Mutation Lys758 → Ile of the Sarcoplasmic Reticulum Ca2+-ATPase Enhances Dephosphorylation ofE 2 P and Inhibits theE 2 to E 1Ca2Transition*

The highly conserved lysine residue Lys758 in the fifth stalk segment of the sarcoplasmic reticulum Ca2+-ATPase was substituted with either isoleucine or arginine by site-directed mutagenesis. The substitution with arginine was without significant effects on Ca2+-ATPase function, whereas multiple changes of functional characteristics were observed with the Lys758→ Ile mutant. These included insensitivity of ATPase activity to the calcium ionophore A23187, an alkaline shift of the pH dependence of ATPase activity, reduced maximum molecular turnover rate and steady-state phosphorylation level, reduced apparent affinities for Ca2+ and inorganic phosphate, as well as increased sensitivity to inhibition by vanadate. Analysis of the partial reaction steps of the enzyme cycle traced these changes to two steps. The rate of dephosphorylation of the ADP-insensitive phosphoenzyme intermediate (E 2 P) was increased, irrespective of variations of pH, K+, Ca2+, and dimethyl sulfoxide concentration. In addition, the rate of conversion of the dephosphoenzyme with low Ca2+ affinity (E 2) to the Ca2+-bound form activated for phosphorylation (E 1Ca2) was reduced in the mutant, and the ATP-induced rate enhancement of this step required higher ATP concentrations in the mutant compared with the wild type.

The Ca 2ϩ -ATPase of sarco(endo)plasmic reticulum is an energy transducing membrane-bound enzyme that pumps Ca 2ϩ at the expense of chemical energy being released by hydrolysis of ATP. As in other P-type ATPases, the proposed enzyme cycle (Fig. 1) comprises a series of consecutive conformational changes that couple vectorial ion transport processes with the formation and breakdown of an aspartyl-phosphorylated intermediate (1)(2)(3)(4). The polypeptide chain of the Ca 2ϩ -ATPase is folded to form two major cytoplasmic loops that are connected through four "stalk segments," S2-S5, 1 to a membrane domain thought to comprise 10 transmembrane helices M1-M10 (2,5,6). The ATP-binding site and the catalytic center for phosphorylation and dephosphorylation are located in the cytoplasmic domain, whereas residues involved in the binding and occlusion of Ca 2ϩ are found in the membrane sector (2,3,(7)(8)(9)(10)(11). Residues involved in the deoccluding transformation of E 1 PCa 2 to E 2 P (Fig. 1) have been found as well cytoplasmic as membranous subdomains, attesting to the global nature of this conformational change (3,(12)(13)(14)(15). Point mutations that affect the dephosphorylation of E 2 P have so far only been identified in the membrane embedded segments M4, M5, and M6 (3, 8, 16 -18). The observation that mutations in the Ca 2ϩ -binding membrane domain can block conformational changes in the catalytic site and the formation or decomposition of the acylphosphate illustrates the functional linkage between these spatially well separated domains and the occurrence of longrange intramolecular signaling during enzyme turnover, although the nature of this interaction remains obscure. The stalk segments S4 and S5 that physically link the largest cytoplasmic loop containing the phosphorylated aspartic acid residue (Asp 351 ) to transmembrane segments M4 and M5 are logical candidates for structural elements being involved in the intramolecular signaling between the membrane domain and the catalytic site, and indeed several of the residues shown to be crucial to the E 1 PCa 2 to E 2 P transition are located in the S4 segment (3,12,14,15). The recent finding (16) that replacement of the tyrosine Tyr 763 at the M5S5 boundary with glycine resulted in a mutant catalyzing ATP hydrolysis without net accumulation of Ca 2ϩ in the microsomal vesicles awards this subdomain a central role in energy coupling or in formation of a gate controlling the cytoplasmic entrance to the pathway for Ca 2ϩ translocation (16), but the role of the remainder of stalk segment S5 has not been elucidated in any detail.
In the present study, we have analyzed the functional consequences of replacing the highly conserved lysine, Lys 758 , in stalk segment S5 of the Ca 2ϩ -ATPase with isoleucine or arginine. Mutation Lys 758 3 Ile results in multiple changes of phenotypic characteristics, which can be traced to an increased rate of dephosphorylation of E 2 P and a decreased rate of the Ca 2ϩ -binding E 2 to E 1 Ca 2 transition. These changes have not previously been observed with a point mutant of the Ca 2ϩ -ATPase.

EXPERIMENTAL PROCEDURES
Mutagenesis and Expression-The construction of mutant Ca 2ϩ -ATPase cDNAs and the expression of wild type and mutants in COS-1 cells have previously been described (12,16). The expression level was quantitated by a specific sandwich enzyme-linked immunosorbent assay (19) and by determination of the active site concentration by phosphorylation (see below), and was related to the total microsomal protein determined by the dye-binding method of Bradford (20).
ATPase Activity-The rate of ATP hydrolysis catalyzed by the microsomal membranes was determined at 37°C by measurement of the amount of P i liberated over a period of 20 min using the Baginski method (21). The hydrolysis of ATP was terminated by dilution of 0.5 ml of the reaction mixture in 1.0 ml of ice-cold 0.5 M HCl, 4 mM ammonium heptamolybdate, 170 mM ascorbic acid. Subsequently, 1.5 ml of 150 mM sodium-m-arsenite, 70 mM sodium citrate, and 0.35 mM CH 3 COOH were added, and the mixture kept at 37°C for 10 min to complete the reaction (21)(22)(23). Measurements of absorbance at 850 nm were related to a standard of known P i concentration. In some cases the rate of ATP hydrolysis was measured spectrophotometrically at 37°C by a NADHcoupled assay (16) in the presence of 0.15 mM NADH, 1 mM phosphoenolpyruvate, 10 IU of lactate dehydrogenase/ml, 10 IU of pyruvate kinase/ml.
