Critical Roles of Interdomain Interactions for Modulatory ATP Binding to Sarcoplasmic Reticulum Ca2+-ATPase*

Background: ATP stimulates dephosphorylation of Ca2+-ATPase. Results: ATP affinities of intermediate states in the dephosphorylation are altered by certain mutations interfering with interactions between A-, P-, and N-domains. Conclusion: Disruption of ATP modulation by mutation is explained by destabilization of the enzyme-phosphoryl transition state with bound nucleotide. Significance: Mechanisms underlying the modulatory effect of ATP and the importance therein of interdomain bonds are elucidated. ATP has dual roles in the reaction cycle of sarcoplasmic reticulum Ca2+-ATPase. Upon binding to the Ca2E1 state, ATP phosphorylates the enzyme, and by binding to other conformational states in a non-phosphorylating modulatory mode ATP stimulates the dephosphorylation and other partial reaction steps of the cycle, thereby ensuring a high rate of Ca2+ transport under physiological conditions. The present study elucidates the mechanism underlying the modulatory effect on dephosphorylation. In the intermediate states of dephosphorylation the A-domain residues Ser186 and Asp203 interact with Glu439 (N-domain) and Arg678 (P-domain), respectively. Single mutations to these residues abolish the stimulation of dephosphorylation by ATP. The double mutation swapping Asp203 and Arg678 rescues ATP stimulation, whereas this is not the case for the double mutation swapping Ser186 and Glu439. By taking advantage of the ability of wild type and mutant Ca2+-ATPases to form stable complexes with aluminum fluoride (E2·AlF) and beryllium fluoride (E2·BeF) as analogs of the E2·P phosphoryl transition state and E2P ground state, respectively, of the dephosphorylation reaction, the mutational effects on ATP binding to these intermediates are demonstrated. In the wild type Ca2+-ATPase, the ATP affinity of the E2·P phosphoryl transition state is higher than that of the E2P ground state, thus explaining the stimulation of dephosphorylation by nucleotide-induced transition state stabilization. We find that the Asp203-Arg678 and Ser186-Glu439 interdomain bonds are critical, because they tighten the interaction with ATP in the E2·P phosphoryl transition state. Moreover, ATP binding and the Ser186-Glu439 bond are mutually exclusive in the E2P ground state.

ATP has dual roles in the reaction cycle of sarcoplasmic reticulum Ca 2؉ -ATPase. Upon binding to the Ca 2 E1 state, ATP phosphorylates the enzyme, and by binding to other conformational states in a non-phosphorylating modulatory mode ATP stimulates the dephosphorylation and other partial reaction steps of the cycle, thereby ensuring a high rate of Ca 2؉ transport under physiological conditions. The present study elucidates the mechanism underlying the modulatory effect on dephosphorylation. In the intermediate states of dephosphorylation the A-domain residues Ser 186 and Asp 203 interact with Glu 439 (N-domain) and Arg 678 (P-domain), respectively. Single mutations to these residues abolish the stimulation of dephosphorylation by ATP. The double mutation swapping Asp 203 and Arg 678 rescues ATP stimulation, whereas this is not the case for the double mutation swapping Ser 186 and Glu 439 . By taking advantage of the ability of wild type and mutant Ca 2؉ -ATPases to form stable complexes with aluminum fluoride (E2⅐AlF) and beryllium fluoride (E2⅐BeF) as analogs of the E2⅐P phosphoryl transition state and E2P ground state, respectively, of the dephosphorylation reaction, the mutational effects on ATP binding to these intermediates are demonstrated. In the wild type Ca 2؉ -ATPase, the ATP affinity of the E2⅐P phosphoryl transition state is higher than that of the E2P ground state, thus explaining the stimulation of dephosphorylation by nucleotide-induced transition state stabilization. We find that the Asp 203 -Arg 678 and Ser 186 -Glu 439 interdomain bonds are critical, because they tighten the interaction with ATP in the E2⅐P phosphoryl transition state. Moreover, ATP binding and the Ser 186 -Glu 439 bond are mutually exclusive in the E2P ground state.
Using the energy liberated by ATP hydrolysis the sarco (endo)plasmic reticulum Ca 2ϩ -ATPase pumps Ca 2ϩ ions into the endoplasmic reticulum, a critical aspect of a variety of important physiological processes in animal cells such as contraction and relaxation of muscle and secretion of hormones, enzymes, and neurotransmitters. The Ca 2ϩ -ATPase is an integral membrane protein made up of 10 transmembrane helixes (M1-M10) connected to a large cytoplasmic headpiece comprised of three distinct and loosely connected domains, N ("nucleotide binding"), P ("phosphorylation"), and A ("actuator"). Ca 2ϩ transport is achieved by means of a reaction cycle (Scheme 1) involving the formation and decay of an aspartylphosphorylated intermediate coupled to protein conformational changes, whereby the Ca 2ϩ binding sites sequentially alter affinity for Ca 2ϩ and exposure to the cytosol and endoplasmic reticulum lumen, thus enabling the translocation of Ca 2ϩ to the lumen against a large Ca 2ϩ gradient. The determination by x-ray crystallography of atomic structures of the Ca 2ϩ -AT-Pase in several different physiologically relevant conformations has, in combination with a wealth of functional studies, provided unique insight into the nanomachinery of the Ca 2ϩ -AT-Pase (reviewed in Refs. 1 and 2). Thus, during the reaction cycle, the A-and N-domains undergo large displacements relative to the P-domain. Phosphorylation from ATP bound mainly to the N-and P-domains in the E1 conformation is triggered by the binding of the two Ca 2ϩ ions in the membrane domain forming the Ca 2 E1 state. Crystal structures indicate that Ca 2ϩ binding to M4 directs a movement of M1-M2 required for bending of the P-domain to form a more compact structure of the ATP site that allows the phosphoryl transfer to the Asp 351 carboxylate group (3,4). The dephosphorylation of E2P occurs as a consequence of the movement of the conserved phosphatase motif of the A-domain, 181 TGES, into close contact with the Asp 351 aspartyl phosphate, allowing the side chain of the glutamate Glu 183 to catalyze a nucleophilic attack of the aspartyl phosphate in a S N 2 reaction, forming a pentacoordinated phosphoryl transition state (5,6). The dephosphorylation proceeds from the E2P ground state 2 through the E2⅐P phosphoryl transition state and into the E2⅐P i product state, and these enzyme intermediate states ("E2P-like states") are mimicked by the complexes of the Ca 2ϩ -ATPase with the respective metal fluorides BeF, AlF, 3 and MgF (6 -9).
