Insulin Sti~ulation of (~a+,K+)-Adenosine Triphosphatase-dependent "Rb+ Uptake in Rat Adipocytes"

Insulin stimulated the uptake of u6Rb+ (a K+ analog) in rat adipocytes and increased the steady state concentration of intracellular potassium. Half-maximal stimulation occurred at an insulin concentration of 200 PM. Both basal- and insulin-stimulated "Rb+ transport rates depended on the concentration of external K+, external Na+, and were 90% inhibited by M ouabain and M KCN, indicating that the hormone was ac- tivating the (Na+,K+)-ATPase. Insulin had no effect on the entry of 22Na+ or exit of "Rb'. Kinetic analysis demonstrated that insulin acted by increasing the max- imum velocity, V,,,, of "Rb' entry. Inhibition of the rate of Rb' uptake by ouabain was best described by a biphasic inhibition curve. Scatchard analysis of ouabain binding to intact cells indicated binding sites with multiple affinities. Only the rubidium transport sites which exhibited a high affinity for ouabain were stirn-ulated by insulin. Stimulation required insulin binding to an intact cell surface receptor, as it was reversible by trypsinization. We conclude that the uptake of "Rb+ by the (Na',K')-ATPase is an insulin-sensitive membrane transport process in the fat cell.

The sodium and potassium ion-activated adenosine triphosphatase is the enzyme responsible for active transport of sodium and potassium ions and the maintenance of a cationic gradient across the plasma membrane of almost all eukaryotic cells (for review, see Ref. 1). The activity of this membranebound transport protein is specifically inhibited by the cardiac glycoside ouabain, which binds to the portion of the enzyme facing the extracellular fluid (2) and thereby inhibits ion transport and ATP hydrolysis.
There are numerous reports in the literature that the levels of intracellular Na' and K' are altered by insulin. As early as 1924, it was observed that the levels of serum potassium decreased in response to administration of insulin in vivo ( 3 ) . More recent in vitro studies document insulin stimulation of K' uptake (4)(5)(6)(7)(8) and Na" efflux (8)(9)(10)(11)(12) in various preparations of intact frog and rat muscle, as well as rat adipose tissue ( 1 3 , mouse fat cells (141, and duck salt gland (15). In several cases the insulin effect was shown to be inhibited by cardiac glycosides (7, 8, 10-12, 14, ls), suggesting that insulin affects the transport activity of the (Na',K')-ATPase. ' However, no * This research was supported by Grant HL08893 from the National Institutes of Health and Grant PLM 78-04364 from the National Science Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "ndwrtisenzent" in accordance with I8 U.S.C. Section IT34 solely to indicate this fact.
$ National Institutes of Health Predoctoral Trainee (GM 138). single study has demonstrated that physiological concentrations of insulin stimulate potassium t,ransport, that this effect is mediated by the (Na',K')-ATPase and correlates with insulin binding to its cell surface receptor. The rat adipocyte is an isolated cell type in which the binding of insulin to its receptor and the effects of the hormone on the membrane transport system for glucose have been studied in great detail (16)(17)(18). We, therefore, decided to investigate the effects of insulin on ion transport in the fat cell in order to further define the kinetic parameters altered by hormonal stimulation of the (Na',K+)-ATPase, to correlate the insulin effect with binding to the cell surface receptor, and to determine the ''Kacti<,,,'' (19) of insulin on K' transport in the intact isolated cell. In addition, we wondered whether there were similarities in the way insulin affects the transport systems for glucose and K' in the same cell. This paper demonstrates that insulin stimulated the uptake of rubidium (a potassium analog) into rat adipocytes and that this increased uptake was a consequence of the increased activity of the (Na',K')-ATPase. The stimulation of "Rb+ uptake serves as another criterion for insulin responsiveness in the intact cell.

EXPERIMENTAL PROCEDURES
Isolation of F a t CeZls-Adipocytes were prepared according to the method of Rodbell (20). Epididymal fat pads were removed from 150t.o 200-g male rats (CD strain, Charles River Breeding Laboratories) immediately following decapitation. Digest.ion was carried out in Krebs-Ringer phosphate buffer, containing 4% (w/v) bovine serum albumin and 1 mg/ml of bacterial collagenase (type I). The buffer contained 140 mM NaC1/5 mM KC1/1.4 mM CaCln/l mM MgS04/10 mM NarHPO, and was adjusted to pH 7. 4 with NaOH after addition of albumin. Following a 1-h digestion a t 37OC in a shaking water bath, the cells were filtered through Japanese silk (8xx mesh) and washed three times with the albumin buffer in an International clinical centrifuge.