To calculate the activity referable to Ca 2ϩ -ATPase, the amount of P i formed in the presence of 4 mM EGTA was subtracted from that formed in the presence of Ca 2ϩ . The molecular turnover rate was calculated as the ratio between the specific Ca 2ϩ -ATPase activity (nanomoles of ATP hydrolyzed per mg of microsomal protein per second) and the active-site concentration (nanomoles of enzyme per mg of microsomal protein) measured by quantitation of phosphorylation from 32 P i as described below.
In experiments examining the Ca 2ϩ concentration dependence, various concentrations of CaCl 2 were added, and the free Ca 2ϩ concentrations were calculated using the computer program MAXC and the stability constants therein (24). To study vanadate inhibition of ATPase activity, the indicated concentrations of monovanadate were obtained from metavanadate as described previously (25).
Relation between Ca 2ϩ -activated ATP Hydrolysis and ATP-driven Ca 2ϩ Uptake-Combined measurements of ATPase activity and Ca 2ϩ transport were performed at 37°C in a medium containing 20 mM MOPS/Tris, pH 7.0, 100 mM KCl, 7 mM MgCl 2 , 5 mM potassium oxalate, 0.5 mM EGTA, and 0.55 mM CaCl 2 . Oxalate was included to act as Ca 2ϩ trap inside the microsomal vesicles (12). The medium used for Ca 2ϩuptake measurements in addition contained 10 5 Bq of 45 Ca 2ϩ /ml to trace Ca 2ϩ accumulation in the microsomes. The reaction was initiated by the addition of ATP to a final concentration of 5 mM. ATP hydrolysis was monitored by the Baginski (21) method, and Ca 2ϩ uptake in the microsomal vesicles was measured by Millipore filtration following quench with LaCl 3 as described previously (12).
Phosphorylation from [␥-32 P]ATP-Phosphorylation with [␥-32 P]ATP was performed on microsomal membranes (usually corresponding to about 0.5 pmol of Ca 2ϩ -ATPase) according to the previously described principles (13,16), at 0°C for 1-30 s in 100 l of the medium described in the figure legends. The reaction solutions contained in Eppendorf tubes immersed in ice water were stirred efficiently by a vertically orientated tiny magnet bar (8 ϫ 1-mm). To examine the pre-steadystate kinetics of phosphorylation starting from enzyme in the E 1 Ca 2 form or the E 2 form, the microsomes were preincubated for at least 10 min in media containing, in addition to other components described in the figure legends, 0.1 mM CaCl 2 or 1 mM EGTA, respectively. Acid quenching of the phosphorylated enzyme was performed at serial time intervals by addition of 1 ml of ice-cold 7% (w/v) trichloroacetic acid containing 1 mM P i . The acid-precipitated microsomal protein was washed by centrifugation and subjected to SDS-polyacrylamide gel electrophoresis at pH 6.0 (13), and the radioactivity associated with the separated Ca 2ϩ -ATPase was quantitated by electronic autoradiography of the dried gels using a Packard InstantImager. Phosphorylation levels calculated as the ratio between the concentration of phosphoenzyme and the active site concentration are presented in the figures, except when otherwise indicated.
Phosphorylation from 32 P i -To determine the total phosphorylation capacity of the expressed Ca 2ϩ -ATPase ("active site concentration") phosphorylation was carried out on microsomal membranes at 25°C for 10 min in a medium containing 100 mM MES/Tris, pH 6.0, 10 mM MgCl 2 , 2 mM EGTA, 0.5 mM 32 P i , and 30% (v/v) dimethyl sulfoxide. Dimethyl sulfoxide was present to increase the affinity for P i and thus ensure that 0.5 mM P i was saturating (26,27). Titration of apparent affinity for P i was performed at varying concentrations of 32 P i in the same medium as described above with 20% dimethyl sulfoxide. To study the time course of dephosphorylation, a phosphorylated sample was cooled to 0°C and subsequently diluted 20-fold in ice-cold medium without dimethyl sulfoxide or containing 5 mM non-radioactive P i to terminate the phosphorylation and permit monitoration of dephosphorylation. When indicated, the microsomes were equilibrated with 1 M of the calcium ionophore A23187 before phosphorylation to allow rapid equilibration across the membrane of Ca 2ϩ present in the dilution medium. The amount of phosphorylated protein was quantitated following acid quenching as described above for phosphorylation with [␥-32 P]ATP.
Curve Fitting and Kinetic Simulation-Data were fitted by nonlinear regression using the Sigmaplot program (Jandel Scientific). The analysis of ligand concentration dependences was based on the Hill equation, Time courses of phosphorylation and dephosphorylation were fitted to first-order kinetic equations whenever possible. Kinetic simulations of the enzyme cycle were carried out on an IBM-compatible PC using a program (SimZyme) based on the 4th order Runge-Kutta approximation method.

RESULTS
Expression-When the cell culture is working optimally, expression levels of 200 -400 pmol of Ca 2ϩ -ATPase/mg of total microsomal protein can be obtained for the exogenous wild-type enzyme, corresponding to several hundredfold the level of the endogenous COS-1 cell Ca 2ϩ -ATPase. Mutants Lys 758 3 Ile and Lys 758 3 Arg were both found to be expressed to a level similar to that of the wild type.
Ionophore Sensitivity and Ca 2ϩ Dependence of ATPase Activity of Mutant Lys 758 3 Ile-Steady-state rates of ATP hydrolysis were determined for the wild type and mutant Lys 758 3 Ile at various Ca 2ϩ concentrations at pH 7.0, in the presence and absence of the calcium ionophore A23187 (Fig. 2). By allowing passive efflux of Ca 2ϩ accumulated inside the microsomal vesicles the calcium ionophore increases the ATPase activity of the wild type 2-3-fold, due to relief of the "back inhibition" of the E 1 PCa 2 to E 2 P transformation imposed by a high lumenal Ca 2ϩ concentration (cf. Fig. 1). By contrast, the calcium ionophore did not increase the ATPase activity of mutant Lys 758 3 Ile, but gave rise to a slight inhibition. This observation was confirmed in measurements where the NADH-coupled spectrophotometric assay was used to study in detail the time course of ATP hydrolysis before and immediately following the addition of ionophore (inset of Fig. 2).
