. the variability (or run-down etc.)
in our experiments is much less than that observed in most previous studies using
isolated cells (but see Hagiwara & Irisawa, 1989; Denyer & Brown, 1990a. b). The
reason(s) for this are not known, but they may be related to the extent to which the
cells have rounded up (i.e. the Ca2+ tolerance, cf. Nakayama et al. 1984) or to the
osmotic strength of the solutions used in preparing and maintaining some of these
isolated cells (cf. DiFrancesco et al. 1986).
In our kinetic analysis of If currents the rate constants for opening and closing of
 SINO-ATRIAL NODE PACEMAKING 79
the channels were determined from the steady-state activation curve in combination
with direct measurements of the time constants of activation or deactivation of If.
With this information (Fig. 10), as well as knowledge of the reversal potential, the
fully activated conductance, and the effects of Cs' blockade (see Fig. 11; and Denyer

60 pA

500

Fig. 13. Comparison of the size and time course of the delayed rectifier current (IK) and
If in a single cell from the rabbit S-A node. Panel A shows the clamp protocol. At the start
of the experiment, spontaneous electrical activity was recorded; thereafter two different
protocols were applied from -30 mV. In the first, a 1-5 s hyperpolarization to -80 mV
was applied to activate I, In the second protocol, this hyperpolarization was preceded by
a 700 ms depolarization to 0 mV in order to activate IK (see panel B). Note that at
-80 mV the decay of IK is much faster than the activation of If. These findings suggest
that both the decay of IK and the activation of If can modulate the initial portion of the
pacemaker potential in the S-A node (see text for further description and discussion).

& Brown 1990b) it was possible to estimate the size and the time course of If during
the 'typical' diastolic depolarization. The data in Fig. 10 indicate that a substantial
fraction of the net inward current which underlies a 'typical' diastolic depolarization
could be attributable to activation of If, and that the influence of If will become large
whenever the cells are hyperpolarized. An important question which remains is: in
the first part of the pacemaker potential, are both the deactivation of IK and the
activation of If likely to be significant current changes? The experiment in Fig. 13
was carried out to examine this directly.
A spontaneously active type III cell was voltage clamped at a holding potential of
-30 mV (Fig. 13A) and two different protocols were applied. In the first protocol a
hyperpolarization to -80 mV which activated If was applied. Every second cycle a
two-pulse protocol was used. The first pulse, Pl, depolarized the cell into the plateau
80 A. C G. VAN GINNEKEN AND IL' GILES
range of potentials (0 mV) resulting in activation of 'Ca and IK; the potential was
then clamped back to -80 mV. The superimposed currents in panel B of Fig. 13
show the significant, relatively fast decline of IK superimposed upon the activation
of If. From these results it appears that both 'K and If modulate the initial portion
of the pacemaker depolarization.
In summary, our results confirm and extend previously available information
regarding If in mammalian cardiac pacemaker tissue: (1) They demonstrate
convincingly that If can be observed in each viable (type III or IV) spontaneously
active single cardiac pacemaker cell from rabbit heart. (2) They rule out any
significant contribution of electrochemical gradient changes to the size of the time
course of If. (3) They provide quantitative data concerning the steady-state voltage
dependence and kinetics of activation and deactivation of this current. (4) They form
the basis for a semiquantitative model which allows the size of If during the
pacemaker depolarization to be estimated.
In combination, these data suggest that If can be an important variable both in
normal 'physiological' conditions and also in any circumstance in which the cells
become hyperpolarized, e.g. exogenous application of acetylcholine (DiFrancesco
et al. 1989; Chang et al. 1990), adenosine (Belardinelli et al. 1988) or changes (increases)
in the extent of intercellular coupling from adjacent, more hyperpolarized crista
terminalis or atrial tissue. In additional work, it will be very important to study
further the selectivity of If and details of its ion transfer mechanism, and to continue
to explore the relationship between the anatomical location of the cells which are
isolated and the autonomic innervation of the S-A node (cf. Roberts et al. 1989).
This work was supported by grants from the National Institutes of Health (HL-33662), the
Canadian MRC, the Canadian Heart and Stroke Foundation. the Canadian Medical Research
Foundation, the Alberta Heritage Foundation for Medical Research (AHFMR) and the
Netherlands Organization for the Advancement of Pure Research (ZWO). Dr Giles held an
Established Investigator's Award from the American Heart Association and is presently an
AHFAIR MIedical Scientist. Dr van Ginneken was supported by an AHFMR Post-doctoral
Fellowship. Part of this work has been presented previously as an abstract (van Ginneken & Giles,
1985). We thank Professor H. Irisawa and Dr Alex West for their comments and criticisms, and
Ms Lenore Doell for skilled secretarial assistance.