ATP is an important mediator of interdomain interactions, with all three cytoplasmic domains contributing to nucleotide binding at various stages of the reaction cycle (9 -17). Aside from being the phosphorylating substrate in the Ca 2 E1 state, ATP also fulfills a role as a non-phosphorylating modulator of the pump cycle, accelerating certain partial reaction steps without being hydrolyzed (boxed ATP in Scheme 1) (16 -29). Some of the residues that contribute to ATP binding in the phosphorylating mode in E1 are also involved in ATP binding in the modulatory mode, in particular N-domain residues. When the A-domain approaches the P-domain in E2 and E2P-like states, the N-domain is partially displaced, leading to a less compact structure of the ATP site and affinities for modulatory ATP 10 -100-fold lower than the affinity of E1 for phosphorylating ATP (17). The apparent affinity for ATP binding to E2P is further lowered in the presence of Mg 2ϩ , because only free ATP and not MgATP can bind to this state, which already contains Mg 2ϩ bound with the phosphoryl group (26,27). The critical nature of the modulatory modes of ATP binding is reflected by the fact that Ca 2ϩ -transport/ATPase activity at the micromolar concentrations of ATP sufficient to saturate the active site in Ca 2 E1, phosphorylating the enzyme, is only a fraction of the activity at physiological (millimolar) ATP concentrations, where binding of modulatory ATP is saturated (e.g. Fig. 1 in Ref. 29). Using the metal fluorides BeF, AlF, and MgF as analogs of the phosphoryl group in the various intermediate states we recently demonstrated that stimulation by ATP of E2P dephosphorylation in the wild type enzyme is reflected in the ATP affinity profiles of the intermediates occurring in the dephosphorylation reaction sequence (17). Hence, the ATP affinity increases from the E2P ground state to the E2⅐P phosphoryl transition state and then decreases again from the E2⅐P i product state to the E2 ground state, suggesting that ATP accelerates E2P dephosphorylation by tightening the interactions around the catalytic site in the E2⅐P phosphoryl transition state, thereby stabilizing this intermediate and lowering the activation energy of the reaction. 2 The structural basis for this variation of ATP affinity along the reaction coordinate of dephosphorylation has not been clear, and in the present work we focus on the important roles of two interdomain interactions: the ionic bond between Asp 203 (A-domain) and Arg 678 (P-domain) and the hydrogen bond between Ser 186 (A-domain) and Glu 439 (N-domain). Common to these two interactions is that they seem to exist only in the E2P class of substates of the pump cycle (i.e. in the E2P ground state and the E2⅐P phosphoryl transition state). Hence, from crystal structures it appears that the Asp 203 -Arg 678 and Ser 186 -Glu 439 bonds break when the enzyme liberates P i and enters the E2 state, and the residues remain apart throughout the rest of the reaction cycle as a consequence of the A-domain movements (cf. Fig. 1 and Table 1). Ser 186 , Asp 203 , Glu 439 , and Arg 678 have been implicated as critical for Ca 2ϩ -ATPase function (16, 28, 30 -33), and both interaction sites are positioned close to the bound nucleotide in the crystal structure of the Ca 2ϩ -ATPase stabilized in the E2⅐AlF⅐AMP-PCP state (9), with the Asp 203 -Arg 678 interaction near the ␥-phosphate of AMP-PCP and the Ser 186 -Glu 439 interaction near the adenine moiety ( Fig.  1). We now demonstrate the importance of the two cytoplasmic domain interactions for the progression of E2P dephosphorylation as well as the ATP modulation of the dephosphorylation reaction, and by analyzing the ATP affinity profiles of the intermediates in mutants we provide insight in the mechanism underlying the modulatory effect.

EXPERIMENTAL PROCEDURES
Site-directed mutagenesis of cDNA encoding the rabbit fast twitch muscle Ca 2ϩ -ATPase (SERCA1a isoform) inserted into the pMT2 vector (34) was carried out using the QuikChange site-directed mutagenesis kit (Agilent Technologies), and the mutant cDNA was sequenced throughout. The cDNA encoding mutants E439A and R678A was the same as used previously (17,28). To express wild type and mutant cDNA, COS-1 cells were transfected using the calcium phosphate precipitation method (35). Microsomal vesicles containing either wild type or mutant Ca 2ϩ -ATPase were isolated by differential centrifugation (36). The concentration of expressed Ca 2ϩ -ATPase was determined by an enzyme-linked immunosorbent assay (37) and by determination of the maximum capacity for phosphorylation with ATP or P i ("active site concentration" (38)).