Transport Assays-The adipocyte suspension was made up to a 30 to 40% packed cell volume in round bottom polypropylene tubes. All transport assays were performed a t 37°C in Krebs-Ringer phosphate albumin buffer, unless otherwise stated. Fol~owing the addition of tracer amounts of radioisotope (final concentration, 2 to 10 pCi/ ml), duplicate 100-pI aliquots of Eel1 suspension were removed at the indicated time intervals. Separation of cells from radioactive medium was achieved using the oil centrifugation technique of Gliemann et al. (21). The cell suspension was centrifuged through dinonyl phthalate in a 4OO-pl polyethylene tube for 20 s in a Beckman Microfuge B. The tube was then frozen in a dry ice/acetone mixture and sliced through the oil layer. The cell layer and aqueous "subnatant" were placed in separate 20-ml scintillation vials and solubilized by vortexing in 0.4 ml of 5% sodium dodecyl sulfate. After 1 h, 4 ml of Aquasol (New England Nuclear) was added, and the vials were capped, vortexed, and counted in a Beckman liquid scintillation counter.
For determination of the K,,* and V,,,,, of ""Rh' transport, fat cells were washed three times with K'-free buffer and resuspended in buffer containing the indicated concentration of RbCI. All tubes contained 0.5 mi of 30% cells and were preincubated for 10 min a t 37°C in the presence or absence of 8 nM insulin. Transport was initiated by the addition of 5 pl of ""RbCI. Ten minutes later, triplicate 20938 This is an Open Access article under the CC BY license. 100-pl aliquots were withdrawn and processed as described above.
To determine the rate of "Na' efflux, 1.0 ml of packed fat cells (prepared in Krebs-Ringer phosphate buffer) was incubated with 20 pCi of "NaC1 for 10 min at 37°C. Insulin and/or ouabain was added during the final 2 min of the incubation. Following the addition of 2.0 mi of Na'-free buffer (140 mM choline chloride/5 ! " KClI1.4 mM CaC1,/1 mM MgSOs/10 mM ~s -p h o s p h a t e /~% bovine serum albumin, pH 7.4) (choline Ringer), the cells were spun in an International clinical centrifuge and resuspended in 1.5 ml of choline Ringer with or without ouabain and/or insulin. The final cell suspension contained 2.5 ml of 40% adipocytes. Washing and resuspension of the cells was accomplished in 30 to 40 s. Time zero was defined as the time of addition of the 1.5 ml of choline Ringer. "' Na' transport was assayed at 37°C as described above. Alternatively, 0.5 ml of packed fat cells was loaded with "Na' as above, and 4 ml of choline Ringer were added at time zero. Both methods yielded nearly identical results.
Intracellular water volume and extracellular trapped volume measurements were made on suspensions of adipocytes containing 1 to 5 pCi/ml of ['H]HaO and ['4C]sucrose. These determinations were always made in parallel with experimental tubes containing "'Rb'/ ['H JH2D or "~Na'/['H]H~O. The spillover of "'Rb' or :"Na+ into the set "H-channel was less than 2%.
Ouabain-binding Analysis-Cells were placed in tubes containing 4 nM insulin or an equivalent volume of buffer and incubated at 37°C for 10 min. Then each tube received 1 X 10.' M [:'H]ouabain and nonradioactive ouabain in a concentration range of 5 X lo-" M to 1 X lo-,' M. Following a 35-min incubation at 37°C (a time sufficient to reach equilibrium), duplicate 200-4 aliquots were withdrawn from all tubes and processed as described.

Insulin Effect on Rb'
Transport-Transport activity of the (Na',K+)-ATPase can be monitored with radioactive rubidium (wRbf), which substitutes for potassium in the activation of ATP hydrolytic activity and the active transport of K+ into the cell (22). The time course of @Rb' uptake by rat adipocytes is shown in Fig. 1. Addition of insulin (8 m) caused a 50% increase in the rate of rubidium uptake (rate increase, 50% 2 20% in 33 experiments). This increase was apparent within 2 min after exposure to insulin, and it occurred when insulin was added to cells before, at the same time as, or 30 min after @'jRb+ addition. The fractional increase in the rate of @'jRb' uptake was independent of the time of addition of insulin. Identical results were obtained when 42K+ was used as the tracer. However, due to its longer isotopic half-life (19 days for @'jRb+ uersus 12 h for 42K+), =Rb+ was far more convenient to handle and was used for all subsequent studies.
