Crystals of sarcoplasmic reticulum Ca(2+)-ATPase

High-resolution structures of the Ca(2+)-ATPase have over the last 5 years added a structural dimension to our understanding of the function of this integral membrane protein. The Ca(2+)-ATPase is now by far the membrane protein where the most functionally different conformations have been described in precise structural detail. Here, we review our experience from solving Ca(2+)-ATPase structures: a purification scheme involving minimum handling of the protein to preserve natural and essential lipids, a rational approach to screening for crystals based on a limited number of polyethyleneglycols and many different salts, improving crystal quality using additives, collecting the data and finally solving the structures. We argue that certain of the lessons learned in the present study are very likely to be useful for crystallisation of eukaryotic membrane proteins in general.     

Proton paths in the sarcoplasmic reticulum Ca(2+) -ATPase

The sarcoplasmic reticulum Ca(2+)-ATPase (SERCA1a) pumps Ca(2+) and countertransport protons. Proton pathways in the Ca(2+) bound and Ca(2+)-free states are suggested based on an analysis of crystal structures to which water molecules were added. The pathways are indicated by chains of water molecules that interact favorably with the protein. In the Ca(2+) bound state Ca(2)E1, one of the proposed Ca(2+) entry paths is suggested to operate additionally or alternatively as proton pathway. In analogs of the ADP-insensitive phosphoenzyme E2P and in the Ca(2+)-free state E2, the proton path leads between transmembrane helices M5 to M8 from the lumenal side of the protein to the Ca(2+) binding residues Glu-771, Asp-800 and Glu-908. The proton path is different from suggested Ca(2+) dissociation pathways. We suggest that separate proton and Ca(2+) pathways enable rapid (partial) neutralization of the empty cation binding sites. For this reason, transient protonation of empty cation binding sites and separate pathways for different ions are advantageous for P-type ATPases in general.     

Characterisation of thapsigargin-releasable Ca(2+) from the Ca(2+)-ATPase of sarcoplasmic reticulum at limiting [Ca(2+)

The Ca(2+) binding sites of the Ca(2+)-ATPase of skeletal muscle sarcoplasmic reticulum (SR) have been identified as two high-affinity sites orientated towards the cytoplasm, two sites of low affinity facing the lumen, and a transient occluded species that is isolated from both membrane surfaces. Binding and release studies, using (45)Ca(2+), have invoked models with sequential binding and release from high- and low-affinity sites in a channel-like structure. We have characterised turnover conditions in isolated SR vesicles with oxalate in a Ca(2+)-limited state, [Ca(2)](lim), where both high- and low-affinity sites are vacant in the absence of chelators (Biochim. Biophys. Acta 1418 (1999) 48-60). Thapsigargin (TG), a high-affinity specific inhibitor of the Ca(2+)-ATPase, released a fraction of total Ca(2+) at [Ca(2+)](lim) that accumulated during active transport. Maximal Ca(2+) release was at 2:1 TG/ATPase. Ionophore, A23187, and Triton X-100 released the rest of Ca(2+) resistant to TG. The amount of Ca(2+) released depended on the incubation time at [Ca(2+)](lim), being 3.0 nmol/mg at 20 s and 0.42 nmol/mg at 1000 s. Rate constants for release declined from 0. 13 to 0.03 s(-1). The rapidly released early fraction declined with time and k=0.13 min(-1). Release was not due to reversal of the pump cycle since ADP had no effect; neither was release impaired with substrates acetyl phosphate or GTP. A phase of reuptake of Ca(2+) followed release, being greater with shorter delay (up to 200 s) following active transport. Reuptake was minimal with GTP, with delays more than 300 s, and was abolished by vanadate and at higher [TG], >5 microM. Ruthenium red had no effect on efflux, indicating that ryanodine-sensitive efflux channels in terminal cisternal membranes are not involved in the Ca(2+) release mechanism. It is concluded that the Ca(2+) released by TG is from the occluded Ca(2+) fraction. The Ca(2+) occlusion sites appear to be independent of both high-affinity cytoplasmic and low-affinity lumenal sites, supporting a multisite 'in line' sequential binding mechanism for Ca(2+) transport.     

Covalent labeling of the cytoplasmic or luminal domains of the sarcoplasmic reticulum Ca(2+)-ATPase with fluorescent azido dyes.

