PHYSIOLOGICAL REVIEWS
Vol. 81, No. 4, October 2001
Printed in U.S.A.
Electrophysiology of the Sodium-Potassium-ATPase
in Cardiac Cells
HELFRIED GÜNTHER GLITSCH
Arbeitsgruppe Muskelphysiologie, Fakultät für Biologie, Ruhr-Universität Bochum, Bochum, Germany
www.prv.org
0031-9333/01 $15.00 Copyright © 2001 the American Physiological Society
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
I. Introduction
A. Definition of electrogenic Na⫹-K⫹ pumping and historical background
B. Physiological significance of electrophysiological studies on the cardiac Na⫹-K⫹ pump
II. Activation of the Cardiac Pump Current by Monovalent Cations
A. Activation of Ip by intracellular Na⫹
B. Activation of Ip by extracellular K⫹ and its congeners
C. Li⫹ binds to extra- and intracellular binding sites of the cardiac Na⫹-K⫹ pump
and activates Ip
D. Pump current densities, pump site densities, and maximum turnover rate
III. The Reversal Potential of the Cardiac Pump Current
A. Theoretical considerations
B. Evidence for backward running Na⫹-K⫹ pump and determination of the Na⫹-K⫹ pump Erev
IV. Voltage Dependence of Cardiac Pump Currents
A. Cardiac steady-state Ip-V relationships
B. Transient pump currents
V. Dependence of Pump Current on Intracellular ATP
VI. Temperature Dependence of Steady-State and Transient Pump Currents
VII. Significance of Electrogenic Sodium-Potassium Pumping for the Membrane Potential
of Cardiac Cells
A. Contribution of electrogenic Na⫹-K⫹ pumping to the cardiac resting potential
B. Importance of electrogenic Na⫹-K⫹ pumping for the cardiac action potential
VIII. Modulation of Cardiac Pump Current by Autonomic Transmitters and Related Compounds
A. Effect of adrenergic agonists on Ip
B. Modulation of cardiac Ip by acetylcholine
IX. Cardiac Glycosides and Cardiac Pump Current
A. Binding of cardiac glycosides to various isoforms of the cardiac Na⫹-K⫹ pump is
species dependent
B. Kinetics of cardiac steroid binding to the cardiac Na⫹-K⫹ pump
C. Cardiac glycosides alter the cardiac Ip-V relationship
X. Modulation of the Cardiac Sodium-Potassium Pump by Hormones
A. Aldosterone
B. Angiotensin-converting enzyme inhibition
C. Thyroid status and cardiac Na⫹-K⫹ pump
D. Insulin changes the cardiac Ip-V curve
XI. Miscellaneous
A. Anisosmolar external solution affects the activity of the cardiac Na⫹-K⫹ pump
B. Dietary cholesterol alters cardiac Na⫹-K⫹ pumping
C. Amiodarone inhibits the cardiac Na⫹-K⫹ pump following acute and chronic treatment
by different mechanisms
XII. Effects of the Cardiac Sodium-Potassium Pump on Ion Transporters and Channels Measured
by Electrophysiological Techniques
A. Modulation of the cardiac Na⫹/Ca2⫹ exchange
B. Interaction between the cardiac Na⫹-K⫹ pump and KATP channels
C. Effects of the cardiac Na⫹-K⫹ pump on IK(Na)
D. Blockade of the Na⫹-K⫹ pump activates IK(ACh) in atrial myocytes
XIII. Conclusion
1792
1792
1794
1794
1794
1796
1798
1799
1799
1799
1800
1801
1801
1805
1807
1808
1808
1808
1809
1810
1810
1812
1812
1812
1814
1816
1816
1816
1817
1817
1818
1818
1818
1819
1819
1819
1819
1820
1821
1821
1821
1791
1792
HELFRIED GÜNTHER GLITSCH
Glitsch, Helfried Günther. Electrophysiology of the Sodium-Potassium-ATPase in Cardiac Cells. Physiol Rev 81:
1791–1826, 2001.—Like several other ion transporters, the Na⫹-K⫹ pump of animal cells is electrogenic. The pump
generates the pump current Ip. Under physiological conditions, Ip is an outward current. It can be measured by
electrophysiological methods. These methods permit the study of characteristics of the Na⫹-K⫹ pump in its
physiological environment, i.e., in the cell membrane. The cell membrane, across which a potential gradient exists,
separates the cytosol and extracellular medium, which have distinctly different ionic compositions. The introduction
of the patch-clamp techniques and the enzymatic isolation of cells have facilitated the investigation of Ip in single
cardiac myocytes. This review summarizes and discusses the results obtained from Ip measurements in isolated
cardiac cells. These results offer new exciting insights into the voltage and ionic dependence of the Na⫹-K⫹ pump
activity, its effect on membrane potential, and its modulation by hormones, transmitters, and drugs. They are
fundamental for our current understanding of Na⫹-K⫹ pumping in electrically excitable cells.
I. INTRODUCTION
The Mg2⫹-dependent, Na⫹- and K⫹-activated ATPase
(EC 3.6.1.37; Ref. 165) is the molecular basis of the
Na⫹-K⫹ pump in animal cell membranes. The elucidation
of the amino acid sequence of the Na⫹-K⫹-ATPase ␣-subunit (105, 163) and of the ␤-subunit (106, 162) of various
species (180) has prompted numerous studies on ATPase
molecules modified by mutagenesis and heterologously
expressed in various cells to investigate the relationship
between structure and function of the Na⫹-K⫹-ATPase
(98, 117, 180). Furthermore, the existence of Na⫹/K⫹
isozymes has been studied in a large variety of species
and tissues (19, 174). The Na⫹-K⫹-ATPase consists of at
least two subunit proteins in stoichiometric amounts, the
␣- and ␤-subunit. The ␣-subunit exhibits a molecular mass
of ⬃110 kDa and probably spans 10 times the cell membrane. Both NH2 and COOH termini face the cytoplasma.
The ␣-subunit contains the binding sites for ATP, Na⫹,
K⫹, cardiac glycosides, specific inhibitors of the enzyme,
and the phosphorylation site. Thus the ␣-subunit is largely
responsible for the catalytic, transport, and pharmacological characteristics of the ATPase. The smaller ␤-unit,
with a molecular mass of ⬃50 kDa (depending on the
degree of glycosylation), has only one transmembrane
domain. The COOH terminus is located at the large
ectodomain of the subunit, whereas the NH2 terminus is
exposed to the cytoplasma. The activity of the Na⫹-K⫹ATPase requires the ␤-subunit. The subunit modulates the
transport characteristics of the ATPase and plays an important role in the maturation and proper membrane insertion of the Na⫹-K⫹-ATPase (19). It is still unclear
whether the enzyme in vivo works as an(␣␤)-monomer or
an (␣␤)2-diprotomer (154). A small, hydrophobic protein
of ⬃12 kDa, termed the ␥-subunit, copurifies with the ␣and ␤-subunits of the Na⫹-K⫹-ATPase. It has been found
in various tissues of different species (124). The physiological function of the ␥-subunit is not yet known. Like
other cellular proteins, ␣- and ␤-subunits are expressed in
Physiol Rev • VOL
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
A. Definition of Electrogenic Naⴙ-Kⴙ Pumping and
Historical Background
various isoforms. At present four ␣-subunits (␣1-␣4) and
three ␤-subunits (␤1-␤3) have been identified. The ␣1, ␣2,
and ␣3 are expressed in a variety of tissues, whereas the
␣4-protein has been detected so far only in the rat testis
(186). Both the ␣- and ␤-isoforms of the Na⫹-K⫹-ATPase
are expressed in a tissue-specific pattern (reviewed in
Ref. 19). As to the ␣-subunits, the ␣1-isoform is expressed
ubiquitously, whereas the ␣2-expression is predominant
in cardiac and skeletal muscle, brain, and adipocytes. The
␣3-isoform is abundant in neural tissues and in the ovary
(19, 186). The tissue-specific expression of ATPase isoforms can be altered during development and by hormones. Any combination between one of the ␣-subunits
␣1-␣3 and one of the ␤-subunits ␤1-␤3 may result in an
active Na⫹-K⫹-ATPase isoform (27). The isoforms differ
in their kinetic characteristics with regard to activation by
Na⫹, K⫹, ATP, and inhibition by cardiac glycosides. For
example, the rat ␣1␤1-isoform expressed in Sf9 insect
cells shows a higher Na⫹ and K⫹ affinity but a lower
affinity for ATP and a lower sensitivity toward cardiac
glycosides than the ␣3␤1-isoform of the Na⫹-K⫹-ATPase
(19). Because of their different kinetic characteristics and
tissue-specific expression, the various Na⫹-K⫹-ATPase
isoforms probably meet different physiological demands.
On the one hand, the ␣1␤1-isoform may be a general,
housekeeping enzyme, since it is ubiquitously expressed
and exhibits suitable kinetic properties with relatively
high Na⫹ and K⫹ affinities. On the other hand, ␣3-isoforms
are especially suited to restore the Na⫹ and K⫹ gradients
across the cell membrane of electrically excitable cells
due to their lower Na⫹ and K⫹ affinities and higher ATP
affinity. With regard to the heart, the expression of the
Na⫹-K⫹-ATPase isoforms is species specific. There is a
marked variation of ␣2- and ␣3-expression among the
species, whereas the ␣1-isoform of the ATPase is present
in cardiac tissue of all species studied. For instance,
ventricular cells from adult human or macaque hearts
express three ␣-isoforms, ventricular myocytes from the
adult rat heart contain mainly ␣1- and ␣2-isoforms of the
Na⫹-K⫹-ATPase, whereas the sheep heart expresses only
the ␣1-isoform (175). Whether guinea pig ventricular cells
exclusively contains the ␣1-isoform (175) or additional ␣2
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
whereas in the P-E2 confirmation the binding sites face the
extracellular space and preferably bind extracellular K⫹
(Ko⫹) (or its congeners). Some binding sites are located at
the bottom of an “access channel” and not just at the
surface of the pump molecule. During the Na⫹ or K⫹
translocation across the cell membrane carried out by
the pump, there are states [(Na3) E1-P and E2 (K2) in
Fig. 1] in which the transported cations are “occluded”
in the pump molecule and unable to interchange with
ions in the surrounding media.
Electrogenicity denotes the characteristic of a biological transport mechanism to produce electrical current. The concept of electrogenic Na⫹-K⫹ pumping slowly
emerged during the 1950s (176). However, a most influential paper on active Na⫹-K⫹ transport in axons of Sepia
and Loligo (94) lent little support to this idea. Connelly
(26) was the first to conclude on the basis of experimental
evidence that the Na⫹-K⫹ pump in nerve fibres is electrogenic. “. . . since 1960 more and more evidence has accumulated showing that the pump is probably always at
least partly electrogenic, with more sodium being extruded than potassium taken up” (176). Today, it is generally accepted that the Na⫹-K⫹ pump of animal cells is
electrogenic and generates the pump current Ip. Since,
under physiological conditions, three Na⫹ are removed
from the cell but only two K⫹ are taken up per pump
cycle, Ip is an outward current. The existence of an electrogenic Na⫹-K⫹ pump in cardiac cells was first suggested
to explain the high temperature sensitivity of the cardiac
resting potential (29). The early experimental data demonstrating electrogenic Na⫹ pumping in nerve and muscle
have been reviewed in detail several years ago (119, 176).
The experimental evidence for the electrogenicity of the
FIG. 1. An electrostatic model of the
Na⫹-K⫹ pump. The conformation E1 of the
pump faces the cytosol and exhibits at the
surface two negatively charged binding
sites for which intracellular Na⫹ and K⫹
compete with different affinities. A third
Na⫹-specific neutral binding site is located
inside the pump molecule. To reach the
binding site the ion has to migrate through
a narrow “access channel.” Phosphorylation of the pump by ATP induces Na⫹ occlusion [(Na3)E1-P]. Transition to the E2
conformation opens the access to the extracellular medium via a narrow access
channel. Binding of two K⫹ probably occurs at the bottom of the channel. It causes
dephosphorylation and K⫹ occlusion
[E2(K2)]. Release of two K⫹ into the cytol
completes the pump cycle. [From Heyse et
al. (89), by copyright permission of The
Rockefeller University Press.]
Physiol Rev • VOL
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
(62) is still controversial. Because the ␣-subunits are specific cardiac steroid receptors exhibiting different sensitivity for these drugs, this controversy echos in contradictory reports on the number of different glycoside
receptors present in guinea pig cardiac ventricular myocytes (see below). For further information about the isoforms of the Na⫹-K⫹-ATPase, the reader is referred to
pertinent reviews (19, 174).
The Na⫹-K⫹ pump maintains the Na⫹ and K⫹ gradients
between the cytosol and extracellular medium. The maintenance of the gradients is a prerequisite for the ionic homeostasis of the cells, for cell volume regulation, and for
secondary active transports of amino acids, sugars, bile
acids, neurotransmitters, ions, and other solutes across the
cell boundary. Furthermore, in electrically excitable cells,
creation and maintenance of Na⫹ and K⫹ gradients across
the membrane are required for the generation of the resting
potential and the generation and propagation of action potentials. It is clear that Na⫹-K⫹ pumping is of prime functional significance in cells displaying relatively frequent electrical discharges over a long period of time, like cardiac
cells. Our present understanding of Na⫹-K⫹ pumping is outlined in the Post-Albers cycle (3, 137–139). For detailled
information about the experimental basis and the scientific
elaboration of the concept, the reader is referred to excellent, introductory reviews (72, 111). The simplified scheme
of the pump cycle, shown in Figure 1, may facilitate the
appreciation of the findings and ideas discussed below. According to the Post-Albers cycle, the Na⫹-K⫹ pump essentially exists in two conformations, E1 and E2, which may be
phosphorylated (E1-P; P-E2 in the Fig. 1) or dephosphorylated. In the E1 ATP conformation, the cation-binding sites
of the pump face the cytoplasm and preferentially bind Na⫹,
1793
1794
HELFRIED GÜNTHER GLITSCH
B. Physiological Significance of
Electrophysiological Studies on
the Cardiac Naⴙ-Kⴙ Pump
The electrogenicity of ion pumps cannot be considered an epiphenomenon which is inevitably linked to ion
pumping. Although the electrical potential difference set
up by the pump across the cell membrane is thermodynamically equivalent to the simultaneously generated osmotic gradient, the former and the latter display quite
different kinetic characteristics. As a consequence, the
pump-generated membrane potential is, under many conditions, more efficient as a driving force for secondary
active transports than the osmotic (ionic) gradient produced by the pump (see Ref. 111, p. 13–14). In addition,
the pump current Ip directly affects the automaticity, the
resting and action potential, and thereby the conduction
of electrical impulses in excitable membranes. Ip is a
direct indicator of Na⫹-K⫹ pumping, since the coupling
ratio 3Na⫹:2K⫹:1ATP per pump cycle remains constant
under a variety of conditions including changes of the
intracellular Na⫹ or extracellular K⫹ concentration and of
the membrane potential. This applies for cardiac (46, 65)
and noncardiac tissues (reviewed in Ref. 32). Measurements of Ip by means of electrophysiological methods
offer the possibility to study the Na⫹-K⫹ pump in its
physiological environment, i.e., in the cell membrane sepPhysiol Rev • VOL
arating two compartments of different ionic compositions
(intra- and extracellular space) and to measure pumpmediated Na⫹ and K⫹ fluxes. In cardiac myocytes as in
most animal cells, a membrane potential exists across the
cell membrane. Because a translocation of electrical
charge across the membrane constitutes the pump current, the Ip amplitude must depend on the membrane
potential. The interaction between the electrogenic
Na⫹-K⫹ pump and the membrane potential is best studied
by electrophysiological techniques. The measured voltage
dependence of external and internal cation binding to the
cardiac pump has inspired our imagination of the molecular shape of the Na⫹-K⫹ pump as a channel-like structure. Furthermore, electrophysiological studies have rendered possible the identification of additional partial
reactions in the pump cycle displaying voltage sensitivity.
They also revealed effects of Na⫹-K⫹ pumping on currents produced by cardiac ionic channels or transporters
in the vicinity of pump molecules. A major advantage of
electrophysiological methods for studies on Na⫹-K⫹
pump-generated Na⫹ and K⫹ fluxes across the sarcolemma over, for instance, tracer measurements is the
much better time resolution (up to the microsecond
range). Corresponding measurements of Ip provided
new insights into the kinetics of the interaction between
the Na⫹-K⫹ pump and drugs with hitherto unrivalled
precision.
This paper reviews some characteristics of the
Na⫹-K⫹ pump as a current-generating molecule in single
cardiac cells. The pump current has been investigated in
cells isolated from various regions of the heart of different
mammalian species. These regions differ in morphology
and function. They include the primary pacemaker (sinoatrial node), the cardiac conducting system (Purkinje fibres), and the working myocardium (atrial and ventricles). Especially the ventricular myocytes are easy to
isolate and exhibit a high density of pump molecules in
the cell membrane. Their cellular geometry is adequate
for studies on the Na⫹-K⫹ pump by means of patch-clamp
techniques. For these reasons, several investigations that
are pivotal for our understanding of the electrogenicity of
the Na⫹-K⫹ pump have been carried out on these cells.
II. ACTIVATION OF THE CARDIAC PUMP
CURRENT BY MONOVALENT CATIONS
A. Activation of Ip by Intracellular Naⴙ
1. Mean affinity constant values for Ip activation by
intracellular Na⫹ solution vary according to the
experimental conditions
Earlier studies in multicellular cardiac preparations
showed that intracellular Na⫹ (Na ⫹
i ) activates Ip (45, 65).
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
cardiac Na⫹-K⫹ pump from multicellular preparations has
been presented in various reviews (45, 47, 63, 65, 181).
The introduction of new methods and techniques
into the electrophysiology of the Na⫹-K⫹-ATPase during
the last two decades has markedly improved our knowledge of electrogenic Na⫹ pumping. The isolation of single
cardiac cells (140) rendered possible Ip measurements
(48) by means of patch-clamp techniques (77). Reliable Ip
measurements require patch pipettes with a large tip diameter of ⬃4 –5 ␮m and a low resistance (⬍2 M⍀). In
addition, membrane currents other than Ip have to be
suppressed by adequate experimental conditions. The
measurements demonstrated the activation of Ip by various intra- and extracellular cations and established the
voltage dependence of Ip under a variety of conditions.
Furthermore, they revealed the existence of transient
pump currents in the Na⫹ and K⫹ limb of the Na⫹-K⫹
pump cycle (130, 135). Clearly, electrophysiological investigations of the Na⫹-K⫹ pump in noncardiac preparations
have likewise produced exciting new insight into active
cation transport. For example, studies of Ip in Xenopus
oocytes (110, 147) and combined electrophysiological and
tracer flux measurements in squid axons (53, 144) provided important data for our current understanding of
structural and functional properties of the Na⫹-K⫹ pump
in animal cells (111, 145, 150, 180).
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
TABLE
the main cation in cardiac cells, the “physiological” K0.5
value for Ip activation by Na⫹
i may be in the range of ⬃20
mM Nao⫹.
Figure 2 compares the Ip activation as a function of
[Na⫹]pip using pipette solutions containing different main
cations in guinea pig ventricular myocytes at 0 mV holding
potential. Obviously, half-maximal Ip activation occurs at
higher [Na⫹]pip if Cs⫹ instead of TEA⫹ or N-methyl-Dglucamine ion (NMDG⫹) is used as main pipette cation.
In summary, cardiac Ip is activated by intracellular
Na⫹. The [Na⫹]pip reported for half-maximal Ip activation
varies widely according to the experimental procedure
and the ionic conditions chosen.
2. Is Ip activation by intracellular Na⫹
voltage dependent?
Whether or not binding of intracellular Na⫹ to the
cardiac Na⫹-K⫹ pump is voltage sensitive is still a point of
controversy. The sensitivity of the pump in guinea pig
ventricular cells to [Na⫹]pip in the range between 3 and 50
mM increased with depolarization (131). An e-fold drop of
the K0.5 value (i.e., [Na⫹]pip for half-maximal Ip activation)
was estimated for a depolarization by 250 mV. However,
the effect was present only in Na⫹-rich superfusion media
but absent in Na⫹-free solution where the apparent affin⫹
ity of the pump to Napip
seemed to be voltage independent. Consequently, it was concluded that cytoplasmic
Na⫹ binding to the cardiac Na⫹-K⫹ pump is voltage insensitive (131). The weak membrane potential dependence of the K0.5 value seen in myocytes superfused with
Na⫹-rich media was ascribed to a voltage-dependent redistribution of enzyme intermediates in the pump cycle.
In line with the findings described above, an increasing
sensitivity of the pump to low [Na⫹]pip (5 mM) was ob-
⫹
1. Activation of Ip by Napip
K0.5 Value, mM
⬃10
11
2.7
⬃3
⬃43 or 1 (KD1) and
29 (KD2 ⫽ KD3)
14
⬃40
21.4
7.8
21.4
18.7
8.6
14
17.4
9
Species
Guinea
Guinea
Guinea
Guinea
pig
pig
pig
pig
Guinea pig
Guinea pig
Rat
Rat
Rat
Rabbit
Rabbit
Rabbit
Rabbit
Chicken
Sheep
Cells
Principal
Internal Cation
Ventricular
Ventricular
Ventricular
Ventricular
Cs⫹
Cs⫹
TEA⫹
TEA⫹, NMDG
Ventricular
Atrial
Ventricular
Ventricular
Ventricular
Ventricular
Ventricular
Ventricular
Sinoatrial node
Embryonic cardiac myocytes
Purkinje cardioballs
K⫹ ⫹ TEA⫹
Cs⫹
K⫹ ⫹ Cs⫹
K⫹
TEA⫹
K⫹ ⫹ TMA⫹
Cs⫹
TEA⫹
Cs⫹
K⫹
Cs⫹
Holding
Potential, mV
0
0
0
0
⫺40
0
⫹20 to ⫺120
0
0
⫺40
0
0
0
⫺70
⫺40
Reference
No.
49
131
84
9
128
109
167
84
84
184
161
84
151
168
66
TEA⫹, tetraethylammonium ion; NMDG, N-methyl-D-glutamate; TMA⫹, tetramethylammonium ion; K0.5, mean affinity constant; Ip, pump
current.
Physiol Rev • VOL
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
However, due to experimental problems, a quantitative
relationship between [Na⫹]i and Ip was difficult to establish. Whole cell recording from single cardiac cells certainly facilitated investigations on this issue, and, during
the last decade, various mean affinity constant (K0.5) values for half-maximal Ip activation by intracellular Na⫹
⫹
(Napip
) have been reported. In view of the variable numbers published, the following points should be kept in
mind. Apart from species differences (including the expression of different Na⫹-K⫹-ATPase isoforms), the experimental procedure chosen for estimation of a K0.5
value is likely to affect the result. A low tip restistance
(⬃1 M⍀) of the patch pipette is a prerequisite for correct
measurements of Ip as a function of [Na⫹]i (121). There is
evidence suggesting the existence of a “fuzzy space”
(112), in which the ionic concentration deviates from that
in the bulk cytosolic solution (23, 159). As to the cardiac
Ip-[Na⫹]pip relationship, it was demonstrated that the subsarcolemmal [Na⫹] is not always controlled by the [Na⫹]
of the pipette solution. This is true not only during strong
Ip activation but also in the steady state, at least in certain
cells (16). Furthermore, a relatively high patch-pipette
resistance might cause an additional Na⫹ gradient across
the cytosol if active Na⫹/K⫹ exchange is strongly activated. Thus a thoughtful procedure is required to obtain
reliable results. In addition, intracellular K⫹ are known to
be competitive inhibitors of Na⫹ at intracellular Na⫹binding sites of the Na⫹-K⫹ pump (73). Consequently, one
would expect a lower [Na⫹]pip for half-maximal Ip activation (K0.5 value) from measurements where the main
cation of the pipette solution is a weaker competitor than
K⫹ [Cs⫹ or even tetraethylammonium ion (TEA⫹)]. It
might be helpful to remember these points when reading
the data presented in Table 1. They were mainly obtained
at 30 –37°C. Since, under physiological conditions, K⫹ is
1795
1796
HELFRIED GÜNTHER GLITSCH
pump of guinea pig ventricular myocytes was clearly
shown in Na⫹-free solution (9). Figure 3 is from this work
and demonstrates that the K0.5 value for Ip activation by
[Na⫹]pip varies with membrane potential in myocytes containing either NMDG⫹ (A) or TEA⫹ (B) as the main cation
species. Thus experimental evidence is accumulating that
the apparent affinity of the cardiac Na⫹-K⫹ pump to intracellular Na⫹ is voltage dependent and increases with
depolarization. This conclusion is supported by observations from experiments on cell-free Na⫹-K⫹-ATPase systems (74, 89, 134, 171).
