Substitutions of serine 775 in the alpha subunit of the Na,K-ATPase selectively disrupt K+ high affinity activation without affecting Na+ interaction.

The functional role of serine 775, predicted to be located in the fifth transmembrane segment of the alpha subunit of the Na,K-ATPase (YTLTSNIPE), was studied using site-directed mutagenesis, expression, and kinetic analysis. Substitutions S775A, S775C, and S775Y were introduced into an ouabain-resistant alpha 1 sheep isoform and expressed in HeLa cells. cDNAs carrying substitutions S775C and S775A produced ouabain-resistant colonies only when extracellular K+ was increased from 5.4 mM to 10 or 20 mM, respectively. No ouabain-resistant colonies were obtained for substitutions S775Y at any tested K+ concentration. Kinetic characterization of S775C and S775A substituted enzymes showed expression levels higher than control enzyme, reduced Vmax and turnover, and normal phosphorylation and high affinity ATP binding. Dephosphorylation experiments indicated that S775A substituted enzyme is insensitive to ADP but readily dephosphorylated by K+. The K+ K1/2 values for the activation of the Na,K-ATPase were markedly altered, with S775C displaying a 13-fold increase and S775A exhibiting a 31-fold increase. These large changes in the Na,K-ATPase affinity for K+ are consistent with the participation of this amino acid in binding K+ during the translocation of this cation. Substitutions of Ser775 did not change Na+ affinity, indicating that this residue is likely not involved in Na+ binding and occlusion. These data show that the electronegative oxygen and the small side chain of Ser775 are required for efficient enzyme function. Moreover, these results suggest Ser775 plays a distinct role in K+ transport and not in Na+ interactions, revealing a possible mechanism for the enzymatic differentiation of these cations by the Na,K-ATPase.

The Na,K-ATPase uses the energy produced by ATP hydrolysis to transport Na ϩ and K ϩ ions across the plasma membrane of eukaryotic cells (1)(2)(3). This enzyme belongs to a class of ion pumps, termed P-type ATPases, that has a common catalytic mechanism in which an aspartyl phosphate intermediate is transiently formed (1, 4 -6). In addition to the Na,K-ATPase, this group includes the gastric H,K-ATPase, the sarco(endo)plasmic reticulum Ca-ATPase, and the plasma membrane Ca-ATPase, among others.
The hydrolytic and transport cycles of the Na,K-ATPase have been described in detail (1)(2)(3). However, because little is known about the structure of the enzyme, the molecular events that take place in the protein during the binding of Na ϩ , K ϩ , and ATP, the hydrolysis of ATP, the transfer of energy to the cation sites, and finally the transport of ions, are largely unknown.
The Na,K-ATPase consists of two major subunits, ␣ (M r ϭ 112,000) and ␤ (M r ϭ 35,000 for the protein component) (1,2,7). Studies using hydropathy analysis, epitope localization, selective proteolytic cleavage, and labeling of putative transmembrane segments with hydrophobic probes suggest a topological arrangement of the ␣ subunit involving 10 transmembrane segments (3,8, see Fig. 3 in Ref. 2). However, the assignment of particular sequences to some transmembrane segments is still questioned; for instance, the exact identity of the fifth and sixth transmembrane segments (8 -10).
A major objective of our studies is to identify the amino acids that constitute the cation binding and occlusion sites of this enzyme. It is accepted that these sites reside in the transmembrane region of the ␣ subunit (2, 8, 10 -12). It has been shown that after removing large cytoplasmic portions of the protein by protease treatment, the remaining transmembrane fragments can still occlude K ϩ (13,14). It is reasonable to expect that carboxyl residues located within the membrane could be part of the occlusion cage neutralizing the cation charges. This concept is supported by findings in the Ca-ATPase from sarcoplasmic reticulum, where intramembrane carboxyl residues seem to play an important role in Ca 2ϩ binding and transport (15,16). In the Na,K-ATPase there are probably seven carboxyl residues within the membrane (Glu 327 , Glu 779 , Asp 804 , Asp 808 , Asp 926 , Glu 953 , and Asp 954 ) 1 and they have been the target of chemical modification and site-directed mutagenesis (17)(18)(19)(20)(21)(22)(23)(24)(25)(26). Chemical modification studies using 4-(diazomethyl)-7-diethylamino) coumarin (DEAC) 2 identified Glu 779 as involved in cation binding (17,18). Although transfectants of ouabain-resistant enzyme containing both E779L and E779D are unable to confer ouabain resistance to ouabain-sensitive HeLa cells, the substitutions E779A and E779Q yield functional enzymes, indicating that neither the carboxyl group nor the electronegative atom are "essential" for enzyme function (19 -22). However, it is possible that an oxygen atom or a hydrogen bonding capability is required in this position for stabilizing the E 1 P(Na) form of the enzyme (21). The other carboxyl residues located in the membrane have also been mutated with mixed results (19 -26). Van Huysse et al. (23,25) demonstrated that carboxyl residues Glu 953 and Glu 954 are not required for enzyme function. Likewise substitutions of Asp 926 did not lead to major alterations in the cation affinity (19). Similar to Glu 779 , substitutions of Glu 327 and Asp 808 display mixed characteristics with either severely impaired activity (E327A, E327D, D808L, D808Q) or slightly altered kinetic parameters (E327L, E327Q, D808E, 3 D808A 3 ) (19,21,24,26). Furthermore, the functional studies that have been performed with some of these mutants do not indicate a clear pattern to support a particular role for these residues or to provide a convincing structural model (based on structure-function relationships). At the present, only Asp 804 appears to be important for enzyme function, since no functional substitution has been found for this residue 3 (19,21). Thus, in spite of considerable work, there is no clear understanding of the "cation binding cage," although it seems probable that other residues (possibly non-carboxyl) must be necessary for cation binding, occlusion, and transport.