The uptake of =Rb+ shown in Fig. 1B exhibited saturation kinetics, indicating that the system was approaching isotopic steady state. We calculated the steady state level of K+ in the cell based on the amount of =Rb+ which accumulated at saturation (Fig. 1B) and the intracellular water volume (determined as the net ["H]H20 volume in the presence of [".C]sucrose). In four separate experiments, the intracellular K' concentration was 177 mM (A17 mM, S.D.) in control untreated cells and 7 m~ (+2 mM) in the presence of 1 m~ ouabain; insulin increased this value to 222 m~ (225 mM). Thus the 50% stimulation of the initial rate of rubidium uptake induced by insulin resulted in an increase in the steady state K+ concentration of roughly 40 m~.
If the insulin effect is mediated through the (Na+,K+)- ml , 30% cells) were incubated for 15 min at 37OC in the presence or absence of 8 nM insulin and/or I r n~ ouabain. The transport "say was initiated by the addition of 2 pCi of "'HbCI to each tube; duplicate loo-$ aliquots were withdrawn at the indicated times and processed as described under "Experimental Procedures." Symbols are as follows: e, albumin buffer; 0 , 8 nM insulin; A, 1 mM ouabain; A, 1 mM ouabain + 8 nM insulin. Halffilled symbols are coincident data points.
All data points represent the average of duplicate determinations. "Rb' uptake is expressed as nanomoles of Rb' (K') per 100 p1 of cell suspension and is calculated based on the specific activity of %RbCI in a medium containing 5 r n~ KCl. The lines drawn in Fig. 1B represent the best fit of the data to a saturation kinetics curve by regression analysis (correlation coeffkient r = 0.98). In A, the difference between filled and open circle raw data points was statistically significant ( p < 0.005), whereas the difference between fiUed and open triangles was not ( p > 0.10) (paired Student t test).
ATPase, then stimulation should be abolished by ouabain, a specific inhibitor of the cation pump. The data in Fig. 1 and Table I verify this inference. In the presence of M ouabain, the basal rate of Rb' transport was inhibited 90%, and insulin stimulation was blocked entirely. This was not due to an alteration of insulin-binding capacity, since M ouabain had no effect on specific binding of '"I-insulin to the cell (data not shown). In addition, treatment of adipocytes with 1.0 mM KCN, which lowers cellular ATP levels to 10% of control values (23), reduced both basal and insulin-stimulated %Rb+ uptake rates to the same level observed in the presence of ouabain ( Table I) (Table I).
An increase in enzyme activity can be mediated by a change in the Michaelis constant, K,, or maximum velocity, V,,,, or both. Fig. 2 demonstrates that the Km of the (Na',K+)-ATPase for Rb' (1.2 mM) remained unchanged in the presence of insulin but that the V,,, of Rb' transport was increased. This effect was not a response to an insulin-induced increase in K' efflux from the cell. When cells were preincubated with =Rb', and ouabain was then added, the leakage of Rb+ out of the cell was quite slow (Fig. 3). Insulin had no apparent effect on the exchange of Rb' for external K' during the time course monitored. Furthermore, if fat cells were preincubated with &"Rb+, then resuspended in K'-free buffer, insulin had no observable effect on the efflux of Rb' .
Adipocyte Ma+ Fluxes-To test the possibility that insulin increases the entry of Na+ into the cell, the uptake of "Na+ was monitored after the addition of insulin. Ouabain was present to prevent efflux of Na+ through the (Na+,K+)-ATPase. No difference in the rate or amount of "Na+ entering the adipocyte in the presence or absence of hormone was observed ( Fig. 4A). Furthermore, when ouabain and insdin were added s~~t a n~u s l y (at time zero) to the adipocyte suspension,

TABLE I Effect of insulin and ATPase inhibitors on WKRb+ uptake
A 30% adipocyte suspension was incubated for 15 min at 37'C in the absence or presence of the indicated compounds. Transport was initiated by the addition of 2 pCi of =RbCl. Ten minutes later, triplicate 100-pl afiquots were withdrawn and processed. Rate of "Rb+ uptake as a function of Rb' concentration, in the presence and absence of insulin. Fat cells were washed and resuspended in KC1-free buffer containing the indicated concentration of RbCi. Each set of tubes contained 0.5 ml of a 30% adipocyte suspension and was incubated for 10 min in the absence (0) or presence (0) of 8 nM insulin. Transport was initiated by the addition of 2.5 pCi of "RbCl (final concentration, 60 p~) .

Rate of =Rb+ uDtake
Ten minutes later, triplicate 100-pl aliquots were withdrawn and processed. Each point is the average of triplicate determinations. Z2Na+ uptake was identical with that observed in cells treated only with ouabain at time zero (data not shown).