Sarcoplasmic reticulum (SR) vesicles were incubated with azido derivatives of Cascade blue (ACB), Lucifer yellow (ALY), 2,7-naphthalene-disulfonic acid (ANDS), and fluorescein (AF) for 0.1-24 h at 2 degrees C. All four dyes gave intense reaction with the cytoplasmic domain of the Ca(2+)-ATPase on photoactivation after brief incubation. The penetration of the dyes into the luminal space of the SR was determined after centrifugation through Sephadex microcolumns to remove the external dye, followed by photolabeling and gel electrophoresis of the photolabeled proteins. The reaction of ACB and ANDS with the Ca(2+)-ATPase and with calsequestrin increased progressively during incubation up to 24 h indicating their slow accumulation in the luminal space, while ALY and AF did not show significant penetration into the vesicles. The distribution of the covalently attached ACB in the Ca(2+)-ATPase was tested by tryptic proteolysis after labeling exclusively from the outside (OS), from the inside (IS) or from both sides (BS). In all cases intense ACB fluorescence was seen in the A fragment with inhibition of ATPase activity. In the OS preparations the A1, while in IS the A2 fragment was more intensely labeled. There was no significant incorporation of ACB into the region of B fragment identified by FITC fluorescence. The crystallization of the Ca(2+)-ATPase by EGTA + decavanadate was completely inhibited in the BS samples after labeling either in the Ca2E1 or E2V conformation. There was no inhibition of crystallization in the OS preparations. In the IS preparations labeled in the Ca2E1 state the crystallization was impaired, while in the E2V state there was only slight disorganization of the crystals. The total amount of ACB photoincorporated into SR proteins after incubation for 24 h was 1.75 nmol/mg protein; 2/3 of this labeling occurred from the outside and 1/3 from the inside. Similar level of labeling was obtained in media that stabilize the E1 or the E2 conformation of the Ca(2+)-ATPase.     

Fluoroaluminate complexes are bifunctional analogues of phosphate in sarcoplasmic reticulum Ca(2+)-ATPase.

The mechanism of inhibition of the sarcoplamc reticulum (SR) Ca(2+)-ATPase by the fluoroaluminate complexes was investigated. First, AlF4- was shown to bind to the Ca(2+)-free conformation of the enzyme by a slow quasi-irreversible process. The rate constants of the reaction are k+ = 16 x 10(3) M-1 s-1 and k- < 1.5 10(-3) s-1. We directly measured a stoichiometry of about 4.8 nmol of AlF4- bound/mg of protein. Mg2+ was a necessary cofactor for the reaction with a dissociation constant of 3 mM. It was demonstrated (Dupont, Y., and Pougeois, R. (1983) FEBS Lett. 156, 93-98) that phosphorylation by P(i) induced a dehydration of the catalytic site. The same process has been shown here to occur upon AlF4- binding either by the use of Me2SO or by demonstration of an increase of bound 2',3'-O-(2,4,6-trinitrocyclohexadienyldene)adenosine triphosphate fluorescence. Phosphorylation by P(i) is inhibited by the binding of AlF4-. Second, a fluoroaluminate complex, presumably AlF4-, was also shown to bind to the Ca(2+)-bound conformation of the Ca(2+)-ATPase in the presence of ADP and stabilize a E1.Ca2.ADP.AlFx complex. The dissociation constant of the nucleotidic site for ADP was shifted to the micromolar range. The Ca2+ ions bound on the external high affinity sites became occluded upon binding of (ADP + AlFx). We propose that AlF4- mimics P(i) binding to the Ca(2+)-free conformation of the ATPase and stabilizes an intermediate similar to the acyl-phosphate derivative; it also acts as an analogue of the gamma-phosphate of ATP and stabilizes an E1.[Ca2].ADP.AlF4 complex where the Ca2+ ions are occluded.     

Vectorially oriented monolayers of detergent-solubilized Ca(2+) -ATPase from sarcoplasmic reticulum.