By analogy with other cation binding proteins, it is likely that the cations are coordinated by six or eight oxygen-containing residues, water, or even backbone carbonyl oxygen, when they are bound or occluded in the Na,K-ATPase (27)(28)(29). Considering this, we have begun testing the putative involvement of other (non-carboxyl) oxygen-containing residues in cation binding and occlusion. Particularly, we have targeted those located in the putative fifth transmembrane segment (Y 771 TLTSNIPEITPFLIFIIANI 791 ). Previously, we have observed that substitutions to alanine of several of these residues yield apparently nonfunctional enzymes (20). Furthermore, the structural connection of this transmembrane segment with the large cytoplasmic loop (where ATP binds and is hydrolyzed) suggests that this segment might play a key role in cation binding and also in the structural transitions required to transport Na ϩ and K ϩ through the plasma membrane.
We report here the functional effects of substitutions of Ser 775 (YTLTSNIPE). Our findings indicate that, while this serine is not involved in Na ϩ interactions with the enzyme, it is very likely part of the K ϩ -binding site. The participation of this serine in the interaction of one cation but not the other provides an initial structural basis for cation selectivity and contributes to the understanding of the transport mechanism of the Na,K-ATPase.

EXPERIMENTAL PROCEDURES
Mutagenesis and Cloning-The eukaryotic expression vector pKC4 containing the sheep Na,K-ATPase ␣1 subunit cDNA, which was modified by substitutions Q111R and N122D to encode a form of the enzyme with low affinity for ouabain (RD ␣1), was used (30). A 763-base pair HindIII-BglII cassette containing the region encoding amino acid residues 691-945 from the sheep ␣1 cDNA was subcloned into the M13mp19 vector and used to introduce the desired single mutation by the method of Kunkel (31). Mutated cassettes were sequenced to verify that no other mutations were produced. The cassettes were then subcloned back into the sheep RD ␣1 cDNA in the pKC4 expression vector. Final constructs were examined by restriction analysis, as well as by sequencing across the mutation site. Nucleotide substitutions were made to produce the following amino acid replacements: S775A, S775C, and S775Y.
Tissue Culture and Transfection of HeLa Cells-HeLa cells were maintained in Dulbecco's modified Eagle's media (DMEM) supplemented with 10% calf serum at 37°C in humidified air at 5% CO 2 . Twenty-five g of the pKC4 expression vectors were transfected into 5 ϫ 10 5 HeLa cells in accordance with the modified calcium-phosphate method of Chen and Okayama (32). Two days after transfection the transfected HeLa cells were selected by inclusion of 1 M ouabain in the culture media for a period of 2-5 weeks (30). Ouabain-resistant colonies were isolated from different transfections and expanded into stable cell lines.
In experiments in which the effect of extracellular K ϩ on cell survival was examined, DMEM (that contains 5.4 mM KCl) was supplemented with KCl to reach either 10 or 20 mM KCl. The cells were adapted to the higher K ϩ concentrations 2 days prior to transfection and maintained at the desired K ϩ concentration during the selection with ouabain. Controls with choline-Cl instead of KCl were performed. Mock transfections were made to determine if the inclusion of KCl had any effect on the selection with ouabain. Colonies obtained after selection in high K ϩ were consistently maintained in the KCl concentration in which the cells were selected.
Membrane Preparations-Crude membranes from HeLa cells were prepared using a NaI treatment (33) as described by Jewell and Lingrel (34). Na,K-ATPase from sheep kidney outer medulla was purified according to Jørgensen (35) with the modifications of Liang and Winter (36). The Na,K-ATPase activity of the purified sheep enzyme was 21 mol P i /mg/min and the preparation appeared Ͼ90% pure by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis. The total protein concentrations were determined by the method of Bradford (37) using bovine serum albumin (BSA) as standard.
Quantification of the Expressed Na,K-ATPase-Aliquots (5 l) of a serial dilution of each membrane preparation were dot-blotted onto nitrocellulose membranes. The blots were stained with the sheep-specific monoclonal antibody M8-P1-A3 and anti-mouse horseradish peroxidase-conjugated secondary antibody. The antibody M8-P1-A3, which was a generous gift from Dr. James Ball (University of Cincinnati), does not recognize the endogenous human isoforms (38). A serial dilution of the purified sheep Na,K-ATPase was also blotted onto each membrane and served as the standard by which the amount of heterologously expressed protein was determined. Only concentrations falling in the linear range of the signal/concentration curves were used for calculations. Membranes prepared from untransfected HeLa cells were spotted as negative controls. No detectable signal was associated with dots of these membranes carrying 100 times more total protein than those corresponding to transfected cells.
