Interaction between the Catalytic Site and the A-M3 Linker Stabilizes E2/E2P Conformational States of Na+,K+-ATPase*

The consequences of mutations Ile265 → Ala, Thr267 → Ala, Gly271 → Ala, and Gly274 → Ala for the partial reaction steps of the Na+,K+-ATPase transport cycle were analyzed. The mutated residues are part of the long loop (“A-M3 linker”) connecting the cytoplasmic A-domain with transmembrane segment M3. It was found that mutation Ile265 → Ala displaces the E1-E2 and E1P-E2P equilibria in favor of E1/E1P, whereas mutations Thr267 → Ala, Gly271 → Ala, and Gly274 → Ala displace these conformational equilibria in favor of E2/E2P. The mutations affect both the rearrangement of the cytoplasmic domains (seen by changes in phosphoenzyme properties and apparent ATP/vanadate affinities) and the membrane sector (indicated by change in K+/Rb+ deocclusion rate). Destabilization of E2/E2P in Ile265 → Ala, as well as a direct effect on the intrinsic affinity of the E2 form for vanadate, may be explained on the basis of the E2 crystal structures of the Ca2+-ATPase, showing interaction of the equivalent isoleucine with conserved residues near the catalytic region of the P-domain. The rate of phosphorylation from ATP was unaffected in Ile265 → Ala, indicating a lack of interference with the catalytic function in E1/E1P. The effects of mutations Thr267 → Ala, Gly271 → Ala, and Gly274 → Ala provide the first evidence in the literature of a relative stabilization of E2/E2P resulting from perturbation of the A-M3 linker region. These mutations may lead to increased strain of the A-M3 linker in E1/E1P, increased stability of the A3 helix of the A-M3 linker in E2/E2P, and/or a change of the orientation of the A3 helix, facilitating its interaction with the P-domain.

(Asp 371 in rat kidney Na ϩ ,K ϩ -ATPase) 2 located in the large cytoplasmic domain of the enzyme. The transport mechanism of the Na ϩ ,K ϩ -ATPase is often described by the simplified reaction sequence E 1 Na 3 3 E 1 P(Na 3 ) 3 E 2 PK 2 3 E 2 (K 2 ) depicted below in Scheme 1, linking, through conformational changes, the ion movement across the membrane to hydrolysis of ATP taking place in the cytoplasmic part of the molecule some 40 Å away from the intramembranous ion binding sites (1)(2)(3). The Na ϩ ,K ϩ -ATPase consists of ten transmembrane helices, M1-M10 (2,4,5), of which at least M4, M5, and M6 contribute to the cation transport pathway (2, 6 -10). The transmembrane part is linked to a cytoplasmic head piece made up by three domains, denoted A-(actuator), N-(nucleotide-binding), and P-(phosphorylation) domains in the terminology based on the high resolution crystal structure of the closely related Ca 2ϩ -ATPase (2,5,11). The A-domain consists of the N-terminal portion of the polypeptide chain in addition to the loop between M2 and M3, whereas the N-and P-domains are composed of the loop between M4 and M5 with the N-domain being inserted into the P-domain (11). Comparing the atomic structures of the Ca 2ϩ -ATPase in Ca 2ϩ -bound E 1 forms (E 1 (Ca 2 )) (11) and E 1 (Ca 2 )⅐Mg⅐AlF 4 ⅐ADP (12)) with those of the Ca 2ϩ -free E 2 forms stabilized by thapsigargin (E 2 (TG)) (13) and E 2 (TG)⅐Mg⅐MgF 4 (14)), it appears that the cytoplasmic domains undergo large rearrangements in relation to the E 1 -E 2 conformational changes, and a key event in active transport may be the extensive rotation of the A-domain parallel to the membrane. Because the A-domain is linked to the transmembrane domain through long cytoplasmic extensions of M1, M2, and M3, it is likely that these linker segments play critical roles in rotation of the A-domain and are essential for the interaction of the A-domain with the P-domain, with concomitant positioning of the conserved A-domain motif 214 TGES 217 in the vicinity of the catalytic site (15)(16)(17) and long range communication with the ion binding sites in the membrane.
Long before the appearance of the Ca 2ϩ -ATPase crystal structures, proteolytic cleavage experiments with Na ϩ ,K ϩ -ATPase had demonstrated that cleavage sites in the cytoplasmic extension linking M3 to the A-domain ("A-M3 linker") are alternately exposed in relation to the E 1 -E 2 and E 1 P-E 2 P conformational transitions. Functional characterization of the chymotrypsin cleaved enzyme showed a block of the translocation of Na ϩ , due to stabilization of the E 1 P(Na 3 ) intermediate (18,19). Proteolytic cleavage in the corresponding region of the Ca 2ϩ -ATPase with excision of a small five amino acids sequence ( 239 MAATE 243 ) likewise seems to affect the conformational changes (20). The cleaved enzyme retained the ability to bind Ca 2ϩ and react with ATP, but the transition from E 1 P to E 2 P was blocked (20). The functional consequences of mutation * This work was supported by grants from the Danish Medical Research Council, the Novo Nordisk Foundation, Denmark, the Lundbeck Foundation, Denmark, and the Research Foundation of Aarhus University. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
in the Na ϩ ,K ϩ -ATPase of Gly 263 next to the proteolytic cleavage point, or mutation of the homologous counterpart, Gly 233 , in the Ca 2ϩ -ATPase, are consistent with effects on the E 1 -E 2 and E 1 P-E 2 P conformational equilibria (21,22). It is therefore very interesting that the crystal structures of the Ca 2ϩ -ATPase show large structural differences in this region. In the E 2 forms (13,14), the whole region connecting domain A with M3 has moved from a peripheral to a more central position closer to domain P, and the rearrangements include the formation of an ␣-helix in the E 2 form, involving residues immediately C-terminal to Gly 233 (Gly 263 of Na ϩ ,K ϩ -ATPase).
