The Mechanism of Ca2+ Transport by Sarco(Endo)plasmic Reticulum Ca2+-ATPases

Ca2+ pumps, together with Ca2+release channels, form ubiquitous Ca2+ regulatory systems in muscle and non-muscle cells. The sarco(endo)plasmic reticulum Ca2+-ATPases (SERCA) 1The abbreviations used are: SERCA, sarco(endo)plasmic reticulum Ca2+-ATPases; 8-azido-TNP-ATP, 8-azido-2′(3′)-O-(2,4,6-trinitrophenyl)adenosine 5′-triphosphate; FTIR, Fourier-transform infrared spectroscopy. and the plasma membrane Ca2+-ATPases have the highest affinity for Ca2+ removal from the cytoplasm and, together, set resting cytoplasmic Ca2+ concentrations. Three differentially expressed genes encode SERCA proteins (1Wu K.D. Lee W.S. Wey J. Bungard D. Lytton J. Am. J. Physiol. 1995; 269: C775-C784Crossref PubMed Google Scholar). SERCA1a and -1b are expressed in fast-twitch skeletal muscle, but loss of SERCA1 function in Brody disease is sufficiently compensated to preserve life (2Odermatt A. Taschner P.E.M. Khanna V.K. Busch H.F.M. Karpati G. Jablecki C.K. Breuning M.H. MacLennan D.H. Nat. Genet. 1996; 14: 191-194Crossref PubMed Scopus (190) Google Scholar). SERCA2a is the cardiac/slow-twitch isoform, whereas SERCA2b, with a C-terminal extension, is expressed in smooth muscle and non-muscle tissues. It is almost certainly an essential gene. SERCA3 is expressed in a limited set of non-muscle tissues, including endothelial, epithelial, and lymphocytic cells and platelets, and its knockout is not lethal (3Liu L.H. Paul R.J. Sutliff R.L. Miller M.L. Lorenz J.N. Pun R.Y.K. Duffy J.J. Doetschman T. Kimura Y. Machennan D.H. Hoying J.B. Shull G.E. J. Biol. Chem. 1997; 272 (in press)Google Scholar). SERCA enzymes are typical of the class of P-type ATPases, which form a phosphoprotein intermediate and undergo conformational changes during the course of ATP hydrolysis (4de Meis L. Vianna A.L. Annu. Rev. Biochem. 1979; 48: 275-292Crossref PubMed Scopus (557) Google Scholar, 5Jencks W.P. Ann. N. Y. Acad. Sci. 1992; 671: 49-56Crossref PubMed Scopus (20) Google Scholar). Some of the conformational states can be stabilized, either by adjustment of reaction conditions or through mutagenesis, and characterized as intermediates in the overall reaction cycle (Fig. 1 A). The phosphorylated intermediate,E 1P(Ca)2, can phosphorylate ADP, whereas E 2P can only react with water. The formation of E 1P requires that two high affinity Ca2+ binding sites be occupied. The enzyme is then phosphorylated by ATP and, concomitantly, the two Ca2+ ions are occluded and can no longer exchange with cytoplasmic Ca2+. The rate-limiting transition toE 2P is accompanied by loss of Ca2+into the lumen, the affinity having fallen by 3 orders of magnitude. Hydrolysis of E 2P and regeneration of the high affinity Ca2+ binding sites (E 1(Ca)2) complete the reversible cycle. High lumenal Ca2+ drives the formation ofE 1P from phosphate (Pi), and its effect on the level of E 1P led Jencks (5Jencks W.P. Ann. N. Y. Acad. Sci. 1992; 671: 49-56Crossref PubMed Scopus (20) Google Scholar, 6Mintz E. Guillain F. Biochim. Biophys. Acta. 1997; 1318: 52-70Crossref PubMed Scopus (91) Google Scholar) to postulate a second set of Ca2+ binding sites on the lumenal surface. Proton countertransport, involving the exchange of one H+ per Ca2+, has been shown during the reaction cycle (7Yu X. Carroll S. Rigaud J.L. Inesi G. Biophys. J. 1993; 64: 1232-1242Abstract Full Text PDF PubMed Scopus (146) Google Scholar), emphasizing the similarity between Ca/H- and Na/K- or H/K-ATPases. Two-dimensional arrays