Moreover, there was a conspicuous 3-fold difference between the maximum molecular turnover numbers for ATP hydrolysis measured with the Lys 758 3 Ile mutant and the wild type, and the apparent calcium affinity displayed by the mutant was approximately 3-fold lower than that of the wild type ( Fig. 2).
At Ca 2ϩ concentrations above 10 M, the ATPase activity of the wild type starts to decline, an effect which is most pronounced in the presence of the calcium ionophore and is, at least under these conditions, partly due to Ca 2ϩ binding at the low affinity lumenal inhibitory sites. In addition, some of the inhibition of the wild type at high Ca 2ϩ concentration may result from the substitution of Ca 2ϩ for catalytic Mg 2ϩ in the phosphorylated complex, slowing the conversion of E 1 PCa 2 to E 2 P (28). In the mutant, the inhibitory effect of high Ca 2ϩ concentration was less pronounced compared with the wild type (Fig. 2).
pH Dependence of Ca 2ϩ -ATPase Activity of Mutant Lys 758 3 Ile-The pH dependence of the Ca 2ϩ -activated ATPase activity was studied in the pH range 6 -9 (Fig. 3). The wild type displays a pH optimum in the physiological pH range around pH 7. The decline in ATPase activity observed for the wild type at the acidic side of the pH optimum is caused primarily by an inhibitory influence of protons on the E 2 to E 1 Ca 2 conversion, whereas on the alkaline side the decline in activity stems predominantly from a decrease in the rate of E 2 P dephosphorylation (29 -35). For the Lys 758 3 Ile mutant, the pH optimum was shifted toward a more alkaline value relative to the pH optimum of the wild type. In the alkaline pH range up to pH 9 there was almost no decrease in the ATPase activity of the mutant, and at pH 9 the turnover rate of the mutant was considerably higher than that of the wild type. Fig. 3 also shows that the activating effect of the calcium ionophore on the wild type was abolished toward the extremes of the pH interval studied, possibly because the E 1 PCa 2 to E 2 P transition sensitive to lumenal Ca 2ϩ becomes less rate-limiting under these conditions. The inhibitory effect of the calcium ionophore on the ATPase activity of the Lys 758 3 Ile mutant was most pronounced at neutral and acidic pH values.
Ca 2ϩ Uptake in Microsomal Vesicles by Mutant Lys 758 3 Ile-The absence of an activating effect of calcium ionophore on the ATP hydrolysis activity of mutant Lys 758 3 Ile is reminiscent of the previous observation with the Tyr 763 3 Gly mutant of the Ca 2ϩ -ATPase (16). Mutant Tyr 763 3 Gly was shown not to accumulate Ca 2ϩ inside the microsomal vesicles, although ATP was hydrolyzed, explaining the absence of ionophore-mediated activation, since no back inhibition could be exerted in the absence of Ca 2ϩ accumulation. To determine whether the same explanation is valid in the case of mutant Lys 758 3 Ile, Ca 2ϩ uptake and ATP hydrolysis were monitored under experimental conditions that were identical except for the addition of 45 Ca 2ϩ as tracer to the uptake medium. As seen from the data in Fig. 4, the mutant was clearly capable of accumulating Ca 2ϩ inside the microsomal vesicles. The molecular turnover rates for ATP hydrolysis and Ca 2ϩ transport by the mutant were both found to be reduced compared with the wild type, consistent with the above described results for the ATPase activity. Because the measurements of Ca 2ϩ uptake and ATP hydrolysis were performed under identical conditions it was possible to calculate the apparent "coupling ratio" between Ca 2ϩ accumulation and ATP hydrolysis (inset of Fig. 4). The result of the calculation is in line with the notion that the theoretical molar ratio of two calcium ions transported per ATP molecule hydrolyzed is attained only transiently in the initial phase of the reaction (36). During the approach to steady state, the coupling ratio declines to lower values due to the increase in Ca 2ϩ efflux resulting from the building up of a Ca 2ϩ gradient. There was no significant difference between the coupling ratios of the wild type and mutant Lys 758 3 Ile at any time, and hence it can be concluded that in the Lys 758 3 Ile mutant the absence of an ionophore-mediated activation of ATP hydrolysis results from insensitivity of the mutant to accumulated Ca 2ϩ rather than inability to accumulate Ca 2ϩ .
Ca 2ϩ and pH Dependence of ATPase Activity of Mutant Lys 758 3 Arg-In contrast to the situation with the Lys 758 3 Ile mutant, the ionophore-mediated activation, the Ca 2ϩ and pH dependences of ATPase activity, and the maximum molecular turnover rate of the mutant with the more conservative amino acid substitution Lys 758 3 Arg were similar to those of the wild type (Fig. 5), as was the coupling ratio between Ca 2ϩ accumulation and ATP hydrolysis (results not shown).
To elucidate the basis for the differences between the characteristics of the overall ATPase activities of the wild type and mutant Lys 758 3 Ile, the partial reaction steps of the enzyme cycle were analyzed under various experimental conditions, as described below.