In this report we have investigated the functional consequences of mutation in this region further. Ile 265 , with its bulky hydrophobic side chain, and Thr 267 were both replaced with alanine. To examine the role of the conformation of the peptide backbone, we furthermore replaced the two glycines Gly 271 and Gly 274 with alanine. Neither of the residues selected for the present study, or their equivalents in other P-type ATPases, have previously been studied by mutagenesis. The isoleucine is particularly well conserved, being present in Na ϩ ,K ϩ -ATPases, sarco(endo)plasmic reticulum Ca 2ϩ -ATPases, and H ϩ ,K ϩ -ATPases. To examine the functional importance of these residues in Na ϩ ,K ϩ -ATPase, we have studied the effects of the mutations on the overall and partial reactions of the enzyme, using steady-state and transient kinetic measurements.

EXPERIMENTAL PROCEDURES
Mutagenesis, Expression, and Enzyme Preparation-Site-directed mutagenesis of the ouabain-resistant rat ␣ 1 -isoform of Na ϩ ,K ϩ -ATPase was carried out by the Kunkel method (23) or by using the QuikChange site-directed mutagenesis kit (Stratagene) to introduce the desired mutations directly into full-length cDNA. The mutant or wild-type cDNA was expressed in COS-1 cells, using 5 M ouabain in the growth medium to select stable transfectants, and a crude plasma membrane fraction was isolated from the cells and made leaky by treatment with sodium deoxycholate or alamethicin (8,24,25).
Functional Analysis-To eliminate the contribution of the endogenous ouabain-sensitive COS-1 cell enzyme (ϳ10% of the total Na ϩ ,K ϩ -ATPase content of the preparation), 10 M ouabain was added to all assays except for the one in which the affinity for ouabain was determined by variation of the ouabain concentration. For the phosphorylation assays, the enzyme was moreover preincubated in the presence of ouabain.
ATP hydrolysis was measured at 37°C as described previously (24,25). The Na ϩ ,K ϩ -ATPase activity associated with the expressed exogenous enzyme was calculated by subtracting the background ATPase activity measured at 10 mM ouabain from the ATPase activity measured at 10 M ouabain.
Studies of the Na ϩ dependence of steady-state phosphorylation from [␥-32 P]ATP and the time course of dephosphorylation, as well as the determination of the active site concentration by phosphorylation in the presence of 150 mM NaCl and oligomycin (20 g/ml) to inhibit dephosphorylation, were carried out at 0°C (26). Deocclusion of K ϩ or Rb ϩ was studied in phosphorylation experiments at 10°C, following formation of K ϩ -or Rb ϩ -occluded enzyme at room temperature as described previously (21,27,28). Rapid kinetic phosphorylation experiments at 25°C with enzyme present in the E 1 Na 3 form were performed using a Bio-Logic quench-flow module according to the previously described "Protocol 1" (21,29). In all the above-described phosphorylation experiments, background phosphorylation was determined in the presence of 50 mM KCl without NaCl.
To examine the vanadate affinity under equilibrium conditions, 10 g of deoxycholate-treated plasma membranes was incubated at 20°C for 30 min in 40 l of medium containing 20 mM Tris (pH 7.5), 5 mM where EP max is the phosphorylation level obtained in the absence of vanadate, and EP ϱ is the phosphorylation level corresponding to infinite vanadate concentration. Data Analysis and Statistics-Data normalization, averaging, and nonlinear regression analysis were carried out as previously (21). The lines in the figures show the best fit to the complete set of normalized data, and the extracted parameters with standard errors are indicated in the tables. ATP and vanadate dependences of the ATPase activity and the Na ϩ dependence of phosphorylation were analyzed using the Hill equation. In the analysis of the ouabain dependence of ATPase activity, the ouabain-inhibited enzyme was represented by the sum of two Hill functions, one corresponding to the exogenous rat enzyme and one corresponding to the endogenous COS-1 cell Na ϩ ,K ϩ -ATPase. The time course of K ϩ deocclusion was analyzed using the biphasic time function described previously (21,27) in which the component corresponding to the rapid phase is at maximum from the beginning. The time dependence of phosphorylation of enzyme in the E 1 Na 3 form was fitted by using a mono-exponential function, and the dephosphorylation time courses following chase with ATP or ADP were fitted by using a bi-exponential function.

Expression, Ouabain Sensitivity, and Catalytic Turnover
Rate-The amino acid substitutions Ile 265 3 Ala, Thr 267 3 Ala, Gly 271 3 Ala, and Gly 274 3 Ala were introduced into the ouabain-resistant rat kidney ␣ 1 -isoform of the Na ϩ ,K ϩ -ATPase, and the mutant enzymes were expressed in COS-1 cells as previously (8,24,25). As in our earlier studies, we used a culture medium containing 5 M ouabain for selection of the stable COS-1 cell lines expressing the mutant ␣ 1 enzymes. This is feasible, provided the mutant enzyme is functional, because the rat kidney Na ϩ ,K ϩ -ATPase is less sensitive to ouabain (K 0.5 for ouabain inhibition Ͼ 100 M, cf. Table I) than the endogenous COS-1 cell Na ϩ ,K ϩ -ATPase (K 0.5 0.3-0.8 M) (24,25). For all mutants the Na ϩ ,K ϩ -transport rate was high enough to support cell growth. The mutants were expressed to an active site concentration (determined by phosphorylation with oligomycin included to maximally stabilize the phosphoenzyme) similar to that determined for the wild-type Na ϩ ,K ϩ -ATPase, ranging between 30 and 60 pmol/mg of total membrane protein. This high expression level enabled us to characterize the functional properties of the mutants by measuring the overall function and by using various phosphorylation and dephosphorylation protocols to provide kinetic information on the individual steps in the catalytic cycle.
