Importance of Stalk Segment S5 for Intramolecular Communication in the Sarcoplasmic Reticulum Ca 2 1 -ATPase*

Sixteen residues in stalk segment S5 of the Ca 2 1 ATPase of sarcoplasmic reticulum were studied by site-directed mutagenesis. The rate of the Ca 2 1 binding transition, determined at 0 °C, was enhanced relative to wild type in mutants Ile 743 3 Ala, Val 747 3 Ala, Glu 748 3 Ala, Glu 749 3 Ala, Met 757 3 Gly, and Gln 759 3 Ala and reduced in mutants Asp 737 3 Ala, Asp 738 3 Ala, Ala 752 3 Leu, and Tyr 754 3 Ala. In mutant Arg 762 3 Ile, the rate of the Ca 2 1 binding transition was wild type like at 0 °C, whereas it was 3.5-fold reduced relative to wild type at 25 °C. The rate of dephosphorylation of the ADP-insen-sitive phosphoenzyme was increased conspicuously in mutants Ile 743 3 Ala and Tyr 754 3 Ala (close to 20-fold in the absence of K 1 ) and increased to a lesser extent in Asn 739 3 Ala, Glu 749 3 Ala, Gly 750 3 Ala, Ala 752 3 Gly, Met 757 3 Gly, and Arg 762 3 Ile, whereas it was reduced in mutants Asp 737 3 Ala, Val 744 3 Gly, Val 744 3 Ala, Val 747 3 Ala, and Ala 752 3 Leu. In mutants Ile 743 3 Ala, Tyr 754 3 Ala, and Arg 762 3 Ile, the apparent affinities for vanadate were enhanced 23-, 30-, and 18-fold, respectively, potassium oxalate (to act as Ca 2 1 trap inside the vesicles), using a Millipore filtration method (12). Phosphorylation from [ g - 32 P]ATP and Dephosphorylation— To study the time course of phosphorylation of enzyme initially present in the Ca 2 1 -deprived form, microsomes preincubated in a medium containing 40 m M MOPS/Tris, pH 7.0, 80 m M KCl, and 1 or 2 m M EGTA were mixed with a medium of a composition producing final concentrations of 40 m M MOPS/Tris, pH 7.0, 80 m M KCl, 2 or 5 m M [ g - 32 P]ATP, 5 m M MgCl 2 , and 0.1 m M free Ca 2 1 , followed by acid quenching at serial time intervals by further mixing with the double volume of 25% (w/v) trichloroacetic acid containing 100 m M H 3 PO 4 . The time course of phosphorylation of en- zyme initially present in the Ca 2 1 -saturated form was studied in the same way except that the medium used for preincubation contained 0.1 m M Ca 2 1 instead of EGTA. For experiments at pH 6.0, 40 m M MES adjusted to pH 6.0 with Tris was present instead of 40 m M MOPS/Tris. The phosphorylation reactions were carried out either at 0 °C, using a manual mixing technique (12), or at 25 °C, using the Bio-Logic quench flow module QFM-5 (Bio-Logic Instruments, Claix, France) as described previously (21). The quench flow technique was also used to study Ca 2 1 dissociation from the enzyme by monitoring the phosphorylation following simulta- neous addition of EGTA and [ g - 32 P]ATP to Ca 2 1 -saturated enzyme (21). Microsomes in m M MOPS/Tris, m M m M 2 , and m M CaCl 2 with an equal volume m M MOPS/Tris, m M

The Ca 2ϩ -ATPase of sarcoplasmic reticulum functions as a Ca 2ϩ pump that actively transports Ca 2ϩ against a concentration gradient in exchange for protons, utilizing energy derived from ATP (1)(2)(3)(4)(5)(6). It belongs to the family of P-type ATPases in which the hydrolysis of ATP is linked with ion translocation through phosphorylation and dephosphorylation of an aspartic acid residue in the ATPase protein (Scheme 1). Transfer of the ␥-phosphoryl group of ATP to the protein is activated by a series of protein conformational changes associated with the binding of two calcium ions at cytoplasmically facing sites in exchange for protons ("the Ca 2ϩ binding transition" H n E 2 3 Ca 2 E 1 ). The dephosphorylation (H n E 2 P 3 H n E 2 ) is triggered by Ca 2ϩ /H ϩ exchange at luminally facing sites. Structural studies of the Ca 2ϩ -ATPase have revealed a large cytoplasmic head, which through a stalk is connected with the transmembrane domain (7). The head is made up of two cytoplasmic loops of the peptide chain, and the phosphorylated aspartic acid residue (Asp 351 ) resides in the largest of these. The membrane domain comprises ten transmembrane helices (M1-M10), encompassing sites for binding and translocation of Ca 2ϩ and H ϩ (3,8,9). The stalk consists of extensions of some of the transmembrane helices continuing into the cytoplasmic head. The tight coupling between the catalytic and vectorial processes, associated with the cytoplasmic head and the membrane domain, respectively, is dependent on intramolecular communication through the interconnecting stalk. Mutations to residues in stalk segment S4 have been shown to impair the transition between the two major phosphoforms of the enzyme, Ca 2 E 1 P 3 H n E 2 P, suggesting an important role of this segment in transmission of ATP-derived energy required in the membrane domain for Ca 2ϩ translocation (3,10,11). Stalk segment S5 could be another important mediator of the intramolecular communication, but its role is less well understood compared with the role of S4. We have previously demonstrated that replacement of Lys 758 in S5 with isoleucine leads to concurrent increase in the rate of dephosphorylation H n E 2 P 3 H n E 2 and decrease in the rate of the Ca 2ϩ binding transition H n E 2 3 Ca 2 E 1 (12). Furthermore, the tight coupling between ATP hydrolysis and Ca 2ϩ translocation has been found to be abolished by replacement of Tyr 763 at the M5S5 boundary with glycine (13). In the present study, we have inquired about the functional importance of many of the other residues in stalk segment S5. Would mutations to other S5 residues give rise to functional consequences similar to those observed for the Lys 758 3 Ile and the Tyr 763 3 Gly mutants, and do stalk segments S4 and S5 play separate roles in intramolecular signaling? In addition to Lys 758 , several other residues in the S5 region are well conserved among the P-type ATPases, such as the charged or polar residues Asp 737 , Asp 738 , Asn 739 , and Arg 751 , the aromatic residues Phe 740 and Tyr 754 , the aliphatic residues Ile 743 , Val 744 , and Val 747 , and Gly 750 (14). We have analyzed the importance of these and other residues located in S5 and at the M5S5 boundary (see Fig. 1), by studying the functional consequences of their replacement for the overall and partial reactions of the enzyme.

