Regulation of endogenous and expressed Na+/K+ pumps in Xenopus oocytes by membrane potential and stimulation of protein kinases

Modulation of the current generated by the Na+/K+ pump by membrane potential and protein kinases was investigated in oocytes of Xenopus laevis. In addition to a positive slope region in the current-voltage (I-V) relationship of the Na+/K+ pump, a negative slope region has been described in these cells (Lafaire & Schwarz, 1986) and has been attributed to a voltage-dependent apparent Km value for pump stimulation by external [K+] (Rakowski et al., 1991). To study this feature in more detail, Xenopus oocytes were used for comparative analysis of the negative slope of the I-V relationship of the endogenous Na+/K+ pump and of the Na+/K+ pump of the electric organ of Torpedo californica expressed in the oocytes. The effects of stimulation of protein kinases A and C on the negative slope were also analyzed. To investigate the negative slope over a wide potential range, experiments were performed in Na(+)-free solution and in the presence of high concentrations of Ba2+ and tetraethylammonium, to block all nonpump related K(+)-sensitive currents. Pump currents and pump-mediated fluxes were determined as differences of currents or fluxes in solutions with and without extracellular K+. The voltage dependence of the Km value for stimulation of the Na+/K+ pump by external [K+] shows significant species differences. Over the entire voltage range from -140 to +20 mV, the Km value for the Na+/K+ pump of Torpedo electroplax is substantially higher than for the endogenous pump and exhibits more pronounced voltage dependence. For the Xenopus pump, the voltage dependence can be described by voltage-dependent stimulation by external [K+] and can be interpreted by voltage-dependent K+ binding, assuming that an effective charge between 0.37 and 0.56 of an elementary charge is moved in the electrical field. An analogous evaluation of the voltage dependence of the Torpedo pump requires the assumption of movement of two effective charges of 0.16 and 1.0 of an elementary charge. Application of 1,2-dioctanoyl-sn-glycerol (diC8, 10-50 microM), which is known to stimulate protein kinase C, reduces the maximum activity of the Xenopus pumps in the oocyte membrane by 40% and modulates the voltage dependence of K+ stimulation. For the endogenous Xenopus pump, the apparent effective charge increased from 0.37 to 0.51 of elementary charge and the apparent Km at 0 mV increased from 0.46 to 0.83 mM. For the Torpedo pump, one of the apparent effective charges increased from 1.0 to 2.5 of elementary charge.(ABSTRACT TRUNCATED AT 400 WORDS)     

Electrophysiological characterization of a TTX-sensitive sodium current in native Xenopus oocytes.

We have studied a fast inward current expressed in oocytes from one Xenopus laevis. This current was characterized as a sodium current. It was activated by depolarizations to -50 mV or higher, peaked within 3-5 ms, and then decayed following a mono-exponential timecourse. When clamped at different holding potentials, the current displayed voltage-dependent inactivation with a V0.5 of -51 mV. The channel responsible for this Na+ entry was blocked by tetrodotoxin with a K0.5 of 8 nM, and was resistant to block by lidocaine at concentrations up to 100 microM. The pharmacological similarities between neuronal and oocyte sodium channels suggest that the two channels share a conserved structure.     

Properties of voltage-gated Ca2+ currents measured from mouse pancreatic beta-cells in situ

We used the single-microelectrode voltage-clamp technique to record ionic currents from pancreatic beta-cells within intact mouse islets of Langerhans at 37 degrees C, the typical preparation for studies of glucose-induced "bursting" electrical activity. Cells were impaled with intracellular microelectrodes, and voltage pulses were applied in the presence of tetraethylammonium. Under these conditions, a voltage-dependent Ca2+ current (I(Cav)), containing L-type and non-L-type components, was observed. The current measured in situ was larger than that measured in single cells with whole-cell patch clamping, particularly at membrane potentials corresponding to the action potentials of beta-cell electrical activity. The temperature dependence of I(Cav) was not sufficient to account for the difference in size of the currents recorded with the two methods. During prolonged pulses, the voltage-dependent Ca2+ current measured in situ displayed both rapid and slow components of inactivation. The rapid component was Ca2+-dependent and was inhibited by the membrane-permeable Ca2+ chelator, BAPTA-AM. The effect of BAPTA-AM on beta-cell electrical activity then demonstrated that Ca2+-dependent inactivation of I(Cav) contributes to action potential repolarization and to control of burst frequency. Our results demonstrate the utility of voltage clamping beta-cells in situ for determining the roles of ion channels in electrical activity and insulin secretion.     

