Multiple and Distinct Effects of Mutations of Tyr122, Glu123, Arg324, and Arg334 Involved in Interactions between the Top Part of Second and Fourth Transmembrane Helices in Sarcoplasmic Reticulum Ca2+-ATPase

We explored, by mutational substitutions and kinetic analysis, possible roles of the four residues involved in the hydrogen-bonding or ionic interactions found in the Ca2+-bound structure of sarcoplasmic reticulum Ca2+-ATPase, Tyr122-Arg324, and Glu123-Arg334 at the top part of second transmembrane helix (M2) connected to the A domain and fourth transmembrane helix (M4) in the P domain. The observed substitution effects indicated that Glu123, Arg334, and Tyr122 contributed to the rapid transition between the Ca2+-unbound and bound states of the unphosphorylated enzyme. Results further showed the more profound inhibitory effects of the substitutions in the M4/P domain (Arg324 and Arg334) upon the isomeric transition of phosphorylated intermediate (EP) (loss of ADP sensitivity) and those in M2/A domain (Tyr122 and Glu123) upon the subsequent processing and hydrolysis of EP. The observed distinct effects suggest that the interactions seen in the Ca2+-bound structure are not functionally important but indicate that Arg334 with its positive charge and Tyr122 with its aromatic ring are critically important for the above distinct steps. On the basis of the available structural information, the results strongly suggest that Arg334 moves downward and forms new interactions with M2 (likely Asn111); it thus contributes to the inclination of the M4/P domain toward the M2/A domain, which is crucial for the appropriate gathering between the P domain and the largely rotated A domain to cause the loss of ADP sensitivity. On the other hand, Tyr122 most likely functions in the subsequent Ca2+-releasing step to produce hydrophobic interactions at the A-P domain interface formed upon their gathering and thus to produce the Ca2+-released form of EP. During the Ca2+-transport cycle, the four residues seem to change interaction partners and thus contribute to the coordinated movements of the cytoplasmic and transmembrane domains.

Sarcoplasmic reticulum Ca 2ϩ -ATPase (SERCA1a) 1 is a representative member of P-type ion-transporting ATPases; it catalyzes Ca 2ϩ transport coupled with ATP hydrolysis ( Fig. 1; Refs. 1 and 2, and for recent reviews, see Refs. 3 and 4). In the catalytic cycle, the enzyme is activated by the binding of two Ca 2ϩ ions (E2 to E1Ca 2 , steps 1 and 2) and then autophosphorylated at Asp 351 by MgATP to form ADP-sensitive phosphoenzyme (E1P, step 3). Upon formation of this EP, the bound Ca 2ϩ ions are occluded in the transport sites. The subsequent isomeric transition to the ADP-insensitive form (E2P) will result in a reduction in affinity and a change in orientation of the Ca 2ϩ -binding sites and, thus, the Ca 2ϩ release into lumen (steps 4 -5). Finally, hydrolysis takes place and returns the enzyme into an unphosphorylated and Ca 2ϩ -unbound form (E2, step 6). E2P can also be formed from P i in the presence of Mg 2ϩ and the absence of Ca 2ϩ by reversal of its hydrolysis.
The enzyme has three cytoplasmic domains (N, P, and A) which are widely separated in the Ca 2ϩ -bound form (E1Ca 2 ) but are associated in the Ca 2ϩ -unbound and thapsigarginbound form E2(TG) (Refs. 5 and 6; Fig. 2). In E2(TG), the A domain has largely rotated, and the P domain has significantly inclined together with the transmembrane helices M4 and M5 toward A domain to associate with A domain. We showed previously in the proteolysis experiments (7,8) that the large rotation of A domain and its gathering with P and N domains most likely occur during the E1P to E2P transition and Ca 2ϩ release (steps 4 -5) to form the most compactly organized single headpiece in the Ca 2ϩ -released form of E2P; we further suggested that the stabilization energy provided by the intimate contacts between three cytoplasmic domains in E2P will provide energy for moving transmembrane helices and release the bound Ca 2ϩ into lumen. To gain further insight into the energy coupling between cytoplasmic and transmembrane domains, it is crucial to find out the structural elements essential for the changes in the cytoplasmic domain organization and the coordinated movements of transmembrane helices and reveal the actual roles of each of the elements in the Ca 2ϩ -transport cycle. In this regard, we have recently identified the Glu 40 -Ser 48 loop as a critical element for the rotation of A domain and coordinated unique motions of M1 during the E1P to E2P transition (9); we also identified the Lys 189 -Lys 205 outermost loop of A domain as making intimate contact with P domain for the subsequent processing and hydrolysis of E2P (10).
