The Na,K-ATPase ()is an integral membrane protein found in nearly all eukaryotic cells that utilizes energy from the hydrolysis of intracellular ATP to transport three sodium ions out of the cell and two potassium ions into the cell (1, 2, 3) . The enzyme is composed of an subunit containing 8-10 transmembrane domains and a subunit which has one transmembrane segment. The catalytic cycle of Na,K-ATPase involves an acid-stable phosphorylated form of the protein, a characteristic shared by other active cation transporters such as the sarcoplasmic reticulum Ca-ATPase, the plasma membrane Ca-ATPase, H,K-ATPase, and Mg-ATPase(4, 5, 6, 7) . Unlike other ATPases, Na,K-ATPase is inhibited by cardiac glycosides such as ouabain.

In recent years, structure-function studies of Na,K-ATPase have focused on identifying the amino acids involved in binding and translocating Na and K ions. Both chemical modification studies and site-directed mutagenesis studies have targeted anionic amino acid residues located in the transmembrane domains of the subunit. These residues contain negatively charged side chains which are thought to neutralize the cations during transport through the hydrophobic lipid bilayer. Six oxygen-containing residues in the transmembrane domains of the sarcoplasmic reticulum Ca-ATPase have been implicated as essential amino acids for Ca transport using site-directed mutagenesis(8) . Three of these residues contain carboxylic acid side chains. These negatively charged residues are conserved in the Na,K-ATPase and include Glu, Glu, and Asp in the sheep 1 isoform.

N,N`-dicyclohexylcarbodiimide (DCCD) labels Na,K-ATPase in a K protectable manner and blocks cation occlusion upon modification(9) . Hence, DCCD is thought to bind to a residue essential for potassium occlusion. By combining DCCD labeling and limited proteolytic digestion, two sites of DCCD modification have been located within the transmembrane domains (10) . Glutamic acid 953 in the COOH-terminal half of the sheep 1 protein has been identified as one site of modification and glutamic acid 327 is hypothesized to be another site of modification. Site-directed mutagenesis studies, involving Glu and a neighboring glutamic acid residue, Glu, have demonstrated that altering the charged side chains at these positions does not dramatically change the cation-dependence properties of ATP hydrolysis by Na,K-ATPase(11) . Therefore, it is possible that the residue labeled by DCCD in the NH-terminal half of the enzyme, Glu, is the carboxyl group important for cation occlusion.

Glutamic acid 327 is located in the fourth transmembrane domain of the Na,K-ATPase and is highly conserved throughout the amino acid sequences of related ATPases(4, 5, 6, 7) . Previously, site-directed mutagenesis was employed to change Glu to Gln, Leu, Ala, and Asp(12, 13, 14) . These mutagenesis studies utilized similar expression schemes which take advantage of species-specific variations in ouabain sensitivity to examine the effects of the mutation on cell viability and on the cation-dependent ATPase activity. These schemes involve introducing the mutation into a cDNA which encodes a ouabain-resistant isoform of Na,K-ATPase (K = 1 10M; i.e. rat 1, sheep 1(RD) ()or rat 2*()), expressing this mutant in a cell line containing a ouabain-sensitive endogenous form of the ATPase (K = 1 10M; COS-1 or HeLa) and selecting for cells expressing the ouabain-resistant protein by growing the cells in the presence of ouabain (0.2, 1, or 5 µM). An active Na,K-ATPase is essential for the survival of mammalian cells due to its role in the establishment and maintenance of the electrochemical gradient across the cell membrane. By growing these transfected cells in the presence of ouabain, the endogenous Na,K-ATPase is inhibited, and only the exogenous ATPase can produce the essential cation gradient. Only if the transfected cDNA encodes for an active enzyme will cells be viable in ouabain. The mutant forms of Na,K-ATPase, E327Q and E327L, were able to transport cations, produce an electrochemical gradient, and maintain cell viability. The mutants E327D and E327A were unable to support cell viability in the presence of ouabain. The cation-dependence of Na,K-ATPase activity was examined in the presence of 1 10M ouabain for mutant proteins E327L and E327Q. The mutant E327L demonstrated a 4-fold higher K for Na and a 2-fold higher K for K than the wild type protein. The mutant E327Q displayed a 2-fold higher K for Na and a 3-6-fold higher K for K than the wild type ATPase. From these small changes in the K values, it was concluded that the charge of this glutamic acid side chain was not essential for ATPase activity. However, based on the cell viability results, the length of the side chain was suggested to be important sterically in the enzymatic cycle. Although these original site-directed mutagenesis studies revealed some important facts about the glutamic acid at position 327, the mechanistic step (i.e. cation binding or conformational changes) that the substitutions E327D and E327A disrupt was not identified, since these mutant enzymes were inactive and could not be examined.