REFERENCES

ARMSTRONG, C. AI. (1990). Potassium channel architecture and channel blockers. In Potassium
Channels: Basic Function and Therapeutic Aspects, ed. COLATSKY. T. J., pp. 1-17. Wiley-Liss
Ltd, New York.
BELARDINELLI, L., GILES, WV. R. & WEST, A. (1988). Ionic mechanism of adenosine actions in
pacemaker cells from rabbit heart. Journal of Physiology 405, 615-633.
BEZANILLA, F. (1985). Gating of sodium and potassium channels. Journal of Membrane Biology 88,
97-111.
BLEEKER, W. K., MACKAAY, A. J. C., MASSON-PEVET, M., BOUMAN, L. N. & BECKER, A. E. (1980).
Functional and morphological organization of the rabbit sinus node: a correlative study.
Circulation Research 46, 11-42.
Bois, P. & LENFANT, J. (1988). Isolated cells of the frog sinus venosus: properties of the inward
current activated by hyperpolarization. Pfiigers Archiv 416, 339-346.
BROWN, H. F. (1982). Electrophysiology of the rabbit sino-atrial node. Physiological Reviews 62,
505-530.
 `SIN.O-ATRIAL NODE PACEMAKING 81
BROWN, H. F. & DIFRANCESCO, D. (1980). Voltage-clamp investigations of membrane currents
underlying pacemaker activity in rabbit sino-atrial node. Journal of Physiology 308, 331-351.
BROWN. H. F.. GILES, WV. & NOBLE, S. J. (1977). Membrane currents underlying activity in frog
sinus venosus. Journal of Physiology 271, 783-816.
BROWN. H. F., KINIURA, J. & NOBLE, S. (1982). The relative contributions of various time-
dependent mnembrane currenits to pacemaker activity in the sino-atrial node. In Cardiac Rate and
Rhythm, ed. BOUMAN, L. N. & JONGSMA, H. J., pp. 63-68. Martinus Nijhoff, The Hague.
CARMELIET, E., COETZEE, WV. & MUIUBAGWA, K. (1982). Pacemaker activity in the rabbit cardiac
pacemaker Purkinje fibres. Journal of Physiology 334, 55-56P.
CHANG, F., GAO, J., TROMBA, C., COHEN, J. & DIFRANCESCO, D. (1990). Acetylcholine reverses
effects of t3-agonists on pacemaker current in canine cardiac Purkinje fibers but has no direct
action. A difference between primary and secondary pacemakers. Circulation Research 66,
633-636.
D)ENYER, J. C. & BROWN, H. F. (1990a). Rabbit sino-atrial node cells: isolation and electro-
physiological properties. Journal of Physiology 428, 405-424.
DENYER, J. C. & BROWN, H. F. (1990b). Pacemaking in rabbit isolated sino-atrial node cells during
Cs' block of the hyperpolarization-activated current, i. Journal of Physiology 429, 401-409.
DiFRANCESCO. D. (1981 a). A new interpretation of the pacemaker current in calf Purkinje fibres.
Journal of Physiology 341, 359-376.
DIFRANCESCO, D. (1981 b). A study of the ionic nature of the pacemaker current in Purkinje fibres.
Journal of Physiology 341, 377-393.
DIFRANcESCO, D. (1985). The cardiac hyperpolarizing-activated current, if. Origins and
developments. Progress in Biophysics and Molecular Biology 46, 163-183.
DIFRANcESco, D. (1986). Characterization of single pacemaker channels in cardiac sino-atrial node
cells. Nature 324, 470-473.
DiFRANCESCO, D. (1987). The pacemaker current in the sinus node. European Heart Journal 8,
suppl. L, 19-23.
DIFRANCESCO, D., DUcOURET, P. & ROBINSON, R. B. (1989). Muscarinic modulation of cardiac rate
at low acetylcholine concentration. Science 243, 669-671.
I)iFRANCESCO, D., FERRONI, A., MAZZANTI, M. & TROMBA, C. (1986). Properties of the
hyperpolarizing-activated current (if) in cells isolated from the rabbit sino-atrial node. Journal
of Physiology 377, 61-88.
D)IFRANCESCO, D. & NOBLE, D. (1985). A model of cardiac electrical activity incorporating ionic
pumps and concentration changes. Philosophical Transactions of the Royal Society B 307,
353-398.
DIFRANCESCO. D. & NOBLE, D. (1989). The current if and its contribution to cardiac pacemaking.
In Cellular and Neuronal Oscillators, ed. JACKLET, J. pp. 31-57. Marcel Dekker, New York.
DIFRANCESCO, D. & TROMBA, C. (1987). Acetylcholine inhibits activation of the cardiac
hyperpolarizing-activated current, if. Pfluigers Archiv 410, 139-142.
DIFRANCESCO, D. & TROMBA. C. (1988). Inhibition of the hyperpolarization-activated current (if)
induced by acetylcholine in rabbit sino-atrial node myocytes. Journal of Physiology 405,
477-491.
DOERR, TH., DENGER, R. & TRAUTWEIN, W. (1989). Calcium currents in single S-A node cells of the
rabbit heart studied with action potential clamp. Pfluigers Archiv 413, 599-603.
DOERR, TH. & TRAUTWEIN, XW. (1990). On the mechanism of the "specific bradycardiac action" of
the verapamil derivative UL-FS 49. Naunyn-Schmiedeberg's Archives of Pharmacology 341,
331-341.
EATON, D. C. & BRODWICK, AI. S. (1980). Effects of barium on the potassium conductance of squid
axon. Journal of General Physiology 75, 727-750.
GADSBY, D. C. (1984). The Na/K pump of cardiac cells. Annual Review of BiophysicalBioengineering
13, 373-398.
GILES, WV. R. & SHIBATA, E. (1985). Voltage clamp of bullfrog cardiac pacemaker cells: a
quantitative analysis of potassium currents. Journal of Physiology 368, 265-292.
GILES, W. R. & VAN GINNEKEN, A. C. G. (1985). A transiernt outward current in isolated cells from
the crista terminalis from the rabbit heart. Journal of Physiology 368, 243-264.
GLITSCH, H. G., PusCH, H. & VERDONCK, F. (1986). The contribution of Na and K ions to the
pacemaker current in sheep cardiac Purkinje fibres. Pfluigers Archiv 406, 464-471.
 82 A. C G. VAN GINNTi\fEKENV AND HIL GILES
Ca"+ of the hvperpolarization
HAGIWARA, N. & IRISAWA, H. (1989). Modulation by intracellular
activated inward current in rabbit single sino-atrial node cells. Journal of Physiology 409,
121-141.
HAGIWARA, N., IRISAWA, H. & KAMEYAMA, M. (1988). Contribution of two types of calcium
currents to the
233-253.
pacemaker potentials of rabbit sino-atrial node cells. Journal ofPhysioloyy395.