Na,K-ATPase Activity Determinations-The standard assay medium for Na,K-ATPase was 0.5 mM EGTA, 130 mM NaCl, 20 mM KCl, 3 mM MgCl 2 , 3 mM ATP, and 50 mM imidazole, pH (20°C) 7.2; 0.3 mg/ml BSA, approximately 1 g/ml of membrane protein, and either 10 or 0.01 mM ouabain. In some experiments ouabain, Na ϩ ,K ϩ or ATP were varied as indicated in the figure legends. Choline-Cl was added to the assay media to maintain the ionic strength, when the cation concentrations were changed (total [monovalent cation] ϭ 200 mM). The assay was performed at 37°C for 30 min and the released P i determined by the colorimetric method of Lanzetta et al. (39). The Na,K-ATPase corresponding to the expressed proteins was calculated by subtracting the activity observed in the presence of 10 mM ouabain from that detected with 0.01 mM of the inhibitor. A ouabain concentration of 0.01 mM inhibits the endogenous human enzyme, while 10 mM ouabain inhibits both the human and the heterologously expressed RD enzymes. Activation curves were fit using the equation v ϭ V max L n /L n ϩ K1 ⁄2 n . Phosphorylation Assays-Na ϩ -activated phosphorylation by ATP was carried out in a medium containing 100 mM NaCl, 0.04 mM EGTA, 75 mM Hepes/imidazole, pH (20°C) 7.2, 1 mM MgCl 2 , 0.01 mM [ 32 P]ATP, 0.01 mM ouabain, and 0.2 mg/ml membrane protein. Blank tubes were identical; however, NaCl was substituted by 100 mM KCl. In experiments where the Na ϩ dependence was studied, imidazole was replaced by Tris and choline-Cl was included to maintain ionic strength in the media. The reaction was initiated by addition of [ 32 P]ATP and stopped after 30 s at 0°C with five volumes of 7.5% trichloroacetic acid, 1 mM ATP, and 1 mM P i . The samples were pelleted by centrifugation and the pellets washed three times with the same solution. The pellets were resuspended in 2% SDS and radioactivity quantified with a liquid scintillation counter.
Dephosphorylation Assays-The effects of EDTA, ADP, or K ϩ on phosphoenzyme level were examined in samples phosphorylated as described above. 5 mM EDTA or 1 mM ADP, 5 mM EDTA, or 10 mM KCl, 5 mM EDTA, or 100 mM KCl, 5 mM EDTA (final concentrations) were added to the medium after 30 s instead of the acid-stopping solution and the samples incubated for a further 5 s at 0°C. The incubations were then stopped with acid, pelleted, resuspended, and counted as described above.

RESULTS
A mutagenesis and expression strategy previously developed in our laboratory (7,19,20,22,23,30) was employed to study the role of Ser 775 . This system uses a cDNA encoding the sheep ␣1 subunit that has been made ouabain insensitive by the introduction of substitutions Q111R and N122D (RD enzyme). This cDNA is specifically mutated and is transfected into HeLa cells which express a human ␣1 isoform with high affinity for the inhibitor. The transfected cells are grown in the presence of 1 M ouabain. This concentration of ouabain leads to the death of the untransfected cells as well as those transfected with cDNAs carrying substitutions that alter Na,K-ATPase function to the extent that the expressed enzyme cannot support cell viability under the cell culture conditions. In previous studies, only cells carrying mutated pumps with relatively mild effects in overall Na,K-ATPase activity have survived the selective pressure of ouabain, and the stable expression has allowed the study of the enzymatic characteristics of these enzymes.
Substitutions of serine at position 775 were designed to remove the hydroxyl group S775A, to make the most conservative alteration changing the hydroxyl to a sulfhydryl, S775C, or to change the size while retaining the oxygen atom S775Y. In initial studies we observed that none of the cells transfected with cDNAs coding for substitutions at Ser 775 were able to produce colonies after selection in regular DMEM which contain 5.4 mM KCl (Fig. 1). Under similar conditions HeLa cells transfected with control RD ␣ cDNA yielded a large number of ouabain-resistant colonies. It was apparent that ␣ subunits substituted at Ser 775 were unable to produce functional enzymes, including the conservative substitution to cysteine. It was possible that these substitutions disrupted the cation binding site of the Na,K-ATPase, supporting our hypothesis that Ser 775 was involved in the coordination of cations. Based on this assumption, if these substitutions diminish the cation (Na ϩ or K ϩ ) affinities, then by increasing the cation concentrations it could be possible to "saturate" the enzyme allowing it to function at (or near) physiological levels, allowing the survival of transfected cells. Experiments were designed to determine if increasing K ϩ in the culture media would allow the cells to survive. In these experiments, substitutions S775C and S775A that were unable to sustain HeLa cell growth in regular DMEM (5.4 mM K ϩ ) produced numerous ouabain-resistant colonies when K ϩ was increased to 10 and 20 mM, respectively (Fig. 1). No colonies were detected in plates transfected with mutant S775Y. While K ϩ does not affect the total ouabain binding capacity, it does lower the ouabain binding rate (40), and therefore high K ϩ concentrations may alter the ouabain selection procedure such that untransfected HeLa cells can survive in the presence of the inhibitor. Mock transfections were performed with no DNA and ouabain selection in high K ϩ . No colonies were detected, indicating that the higher K ϩ concentrations did not affect the ouabain selection method. Likewise, colonies were not observed in controls using choline-Cl instead of KCl. These results indicate that although enzymes with the substitutions S775A and S775C were severely impaired, they could function enough to support cell growth under facilitating conditions. Furthermore, these findings predicted that the interaction between K ϩ and the enzyme was affected by substitutions at Ser 775 .