In order to study both inactive and active mutant enzymes, we have examined the cation-enzyme interactions of Na,K-ATPase by observing the effects of Na and K on ouabain binding to the enzyme. The affinity of the Na,K-ATPase for cardiac glycosides is closely linked to enzyme cycling. Two assay environments which are known to induce high affinity ouabain binding to the enzyme are Mg, ATP, and Na or Mg and P(15, 16) . Both of these conditions promote formation of a phosphorylated form of the enzyme (E2-P) which is cation-sensitive. Thus, the role specific amino acids play in binding cations during the catalytic cycle of Na,K-ATPase may be probed by studying the effects of cations on [H]ouabain binding to mutant forms of the Na,K-ATPase.

EXPERIMENTAL PROCEDURES

Materials

Site-directed Mutagenesis

Transfection of NIH 3T3 Cells

Sequence Analysis of Genomic DNA Isolated from Clonal NIH 3T3 Cell Lines

Isolation of Crude Plasma Membranes from NIH 3T3 Cells

The crude plasma membrane preparations from NIH 3T3 cell lines were analyzed by Western immunoblotting. The proteins contained in the membrane preparations were resolved on a 10% SDS-polyacrylamide gel, electrophoretically transferred onto nitrocellulose, and stained with the monoclonal antibody M7-PB-E9 and an antimouse horseradish peroxidase-conjugated secondary antibody. M7-PB-E9 was raised against lamb kidney Na,K-ATPase and is subunit specific such that it distinguishes the transfected sheep 1 isoform from the endogenous mouse 1 subunit(23, 24) . M7-PB-E9 was a generous gift from the laboratory of Dr. W. J. Ball (University of Cincinnati).

[H]Ouabain Binding to Crude Membrane Fragments

Figure 2: Ouabain competition curves. [H]Ouabain binding was measured in the absence and presence of various concentrations of unlabeled ouabain as shown on the x axis. The symbols represent the mean of triplicate determinations. The error was calculated for each point and is shown unless it is smaller than the symbol size. Assay conditions were as described under ``Experimental Procedures,'' except for E327D which was assayed in the presence of 15 mM Tris-P and 10 mM MgCl. Symbol representation is as follows: , wild type sheep 1; , E327Q; , E327D; , E327L; and , E327A. The data were fit to a simple self-competition model:

where [I] is the concentration of unlabeled ouabain, E is the amount of enzyme, [O] is the concentration of [H]ouabain, and NS is the proportionality constant for nonspecific binding. Three adjustable parameters were calculated for each curve and include K (dissociation constant), Eand NS. The fitted parameters were calculated: wild type sheep 1 (K = 1.51 ± 0.18 nM; E = 0.106 ± 0.003 nM; NS = 0.000343 ± 0.000012); E327Q (K = 4.06 ± 0.35 nM; E = 0.130 ± 0.004 nM; NS = 0.000322 ± 0.0000096); E327D (K = 12.5 ± 0.8 nM; E = 0.268 ± 0.009 nM; NS = 0.000379 ± 0.0000146); E327L (K = 2.87 ± 0.31 nM; E = 0.142 ± 0.004 nM; NS = 0.000470 ± 0.0000143); E327A (K = 1.67 ± 0.13 nM; E = 0.110 ± 0.004 nM; NS = 0.000365 ± 0.000012).

Figure 3: Na competition curves. [H]Ouabain binding was measured for isolated membranes in the absence and presence of various concentrations of NaCl as shown on the x axis. The symbols represent the mean of duplicate determinations. The error bars represent the range of the duplicate determinations and are not shown if smaller than the symbol size. The assay conditions were as described under ``Experimental Procedures,'' except for E327D which was assayed in the presence of 15 mM Tris-P and 10 mM MgCl. The symbol representation is as follows: , wild type sheep 1; , E327Q; , E327D; , E327L; and , E327A. The curves displaying the data of E327Q and the wild type were fit to a competitive model involving three Na sites:

where K is the dissociation constant for ouabain obtained from the ouabain competition experiments (Fig. 2) and [O] is the concentration of [H]ouabain. E327D, E327L, and E327A were fit to a similar competitive equation involving two Na sites such that the fourth term in the denominator of this equation was eliminated. The two adjustable parameters calculated were: K, (Na inhibition constant) and E (amount of ouabain-binding sites). The fitted values obtained were: wild type sheep 1 (K = 15.7 ± 0.1 mM; E = 0.241 ± 0.003 nM); E327Q (K = 20.8 ± 0.3 mM; E = 0.167 ± 0.003 nM); E327D (K = 7.51 ± 0.16 mM; E = 0.359 ± 0.008 nM); E327L (K = 19.2 ± 0.4 mM; E = 0.106 ± 0.002 nM); and E327A (K = 24.2 ± 0.4 mM; E = 0.289 ± 0.003 nM).