The ouabain concentration dependence of Na ϩ ,K ϩ -ATPase activity showed that the K 0.5 value for ouabain inhibition of mutants Thr 267 3 Ala, Gly 271 3 Ala, and Gly 274 3 Ala was very similar to that of the wild type, whereas mutant Ile 265 3 Ala displayed a 2.7-fold increased K 0.5 value (Table I). To avoid any contribution from the endogenous COS-1 cell enzyme, all measurements of ATPase activity and phosphoenzyme described below were carried out in the presence of 10 M ouabain.
The catalytic turnover rate was calculated as the ratio between the maximum Na ϩ ,K ϩ -ATPase activity (determined at 37°C and saturating substrate concentrations) and the active site concentration. As seen in Table I, the catalytic turnover rate was reduced to 61% in mutant Ile 265 3 Ala, and to 80 and 86%, relative to wild type, in Thr 267 3 Ala and Gly 271 3 Ala, respectively, whereas it was wild type-like in Gly 274 3 Ala.
ATP Dependence of Na ϩ ,K ϩ -ATPase Activity- Fig. 1 shows results of experiments in which the ATP dependence of the Na ϩ ,K ϩ -ATPase activity was studied. The K 0.5 value for ATP extracted from the data is listed in Table I. Relative to the wild type, Ile 265 3 Ala exhibited a 5.3-fold increase of the apparent affinity for ATP (K 0.5 reduced), whereas in mutants Thr 267 3 Ala, Gly 271 3 Ala, and Gly 274 3 Ala, the apparent affinity for ATP was reduced (K 0.5 increased 1.1-to 1.6-fold). In the wildtype Na ϩ ,K ϩ -ATPase, ATP binds to E 2 (K 2 ) with low affinity, accelerating the rate-limiting transition between E 2 (K 2 ) and E 1 Na 3 , cf. Scheme 1, boxed ATP (3,31). The finding that mutant Ile 265 3 Ala exhibits an increased apparent affinity for ATP in the activation of Na ϩ ,K ϩ -ATPase activity, relative to wild type, may be explained by an acceleration of E 2 (K 2 ) 3 E 1 Na 3 induced by the mutation independently of ATP. Alternatively, the Ile 265 3 Ala mutation might increase the intrinsic affinity of the low affinity binding site on E 2 (K 2 ) for ATP. For mutants Thr 267 3 Ala, Gly 271 3 Ala, and Gly 274 3 Ala, on the other hand, the data are consistent with either a stabilization of the K ϩ -occluded intermediate, E 2 (K 2 ), occurring independently of ATP, or a reduced intrinsic affinity of E 2 (K 2 ) for ATP.
K ϩ Occlusion-From the data in Fig. 1 it can be calculated that an increase of ATP concentration from 10 M to 3 mM enhances the turnover rate 37-fold in the wild type, whereas the enhancement is only 8.5-fold in Ile 265 3 Ala. At 10 M ATP, the turnover rate is 230 and 606 min Ϫ1 for the wild type and Ile 265 3 Ala, respectively. At this low ATP concentration, K ϩ deocclusion, E 2 (K 2 ) 3 E 1 ϩ 2K ϩ , is rate-limiting for the overall ATPase reaction. Therefore, the increased turnover rate of Ile 265 3 Ala, relative to wild type, determined under these conditions suggests that the rate of K ϩ deocclusion is increased in the mutant.
To examine the rate of K ϩ deocclusion directly, the previously described phosphorylation assay was applied (21,27,28). Briefly, the enzyme is equilibrated with K ϩ in the absence of Na ϩ and ATP to form E 2 (K 2 ). The phosphorylation kinetics is then determined following initiation of the phosphorylation reaction by a 10-fold dilution of the enzyme in a solution containing 100 mM Na ϩ and 1 M [␥-32 P]ATP. Oligomycin is added just prior to phosphorylation to prevent decay of the phosphoenzyme. This procedure allows estimation of the level of E 2 (K 2 ) initially present, as well as the rate constant for the K ϩ deocclusion reaction E 2 (K 2 ) 3 E 1 ϩ 2K ϩ . The data presented in Fig. 2 were analyzed by fitting a biphasic time function (21,27), in which the ordinate intercept reflects the non-occluded enzyme pool ready to bind Na ϩ and phosphorylate immediately, whereas the exponential part (shown by the line in Fig. 2) corresponds to deocclusion of K ϩ from E 2 (K 2 ). The fraction of enzyme initially present as E 2 (K 2 ) equals 100% minus the ordinate intercept. The parameters determined by this procedure are listed in Table II for all mutants. As is seen in Fig. 2 (upper panel) and Table II, Ile 265 3 Ala differed significantly from the other mutants and wild type with respect to both the deocclusion rate (9-fold enhanced, relative to wild type) and the level of the K ϩ -occluded intermediate initially present (65%  Table I.  I Na ϩ ,K ϩ -ATPase turnover rate and apparent affinities determined by measurement of ligand concentration dependences of Na ϩ ,K ϩ -ATPase activity a Determined by ouabain titration of Na ϩ ,K ϩ -ATPase activity. Measurements of ATP hydrolysis were carried out at 37°C in the presence of 130 mM NaCl, 20 mM KCl, 3 mM ATP, 3 mM MgCl 2 , 30 mM histidine buffer (pH 7.4), 1 mM EGTA, and various concentrations of ouabain. A function with the ouabain-inhibited enzyme represented by the sum of two hyperbolic components, one corresponding to the expressed exogenous rat enzyme and one corresponding to the endogenous COS-1 cell Na ϩ ,K ϩ -ATPase, was fitted to the data points (average values corresponding to three experiments). The K 0.5 value determined for the endogenous enzyme was Յ1 M, and the K 0.5 value of the expressed exogenous enzyme is listed here.