HAMILL, 0. P., MARTY, A., NEHER, E., SAKMANN, B. &SIGWORTH. F. (1981). Improved patch-
clamp techniques for high-resolution current recording from cells and cell-free membrane
patches. Pfluigers Archiv 391, 85-100.
HARTZELL, H. C. (1988). Regulation of cardiac ion channels by catecholamines, acetvlcholine and
second messenger systems. Progress in Biophysics and MIolecular 1655-47.
Biology 52,
HUME, J. R. & GILES, W. (1983). Ionic currents in single isolated bullfrog atrial cells. Journal of
General Physiology 81, 153-194.
IRISAWA.
375-388.
H. (1984). Electrophysiology of single cardiac cells. Japanese Journal of Physiology 34,

IRISAWA,H. (1987). Membrane currents in cardiac pacemaker tissue. Experientia 43, 1131-1135.
IRISAWA, H., BROWN, H. F. & GILES, W. (1991). Sino-atrial pacemaker mechanisms. Physiological
Reviews (in the Press).
KREITNER, D. (1985). Electrophysiological study
sinus node. Cardiovascular Research 19, 304-318.
of the two pacemaker mechanisms in the rabbit

MASSON-PEVET, M., BLEEKER, W. K., BESSELSEN,E., TREYTEL, B. XW., JONGSMA, H. J. &

BOUMAN, L. N. (1984). Pacemaker cell types in the rabbit sinus node: A correlative
ultrastructural and electrophysiological study. Journal of Molecular and Cellular Cardiology 16,
53-63.
MASSON-PIiVET, M., GROS, D. & BESSELSEN, E. (1980). The caveolae in rabbit sinius
Cell and Tissue Research 208, 183-196.
nodeand atrium.

MASSON-PEVET, M., JONGSMA, H. J., BLEEKER, W. K., TSJERNINA, L., VAN GINNEKEN. G., A. C.
TREIJTEL, B. W. &BOUMAN, L. N. (1982). Intact isolated
heart. Journal of Molecular and Cellular Cardiology 14, 295-299.
sinus node cells from the adult rabbit

