Use of a Monoclonal Antibody to Quantify (Na+,K+)-ATPase Activity and Sites in Normal and Regenerating Rat Liver*

Quantitative measurements of (Na+,K+)-ATPase activity and numbers of (Na+,K+)-ATPase sites in mem- branes from quiescent and regenerating rat liver have been made using an anticatalytic monoclonal antibody (9-A5) that binds to a subunits of the sodium pump (Schenk, D. B., and Leffert, H. L. (1983) Proc. Nutl. Acad. Sei. U. S. A. 80, 5281-5285). To validate the measurements, kinetic properties of 9-A5 binding to plasma membrane sodium pumps, specificity and re- quirements of the reactions, and mechanisms by which 9-A5 inhibits (Na+,K+)-ATPase were analyzed. "'1-9- A5 binding is saturable and reversible (kl = 1.8 X los*M"*S"; k2 = 2.7 X 10-4*s-'). At equilibrium, 9-A5 binds to a single class of sites revealed by Scatchard plots ( K D , ~ ~ ~ ~ = 0.64 nM, Bmmx = 29.3 pmol/mg of proteins; = 238,000 sites-cell-'). This binding requires monovalent cations (sodium, potassium, or lithium); is blocked by purified (Na+,K')-ATPase; is inhibited noncompetitively by ATP (Knappl = 0.5

Quantitative measurements of (Na+,K+)-ATPase activity and numbers of (Na+,K+)-ATPase sites in membranes from quiescent and regenerating rat liver have been made using an anticatalytic monoclonal antibody (9-A5) that binds to a subunits of the sodium pump  [5281][5282][5283][5284][5285]. To validate the measurements, kinetic properties of 9-A5 binding to plasma membrane sodium pumps, specificity and requirements of the reactions, and mechanisms by which 9-A5 inhibits (Na+,K+)-ATPase were analyzed. "'1-9-A 5 binding is saturable and reversible (kl = 1.8 X los*M"*S"; k2 = 2.7 X 10-4*s-'). At equilibrium, 9-A5 binds to a single class of sites revealed by Scatchard plots ( K D ,~~~~ = 0.64 nM, Bmmx = 29.3 pmol/mg of proteins; = 238,000 sites-cell-'). This binding requires monovalent cations (sodium, potassium, or lithium); is blocked by purified (Na+,K')-ATPase; is inhibited noncompetitively by A T P (Knappl = 0.5 mM); and is unaffected by ouabain. 9-A5 inhibits ATP-stimulated (Na+,K+)-ATPase noncompetitively by blocking sodium-dependent phosphorylation of (Y subunits of liver or kidney membrane (Na+,K')-ATPase, Twelve h after 67% hepatectomy, maximal lZ5I-9-A5 binding to plasma membranes from regenerating liver falls 30 f 7% compared to sham-operated controls (p < 0.01). In contrast, (Na+,K+)-ATPase activity in regenerating liver membranes rises 58 f 12% compared to controls (p < 0.03). Similar experiments with particulate fractions from regenerating liver show insignificant decreases in maximal lZ5I-9-A5 binding (22 f 12%) but large increases in (Na',K*)-ATPase activity (325 f 14%) compared to controls (p < 0.001). No differences among groups are seen in KO values for 9-A5 binding or in the activities of plasma membrane 5'-nucleotidase (EC 3.1. 3.5). Thus, stimulation of the sodium pump during the late prereplicative phase of liver regeneration is not accompanied by increases in the numbers of (Na+,K')-ATPase sites. Instead, it appears that preexisting (Na+,K')-ATPases are activated specifically before DNA replication starts, Several events occurring at the cell surface are thought to * This work was supported by United States Public Health Service Grants AM 28215, AM 28392, and GM 07752 and the American Liver 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 "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
To whom correspondence should be addressed.
regulate the initiation of animal cell proliferation. Some of these include accelerated rates of amiloride-sensitive sodium (2, 3) and ouabain-sensitive *6Rb+ influx (3-7), membrane potential hyperpolarization (8,9), and the stimulation of Na+dependent amino acid transport (2,10,11). All are influenced by asymmetric sodium and potassium gradients across the plasma membrane. These gradients are generated, in part, by (Na+,K+)-ATPase (EC 3.6.1.3) (12). To understand how sodium pump-mediated changes in cation transport rates might influence the initiation of DNA replication, particularly during liver regeneration, it is important to learn how hepatic (Na',K+)-ATPase activity is regulated during the prereplicative phase of the "cell cycle." Previous investigations with isolated (7, 13,14) and cultured (3) hepatocytes, and regenerating liver (9,(15)(16)(17), suggest that ouabain-sensitive %Rb+ influx (a reflection of sodium pumpdependent potassium influx (18)) and (Na+,K')-ATPase activity rise during transitions from quiescent to proliferating states. Since (Na+,K+)-ATPase sites were not quantitated in these studies, the molecular basis for such increases in pump activity, for example, increases in the numbers of sodium pump sites as opposed to increases in enzyme turnover number, remains unknown. Various approaches have been employed to quantitate (Na',K+)-ATPase sites. In binding studies unrelated to mitogenesis, [3H]ouabain has been used since it inhibits sodium pump activity by binding to the (Na+,K+)-ATPase catalytic a subunit (19). Detailed binding experiments are difficult to make in mitogen-related work, however, since animal cell growth control systems often consist of rodent cells whose sodium pump activity is relatively insensitive to ouabain ( K I [~~~~ = 0.2 mM; Ref. 20). In addition, studies with [3H] ouabain frequently generate curvilinear Scatchard plots (21,22) and are hampered by high nonspecific binding (23), problems that complicate interpretation of results. Another approach, based upon the quantitation of ouabain-dependent uptake of 32Pi into a subunits, has several advantages (22) but potentially underestimates the numbers of (Na+,K+)-ATPase sites, since it requires the detection of alkali labile acylphosphate bonds after gel electrophoresis (22). Many of these problems might be circumvented with a new (Na+,K')-AT-Pase specific ligand.