Slow and incomplete inactivations of voltage-gated channels dominate encoding in synthetic neurons.

Electrically excitable channels were expressed in Chinese hamster ovary cells using a vaccinia virus vector system. In cells expressing rat brain IIA Na+ channels only, brief pulses (< 1 ms) of depolarizing current resulted in action potentials with a prolonged (0.5-3 s) depolarizing plateau; this plateau was caused by slow and incomplete Na+ channel inactivation. In cells expressing both Na+ and Drosophila Shaker H4 transient K+ channels, there were neuron-like action potentials. In cells with appropriate Na+/K+ current ratios, maintaining stimulation produced repetitive firing over a 10-fold range of frequencies but eventually led to "lock-up" of the potential at a positive value after several seconds of stimulation. The latter effect was due primarily to slow inactivation of the K+ currents. Numerical simulations of modified Hodgkin-Huxley equations describing these currents, using parameters from voltage-clamp kinetics studied in the same cells, accounted for most features of the voltage trajectories. The present study shows that insights into the mechanisms for generating action potentials and trains of action potentials in real excitable cells can be obtained from the analysis of synthetic excitable cells that express a controlled repertoire of ion channels.     

Voltage-dependent stimulation of Na+/K(+)-pump current by external cations: selectivity of different K+ congeners.

Currents generated by the endogenous Na+/K+ pump in the oocytes of Xenopus laevis were determined under voltage-clamp as currents activated by different K+ congeners. The voltage dependence of the pump current reflects voltage-dependent steps in the reaction cycle. The decrease of K(+)-activated pump current at positive potentials has been attributed to voltage-dependent stimulation by the external K+ (Rakowski, Vasilets, LaTona and Schwarz (1991) J. Membr. Biol. 121, 177-187). In Na(+)-free solution, activation of the pump by external cations seems to be the dominating voltage-dependent and rate-determining step in the reaction cycle. Under these conditions, the voltage dependence of apparent Km values for pump activation can be analyzed. The dependence suggests voltage-dependent binding of extracellular cations assuming that an effective charge of about 0.4 of an elementary charge is moved in the electrical field during a step associated with the cation binding. The apparent Km values at 0 mV differ for various cations that stimulate pump activity. The values are in mM: 0.10 for Tl+, 0.63 for K+, 0.71 for Rb+, 9.3 for NH4+, and 12.9 for Cs+. The corresponding apparent affinities follow the same sequence as the cation permeability of the K(+)-selective delayed rectifier channel of nerve cells. The results are compatible with the interpretation that the cations have to pass an ion-selective access channel to reach their binding sites in the pump molecule.     

Modification of K channel inactivation by papain and N-bromoacetamide.

The whole-cell variation of the patch clamp technique was used to study macroscopic K current in voltage clamped GH3 cells. An inactivating, voltage-dependent K current was studied in isolation by inhibiting Ca-activated K currents with internal Ca chelators and external tetraethylammonium ions. Under control conditions, the K current inactivated in two phases with time constants of 25 and 79 ms. After treatment with either a proteolytic enzyme such as papain or the amino acid reagent N-bromoacetamide, the K current no longer inactivated rapidly, but decayed very slowly with a time constant of 500 to 750 ms. The action of papain or N-bromoacetamide on K channels is comparable to their action on Na channels, suggesting that inactivation in Na and K channels occurs by a similar mechanism.     

