Characterization of the adipocyte ghost (Na+,K+) pump. Insights into the insulin regulation of the adipocyte (Na+,K+) pump.

The (Na+,K+) ATPase in plasma membranes isolated from rat adipocytes is insensitive to insulin (Lytton J., Lin, J.C., and Guidotti, G. (1985) J. Biol. Chem. 260, 1177-1184). For this reason, the characteristics of the (Na+,K+) pump in adipocyte ghosts, prepared by hypotonic lysis of adipocytes (Rodbell, M. (1967) J. Biol. Chem. 242, 5744-5750), were studied. Herein it is demonstrated that the (Na+,K+) pump in ghosts is identical to that described in isolated plasma membranes, sharing the following characteristics: 1) the Ki values for ouabain are 1.3 x 10(-7) M and 4.5 x 10(-5) M for the alpha 2 and alpha 1 isozymes, respectively; 2) the K0.5 values for sodium are 11.4 +/- 1.6 and 7.2 +/- 3.8 mM for the alpha 2 and alpha 1 isozymes, respectively; 3) both forms of the (Na+,K+) pump are insensitive to insulin stimulation, presumably because the activities are already maximal. The ghosts are not in an insulin-stimulated state because the activity of the glucose transporter is not increased as it is in ghosts prepared from insulin-treated cells. In addition, presented evidence demonstrates that ghost internal sodium concentration, [Na+]i, is very sensitive to changes in the activity of the (Na+,K+) pump. If the [Na+]i, of adipocytes is also very sensitive to the activity of the (Na+,K+) pump, the mechanism of insulin stimulation of the adipocyte (Na+,K+) pump requires reexamination.     

Orientation of vesicles isolated from baso-lataral membranes of renal cortex

Baso-lateral membranes were isolated from the canine and porcine kidney cortex by several different methods currently in use. Sidedness of the isolated membrane vesicles was determined by procedures using 1. ouabain-sensitive (Na+K+)ATPase assays in the presence and in the absence of sodium dodecylsulfate or digitoxigenin plus monensin, 2. (Na+, K+, Mg2+)ATPase assays with valinomycin, 3. sialidase accessibility, and 4. binding of hydrophilic and lipophilic cardiac glycosides. The (Na+K+)ATPase activity in the membrane preparation was increased 10-fold of that found in the crude homogenate. Isolated membrane vesicles, prepared by different techniques, were all found to be overwhelmingly of right-side-out orientation;namely, right-side-out = 51-68%, inside-out = 4-13%, and unsealed vesicles = 26-42%. Results of sidedness determinations by different methods showed a good agreement. Thus, predominantly right-side-out oriented vesicles are formed during conventional isolation procedures for membranes of the kidney cortex.     

On the mechanism of shrinkage-induced potassium influx in rat and human erythrocytes.

The rates of 86Rb influx into human and rat erythrocytes were studied in media of various tonicity. At sucrose concentrations below 0.3 mol/l, the ouabain-insensitive, furosemide-inhibited component of influx increased in rat but not in human erythrocytes; this may be explained by a rise in the rate of Na+, K+, Cl-- and/or K+, Cl-cotransport. An increase in osmolarity resulted in a reduction of this as well as of the ouabain and furosemide-insensitive component in rat erythrocytes. At the same conditions a drastic inhibition of Na+, K(+)-pump occurred both in rat and human erythrocytes. We failed to observe a lag-phase in the activation of the cotransport in rat erythrocytes; i. e. the process of activation parallels the shrinkage of cells. In rat erythrocyte ghosts, the shrinkage-induced stimulation of the cotransport was lost, and the direction of their osmotic reaction (inhibition of transport pathways) was similar to that in human erythrocyte ghosts. It is suggested that the mechanism of volume regulation of ion transport in intact cells involves a step of physical amplification via a change in interactions between the protein carcass and the lipid bilayer.     

