ATP and Acetyl Phosphate Induces Molecular Events near the ATP Binding Site and the Membrane Domain of Na+,K+-ATPase

The addition of ATP to Mg2+-Na+-bound-probe labeled Na+,K+-ATPase preparations containing ∼0.5 mol of pyridoxal 5′-diphospho-5′-adenosine (AP2PL) probe at Lys-480 and ∼0.9 mol of fluorescein 5′-isothiocyanate (FITC) probe at Lys-501 showed a decrease and an increase in the AP2PL fluorescence intensity with neither significant ATP-dependent phosphorylation nor FITC fluorescence change. The rate constants for the fluorescence change increased nearly linearly with increasing ATP concentrations. The substitution of AcP for ATP decreased the FITC fluorescence rather monophasically, 8.5/s, which was followed by the half-site phosphorylation with same amount of components with different rate constant, 7.2 and 4.6/s, followed by a much slower increase in the two components of AP2PL fluorescence, 1.4 and 0.2/s. The addition of Na+ with increasing concentrations of ATP to the K+-bound AP2PL-FITC enzymes induced accelerations in the decrease and an increase in the AP2PL fluorescence intensity with two different increases in the FITC fluorescence intensity, showing that the same concentration of ATP is capable of inducing four different fluorescence changes. The addition of ATP to the Mg2+-Na+-bound enzymes modified withN-[p-(2-benzimidazolyl)phenyl]-maleimide (BIPM) at Cys-964 and retaining full Na+,K+-ATPase activity induced two different increases in BIPM fluorescence intensity. Each rate constant for the BIPM fluorescence change versus concentrations of ATP gave two intersecting straight lines. These data and the stoichiometries of fluorescence probe bindings and ATP- and AcP-dependent phosphorylation provide strong support for the conclusion that the functional membrane-bound Na+,K+-ATPase is a tetramer.

The transport of sodium and potassium ions coupled with the hydrolysis of ATP is performed by Na ϩ ,K ϩ -ATPase (1-6), which shows high affinity ATP binding for phosphorylation in the presence of Mg 2ϩ and Na ϩ and low affinity binding for the deocclusion of K ϩ (2,4). To understand the mechanism of energy transduction for this enzyme, we have studied the ATP-induced conformational events for the enzymes, modified with various fluorescence probes (7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19). Recently, the half-site stoichiometry for phosphorylation by ATP for the case of Na ϩ ,K ϩ -ATPase, which is independent of specific activities, while retaining nearly full-site ATP binding capacity in the presence of CDTA 1 (18) has been unequivocally demonstrated (19). The data also showed that enzyme preparations that contained ϳ0.5 mol of pyridoxal 5Ј-diphospho-5Ј-adenosine (AP 2 PL) probe at Lys-480 of the ␣-chain reduced the phosphorylation stoichiometry to ϳ0. 25. When the AP 2 PL enzymes were further treated with a fluorescein 5Ј-isothiocyanate (FITC), the resulting AP 2 PL-FITC enzymes contained ϳ0.9 mol of the FITC probe at Lys-501, which reduced the value to ϳ0.03. The FITC treatment had a negligible effect on the rate and the extent of individual ATP-and AcP-dependent fluorescence changes due to the AP 2 PL chromophore and the amount of AcP-dependent half-site phosphorylation in the presence of Mg 2ϩ and Na ϩ (18,19). These data are consistent with the hypothesis that phosphorylation is not required for ATP-induced AP 2 PL fluorescence change of the AP 2 PL and AP 2 PL-FITC enzymes and that four ATP binding sites are present in Na ϩ ,K ϩ -ATPase. In this paper, ATP-or AcP-induced conformational events were followed by measuring fluorescence changes of the AP 2 PL probe at Lys-480 and the FITC probe at Lys-501 using the AP 2 PL-FITC enzymes. Since the N-(p- (2-benzimidazolyl)phenyl) maleimide (BIPM) treatment had little effect on both Na ϩ ,K ϩ -ATPase activity and the phosphorylation capacity (7), ATPinduced BIPM fluorescence changes at Cys-964 of the ␣-chain, which is near the transmembrane domain M9 of the enzyme (20), were also followed whether or not ATP induced multiple conformational changes in the BIPM enzyme.