The likely importance of the hydrogen bond between Tyr 122 at the top part of M2 (connected to the A domain) and Arg 324 at the top part of M4 (in the P domain) was identified as stabilizing the structure in E1Ca 2 (Ref. 6; Fig. 2). The Glu 123 -Arg 334 ionic bond is also very likely in this region in E1Ca 2 . These interactions are likely lost during the catalytic cycle, because they are not present in E2(TG); therefore, it is possible that the above four residues at the M2/A-M4/P domain interface may function in a certain specific catalytic step(s) in which the large motions of the domains take place.
In the present study, therefore, we explored the possible roles of the four residues (Tyr 122 , Glu 123 , Arg 324 , and Arg 334 ) by mutational substitutions and kinetic analysis. Results indicate that Arg 334 , Glu 123 , and Tyr 122 contribute to the rapid E2 to E1Ca 2 transition. We further found the more profound inhibitory effects of the substitutions in M4/P domain (especially Arg 334 ) on the E1P to E2P transition (step 4) and those in M2/A domain (especially Tyr 122 ) on the subsequent processing and hydrolysis of E2P (steps 5-6). The results provide novel information on how the changes in organization between A and P domains proceed by the contribution of the newly identified essential structural elements during the E1P to E2P transition and the E2P processing to release Ca 2ϩ , and refine the picture of global conformational changes of the enzyme in these processes. During the Ca 2ϩ -transport cycle, the residues on the M2/A domain and on M4/P domain seem to move relative to each other, change the interaction partners, and hence contribute to the proper domain motions and the cross-talk between the cytoplasmic and transmembrane domains.

EXPERIMENTAL PROCEDURES
Mutagenesis and Expression-The Stratagene QuikChange TM sitedirected mutagenesis method (Stratagene, La Jolla, CA) was utilized for the substitution of residues in the rabbit SERCA1a cDNA. The ApaI-KpnI or KpnI-SalI restriction fragments with the desired mutation were excised from the plasmid and ligated back into the corresponding region in the full-length SERCA1a cDNA in the pMT2 expression vector (11). The pMT2 DNA was transfected into COS-1 cells by the liposome-mediated transfection method. Microsomes were prepared from the cells as described previously (12). The "control microsomes" were prepared from COS-1 cells transfected with the pMT2 vector containing no SERCA1a cDNA. The amount of expressed SERCA1a was quantified by a sandwich enzyme-linked immunosorbent assay as described previously (13). Expression level of wild-type SERCA1a was 2-3% of total microsomal proteins, and all the mutants were expressed at levels comparable with that of the wild type.
Ca 2ϩ -ATPase Activity-The rate of ATP hydrolysis was determined at 25°C in a mixture containing 20 g/ml microsomal protein, 5 mM [␥-32 P]ATP, 1 M A23187, 0.1 M KCl, 7 mM MgCl 2 , 0.5 mM CaCl 2 , 0.4 mM EGTA, and 50 mM MOPS/Tris (pH 7.0). The Ca 2ϩ -ATPase activity was obtained by subtracting the Ca 2ϩ -independent ATPase activity, which was determined in the presence of 5 mM EGTA without added CaCl 2 , otherwise as above. The specific ATPase activity/mg of expressed SERCA1a protein was calculated from the amount of expressed SERCA1a and the Ca 2ϩ -ATPase activity of expressed SERCA1a, which was obtained by subtracting the Ca 2ϩ -ATPase activity of the control microsomes from that of the microsomes expressing SERCA1a. This background level with the control microsomes was as low as 3% of the activity of microsomes expressing the wild-type SECRA1a.
Formation and Hydrolysis of EP-Phosphorylation of SERCA1a in microsomes with [␥-32 P]ATP or 32 P i and dephosphorylation of 32 P-labeled SERCA1a were performed under conditions described in the legends to figures. The reactions were quenched with ice-cold trichloroacetic acid containing P i . Rapid kinetics measurements of phosphorylation and dephosphorylation were performed with a handmade rapid mixing apparatus (14), otherwise as above. The precipitated proteins were separated by 5% SDS-PAGE at pH 6.0 according to Weber and Osborn (15). The radioactivity associated with the separated Ca 2ϩ -ATPase was quantitated by digital autoradiography as described (16). The amount of EP formed with the expressed SERCA1a was obtained by subtracting the background radioactivity with the control microsomes. This background was less than 5% of the radioactivity of EP formed with the expressed wild-type SERCA1a. The amount of EP/mg of SERCA1a protein was calculated from the amount of EP thus obtained and the amount of expressed SERCA1a.
Miscellaneous-Protein concentrations were determined by the method of Lowry et al. (17), with bovine serum albumin as a standard. Data were analyzed by nonlinear regression using the program Origin (Microcal Software, Inc., Northampton, MA). Three-dimensional models of the enzyme were reproduced by the program VMD (18).