b The catalytic turnover rate calculated as the ratio between the maximal Na ϩ ,K ϩ -ATPase activity (measured in the presence of 3 mM ATP and optimal concentrations of Na ϩ and K ϩ , and 10 M ouabain) and the corresponding active site concentration determined on the same membrane preparation by phosphorylation at 0°C with ͓␥-32 P͔ATP in the presence of the optimal Na ϩ concentration together with oligomycin (20 g/ml). versus 88% for wild type). Because the latter parameter corresponds to enzyme without ATP, it may be concluded that the Ile 265 3 Ala mutation induces a shift of the conformational equilibrium in favor of E 1 independent of ATP. For Thr 267 3 Ala, Gly 271 3 Ala, and Gly 274 3 Ala, the amount of E 2 (K 2 ) was increased relative to wild type, amounting to 90 -97%, and the rate constant for the deocclusion reaction was reduced correspondingly (1.4-to 1.6-fold) relative to wild type.
For Ile 265 3 Ala, additional experiments similar to those described above were carried out, in which the enzyme was equilibrated with 8 mM Rb ϩ in place of 1 mM K ϩ (Fig. 2, lower  panel). Under these conditions, the level of occluded enzyme (E 2 (Rb 2 )) initially present was higher, now constituting as much as 90% in the mutant versus 99% in the wild type. The rate of deocclusion was 15-fold increased in the mutant, relative to wild type (Table II).
Vanadate Dependence of Na ϩ ,K ϩ -ATPase Activity-Vanadate inhibits pump activity by binding in competition with P i to the phosphorylation site of the E 2 and E 2 (K 2 ) conformations, thereby forming a stable dead-end state, probably resembling the E 2 ⅐P i complex from which phosphate normally is released (32). To obtain information about the steady-state concentration of E 2 /E 2 (K 2 ) accumulated during ATP hydrolysis, we investigated the vanadate dependence of Na ϩ ,K ϩ -ATPase activity. As shown in Fig. 3 and Table I, Ile 265 3 Ala was found ϳ23-fold less sensitive to vanadate inhibition than wild-type Na ϩ ,K ϩ -ATPase. By contrast, the other mutants showed higher sensitivity to vanadate inhibition than the wild type, 2-fold for Thr 267 3 Ala and Gly 274 3 Ala and about 1.4-fold for Gly 271 3 Ala. This result is consistent with an increased accumulation of E 2 (K 2 ), caused by the decrease in the K ϩ deocclusion rate described above for these mutants. The reduced sensitivity of Ile 265 3 Ala to vanadate inhibition is consistent with the depletion of the E 2 (K 2 ) intermediate during enzyme cycling caused by the increase in the K ϩ deocclusion rate. Another contributing factor could be a direct effect of the mutation on the vanadate-binding site, i.e. a change of the intrinsic affinity for vanadate. This was examined in the experiments described in the next section.
Vanadate Binding at Equilibrium-To gain further insight into the effect of the Ile 265 3 Ala mutation on the vanadatebinding properties of the E 2 form and the equilibrium between E 2 and E 1 , we determined the vanadate affinity in a phosphorylation assay allowing the enzyme⅐vanadate complex to be formed under equilibrium conditions without enzyme cycling (16). The enzyme is incubated at 20°C with various concentrations of vanadate in the absence of Na ϩ and ATP and for an extended period of time to attain equilibrium. Subsequently phosphorylation from ATP is carried out in the presence of Na ϩ . The rationale is that no high affinity ATP binding can take place when vanadate is bound to the enzyme. Consequently, the phosphorylation level equals the enzyme fraction that has not reacted with vanadate during the preincubation period and, thus, reflects the equilibrium between the free and vanadate-bound enzyme forms existing before initiation of phosphorylation. Care is taken to minimize dissociation of vanadate from the preformed enzyme⅐vanadate complex during the phosphorylation reaction by reducing the temperature to 0°C and keeping the phosphorylation time short, yet long enough to ensure full phosphorylation of all vanadate-free enzyme.