and helical tubes of SERCA1a were first produced by treatment of native membranes with decavanadate, and in a negative stain, these yielded a three-dimensional structure with a resolution of 25 Å (8Martonosi A.N. Biosci. Rep. 1995; 15: 263-281Crossref PubMed Scopus (39) Google Scholar). At high Ca2+ concentration, thin plates were obtained, which have given a 6-Å projection map (8Martonosi A.N. Biosci. Rep. 1995; 15: 263-281Crossref PubMed Scopus (39) Google Scholar, 9Stokes D.L. Green N.M. J. Mol. Biol. 1990; 213: 529-538Crossref PubMed Scopus (58) Google Scholar). Thapsigargin, a SERCA-specific inhibitor (10Lytton J. Westlin M. Hanley M.R. J. Biol. Chem. 1991; 266: 17067-17071Abstract Full Text PDF PubMed Google Scholar) that appears to bind to the M3 transmembrane sequence (11Norregaard A. Vilsen B. Andersen J.P. J. Biol. Chem. 1994; 269: 26598-26601Abstract Full Text PDF PubMed Google Scholar), stabilizes theE 2 state of the pump and promotes formation of helical tubes (12Stokes D.L. Lacapere J.J. J. Biol. Chem. 1994; 269: 11606-11613Abstract Full Text PDF PubMed Google Scholar). These are also compatible with bound nucleotides, but they are disrupted by low Ca2+. In contrast, the thin plates, probably corresponding toE 1(Ca)2, are disrupted by thapsigargin and by nucleotides. Current modeling is based on a 14-Å structure (Fig. 1 B) obtained by cryoelectron microscopy of decavanadate tubes (13Toyoshima C. Sasabe H. Stokes D.L. Nature. 1993; 362: 467-471Crossref PubMed Scopus (199) Google Scholar). A large cytoplasmic head is linked to the membrane by a narrow stalk. The protein within the membrane is divided into two major densities,A1 and A2, lying beneath the stalk, and two minor densities, B and C, to one side. A recent structure for the CrATP-bound complex (14Yonekura K. Stokes D.L. Sasabe H. Toyoshima C. Biophys. J. 1997; 72: 997-1005Abstract Full Text PDF PubMed Scopus (40) Google Scholar) locates the nucleotide binding site in the groove on the underside of the head (Fig.1 B). Two main segments of sequence (Fig. 2) form the cytoplasmic head and stalk (Fig. 1 B) (15MacLennan D.H. Brandl C.J. Korczak B. Green N.M. Nature. 1985; 316: 696-700Crossref PubMed Scopus (855) Google Scholar). The segment of about 130 residues following the M1/M2 hairpin, likely to form a 7–8-membered β-strand domain, is linked by long, amphipathic helices, S2 and S3, to M2 and M3. A central segment of about 440 residues forms the main head region. It includes the phosphorylation site at Asp351 and widely separated residues such as Lys492, Lys515, Cys674, Lys684, Asp703, and Asp707, which are affinity labeled by various nucleotide analogues in either Ca2+ or homologous Na/K-ATPases (16Moller J.V. Juul B. le Maire M. Biochim. Biophys. Acta. 1996; 1286: 1-51Crossref PubMed Scopus (667) Google Scholar). The N and C termini of the central domain form stalk helices S4 and S5, which link M4 to the phosphorylation site and M5 to the hinge region. The bulk of the central domain is predicted to be a mixture of α-helices and β-strands, which alternate fairly regularly in the C-terminal half, a characteristic associated with nucleotide binding, but less regularly in the N-terminal half, referred to as the phosphorylation domain. The phosphorylation domain extends to the variable sequence preceding Lys492, which is labeled by 8-azido-TNP-ATP (17McIntosh D.B. Woolley D.G. Berman M.C. J. Biol. Chem. 1992; 267: 5301-5309Abstract Full Text PDF PubMed Google Scholar), and is likely to be part of the nucleotide binding domain. On the basis of a 20-residue Walker B region and an overall β-α-β pattern of secondary structure, kinase-related folds in the nucleotide binding region were proposed (18Taylor W.R. Green N.M. Eur. J. Biochem. 