Functional consequences of alterations to hydrophobic amino acids located in the M4 transmembrane sector of the Ca(2+)-ATPase of sarcoplasmic reticulum.

Those hydrophobic residues between Ile298 and Ile315 in transmembrane segment M4 of the Ca(2+)-ATPase of sarcoplasmic reticulum, not previously mutated, were mutated systematically in ways that would alter their size or polarity, and functional consequences were measured. Fourteen residues in this sequence are organized as juxtapositions of large, hydrophobic (Val, Leu, Ile) and small (Ala, Gly) residues, and these were altered so that large residues were substituted for small and vice versa. Several mutants exhibited diminished Ca2+ transport, but mutants A305V and A306V lost all Ca2+ transport function. In both cases, the mutants were phosphorylated with ATP in the presence of Ca2+ and with inorganic phosphate only in the absence of Ca2+, indicating that the Ca(2+)-binding sites were intact. Reduced Ca2+ affinity, as measured by Ca2+ dependence of phosphorylation from ATP, was observed for mutant A305V. In both mutants, the ADP-insensitive phosphoenzyme intermediate (E2P) decayed slowly relative to the wild-type enzyme, suggesting that the E2P to E2 conformational transition was impaired, slowing the rate of the phosphatase reaction. Double mutants which reversed the order of Val304 and Ala305 and Ala306 and Ile307, resulted in the same phenotype as the single Ala mutations. These results, combined with our previous demonstration that Glu309 is a Ca2+ binding residue, that Pro312 is involved in E1P to E2P conformational changes, and that Gly310 is involved in E2P to E2 conformational changes, support the hypothesis that transmembrane segment M4 plays a key role in the Ca2+ transport function of the Ca(2+)-ATPase through its involvement in both the binding of Ca2+ and the subsequent conformational changes which bring about the translocation of Ca2+ to the lumen of the membrane.

The Ca2+-ATPase of sarcoplasmic reticulum pumps Ca2+ from the sarcoplasm to the lumen at the expense of ATP hydrolysis. The molecule has been the focus of extensive structural and functional studies (MacLennan et al., 1970;Inesi, 1985;Inesi et al., 1990;Inesi and de Meis, 1985;Jorgensen and Andersen, 1988). The enzyme is a 110-kDa amphipathic protein with a cytoplasmic headpiece and stalk protruding from a globular basepiece, which forms part of the sarcoplasmic reticulum membrane structure. In the course of Ca2+ transport, Ca2+ and ATP are bound to the Ca2+-ATPase with high affinity and ATP is hydrolyzed, forming a phosphoenzyme intermediate. Ensuing conformational transitions cause vectorial displacement of calcium from its binding sites so that Ca", initially bound from the cytoplasmic surface, is released to the lumen.
The cloning of Ca2+-ATPase cDNA (MacLennan et al., 1985;Brand1 et al., 1986) and the functional expression of full-length cDNA in COS-1 cells (Maruyama and MacLennan, 1988) have made it feasible to study structure-function relationships in the protein by studies of the functional consequences of site-specific mutagenesis. Using this approach, we have identified a number of amino acids which, when mutated, disrupt Ca2+ transport function. By analyzing the partial reactions of the transport-defective mutants, we have identified amino acids involved in the ATP binding site (Maruyama and MacLennan, 1988;Maruyama et al., 1989;Clarke et al., 199Oc), in Ca2+ binding (Clarke et al., 1989a) and in conformational changes Vilsen et al., 1989;Clarke et al., 1990b;Andersen et al., 1992).
An important concept emerging from these studies is that 6 residues, G~u~~, G~u~~' , AsnIg6, Thr7", Aspaoo, and Glugm, which reside in transmembrane helices M4, M5, Ms, and Ms, are associated in formation of the Ca2+ binding sites. We have suggested that the four helices interact to form all or part of the pathway through which Ca2+ is actively transported. Conformational changes in one or more of these transmembrane segments may play a central role in causing the bound Ca2+ ions to become accessible to the membrane lumen. In support of this hypothesis, it has been found that some mutations to Glu309 (Clarke et al., 1990a) and Pro312 , predicted to be located within transmembrane segment M4, block the conformational transition between E,P' and E2P, while mutation of Gly3lo blocks the conformational transition between EzP and Ez.