Effects of Substitutions on Ca 2ϩ -ATPase Activity-
The specific Ca 2ϩ -ATPase activities of the expressed mutant and wildtype SERCA1a were determined at 25°C. All the substitutions resulted in the reduction of the activity (4.30 Ϯ 0.27 mol/ min/mg SERCA1a protein (n ϭ 4) observed with the wild type). The substitution of Tyr 122 with phenylalanine moderately reduced the activity to 51% of the activity of the wild type, and the substitution with alanine almost completely diminished it (to 1%). The alanine substitutions of Glu 123 and Arg 324 slightly or moderately reduced the activity to 75 and 86%, respectively, and the introduction of the opposite charge to Glu 123 and Arg 324 caused a much stronger inhibition to 30 and 41%, respectively. The substitutions of Arg 334 (especially with glutamate) resulted in strong or almost complete inhibition (to 6% of the activity of wild type). The result with R334A (reduction to 45%) is consistent with the previously observed inhibition of the activity with this mutant (actually the reduction in the previous study (to 22%) was even larger; Ref. 19).
Ca 2ϩ Concentration Dependence of EP Formation from ATP and the E2 to E1Ca 2 Transition-In Fig. 3, the amount of EP formed from ATP was determined at steady state (at 15 s) after the addition of ATP to the enzyme preincubated with various concentrations of Ca 2ϩ . All the mutants formed EP, and the amount formed with a saturating Ca 2ϩ was comparable with that of wild type. Furthermore, the dissociation constants for Ca 2ϩ and Hill coefficient obtained by least squares fit with these mutants (0.08 -0.19 and 1.7-2.2 M, respectively) were nearly the same as those obtained with the wild type (0.13 and 1.7 M, respectively). The results indicate that the Tyr 122 -Arg 324 and Glu 123 -Arg 334 interactions found in E1Ca 2 and these four residues do not contribute significantly to the steady-state equilibrium between E2 and E1Ca 2 . In the kinetics of the EP formation at the saturating Ca 2ϩ (100 M), we further found that the first-order rate in all the mutants (2.0 -2.9 s Ϫ1 ) is almost the same as that of the wild type (2.5 s Ϫ1 ) (E1Ca 2 to E1PCa 2 in Table I). Thus, the substitutions have essentially no effect upon EP formation from E1Ca 2 in step 3.
We then examined the rate of the E2 to E1Ca 2 transition in steps 1 and 2. For this experiment, the mutants and wild type were preincubated in the absence of Ca 2ϩ at pH 6, where the equilibrium between E1 and E2 is most shifted to E2 (20), and then phosphorylated by the simultaneous addition of saturat-ing Ca 2ϩ and ATP (Fig. 4). The time course of EP formation was well described by the first-order kinetics (although the maximum level of EP formed with R334A was somewhat lower than that at pH 7). The rates obtained are summarized in Table I. When ATP was added to the enzyme preincubated with Ca 2ϩ otherwise as above, the EP formation was much faster and almost completed within 1 s for all the mutants and wild type; therefore, the rates obtained above actually reflect the ratelimiting E2 to E1Ca 2 transition.
The substitutions of either of the residues of the Glu 123 -Arg 334 interaction pair reduced the rate of the E2 to E1Ca 2 transition with the larger magnitudes in the Arg 334 mutants (Fig. 4A). In each of Glu 123 and Arg 334 , the introduction of an opposite charge had a stronger effect than the alanine substitution. The results indicate that the side chains of these residues, especially of Arg 334 , are important for the rapid E2 to E1Ca 2 transition. The substitutions of Tyr 122 in the other Tyr 122 -Arg 324 interaction pair also reduced the rate of the E2 to E1Ca 2 transition, with the larger magnitude in the alanine substitution than in the phenylalanine substitution (Fig. 4B). In contrast, both the alanine-and glutamate-substitutions of Arg 324 rather increased the rate by ϳ2-fold (and thus Arg 324 , with its native structure, somewhat slows this process). The results indicate that the Tyr 122 -Arg 324 hydrogen bond found in E1Ca 2 is likely not important for the rapid E2 to E1Ca 2 transition, but Tyr 122 to some extent contributes to the rapid transition. It should also be noted that the substitutions of the four residues caused the changes in the rate of the E2 to E1Ca 2 transition without significant change in the Ca 2ϩ affinity (see Fig. 3). It is likely that the substitutions affected the activation energy for this process (and/or the energy levels of both E2 and E1Ca 2 to the same extent).