2. Activation of pump current (Ip) as a function of pipette Na⫹
concentration ([Na]pip) in guinea pig ventricular myocytes. Ip normalized to the respective amplitudes at 50 mM [Na]pip. The curves obey a
Hill equation. The principal cation in the pipette solution affects Ip
activation. [Data from Nakao and Gadsby (131) (F) and Barmashenko et
al. (9) (‚, 䊐).]
FIG.
served with depolarization in sheep cardiac Purkinje cells
superfused with Na⫹-rich medium (66). Similarly, an increased apparent Na⫹ affinity of the pump to low [Na⫹]pip
with depolarization was reported for guinea pig atrial
myocytes superfused with Na⫹-containing or Na⫹-free
solution (109). However, little evidence for voltage-depen⫹
dent internal Na⫹ binding between 20 and 85 mM Napip
was found during whole cell recording from rat ventricular cells in Na⫹-containing medium (167). The results
described so far were obtained with pipette solutions
containing K⫹ or Cs⫹, which may compete with Na⫹ for
internal cation binding sites of the Na⫹-K⫹ pump and may
thereby obscure the mechanism of Na⫹ binding. In a more
recent study, TEA⫹ or NMDG⫹ were used as noncompetitive main cations in the pipette solution, and at ⱕ5 mM
⫹
Napip
, voltage-sensitive binding of internal Na⫹ to the
1. Measurements in Na⫹-containing solution
Under physiological conditions, the Na⫹ pump of
animal cells is activated by extracellular K⫹. The activation of the pump by external K⫹ follows sigmoid saturation kinetics that can be described by a Hill equation.
Earlier studies on the pump activation by extracellular K⫹
in multicellular cardiac preparations have been reviewed
in useful articles (45, 65). The majority of the data suggest
half-maximal Ip activation by [K⫹]o in the low millimolar
range. Similar K0.5 values were obtained from experiments on single cardiac myocytes. The data obtained at
30 –37°C are presented in Table 2. The Ip activation by two
K⫹ congeners, Tl⫹ and NH⫹
4 , was tested in rabbit Purkinje
cells (15). The authors derived K0.5 values of 0.4 mM Tl⫹
and 5.7 mM NH⫹
4 and noted no alteration in the shape of
the Ip-Voltage (V) curve if equipotent concentrations of
K⫹, Tl⫹, or NH⫹
4 were applied. The following relative
potency of extracellular monovalent cations to activate Ip
was observed in rabbit sinoatrial node cells: K⫹ⱖ Rb⫹
⬎ Cs⫹ ⬎⬎ Li⫹ (151). The order of potency is the same as
that observed for the activation of the isolated Mg2⫹dependent, Na⫹- and K⫹-activated ATPase, which is the
FIG. 3. Mean affinity constant (K0.5) val⫹
ues for Ip activation by Napip
are voltage
dependent. K0.5 values for activation of Ip by
⫹
Napip in cells containing either N-methyl-Dglucamine ion (NMDG⫹) (A) or tetraethylammonium ion (TEA⫹) (B) are plotted versus membrane potential (V). In both cases
K0.5 values decrease monotonically with depolarization. The fitted curves obey a Boltzmann equation [analogous to Equation 1
with ␣-values of 0.16 for NMDG⫹ (A) and
0.15 for TEA⫹ (B); r 2 ⫽ 0.948 (A) and 0.895
(B)]. [From Barmashenko et al. (9). Copyright 1999 The Physiological Society.]
Physiol Rev • VOL
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
B. Activation of Ip by Extracellular Kⴙ and
Its Congeners
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
TABLE
1797
⫹
2. Activation of Ip by K⫹
o in Na -containing solution
K0.5 Value,
mM
Species
1.5
1.1
1.2
1.2
3.5
0.9
1.9
1.4
1.9
0.8
Guinea pig
Guinea pig
Guinea pig
Rat
Rat
Rat
Chicken
Rabbit
Rabbit
Dog
Cells
Internal [Na⫹],
mM
Holding
Potential,
mV
Reference
No.
Ventricular
Ventricular
Atrial
Ventricular
Ventricular
Ventricular
Embryonic cardiac cells
Sinoatrial node
Purkinje cells
Purkinje cells
50pip
50pip
50pip
10pip
85pip
100pip
Augmented
30pip
10pip
Not determined
0
⫺20
0
⫺40
⫺40
⫺20
⫺70
0
⫺20
⫺30
131
84
109
107
107
84
168,170
151
15
25
molecular basis of the Na⫹-K⫹ pump (155). In addition,
the order of potency confirmed the sequence deduced
from voltage-clamp measurement on guinea pig papillary
muscles and sheep Purkinje fibers (39).
2. K0.5 values for Ip activation by external cations
in Na⫹-free media
Since the early observation on erythrocytes it is
known that external Na⫹ and K⫹ compete for common
binding sites (138). As a consequence, the K0.5 value for
pump activation by extracellular K⫹ is appreciably lower
in Na⫹-free than in Na⫹-containing media. Thus one
would expect for cardiac cells a lower K0.5 value for the Ip
activation by K⫹ and its congeners in Na⫹-free than in
Na⫹-containing solution. In fact, a K0.5 value of 0.22 mM
was derived for the activation of Ip by Ko⫹ in guinea pig
ventricular myocytes superfused with Na⫹-free medium
(holding potential 0 mV) (131). This is shown in Figure 4,
where normalized pump currents are plotted versus
[K⫹]o. The holding potential is 0 mV. The K0.5 value for Ip
activation by Ko⫹ in these cells decreases from 1.5 mM Ko⫹
in Na⫹-containing solution to 0.22 mM Ko⫹ in Na⫹-free
medium (131). Similarly, a K0.5 value of 0.2 mM was
reported for the Ip activation by Ko⫹ in rat and guinea pig
⫹
⫹
⫹
⫹
FIG. 4. Activation of Na -K pump current in guinea pig ventricular myocytes at 0 mV by [K ]o at high [Na ]o (A)
or at zero [Na⫹]o (B). In both cases, pump current amplitude at each [K⫹]o was normalized to control runs at 5.4 mM
Ko⫹. The graphs show mean values ⫾ SE of the resulting relative pump currents plotted against [K⫹]o, and the curves
show least-squares fits of the Hill equation to the unweighted data. [Na⫹]pip was 50 mM in all cells. A: data from 17 cells
at high [Na⫹]o; best-fit parameters: maximal relative current ⫽ 1.30 ⫾ 0.10; K0.5 ⫽ 1.54 ⫾ 0.31 mM Ko⫹, nH ⫽ 0.96 ⫾ 0.13.
B: data from 4 cells at zero [Na⫹]o; best-fit Hill equation parameters: maximal relative current ⫽ 1.03 ⫾ 0.05, K0.5 ⫽
0.22 ⫾ 0.03 mM Ko⫹, nH ⫽ 1.12 ⫾ 0.14. [Adapted from Nakao and Gadsby (131).]
Physiol Rev • VOL
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
The subscript pip refers to the Na⫹ concentration of the patch pipette solution.
1798
HELFRIED GÜNTHER GLITSCH
3. Activation of Ip by K⫹
o in cardiac cells is
voltage dependent
The negative slope of the Ip-V relationship in Xenopus oocytes is probably due to voltage-sensitive binding
of extracellular K⫹ to the Na⫹-K⫹ pump (147). It is considered that Ko⫹ binding to the pump occurs at the bottom
of a “high field, narrow access channel” (see Ref. 111, p.
74 – 83). To reach their binding sites, the K⫹ have to
migrate through the channel within the electrical field
across the cell membrane. The membrane potential
thereby affects the local K⫹ concentration at the binding
sites which, in general, differ from [K⫹] of the extracellular medium. Depolarization of the cell membrane de-
FIG. 5. K0.5 values for Ip activation of rabbit cardiac Purkinje cells by
external cations at various clamp potentials. The curves fitted to the data
obey Equation 1 with parameters listed in Table 3. A: F, K⫹; r 2 ⫽ 0.93;
⫹
2
⫹
2
⫹
2
E, Tl ; r ⫽ 0.79. B: ■, NH4 ; r ⫽ 0.98; 䊐, Cs ; r ⫽ 0.77. Note the
different scales of the ordinates in A and B. [From Bielen et al. (18).
Copyright 1993 The Physiological Society.]
Physiol Rev • VOL
3. K0.5(Vc⫽0 mV) and ␣-values for Ip activation by
external cations in Na⫹-free solution
TABLE
K0.5(Vc⫽0 mV), mM
␣
Tl⫹
K⫹
NH⫹
4
Cs⫹
0.05
0.32
0.08
0.24
0.4
0.29
1.5
0.18
[From Bielen et al. (18). Copyright 1993 The Physiological Society.]
creases the local [K⫹] and increases the estimated (apparent) K0.5 value for Ip activation by Ko⫹. Vice versa,
hyperpolarization increases the local [K⫹] at the binding
sites and decreases the K0.5 value. Under certain experimental conditions, the cardiac Ip-V relationship also exhibits a region of negative slope. This is true for myocytes
superfused with Na⫹-containing solution (13, 15, 84, 167)
or Na⫹-free (poor) media (13, 15, 18, 52), especially at low
(⬍K0.5 value) concentrations of Ko⫹ or its congeners. A
detailled study on the activation of Ip by Ko⫹, Tlo⫹, extra⫹
cellular NH⫹
4 , and Cso at various membrane potentials
(Vc) was carried out with rabbit cardiac Purkinje cells in
Na⫹-free solution (18) (Fig. 5; see also Table 3). Figure 5
shows that the apparent affinity of the cardiac Na⫹-K⫹
pump to Ko⫹, Tlo⫹, Cso⫹, or extracellular NH⫹
4 depends on
voltage. The affinity decreases with depolarization,
whereas the corresponding K0.5 values increase according
to the Boltzmann equation
K 0.5 ⫽ K0.5共Vc⫽0 mV兲 䡠 exp共␣ 䡠 FVc /RT兲
(1)
where K0.5(Vc⫽0 mV) denotes the K0.5 value at zero potential. ␣ is a steepness factor, and F, Vc, R, and T have their
usual meanings. K0.5(Vc⫽0 mV) values for Ip activation by
various external cations and the respective ␣-values are
collected in Table 3 (from Ref. 18). Because the steps
subsequent to Ko⫹ binding in the K⫹ limb of the pump
cycle are supposed to be voltage independent (135), it is
assumed that the binding of Ko⫹ or its congeners to the
Na⫹-K⫹ pump probably is the voltage-sensitive process
involved. From the ␣-values and the Hill coefficients derived, it can be estimated that the external activator cations sense ⬃0.2 of the membrane potential at their binding sites (see Ref. 150). In summary, the binding of Ko⫹ and
ist congerners to the cardiac Na⫹-K⫹ pump is most probably voltage dependent.
C. Liⴙ Binds to Extra- and Intracellular Binding
Sites of the Cardiac Naⴙ-Kⴙ Pump and
Activates Ip
It is known from earlier studies on multicellular cardiac preparations that Li⫹ is a weak external pump activator cation (see Ref. 64 for references). As to isolated
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
ventricular cells (84). Furthermore, the maximum Ip
amplitude that can be activated in Na⫹-free media by
Ko⫹ and congeners declines with increasing K0.5 values
of the various monovalent cations. K0.5 values for the Ip
activation by external K⫹ or Tl⫹ of ⬃0.22 mM Ko⫹ and
⬃0.08 mM Tlo⫹ were estimated in rat ventricular myocytes superfused with Na⫹-free solution (holding potential 0 mV) (13). In Na⫹-free media, the K0.5 values for
the activation of the pump current by Ko⫹ or K⫹ congeners in rabbit Purkinje cells amount to 0.05 mM Tlo⫹,
⫹
0.08 mM Ko⫹, 0.4 mM extracellular NH⫹
4 , and 1.5 mM Cso
(holding potential 0 mV) (18). The authors concluded
that in general the K0.5 values for the Na⫹-K⫹ pump
activation in cardiac cells superfused with Na⫹-free
solution are lower by a factor of 10 –20 than those
obtained in Na⫹-containing media. The same is true for
the K0.5 values reported from Ip measurements in squid
axons (144) and Xenopus oocytes (133), although the
absolute numbers are somewhat larger than those estimated for cardiac cells.
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
D. Pump Current Densities, Pump Site Densities,
and Maximum Turnover Rate
Usually, in experiments designed to measure Ip, the
Na⫹-K⫹ pump of cells is strongly activated by nearly
⫹
saturating concentrations of Napip
and [K⫹]o. The measured Ip densities depend not only on these ionic concentrations, but also on the membrane potential and the
temperature at which the measurements were carried out.
In addition, cells in culture often exhibit lower Ip densities
than freshly isolated cells. Furthermore, the species of the
main intracellular cation used (K⫹, Cs⫹, or NMDG⫹) affects the Ip density. Ionic species that do not compete
with Na⫹ for intracellular binding sites of the pump (for
instance NMDG⫹) tend to evoke higher Ip densities (166).
It might be helpful for the reader of the data presented
below to keep these points in mind. The first measurement of the Ip density in isolated guinea pig ventricular
myocytes yielded 1–1.5 ␮A/cm2 (48). The maximum Ip
density of these cells were estimated to be ⬃1.6 ␮A/cm2
(131). In sheep Purkinje cardioballs (1–3 days in culture),
an Ip density of 1.1 ␮A/cm2 has been observed (66). A
maximum Ip of ⬃1.9 ␮A/cm2 has been derived from Ip
measurements in rabbit sinoatrial node cells (151).
Guinea pig atrial cardioballs (1 day in culture) exhibit an
Ip density of 0.66 ␮A/cm2 (109), whereas rabbit ventricular myocytes display Ip densities between ⬃1.7 and ⬃2.2
Physiol Rev • VOL
␮A/cm2 (14, 75, 76). Ip densities reported for rat ventricular cells vary between 1.6 ␮A/cm2 (166) and ⬃4.3 ␮A/cm2
(101). All these numbers were obtained from measurements at membrane potentials greater than ⫺45 mV and
near body temperature. They agree reasonably well with
earlier data for the active Na⫹ efflux via the cardiac Na⫹
pump obtained by various nonelectrophysiological methods (63). (An Ip density of 1 ␮A/cm2 translates to a pumpmediated flux of ⬃30 pmol䡠cm⫺2䡠s⫺1). They are also in
line with previous estimates of Ip density in multicellular
cardiac preparations (28). For comparison, simultaneous
measurements of Ip density and active 22Na efflux in squid
axons yielded 0.89 ␮A/cm2 and ⬃ 25 pmol䡠cm⫺2䡠s⫺1 (144),
again in accordance with earlier flux data (94).
From the maximum quantity of movable charge derived from measurements of transient pump currents in
single cardiac cells (see sect. IVB), the pump site density
of the myocytes has been estimated assuming a single
charge (1.6 ⫻ 10⫺19 C) to be transfered per pump molecule and the specific membrane capacitance to be 1 ␮F/
cm2. The first number obtained by this procedure for the
pump site density of single cardiac cells was published by
Nakao and Gadsby (130). According to the authors, the
pump site density of guinea pig ventricular myocytes is
⬃1,200/␮m2. Later estimates include slightly higher numbers of 2,200 to 2,800 pumps/␮m2 for guinea pig ventricular cells (108) and ⬃ 2,600 sites/␮m2 for rat ventricular
myocytes (35). These values are higher by a factor 2 to 4
than an earlier number derived from Ip measurements in
a guinea pig multicellular ventricular preparation (28).
The maximum turnover rate of the charge transfer by
the cardiac Na⫹-K⫹ pump can be calculated if the maximum Ip density and the pump site density are known.
Accordingly, a maximum turnover rate of ⬃80/s has been
obtained from the data mentioned above for guinea pig
ventricular myocytes (51). A higher maximum turnover
rate of ⬃200/s has been derived from measurements of
transient pump currents in excised patches of rat ventricle cells (42). Both numbers apply to turnover rates at 0
mV and 36°C.
The pump site densities reported above are within
the range of earlier estimates in a variety of cell species by
means of different methods, mainly cardiac glycoside
binding (see Table 3 in Ref. 32). These estimates vary
considerably among the cell types, whereas the calculated
turnover rates are much more similar and amount to
⬃100/s at body temperature (32).
III. THE REVERSAL POTENTIAL OF THE
CARDIAC PUMP CURRENT
A. Theoretical Considerations
The free energy of intracellular ATP hydrolysis
(⌬GATP) fuels the active Na⫹/K⫹ transport by the Na⫹-K⫹
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
cardiac cells, Ip is activated by Lio⫹ in single rabbit Purkinje cells (15). In isolated sinoatrial cells, Lio⫹ is an
⬃14-fold weaker external activator than Ko⫹ (151). Inter⫹
⫹
⫹
estingly, intracellular Li⫹ activates Li⫹
i /Ko and Lii /Lio
⫹ ⫹
exchange via the Na -K pump in rabbit ventricular myocytes (83). However, internal Li⫹ is a much less effective
activator than intracellular Na⫹. The pump current den⫹
sity elicited by 160 mM Lipip
was clearly lower than the
current density measured at high [Na⫹]pip. A K0.5 value for
Ip activation by [Li⫹]pip of 36 mM was found for guinea pig
ventricular myocytes superfused with Na⫹-free (choline⫹) medium containing 1 mM K⫹ (holding potential
⫺20 mV) (87). Half-maximal Ip activation was observed at
23 or 73 mM Lio⫹ in myocytes containing either 50 to 100
mM [Na⫹]pip or 100 mM [Li⫹]pip, respectively. Binding of
Lio⫹ to the pump is voltage sensitive. At the locus of their
binding, Li⫹ sense ⬃0.2 of the membrane potential across
the sarcolemma. Thus Lio⫹, like the other K⫹ congeners
and Ko⫹ itself, probably binds to the Na⫹-K⫹ pump at the
bottom of an access channel. Regions of positive and/or
negative slope persist in the cardiac Ip-V curve under
conditions where the intra- and extracellular cation binding sites of the Na⫹-K⫹ pump should be saturated by Li⫹.
Therefore, mechanisms apart from binding/rebinding of
the transported cations are voltage dependent and affect
the shape of the cardiac Ip-V relationship (87).
1799
1800
HELFRIED GÜNTHER GLITSCH
pump. It amounts to about ⫺60 kJ/mol in many animal
cells (31). The energy is used to transport 3 mol Na⫹ and
2 mol K⫹ against their respective concentration gradient
(osmotic work) and the electric charge of 1 mol Na⫹
against the electrical field across the cell membrane (electrical work). Thus the physiological active Na⫹/K⫹ transport proceeds as long as
o
i
兩⌬G ATP兩 ⬎ 3 ⫻ RT ln共aNa
/aNa
兲 ⫹ 2 䡠 RT ln共aiK/aoK兲 ⫺ F 䡠 Em (2)
E rev ⫽ ⌬GATP /F ⫹ 3ENa ⫺ 2EK
(3)
where ENa and EK stand for the Nernst potential of the
respective cation. From Equation 3, Erev can be calculated to be
⫹ 3 ⫻ 0.07 V ⫹ 2 ⫻ 0.09 V ⫽ ⫺0.232 V
(3a)
Thus under physiological conditions Erev is beyond the
membrane potentials that are experimentally accessible.
However, according to Equation 3, it is possible to shift Erev
to more positive membrane potentials by lowering ⌬GATP
and steepening the ionic gradients. ⌬GATP is given by
o
⌬G ATP ⫽ ⌬GATP
⫹ RT ln共关ADP兴 䡠 关Pi兴/关ATP兴兲
(4)
o
where ⌬GATP
denotes the standard free energy of ATP
hydrolysis and [ATP], [ADP], and [Pi] represent the intracellular concentration of adenosine triphosphate, adenosine diphosphate, and inorganic phosphate, respectively.
Thus a lower (less negative) ⌬GATP can be obtained by
increasing the [ADP] 䡠 [Pi]/[ATP] ratio in the cell studied.
B. Evidence for Backward Running Naⴙ-Kⴙ Pump
and Determination of the Naⴙ-Kⴙ Pump Erev
The procedure outlined above was first applied to squid
giant axons, and an inwardly directed Ip was demonstrated
(34). Under similar experimental conditions, an inwardly
directed Ip was measured in guinea pig ventricular myocytes
over a range of membrane potentials between ⫹40 and
⫺120 mV (6). Figure 6 presents some of the results . Figure
6A (bottom trace) demonstrates that the strophanthidin-
FIG. 6. Voltage dependence of inward
current generated by the backward-running
Na⫹-K⫹ pump. A: chart recording of membrane potential (top) and membrane current
(bottom); holding potential, ⫺40 mV. The
horizontal bar marks exposure to 0.5 mM
strophanthidin (str). The K⫹-free external
solution contained 5 mM Ba2⫹; the pipette
solution was Na⫹ and Cs⫹ free and contained 145 mM K⫹, 5 mM MgATP, 5 mM
Tris2ADP, and 5 mM phosphate. B: superimposed records of strophanthidin-sensitive
currents for 80-ms pulses to ⫹40, 0, ⫺60,
and ⫺100 mV, obtained by subtracting each
trace recorded in the presence of strophanthidin from the average of control traces
recorded during pulses to the same potential
just before and just after the exposure to
strophanthidin. C: whole cell current-voltage
relationships from the experiment in A, determined before (E), during (‚), and after (F)
exposure to strophanthidin. Ordinate, steady
current levels; abscissa, membrane potential.
D: current-voltage relationship of the backward-running Na⫹-K⫹ pump. Ordinate, steady
levels of the strophanthidin-sensitive currents
represented in B; abscissa, membrane potential. [Adapted from Bahinski et al. (6).]
Physiol Rev • VOL
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
where ao and ai represent the activities of the transported
ions at both sides of the cell membrane, Em denotes the
membrane potential, and R, T, and F have their usual
meanings. According to Equation 2, there should be a
membrane potential where ⌬GATP equals the energy required for the active cation transport. This potential is
called the reversal potential (Erev) of the Na⫹-K⫹ pump.
No active Na⫹/K⫹ fluxes occur at Erev, and Ip vanishes. If
the membrane potential becomes more negative than
Erev, the Na⫹-K⫹ pump runs backward producing an active Na⫹ influx into the cell and generating an inwardly
directed Ip. Erev can be derived from Equation 2 by
E rev ⫽ ⫺60 ⫻ 103 共V 䡠 A 䡠 s/mol兲/9.65 ⫻ 104 共A 䡠 s/mol兲
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
IV. VOLTAGE DEPENDENCE OF CARDIAC
PUMP CURRENTS
A. Cardiac Steady-State Ip-V Relationships
1. Basic characteristics of the Ip-V curve
Whole cell recording is a mode of the patch-clamp
technique (77) which permits a much better control of the
Physiol Rev • VOL
membrane potential and the ionic composition of the
intracellular compartment than earlier voltage-clamp
techniques applied to multicellular preparations. Gadsby
et al. (48) were the first to study the voltage dependence
of Ip by whole cell recording from isolated, single guinea
pig ventricular cells in Na⫹-containing solution after minimizing passive Na⫹, Ca2⫹, and K⫹ currents. According to
the authors, the Ip-V relationship of the myocytes is sigmoid in shape with a steep positive slope between ⫺100
and 0 mV, a less steep slope at more negative potentials,
and nearly no voltage dependence of Ip at positive membrane potentials (see also Fig. 7A, circles). A further
careful study by the authors confirmed these characteristics of the cardiac Ip-V curve (51). However, a region of
negative slope in the Ip-V curve was not observed. A
decrease of Ip with hyperpolarization in guinea pig ventricular cells was also reported in 1987 by others (123) in
abstract form. The pump current of adult isolated rat
cardiac myocytes displays a similar voltage dependence
as in guinea pig myocytes (167). This is also true for the
voltage dependence of Ip in rabbit ventricular cells dialyzed with a pipette solution containing 80 mM Na⫹ (79).