In this report, we describe the functional consequences of The abbreviations used are: EIP, ADP-sensitive phosphoenzyme intermediate; E2P, ADP-insensitive phosphoenzyme intermediate; MOPS, 3-(N-morpholino)propanesulfonic acid. of the Ca2+-ATPase Transmembrane Domain mutations to nonpolar residues which are, for the most part, juxtaposed in a series of alternating large and small size in transmembrane segment M4. These residues are contiguous with the putative Ca2+ binding ligand Glu309. We found that changes to hydrophobic residues predicted to form one surface of predicted transmembrane segment M4 usually resulted in mutants with reduced Ca2+ transport activity, but that mutants A305V and A306V on the opposite predicted surface lacked any detectable Ca2+ transport function. These mutants were blocked in the E2P to E2 transition. Modelling suggests that Ala305, Ala306, Glu309, Gly3lo, and Pro312 form an active site occupying part of one surface of transmembrane helix M4, while the large hydrophobic residues occupy the opposite surface.

EXPERIMENTAL PROCEDURES
Oligonucleotide-directed Mutagenesis and cDNA Expression in COS-I Cells-The methods employed have been described in detail (Maruyama and MacLennan, 1988;Clarke et al., 1989b). A summary is as follows. Mutagenesis using synthetic oligonucleotides was performed in fragments of the rabbit fast-twitch skeletal muscle Ca2+-ATPase cDNA by the method of Kunkel (1985). The fragment containing the desired mutation was subcloned back into its original position, and the entire cDNA was cloned into the EcoRI site of vector p91023(B) (Wong et al., 1985) for expression in COS-1 cells (Gluzman, 1981). Microsomes were prepared from the transfected cells and suspended in a solution containing 0.25 M sucrose, 0.15 M KC1, 3 mM 2-mercaptoethanol, 20 FM CaC12, and 10 mM Tris-HC1, pH 7.5. Ca2+ transport activity was assayed in a reaction mixture containing 20 mM MOPS, pH 6.8, 100 mM KC1, 5 mM MgCl,, 5 mM ATP, 0.45 mM CaClz (containing 1 mCi/ml 45Caz+), 0.5 mM EGTA, and 5 mM potassium oxalate. For the assay of Ca2+ dependence of various functions, free Ca" concentrations were calculated by the computer program of Fabiato and Fabiato (1979). A sandwich, enzyme-linked immunosorbent assay using monoclonal antibody A52 (Zubrzycka-Gaarn et al., 1984) was used to quantify the amount of Cazf-ATPase expressed in each microsomal preparation. Protein concentration was determined using a dye binding assay (Bradford, 1976) with bovine serum albumin as standard.
Assays of dephosphorylation of the phosphoenzyme intermediates were carried out as described by Andersen et al. (1989) and Clarke et al. (1990~). Labeled proteins were separated by gel electrophoresis using the method of Weber and Osborn (1969) using 6% acrylamide gel and running buffer adjusted to pH 6.3. Radioactivity was detected using Kodak X-Omat XAR5 film. Fig. 1 shows an a-helical representation of transmembrane segment M4. In previous studies of the functional consequences of mutations to glycine, prolines, and charged residues in this sequence, we observed that substitutions for Pro308, Glu309, Gly3lo, and Pro312 inhibited function of the enzyme (Clarke et al., 1989a;Vilsen et al., 1989;Andersen et al., 1992). A first question was whether mutation of additional amino acids in this transmembrane sequence would lead to loss of function. Within the stretch of 18 amino acids between IleZg8 and Ile315 there are 6 alanines, 3 valines, 2 leucines, 3 isoleucines, and 1 glycine. Each of the 6 alanines is juxtaposed with a larger hydrophobic residue, either leucine, isoleucine, or valine, while the single glycine is juxtaposed with leucine. Thus, a second question was whether this juxtaposition leads to meaningful packing of the transmembrane sequence and, to test this, large residues were replaced with small and vice The functional consequences of mutation of and , Glu309 (Clarke et al., 1989a), and Gly310 (Andersen et al., 1992) have been described in earlier papers. uersa. A third question was whether the polarity of these residues was important, and, to evaluate this possibility, some of the hydrophobic residues were replaced with the more polar residue, serine.

