During each complete reaction cycle, the Na/K pump transports three Na ions out across the cell membrane and two K ions in. The resulting net extrusion of positive charge generates outward membrane current but, until now, it was unclear how that net charge movement occurs. Reasonable possibilities included a single positive charge moving outwards during Na translocation; or a single negative charge moving inwards during K translocation; or either positive or negative charges moving during both translocation steps, but in unequal quantities. Any step that involves net charge movement through the membrane must have voltage-dependent transition rates. Here we report measurements of transient, voltage-dependent, displacement currents generated by the pump when its normal Na/K transport cycle has been interrupted by removal of external K and it is thus constrained to carry out Na/Na exchange. The quantity and voltage sensitivity of the charge moved during these transient currents suggests that Na translocation includes a voltage-dependent transition involving movement of one positive charge across the membrane. This single step can thus fully account for the electrogenic nature of Na/K exchange. The result provides important new insight into the molecular mechanism of active cation transport.

Na/K pump current was determined between -140 and +60 mV as steady-state, strophanthidin-sensitive, whole-cell current in guinea pig ventricular myocytes, voltage-clamped and internally dialyzed via wide-tipped pipettes. Solutions were designed to minimize all other components of membrane current. A device for exchanging the solution inside the pipette permitted investigation of Na/K pump current-voltage (I-V) relationships at several levels of pipette [Na] .5 ---0.2 mM, nH -------1.1. The voltage-independent activation of Na/K pump current by both intracellular Na ions and extracellular K ions, at zero [Na]o, suggests that neither ion binds within the membrane field. Extracellular Na ions, however, seem to have both a voltage-dependent and a voltage-independent influence on the Na/K pump: they inhibit outward Na/K pump current in a strongly voltage-dependent fashion, with higher apparent affinity at more negative potentials (K0.5 == 90 mM at -120 mV, and -170 mM at -80 mV), and they compete with extracellular K ions in a seemingly voltage-independent manner. Possibly, Na ions are released from the Na/K

Whole-cell currents were recorded in guinea pig ventricular myocytes at ~36~ before, during, and after exposure to maximally effective concentrations of strophanthidin, a cardiotonic steroid and specific inhibitor of the Na/K pump. Wide-tipped pipettes, in combination with a device for exchanging the solution inside the pipette, afforded reasonable control of the ionic composition of the intraceUular solution and of the membrane potential. Internal and external solutions were designed to minimize channel currents and Na/Ca exchange current while sustaining vigorous forward Na/K transport, monitored as strophanthidinsensitive current. 100-ms voltage pulses from the -40 mV holding potential were used to determine steady-state levels of membrane current between -140 and + 60 mV. Control experiments demonstrated that if the Na/K pump cycle were first arrested, e.g., by withdrawal of external K, or of both internal and external Na, then neither strophanthidin nor its vehicle, dimethylsulfoxide, had any discernible effect on steady-state membrane current. Further controls showed that, with the Na/K pump inhibited by strophanthidin, membrane current was insensitive to changes of external [K] between 5.4 and 0 mM and was little altered by changing the pipette [Na] from 0 to 50 mM. Strophanthidin-sensitive current therefore closely approximated Na/K pump current, and was virtually free of contamination by current components altered by the changes in extracellular [K] and intracellular [Na] expected to accompany pump inhibition. The steady-state Na/K pump current-voltage (I-V) relationship, with the pump strongly activated by 5.4 mM external K and 50 mM internal Na (and 10 mM ATP), was sigmoid in shape with a steep positive slope between about 0 and -100 mV, a less steep slope at more negative potentials, and an extremely shallow slope at positive potentials; no region of negative slope was found. That shape of I-V relationship can be generated by a two-state cycle with one pair of voltage-sensitive rate constants and one pair of voltage-insensitive rate constants: such a two-state scheme is a valid steadystate representation of a multi-state cycle that includes only a single voltage-sensitive step.

The voltage dependence of steady and transient changes in Na +/K+ pump current, in response to step changes in membrane potential, was investigated in guinea pig ventricular myocytes voltage clamped and internally dialyzed under experimental conditions designed to support four separate modes of Na+/K+ pump activity. Voltage MATERIALS AND METHODS Hearts were rapidly excised from guinea pigs (300-500 g) fully anesthetized with sodium pentobarbital (-50 mg/kg, i.p.), their aortas were cannulated, and retrograde perfusion of the coronary arteries was begun, first with normal Tyrode's solution (145 mM NaCI/5.4 mM KCI/1.8 mM CaC12/0.5 mM MgCl2/5.5 mM dextrose/5 mM Hepes' NaOH, pH 7.4), then with nominally Ca2"-free Tyrode's solution for 3 min, and then for -15 min with the latter solution containing collagenase (type I, Sigma) at 0.5 mg/ml. The collagenase was washed out of the partially digested heart, which was cut open and kept in a high K+/low Ca' medium (17). Cells harvested from fragments of myocardium were allowed to settle onto the glass coverslip forming the bottom of the experimental chamber on the movable stage of an inverted microscope (Nikon Diaphot), before beginning superfusion with normal Tyrode's solution at 36°C. Giga-ohm seals were obtained with wide-tipped (z5 ,um), fire-polished, pipettes (resistance, -1 MI) filled with normal Tyrode's solution that was exchanged (15) just before rupture of the cell membrane for pipette solution (50 mM NaOH/-85 mM CsOH/85 mM aspartic acid/5 mM pyruvic acid/2 mM MgCl2/10 mM MgATP/5 mM Tris2 creatine phosphate/20 mM tetraethylammonium chloride/5.5 mM dextrose/10 mM EGTA/10 mM Hepes, pH 7.4). The voltage-clamped cell was then superfused with modified Ca2`-free Tyrode's solution (150 mM NaCl/2.3 mM MgC12/2 mM BaCl2/0.5 mM CdCl2/5.5 mM dextrose/5 mM Hepes-NaOH, pH 7.4). Variations of these intracellular and extracellular solutions, as specified in text and legends, were used to sustain four modes of Na+/K+ pump activity (e.g., ref. 18), namely Na+/Na+ exchange, forward or backward Na+/K+ exchange, or K + /K + exchange. The osmolarity of all solutions was -300 mosM, and all solutions were designed to minimize ionchannel currents (3, 13) and Na+/Ca2+ exchange current (19). Na+/K+ pump current was determined as the difference between whole-cell current recorded in the absence and in the presence of 0.5-2 mM strophanthidin. Strophanthidin was added from a 0.5 M solution in dimethyl sulfoxide; control measurements showed that up to 0.5% (by volume) dimethyl sulfoxide had no effect on membrane currents (see Fig. 3C).Current and voltage signals were low-pass filtered at 2 kHz (6-pole Bessel filter), digitized (12-bit resolution) on line at 8 kHz, and stored in a computer for later analysis. Up to 55% of the series resistance between pipette interior and cell membrane was compensated by summing a fraction of the clamp output to the command potential. The