In this report, the binding characteristics and anticatalytic properties of such a ligand, a high-affinity mouse monoclonal antibody (9-A5) directed against the CY subunit of rat (Na+,K+)-ATPase (l), are described. By using 9-A5 as a specific ligand to (Na+,K+)-ATPase, we find that sodium pump activation in regenerating liver occurs without detectable increases in the numbers of (Na+,K+)-ATPase sites.
Radiolubeling of Antibody 9-A5-9-A5 was labeled either enzymatically with NalZ51 or metabolically with [35S]methionine. Enzymatic radioiodination' of 9-A5 was performed with lactoperoxidase and glucose oxidase (25). All buffers and reagents used for iodination reactions and lZ5I-9-A5 purification were passed through 0.45-pm Nalgene@ type S filters (Rochester, NY). Labeled antibodies were separated from other reaction mixture components by Sephadex G-100 chromatography at 21 "C in a buffer containing 50 mM sodium phosphate (adjusted to pH 7.5 by mixing mono-and dibasic sodium phosphate in appropriate ratios) and 0.1% (w/v) gelatin. All phosphate buffers made subsequently were adjusted to the indicated pH values in a similar manner. The pooled void volume containing ' ' ' Ilabeled antibodies was supplemented with 2% (w/v) bovine serum albumin and stored in 1-ml aliquots at -70 'C. The resulting '=I-9-A5 had a specific activity of 15-18 pCi.pg of proteins" (5.7 X lo3 cpm/fmol of 9-A5) and was 295% precipitable with ice-cold 10% (w/ v) trichloroacetic acid. The radiolabeled antibody was diluted to a specific activity of 30-90 cpm/fmol of lZ5I-9-A5 in the binding assays described below. Autoradiographs of stock solutions of lZ5I-9-A5, analyzed by SDS-PAGE (26) in the presence of 1% (v/v) p-mercaptoethanol, revealed two bands of M , = 31,000 and 55,000 (Ref. 1).
male Fischer/344 rats (250-300 g) were perfused via the portal vein with 20 ml of 0.15 M sodium chloride (10 ml.min-') at 37 "C. The remaining steps were performed at 4 "C. The livers were excised and minced in histidine/sucrose buffer. Minced tissue was homogenized with 10 up and down strokes of a loose fitting Dounce homogenizer (clearance = 0.1 mm) in a 5-fold volume of buffer. The homogenate was filtered through four layers of buffer soaked gauze (U.S.P. type VII, Johnson and Johnson, New Brunswick, NJ) and centrifuged at 280 X g for 10 min to remove nuclei and intact cells. The supernatant fraction was retained. The pellet was rehomogenized and centrifuged to obtain a second supernatant. The two supernatant fractions were pooled and centrifuged at 40,000 X g for 30 min. The final pellets were resuspended (at 15 mg of proteins.ml") and stored at -70 "C. Ouabain-sensitive ATPase activity in these preparations was stable for at least 8 weeks.
Hepatocytes from 13-day-old quiescent primary cultures were generated by standard procedures (30). The viability, yields, and differentiated functions of such cell preparations have been described elsewhere (30,31). Particulate fractions of cultured cells were used instead of freshly isolated cells (to avoid artifactual proteolysis of membrane proteins) and were prepared as follows. Culture media were aspirated from the dishes and the monolayers were washed twice at 4 "C with 2 ml of Rat Ringer's buffer (140 mM sodium chloride, 5 mM potassium chloride, 1 mM magnesium chloride, 1.3 mM sodium phosphate (monobasic), 10 mM sodium bicarbonate, 10 mM D-glucose, and 2 mM calcium chloride (pH 7.4)). Ice-cold histidine/sucrose buffer was added (1 ml.culture") and the cells were removed from the 35mm dishes with a rubber spatula. The disrupted cells were homogenized and crude particulate fractions prepared exactly as described above for liver tissue. Pellets obtained after the last centrifugation step were resuspended (at 4 mg of proteins.ml") and stored at Enzyme Activity Determinations-(Na+,K+)-ATPase activity was measured in various fractions (30 pg of proteins. ml-', or as noted) with an NADH-coupled enzyme assay (24) employing NADH, pyruvate kinase, lactate dehydrogenase, and unless otherwise noted, 3 mM ATP, 100 mM sodium chloride, 20 mM potassium chloride, and 5 mM magnesium chloride. Linear initial rates of ATP hydrolysis were monitored for 3-4 min by following decreases in with a Beckman DU-8 spectrophotometer equipped with a Kinetics I1 Compuset@.
5"Nucleotidase activity was measured in liver membrane or crude particulate fractions (2-20 pg of proteins.ml-') at 37 "C for 30 min in a buffer containing 50 mM Tris.HC1 (pH 7.5), 10 mM MgClz, and 5 mM AMP. The reactions were terminated by the addition of icecold 10% (w/v) trichloroacetic acid. Samples were centrifuged at 2000 x g for 10 min at 4 "C. Chen's (32) reagent was added and the nanomole of inorganic phosphate released per mg-'.min" into the supernatant were quantitated by readings at ASZO,,~.