Erythrocyte membrane ATPase and calcium pumping activities in porcine malignant hyperthermia

To investigate possible abnormalities in erythrocyte membrane enzyme activities in the pharmacogenetic disorder MH, membrane ATPase activities have been examined in erythrocyte ghosts prepared from red blood cells of MHS and normal swine. While no differences were noted in Mg2+-ATPase activities, the (Na+, K+)-ATPase activity of MHS erythrocyte ghosts was less than that of normal ghosts. Ca2+-ATPase activity exhibited low- and high-affinity Ca2+-binding sites in both types of erythrocyte ghost. While the Km for Ca2+ was greater for normal than for MHS erythrocyte ghosts at the high-affinity Ca2+-binding site, the reverse was true at the low-affinity Ca2+-binding site. Irrespective of the type of calcium binding site occupied, the Vmax for normal erythrocyte ghost Ca2+-ATPase activity was greater than that for MHS ghosts. In the presence of calmodulin, there was now no difference between MHS and normal erythrocyte ghosts in either the Km for Ca2+ or the Vmax of the Ca2+-ATPase activity. To determine if the calcium pumping activity of intact MHS and normal pig erythrocytes differed, calcium efflux from the 45Ca-loaded erythrocytes was determined; this activity was significantly greater for MHS than for normal erythrocytes. Thus, the present study confirms that there are abnormalities in the membranes of MHS pig red blood cells. However, we conclude that these abnormalities are unlikely to result in an impaired ability of MHS erythrocytes to regulate their cytosolic Ca2+ concentration.     

Rat brain has the alpha 3 form of the (Na+,K+)ATPase

Multiple forms of the catalytic subunit of the (Na+,K+)ATPase have been identified in rat brain. While two of them (alpha 1 and alpha 2) have been well characterized, the third form (alpha 3) of these catalytic subunits only recently has been described by cDNA cloning; the corresponding polypeptide has not been isolated. In this paper it is shown that rat brain contains the alpha 3 chain. The catalytic subunits of the (Na+, K+)ATPase from rat brain axolemma were purified by SDS-PAGE and subjected to formic acid cleavage. Amino acid sequence analysis of the resulting fragments revealed that axolemma has the alpha 3 form of the catalytic subunit. In addition, alpha 3-specific antiserum was raised in rabbits immunized with a synthetic peptide. Immunoblotting with this antiserum revealed that the alpha 3 form of the (Na+,K+)ATPase is present also in whole brain microsomes. In SDS-PAGE, the mobilities of the three catalytic subunits of brain (Na+, K+)ATPase follow the order alpha 1 greater than alpha 2 greater than alpha 3. Determination of the ouabain-inhibitable ATPase activity indicates that if the alpha 3 form of the (Na+,K+)ATPase is able to hydrolyze ATP, it is present in a form of the enzyme with a high affinity for this cardiac glycoside and is similar to the alpha 2 form in this respect.     

Novel regulation of Na, K-ATPase by Src tyrosine kinases in cortical neurons

The Na+, K+-ATPase or Na+, K+-pump plays a critical role in ion homeostasis and many cellular events. The Na+, K+-pump activity is regulated by serine/threonine phosphorylation, the role of tyrosine kinases in the regulation, however, is obscure. We now present novel evidence showing that tyrosine phosphorylation activates the Na+, K+-pump in cortical neurons. The electrogenic activity of the Na+, K+-pump was measured using whole-cell voltage clamp. A tonic activity was revealed by an inward current induced by the specific inhibitor ouabain or strophanthidin; an outward current due to activation of the pump was triggered by raising extracellular K+. The inward and outward currents were attenuated by the tyrosine kinase inhibitor genistein, herbimycin A, or lavendustin A, while blocking tyrosine phosphatases increased the pump current. Down-regulation of the pump current was also seen with the Src inhibitor PP1 and intracellularly applied anti-Lyn or anti-Yes antibody. Consistently, intracellular application of Lyn kinase up-regulated the pump current. Immunoprecipitation and western blotting showed tyrosine phosphorylation and a direct interaction between Lyn and the alpha3 subunit of the Na+, K+-pump. The tyrosine phosphorylation of the alpha3 subunit was reduced by serum deprivation. These data suggest that the Na+, K+-ATPase activity in central neurons is regulated by specific Src tyrosine kinases via a protein-protein mechanism and may play a role in apoptosis.     