EXPERIMENTAL PROCEDURES
Mutagenesis and Expression-The principles for construction of mutant Ca 2ϩ -ATPase cDNAs and expression of wild type and mutants in COS-1 cells have been described previously (15). Microsomes containing expressed wild-type or mutant Ca 2ϩ -ATPase were harvested by differential centrifugation, and expression levels were quantified by a specific sandwich enzyme-linked immunosorbent assay (16), using as standard expressed wild type for which the concentration was determined by measurement of the capacity for phosphorylation by inorganic phosphate in the presence of 30% dimethyl sulfoxide (12). Total microsomal protein was determined by the dye binding method of Bradford (17).
ATP Hydrolysis and Ca 2ϩ Uptake-For determination of Ca 2ϩ -activated ATPase activity, microsomes were incubated for 10 min at 37°C either in the presence or absence of 1 M calcium ionophore A23187 in a medium containing 50 mM TES/Tris, 1 pH 7.0, 100 mM KCl, 7 mM MgCl 2 , 5 mM ATP, 1 mM EGTA, and various concentrations of CaCl 2 to set the desired free Ca 2ϩ concentration, as calculated using the MAXC computer program and constants therein (18). The ATP hydrolysis catalyzed by the microsomes was determined by measuring the amount of P i liberated (12) using the method of Baginski et al. (19). ATP hydrolysis referable to Ca 2ϩ -ATPase activity was calculated following subtraction of the amount of P i liberated in the presence of 4 mM EGTA without added Ca 2ϩ . The turnover rate was calculated as mol P i liberated/mol Ca 2ϩ -ATPase/s. To study the vanadate inhibition of ATPase activity, the indicated concentrations of monovanadate were obtained from metavanadate as described previously (20). 45 Ca 2ϩ uptake in the microsomal vesicles was determined in the absence of ionophore and in the presence of 5 mM potassium oxalate (to act as Ca 2ϩ trap inside the vesicles), using a Millipore filtration method (12).
Phosphorylation from [␥-32 P]ATP and Dephosphorylation-To study the time course of phosphorylation of enzyme initially present in the Ca 2ϩ -deprived form, microsomes preincubated in a medium containing 40 mM MOPS/Tris, pH 7.0, 80 mM KCl, and 1 or 2 mM EGTA were mixed with a medium of a composition producing final concentrations of 40 mM MOPS/Tris, pH 7.0, 80 mM KCl, 2 or 5 M [␥-32 P]ATP, 5 mM MgCl 2 , and 0.1 mM free Ca 2ϩ , followed by acid quenching at serial time intervals by further mixing with the double volume of 25% (w/v) trichloroacetic acid containing 100 mM H 3 PO 4 . The time course of phosphorylation of enzyme initially present in the Ca 2ϩ -saturated form was studied in the same way except that the medium used for preincubation contained 0.1 mM Ca 2ϩ instead of EGTA. For experiments at pH 6.0, 40 mM MES adjusted to pH 6.0 with Tris was present instead of 40 mM MOPS/Tris. The phosphorylation reactions were carried out either at 0°C, using a manual mixing technique (12), or at 25°C, using the Bio-Logic quench flow module QFM-5 (Bio-Logic Instruments, Claix, France) as described previously (21).
The quench flow technique was also used to study Ca 2ϩ dissociation from the enzyme by monitoring the phosphorylation following simultaneous addition of EGTA and [␥-32 P]ATP to Ca 2ϩ -saturated enzyme (21). Microsomes suspended in 40 mM MOPS/Tris, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , and 100 M CaCl 2 were mixed with an equal volume of 40 mM MOPS/Tris, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 4 mM EGTA, and 10 M [␥-32 P]ATP, followed by acid quenching 34 ms later.
Dephosphorylation of enzyme phosphorylated by [␥-32 P]ATP was studied at 0°C using the manual mixing technique (12). Microsomes were phosphorylated for 15 s at 0°C in a medium containing 40 mM MOPS/Tris, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 100 M CaCl 2 , 2 M [␥-32 P]ATP, and 1 M calcium ionophore A23187, followed by addition, 5 s prior to acid quenching, of either nonradioactive ATP (purified by ionic exchange chromatography to remove contaminant ADP) to a final concentration of 600 M, or ADP with EGTA to final concentrations of 1 mM each. In all the above described phosphorylation experiments, the acidprecipitated protein was washed by centrifugation and subjected to SDS-polyacrylamide gel electrophoresis according to Weber and Osborn (22) in a 7% polyacrylamide gel at pH 6.0. Radioactivity associated with the separated Ca 2ϩ -ATPase band was quantified by imaging using a Packard Cyclone Storage Phosphor System.
Phosphorylation by 32 P i and Dephosphorylation-For phosphorylation from 32 P i , microsomes were incubated 10 min at 25°C in a medium containing 100 mM MES/Tris, pH 6.0, 10 mM MgCl 2 , 2 mM EGTA, 30% (v/v) dimethyl sulfoxide, and 0.5 mM 32 P i . Dimethyl sulfoxide was present to increase the P i affinity, thus ensuring that 0.5 mM P i was saturating in the wild type (23,24). To study dephosphorylation, the phosphorylated microsomes were cooled to 0°C and subsequently diluted 20-fold into ice-cold medium containing 50 mM TES/Tris, pH 7.0, 2 mM MgCl 2 , 2 mM EGTA, and 5 mM nonradioactive P i , with or without 100 mM KCl, followed by acid quenching at serial intervals (2, 4, 6, and 10 s after the dilution). The acid-quenched protein was analyzed as described above.