Such experiments were performed with Ile 265 3 Ala and wild type (Fig. 4), and the parameters, determined assuming a simple one-site binding model (see "Experimental Procedures"), are summarized in Table III. The equilibrium binding affinity of wild type for vanadate was found about 30-fold higher than the apparent affinity determined during enzyme cycling with ATP present. This is presumably due to the lack of competition from ATP binding and phosphorylation during the equilibration with vanadate. Fig. 4 (upper panel) shows the results obtained following equilibration with vanadate in the absence of any of the transported cations. Under these conditions, Ile 265 3 Ala exhibited a 28-fold reduction of the apparent affinity for vanadate relative to wild type. As seen in Fig. 4 (lower panel) the vanadate affinity was also determined in the presence of 8 mM Rb ϩ , i.e. conditions where 90% of the enzyme is E 2 (Rb 2 ) as revealed in the measurements of occlusion and deocclusion of Rb ϩ discussed above (Fig. 2, lower panel), thereby eliminating the influence of the shift of the E 1 -E 2 equilibrium induced by the mutation. Under these conditions mutant Ile 265 3 Ala still displayed a 5-fold reduced affinity for vanadate relative to wild type. This indicates that the Ile 265 3 Ala mutation affects not only the E 1 -E 2 conformational equilibrium, but also the intrinsic affinity of E 2 for vanadate.
K ϩ and Na ϩ Affinities-The mutational effects on K ϩ bind- Each line shows the best fit to the complete set of normalized data of the time function described previously (21,27), in which the ordinate intercept reflects the non-occluded enzyme pool ready to bind Na ϩ and phosphorylate immediately, and the slow, exponential part corresponds to deocclusion of K ϩ /Rb ϩ . The deocclusion rate and the amount of E 2 (K 2 ) or E 2 (Rb 2 ) initially present are listed in Table  II along with the standard errors. ing at the extracellularly facing sites and Na ϩ binding at the cytoplasmically facing sites were examined by studying the K ϩ dependence of Na ϩ ,K ϩ -ATPase activity and the Na ϩ dependence of phosphorylation from ATP (in the absence of K ϩ ). The K 0.5 (K ϩ ) and K 0.5 (Na ϩ ) values extracted from the K ϩ -and Na ϩ -activation profiles are listed in Tables I and II, respectively. There was little difference between the mutants and the wild type with respect to apparent affinity for extracellular K ϩ . The apparent affinity for cytoplasmic Na ϩ was slightly increased (1.6-fold) in Ile 265 3 Ala relative to wild type and was wild type-like in Thr 267 3 Ala, Gly 271 3 Ala, and Gly 274 3 Ala. Because only the E 1 form binds cytoplasmic Na ϩ with high affinity, the effect of the Ile 265 3 Ala mutation on the apparent Na ϩ affinity is consistent with a shift of the E 1 -E 2 conformational equilibrium of the dephosphoenzyme in favor of E 1 .
Time Course of Phosphorylation with ATP-The phosphoryl-ation reaction E 1 Na 3 3 E 1 P(Na 3 ) and the ATP affinity of the E 1 Na 3 intermediate were tested in experiments studying the time course of phosphorylation with 2 M [␥-32 P]ATP at 25°C in the presence of oligomycin to inhibit dephosphorylation, using the quench-flow module QFM-5 as described previously (21,29). The presence of oligomycin reduces the dephosphorylation rate, because of the stabilizing effect exerted by oligomycin on the E 1 P(Na 3 ) form (1). For all mutants, the observed rate constant of phosphorylation was very similar to the value of 27 s Ϫ1 of the wild-type enzyme (Table II). Because the phosphorylation reaction is less than half saturated in the wild type in the presence of 2 M ATP (21,29), any difference between mutant and wild type with respect to the affinity of E 1 Na 3 for ATP should be reflected in the observed rate constant of phosphorylation. Thus, the phosphorylation reaction and the binding of ATP to E 1 Na 3 seem to be unaffected by the mutations. Oligomycin Effect on Phosphoenzyme Level- Table II moreover shows the steady-state phosphorylation level obtained with [␥-32 P]ATP at low temperature (0°C) and a saturating Na ϩ concentration of 150 mM in the absence of oligomycin as a percentage of the phosphorylation level reached in the presence of oligomycin. The presence of oligomycin ensures a level of phosphoenzyme equal to the active site concentration, because of the reduced dephosphorylation rate. It is seen that the level of phosphoenzyme reached in the wild type in the absence of oligomycin constituted 82% of that reached in the presence of oligomycin. In mutant Ile 265 3 Ala, the ratio between the phosphorylation levels reached in the absence and presence of oligomycin was close to 1, whereas in the other mutants it was wild type-like. The data suggest that the Ile 265 3 Ala mutation mimics the effect of oligomycin by reducing the dephosphorylation rate. This hypothesis was confirmed in further studies of the characteristics of the phosphoenzyme as described below.
ADP Sensitivity of the Phosphoenzyme-The phosphoenzyme consists of two major forms designated E 1 P and E 2 P (3, 33). E 1 P is K ϩ -insensitive but ADP-sensitive, i.e. able to donate the phosphoryl group back to ADP forming ATP. E 2 P is ADPinsensitive and K ϩ -sensitive. Hence, the presence of K ϩ at the extracellularly facing sites induces rapid hydrolysis of E 2 P. Fig.  5 presents results of experiments in which the ADP sensitivity was studied. Phosphorylation with [␥-32 P]ATP was carried out at conditions known to promote accumulation of the E 2 P inter-  Table I.  From Fig. 2, calculated as 100% minus the ordinate intercept of the fitted line corresponding to the slow phase. Data obtained with Rb ϩ are shown in parentheses.