The data indicate the occurrence of four ATP binding induced out of phase fluorescence changes not only near the ATP binding domain, but also in the membrane domain.

MATERIALS AND METHODS
Experimental methods were exactly as described in previous papers unless otherwise indicated (18,19).
Steady-state fluorescence measurements were performed using a Shimadzu RF-503 difference spectrofluorophotometer at 25°C (8) with a sample of 3.2 ml (50 g of protein/ml) under the same reaction conditions as for the Na ϩ -ATPase measurements as described in Fig.  1 legend, except in some experiments, ATP was replaced with AcP. The fluorescence excitation was at 320 (AP 2 PL) and 470 (FITC) nm, and the emission was detected at 390 (AP 2 PL) and 520 (FITC) nm, respectively.
Transient fluorescence measurements were performed using an Ap-plied Photophysics DX 17MV stopped-flow spectrofluorimeter at 25°C (11). The experiments were started by mixing equal volumes (25 l) of solutions as described for the Na ϩ -ATPase measurements except that in some experiments, 16 mM NaCl was replaced with 1.6 mM KCl. In addition, one volume contained 60 g/ml enzyme protein and the other contained various concentrations of ATP or AcP without or with 5 or 160 mM NaCl or corresponding concentrations of Tris-HCl (pH 7.4) or choline chloride as a control. The sample was excited at 305 (BIPM), 320 (AP 2 PL), and 470 (FITC) nm and the emission was detected at 360 (BIPM), 390 (AP 2 PL), and 520 (FITC) nm after passage through Nippon Shinku Kogaku interference filters as described previously (19). The experiments were repeated at least 10 times. The ratios of accumulated data in the presence of substrates to those obtained for the controls were fitted to a single or a double exponential curve in order to estimate the apparent rate constants and the extent of fluorescence changes using a non-linear least squares fit analysis. The experiments were performed using several different enzyme preparations, and the data shown are typical examples. The starting positions of all the curves for transient fluorescence changes (Figs. 2A, 3A, 3B, 5, 6, and 7A) were offset in order to see the shape of each curve more clearly. Thus, the change in fluorescence intensity after addition of ligand(s) in each curve is the change from the corresponding starting position of the curve. The time-dependent phosphorylation was followed using a Bio-Logic Quench-Flow Module (QFM-5) at 25°C. The experiments were started by mixing equal volumes (60 l) of the solutions used for the Na ϩ -ATPase measurement except the [ 32 P]ATP was replaced with [ 32 P]AcP. After mixing, the reaction was quenched by adding 1 M HCl, 10 mM H 3 PO 4 , and 1 mM AcP or ATP (11).

Steady State Fluorescence Measurement of ATP-and AcPinduced Fluorescence
Changes in the Mg 2ϩ -Na ϩ -bound AP 2 PL-FITC Enzyme-Previous stopped flow experiments showed that the addition of ATP and AcP to the Mg 2ϩ -Na ϩ -bound AP 2 PL-FITC enzymes containing ϳ0.5 mol of AP 2 PL probe at Lys-480 and ϳ0.9 mol of FITC probe at Lys-501/␣-chain, respectively, induced a similar extent of increase in the AP 2 PL fluorescence intensity without and with phosphorylation (18). To investigate this point further, steady state fluorescence measurements were performed. The addition of 10 M ATP to the Mg 2ϩ -Na ϩ -bound AP 2 PL-FITC enzymes increased the AP 2 PL fluorescence with no observable FITC fluorescence change (Fig. 1A, top and bottom). The AP 2 PL fluorescence decreased to the initial level over a period of time and the readdition of ATP caused a reversible increase in the fluorescence (Fig. 1A, top). However, no significant FITC fluorescence change was observed, even after the addition of 100 M ATP, which kept the increased level of AP 2 PL fluorescence up to 1 h (data not shown). The increase in AP 2 PL fluorescence by ATP in the presence of Mg 2ϩ and Na ϩ (Fig. 1A, top) without significant phosphorylation is indicative of an enzyme state just after ATP hydrolysis (see Ref. 18 and Table I of Ref. 19). The reason for time-dependent decrease in AP 2 PL fluorescence was shown to be mainly due to the decrease of ATP concentration by Na ϩ -ATPase activity of the AP 2 PL-FITC enzyme preparation, which still contained ϳ6% of Na ϩ -dependent phosphorylation capacity from ATP as shown (19).