Thallium inhibition of ouabain-sensitive sodium transport and of the (Na+ plus K+)-ATPase in human erythrocytes.

The influence of Tl+ on Na+ transport and on the ATPase activity in human erythrocytes was studied. 0.1-1.0 mM Tl+ added to a K+-free medium inhibited the ouabain-sensitive self-exchange of Na+ and activated both the ouabain-sensitive 22Na outward transport and the transport related ATPase. 5-10mM external Tl+ caused inhibition of the ouabain-sensitive 22Na efflux as well as the (Na+ plus Tl+)-ATPase. Competition between the internal Na+ and rapidly penetrating thallous ions at the inner Na+-specific binding sites of the erythrocyte membrane could account for the inhibitory effect of Tl+. An increase of the internal Na+ concentration in erythrocytes or in ghosts protected the system against the inhibitory effect of high concentration of Tl+. A protective effect of Na+ was also demonstrated on the (Na+ plus Tl+)-ATPase of fragmented erythrocyte membranes studied at various Na+ and Tl+ concentrations.     

The control by Ca2+ of the polyphosphoinositide phosphodiesterase and the Ca2+-pump ATPase in human erythrocytes

1. Both the Ca(2+)-pump ATPase and the polyphosphoinositide phosphodiesterase of the erythrocyte membrane can, when assayed under appropriate conditions, be activated by Ca(2+) in the micromolar range. We have therefore compared the mechanisms and affinities for Ca(2+) activation of the two enzymes in human erythrocyte membranes, to see whether the polyphosphoinositide phosphodiesterase would be active in normal healthy erythrocytes. 2. At physiological ionic strength and in the presence of calmodulin, the Ca(2+)-pump ATPase was activated by Ca(2+) in a highly co-operative manner, with half-maximal activation occurring at about 0.3mum-Ca(2+). At an optimal Ca(2+) concentration, calmodulin stimulated the Ca(2+)-sensitive ATPase activity about 10-fold. 3. Ca(2+) activated the polyphosphoinositide phosphodiesterase in a non-co-operative manner. The Ca(2+) requirements for breakdown of phosphatidylinositol 4-phosphate and phosphatidylinositol 4,5-bisphosphate were identical, which supports our previous conclusion that Ca(2+) activates a single polyphosphoinositide phosphodiesterase that degrades both lipids with equal facility. Added calmodulin did not affect the activity of the polyphosphoinositide phosphodiesterase. 4. At low ionic strength in the absence of Mg(2+), half-maximal activation of the phosphodiesterase was at about 3mum-Ca(2+). The presence of 1mm-Mg(2+) shifted the Ca(2+) activation curve to the right, as did elevation of the ionic strength. When the Ca(2+)-pump ATPase and the polyphosphoinositide phosphodiesterase were assayed in the same incubations and under conditions of intracellular ionic strength and Mg(2+) concentration, the ATPase was fully activated at 3mum-Ca(2+), whereas no polyphosphoinositide phosphodiesterase activity was detected below 100mum-Ca(2+). 5. The Ca(2+)-pump ATPase of the erythrocyte membrane normally maintains the Ca(2+) concentration of healthy erythrocytes below approx. 0.1mum. It therefore seems unlikely that the polyphosphoinositide phosphodiesterase of the erythrocyte membrane ever expresses its activity in a healthy erythrocyte.     