Titration of Phosphorylation from ATP and P i -The steadystate level of phosphoenzyme formed from ATP was examined at 0°C, where the ratio between the rates of phosphorylation and dephosphorylation favors accumulation of phosphoenzyme in the wild type even at rather low ATP concentrations. The phosphorylation from ATP is activated by Ca 2ϩ binding at the high-affinity cytoplasmically facing sites (Fig. 1). As seen in for activation of phosphorylation than the wild type, the K 0.5 (Ca 2ϩ ) values being 1.2 and 0.4 M, respectively, in accordance with the Ca 2ϩ titrations of ATPase activity presented in Fig. 2. In Fig. 6A the results for wild type and mutant were normalized separately taking the values reached at saturating Ca 2ϩ concentration as 100%. Fig. 6B compares the ATP concentration dependences of the steady-state phosphorylation levels calculated by relating the concentration of phosphoenzyme to the active site concentration. Contrary to the wild type, which was 100% phosphorylated at 10 M ATP, the Lys 758 3 FIG. 4. Relation between Ca 2؉ -activated ATP hydrolysis and ATP-driven Ca 2؉ uptake in the microsomal vesicles. Measurements were accomplished as described under "Experimental Procedures" for the wild-type Ca 2ϩ -ATPase (q, E) and mutant Lys 758 3 Ile (ç, É). The left ordinate (open symbols) shows the amount of P i liberated, the maximal value found with the wild type taken as 100%. The right ordinate (closed symbols) shows the calcium accumulation, the maximal value found with the wild type taken as 100%. In these assays, equal active site concentrations of the wild type and the mutant were inserted. The inset shows the molar ratio between Ca 2ϩ uptake and ATP hydrolysis for wild type (E) and Lys 758 3 Ile mutant (É). Ile mutant was only 60% phosphorylated at this ATP concentration, even though the data obtained with the mutant could be fitted to a hyperbolic dependence on the ATP concentration with a K 0.5 value of only 0.34 M.
The E 2 form of the Ca 2ϩ -ATPase can be phosphorylated backwards from P i (Fig. 1). The apparent affinity for P i was determined under optimal conditions for accumulation of E 2 , in the absence of Ca 2ϩ at 25°C, pH 6.0, and in the presence of 20% dimethyl sulfoxide (Fig. 6C). The apparent affinity for P i displayed by the mutant was almost 5-fold reduced relative to the affinity of the wild type.
Dephosphorylation of Phosphoenzyme Formed from 32 P i -The reduced apparent affinity for P i found with the Lys 758 3 Ile mutant is the result of either a reduced rate of phosphorylation from P i and/or an increased rate of dephosphorylation. The dephosphorylation kinetics of the E 2 P phosphoenzyme intermediate formed from 32 P i in the absence of Ca 2ϩ was monitored at 0°C following termination of phosphorylation by dilution of the phosphorylated enzyme. The composition of the dilution medium was varied to examine the characteristics of the mutant phosphoenzyme. In a medium with 100 mM K ϩ and without Ca 2ϩ at pH 7.0 (Fig. 7A), the dephosphorylation was relatively fast for the wild type as well as the mutant, corresponding to half-lives of the phosphoenzymes below 1 s. The dephosphorylation rate of the mutant was higher than that of the wild type, but since both rates approached the experimental limit set by the time resolution of the mixing system, the magnitude of the difference could not be estimated with certainty. Therefore additional experiments were carried out under conditions known to result in a lower rate of dephosphorylation of the wild-type phosphoenzyme.
When K ϩ was omitted from the dilution medium (Fig. 7B), the half-life of the wild-type phosphoenzyme increased at least 3-fold, in accordance with the literature (37), while there was no detectable change in the dephosphorylation rate of the mutant. Thus, there was a marked difference between the dephosphorylation rates of the Lys 758 3 Ile mutant and wild type under these conditions.
As seen in Fig. 7C, addition of 5 mM Ca 2ϩ in the presence of calcium ionophore at pH 7.0 reduced the rate of dephosphorylation of the wild type severalfold, due to the back inhibition by Ca 2ϩ binding at the low affinity lumenal sites described above. In contrast to the conspicuous effect of Ca 2ϩ on the wild type, there was only a minor increase in the half-life of the mutant phosphoenzyme when dephosphorylation proceeded in the presence of 5 mM Ca 2ϩ and ionophore, and hence the difference between the dephosphorylation rates of wild type and mutant amounted to more than 40-fold under these conditions. This is consistent with the insensitivity of the ATPase activity of the mutant to back inhibition by accumulated Ca 2ϩ described above.
The dephosphorylation of E 2 P was in addition examined at pH 7.0 in the absence of K ϩ and Ca 2ϩ and presence of 15% dimethyl sulfoxide (Fig. 7D). In these experiments, the dilution medium contained 5 mM nonradioactive P i to ensure that phosphorylation from 32 P i was efficiently terminated despite the presence of dimethyl sulfoxide. Under these conditions dephosphorylation was very slow in the wild type, in accordance with the previously reported effects of dimethyl sulfoxide (26,27), and a 10-fold difference was seen between the half-lives of mutant and wild-type phosphoenzymes.
Yet another condition known to impede the progression of phosphoenzyme decomposition of the wild-type Ca 2ϩ -ATPase is high pH in the presence of Mg 2ϩ (26, 29 -31, 34). Although the dephosphorylation rate of the Lys 758 3 Ile mutant decreased considerably when the pH was increased to 8.35 in the presence of 10 mM Mg 2ϩ , the wild type responded even stronger, and a large difference (more than 10-fold) between the dephosphorylation rates of mutant and wild type was thus manifested under these conditions (Fig. 8A). This accords with the above described lack of inhibition of the overall turnover rate of the mutant at high pH (cf. Fig. 3). Fig. 8B shows that addition of 1 mM Ca 2ϩ at pH 8.35 increased the half-life of the mutant phosphoenzyme more than 10-fold. The synergistic effects of high pH and Ca 2ϩ may be ascribable to the increased affinity of the lumenal sites for Ca 2ϩ at high pH (26) and suggests that the lumenal sites are still functioning and can mediate back inhibition in the mutant, although less efficiently than in the wild type. The initial fast phase of dephosphorylation seen with both the wild type and the mutant in Fig. 8B calls for some explanation. We believe that it represents the dephosphorylation occurring during the time it takes Ca 2ϩ present in the dilution medium to reach the inhibitory Ca 2ϩ sites in the vesicular lumen.
When the dephosphorylation rate of the Lys 758 3 Arg mutant was examined by the same methods as described above for Figs. 7 and 8, there was no significant difference from the wild type (results not shown).