The addition of 1 mM AcP to the Mg 2ϩ -Na ϩ -bound AP 2 PL-FITC enzymes increased AP 2 PL fluorescence and decreased FITC fluorescence (Fig. 1B, top and bottom). In addition the phosphoenzyme fully accumulated to a level of ϳ0.5 mol/mol of ␣-chain as is shown in the accompanying article (19). The addition of ouabain further increased AP 2 PL fluorescence to accumulate ouabain-bound phosphoenzyme but had little effect on FITC fluorescence (Fig. 1B, top and bottom). Further addition of ATP to ouabain-bound phosphoenzyme had a negligible effect on AP 2 PL fluorescence intensity (Fig. 1B, top).
The AP 2 PL fluorescence intensity at Lys-480, which accompanied the accumulation of ouabain-bound phosphoenzyme was higher than that of K ϩ -sensitive phosphoenzyme from AcP ( Fig. 1B, top). Such a conformational difference detected by fluorescence change has not been detected with the BIPM probe at Cys-964 (8) or the FITC probe at Lys-501 (14) to date. The FITC fluorescence change induced by AcP in the presence of Mg 2ϩ and Na ϩ is known to be reversible (17), as in the case for AP 2 PL fluorescence change by ATP described above.
Stopped-flow Measurement of ATP-induced AP 2 PL Fluorescence Changes of the Mg 2ϩ -Na ϩ -bound AP 2 PL-FITC Enzyme-To follow the initial ATP-induced phase of the AP 2 PL fluorescence change of the Mg 2ϩ -Na ϩ -bound AP 2 PL-FITC enzymes, various concentrations of ATP were added to the AP 2 PL-FITC enzyme ( Fig. 2A). The addition of 1 and 3 M ATP induced a biphasic AP 2 PL fluorescence change similar to that observed for the AP 2 PL enzyme (19), namely a rapid decrease, followed by a slow increase. The rapid decrease apparently disappeared because of an acceleration in the slow increase in the presence of up to 10 M ATP. However, the extent of the fluorescence increase was near saturation at around 10 M ATP (Fig. 2B, inset).
The rate constants for the rapid decrease and the slow increase increased in a nearly linear fashion with increasing concentrations of ATP (Fig. 2B), suggesting the participation of at least two different ATP bindings. The slopes of the two straight lines (Fig. 2B) permitted a semiquantitative estimation of the apparent second order rate constants (k ϩ1 ) to be 1.8 ϫ 10 6 and 0.1 ϫ 10 6 M Ϫ1 s Ϫ1 , respectively, for ATP for the formation of the two different conformational states with negative and positive AP 2 PL fluorescence intensity, compared with the Na ϩ -bound enzyme. The dissociation rate constant (k Ϫ1 ) and the dissociation constant (K d ϭ k Ϫ1 /k ϩ1 ) for the latter were estimated from the intercept, 0.82 s Ϫ1 , and the ratio, 7.5 M. These data suggest that at least two different ATP bindings with a ϳ20-fold difference in k ϩ1 value induced quite different AP 2 PL fluorescence changes in the Mg 2ϩ -Na ϩ -bound AP 2 PL-FITC enzymes without significant phosphorylation (18,19). ATP appears to be a better substrate than AcP in terms of increasing AP 2 PL fluorescence, as evidenced by the fact that each k 1 and K d was at least 100-fold and 1/40 of that for AcP (19).