Expression Level, Na,K-ATPase Activity, and Turnover Number of Substitutions of Ser 775 -The Ser 775 substitutions were studied in the RD enzyme background. Since the difference in ouabain affinities between the endogenous human Na,K-ATPase and the expressed RD-substituted enzymes is the basis for selection and the assay of enzymatic characteristics, the sensitivities of these enzymes to ouabain inhibition were determined. No significant difference in ouabain affinity was detected between the control RD and the Ser 775 -substituted enzymes. The ouabain IC 50 for these enzymes was approximately 500-fold higher than that of the endogenous human isoform (not shown). Also from these curves it was evident that the amount of endogenous human enzyme, i.e. the Na,K-ATPase activity inhibited by 0.01 mM ouabain, was less than 10% in all membrane preparations.
In site-directed mutagenesis studies, the expression system and the characteristics of a particular protein can yield distinct expression levels for different substitutions. These differences in expression levels are relevant when parameters such as specific activity are necessary to understand functional alterations. We have estimated the level of expression of the two mutants S775A and S775C and the control RD enzyme by quantifying immunostained dot-blots of serial dilutions of each membrane preparation. A purified sheep enzyme was used as a standard. Table I shows the level of each of these enzymes in our membrane preparations. It was determined that the mutated Na,K-ATPases S775A and S775C were expressed in significantly larger quantities than the control RD protein. However, reduced V max were found in the S775A-and S775Csubstituted enzymes when their activities were compared to the RD control (Na,K-ATPase activities measured at saturating concentrations of all ligands and expressed per milligram of heterologous enzyme). These results imply that the substitutions of Ser 775 produce a major alteration in the enzymatic activity. Keeping in mind that purified sheep enzyme has been used as a standard, the activity of the control RD enzyme (calculated per milligram of heterologous enzyme) (Table I) is lower than those usually obtained for purified Na,K-ATPase. It is known that a significant pool of unassembled (likely nonfunctional) ␣ subunits is normally present in cells (41)(42). These subunits are probably present in our preparations and may be detected in our immunoblot determinations, lowering the calculated activity values (expressed per milligram of heterologous enzyme). Other factors including overestimation of the expression level due to the immunostaining method or partial inactivation by NaI during membrane preparations may have also decreased the calculated activity. In any case, our quantification demonstrates that the substituted enzymes are expressed in larger quantities than the control RD enzyme.
To determine the amount of functional enzyme, the phosphorylation levels were measured. Under the phosphorylation conditions used, the E 1 7E 1 P equilibrium is largely displaced toward the phosphoenzyme form, thus the level of E 1 P (or E 2 P) is nearly equal to the total amount of functional enzyme. Similar phosphorylation levels (calculated per milligram of heterolo- Role of Serine 775 in the Na,K-ATPase K ϩ -binding Site 22766 gously expressed protein) were detected in the three enzymes (Table I), indicating that the fraction of functional ␣ subunit was comparable for the three proteins and that the activity values (per milligram of heterologous enzyme) of S775C and S775A enzymes are lower than that of RD enzyme. This last point is more evident when the turnover number, which is independent of the protein concentration, is calculated. As expected, the turnover numbers of the substituted enzymes were significantly lower than that of RD control enzyme (Table I) showing again the effects of these replacements in the overall enzyme activity.
The turnover value of the control RD enzyme is in good agreement with others reported for purified Na,K-ATPase (43,44); however, it is much lower than those described for heterologously expressed sodium pump (21,22,24,45). This discrepancy is most likely based in the different procedure used to determine the phosphoenzyme level. Previous studies have separated the phosphorylated Na,K-ATPase from other phosphoproteins in the preparations using acidic SDS-polyacrylamide gel electrophoresis gels. A common problem of these gels is that not all the protein (after being precipitated with trichloroacetic acid and partially resuspended with SDS) enters into the gels. This eventually leads to an underestimation of the phosphoenzyme and, consequently, an overestimation of the turnover number. In place of the acidic gels, we directly measure the phosphorylated Na,K-ATPase by subtracting the level of phosphorylated protein formed in the presence of K ϩ (absence of Na ϩ ) from the phosphoenzyme formed in the presence of Na ϩ (no K ϩ ). Controls using gels to separate the different phosphorylated proteins showed that the only K ϩ -or Na ϩsensitive phosphoprotein in our preparations is the Na,K-ATPase (22). Control experiments were also performed using membranes from untransfected HeLa cells where the human isoform is not phosphorylated if it is preincubated in the presence of 0.01 mM ouabain. After preincubation with ouabain the phosphoprotein was insensitive to K ϩ and its level identical to that obtained for membranes from transfected cells at saturating K ϩ (no Na ϩ ) (not shown). The background levels in the presence of K ϩ were never higher than 50% of the phosphoprotein produced in the presence of Na ϩ .