Ca 2ϩ -induced Back Conversion of E 2 P to E 1 PCa 2 -To examine whether a high lumenal Ca 2ϩ concentration is able to promote conversion of E 2 P backwards to produce E 1 PCa 2 in the Lys 758 3 Ile mutant as in the wild type, the enzyme was phosphorylated from 32 P i in the presence of calcium ionophore, and the phosphoenzyme was subsequently diluted into a high pH medium containing 1 mM Ca 2ϩ and 2 mM ADP to dephosphorylate E 1 PCa 2 . As seen in Fig. 9, both the wild type and the mutant were able to dephosphorylate rapidly in the presence of FIG. 7. Dephosphorylation at pH 7.0 of the phosphoenzyme formed from 32 P i . Phosphorylation of wild type (E) and Lys 758 3 Ile mutant (É) was carried out at 25°C for 10 min in the presence of 100 mM MES/Tris, pH 6.0, 10 mM MgCl 2 , 2 mM EGTA, 30% (v/v) dimethyl sulfoxide, 1 M calcium ionophore A23187, and 0.25 mM 32 P i . Following cooling of the sample to 0°C, the phosphorylation was terminated by a 20-fold dilution of an aliquot into an ice-cold medium of the composition described below, and acid quenching was performed at serial time intervals. A, the dilution medium contained 50 mM TES/Tris, pH 7.0, 100 mM KCl, and 2 mM MgCl 2 ; B, same as A without KCl; C, same as A with 5 mM CaCl 2 included; D, same as B but with 15% (v/v) dimethyl sulfoxide and 5 mM non-radioactive P i included. The amount of phosphoenzyme present at the end of the 10-min phosphorylation period was taken as 100%.
ADP but not in the absence of ADP, the latter condition being equivalent to that of Fig. 8B. These findings show that at least at high pH, Ca 2ϩ is able to drive the E 1 PCa 2 to E 2 P conversion backwards in the Lys 758 3 Ile mutant by a mechanism similar to that pertaining to the wild type, again attesting that under appropriate conditions the lumenal Ca 2ϩ -binding sites are able to function in the mutant.
The E 1 PCa 2 to E 2 P Conversion-The accelerated dephosphorylation of E 2 P in the Lys 758 3 Ile mutant can account for the fact that the overall ATPase turnover rate is higher in the mutant than in the wild type at alkaline pH where the dephosphorylation of E 2 P tends to be rate-limiting. However, the reduced turnover rate of the Lys 758 3 Ile mutant relative to the wild type observed at neutral pH demands that another partial reaction of the cycle must be inhibited by the mutation. It is well recognized that in the wild-type Ca 2ϩ -ATPase, the E 1 PCa 2 to E 2 P transition contributes to rate limitation of the overall cycle at neutral pH in the presence of K ϩ (29,38,39). In fact, under these conditions and at 0°C this is the major ratedetermining step of the cycle (28,29). Because it was demonstrated above that dephosphorylation of the mutant and wildtype E 2 P forms proceeded rapidly at 0°C and neutral pH in the presence of K ϩ (Fig. 7A), the dephosphorylation of E 2 P was not rate-limiting under these conditions, and the rate of the E 1 PCa 2 to E 2 P conversion could therefore be examined by monitoring the dephosphorylation in the forward direction of phosphoenzyme accumulated as E 1 PCa 2 at steady state. From the results presented in Table I it is clear that there was no significant difference between the decay rates of the mutant and wild-type phosphoenzymes measured following termination of phosphorylation with EGTA. A similar result was obtained when the E 1 PCa 2 phosphoenzyme was chased with 400 M unlabeled ATP (not shown). Table I, furthermore, shows that the addition of ADP resulted in a very rapid dephosphorylation, confirming that the phosphoenzyme studied was initially accumulated in the ADP-sensitive E 1 PCa 2 form. Taken together, these data demonstrate that the rates of E 1 PCa 2 to E 2 P conversion in wild type and mutant are similar.
The Rates of the E 2 to E 1 Ca 2 Conversion and the Phosphorylation of E 1 Ca 2 -As shown in Fig. 10, the Ca 2ϩ -binding E 2 to E 1 Ca 2 transition was examined at pH 7.0 at 0°C by monitoring the time dependence of phosphorylation from ATP under two different sets of conditions (32,40). The enzyme was preincubated either in the presence of Ca 2ϩ to accumulate the E 1 Ca 2 form, followed by initiation of phosphorylation by addition of 2 M [␥-32 P]ATP, or in the absence of Ca 2ϩ (presence of EGTA) and Mg 2ϩ to accumulate the E 2 form, followed by initiation of phosphorylation by simultaneous addition of Ca 2ϩ , Mg 2ϩ , and [␥-32 P]ATP.
Addition of ATP to the enzyme in the E 1 Ca 2 form resulted in the appearance of an overshoot of phosphorylation in the Lys 758 3 Ile mutant, whereas such an overshoot was not seen with the wild type under the present conditions (Fig. 10A). The overshoot observed with the mutant is composed of an initial fast phosphorylation, reaching almost the same phosphorylation level as the wild type, followed by a decline to a steady-state level of phosphorylation corresponding to 40 -50% of the active site concentration, consistent with the steady-state phosphorylation data presented above (Fig. 6B).
The simultaneous addition of Ca 2ϩ , Mg 2ϩ , and [␥-32 P]ATP to enzyme accumulated in the E 2 form (Fig. 10B) allows evaluation of the rate at which the enzyme becomes available for phosphorylation from ATP, i.e. the rate of the E 2 to E 1 Ca 2 FIG. 8. Ca 2؉ dependence of dephosphorylation at alkaline pH of the phosphoenzyme formed from 32 P i . Phosphorylation of wild type (E) and Lys 758 3 Ile mutant (É) was carried out as described for Fig. 7. Following cooling of the sample to 0°C, phosphorylation was terminated by a 20-fold dilution of an aliquot into an ice-cold medium of the composition described below, and acid quenching was performed at serial time intervals. A, the dilution medium contained 100 mM TES/ Tris, pH 8.35, 10 mM MgCl 2 , and 2 mM EGTA; B, same as A with 1 mM CaCl 2 and without EGTA. The amount of phosphoenzyme present at the end of the phosphorylation period was taken as 100%.