When the concentration of added ATP was reduced to 0.3 M ATP, the concentration of which was ϳ3 fold larger than the concentration of the ␣-chain present ( Fig. 2A, inset), the rate A Na ϩ ,K ϩ -ATPase preparation from pig kidney was treated with 50 M AP 2 PL or further treated with 15 M FITC as described previously (19). The final reaction mixture for the fluorescence measurements contained 30 g protein/ml with 25 mM imidazole-HCl (pH 7.4), 4 mM MgCl 2 , 25 mM sucrose, 16 mM NaCl, and 0.1 mM EDTA-Tris (pH 7.4) with 10 M ATP or 1 mM AcP. Each fluorescence change was monitored after the addition of 10 M ATP (A) or 1 mM AcP followed by 10 M ouabain (B) to the AP 2 PL-FITC enzyme. A difference spectrofluorophotometer was used to monitor the change at 25°C. The fluorescence intensities shown are the relative intensity of the initial enzyme states. constant for the decrease and the increase in AP 2 PL fluorescence were determined to be, respectively, 1.3/s and 0.06/s. This very slow increase seemed to be due to the slow turn over of the AP 2 PL-FITC enzyme just after hydrolysis of ATP (18,19) as it reached the original fluorescence level rather than to the second ATP binding-induced increase in the fluorescence, the extent of which would be higher than the initial fluorescence level as observed in the presence of up to 3 M ATP (Fig. 2

, A and B, inset, open circles).
Stopped-flow Measurement of ATP-and Na ϩ -induced Fluorescence Changes of K ϩ -bound AP 2 PL-FITC Enzyme-High concentrations of ATP are known to accelerate the transition of K ϩ -bound enzyme to Na ϩ -bound enzyme (21). The ATP binding capacity of the AP 2 PL-FITC enzyme was already shown to be at least 50% of that of the control enzyme (18). To investigate whether ATP accelerated the transition of K ϩ -bound AP 2 PL-FITC enzyme to the Na ϩ -bound enzyme, both AP 2 PL and FITC fluorescence changes were followed immediately after the addition of ATP and/or Na ϩ . The addition of 5 mM ATP induced fluorescence changes in neither AP 2 PL (Fig. 3A, second trace) nor FITC (Fig. 3B, third trace). The addition of 160 mM Na ϩ induced a decrease in the AP 2 PL fluorescence (Fig. 3A, third trace), 3.6/s, and a biphasic increase in FITC fluorescence (Fig.  3B, second trace), 4.4 and 0.8/s. The addition of 160 mM Na ϩ with 0.5 mM ATP induced a biphasic AP 2 PL change (Fig. 3A, bottom), a rapid decrease (21/s), followed by a slow increase (0.7/s) with only slight effect of ATP to increase in both the extent and the rate of the FITC fluorescence (trace not shown). The extent of the initial rapid decrease in AP 2 PL fluorescence with increasing the concentrations of ATP could not be detected accurately due to the subsequent increase (Fig. 4B, closed circles).
The addition of 160 mM sodium with 5 mM ATP induced a monophasic increase in AP 2 PL fluorescence (Fig. 3A, top), 6.4/s, with a biphasic increase in FITC fluorescence (Fig. 3B, top), 7.0 and 1.2/s, with slight increase in the extent. Although the addition of 5 mM Na ϩ induced slight increases in FITC fluorescence (Fig. 3B, bottom), the simultaneous addition of both 5 mM Na ϩ and 5 mM ATP induced larger biphasic increases in FITC fluorescence (Fig. 3B, fourth trace), 5.8 and 1.8/s, which showed more clearly the effects of ATP on the extent of FITC fluorescence change. Fig. 4 (A and B, respectively) shows the dependence of ATP concentration on the rate constants and the corresponding extent of both fluorescence changes in the presence of 160 mM Na ϩ . Each rate constant for the decrease and the increase of AP 2 PL fluorescence increased with increasing concentrations of ATP accompanied by some deviation from linearity (Fig. 4A, closed and open circles). The concentration dependence of ATP on FITC fluorescence changes that accompanied the transition from the K ϩ form to the Na ϩ form also suggest two different extents of ATP-induced FITC fluorescence increases with only a small increase in the rate constants (Fig. 4, A and B, open and closed triangles). These data showed that level of ATP up to submillimolar levels exerted two different actions on each of the AP 2 PL and FITC fluorescence changes for the K ϩ -bound AP 2 PL-FITC enzymes, namely that the same concentrations of ATP resulted in four different fluorescence changes with different rate constant and/or extents. However, approximately micromolar ATP resulted in two actions on the AP 2 PL fluorescence change of the Na ϩ -bound AP 2 PL-FITC enzymes (Fig. 2B) with little FITC fluorescence change.