b From Fig. 2, rate constant corresponding to the slow phase indicated by the fitted line. Data obtained with Rb ϩ are shown in parentheses. c Determined by Na ϩ titration of phosphorylation by ͓␥-32 P͔ATP. Phosphorylation was carried out for 15 s at 0°C in a medium containing 2 M ͓␥-32 P͔ATP, 20 mM Tris (pH 7.5), 3 mM MgCl 2 , 1 mM EGTA, 10 M ouabain, oligomycin (20 g/ml), and various concentrations of NaCl and N-methyl-D-glucamine (keeping the ionic strength constant at 150 mM). A Hill function was fitted to the data points (average values corresponding to at least four independent experiments), giving the K 0.5 and Hill numbers listed. d Determined in phosphorylation experiments at 25°C, using a Bio-Logic quench-flow module. Phosphorylation was carried out in the presence of 2 M ͓␥-32 P͔ATP and oligomycin according to "Protocol 1" previously described (21,29). The enzyme was preincubated in a medium containing 100 mM NaCl, 20 mM Tris (pH 7.5), 3 mM MgCl 2 , 1 mM EGTA, 10 M ouabain, and oligomycin (20 g/ml), and mixed with an equal volume of the same buffer containing 4 M ͓␥-32 P͔ATP, followed by acid quenching at various time intervals. A mono-exponential function was fitted to the data points (average values corresponding to at least two independent experiments), giving the rate constants listed.
e Ratio between phosphoenzyme levels obtained without and with oligomycin (20 g/ml) following incubation with 2 M ͓␥-32 P͔ATP for 15 s at 0°C in the presence of 150 mM NaCl, 20 mM Tris (pH 7.5), 3 mM MgCl 2 , 1 mM EGTA, and 10 M ouabain. f From Fig. 5. g From Fig. 6. h ND, not determined. mediate in the wild type (presence of 20 mM Na ϩ and absence of K ϩ ), and the time course of dephosphorylation was followed upon addition of ADP. The extents of the rapid and slow decay components, which reflect the initial amounts of the ADPsensitive E 1 P and the ADP-insensitive E 2 P, respectively, were estimated by fitting a bi-exponential function to the data points as described previously (21,29). The results of this analysis are listed in Table II for all mutants. It is seen that for wild-type Na ϩ ,K ϩ -ATPase 30% of the phosphoenzyme was ADP-sensitive E 1 P. For Ile 265 3 Ala, the E 1 P level was significantly higher, 89%, i.e. a 3-fold increase relative to wild type. For Thr 267 3 Ala, Gly 271 3 Ala, and Gly 274 3 Ala, on the other hand, the E 1 P level was only 14, 14, and 19%, respectively (Table II), indicative of a shift in the distribution between E 1 P and E 2 P in favor of E 2 P. It is furthermore interesting to note that the rate constant corresponding to the slow phase, reflecting the E 2 P dephosphorylation, was increased 1.6-to 1.7-fold in Thr 267 3 Ala and Gly 274 3 Ala, whereas in Gly 271 3 Ala it was wild type-like (for rate constants, see legend to Fig. 5).
Kinetics of E 1 P Phosphoenzyme Turnover-Because the above-described findings reveal a shift in the distribution between E 1 P and E 2 P in the mutants, further experiments were conducted examining the rate of the E 1 P 3 E 2 P interconversion ( Fig. 6 and Table II). To promote accumulation of the E 1 P form even in the wild type, phosphorylation was carried out in the presence of a high NaCl concentration (21,26). More than 90% of the phosphoenzyme formed under these conditions is the ADP-sensitive E 1 P intermediate, as revealed by its decay within 2 s upon addition of ADP (data not shown). Dephosphorylation proceeding in the forward direction through the E 1 P 3 E 2 P transition and the subsequent dephosphorylation of E 2 P was then monitored following addition of a solution producing a final concentration of Na ϩ of either 200 mM (Fig. 6, upper panel) or 600 mM (Fig. 6, lower panel), 1 mM unlabeled ATP, and 20 mM KCl (to ensure rapid dephosphorylation of E 2 P such that E 1 P 3 E 2 P is rate-limiting for the decay of E 1 P, cf. Scheme 1). Under both sets of dephosphorylation conditions, the halflife (T 0.5 ) was increased in Ile 265 3 Ala, corresponding to a 10-fold reduction in the rate of E 1 P 3 E 2 P interconversion at 200 mM NaCl and an 11-fold reduction at 600 mM NaCl. By contrast, in Thr 267 3 Ala, Gly 271 3 Ala, and Gly 274 3 Ala the T 0.5 for dephosphorylation was shortened relative to wild type, 4-, 2.7-, and 1.5-fold, respectively, at 600 mM NaCl, indicating that the E 1 P 3 E 2 P conversion is accelerated in these mutants, relative to wild type. DISCUSSION In the present study, we have examined the functional consequences of alanine substitution of either of the residues Ile 265 , Thr 267 , Gly 271 , and Gly 274 in the A-M3 linker segment of the Na ϩ ,K ϩ -ATPase. Our results demonstrate that Ile 265 3  Table III.  From Fig. 4, upper panel. b From Fig. 4, lower panel. Ala displaces the E 1 -E 2 and E 1 P-E 2 P equilibria in favor of E 1 /E 1 P, whereas Thr 267 3 Ala, Gly 271 3 Ala, and Gly 274 3 Ala displace these conformational equilibria in the opposite direction, in favor of E 2 /E 2 P. The mutations affect both the rearrangement of the cytoplasmic domains (seen by changes in phosphoenzyme properties and apparent ATP/vanadate affinities) and the conformational change of the membrane sector (indicated by the change of K ϩ /Rb ϩ deocclusion rate).