A slowly inactivating potassium current in native oocytes of Xenopus laevis

Membrane currents were recorded in voltage-clamped oocytes of Xenopus laevis in response to voltage steps. We describe results obtained in oocytes obtained from one donor frog, which showed an unusually large outward current upon depolarization. Measurements of reversal potentials of tail currents in solutions of different K+ concentration indicated that this current is carried largely by K+ ions. It was strongly reduced by extracellular application of tetraethylammonium, though not by Ba2+ or 4-aminopyridine. Removal of surrounding follicular cells did not reduce the K+ current, indicating that it arises across the oocyte membrane proper. Activation of the K+ conductance was first detected with depolarization to about -12 mV, increased with a limiting voltage sensitivity of 3 mV for an e-fold change in current, and was half-maximally activated at about +10 mV. The current rose following a single exponential timecourse after depolarization, with a time constant that shortened from about 400 ms at -10 mV to about 15 ms at +80 mV. During prolonged depolarization the current inactivated with a time constant of about 4 s, which did not alter greatly with potential. The K+ current was independent of Ca2+, as it was not altered by addition of 10 mM Mn2+ to the bathing medium, or by intracellular injection of EGTA. Noise analysis of K+ current fluctuations indicated that the current is carried by channels with a unitary conductance of about 20 ps and a mean open lifetime of about 300 ms (at room temperature and potential of +10 to +20 mV).     

Two types of potassium channels in murine T lymphocytes

The properties of two types of K+ channels in murine T lymphocytes are described on the basis of whole-cell and isolated-patch recordings using the gigohm-seal technique. Type l (standing for "lpr gene locus" or "large") channels were characterized mainly in T cells from mutant MRL/MpJ-lpr/lpr mice, in which they are present in large numbers. Type n ("normal") K+ channels are abundant and therefore most readily studied in concanavalin A-activated T cells from four strains of mice, MRL-+/+, CBA/J, C57BL/6J, and BALB/c. Type l channels, compared with type n, are activated at potentials approximately 30 mV more positive, and close much more rapidly upon repolarization. Type l channels inactivate more slowly and less completely than type n during maintained depolarization, but recover from inactivation more rapidly, so that little inactivation accumulates during repetitive pulses. Type l channels have a higher unitary conductance (21 pS) than type n (12 pS) and are less sensitive to block by external Co++, but are 100-fold more sensitive to block by external tetraethylammonium (TEA), with half-block of type l channels at 50-100 microM TEA compared with 8-16 mM for type n. TEA blocks both types of channels by reducing the apparent single channel current amplitude, with a dose-response relation similar to that for blocking macroscopic currents. Murine type n K+ channels resemble K+ channels in human T cells.     

Ca(2+)-dependent inactivation of a cloned cardiac Ca2+ channel alpha 1 subunit (alpha 1C) expressed in Xenopus oocytes.

The alpha 1 subunit of cardiac Ca2+ channel, expressed alone or coexpressed with the corresponding beta subunit in Xenopus laevis oocytes, elicits rapidly inactivating Ca2+ currents. The inactivation has the following properties: 1) It is practically absent in external Ba2+; 2) it increases with Ca2+ current amplitudes; 3) it is faster at more negative potentials for comparable Ca2+ current amplitudes; 4) it is independent of channel density; and 5) it does not require the beta subunit. These findings indicate that the Ca2+ binding site responsible for inactivation is encoded in the alpha 1 subunit and suggest that it is located near the inner channel mouth but outside the membrane electric field.     