AcP-induced Fluorescence Changes and Phosphorylation of Mg 2ϩ -Na ϩ -bound AP 2 PL-FITC Enzyme-To investigate the relationship between the AcP-dependent phosphorylation and fluorescence changes, 1 mM AcP was added, the concentration of which is sufficient to saturate phosphorylation (19). This caused a decrease in the FITC fluorescence with rate constants of 8.5 and 1.2/s, and a decrease of 5.5 and 0.8% respectively (Fig. 5, second trace). The data could be also fitted to a single exponential curve with 7.1/s and 5.7%. Phosphorylation occurred with a rapid (7.2/s) and slow (4.6/s) phase, and showed nearly the same amount of components, namely 24 and 26%, to give half-site phosphorylation (Fig. 6, open circles), followed by two different AP 2 PL fluorescence increases with rate constants of 1.4 and 0.2/s and an extent of 2.4 and 0.7%, respectively (Fig. 6, top trace). Such double exponential AP 2 PL fluorescence changes cannot be obtained for the case of AcP-induced AP 2 PL fluorescence of AP 2 PL enzyme (19). The addition of 10 M ATP induced neither FITC fluorescence change (Fig. 5, top) nor significant phosphorylation (18,19) despite an increase in AP 2 PL fluorescence (Fig. 2). However, the less bulky substrate, AcP, induced both AP 2 PL (13, 18) and FITC (10,11,14) fluorescence change and phosphorylation as was described above.
The extent of FITC fluorescence decrease of the AP 2 PL-FITC enzyme induced by 1 mM AcP was 6.3 (ϭ 5.5 ϩ 0.8) % (Fig. 5,  second trace), while that of the FITC enzyme, which was not treated with AP 2 PL, was 10.5 (ϭ 10.1 ϩ 0.4) % (Fig. 5, fourth  trace), with similar rate constants. The relative FITC fluores-cence changes accompanying the transition from the K ϩ form AP 2 PL-FITC enzyme to the Na ϩ form enzyme with or without ATP were also approximately half of those of the FITC enzyme (16). These data suggest that the half-site modification of Lys-480 with AP 2 PL reduced the extent of the FITC fluorescence change at Lys-501 by nearly 50%.
When the FITC enzyme, which contained ϳ0.9 mol of FITC probe at Lys-501, was treated with AP 2 PL, the resulting FITC-AP 2 PL enzyme showed neither an AcP-nor an ATP-induced AP 2 PL fluorescence change (Fig. 6, second and third trace) but the FITC fluorescence changes were retained, similar to those of the FITC enzyme (Fig. 5, third and bottom traces). The data show that it is not possible to modify Lys-480 with AP 2 PL, when Lys-501 is already modified with FITC.
Stopped-flow Measurement of ATP-induced Fluorescence Changes of the Mg 2ϩ -Na ϩ -bound BIPM Enzyme-The ATPinduced biphasic AP 2 PL and FITC fluorescence changes (Figs. [2][3][4] and other experiments (14,16,18,19) strongly suggest that the enzyme is oligomeric in nature. However, one may ask whether the complexity observed above is due to the results of chemical modification in the vicinity of the ATP protectable Lys residues. To investigate this point, ATP-induced conformational changes using the BIPM-modified enzyme containing ϳ1 mol of BIPM probe at Cys-964/␣-chain, which retained the full, original Na ϩ ,K ϩ -ATPase activity (7), were followed.