1989; 179: 241-248Crossref PubMed Scopus (136) Google Scholar). However, the inclusion of Lys492 implicated an extra antiparallel β-strand in this domain, which together with immunological evidence for an exposed epitope on the central strand of the β-sheet (19Mata A.M. Matthews I. Tunwell R.E. Sharma R.P. Lee A.G. East J.M. Biochem. J. 1992; 286: 567-580Crossref PubMed Scopus (31) Google Scholar, 20Lee A.G. Lee A.G. Biomembranes. 5. JAI Press, Amsterdam1996Google Scholar) suggests that this region has a new fold specific to P-type pumps. The C terminus of the nucleotide binding domain, close to the top of the stalk, is highly conserved and may form a subdomain, which by analogy with some kinases could form a hinge between phosphorylation and nucleotide binding domains. Several sites in this hinge region are labeled by γ-phosphate-linked affinity labels, suggesting that it is close to the phosphorylation site (16Moller J.V. Juul B. le Maire M. Biochim. Biophys. Acta. 1996; 1286: 1-51Crossref PubMed Scopus (667) Google Scholar). The sequences of all P-type ion pumps, except for the shorter ones mostly specific for copper or cadmium, show a common pattern of hydrophobicity (21Stokes D.L. Taylor W.R. Green N.M. FEBS Lett. 1994; 346: 32-38Crossref PubMed Scopus (60) Google Scholar). A 10-transmembrane helix model for SERCA1a and SERCA2a (Fig. 3) was proposed on the basis of analysis of hydrophobicity (15MacLennan D.H. Brandl C.J. Korczak B. Green N.M. Nature. 1985; 316: 696-700Crossref PubMed Scopus (855) Google Scholar). It has received growing support (16Moller J.V. Juul B. le Maire M. Biochim. Biophys. Acta. 1996; 1286: 1-51Crossref PubMed Scopus (667) Google Scholar, 20Lee A.G. Lee A.G. Biomembranes. 5. JAI Press, Amsterdam1996Google Scholar) from well controlled experiments with proteases, antibodies, and sulfhydryl labels, which showed the N and C termini to be cytoplasmic and the M7/M8 loop to be lumenal, and from proteolysis of intact vesicles with trypsin, which led to the isolation of hydrophobic fragments corresponding to M1/M2, M3/M4, M5/M6, and M7/M8 (22Shin J.M. Kajimura M. Arguello J.M. Kaplan J.H. Sachs G. J. Biol. Chem. 1994; 269: 22533-22537Abstract Full Text PDF PubMed Google Scholar). Evidence for the M9/M10 hairpin comes from experiments involvingin vitro insertion of this hairpin into membranes using eukaryotic expression vectors encoding a series of membrane inserts (23Bayle D. Weeks D. Sachs G. J. Biol. Chem. 1995; 270: 25678-25684Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar). The volume of the transmembrane domain, as deduced from a 14-Å structure (21Stokes D.L. Taylor W.R. Green N.M. FEBS Lett. 1994; 346: 32-38Crossref PubMed Scopus (60) Google Scholar), will accommodate 10-transmembrane helices. Tentative assignments of the 10-transmembrane helices (Fig. 3) to the four transmembrane regions, A1, A2, B, and C (Fig. 1 B), have been made (13Toyoshima C. Sasabe H. Stokes D.L. Nature. 