ATP and Protein Measurements-Liver membranes (50-100 pg of proteins) were precipitated with 0.6 N HClOl and centrifuged at 2000 X g for 10 min at 4 "C. The supernatants were adjusted to pH 7.0 with 6 N KOH, and their ATP contents were measured with luciferinluciferase (33). Proteins were measured by the method of Lowry (341, using bovine serum albumin as a standard. Sodium-dependent Phosphorylation of (Na+,K+)-ATPase-Liver or kidney (Na+,K+)-ATPase was phosphorylated in reaction mixtures of 200 pl containing (per ml) 200 pCi of [Y-~'P]ATP (3000 Ci/mmol), 30 p~ Tris. ATP, 100 pg of proteins, and either (a) "sodium" (100 mM NaC1,5 mM MgC12, 50 mM sodium phosphate) buffer (pH 7.5) or (b) "potassium" (100 mM KCl, 50 mM potassium phosphate) buffer (pH 7.5). When noted, reaction mixtures were preincubated for 1 h at 37 "C with either 9-A5 (200 pg/ml-') or ouabain (1 mM). Phosphorylation reactions were initiated with [y3'P]ATP and incubated at 4 "C for 20 s (steady-state conditions). Reactions were terminated with 400 pl of ice-cold trichloroacetic acid to a final concentration of 10% (w/v). The phosphoprotein-containing samples were centrifuged at 2000 X g for 10 min at 4 "C, washed with ice-cold 10% (w/v) trichloroacetic acid, resuspended in electrophoresis sample buffer at 4 "C (see below), and immediately analyzed by SDS-PAGE under acidic conditions. Acidic 10% polyacrylamide gels were prepared as described elsewhere (35). The sample buffer contained 120 mM Tris'HCl adjusted to pH 2.9 with citric acid, 1% (w/v) SDS, 1% (v/v) 0-mercaptoethanol, and 4% (v/v) glycerol. Since acidic polyacrylamide gels adhere to glass, the electrophoresis plates were lined with Saran Wrap" @ow, Indianapolis, IN) to facilitate removal of the gels. Electrophoresis was performed at constant current (35 mA) for 6-7 h at 4 "C. The gels were stained with Coomassie Blue, dried, and subjected to autoradiography for 1-2 days at -70 "C. The developed films were scanned under white light with a densitometer to determine relative proportions of phosphoprotein species (1).
Measurements of Radiolubeled 9-A5 Binding to Liver Membrane or Partieuhte Fractions-Unless stated otherwise, standard binding reactions were performed as follows. To measure total T-9-A5 binding, varying concentrations of '%1-9-A5 (0.10-30.0 nM) were added to one set of polystyrene tubes (12 X 75 mm) in buffer containing 100 mM NaCl, 20 mM KC1, 20 mM Tris.HC1, 1 mM EDTA, adjusted to pH 7.5 (with 1 N HCI), and 2% (w/v) bovine serum albumin. Under these conditions, adsorption of radioligand to the tubes was negligible. To measure nonspecific ImI-9-A5 binding, a 100-fold excess of unlabeled 9-A5 was added to a second set of identical tubes (both sets of tubes were prepared in duplicate). Binding reactions were initiated by adding either purified liver membranes (10 pg of proteins. tube") or crude liver particulate fractions (100 pg of proteins. tube") to the mixtures (final total volume = 250 pl) and allowed to proceed for 90 min at 37 "C. The reactions were stopped by adding 2 ml of ice-cold binding buffer. Unbound ("free") '"1-9-A5 was separated from membrane-bound Iz6I-9-A5 by centrifugation at 2000 X g for 10 min at 4 "C. Unbound '%1-9-A5 (supernatant) was removed by aspiration. The reaction pellets were counted in a Searle Model 1196 y counter (efficiency for '%I = 90%).
fmol '"1-9-A5 bound/mg proteins total cpm bound/mg proteins -nonspecific cpm bound/mg proteins Under standard binding reaction conditions, nonspecific binding varied linearly from 8 to 30% of total counts/min bound at 0.10-30 nM '%1-9-A5, respectively. When centrifugation times were increased beyond 10 min, specific lz5I-9-A5 binding was unaltered. Free '"1-9-A5 concentrations, used for Scatchard analyses of binding data (36), were calculated from the difference of total counts/min addedtotal counts/min bound after 90 min. Similar methods and data analyses were used for experiments employing [35S]9-A5. Surgical Induction of Liver Regeneration-Male Fischer/344 rats (250-300 g) maintained on 12 h light (700-1900) and dark (1900-700) cycles, were sham or 67% hepatectomized under ether anesthesia between 800 and 1000 h (time t = to, Ref. 37). Water was supplied ad libitum without food after surgery. In some instances, see "Results", groups of animals were etherized at t = to (no surgery, anesthesia control) or t = tlz (no surgery, 12-h diurnal variation control). For all groups, the livers were perfused (as above) and the caudate and accessory lobes were removed 12 h postoperatively ( t = 2000-2200 h). Plasma membranes or crude particulate fractions were prepared as described above.