In the absence of ouabain, the rate of "Na+ tracer equilibration was extremely rapid (less than 2 rnin), and the level of 22Nac in the cell remained constant for at least 2% h (Fig. 43). Using the data of Fig. 4 3 and a concomitant measure of net intracellular water volume, we calculated that the internal Na+ concentration was approximately 15 mM ( k 5 mM, S.D.).
The amount of tracer in the insulin-treated cells was slightly lower, but this difference is not statistically significant. In contrast, in the presence of ouabain (Fig. 4B, upper curve), the Na+ level rose to nearly 120 m~ by 3 h, suggesting that the Na+ concentration gradient normally maintained by the (Na',K')-ATPTe had been dissipated.
The next set of experiments was designed to determine the effect of insulin on (Na',K')-ATPase-mediated Na' efflux.
When cells were preincubated with "Na+ and resuspended in nonradioactive Krebs-Ringer buffer, no difference in the loss of "Na+ from the cells was observed in the presence or absence of 1 mM ouabain (data not shown). One explanation for this observation is that a considerable portion of the isotope "loss" is due to the rapid exchange of internal radioactive Na' for external unlabeled Na+. Accordingly, adipocytes were loaded with 2'Na+ in Krebs-Ringer buffer and then resuspended in Na+-free choline Ringer buffer (see under "Experimental Procedures") in order to monitor net isotope efflux. The results of this experiment are shown in Fig. 5. The half-time for 'lNa+ efflux was 2 min; in the presence of 1 mM ouabain, the t l / z efflux was 9 to 10 min. Unexpectedly, no effect on the rate of ouabain-inhibitable "' Na' efflux was detected when insulin was added. We then tested whether adipocytes remain insulin responsive under the experimental conditions used to minotor **Na+ efflux. L-Arabinose is a nonmetabolizable sugar whose transport has been shown to occur through the D-glucose carrier and to be stimulated by insulin (25). When adipocytes were loaded with ~-[l-'~C]arabinose in Krebs-Ringer buffer and then resuspended in choline Ringer buffer, no increase in arabinose efflux was observed upon addition of insulin (data not shown). However, cells incubated for an additional 20 min in choline Ouabain was present to prevent efflux of Na' through the (Na+,K') pump. At time zero, 30 pCi of '!NaCl (carrier free) was added to both tubes, and one tube received 3 nM insulin. Aliquots (100 pl each) were withdrawn at the indicated times. "Na" uptake is expressed as a percentage (*'Na+ counts per min in ceUs/total r2Na+ counts per min) per 100-p1 aliquot, after subtraction of trapped volume. Each data point represents the average of duplicate determinations (0, control; 0, 3 nM insulin). Statistical analysis of individual real data points revealed no significant difference in uptake values ( p > 0.10 by paired Student t test). B, "Na" tracer equilibration levels. Adipocytes (2.0 ml, 30% cells) were incubated for 15 min at 37OC in the presence or absence of 8 nM insulin or 1 m M ouabain. At time zero, 30 pCi of 22NaC1 was added to each tube, and '2Na+ uptake was determined as described above. Each data point represents the average of duplicate determinations. Symbols are the same as in Fig. 1.

Ringer buffer did exhibit insulin-stimulated arabinose flux.'
These observations agree with the results of Vega and Kono (26), who reported that mechanical agitation (such as shaking or centrifugation of the cells) transiently blunts the insulin effect on sugar transport in fat cells. Since it was not possible to resuspend cells in Na+-free buffer without c e n t~g a t i o n or shaking nor was it possible to allow the cells to "recover" for 15 min following c e n t~g a t i o n (at which time all the internal Na' would have been lost), an experiment to measure insul i n -s t i m~a~d Na' efflux was technically not feasible. This difficulty could be overcome if we could find an inhibitor of the ouabain-insensitive component of Na' flux. To date, no inhibition of 22Na+ uptake has been obtained with 0.5 mM amiloride, 1 PM tetrodotoxin, or 1 II~M tetracaine." Cation Activation of Rb' Uptake-The transport activity of the (Na',K')-ATPase in red blood cell ghosts has been shown to be sensitive to the internal and external concentrations of Na' and K' (1). Table I1 shows that the rate of =Rb+ uptake in the rat adipocyte increased as the concentration of external Na' was increased and that a maximum rate was achieved a t 145 mM Na' . These data are consistent with the interpretation that the concentration of extracellular Na' determines the level of intracellular sodium ion which is a substrate for the (Na',K')-ATPase; in the absence of external Na', the supply of internal Na' was depleted by the pump in less than 5 min (Fig. 5) and the pump shut down. It is conceivable that insulin stimulates the transport activity of the (Na',K')-ATPase by increasing the supply of internal Na'. To test this possibility, the rate of %Rb' uptake was monitored in the presence of various concentrations of monensin, a carboxylic acid ionophore which catalyzes the electroneutral exchange of Na" and protons across the cell mem-

TABLE I1
Effect of external NaCl on the rate of %Rb+ uptake Fat cells were washed and resuspended in Ringer buffer containing the indicated concentrations of NaCI. The Na' concentration was varied by iso-osmotic substitution of choline chloride for NaCl. The sodium-free buffer contained 10 m~ Tris-phosphate, pH 7.4, instead of 10 m sodium phosphate. Following a 15-min incubation at 37°C in the presence or absence of 8 rn insulin, 2 pCi/ml of =RbCl was added to all tubes. Duplicate 100-4 aliquots were withdrawn 5 and IO min later. The ouabain-resistant rate has been subtracted from all values. [NaCl] Rate of =Rb+ uptake centration was approximately 40% higher than in control untreated cells (data not shown).