MAYLIE, J. & MORAD, M. (1984). Ionic currents responsible for the generation of pacemaker current
in the rabbit sino-atrial node. Journal of Physiology 355. 215-235.
MAYLIE, J., MORAD, M. & WEISS, J. (1981). A study of the pace-maker potential in rabbit sinio-
atrial node: measurement of the potassium activity under voltage-clamp conditions. Journal of
Physiology 311, 161-178.
NAKAO, M. & GADSBY, D. C. (1989). [Na] and [K] dependence of the Na/K pump current-voltage
relationship in guinea pig ventricular myocytes. Journal of General Physiology 94, 539-565.
potential
NAKAYAMA, T., KURACHI, Y., NOMA, A. & IRISAWA, H. (1984). Action
currents of single pacemaker cells of the rabbit heart. Pfluigiers
and membrane
Archir, 402. 248-257.
NATHAN, R. D. (1986). Two electrophysiologically
rabbit sinoatrial node. American Journal of
distinct typesH325-330.
Physiology 250.
of cultured pacemaker cells from
NOBLE. D. & NOBLE, S. J. (1984). A model of sino-atrial node electrical activity based on a
B
modification of the DiFrancesco-Noble (1984) equations. Proceedings of the Royal Society 222,
295-304.
NOBLE, D. & TsIEN, R. W. (1968). The kinetics and rectifier properties of the slow potassium
current in cardiac Purkinje fibres. Journal of Physiology 195. 185-214.
NOMA, A., MORAD, M. & IRISAWA, H. (1983). Does the 'pacemaker current' generate the diastolic
depolarization in rabbit SA node cells? Pfiigers Archiv 397, 190-194.
IRISAWA, conductanices
NOMA, A., NAKAYAMA, T., KURACHI, Y. & H. (1984). Resting K in
pacemaker
245-254.
and non-pacemaker heart cells from the rabbit. Japanese Journal of Physiology 34,
PROVENCHER, S. W. (1976). A Fourier method for analysis of exponential decay curves. Biophysical
Journal 16, 27-41.
RASMUSSON, R., CLARK, J. W., GILES, WV. R., SHIBATA, E. F. & CAMPBELL, D. L. (1990). A
mathematical model of a bullfrog cardiac pacemaker cell. American Journal of Physiology 259.
H352-369.
ROBERTS, L. A., SLOCUM, G. R. & RILEY, D. A. (1989). Morphological study of the innervation
pattern of the rabbit sinoatrial node. American Journal of Anatomy 185, 74-88.
 SINO-A TRIAL NODE PACEMAKINV 83
ROBINSON. K. & GILES, W. (1986). Menu-driven A/D conversion, display, and plotting programs
for the IBM PC in electrophysiological studies. Computer iMethods and Programs in Biomedicine
23, 319-327.
SCHEID, C. R. & FAY, F. (1980). Control of ion distribution in isolated smooth muscle cells. I.
Potassium. Journal of General Physiology 75, 163-182.
SEYAMA, I. (1979). Characteristics of the anion channel in the sino-atrial node cell of the rabbit.
Journal of Physiology 294, 447-460.
SIHIBASAKI. T. (1987). Conductance and kinetics of delayed rectifier potassium channels in nodal
cells of the rabbit heart. Journal of Physiology 387. 227-250.
SIIIBATA, E. F.. DRuRY, T., REFSUM, H., ALDRETE, V. & GILES, W. (1989). Contributions of a
transient outward current to repolarization in human atrium. American Journal of Physiology
257, H1773-1781.
SHIBATA, E. F. & GILES, W. R. (1985). Ionic currents which generate the spontaneous diastolic
depolarization in individual cardiac pacemaker cells. Proceedings of the National Academy of
Sciences of the USA 82, 7796-7800.
SHIBATA. E. F., GILES, W. & POLLACK, G. H. (1985). Threshold effects of acetylcholine on primary
pacemaker cells of the rabbit sino-atrial node. Proceedings of the Royal Society B 223, 355-378.
TANIGUCHI, J., KOKOBUN. S., NOMA, A. & IRISAWA, H. (1981). Spontaneously active cells isolated
from the sino-atrial and atrio ventricular node of the rabbit heart. Japanese Journal of
Physiology 31, 547-558.
VAN BOGAERT, P. P. & CARMELIET, E. (1985). Chloride sensitivity of the if pacemaker current in
cardiac Purkinje fibres from sheep and atrial appendage fibres from human hearts. Archives
Internationales de Physiologie et de Biochimie 93, 14-15.
VAN GINNEKEN, A. C. G., BOUMAN, L. N.. JONGSMA, H. J., DUIVENVOORDEN, J. J., OPTHOF, T. &
GILES, W. R. (1987). Alinidine as a model of the mode of action of specific bradyeardic agents
on SA node activity. European Heart Journal 8, suppl. I, 25-33.
VAN GINNEKEN, A. C. G. & GILES, W. R. (1985). If in isolated cells from the rabbit SA-node.
Biophysical Journal 47, 496a.
YANAGIIIARA, K. & IRISAWA, H. (1980). Inward current activated during hyperpolarization in the
rabbit sino-atrial node cell. Pfluigers Archiv 385, 11-19.
YANAGIHARA, K., NOMA, A. & IRISAWA, H. (1980). Reconstruction of sino-atrial node pacemaker
potential based on voltage clamp experiments. Japanese Journal of Physiology 30, 841-857.