Figure 4: K-Induced effects on ouabain binding. [H]Ouabain binding was measured in NaI-treated membranes in the presence of various concentrations of KCl as shown on the x axis. The symbols represent the mean of duplicate determinations. The error bars represent the range of the duplicate determinations and are not shown if smaller than the symbol size. The assay conditions were (5 mM P and 5 mM Mg) as described under ``Experimental Procedures.'' The assay conditions for the inset data were the same with the exception that the Mg concentration was 10 mM and the Tris-P concentration was 15 mM. The symbol representation is as follows: , wild type sheep 1; , E327Q; , E327D; , E327L; and , E327A. The amount of bound [H]ouabain (B) as a function of K concentration was calculated with the following equation:

where E is the total enzyme concentration, [O] is the concentration of [H]ouabain, is the interaction factor of inhibition, and NS is the proportionality constant for nonspecific binding. Fitted parameters for the inhibition phase were: wild type sheep 1 (at 5 mM P and 5 mM Mg) (K = 1.02 ± 0.03 mM; E = 0.115 ± 0.002 nM; = 10.1 ± 0.2); E327A (at 5 mM P and 5 mM Mg) (K = 0.879 ± 0.009 mM; E = 0.117 ± 0.003 nM; = 2.76 ± 0.11); E327L (at 5 mM P and 5 mM Mg) (K = 1.18 ± 0.07 mM; E = 0.139 ± 0.003 nM; = 3.35 ± 0.11); wild type sheep 1 (at 15 mM P and 10 mM Mg) (K = 1.17 ± 0.06 mM; E = 0.112 ± 0.002 nM; = 12.9 ± 0.3). Fitted parameters for the activation phases: E327A (K = 84 ± 28 mM; = 1.40 ± 0.31); E327D (at 5 mM P and 5 mM Mg) (AC = 2.71 ± 0.54 mM); E327D (at 15 mM P and 10 mM Mg) (AC = 2.23 ± 0.51 mM). K values for E327A were calculated by fitting both the inhibition and the activation phase of these data to the following equation:

where is the interaction factor of activation. Data describing the KCl effects on E327Q were not fit to any activation or inhibition models, and a smooth curve was drawn through the data.

[H]Ouabain Binding to Whole NIH 3T3 Cells

RESULTS

Expression of Glu Mutants

The mutations were constructed in a sheep 1 cDNA which encodes a ouabain-sensitive enzyme, and these cDNAs were transfected into NIH 3T3 cells containing an endogenous protein with a low affinity for ouabain. The exogenous proteins bind ouabain with a 1000-fold higher affinity than the endogenous enzyme. Thus, at the low concentrations of [H]ouabain utilized in these experiments, there is no interference from the binding of [H]ouabain to the endogenous Na,K-ATPase(21) . Since the exogenous protein is ouabain sensitive, no selectable function is conferred to the cells upon expression of these cDNAs; therefore, an alternative to ouabain selection was required. The mutated cDNAs were cotransfected with a gene which codes for a neomycin resistance protein and selected in 400 µg/ml of G418. cDNAs encoding all four substitutions (E327A, E327D, E327Q, and E327L) were transfected into NIH 3T3 cells, and clonal cell lines were established for each mutant. Western analysis of crude membrane preparations from these cell lines indicated that the sheep 1 proteins (wild type and mutants) were being expressed. To establish that the transfected cDNAs were integrated into the cell genome, genomic DNA was isolated from each clonal cell line. Using sheep 1-specific primers, PCR was employed to amplify a region of the DNA which encodes amino acids 230-512 of the exogenous DNA. The DNA fragment from each cell line was sequenced and revealed a single altered codon which encoded the desired amino acid at position 327. No additional mutations were detected in the DNA fragment surrounding the mutated codon.