Formation of the complexes of wild type or mutant Ca 2ϩ -ATPase in the E2 state with AlF or BeF prior to photolabeling was achieved by pre-equilibration of the enzyme for 30 min at 2 Throughout this text the terms "E2P ground state" and "E2⅐P phosphoryl transition state" are used to designate the enzyme intermediates with the bound phosphoryl group in the ground state and transition state, respectively, as occurring during dephosphorylation (5,6). According to classic transition state theory, the catalytic rate of an enzyme increases with the stability of the enzyme-substrate transition state complex, due to lowering of the activation energy of the reaction. 3 The abbreviations used are: AlF, complex of Al 3ϩ and fluoride; AMP-PCP, adenosine 5Ј-(␤,␥-methylene)-triphosphate; EPPS, N-2-hydroxyethylpiperazine-NЈ-3-propanesulfonic acid; K 0.5 , ligand concentration giving half-maximum effect; TMAH, tetramethyl ammonium hydroxide; TNP, trinitrophenyl; TNP-8N 3 -ATP, 2Ј,3Ј-O-(2,4,6-trinitrophenyl)-8-azidoadenosine 5Ј-triphosphate; TNP-AMP, 2Ј,3Ј-O-(2,4,6-trinitrophenyl)-adenosine 5Ј-monophosphate. SCHEME 1. Ca 2؉ -ATPase reaction cycle. Major conformational changes and substrate binding and dissociation steps are shown. Boxed ATP indicates steps for which the rate is enhanced by additional binding of ATP or MgATP in a non-phosphorylating mode, i.e. without being hydrolyzed ("modulatory ATP"). Encircled AlF, VO 4 3Ϫ (orthovanadate), and BeF are indicated below the reaction intermediates, for which stable analogs can be formed by incubation of Ca 2ϩ -deprived enzyme with the respective phosphate analogs in the presence of Mg 2ϩ .  The enzyme-inhibitor complexes were formed immediately prior to the initiation of the photolabeling experiments and were kept on ice throughout (Ͻ1 h). The high stability of the E2⅐AlF, E2⅐BeF, and E2⅐orthovanadate complexes during the course of the photolabeling experiments has been documented previously (17).
The synthesis of the [␥-32 P]TNP-8N 3 -ATP photolabel, its application as a specific photolabel of the Ca 2ϩ -ATPase, the competitive inhibition by ATP of [␥-32 P]TNP-8N 3 -ATP photolabeling, and the quantification of 32 P-labeled bands by electronic autoradiography following SDS-PAGE were carried out using the previously established procedures (13,39). The presently applied experimental setup for the [␥-32 P]TNP-8N 3 -ATP photolabeling of the Ca 2ϩ -ATPase is described in detail in Ref. Measurements of phosphorylation from [␥-32 P]ATP or 32 P i were carried out by acid quenching followed by acid SDS-polyacrylamide gel electrophoresis and quantification of the radioactivity associated with the Ca 2ϩ -ATPase band using the previously established procedures (16,28,40). For studies of the ATP dependence of dephosphorylation of E2P, phosphorylation with 0.5 mM 32 P i was carried out for 10 min at 25°C in 100 mM MES/Tris (pH 6.0), 10 mM MgCl 2 , 2 mM EGTA, and 30% (v/v) dimethyl sulfoxide. The phosphorylated sample was chilled in ice water, and dephosphorylation was followed at 0°C by a 19-fold dilution into ice-cold medium containing 50 mM MOPS/Tris (pH 7.0), 2 mM EGTA, 10 mM EDTA, 5 mM H 3 PO 4 , and various concentrations of ATP.
The data were analyzed by nonlinear regression using the SigmaPlot program (SPSS, Inc.). The analysis of the TNP-8N 3 -ATP photolabeling data were based on the hyperbolic func- in which Y is the amount of photolabeled Ca 2ϩ -ATPase, Y max is the maximum amount of photolabeled Ca 2ϩ -ATPase, K 0.5 is the concentration of TNP-8N 3 -ATP giving half-maximum labeling, and m⅐[TNP-8N 3 -ATP] is a linear background component, which has been subtracted from the data shown (13). The analysis of the data obtained from ATP inhibition of TNP-8N 3 -ATP photolabeling was based on the Hill equation modified to describe inhibition, in which Y and Y max are defined as above, K 0.5 is the concentration of ATP giving half-maximum effect, and n is the Hill coefficient (varying between 0.74 and 1.02 for the present data). The "true" dissociation constant, K D , for ATP binding was calculated from the measured K 0.5 values using the validated equation for competitive inhibition (13). For analysis of the modulatory effect of ATP on the rate of E2P dephosphorylation, the ATP concentration dependence of the rate constant was analyzed according to the hyperbolic function, k obs ϭ k 0 ϩ (k max Ϫ k 0 ) [ATP]/(K 0.5 ϩ [ATP]), in which k obs is the rate constant observed at the indicated ATP concentration, k 0 is the rate constant in the absence of ATP ("basal rate"), and k max is the extrapolated value of the rate constant corresponding to infinite ATP concentration (16). The experiments were conducted at least twice on independent microsomal preparations, and average values are shown on the graphs, with error bars when larger than the size of the symbols.

Design and Expression of the Mutant Ca 2ϩ -ATPases-To
address the importance of the Asp 203 -Arg 678 (A-P domains) and Ser 186 -Glu 439 (A-N domains) interaction sites in the E2P states, we analyzed four mutations replacing these residues individually with the residue corresponding to the interaction partner (mutants D203R, R678D, S186E, E439S) as well as two double mutations, D203R/R678D and S186E/E439S, in which the side chains of the interaction partners were swapped. In addition, the point mutations R678Q, S186A, S186P, and Q202A as well as the double mutation Q202A/D203A were included in the study, the latter two motivated by previous reports that replacement of Gln 202 with alanine causes reduced Ca 2ϩ transport (32) and ATPase activity (33). The wild type and mutant Ca 2ϩ -ATPase constructs were expressed in COS-1 cells to similar high levels, allowing the study of the kinetics of the partial reactions of the pump cycle as well as direct measurements of nucleotide binding.
Rate and ATP Dependence of E2P Dephosphorylation-The rate of dephosphorylation of E2P phosphoenzyme was measured by first phosphorylating Ca 2ϩ -deprived enzyme by 32 P i in the backward direction of the normal reaction cycle (cf. Scheme 1) and then chasing the E2P phosphoenzyme with EDTA to remove free Mg 2ϩ , thereby terminating phosphorylation. Fig. 2 shows the time courses observed in the absence of ATP in the dephosphorylation medium, with the extracted rate constants (the basal rate "k 0 ") listed in Table 2. Fig. 3 shows the E2P dephosphorylation rate constants obtained for wild type and mutants at varying concentrations of ATP in the dephosphorylation buffer. The apparent affinities for modulatory ATP (K 0.5 ) extracted from the analysis are listed in Table 2 together with the ratio between the dephosphorylation rates at 1 and 0 mM ATP (k 1 /k 0 ).