The addition of 0.01 and 0.03 M ATP to the BIPM enzyme in the presence of Na ϩ and Mg 2ϩ induced a single exponential increase in BIPM fluorescence (Fig. 7A). However, with increasing concentrations of added ATP, the BIPM fluorescence change deviated from a single exponential curve and could be fit to a double exponential curve. The extent of BIPM fluorescence changes increased to give maximum values, ϳ4 and ϳ3% with K 0.5 values for ATP in the submicromolar range and then decreased with increasing concentrations of ATP, respectively (Fig. 7, A, B, and inset). The apparent rate constant for the high fluorescence intensity was approximately 1 order of magnitude larger than that of the low intensity at the same ATP concentrations (Fig. 7B, open and closed circles). These apparent rate constants were significantly larger than the turnover number of Na ϩ -ATPase for the unmodified control enzyme, 0.35/s, (19) indicating that these conformational changes were sufficiently rapid to be candidates for reaction intermediates. Each plot of the apparent rate constant versus ATP concentrations gave two intersecting straight lines (Fig. 7B). These data suggest that each of the two different ATP bindings to the enzyme in the presence of low and high concentrations of ATP induced, respectively, the formation of two different conformational states that can be associated with the high and low BIPM fluorescence. The extents of both the high and low BIPM fluorescence decreased with increasing concentrations of up to micromolar ATP (Fig. 7B, inset) which was also detected under steady state fluorescence measurement using a difference spectrophotometer (not shown). The k ϩ1 values for ATP were semiquantitatively estimated, assuming that these plots gave straight lines (Fig. 7B). The k ϩ1 values for the formation of the conformation with higher BIPM fluorescence in the presence of low and high ATP concentrations were 18 ϫ 10 6 and 4.2 ϫ 10 6 M Ϫ1 s Ϫ1 , the corresponding k Ϫ1 was 9.4 and 77 s Ϫ1 , and the K d was 0.52 and 19 M, respectively. The k ϩ1 values of the conformation with lower BIPM fluorescence in the presence of low and high concentrations of ATP was 3.9 ϫ 10 6 and 0.69 ϫ 10 6 M Ϫ1 s Ϫ1 , and the corresponding k Ϫ1 was 2.7 and 5.1 s Ϫ1 and K d was 0.69 and 7.5 M, respectively. These data suggest the occurrence of at least two different ATP bindings, respectively, which induce different BIPM fluorescence changes at Cys-964 in the presence of low and high concentrations of ATP. It has already been shown that a dynamic BIPM fluorescence change at Cys-964 requires the simultaneous presence of ATP, Na ϩ , and Mg 2ϩ or Ca 2ϩ (9) accompanying or followed by half-site phosphorylation (19), respectively. ATP can be replaced by neither AMP-PNP nor ADP (9).

Do the Entire AP 2 PL Probe at Lys-480 and FITC Probe at Lys-501 in Each ␣-Chain Participate in the Fluorescence
Change?-Binding stoichiometries of the AP 2 PL and FITC probe/␣-chain (18) and phosphorylation stoichiometries (19) suggest the presence of nearly equal amounts of two main populations of AP 2 PL-FITC-and FITC-modified ␣-chains in the AP 2 PL-FITC enzymes (Fig. 8A). The same concentrations of ATP induced two quite different AP 2 PL fluorescence changes in the AP 2 PL-FITC enzyme (Fig. 2B), and these rate constants increased almost linearly with increasing concentrations of ATP, indicating that both AP 2 PL probes are responsible for the change.