[H]Ouabain binding to intact 3T3 cells was performed to establish that the mutant sheep 1 enzymes were located in the plasma membrane. Fig. 1presents the amount of [H]ouabain (60 nM) bound/mg of total protein for untransfected NIH 3T3 cells and for two clonal cell lines of each mutant and wild type sheep 1 Na,K-ATPase. The nonspecific binding (NS) of [H]ouabain (60 nM) to each cell line was measured in the presence of 1 µM unlabeled ouabain. From these data, one can see that the cell lines which express sheep 1 proteins (mutant or wild type) bind 5-10-fold more [H]ouabain/mg of total protein than the untransfected 3T3 cells and at least 4-fold more than nonspecific binding. The presence of [H]ouabain binding in the mutant cell lines reveals that at least some of the translated mutant sheep 1 protein is located in the plasma membranes of these clonal cell lines. Moreover, the [H]ouabain binding to intact cells indicates that the mutant proteins are assembled with an extracellular protein conformation sufficiently similar to the native sheep 1 protein to permit ouabain binding. The variation in maximum binding levels for each mutant cell line may be due to different expression levels, secondary to variable insertional sites of the transfected cDNAs in the genomic DNA. It is interesting to note that E327L has the lowest expression level in 3T3 cells as measured by [H]ouabain binding to whole cells, yet it is translated at high enough levels in HeLa cells to support viability. Thus, it does not appear that the lower expression levels of the mutant sheep 1 proteins are the result of the amino acid substitution interrupting the protein processing steps of Na,K-ATPase.

Figure 1: [H]Ouabain binding to intact NIH 3T3 cells. The amount of [H]ouabain/mg of total protein is displayed for the clonal 3T3 cells expressing wild type sheep 1 protein (WT), untransfected NIH 3T3 cells (3T3), and clonal 3T3 lines expressing E327A, E327Q, E327D, and E327L. The maximum values represent the amount of 60 nM [H]ouabain bound to intact cells in the presence of K/Ca-free buffer with respect to the total protein within these cells. Nonspecific binding (NS) to the intact 3T3 cells was also measured in identical conditions with the addition of 1 µM unlabeled ouabain. WT-NS, A-NS, Q-NS, D-NS, and L-NS display the nonspecific binding observed for the clonal cell lines expressing wild type sheep 1 Na,K-ATPase, E327A, E327Q, E327D, and E327L, respectively. The values were obtained by averaging the amounts bound from three separate experiments performed in triplicate for two clonal cell lines of each mutant.

Ouabain Competition Curves in Glu Mutants

Sodium Inhibition Curves for Wild Type and Glu Mutant Forms of Na,K-ATPase

The effects of increasing ionic strength on [H]ouabain binding were examined in a series of equilibrium binding assays with increasing concentrations of choline chloride. Similar to previous studies(26) , choline chloride concentrations above 150 mM inhibited [H]ouabain binding to both mutant and wild type Na,K-ATPase membrane preparations (data not presented). IC values obtained from fitting these data to a simple Hill equation were approximately 195-210 mM for both the wild type and the mutant forms of the ATPase. Therefore, the effects of Na and K observed at cation concentrations 100 mM can only be explained through specific interactions of these monovalent cations with the enzyme and do not reflect inhibition due to elevated ionic strength.

Potassium Competition Curves for Wild Type and Glu Mutant Forms of Na,K-ATPase

Previously, a partially competitive model for K effects on [H]ouabain binding to wild type sheep 1 protein was used to describe the incomplete inhibition of ouabain binding at saturating K concentrations(26) . The data obtained for the wild type sheep 1 enzyme and for the K inhibition phase of E327L and E327A fit this model (see Fig. 4legend). The Hill coefficients and the K values characterizing the K inhibition of ouabain binding to E327L and E327A are presented in Table 1. The potassium K values are approximately 1 mM for E327L and E327A, similar to the K value for the wild type Na,K-ATPase. Unlike the wild type and E327L proteins, higher concentrations of K (10 mM) stimulate [H]ouabain binding to the mutant E327A protein. This increase in [H]ouabain binding is characterized by a K value of 84 ± 28 mM and a Hill coefficient of 0.97 ± 0.35. Generally, it appears that both E327A and E327L bind low concentrations of K in a manner similar to the wild type (similar K values) and undergo a K-induced conformational change which reduces their affinity for [H]ouabain.

In contrast to the results with E327A and E327L, K did not inhibit [H]ouabain binding to the E327Q and E327D mutants. The effect of K on E327Q was not fit to any model. The data describing K interactions with E327D demonstrated an increase in [H]ouabain binding with increasing concentrations of K and were fit to a simple Hill equation yielding a Hill coefficient of 1.9 ± 0.3 and an AC of 3.14 ± 0.19 mM. Overall, it appears that either the ability of E327D and E327Q to bind K has been altered or that these mutants do not undergo the K-induced conformational change which normally reduces the affinity of the enzymes for [H]ouabain.