Mutations to the Asp 203 -Arg 678 interaction site inhibited E2P dephosphorylation, most markedly for D203R, which displayed a ϳ5-fold lower basal dephosphorylation rate compared with the wild type. Importantly, the swap mutant D203R/R678D displayed a basal dephosphorylation rate ϳ2-fold higher than that of the D203R point mutant, corresponding to a partial restoration of the dephosphorylation rate, which may result from ionic interaction between the A-and P-domains through the swapped side chains at positions 203 and 678.
For the wild type, the E2P dephosphorylation rate was ϳ3-fold higher in the presence of 1 mM ATP, relative to the basal rate in the absence of ATP, and the ATP concentration required for half-maximal stimulation of E2P dephosphorylation was 65 M (Fig. 3 and Table 2). Mutant D203R displayed a conspicuous reduction of the ATP sensitivity, with very little enhancement of the dephosphorylation rate up to 200 M ATP, implying a strongly reduced affinity of mutant D203R for modulatory ATP binding (K 0.5 estimate Ͼ1 mM, Fig. 3 and Table 2). The mutants with single substitution of Arg 678 displayed little or no stimulatory effect of ATP on the dephosphorylation rate in the 0 -1 mM ATP concentration range (reflected by the k 1 /k 0 ATP enhancement factors being close to 1, cf. Fig. 3 and Table  2). This indicates that Arg 678 is critical for modulatory ATP binding, or, at least, for making the binding productive in terms of activation of dephosphorylation. Remarkably, in the swap mutant D203R/R678D ATP modulation of E2P dephosphorylation was much less affected as compared with the respective mutants with single substitutions D203R and R678D ( Fig. 3 and Table 2). Hence, for D203R/R678D, the dephosphorylation rate increased by a factor of 2.2 over the 0 -1 mM ATP concentration range with a K 0.5(ATP) of 181 M, thus clearly demonstrating a gain of function relative to the D203R and R678D point mutants, which might result from the interaction between the swapped side chains.
Also Q202A and Q202A/D203A displayed markedly reduced basal E2P dephosphorylation rates. However, only the double mutant involving Asp 203 showed a significantly reduced modulatory effect of ATP, whereas Q202A displayed a slightly higher affinity for modulatory ATP than the wild type and wild type-like k 1 /k 0 enhancement factor ( Fig. 3 and Table 2).
Mutations to individual residues Ser 186 and Glu 439 generally resulted in substantial increases of the basal E2P dephosphorylation rate, 9-fold in the case of E439S and as much as 18-fold for S186P, relative to the wild type. S186E was an exception, showing little effect on the basal dephosphorylation rate. The swap mutation S186E/E439S restored the basal dephosphorylation rate to a level only ϳ2-fold faster than that of the wild type ( Fig. 2 and Table 2). However, for S186E/E439S as well as the mutants with single substitutions of Ser 186 and Glu 439 no or little stimulation of the dephosphorylation by ATP was seen FIGURE 2. Rate of E2P dephosphorylation in the absence of ATP. The microsomes containing the enzyme were incubated with 32 P i in the presence of Mg 2ϩ , as described under "Experimental Procedures." The accumulated E2P phosphoenzyme was then chased by dilution into ATP-free dephosphorylation medium containing excess EDTA to remove Mg 2ϩ , followed by acid quenching after various time intervals as indicated on the abscissa. The dephosphorylation rate constants, obtained by fitting of an exponential decay function to the data, are listed in Table 2 ("basal rate", k 0 ). The broken lines reproduce the wild type data from the upper left panel for direct comparison.

TABLE 2 Basal rates, affinity constants, and enhancement factors for the modulation by ATP of E2P dephosphorylation
The basal dephosphorylation rate constant (k 0 ), the ATP concentration giving halfmaximum activation (K 0.5(ATP) ), and the dephosphorylation rate constant at 1 mM ATP (k 1 ) were derived from the data shown in Figs     in the 0 -1 mM concentration range (Fig. 3 and Table 2). Hence, although the Ser 186 -Glu 439 swap mutation rescues the basal dephosphorylation rate of the E439S mutant to a wild type-like level, it does not rescue the stimulation by ATP of dephosphorylation. Nucleotide Affinity in Analog States of the E2P Dephosphorylation Reaction Sequence-To understand the mechanism underlying the altered ATP sensitivity of the E2P dephosphorylation rate seen for some of the mutants we proceeded to study the ATP affinity of individual states of the dephosphorylation reaction sequence, the E2P ground state represented by E2⅐BeF (7,9), and the E2⅐P phosphoryl transition state represented by E2⅐AlF (6, 7). Like the wild type enzyme, the mutants formed stable complexes of the E2 state with the metal fluorides, as proved by measuring the phosphorylation from [␥-32 P]ATP in the presence of Ca 2ϩ following pre-equilibration with AlF or BeF in the absence of Ca 2ϩ . Because of the mutually exclusive binding of these metal fluoride phosphoryl analogs and the ␥-phosphate of ATP at the phosphorylation site, and the low rate of dissociation of the metal fluorides, the 32 P incorporation determined under these conditions was inhibited. Hence, as demonstrated below, we were able to determine the affinity of these complexes for the modulatory ATP by use of our previously validated TNP-8N 3 -ATP photolabeling assay (13,17). Fig. 4 shows the [␥-32 P]TNP-8N 3 -ATP concentration dependence of [␥-32 P]TNP-8N 3 -ATP photolabeling of the wild type and mutants in E2⅐AlF (circles) and E2⅐BeF (triangles) states, with the extracted K 0.5 values for the [␥-32 P]TNP-8N 3 -ATP binding listed in Table 3. The binding site for TNP-8N 3 -ATP overlaps sufficiently with the binding site for modulatory ATP in the E2 states to allow competition binding assays in which the affinity for modulatory ATP is obtained from the ATP concentration