Hence, Ile 265 3 Ala led to increased apparent affinity for ATP, reduced apparent affinity for vanadate, reduced equilibrium concentration of the occluded E 2 (K 2 ) form even in the absence of ATP, increased K ϩ deocclusion rate, and increased apparent affinity for Na ϩ in activation of phosphorylation, i.e. properties consistent with a destabilization of the E 2 form relative to E 1 . Moreover, the increased fractional amount of the ADP-sensitive E 1 P phosphoenzyme intermediate and the 10fold reduced rate of decay of E 1 P imply that the E 1 P-E 2 P equilibrium is displaced in favor of E 1 P in Ile 265 3 Ala. The reduced apparent affinity for ouabain seen for Ile 265 3 Ala may accordingly be interpreted in terms of a reduced level of E 2 P, because ouabain is thought to bind preferentially to E 2 P (34).
By contrast, the Thr 267 , Gly 271 , and Gly 274 mutations led to decreased apparent affinity for ATP, increased apparent affin-ity for vanadate, increased equilibrium level of the occluded E 2 (K 2 ) form even in the absence of ATP, and decreased K ϩ deocclusion rate, i.e. properties consistent with a stabilization of the E 2 form relative to E 1 . These mutations furthermore increased the relative amount of the ADP-insensitive E 2 P phosphoenzyme intermediate and enhanced the rate of decay of E 1 P, indicating destabilization of E 1 P. Notably, this is the first evidence in the literature of a displacement of the E 1 -E 2 and E 1 P-E 2 P equilibria in favor of E 2 /E 2 P resulting from perturbation of the A-M3 linker region. Previously reported effects of proteolytic cleavage or mutation in this region are all consistent with an accumulation of E 1 /E 1 P similar to that seen here for Ile 265 3 Ala (18 -22).
The apparent affinity of Ile 265 3 Ala for the inhibitor vanadate, which binds preferentially to the E 2 form (32), was found 28-fold decreased, relative to wild type, under equilibrium conditions where the enzyme is not cycling. This effect is more pronounced than the 9-to 15-fold increase of the rate of E 2 3 E 1 determined in the K ϩ and Rb ϩ deocclusion experiments. Moreover, the affinity for vanadate was still found 5-fold reduced, relative to wild type, when determined in the presence of a high concentration of Rb ϩ , forcing the enzyme into the E 2 form and, thereby, eliminating the influence of the shift of the E 1 -E 2 equilibrium induced by the mutation. This behavior differs from that of the previously described Gly 263 3 Ala mutant, which showed normal vanadate affinity upon the addition of Rb ϩ (16), as expected for a mutation affecting only the conformational equilibrium and not the intrinsic affinity. The retention of a significant reduction of vanadate affinity in Ile 265 3 Ala, even after the enzyme had been forced into the E 2 form, suggests that part of the interference of the mutation with vanadate binding owes to a direct effect on the intrinsic affinity of the E 2 form for vanadate, similar to the effect of mutation of Thr 214 in the highly conserved TGES loop that becomes an integral part of the catalytic site in the E 2 /E 2 P forms (16). Hence, it may possibly be inferred that Ile 265 is close to the catalytic site in the E 2 /E 2 P forms. On the other hand, the normal rate of phosphorylation from ATP indicates that the Ile 265 3 Ala mutation does not interfere with the function of the catalytic site in E 1 /E 1 P.
Topology analysis, using site-directed chemical labeling and other techniques, has provided considerable evidence that the 30% identity and 65% similarity in amino acid sequence of sarco(endo)plasmic reticulum Ca 2ϩ -ATPase and Na ϩ ,K ϩ -ATPase results in similar overall folds of these proteins (2,5). Hence, although caution clearly is needed when interpreting the Na ϩ ,K ϩ -ATPase data in light of the Ca 2ϩ -ATPase crystal structures, the Ca 2ϩ -ATPase crystal structures may provide useful provisional models for understanding Na ϩ ,K ϩ -ATPase structure-function relationship. Because Ile 265 is conserved between Na ϩ ,K ϩ -ATPase and sarco(endo)plasmic reticulum Ca 2ϩ -ATPase, the relations of this residue in the Ca 2ϩ -ATPase crystal structures are particularly relevant to understanding the present findings. The other residues studied here are less well conserved. Thr 267 is replaced either by the closely related serine or by aspartate in the various sarco(endo)plasmic reticulum Ca 2ϩ -ATPase isoforms, and the Ca 2ϩ -ATPase counterparts of Gly 271 and Gly 274 are alanine and glutamine/proline, respectively. Fig. 7 shows the structural models of the crystalline E 1 and E 2 forms of Ca 2ϩ -ATPase (11)(12)(13)(14) with the Na ϩ ,K ϩ -ATPase residues Ile 265 , Thr 267 , Gly 271 , and Gly 274 mutated in the present study indicated in parentheses following the corresponding Ca 2ϩ -ATPase residues (shown in yellow). Interestingly, in the E 2 forms (Fig. 7, lower panels) (13), and 1WPG (lower right panel) (14) are shown. In the left panels, Ca 2ϩ -ATPase residue numbering is shown with the numbers of the corresponding Na ϩ ,K ϩ -ATPase residues in parentheses. Residues equivalent to the Na ϩ ,K ϩ -ATPase residues mutated in the present study are highlighted in yellow. Asn 706 of the catalytic site, which is equivalent to Asn 715 in Na ϩ ,K ϩ -ATPase, is highlighted in purple. Val 705 , equivalent to Val 714 in Na ϩ ,K ϩ -ATPase, is highlighted in blue. The white residues in the transmembrane part are those thought to coordinate the transported ions. The dashed ellipse indicates the catalytic site, including the AlF 4 (green molecule) mimicking the phosphoryl transition state. In the lower right panel, MgF 4 , mimicking P i , is shown with the Mg atom red and the fluorine atoms green. Na ϩ ,K ϩ -ATPase (blue) and the asparagine equivalent to Asn 715 in Na ϩ ,K ϩ -ATPase (purple), thus allowing van der Waals interaction, whereas the distance is as large as 25 Å in the E 1 structures (Fig. 7, upper panels). The asparagine is conserved in all P-type ATPases and related phosphatases (35) and is part of the catalytic site as indicated by the dashed ellipse in the E 1 (Ca 2 )⅐Mg⅐AlF 4 ⅐ADP crystal structure in Fig. 7. Both in the latter structure and in E 2 (TG)⅐Mg⅐MgF 4 (Fig. 7, lower right  panel), thought to mimick E 2 ⅐MgP i , the asparagine is seen to interact with a fluorine atom corresponding to an oxygen atom of the phosphoryl group in the catalytic site. Interference with the van der Waals interaction of the isoleucine with residues in/close to the catalytic site (Val 714 and Asn 715 in Na ϩ ,K ϩ -ATPase) in the E 2 conformation seems to explain the presently observed effect of replacement of Ile 265 with the smaller alanine on the intrinsic affinity for vanadate. Furthermore, disruption of the interaction of Ile 265 with Val 714 and Asn 715 may explain the destabilization of E 2 /E 2 P induced by substitution of Ile 265 .