Stimulation of [H]Ouabain Binding by Mg and P

Apparent AC values for P stimulation of [H]ouabain binding in the presence of 5 mM Mg are also presented in Table 1. These values range from 0.08 to 1.25 mM. Cation competition data presented in Fig. 3and Fig. 4were obtained in 5 mM P. Thus, all the inorganic phosphate site(s) which stimulate ouabain binding were occupied for the mutant proteins E327A, E327L, and E327Q in the absence of monovalent cations. The E327D mutant demonstrated the largest change in its AC value for P which was approximately 10-fold higher than the wild type Na,K-ATPase. The ouabain, Na, and K competition curves were therefore repeated in the presence of 15 mM P and 10 mM Mg for the E327D mutant enzyme, and these calculated kinetic constants are reported in Table 1. The K competition constants were not significantly changed by the higher P and Mg concentrations for either the E327D or wild type proteins (see inset of Fig. 4). This is consistent with the concept that the inhibition of K is a direct effect and not an indirect effect due to a change in the degree of saturation by Mg and/or P. Additional K and Na inhibition curves at various Mg and P concentrations must be done to fully understand the interaction of all these ions (monovalent or divalent cations or P) with the Na,K-ATPase. Since all Mg and P activation sites were saturated in the absence of monovalent cations under these equilibrium condition, we conclude that the inhibition and activation of [H]ouabain binding observed in Fig. 2Fig. 3Fig. 4were due to ouabain, Na, and K, respectively.

DISCUSSION

The exact role that glutamic acid 327 plays in the sheep 1 Na,K-ATPase is unknown. ATPase activity studies characterizing proteins containing amino acid substitutions at this site have been limited by the nonfunctional character of some of the mutants. In this work, the cation-protein interactions of four mutant proteins, E327Q, E327D, E327L, and E327A, were examined using Na and K inhibition of [H]ouabain binding. This radiolabeling technique can in principle detect any mutation in the Na,K-ATPase which alters either the number of cations associated with the enzyme or which alters the stability of an intermediate within the Na,K-ATPase pathway which is normally sensitive to cation binding.

Effects of Substitutions on Na,K-ATPase Affinity for Ouabain

Na Inhibition of [H]Ouabain Binding

In addition to altered Hill coefficients, the K values for Na inhibition of ouabain binding were increased nearly 2-fold for E327L, E327A, and E327Q. Previously, Na interactions with E327Q and E327L were investigated using cation-dependent ATPase assays in the presence of ATP, Mg, and saturating concentrations of K(13, 14) . The K values for the Na dependence of ATPase activity for E327Q and E327L were reported to be 2-fold higher than that of the wild type ATPase. This similar increase in K values and K values for Na is consistent with the concept that the Na sites which activate the ATPase activity may be identical to the Na sites which inhibit ouabain binding. It appears that substitution of the carboxyl side chain of Glu disrupts Na binding to the enzyme by either reducing the number of Na inhibition sites (E327D, E327L, and E327A) or by increasing the inhibition constant of Na (E327Q, E327L, and E327A). No correlation exists between the inability of E327A and E327D to support cell viability and their ability to interact with Na as measured by inhibition of ouabain binding.

K Effects on [H]Ouabain Binding

The K inhibition profile for E327L is similar to that observed for wild type Na,K-ATPase. The inhibition constant for K is approximately 2-fold higher than the K value for K inhibition of the wild type and is similar to the K value determined for this mutant by K-dependent ATPase measurements (K = 1.24 ± 0.05 mM, K = 1.25 ± 0.13 mM)(13) . When the K inhibition data for E327L was fit to the partially competitive equation (see legend of Fig. 4), the interaction factor () was determined to be 3-fold lower than the for wild type. This change in the interaction factor is evident in the K inhibition profile of E327L as a higher level of [H]ouabain is bound at saturating K concentrations. This is consistent with the concept that the K-complexed mutant enzyme has a higher affinity for ouabain than does the K-complexed wild type Na,K-ATPase. Thus, upon the association of two K ions with E327L, the mutant undergoes only a portion of the conformational changes normally induced by K and retains an extracellular surface to which ouabain can bind more readily than to the wild type. E327L with two K ions bound also retains sufficient structural integrity in the membrane such that this mutant can support cell viability, possibly because other ligand-enzyme interactions compensate for the defect in the K-induced conformational change. For example, E327L exhibits a 5-fold decrease in the K constant for ATP. ()

[H]Ouabain binding to E327Q is not inhibited by K. Previously, it was reported that E327Q demonstrated a 3-6-fold increase in its K value for K-stimulated ATPase activit