1993; 362: 467-471Crossref PubMed Scopus (199) Google Scholar, 21Stokes D.L. Taylor W.R. Green N.M. FEBS Lett. 1994; 346: 32-38Crossref PubMed Scopus (60) Google Scholar) but will be subject to refinement. M2, M3, M4, and M5, which underlie the stalk, must occupy part of A1, but differing slopes of helices below the stalk would permit crossover to other regions. Since M4, M5, and M6 must cluster to form the Ca2+binding sites (see below), M6 must lie close, making it a strong candidate for A2. This places the entrance to the Ca2+binding sites to one side of the stalk on the boundary between A1 and A2. The small lumenal domain joins the A2 region to the B region. Since the only long loop (38 residues) on that side of the membrane joins M7 to M8, the M7/M8 hairpin can be associated with both A2 and B. Cytoplasmic densities near the membrane surface, above the A1 and C regions, could represent N and C termini, respectively. If so, M9 and M10 might occupy the C region, whereas M1 might be located in A1. A second approach to helix arrangement involves analysis of conserved (internal) and variable (lipid-exposed) sites in the transmembrane sequences of diverse Ca2+ pumps (21Stokes D.L. Taylor W.R. Green N.M. FEBS Lett. 1994; 346: 32-38Crossref PubMed Scopus (60) Google Scholar). Site-directed mutagenesis has provided key insights into structure/function relationships in SERCA1 and SERCA2 (24MacLennan D.H. Clarke D.M. Loo T.W. Skerjanc I.S. Acta Physiol. Scand. 1992; 607: 141-150Google Scholar, 25Andersen J.P. Biosci. Rep. 1995; 15: 243-261Crossref PubMed Scopus (118) Google Scholar). In these experiments, mutated cDNA is expressed transiently in heterologous cell culture, microsomal vesicles are isolated, and overall and partial reactions of ATP-dependent Ca2+ transport are assayed. In Fig. 3, loss of function, reduced function, and unaltered function mutants in SERCA1 or SERCA2 are located relative to predicted structural domains. Two principles were key to the characterization of two Ca2+ binding sites through mutagenesis. These were: (i) that binding of Ca2+ to the first site (site I), presumably the more distal from the cytoplasm, leads to cooperative binding to the second, presumably more proximal site (site II) (6Mintz E. Guillain F. Biochim. Biophys. Acta. 1997; 1318: 52-70Crossref PubMed Scopus (91) Google Scholar, 26Inesi G. Sumbilla C. Kirtley M.E. Physiol. Rev. 1990; 70: 749-760Crossref PubMed Scopus (169) Google Scholar); and (ii) occupation of both Ca2+binding sites I and II is required for "forward" phosphorylation from ATP, whereas occupation of site I alone is sufficient to convert Pi-reactive E 2 conformations to non-reactive E 1, thereby depleting the substrate for "reverse" phosphorylation from Pi (25Andersen J.P. Biosci. Rep. 1995; 15: 243-261Crossref PubMed Scopus (118) Google Scholar, 27Andersen J.P. Vilsen B. FEBS Lett. 1995; 359: 101-106Crossref PubMed Scopus (116) Google Scholar) (Fig.1). The initial mutagenic screen identified Glu309 in M4, Glu771 in M5, Asn796, Thr799, and Asn800 in M6, and Glu908 in M8 as potential Ca2+ binding ligands (28Clarke D.M. Loo T.W. Inesi G. MacLennan D.H. Nature. 1989; 339: 476-478Crossref PubMed Scopus (510) Google Scholar). Mutants were Ca2+transport negative and not phosphorylated by ATP plus Ca2+; for all but N796A, high Ca2+ did not prevent reverse phosphorylation, suggesting that mutation of any of these residues would lead to the loss of at least one Ca2+ binding site. The first measurements of reverse phosphorylation were carried out at pH 6.4, but later measurements, carried out at neutral pH, showed normal Ca2+ inhibition of phosphorylation from Pi for mutants E309Q as well as for N796A (29Andersen J.P. Vilsen B. J. Biol. Chem. 1992; 267: 19383-19387Abstract Full Text PDF PubMed Google Scholar, 30Andersen J.P. Vilsen B. J. Biol. Chem. 1994; 269: 15931-15936Abstract Full Text PDF PubMed Google