For unknown reasons, some authors were unable to detect any voltage dependence of Ip in guinea pig ventricular
myocytes (127). In the cardiac conducting system, the
Ip-V curve shows little voltage dependence at membrane
potenials positive to ⫺20 mV. The pump current declines
with hyperpolarization (66).
2. Effects of internal Na⫹ on the Ip-V relationship
of cardiac cells
Already in 1987 it was reported that lowering internal
[Na⫹] diminishes the pump current of guinea pig ventricular cells in Na⫹-rich solution and shifts the Ip-V curve to
the right (to more positive potentials), whereas the voltage dependence of Ip persists (50). However, in the absence of external Na⫹, lowering [Na⫹]pip from 50 to ⱕ17
mM scaled down the Ip-V relationship of these cells without a marked shift toward more positive potentials (131).
A similar shift of the normalized Ip-V curve to the right
and a reduction of Ip if [Na]pip was lowered from 50 to 5
mM occurred in sheep cardiac Purkinje cells superfused
with Na⫹-rich medium (66). Reducing [Na⫹]pip from 85 to
20 mM scales down the Ip-V relationship of rat myocytes
at 145 mM Nao⫹ without any shift of the Ip-V curve along
the voltage axis (167). It seems likely that [Na⫹]pip in
these experiments was not low enough to produce this
shift. By way of contrast, both a scaling down and a
rightward shift of the Ip-V relationship were observed in
guinea pig atrial myocytes superfused with Na⫹-free or
Na⫹-containing media (109). Recently, a scaling down and
a shift of the normalized Ip-V curve toward more positive
potentials were reported from measurements on guinea
pig ventricular cells in Na⫹-free solution after lowering
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
inhibited pump current Ip of a guinea pig ventricular myocyte is inwardly directed under the conditions chosen. As
shown in Figure 6B, Ip remains an inward current even at
positive potentials. Figure 6C displays I-V curves before,
during, and after application of the cardiac steroid, a specific
inhibitor of the Na⫹-K⫹ pump. The I-V relationships were
derived from the experiment illustrated in Figure 6A. Figure
6D exhibits the Ip-V curve of the backward running Na⫹-K⫹
pump obtained from the strophanthidin-sensitive current Ip
at various potentials, as illustrated in Figure 6B. Ip was small
at ⫹40 mV, increased with hyperpolarization, and reached
an apparent plateau level of about ⫺0.32 ␮A/cm2 near ⫺100
mV. By means of a similar approach, an inward Ip between
⫹30 mV and ⫺110 mV that did not reach a plateau at the
most negative membrane potential tested was observed in
cardiac Purkinje cells. The Ip density amounted to ⫺0.13
␮A/cm2 at ⫺95 mV (71). Furthermore, Erev shifted to more
positive potentials at less negative ⌬GATP values. Shifting
⌬GATP to less negative values also diminished Ip over the
entire voltage range studied. However, the concentration of
extracellular Cs⫹ required for half-maximal Ip activation
remained unchanged. Equation 3 predicts changes of Erev
by variation of the transmembrane gradients of the pumped
cations at constant ⌬GATP. This prediction was verified (69).
Flattening the ionic gradients increased Ip over the entire
voltage range studied and shifted Erev toward more negative
potentials. Conversely, steepening the gradients diminished
Ip and shifted Erev to more positive potentials.
The Ip-V curve of the backward running Na⫹-K⫹ pump
was also studied in internally dialyzed squid axons under
voltage clamp (143). There was a steady decline of Ip density
from an apparent plateau at ⫺80 to ⫺100 mV (⫺0.24 ␮A/
cm2) to practically zero at ⫹30 mV. In contrast to earlier
observations (34), a negative slope of the Ip-V relationship
was not found. In K⫹-free solution, an inwardly directed Ip
was measured in Xenopus oocytes with a reduced intracellular [Na] and an augmented [ADP] 䡠 [Pi]/[ATP] ratio. The Ip
density amounted to about ⫺0.1 ␮A/cm2 at ⫺100 mV without an apparent plateau and declined with depolarization.
Under these conditions the Ip reversal potential was obviously at positive membrane potentials (38).
In summary, the studies demonstrate under suitable
conditions in various cell species a backward running
Na⫹-K⫹ pump generating an inward Ip, in line with thermodynamic considerations.
1801
1802
HELFRIED GÜNTHER GLITSCH
[Na⫹]pip from 50 to ⱕ5 mM (main cation in pipette: TEA⫹
or NMDG⫹) (9). Some of the findings are displayed in
Figure 7, B and C. Lowering [Na⫹]pip from 50 to 5 mM or
below reveals the effect of [Na⫹]pip on the cardiac Ip-V
curve in myocytes containing NMDG⫹ (Fig. 7B) or TEA⫹
(Fig. 7C) as main cation. The relative Ip activation at low
[Na⫹]pip increases with depolarization, i.e., the K0.5 value
for Ip activation by [Na⫹]pip declines (Fig. 3). In Na⫹-free
pipette solutions, Ip is absent in the entire voltage range
tested.
In summary, reducing [Na⫹]pip to values well below
the K0.5 value reveals not only a scaling down of the Ip-V
relationship in cardiac cells but also a rightward shift of
the Ip-V curve to more positive membrane potentials. The
effect is observed in Na⫹-free and Na⫹-containing solution. The reason why Nakao and Gadsby (131) did not
detect the shift of the Ip-V curve to the right at low
[Na⫹]pip and [Na⫹]o remains unclear. The shift may be
produced by the voltage sensitivity of internal Na⫹ binding to the Na⫹-K⫹ pump. At [Na⫹]pip, lower than saturating concentrations, negative potentials will inhibit the
binding of internal Na⫹ to the pump and thereby the
activation of Ip. As a consequence, the normalized Ip-V
relationship is shifted toward more positive membrane
potentials. However, without additional information it
cannot be excluded that the voltage dependence of interPhysiol Rev • VOL
nal Na⫹ binding is only apparent and is due to the voltage
sensitivity of any slower step subsequent to Na⫹ binding
in the pump cycle.
3. The voltage dependence of Ip varies with the
extracellular Na⫹ concentration
The sigmoid shape of the cardiac Ip-V relationship in
Na⫹-containing solution is characterized by a steep positive slope of the curve at negative membrane potentials
up to ⫺100 mV and a less positive slope at even more
negative potentials. Lowering [Na⫹]o reduces the voltage
dependence of Ip in cardiac cells in a concentrationdependent manner, as first reported from measurements
on isolated guinea pig ventricular myocytes (50, 131).
Figure 7A shows, for example, that Ip of a guinea pig
ventricle cell in Na⫹-poor solution (1.5 mM Nao⫹) is clearly
less voltage dependent than in Na⫹-rich control medium
(150 mM Nao⫹). The effect of Nao⫹ on the Ip-V curve was
also observed in rabbit cardiac Purkinje cells (15), rat
ventricular myocytes (167), Xenopus oocytes (158), and
squid axons (144). Most probably, a step late in the Na⫹
translocation by the pump is affected by Nao⫹. The Nao⫹
effect on the potential dependence of Ip is voltage dependent (131). The K0.5 value (i.e., [Na⫹]o exerting the halfmaximum effect) increases fourfold from 91 to 357 mM
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
⫹
⫹
FIG. 7. The effect of Nao and Napip on
the Ip-V relationship of guinea pig ventricular myocytes. A: Na⫹-K⫹ pump current-voltage relationships determined in a single cell
⫹
⫹
at 50 mM Na⫹
i , 5.4 mM Ko , and [Na ]o of 150
mM (E), then 1.5 mM (‚), and then 150 mM
again (F). Steady current levels were measured near the end of 160-ms voltage pulses
from the holding potential ⫺40 mV. Pump
currents were obtained by subtracting currents recorded in the presence of 2 mM strophanthidin from the average of currents recorded just before, and just after, the brief
exposure to strophanthindin. [Adapted from
Gadsby and Nakao (50).] B and C: Ip-V relationships at various [Na⫹]pip in Na⫹-free solution. Na⫹-K⫹ pump currents normalized to
cell capacitance are plotted versus membrane potential. B: Ip-V relationships at 0
mM (ƒ), 0.5 mM ({), 2 mM (‚), 5 mM (E), and
⫹
50 mM (䊐) Napip
in cells dialyzed with
NMDG⫹-containing pipette solution (n ⫽ 4 –
8). C: corresponding Ip-V curves for myocytes containing TEA⫹ as the principal cation (n ⫽ 4 –7). Symbols have the same
meaning as in B. Note the absence of Ip in
Na⫹-free pipette solutions and the voltage
⫹
dependence of Ip at low (ⱕ5 mM) Napip
. In
⫹
contrast, Ip activated by 50 mM Napip
is voltage insensitive. [From Barmashenko et al.
(9). Copyright 1999 The Physiological Society.]
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
1803
4. Effect of K ⫹
o on the Ip-V relationship
In contrast to the sigmoid Ip-V relationship of cardiac
cells (48), endogenous Na⫹-K⫹ pumps of Xenopus oocytes (110, 158) and Torpedo pumps expressed in Xenopus oocytes (157) display Ip-V curves exhibiting a maximum and a region of negative slope positive to ⫹20 mV.
The authors concluded that their measurements suggest
at least two voltage-dependent partial reactions oppositely affected by the membrane potential in the Na⫹-K⫹
pump cycle (110). It was considered that the region of
negative slope in the Ip-V curve might possibly be an
artifact caused by passive currents overlapping Ip (146). A
later study finally established beyond any reasonable
doubt that the Ip-V relationship of Xenopus oocytes
shows a region of negative slope (147). According to the
authors, the negative slope of the Ip-V curve is most
probably due to voltage-dependent binding of Ko⫹ to the
Na⫹-K⫹ pump. The extent of the region depends on the
experimental conditions chosen.
Since then, Ip-V curves displaying a region of negative slope were also observed under special conditions in
cardiac cells. A region of negative slope in the Ip-V curve
of rabbit cardiac Purkinje cells was measured in Na⫹containing media at low [K⫹]o (less than K0.5 value) and
positive membrane potentials. The region extends to negative voltages in Na⫹-free solution containing low concen⫹
tration of K⫹ or its congeners Tl⫹, NH⫹
4 , and Cs (15, 18).
An example is presented in Figure 8. Ip-V curves of rabbit
Purkinje cells in Na⫹-free (choline) solution containing
10.8 or 0.05 mM K⫹ are shown. Figure 8A displays Ip-V
curves from three cells. Ip is presented in absolute values
Physiol Rev • VOL
⫹
FIG. 8. Ip-V relationships of Purkinje cells in Na -free, cholinecontaining media with different K⫹ concentrations. A: mean Ip-V curves
of cells superfused with solutions containing 0.05 mM K⫹ with (F) or
without (E) 0.2 mM dihydroouabain (DHO) or 10.8 mM K⫹ with (■) or
without (䊐) 0.2 mM DHO. Ip is expressed in pA (n ⫽ 3). B: mean Ip-V
relationships of cells in bathing fluids containing either 0.05 (E) or 10.8
(䊐) mM K⫹. Ip normalized to the corresponding Ip amplitudes at ⫺20 mV
(n ⫽ 5–9). [From Bielen et al. (15). Copyright 1991 The Physiological
Society.]
(pA). Figure 8B exhibits normalized mean Ip-V relationships. In Figure 8, A and B, the Ip-V curves of cells in
K⫹-poor solution have an extended region of negative
slope. The solid symbols in Figure 8A indicate that the
currents measured are blocked by dihydroouabain (DHO)
and, therefore, represent Ip. Similarly, an extended region
of negative slope in the Ip-V relationship of guinea pig
ventricular myocytes was observed at low [Na⫹]o and
[K⫹]o (52). Furthermore, an increasingly negative slope of
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
Nao⫹ between ⫺120 and ⫺40 mV. The authors were able to
fit their Ip-V data obtained at various [Na⫹]o by a kinetic
equation derived within the framework of the pseudo
two-state model of Na⫹-K⫹ pumping (81; see sect. IVA6).
The mechanism by which Nao⫹ affects the voltage dependence of Ip is probably voltage-sensitive Nao⫹ binding to
the Na⫹-K⫹ pump. The voltage dependence is not a characteristic of the binding per se but is rather due to a
voltage-dependent variation of [Na⫹] at the bottom of a
“narrow, high-field access channel” (see Ref. 111, p. 75–
83) where Nao⫹ binding to the pump occurs (147). The
increasing inhibiton of Ip by Nao⫹ at increasingly more
negative potentials is probably not mediated via an increasing unidirectional Na⫹ influx as a consequence of
enhanced voltage-dependent Nao⫹ binding. The reason is
that both unidirectional active 22Na efflux through the
pump and pump current in the squid giant axon exhibit the
same voltage dependence in Na⫹-free and Na⫹-containing
artificial sea water (144). Of course, the voltage dependence
of efflux and current is steeper in the latter solution. Thus an
almost irreversible partial reaction seems to be present in
the Na⫹ transport limb of the Na⫹-K⫹ pump.
1804
HELFRIED GÜNTHER GLITSCH
the Ip-V curve of rat ventricular cells was measured in
Na⫹-containing media at positive membrane potentials as
[K⫹]o was lowered (167).
5. The 3Na⫹:2K⫹ stoichiometry of the Na⫹-K⫹ pump
is voltage independent
6. Interpretation of Ip-V relationships
The sigmoid shape of the Ip-V curve in guinea pig
ventricular myocytes (48) was interpreted in the following way (51). Since the pump rate (and therefore Ip) is
nearly voltage independent at positive membrane potentials, a voltage-independent partial reaction rate limits the
pump cycle in this potential range. At more negative
potentials a step with voltage-sensitive transition rates
limits the cycle rate, either directly, or via the control of
the level of an intermediate that enters the rate-limiting
partial reaction (see also Ref. 33) . In view of the experimental data only a single voltage-dependent partial reaction was assumed. Several studies on cells and cell-free
systems (see Refs. 33, 142) suggested that the Na⫹-extruding limb of the pump cycle contains this voltage-sensitive
step. More precisely, it seems likely that the deocclusion
and release/rebinding of Na⫹ to/from the extracellular
space are voltage dependent (see Ref. 4). This step in turn
controls the concentration of an intermediate participating in the rate-limiting partial reaction, which, under the
experimental conditions chosen, is probably the K⫹ translocation to the cell interior (see Ref. 33). Thus the shape
of the Ip-V relationship in guinea pig ventricle cells was
explained by assuming that depolarization of the sarcolemma enhances the voltage-dependent step of the Na⫹
translocation such that the voltage-insensitive K⫹ import
becomes rate limiting for the Na⫹-K⫹ pump cycle at positive membrane potentials (6).
The quantitative kinetic analysis of the data (51) was
based on a model mentioned above (81). According to this
Physiol Rev • VOL
(Scheme 1)
where E1 and E2 represent the two states of the pump
cycle, ␣ and ␤ denote empirical voltage-dependent rate
constants, and c and d signify lumped empirical rate
constants that are voltage insensitive. If an asymmetrical
Eyring barrier exists for charge translocation (see Ref.
111, p. 68 – 69) and a single charge is moved in the pump
cycle (130), the voltage-sensitive rate constants are given
by ␣ ⫽ ␣0 䡠 exp[␦ 䡠 V 䡠 F/R 䡠 T] and ␤ ⫽ ␤0 䡠 exp[⫺(1
⫺ ␦)V 䡠 F/R 䡠 T], where ␣0 and ␤0 represent the forward
and backward rate constant at zero potential, respectively; ␦ indicates the location of the barrier in the cell
membrane; and V, F, R, and T have their usual meaning.
By assuming numbers for the rate constants suggested by
the experimental data, and ␦ ⫽ 0.1, it was possible to fit
the turnover rate-voltage relationship derived from the
Ip-V curve by means of an equation deduced on the basis
of the pseudo two-state model for the pump cycle (51).
As mentioned above, this interpretation assumes a
single voltage-sensitive partial reaction in the Na⫹-K⫹
pump cycle. However, to understand Ip-V curves displaying a region of negative slope, an additional potentialdependent step in the pump cycle has to be postulated
(147). It was hypothesized that binding of extracellular K⫹
to the Na⫹-K⫹ pump might occur within the cell membrane. To reach their binding sites, the K⫹ have to cross
a part of the electrical field over the membrane. As a
consequence, the membrane potential affects the local K⫹
concentration at the binding sites. Depolarization decreases the K⫹ concentration at the sites and thereby
diminishes the pump cycle rate and Ip. This mechanism
produces the negative slope of the Ip-V relationship. The
reader is referred to a brilliant, introductory review (Ref.
111, p. 75– 83) for a detailed discussion of this “high field,
narrow access channel hypothesis.” Some early observations (6, 74) suggested that the steps of K⫹ translocation
by the pump beyond K⫹ binding may be voltage independent, and recent experimental evidence (135) supports
this view.
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
For an adequate interpretation of the shape of Ip-V
relationships it is essential to know whether the 3Na⫹:
2K⫹ stoichiometry of the Na⫹-K⫹ pump is voltage dependent. Two studies of Ip-V curves in isolated noncardiac
cells have provided relevant informations. Simultaneous
measurements of 22Na efflux and pump current revealed
that the Na⫹:K⫹ stoichiometry of Torpedo californica
pumps expressed in Xenopus oocytes is voltage independent between ⫹50 and ⫺100 mV (157, 179). Comparable
experiments on squid giant axons also demonstrated that
the Na⫹:K⫹ stoichiometry is voltage independent (144).
Both the pumped Na⫹ efflux and the Na⫹-K⫹ pump current declined by roughly the same amount upon hyperpolarization. This voltage dependence implies an effect of
membrane potential on Na⫹ release/rebinding at the extracellular face of the cell membrane rather than on a
hypothetical reverse Na⫹ transport via Na⫹-K⫹ pump.
model for the interpretation of Ip-V curves, any multistate
(unbranched) pump cycle containing a single voltagesensitive step can be treated as a pseudo two-state cycle
(scheme 1)
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
1805
B. Transient Pump Currents
1. Identification of transient pump currents during
electroneutral Na⫹/Na⫹ exchange in cardiac cells
Physiol Rev • VOL
FIG. 9. Voltage dependence of charge movement (A) and of decay
rate constant (B) during transient, strophanthidin-sensitive currents
(inset in A). Holding potential, ⫺40 mV; temperature, 35°C. A, inset:
superimposed sample records of strophanthidin-sensitive currents elicited by pulses to ⫹60, ⫹20, ⫺20, and ⫺60 mV; graph, time integral [Q(V)]
of current transients initiated by each pulse, plotted against pulse potential (V). E, Charge measured directly from the records; F, charge
estimated by extrapolation of the exponential current decay. Only “on”
charge movement is plotted. “Off” and “on” charge movements appear to
be the same. The smooth curve, derived from the Boltzmann relation,
shows ⌬Q(V)/⌬Qmax ⫽ 1/[1 ⫹ exp(V⬘ ⫺ V)/k], where ⌬Q(V) ⫽
Q(V) ⫺ Qmin, ⌬Qmax ⫽ Qmax ⫺ Qmin; V⬘ is the potential at the midpoint,
and k provides a measure of the equivalent charge. Vertical bar, 250 pA;
horizontal bar, 25 ms. B: rate constants of exponential fits to transient
currents plotted against membrane potential: E, rate constants of “on”
transients; F, rate constants of “off” transients, at ⫺40 mV, following
repolarization from ⫺60 mV ⬍ V ⬍ ⫹80 mV. C: simplified kinetic scheme
for ATP-dependent Na⫹/Na⫹ exchange emphasizing the voltage-sensitive step A7B, governed by voltage-dependent forward and backward
rate constants, ␣ (V) and ␤ (V). [Adapted from Nakao and Gadsby (130).]
2. Hypothetical mechanism of the transient
pump currents
The characteristics of the transient pump currents
were explained by the hypothesis that the Na⫹ binding
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
To identify voltage-dependent partial reactions in the
pump cycle, it is extremely helpful to constrain the active
cation transport to only a few steps of the cycle. The
physiological Nao⫹/K⫹
exchange via the pump can be
i
blocked by application of K⫹-free solution and thereby
the transporter is forced to carry out Nao⫹/Na⫹
i exchange
(130). Under these conditions a steady-state pump current
is absent, i.e., the Na⫹/Na⫹ exchange is electroneutral.
Nevertheless, cardiac glycoside-sensitive transient currents were observed after abrupt changes in membrane
potential. These currents were outwardly directed upon
depolarization and inwardly directed upon hyperpolarization. They declined monoexponentially to zero current
(Fig. 9A, inset). They were blocked by Na⫹-free internal
or external media, cardiac glycosides, or oligomycin B
and required metabolic energy. The characteristics of the
currents strongly suggest that they are generated by the
Na⫹ pump and represent transient pump currents. Since
the physiological Na⫹/K⫹ exchange mode of the pump
does not exist under the experimental conditions chosen,
the observed charge translocation most probably occurs
in the Na⫹ limb of the Na⫹-K⫹ pump (Fig. 9C). More
specifically, it seems likely that the charge translocating
step is Na⫹ release/rebinding to/from the exterior,
whereas binding and occlusion of intracellular Na⫹ do not
cause charge translocation under comparable conditions
(5). The voltage dependence of the charge moved by the
transient pump currents could be fitted by a BoltzmannFermi relation assuming that a single positive electrical
charge crosses the entire membrane dielectric (Fig. 9A).
The rate constant of transient current decline depends on
membrane potential. It remains nearly constant at positive membrane potentials and increases steeply at negative voltages (Fig. 9B). The same rate constant was found
for the electrogenic Na⫹ translocation by analyzing transient pump currents at 24°C and zero potential when
either voltage or ATP concentration jumps were applied
to the same giant patch from a guinea pig ventricular
myocyte (43). The possible effect of ␤-adrenergic stimulation on the transient pump current was studied in adult
rat cardiac myocytes (35). According to the authors,
10 –50 ␮M norepinephrine did not alter the total charge
transferred during the transient pump current. The density of pump molecules was calculated from the maximal
charge moved during the transient current. It remained
unchanged in the presence of the drug. Most recently, an
altered voltage dependence of the charge transfer by the
the Na⫹/K⫹ pump was observed upon application of forskolin to guinea pig ventricular cells (108; see sect. VIIIA1).
1806
HELFRIED GÜNTHER GLITSCH
Physiol Rev • VOL
potential steps revealed additional current components
probably related to Na⫹ release and rebinding from or to
the E2P Na3/E2P Na2 intermediates of the pump cycle
(92). Similarly, measurements on squid axons succeeded
in differentiating at least three components of transient
pump current which might be related to Na⫹ release to
the exterior (ultrafast component) and two distinct transitions involved in the Na⫹ deocclusion/release (142).
High-speed voltage jumps revealed that the pre-steadystate charge movements relax in three phases that represent the deocclusion and release of the three Na⫹ in a
strictly sequential order (95).
3. Identification and properties of transient pump
currents during electroneutral K⫹/K⫹ exchange
Transient pump currents might likewise occur under
conditions that constrain the Na⫹-K⫹ pump to Ko⫹/K⫹
i
exchange. Although transient currents in guinea pig ventricular myocytes were observed when the pump carried
out forward (physiological) or backward Na⫹/K⫹ transport or Na⫹/Na⫹ exchange, they were never recorded
under conditions permitting only Ko⫹/K⫹
i exchange (6).