Minus insulin Plus insulin
The hypothesis that insulin activates =Rb' uptake by increasing Na' entry is untenable in the rat adipocyte for the following reasons. No change in "Na+ entry was detected in the presence: of insulin (Fig. 4A). The basal rate of Rb' uptake was not increased when the external Na' concentration was raised from 145 mM to 195 mM, and the insulin-stimulated rate at both Na+ concentrations was identical (Table 11). If insulin acted by increasing the membrane permeability to Na+ ions, we would have expected to see a much larger insulin stimulation of the rate of Rb' uptake with a 13:l Na' concentration gradient (195 m~) compared to that observed at 145 mM Na' . Finally, although the Na' ionophore monensin stimulates ' ' Rb' uptake to the same extent as insulin, it does so by increasing the intracellular Na' concentration. Clearly, no such increase in Na' content was observed in the presence of insulin (Fig. 4B). Taken together, these results suggest that insulin acted to increase the activity of the (Na',K+)-ATPase and that this effect was not due to an alteration of intrinsic membrane permeability to Na' or K' , nor to the opening of an ion-specific gate or "channel" in the membrane (see under "Discussion").
Requirement for the Insulin Receptor-Several lines of evidence support the claim that stimulation of Rb' uptake requires intact insulin and an intact insulin receptor. Insulin was briefly incubated at 100°C and then repeatedly frozen and thawed. This "heat-inactivated'' insulin no longer stimulated Rb' uptake, whereas native untreated hormone did. When epididymal fat pads were treated with high concentrations of insulin (400 n~) during collagenase digestion, the cells isolated from this tissue exhibited a stimulated rate of Rb' uptake. Subsequent addition of 3 n~ insulin was without further effect.
The insulin concentration dependence of this stimulation is depicted in Fig. 6 Treatment with 1 mg/ml of trypsin for 15 min at 37"C, followed by addition of 3 mg/ml of soybean trypsin inhibitor, destroyed the "'I-insulin binding capacity of adipocytes (data not shown and compare with Ref. 30). Subsequent exposure of these cells to insulin did not result in a stimulated Rb' uptake rate. This concentration of trypsin had no effect on the basal rate of Rb' uptake ( Fig. 7). We conclude that stimulation of Rb' uptake requires insulin binding to the intact cell surface receptor. Furthermore, treatment of insulinactivated cells with trypsin (Fig. 7) resulted in destruction of the insulin binding with a concomitant lowering of the stimulated Rb' uptake rate to that of control cells. This implies that insulin activation of the (Na',K')-ATPase is a reversible phenomenon.
Ouabain Binding a n d Inhibition of Transport-The increase in the VmaX of Rb' uptake could be the result of an increase in the number of ion pumps or an activation of existing pumps. To distinguish between these two possibilities it was necessary to quantitate the number of (Na',K+)-ATP- ases on the cell surface. Fig. 8 shows a Scatchard plot (31) of ["Hlouabain binding to intact cells and indicates the presence of binding sites with multiple affinities for the inhibitor. There is a class of sites with a high affinity for ouabain; these sites exhibit an apparent Kd of 1.5 X lo-? M and number approximately 2 X lo6 per cell. Insulin had no apparent effect on the affinity for ouabain or the number of these sites. In addition, a larger number of lower affinity sites was detected including nonspecific binding sites. Unfortunately, analysis of this region of the curve is extremely difficult and precludes an accurate determination of binding constants.