Data Analysis and Presentation-All data presented in the figures and table or described in the text are average values of 2-5 experiments. Standard deviations are given in the table and are shown in the figures when larger than the size of the symbol. Time courses of phosphorylation or dephosphorylation were either fitted to monoexponential functions corresponding to first-order kinetics, using the SigmaPlot program (SPSS, Inc.), or were fitted by computer simulation of a simplified reaction cycle as described previously (21).

RESULTS
Expression-Mutants analyzed in the present study are presented in Fig. 1. An expression level of 200 -400 pmol Ca 2ϩ -ATPase/mg of total microsomal protein was achieved for the wild-type Ca 2ϩ -ATPase, i.e. several hundred-fold higher than that of the endogenous COS cell Ca 2ϩ -ATPase. The various mutants represented by single-letter code without parentheses in Fig. 1 were expressed to wild-type level, whereas the expression levels were very low for the mutations shown in parentheses. The low levels of expression of mutants Phe 740 3 Ala, Ile 743 3 Gly, Val 747 3 Gly, and Gly 750 3 Leu precluded functional characterization but less drastic alterations to the same residues, mutations Phe 740 3 Leu, Ile 743 3 Ala, Val 747 3 Ala, and Gly 750 3 Ala resulted in expression to wild-type level, thereby enabling characterization of the functional importance of these residues. For Arg 751 , only the conservative replacement with lysine gave rise to a protein that could be expressed to a level comparable with the wild-type level, whereas insignificant expression was found for mutants with alanine, isoleucine, or glutamate substitutions, either by enzyme-linked immunosorbent assay or by Western blotting (Fig. 2). Two inversion mutants moving Arg 751 one position in either direction were also produced, Gly 750 -Arg 751 3 Arg 750 -Gly 751 and Arg 751 -Ala 752 3 Ala 751 -Arg 752 , but as illustrated in Fig. 2, only the Arg 751 -Ala 752 3 Ala 751 -Arg 752 mutant could be expressed to a significant level.
Ca 2ϩ and Ionophore Dependence of ATPase Activity-To determine the overall functional consequences of the various amino acid substitutions, we measured the steady-state turnover number for ATPase activity at 37°C in the presence of 5 mM MgATP and various Ca 2ϩ concentrations, with and with- out the calcium ionophore A23187 (25). The calcium ionophore allows passive efflux of Ca 2ϩ that has been actively transported into the microsomal vesicles, and for the wild-type enzyme, this leads to a 2-fold increase in the maximum turnover rate, because the luminal Ca 2ϩ concentration is diminished, thereby relieving the "back inhibition" of the rate-limiting Ca 2 E 1 P 3 H n E 2 P transition caused by Ca 2ϩ binding at luminal low affinity sites. All the mutants that could be expressed to a normal level, except Arg 751 3 Lys and the inversion mutant Arg 751 -Ala 752 3 Ala 751 -Arg 752 , displayed ATPase activity, and the turnover rates determined at optimum Ca 2ϩ concentration are given in Table I, both for the presence and absence of ionophore. The maximal turnover rates of mutants Val 744 3 Gly, Val 747 3 Ala, and Arg 762 3 Ile were considerably lower than that of the wild type. Furthermore, mutants Val 744 3 Gly, Val 747 3 Ala, and Gly 750 3 Ala showed diminished activation by ionophore, and for mutant Arg 762 3 Ile we observed an anomalous, slightly inhibitory, effect of the ionophore on the ATPase activity, similar to that previously described for mutant Lys 758 3 Ile (12).
The disappearance of the activating effect of the calcium ionophore could be the consequence of uncoupling of ATP hydrolysis from Ca 2ϩ transport (13). To test for uncoupling, 45 Ca 2ϩ uptake in the microsomes was determined. All mutants displaying ATPase activity, including Val 744 3 Gly, Val 747 3 Ala, Gly 750 3 Ala, and Arg 762 3 Ile, were able to transport Ca 2ϩ actively, and the ratios between Ca 2ϩ transport and ATPase activity did not differ significantly among mutants and wild type (data not shown). Hence, there was no indication of an uncoupling of ATPase activity from Ca 2ϩ transport in any of the mutants. In this situation, the reduced activation by the calcium ionophore may be explained by a shift of the ratelimiting step away from the Ca 2 E 1 P 3 H n E 2 P transition sensitive to back inhibition by accumulated Ca 2ϩ . In fact, an inhibitory effect of the ionophore on the turnover rate, as observed for mutant Arg 762 3 Ile, might be expected under conditions where the Ca 2ϩ binding transition, H n E 2 3 Ca 2 E 1 , is rate-limiting, because the ionophore exerts a direct inhibitory effect on this step (12). When added in large doses, the ionophore causes strong inhibition of the Ca 2ϩ binding transition of the wild type (26), and even the low ionophore concentration used here to make the membranes leaky should exert some inhibitory effect. In the wild type, the effect on the steady-state ATPase activity of this inhibition is masked because of the simultaneous relief of the back inhibition of the Ca 2 E 1 P 3 H n E 2 P transition, but if the Ca 2ϩ binding transition has be-  a The maximum turnover rate (mol P i liberated/mol Ca 2ϩ -ATPase/s) shown relative to the maximum turnover rate of the wild type determined in the presence of calcium ionophore A23187 (cf. Fig. 3). The value corresponding to 100% is 129 s Ϫ1 .
b The Ca 2ϩ concentration giving half-maximum activation of ATP hydrolysis (cf. Fig. 3). c Rate constant corresponding to phosphorylation of enzyme preincubated in the presence of EGTA (cf. Fig. 4). d Apparent rate constants corresponding to dephosphorylation of phosphoenzyme formed in the presence of [␥-32 P]ATP. The phosphoenzyme was chased with 600 M nonradioactive ATP (ATP chase) or with 1 mM ADP (ADP chase). The rate constants were calculated from the fraction of phosphoenzyme remaining after a 5-s chase.
e Rate constants corresponding to dephosphorylation of phosphoenzyme formed in the presence of 32 P i . To study dephosphorylation, the phosphoenzyme was diluted 20-fold into a medium with (ϩK ϩ ) or without (ϪK ϩ ) K ϩ . f ND, not determined. g The rate was too high for an accurate determination of the rate constant and only the lower limit of 2 s Ϫ1 could be estimated.