The alterations described in Table I indicate that Ser 775 is an important residue required for normal enzyme function. On the other hand, the data in Table I also point out that the cells are compensating for the lower activity of the S775A and S775C substitutions by producing large quantities of these enzymes to reach the amount of Na,K-ATPase activity (and consequent ionic homeostasis) required for cell survival. In this way, no difference between the mutants and the control enzymes is observed when comparing the total Na,K-ATPase activities (expressed per milligram of total protein in the membrane preparation) at the K ϩ concentrations at which the cells were selected (RD: 12.4 Ϯ 2.4 mol/h/mg (5.4 mM K ϩ ); S775C: 13.9 Ϯ 2.5 mol/h/mg (10 mM K ϩ ); S775A 10.8 Ϯ 2.2 mol/h/mg (20 mM K ϩ )). Furthermore, if human Na,K-ATPase from untransfected HeLa cells is isolated and measured under similar conditions, the activity is comparable to the one produced by the expressed proteins (13.5 Ϯ 3.4 mol/h/mg (5.4 mM K ϩ )). This demonstrates that in this expression system the activity of the heterologous protein must almost completely replace the endogenous Na,K-ATPase activity to maintain cell viability.
Cation Activation of the Na,K-ATPase Activity-The marked effects of external K ϩ in the survival of cells expressing Ser 775substituted enzymes point to possible alterations in the interactions of this cation with the mutated enzymes. This became evident when the K ϩ dependence of the Na,K-ATPase activity of the control RD enzyme and both mutants S775A and S775C were determined ( Fig. 2A). A large decrease in the apparent affinity for K ϩ was detected in both substituted enzymes. S775A shows a 30-fold lower K ϩ affinity compared to control RD, while the K ϩ affinity decreases 13-fold when the serine is replaced by cysteine. The magnitude of these alterations produced by the substitutions demonstrates the important putative role of Ser 775 as a part of the cation binding site. Furthermore, these results explain the observed effect of extracellular K ϩ on cell survival (Fig. 1). It is clear from Fig. 2 that at 5.4 mM K ϩ (K ϩ concentration in regular DMEM) the substituted enzymes are not able to transport Na ϩ and K ϩ at a significant level and that the alterations in K ϩ K1 ⁄2 of both mutants correlate with the extracellular K ϩ level required to obtain colonies.
The curves shown in Fig. 2A were performed at 30 mM NaCl which saturates the intracellular Na ϩ sites (E 1 cation sites). It is known that Na ϩ can compete with K ϩ for the extracellular cation binding sites (E 2 cation sites) (46,47). We determined if this competitive effect was present in S775A, assuming initially that if K ϩ binding to the external sites has been significantly affected by the replacements at Ser 775 , the interaction of Na ϩ at those sites would be largely removed (i.e. it would not compete). However, it was observed that 100 mM NaCl was able to displace to the right both curves, and the K ϩ K1 ⁄2 is higher in both RD and S775A enzymes (compared to that at 30 mM Na ϩ ) such that the relative increase (near 30-fold) is maintained (Fig. 2B).
Previously, amino acid substitutions in the Na,K-ATPase have, to a smaller or larger extent, affected the binding of both Na ϩ and K ϩ (19,(21)(22)(23)(24)(25)(26). Considering the simplest mechanism where the same "sites" that are involved in transporting Na ϩ outward bind and transport K ϩ inward, it was expected that a substitution affecting K ϩ interactions would also alter Na ϩ binding. Fig. 3 shows the dependence of the Na,K-ATPase on Na ϩ concentration. The K1 ⁄2 for Na ϩ activation of the Na,K-ATPase activity was not affected by the substitutions of Ser 775 . Since external K ϩ sites of S775A are not saturated at 20 mM KCl, the Na ϩ activation for this substitution was also measured at 100 mM KCl; again under this condition the substitu- tion had no effect on Na ϩ activation. Na ϩ K1 ⁄2 (mM) were: RD enzyme 16.50 Ϯ 2.81; S775A 21.22 Ϯ 4.72 (curves not shown). A possible explanation for this unique result was that an altered affinity for Na ϩ was masked by the turnover condition at which the Na ϩ -enzyme interactions were examined. To test this alternative, the Na ϩ K1 ⁄2 for the activation of the phosphorylation by ATP was determined for the RD control and the S775A-substituted enzymes (Fig. 4). Again no alteration in the Na ϩ activation of phosphorylation by ATP was observed. These data strongly suggest that Ser 775 does not participate in the coordination of Na ϩ when it interacts with the enzyme.