RESULTS
The mutants were transfected into COS-1 cells and microsomes were prepared for compositional and functional analysis. Each of the mutant proteins was readily synthesized and incorporated into the endoplasmic reticulum in COS-1 cells. The average expression level for each of the mutants in 3 to 6 microsomal preparations was estimated to lie between 80 and 270% of the average wild-type expression level (Table I). Immunoblots of each of the mutants demonstrated the presence of a single intact polypeptide of molecular weight 110,000 that reacted with antibody A52, specific for the Ca2+-ATPase (Maruyama and MacLennan, 1988). These observations indicated that none of the mutations affected incorporation of the protein into the endoplasmic reticulum or caused any major structural perturbation which would subject the protein to enhanced levels of proteolytic breakdown. Table I also shows the Ca2+ transport rates of the mutants relative to wild-type Ca2+-ATPase at pCa 5 in the presence of 5 mM ATP and 5 mM oxalate. Mutants I298A, A299V, A301V, A313V, and A313S all transported Ca2+ at maximal rates, close to those observed with the wild-type enzyme expressed and assayed under comparable conditions. The mutants V300.4, V300S, L302A, A303V, V304A, V304L, V304S, A305G, A306G, I307A, I307S, G310A, L311A, L311S, V314A, V314L, and I315A displayed reduced Ca2+ transport function ranging between 15% and 62% of the rate obtained with wildtype enzyme. No measurable Ca2+ transport activity was observed for mutants A305V, A306V, or G310V.
An estimate of the Ca2+ affinity of the mutants retaining Ca2+ transport function was determined by measurement of the Ca2+ dependence of Ca2+ transport. Mutants A303V and 1307s displayed significantly lower affinities of 1.6 p M and 1.2 WM, respectively) than the wild-type enzyme (KO, of 0.3 p~) . A small decrease in Ca2+ affinity was observed for mutants L302A, V304L, A305G, A306G, I307A, G310A, L311A, V314A, and V314L, while a small increase in Ca2+ Calculated from the content of mutant and wild-type Ca2+-ATPase proteins, determined by enzyme-linked immunosorbent assay, per mg of total microsomal protein.
Calcium transport rates were calculated relative to wild-type Caz+ uptake rates which were included as a control for each experiment. The average Ca2+ transport rate was 6.1 pmollminlmg of Ca2+-ATPase protein. Each mutant was carried through the transfection, expression, and Ca2+ uptake procedures at least twice.
Determined from the CaZ+ dependence of Ca2+ uptake. Estimated from Caz+ dependence of phosphorylation from ATP. Andersen et al., 1992. affinity was observed for mutants V300A, V300S, V304A, V3045, A313S, and I315A. The finding of a complete loss of Ca2+ transport activity for the A305V and A306V mutants led us to examine their partial reactions of Ca2+ transport. Since the binding of 2 Ca2+ ions with high affinity is required to activate the transfer of the phosphoryl group from ATP to the enzyme (de Meis and Vianna, 1979;Petithory and Jencks, 1988), phosphorylation of the enzyme from ATP also defines whether both Caz+ binding sites are intact in the mutant enzymes. Both mutants showed about the same level of phosphorylation from ATP as the wild-type enzyme when the reaction was carried out a t a free ca2+ concentration of approximately 100 p M (Fig. 2).
The Ca2+-dependent phosphorylation of the transport-defective mutants with ATP indicates that both of the high affinity Ca2+-binding sites were occupied in the presence of 100 p~ free Ca2+. When the free Ca2+ concentrations were varied below this level, however, a striking difference between mutant and wild-type enzymes was observed. As shown in Fig. 3, the wild-type Ca2+-ATPase was maximally phosphorylated at pCa 6, whereas phosphorylation of the A305V mutant was still increasing at pCa 5. On the other hand, the A306V mutant showed a higher apparent affinity for Ca2+ relative to  4 ) and with cDNAs encoding the + Val (lanes 2 and 5 ) and Ala""fi + Val (lanes 3 and 6 ) mutations which yielded products lacking Ca'+ transport function. Phosphoenzyme formation with either ATP or Pi was carried out in the presence (+) or absence (-) of 100 p~ Ca2+ as described under "Experimental Procedures." Samples containing an equivalent amount of expressed ATPase were separated by the method of Weber and Osborn (1969) using 6% acrylamide gels and a running buffer adjusted to pH 6.3. Radioactivity was detected by autoradiography. . Phosphorylation of the Ca2+-ATPase in microsomes prepared from COS-1 cells was performed a t 0 "C for 10 s with 2 p~ [y-"'PIATP in a reaction mixture containing 20 mM MOPS, pH 7.0, 80 mM KC1, 5 mM MgCl,, 100 p M CaCl,, and various concentrations of EGTA to give the free Ca'+ concentrations indicated. The free Caz+ concentration (pCa) was calculated according to Fabiato and Fabiato (1979). The acid-quenched samples were separated as described in the legend to Fig. 2, and radioactivity was detected by autoradiography.
that of the wild type, with maximal phosphorylation occurring at pCa 6.25-6.5.