come rate-limiting for the cycle, the inhibitory effect of the ionophore should manifest itself as a decrease in the turnover rate, as actually seen for mutant Arg 762 3 Ile. Examples of Ca 2ϩ titration of the ATPase activity in the presence of ionophore are shown in Fig. 3, and the K 0.5 values for Ca 2ϩ activation are indicated for all the mutants in Table I. The activation by Ca 2ϩ reflects Ca 2ϩ binding at the cytoplasmically facing high affinity sites, whereas the inhibition seen at high Ca 2ϩ concentrations reflects in part the back inhibition by Ca 2ϩ binding at the luminal sites and in part the effect of replacing MgATP by CaATP (27)(28)(29). In mutants Val 744 3 Gly, Val 744 3 Ala, and Val 747 3 Ala, the apparent affinity for Ca 2ϩ at the activating sites was 1.5-2-fold increased (K 0.5 decreased) relative to wild type. By contrast, the apparent affinity was reduced 2-and 4-fold, respectively, in mutants Gly 750 3 Ala and Arg 762 3 Ile. In the other mutants, the K 0.5 values for Ca 2ϩ activation did not deviate or deviated only insignificantly from wild type.
Phosphorylation by [␥-32 P]ATP-Having established the impact of the various mutations on the overall enzyme function, we studied the partial reactions. First, the Ca 2ϩ -activated phosphorylation by [␥-32 P]ATP was determined after 15 s of incubation with 2 M [␥-32 P]ATP in the presence of 100 M Ca 2ϩ at 0°C. All the mutants exhibiting ATPase activity were able to form a phosphoenzyme intermediate under these conditions, whereas the two expressed mutants that were inactive with respect to ATP hydrolysis, Arg 751 3 Lys and Arg 751 -Ala 752 3 Ala 751 -Arg 752 , showed no measurable phosphorylation. Further examination showed that the latter mutants were unable to phosphorylate also in the presence of 10 mM Ca 2ϩ at 0°C as well as in the presence of 100 M or 10 mM Ca 2ϩ at 25°C (data not shown).
The kinetics of the reaction sequence consisting of dissociation of protons from H n E 2 , Ca 2ϩ and ATP binding, and subsequent phosphorylation to form Ca 2 E 1 P and H n E 2 P (Scheme 1) were studied at 0°C at pH 7.0. The enzyme was preincubated in the absence of Ca 2ϩ (presence of EGTA), and phosphorylation was initiated by the simultaneous addition of CaCl 2 and [␥-32 P]ATP to yield final concentrations of 100 M Ca 2ϩ and 2 M [␥-32 P]ATP. Examples of the data are shown in Fig. 4. The time courses of phosphorylation could be fitted to a monoexponential function, and the derived rate coefficient is given in Table I for all the mutants. A significant enhancement of the rate of phosphorylation relative to wild type was found for six mutants, more than 2-fold in mutants Val 747 3 Ala (shown in  Fig. 4), and Tyr 754 3 Ala, a significant decrease in the phosphorylation rate was found (in Asp 738 3 Ala and Ala 752 3 Leu the rate was about half that of wild type).
Under the experimental conditions applied here, the phosphorylation rate of the wild type is limited by one or more of the steps in the Ca 2ϩ binding transition (H n E 2 3 Ca 2 E 1 ) and not by ATP binding or phosphoryl transfer (Ca 2 E 1 3 Ca 2 E 1 P) (12). Therefore, it may be concluded that the rate of the Ca 2ϩ binding transition is enhanced relative to wild type in mutants Ile 743 3 Ala, Val 747 3 Ala, Glu 748 3 Ala, Glu 749 3 Ala, Met 757 3 Gly, and Gln 759 3 Ala.
To locate the step limiting the rate in the slowly phosphorylating mutants, Asp 737 3 Ala, Asp 738 3 Ala, Ala 752 3 Leu, and Tyr 754 3 Ala, these mutants were further examined in experiments where 2 M [␥-32 P]ATP was added to enzyme preincubated with 100 M Ca 2ϩ (i.e. initially present in the Ca 2ϩsaturated Ca 2 E 1 form), under conditions otherwise similar to those in Fig. 4. As a consequence of the preincubation with Ca 2ϩ , the rate of phosphorylation increased considerably, so that more than 80% of the maximal phosphorylation level was reached after 1 s, the lower time limit for data collection by the manual mixing technique (data not shown). Under these conditions, the Ca 2ϩ binding transition is excluded from the reaction sequence studied, and it appears, therefore, that the slow step limiting the phosphorylation rate of these mutants in the experiments carried out after preincubation in the absence of Ca 2ϩ must be the Ca 2ϩ binding transition and not ATP binding or phosphoryl transfer.
Dephosphorylation of Phosphoenzyme Formed from [␥-32 P]ATP-The Ca 2 E 1 P phosphoenzyme intermediate can dephosphorylate either by donation of the bound phosphoryl group back to ADP, producing ATP or through the Ca 2 E 1 P 3 H n E 2 P transition and subsequent hydrolysis of the aspartyl phosphoryl bond, H n E 2 P 3 H n E 2 (Scheme 1). The H n E 2 P phosphoenzyme intermediate reacts with water liberating P i but cannot react with ADP. We examined the dephosphorylation of phosphoenzyme formed in the presence of [␥-32 P]ATP under conditions (0°C, presence of K ϩ , neutral pH) where the Ca 2 E 1 P form tends to accumulate in the wild type as a consequence of a relatively slow Ca 2 E 1 P 3 H n E 2 P transition (15, 28 -30). The fraction of phosphoenzyme remaining 5 s after chase with an excess of either nonradioactive ATP or ADP was determined and the derived rate constants for the two types of chase are FIG. 3. Ca 2؉ dependence of Ca 2؉ -ATPase activity. The rate of ATP hydrolysis in the wild type (q) and mutants Val 747 3 Ala (E), Gly 750 3 Ala (OE), and Arg 762 3 Ile (ƒ) was determined in the presence of 5 mM ATP at pH 7.0 and 37°C at various Ca 2ϩ concentrations in the presence of calcium ionophore A23187 as described under "Experimental Procedures," and the turnover rates (mol P i liberated/mol Ca 2ϩ -ATPase/s) are shown relative to the maximum turnover rate of the wild type, which was 129 s Ϫ1 . The K 0.5 values determined from these data are given in Table I  shown in Table I. It can be seen that the dephosphorylation of the wild type occurred more rapidly upon addition of ADP as compared with ATP, consistent with the presence of a large fraction of the wild-type phosphoenzyme as ADP-sensitive Ca 2 E 1 P. For some of the mutants, notably Val 744 3 Gly and Val 747 3 Ala, the data shown in Table I indicate a significant reduction of the dephosphorylation rate relative to wild type, in the presence of ATP as well as in the presence of ADP, suggesting that the reaction H n E 2 P 3 H n E 2 occurs at a reduced rate with resulting accumulation of the ADP-insensitive H n E 2 P form. None of the mutants showed the combination of a significantly reduced dephosphorylation rate in the presence of ATP and a normal or increased dephosphorylation rate in the presence of ADP that would be consistent with a block of the Ca 2 E 1 P3 H n E 2 P transition (cf. Refs. 10, 11 and 15).