Scholar). These results are consistent with retention of Ca2+ binding site I, implying that both Glu309 and Asn796contribute to site II. This conclusion was supported by the direct demonstration that mutant E309Q retains a single Ca2+binding site and that, at pH 6.4, Ca2+ can gain access to the site (presumably from the lumenal side) in detergent-disrupted membranes but not in intact vesicles (31Skerjanc I.S. Toyofuku T. Richardson C. MacLennan D.H. J. Biol. Chem. 1993; 268: 15944-15950Abstract Full Text PDF PubMed Google Scholar). By contrast, mutants E771Q, T799A, and E908A showed similar, very low Ca2+ affinity in both forward and reverse phosphorylation assays, implying that site I was disrupted (29Andersen J.P. Vilsen B. J. Biol. Chem. 1992; 267: 19383-19387Abstract Full Text PDF PubMed Google Scholar, 30Andersen J.P. Vilsen B. J. Biol. Chem. 1994; 269: 15931-15936Abstract Full Text PDF PubMed Google Scholar). Mutant D800N showed reduced Ca2+ affinity in both assays but to different extents, suggesting that Asp800 was contributing to both sites. In the plasma membrane Ca2+-ATPases, which transport only a single Ca2+, the residues homologous to Glu771, Thr799, and Glu908, all assigned to site I, are replaced by Ala, Met, and Gln, respectively (32Shull G.E. Greeb J. J. Biol. Chem. 1988; 263: 8646-8657Abstract Full Text PDF PubMed Google Scholar). Mutants E309N, E771Q, N796A, T799A, D800N, as well as G310P and G801V (but not G770A) lost the ability to occlude Ca2+ in the presence of CrATP (25Andersen J.P. Biosci. Rep. 1995; 15: 243-261Crossref PubMed Scopus (118) Google Scholar, 30Andersen J.P. Vilsen B. J. Biol. Chem. 1994; 269: 15931-15936Abstract Full Text PDF PubMed Google Scholar, 33Vilsen B. Andersen J.P. J. Biol. Chem. 1992; 267: 25739-25743Abstract Full Text PDF PubMed Google Scholar). By contrast, E908Q retained full function (mutants E309Q, E771Q, and D800N are non-functional), and the mutant E908A retained Ca2+ occlusion and transport with low affinity (25Andersen J.P. Biosci. Rep. 1995; 15: 243-261Crossref PubMed Scopus (118) Google Scholar). These observations, together with the fact that Glu908 is the only mutation-sensitive residue in all of M8 (34Rice W.J. MacLennan D.H. J. Biol. Chem. 1996; 271: 31412-31419Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar), suggest, at best, a peripheral role in Ca2+ binding and transport for Glu908 and for helix M8. The location of Ca2+ ligands on separate helices means that correct helix orientation will be crucial for the formation of high affinity Ca2+ binding sites and that reorientation coupled to movements of the cytoplasmic domains could cause occlusion and changes in Ca2+ affinity. The packing of these helices is being studied by introduction of pairs of cysteines into selected positions and assay of the expressed products for cross-linking (35Rice W.J. Green N.M. MacLennan D.H. J. Biol. Chem. 1997; (in press): 272Google Scholar). Cross-links observed at different tiers of helices M4 and M6 (A305C/L793C, E309C/N796C, T317C/A804C) are in relative positionsi, i + 4, i + 12, favoring packing of M4 and M6 as a right-handed supercoil at an angle of about 40°. It would normally be difficult to maintain such a large angle over several turns of helix, but the presence of four prolines and three glycines could permit sufficient curvature. Insights from cross-linking data provide a possible solution to a problem raised by the assignment of Ca2+ ligands to two sites in a model with the two sites stacked one above the other (Fig.3 