FIG. 9. Dephosphorylation at alkaline pH induced by addition of ADP and Ca 2؉ to the phosphoenzyme formed from 32 P i . Wildtype Ca 2ϩ -ATPase and the Lys 758 3 Ile mutant were phosphorylated and cooled as described for Fig. 7. Dephosphorylation was monitored at the indicated time intervals following a 20-fold dilution of an aliquot into an ice-cold medium containing 100 mM TES/Tris, pH 8.35, 10 mM MgCl 2 , and 1 mM CaCl 2 . In addition, 2 mM ADP was included when indicated ("ϩADP"). The electronic autoradiographs of the radioactivity associated with the ATPase band on the gel are shown. conversion. For the wild type, this procedure resulted in a slightly reduced rate of phosphoenzyme formation relative to the rate observed when starting from enzyme accumulated in E 1 Ca 2 form. Importantly, under these conditions, mutant Lys 758 3 Ile phosphorylated at a rate that was 5-6-fold lower than that of the wild type, and no overshoot was seen. Again the mutant reached a steady-state level of phosphorylation corresponding to 40 -50% of the active site concentration. Because the mutant was able to phosphorylate as rapidly as the wild type from E 1 Ca 2 , it can be concluded that the E 2 to E 1 Ca 2 transition proceeded at a considerably reduced rate in the mutant compared with the wild type. It is well known that ATP in micromolar concentration enhances the E 2 to E 1 Ca 2 transition rate in the wild type by binding in a regulatory mode to E 2 (1,32,(41)(42)(43), and to see whether this modulatory effect of ATP could be observed also in the mutant, phosphorylation was monitored following concurrent addition of Ca 2ϩ , Mg 2ϩ , and 10 M [␥-32 P]ATP to Ca 2ϩdepleted enzyme by the same procedure as described above (Fig. 11). Comparison of the results in Figs. 10B and 11 reveals that ATP was able to accelerate the E 2 to E 1 Ca 2 transition in both the wild type and the mutant. The increase in ATP concentration from 2 to 10 M produced a 5-fold increment of the rate of phosphorylation in the mutant, and the steady-state phosphorylation level of the mutant increased to about 60% of the active site concentration in accordance with the data in Fig. 6B.
The Modulatory Effect of ATP on Steady-state ATPase Activity at 37°C- Fig. 12 compares the ATP concentration dependences of ATPase activity of the wild type and mutant Lys 758 3 Ile measured in the presence of calcium ionophore. Both the wild type and the mutant exhibited a complex activation profile indicative of several phases of activation, as described previously for the wild type (10,38,42). The basal activation of ATP hydrolysis below 10 M ATP reflects the saturation of the catalytic site in the E 1 Ca 2 form (cf. Fig. 6B). This was not studied in any detail here due to the relatively low signal to background ratio in the low concentration range. Activation at higher ATP concentrations reflects ATP-induced rate enhancement of conformational changes that are rate-determining for the overall reaction. In the 10 -100 M ATP concentration range, the rate enhancement in the wild type is primarily due to the effect of ATP on the rate of E 2 to E 1 Ca 2 transition of the dephosphoenzyme described above (32,(41)(42)(43). At higher concentrations, ATP modulates the rates of the E 1 PCa 2 to E 2 P transition and the dephosphorylation of E 2 P (28,31,38,42). It can be seen in Fig. 12 that the difference between the rates of ATP hydrolysis of mutant and wild type was largest at ATP concentrations below 100 M, the activation profile of the mutant being steeper than that of the wild type between 100 M and 1 mM ATP.
Inhibition by Vanadate-From the above described data showing an increased rate of dephosphorylation and a decreased rate of the Ca 2ϩ -binding E 2 to E 1 Ca 2 transition in the Lys 758 3 Ile mutant, it may be rationalized that an increased amount of mutant enzyme should accumulate in the E 2 form at steady state. To examine the level of E 2 intermediate present at steady state, vanadate inhibition of ATPase activity was investigated. Vanadate, acting as a transition state analog of the phosphoryl group, binds to the enzyme in the E 2 conformation, thereby impeding the continuation of the enzyme cycle (25,44). From the result shown in Fig. 13 it is clear that the mutant is inhibited by vanadate with much higher apparent affinity than the wild type, K 0.5 values being 0.5 and 30 M for the mutant and wild type, respectively. This finding is consistent with the hypothesis that the mutant accumulates in the E 2 form (or an "E 2 -like" form) at steady state to a greater extent than the wild type.
It should be noted that mainly for technical reasons the data in Fig. 13 were obtained at 37°C, whereas the phosphorylation and dephosphorylation studies from which the result in Fig. 13 was predicted were carried out at 0°C, demonstrating that the temperature is of minor importance for the general conclusions. DISCUSSION The present study is the first to describe a mutant that displays a higher rate of dephosphorylation of E 2 P than the wild type. An increased dephosphorylation rate of the Lys 758 3 Ile mutant relative to that of the wild type was seen independent of variation of pH, K ϩ , and Ca 2ϩ concentration. Even in the presence of 15% dimethyl sulfoxide, which is a very potent inhibitor of dephosphorylation (26,27), the dephosphorylation rate of the mutant was much faster than that of the wild type.