Therefore, the authors concluded that K⫹ translocation is
not associated with a net charge transport across the
membrane dielectric and is, most probably, voltage independent. Furthermore, they made the interesting point
that the transient pump currents observed during forward
and backward Na⫹/K⫹ transport and Na⫹/Na⫹ exchange
indicate that the forward and the backward mode of
Na⫹/K⫹ transport is rate-limited not by a voltage-dependent partial reaction but by a voltage-insensitive step of
the pump cycle, probably K⫹ translocation. However, the
data presented were, on the one hand, consistent with the
notion that the actual E1K2 7 E2K2 K⫹ translocation is
probably electroneutral, but, on the other hand, they did
not exclude voltage-dependent Ko⫹ binding to the pump,
since the experiments were conducted at saturating [K⫹]o
(Ref. 111, p. 216). In fact, ouabain-sensitive transient
(pump) currents were more recently measured under Ko⫹/
⫹
⫹
K⫹
i exchange conditions at [K ]o (or [Tl ]o) submaximally activating the pump of rat ventricular myocytes
(135). The currents were observed at the beginning and
immediately after the end of steep changes in membrane
potential. The maximum charges moved during “on” and
“off” transients amounted to 9.6 and 9.1 fC/pF with an
effective valency of 0.48 (compare this value to the effective valency of 1.0 derived from the Q-V curve for Na⫹/
Na⫹ exchange, Ref. 130). The rate constant of current
decline remained constant at positive membrane potentials but increased with increasingly more negative potentials, reminiscent of characteristics of the kinetics of transient pump currents during Na⫹/Na⫹ exchange (130) (see
Fig. 9B). The transient pump currents seen under conditions of K⫹/K⫹ exchange were absent in K⫹-free (Tl⫹-
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
sites of the pump molecule provide two negative charges
and thus exhibit one positive charge if three Na⫹ are
bound (130). As a consequence, the rate of Na⫹ translocation by the pump is expected to be voltage sensitive.
Since a positive charge is moved, in general the forward
rate constant of the translocation should be increased
upon depolarization, whereas the backward rate constant
should be diminished. Vice versa, a hyperpolarization of
the cell membrane should reduce the forward rate constant and augment the backward rate constant. Thus the
steady-state concentrations of the enzyme intermediates
participating in the translocation step depend on membrane potential. After a voltage jump, the rate constants
immediately reach their new values, whereas the concentrations of the intermediates approach their new steadystate more slowly. During this approach the forward and
backward fluxes of the intermediates differ and cause a
transient pump current that declines monoexponentially
with a rate constant that is the sum of the new forward
and backward rate constants at the new membrane potential. The rate constant of decline of the transient pump
current increases steeply with hyperpolarization but is
nearly constant at positive potentials (Fig. 9B). Since at
positive voltages the backward rate constant of the translocation step should be small, the rate constant of current
decline will be dominated by the forward rate constant,
which seems to be nearly voltage independent. Thus the
observed potential dependence of the rate constant of
current decline is largely due to the voltage dependence
of the backward rate constant. By comparing the forward
and backward rate constants near zero potential with data
in the literature, the authors suggested that the studied
voltage-dependent step of the pump cycle might be the
translocation/deocclusion E1P (Na3) 7 E2P Na3. Furthermore, they discussed the possibility that the voltage dependence of this partial reaction might fully account for
the electrogenicity and voltage sensitivity of the physiological Na⫹/K⫹ exchange carried out by the Na⫹-K⫹ pump
(130). By application of a giant-patch technique (90) to
guinea pig ventricular myocytes, two components of transient pump current were distinguished, suggesting Na⫹
release/rebinding from/to E2P in two partial reactions
with different rate constants, amounts of charge Q moved
and slope factors of the corresponding Boltzmann-Fermi
equations (91). Both Qfast (within the first 100 ␮s during a
potential step) and Qslow (in the millisecond range) depend on [Na⫹]o. These observations were interpreted to
mean that the Na⫹ binding sites of the pump open to the
exterior in two voltage-sensitive partial reactions. Deocclusion of the first Na⫹ rate limits the strongly voltagedependent release of this ion constituting Qslow, whereas
a further electroneutral conformational change enables
the last two Na⫹ to dissociate from their binding sites in
a weakly voltage-sensitive partial reaction (Qfast). The
further analysis of transient pump currents during 50-␮s
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
1807
V. DEPENDENCE OF PUMP CURRENT
ON INTRACELLULAR ATP
The molecular basis of the Na⫹-K⫹ pump is the Mg2⫹dependent, Na⫹-K⫹-activated ATPase. Thus one would
expect that the activity of the pump, including the generation of Ip, depends on intracellular ATP. In 1983 it was
verified that in dog cardiac preparations both Na⫹-K⫹
pump and sarcolemmal Na⫹-K⫹-ATPase display identical
dependencies on Na⫹ and ATP (136). The ATP concentration required for half-maximum pump activation in the
sarcolemmal vesicles studied was estimated to be ⬃210
␮M. From studies on the steady-state Ip in isolated cardiac Purkinje cells it seems likely that the Na⫹-K⫹ pump
is preferentially fuelled by glycolytic ATP synthesis (70),
although other ATP sources contribute to the ATP supply
for the cardiac Na⫹-K⫹ pump (see Ref. 141). The pump
current of giant patches from guinea pig ventricular cells
is half-maximally activated at ⬃80 ␮M cytosolic MgATP
(93). Similarly, a saturable ATP dependence of Ip with a
Michaelis constant (Km) value of ⬃150 ␮M was observed
in giant patches from rat and guinea pig ventricular myocytes (42). Figure 10 illustrates these findings. Figure 10A
displays outward (pump) currents (bottom trace) evoked
by various MgATP concentrations (top trace) at 0 mV in a
giant patch from a rat ventricular myocyte. The pump
current increases with increasing [MgATP]. Figure 10B
shows normalized mean Ip values as a function of
[MgATP]. The experiments were carried out on giant
patches from rat and guinea pig ventricular cells under
Physiol Rev • VOL
FIG. 10. Recording of patch current in response to superfusion of
the patches with solutions containing different MgATP concentrations
as indicated in the top panel of A. At the beginning and the end of the
current recording, a 10-mV voltage pulse was applied to test the seal
resistance. Excised patch was from a rat ventricular myocyte. Temperature was 24°C. B: dependence of the normalized patch current on ATP
concentration; data plotted are means ⫾ SD. ■, Data from rat cells,
Na⫹-free, NMDG-containing pipette solution including 1 mM CsCl instead of 5 mM KCl; 䊐, data from rat cells, Na⫹-free, NMDG-containing
pipette solution including 5 mM KCl; }, data from guinea pig myocytes
with pipette solution containing NMDG and KCl. Included is a leastsquares fit of a Michaelis-Menten equation to the data (■) (dashed line),
leading to a Vmax of 1.26 ⫾ 0.06 (current at 600 ␮M ATP was set to 1.0)
and an apparent Km of 165 ⫾ 20 ␮M ATP. Apparent Km for (䊐) data is
146 ⫾ 23 ␮M ATP, and for data from guinea pig cells (}) is 163 ⫾ 13 ␮M
ATP. Temperature ⫽ 24°C. [Adapted from Friedrich et al. (42).]
various ionic conditions at zero potential. The dashed
curve represents the fit of a Michaelis-Menten equation to
the data. The K0.5 values were estimated to be 146 to 165
␮M MgATP and suggest ATP binding to the low-affinity
ATP binding site of the Na⫹-K⫹ pump. A strong reduction
of the transient pump currents evoked by voltage jumps
was noted in guinea pig cells upon internal perfusion with
ATP-poor or -free solutions (130). ATP concentration
jumps induced by photolytic ATP release from caged ATP
activate cardiac Ip. The kinetics of the transient pump
currents induced by ATP jumps or voltage jumps were
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
free) external solution or at 2 mM Tlo⫹ (135), a [Tl⫹]o
saturating the external cation binding sites of the pump
(18). These properties of the transient currents strongly
suggest voltage-dependent Ko⫹(Tlo⫹) binding to the pump.
Lowering [Tl⫹]o from 0.1 to 0.025 mM did not affect the
rate constant of current decline at positve potentials but
decreased the rate constant at negative membrane potentials. Again, this finding is reminiscent of the effect of
lowering [Na⫹]o on the rate constant of current decline
during Na⫹/Na⫹ exchange (52). By comparing the potential- and temperature-dependent characteristics of the
transient Ip under conditions of Na⫹/Na⫹ and K⫹/K⫹ exchange, it was concluded that binding of extracellular K⫹
and Na⫹ to the Na⫹-K⫹ pump occurs at different sites in
the ATPase or to different conformations of the enzyme
(135). The maximal charge moved, the effective valency,
and the activation energy of the transient charge movements are all larger under Na⫹/Na⫹ exchange conditions.
Most probably, the voltage-dependent Nao⫹ and Ko⫹ binding to the Na⫹-K⫹ pump is the main mechanism underlying the voltage dependence of the active Na⫹/K⫹ transport in cardiac cells containing a high intracellular Na⫹
and K⫹ concentration.
1808
HELFRIED GÜNTHER GLITSCH
studied in giant patches from guinea pig cardiomyocytes
superfused with K⫹-free media (43). The rate constants of
the fast component from the transient pump currents
evoked by an ATP concentration jump were very similar
to the rate constants of the transient pump currents induced by voltage jumps to zero or positive potentials
(⬃200 s⫺1 at 24°C) and displayed the same activation
energy. Furthermore, the two techniques revealed a comparable amount of charge moved during the transient
pump currents in the same giant patch. This result implies
that the same charge-carrying mechanism was studied by
both techniques.
Because sarcolemmal Na⫹-K⫹ transport by the pump
is performed by a vectorial enzyme reaction, a marked
temperature dependence of the Na⫹-K⫹ pump activity is
anticipated. As to the steady-state Ip of single rabbit sinoatrial node cells, a Q10 value of 2.1 in a temperature range
between 25 and 37°C has been reported (151). This means
a change of the Ip amplitude by a factor of 2.1 following an
alteration of the temperature by 10°C. Similar Q10 values
were found for the steady-state Ip of isolated sheep cardiac Purkinje cells (Q10 ⫽ 2.9) and single guinea pig
ventricular myocytes (Q10 ⫽ 2.2) (187).
Concerning transient pump currents, the rate constants of current decline during Nao⫹/Na⫹
i exchange in
guinea pig ventricle cells exhibited a voltage-independent
Q10 value of ⬃3 (130). In voltage-jump experiments on
excised giant patches from guinea pig ventricular myocytes, the activation energy for the forward rate constants
of the transient Nao⫹/Na⫹
i pump currents was estimated to
be ⬃80 kJ/mol. Nearly the same value (84 kJ/mol) was
observed for the rate of the electrogenic step (or an
electroneutral partial reaction preceding and rate-limiting
the faster electrogenic step) in these patches during ATP
concentration jumps (43). Interestingly, an Arrhenius plot
yielded an activation energy of 27.5 kJ/mol for the off-rate
constants of transient pump currents observed during
Tlo⫹/K⫹
i exchange in rat cardiac ventricular myocytes. A
clearly higher activation energy of 87 kJ/mol was found
for the off-rate constants of the current transients during
Nao⫹/Na⫹
i transport via the pump in the same cell species
(135). Therefore, it was concluded that the Nao⫹- and
Tlo⫹-evoked charge movements do not reflect the characteristics of the same partial reaction in the Na⫹-K⫹ pump
cycle. It seems likely that different conformational states
of the Na⫹-K⫹-ATPase control the rate of the electrogenic
transitions and/or the binding of Nao⫹ and Ko⫹ to the pump
(135).
Physiol Rev • VOL
A. Contribution of Electrogenic Naⴙ-Kⴙ Pumping
to the Cardiac Resting Potential
1. Theoretical note
The Na⫹-K⫹ pump directly generates a potential
difference across the cell membrane. Under physiological conditions this potential difference Ep hyperpolarizes the cell membrane and, therefore, increases the
absolute value of the resting potential. According to
Ohm’s law, the magnitude of Ep depends on both the Ip
amplitude and the membrane resistance. The Ip amplitude is determined by the pump molecule density in the
cell membrane and by the turnover rate of the pump.
The membrane resistance depends on the density and
characteristics of mainly the ionic channels in the membrane. Accordingly, the contribution of electrogenic
Na⫹-K⫹ pumping to the resting potential of various
cells varies widely. There are animal cells in which the
resting potential is predominantly generated by the
Na⫹-K⫹ pump. These cells include T lymphocytes of
mice (100), rat mast cells (21), and vomeronasal chemoreceptor neurons of the frog (177). In cells where
the passive Na⫹ and K⫹ fluxes crucially contribute to
the setting of the resting potential, the effect of electrogenic Na⫹-K⫹ pumping on the resting potential (Em)
may be described by the Mullins-Noda equation (129)
Em ⫽
R 䡠 T PNa 䡠 关Na⫹兴o ⫹ r 䡠 PK 䡠 关K⫹兴o
ln
F
PNa 䡠 关Na⫹兴i ⫹ r 䡠 PK 䡠 关K⫹兴i
(5)
where r denotes the coupling ratio between the pumped
Na⫹ and K⫹ fluxes (r ⫽ 1.5 under physiological conditions) and the other symbols have their usual meaning.
According to Ascher (see Ref. 176), the potential difference Ep set up by the Na⫹-K⫹ pump in these cells cannot
exceed
Ep ⫽
R䡠T 1
ln
F
r
(6)
or 10 to 11 mV in resting cells. Although the passive
Na⫹ and K⫹ fluxes are certainly determining factors for
the cardiac resting potential, it seems questionable
whether the Mullins-Noda equation adequately describes the contribution of electrogenic Na⫹-K⫹ pumping to the cardiac resting potential. The assumption
implied in the equation of voltage-independent Na⫹ and
K⫹ permeabilities probably does not hold for cardiac
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
VI. TEMPERATURE DEPENDENCE OF STEADYSTATE AND TRANSIENT PUMP CURRENTS
VII. SIGNIFICANCE OF ELECTROGENIC
SODIUM-POTASSIUM PUMPING FOR
THE MEMBRANE POTENTIAL OF
CARDIAC CELLS
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
cell membranes. More specifically, due to the typical
N-shaped steady-state I-V relationship of heart cells,
the inhibition of the Na⫹-K⫹ pump probably causes a
much larger depolarization of the sarcolemma than
the predicted 10 mV (45). Similarly, the strong inward
rectification of the cardiac steady-state I-V curve might
evoke a considerably stronger depolarization upon
pump blockade than expected from the slope conductance of the sarcolemma at the resting potential (10).
2. Experimental data
B. Importance of Electrogenic Naⴙ-Kⴙ Pumping
for the Cardiac Action Potential
As mentioned above, the potential difference Ep
contributed by the electrogenic Na⫹-K⫹ pump to the
cardiac membrane potential depends on the Ip amplitude and the membrane resistance. Compared with the
values at resting potential, both factors are increased at
the plateau of the cardiac action potential. The higher
membrane restistance is mainly due to a reduced potassium conductance at the plateau level (see Ref. 132,
p. 28). The cardiac Ip-V relationship predicts an increase of Ip by a factor ⬃2 at the action potential
plateau (15, 131) (see also Fig. 7A). Furthermore, the
enhanced Na⫹ influx during an action potential additionally augments Ip via an increased subsarcolemmal
Na⫹ concentration. For these reasons one would expect that electrogenic Na⫹-K⫹ pumping markedly affects the shape of the cardiac action potential during
the plateau phase. Indeed, in the first voltage-clamp
study of the cardiac Ip in a multicellular preparation, a
prolongation of the action potential was observed in
Purkinje fibers as an early effect of the Na⫹-K⫹ pump
inhibition by the cardiac glycoside DHO (99). Vice
versa, the activation of the pump in Purkinje fibers
following a short period of increased stimulation frequency or in K⫹-free solution considerably shortens the
action potential duration (47).
In isolated cardiac cells too, an early effect of
cardiac glycosides is a lengthening of the action potential. Figure 11 shows an example. The action potential
of an isolated guinea pig ventricular myocyte is prolonged by ouabain in a concentration-dependent manner. The higher the concentration applied, the stronger
is the Na⫹-K⫹ pump inhibition and the longer is the
action potential duration (at 90% repolarization; APD90)
since the repolarizing Ip is increasingly diminished.
Comparable effects of other cardiac steroids have been
reported (e.g., Ref. 113). The effects are in accordance
with calculations based on the membrane resistance
and Ip amplitude at the plateau level of the action
potential (113, 115, 116). These calculations predict a
⫹
⫹
FIG. 11. Early effects of Na -K pump inhibition on
the action potential of a guinea pig ventricular myocyte.
Within 10 s after application, ouabain, a specific inhibitor
of the Na⫹-K⫹-ATPase, prolongs the action potential at
90% repolarization (APD90) and raises the action potential
plateau. Note the depolarization of the resting sarcolemma. Temperature ⫽ 35°C. The effects are concentration dependent. (From S. Zillikens and H. G. Glitsch, unpublished data.)
Physiol Rev • VOL
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
Experimental estimations of the pump contribution
to the cardiac resting potential in multicellular preparations yielded 5–10% of the resting potential (reviewed in
Refs. 40, 65). As to single cells, an Ep of 4.2 mV was
derived contributing to the resting potential of canine
cardiac Purkinje cells (25). In chick cardiac myocytes, Ip
hyperpolarizes the sarcolemma by 6.5 mV (170). An Ep
value of only 0.4 mV was reported from measurements on
guinea pig ventricular myocytes (114). A much higher Ep
of ⬃20 mV can be calculated from data obtained in experiments on isolated rabbit sinoatrial node cells (151).
Although, in general, the contribution of electrogenic
Na⫹-K⫹ pumping to the cardiac resting potential amounts
to only a few millivolts, it is of physiological significance.
This is because the steady-state inactivation of Na⫹ (and
Ca2⫹) channels and thus their availability for the production of action or pacemaker potentials depends steeply on
membrane voltage near the resting (or maximal diastolic)
potential.
1809
1810
HELFRIED GÜNTHER GLITSCH
plateau depolarization of up to 9 –16 mV following a
blockade of Ip. The significance of Ip and its modulation
by adrenergic transmitters for the shape of the cardiac
action potential under physiological conditions have
recently been reemphasized (183).
VIII. MODULATION OF CARDIAC PUMP
CURRENT BY AUTONOMIC
TRANSMITTERS AND RELATED
COMPOUNDS
Physiol Rev • VOL
A. Effect of Adrenergic Agonists on Ip
1. ␤-Adrenergic stimulation of cardiac Ip
A stimulatory action of these substances on the
Na⫹-K⫹ pump in multicellular cardiac preparations has
been reported by various authors during the last 45 years
(for references, see Ref. 67). However, the mechanism of
action remained unclear. More recent studies on isolated
cardiac cells failed so far to elucidate the mechanism.
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
It seems likely that autonomic transmitters adjust the
Na⫹-K⫹ pump activity to the functional demands of the
cells and the organism. In fact, various effects of the
transmitters (and related compounds) on Na⫹-K⫹ pumping of single cardiac cells via adrenergic and muscarinergic acetylcholine receptors have been described. The
mechanisms of the intracellular signal transduction from
receptor stimulation to the cardiac Na⫹-K⫹ pump are
outlined in scheme 2, which is partially hypothetical.
␤-Adrenergic agonists exert their effects via ␤-receptors
in the sarcolemma. The receptors are coupled to Gs proteins which activate adenyl cyclase (AC) to facilitate the
synthesis of cAMP
cAMP stimulates protein kinase A (PKA) which enhances active Na⫹/K⫹ exchange by phosphorylation of
Na⫹-K⫹ pump molecules. Stimulation of adrenergic ␣1receptors causes, via a G protein, the activation of
phospholipase C (PLC) which cleaves phosphatidylinositol 4,5-diphosphate (PIP2) to diacylglycerol (DAG)
and inositol trisphosphate (IP3). DAG and Ca2⫹ (released from the sarcoplasmic reticulum by IP3) activate
protein kinase C, which in turn stimulates cardiac
Na⫹-K⫹ pumping by pump phosphorylation. Binding of
acetylcholine to a muscarinergic acetylcholine receptor
(M2 receptor) causes inhibition of adenylyl cyclase,
decrease of [cAMP], and thereby inhibition of PKA and
Na⫹-K⫹ pump activity.
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
Physiol Rev • VOL
served at nanomolar and subnanomalar [Ca2⫹]i and was
mediated, at least in part, by facilitation of a partial reaction in the Na⫹ limb of the pump cycle, different from Na⫹
i
binding (108).
In contrast to the just described observations on
guinea pig ventricle myocytes, a study on rat ventricular
cells did not detect any ␤-adrenergic regulation of the
Na⫹-K⫹-ATPase (101). Surprisingly, other experiments
using the same cell species, similar solutions, and electrophysiological techniques but a different protocol
showed an Ip stimulation by norepinephrine and isoprenaline (35). Figure 12 illustrates some of these results.
Under the conditions chosen, the membrane current of a
rat myocyte clamped at ⫺40 mV is nearly identical to the
ouabain-sensitive Ip (Fig. 12, A and B). Norepinephrine
(10 ␮M) and isoprenaline (10 ␮M) increase Ip. This is
shown in Figure 12, A and B, respectively. Mean values of
FIG. 12. Stimulation of Ip by norepinephrine (NA) and isoprenaline
(Iso) in a rat cardiac myocyte. A: example recording of the holding
current (Ih) at ⫺40 mV as affected by 10 ␮M NA (marked by arrows) and
1 mM ouabain (Ou). The peak effect of NA was measured as shown by
arrow labeled INa. Ouabain was applied to the cell for a short time before
and after NA application to measure Ip as shown by arrow Ip. B: example
record of Ih from a different myocyte illustrating the whole cell current
response to 10 ␮M Iso. Protocol, conditions, and labels are the same as
in A. C: mean amplitudes ⫾ SE of Ip before and during the application
of NA (1–10 ␮M, columns on left) or Iso (1–10 ␮M, columns on right); n
indicates the number of cells studied. [Adapted from Dobretsov et al.
(35).]
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
Even worse, according to the information available at
present, it seems questionable whether a universal mechanism of catecholamine action exists for cardiac cells of
different animal species. Apart from species differences,
technical problems [impaired receptor function following
cell isolation, use of patch pipettes with relatively high
resistance (⬎2 M⍀)] may also contribute to the inconsistent observations.
Isoprenaline (10⫺7 to 10⫺6 M) induced a decrease of
i
the intracellular Na⫹ activity (aNa
) by ⬃25% in isolated
i
rabbit ventricular myocytes and an enhanced aNa
decline
⫹
following an increase of the extracellular K concentration (30). Unfortunately, the sensitivity of the effects to
cardiac glycosides, specific inhibitors of the Na⫹-K⫹
pump, was not tested. The authors concluded from their
results that isoprenaline directly stimulates cardiac active
Na⫹/K⫹ transport and not indirectly via a drug-induced
accumulation of external K⫹. The latter mechanism had
previously been considered to explain the stimulation of
the cardiac Na⫹-K⫹ pump by catecholamines in multicellular preparations. In contrast to the experiments just
mentioned, more indirect measurements failed to demonstrate any direct ␤-adrenergic stimulation of the Na⫹-K⫹
pump in guinea pig ventricular myocytes (120). However,
other observations on the same cell species revealed a
direct ␤-adrenergic effect (56). The ␤-adrenergic modulation of Ip was not due to drug-induced changes of the
intracellular Na⫹ or extracellular K⫹ concentration but
directly to the variation of the maximum pump turnover
rate. Interestingly, Ip was reduced by isoprenaline at low
(⬍150 nM) intracellular [Ca2⫹], but increased at higher
[Ca2⫹]i. In a further patch-clamp study (54) it was shown
that the ␤-adrenergic increase of Ip at high [Ca2⫹]i was
mediated by a phosphorylation step via the cAMP-PKA
cascade. A study of the ␤-adrenergic effects of isoprenaline on the Ip-V relationship in guinea pig ventricular cells
(59) showed that the increase of Ip at high [Ca2⫹]i (1.4
␮M) was voltage dependent. At positive voltages isoprenaline had little effect on the Ip amplitude, whereas the
drug increased Ip at negative membrane potentials. However, the inhibition of Ip caused by isoprenaline at low
[Ca2⫹]i (15 nM) was voltage independent. The authors
suggested two effects of intracellular Ca2⫹ on the ␤-adrenergic modulation of cardiac Ip. First, it counteracts the
inhibiton of Ip induced by protein phosphorylation via
PKA activation (55), and second, it shifts the Ip-V curve in
the negative direction during the ␤-adrenergic, PKA-mediated phosphorylation of the cardiac Na⫹-K⫹ pump. The
physiological significance of the increase of Ip during
␤-adrenergic stimulation at high [Ca2⫹]i may consist of a
partial compensation for ␤-adrenergic effects on membrane conductances, which tend to prolong the cardiac
action potential. Most recently, an increase of Ip by the
adenylyl cyclase activator forskolin in guinea pig ventricular myocytes has been reported. The increase was ob-
1811
1812
HELFRIED GÜNTHER GLITSCH
2. ␣-Adrenergic stimulation of Ip in cardiac cells
In canine Purkinje myocytes, the ␣-adrenergic drug
phenylephrine increases the Na⫹-K⫹ pump activity and
thereby Ip and decreases the background K⫹ conductance. Both effects are mediated by ␣1-adrenoceptors and
are blocked by a pretreatment of the Purkinje fibers with
pertussis toxin. The ␤-antagonist propranolol is ineffective. The effect of phenylephrine is abolished by 10⫺4 M
DHO, a specific blocker of the Na⫹-K⫹ pump, in Ba2⫹containing external solution. The known negative chronotropic effect of phenylephrine might be due to activation of electrogenic Na⫹-K⫹ pumping. The pump seems to
be coupled to a pertussis toxin-sensitive G protein (160).