The E1P to E2P Transition and the Subsequent Processing of E2P-We then analyzed the steps after formation of EP, i.e. the isomeric transition and decay of EP through steps 4 -6, and found that the substitutions in M4/P domain (Arg 324 and Arg 334 ) and those in M2/A domain (Tyr 122 and Glu 123 ) both have serious effects, but on the distinct steps. We first determined the fraction of ADP-insensitive EP (E2P) accumulated at 15 s (nearly the steady state) after the addition of ATP to the enzyme preincubated with Ca 2ϩ (as described in the legend to Fig. 5). In the presence of K ϩ , which accelerates the decay of E2P and thus causes its low accumulation in the wild type (18% of the total amount of EP under the present conditions; Ref. 21), the fraction of E2P accumulated in the Arg 324 and Arg 334 mutants was even lower (being 8% (R324A), 3% (R324A), 4% (R334A and R334E) of the total amount of EP), but it was considerably higher in the Tyr 122 and Glu 123 mutants (except for E123A, for which it was 29% (Y122F), 49% (Y122A), 15% (E123A), and 57% (E123K). In the absence of K ϩ , the fraction of accumulated E2P was again higher in the Tyr 122 and Glu 123 mutants and was also at the higher level in the wild type (58% (Y122F), 50% (Y122A), 31% (E123A), 77% (E123K), and 43% (wild type)). In contrast, the fraction of accumulated E2P in the Arg 324 mutants was still low (25% (R324A) and 14% (R324A)), or very low in the Arg 334 mutants (5% in R334A and R334E). The results indicate that the E1P to E2P transition in step 4 is inhibited in the Arg 324 and Arg 334 mutants (especially the Arg 334 mutants) but not in the Tyr 122 and Glu 123 mutants.
We then determined the time course of E2P formation from E1Ca 2 with ATP in the absence (Fig. 5A) and presence (Fig. 5B) of K ϩ with the Tyr 122 and Glu 123 mutants, which were shown above to accumulate the large amount of E2P. Under both sets of conditions, the total amount of EP reached its maximum level very rapidly (within ϳ1 s) and remained unchanged during the experiments (60 s); thus, the time course actually reflects the E2P accumulation from E1P in step 4. In the absence of K ϩ (Fig. 5A), a large amount of E2P accumulated in the wild type as well as in the mutants, and thus the time courses can be clearly compared. The E2P accumulation apparently proceeded with first-order kinetics (the rates obtained are summarized in Table I). The rate in the mutants was similar (Y122A and Y122F) or even faster (E123A and E123K, by ϳ3-fold), as compared with that of the wild type. In each of the mutants, which accumulate a fair amount of E2P also in the presence of K ϩ (Y122F, Y122A, E123K), the apparent rate obtained in the presence of K ϩ was similar to that obtained in its absence ( Fig.  5B and Table I). The substitutions of Tyr 122 and Glu 123 thus did not inhibit the rapid E1P to E2P transition. The results suggest that the substitutions caused the E2P accumulation because of possible inhibition of the subsequent E2P decay.
For the Arg 324 and Arg 334 mutants, which do not accumulate E2P but accumulate mostly E1P, the decay of EP formed from ATP was determined in the presence of K ϩ instead of the E2P formation (Fig. 6A). This EP decay most likely reflects the rate-limiting E1P to E2P transition (as also known for the wild type; Ref. 22). In Fig. 6, the EP decay was determined by first phosphorylating with [␥-32 P]ATP in the presence of K ϩ and Ca 2ϩ for 15 s and then terminating phosphorylation by adding excess EGTA to prevent further phosphorylation and thus allow for decay of 32 P-labeled EP. The EP decay was markedly slowed by the substitutions, especially of Arg 334 . The time courses were fitted well with a single exponential, and the rates obtained are summarized in Table I. The rate was reduced in R324A to ϳ40% of that in the wild type, further reduced to ϳ20% in R324E and R334A, and almost completely inhibited in R334E (6%). In each of the time points in the EP decay, almost all of the total amount of EP was E1P (data not shown). The results indicate that these two residues on the M4/P domain are important for the rapid E1P to E2P transition in step 4, and that Arg 334 is especially crucial.
The same sets of experiments were done with the Tyr 122 and Glu 123 mutants (Fig. 6B), but in this case, to examine the possible inhibition of processing of E2P after the E1P to E2P transition. The time courses were apparently fitted very well with a single exponential, although a substantial fraction of the total amount of EP was ADP-insensitive in these mutants at the start of the decay reaction (see Fig. 5). The EP decay was almost completely blocked in Y122A and substantially slowed in E123A and E123K (Table I). In Y122F, the rate was similar to that in the wild type.