Effects of (Na+,K+)-ATPase Preparations on lZ5I-9-A5 Binding to Liver
Membranes-Two different kinds of highly purified preparations of rat kidney (Na+,K+)-ATPase blocked lZ5I-9-A5 specific binding to membranes (Table f). These effects were obtained by either ( a ) preincubating '"I-9-A5 with excess particulate kidney (Na+,K+)-ATPase, followed by measuring the residual unbound antibody available for binding to liver membranes, or ( b ) directly competing for lZ5I-9-A5 binding to membranes by adding ClzEs-solubilized (Na+,K+)-ATPase to reaction mixtures. Control experiments indicated that lZ5I-9-A5 binding was unaffected by &E8. Fig. 1 shows the amount of specific "'1-9-A5 binding as a function of membrane protein concentration. Binding was directly proportional to protein concentrations between 4 and 40 pg of proteins-tube" (slope = 1.0, p < 0.05). Nonlinear binding occurred above 60 pg of proteins. Effects of purified (Na+,K+)-ATPase preparations on specific lmI-9-A5 binding to liver plasm membranes Binding assays were performed under equilibrium conditions for 90 min at 37 "C in the presence of 100 mM sodium chloride and 20 mM potassium chloride. lz6I-9-A5 and liver membranes were present at 2 nM and 40 pg of proteins. ml-', respectively. Under these conditions, 20 pmol of '251-9-A5.mg of proteins" represented maximal specific binding (100%). Effects of membrane-bound purified kidney (Na+,K+)-ATPase were studied by preincubation of excess kidney enzyme for 1 h at 25 "C with 2 nM '"1-9-A5 followed by centrifugation at 4 "C for 10 min at 12,000 X g to remove antibody-enzyme complexes from the supernatant. The supernatant was then assayed for "'1-9-A5 available for binding by incubation with liver membranes under standard binding conditions. Kidney (Na+,K+)-ATPase, solubilized with ClzEa to achieve maximal enzyme activity (1, 28), was added directly to liver membrane binding reaction mixtures. The data are average values of three separate experiments (S.E. 510% in all cases). Associution Kinetics- Fig. 2A shows the kinetics of association of 1.5 nM lz5I-9-A5 with specific membrane sites at 37 "C. The association was time-dependent and saturable. Half-maximal specific binding occurred at 5 min; apparent equilibrium occurred within 45 min. The data were plotted according to the equation

9-A5
Activation of (Nai,K+)-ATPase during Rat Liver Regeneration Kinetic studies of specific I2'I-9-A5 binding to rat liver plasma membranes. A , time course of association of Iz5I-9-A5. Membranes (10 pg of proteins. tube") were incubated a t 37 "C with 1.5 nM Iz5I-9-A5 (73 cpm/fmol) in the absence (total binding) or presence ("nonspecific" binding) of 200 nM unlabeled 9-A5. Aliquots (250 pl) were removed at the indicated times and the amount of specifically bound radioligand (total counts/min nonspecific counts/min) was determined (0). The data are average values of three separate experiments (S.E. 5 8% in all cases). Iz51-9-A5 specifically bound to membranes that were preincubated in binding buffer for 90 min a t 37 "C (to see if binding site degradation occurred) (0). Inset, semilogarithmic transformation of the data (O), fit to a linear curve were centrifuged, washed, and resuspended in 10 ml of assay buffer containing 200 nM unlabeled 9-A5 at 37 "C (time to = B,). The dissociation of bound "'1-9-A5 was followed at the indicated times by removing 250.~1 aliquots (5 pg of proteins) from the two sets of samples and washing as described under "Materials and Methods." Specifically bound radioligand (total counts/minnonspecific counts/min) was plotted as a function of time after dilution and addition of unlabeled 9-A5. The data are average values of three separate experiments (S.E. 5 10% in all cases). Inset, semilogarithmic transformation of the data, fit to a linear curve (p < 0.04) described by the equation ln(B,/B,) = k,, was made to calculate kz (slope; see "Results").
where Be, and B, are the amounts of specific ligand bound at equilibrium and at time = t , respectively; k, and k2 are association and dissociation rate constants, and ['2sI-9-A5] is the ligand concentration under the experimental conditions (38). A graph of ln(B,,/[B,, -Bt]) versus time gave a linear curve with slope (a pseudo first-order observed association rate constant, kOb) = 2.85 X 10-3.s-1 (Fig. 2 A , inset). Since low concentrations of antibody were used sometimes (e.g. 0.1 nM), 90-min incubation periods were chosen for routine binding conditions to achieve apparent equilibrium.
Dissociation Kinetics- Fig. 2B shows the kinetics of dissociation of specifically bound lZ5I-9-A5 from membranes, in the presence of 200 nM unlabeled 9-A5, following 90 min preincubation with 1.5 nM lZ5I-9-A5 a t 37 "C. Half of the radioligand dissociated within 30 min. Almost full dissociation oc-curred within 120 min, indicating reversibility of the reaction. Based on the equation a plot of ln(B,/B,) uersus time gave a linear curve with slope (a dissociation rate constant, k2) = 2.7 X 10-4.s" (Fig. 2B, inset).
Equilibrium Binding of l2'1-9-A5 to Liver Membranes- Fig.  3 shows the amount of specific lZ5I-9-A5 binding at equilibrium as a function of I2'I-9-A5 concentration. The binding curve was hyperbolic, suggesting that the radioligand bound to a single class of sites. This interpretation was supported by Scatchard analysis (Fig. 3, inset), which generated a linear curve from which an apparent dissociation constant (KD[app,) of 0.64 nM was derived. Maximal Iz5I-9-A5 binding (Bmax) was 29.3 pmol of '251-9-A5 bound per mg of proteins (Fig. 3, inset). Using a different liver plasma membrane sample derived from caudate and accessory lobes (as opposed to the whole liver), similar experiments with [35S]9-A5, also shown in Fig. 3, revealed a similar KD[appl (0.89 nM) but a decreased Bmax (13 pmol of lZ5I-9-A5 bound per mg of proteins). The decreased B,,, in this latter preparation was consistent with its lower (Na',K')-ATPase activity, and was seen when either [35S]9-A5 or 12'I-9-A5 were used to quantify (Na',K')-ATPase bind- ing sites (see below). The equality of [35S]9-A5 and lZ5I-9-A5 dissociation constants suggests that radioiodination of 9-A5 had little or no effect on its interaction with (Na+,K+)-AT-Pase.
Stability of lZ5I-9-A5 and Plasma Membrane (Na+,K+)-AT-Pase during the Binding Reaction-To see if lZ5I-9-A5 degradation occurred during the binding reaction, samples containing lZ5I-9-A5 were withdrawn at the beginning (t = 0) and end ( t = 90 min) and analyzed by SDS-PAGE. Autoradiographs of electrophoresed radioligands showed only bands corresponding to the light and heavy chain of 9-A5 (Mr = 31,000 and 55,000). No lZ51-labeled proteolytic fragments were detected. Similar results were seen on autoradiographs from gels containing lZ5I-9-A5 that had been allowed to bind, and then eluted from membranes with 0.1 M glycine, adjusted to pH 2.8 with 6 N HC1 (data not shown).
Evidence of binding site stability during the reaction was obtained from two observations. First, liver membranes preincubated for 90 min at 37 "C were capable of binding similar amounts of lZ51-9-A5 in subsequent rounds of binding (Fig.  2 A ) . Second, (Na+,K+)-ATPase activity was only slightly reduced (15%) after the 90 min preincubation a t 37 "C (data not shown).