The Lys 758 3 Ile mutant, furthermore, exhibited a conspicuous reduction of the rate of the Ca 2ϩ -binding E 2 to E 1 Ca 2 conversion of the dephosphoenzyme (5-6-fold at neutral pH, low ATP concentration, 0°C), as seen in the measurements of the phosphorylation rate with enzyme preaccumulated in the E 2 form. The E 2 to E 1 Ca 2 conversion comprises several substeps, including the sequential dissociation of two or three protons, the sequential binding of two calcium ions, and associated (or intervening) conformational changes (2,45), and one or more of these substeps may be affected by the Lys 758 3 Ile mutation. Like the increased dephosphorylation rate, the reduced rate of E 2 to E 1 Ca 2 conversion in the Lys 758 3 Ile mutant represents a new feature not previously reported for any Ca 2ϩ -ATPase mutant. Together these two changes of partial reaction rates may explain most of the phenotypic characteristics of the mutant.
In interpreting the data it should be kept in mind that different reaction steps may become rate-limiting for the overall Ca 2ϩ -ATPase cycle depending on the experimental conditions, including temperature, pH, K ϩ , Mg 2ϩ , ATP concentration, and the Ca 2ϩ concentration on the two sides of the membrane (29,38,39). In the wild-type Ca 2ϩ -ATPase, the ATPase activity measured at saturating substrate concentration in the presence of K ϩ at neutral pH and 37°C increases 2-3-fold upon addition of the calcium ionophore A23187, due to relief of the back inhibition of the E 1 P Ca 2 to E 2 P transition by accumulated Ca 2ϩ present at millimolar concentration inside the microsomes. This effect was not seen with the mutant enzyme despite its ability to accumulate Ca 2ϩ , but on the contrary the presence of the calcium ionophore slightly inhibited the ATPase activity of the mutant. Absence of ionophoremediated activation was previously observed with the Tyr 763 3 Gly mutant and was shown to be a consequence of uncoupling of Ca 2ϩ transport from ATP hydrolysis induced by the Tyr 763 3 Gly mutation (16). However, because Ca 2ϩ accumulation in the vesicles could be observed with the Lys 758 3 Ile mutant, and the ratio between Ca 2ϩ transport and ATP hydrolysis was the same as in the wild type, the lack of activation by ionophore cannot in the present case be explained by uncoupling. Moreover, because we demonstrated that addition of 1 mM Ca 2ϩ at alkaline pH in the presence of calcium ionophore reduced the rate of dephosphorylation of E 2 P considerably and led to back conversion of E 2 P into ADP-sensitive phosphoenzyme (Figs. 8 and 9), the lumenal inhibitory Ca 2ϩ -binding sites seem to function in the mutant, and the increased rate of dephosphorylation is probably not accomplished through an alternative reaction path in which dephosphorylation can occur with Ca 2ϩ remaining bound to the phosphoenzyme. The two fundamental kinetic effects of the mutation, the increased rate of dephosphorylation of E 2 P (without bound Ca 2ϩ ) and the reduced rate of the E 2 to E 1 Ca 2 conversion, may, however, both be contributing to the absence of ionophore-induced activation of the ATP hydrolysis in the Lys 758 3 Ile mutant, as can be demonstrated by computer simulation of the reaction cycle. An increased rate of dephosphorylation of E 2 P effectively increases the Ca 2ϩ concentration required to drive the E 1 PCa 2 to E 2 P interconversion backwards, and inhibition of the E 2 to E 1 Ca 2 conversion renders this step more rate determining for the overall reaction than the E 1 PCa 2 to E 2 P conversion sensitive to lumenal Ca 2ϩ . The 3-fold lower maximum turnover rate of the mutant relative to the wild type seen at 37°C at neutral pH (Fig. 2) shows clearly that rate limitation in the mutant is imposed by the E 2 to E 1 Ca 2 conversion under these conditions. The slight inhibitory effect of the calcium ionophore on ATPase activity in the mutant is also noteworthy. A23187 causes strong inhibition of the E 2 to E 1 Ca 2 conversion when added in a much larger dose FIG. 12. ATP dependence of Ca 2؉ -ATPase activity. The rates of ATP hydrolysis of the wild type (E) and the Lys 758 3 Ile mutant (É) were measured at 37°C at the indicated concentrations of ATP (present almost entirely as MgATP) in the presence of 25 mM TES, pH 7.5, 100 mM KCl, 100 M CaCl 2 , 1 mM free Mg 2ϩ , and 1 M calcium ionophore A23187, using the NADH-coupled assay. The molecular turnover rates for wild type and mutant were normalized separately taking the maximal values at 5 mM ATP as 100%. The lines are theoretical curves computed on the basis of the reaction cycle in Fig. 1, with ATP-induced rate enhancements incorporated corresponding to the E 2 to E 1 Ca 2 transition, the E 1 PCa 2 to E 2 P transition, and the dephosphorylation of E 2 P. The basic and ATP-modulated rates of the E 2 to E 1 Ca 2 transition were, respectively, 30-and 12-fold reduced in the mutant relative to the wild type, and the affinity of the E 2 form for ATP was 10-fold reduced in the mutant (K d (ATP) values of 500 and 50 M for mutant and wild type E 2 forms, respectively), whereas the ATP affinities of the other intermediates in the enzyme cycle were identical in mutant and wild type. than that applied in the present study to make the membrane leaky (46), and it can be predicted that even the low concentration used here should exert some inhibition of the E 2 to E 1 Ca 2 conversion. Any effect of this inhibition on ATPase activity would be masked in the wild type due to the simultaneous relief of the back inhibition of the rate-determining E 1 PCa 2 to E 2 P conversion, but if the E 2 to E 1 Ca 2 conversion limits the overall rate of the enzyme cycle in the mutant, the inhibitory effect of the ionophore should be manifested as a decrease of the ATPase activity in this case, consistent with the experimental finding.
The Lys 758 3 Ile mutant displayed a pH dependence of ATPase activity which was alkaline shifted, so that the turnover rate of the mutant was increased relative to that of the wild type at high pH. This can be explained by the effect of the mutation on the dephosphorylation of E 2 P. In the wild type, alkaline pH inhibits the dephosphorylation severely in the presence of Mg 2ϩ (26, 29 -31, 34, 39), but in the mutant the dephosphorylation was less strongly inhibited under these conditions. In addition, the rate of the E 2 to E 1 Ca 2 transition increases with pH (32,35), so that this step becomes less limiting for the rate of the overall cycle compared with the situation at neutral pH.
