Charge translocation by the NaK-ATPase from shark rectal gland was measured by adsorption of proteoliposomes to a planar lipid membrane. The proteoliposomes were prepared by reconstitution of purified NaK-ATPase into liposomes consisting of E. coli lipids. The protein was activated by applying an ATP concentration jump produced by photolysis of a protected derivative of ATP, caged ATP. K+ titrations were used to study the effect of K+ on the charge translocation kinetics of the protein. The time-dependent currents obtained after activation of the enzyme with caged ATP were analyzed with a simplified Albers-Post model (E1 (k1)-->E1ATP (k2)-->E2P (k3)-->E1) taking into account the capacitive coupling of the protein to the measuring system. The results of the K+ titrations show a strong dependence of the rate constant k3 on the K+ concentration at the extracellular side of the protein, indicating the K+ activated dephosphorylation reaction. In contrast, k1 and k2 remained constant. The K+ dependence of the rate k3 could be well described with a K+ binding model with two equivalent binding sites (E2P + 2K+ <==> E2P(K) + K+ <==> E2 P(2K)) followed by a rate limiting reaction (E2P(2K) --> E1(2K)). The half saturating K+ concentration K3,0.5 and the microscopic dissociation constant K3 for the K+ dependence of k3 were 4.5mM and 1.9mM respectively. At saturating K+ concentration the rate constant k3 was approximately 100 s(-1). The relative amount of net charge transported during the Na+ and the K+ dependent reactions could be determined from the experiments. Our results suggest electroneutral K+ translocation and do not support electrogenic K+ binding in an extracellular access channel. This is compatible with a model where 2 negative charges are cotransported with 3Na+ and 2K+ ions. Error analysis gives an upper limit of 20% charge transported during K+ translocation or during electrogenic K+ binding in a presumptive access channel compared to Na+ translocation.