Phosphoenzyme Formed from 32 P i -All the mutants exhibiting ATPase activity and phosphorylation from ATP were able to form a phosphoenzyme also in the backward reaction with 32 P i in the absence of Ca 2ϩ (the ADP-insensitive H n E 2 P intermediate; Scheme 1), whereas no phosphorylation from 32 P i could be detected for the two expressed mutants that were inactive with respect to ATP hydrolysis and phosphorylation from ATP, Arg 751 3 Lys and Arg 751 -Ala 752 3 Ala 751 -Arg 752 (data not shown).
The dephosphorylation of the phosphoenzyme formed from 32 P i was studied upon a 20-fold dilution of the phosphorylation mixture into a chase medium containing nonradioactive P i . For many of the mutants the dephosphorylation was examined both in the presence and absence of K ϩ in the chase medium (Table I). K ϩ is known to stimulate dephosphorylation in the wild-type Ca 2ϩ -ATPase by binding to sites facing the cytoplasmic side of the membrane without being transported (31,32). The data could be fitted to monoexponential decay functions as illustrated by a few examples in Fig. 5, and the derived rate constants are given in Table I. The smaller rate constants relative to wild type determined for Val 744 3 Gly and Val 747 3 Ala in the presence of K ϩ confirm that the reason for the reduced ADP sensitivity of the phosphoenzyme of these mutants noted above is a reduced rate of dephosphorylation of H n E 2 P with resulting accumulation of this intermediate. Likewise, the dephosphorylation rate was found to be somewhat reduced relative to wild type in mutants Asp 737 3 Ala, Val 744 3 Ala, and Ala 752 3 Leu in the presence of K ϩ (Table I), also consistent with the data for the ADP sensitivity of the phosphoenzyme formed from ATP.
The mutants Ile 743 3 Ala and Tyr 754 3 Ala displayed a conspicuous increase of the dephosphorylation rate relative to wild type both in the presence and absence of K ϩ . In the presence of K ϩ , the rate was too high for accurate measurement in these mutants, but in the absence of K ϩ the rate constants could be determined and were close to 20-fold higher than that corresponding to wild type (Table I and Fig. 5). In the absence of K ϩ , the rate constants corresponding to mutants Asn 739 3 Ala, Glu 749 3 Ala, Gly 750 3 Ala, Met 757 3 Gly, and Arg 762 3 Ile were found to be 2-10-fold higher than that of the wild type. For Glu 749 3 Ala, Gly 750 3 Ala, Met 757 3 Gly, and Arg 762 3 Ile, the rate constant was determined in the presence of K ϩ , as well, and under these conditions it was 1.5-2-fold enhanced relative to wild type (Table I).
Rapid Kinetic Studies of the Ca 2ϩ Binding Transition-Because the simplest explanation of the anomalous ionophore effect described above for mutant Arg 762 3 Ile is a shift of the rate-limiting step away from the Ca 2 E 1 P 3 H n E 2 P transition to the Ca 2ϩ binding transition, we were puzzled over the apparently normal rate of the Ca 2ϩ binding transition determined for this mutant (0.27 s Ϫ1 ; Table I). Because this result had been obtained at 0°C, the kinetics of the Ca 2ϩ binding transition in mutant Arg 762 3 Ile and other selected mutants were further analyzed at 25°C, using the quench flow technique described previously (21). Except for the higher temperature and a higher ATP concentration of 5 M instead of 2 M, the experimental protocol for the left panels of Fig. 6 was similar to that used for Fig. 4 and Table I, phosphorylation being initiated by simultaneous addition of Ca 2ϩ and [␥-32 P]ATP to Ca 2ϩ -deprived enzyme. By fitting the data in the FIG. 5. Dephosphorylation of phosphoenzyme formed in the presence of 32 P i . Wild type (q) and mutants Ile 743 3 Ala (‚), Tyr 754 3 Ala (f), and Arg 762 3 Ile (ƒ) were phosphorylated for 10 min at 25°C in the presence of 0.5 mM 32 P i . Subsequently, the phosphorylation mixture was cooled to 0°C, and dephosphorylation was studied at 0°C and pH 7.0 in the absence of K ϩ , by a 20-fold dilution of the phosphorylated enzyme into an ice-cold medium, using a manual mixing procedure (see "Experimental Procedures"). The data were analyzed as monoexponential decays, and the extracted rate coefficients are shown in Table I  Procedures"), following preincubation either in the absence of Ca 2ϩ (to study H n E 2 3 Ca 2 EP, left panels) or in the presence of 100 M Ca 2ϩ (to study Ca 2 E 1 3 Ca 2 EP, right panels). The time course of phosphorylation of enzyme preincubated in the absence of Ca 2ϩ was fitted to a monoexponential function with rate coefficients of 21, 19, 11, and 6 s Ϫ1 for the wild type and mutants Val 747 3 Ala, Ala 752 3 Leu, and Arg 762 3 Ile, respectively. The time course of phosphorylation of enzyme preincubated in the presence of Ca 2ϩ was fitted by computer simulation of a simplified reaction cycle as described previously (21). In each case the maximum level of phosphorylation was set to 100%. left panels to monoexponential functions, rate constants of 21, 19, 11, and 6 s Ϫ1 were obtained for the wild type and mutants Val 747 3 Ala, Ala 752 3 Leu, and Arg 762 3 Ile, respectively. Hence, the observed rate constant for mutant Arg 762 3 Ile was 3.5-fold lower than that of the wild type at 25°C and 5 M ATP (Fig. 6, left panel), whereas it was wild type like at 0°C and 2 M ATP (Table I). The observed rate constant for mutant Val 747 3 Ala was wild type like at 25°C and 5 M ATP (Fig. 6, left panels), whereas it was more than 2-fold enhanced relative to wild type at 0°C and 2 M ATP (Table I). For mutant Ala 752 3 Leu, the observed rate constant was about half that of the wild type under both sets of conditions.