Hydrolysis of E2P Formed from P i -To examine hydrolysis of E2P without Ca 2ϩ in the presence of K ϩ (step 6), the enzyme was first phosphorylated with 32 P i in the absence of Ca 2ϩ and K ϩ and in the presence of 35% (v/v) Me 2 SO, which favors E2P formation (23); then, the phosphorylated samples were diluted at 0°C with a large volume of solution containing K ϩ and non-radioactive P i without Ca 2ϩ (Fig. 7, A and B). Thus, the conditions were made in all other respects identical to those used for the E2P formation from ATP (Fig. 5) and the decay of EP formed from ATP in the presence of Ca 2ϩ (Fig. 6). Hydrolysis of 32 P-labeled E2P proceeded with first-order kinetics, and the rates obtained were summarized in Table I. In the Arg 324 and Arg 334 mutants, the rate of E2P hydrolysis was similar to or even faster (by ϳ3-fold) than that of wild type. On the other hand, in the Tyr 122 and Glu 123 mutants, the E2P hydrolysis was markedly slowed (except E123A, which showed the enhanced rate) and almost completely inhibited in Y122A.
Inhibition of Processing of E2P Formed from ATP in Tyr 122 and Glu 123 Mutants-In the above experiments performed at 0°C with the mutant Y122A (Figs. 6B and 7B), the decay of EP formed from ATP with Ca 2ϩ and the E2P hydrolysis without Ca 2ϩ were both not detected in the experimental time scale and thus could not be compared with each other. Therefore, we performed the same set of experiments at a higher temperature (25°C; Fig. 7C) and actually could observe both the EP decay and hydrolysis, which were, again, very strongly inhibited and extremely slow, as compared with those in the wild type (which were complete within 1 s). The results with Y122A revealed that the decay of EP formed from ATP in the presence of Ca 2ϩ was definitely slower than the hydrolysis of E2P without Ca 2ϩ . show the least squares fit to a single exponential, and the apparent rates to reach the steady-state E2P level thus obtained are given in Table I. Wild type (q) and mutants Y122F (E), Y122A (‚), E123A (Ⅺ), and E123K (ƒ) are indicated.

TABLE I Rate constants for partial reaction steps
The rate constants for the partial reaction steps were obtained in the experiments shown in Fig. 4 (formation of EP from E2, i.e. the rate-limiting E2 to E1Ca 2 transition in steps 1-2), Fig. 5 (formation of E2P from ATP, i.e. E1PCa 2 to E2PCa 2 in step 4), Fig. 6 (decay of EP formed from ATP), and Fig. 7 (hydrolysis of E2P formed from P i in step 6). The rate constants for the E1PCa 2 formation from ATP (E1Ca 2 to E1PCa 2 in step 3) were also determined (see text). In parenthesis, the values obtained with the wild type are normalized to 100%. We also observed that the formation of the ADP-insensitive EP from ATP in Y122A was completed within 1 s after the addition of ATP at 25°C (data not shown). The results indicate that the step after the loss of ADP-sensitivity (in step 4) but before the E2P hydrolysis without Ca 2ϩ (in step 6) is also strongly inhibited in Y122A. It is noteworthy that the effects of the alanine-substitution of Tyr 122 are very similar to those of substitutions of Val 200 on the Lys 189 -Lys 205 loop on the A domain (10), in which both the rate of processing of E2P after the loss of ADP-sensitivity and the rate of E2P hydrolysis are dramatically reduced (but not the rate of loss of ADP-sensitivity). The kinetic results with the Glu 123 mutants showed that the decay of EP formed from ATP is slower than the E2P hydrolysis (step 6) and also substantially slower than the accumulation of E2P from E1P (step 4) in these mutants (Figs. 5-7 and Table I). Especially with the mutant E123K, it was clearly shown in the presence of K ϩ that the large amount of E2P accumulation occurred with the much faster rate than that of decay of EP. The results are consistent with the view that the step after the loss of ADP-sensitivity but before the E2P hydrolysis is inhibited in these mutants. DISCUSSION In the present study, we explored the possible roles of the residues involved in the ionic and hydrogen-bonding interactions present in the crystal structure E1Ca 2 between the top part of M4 in P domain (Arg 324 and Arg 334 on M4/P domain) and that of M2 (or the loop) connected to A domain (Tyr 122 and Glu 123 on M2/A domain; Fig. 2). We found that Arg 334 (especially) and Glu 123 are important for the rapid E2 to E1Ca 2 transition, and that Tyr 122 also contributes to some extent. Furthermore, we found the more profound inhibitory effects of the substitutions in each of the two regions on the successive but distinct catalytic steps after the formation of E1P, i.e. the substitutions in M4/P domain (particularly of Arg 334 ) seriously affect the loss of ADP-sensitivity, and those in the M2/A domain (particularly of Tyr 122 ) seriously affect the subsequent processing and hydrolysis of E2P. The results likely reflect the distinct structural roles of the residues in the specific steps as discussed in the following, but the interactions seen in the E1Ca 2 structure are not functionally important, as may be deduced from the pattern of functional changes induced by mutations.