Although very close to Ile 265 in the primary structure, Thr 267 apparently plays a different role according to our functional analysis. This difference may also be understood by consideration of the Ca 2ϩ -ATPase E 2 crystal structures, where the corresponding residue, Asp 237 , is seen to point away from the catalytic site toward the water phase. Similarly, the residues corresponding to the presently mutated Gly 271 and Gly 274 are located in a region that has access to the water phase and is far away from the catalytic site.
Another feature of relevance for interpretation of the present results is the difference between the E 2 and E 1 forms with respect to the ␣-helix content of the A-M3 linker segment. In the E 2 forms, a large ␣-helix (third helix of the A-domain, denoted "A3 helix") is present in the region containing the residues equivalent to Na ϩ ,K ϩ -ATPase residues Ile 265 and Thr 267 , with the alanine corresponding to Gly 271 located near the C-terminal end of the helix. The hydrophobic surface of the A3 helix containing Ile 265 is in close contact with the sixth helix of the P-domain ("P6 helix"), which contains at its N-terminal end the above-mentioned asparagine of the catalytic site (Fig.  7, lower panels). The secondary structure of this region is quite different in the E 1 forms, where the A3 helix is unwound with the corresponding peptide chain forming a long flexible loop. Hence, side-chain substitutions that alter the ␣-helix propensity might influence the E 1 -E 2 conformational equilibrium. Generally, glycine and threonine residues show less ␣-helix propensity than alanine (36), and the Thr 267 3 Ala, Gly 271 3 Ala, and Gly 274 3 Ala mutations might therefore affect the E 1 -E 2 and E 1 P-E 2 P equilibria by stabilizing and extending the ␣-helical structure characteristic of E 2 /E 2 P forms.
Comparison of the E 1 and the E 2 crystal structures of the Ca 2ϩ -ATPase shows that the transition between these states involves a rotation of the A-domain parallel to the membrane, bringing the A-domain into intimate contact with the P-domain. The rotation of the A-domain occurring in relation to E 1 P 3 E 2 P transition of the phosphoenzyme allows ADP in the catalytic site to be replaced by the TGES loop of the A-domain (14). Concurrent movements of the transmembrane segments must lead to opening of a transport pathway toward the extracellular side in the E 2 P state. The rotation of the A-domain may be the initial event in the conformational rearrangement, which due to the links between the A-domain and the transmembrane segments M1, M2, and M3 and the contact formed between the A-domain and the P-domain (directly linked to M4 and M5) leads to rearrangement of the transmembrane sector. Because proteolytic cleavage of the A-M3 linker blocks the E 1 P 3 E 2 P transition (18 -20), it is conceivable that strain exerted on the A-M3 linker contributes to the force driving rotation of the A-domain (14). The mutational effects on the E 1 -E 2 and E 1 P-E 2 P equilibria observed in the present study are consistent with this hypothesis. The interaction of the side chain of the isoleucine of the A-M3 linker with P-domain residues in/close to the catalytic site seen in the E 2 crystal structures of Ca 2ϩ -ATPase, together with the other contacts between the A3 helix and the P6 helix, could be critical for controlling the A-domain rotation and, thus, correct insertion of the TGES loop in place of the leaving ADP molecule. Following dissociation of ADP from the catalytic site, the P-domain residues can come into contact with the isoleucine of the A-M3 linker. This interaction might influence the position of the P-domain, which inclines downward in the E 2 structure, thereby relieving the strain of the A-M3 linker and enforcing movements of the transmembrane segments. The interaction of the hydrophilic side chain of Thr 267 with the water phase, and the presence of the two glycines Gly 271 and Gly 274 , imposing flexibility, could be important for the conformation and direction of the backbone of the A-M3 linker in Na ϩ ,K ϩ -ATPase. The displacement of the E 1 -E 2 and E 1 P-E 2 P equilibria in favor of E 2 /E 2 P induced by substitution of either of these residues with alanine might be a consequence of increased strain of the A-M3 linker in E 1 /E 1 P, increased stability of the A3 helix in E 2 /E 2 P, and/or a change of the orientation of the A3 helix, facilitating its interaction with the P6 helix.