The right panels in Fig. 6 show the results of initiating the phosphorylation by [␥-32 P]ATP addition to enzyme preincubated in the presence of Ca 2ϩ , under conditions otherwise similar to those applied corresponding to the left panels. After the preincubation with Ca 2ϩ , the initial rates of phosphorylation were very similar for wild type and mutants and higher than the rates observed following preincubation in the absence of Ca 2ϩ . As already noted in connection with Fig. 4, this means that the time courses observed following preincubation in the absence of Ca 2ϩ (i.e. those displayed in the left panels of Fig. 6) reflect the Ca 2ϩ binding transition and not ATP binding or phosphoryl transfer. The right panels of Fig. 6 furthermore show an initial overshoot of phosphorylation. This is caused by the occurrence of a relatively slow step between dephosphorylation and rephosphorylation. The overshoot is seen to be largest for mutant Arg 762 3 Ile, consistent with a very low rate of the Ca 2ϩ binding transition in this mutant.
Additional phosphorylation experiments conducted at 25°C as in Fig. 6 (left panels) Table I and Fig. 6 (left panels) with respect to the rates of mutants Val 747 3 Ala and Arg 762 3 Ile relative to wild type is caused by the temperature difference and not by the difference in ATP concentration (data not shown).
Hence, it may be concluded that the temperature dependence of the rate of the Ca 2ϩ binding transition is less steep for mutants Val 747 3 Ala and Arg 762 3 Ile compared with the wild type and mutant Ala 752 3 Leu. This would be explained if the energy of activation of a rate-limiting substep is changed by the mutation or if a different substep has taken over the role of being rate-limiting. According to Forge et al. (5), the Ca 2ϩ binding transition of the wild type comprises at least four substeps between H 3 E 2 and Ca 2 E 1 in a branched scheme containing three or more Ca 2ϩ -free species with different degrees of protonation. At room temperature in the absence of Ca 2ϩ , almost all the enzyme resides in the H 3 E 2 form at pH 6.0, whereas HE 1 prevails at neutral pH (5). To look for possible changes to the proton dissociation steps in the Val 747 3 Ala mutant, we examined the kinetics of the wild type and the Val 747 3 Ala mutant at pH 6.0 under conditions otherwise similar to those corresponding to the left panels of Fig. 6, and the rate coefficients were 3.8 and 8.4 s Ϫ1 , respectively (data not shown), which should be compared with the corresponding values of 21 and 19 s Ϫ1 obtained at pH 7.0 (see above). This demonstrates that the Val 747 3 Ala mutant is less sensitive to the pH change than the wild type, which is reminiscent of the lower temperature sensitivity of the Val 747 3 Ala mutant relative to the wild type. The finding that for mutant Val 747 3 Ala the rate of the Ca 2ϩ binding transition is wild-type like at 25°C at pH 7.0, whereas it is enhanced relative to wild type both at 0°C at pH 7.0 and at 25°C at pH 6.0, would be consistent with the hypothesis that the same step (possibly H 3 E 2 3 HE 1 ; cf. Ref. 5) limits the rate of the Ca 2ϩ binding transition in the latter two conditions and that this step is accelerated in the mutant relative to wild type, whereas a different step, which is unaffected by the mutation, is rate-limiting at 25°C at pH 7.0 both in wild type and mutant.
Inhibition by Vanadate-As a transition state analog of the phosphoryl group, vanadate binds preferentially to the H n E 2 form of the enzyme (20,33), and the apparent affinity for vanadate depends on the concentration of H n E 2 relative to the other enzyme intermediates, as well as on the intrinsic affinity of H n E 2 for vanadate. The intrinsic affinity is related to the stability of the transition state in the H n E 2 P 3 H n E 2 dephosphorylation reaction and is expected to increase under conditions where the rate of dephosphorylation is enhanced (34 -36). Hence, an increased dephosphorylation rate or a reduced rate of the Ca 2ϩ binding transition (resulting in accumulation of H n E 2 ) would both be expected to be associated with an increased apparent affinity for vanadate. To obtain further information about these partial reactions in selected mutants at 37°C under the conditions used to study the overall functional performance of the enzyme, the vanadate affinity was examined by titrating the vanadate inhibition of ATPase activity, and the results are shown in Fig. 7. In mutants Ile 743 3 Ala, Tyr 754 3 Ala, and Arg 762 3 Ile, the apparent affinities for vanadate were enhanced 23-, 30-, and 18-fold, respectively, relative to wild type, consistent with stabilization of the transition state in the dephosphorylation reaction and/or accumulation of H n E 2 in these mutants at 37°C. Only a 2-fold increase in apparent vanadate affinity relative to wild type was found for the Ala 752 3 Leu mutant, which could reflect a balance between effects related to a reduced dephosphorylation rate (Table I) and a reduced rate of the Ca 2ϩ binding transition (Fig. 6). For the Val 747 3 Ala mutant, the apparent vanadate affinity was 1.4-fold reduced relative to wild type, possibly reflecting a reduced dephosphorylation rate (Table I).