Modifications in the activities of membrane-bound enzymes during in vivo ageing of human and rabbit erythrocytes

Erythrocyte plasma membranes were isolated from a homogeneous population of human or rabbit erythrocytes fractionated into classes representing young, middle-age and old age in vivo. Lipid analyses of human erythrocyte plasma membranes reveal a decrease of the cholesterol to phospholipid molar ratio, followed by a marked decrease in the activities of the membrane-bound enzymes (Na+,K+)-stimulated ATPase, acetylcholinesterase and NAD+ase from young to old age. Such changes were not observed between young and middle-age rabbit erythrocytes. Incubation of rabbit young erythrocytes with phosphatidylcholine vesicles (liposomes) to obtain partial depletion of their membrane cholesterol, indicated that cholesterol depletion causes a statistically significant decrease of the (Na+,K+)-stimulated ATPase and acetylcholinesterase activities, but the NAD+ase activity remained almost unchanged. The biological significance of these data are discussed in terms of the differences and modifications in the interaction of membrane-bound enzymes with membrane lipids during in vivo ageing of erythrocytes.     

Labeling of the glycoprotein subunit of (Na,K)ATPase with fluorescent probes

Sodium plus potassium activated adenosinetriphosphatase [(Na,K)ATPase] is composed of a catalytic subunit (alpha) and a glycoprotein subunit (beta) of unknown function. A method has been developed to label the beta subunit of purified dog kidney (Na,K)ATPase with fluorescent probes. The method consists of oxidation of beta-subunit oligosaccharides, reaction of the resulting aldehydes with fluorescent hydrazides, and reduction of the hydrazones and unreacted aldehydes with NaBH4. Two oxidation methods were compared. Simultaneous treatment with neuraminidase and galactose oxidase did not inhibit significantly (Na,K)ATPase activity and allowed insertion of up to 11 mol of probe per mol of beta. In contrast, oxidation of (Na,K)ATPase oligosaccharides with periodate resulted in 50-80% inhibition of the (Na,K)ATPase activity with low or undetectable labeling. Eleven commercial probes and two novel hydrazides were tested for labeling of (Na,K)ATPase treated with galactose oxidase and neuraminidase. Eight probes did not label (Na,-K)ATPase but labeled red cell ghosts oxidized with periodate. Four probes labeled beta specifically but either adsorbed to the membrane tightly, or cross-linked the beta subunits, or formed unstable adducts. Lucifer yellow CH labeled beta specifically without membrane adsorption. Labeling stoichiometries from 1 to 11 mol of lucifer yellow CH per mol of beta were obtained without inhibition of (Na,K)ATPase activity and without significant alteration of the anthroylouabain binding capacity or its association and dissociation kinetics. Anthroylouabain specifically bound to the lucifer-labeled (Na,K)ATPase had a decreased quantum yield, probably due to resonance energy transfer. This suggests that the sites of lucifer attachment on beta are within energy transfer distance from the cardiac glycoside site on alpha.(ABSTRACT TRUNCATED AT 250 WORDS)     

Radiation inactivation experiments predict that a large aggregate form of the insulin receptor is a highly active tyrosine-specific protein kinase

The technique of radiation inactivation has been used on a highly purified insulin receptor in order to determine the functional molecular size responsible for tyrosine-specific protein kinase activity. When both insulin binding and kinase activities were analyzed with the same receptor preparations, the functional size for kinase activity was found to be larger than that for insulin binding activity. The radiation inactivation curve for kinase activity was multiphasic. This indicates that at least two components contribute to total kinase activity. The average minimal functional size for the kinase was 370,000 +/- 60,000 daltons (n = 7) which corresponds to the alpha 2 beta 2 form of the insulin receptor. The average functional size for larger forms was estimated to be approximately 4 X 10(6) daltons. (To minimize the complexity of the model used in this analysis, we have analyzed the radiation inactivation curves of the insulin receptor kinase activity with a two-component model. However, we believe that the larger component, greater than 1 X 10(6) daltons, is probably not a single molecular weight species but rather represents a continuum of sizes or aggregates of the alpha 2 beta 2 form of the receptor.) These larger forms contributed 93% of the total activity. Mild reduction of the insulin receptor preparation with dithiothreitol (DTT) activated the total kinase activity by 3.5-fold. Under this condition, the minimal functional kinase size was 380,000 +/- 30,000 daltons (n = 6) while the average functional size for the larger forms was approximately 3 X 10(6) daltons.(ABSTRACT TRUNCATED AT 250 WORDS)     