A) (25Andersen J.P. Biosci. Rep. 1995; 15: 243-261Crossref PubMed Scopus (118) Google Scholar, 26Inesi G. Sumbilla C. Kirtley M.E. Physiol. Rev. 1990; 70: 749-760Crossref PubMed Scopus (169) Google Scholar). Stacking leads to the placement of Asn796 in the more cytoplasmic site (site II), even though it is the most lumenal ligand. Andersen (25Andersen J.P. Biosci. Rep. 1995; 15: 243-261Crossref PubMed Scopus (118) Google Scholar) suggested that this might be resolved if M6 were not fully helical. However, if M4 and M6 are oriented to optimize cross-links between them, as indicated in Fig.3 B (cross-section), then Glu309, Asn796, and Asp800 would be positioned near the M4/M6 contact, whereas Thr799 would lie to the right of this contact. If M5 is placed so that Glu771 is apposed to Thr799 in M6, then the ligands between M4 and M6 are those assigned to site II (Glu309, Asn796), whereas the ligands between M5 and M6 are those assigned to site I (Glu771, Thr799). Asp800 in M6 would be in a position to contribute to both sites. This new "side-by-side sites" model has the advantage that Asn796, which lies below Asp800 in the helix (Fig. 3 B, oblique view), can contribute to site II without distortion of the M5/M6 helices. Positioning of M8 so that Glu908 would be apposed to site I would permit its peripheral contribution to that site. The side-by-side sites model would mean that the pathway of Ca2+ would an course (Fig.3 B, large a direct pathway (Fig. 3 In both occupation of site I by Ca2+ cooperative binding to site in Ca2+ at site I, but in the side-by-side sites model Ca2+ to site II could be through an pathway could are following and could be or through a single pathway or large evidence for during occupation of the more lumenal site (presumably site from the more cytoplasmic the of the at the cytoplasmic (6Mintz E. Guillain F. Biochim. Biophys. Acta. 1997; 1318: 52-70Crossref PubMed Scopus (91) Google Scholar). mutagenesis that small residues lying within on the helix in the three turns the Ca2+ ligands in M4, M5, and M6 pump at a of different in the reaction cycle (34Rice W.J. MacLennan D.H. J. Biol. Chem. 1996; 271: 31412-31419Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar). By contrast, only a of the residues in the top and two turns of M4, M5, and M6 are to mutation and, are in This of small residues in the of the membrane could provide a with the residues access to In M5 the of the above Glu771 and near the top of M5, by the an mutant J.P. J. Biol. Chem. 1995; 270: Full Text Full Text PDF PubMed Scopus Google Scholar) in which Ca2+ to the cytoplasm it can be (Fig. At the of M4, the mutation-sensitive (28Clarke D.M. Loo T.W. Inesi G. MacLennan D.H. Nature. 1989; 339: 476-478Crossref PubMed Scopus (510) Google Scholar, L. Sumbilla C. D. L. C. Kirtley M.E. Inesi G. J. Biol. Chem. 1996; 271: Full Text Full Text PDF PubMed Scopus Google Scholar) has a and position to function as a residue for of Ca2+ to the A in the loop loss of and loss of the ability of Ca2+ to prevent phosphorylation from Pi T. F. L. A. S. F. J.V. le Maire M. J. Biol. Chem. 1997; Full Text Full Text PDF PubMed Scopus Google Scholar). This loop could form part of the to Ca2+ binding sites I and II or be part of the during A in the mutation-sensitive of M4 and M6, suggests a sequence (34Rice W.J. MacLennan D.H. J. Biol. Chem. 1996; 271: 31412-31419Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar). A sequence in the region of M5 is However, is no in the ligands by M4, M5, and M6 to the Ca2+ binding and transport