Dissociation of Ca 2ϩ -To understand the basis for the altered apparent affinity for Ca 2ϩ observed for mutants Val 747 3 Ala, Gly 750 3 Ala, and Arg 762 3 Ile in the Ca 2ϩ titrations of ATPase activity (Fig. 3), we analyzed the rate of Ca 2ϩ dissociation in these mutants. Upon addition of EGTA to Ca 2ϩ -saturated enzyme, the ability to phosphorylate disappears at a rate corresponding to the dissociation of the calcium ion that is first to leave toward the cytoplasmic side in the sequential mechanism (37), and it is therefore possible to determine this rate by phosphorylation measurements (37, 38). We have previously described a simple approach in which the apparent rate constant for Ca 2ϩ dissociation (k ϪCa ) is calculated from the ratio EP ATPϩEGTA /EP ATP , where EP ATPϩEGTA is the amount of phosphoenzyme measured 34 ms after simultaneous addition of [␥-32 P]ATP and EGTA and EP ATP is the amount of phosphoenzyme measured after 34 ms of incubation of the Ca 2ϩ -saturated enzyme with [␥-32 P]ATP in the absence of EGTA (21). Results obtained by this procedure are displayed in Fig. 8, and it is seen that the rate of Ca 2ϩ dissociation is enhanced about 11-fold in the Gly 750 3 Ala mutant, relative to wild type, whereas it is wild type like in mutant Arg 762 3 Ile and about half that of the wild type in Val 747 3 Ala. DISCUSSION In the present study, we have found that various mutations to residues in stalk segment S5 of the Ca 2ϩ -ATPase have effects on the rate of the reaction sequence H n E 2 3 Ca 2 E 1 associated with Ca 2ϩ binding at the cytoplasmically facing sites (the Ca 2ϩ binding transition) and on the rate of the dephosphorylation of the ADP-insensitive phosphoenzyme, H n E 2 P 3 H n E 2 , normally triggered by Ca 2ϩ /H ϩ exchange at the luminal sites. Both accelerating and decelerating effects on these two reaction sequences were observed. At 0°C, the rate of the Ca 2ϩ binding transition was enhanced in six mutants (more than 2-fold in Val 747 3 Ala and Met 757 3 Gly), whereas a decrease in the rate was noted in another four mutants (most significant in Asp 738 3 Ala and Ala 752 3 Leu). In mutant Arg 762 3 Ile, the rate of the Ca 2ϩ binding transition was wild type like at 0°C, whereas it was as much as 3.5-fold reduced relative to wild type at 25°C. The dephosphorylation of H n E 2 P was enhanced in eight mutants (as much as 20-fold in mutants Ile 743 3 Ala and Tyr 754 3 Ala, as determined in the absence of K ϩ ) and reduced in five mutants, with resulting accumulation of H n E 2 P at steady state (particularly in Val 744 3 Gly and Val 747 3 Ala). In addition, a conspicuous 11-fold increase and a 2-fold decrease in the rate of Ca 2ϩ dissociation were observed for mutants Gly 750 3 Ala and Val 747 3 Ala, respectively, thus highlighting the functional interaction between S5 and the Ca 2ϩ binding structure in the membrane (Fig. 8).
Taken together, the observed mutational effects suggest that S5 plays an important role in controlling events in the Ca 2ϩ /H ϩ transport sites and in communicating changes in their occupancy to the catalytic site. The conspicuous acceleration of dephosphorylation seen for some of the S5 mutants in the present study and previously for the Lys 758 3 Ile mutant (12) may represent a bypass of the requirement for luminal Ca 2ϩ /H ϩ exchange to trigger dephosphorylation (note that in Ref. 12 the Lys 758 3 Ile mutant was shown to dephosphorylate rapidly even at alkaline pH and in the presence of a high Ca 2ϩ concentration on both sides of the membrane). It is furthermore noteworthy that in the present study the enhancement of the dephosphorylation rate was seen for some of the mutants, notably Met 757 3 Gly and Arg 762 3 Ile, to be more pronounced in the absence of K ϩ than in its presence, suggesting that the normal regulation of the dephosphorylation by K ϩ , binding at cytoplasmically facing sites without being transported (31,32), is bypassed to some extent.
Most of the data on the overall functional performance of the mutants measured as steady-state ATP hydrolysis activity (Figs. 3 and 7 and Table I) can be explained on the basis of changes to the above mentioned partial reactions. The maximal turnover rate for ATP hydrolysis was found to be considerably reduced relative to wild type in mutants Val 744 3 Gly, Val 747 3 Ala, and Arg 762 3 Ile. In mutants Val 744 3 Gly and Val 747 3 Ala, the rate of ATP hydrolysis is likely to be limited by the slow dephosphorylation of H n E 2 P. For mutant Arg 762 3 Ile, the Ca 2ϩ binding transition is a good candidate for a rate-limiting reaction sequence at 37°C under the conditions used for determination of the ATPase activity. Although the rate of the Ca 2ϩ binding transition of this mutant was wild type like at 0°C, it was as much as 3.5-fold reduced relative to wild type at 25°C, and given this less steep temperature dependence relative to wild type, the rate of the Ca 2ϩ binding transition at 37°C may be considerably reduced in the Arg 762 3 Ile mutant compared with wild type. This would also be consistent with our observation of an anomalous, slightly inhibitory, effect of the calcium ionophore on the ATPase activity of this mutant.
The titration of the Ca 2ϩ dependence of the turnover rate for ATP hydrolysis demonstrated an increased apparent Ca 2ϩ affinity relative to wild type in mutants Val 744 3 Gly, Val 744 3 Ala, and Val 747 3 Ala, whereas the apparent Ca 2ϩ affinity was reduced in Gly 750 3 Ala and Arg 762 3 Ile. For mutants Val 747 3 Ala, Gly 750 3 Ala, and Arg 762 3 Ile, we determined the rate of Ca 2ϩ dissociation (Fig. 8) and in case of Val 747 3 Ala and Gly 750 3 Ala the result of this measurement, showing a reduced and an increased rate, respectively, explains the change in apparent Ca 2ϩ affinity. For Arg 762 3 Ile, the Ca 2ϩ dissociation rate was found to be normal, but the reduced rate of the Ca 2ϩ binding transition explains the reduced apparent Ca 2ϩ affinity.