The half-site modification of Lys-480 with AP 2 PL reduced the extent of FITC fluorescence change induced by AcP (Fig. 5) and that by Na ϩ with or without ATP accompanying transition from the K ϩ -bound to the Na ϩ -bound enzyme (see Fig. 3 of Ref. 16) by nearly 50% compared with those of the FITC enzyme. Neither AcP-nor ATP-induced AP 2 PL fluorescence changes were detected (Fig. 6, second and bottom traces) in the FITC-AP 2 PL enzyme, which showed an AcP-induced phosphorylation and an FITC fluorescence change (Fig. 5, third trace). The data showed that the modification of Lys-501 with FITC prevented the incorporation of AP 2 PL probe into Lys-480. The distance between Lys-480 and Lys-501 has been shown to be 14 Å (22). These data also suggest that the three-dimensional structure near the ATP binding and phosphorylation domain becomes rather crowded, as shown by the half-site modification of Lys-480 with AP 2 PL, which prohibits the further modification of the remaining Lys-480, and that full-site modification of Lys-501 with FITC prevents the modification of Lys-480 with AP 2 PL. Furthermore, the absence of ATP-dependent phosphorylation despite presence of AcP-dependent phosphorylation was also observed. These data support the view that the AP 2 PL probe bound to Lys-480 resulted in a very little change in the microenvironment of the FITC probe at Lys-501 (Fig. 8A, striped ovals) in the same ␣-chain, showing that only the half of the FITC probe at Lys-501 is responsible for the fluorescence change.
Oligomeric Interaction between Subunits-The stoichiometries described above and differences in the extent of AP 2 PL fluorescence changes at the same Lys-480 (Fig. 2B, inset), the FITC fluorescence changes at the Lys-501 (Fig. 4B) and the BIPM fluorescence changes at Cys-964 (Fig. 7B, inset) are Simple hypothetical oligomer models of membrane-bound Na ϩ ,K ϩ -ATPase preparations are represented as a teraprotomer in which the half site modified with AP 2 PL at Lys-480 and the full site modified with FITC at Lys-501 without phosphorylation by ATP (A) and a tetraprotomer with half site phosphorylated at Asp-369 (B), where P designates phosphorylation by ATP. Phosphorylation by AcP occurred on the half site irrespective of AP 2 PL and/or FITC modification as shown (18,19). To simplify the models, ␤-subunits are not shown. Details are described under "Discussion." clear evidence in favor of the oligomeric nature of the enzyme. Rather lower concentrations of ATP induced two different effects on the rate and extent of not only AP 2 PL fluorescence of the Mg 2ϩ -Na ϩ -bound AP 2 PL-FITC enzymes ( Fig.  2) but also BIPM fluorescence changes in the Mg 2ϩ -Na ϩbound BIPM enzyme (Fig. 7). Plots of apparent rate constants against concentrations of ATP gave two straight lines for AP 2 PL fluorescence (Fig. 2B) and two intersecting straight lines in the presence of low and high concentrations of ATP for the BIPM fluorescence (Fig. 7B). The linear increase in the apparent rate constant with increasing ATP concentrations has been explained to be the result of a large dissociation constant for ATP of a nonfluorescent precursor of fluorescent enzyme form, wherein the conformation (fluorescence) change is slow (23,24). However, the data obtained favor another possibility that the ATP affects other ATPinduced fluorescence changes in the same molecule of a different subunit such as observed in the Mg 2ϩ -Na ϩ -bound AP 2 PL enzyme and the Mg 2ϩ -Na ϩ -bound BIPM enzyme. More direct evidence for four different effects of ATP were detected with respect to both the rate and extent of AP 2 PL and FITC fluorescence changes which accompanied the transition of the K ϩ -bound form to the Na ϩ -bound form (Figs. 3 and 4) of the AP 2 PL-FITC enzyme. In actual fact, the enzyme preparations were shown to retain up to 50% of the ATP binding capacity with affinity reduced by 2 orders of magnitude (18). The decrease in the extents of BIPM fluorescence at the same Cys-964 with increasing concentrations of ATP (Fig. 7B, inset) also reflect some cooperative interaction between the ␣-chains which reduce the BIPM fluorescence intensity. These data are consistent with a hypothesis that all ␣-chains are capable of accepting ATP and interacting with adjacent ␣-chains, thus accelerating ATP-induced conformational changes. A similar decrease in Trp fluorescence with an increase in the amount of phosphoenzyme in H ϩ ,K ϩ -ATPase has also been shown (25).