To determine if insulin would stimulate the pump sites with high affinity for ouabain, the rate of rubidium transport was monitored as a function of ouabain concentration, in the presence and absence of insulin. If insulin was activating the high affinity pump sites, stimulation should be abolished at low ouabain concentrations. Conversely, if the low affinity sites were affected by insulin, the stimulated rate of Rb' uptake should approximate the control rate only at relatively high concentrations of ouabain. Fig. 9 supports the first hypothesis. Ouabain inhibition of the rate of Rb' uptake is best described by a biphasic inhibition curve. Insulin stimulation occurred only at ouabain concentrations less than lomti M. Between and M ouabain, the insulin-treated and control rates were identical. This result suggests that the effect of insulin is on the pumps which have a high affinity for ouabain (Fig. 8). These pump sites were responsible for approximately 30 to 40% of the active transport of Rb' and exhibited a KI (concentration at which half-maximal inhibition is achieved) for ouabain of 1 X lo-' M. The lower affinity sites constituted the remainder of the transport activity with a KI of 1 X loF5 M ouabain. Based on these results, it is evident that the rat adipocyte contains at least two types of (Na+,K")-ATPases, as defined by their affinity and sensitivity to ouabain. The insulin stimulation of *'jRb+ transport described in this paper is mediated by the activation of a subset of cation pump sites which constitute the high affinity ouabain-binding sites on the cell surface.

DISCUSSION
Insulin Stimulation of K' Transport-The results presented here indicate that insulin stimulates the uptake of *6Rb' mediated by the (Na',K")-ATPase in the rat adipocyte. Within 2 min after exposure to the hormone, a 50% increase in the rate of wRb' uptake was observed in isolated fat cells. As a consequence, a higher level of intracellular potassium was achieved (basal = 180 mM; insulin-stimulated = 220 mM K"). The uptake of "Rb" in the presence and absence of insulin was activated by external K" and external Na', was energy dependent, and was ouabain inhibitable. These observations strongly suggest that insulin stimulates the (Na',K')-ATPase activity of the rat adipocyte. The fat cell contained at least two types of (Na',K')-ATPases, only one of which was stimulated by insulin. Insulin acted by increasing the V,,, of Rb' entry into the cell, with no immediate effect on the exit of Rb' or entry of Na'. The hormone effect was mediated by physiological concentrations of insulin, required insulin binding to i t s intact cell surface receptor, and was reversible by trypsinization.
Insulin activation of monovalent cation transport has been reported in several other systems (4-15). Exposure of rat soleus muscle (8, 321, duck salt gland (15), and rat uterine muscle (11) to 100 m~i u n i t s /~ of insulin (740 nM) results in 20 to 30% increases in ouabain-i~ibitable *' Kc (or %Rb+) influx. Relatively few studies have investigated hormone-sensitive potassium accumulation in adipose tissue. The K' content of rat fat pads increases when the tissue is incubated for 3 h at 37'C in the presence of 50 milliunits/ml of insulin and 8 mM glucose, with no change evident in the level of Na' (13). In isolated mouse fat cells, a decrease in 4"K' uptake induced by adrenaline is prevented when insulin is present; however, insulin alone does not alter net K' uptake (14). Clausen et at. (33) reported that rat fat cell ghosts take up *'K+ by an energy-dependent ouabain-inhibitable process but did not investigate the effect of insulin. Prior to this report, therefore, no study had demonstrated a direct effect of insulin on K' uptake in an intact, isolated cell preparation. Moreover, the insulin concentration dependence for this effect was not de-termined in any tissue type previously investigated.
Stimulation of the (Na',K')-ATPase in quiescent mouse 3T3 cells by serum has been documented (34) and exhibits kinetic characteristics similar to those observed in rat adipocytes. Serum stimulates the V,,, of Rb' influx without affecting Rb' efflux. The stimulation observed is rapid (within 2 rnin after addition of serum) and is blocked by ouabain (34). However, the primary effect of serum on 3T3 cells appears to be an increase in Na' entry into the cells (29), which then activates the (Na+,K+) pump, resulting in an increase in Rb' uptake. Although increasing the intracellular Na' concentration (in the presence of monensin) will stimulate the uptake of Rb' in adipocytes, this mechanism is probably not involved in the stimulation of fat cell (Na',K')-ATPase by insulin. When Na' efflux through the pump was prevented by ouabain, no increase in the rate or amount of ' "a' entering the adipocyte was detected in the presence of insulin (Fig. 4A).