The titration of steady-state phosphorylation at 0°C as a function of ATP concentration below 10 M showed only a slight reduction of apparent affinity for ATP in the Lys 758 3 Ile mutant relative to the wild type (Fig. 6B). Moreover, on the basis of the wild-type-like phosphorylation rate observed at 2 M ATP, when the mutant enzyme had been preaccumulated in the E 1 Ca 2 form (Fig. 10A), it seems unlikely that the intrinsic ATP affinity of the catalytic site in E 1 Ca 2 or the rate constant for phosphorylation was significantly lower in the mutant than in the wild type. The presence of a phosphorylation overshoot and a low steady-state level of phosphorylation in the mutant can be accounted for by the low rate of E 2 to E 1 Ca 2 conversion in combination with a high rate of dephosphorylation.
In the Lys 758 3 Ile mutant, as well as in the wild type, titration of the ATP concentration dependence of steady-state ATP hydrolysis activity at 37°C (Fig. 12) showed a complex activation profile indicative of secondary activation by ATP that binds with different affinities to the various intermediates, as described previously for the wild type (10,38,42). The difference between the ATPase activities of mutant and wild type was largest at ATP concentrations below 100 M, where the secondary activation seen in the wild type predominantly reflects the modulation by ATP of the E 2 to E 1 Ca 2 conversion (32,(41)(42)(43). At higher ATP concentrations the rise in activity was steeper in the mutant than in the wild type, suggesting that a modulatory effect of ATP exerted between 100 M and 1 mM ATP was partly able to compensate for the intrinsically low rate of E 2 to E 1 Ca 2 transition in the mutant. We have carried out a series of computer simulations of the enzyme cycle with ATP-induced rate enhancements incorporated corresponding to the E 2 to E 1 Ca 2 transition, the E 1 PCa 2 to E 2 P transition, and the dephosphorylation of E 2 P (for further discussion of these modulatory effects, see Ref. 10, 38, and 42). To reproduce the data in Fig. 12, it was necessary to assume not only that the basic rate of the E 2 to E 1 Ca 2 conversion at low ATP concentration is severalfold reduced in the mutant relative to that of the wild type, but also that the ATP-induced rate enhancement of this conversion requires higher ATP concentration in the mutant than in the wild type (see legend to Fig. 12). Hence, it is possible that the ATP affinity of the E 2 form or another intermediate in the path between E 2 and E 1 Ca 2 is lower in the mutant than in the wild type. In this connection it should be noted that a 5-fold rise in ATP concentration from 2 to 10 M produced a proportional 5-fold increase in the E 2 to E 1 Ca 2 conversion rate in the mutant (compare Figs. 10B and 11), indicating that the activating effect of ATP was far from being saturated in this concentration range.
The changes in phenotypic characteristics of the Ca 2ϩ -ATPase induced by the Lys 758 3 Ile mutation, furthermore, included reduced apparent affinities for Ca 2ϩ and P i , as well as an increased apparent affinity for vanadate. The reduced apparent affinity for P i follows directly as a consequence of the increased dephosphorylation rate (cf. Fig. 1) and demonstrates that the enhancement of the "off rate" for P i is not matched by a comparable enhancement of the "on rate." The 3-fold reduction in the apparent Ca 2ϩ affinity of the mutant relative to the wild type observed in the Ca 2ϩ titrations of ATPase activity and phosphorylation from ATP can be explained as a consequence of the reduced rate of the Ca 2ϩ -binding E 2 to E 1 Ca 2 conversion. Moreover, because vanadate binds to the E 2 form as a transition-state analog of the phosphoryl group, the accumulation of E 2 at steady state may explain the increased apparent affinity for vanadate in the mutant.
The functional properties found with the Lys 758 3 Ile mutant resemble at least partially those of the non-muscle isoform of the Ca 2ϩ -ATPase (SERCA3). Hence, the functional differences found between SERCA3 and the wild type of SERCA1 (the isoform used in the present study) included a decreased apparent affinity for Ca 2ϩ , an increased inhibition by vanadate, and an alkaline shift in pH optimum (47). The phosphorylation from P i and the dephosphorylation rate of E 2 P have not been characterized for SERCA3. The functional differences between SERCA3 and SERCA1 cannot be associated to Lys 758 , which is present in both isoforms, but there are multiple amino acid differences between SERCA3 and SERCA1 in the cytoplasmic region just N-terminal to S5 (for sequence comparison, see Ref. 48). Interestingly, the literature contains other examples of functional perturbations of the type seen with the Lys 758 3 Ile mutant. Hence, both nonylphenol (49) and jasmone, a component of peppermint oil (50,51), caused enhanced dephosphorylation of E 2 P as well as a reduced rate of E 2 to E 1 Ca 2 conversion, when incorporated into Ca 2ϩ -ATPase membranes. The group of sesquiterpene lactones, to which thapsigargin belongs, as well as 2,5-di(tert-butyl)-1,4-benzohydroquinone, also appear to stabilize an E 2 -like state of the Ca 2ϩ -ATPase with low ATP affinity and destabilize E 2 P (10, 52-54). Other hydrophobic/amphiphilic molecules such as C 12 E 8 (38) and procaine (39) exert the opposite effect, combining enhancement of the E 2 to E 1 Ca 2 transition with inhibition of E 2 P dephosphorylation. Hence, there seems to exist an obligatory linkage between the stabilization/destabilization of E 2 and E 2 -like states and the destabilization/stabilization of E 2 P, suggesting that the above mentioned perturbations all exert their effects by displacing a conformational equilibrium of the protein. The present results define amino acid residue at position 758 as a significant factor in the control of this conformational equilibrium.