dependence of inhibition of TNP-8N 3 -ATP photolabeling (17). Results of such competition experiments are shown in Fig. 5, with the calculated ATP affinity constants (K D ) listed in Table 3. Two mutants, R678D and R678Q, showed poor labeling with TNP-8N 3 -ATP in either metal fluoride complexed state, thus preventing the determination of their affinity for the photolabel or for ATP. For the remaining mutants, the levels of photolabeling at saturating concentrations of TNP-8N 3 -ATP were as high as those obtained with the wild type in the E2⅐AlF state as well as the E2⅐BeF state, thus allowing determination of the K 0.5 for TNP-8N 3 -ATP and the K D for ATP. Importantly, this was also the case for the swap mutant D203R/R678D. Furthermore, whereas the point mutation D203R markedly lowered TNP-8N 3 -ATP affinity of both E2⅐AlF and E2⅐BeF (4-and 9-fold, respectively), as well as ATP affinity (12-and 30-fold, respectively), the affinity obtained with D203R/R678D was wild type-like for both TNP-8N 3 -ATP and ATP in the case of E2⅐AlF and only slightly reduced for E2⅐BeF (2.1-fold for TNP-8N 3 -ATP and 1.4-fold for ATP, relative to wild type), thus once again implying that the Asp 203 -Arg 678 side chain swap is much less disruptive than each of the individual D203R and R678D mutations (Figs. 4 and 5, and Table 3).
The affinity of the wild type for TNP-8N 3 -ATP and ATP is 4and 3-fold higher, respectively, in the E2⅐AlF state than in E2⅐BeF (Figs. 4 and 5, and Table 3, BeF/AlF affinity constant ratio 4 and 3, respectively, for the wild type). In the wild type, ATP binding therefore appears to stabilize the E2⅐P phosphoryl transition state (mimicked by E2⅐AlF) relative to the E2P ground state (mimicked by E2⅐BeF), thus accounting for the stimulatory effect of ATP on dephosphorylation. Mutants with alterations to the Ser 186 -Glu 349 interaction site differed markedly from the wild type enzyme in this respect. Hence, for S186A and S186P there was only a marginal difference between the ATP affinities of E2⅐AlF and E2⅐BeF states, and for mutants S186E and E439S, as well as S186E/E439S, the ATP affinity was actually higher in E2⅐BeF than in E2⅐AlF (Table 3, BeF/AlF  Table 3. K D(ATP) ratio Ͻ 1), thus explaining the lack of ATP stimulation of E2P dephosphorylation seen for these mutants (Fig. 3). For the Ser 186 and Glu 439 single substitutions, both an ATP affinity increase in E2⅐BeF and an ATP affinity decrease in E2⅐AlF, relative to wild type, contributed to make the affinity constant ratio differ markedly from that of the wild type. For the swap mutant the situation was different, because a marked 25-fold increase of the ATP affinity in E2⅐BeF constituted the major reason for the higher ATP affinity in E2⅐BeF compared with E2⅐AlF. In fact, the ATP affinity of E2⅐AlF was slightly higher for S186E/E439S than for the wild type and, notably, 4-fold higher than for the point mutants S186E and E439S, thus showing gain of a wild type-like function in the swap mutant with respect to ATP binding to E2⅐AlF, i.e. the E2⅐P phosphoryl transition state with bound ATP was more stable in the swap mutant than in the point mutants. However, due to the very high affinity of the E2⅐BeF state for ATP the ratio between the K D(ATP) values of E2⅐BeF and E2⅐AlF in S186E/E439S was still only a fraction of that of the wild type (Table 3). For ATP binding to E2⅐BeF, no gain of a wild type-like function was obtained with S186E/ E439S relative to the point mutants, because the ATP affinity was increased even more in S186E/E439S (25-fold) than in S186E (7-fold) and E439S (14-fold), thus showing that the ground state with bound ATP was more stable in the swap mutant than in the point mutants.
The ATP affinity was unaffected by the Q202A mutation in the E2⅐AlF state and only ϳ2-fold increased relative to wild type in the E2⅐BeF state, which is in accordance with the wild typelike ATP modulation of dephosphorylation in this mutant (Fig.  3). The double mutation Q202A/D203A lowered ATP affinity as much as 7-fold in E2⅐AlF but only 1.5-fold in E2⅐BeF, thus explaining the reduced sensitivity of the dephosphorylation rate to ATP modulation in this mutant.
To confirm the gain of wild type-like function observed for the D203R/R678D swap mutant, in comparison with the corresponding point mutants, we also carried out TNP-8N 3 -ATP photolabeling experiments with wild type and mutants D203R, R678D, and D203R/R678D stabilized in the E2⅐P phosphoryl transition state-like conformation by use of another phosphoryl transition state analog, orthovanadate (17,41,42). As seen in Fig. 6A, the D203R and R678D point mutations reduced the affinity for orthovanadate 16-and 17-fold, respectively, whereas the D203R/R678D swap mutant displayed 43-fold lower affinity for orthovanadate than the wild type enzyme. Fig.  6B shows the TNP-8N 3 -ATP concentration dependence of photolabeling of wild type and mutants inhibited by a saturating concentration (0.5 mM) of orthovanadate. Similar to the situation with R678D in E2⅐AlF and E2⅐BeF states, the E2⅐vanadate state of this mutant was poorly labeled with TNP-8N 3 -ATP (not shown). In contrast, the E2⅐vanadate state of the D203R/R678D swap mutant was labeled to the same extent as the wild type, and indeed with an affinity for TNP-8N 3 -ATP almost identical to that of the wild type (Fig. 6B). The gain-offunction of the D203R/R678D swap mutant with respect to TNP-8N 3 -ATP binding is further supported by the fact that point mutation D203R lowered the affinity for the photolabel ϳ8-fold relative to the wild type in the E2⅐vanadate state. A similar picture emerged from the titrations of the competitive inhibition by ATP of TNP-8N 3 -ATP photolabeling shown in Fig. 6C. Hence, mutant D203R displayed 48-fold reduced ATP affinity relative to wild type in the E2⅐vanadate state, contrasting the wild type-like ATP affinity of the D203R/R678D swap mutant.