For mutants Ile 743 3 Ala, Tyr 754 3 Ala, and Arg 762 3 Ile, we found significant increases in the apparent affinity for vanadate determined under steady-state ATP hydrolysis conditions at 37°C (Fig. 7). Because vanadate binds preferentially to the H n E 2 form of the enzyme as a transition state analog of the phosphoryl group, the data are consistent with the increased dephosphorylation rate seen particularly for Ile 743 3 Ala and Tyr 754 3 Ala and with the reduced rate of the Ca 2ϩ binding transition of Arg 762 3 Ile.
Although the functional effects of several hundred different mutations to the sarcoplasmic reticulum Ca 2ϩ -ATPase have been examined, the combination of changes to the Ca 2ϩ binding transition of the dephosphoenzyme and the dephosphorylation of H n E 2 P, as observed in the present study, has only been previously reported for another S5 mutant, Lys 758 3 Ile (12), and for the mutant Glu 309 3 Asp, in which the critical Ca 2ϩ liganding residue in M4 is altered (39). Certain mutations in other regions such as the smaller cytoplasmic domain ("␤domain") and stalk segment S4 have been shown to affect the rate of the Ca 2ϩ binding transition, but in these cases there was a simultaneous block of the Ca 2 E 1 P 3 H n E 2 P transition (21). In the light of the previous findings of reduced Ca 2 E 1 P 3 H n E 2 P transition rates in mutants with alterations to residues in stalk segment S4 (3, 10, 11), it is remarkable that no block of FIG. 8. Rate of Ca 2؉ dissociation. For the wild type and each mutant, the phosphoenzyme level measured at 25°C after 34 ms of exposure of Ca 2ϩ -saturated enzyme to excess EGTA, and [␥-32 P]ATP as described under "Experimental Procedures" is shown as the percentage of the amount of phosphoenzyme determined after 34 ms of incubation with [␥-32 P]ATP in the presence of Ca 2ϩ without EGTA. To the right are shown the rate constants for Ca 2ϩ dissociation (k ϪCa ), calculated as described previously (21). The calculation was based on a rate constant of 35 s Ϫ1 for the Ca 2 E 1 3 Ca 2 E 1 P reaction in wild type as well as mutants (21), because experiments performed as in Fig. 6 (right panels) showed no significant differences between the initial phosphorylation rates.
the Ca 2 E 1 P 3 H n E 2 P transition was seen for any of the S5 mutants examined in the present study. Instead, the rate of dephosphorylation of H n E 2 P was altered in the S5 mutants, and in several cases it was enhanced. This difference points to a mechanism in which S4 and S5 play distinct roles in energy transduction. S4 takes part in the conformational changes associated with the Ca 2 E 1 P 3 H n E 2 P transition, whereas S5 may be instrumental in transmitting conformational changes from the Ca 2ϩ /H ϩ -binding sites in the membrane sector to the phosphorylation domain, controlling phosphorylation and dephosphorylation.
One S5 residue, Arg 751 , appeared to be crucial both to the structural and the functional integrity of the enzyme. Arg 751 is very highly conserved within the P-type ATPase family and according to the present findings it is important for the expression of the ATPase protein in the COS cells. Probably, the arginine is needed to attain the conformation required for proper membrane insertion. Only the conservative substitution Arg 751 3 Lys and the inversion mutation Arg 751 -Ala 752 3 Ala 751 -Arg 752 were compatible with expression at a significant level, and even these mutants were unable to hydrolyze ATP or become phosphorylated by ATP or P i . Also the substitution of the neighboring Gly 750 with leucine reduced enzyme expression and disrupted function. This could be explained by interference of the bulky leucine side chain with that of Arg 751 , because the Gly 750 3 Ala mutant was expressed to wild-type level. In this connection, it is interesting that one of the missense mutations identified in patients with Darier-White disease results in replacement of Gly 750 with arginine (40). The present findings are in accordance with the hypothesis that this mutation disrupts function and probably also the expression, of the Ca 2ϩ -ATPase.
The very recently published 2.6 Å resolution structure of the Ca 2ϩ -ATPase crystallized in Ca 2 E 1 form (41) provides important clues to understanding the present and previous (12) observations on S5 mutants. M5 and S5 seem to constitute a continuous 60 Å long and fairly straight ␣-helical structure located in the center of the protein. S5 is linked to the transmembrane domain and the Ca 2ϩ binding structure comprised by M4, M5, M6, and M8, not only through its connection with M5 but also through a network of hydrogen bonds and van der Waals interactions with residues in the loop between transmembrane segments M6 and M7 ("L6 -7") and with residues in M7 and M8 (41). Of particular importance for the interpretation of the present results is that Arg 751 participates in several interactions (including one or two hydrogen bonds) with residues between Ile 816 and Pro 821 in the L6 -7 loop and that Arg 762 interacts with Arg 836 and Asn 914 in M7 and M8, respectively. Lys 758 , studied previously (12), is hydrogen-bonded to the backbone carbonyl oxygen of Leu 828 in the L6 -7 loop. Thus, the key roles that we have demonstrated for these residues may be understood in terms of the contacts they form between the membranous and cytoplasmic domains through these bonds, stabilizing the overall structure of the protein and establishing a functional linkage between the Ca 2ϩ binding structure and the catalytic site. The difference between the functional roles of S4 and S5 may be related to the bonds between S5 and the L6 -7 loop. These bonds may allow changes in the position of the L6 -7 loop, elicited by changes in the occupancy of the ion binding sites, to be transmitted through S5 to the catalytic domain. It is noteworthy in this connection that Gly 750 is likely to introduce a certain flexibility in S5 that may be required for the correct positioning of the side chain of Arg 751 and, thus, the L6 -7 loop, to ensure a proper function of the Ca 2ϩ sites. Thus, the enhanced Ca 2ϩ dissociation rate in mutant Gly 750 3 Ala may be the consequence of a slight change in the position of the Arg 751 side chain with resulting distortion of the L6 -7 loop. Likewise, the effects of glycine and alanine substitutions of the larger hydrophobic/aromatic side chains of Phe 740 , Ile 743 , Val 744 , Val 747 , Tyr 754 , and Met 757 , which do not have particular interactions with side chains in other domains in the crystallized Ca 2 E 1 form (41), may be ascribed to destabilization of the helix structure, disrupting interactions involving S5 residues different from those being substituted.