ATP Binding to Substituted Enzymes-It is known that ATP interacts with the Na,K-ATPase with two different apparent affinities. The high affinity binding is associated with the phosphorylation of the E 1 form of the enzyme by ATP at the catalytic site, while the low affinity ATP binding accelerates the deocclusion of K ϩ and the E 2 3 E 1 conformational transition (1, 48 -50). The interactions of ATP with high and low affinities were determined to assess the possible long range effects of these substitutions on the nucleotide-binding site. No alter-ation was observed in the binding of the nucleotide with high affinity when the ATP K m was determined in the sodiumactivated phosphorylation of the S775A-substituted enzyme (Fig. 5). Therefore, it appears that the replacement of Ser 775 did not affect the structure of the ATP-binding domain. The Na,K-ATPase activity dependence on ATP was measured to observe the interaction of the nucleotide at the low affinity binding site (Fig. 6). An increase affinity for ATP at the low affinity binding site was observed in both mutated ATPases, suggesting either a displacement of the E 2 7 E 1 toward E 1 or the instability of E 2 (K), such that less ATP is required to  FIG. 3. Na ؉ dependence of Na,K-ATPase activity of RD and Ser 775 -substituted enzymes. The Na,K-ATPase activity was determined as indicated under "Experimental Procedures." K ϩ concentration was kept at 20 mM, while Na ϩ concentration was varied. The values are the mean of results obtained with membrane preparations from four independent clones, each one measured in duplicate. Standard errors were lower than 5% and error bars are not plotted for simplicity. The Na,K-ATPase activities corresponding to 100% were similar to the V max values presented in Table I. The Na ϩ K1 ⁄2 (mM) and Hill coefficients were as follows: RD enzyme 7.06 Ϯ 0.43, n ϭ 2.36 (E); S775C 5.56 Ϯ 0.53, n ϭ 1.95 (å); S775A 5.98 Ϯ 1.01, n ϭ 1.65 (q).
FIG. 4. Na ؉ dependence of phosphorylation by ATP of RD-and S775A-substituted enzymes. The sodium-activated ATP phosphorylation was measured at different sodium concentrations as indicated under "Experimental Procedures." The values are the mean of results obtained with membrane preparation from four independent clones, each one measured in duplicate. Standard errors were between 5-10%, and error bars are not plotted for simplicity. The maximum phosphorylation levels corresponding to 100% were similar to the phosphorylation values presented in Table I. The Na ϩ K1 ⁄2 and Hill coefficients were as follows: RD enzyme 0.88 Ϯ 0.24 mM, n ϭ 0.95 (E); S775A 0.91 Ϯ 0.14 mM, n ϭ 1.46 (q).
Role of Serine 775 in the Na,K-ATPase K ϩ -binding Site 22768 accelerate the conformational transition.
Dephosphorylation of S775A-substituted Enzyme-The Na,K-ATPase phosphoenzyme contains both K ϩ -sensitive ADP-insensitive (E 2 P) and ADP-sensitive K ϩ -insensitive (E 1 P) components (51,52). We have studied the presence of these two forms after phosphorylation of RD control and S775A substituted enzymes (Table II). Addition of ADP or K ϩ lead to a reduction of the RD phosphoenzyme level to 48 and 81-76% respectively in 5 s at 0°C. The S775A phosphoenzyme was very insensitive to the addition of ADP, remaining 91% phosphorylated after 5 s at 0°C. The effect of two K ϩ concentrations on the S775A phosphoenzyme was studied, 10 and 100 mM, due to the lower apparent affinity for K ϩ of this enzyme. A fast dephos-phorylation of S775A enzyme was observed after the addition of K ϩ . These results suggest that the equilibrium E 1 P 7 E 2 P is displaced to the right in the substituted enzyme.

DISCUSSION
Expression of Na,K-ATPase ␣ Subunit Substituted at Ser 775 -To study the role of Ser 775 , we have used a mutagenesis and expression system in which selective pressure with ouabain leads to the death of the untransfected ouabain-sensitive cells and those carrying substituted Na,K-ATPase that can not support cell viability (7,19,20,22,23,30). A drawback of this approach has been that only cells carrying substitutions with relatively mild effects in the overall Na,K-ATPase activity survive the selection with ouabain. An alternative methodology has been developed to study substituted enzymes unable to maintain cell viability (26). In this system the displacement of ouabain binding by Na ϩ or K ϩ , or its stimulation by phosphorylation, has been examined using ouabain binding to the heterologous expressed pumps in the low ouabain affinity background of mouse cells. We describe in this report another approach that permits the rescue of apparently nonfunctional mutated enzymes. This is the modification of the culture media to meet the requirements of the expressed enzyme and therefore the ionic balance of the cell. This manipulation is not applicable to all mutants and clearly depends on the functional alteration that the mutation produces. In this sense the results in this report show correlation between the K ϩ K1 ⁄2 of each mutant and the external [K ϩ ] required for survival.