There was no significant change in the intracellular steady state Na' level when insulin was present (Fig. 4B). Finally, increasing the Na' concentration gradient (by altering the external NaCl concentration) did not magnify insulin stimulation of Rb' uptake. Thus, although the activity of the adipocyte (Na',K+)-ATPase was dependent on the external: internal Na' concentration gradient and the intracellular Na' concentration, it did not appear that the mechanism of insulin stimulation of the pump involved a noticeable alteration in the cell's internal level of sodium. Other reports that have documented increases in K' content in rat uterus and fat pads have not observed any change in Na' levels with insulin (11,13).
In principle, it should be possible to detect a concomitant increase in Na' efflux from a cell or tissue type in which insulin stimulates the (Na',K')-ATPase. Such increases in Na' efflux have been reported for various muscle preparations (8,10,12). The rat adipocyte contains 15 fmol of Na'/cell (Fig. 4B). Assuming a pumping rate of 2 X loy ions/cell/min (Fig. 2), we calculate that it would take 4 to 5 min for a fat cell to pump out all the available internal Na'. This is what was found experimentally (Fig. 5). As mentioned under "Results," rapid resuspension of adipocytes in Na'-Eree buffer transiently obscured insulin stimulation of membrane transport events and, therefore, no hormone effect on net P2Na+ efflux could be observed.
It must be emphasized that accurate measurements of intracellular concentrations of Na' are extremely difficult due to the miniscule intracellular water volume of the fat cell (about lo"* literslcell) (21). Although the amount of Z"Na' in the insulin-treated cells was slightly lower than that of control cells (Fig. 4B), the variance in the data makes it impossible to state that this difference is significant ( p > 0.10 by the Student t test). In order to be considered statistically significant, the decrease in sodium content of insulin-treated cells would have had to be greater than 5 mM, when compared to the Na' content of control cells (15 mM, *5 S.D.).
The results presented in this paper indicate that insulin stimulated the activity of the adipocyte (Na',K')-ATPase. In the presence of hormone, the initial rate of uptake of K' was increased by 50%, there was no change in K' efflux, and as a consequence, the steady state intracellular K' concentration was increased by 40 mM. We did not detect a decrease in the steady state level of intracellular Na+. One possible explanation for this observation is that anything less than a 30% decrease in internal Na' content would have been obscured by the variance in the data (i.e. a change from 15 mM to 10 mM Na' could not be considered significant). If the intracellular Na' concentration was indeed lower in insulin-treated cells, then, assuming no change in the rate of Na' influx ( Fig.   4A), one would expect that the activity of the Na' pump would fall until a new steady state was reached. At this point, rubidium entry and sodium efflux rates would approximate those rates observed before insulin was added. This hypothesis predicts that insulin stimulation of Rb' entry would be observable only as a transient effect. However, we have observed increased rates of Rb' influx in the presence of insulin for as long as 60 min. Since it is unlikely that it would take more than 5 min to readjust intracellular Na' levels ( Fig. 5) and thus the activity of the pump, the effect of insulin is probably to increase the steady state rate of Rb' influx through the (Na',K')-ATPase.
Alternatively, insulin stimulation of the pump may effect a transient decrease in the internal Na' concentration, which would increase the externakinternal Na' gradient. (For example, decreasing the internal Na' concentration from 15 mM to 10 mM would increase the Na' concentration gradient from 9.67:l to 14.51, a 50% higher value.) Nonpump-mediated Na' influx might then rapidly restore the original intracellular Na' levels, and thus no change in the steady state concentration of Na' would be detected. Thus, an increased rate of Na' entry would be observed only in the presence of an insulinstimulated active Na' pump. Given the resolution of current methods employed to monitor Na' transport, we would not be able to detect a rapid transient effect of insulin on ?'Na+ fluxes.
There is an additional complication concerning the observation that during stimulation by insulin the K' (Rb') concentration in the fat cell increased while there was not a measurable similar decrease in the Na' concentration. This new distribution of ions requires that either the volume of the fat cell increases, or that there is sequestration of K' ions into intracellular c o m p~m e n t s , or that there is excretion of some other solute to decrease the osmolarity of the cell, or that the cell membrane is not permeable to water. Since the volume of the fat cell did not change after exposure to insulin (control, 0.86 picoliter/cell * 0.08, S.D.; insulin treated, 0.83 picoliter/ cell +-0.07, S.D.), and the possibility that the membrane is not permeable to water is unlikely, one must presume that osmotic balance is achieved either by intracellular sequestration of K' ions or by loss of some unidentified solute.