DISCUSSION
During the dephosphorylation of the E2P state of the Ca 2ϩ -ATPase, the conserved Glu 183 of the 181 TGES motif of the A-domain positions the water molecule responsible for aspartyl phosphate hydrolysis, and the optimal geometry for the nucleophilic attack of the aspartyl phosphate by water is reached in the pentacoordinated E2⅐P phosphoryl transition state (5,6). ATP seems to accelerate E2P dephosphorylation by tightening the interactions around the catalytic site in the E2⅐P phosphoryl transition state, thereby lowering the activation energy of the reaction (17). This modulatory effect of ATP is a consequence of acceleration of the forward dephosphorylation reaction and Affinity for TNP-8N 3 -ATP and ATP of E 2 ⅐AlF and E 2 ⅐BeF states K 0.5 values for TNP-8N 3 -ATP binding and K D values for the competitive inhibition by ATP of TNP-8N 3 -ATP photolabeling were derived from the data shown in Figs. 4 and 5, respectively, and are indicated relative (in %) to that of the wild type obtained under the same conditions (for wild type the absolute K 0.5 and K D values are shown bracketed). The S.E. is indicated with the number of experiments in parentheses. The two columns denoted "BeF/AlF" list the respective TNP-8N 3 -ATP and ATP affinity constants obtained in E 2 ⅐BeF state relative to those obtained in E 2 ⅐AlF state. 70 Ϯ 6 (n ϭ 2) 176 Ϯ 7 (n ϭ 2) 9.67 89 Ϯ 8 (n ϭ 2) 61 Ϯ 5 (n ϭ 3) 2.08 Q202A/D203A 281 Ϯ 20 (n ϭ 2) 208 Ϯ 12 (n ϭ 2) 2.84 678 Ϯ 75 (n ϭ 2) 148 Ϯ 13 (n ϭ 2) 0.66 S186A 97 Ϯ 6 (n ϭ 2) 46 Ϯ 2 (n ϭ 2) 1.81 128 Ϯ 7 (n ϭ 2) 49 Ϯ 3 (n ϭ 3) 1.14 S186E

K 0.5 (TNP-8N 3 -ATP) K D(ATP
408 Interdomain Interactions in SERCA OCTOBER 17, 2014 • VOLUME 289 • NUMBER 42 is not due to prevention by ATP of reversal of the dephosphorylation, because it is observable in the absence of a significant concentration of P i and both by transient kinetic studies and at steady state (28). We present here the results of investigating the importance for the dephosphorylation reaction and its modulation by ATP of two cytoplasmic domain interactions seen in the E2P-like conformations, where they connect the A-domain to the P-and N-domains, respectively (Fig. 1). Mutations to the four residues (Asp 203 , Arg 678 , Ser 186 , and Glu 439 ) that make up the two interaction sites markedly affected the basal rate of E2P dephosphorylation as well as the stimulation of this partial reaction step by FIGURE 5. ATP concentration dependence of inhibition of TNP-8N 3 -ATP photolabeling in stable E2P-like states. The enzyme was preincubated with AlF or BeF, as described under "Experimental Procedures" and subjected to photolabeling at a TNP-8N 3 -ATP concentration of 3 times the K 0.5 for TNP-8N 3 -ATP in the presence of the indicated concentrations of ATP. In each case, the maximum level of specific labeling was defined as 100%. Symbols for all panels are indicated in the upper left panel. The broken lines reproduce the wild type data from the upper left panel for direct comparison. The affinity constants extracted from fits of the Hill equation for inhibition to the data (cf. "Experimental Procedures") are listed in Table 3. Note the reversed order of ATP affinity in E2⅐AlF and E2⅐BeF of mutants S186E, E439S, and S186E/E439S, relative the wild type enzyme.  (n ϭ 2). B, TNP-8N 3 -ATP concentration dependence of photolabeling of enzyme in the E2⅐orthovanadate state. The enzyme was preincubated with orthovanadate as described under "Experimental Procedures," at a saturating concentration of 0.5 mM, and subjected to TNP-8N 3 -ATP photolabeling at the indicated concentrations of TNP-8N 3 -ATP. In each case, the maximum level of specific labeling was defined as 100%. The lines show the best fits of a hyperbolic function to the data giving the following affinity constants: wild type, K 0.5 ϭ 8.3 Ϯ 0.7 nM (n ϭ 2); D203R, K 0.5 ϭ 70.0 Ϯ 8.8 nM (n ϭ 2); D203R/R678D, K 0.5 ϭ 6.6 Ϯ 0.4 nM (n ϭ 2). For R678D, no specific labeling by TNP-8N 3 -ATP was obtained. C, ATP concentration dependence of inhibition of TNP-8N 3 -ATP photolabeling of enzyme in the E2⅐orthovanadate state. Enzyme preincubated with 0.5 mM orthovanadate was subjected to TNP-8N 3 -ATP photolabeling at 3 times the K 0.5 for TNP-8N 3 -ATP in the presence of the indicated concentrations of ATP. For R678D, the lack of specific TNP-8N 3 -ATP labeling precluded determination of ATP dependence. In each case, the maximum level of specific labeling was defined as 100%. The lines show the best fits of the Hill equation for inhibition to the data giving the following affinity constants: wild type, K D ϭ 0.86 Ϯ 0.03 M (n ϭ 2); D203R, K 0.5 ϭ 41.1 Ϯ 8. ATP. Dissection of the mutational effects on ATP stimulation of E2P dephosphorylation, resolved by studying the nucleotide affinity of analogs of the intermediate states in the dephosphorylation reaction sequence, revealed a clear difference between the functions of the two interaction sites. The point mutations to Asp 203 and Arg 678 (A-P domains interaction), but importantly not the D203R/R678D swap mutation, were detrimental to ATP binding both in the E2P ground state (mimicked by E2⅐BeF) and in the E2⅐P phosphoryl transition state (mimicked by E2⅐AlF) with consequent disruption of the stimulatory effect of ATP on dephosphorylation. The point mutations to Ser 186 and Glu 439 (A-N domains interaction), on the other hand, increased the ATP affinity of the E2P ground state and lowered the ATP affinity of the E2⅐P phosphoryl transition state. Unlike the D203R/R678D swap mutation the S186E/E439S swap mutation was unable to rescue the abolishment of the modulatory effect of ATP, because S186E/E439S increased the affinity of the E2P ground state for ATP even more than the corresponding point mutations.