Dephosphorylation of the high energy phosphoenzyme intermediate, EIP, in the mutants A305V and A306V was examined by phosphorylating each enzyme a t 0 "C in the presence of 2 p~ ATP and 0.1 mM Ca2+ at pH 7.0, stopping the reaction after 10 s by the addition of EGTA, and measuring decay of the phosphoenzyme intermediate through EzP and subsequent hydrolysis of E2P to EP. Dephosphorylation was also measured by the addition of both EGTA and ADP to the phosphoenzyme. In this case, dephosphorylation of EIP occurs through the rephosphorylation of ADP by the high energy phosphoenzyme intermediate. As shown in Fig. 4, the wildtype phosphoenzyme intermediate decayed rapidly after the addition of EGTA in the presence or absence of ADP, but, by contrast, the phosphorylated intermediates formed in mutants A305V and A306V were relatively stable under either condition. This suggests, first, that the phosphoenzymes formed in these mutants are not readily dephosphorylated and, second, that it is the E2P forms of the phosphoenzymes that are stabilized.
A phosphorylated intermediate of the Ca2+-ATPase can also be formed from Pi, but only in the absence of Ca2+, since Ca2+ binding appears to drive the native enzyme to a conformation incompatible with phosphorylation by Pi (Masuda and de Meis, 1973;de Meis and Masuda, 1974). As shown in Fig.  2, both of the mutant proteins formed phosphoenzyme intermediates from Pi in the absence of Ca2+. On the other hand, phosphorylation of the wild-type and both mutant enzymes by Pi was inhibited in the presence of 0.1 mM Ca2+. Dephosphorylation of the E2P phosphoenzyme intermediates was also examined by first phosphorylating the enzyme with [32P]Pi at 25 "C in the presence of 2 mM EGTA and 20% (v/v) dimethyl sulfoxide at pH 6.4 and then diluting the samples with 10 volumes of 50 mM MOPS, pH 7,80 mM KC1, 5 mM MgCl,, 1 mM nonradioactive Pi, and 0.1 mM CaClz at 0 "C to terminate the phosphorylation reaction and allow dephosphorylation to proceed. As shown in Fig. 5, decay of the E2P intermediates of both mutants occurred at rates slower than that observed with the wild-type enzyme. The E2P intermediate of the wild-type enzyme was virtually complete within 7 s, whereas a substantial level of E2P was observed for each of the mutants 7 s after the initiation of dephosphorylation.
The loss of function induced by mutation of Ala3' ' or Ala3'"j to larger residues could have been caused by increased bulk in the transmembrane sequence. To test this possibility, we carried out double mutations by replacing Va1304-Ala305 with Ala-Val and Ala3'"j-Ile307 with Ile-Ala. The double mutants were also devoid of Ca2+ transport activity (Table I). We then examined their partial reactions. Both double mutants were phosphorylated by ATP in the presence of Ca2+ and by Pi in the absence of Ca2+ (Fig. 6). As a control, we included G310V, whose properties were reported previously (Andersen et al., 1992). Measurement of the Ca2+ dependence of phosphorylation from ATP revealed that the VA305AV mutant had a lower affinity for Ca2+ than wild-type, while the AI307IA had a Ca2+ affinity very comparable to that of wild-type (Fig. 7).  I and 2), wild-type (lanes 3 and 4 ) , mutant Val3"-Ala305 + Ala-Val (lanes 5 and 6), mutant Ala3w-Ile307 + Ile-Ala (lanes   7 and B ) , and GIY1' + Val (lanes 9 and IO). Protocols  When dephosphorylation of the mutants was examined, their properties were found to be similar to those of the single A305V, A306V, and G310V mutants. The phosphorylated intermediates formed from ATP were relatively stable, both in the presence of ADP plus EGTA or of EGTA alone (Fig.  8) and appeared to be blocked in the E2P to E transition. This view was supported by the experiments shown in Fig. 9, in which these mutant proteins were phosphorylated by Pi and then transferred to a buffer favoring rapid dephosphorylation of the wild-type protein. Under these conditions, the mutants were seen to dephosphorylate slowly in comparison with the wild-type enzyme. These studies suggested that it was not only increased bulk within the membrane that affected function. Further analysis revealed that mutation of Ala305 to Gly resulted in reduction of transport activity to 28% of wild-type, while mutation of Ala3ffi to Gly resulted in a mutant with 62% activity. Thus, the precise size of these 2 residues is critical to proper functioning of the protein. It is also of interest that mutations of Val3" to Ala reduced Ca2+ transport activity to 25% while mutation of VaPo4 to Leu, of similar size, reduced activity to 35% of control values (Table I). Clearly, the precise alignment of these residues in this section of the transmembrane sequence is critical for full function.