To understand how these cells survive with a severely impaired Na,K-ATPase, it is necessary to examine the expression levels in combination with the activity of these pumps. The cells appear to compensate for the low activity of the substituted enzymes by overproducing them to a level which can support ionic homeostasis compatible with cell growth. Even though our experiments were not designed to study the underlying induction mechanism, it is interesting to consider our findings in relation to other studies. It has been known for some time that the extracellular K ϩ level influences both the Na,K-ATPase expression and activity (53)(54)(55). The diminution of extracellular K ϩ leads to slower pump activity and disruption of the transmembrane ionic gradients. In this case the cells increase the number of Na,K-ATPases in the membrane as a compensatory mechanism to the lower activity. A similar situation is observed with substitutions at Ser 775 , where the extracellular K ϩ (almost under the K1 ⁄2 ) and the slower turnover leads to low Na,K-ATPase activity and, in turn, to a large production of enzyme.
Kinetic Characteristics of Substitutions at Ser 775 -In structure-function studies using site-directed mutagenesis (or chemical modification) techniques, it is important to determine whether substitutions that result in altered ligand interactions affect specifically the binding site or whether the enzyme is held in a conformational state which does not bind the ligand (17,56). We have measured diverse kinetic characteristics of the enzymes carrying substitutions in Ser 775 to understand the  Role of Serine 775 in the Na,K-ATPase K ϩ -binding Site 22769 functional role of this residue. These studies were designed to assess the interactions of the cations Na ϩ and K ϩ and ATP with the enzyme and discriminate conformational effects. Measuring the K ϩ activation of the Na,K-ATPase activity made it clear that removal of the hydroxyl group has a major effect in the K ϩ interaction with the enzyme. Even the conservative replacement of serine with cysteine results in a significant reduction in K ϩ affinity. On the contrary no alteration was observed in the Na ϩ activation of the Na,K-ATPase activity or phosphorylation by ATP. This indicates that this serine is not participating in the coordination of Na ϩ when it is translocated by the enzyme. It is clear that the substitutions at Ser 775 did not produce long range conformational effects that affect, for instance, the binding of ATP with high affinity. However, these enzymes have reduced turnover numbers indicating that one of the partial reactions of their catalytic cycle is slower compared to the RD control enzyme. Our determinations do not allow us to describe the affected partial reaction, although neither the dephosphorylation step (K ϩ can readily dephosphorylate E 2 P) nor the E 2 3 E 1 transition (increase in the apparent affinity for ATP at the low affinity nucleotide-binding site) appear to be slower in the substituted enzymes.
A decrease in the affinity for K ϩ could be due to either a specific alteration of the binding site or to a tendency of the enzyme to remain in an E 1 conformation with low affinity for K ϩ . The Ser 775 -substituted enzymes do not have any apparent preference to remain in the E 1 conformation. This is evident from the unaltered Na ϩ K1 ⁄2 measured under turnover conditions. A stabilization of E 1 would decrease this K1 ⁄2 , as observed for the L332A mutant (57). It could be argued that in our case this alteration could be masked by a real increase in the Na ϩ K1 ⁄2 (observable under non-turnover conditions). We did not see such a change when the Na ϩ activation of the ATP phosphorylation was analyzed. Furthermore, if the mutated enzymes had a tendency to remain in E 1 , then a nonspecific reduction in the interaction of any cation with E 2 should have been observed. On the contrary, the competition of Na ϩ with K ϩ for the extracellular (E 2 ) cation binding site seems largely unaltered.
However, the binding of ATP with low affinity is affected, and this could indicate a displacement in the equilibrium between conformational intermediates. ATP binds with low affinity to E 2 (K) (K ϩ occluded). This binding accelerates the deocclusion of K ϩ and the E 2 3 E 1 conformational transition (1, 48 -50). Other substitutions, such as E779A, E779Q, and E327Q (21,22,24), also produce increases in the affinity of the low affinity nucleotide binding site. However, the changes in cation affinity observed in those enzymes (2-5-fold) are minor compared to the change in K ϩ K1 ⁄2 of S775C and S775A enzymes (10 -30-fold). Therefore, the changes in the low affinity ATP binding of S775C and S775A enzymes cannot justify the large increments observed in K ϩ K1 ⁄2 . On the other hand, since we observed that the K ϩ enzyme interaction (probably binding and occlusion) is affected by the substitution of Ser 775 , then this increase in the affinity of ATP might be more likely associated to an instability of E 2 (K) (that would accelerate deocclusion and increase the apparent ATP affinity) than to a tendency of the substituted Na,K-ATPase to go to an E 1 form.
Considering that a slow E 2 P dephosphorylation rate or a displacement of E 1 P 7 E 2 P toward E 1 P would lead to a reduced affinity for K ϩ , we also analyzed the status of the phosphorylated forms of the S775A enzyme. Our results indicate a displacement of the equilibrium E 1 P 7 E 2 P toward the E 2 P form (i.e. S775A is insensitive to ADP). This conformational effect of the substitution also affects the K ϩ -induced dephosphorylation of the E 2 P form, which is apparently faster (compared to con-trol RD enzyme) at saturating K ϩ . The consequence, if any, of this conformational effect on the K ϩ affinity should be to increase the affinity for the cation. On the contrary we detected a decrease in affinity; thus, the conformational effect is perhaps masking an even bigger alteration at the cation site.