Position-dependent attenuation by Kv1.6 of N-type inactivation of Kv1.4-containing channels

Assembly of distinct α subunits of Kv1 (voltage-gated K(+) channels) into tetramers underlies the diversity of their outward currents in neurons. Kv1.4-containing channels normally exhibit N-type rapid inactivation, mediated through an NIB (N-terminal inactivation ball); this can be over-ridden if associated with a Kv1.6 α subunit, via its NIP (N-type inactivation prevention) domain. Herein, NIP function was shown to require positioning of Kv1.6 adjacent to the Kv1.4 subunit. Using a recently devised gene concatenation, heterotetrameric Kv1 channels were expressed as single-chain proteins on the plasmalemma of HEK (human embryonic kidney)-293 cells, so their constituents could be arranged in different positions. Placing the Kv1.4 and 1.6 genes together, followed by two copies of Kv1.2, yielded a K(+) current devoid of fast inactivation. Mutation of critical glutamates within the NIP endowed rapid inactivation. Moreover, separating Kv1.4 and 1.6 with a copy of Kv1.2 gave a fast-inactivating K(+) current with steady-state inactivation shifted to more negative potentials and exhibiting slower recovery, correlating with similar inactivation kinetics seen for Kv1.4-(1.2)(3). Alternatively, separating Kv1.4 and 1.6 with two copies of Kv1.2 yielded slow-inactivating currents, because in this concatamer Kv1.4 and 1.6 should be together. These findings also confirm that the gene concatenation can generate K(+) channels with α subunits in pre-determined positions.     

Low molecular weight poly(A)+ mRNA species encode factors that modulate gating of a non-Shaker A-type K+ channel

Voltage-dependent K+ channels control repolarization of action potentials and help establish firing patterns in nerve cells. To determine the nature and role of molecular components that modulate K+ channel function in vivo, we coinjected Xenopus oocytes with cRNA encoding a cloned subthreshold A-type K+ channel (mShal1, also referred to as mKv4.1) and a low molecular weight (LMW) fraction (2-4 kb) of poly(A)+ mRNA (both from rodent brain). Coinjected oocytes exhibited a significant (fourfold) increase in the surface expression of mShal1 K+ channels with no change in the open-channel conductance. Coexpression also modified the gating kinetics of mShal1 current in several respects. Macroscopic inactivation of whole oocyte currents was fitted with the sum of two exponential components. Both fast and slow time constants of inactivation were accelerated at all membrane potentials in coinjected oocytes (tau f = 47.2 ms vs 56.5 ms at 0 mV and tau s = 157 ms vs 225 ms at 0 mV), and the corresponding ratios of amplitude terms were shifted toward domination by the fast component (Af/As = 2.71 vs 1.17 at 0 mV). Macroscopic activation was characterized in terms of the time-to-peak current, and it was found to be more rapid at all membrane potentials in coinjected oocytes (9.9 ms vs 13.5 ms at 0 mV). Coexpression also leads to more rapid recovery from inactivation (approximately 2.4-fold faster at -100 mV). The coexpressed K+ currents in oocytes resemble currents expressed in mouse fibroblasts (NIH3T3) transfected only with mShal1 cDNA. These results indicate that mammalian regulatory subunits or enzymes encoded by LMW mRNA species, which are apparently missing or expressed at low levels in Xenopus oocytes, may modulate gating in some native subthreshold A-type K+ channels.     

External pore collapse as an inactivation mechanism for Kv4.3 K+ channels

Kv4 channels are thought to lack a C-type inactivation mechanism (collapse of the external pore) and to inactivate as a result of a concerted action of cytoplasmic regions of the channel. To investigate whether Kv4 channels have outer pore conformational changes during the inactivation process, the inactivation properties of Kv4.3 were characterized in 0 mM and in 2 mM external K+ in whole-cell voltage-clamp experiments. Removal of external K+ increased the inactivation rates and favored cumulative inactivation by repetitive stimulation. The reduction in current amplitude during repetitive stimulation and the faster inactivation rates in 0 mM external K+ were not due to changes in the voltage dependence of channel opening or to internal K+ depletion. The extent of the collapse of the K+ conductance upon removal of external K+ was more pronounced in NMG+-than in Na+-containing solutions. The reduction in the current amplitude during cumulative inactivation by repetitive stimulation is not associated with kinetic changes, suggesting that it is due to a diminished number of functional channels with unchanged gating properties. These observations meet the criteria for a typical C-type inactivation, as removal of external K+ destabilizes the conducting state, leading to the collapse of the pore. A tentative model is presented, in which K+ bound to high-affinity K+-binding sites in the selectivity filter destabilizes an outer neighboring K+ modulatory site that is saturated at approximately 2 mM external K+. We conclude that Kv4 channels have a C-type inactivation mechanism and that previously reported alterations in the inactivation rates after N- and C- termini mutagenesis may arise from secondary changes in the electrostatic interactions between K+-binding sites in the selectivity filter and the neighboring K+-modulatory site, that would result in changes in its K+ occupancy.     