Hydrophobic interactions are clearly important within transmembrane sequences. We therefore tested the functional consequences of mutation to Ser of a series of 4 hydrophobic residues, V a P , Va1304, Ile307, and Leu311, predicted to form a hydrophobic surface opposite to the more hydrophilic surface on which Ala305, Ala306, Gldm, Gly3lo, Pro312, and Ala313 (also mutated to Ser) are predicted to reside ( Table I). Each of these mutations reduced Ca2+ transport activity to between 27 and 48% of wild-type activity, suggesting the essentiality of large, hydrophobic residues on this surface of the a-helical sequence. The postulate that the precise alignment of these hydrophobic residues on one predicted surface is critical to function is supported by the fact that mutation of any of these 4 residues to Ala resulted in loss or reduction of activity to 18-48% of control activity. Mutation of Va1304 or Val314 to Leu also resulted in reduction of activity to about 35% of control value.

DISCUSSION
In previous studies, we demonstrated that 3 residues, G~u~~, Gly31°, and Pro312, located in transmembrane segment MA, play critical roles in Ca2+ transport by the Ca2+-ATPase. Changes to these residues had profound effects on either Ca2+ binding or subsequent conformational changes which accompany translocation of Ca2+ ions across the membrane. In the present study, we investigated the functional consequences of perturbing juxtaposed nonpolar residues of alternating large and small size in the transmembrane segment MI. Although most of the changes would result in an alteration in the bulk of the side chain or a decrease in hydrophobicity, none of the mutations actually caused a measurable decrease in structural integrity, since all mutant proteins were stably incorporated into the endoplasmic reticulum membrane. In previous studies of changes made to some 250 residues in the Ca2+-ATPase, very few have resulted in noticeable disruption of structural integrity of the expressed protein.
In Fig. 1, we have illustrated the results of the present study and integrated them with results from earlier investigations. The transmembrane segment, M4, and the stalk segment, SI, are predicted to form a long a-helix. In earlier studies (Vilsen et al., 1991), we attempted to define the border between S4 and M4. The triple mutant Ile-Thr-Thr317 + Arg-Asp-Asp was stably incorporated into the membrane, but was functionally inactive. This was considered to be consistent with the location of these residues at the S4/M4 boundary in at least one conformation of the enzyme. Lack of penetration of these residues into the membrane during transition to other conformations may, however, have been a factor in the loss of function of this triple mutant.
In the present study, we examined the functional consequences of mutations of the small Ala and Gly residues and large, hydrophobic Leu, Ile, and Val residues in MI. It is of interest that 7 of the 8 large hydrophobic residues in this sequence are juxtaposed with a small residue. Examination of Fig. 1 reveals three functional regional domains within M4. The first, consisting of IleZg8, Ala2%, and Val3'', are relatively unaffected by mutation. We can assume that this face of the transmembrane helix, near the lumenal boundary, is of little functional significance.
A second group of residues, consisting of Ala305, Alam, G~u~~, Gly3lo, and Pro3", create a second functional domain occupying a face of the helix in the middle of the transmembrane sequence. This is an essential surface for the Ca2+ translocation site. Mutations of these 5 residues invariably resulted in dramatic reductions in Ca2+ transport. For the Ala3OS or Ala3ffi to Gly mutations, a small decrease in size, or for the Gly' " to Ala mutation, a small increase in size led to loss of 40-70% of Ca2+ transport activity. Mutation of these residues to Val led to total loss of activity and here the effect was a block of conformational changes which, through long range interactions, permit dephosphorylation of Asp351 in the cytoplasmic domain. This was also true for the mutation of G I U~~ to Asp (Clarke et al., 1990a). The mutation of Pro3" to Ala or Gly also led to loss of Ca2+ transport activity, but here the block was in the transition of the EIP conformation to the E2P conformation, again through long range interactions . Mutation of each of these residues also altered Ca2+ affinity. The G~u~~ to Gln mutation provided the most striking loss of Ca2+ affinity, since Ca2+, even at 100 p~, could not prevent phosphorylation of the Ca2+-ATPase by Pi at acid pH (Clarke et al., 1989a). Mutation of Gly3lo to Ala (Andersen et al., 1992) or of Ala"' to Val also led to large reductions in Ca2+ affinity. By contrast, mutation of Pro312 to Ala or Gly enhanced Ca2+ affinity by %fold , and this was also true of the mutation of Ala3ffi to Val (Fig. 3). Thus, each of these residues appears to be intimately involved with Ca2+ binding and with the conformational changes that precede and accompany the translocation of Ca2+ across the membrane. It is no wonder, then, that they should be so sensitive to substitutions that alter their spatial environment by even small amounts.