An important distinction that must be made when interpreting these data is whether the mutational effects occur as a consequence of destabilization of the respective domain contacts, or the individual amino acid side chains play direct roles in interaction with nucleotide.
Importance of Asp 203 and Arg 678 and Their Ion Bond Interaction in E2P States-A key finding in our study is the remarkable gain of wild type-like function of the D203R/R678D swap mutant with respect to nucleotide affinity in the E2P-like states, relative to the respective point mutants, D203R and R678D. Thus, D203R displayed 12-and 48-fold reduced ATP affinity relative to the wild type in E2⅐AlF and E2⅐orthovanadate states (E2⅐P phosphoryl transition state analogs), respectively, and 30-fold reduced ATP affinity in the E2⅐BeF state (E2P ground state analog), and R678D was unable to become photolabeled with TNP-8N 3 -ATP in either of the three E2P-like states, suggesting severely reduced nucleotide affinity. In sharp contrast hereto the D203R/R678D swap mutant displayed wild type-like affinities for TNP-8N 3 -ATP and ATP in both E2⅐AlF and E2⅐orthovanadate states and only slightly reduced affinity for the two nucleotides in the E2⅐BeF state (Figs. 4 -6 and Table 3). The results of the nucleotide binding measurements with D203R, R678D, and D203R/R678D rationalize the finding that the two point mutations abolished the stimulatory effect of ATP on E2P dephosphorylation, whereas the swap mutant retained significant ATP modulation with reasonable apparent ATP affinity for activation ( Fig. 3 and Table 2). Hence, the nucleotide binding measurements indicate that the effects of the point mutations on ATP modulation are caused by reduced binding affinity for ATP, and not merely by interference with the transmission to the catalytic site of the conformational changes elicited by ATP binding. In the crystal structure of the E2⅐AlF⅐AMP-PCP state (Fig. 1, upper left panel) the side chain of Arg 678 seems poised for ion bond interaction with the ␥-phosphate of the nucleotide, the guanidinium group of Arg 678 pointing toward the ␥-phosphate at a distance of 4.2 Å, and Asp 203 contributes to position Arg 678 . Hence, at least part of the effects of the Arg 678 and Asp 203 point mutations on the affinity for ATP could arise from disruption of such binding interaction. However, because of the gain of function seen with the swap mutant, the Asp 203 -Arg 678 ionic interaction appears more important for the ATP affinity than the exact positioning of the arginine side chain. The Asp 203 -Arg 678 interaction might contribute to stabilization of the ATP binding pocket as a whole, including tightening of other ATP-protein interactions than that between the ␥-phosphate and Arg 678 , which is supported also by the finding that TNP-8N 3 -ATP binding is profoundly affected in the Asp 203 and Arg 678 point mutants, but not in the swap mutant, although TNP-nucleotides are positioned in the binding pocket rather differently from nucleotides without the TNP moiety, as seen in various crystal structures (43) (see Fig. 1, and further discussion below).
Role of Gln 202 in E2P Dephosphorylation-The Q202A mutant displayed the lowest basal rate of E2P dephosphorylation of all the mutants studied in the present work. The stimulation by ATP of the dephosphorylation reaction, however, proceeded with wild type-like K 0.5(ATP) and rate enhancement factor (k 1 /k 0 ) for Q202A (Table 2). Accordingly, Q202A also behaved in a wild type-like manner with respect to the affinities for TNP-8N 3 -ATP and ATP in E2⅐BeF and E2⅐AlF states ( Table  3), suggesting that the very slow dephosphorylation rate of Q202A is unrelated to the actions of modulatory ATP. In the various published crystal structures of E2P-like states (e.g. Refs. 6, 8, and 9), the Gln 202 side chain amide is within hydrogen bonding distance (ϳ3 Å) of the backbone carbonyl of Gly 626 , which is a critical residue at the catalytic site (15,44). Because of the A-domain rearrangements during the Ca 2ϩ -transport cycle, Gln 202 and Gly 626 are much further apart in E2 (10 Å) than in the E2P-like states and more than 30 Å apart in Ca 2 E1, excluding the possibility of Gln 202 -Gly 626 interaction in the latter states. The close contact of Gln 202 with the Gly 626 backbone carbonyl in the E2P-like states appears to explain the marked effect by the Q202A mutation on the E2P dephosphorylation rate.