Influence of some phospholipase A2 and cytochrome P450 inhibitors on rat arterial smooth muscle K+ currents.

The hyperpolarizing factor that is liberated by vascular endothelial cells in response to various agonists, and known to induce relaxation by opening of smooth muscle K+ channels, has been suggested to be a product of cytochrome P450 dependent arachidonic acid metabolism. In this study, the direct influence of two phospholipase A2 inhibitors and of five structurally and mechanistically different cytochrome P450 inhibitors on K+ currents in freshly isolated vascular smooth muscle cells from the rat aorta was investigated. On stepping the cell membrane potential from -70 mV to a series of depolarized test potentials, a noisy outward current developed at test potentials > +10 mV, which showed no appreciable inactivation during the voltage pulse. It was largely abolished by 3 mM external tetraethylammonium chloride (TEA), suggesting that it predominantly consisted of current through large-conductance Ca(2+)-activated K+ channels. The phospholipase A2 inhibitor quinacrine considerably inhibited this TEA-sensitive current, while 4-bromophenacylbromide exerted no effect. The cytochrome P450 inhibitors proadifen and miconazole reversibly decreased the amplitude of I(K), while clotrimazole and 1-aminobenzotriazole had no effect. Conversely, 17-octadecynoic acid increased whole-cell I(K). These results show that some phospholipase A2 and cytochrome P450 inhibitors may interfere with K+ channel activation in the rat arterial smooth muscle cell. The relevance of these findings to studies on the involvement of cytochrome P450 dependent metabolism in the generation of the endothelium-derived hyperpolarizing factor in intact arteries is discussed.     

Characterization of two distinct depolarization-activated K+ currents in isolated adult rat ventricular myocytes

Depolarization-activated outward K+ currents in isolated adult rat ventricular myocytes were characterized using the whole-cell variation of the patch-clamp recording technique. During brief depolarizations to potentials positive to -40 mV, Ca(2+)-independent outward K+ currents in these cells rise to a transient peak, followed by a slower decay to an apparent plateau. The analyses completed here reveal that the observed outward current waveforms result from the activation of two kinetically distinct voltage-dependent K+ currents: one that activates and inactivates rapidly, and one that activates and inactivates slowly, on membrane depolarization. These currents are referred to here as Ito (transient outward) and IK (delayed rectifier), respectively, because their properties are similar (although not identical) to these K+ current types in other cells. Although the voltage dependences of Ito and IK activation are similar, Ito activates approximately 10-fold and inactivates approximately 30-fold more rapidly than IK at all test potentials. In the composite current waveforms measured during brief depolarizations, therefore, the peak current predominantly reflects Ito, whereas IK is the primary determinant of the plateau. There are also marked differences in the voltage dependences of steady-state inactivation of these two K+ currents: IK undergoes steady-state inactivation at all potentials positive to -120 mV, and is 50% inactivated at -69 mV; Ito, in contrast, is insensitive to steady-state inactivation at membrane potentials negative to -50 mV. In addition, Ito recovers from steady-state inactivation faster than IK: at -90 mV, for example, approximately 70% recovery from the inactivation produced at -20 mV is observed within 20 ms for Ito; IK recovers approximately 25-fold more slowly. The pharmacological properties of Ito and IK are also distinct: 4-aminopyridine preferentially attenuates Ito, and tetraethylammonium suppresses predominantly IK. The voltage- and time-dependent properties of these currents are interpreted here in terms of a model in which Ito underlies the initial, rapid repolarization phase of the action potential (AP), and IK is responsible for the slower phase of AP repolarization back to the resting membrane potential, in adult rat ventricular myocytes.     