A third group of residues consisting of Val3m, Leu3'', Ala303, V a P , Ile307, Pro3'', Leu311, Va1314, and Ile315 can be envisioned, in three dimensions, as forming a hydrophobic surface on the opposite side of the transmembrane helix from the surface containing the active site residues Ala3' ' , Ala3ffi, G~u~~, Gly3lo, and Pro3" (Fig. 1). Mutations to residues on this hydrophobic surface, leading to reduction in size or polarity, led to considerable, but not complete, loss of function (Table I). In the case of the Ala303 to Val, Ile307 to Ser, and Pro308 to Ala (Vilsen  et al., 1989) mutations, a decrease in Ca2+ affinity of 3-to 5fold was observed. We suggest that maintenance of this pre-cise hydrophobic surface is important for the efficient functioning of the Ca2+ pump. This point is emphasized by the fact that mutations of Va1304 and VaP4 to Leu, involving a relatively small change in size and hydrophobicity, led to reduction of Ca2+ transport function to about 35% of control values (Table I).
An important question in designing the present experiments was whether there is significance in the alteration of the small amino acids Ala or Gly, with the large bulky residues Val, Leu, or Ile in the transmembrane segment. Table I illustrates that there was relatively little significance to the large and small juxtaposition at the transmembrane boundaries for the mutations I298A, A299V, A301V, A313V, or A313S. By contrast, replacement of Ala305 or Ala306 with Val led to complete loss of function. The inverse double mutations of residues 304 and 305 or of residues 306 and 307, resulting in a restoration of the original bulk, but in different juxtaposition, did not lead to restoration of function. The resulting proteins had no activity and were similar in character to the single mutations of Ala305 or Ala306 to Val. Even the alteration of Ala305 or of Ala3% to Gly led to very large losses in function and to a slightly lower Caz+ affinity, suggesting that the sizes of these 2 residues are critical to function. This indicates that the region near the center of transmembrane sequence 4 is precisely packed, in line with its central role in the conformational changes leading to Ca2+ transport, and that any alteration in the arrangement of the small and bulky residues in this part of the molecule is significant. Mutation of Ala305 or Ala306 to Gly did not prevent turnover of the enzyme, but it did result in lower activity and a slightly lower calcium affinity.
Alteration of Ala305, Ala306, or Gly310 to Val led to a phenotype in which the enzyme could be phosphorylated from either ATP and Ca2+ or from Pi, forming a relatively stable E2P phosphoenzyme intermediate in each case. The major block in the Ca2+ transport reaction cycle was found at the phosphatase step. We do not propose that transmembrane sequences are responsible for phosphatase function. Rather we believe that this phenotype represents a subset of conformational change mutants in which the phosphoryl group in the E2P conformation does not become accessible to phosphatase activity which must reside within the cytoplasmic domains where the phosphoryl group is formed. We believe that these blocks represent long range interactions in which the influence of the alteration of a residue in one active site (the Ca2+ binding domain) can be observed at a second active site (the ATP hydrolytic domain). We believe that these sites are separated by 40-50 A in space, but communicate readily through conformational changes.
In earlier studies, Serrano (1988) proposed that a phosphatase activity might reside in the 8-strand domain, since mutation of a Gly in the P-strand domain of the H+-ATPase corresponding to Gly233 in the CaZ'-ATPase appeared to block phosphatase activity. Later studies  showed that it was the EIP form of the enzyme, rather than the E2P form, that was stabilized. Accordingly, Gly233 is not involved in phosphatase function. The question of whether the p-strand domain contains phosphatase activity is, however, still not answered.