A method for tethering proteins to solid surfaces has been utilized to form vectorially oriented monolayers of the detergent-solubilized integral membrane protein Ca(2+) -ATPase from the sarcoplasmic reticulum (SR). Bifunctional, organic self-assembled monolayers (SAMs) possessing "headgroup" binding specificity for the substrate and "endgroup" binding specificity for the enzyme were utilized to tether the enzyme to the substrate. Specifically, an amine-terminated 11-siloxyundecaneamine SAM was found to bind the Ca(2+)-ATPase primarily electrostatically. The Ca(2+)-ATPase was labeled with the fluorescent probe 5-(2-[(iodoacetyl)amino]ethyl)aminonaphthalene-1-sulfonic acid before monolayer formation. Consequently, fluorescence measurements performed on amine-terminated SAM/enzyme monolayers formed on quartz substrates served to establish the nature of protein binding. Formation of the monolayers on inorganic multilayer substrates fabricated by molecular beam epitaxy made it possible to use x-ray interferometry to determine the profile structure for the system, which was proved correct by x-ray holography. The profile structures established the vectorial orientation of the Ca(2+)-ATPase within these monolayers, to a spatial resolution of approximately 12 A. Such vectorially oriented monolayers of detergent-solubilized Ca(2+)-ATPase from SR make possible a wide variety of correlative structure/function studies, which would serve to elucidate the mechanism of Ca(2+) transport by this enzyme.     

Serine phosphorylation of the sarcoplasmic reticulum Ca(2+)-ATPase in the intact beating rabbit heart.

Recent studies have demonstrated that Ca(2+)/calmodulin-dependent protein kinase phosphorylates the Ca(2+)-pumping ATPase of cardiac sarcoplasmic reticulum (SR) in vitro. Also, evidence from in vitro studies suggested that this phosphorylation, occurring at Ser(38), results in stimulation of Ca(2+) transport. In the present study, we investigated whether serine phosphorylation of the SR Ca(2+)-ATPase occurs in the intact functioning heart. Hearts removed from anesthetized rabbits were subjected to retrograde aortic perfusion of the coronary arteries with oxygenated mammalian Ringer solution containing (32)P(i) and contractions were monitored by recording systolic left ventricular pressure development. Following 45-50 min of (32)P perfusion, the hearts were freeze-clamped, SR isolated, and analyzed for protein phosphorylation. SDS-polyacrylamide gel electrophoresis and autoradiography showed phosphorylation of several peptides including the Ca(2+)-ATPase and Ca(2+) release channel (ryanodine receptor). The identity of Ca(2+)-ATPase as a phosphorylated substrate was confirmed by Western immunoblotting as well as immunoprecipitation using a cardiac SR Ca(2+)-ATPase-specific monoclonal antibody. The Ca(2+)-ATPase showed immunoreactivity with a phosphoserine monoclonal antibody indicating that the in situ phosphorylation occurred at the serine residue. Quantification of Ca(2+)-ATPase phosphorylation in situ yielded a value of 208 +/- 12 pmol (32)P/mg SR protein which corresponded to the phosphorylation of approximately 20% of the Ca(2+) pump units in the SR membrane. Since this phosphorylation occurred under basal conditions (i.e., in the absence of any inotropic intervention), a considerable steady-state pool of serine-phosphorylated Ca(2+)-ATPase likely exists in the normally beating heart. These findings demonstrate that serine phosphorylation of the Ca(2+)-ATPase is a physiological event which may be important in the regulation of SR function.     

Characterization of phenylmaleimide inhibition of the Ca(2+)-ATPase from skeletal-muscle sarcoplasmic reticulum

The Ca(2+)-ATPase from sarcoplasmic reticulum reacts with phenylmaleimide, producing the inhibition of the ATPase activity following a pseudo-first-order kinetic with a rate constant of 19 M(-1) s(-1). Calcium and ATP binding are not altered upon phenylmaleimide inhibition. However, the presence of millimolar calcium, and to a lesser extent magnesium, in the inhibition medium enhances the effect of phenylmaleimide, causing a higher degree of inhibition. Solubilization with C(12)E(8) does not affect the ATPase inhibition, excluding any kind of participation of the lipid bilayer. Phosphorylation with ATP in steady-state conditions as well as phosphorylation with inorganic phosphate in equilibrium conditions were strongly inhibited. Conversely, we have found that the occupancy of the phosphorylation site by ortovanadate fully protects against the inhibitory effect of phenylmaleimide, indicating a conformational transition associated with the phosphorylation reaction.     