A voltage-dependent role for K+ in recovery from C-type inactivation.

Recovery from C-type inactivation of Kv1.3 can be accelerated by the binding of extracellular potassium to the channel in a voltage-dependent fashion. Whole-cell patch-clamp recordings of human T lymphocytes show that Ko+ can bind to open or inactivated channels. Recovery is biphasic with time constants that depend on the holding potential. Recovery is also dependent on the voltage of the depolarizing pulse that induces the inactivation, consistent with a modulatory binding site for K+ located at an effective membrane electrical field distance of 30%. This K(+)-enhanced recovery can be further potentiated by the binding of extracellular tetraethylammonium to the inactivated channel, although the tetraethylammonium does not interact directly with the K(+)-binding site. Our findings are consistent with a model in which K+ can bind and unbind slowly from a channel in the inactivated state, and inactivated channels that are bound by K+ will recover with a rate that is fast relative to unbound channels. Our data suggest that the kinetics of K+ binding to the modulatory site are slower than these recovery rates, especially at hyperpolarized voltages.     

Fast inactivation causes rectification of the IKr channel

The mechanism of rectification of HERG, the human cardiac delayed rectifier K+ channel, was studied after heterologous expression in Xenopus oocytes. Currents were measured using two-microelectrode and macropatch voltage clamp techniques. The fully activated current- voltage (I-V) relationship for HERG inwardly rectified. Rectification was not altered by exposing the cytoplasmic side of a macropatch to a divalent-free solution, indicating this property was not caused by voltage-dependent block of outward current by Mg2+ or other soluble cytosolic molecules. The instantaneous I-V relationship for HERG was linear after removal of fast inactivation by a brief hyperpolarization. The time constants for the onset of and recovery from inactivation were a bell-shaped function of membrane potential. The time constants of inactivation varied from 1.8 ms at +50 mV to 16 ms at -20 mV; recovery from inactivation varied from 4.7 ms at -120 mV to 15 ms at -50 mV. Truncation of the NH2-terminal region of HERG shifted the voltage dependence of activation and inactivation by +20 to +30 mV. In addition, the rate of deactivation of the truncated channel was much faster than wild-type HERG. The mechanism of HERG rectification is voltage-gated fast inactivation. Inactivation of channels proceeds at a much faster rate than activation, such that no outward current is observed upon depolarization to very high membrane potentials. Fast inactivation of HERG and the resulting rectification are partly responsible for the prolonged plateau phase typical of ventricular action potentials.     

Effects of gadolinium on ion channels in the myelinated axon of Xenopus laevis: four sites of action.

The action of gadolinium (Gd3+) on ion currents in myelinated axons of Xenopus laevis was investigated with the voltage clamp technique. The analysis revealed the following effects. (i) The potential-dependent parameters of both Na and K channels were shifted. The shift was equally large for activation, inactivation, and activation time constant curves (+9 mV for 100 microM Gd3+). The effects could be explained by screening of fixed surface charges at a density of -1.2 e nm-2. (ii) The rate of gating for both Na and K channels was reduced more than predicted from the shift. This effect could be quantified as a scaling (by a factor 3 and 5 respectively at 100 microM Gd3+) of the activation time constant curves. (iii) An activation- and inactivation-independent block of both Na and K channels, obeying 1:1 stoichiometry with a Kd value of about 70 microM potential-independent block of leakage current, obeying 1:2 stoichiometry with a Kd value of 600 microM. (iv) The analysis suggests separate binding sites for the effects, comprising high affinity modulatory and blocking sites on the channel proteins and low affinity receptors on phospholipids, associated with the effect on the leakage current.     