In contrast to these results, monovalent cation-dependent 12'1-9-A5 binding was blocked 99% by 3 mM ATP (Table 11). Fig. 4A shows that this effect was ATP-concentration dependent, with an apparent KI of 0.5 mM (= Km~ATpl for enzyme activity (see Fig. 5)). Fig. 4B shows the concentration dependence of '251-9-A5 binding to membranes in the absence or TABLE I1 Effects of (Na+,K+)-ATPase substrates and inhibitors on specific l2'1-9-A5 binding to liver p l a s m membranes Binding assays were performed as described in Table I (20 pmol of lZ51-9-A5. mg of proteins". 90 min" = maximal specific binding = 100%). Compounds tested were present throughout the 0-90-min reaction.   FIG. 4. Effects of ATP on specific "'1-9-A5 binding to liver plasma membranes. A, ATP concentration dependence of inhibition of lZ5I-9-A5 binding. Membranes (20 pg of proteins. tube") were incubated with lZ5I-9-A5 (2 nM; 73 cpm/fmol) for 90 min at 37 "C in the presence of the indicated concentrations of ATP plus an ATPregenerating system (see Table 11). The amounts of specifically bound lZ5I-9-A5 (fmol. mg of proteins") were calculated from the differences in bound radioactivity in the absence or presence of 200 nM unlabeled presence of 0.6 mM ATP. As expected (see Fig. 3), both binding curves were hyperbolic. The E,,, values obtained in the absence or presence of ATP were 29 or 9 pmol of 1251-9-A5 of proteins, respectively. For both curves, Scatchard analyses gave KD~appl values = 0.9 nM (data not shown).
The inhibitory effects of ATP were not due to degradation products like adenosine or AMP, or to ATP-regenerating system components (phosphoenolpyruvate or pyruvate kinase) since none of these molecules inhibited specific binding (Table 11). However, ADP partially blocked lZ5I-9-A5 binding (51% at 3 mM ADP). This effect was not due to ATP contamination of the ADP preparation or to ATP contamination of the membranes (which contained less than 1.0 nmol of ATP/ mg of proteins). The addition of 25 mM inorganic phosphate reduced the inhibitory effect of 3.0 mM ATP on antibody binding by 30% but had no effect on ADP inhibition of binding. These observations are consistent with the ability of inorganic phosphate to bind to (Na+,K+)-ATPase and block ATP but not ADP binding (40).
Effects of 9-A5 on Sodium-dependent Phosphorylation of Liver and Kidney (Na+,K+)-ATPase- Fig. 6 A shows sodiumdependent phosphorylation of kidney (Na+,K+)-ATPase as revealed by autoradiographs of acidic SDS-polyacrylamide gels. When 100 mM sodium chloride was present, two "P-  were seen (lane 1 ) . The M , = 110,000 phosphoprotein predominated, as judged by densitometric scanning (= 50% of total absorbance; data not shown). When the membranes were preincubated with 9-A5, in the presence of sodium, significant decreases were seen (greater than 90%) in the phosphorylation of the M , = 110,000 protein and, to a lesser extent, the ' The use of kidney enzyme in these studies was indicated because of its high specific activity. In addition, no structural differences between kidney and liver cy subunits (purified by affinity chromatography over 9-A5-Sepharose columns) have been detected by peptide mapping. 12,000 phosphoprotein is unclear but it may have arisen from a subunit degradation (41). As seen with the kidney enzyme ( Fig. 6A), phosphorylation of the M, = 110,000 liver protein, which co-migrated with (Na+,K+)-ATPase a subunits, was blocked by potassium chloride (Fig. 6B, lane 3). These results show that 9-A5 inhibited sodium-dependent phosphorylation of the liver membrane (Na+,K+)-ATPase a subunit.
Comparisons between lz5Z-9-A5 Binding to Liver Membranes and Crude Particulate Fractions-To exclude the possibility that artifacts of liver membrane isolation (e.g. preferential sodium pump losses or recoveries) might influence binding of '*'1-9-A5 to (Na+,K+)-ATPase, BmnX and KDIappl values obtained with liver plasma membrane and total crude particulate fractions (whole liver or cultured hepatocytes) were compared. The results are given in Table 111.
Based upon these results, and conversion factors for proportions of liver membrane and particulate proteins that constitute hepatocyte plasma membrane proteins (see legend to Table HI), the numbers of (Na+,K')-ATPase sites.hepatocyte" were estimated to be 264,000 (plasma membranes), 217,000 (tissue crude particulate fractions), and 232,000 (cultured hepatocyte crude particulate fractions).

Quantitation of (Na+,K+)-ATPase Activity and Binding Sites in Tissue Preparations from Normal and Regenerating Rat Liver
Plasma Membrane Preparations- Fig. 7 A shows that 12 h after surgery, 9-A5-sensitive VmaX ATPase activity rose 58 f 12% in membranes from regenerating liver, compared to Similar enrichment ratios of enzyme activity and Bmax values have not been determined for crude particulate fractions prepared from 12-day-old cultured hepatocytes (whose specific enzyme activities were consistently 4-5-fold higher than adult liver tissue levels) because of technical difficulties in preparing large enough quantities of plasma membranes from such samples. mM Hepes (pH 7.5)). ATPase activity (left y axis) was determined with an NADH-coupled ATPase assay. The data are the averages (n = 3 ? S.E.) of the inhibitor-sensitive ATPase activities (nanomole of ATP hydrolyzed per mg" .min"). 5'-Nucleotidase activities (right y axis) were determined in the same membrane fractions (n = 3 f S.E.) as described under "Materials and Methods." B, maximal lz5I-9-A5 binding. Identical membrane preparations (above) were used to quantitate the numbers of (Na+,K+)-ATPase sites (Bum values) obtained from Scatchard plots (as described in the legend to Fig. 3). Average membranes from sham-operated rats (138 nmol of ATP hydrolyzed per mg"-min"; p < 0.03). Identical results were obtained for increases in ouabain-sensitive ATPase activity. These results confirmed and extended previous reports (15)(16)(17) and, as others observed, reflected specific membrane enzyme changes since, compared to controls, regenerating liver membranes showed only slight insignificant increases (12%) in the activity of another plasma membrane enzyme, 5'nucleotidase (Fig. 7A).