Reversible inhibition of (Na+, K+) ATPase by Mg2+, adenosine triphosphate, and K+.

Adenosine triphosphate (ATP) hydrolysis catalyzed by the plasma membrane (Na+,K+)ATPase isolated from several sources was inhibited by Mg+, provided that K+ and ATP were also present. Phosphorylation of the adenosine triphosphatase (ATPase) by ATP and by inorganic phosphate was also inhibited, as was p-nitrophenyl phosphatase activity. (Ethylenedinitrilo)tetraacetic acid (EDTA) and catecholamines protected from and reversed the inhibition of ATP hydrolysis by Mg2+, K+ and ATP. EDTA was protected by chelation of Mg2+ but catecholamines acted by some other mechanism. The specificities of various nucleotides as inhibitors (in conjunction with Mg2+ and K+) and as substrates for the (Na+, K+) ATPase were strikingly different. ATP, ADP, beta,gamma-CH2-ATP and alpha,beta-CH2-ADP were active as inhibitors, whereas inosine, cytidine, uridine, and guanosine triphosphates (ITP, CTP, UTP, and GTP) and adenosine monophosphate (AMP) were not. On the other hand, ATP and CTP were substrates and beta,gamma-NH-ATP was a competitive inhibitor of ATP hydrolysis, but not an inhibitor in conjunction with Mg2+ and K+. The Ca2+-ATPase from sarcoplasmic reticulum and F1, the Mg2+-ATPase from the inner mitochondrial membrane, were also inhibited by Mg2+. Catecholamines reversed inhibition of the Ca2+-ATPase, but not that of F1.     

Further characterization of HeLa S3 plasma membrane ghosts

A plasma membrane fraction of HeLa S3 cells, consisting of ghosts, is characterized more fully. A simple procedure is described which permits light and electron microscope study of the plasma membrane fraction through the entire depth of the final product pellet and through large areas parallel to the surface. Contamination by nuclei is 0.14%, too little for DNA detection by the diphenylamine reaction. Contamination by rough endoplasmic reticulum and ribosomes is small, a single ghost containing about 3% of the RNA in a single cell. Mitochondria were not encountered. Electron microscopy also shows (a) small vesicles associated with the outer surface of the ghosts, and (b) a filamentous web at the inner face of the ghost membrane. Sodium dodecyl sulfate (SDS)-polyacrylamide gel analysis shows that of the many Coomassie Blue-stained bands two were prominent. One, 43,000 daltons, co-migrated with purified rabbit muscle actin and constituted about 7.5% of the plasma membrane protein. The other major band, 34,000 daltons, was concentrated in the plasma membrane fraction. Two major glycoproteins detected by autoradiography of [14C]fucose-labeled glycoproteins on the gels, had apparent molecular weights of 35,000 daltons and 32,000 daltons. These major bands did not stain with Coomassie Blue. There were many other minor glycoprotein bands in the 200,000- to 80,000-dalton range. Ouabain-sensitive, Na+, K+-adenosine triphosphatase (ATPase) activity of the ghost fraction is purified 9.1 (+/- 2.2) times over the homogenate; recover of the activity is 12.0 (+/- 3.8%) of the homogenate. Enrichment and recovery of fucosylglycoprotein parallel those for ouabain-sensitive Na+, K+-ATPase activity. Fucosyl glycoprotein is recovered more than the enzyme activity in a smooth membrane vesicle fraction probably containing the bulk of plasma membrane not recovered as ghosts.     