Norepinephrine and the ␣-adrenergic agonists phenylephrine, methoxamine, and metaraminol increase Ip of guinea
pig ventricular myocytes in propranolol-containing media.
The increase is blocked by the ␣1-antagonist prazosin and
is unaffected by the ␣2-antagonist yohimbine. The stimulation of Ip is not due to accumulation of intracellular Na⫹
or extracellular K⫹ and is independent of voltage. The
norepinephrine concentration required for half-maximal
Ip stimulation (K0.5 value) depends on [Ca2⫹]i and decreases from 219 nM at 15 nM Ca2⫹
to 3 nM at 1.4 ␮M
i
Physiol Rev • VOL
Ca2⫹
(183). Ip measurements during application of (noni
specific) activators or inhibitors of PKC (60, 61) are in line
with the hypothesis that cardiac Na⫹/K⫹ pump is coupled
to ␣1-adrenoceptors via PKC. The sensitivity of the coupling depends on [Ca2⫹]i. The maximal increase in Ip is
independent of membrane potential and [Ca2⫹]i (183). An
electropharmacological study demonstrated that the ␣1adrenoceptor stimulation of Ip in rat ventricular myocytes
is mediated by the ␣1b-subtype of the ␣1-adrenergic receptors (185). It is helpful to note that adrenergic modulation
of the Na⫹-K⫹ pump (Na⫹-K⫹-ATPase) might result in
different effects, depending on cell species and/or experimental conditions. For further discussion and additional
references, the reader is referred to a relevant review
(180). In a most recent paper, evidence is presented that
two isoforms of the ␣-subunit of the Na⫹-K⫹-ATPase (␣1,
82%; ␣2, 18%) are present in guinea pig ventricles. The
␤-adrenergic, PKA-mediated effects on the cardiac
Na⫹-K⫹ pump are targeted to the ␣1-isoform, whereas the
␣-adrenergic, PKC-dependent effects are targeted to the
␣2-isoform (62).
B. Modulation of Cardiac Ip by Acetylcholine
To our knowledge, so far only one report has been
published on the action of acetylcholine (ACh) on the
cardiac Ip (58). According to this study ACh does not
modulate the basal Ip at any voltage in guinea pig ventricular myocytes containing a high or low [Ca2⫹]i. However,
it reverses the effect of isoprenaline on Ip, regardless of
[Ca2⫹]i, with a K0.5 value of ⬃16 nM. Figure 13 presents an
original record of membrane current from a guinea pig
ventricular cell clamped at ⫺60 mV (58). Ip is estimated as
current blocked by 1 mM DHO. At low internal [Ca2⫹] (15
nM), isoprenaline (1 ␮M) decreases Ip. This effect is
abolished by acetylcholine (ACh; 1 ␮M). The ACh effect is
mediated by muscarinic receptors since it is blocked by
atropine. The ACh-induced stimulation of the receptor
leads most probably via a Gi protein to an inhibition of
adenylyl cyclase and in consequence to a decrease in
[cAMP] (see scheme 2). Thus a high vagal tone per se
does not modulate the cardiac Na⫹-K⫹ pump, but activation of muscarinic receptors reverses the modulation of
the pump induced by a high sympathetic tone.
IX. CARDIAC GLYCOSIDES AND CARDIAC
PUMP CURRENT
A. Binding of Cardiac Glycosides to Various
Isoforms of the Cardiac Naⴙ-Kⴙ Pump Is
Species Dependent
As mentioned in section I, different isoforms of the
␣-subunit of the Na⫹-K⫹-ATPase are expressed in the
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
the Ip activation by the two catecholamines are presented
in Figure 12C. Both substances augment Ip by ⬃40%. The
stimulation of Ip occurred at low [Ca2⫹]i (20 nM), was
⫹
voltage independent, and was not due to Na⫹
i or Ko
⫹
⫹
accumulation, changes in the pump affinity to Nai or Ko ,
or in the pump site density. The effect of isoprenaline
increased with increasing the intracellular K⫹ concentration from 0 to 100 mM K⫹. From simulation of the experimental data with a simplified Post-Albers cycle, the authors suggested that the ␤-adrenergic Ip stimulation in rat
myocytes might be caused by a drug-induced facilitation
of K⫹ deocclusion and K⫹ release into the cell interior.
Furthermore, the adrenergic modulation of Ip was examined in adult rat cardiac myocytes in short-term culture
(up to 4 days). The Ip stimulation by norepinephrine and
isoprenaline (at 20 nM free Ca 2⫹
i ) displayed K0.5 values of
36 and 1.5 nM, respectively, and was ␤-receptor mediated
(166). In contrast to experiments on guinea pig ventricular cells (56, 59), a reduction rather than an increase of the
adrenergic Ip stimulation with increasing [Ca2⫹]i was observed, and no effect of adrenergic stimulation on the
shape or position of the Ip-V relationship could be detected. Because there was no change in the voltage dependence of Ip during ␤-adrenergic stimulation of the
Na⫹-K⫹ pump, it was concluded that the Na⫹ release
from the pump to the extracellular medium is not modulated by ␤-agonists and that the mechanism of ␤-adrenergic Ip modulation in rat ventricular cells is qualitatively
different from that proposed for guinea pig cardiac myocytes (166).
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
1813
cardiac cells of various animal species. In general, two of
three ␣-subunit isoforms are detected in the heart. There
is agreement in the literature that rat myocytes (e.g., Refs.
12, 164, 175) and myocytes from many other species,
including human cells (e.g., Refs. 175, 182), express more
than one Na⫹-K⫹ pump isoform (see Ref. 175). Because
the ␣-isoforms represent the receptors for cardiac glycosides and normally differ in their sensitivity for the drugs
(see Ref. 19, but also Ref. 27), more than one receptor
class for cardiac steroids should be found in heart cells.
However, only a single ␣-isoform is expressed in sheep
cardiac cells (175) and, perhaps, in guinea pig ventricular
myocytes (175; but see Refs. 11, 62). Thus only one receptor class may be present in the heart cells of these two
species. The sensitivity of the ␣1-isoforms for cardiac
steroids varies widely among the animal species. For
instance, the rat ␣1-subunit is known to be extremely
resistant to the drugs.
The chapter below describes the interaction between
cardiac glycosides and the Na⫹-K⫹ pump as studied by
electrophysiological techniques in isolated cardiac cells.
It is not in the focus of the present article to review the
immense literature on cellular and subcellular actions of
cardiac steroids in other preparations and/or investigated
by other physiological or by biochemical methods.
Clearly, this literature is at least as important for our
understanding of the glycoside action as the few observations reported here. For a more extensive overview the
reader is referred to relevant reviews (2, 40, 78, 116).
Ever since the pump current was first described in a
cardiac preparation, cardiac glycosides, which are specific inhibitors of the Na⫹-K⫹ pump (153), have been used
to identify cardiac Ip (99). For example, in the first paper
on the voltage dependence of Ip in single cardiac cells, the
pump current of guinea pig ventricular myocytes was
Physiol Rev • VOL
measured as the ouabain-blockable current (48). As just
described, it is still controversial whether in these cells
more than one ␣-subunit isoform of the Na⫹-K⫹-ATPase
(more than one class of cardiac steroid-receptors) is expressed and functioning. On the one hand, two components of the total Ip with different sensitivities to cardiac
glycosides have been described (57, 127). According to
the authors, the highly sensitive component contributes
up to 30 – 40% of the total Ip under quasi-physiological
conditions. This is in line with a biochemical-immunological study (11) which described a component with high
affinity toward cardiac glycosides (dissociation constant
⬃10 nM ouabain, DHO, or digitoxigenin) that contributes
as much as 55% to the total Na⫹-K⫹-ATPase activity of
guinea pig ventricular cells. Furthermore, a recent paper
(62) showed by a molecular biological approach that 18%
of the pump mRNA consists of mRNA for the ␣2-subunit
of the pump which exhibits a high affinity for cardiac
glycosides. On the other hand, Ip inhibition by strophanthidin occurs as reversible 1:1 binding to a single class of
pumps in guinea pig ventricular cells (114). Similarly, the
Ip inhibition by DHO in rat and guinea pig cardiac myocytes in a range between ⬃20 and 95% inhibition can be
analyzed by assuming simple saturation kinetics and a
single population of pump molecules (86). Of course, this
does not exclude the presence of a minor class of pumps
(additional ␣-isoform), displaying a different affinity to
cardiac glycosides. However, application of immunologic
methods did not detect any ␣2-isoenzyme of the Na⫹-K⫹
pump in guinea pig ventricular myocytes (122, 175). Finally, the Ip inhibition-concentration curve (range 10⫺8 to
5 ⫻ 10⫺5 M) for guinea pig ventricular myocytes could not
be fitted by assuming two isozymes showing different
sensitivities toward DHO (84). So far, the reason for the
conflicting results with guinea pig cardiac cells is com-
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
2⫹
FIG. 13. Acetylcholine (ACh) reverses the effects of Iso on Ip at low [Ca ]i and ⫺60 mV. The guinea pig ventricular
myocyte was held at ⫺60 mV, and the pipette solution contained 15 nM Ca2⫹. The continuous line indicates the
application of Iso (1 ␮M), and the dashed line indicates the application of Iso plus 1 ␮M ACh (Iso ⫹ ACh). The vertical
line labeled Ip indicates the size of Ip and how it is measured. The effect of ACh on Ip was reversible. [Adapted from Gao
et al. (58).].
1814
HELFRIED GÜNTHER GLITSCH
pletely unclear. A qualitatively or quantitatively different
expression of ␣-subunit isoforms of the Na⫹-K⫹ pump in
various strains of guinea pig may be involved.
By way of contrast, two components were clearly
identified in the Ip inhibition-concentration curve in rat
myocytes. The high-affinity component amounted to
⬃32% of the total Ip in cells from young animals but only
to ⬃15% in myocytes from adult rats (84).
B. Kinetics of Cardiac Steroid Binding
to the Cardiac Naⴙ-Kⴙ Pump
To study the inhibition of the cardiac Ip by cardiac
glycosides, DHO is often used since its action on the
Na⫹-K⫹ pump is readily reversible. The inhibition of Ip by
DHO was investigated in isolated canine Purkinje myocytes at 8 mM Ko⫹ by means of a discontinuous voltageclamp technique (25). The data were analyzed by assuming a reversible one-to-one interaction between the
cardiac steroid and a single receptor class R on the
Na⫹-K⫹ pump (28)
k1
DHO ⫹ R N DHO 䡠 R
(7)
k2
The equilibrium dissociation constant KD of this reaction
is given by
KD ⫽
k2
k1
(8)
The association rate constant k1 is related to ␶on, the time
constant of the process of DHO binding, by
1
⫽ k 1 䡠 关DHO兴 ⫹ k2
␶ on
(9)
whereas the dissociation rate constant k2 is defined as
k2 ⫽
1
␶off
(10)
␶off is the time constant of DHO unbinding. By applying
the above equations to their data, an apparent KD value
(i.e., the [DHO] which causes half-maximal inhibition of
Ip) of 3.7 ⫻ 10⫺6 M DHO (K0.5 value), an apparent
association rate constant k⬘1 of 2.5 ⫻ 103 (M ⫻ s)⫺1 and
a dissociation rate constant k⬘2 of 0.009 s⫺1 were calculated (25). The agreement between measured steadystate Ip inhibition as a function of [DHO] and calculated
Physiol Rev • VOL
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
1. Reversible one-to-one binding to a single class
of receptors on the pump
(k⬘2/k⬘1) K0.5 value suggested that the kinetic model outlined above is reasonable (25). A similar but more
detailed study was carried out on single rabbit cardiac
Purkinje cells (17). Half-maximal inhibition of Ip by
1 ⫻ 10⫺5 M DHO was found in a medium containing 2
mM Ko⫹. This K0.5 value increased to 3.5 ⫻ 10⫺5 M DHO
at 10.8 mM Ko⫹. The antagonistic effect of Ko⫹ on the Ip
inhibition by DHO was studied in some details as
shown in Figure 14 (17). Figure 14A, top panel, displays
the membrane current of a Purkinje cell at 2 mM Ko⫹
before, during, and after application af 10⫺4 M DHO.
The steroid inhibits Ip and therefore shifts the membrane current in the inward directions. A new stable
current level is reached in ⬃10 s. Short intermittent test
pulses of K⫹-free medium block Ip and indicate the
current level in the absence of Na⫹-K⫹ pumping. It is
clear from the current trace that 10⫺4 M DHO inhibits
⬃90% of Ip at 2 mM Ko⫹. Figure 14A, bottom panel,
demonstrates that at 10.8 mM Ko⫹ the Ip inhibition by
10⫺4 M DHO proceeds more slowly to a final level of
only ⬃65% of Ip in drug-free solution. Figure 14B emphasizes that the time constant of Ip decline in DHOcontaining solution is smaller at 2 mM Ko⫹ than at 10.8
mM Ko⫹. By way of contrast, washout of the drug is very
similar at both [K⫹]o (Fig. 14C). Application of the
reaction scheme presented above yielded apparent association rate constants (k⬘1) of 1 ⫻ 104 (M ⫻ s)⫺1 at 2
mM Ko⫹ and 2.8 ⫻ 103 (M ⫻ s)⫺1 at 10.8 mM Ko⫹; the
apparent dissociation rate constant k⬘2 turned out to be
almost Ko⫹ independent and amounted to ⬃0.06 s⫺1 at
both [K⫹]o. Increasing [Na⫹] of the patch-pipette solution (for intracellular perfusion) from 5 to 50 mM increased the k⬘1 value but left the k⬘2 value unchanged. As
a consequence, the K0.5 value decreased by a factor of
3–5. The main changes occurred after an increase of
[Na⫹]pip from 5 to 15 mM. The observed dependence of
K0.5 on [Na⫹]pip is reminiscent of an earlier paper (169)
which demonstrated a decrease of the K0.5 value of
ouabain for the pump inhibition in chick cardiac myocytes upon an increase in the intracellular Na⫹ concentration. In accordance with a model published previously (88), it was concluded that an increase of Na⫹
i
augments the probability for a distinct pump molecule
to take on a conformation that binds ouabain with high
affinity. The K0.5 value for Ip inhibition in rat and guinea
pig ventricular myocytes amounted to 2.4 ⫻ 10⫺3 and
1.4 ⫻ 10⫺5 M DHO, respectively (Ko⫹, 5.4 mM; holding
potential, ⫺20 mV; Napip, 50 or 100 mM). The higher
K0.5 value for binding of DHO to rat cells was due to a
smaller apparent association rate constant and a larger
dissociation rate constant. There was little evidence for
the presence of an isoform with higher DHO affinity in
guinea pig myocytes, and only a few percent of the
pump molecules in rat cells displayed a higher affinity
to DHO (86). Half-maximal Ip activation by Ko⫹ re-
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
1815
mained nearly unchanged, but Ip(max) decreased following addition of [DHO] near the respective K0.5 value at
5.4 mM Ko⫹. The K0.5 value of DHO for Ip inhibition and
the activation of Ip showed a similar dependence on Ko⫹
(86). The authors also reported that the K0.5 value was
larger in Na⫹-free than in Na⫹-containing media under
conditions where Ip activation by Ko⫹ was nearly the
same, suggesting that DHO binding occurred preferentially to a Na⫹-bound conformational state of the cardiac Na⫹-K⫹ pump. In line with earlier biochemical
evidence (reviewed in Ref. 78), the data were interpreted to mean that the antagonism between DHO and
Ko⫹ as to the activation of Ip is noncompetitive. Ko⫹ and
the cardiac glycoside bind to different conformational
states of the pump transiently exposed to the external
face of the cell membrane during the pump cycle. DHObound states are unable to generate Ip (86).
2. Binding of cardiac glycosides to two classes
of pump molecules
In line with reports on two isoforms of the ␣-subunit
present in guinea pig ventricular myocytes (11, 62), inhibition of Ip in these cells has been occasionally described
to display a biphasic concentration dependence (127).
The data were fitted best by the equation
I DHO ⫽
IDHO max h ⫻ 关DHO兴 IDHO max l ⫻ 关DHO兴
⫹
KDh ⫹ 关DHO兴
KDl ⫹ 关DHO兴
Physiol Rev • VOL
(11)
where IDHO denotes the current blocked by a distinct
[DHO], IDHO max h and IDHO max l represent the DHO-blockable current of pump isoforms showing a high or low
affinity toward DHO, respectively, and KDh and KDl are
the respective equilibrium dissociation constants of the
interaction between DHO and the two classes of pumps
displaying either high or low affinity to DHO. KDh was
calculated to be 5 ⫻ 10⫺8 M DHO, and KDl was derived to
be 6.5 ⫻ 10⫺5 M DHO. Under the conditions chosen
⫹
(temperature, 34°C; Napip
, 30 mM; Ko⫹, 5.4 mM; holding
potential, ⫺40 mV), the pumps with a high DHO affinity
contributed as much as 40% to the total Ip. A detailed
analysis of electrogenic Na⫹-K⫹ pumping in guinea pig
ventricular myocytes also indicated the presence of more
than one ␣-subunit isoform (57). Obviously, two classes of
pump molecules displaying different affinities to DHO
contribute to the total pump current at ⫺60 mV holding
potential. The high-affinity pumps are half-maximally inhibited by 7.5 ⫻ 10⫺7 M DHO and the low-affinity pump
molecules by 7.2 ⫻ 10⫺5 M DHO. Although the latter K0.5
value agrees nicely with the corresponding KDl value
above (127), the former is larger by a factor of 15. The
high-affinity pumps produced ⬃30% of the total pump
current. They were half-maximally activated by 0.4 mM
Ko⫹, whereas the low-affinity pumps required 3.7 mM Ko⫹
for half-maximal activation.
As reported above, a well-established kinetic competition exists between Ko⫹ and binding of cardiac glycosides to the cardiac Na⫹-K⫹ pump (8; reviews in
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
FIG. 14. Binding and unbinding of DHO at
two external K⫹ concentrations. A: membrane current of a rabbit cardiac Purkinje cell
at 2 mM Ko⫹ (top trace) and at 10.8 mM Ko⫹
(bottom trace) during binding and unbinding
of 10⫺4 M DHO. Holding potential was ⫺20
mV. Inward deflections of the current are
caused by pulses of K⫹-free solution applied
to block Ip. Note the different time calibration. B: time course of DHO binding during
the experiment shown in A. Plot of the logarithm of the difference between actual Ip (Ipt)
and final Ip amplitude (Ip⬁) in the drug-containing solution versus time. ␶on is larger at
10.8 mM Ko⫹ (F) than at 2 mM Ko⫹ (E). C: time
course of DHO unbinding during the experiment depicted in A. Plot of the logarithm of
the difference between final Ip amplitude
(Ip⬁) in drug-free medium and actual Ip (Ipt)
during unbinding versus time. ␶off is similar at
both [K⫹]o (E, 2 mM Ko⫹; F, 10.8 mM Ko⫹).
Lines fitted to the data in B and C are regression lines; r represents the correlation coefficient. [Adapted from Bielen et al. (17).]
1816
HELFRIED GÜNTHER GLITSCH
Refs. 2, 40, 156). Surprisingly, the Ip inhibition observed
at various [DHO] in guinea pig (and canine) ventricular
myocytes in media containing 1 or 4.6 mM Ko⫹ could not
be fitted by assuming this type of competition between
Ko⫹ and DHO. In contrast to observations on cultured
chick cardiac myocytes (169) and rabbit cardiac Purkinje cells (17), increasing [Na⫹]pip from 10 to 50 mM
had little effect on the K0.5 value of DHO for halfmaximal Ip inhibition (57).
3. At therapeutic concentrations cardiac glycosides
preferentially bind to high-affinity pump molecules
in human cardiac tissue
4. Evaluation of new cardiac steroids as Na⫹-K⫹
pump modulators by measurements of the
cardiac Ip
A convenient method to study the modulating effect
of new cardiac steroids on the cardiac Na⫹-K⫹ pump is to
measure the change of Ip caused by the drugs in cardiac
cells. This method was used in an attempt to identify
C-22-substituted derivatives of digitoxigenin, which might
be suitable for affinity labeling of the binding site(s) of the
lactone ring on the Na⫹-K⫹-ATPase. Whole cell recording
from sheep cardiac Purkinje cells revealed that esters
derived from 22-hydroxy-digitoxigenin are useful for this
purpose (41).
Physiol Rev • VOL
Alterations of the cardiac Ip-V curve by cardiac steroids were shown when the interaction between DHO or
ouabain, and the Na⫹-K⫹ pump of rat and guinea pig
ventricular myocytes was studied at various [K⫹]o and
membrane potentials by means of whole cell recording
(85). The glycoside-induced changes of the cardiac Ip-V
relationship are probably mediated by potential-evoked
alterations of the local [K⫹]o, near the K⫹ binding sites of
the Na⫹-K⫹ pump at the bottom of an extracellular access
channel. The alterations were estimated by means of the
Boltzmann-Fermi equation and compared with the effect
of various [K⫹]o on the binding of DHO to the cardiac
Na⫹-K⫹ pump at 0 mV. Upon a change of [K⫹] in the
access channel, a new equilibrium of glycoside binding to
the pump is reached. The time required to establish the
new binding equilibrium depends on the glycoside, the
concentration of the cardioactive steroid, and on the sensitivity of the cells toward the drug. The findings lend
further support to the access-channel hypothesis and thus
suggest that most probably Ko⫹ binding to the cardiac
Na⫹-K⫹ pump occurs within the electrical field across the
sarcolemma (85).
X. MODULATION OF THE CARDIAC
SODIUM-POTASSIUM PUMP BY HORMONES
A. Aldosterone
A short-term exposure of cardiac cells to the mineralocorticoid aldosterone affects the Na⫹-K⫹ pump. This
was shown in isolated rabbit ventricular myocytes by
i
whole cell recording of Ip and measurements of aNa
and
pH in right ventricular papillary muscles using ion-sensitive microelectrodes (125). After exposure to 10 nM aldoi
sterone for 20 min, aNa
was significantly higher (9.9 ⫾ 0.7
mM) than in controls (7.7 ⫾ 0.7 mM), whereas intracellular pH remained unchanged (even after 60 min in 100 nM
aldosterone). Furthermore, after blockade of the Na⫹-K⫹
i
pump by DHO, the rise of aNa
was enhanced in aldosterone-treated papillary muscles. The effect of aldosterone
on Ip varied with the transmembranal Na⫹ and Cl⫺ gradients. There was an aldosterone-induced increase of Ip if
these gradients were steeply inward. Flattening the Na⫹
and Cl⫺ gradients resulted in an aldosterone-evoked decrease of Ip. Bumetanide (10 ␮M), an inhibitor of the
Na⫹-K⫹-2Cl⫺ transporter, abolished the increase of Ip and
Na⫹ influx evoked by aldosterone. Similarly, the mineralocorticoid receptor blocker potassium canrenoate completely blocked the aldosterone-induced Ip increase. The
data strongly suggest that stimulation of Ip and Na⫹ influx
in cardiac myocytes exposed to aldosterone in vitro is
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
Studies on electrogenic Na⫹-K⫹ pumping of human
atrial tissue in media containing various acetylstrophanthidin concentrations (149) or of atrial tissue from digoxin-treated patients (148) suggest that at least two classes
of pump molecules displaying different glycoside sensitivity are present in the human heart. If administered
therapeutically, cardiac glycosides seem to bind mainly to
high-affinity pumps that may represent up to ⬃33% of the
pumps (149). These findings are in qualitative agreement
with observations on Ip inhibition by cardiac glycosides in
isolated ventricular myocytes from the human failing
heart (182). According to the latter authors, the Ip inhibition-concentration curve for DHO is biphasic. The K0.5
value for inhibition of the high-affinity pumps amounted
to 6 ⫻ 10⫺9 M DHO. This class of pump molecules generated ⬃14% of the total Ip under the conditions chosen.