Arg 334 and Arg 324 on M4/P Domain for the E1P to E2P Transition-Because the substitutions of Arg 324 and Arg 334 inhibited strongly the E1P to E2P transition and, in sharp contrast, those of Tyr 122 and Glu 123 did not affect them seriously (or rather, they increased the E2P accumulation; Figs. 5-6, Table I, and "Results"), the Tyr 122 -Arg 324 and Glu 123 -Arg 334 interactions present in E1Ca 2 are likely lost during or before the E1P to E2P transition. In fact, in E2V, which is analogous to E2P (5, 7), the residues in each of the interaction pairs are separated by the relative downward movement of M4/P domain by ϳ10 residues (Fig. 8A).
With respect to the structural role of Arg 334 , the strong or almost complete inhibition of the loss of ADP sensitivity by its alanine-and glutamate-substitutions indicates that Arg 334 , with its positive charge, is crucial for the E1P to E2P transition (step 4), and further, it suggests that such roles may involve interactions with a certain polar or negatively charged residue(s). In E2V, Arg 334 in the P domain actually protrudes onto M2 and is in close contact with Asn 111 (Fig. 8A), of which alanine-substitution was previously shown to suppress almost completely the activity without inhibiting phosphorylation with ATP (24). (In E2(TG), Arg 334 is very close to Asn 114 , of which substitution was also shown to reduce strongly the activity (24).) Therefore, Arg 334 likely moves on M2 from Glu 123 to Asn 111 . For such movement of Arg 334 , it is of interest that there are several negatively charged (Glu 123 , Glu 121 , Glu 117 , and Glu 113 ) and polar (Asn 114 and Asn 111 ) residues situated on M2 between Glu 123 and Asn 111 (Fig. 8A). It is tempting to speculate that those residues on M2 (and possibly the neighboring Glu 40 , Glu 44 , and Glu 45 ) together effectively guide Arg 334 to the final interaction site during the large rotation of the A domain in the E1P to E2P transition (hence the limited substitutions of each of the single glutamates on M2 may possibly be compensated for by other glutamates, which, in fact, we found (24)). In any case, the downward movement of Arg 334 and its interaction with the new partner (Asn 111 ) would cause the inclination and large shift of M4/P domain toward the M2/A domain. In fact, P domain and M4/M5, which are directly linked with P domain, are significantly inclined (or bent) forward in E2V (as well as in E2(TG)). The possible interactions of Arg 334 with the negatively charged residues may also be important during the E2 to E1Ca 2 transition (moving back from Asn 114 (E2(TG)) to Glu 123 (E1Ca 2 )), because the rate of this transition was also markedly reduced by the Arg 334 substitutions (Fig. 4 and Table I).
With respect to Arg 324 on M4, the observed inhibition of the  Fig. 3. Phosphorylation was terminated by the addition of an equal volume of a buffer containing 10 mM EGTA, 0.1 M KCl, 7 mM MgCl 2 , and 50 mM MOPS/Tris (pH 7.0) at 0°C; the decay reaction was quenched with trichloroacetic acid at different times, as indicated after the addition of EGTA. The amounts of EP obtained at zero time (i.e. immediately before the addition of EGTA) are normalized to 100%. Solid lines show the least squares fit to a single exponential, and the decay rate constants thus obtained are given in Table I. In A, wild type (q) and mutants R324A (E), R324E (‚), R334A (Ⅺ), and R334E (ƒ) are indicated. In B, wild type (q) and mutants Y122F (E), Y122A (‚), E123A (Ⅺ), and E123K (ƒ) are indicated. E1P to E2P transition by its substitutions is consistent with the view that the large motions of P domain and the top part of M4 occur during the loss of ADP sensitivity in step 4, although how this residue contributes to the process cannot be stated clearly on the basis of available structures (as its side chain is already exposed to solvent in E2V). Interestingly, the E2 to E1Ca 2 transition was rather accelerated in the Arg 324 mutants. Thus, the structural change for the E2 to E1Ca 2 transition and that for the E1P to E2P transition are not simply reverse to each other.
Tyr 122 and Glu 123 on M2/A Domain for the Processing and Hydrolysis of E2P-The results on Tyr 122 indicate that the hydrophobic aromatic ring of Tyr 122 is critically important for decay and hydrolysis of E2P after the loss of ADP-sensitivity in step 4 (Figs. 5-7). Such a structural role may involve some hydrophobic interactions. In E2V, Tyr 122 protrudes into and is situated at the center of the hydrophobic cluster composed of the residues on the P domain (Val 705 and Val 726 ), the A domain (Ile 179 and Ile 232 ), and the M2/A domain (Leu 119 and Tyr 122 ) (Fig. 8A). The gathering of these residues is obviously realized as a consequence of the large rotation of the A domain and the inclination of the P domain. Importantly, such motions and docking of the A and P domains most likely occur during the E1P to E2P transition in step 4 (see the above discussion on Arg 334 and Refs. 7-9), and the serious effect of the Tyr 122 substitution appeared after this transition. Therefore, the interactions involving the key residue Tyr 122 in the hydrophobic cluster most likely occurred after the E1P to E2P transition in step 4 and are essential for the subsequent processing and hydrolysis of E2P. For the rapid E2P hydrolysis (step 6), the 4Ј-OH group of Tyr 122 also likely contributes to some extent (by the possible hydrogen bond within the A domain (Lys 158 )), because the rate was appreciably reduced by the removal of the OH group in the mutant Y122F.