FIG. 8. Quantitation of (Na+,K+)-ATPase activity and bind-
ing sites in crude particulate fractions of normal and regenerating rat liver. Rats were operated upon ( n = 3.group"), killed 12 h later, and pooled particulate fractions prepared as described under "Materials and Methods." A, (Na+,K+)-ATPase and 5'-nucleotidase activities. Conditions were identical to those described in the legend to Fig.7A (sham, open bars; 67% hepatectomy, shaded bars) except that 25 pg of proteins.tube-' were used for enzyme assays (final volume = 1 ml). B, maximal lZ5I-9-A5 binding. Conditions were identical to those described in the legend to Fig. 7B (sham, open bars or 0 , 67% hepatectomy, shaded bars or 0 ) except that 100 pg of proteins. tube" were used for Scatchard analyses. proteins; p < 0.01). However, no differences between the two groups were seen with respect to &[app] values (= 0.9 nM; see Fig. 7B). 6 Other variables like surgical stress (anesthesia or blood loss) or diurnal variation could not account for these observations, since results essentially identical to sham-operated controls (with respect to enzyme activity and maximal 9 -9 -A5 binding) were obtained with membranes prepared from rats etherized (no previous surgery) at t = to or t = tI2 (data not shown). The possibility that soluble activators and/or inhibitors of (Na+,K+)-ATPase activity and lz51-9-A5 binding might have affected the results seems unlikely from findings that enzyme activities and Elmax values were directly proportional to the ratios of hepatectomized uersus sham membranes in appropriate mixing and serial dilution experiments (data not shown).
Crude Particulate Fractions-To exclude the possibility that the results observed in Fig. 7 were due to differential The reasons for the absolute decreases in B, , , values obtained in control tissue preparations in experiments shown in Fig. 7B, compared to Bma= values shown in Fig. 3, are unclear. Similar differences were found in membrane preparations from anesthesia and diurnal variation controls. Contributing factors might be intrinsic binding site concentration differences with respect to lobular structure (e.g. caudate plus accessory lobes (Fig. 7B) uersm whole liver (Fig. 3)). membrane (Na+,K+)-ATPase recovery and/or stability, similar experiments were performed with crude particulate fractions of liver caudate and accessory lobes. Particulate fractions were chosen since their yields of (Na',K+)-ATPase activity are greater than 80% (where activity of starting homogenates = loo%), compared to 1-16% yields for liver plasma membranes (29,42). The results are shown in Fig. 8. The same trends were found 1) (Na+,K+)-ATPase activity in regenerating tissue was elevated (325 f 14%) compared to sham-operated rats (9.0 nmol of ATP hydrolyzed per mg" .

Biphasic increases in
sodium pump activities are implicated in the repertoire of "early" and "late" prereplicative events that initiate animal cell proliferation (2,3). Mitogen-stimulated passive sodium influxes account for much of this early pump activation (2, 3, 5-7). Less is known about mechanisms which regulate the later changes. In an attempt to understand their molecular basis, quantitative measurements of (Na',K+)-ATPase activity and numbers of (Na',K+)-AT-Pase sites have been made in regenerating liver, 12 h after partial hepatectomy, with an anticatalytic mouse monoclonal antibody (9-A5) directed against the sodium pump (1). The results of these experiments, using plasma membranes or crude particulate fractions from normal and regenerating liver, show that (Na+,K+)-ATPase is stimulated in regenerating liver without detectable increases in the numbers of 9-A5 binding sites. This indicates that late prereplicative increases in (Na+,K+)-ATPase activity are due, directly and/or indirectly, to specific activation of pre-existing pump molecules. This activation is apparently manifested by increases in (Na+,K+)-ATPase turnover number.
To validate the measurements, initial studies were designed to elucidate the specificity and nature of interactions between 9-A5 and its putative (Na+,K+)-ATPase binding sites in native membranes. Six kinds of investigations indicated that 9-A5 binding, which is time-dependent, saturable, and reversible, occurs with absolute specificity for a single class of (Na+,K+)-ATPase binding sites.
First, purified (Na+,K+)-ATPase, from solubilized or intact kidney membrane preparations, blocked '251-9-A5 specific binding to liver membranes. These results are consistent with previous observations that 9-A5: (a) immunoblots specifically to a subunits of (Na+,K')-ATPase, (b)directly inhibits liver membrane (Na+,K+)-ATPase V,,, activity ( K~l~~~l = 30 nM), and ( c ) localizes to hepatocyte bile canalicular (1) and kidney epithelium basolateral membranes as revealed by immunofluorescent ~t a i n i n g .~ Second, the kinetics of association and dissociation of radioligand binding to liver membranes were pseudo first-order and zero-order, respectively, with a low K D [~~~~ (0.15 nM). Third, Scatchard plots from equilibrium binding studies, made over a broad Iz5I-9-A5 concentration range (0.10-100 nM), showed strictly linear curves. In both cases, this behavior is expected for ligand interactions with homogeneous classes of binding sites (36). These results are unlikely to have been affected by 9-A5 or (Na+,K+)-ATPase binding site degradation because (a) autoradiographs showed that Iz5I-9-A5 (analyzed by SDS-PAGE) was intact after prolonged incubations in binding reaction mixtures, and (b) (Na',K')-ATPase activity levels and (Na+,K+)-ATPase binding site numbers remained constant throughout the binding reactions. In addition, the binding properties of lZ5I-9-A5 were probably equivalent to unlabeled 9-A5, inasmuch as [35S]9-A5 showed identical properties with respect to its K D and Bmax values. Although the KD from Scatchard analysis (0.64 nM) was higher than the KDlappl from nonequilibrium kinetic data (kz/kl = 0.15 nM), this "discrepancy" has been seen in other antibody binding studies (43) and discussed elsewhere (44).