The low-affinity pumps displayed a K0.5 value of 2 ⫻ 10⫺6 M
DHO, 10 times lower than for pumps in guinea pig ventricular cells. The difference is obviously due to a larger
dissociation rate constant for DHO in guinea pig myocytes. The K0.5 values of digoxin for high-affinity Na⫹-K⫹
pumps were estimated to be ⬃10⫺9 M and for low-affinity
pumps 3 ⫻ 10⫺7 M, respectively (at [Na⫹]pip ⫽ 100 mM).
Thus this report suggests that therapeutic doses probably
block preferentially high-affinity pump molecules and inhibit 7–10% of the total Ip or Na⫹-K⫹ pump activity (182).
C. Cardiac Glycosides Alter the Cardiac
Ip-V Relationship
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
B. Angiotensin-Converting Enzyme Inhibition
By means of the same methods and cardiac preparai
tions, a decrease of aNa
from 8.2 ⫾ 0.4 mM (controls) to
3.6 ⫾ 0.4 mM, but no change in intracellular pH was found
in ventricular papillary muscles of rabbits treated with the
angiotensin-converting enzyme (ACE) inhibitor captopril
for 8 days (97). This treatment also caused an increased Ip
of isolated ventricular myocytes at near-physiological
[Na⫹]pip. No captopril effect was seen at [Na⫹]pip ⬍2.5
mM and ⬎10 mM. The effect on Ip was imitated by enalapril, an ACE inhibitor containing no sulfhydryl group,
but not by the vasodilator hydralazine or by an exposure
to captopril in vitro (20 –55 min). According to the authors, captopril induces an intrinsic increase in cardiac
Na⫹/K⫹ pumping at near-physiological [Na⫹]i. One year
later, the same group showed that the modulation of the
cardiac Na⫹-K⫹ pump activity by ACE inhibitors is due to
an interference with the physiological mechanism of angiotensin metabolism (96). Compared with controls, myocytes from rabbits treated with the angiotensin II type 1
receptor antagonist losartan exhibited an increased Ip (at
approximately physiological [Na⫹]i) similar to Ip of myocytes from captopril-treated animals. In addition, the latter cells displayed an Ip near the control level following
exposure to angiotensin II (10 nM) for 45 min. There was
no change in the mRNAs for the subunits (␣1, ␣2, ␤1) of
the Na⫹-K⫹ pump in myocytes of captopril-treated rabbits
compared with control. This and other findings suggest an
action of captropril on already existing pump molecules
rather than an interference with the pump synthesis. The
reduction of Ip by angiotensin in ventricular myocytes
from rabbits was blocked by pertussis toxin and the PKC
inhibitors staurosporine and bis-indolylmaleimide I. It
was mimicked by the PKC activator phorbol 12-myristate
13-acetate (PMA). Thus modulation of Ip (and cardiac
Na⫹-K⫹ pump activity) by angiotensin II is mediated via
the angiotensin II type 1 receptor, a G protein, and PKC.
Physiol Rev • VOL
Figure 15 is from a recent paper on the subject (22) and
shows that the increase of Ip caused by captopril or
losartan can be blocked by PMA-induced PKC stimulation, by inclusion of the PKC fragment PKCF (see legend
to Fig. 15) in the patch pipette solution, or by exposure of
myocytes from captropril-treated rabbits to angiotensin
II. The modulation of cardiac Ip by angiotensin II is probably based on a decrease of the pump’s selectivity for Na⫹
i
over K⫹
i at binding sites near the cytosolic surface of the
sarcolemma. The change in selectivity includes a phosphorylation by PKC. The Ip-V curve is not affected by the
hormone.
C. Thyroid Status and Cardiac Naⴙ-Kⴙ Pump
The dependence of the cardiac Na⫹-K⫹ pump activity
on the thyroid status was studied in isolated ventricular
myocytes and papillary muscles from the rabbit (37). Ip of
single ventricular cells was measured by means of whole
i
cell recording and aNa
in papillary muscles by means of
⫹
Na -sensitive microelectrodes. At ⫺40 mV and 10 mM
⫹
Napip
Ip amounted to 0.24 ⫾ 0.02 pA/pF in myocytes from
hypothyroid rabbits and to 0.48 ⫾ 0.05 pA/pF in cells from
animals treated with 3,3⬘,5-triiodothyronine (T3). In cells
i
from hypothyroid rabbits, aNa
was 5.2 ⫾ 0.42 mM,
i
whereas aNa amounted to 7.6 ⫾ 0.69 mM in papillary
muscles from T3-treated rabbits. The increase in Ip was
not due to an increased passive Na⫹ influx not compensated for by the intracellular dialysis. Using the Oxsoft
Heart computer model (Oxsoft Heart Program Manual;
Oxford, UK; Oxsoft, 1993), the effect of the T3-induced
FIG. 15. Effect of protein kinase C (PKC) activation and angiotensin
II (ANG II) on Ip. Myocytes were isolated from rabbits treated with
captopril (capt) or losartan (los), and PKC was activated by incubation
with phorbol 12-myristate 13-acetate (PMA) in pipette solutions. A
[K⫹]pip of 70 mM was used in all experiments. Number of myocytes in
each group is in parentheses. * Significantly different vs. captopril treatment alone. ** Significantly different vs. losartan treatment alone.
[Adapted from Buhagiar et al. (22).]
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
most probably due to activation of the Na⫹-K⫹-2Cl⫺ transporter by the hormone (125).
Interestingly, when the hormone (50 ␮g/kg body wt)
was administered to rabbits for 7 days, Ip of isolated
ventricular cells decreased at 10 mM but not at 80 mM
⫹
i
Napip
. The aNa
of papillary muscles from treated animals
increased, whereas the Na⫹ influx remained unchanged,
in contrast to the observations during short-term exposure (see above). Aldosterone had no effect on the density
of pump molecules in the sarcolemma. The aldosterone
antagonist spironolactone blocked both the Ip decrease at
⫹
i
10 mM Napip
and the increase of aNa
in papillary muscles.
Thus the mineralocorticoid decreased the Na⫹ affinity of
the cardiac Na⫹-K⫹ pump but did not affect Na⫹ influx
following infusion for several days (126).
1817
1818
HELFRIED GÜNTHER GLITSCH
i
changes in Ip and aNa
on the cardiac action potential was
simulated and found to be similar to the action potential
shortening observed in papillary muscles. A subsequent
i
paper by the same group demonstrated that the aNa
increase in papillary muscles of T3-treated rabbits was
caused by an enhanced Na⫹/H⫹ exchange due to an augmented acid production (36).
D. Insulin Changes the Cardiac Ip-V Curve
XI. MISCELLANEOUS
A. Anisosmolar External Solution Affects the
Activity of the Cardiac Naⴙ-Kⴙ Pump
The role of the Na⫹-K⫹-ATPase in the short-term
volume regulation of cardiac and other cells is controversial. The effect of hypo- and hyperosmolar bathing solui
tions on Ip and aNa
was examined in cardiac preparations
by whole cell recording from isolated ventricular myoi
cytes and aNa
recording by means of Na⫹-sensitive microelectrodes in right ventricular trabeculae of the guinea pig
Physiol Rev • VOL
FIG. 16. Dependence of Ip on [Na]pip in isosmolar (E) and hyposmolar (F) solutions. Values are means ⫾ SE of Ip normalized for cell
capacitance plotted against [Na⫹]pip. Data are from a total of 37 rabbit
ventricular myocytes exposed to isosmolar solution and 38 myocytes
exposed to hyposmolar solution. Numbers in parentheses are number of
cells studied at each [Na⫹]pip. Four experiments using 0 mM [Na⫹]pip
were performed at each osmolarity. [Na⫹] in both isosmolar and hyposmolar solutions was reduced to 70 mM. Na⫹ in isosmolar solution was
substituted with 140 mM sucrose. Curves represent least-squares fit of
the Hill equation to data. Note that [Na⫹]pip vs. Ip relationship is shifted
to left in hyposmolar solution relative to that in isosmolar solution.
[Na⫹]pip for half-maximal current (K0.5) values derived from the fit were
12.8 ⫾ 2.0 and 21.4 ⫾ 3.0 mM, respectively. Maximal Ip was similar in
both solutions (1.25 ⫾ 0.10 vs. 1.24 ⫾ 0.17 pA/pF). [Adapted from
Whalley et al. (184).]
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
Insulin (100 mU/ml) flattens the Ip-V curve of rabbit
ventricular myocytes at negative and positive membrane
⫹
potentials (80). The flattening, observed at 10 mM Napip
,
was absent if the pipette solution contained 100 ␮M tyrphostin A-25, a tyrosine kinase inhibitor. It seems likely
that the tyrosine kinase activity of the insulin receptor
modulates the voltage dependence of cardiac Ip. The
topic was further examined in a most recent paper (79).
The flattening effect of insulin on the Ip-V relationship of
the rabbit cardiac myocytes depends on [Na⫹]pip and
voltage. When [Na⫹]pip was 80 mM and thus nearly saturating the internal Na⫹ binding sites of the pump, or when
the membrane potential was positive to ⫹20 mV, the
insulin effect on the Ip-V curve disappeared. However, the
flattening persisted in Nao⫹- or K⫹i -poor media at high
[K⫹]o, suggesting that competition of Na⫹ and K⫹ at
external or internal cation binding sites of the pump was
not involved. Apart from the inhibiton of the insulin effect
by tyrphostin A-25, wortmannin (1 ␮Mpip a phosphatidylinositol 3-kinase inhibitor) abolished the Ip stimulation by
insulin at the holding potential of ⫺40 mV. The hormone
most probably facilitates the voltage-dependent binding
of Na⫹i to the highly specific Na⫹-binding site within the
internal access channel of the Na⫹-K⫹ pump via activation of the insulin receptor tyrosine kinase, phosphatidylinositol 3-kinase and protein phosphatase 1. PKC and
protein phosphatase 2 are unlikely to be involved.
The effects of catecholamines on cardiac electrogenic Na⫹-K⫹ pumping have already been reviewed in
section VIIIA.
heart, respectively (184). Figure 16 displays Ip density as
a function of [Na⫹]pip in ventricular myocytes at different
osmolarities of the external media. As shown in Figure 16,
Ip measurements revealed an increase in the apparent
Na⫹i affinity of the Na⫹-K⫹ pump in myocytes from a K0.5
⫹
value of 21.4 mM (isosmolar medium) to 12.8 mM Napip
during cell swelling in hyposmolar solution. Half-maxi⫹
mum Ip activation occurred at 39 mM Napip
during cell
shrinkage in hyperosmolar medium (data not shown).
Maximum Ip remained unaltered, and Hill coefficients
were similar. In ventricular trabeculae, exposure to the
hyposmolar external medium for 20 min caused a sustained Na⫹i decrease. Na⫹i was augmented in hyperosmolar solution. Thus osmotic myocyte swelling stimulated
the Na⫹-K⫹ pump near physiological Na⫹ levels, whereas
shrinkage inhibited the pump. The mechanism of the
changes in the apparent Na⫹i affinity was not investigated.
However, the changes were not caused by alterations in
[Ca2⫹]i, [K⫹]i, or intracellular pH. The mechanism was
further studied in a more recent paper (14). Whole cell
recording of Ip from isolated rabbit ventricular myocytes
confirmed an increase in Ip at near-physiological [Na⫹]pip
⫹
and no effect on Ip at 80 mM Napip
in hyposmolar, Ca2⫹free solution at ⫺40 mV holding potential. Cell swelling
also induced a flattening of the Ip-V curve at negative
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
B. Dietary Cholesterol Alters Cardiac
Naⴙ-Kⴙ Pumping
A modest diet-induced increase in serum cholesterol
⫹
(from ⬃0.9 to ⬃4 – 6 mM) augmented Ip at 10 mM Napip
and ⫺40 mV from ⬃0.3 to ⬃0.5 pA/pF in isolated rabbit
ventricular myocytes (75). The Ip-[Na⫹]pip relationship
⫹
revealed an increased Ip in a range of Napip
between 2.5
⫹
and ⬍30 mM. At higher or lower [Na ]pip, the dietary
cholesterol was without effect. Thus a modest, but not a
strong, increase (⬎12 mM) in serum cholesterol raises the
apparent affinity of the pump for cytoplasmic Na⫹. The
apparent pump affinity for Ko⫹ was unaffected, as was the
Ip-V curve. In right ventricular papillary muscles from
i
rabbits displaying a serum cholesterol of 4.9 mM, aNa
i
decreased to 5 from 7.7 mM aNa in untreated controls
(serum cholesterol 1.1 mM) (75). These observations
clearly suggest that dietary cholesterol can alter the
Na⫹-K⫹ pump activity in quiescent cardiac preparations.
C. Amiodarone Inhibits the Cardiac Naⴙ-Kⴙ Pump
Following Acute and Chronic Treatment by
Different Mechanisms
The antiarrhythmic agent amiodarone produces
(among other effects) a state of cellular hypothyroidism.
Physiol Rev • VOL
Because the cardiac Na⫹-K⫹ pump is sensitive to changes
of the thyroid status (see above), the effect of the antiarrhythmic drug on the pump was studied by whole cell
recording of Ip from isolated rabbit ventricular cells and
i
by aNa
measurents in rabbit ventricular papillary muscles
(76). Treatment for 4 wk resulted in an Ip decrease regardless of whether [Na⫹]pip was 10 or 80 mM. However,
acute exposure to ⬃1 ␮M amiodarone reduced Ip at 10
⫹
mM but not at 80 mM Napip
. The drug had no effect on the
Ip-V curve, and the pump’s apparent affinity to Ko⫹. The
observations suggest two mechanisms of amiodarone action: chronic treatment reduces the Na⫹-K⫹ pump capacity, whereas acute exposure to the drug decreases the
⫹
apparent affinity of the pump to Napip
. The inhibiton of Ip
may contribute to the prolongation of the cardiac action
potential evoked by amiodarone. The drug might exert its
effects on the Na⫹-K⫹ pump via “a modification of the
physicochemical properties of the lipid membrane” (76).
XII. EFFECTS OF THE CARDIAC
SODIUM-POTASSIUM PUMP ON ION
TRANSPORTERS AND CHANNELS
MEASURED BY ELECTROPHYSIOLOGICAL
TECHNIQUES
A. Modulation of the Cardiac Naⴙ/Ca2ⴙ Exchange
During the past decade, the interaction of Na⫹-K⫹
pumping with currents generated by other cardiac ion
transporters and channels has been studied by patchclamp methods. In internally dialyzed isolated cardiac
cells, alterations of the Na⫹-K⫹ pump activity cause
changes not only of the pump current Ip but also of the
Na⫹/Ca2⫹ exchange current INa/Ca. Figure 17 presents an
example. Figure 17A shows the membrane current of a
rabbit cardiac Purkinje cell at ⫺20 mV holding potential.
After 8 min in K⫹-free solution, electrogenic Na⫹-K⫹
pumping is initiated by application of a medium containing 10.8 mM K⫹ (uppermost line). The K⫹-rich solution
immediately evokes an outward current (largely Ip) that
declines from an initial peak toward a steady-state level.
The Na⫹/Ca2⫹ exchanger current INa/Ca is simultaneously
measured as current blocked by 5 mM Nio2⫹. Both the
direction and the amplitude of the exchanger current vary
during the decline of Ip. The INa/Ca amplitude is linearly
related to the pump current (Fig. 17B). The reason probably is that the subsarcolemmal [Na⫹] varies with the
activity of the Na⫹-K⫹ pump and directly affects the
currents generated by both transporters (16). A similar
conclusion was reached from measurements of the Na⫹/
Ca2⫹ exchange in the same cell species (120). A detailed,
quantitative study on guinea pig ventricular myocytes
pointed out that the activites of the Na⫹/Ca2⫹ exchanger
and Na⫹-K⫹ pump are tightly coupled via the intracellular
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
potentials and a region of negative slope at positive voltages. Thus Ip stimulation by hypotonic solution is voltage
dependent. The effect on Ip at the holding potential was
independent of [Ca2⫹]o and [Na⫹]o. In contrast to the
decrease of Ip upon reexposure to isosmolar solution, Ip
stimulation in hyposmotic media was not mediated by
PKC. It was blocked by the tyrosine kinase inhibitor
tyrphostin A-25, which also completely inhibited the variation of the Ip-V relationship observed in hypotonic solution. Furthermore, LY-294002, a specific inhibitor of phosphatidylinositol 3-kinase, and okadaic acid, an inhibitor of
protein phosphatase 1 (PP1) and 2A (PP2A), also blocked
the Ip stimulation at ⫺40 mV. Tyrosine kinase has been
shown to activate protein phosphatase 1, whereas phosphatidylinositol 3-kinase is activated in tyrosine kinasedependent intracellular signaling. The results suggest that
Ip stimulation by cell swelling involves the activation of
tyrosine kinase, phosphatidylinositol 3-kinase, and protein phosphatase 1 and may be mediated by a dephosphorylation of the Na⫹-K⫹-ATPase. In contrast to the
findings reported above (184), an Ip stimulation at high
[Na⫹]pip (⬎35 mM) was observed in about one-half of the
isolated guinea pig ventricular myocytes exposed to hyposmotic medium and an Ip decrease at comparable
[Na⫹]pip in 50% of the myocytes superfused with hypertonic solution. However, only 1 of 10 experiments with
⫹
cells dialyzed with 10 mM Napip
showed a modest Ip
increase (27%) in hypotonic solution (152).
1819
1820
HELFRIED GÜNTHER GLITSCH
Na⫹ concentration in an interactive space representing
⬃14% of the total cell volume (44). This space seems to be
much larger than the subsarcolemmal “fuzzy space” originally proposed (112). Furthermore, electrophysiological
measurements on isolated mouse ventricular myocytes
showed a similar dependence of Ip and Na⫹/Ca2⫹
exchanger current INa/Ca on prior Na⫹/K⫹ activity, presumably because both ion transporters generating the
currents “sense” a similar subsarcolemmal Na⫹ concentration that differs from the bulk [Na⫹] of the cytosol. The
data suggest the “functional presence of so-called Na⫹
fuzzy space” (172), although several questions as to the
existence of such a space remain to be answered. Furthermore, the findings are consistent with the view that
the exchanger and the pump molecules may be colocalized in the sarcolemma (see Ref. 122). The colocalization
of special isoforms of the Na⫹-K⫹ pump and the Na⫹/
Ca2⫹ exchanger may have important physiological and
pharmacological implications. Generally speaking, any
modulation of the cardiac pump activity affects the contractile function of the heart by modifying the Na⫹ /Ca2⫹
exchange. It is widely accepted that cardiac steroids
cause their positive inotropic effect by increasing the
amount of Ca2⫹ released from the sarcoplasmic reticulum
and available for contraction during an action potential.
The increase of [Ca2⫹]i is probably mainly brought about
by an inhibition of the Ca2⫹ efflux via the Na⫹/
Ca2⫹exchanger following Na⫹-K⫹ pump inhibition by the
drugs. The pump inhibition induces an increase in [Na⫹]i
and thereby a reduction of the driving force for the Ca2⫹
efflux (7, 68). Additional cellular mechanisms of cardiac
Physiol Rev • VOL
steroid action and the relation between Na⫹-K⫹ pump
inhibition and cardiac contraction have been discussed
elsewhere in detail (116). It has recently been shown in
rat smooth muscle cells of mesenteric arteries, neurons,
and astrocytes that highly cardiac glycoside-sensitive ␣3and ␣2-subunits (astrocytes) of the Na⫹-K⫹ pump and the
Na⫹/Ca2⫹ exchanger are expressed in a reticular distribution within the cell membrane overlying functional sarcoplasmic (endoplasmic) reticulum. Cardiac steroids (and
endogenous digitalis-like factors) may exert their pharmacological (hormonal) effect by binding to these specialized plasmalemmal domains, modulating [Na⫹]i and
[Ca2⫹]i only in the cleft between cell membrane and sarcoplasmic reticulum (“plasmerosome”) and thereby the
Ca2⫹ content of the adjacent sarcoplasmic reticulum. By
way of contrast, Na⫹-K⫹ pump molecules containing the
cardiac glycoside-insensitive “housekeeping” ␣1-isoform
are uniformly expressed in the rat cell membranes studied
(20, 102).
B. Interaction Between the Cardiac Naⴙ-Kⴙ Pump
and KATP Channels
The effect of altered Na⫹-K⫹ pump activity on IK(ATP)
was investigated by means of patch-clamp measurements
on guinea pig ventricular myocytes (141). Inhibition of the
forward-running pump (i.e., the physiological mode) by
the cardioactive steroid strophanthidin, K⫹-free bathing
solution, or strong hyperpolarization decreased IK(ATP).
Correspondingly, blocking the backward-running (ATP-
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
⫹
FIG. 17. Effect of alterations in transmembrane Na
efflux on Na⫹-K⫹ pump current and Na⫹/Ca2⫹ exchange
current in a single rabbit cardiac Purkinje cell. A: chart
recording of whole cell current at ⫺20 mV holding potential. Top line shows presence of 10.8 mM Ko⫹, short bars
mark the presence of Nio2⫹ (5 mM). Reactivation of the
pump current by 10.8 mM Ko⫹ after a period of 8 min in
K⫹-free medium. B: relation between Nio2⫹-sensitive INa/Ca
exchange current and Na⫹-K⫹ pump current Ip during
pump reactivation. Data are taken from A. [Adapted from
Bielen et al. (16).]
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
C. Effects of the Cardiac Naⴙ-Kⴙ Pump on IK(Na)
A K⫹ channel that is gated by intracellular [Na⫹] ⬎20
mM exists in guinea pig ventricular myocytes. The probability of opening for this channel depends on Na⫹i with a
K0.5 value of 66 mM, but not on Ca2⫹
i or voltage. In passing
it was noted that the inhibition of the Na⫹-K⫹ pump might
be important for the activation of IK(Na) through the channel under pathological conditions (104). This aspect was
emphasized by the demonstration that inhibition of cardiac Na⫹-K⫹ pumping either by ouabain or in K⫹-free
solution activated IK(Na) even at a cytosolic [Na⫹] too low
for the gating of the channel (118). It seems likely that
[Na⫹] at the inner mouth of the channel is, in general,
different from bulk [Na⫹]i. The activity of the Na⫹-K⫹
pump is one important factor governing [Na⫹] at the
channel mouth. [Na⫹] increases following pump inhibiton
Physiol Rev • VOL
and augments the open probability of the IK(Na) channel
(24). The mechanism tends to stabilize the resting potential after an impairment of Na⫹-K⫹ pumping.
D. Blockade of the Naⴙ-Kⴙ Pump Activates IK(ACh)
in Atrial Myocytes
A novel gating mechanism of the cardiac muscarinic
K⫹ channel, independent of G protein activation, was
identified in chick atrial cells (173). The mechanism involves two steps. First, the functional state of the channel
is modified by ATP hydrolysis. Second, [Na⫹]i ⬎3 mM
gates the modified channel with a K0.5 value of ⬃40 mM.
Inhibition of the Na⫹-K⫹ pump either in K⫹-free medium
or by ouabain (5 ⫻ 10⫺4 M) activates IK(ACh) in the same
manner as an increase of [Na⫹]i and implies that the
blockade of the pump activates IK(ACh) via an augmented
subsarcolemmal [Na⫹]. The findings demonstrate for the
first time a Na⫹i -mediated effect of cardiac glycosides on
an atrial ionic channel that is involved in the regulation of
the cardiac rhythm.