Variation of scallop sarcoplasmic reticulum Ca2+-ATPase activity with temperature

Methods for preparing native scallop sarcoplasmic reticulum vesicles, largely purified membranous scallop sarcoplasmic reticulum Ca2+-ATPase, and nonionic detergent-solubilized sarcoplasmic reticulum Ca2+-ATPase are described. The effect of a range of polyoxyethylene-based detergents on the solubilized Ca2+-ATPase was tested. Decaethylene glycol dodecyl ether (C12E10) supported the highest levels of activity, although C12E8 and C12E9 were more routinely used. Arrhenius plots of Ca2+-ATPase activity, where the assays were carried out with the same pH at all temperatures (7.4), showed a region of nonlinearity at 10 degrees C. A very similar plot was obtained when no compensation was made for pH variation with temperature. Both the break in the Arrhenius plot and the activation energies for the scallop sarcoplasmic reticulum above and below the break were very similar to those found for lobster sarcoplasmic reticulum (Madeira, V. M. C., Antunes-Madeira, M. C., and Carvalho, A. R. (1974) Biochem. Biophys. Res. Commun. 65, 997-1003). The Arrhenius plot of the scallop Ca2+-ATPase in C12E8 no longer showed the nonlinearity at 10-12 degrees C seen with the native sarcoplasmic reticulum, but instead a break now appeared at 20-21 degrees C. This is close to the Arrhenius break temperature of rabbit Ca2+-ATPase in C12E8 and of a perturbation in C12E8 (Dean, W. L. (1982) Biophys. J. 37, 56-57).     

Thermal instability of rat muscle sarcoplasmic reticulum Ca(2+)-ATPase function

To examine the thermal instability and the role of sulfhydryl (SH) oxidation on sarcoplasmic reticulum (SR) Ca(2+)-ATPase function, crude homogenates were prepared from the white portion of the gastrocnemius (WG) adult rat muscles (n = 9) and incubated in vitro for < or =60 min either at a normal resting body temperature (37 degrees C) or at a temperature indicative of exercise-induced hyperthermia (41 degrees C) with DTT and without DTT (CON). In general, treatment with DTT resulted in higher Ca(2+)-ATPase and Ca(2+) uptake values (nmol. mg protein(-1). min(-1), P < 0.05), an effect that was not specific to time of incubation. Incubations at 41 degrees C resulted in lower (P < 0.05) Ca(2+) uptake rates (156 +/- 18 and 35.9 +/- 3.3) compared with 37 degrees C (570 +/- 54 and 364 +/- 26) at 30 and 60 min, respectively. At 37 degrees C, ryanodine (300 microM), which was used to block Ca(2+) release from the calcium release channel, prevented the time-dependent decrease in Ca(2+) uptake. A general inactivation (P < 0.05) of maximal Ca(2+)-ATPase activity (V(max)) in CON was observed with incubation time (0 > 30 > 60 min), with the effect being more pronounced (P < 0.05) at 41 degrees C compared with 37 degrees C. The Hill slope, a measure of co-operativity, and the pCa(50), the cytosolic Ca(2+) concentration required for half-maximal activation of Ca(2+)-ATPase activity, decreased (P < 0.05) at 41 degrees C only. Treatment with DTT attenuated the alterations in enzyme kinetics. The increase in V(max) with the Ca(2+) ionophore A-23187 was less pronounced at 41 degrees C compared with 37 degrees C. It is concluded that exposure of homogenates to a temperature typically experienced in exercise results in a reduction in the coupling ratio, which is mediated primarily by lower Ca(2+) uptake and occurs as a result of increases in membrane permeability to Ca(2+). Moreover, the decreases in Ca(2+)-ATPase kinetics in WG with sustained heat stress result from SH oxidation.     

Lamellar stacking in three-dimensional crystals of Ca(2+)-ATPase from sarcoplasmic reticulum.