With the Glu 123 mutants, the kinetic results suggest that the native structure at Glu 123 juxtaposed to Tyr 122 on M2/A domain is also important for the processing of E2P before its hydrolysis. It is also of interest that the strong block of the hydrolysis of E2P (without Ca 2ϩ ) was found in E123K, as opposed to the enhanced rate in E123A. Because Glu 123 is more exposed to solvent in E2V than in E2(TG) but closer in E2(TG) than in E2V to the Glu 40 backbone on the A domain, the conformational changes during the transition from E2P to E2 may possibly involve desolvation at this residue and formation of the interaction within A domain, which is not favored with the introduced positive charge in E123K.
Integrated Picture of Conformational Changes in the E1P to E2P Transition and Subsequent Processing of E2P-The above findings indicate that two distinct structural groups operate in distinct steps during the E1P to E2P transition and subsequent E2P processing (to contribute altogether for the final release of Ca 2ϩ ). We have also recently found (9, 10) that the Glu 40 -Ser 48 loop is critical for the E1P to E2P transition in step 4 and that the Lys 189 -Lys 205 outermost loop (Val 200 loop in Fig. 8) on A domain is essential in the subsequent processing of E2P for the intimate contact of A and P domains. In other previous studies (25)(26)(27)(28)(29)(30)(31)(32), several residues were found to be crucial for the E1P to E2P transition. By gathering all these findings, we can speculate and gain an integrated picture of the conformational changes required for the processes to release Ca 2ϩ . Although these processes in the wild-type enzyme may occur kinetically as a single step (E1PCa 2 3 E2P ϩ 2Ca 2ϩ ), and the existence of the E2PCa 2 intermediate is a controversial issue in the literature on the wild type, the reaction scheme with E2PCa 2 (Fig. 1) can be conveniently used in a discussion about our findings upon the distinct functions of the identified residues. It should FIG. 7. Hydrolysis of E2P formed from P i . Microsomes were phosphorylated with 32 P i at 25°C for 10 min in 50 l of a mixture containing 1 g of microsomal protein, 0.1 mM 32 P i , 4 mM EGTA, 10 mM MgCl 2 , 50 mM MOPS/Tris (pH 7.0), and 35% (v/v) Me 2 SO. The mixture was then cooled and diluted at 0°C by the addition of 0.95 ml of a mixture containing 2.1 mM non-radioactive P i , 105 mM KCl, 4 mM EGTA, 7 mM MgCl 2 , and 50 mM MOPS/Tris (pH 7.0). At different times after the dilution, hydrolysis was quenched with trichloroacetic acid. In A, wild type (q) and mutants R324A (E), R324E (‚), R334A (Ⅺ), and R334E (ƒ) are indicated. In B, wild type (q) and mutants Y122F (E), Y122A (‚), E123A (Ⅺ), and E123K (ƒ) are indicated. The amounts of EP formed with 32 P i at zero time are normalized to 100%. Solid lines show the least squares fit to a single exponential, and the rate constants thus obtained are given in Table I. C, with the mutant Y122A, the hydrolysis of E2P formed from P i in the absence of Ca 2ϩ was determined at 25°C (‚); otherwise, hydrolysis was determined as above and compared with the decay of EP formed from ATP in the presence of Ca 2ϩ (OE), which was determined at 25°C, the same conditions as for the E2P hydrolysis (for details, see the legend to Fig. 6). Solid lines show the least squares fit to a single exponential, and the rate constants (s Ϫ1 ) thus obtained are 0.24 Ϯ 0.02 for the E2P hydrolysis without Ca 2ϩ (‚) and 0.12 Ϯ 0.01 for the decay of EP formed from ATP (OE). In the wild type, the E2P hydrolysis and the decay of EP formed from ATP were completed within 1 s (data not shown).
be noted that accumulation of E2PCa 2 is a hypothesis and not a fact (as it is impossible to measure Ca 2ϩ binding and dissociation in the expressed enzyme that comprises only ϳ2-3% of total proteins in the microsomes).