XIII. CONCLUSION
The Na⫹-K⫹ pump of animal cells generates the
pump current Ip. Under physiological conditions Ip is an
outward current. It can be measured by electrophysiological methods. The introduction of the patch-clamp techniques has rendered possible a hitherto unequaled experimental control of membrane potential and composition
of the intracellular medium in single cells. The techniques
of whole cell recording and recording from giant patches
permit the study of Na⫹-K⫹ pump characteristics by measuring Ip. For this purpose, cardiac cells are especially
suitable, since they exhibit a high density of sarcolemmal
pump molecules and an adequate cellular geometry and
are easily obtainable. In fact, measurements of the cardiac
Ip have been pivotal for the understanding of Na⫹-K⫹
pumping in electrically excitable cells. The cardiac
Na⫹-K⫹ pump can be studied in its physiological environment, i.e., in the sarcolemma which separates cytosol and
extracellular medium with their different ionic compositions and across which a membrane potential exists. The
effect of intra- and extracellular cations on the cardiac
pump, the interaction between membrane potential and
Na⫹-K⫹ pump, and the modulation of the pump activity by
transmitters, hormones, and drugs have been studied with
unrivaled precision and time resolution by means of
patch-clamp techniques. Through the recording of Ip and
transient pump currents, several voltage-dependent partial reactions of the pump cycle have been identified,
including the binding and unbinding of monovalent cations to or from the pump in an access channel. The
fundamental mechanism of ion translocation by the
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
synthesizing) pump opened the KATP channels. From the
effect of bath-applied strophanthidin on single KATP channels in cell-attached patches, it was deduced that the
interaction between the KATP channels and the Na⫹-K⫹
pump is not restricted to a fuzzy space of submicrometric
dimensions and that the cytosolic ATP concentration controls the subsarcolemmal ATP concentration. A similar
study on the modulation of IK(ATP) by the Na⫹-K⫹ pump
was carried out in giant patches and single guinea pig
ventricular cells (103). Inhibition of the Na⫹-K⫹ pump by
replacing Li⫹ or NMDG⫹ for cytosolic Na⫹ decreased
IK(ATP) in giant patches at 1 mM ATP by ⬃30% and lowered the K0.5 value of ATP for the opening of K(ATP)
channels by ⬃40%. Inhibiton of the pump by Na⫹ substitution had little effect on IK(ATP) at low (⬍0.1 mM) or high
(10 mM) ATP concentration, presumably because the activity of either the pump or of the K(ATP) channels, respectively, is low under these conditions. Similarly, whole cell
recording revealed that inhibition of the Na⫹-K⫹ pump by
5 ⫻ 10⫺5 M ouabain strongly decreased an inward rectifying current at 1 mM but not at 10 mM ATP in the pipette
solution. Under the latter condition, IK(ATP) is absent. The
pump activity probably modulates IK(ATP) via variation of
[ATP] in a subsarcolemmal space, where ATP diffusion is
restricted. Modulation of IK(ATP) by the Na⫹-K⫹ pump
through changes of the subsarcolemmal [ATP] was also
proposed on the basis of patch-clamp measurements on
rabbit ventricular myocytes as the mechanism of the attenuation by digoxin of the infarct size-limiting effect of
ischemic preconditioning (82). The same mechanism
seems to govern the effect of the inhibiton of the Na⫹-K⫹
pump by ouabain on IK(ATP) of guinea pig ventricular
myocytes. Furthermore, the mechanism might be the basis for the time-dependent alleviation by DHO of the
action potential shortening in coronary perfused guinea
pig ventricular muscles during metabolic inhibiton (1).
1821
1822
HELFRIED GÜNTHER GLITSCH
pump, however, remains to be clarified. Furthermore, the
quantitative contribution of electrogenic Na⫹-K⫹ pumping to the cardiac action potential under physiological
conditions has still to be assessed. The knowledge of
pump modulation by transmitters, hormones, and drugs
in cardiac cells is preliminary and requires further detailed studies. Finally, the interaction between Na⫹-K⫹
pump molecules and ion channels or transporters colocalized in the sarcolemma is a promising field of future
research.
REFERENCES
1. ABE T, SATO T, KIYOSUE T, SAIKAWA T, SAKATA T, AND ARITA M.
Inhibiton of Na⫹-K⫹ pump alleviates the shortening of action potential duration caused by metabolic inhibition via blockade of
KATP channels in coronary perfused ventricular muscles of guineapigs. J Mol Cell Cardiol 31: 533–542, 1999.
2. AKERA T AND BRODY TM. The role of Na⫹,K⫹-ATPase in the inotropic
action of digitalis. Pharmacol Rev 29: 187–220, 1978.
3. ALBERS RW, FAHN S, AND KOVAL GJ. The role of sodium ions in the
activation of Electrophorus electric organ adenosine triphosphatase. Proc Natl Acad Sci USA 50: 474 – 481, 1963.
4. APELL H-J. Electrogenic properties of the Na,K pump. J Membr Biol
110: 103–114, 1989.
5. APELL HJ, BORLINGHAUS R, AND LÄUGER P. Fast charge translocations
associated with partial reactions of the Na,K-pump. II. Microscopic
analysis of transient currents. J Membr Biol 97: 179 –191, 1987.
6. BAHINSKI A, NAKAO M, AND GADSBY DC. Potassium translocation by
the Na⫹/K⫹ pump is voltage insensitive. Proc Natl Acad Sci USA 85:
3412–3416, 1988.
7. BAKER PF, BLAUSTEIN MP, HODGKIN AL, AND STEINHARDT RA. The
influence of calcium on sodium efflux in squid axons. J Physiol
(Lond) 200: 431– 468, 1969.
8. BAKER PF AND WILLIS JS. Potassium ions and the binding of cardiac
glycosides to mammalian cells. Nature 226: 521–523, 1970.
9. BARMASHENKO G, KOCKSKÄMPER J, AND GLITSCH HG. Depolarization
increases the apparent affinity of the Na⫹-K⫹ pump to cytoplasmic
Na⫹ in isolated guinea-pig ventricular myocytes. J Physiol (Lond)
517: 691– 698, 1999.
10. BAUMGARTEN CM AND FOZZARD HA. Cardiac resting and pacemaker
potentials. In: The Heart and Cardiovascular System (2nd ed.),
edited by Fozzard HA, Haber E, Jennings RB, and Katz AM. New
York: Raven, 1992, vol. 1, p. 963–1001.
11. BERREBI-BERTRAND I, MAIXENT J-M, GUEDE FG, GERBI A, CHARLEMAGNE
D, AND LELIEVRE LG. Two functional Na⫹/K⫹-ATPase isoforms in the
left ventricle of guinea pig heart. Eur J Biochem 196: 129 –133,
1991.
12. BERLIN JR, FLEMING AJ, AND ISHIZUKA N. Identification of low and
high affinity ouabain-sensitive Na pump current in voltage-clamped
rat cardiac myocytes. Ion-motive ATPases: structure, function, and
regulation. Ann NY Acad Sci 671: 440 – 442, 1992.
Physiol Rev • VOL
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
I am grateful to Profs. H. C. Lüttgau (Bochum), W. Schwarz
(Frankfurt), and F. Verdonck (Kortrijk/Leuven) for friendly support and advice over many years and for critical reading of an
earlier draft of the manuscript. I also thank Drs. S. Erlenkamp
and S. Zillikens for their pertinent comments on the manuscript.
I owe a pleasant, excellent cooperation for more than 30 years
to A. Balzer-Ferrai, W. Grabowski, P. Greger, and Dr. H. Pusch
(Dept. of Cell Physiology, Ruhr-University Bochum). Last but
not least, I thank P. Lerch for invaluable secretarial help.
The financial support by the Deutsche Forschungsgemeinschaft for nearly three decades is gratefully acknowledged.
Address for reprint requests and other correspondence:
H. G. Glitsch, Buchenberger Str. 6, 78126 Königsfeld, Germany.
13. BERLIN JR AND PELUFFO RD. Mechanism of electrogenic reaction
steps during K⫹ transport by the Na,K-ATPase. Na/K-ATPase and
related transport ATPases: structure, mechanism, and regulation.
Ann NY Acad Sci 834: 251–259, 1997.
14. BEWICK NL, FERNANDES C, PITT AD, RASMUSSEN HH, AND WHALLEY
DW. Mechanisms of Na⫹-K⫹ pump regulation in cardiac myocytes
during hyposmolar swelling. Am J Physiol Cell Physiol 276: C1091–
C1099, 1999.
15. BIELEN FV, GLITSCH HG, AND VERDONCK F. Dependence of Na⫹ pump
current on external monovalent cations and membrane potential in
rabbit cardiac Purkinje cells. J Physiol (Lond) 442: 169 –189, 1991.
16. BIELEN FV, GLITSCH HG, AND VERDONCK F. Changes of the subsarcolemmal Na⫹ concentration in internally perfused cardiac cells.
Biochim Biophys Acta 1065: 269 –271, 1991.
17. BIELEN FV, GLITSCH HG, AND VERDONCK F. The kinetics of the inhibition by dihydroouabain of the sodium pump current in single
rabbit cardiac Purkinje cells. Naunyn-Schmiedeberg’s Arch Pharmacol 345: 100 –107, 1992.
18. BIELEN FV, GLITSCH HG, AND VERDONCK F. Na⫹ pump current-voltage
relationships of rabbit cardiac Purkinje cells in Na⫹-free solution.
J Physiol (Lond) 465: 699 –714, 1993.
19. BLANCO G AND MERCER RW. Isozymes of the Na-K-ATPase: heterogeneity in structure, diversity in function. Am J Physiol Renal
Physiol 275: F633–F650, 1998.
20. BLAUSTEIN; MP, JUHASZOVA M, AND GOLOVINA VA. The cellular mechanism of action of cardiotonic steroids: a new hypothesis. Clin Exp
Hypertens 20: 691–703, 1998.
21. BRONNER C, MOUSLI M, ELENO N, AND LANDRY Y. Resting plasma
membrane potential of rat peritoneal mast cells is set predominantly by the sodium pump. FEBS Lett 255: 401– 404, 1989.
22. BUHAGIAR KA, HANSEN PS, GRAY DF, MIHAILIDOU AS, AND RASMUSSEN
HH. Angiotensin regulates the selectivity of the Na⫹-K⫹ pump for
intracellular Na⫹. Am J Physiol Cell Physiol 277 : C461–C468, 1999.
23. CARMELIET E. A fuzzy subsarcolemmal space for intracellular Na⫹ in
cardiac cells? Cardiovasc Res 26: 433– 442, 1992.
24. CARMELIET E AND LUK H-N. The Na⫹-activated K⫹ channel in cardiac
cells. In: Intracellular Regulation of Ion Channels, edited by Morad M and Agus Z. Berlin: Springer, 1992, vol. H60, p. 53–59. (NATO
ASI Ser.)
25. COHEN IS, DATYNER NB, GINTANT GA, MULRINE NK, AND PENNEFATHER
P. Properties of electrogenic sodium-potassium pump in isolated
canine Purkinje myocytes. J Physiol (Lond) 383: 251–267, 1987.
26. CONNELLY CM. Recovery processes and metabolism of nerve. Rev
Mod Phys 31: 475– 484, 1959.
27. CRAMBERT G, HASLER U, BEGGAH AT, YU C, MODYANOV NN, HORISBERGER J-D, LELIEVRE L, AND GEERING K. Transport and pharmacological properties of nine different human Na,K-ATPase isozymes.
J Biol Chem 275: 1976 –1986, 2000.
28. DAUT J AND RÜDEL R. The electrogenic sodium pump in guinea-pig
ventricular muscle: inhibiton of pump current by cardiac glycosides. J Physiol (Lond) 330: 243–264, 1982.
29. DÉLÈZE J. Possible reasons for drop of resting potential of mammalian heart preparations during hypothermia. Circ Res 8: 553–557,
1960.
30. DÉSILETS M AND BAUMGARTEN CM. Isoproterenol directly stimulates
the Na⫹-K⫹ pump in isolated cardiac myocytes. Am J Physiol
Heart Circ Physiol 251: H218 –H225, 1986.
31. DE WEER P. The electrogenic sodium pump: thermodynamics and
kinetics. In: Membrane Control of Cellular Activity, edited by
Lüttgau HC. Stuttgart: Fischer, 1986, vol. 33, p. 387–399.
32. DE WEER P. Cellular sodium-potassium transport. In: The Kidney:
Physiology and Pathophysiology (2nd ed.), edited by Seldin DW
and Giebisch G. New York: Raven, 1992, p. 93–112.
33. DE WEER P, GADSBY DC, AND RAKOWSKI RF. Voltage dependence of
the Na-K pump. Annu Rev Physiol 50: 225–241, 1988.
34. DE WEER P AND RAKOWSKI RF. Current generated by backwardrunning electrogenic Na pump in squid giant axons. Nature 309:
450 – 452, 1984.
35. DOBRETSOV M, HASTINGS SL, AND STIMERS JR. Na⫹-K⫹ pump cycle
during ␤-adrenergic stimulation of adult rat cardiac myocytes.
J Physiol (Lond) 507: 527–539, 1998.
36. DOOHAN MM, GRAY DF, HOOL LC, ROBINSON BG, AND RASMUSSEN HH.
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
37.
38.
39.
40.
41.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
Physiol Rev • VOL
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
line on the Na⫹-K⫹ pump current in guinea-pig ventricular myocytes. J Physiol (Lond) 501: 527–535, 1997.
GAO J, MATHIAS RT, COHEN IS, SHI J, AND BALDO GJ. The effects of
␤-stimulation on the Na⫹-K⫹ pump current-voltage relationship in
guinea-pig ventricular myocytes. J Physiol (Lond) 494: 697–708,
1996.
GAO J, MATHIAS RT, COHEN IS, SUN X, AND BALDO GJ. Modulators of
PKC affect Na/K pummp current in guinea pig ventricular myocytes
(Abstract). Biophys J 72: A51, 1997.
GAO J, MATHIAS RT, COHEN IS, WANG Y, SUN X, AND BALDO GJ.
Activation of PKC increases Na⫹-K⫹ pump current in ventricular
myocytes from guinea pig heart. Pflügers Arch 437: 643– 651, 1999.
GAO J, WYMORE R, WYMORE RT, WANG Y, MCKINNON D, DIXON JE,
MATHIAS RT, COHEN IS, AND BALDO GJ. Isoform-specific regulation of
the sodium pump by ␣- and ␤-adrenergic agonists in the guinea-pig
ventricle J. Physiol (Lond) 516: 377–383, 1999.
GLITSCH HG. Characteristics of active Na transport in intact cardiac
cells. Am J Physiol Heart Circ Physiol 236: H189 –H199, 1979.
GLITSCH HG. Electrogenic Na pumping in the heart. Annu Rev
Physiol 44: 389 – 400, 1982.
GLITSCH HG AND KRAHN T. The cardiac electrogenic Na pump. In:
Membrane Control of Cellular Activity, edited by Lüttgau HC.
Stuttgart: Fischer, 1986, vol. 33, p. 401– 417.
GLITSCH HG, KRAHN T, AND PUSCH H. The dependence of sodium
pump current on internal Na concentration and membrane potential in cardioballs from sheep Purkinje fibres. Pflügers Arch 414:
52–58, 1989.
GLITSCH HG, KRAHN T, PUSCH H, AND SULEYMANIAN M. Effect of
isoprenaline on active Na transport in sheep cardiac Purkinje fibres. Pflügers Arch 415: 88 –94, 1989.
GLITSCH HG, REUTER H, AND SCHOLZ H. The effect of the internal
sodium concentration on calcium fluxes in isolated guinea-pig auricles. J Physiol (Lond) 209: 25– 43, 1970.
GLITSCH HG, SCHWARZ W, AND TAPPE A. Cardiac Na⫹ pump currentvoltage relationships at various transmembrane gradients of the
pumped cations. Biochim Biophys Acta 1278: 137–146, 1996.
GLITSCH HG AND TAPPE A. The Na⫹/K⫹ pump of cardiac Purkinje
cells is preferentially fuelled by glycolytic ATP production. Pflügers
Arch 422: 380 –385, 1993.
GLITSCH HG AND TAPPE A. Change of Na⫹ pump current reversal
potential in sheep cardiac Purkinje cells with varying free energy of
ATP hydrolysis. J Physiol (Lond) 484: 605– 616, 1995.
GLYNN IM. “All hands to the sodium pump.” J Physiol (Lond) 462:
1–30, 1993.
GLYNN IM AND KARLISH SJD. The sodium pump. Annu Rev Physiol
37: 13–55, 1975.
GOLDSHLEGGER R, KARLISH SJD, REPHAELI A, AND STEIN WD. The effect
of membrane potential on the mammalian sodium-potassium pump
reconstituted into phospholipid vesicles. J Physiol (Lond) 387:
331–355, 1987.
GRAY DF, HANSEN PS, DOOHAN MM, HOOL LC, AND RASMUSSEN HH.
Dietary cholesterol affects Na⫹-K⫹ pump function in rabbit cardiac
myocytes. Am J Physiol Heart Circ Physiol 272 : H1680 –H1689,
1997.
GRAY DF, MIHAILIDOU AS, HANSEN PS, BUHAGIAR KA, BEWICK NL,
RASMUSSEN HH, AND WHALLEY DW. Amiodarone inhibits the Na⫹-K⫹
pump in rabbit cardiac myocytes after acute and chronic treatment.
J Pharmacol Exp Ther 284: 75– 82, 1998.
HAMILL OP, MARTY A, NEHER E, SAKMANN B, AND SIGWORTH FJ. Improved patch-clamp techniques for high-resolution current recording from cell and cell-free membrane patches. Pflügers Arch 391:
85–100, 1981.
HANSEN O. Interaction of cardiac glycosides with (Na⫹⫹K⫹)-activated ATPase. A biochemical link to digitalis-induced inotropy.
Pharmacol Rev 36: 143–163, 1984.
HANSEN PS, BUHAGIAR KA, GRAY DF, AND RASMUSSEN HH. Voltagedependent stimulation of the Na⫹-K⫹ pump by insulin in rabbit
cardiac myocytes. Am J Physiol Cell Physiol 278: C546 –C553,
2000.
HANSEN PS, GRAY DF, AND RASMUSSEN HH. Insulin regulates the
voltage dependence of the sarcolemmal Na⫹-K⫹ pump (Abstract).
Circulation 92, Suppl I: I-638, 1995.
HANSEN U-P, GRADMANN D, SANDERS D, AND SLAYMAN CL. Interpreta-
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
42.
Thyroid status and regulation of intracellular sodium in rabbit
heart. Am J Physiol Heart Circ Physiol 272: H1589 –H1597, 1997.
DOOHAN MM, HOOL LC, AND RASMUSSEN HH. Thyroid status and
Na⫹-K⫹ pump current, intracellular sodium, and action potential
duration in rabbit heart. Am J Physiol Heart Circ Physiol 268:
H1838 –H1846, 1995.
EFTHYMIADIS A AND SCHWARZ W. Conditions for a backward-running
Na⫹/K⫹ pump in Xenopus oocytes. Biochim Biophys Acta 1068:
73–76, 1991.
EISNER DA AND LEDERER WJ. The role of the sodium pump in the
effects of potassium-depleted solutions on mammalian cardiac
muscle. J Physiol (Lond) 294: 279 –301, 1979.
EISNER DA AND SMITH TW. The Na-K pump and its effectors in
cardiac muscle. In: The Heart and Cardiovascular System (2nd
ed.), edited by Fozzard HA, Haber E, Jennings RB, and Katz AH.
New York: Raven, 1992, vol. 1, p. 863–902.
ERLENKAMP S, GRETZER B, ZILLIKENS S, GLITSCH HG, PUSCH H, STA⫹
⫹
ROSKE T, AND WELZEL P. Na /K pump inhibition and positive inotropic effect of digitoxigenin and some C-22-substituted derivatives
in sheep cardiac preparations. Naunyn-Schmiedeberg’s Arch Pharmacol 357: 54 – 62, 1998.
FRIEDRICH T, BAMBERG E, AND NAGEL G. Na⫹,K⫹-ATPase pump currents in giant excised patches activated by an ATP concentration
jump. Biophys J 71: 2486 –2500, 1996.
FRIEDRICH T AND NAGEL G. Comparison of Na⫹/K⫹-ATPase pump
currents activated by ATP concentration or voltage jumps. Biophys
J 73: 186 –194, 1997.
FUJIOKA Y, MATSUOKA S, BAN T, AND NOMA A. Interaction of the
Na⫹-K⫹ pump and Na⫹-Ca2⫹ exchange via [Na⫹]i in a restricted
space of guinea-pig ventricular cells. J Physiol (Lond) 509: 457–
470, 1998.
GADSBY DC. The Na/K pump of cardiac cells. Annu Rev Biophys
Bioeng 13: 373–398, 1984.
GADSBY DC. The Na/K pump of cardiac myocytes. In: Cardiac
Electrophysiology and Arrhythmias, From Cell to Bedside, edited
by Zipes DP and Jalife J. New York: Saunders, 1991, p. 35–51.
GADSBY DC AND CRANEFIELD PF. Effects of electrogenic sodium
extrusion on the membrane potential of cardiac Purkinje fibers. In:
Normal and Abnormal Conduction in the Heart, edited by Paes de
Carvalho A, Hoffman BF, and Lieberman M. Mount Kisko, NY:
Futura, 1982, p. 225–247.
GADSBY DC, KIMURA J, AND NOMA A. Voltage dependence of Na/K
pump current in isolated heart cells. Nature 315: 63– 65, 1985.
GADSBY DC AND NAKAO M. Dependence of Na/K pump current on
intracellular [Na] in isolated cells from guinea-pig ventricle (Abstract). J Physiol (Lond) 371: 201P, 1986.
GADSBY DC AND NAKAO M. [Na] dependence of the Na/K pump
current-voltage relationship in isolated cells from guinea-pig ventricle (Abstract). J Physiol (Lond) 382: 106P, 1987.
GADSBY DC AND NAKAO M. Steady-state current-voltage relationship
of the Na/K pump in guinea pig ventricular myocytes. J Gen Physiol
94: 511–537, 1989.
GADSBY DC, NAKAO M, BAHINSKI A, NAGEL G, AND SVENSON M. Charge
movements via the cardiac Na,K-ATPase. Acta Physiol Scand 146:
111–123, 1992.
GADSBY DC, RAKOWSKI RF, AND DE WEER P. Extracellular access to
the Na,K pump: pathway similar to ion channel. Science 260: 100 –
103, 1993.
GAO J, COHEN IS, MATHIAS RT, AND BALDO GJ. Regulation of the
␤-stimulation of the Na⫹-K⫹ pump current in guinea-pig ventricular
myocytes by a cAMP-dependent PKA pathway. J Physiol (Lond)
477: 373–380, 1994.
GAO J, COHEN IS, MATHIAS RT, AND BALDO GJ. The inhibitory effect of
␤-stimulation on the Na/K pump current in guinea pig ventricular
myocytes is mediated by a cAMP-dependent PKA pathway.
Pflügers Arch 435: 479 – 484, 1998.
GAO J, MATHIAS RT, COHEN IS, AND BALDO GJ. Isoprenaline, Ca2⫹ and
the Na⫹-K⫹ pump in guinea-pig ventricular myocytes. J Physiol
(Lond) 449: 689 –704, 1992.
GAO J, MATHIAS RT, COHEN IS, AND BALDO GJ. Two functionally
different Na/K pumps in cardiac ventricular myocytes. J Gen
Physiol 106: 995–1030, 1995.
GAO J, MATHIAS RT, COHEN IS, AND BALDO GJ. Effects of acetylcho-
1823
1824
82.
83.
84.
85.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
tion of current-voltage relationships for “active” ion transport systems: I. Steady-state reaction-kinetic analysis of class-I mechanisms. J Membr Biol 63: 165–190, 1981.
HARUNA T, HORIE M, KOUCHI I, NAWADA R, TSUCHIYA K, AKAO M, OTANI
H, MURAKAMI T, AND SASAYAMA S. Coordinate interaction between
ATP-sensitive K⫹ channel and Na⫹,K⫹-ATPase modulates ischemic
preconditioning. Circulation 98: 2905–2910, 1998.
HEMSWORTH PD, WHALLEY DW, AND RASMUSSEN HH. Electrogenic
Li⫹/Li⫹ exchange mediated by the Na⫹-K⫹ pump in rabbit cardiac
myocytes. Am J Physiol Cell Physiol 272 : C1186 –C1192, 1997.
HERMANS AN. Ion Translocation by the Na⫹/K⫹ Pump: Influence of
Membrane Potential and Cardiac Glycosides (MD thesis). Leuven,
Belgium: Acta Biomed Lovan, 1997, vol. 143, p. 1–140.