Evidence for a high and specific concentration of (Na+,K+)ATPase in the plasma membrane of the osteoclast

During bone resorption, the osteoclast actively acidifies a limited extracellular compartment. We hypothesized that, like other cells engaged in ion transport and proton translocation, the osteoclast's membrane might be highly enriched in sodium pumps. Using monoclonal antibodies to both the alpha and the beta subunits, immunoblot analysis, and [3H]ouabain binding, we have demonstrated that the osteoclast plasma membrane is both highly and specifically enriched in (Na+,K+)ATPase, compared with other bone cells, monocytes, macrophages, and other blood and bone marrow cells. The density of binding sites on the osteoclast is equivalent to that of kidney tubule cells. This observation is consistent with the hypothesis that the (Na+,K+)ATPase plays a role in the mechanism of bone resorption, possibly coupled with secondary active calcium and/or proton transport. Monoclonal antibodies against the (Na+,K+)ATPase can therefore be used as specific markers for the osteoclast in bone and bone marrow preparations.     

Kinetic characterization of Na,K-ATPase from rabbit outer renal medulla: properties of the (alpha beta)(2) dimer

We describe and compare the main kinetic characteristics of the (alpha beta)(2) form of rabbit kidney Na,K-ATPase. The dependence of ATPase activity on ATP concentration revealed high (K(0.5)=4 microM) and low (K(0.5)=1.4 mM) affinity sites for ATP, exhibiting negative cooperativity and a specific activity of approximately 700 U/mg. For p-nitrophenylphosphate (PNPP) as substrate, a single saturation curve was found, with a smaller apparent affinity of the enzyme for this substrate (K(0.5)=0.5 mM) and a lower hydrolysis rate (V(M)=42 U/mg). Stimulation of ATPase activity by K(+) (K(0.5)=0.63 mM), Na(+) (K(0.5)=11 mM) and Mg(2+) (K(0.5)=0.60 mM) all showed V(M)'s of approximately 600 U/mg and negative cooperativity. K(+) (K(0.5)=0.69 mM) and Mg(2+) (K(0.5)=0.57 mM) also stimulated PNPPase activity of the (alpha beta)(2) form. Ouabain (K(0.5)=0.01 microM and K(0.5)=0.1 mM) and orthovanadate (K(0.5)=0.06 microM) completely inhibited the ATPase activity of the (alpha beta)(2) form. The kinetic characteristics obtained constitute reference values for diprotomeric (alpha beta)(2)-units of Na,K-ATPase, thus contributing to a better understanding of the biochemical mechanisms of the enzyme.     

The lateral mobility of the (Na+,K+)-dependent ATPase in Madin-Darby canine kidney cells

Fluorescence microphotolysis (recovery after photobleaching) was used to determine the lateral mobility of the (Na+,K+)ATPase and a fluorescent lipid analogue in the plasma membrane of Madin-Darby canine kidney (MDCK) cells at different stages of development. Fluorescein-conjugated Fab' fragments prepared from rabbit anti-dog (Na+,K+)ATPase antibodies (IgG) and 5-(N-hexadecanoyl)aminofluorescein (HEDAF) were used to label the plasma membrane of confluent and subconfluent cultures of MDCK cells. Fractional fluorescence recovery was 50% and 80-90% for the protein and lipid probes, respectively, and was independent of developmental stage. The estimated diffusion constants of the mobile fraction were approximately 5 X 10(-10) cm2/s for the (Na+,K+)ATPase and approximately 2 X 10(-9) cm2/s for HEDAF. Only HEDAF diffusion showed dependency on developmental stage in that D for confluent cells was approximately twice that for subconfluent cells. These results indicate that (Na+,K+)ATPase is 50% immobilized in all developmental stages, whereas lipids in confluent MDCK cells are more mobile than in subconfluent cells. They suggest, furthermore, that the degree of immobilization of the (Na+,K+)ATPase is insufficient to explain its polar distribution, and they support restricted mobility of the ATPase through the tight junctions as the likely mechanism for preventing the diffusion of this protein into the apical domain of the plasma membrane in confluent cell cultures.     