In the E1P to E2P transition (step 4), the A domain largely rotates and then docks with the P domain, which has inclined forward to the A domain, likely by forming the interactions involving the TGES 184 loop on the A domain (34), the residues surrounding Asp 351 on the P domain (Asp 601 -Pro 603 , Asp 627 , Asp 703 , Asn 706 , Lys 352 -Asn 359 , or at least some of them), and Arg 560 on the N domain at the three-domain interface at the central part of the molecule (Fig. 8B, the semitransparent blue cylinder on E2V, an E2P analogue). The Glu 40 -Ser 48 loop connecting the A domain and M1, located at the outer side, is most  (10) and Tyr 122 (this study) are most likely essential for the intimate A-P domain contact in the subsequent processing of E2P and are located at the opposite ends of the A-P domain interface, as shown with the dotted blue circles in E2V, top view (right). The forward shift of the whole P domain is due to its inclination toward the A domain (from E1Ca 2 to E2V, revealed by manually fitting these structures with M8-M10, which do not move in these structures) and is represented by the movement of Asp 351 with the dotted orange arrow. The likely rotation of the P domain (by ϳ25°), another component of the motion of the P domain (from E1Ca 2 to E2V), is depicted by the solid orange arrow on the circle. Dotted red arrow indicates the rotation of the A domain. likely crucial for the large rotation of the A domain and the coordinated unique motions of M1 (upward movement and bending at the membrane surface forming M1Ј) to cause the appropriate domain docking (and possibly contributing to the movement of M3 by van der Waals interactions of M1; Ref. 9). Arg 334 on the M4/P domain moves downward to Asn 111 on M2 during the E1P to E2P transition and hence contributes to the inclination of the P domain with the top part of M4/M5 and, thus, anchors the inclined P domain on the A domain, as discussed above. The Gly 233 -Pro 248 loop connecting the A domain and M3 also functions to anchor the inclined P domain by the interaction with helix P6 (33). For the appropriate domain docking and anchoring, the positioning of three regions is necessary: the TGES 184 loop for docking at the central part, and the other two (Arg 334 and the Gly 233 -Pro 248 loop) for docking at the lower and outer parts but separated at each side of the P domain for anchoring in the inclined state (Fig. 8B, dotted red  circles). The Glu 40 -Ser 48 loop connected to M1Ј at the membrane surface may also stabilize such a state.
The movement of the top part of M2 is large as the directly connected A domain rotates; therefore, the interaction of Arg 334 upon the M4/P domain with the M2/A domain is likely crucial for the large forward movement of the top part of M4/M5 during the E1P to E2P transition. The previously observed blocking of the E1P to E2P transition by mutations on M4, M5, and M8 (26,30,31) is consistent with the view that the transmembrane helices are significantly rearranged with the motions of A and P domains during this transition in step 4, although these movements are likely not enough to open the Ca 2ϩ pathway to lumen.
It should be noted that E1PCa 2 has the P and N domains, with their most closed configuration as realized in the E1Ca 2 ATP complex and the loops connecting the A domain with M2 and M3 being repositioned close to the P domain (thus, they are ready for the subsequent A domain rotation; Ref. 8).
During the E1P to E2P transition, therefore, the P and N domains should be opened to some extent, and, hence, the A domain can rotate in and associate with the P and N domains (thus to interfere with the access of ␤-phosphate of ADP to the acylphosphate; Ref. 8).
After the loss of ADP sensitivity, the final process of gathering the A and P domains is accomplished (in the subsequent step 5) by the formation of interactions at two distinct regions ( . These interactions will likely further distort the P domain (further inclination and/or possible rotation; see the following paragraph) and rearrange more the transmembrane helices to release Ca 2ϩ . The intimate A-P domain contact also likely produces an appropriate conformation and hydrophobic atmosphere around the phosphorylation site, in which a specific water molecule can attack and hydrolyze the acylphosphate bond with essential residues.
On the basis of atomic structures E1Ca 2 and E2(TG) (5, 6), the rearrangements of transmembrane helices to release Ca 2ϩ into lumen seem to involve the large forward shift and downward movement of the top parts of M4 and M5 because of their tilting at the pivoting point Gly 770 , but also likely involve twisting in the arrangement of transmembrane helices. During the E1P to E2P transition and Ca 2ϩ release, such intricate movements should be properly caused by the motions of cytoplasmic domains. In this regard, it is noteworthy that the motion of the P domain (from E1Ca 2 to E2V) most likely involves not only its inclination toward the A domain but also its rotation (though small) (Fig. 8B, dotted and solid orange arrows, respectively). These motions would generate both the tilting of M4/M5, which is directly linked with the P domain, and also the rotation at their top in the tilted state and, thus, twisting of the transmembrane helices. In any case, the exquisite positioning of the two interaction regions (Val 200 loop and Tyr 122 ) at each side of the A-P domain interface may be most suitable for appropriately causing the further forward inclination of the P domain, the tilting of M4/M5 and the possible rotation of P domain, and the twisting of transmembrane helices for Ca 2ϩ release.