HERMANS AN, GLITSCH HG, AND VERDONCK F. The effect of cardiac
glycosides on the Na⫹ pump current-voltage relationship of isolated rat and guinea-pig heart cells. J Physiol (Lond) 481: 279 –291,
1994.
HERMANS AN, GLITSCH HG, AND VERDONCK F. The antagonistic effect
of Ko⫹ and dihydro-ouabain on the Na⫹ pump current of single rat
and guinea-pig cardiac cells. J Physiol (Lond) 484: 617– 628, 1995.
HERMANS AN, GLITSCH HG, AND VERDONCK F. Activation of the
Na⫹/K⫹ pump current by intra- and extracellular Li ions in single
guinea-pig cardiac cells. Biochim Biophys Acta 1330: 83–93, 1997.
HERZIG S, LÜLLMANN H, AND MOHR K. On the cooperativity of
ouabain-binding to intact myocardium. J Mol Cell Cardiol 17: 1095–
1104, 1985.
HEYSE S, WUDDEL I, APELL H-J, AND STÜRMER W. Partial reactions of
the Na,K-ATPase: determination of rate constants. J Gen Physiol
104: 197–240, 1994.
HILGEMANN DW. Giant excised cardiac sarcolemmal membrane
patches: sodium and sodium-calcium exchange currents. Pflügers
Arch 415: 247–249, 1989.
HILGEMANN DW. Channel-like function of the Na,K pump probed at
microsecond resolution in giant membrane patches. Science 263:
1429 –1432, 1994.
HILGEMANN DW. Recent electrical snapshots of the cardiac Na,K
pump. Na/K-ATPase and related transport ATPases: structure,
mechanism, and regulation. Ann NY Acad Sci 834: 260 –269, 1997.
HILGEMANN DW, NAGEL GA, AND GADSBY DC. Na/K pump current in
giant membrane patches excised from ventricular myocytes. In:
The Sodium Pump: Recent Developments, edited by Kaplan JH and
De Weer P. New York: Rockefeller Univ. Press, 1991, p. 543–547.
HODGKIN AL AND KEYNES RD. Active transport of cations in giant
axons from Sepia and Loligo. J Physiol (Lond) 128: 28 – 60, 1955.
HOLMGREN M, WAGG J, BEZANILLA F, RAKOWSKI RF, DE WEER P, AND
GADSBY DC. Three distinct and sequential steps in the release of
sodium ions by the Na⫹/K⫹-ATPase. Nature 403: 898 –901, 2000.
HOOL LC, GRAY DF, ROBINSON BG, AND RASMUSSEN HH. Angiotensinconverting enzyme inhibitors regulate the Na⫹-K⫹ pump via effects
on angiotensin metabolism. Am J Physiol Cell Physiol 271: C172–
C180, 1996.
HOOL LC, WHALLEY DW, DOOHAN MM, AND RASMUSSEN HH. Angiotensin-converting enzyme inhibiton, intracellular Na⫹, and Na⫹-K⫹
pumping in cardiac myocytes. Am J Physiol Cell Physiol 268:
C366 –C375, 1995.
HORISBERGER J-D, LEMAS V, KRAEHENBÜHL J-P, AND ROSSIER BC. Structure-function relationship of Na,K-ATPase. Annu Rev Physiol 53:
565–584, 1991.
ISENBERG G AND TRAUTWEIN W. The effect of dihydro-ouabain and
lithium-ions on the outward current in cardiac Purkinje fibres.
Pflügers Arch 350: 41–54, 1974.
ISHIDA Y AND CHUSED TM. Lack of voltage sensitive potassium
channels and generation of membrane potential by sodium potassium ATPase in murine T lymphocytes. J Immunol 151: 610 – 620,
1993.
ISHIZUKA N AND BERLIN JR. ␤-Adrenergic stimulation does not regulate Na pump function in voltage-clamped ventricular myocytes of
the rat heart. Pflügers Arch 424: 361–363, 1993.
JUHASZOVA M AND BLAUSTEIN MP. Na⫹ pump low and high ouabain
affinity ␣ subunit isoforms are differently distributed in cells. Proc
Natl Acad Sci USA 94: 1800 –1805, 1997.
KABAKOV AY. Activation of KATP channels by Na/K pump in isolated
Physiol Rev • VOL
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
cardiac myocytes and giant membrane patches. Biophys J 75:
2858 –2867, 1998.
KAMEYAMA M, KAKEI M, SATO R, SHIBASAKI T, MATSUDA H, AND IRISAWA
H. Intracellular Na⫹ activates a K⫹ channel in mammalian cardiac
cells. Nature 309: 354 –356, 1984.
KAWAKAMI K, NOGUCHI S, NODA M, TAKAHASHI H, OHTA T, KAWAMURA
M, NOJIMA H, NAGANO K, HIROSE T, INAYAMA S, HAYASHIDA H, MIYATA
T, AND NUMA S. Primary structure of the ␣-subunit of Torpedo
californica (Na⫹ ⫹ K⫹) ATPase deduced from cDNA sequence.
Nature 316: 733–736, 1985.
KAWAKAMI K, NOJIMA H, OHTA T, AND NAGANO K. Molecular cloning
and sequence analysis of human Na, K-ATPase ␤-subunit. Nucleic
Acids Res 14: 2833–2844, 1986.
KINARD TA, LIU X-Y, LIU S, AND STIMERS JR. Effect of Napip on Ko
activation of the Na-K pump in adult rat cardiac myocytes. Am J
Physiol Cell Physiol 266: C37–C41, 1994.
KOCKSKÄMPER J, ERLENKAMP S, AND GLITSCH HG. Activation of the
cAMP-protein kinase A pathway facilitates Na⫹ translocation by
the Na⫹-K⫹ pump in guinea-pig ventricular myocytes. J Physiol
(Lond): 523: 561–574, 2000.
KOCKSKÄMPER J AND GLITSCH HG. Sodium pump of cultured guinea
pig atrial myocytes. Na/K-ATPase and related transport ATPases:
structure, mechanism, and regulation. Ann NY Acad Sci 834: 354 –
356, 1997.
LAFAIRE AV AND SCHWARZ W. Voltage dependence of the rheogenic
Na⫹/K⫹ ATPase in the membrane of oocytes of Xenopus laevis. J
Membr Biol 91: 43–51, 1986.
LÄUGER P. Electrogenic Ion Pumps. Sunderland, MA: Sinauer, 1991,
p. 1–313.
LEDERER WJ, NIGGLI E, AND HADLEY RW. Sodium-calcium exchange
in excitable cells: fuzzy space. Science 248: 283, 1990.
LEVI AJ. The effect of strophanthidin on action potential, calcium
current and contraction in isolated guinea-pig ventricular myocytes. J Physiol (Lond) 443: 1–23, 1991.
LEVI AJ. The electrogenic sodium/potassium pump and passive
sodium influx of isolated guinea pig ventricular myocytes. J Cardiovasc Electrophysiol 3: 225–238, 1992.
LEVI AJ. A role for sodium/calcium exchange in the action potential
shortening caused by strophanthidin in guinea pig ventricular myocytes. Cardiovasc Res 27: 471– 481, 1993.
LEVI AJ, BOYETT MR, AND LEE CO. The cellular actions of digitalis
glycosides on the heart. Prog Biophys Mol Biol 62: 1–54, 1994.
LINGREL JB, ARGÜELLO JM, VAN HUYSSE J, AND KUNTZWEILER TA.
Cation and cardiac glycoside binding sites of the Na, K-ATPase.
Na/K-ATPase and related transport ATPases: structure, mechanism, and regulation. Ann NY Acad Sci 834: 194 –206, 1997.
LUK H-N AND CARMELIET E. Na⫹-activated K⫹ current in cardiac
cells: rectification, open probabilitiy, block and role in digitalis
toxicity. Pflügers Arch 416: 766 –768, 1990.
LÜTTGAU HC AND GLITSCH H. Membrane physiology of nerve and
muscle fibres. Prog Zool 24: 1–132, 1976.
MAIN MJ, GRANTHAM CJ, AND CANNELL MB. Changes in subsarcolemmal sodium concentration measured by Na-Ca exchanger activity
during Na-pump inhibition and ␤-adrenergic stimulation in guineapig ventricular myocytes. Pflügers Arch 435: 112–118, 1997.
MATHIAS RT, COHEN IS, AND OLIVA C. Limitations of the whole cell
patch clamp technique in the control of intracellular concentrations. Biophys J 58: 759 –770, 1990.
MCDONOUGH AA, ZHANG Y, SHIN V, AND FRANK JS. Subcellular distribution of sodium pump isoform subunits in mammalian cardiac
myocytes. Am J Physiol Cell Physiol 270 : C1221–C1227, 1996.
MEHRKE G, DAUT J, HARMANN S, AND DISCHNER A. The effects of
dihydro-ouabain on the current-voltage relation of guinea-pig cardiomyocytes (Abstract). Pflügers Arch 408 Suppl 1: R10, 1987.
MERCER RW, BIEMESDERFER D, BLISS DP, COLLINS JH, AND FORBUSH B.
Molecular cloning and immunological characterization of the
␥-polypeptide, a small protein associated with the Na,K-ATPase.
J Cell Biol 121: 579 –586, 1993.
MIHAILIDOU AS, BUHAGIAR KA, AND RASMUSSEN HH. Na⫹ influx and
Na⫹-K⫹ pump activation during short-term exposure of cardiac
myocytes to aldosterone. Am J Physiol Cell Physiol 274: C175–
C181, 1998.
MIHAILIDOU AS, BUNDGAARD H, MARDINI M, HANSEN PS, KJELDSEN K,
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
86.
HELFRIED GÜNTHER GLITSCH
ELECTROGENIC Na⫹-K⫹ PUMPING IN CARDIAC CELLS
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.
Physiol Rev • VOL
149. RASMUSSEN HH, TEN EICK RE, OKITA GT, HARTZ RS, AND SINGER DH.
Inhibiton of electrogenic Na-pumping attributable to binding of
cardiac steroids to high-affinity pump sites in human atrium.
J Pharmacol Exp Ther 235: 629 – 635, 1985.
150. SAGAR A AND RAKOWSKI RF. Access channel model for the voltage
dependence of the forward-running Na⫹/K⫹ pump. J Gen Physiol
103: 869 – 894, 1994.
151. SAKAI R, HAGIWARA N, MATSUDA N, KASANUKI H, AND HOSODA S.
Sodium-potassium pump current in rabbit sino-atrial node cells.
J Physiol (Lond) 490: 51– 62, 1996.
152. SASAKI N, MITSUIYE T, WANG Z, AND NOMA A. Increase of the delayed
rectifier K⫹ and Na⫹-K⫹ pump currents by hypotonic solutions in
guinea pig cardiac myocytes. Circ Res 75: 887– 895, 1994.
153. SCHATZMANN H-J. Herzglykoside als Hemmstoffe für den aktiven
Kalium- und Natriumtransport durch die Erythrocytenmembran.
Helv Physiol Pharmacol Acta 11: 346 –354, 1953.
154. SCHONER W, THÖNGES D, HAMER E, ANTOLOVIC R, BUXBAUM E, WILLEKE
M, SERPERSU EH, AND SCHEINER-BOBIS G. Is the sodium pump a
functional dimer? In: The Sodium Pump-Structure Mechanism,
Hormonal Control and Its Role in Disease, edited by Bamberg E
and Schoner W. New York: Springer, 1994, p. 332–341.
155. SCHUURMANS STECKHOVEN F AND BONTING SL. Transport adenosine
triphosphatases: properties and functions. Physiol Rev 61: 1–76,
1981.
156. SCHWARTZ A, LINDENMAYER GE, AND ALLEN JC. The sodium-potassium
adenosine triphosphatase: pharmacological, physiological and biochemical aspects. Pharmacol Rev 27: 3–134, 1975.
157. SCHWARZ W AND GU Q. Characteristics of the Na⫹/K⫹-ATPase from
Torpedo californica expressed in Xenopus oocytes: a combination
of tracer flux measurements with electrophsiological measurements. Biochim Biophys Acta 945: 167–174, 1988.
158. SCHWEIGERT B, LAFAIRE AV, AND SCHWARZ W. Voltage dependence of
the Na-K ATPase: measurements of ouabain-dependent membrane
current and ouabain binding in oocytes of Xenopus laevis. Pflügers
Arch 412: 579 –588, 1988.
159. SEMB SO AND SEJERSTED OM. Fuzzy space and control of Na⫹,K⫹pump rate in heart and skeletal muscle. Acta Physiol Scand 156:
213–225, 1996.
160. SHAH A, COHEN IS, AND ROSEN MR. Stimulation of cardiac alpha
receptors increases Na/K pump current and decreases gK via a
pertussis toxin-sensitive pathway. Biophys J 54: 219 –225, 1988.
161. SHATTOCK MJ AND MATSUURA H. Measurement of Na⫹-K⫹ pump
current in isolated rabbit ventricular myocytes using the whole-cell
voltage-clamp technique. Circ Res 72: 91–101, 1993.
162. SHULL GE, LANE LK, AND LINGREL JB. Amino-acid sequence of the
␤-subunit of the (Na⫹ ⫹ K⫹) ATPase deduced from a cDNA. Nature
321: 429 – 431, 1986.
163. SHULL GE, SCHWARTZ A, AND LINGREL JB. Amino-acid sequence of the
catalytic subunit of the (Na⫹ ⫹ K⫹) ATPase deduced from a
complementary DNA. Nature 316: 691– 695, 1985.
164. SHYJAN AW AND LEVENSON R. Antisera specific for the ␣1, ␣2, ␣3, and
␤ subunits of the Na,K-ATPase: differential expression of ␣ and ␤
subunits in rat tissue membranes. Biochemistry 28: 4531– 4535,
1989.
165. SKOU JC. The influence of some cations on an adenosine triphosphatase from peripheral nerves. Biochim Biophys Acta 23: 394 –
401, 1957.
166. STIMERS JR AND DOBRETSOV M. Adrenergic stimulation of Na/K pump
current in adult rat cardiac myocytes in short-term culture. J
Membr Biol 163: 205–216, 1998.
167. STIMERS JR, LIU S, AND KINARD TA. Effect of Nai on activity and
voltage dependence of the Na/K pump in adult rat cardiac myocytes. J Membr Biol 135: 39 – 47, 1993.
168. STIMERS JR, LIU S, AND LIEBERMAN M. Apparent affinity of the Na/K
pump for ouabain in cultured chick cardiac myocytes. J Gen
Physiol 98: 815– 833, 1991.
169. STIMERS JR, LOBAUGH LA, LIU S, SHIGETO N, AND LIEBERMAN M. Intracellular sodium affects ouabain interaction with the Na/K pump in
cultured chick cardiac myocytes. J Gen Physiol 95: 77–95, 1990.
170. STIMERS JR, SHIGETO N, AND LIEBERMAN M. Na/K pump current in
aggregates of cultured chick cardiac myocytes. J Gen Physiol 95:
61–76, 1990.
171. STÜRMER W, BÜHLER R, APELL H-J, AND LÄUGER P. Charge transloca-
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
RASMUSSEN HH. Hyperaldosteronemia in rabbits inhibits the
cardiac sarcolemmal Na⫹-K⫹ pump. Circ Res 86: 37– 42, 2000.
MOGUL DJ, RASMUSSEN HH, SINGER DH, AND TEN EICK RE. Inhibition
of Na-K pump current in guinea pig ventricular myocytes by dihydroouabain occurs at high- and low-affinity sites. Circ Res 64:
1063–1069, 1989.
MOGUL DJ, SINGER DH, AND TEN EICK RE. Dependence of Na-K pump
current on internal Na⫹ in mammalian cardiac myocytes. Am J
Physiol Heart Circ Physiol 259: H488 –H496, 1990.
MULLINS LJ AND NODA K. The influence of sodium-free solutions on
the membrane potential of frog muscle fibers. J Gen Physiol 47:
117–132, 1963.
NAKAO M AND GADSBY DC. Voltage dependence of Na translocation
by the Na/K pump. Nature 323: 628 – 630, 1986.
NAKAO M AND GADSBY DC. [Na] and [K] dependence of the Na/K
pump current-voltage relationship in guinea pig ventricular myocytes. J Gen Physiol 94: 539 –565, 1989.
NOBLE D. The Initiation of the Heartbeat. Oxford UK: Clarendon,
1975, p. 1–156.
OMAY HS AND SCHWARZ W. Voltage-dependent stimulation of Na⫹/
K⫹-pump current by external cations: selectivity of different K⫹
congeners. Biochim Biophys Acta 1104: 167–173, 1992.
OR E, GOLDSHLEGER R, AND KARLISH SJD. An effect of voltage on
binding of Na⫹ at the cytoplasmic surface of the Na⫹-K⫹ pump.
J Biol Chem 271: 2470 –2477, 1996.
PELUFFO RD AND BERLIN JR. Electrogenic K⫹ transport by the
Na⫹-K⫹ pump in rat cardiac ventricular myocytes. J Physiol
(Lond) 501: 33– 40, 1997.
PHILIPSON KD AND NISHIMOTO AY. ATP-dependent Na⫹ transport in
cardiac sarcolemmal vesicles. Biochim Biophys Acta 733: 133–141,
1983.
POST RL, HEGYVARY C, AND KUME S. Activation by adenosine triphosphate in the phosphorylation kinetics of sodium and potassium ion
transport adenosine triphosphatase. J Biol Chem 247: 6530 – 6540,
1972.
POST RL, MERRIT CR, KINSOLVING CR, AND ALBRIGHT CD. Membrane
adenosine triphosphatase as a participant in the active transport of
sodium and potassium in the human erythrocyte. J Biol Chem 235:
1796 –1802, 1960.
POST RL, SEN AK, AND ROSENTHAL AS. A phosphorylated intermediate in adenosine triphosphatase-dependent sodium and potassium
transport across kidney membranes. J Biol Chem 240: 1437–1445,
1965.
POWELL T AND TWIST VW. A rapid technique for the isolation and
purification of adult cardiac muscle cells having respiratory control
and a tolerance to calcium. Biochem Biophys Res Commun 72:
327–333, 1976.
PRIEBE L, FRIEDRICH M, AND BENNDORF K. Functional interaction
between KATP channels and the Na⫹-K⫹ pump in metabolically
inhibited heart cells of the guinea-pig. J Physiol (Lond) 492: 405–
417, 1996.
RAKOWSKI RF, BEZANILLA F, DE WEER P, GADSBY DC, HOLMGREN M,
AND WAGG J. Charge translocation by the Na/K pump. Na/K-ATPase
and related transport ATPases: structure, mechanism, and regulation. Ann NY Acad Sci 834: 231–243, 1997.
RAKOWSKI RF, DE WEER P, AND GADSBY DC. Current-voltage relationship of the backward-running Na/K pump in voltage-clamped internally-dialyzed squid giant axons (Abstract). J Biophys 53: 223a,
1988.
RAKOWSKI RF, GADSBY DC, AND DE WEER P. Stoichiometry and
voltage dependence of the sodium pump in voltage-clamped, internally dialyzed squid giant axon. J Gen Physiol 93: 903–941, 1989.
RAKOWSKI RF, GADSBY DC, AND DE WEER P. Voltage dependence of
the Na/K pump. J Membr Biol 155: 105–112, 1997.
RAKOWSKI RF AND PAXSON CL. Voltage dependence of Na/K pump
current in Xenopus oocytes. J Membr Biol 106: 173–182, 1988.
RAKOWSKI RF, VASILETS LA, LATONA J, AND SCHWARZ W. A negative
slope in the current-voltage relationship of the Na⫹/K⫹ pump in
Xenopus oocytes produced by reduction of external [K⫹]. J Membr
Biol 121: 177–187, 1991.
RASMUSSEN HH, OKITA GT, HARTZ RS, AND TEN EICK RE. Inhibiton of
electrogenic Na⫹-pumping in isolated atrial tissue from patients
treated with digoxin. J Pharmacol Exp Ther 252: 60 – 64, 1990.
AND
1825
1826
172.
173.
174.
175.
176.
178.
179.
tion by the Na,K pump. II. Ion binding and release at the extracellular face. J Membr Biol 121: 163–176, 1991.
SU Z, ZOU A, NONAKA A, ZUBAIR I, SANGUINETTI MC, AND BARRY WH.
Influence of prior Na⫹ pump activity on pump and Na⫹/Ca2⫹
exchange currents in mouse ventricular myocytes. Am J Physiol
Heart Circ Physiol 275: H1808 –H1817, 1998.
SUI JL, CHAN KW, AND LOGOTHETIS DE. Na⫹ activation of the muscarinic K⫹ channel by a G-protein-independent mechanism. J Gen
Physiol 108: 381–391, 1996.
SWEADNER KJ. Isozymes of the Na⫹/K⫹-ATPase. Biochim Biophys
Acta 988: 185–200, 1989.
SWEADNER KJ, HERRERA VLM, AMATO S, MOELLMANN A, GIBBONS DK,
⫹ ⫹
AND REPKE KRH. Immunologic identification of Na ,K -ATPase
isoforms in myocardium-isoform change in deoxycorticosterone
acetate-salt hypertension. Circ Res 74: 669 – 678, 1994.
THOMAS RC. Electrogenic sodium pump in nerve and muscle cells.
Physiol Rev 52: 563–594, 1972.
TROTIER D AND DØVING KB. Direct influence of the sodium pump on
the membrane potential of vomeronasal chemoreceptor neurones
in frog. J Physiol (Lond) 490: 611– 621, 1996.
TSUCHIYA K, HORIE M, HARUNA T, AI T, NISHIMOTO T, FUJIWARA H, AND
SASAYAMA S. Functional communication between cardiac ATP-sensitive K⫹ channel and Na/K ATPase. J Cardiovasc Electrophysiol 9:
415– 422, 1998.
VASILETS LA AND SCHWARZ W. Regulation of endogenous and expressed Na⫹/K⫹ pumps in Xenopus oocytes by membrane potential
and stimulation of protein kinases. J Membr Biol 125: 119 –132,
1992.
Physiol Rev • VOL
180. VASILETS LA AND SCHWARZ W. Structure-function relationships of
cation binding in the Na⫹/K⫹-ATPase. Biochim Biophys Acta 1154:
201–222, 1993.
181. VASSALLE M. The role of the electrogenic sodium pump in controlling excitability in nerve and cardiac fibers. In: Current Topics in
Membranes and Transport, edited by Slayman CL. New York:
Academic, 1982, vol. 16, p. 467– 483.
182. VERDONCK F, VANHAECKE J, MUBAGWA K, FLAMENG W, STANKOVICOVA T,
⫹
⫹
AND SIPIDO K. Sensitivity of the Na /K pump current to cardiac
glycosides in ventricular myocytes from human failing heart (Abstract). Biophys J 74: A354, 1998.
183. WANG Y, GAO J, MATHIAS RT, COHEN IS, SUN X, AND BALDO GJ.
␣-Adrenergic effects on Na⫹-K⫹ pump current in guinea-pig ventricular myocytes. J Physiol (Lond) 509: 117–128, 1998.
184. WHALLEY DW, HOOL LC, TEN EICK RE, AND RASMUSSEN HH. Effect of
osmotic swelling and shrinkage on Na⫹-K⫹ pump activity in mammalian cardiac myocytes. Am J Physiol Cell Physiol 265: C1201–
C1210, 1993.
185. WILLIAMSON AP, KENNEDY RH, SEIFEN E, LINDEMANN JP, AND STIMERS
JR. ␣1b-Adrenoceptor-mediated stimulation of Na-K pump current
in adult rat ventricular myocytes. Am J Physiol Heart Circ Physiol
264: H1315–H1318, 1993.
186. WOO AL, JAMES PF, AND LINGREL JB. Characterization of the fourth
␣ isoform of the Na,K-ATPase. J Membr Biol 169: 39 – 44, 1999.
187. ZILLIKENS S. Die Wirkung von Digitoxigenin und DigitoxigeninDerivaten auf die Na⫹/K⫹-ATPase des Warmblüterherzens (PhD
thesis). Bochum, Germany: Ruhr-Universität, 1999, p. 1–141.
81 • OCTOBER 2001 •
www.prv.org
Downloaded from http://physrev.physiology.org/ by 10.220.32.247 on July 4, 2017
177.
HELFRIED GÜNTHER GLITSCH