Importance of Ser 186 and Glu 439 and Their Hydrogen Bond Interaction in E2P States-With the exception of S186E, the point mutations to Ser 186 and Glu 439 substantially increased the basal E2P dephosphorylation rate determined in the absence of ATP, and the swap mutation S186E/E439S almost restored the rate to a wild type-like level. These findings are in accordance with those of Liu et al. (30), who from analysis of the E2⅐BeF crystal structures (9,45) pointed to a role for the hydrogen bond between Ser 186 and Glu 439 in stabilization of the interaction between domains A, N, and P in the E2P ground state. The pK a of the introduced glutamate in S186E was estimated to be extremely high, consistent with protonation and hydrogen bond formation between this glutamate and Glu 439 , thereby explaining the minimal effect of the mutation S186E (30). Structural modeling suggested that in the swap mutant, serine and glutamate may still be able to interact by hydrogen bonding in the E2⅐BeF state, thus explaining the gain of function seen for S186E/E439S with respect to attaining a wild type-like basal E2P dephosphorylation rate. The present study, furthermore, encompasses mutant S186P, which was found to display the highest basal E2P dephosphorylation rate of the Ser 186 mutants, likely a consequence of the influence of the unique properties of proline on the backbone conformation of the 181 TGES loop.
Also in accordance with Liu et al. (30) we found that S186A, S186E, and E439S, as well as the S186E/E439S swap mutation, abolished the stimulatory effect of ATP on E2P dephosphorylation ( Fig. 3 and Table 2). The results of our direct measurements of nucleotide affinity (Fig. 5 and Table 3) now provide an explanation of this loss of ATP modulation. The mutants with alterations to Ser 186 or Glu 439 displayed a markedly reduced ratio between the K D(ATP) values of E2⅐BeF and E2⅐AlF, relative to the ratio of 3 seen for the wild type (Table 3). In several of these mutants, the ATP affinity was actually higher in E2⅐BeF than in E2⅐AlF. Such a destabilization of the ATP-bound form of the E2⅐P phosphoryl transition state relative to that of the E2P ground state is indeed expected to lead to the observed abolition of the enhancing effect of ATP on the dephosphorylation rate. The swap mutation S186E/E439S was not different from the individual point mutants with respect to the lack of ATP modulation (Fig. 3), which was mirrored in its reduced ratio between the K D(ATP) values of E2⅐BeF and E2⅐AlF (Table  3). This should, however, not mislead one to conclude that there was no gain of function for the swap mutant relative to the point mutants in relation to the binding of modulatory ATP in the E2⅐P phosphoryl transition state. Although the individual Ser 186 and Glu 439 point mutants displayed reduced affinity of E2⅐AlF for ATP (higher K D(ATP) value), relative to the wild type, the swap mutant showed a slightly enhanced affinity of E2⅐AlF for ATP relative to the wild type and indeed a 4-fold higher affinity relative to the S186E and E439S point mutants, thus clearly indicating that the hydrogen bond between Ser 186 and Glu 439 stabilizes the E2⅐P phosphoryl transition state in the form with bound ATP. For the E2P ground state analog, E2⅐BeF, the situation is quite different. Here we found an enhanced ATP affinity for the individual Ser 186 and Glu 439 point mutants (lower K D(ATP) value), relative to that of the wild type, and the simultaneous presence of S186E and E439S in the swap mutant did not lead to any compensation; on the contrary, the ATP affinity was even more enhanced in the swap mutant, thus leading to the reduced ratio between the K D(ATP) values of E2⅐BeF and E2⅐AlF, relative to that of the wild type, even though the K D(ATP) actually was reduced also in E2⅐AlF of the swap mutant (Table 3). Hence, the S186E and E439S mutations work independently in an additive way to generate a more stable nucleotide-bond E2⅐BeF state. These findings indicate that the ATP binding mode differs considerably between E2⅐AlF and E2⅐BeF, which is in line with structural data for TNP-AMP bound forms (43). Hence, in E2⅐BeF crystallized with TNP-AMP the TNP moiety of the nucleotide is slotted into a groove formed between the A-and N-domains, with the Ser 186 and Glu 439 side chains within interaction distance (3-4 Å) of TNP, whereas in the corresponding structure of E2⅐AlF, TNP-AMP is displaced relative to its position in the E2⅐BeF state, with the TNP moiety now 5-7 Å distant from the Ser 186 and Glu 439 side chains (Fig.  1, right panels). No structures of the E2⅐BeF state have so far been solved with bound ATP, AMP-PCP, or any other nucleotide devoid of the TNP moiety, but a comparison of the TNP-AMP bound structures with the structure of E2⅐AlF with bound AMP-PCP (Fig. 1, upper left panel) shows that the TNP moiety of TNP-AMP is found roughly at the location where the adenine of AMP-PCP binds, near Phe 487 . It is, thus, conceivable that the marked increase of ATP affinity in the E2P ground state of the Ser 186 and Glu 439 mutants is related to altered interaction of Ser 186 and Glu 439 directly with the adenine part of ATP. In the wild type E2P ground state, the binding of ATP might cause destabilization by disrupting the hydrogen bond between Ser 186 and Glu 439 . Such a destabilization will not occur if the Ser 186 -Glu 439 bond has already been disrupted by mutation. Furthermore, both single and swap mutations might lead to favorable interaction of the substituents with the adenine ring, thereby enhancing the ATP affinity.
Conclusion-The Asp 203 -Arg 678 (A-P domains) and the Ser 186 -Glu 439 (A-N domains) interactions are both critical for the binding of the modulatory ATP that enhances dephosphorylation of Ca 2ϩ -ATPase. The Asp 203 -Arg 678 ionic bond stabilizes the nucleotide binding pocket in the E2⅐P phosphoryl transition state as well as in the E2P ground state, and the stabilizing effect on the E2⅐P phosphoryl transition state contributes to the accelerating effect of ATP on the dephosphorylation. The Ser 186 -Glu 439 hydrogen bond likewise stabilizes the nucleotide binding pocket in the E2⅐P phosphoryl transition state, whereas in the E2P ground state, ATP binding and the Ser 186 -Glu 439 hydrogen bond are mutually exclusive. Both of these effects of the Ser 186 -Glu 439 bond contribute to the mechanism underlying the accelerating effect of ATP on the dephosphorylation.