Fourth, (Na+,K')-ATPase substrates significantly influenced lZ5I-9-A5 binding to liver membranes. Monovalent cations that bind to the sodium pump (sodium, lithium, or potassium) (18) were required for maximal 9-A5 binding. Antibody binding did not seem to be affected by different enzyme conformations, since either sodium or potassium supported maximal binding yet each cation stabilizes distinct conformations (sodium = El, potassium = E,; Ref. 40). The cation requirement was specific, however, since choline+ did not replace the cations tested. In contrast to these stimulatory effects, ATP, a substrate required for active potassium and sodium transport (45), blocked (K,[appl = 0.5 mM) ion-dependent "'II-9-A5 binding (as did ADP, although less potently). Since ATP affected only lZ5I-9-A5 maximal binding and not its apparent affinity this blockade was noncompetitive. The large positive Hill coefficient seen for inhibition of lZ5I-9-A5 binding by ATP (nH = 3.2; Fig. 4A, inset) suggests that several ATP molecules were involved in this interaction. These findings are consistent with 9-A5's ability to block (Na+,K+)-ATPase activity in the presence of ATP, since anticatalytic studies were performed with enzyme preparations preincubated with 9-A5 (before adding ATP) and negligible amounts of antibody dissociated from 9-A5. enzyme complexes under these conditions.' In addition, both kinetic constants (Kllappl; n H ) were similar to those obtained in the presence of potassium for low-affinity ATP binding sites on (Na+,K')-ATPase (46).
These kinetic observations (suggesting cooperative binding of ATP to the enzyme) and other results (suggesting (Na+,K+)-ATPase activation as an hyperbolic function of ATP concentration (Fig. 5 ) ) have been reported by previous investigators (18,46). One interpretation of these findings in light of the effects of ATP on lZ5I-9-A5 binding to the purified enzyme ( Fig. 4B and Table 11) is that ATP and lZ5I-9-A5 bind to distinct sites on different enzyme intermediates. This interpretation is supported (in a fifth kind of investigation) by preliminary findings that 9-A5 behaved like a noncompetitive inhibitor of ATP-dependent stimulation of (Na+,K+)-AT-Pase, as revealed by Lineweaver-Burk analyses (Fig. 5 ) . However, further kinetic studies are needed to substantiate this conclusion since simple noncompetitive inhibition can arise from removal from the incubation system of active enzyme species and, therefore, is not necessarily indicative of "distinct" sites. In addition, the interpretation is further complicated because (i) there is some uncertainty in the K, and V,,, values, since appropriate weighting factors (47) were not used to make these calculations from the curves (Fig. 5), and (ii) the fact that ATP inhibits Iz51-9-A5 binding to the enzyme in a cooperative manner. Sixth, direct biochemical investigations of its anticatalytic properties showed that 9-A5 specifically blocked sodium-dependent phosphorylation of a subunits of (Na+,K+)-ATPase (in contrast, ouabain stabilized phosphoenzyme intermediates). Notably, with regard to their M , after acidic SDS-PAGE, their ion requirements of phosphorylation, and their response to 9-A5, phosphorylated CY subunits were indistinguishable in liver and kidney preparations. These and other results3 suggest a high degree of structural homology between a subunits from these different tissues; and, they validate the assumption (see e.g. Fig. 5) that the mechanisms of inhibitory effects of 9-A5 on kidney (Na+,K+)-ATPase are similar with respect to liver (Na',K')-ATPase.
The ability to quantitate the numbers of (Na',K')-ATPase sites in hepatic tissue extracts provides a way to estimate the numbers of sodium pumps. hepatocyte". Calculations along these lines yield -238,000 pumps.cel1-l. It is reasonable to equate 9-A5 binding sites with sodium pumps ( i e . at least, cup dimers (41)) since 9-A5 binds specifically to a subunits; electrogenic ion pumping by this protein requires both catalytic functions that 9-A5 inhibits (ATPase activity and sodiumdependent phosphorylation of the CY subunit); and, the numbers of 9-A5 binding sites, in any single tissue preparation examined, are strictly proportional to 9-A5 (and ouabain)sensitive (Na',K+)-ATPase activity." Furthermore, 9-A5 also (a) inhibits ATP-driven cation fluxes across membranes of everted heart vesicles: and (b) co-localizes with "hybridoma 24" (a mouse monoclonal antibody directed against /3 subunits of chicken (Na+,K+)-ATPase (48)) at basolateral membrane sites in chicken kidney tubular epithelial cells, as revealed by immunofluorescent light microscopic studies." The intermediate value we estimated (2232,000 sites. hepatocyte") is calculated from crude particulate tissue obtained from cultured hepatocytes, where conversion factors for cell numbers and crude particulate total proteins are experimentally determined. This value agrees with calculated values for liver tissue crude particulate (-217,000 sites. hepatocyte") and plasma membrane fractions (-264,000 sites. hepatocyte"), indicating that the assumptions made in the latter calculations are valid (see Table 111). These data imply that most hepatocyte sodium pumps detected by 9-A5 are membrane bound, since the number of membrane sodium pumps (264,000 sites. hepatocyte") are within experimental error of the numbers of total cellular sodium pumps (217,000-232,000).
If it is assumed that 9-A5 binding reflects sodium pump concentrations in intact hepatocytes, whose average surface area is -3,000 pm' (49), then a first approximation of pump site density yields -80 sodium pumpslpm' of plasma membrane. This estimate may be low, because sodium pumps are distributed nonrandomly within hepatocyte membranes (I), as observed in other transporting epithelia (41). Immunocytochemical work at both light microscopic (1) and ultrastructural levels7 indicates that hepatocyte bile canalicular membranes contain larger numbers of sodium pumps than basolateral (sinusoidal) membrane faces. Because this asymmetry may be and because canalicular membrane area is "13% of the total surface areas (50) , which might lead to underestimations of binding site numbers in studies with this ligand), the inhibition of (Na+,K+)-ATPase activity by 9-A5 does not require a phosphorylated enzyme intermediate (see Fig. 6). Thus, lZ5I-9-A5 has both practical and theoretical advantages over [3H]ouabain as a radioligand for (Na+,K+)-ATPase, especially for site quantitation studies in tissues with low levels of sodium pumps.