Na+, K+-ATPase activity in erythrocytes after the effect of laser radiation

An in vitro single radiation of helium-neon laser (power flux density being 2 mW/cm2 exposure--1 and 3 min) does not change the concentration of Na+ and K+, activity of Na+, K+-dependent ATPase in erythrocytes and does not affect the intensity of active Na transport through their membrane in the donor blood. The 5 min laser action decreases the level of K+ and increases that of Na+ in the erythrocytes, activates Na+, K+-ATPases and intensifies the active Na+ transport.     

Papain fragmentation of the (Na+,K+)-ATPase beta subunit reveals multiple membrane-bound domains

Purified dog kidney (Na+,K+)-ATPase was reacted with tritiated sodium borohydride after treatment with neuraminidase and galactose oxidase. This procedure did not affect the ATPase activity of the enzyme, and all of the covalently bound radioactivity was found in the beta subunit (Mr 54 000). Papain digestion of the tritiated enzyme produced two labeled fragments of Mr 40 000 and 16 000. Further proteolysis generated an Mr 31 000 peptide from the larger fragment. Unlike the tryptic and chymotryptic sites of the alpha subunit, the sites of papain hydrolysis were insensitive to conformations of the (Na+,K+)-ATPase. Determination of the NH2-terminal sequences was used to arrange the fragments within the linear map of the beta chain. Finally, none of the labeled peptides was released from the membrane under nondenaturing conditions. These results are consistent with a model of the beta subunit containing a 40 000-dalton NH2-terminal piece and a 16 000-dalton COOH-terminal piece. Both fragments have extracellularly exposed carbohydrate and at least one membrane-bound domain.     

ATP inactivates hydrolysis of the K+-sensitive phosphoenzyme of kidney Na+,K+-transport ATPase and activates that of muscle sarcoplasmic reticulum Ca2+-transport ATPase

In order to characterize low affinity ATP-binding sites of renal (Na+,K+) ATPase and sarcoplasmic reticulum (Ca2+)ATPase, the effects of ATP on the splitting of the K+-sensitive phosphoenzymes were compared. ATP inactivated the dephosphorylation in the case of (Na+,K+)ATPase at relatively high concentrations, while activating it in the case of (Ca2+)ATPase. When various nucleotides were tested in place of ATP, inactivators of (Na+,K+)ATPase were found to be activators in (Ca2+)ATPase, with a few exceptions. In the absence of Mg2+, the half-maximum concentration of ATP for the inhibition or for the activation was about 0.35 mM or 0.25 mM, respectively. These values are comparable to the previously reported Km or the dissociation constant of the low affinity ATP site estimated from the steady-state kinetics of the stimulation of ATP hydrolysis or from binding measurements. By increasing the concentration of Mg2+, but not Na+, the effect of ATP on the phosphoenzyme of (Na+,K+)ATPase was reduced. On the other hand, Mg2+ did not modify the effect of ATP on the phosphoenzyme of (Ca2+)ATPase. During (Na+,K+)ATPase turnover, the low affinity ATP site appeared to be exposed in the phosphorylated form of the enzyme, but the magnesium-complexed ATP interacted poorly with the reactive K+-sensitive phosphoenzyme, which has a tightly bound magnesium, probably because of interaction between the divalent cations. In the presence of physiological levels of Mg2+ and K+, ATP appeared to bind to the (Na+,K+)ATPase only after the dephosphorylation, while it binds to the (Ca2+)-ATPase before the dephosphorylation to activate the turnover.