The potassium channel MBK1 (Kv1.1) is expressed in the mouse retina

1. The neurons of the retina have electrical properties that are different from those of most of the other neurons of the central nervous system. To identify the voltage-gated ion channels found in the retina, we screened mouse retinal cDNA libraries with oligonucleotide probes homologous to the mammalian K+ channel MBK1 (Kv1.1) and ligated two partial clones to produce a full-length clone with no significant differences from MBK1. 2. Expression of MBK1 mRNA was determined by RNAse protection. MBK1 mRNA was detected in retinal RNA and was also detected in brain, liver, and heart RNAs. 3. We transcribed the full-length clone, injected it into oocytes of Xenopus laevis, and measured the membrane currents 2 to 6 days later. Depolarization from a holding voltage of -90mV induced a slowly activated outward current with a peak value as large as 20 microA. The current inactivated very slowly with a single exponential time course [mean time constant, 6.5 +/- 0.4 sec (SEM) for activation voltage of -10mV]. 4. The outward current was reduced to half-maximal by 0.42 mM tetraethylammonium, 1.1 mM 4-aminopyridine, and 3.2 mM Ba2+ but was not significantly attenuated by Co2+ (1 mM). 5. The reversal potential (measured with tail currents) changed by 53mV per decade change of [K+] from 1 to 77 mM. 6. The voltage for half-maximal activation of the conductance was -26.6mV (+/- 1.7mV), and the voltage required for an e-fold increase in conductance was 6.9mV (+/- 0.5mV). 7. Thus, the mRNA for MBK1 found in the mouse retina causes the expression of a voltage-dependent K+ current which has properties suitable for may retinal neurons.     

Access channel model for the voltage dependence of the forward-running Na+/K+ pump

The voltage dependence of steady state current produced by the forward mode of operation of the endogenous electrogenic Na+/K+ pump in Na(+)- loaded Xenopus oocytes has been examined using a two-microelectrode voltage clamp technique. Four experimental cases (in a total of 18 different experimental conditions) were explored: variation of external [Na+] ([Na]o) at saturating (10 mM) external [K+] ([K]o), and activation of pump current by various [K]o at 0, 15, and 120 mM [Na]o (tetramethylammonium replacement). Ionic current through K+ channels was blocked by Ba2+ (5 mM) and tetraethylammonium (20 mM), thereby allowing pump-mediated current to be measured by addition or removal of external K+. Control measurements and corrections were made for pump current run-down and holding current drift. Additional controls were done to estimate the magnitude of the inwardly directed pump-mediated current that was present in K(+)-free solution and the residual K(+)- channel current. A pseudo two-state access channel model is described in the Appendix in which only the pseudo first-order rate coefficients for binding of external Na+ and K+ are assumed to be voltage dependent and all transitions between states in the Na+/K+ pump cycle are assumed to be voltage independent. Any three-state or higher order model with only two oppositely directed voltage-dependent rate coefficients can be reduced to an equivalent pseudo two-state model. The steady state current-voltage (I-V) equations derived from the model for each case were simultaneously fit to the I-V data for all four experimental cases and yielded least-squares estimates of the model parameters. The apparent fractional depth of the external access channel for Na+ is 0.486 +/- 0.010; for K+ it is 0.256 +/- 0.009. The Hill coefficient for Na+ is 2.18 +/- 0.06, and the Hill coefficient for K+ (which is dependent on [Na]o) ranges from 0.581 +/- 0.019 to 1.35 +/- 0.034 for 0 and 120 mM [Na]o, respectively. The model provides a reasonable fit to the data and supports the hypothesis that under conditions of saturating internal [Na+], the principal voltage dependence of the Na+/K+ pump cycle is a consequence of the existence of an external high- field access channel in the pump molecule through which Na+ and K+ ions must pass in order to reach their binding sites.