the of the similarity in this of binding sequences is not of highly conserved sequences in the large cytoplasmic domain between M4 and M5 showed that mutants did not form intermediates from either ATP or consistent but not in ATP binding K. Clarke D.M. J. Inesi G. Loo T.W. MacLennan D.H. J. Biol. Chem. 1989; Full Text PDF PubMed Google Scholar, D.M. Loo T.W. MacLennan D.H. J. Biol. Chem. 1990; Full Text PDF PubMed Google Scholar). of ATP binding affinity through inhibition of is a recent assay D.B. Woolley D.G. Vilsen B. Andersen J.P. J. Biol. Chem. 1996; 271: Full Text Full Text PDF PubMed Scopus Google Scholar). this mutants of and Lys492 were to have ATP of in and high ATP in both and regulatory ATP Mutants in also to show Ca2+ The crucial normally its and its Ca2+ within a but the intermediate can be A of in all the major domains this (24MacLennan D.H. Clarke D.M. Loo T.W. Skerjanc I.S. Acta Physiol. Scand. 1992; 607: 141-150Google Scholar, 25Andersen J.P. Biosci. Rep. 1995; 15: 243-261Crossref PubMed Scopus (118) Google Scholar), that the conformational E 1P E 2P all of these domains. Mutants that hydrolysis of E 2P have so been only in M4, M5, and M6, implying that a long between cytoplasmic and transmembrane sites this The of phosphorylation sites and within the long of conformational whereas the domains and long helices provide a for the and of conformational changes be without a structure, a of and have been to conformational changes and, in some to of the hinge domain to the N terminus of the nucleotide binding domain 1P E 2P D.B. J. Biol. Chem. 1992; 267: Full Text PDF PubMed Google Scholar), evidence for essential domain of following the of ATP D. D.H. A. Biophys. J. 1985; 48: Full Text PDF PubMed Scopus Google Scholar) or of structural of the enzyme in two different conformations H. Toyoshima C. Stokes D.L. Biophys. J. 1996; 70: Full Text PDF PubMed Scopus Google Scholar) provide evidence for large changes consistent with domain domain are not in changes in J.L. Y. FEBS Lett. 1992; PubMed Scopus Google Scholar), in or in between bound Inesi G. Biochim. Biophys. Acta. 1992; PubMed Scopus Google Scholar). in which can be assigned to specific (8Martonosi A.N. Biosci. Rep. 1995; 15: 263-281Crossref PubMed Scopus (39) Google Scholar, A. F. J. Biol. Chem. 1996; 271: Full Text Full Text PDF PubMed Scopus Google Scholar, A. K. Y. Biophys. J. 1996; Full Text PDF PubMed Scopus Google Scholar) show that residues may be and that the observed are of the results provide direct evidence for small conformational changes Ca2+ binding, ATP 1P E and hydrolysis of E A for a P-type in into the of the transport and the structure of the pump (24MacLennan D.H. Clarke D.M. Loo T.W. Skerjanc I.S. Acta Physiol. Scand. 1992; 607: 141-150Google Scholar). The Ca2+ binding and sites are in a between M4, M5, and M6, they are by the of Ca2+ binding residues located in these three helices. to the is controlled by between the residues near the cytoplasmic of the helices. The domain movements that close cytoplasmic access to the will If such movements M4, M5, or M6, they might also be to the placement of the ligands required to form high affinity so that, the two Ca2+ ions would be less bound and of consistent with D. Guillain F. E. J. Biol. Chem. 1996; 271: Full Text Full Text PDF PubMed Scopus Google Scholar). long domain movements will the of bound Ca2+ to the conformational changes will in of E 2P and of E 1(Ca)2, the Ca2+ transport cycle. L. Stokes for of the and of recent structural and Y. for in Fig.