Electron microscopy of multilamellar crystals of CA(2+)-ATPase currently offers the best opportunity for obtaining a high-resolution structure of this ATP-driven ion pump. Under certain conditions small, wormlike crystals are formed and provide views parallel to the lamellar plane, from which parameters of lamellar stacking can be directly measured. Assuming that molecular packing is the same, data from these views could supplement those obtained by tilting large, thin platelike crystals. However, we were surprised to discover that the lamellar spacing was variable and depended on the amount of glycerol present during crystallization (20% versus 5%). Projection maps (h,0,l) from these womklike crystals suggest different molecular contacts that give rise to the different lamellar spacings. Based on an orthogonal projection map (h,k,0) from collapsed, wormlike crystals and on x-ray powder patterns, we conclude that molecular packing within the lamellar plane is the same as that in thin, platelike crystals and is unaffected by glycerol. Finally, the orientation of molecules in the lamellar plane was characterized from freeze-dried, shadowed crystals. Comparing the profile of molecules in these multilamellar crystals with that previously observed in helical tubes induced by vanadate gives structural evidence of the conformational change that accompanies binding of calcium of Ca(2+)-ATPase.     

Interactions of hexachlorocyclohexanes with the (Ca2+ + Mg2+)-ATPase from sarcoplasmic reticulum

Hexachlorocyclohexanes have been shown to inhibit the (Ca2+ + Mg2+)-ATPase of muscle sarcoplasmic reticulum reconstituted into bilayers of dioleoylphosphatidylcholine. However, for the ATPase reconstituted into bilayers of dimyristoleoylphosphatidylcholine, a pattern of activation at low concentration followed by inhibition at higher concentration is seen for hexachlorocyclohexanes and alkanes such as decane and hexadecane. The ATPase in sarcoplasmic reticulum vesicles is also inhibited by the hexachlorocyclohexanes. The effects of hexachlorocyclohexanes on activity are largely independent of concentrations of Ca2+ and ATP. Inhibition is more marked at lower temperatures. The hexachlorocyclohexanes quench the tryptophan fluorescence of the ATPase, and the quenching can be used to obtain partition coefficients into the membrane system. As for simple lipid bilayers, partition exhibits a negative temperature coefficient. Binding is related to effects on ATPase activity.     

Regulation of cardiac sarcoplasmic reticulum (Ca2+ + Mg2+)-ATPase

Summary The two high affinity calcium binding sites of the cardiac (Ca²⁺ + Mg²⁺)-ATPase have been identified with the use of Eu³⁺. Eu³⁺ competes for the two high affinity calcium sites on the enzyme. With the use of laser-pulsed fluorescent spectroscopy, the environment of the two sites appear to be heterogeneous and contain different numbers of H₂O molecules coordinated to the ion. The ion appears to be occluded even further in the presence of ATP. Using non-radiative energy transfer studies, we were able to estimate the distance between the two Ca²⁺ sites to be between 9.4 to 10.2 A in the presence of ATP. Finally, from the assumption that the calcium site must contain four carboxylic side chains to provide the 6–8 ligands needed to coordinate calcium, and based on our recently published data, we predict the peptidic backbone of the two sites.     

Evidence for the cytoplasmic location of the N- and C-terminal segments of sarcoplasmic reticulum (Ca2+-Mg2+)-ATPase

Antibodies were produced against 5 peptides corresponding to segments of the (Ca2+-Mg2+)-ATPase of fast-twitch rabbit skeletal muscle sarcoplasmic reticulum (SR) including the N- and C-terminal regions. With the exception of antibodies directed against the peptide corresponding to residues 567-582 all antibodies bound strongly to the ATPase in intact SR vesicles, indicating that the epitopes were located on the cytoplasmic face of the SR. When the vesicles were disrupted, by solubilisation in SDS, binding of these antibodies was unchanged, further supporting the idea that these epitopes were located on the cytoplasmic face of SR. This is the first demonstration of the location of the N- and C-terminal regions of SR (Ca2+-Mg2+)-ATPase. These observations are discussed in the light of current structural models of the ATPase.     

Mapping epitopes on the (Ca(2+)-Mg2+)-ATPase of sarcoplasmic reticulum using fusion proteins.

Epitopes for a number of monoclonal antibodies (mAbs) binding (Ca(2+)-Mg2+)-ATPase purified from skeletal muscle sarcoplasmic reticulum have been defined by studying binding to fusion proteins generated from cDNA fragment libraries. Comparison of these results with those of previous studies of binding of mAbs to proteolytic fragments of the ATPase have allowed the definition of the epitopes to within approx. 100 residues and for one (mAb 1/2H7) to within 45 residues. The experiments suggest considerable exposure of the nucleotide binding domain of the ATPase on the top surface of the protein. Those mAbs that were found to inhibit steady-state ATPase activity were found to bind to epitopes in the nucleotide binding domain of the ATPase.     