Several lines of evidence support the conclusion that, during liver regeneration, direct and/or indirect sodium pump activation accounts for late prereplicative phase increases in (Na+,K+)-ATPase. First, Scatchard analyses of binding of lZ5I-9-A5 to regenerating liver membranes show that a small, statistically significant decrease occurs in Bmax values, in comparison to sham-hepatectomized controls. The increases in (Na+,K+)-ATPase activity and concomitant decreases in the numbers of enzyme sites cannot be attributed to generalized nonspecific changes in membrane enzymes since 5'nucleotidase activities are similar in membrane preparations isolated from either 67% or sham-hepatectomized rats. In addition, the sites available for inhibition of enzyme activity by 9-A5 are the same population of sites that are quantitated in the binding studies (see above). Second, in agreement with other reports (15)(16)(17), the increases (258%) in ouabain-sensitive ATPase activity in regenerating liver membranes are quantitatively identical to the increases in (Na+,K')-ATPase activity seen when 9-A5 is used as the specific inhibitor. Since ouabain and 9-A5 bind to distinct (Na+,K+)-ATPase sites, which probably exist on contralateral membrane surfaces: the increases in enzyme activity must reflect intrinsic membrane alterations and not artifactual differences in inhibitor accessibility to (Na+,K+)-ATPase binding sites. Third, similar trends, increases in (Na',K+)-ATPase activity (325%) and small decreases in numbers of sodium pump sites (22%), are seen when crude particulate fractions are used as the enzyme source. These results exclude the possibility that the observed differences are due to membrane isolation artifacts (e.g. enrichment or exclusion of a subpopulation of enzyme molecules) since crude particulate fractions contain almost all (280%) of the enzymatic activity present in total rat liver homogenates. This latter point is important because the yields of plasma membrane and membrane (Na+,K+)-ATPase activity are typically 5 2 % and 516%, respectively (29,42). Underlying causes of the larger relative increases in regenerating liver (Na+,K+)-ATPase activity seen in crude particulate fractions compared to membranes (325 uersus 58%, respectively) are as yet unknown. Either endogenous (Na+,K+)-ATPase activators, present in crude particulate fractions, are lost during membrane isolation, or "activated" forms of (Na+,K+)-ATPase are labile under current isolation procedures. Further work is needed to distinguish between these or other alternatives.
Our studies and earlier ones (15)(16)(17) do not exclude the possibility that increases in (Na+,K+)-ATPase activity in regenerating liver tissues arise from alterations in the K,,, for one or more (Na+,K+)-ATPase substrate (ATP, sodium, or potassium). This possibility seems unlikely since in several cases where increases in (Na+,K+)-ATPase activity have been reported, without corresponding increases in the numbers of sodium pump "sites," no differences in substrate K,,, values were detected (21,54,59). The experiments needed to address this problem, however, are difficult to perform accurately, especially under sub-V,, conditions, because liver tissue preparations contain very low levels of enzyme activity. Once the hepatic enzyme is purified and concentrated? this problem should be fairly simple to resolve. The present experiments also do not directly exclude the possibility that the observed (Na+,K+)-ATPase activation, without corresponding increases in the numbers of enzyme sites, might be due to subtle changes in the sodium pump itself (see below), reflected by alterations in rate constants of binding of lZ5I-9-A5 to the enzyme. This seems unlikely because measurements of the apparent KO values (of lZ5I-9-A5 binding, at equilibrium), which reflect the ratio of these rate constants (kZ/kl), gave identical results with liver tissues from both sham-and 67%hepatectomized rats (Figs. 7 and 8). Moreover, these K D values were in excellent agreement with those determined directly (in normal tissues) from kinetic studies (Fig. 2).
Calculations of the apparent turnover number of hepatic (Na+,K+)-ATPase" show that it increases 4-fold, from a value of 150 ATP molecules hydrolyzed per lZ5I-9-A5 binding site". s-'. in crude particulate fractions from quiescent liver to 625 ATP molecules hydrolyzed per lZ5I-9-A5 binding site". s" in similar fractions from 67% hepatectomized liver (see Fig. 8). The findings of increased (Na+,K+)-ATPase enzyme turnover number during rat liver regeneration support the hypothesis that this might be a general mechanism responsible for increases in ouabain-sensitive ssRb+ influx, membrane hyperpolarization, and ATPase activity in other proliferating cells (60) midway to late in the prereplicative phase (6-14 h poststimulus).
The structural modifications that cause late G1 sodium pump activation are unknown. If the activation mechanism is direct, one possibility is that the enzyme is phosphorylated by a CAMP-dependent (37) or -independent protein kinase, or dephosphorylated by a specific phosphatase. Recent studies have shown that CAMP-elevating agents lead to increases in (Na+,K+)-ATPase activity (61, 62), and during transitions to growth arrest in cultured Friend erythroleukemia cells, the (Y subunit of (Na+,K+)-ATPase appears to be phosphorylated on a threonine residue as (Na+,K+)-ATPase activity subsides (63). Alternatively, if the activation mechanism is indirect, mitogen-dependent increases in membrane fluidity, which increase membrane (Na+,K+)-ATPase activity (17), might be involved. The latter possibility seems especially intriguing since decreases in plasma membrane cholesterol (that occur during liver regeneration (17)) and in lipid acyl chain order parameters have been shown to stimulate (Na+,K')-ATPase activity in a cholesterol-biosynthetic mutant of Chinese hamster ovary cells (64).