Conformational Change Occurring Out of Phase-From previous conformational studies of Na ϩ ,K ϩ -ATPase modified with or without fluorescence probe, we found that dynamic fluorescence changes of intrinsic and extrinsic fluorescence probes occur out of phase (8,10,11,26). These data were explained by assuming the sequential appearance of reaction intermediates during ATP hydrolysis (9,11). The data described in this paper provide clear evidence for the oligomeric nature of the enzyme and permit an explanation for a portion of the sequential molecular event induced by AcP in each ␣-chain of the Mg 2ϩ -Na ϩbound AP 2 PL-FITC enzymes (Fig. 8A). The binding of 1 mol of AcP to each ␣-chain that contains an FITC probe at Lys-501 without or with an AP 2 PL probe at Lys-480 induced an FITC fluorescence decrease (ϳ8/s) in the FITC labeled-␣-chain with a quarter-site phosphorylation (7, 2/s) followed by another quarter-site phosphorylation (4.6/s) in the AP 2 PL-FITC labeled ␣-chain. The possibility of that of 1 mol of AcP binding to each ␣-chain that contains only an FITC probe at Lys-501 was rejected, since in such a case, the extent of FITC fluorescence change would be the same as the FITC enzyme. In actually, it was 50% (Fig. 5, second and bottom traces). Further experiments will be required to explain the subsequent slow increases (1.4 and 0.2 s Ϫ1 ) in AP 2 PL fluorescence that occur for the case of two AP 2 PL-FITC labeled ␣-chains, and ATP induced a quarter-site phosphorylation followed by biphasic AP 2 PL fluorescence change in the AP 2 PL enzyme (19). However, the present experiments demonstrate that sequential molecular events in each subunit occur out of phase in real time not only in the vicinity of the ATP binding domain, Lys-480 and Lys-501 (18,22), but also in the vicinity of the transmembrane segment, Cys-964 (7,20). Such dynamic conformational differences would not have been detected by proteolytic digestion patterns, which showed clear differences between the Na ϩ -bound enzyme and the K ϩ -bound or phosphorylated enzyme form (27).
Recently, a considerable body of experimental evidence that are consistent with multiple active sites have been reported, as was described in the accompanying article (19). The present data are consistent with the same concentrations of submillimolar ATP inducing four different fluorescence changes that accompany the transition from the K ϩ form to the Na ϩ form of the AP 2 PL-FITC enzyme (Fig. 4) and that up to 1 order of magnitude lower concentrations of ATP induce four different conformational changes of Mg 2ϩ -Na ϩ -bound BIPM enzyme in the absence of K ϩ with lower K d values, ϳ0.5-19 M as detected by the BIPM fluorescence change at Cys-964 (Fig. 7). One may ask if there is some heterogeneity in the enzyme preparations such as due to the presence of denatured enzyme, isozymes, and heterogeneous labeling other than Lys-480, Lys-501, and Cys-964. The former two possibilities have been shown in the accompanying article to be unlikely (19). The present data could not exclude some possibility of heterogeneous labeled enzymes as minor components to give a small amplitude of some of the ATP effect. However, the following data of ours also strongly support such a conclusion independent of the presence of such heterogeneity: (a) quarter-site reactivities of the phosphoenzyme (10, 14) and a quarter-site phosphorylation from P i in the presence of Mg 2ϩ , Na ϩ , ADP, and ouabain (8) and a reduction of phosphorylation stoichiometry from half to a quarter after half-site modification of AP 2 PL at Lys-480 (19); (b) half-site phosphorylation capacity (19) with showing nearly full-site ATP binding capacity (14,18) and half-site modification of AP 2 PL at Lys-480 (18); (c) full-site ouabain binding capacity (14) and full-site reactivity of FITC at Lys-501 (18).
In conclusion, the present data represent the first direct conformational evidence to support the hypothesis that the functional unit of membrane-bound Na ϩ ,K ϩ -ATPase is a tetraprotomer, composed of ␣␤ protomers (Fig. 8B).