Ouabain-binding Sites-Ouabain is a specific inhibitor of the (Na',K')-ATPase. The affinity of the pump for the cardiac glycoside depends on the tissue type and the nature of the ligands present (35). The (Na',K')-ATPase of rat tissues is relativeiy insensitive to ouabain; its affinity for this cardiac glycoside is 10'-to lO'-fold lower than that of the enzymes of other organisms (36,37). In agreement with the data obtained from other rat tissues, complete inhibition of rat adipocyte ATPase was not achieved until the ouabain concentration reached lo-'' M. However, a more detailed kinetic analysis revealed a subset of pumps with high affinity for ouabain (Fig.  9). The biphasic nature of the ouabain i~i b i t i o n curve implies that, in the adipocyte, at least two different. pump sites are actively pumping Rb', with K I values for ouabain of 1 X lo-' M and 1 x M, respectively. A recent publication from this laboratory documents the existence of two distinct molecular forms of the (Na',K')-ATPase in brain tissue (381, with Kt values for the inhibition of ATPase activity by strophanthidin of 2 x IO-: M and 2 x M. These two forms are localized in different subfractions of nervous tissue. The two (Na',K')-ATPases described in this paper appear to be in the same Cell type. The procedure utilized for separation of cells from medium selects only for those cells which are of density less than that of the oil layer (specific gravity 0.98 (21)); any contaminating cell types would pellet under these conditions. We conclude that the rat adipocyte contains two types of

Insulin Stimulation
of Adipocyte (Na',K')-ATPase (Na',K')-ATPase which can be distinguished by their different affinities for ouabain and sensitivity to insulin. The Scatchard plot of ouabain binding to intact adipocytes ( Fig. 8) is nonlinear, consistent with the existence of more than one class of ouabain-binding site. The area of interest is the high affinity portion of the curve, since this corresponds to the class of insulin-stimulated pump sites. These sites exhibit a Kd of 1.5 X M for ouabain and number 2.2 X lo6 sites/cell. Clausen and Hansen (39) obtain a similar number, 1.2 X lo6 ouabain binding sites/cell, with an apparent K d of 1.7 x IO-@ M, in potassium-free medium. The difference in Kd values can be explained by the effect of potassium, which lowers the affinity of the pump for ouabain without affecting the number of binding sites (35). In agreement with the results of Clausen and Hansen (40), we found no change in the number of (Na',K')-ATPase molecules in the fat cell in the presence of insulin.
If one assumes that the (Na',K')-ATPase has a turnover number for ATP of 1OO/s (41) and that 2 K' ions are transported for each ATP molecule hydrolyzed (42), we estimate that there are 2 X lo5 enzyme molecules per cell that are actively pumping K' . This means that less than 10% of the high affinity ouabain-binding sites are active pump sites. If there is a change in the number of active pumps in the presence of insulin, it will be obscured by the large number of inactive binding sites, and thus the change will be impossible to detect by Scatchard analysis. On the other hand, it is also possible that the rat (Na',K+)-ATPase has a much lower turnover number than the enzyme from other organisms or that the adipocyte ion pumps are operating a t 10% of the maximal rate. Whether insulin increases the number of (Na+,K+) pumps or increases the turnover number of the already active pumps remains an unresolved question.
Insulin Stimulation of Membrane Transport Actiuity-The most widely characterized effect of insulin on the fat cell is the stimulation of glucose transport (16)(17)(18). It is of interest to compare this process with the stimulation of Rb' transport described in this paper. Both effects occur on a similar time scale for activation and require only 2 to 5% occupancy of the insulin receptors for half-maximal stimulation (19, 43). In addition, both effects require insulin binding to an intact cell surface receptor (30) but can be mimicked by addition of antiadipocyte membrane antibodies3* (44). However, the stimulation of Rb' uptake occurs in the absence of glucose in the medium and is not inhibited by cytochalasin B, a specific inhibitor of basal and insulin-stimulated glucose uptake. 3 The presence of M ouabain does not immediately affect basal or insulin-stimulated hexose uptake in adipocytes' or soleus muscle (45), although prolonged exposure (90 min) to ouabain has been reported to stimulate 3-0-methylglucose efflux in muscle (45). L-Arabinose transport is unaffected when adipocytes are suspended in Na'-free medium,' but basal and insulin-stimulated =Rb+ uptake rates are completely inhibited under these conditions. Clearly, uptakes of monovalent cations and sugars are mediated by two separate membrane transport systems. Given the similarity in the way that insulin affects these two transport systems, it is reasonable to speculate that hormone interaction with the insulin receptor generates a common signal which serves to activate two different membrane transport proteins. The nature of this putative coupling mechanism is currently under investigation.