Significance and Functional Role of Ser 775 -The alterations in K ϩ affinities produced by substitutions of Ser 775 are the largest changes in affinity observed thus far for amino acid substitutions in the Na,K-ATPase. The degree to which the binding affinities of Na ϩ or K ϩ would be changed by substitution of one of the coordinating residues is unknown. Comparison with observations made in the Ca-ATPase is inadequate because the high affinity for the transported cation (in the low micromolar range) facilitates the detection of large changes in affinity. Instead our results can be compared with those obtained after replacements of carboxyl residues thought to be involved in cation binding in the Na,K-ATPase. Substitutions of these residues removing the carboxyl group (E779A, E327L, D926L, D953L, and D954L) result in 2-and 5-fold decreases in cation affinities, under turnover conditions (19,(21)(22)(23)(24). Only E779A produces an 8-fold increase in the Na ϩ K1 ⁄2 for the activation of phosphorylation by ATP (21). Similarly, when residues located in the cytoplasmic loop and thought to be involved in ATP binding are substituted, a 2-5-fold change in cation affinity is frequently observed, 4 indicating that these effects are easily achievable through an indirect conformational effect. These comparisons set apart Ser 775 , pointing out the major role of this residue in enzyme function.
Another distinctive feature of the mutations of Ser 775 is the correlation between the structural modification and the kinetic effect observed. The most disrupting substitution is the replacement S775Y in which the introduction of the bulky phenol ring yields largely inactive enzyme (taking into account that no ouabain-resistant colonies were obtained for this substitution). Mutation S775A, where the hydroxyl group is removed, shows large functional alterations, while S775C yields an intermediate phenotype between S775A and RD enzyme. These data strongly suggest that the enzyme requires a hydroxyl group as well as a relatively small side chain at this position for efficient cation transport.
Ser 775 is conserved in all the sequenced Na,K-ATPases, and it is also conserved in the sarcoplasmic reticulum Ca-ATPase (7,58). The equivalent of Ser 775 in the sarcoplasmic reticulum Ca-ATPase (S767) has been subjected to site-directed mutagenesis studies, and while no major alteration in Ca 2ϩ affinity was observed, the mutated enzyme exhibited low activity (16). It is interesting that Ser 775 (YTLTSNIPE) is replaced by a lysine (YTLTK 792 NIPE) in the H,K-ATPase (59). This structural difference in these two K ϩ -transporting enzymes is probably based in a different stoichiometry for K ϩ transport or in the different counter ion being transported. The H,K-ATPase stoichiometry has not been satisfactorily determined. Previous studies indicate that one K ϩ is transported in each cycle of the H,K-ATPase (60); thus, requiring fewer coordinating residues than the Na,K-ATPase. It is also possible that the H,K-ATPase uses the amino group of this lysine to coordinate protons being transported as hydronium ions (61) to form a proton relay (62) or to provide the proper environment for carboxyl group deprotonation (63), as it has been proposed for other proton-transporting proteins.
The role of Ser 775 in cation binding fits with the significance attributed to the fifth transmembrane segment, which connects with the cytoplasmic loop containing the phosphorylation site (18). Thus, Ser 775 is likely to perceive conformational changes produced by ATP hydrolysis or translate changes in the binding of K ϩ affecting the phosphoenzyme intermediate. Furthermore, its relative position within the membrane, i.e. the cytoplasmic half, agrees with the proposed model for the cation binding path as the ion well where the occlusion of cations takes place in the cytoplasmic half of the membrane (64,65).
The kinetic mechanism of the Na,K-ATPase indicates that Na ϩ and K ϩ are transported alternately through the membrane. The simplest model would suggest that the same sites that bind Na ϩ from the intracellular milieu are then accessible to bind K ϩ from the extracellular media. Even if both cations share some coordinating groups, others must be different to accommodate the stoichiometry of the transport and the different coordination requirements. Thus, some structural differences are necessary to confer selectivity. Keeping in mind the functional differences, it is interesting to consider the crystal structures of dialkylglycine decarboxylase as an example of a single "site" using different residues to coordinate Na ϩ or K ϩ . This enzyme binds both cations at the same "pocket"; however, a serine that coordinates K ϩ is displaced when Na ϩ binds the enzyme, while an extra water molecule takes up the sixth coordinating position (28). Taking into account that substitutions of Ser 775 affect to a large extent K ϩ but not Na ϩ binding, it is tempting to hypothesize that this serine confers the ion selectivity at the extracellular facing cation sites.
In summary, the functional effects of minimal structural alterations produced by substitutions of Ser 775 in the ␣ subunit of the Na,K-ATPase have been studied. Our findings indicate that this is an important amino acid which is required for the binding of K ϩ with high affinity. Thus, it is likely that Ser 775 participates in the coordination of K ϩ when the cation is bound, occluded, or transported. Furthermore, the Na,K-ATPase probably requires the hydroxyl group of a small amino acid to function properly. On the other hand, Ser 775 does not participate in the coordination of Na ϩ , and its removal seems not to have any effect in the binding of this cation. These results are the first report of a large change in the enzyme affinity for a cation. Furthermore, the critical role of Ser 775 in the binding of only one of the two transported cations is the first evidence that suggests a possible mechanism for the differentiation of Na ϩ and K ϩ by the Na,K-ATPase.