Structural basis of ion pumping by Ca(2+)-ATPase of sarcoplasmic reticulum

The structures of the Ca(2+)-ATPase (SERCA1a) have been determined for five different states by X-ray crystallography. Detailed comparison of the structures in the Ca(2+)-bound form and unbound (but thapsigargin-bound) form reveals that very large rearrangements of the transmembrane helices take place accompanying Ca2+ dissociation and binding and that they are mechanically linked with equally large movements of the cytoplasmic domains. The meanings of the rearrangements of the transmembrane helices and those of the cytoplasmic domains, and the mechanistic roles of the phosphorylation are now becoming clear.     

CrATP-induced Ca2+ occlusion in mutants of the Ca(2+)-ATPase of sarcoplasmic reticulum.

Use of the nonphosphorylating beta,gamma-bidentate chromium(III) complex of ATP to induce a stable Ca(2+)-occluded form of the sarcoplasmic reticulum Ca(2+)-ATPase was combined with molecular sieve high performance liquid chromatography of detergent-solubilized protein to examine the ability of the Ca(2+)-ATPase mutants Gly-233-->Glu, Gly-233-->Val, Glu-309-->Gln, Gly-310-->Pro, Pro-312-->Ala, Ile-315-->Arg, Leu-319-->Arg, Asp-703-->Ala, Gly-770-->Ala, Glu-771-->Gln, Asp-800-->Asn, and Gly-801-->Val to occlude Ca2+. This provided a new approach to identification of amino acid residues involved in Ca2+ binding and in the closure of the gates to the Ca2+ binding pocket of the Ca(2+)-ATPase. The "phosphorylation-negative" mutant Asp-703-->Ala and mutants of ADP-sensitive phosphoenzyme intermediate type were fully capable of occluding Ca2+, as was the mutant Gly-770-->Ala. Mutants in which carboxylic acid-containing residues in the putative transmembrane segments had been substituted ("Ca(2+)-site mutants") and mutant Gly-801-->Val were unable to occlude either of the two calcium ions. In addition, the mutant Gly-310-->Pro, previously classified as ADP-insensitive phosphoenzyme intermediate type (Andersen, J.P., Vilsen, B., and MacLennan, D.H. (1992). J. Biol. Chem. 267, 2767-2774), was unable to occlude Ca2+, even though Ca(2+)-activated phosphorylation from MgATP took place in this mutant.     

Arsenate-induced fluorescence changes in the Ca(2+)-ATPase of sarcoplasmic reticulum membranes.

Arsenate, an analogue of inorganic phosphate, causes an increase in the intrinsic fluorescence of the Ca(2+)-ATPase of sarcoplasmic reticulum membranes. This increase in fluorescence is observed regardless of whether Ca(2+)-loaded or leaky vesicles are assayed. The maximal fluorescence change (2-3%) is observed at pH 6.0 in the presence of Mg2+ and is abolished by the addition of micromolar Ca2+ concentrations. Dimethyl sulfoxide (20% v/v) increases the enzyme's affinity for arsenate one order of magnitude. It is concluded that arsenate, after binding, promotes the same conformational change of the enzyme as that produced by Pi.     

Quercetin interaction with the (Ca2+ + Mg2+)-ATPase of sarcoplasmic reticulum

In sarcoplasmic reticulum vesicles or in the (Ca2+ + Mg2+)-ATPase purified from sarcoplasmic reticulum, quercetin inhibited ATP hydrolysis, Ca2+ uptake, ATP-Pi exchange, ATP synthesis coupled to Ca2+ efflux, ATP-ADP exchange, and steady state phosphorylation of the ATPase by inorganic phosphate. Steady state phosphorylation of the ATPase by ATP was not inhibited. Quercetin also inhibited ATP and ADP binding but not the binding of Ca2+. The inhibition of ATP-dependent Ca2+ transport by quercetin was reversible, and ATP, Ca2+, and dithiothreitol did not affect the inhibitory action of quercetin.