The Average Conformation at Micromolar [Ca2+] of Ca2+-ATPase with Bound Nucleotide Differs from That Adopted with the Transition State Analog ADP·AlFx or with AMPPCP under Crystallization Conditions at Millimolar [Ca2+]
2005; Elsevier BV; Volume: 280; Issue: 19 Linguagem: Inglês
10.1074/jbc.m501596200
ISSN1083-351X
AutoresMartin Picard, Chikashi Toyoshima, Philippe Champeil,
Tópico(s)ATP Synthase and ATPases Research
ResumoCrystalline forms of detergent-solubilized sarcoplasmic reticulum Ca2+-ATPase, obtained in the presence of either a substrate analog, AMPPCP, or a transition state complex, ADP·fluoroaluminate, were recently described to share the same general architecture despite the fact that, when studied in a test tube, these forms show different functional properties. Here, we show that the differences in the properties of the E1·AMPPCP and the E1·ADP·AlFx membraneous (or solubilized) forms are much less pronounced when these properties are examined in the presence of 10 mm Ca2+ (the concentration prevailing in the crystallization media) than when they are examined in the presence of the few micromolar of Ca2+ known to be sufficient to saturate the transport sites. This concerns various properties, including ATPase susceptibility to proteolytic cleavage by proteinase K, ATPase reactivity toward SH-directed Ellman's reagent, ATPase intrinsic fluorescence properties (here described for the E1·ADP·AlFx complex for the first time), and also the rates of 45Ca2+-40Ca2+ exchange at site "II." These results solve the above paradox at least partially and suggest that the presence of a previously unrecognized Ca2+ ion in the E1·AMPPCP crystals should be re-investigated. A contrario, they emphasize the fact that the average conformation of the E1·AMPPCP complex under usual conditions in the test tube differs from that found in the crystalline form. The extended conformation of nucleotide revealed by the E1·AMPPCP crystalline form might be only indicative of the requirements for further processing of the complex, toward the transition state leading to phosphorylation and Ca2+ occlusion. Crystalline forms of detergent-solubilized sarcoplasmic reticulum Ca2+-ATPase, obtained in the presence of either a substrate analog, AMPPCP, or a transition state complex, ADP·fluoroaluminate, were recently described to share the same general architecture despite the fact that, when studied in a test tube, these forms show different functional properties. Here, we show that the differences in the properties of the E1·AMPPCP and the E1·ADP·AlFx membraneous (or solubilized) forms are much less pronounced when these properties are examined in the presence of 10 mm Ca2+ (the concentration prevailing in the crystallization media) than when they are examined in the presence of the few micromolar of Ca2+ known to be sufficient to saturate the transport sites. This concerns various properties, including ATPase susceptibility to proteolytic cleavage by proteinase K, ATPase reactivity toward SH-directed Ellman's reagent, ATPase intrinsic fluorescence properties (here described for the E1·ADP·AlFx complex for the first time), and also the rates of 45Ca2+-40Ca2+ exchange at site "II." These results solve the above paradox at least partially and suggest that the presence of a previously unrecognized Ca2+ ion in the E1·AMPPCP crystals should be re-investigated. A contrario, they emphasize the fact that the average conformation of the E1·AMPPCP complex under usual conditions in the test tube differs from that found in the crystalline form. The extended conformation of nucleotide revealed by the E1·AMPPCP crystalline form might be only indicative of the requirements for further processing of the complex, toward the transition state leading to phosphorylation and Ca2+ occlusion. After the initial description of the high resolution structures of two crystalline forms of the sarcoplasmic reticulum calcium pump (the membranous Ca2+-dependent P-type ATPase SERCA1a) (1Toyoshima C. Nakasako M. Nomura H. Ogawa H. Nature. 2000; 405: 633-634Crossref PubMed Scopus (1613) Google Scholar, 2Toyoshima C. Nomura H. Nature. 2002; 418: 605-611Crossref PubMed Scopus (808) Google Scholar), additional forms of this Ca2+-ATPase were recently crystallized, with the hope of characterizing as many as possible of the different intermediates formed in sequence during the catalytic cycle of this enzyme and therefore to provide a structural basis for the mechanistic description of ion pumping (3Toyoshima C. Nomura H. Sugita Y. FEBS Lett. 2003; 555: 106-110Crossref PubMed Scopus (58) Google Scholar, 4Sørensen T. L.-M. Møller J.V. Nissen P. Science. 2004; 304: 1672-1675Crossref PubMed Scopus (372) Google Scholar, 5Toyoshima C. Mizutani T. Nature. 2004; 430: 529-535Crossref PubMed Scopus (382) Google Scholar, 6Toyoshima C. Nomura H. Tsuda T. Nature. 2004; 442: 361-368Crossref Scopus (382) Google Scholar, 7Olesen C. Sørensen T. L.-M. Nielsen R.C. Møller J.V. Nissen P. Science. 2005; 306: 2251-2255Crossref Scopus (235) Google Scholar). Among these forms, one has its two transport sites occupied by Ca2+, and its nucleotide binding site occupied by a non-hydrolyzable analog of ATP, AMPPCP 1The abbreviations used are: AMPPCP, adenosine 5′-(β,γ-methylene)triphosphate; SR, sarcoplasmic reticulum; ATPase, adenosine triphosphatase; quin2, 2-[(2-amino-5-methylphenoxy)methyl]-6-methoxy-8-aminoquinoline-N,N,N′,N′-tetraacetic acid; Mes, 2-(N-morpholino)ethanesulfonic acid; Mops, 4-morpholinepropanesulfonic acid; DTNB, Ellman's reagent, 5,5′-dithio-bis(2-nitrobenzoic acid); PK, proteinase K; C12E8, octaethyleneglycol monododecyl ether; DDM, dodecylmaltoside; ATPγS, (adenosine 5′-[γ-thio]triphosphate); Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine.; it is referred to as "E1·AMPPCP." Another one, referred to as "E1·AlFx·ADP," also has its two transport sites occupied by Ca2+, and it has been obtained by crystallization of the quasi-irreversible complex formed by ATPase with ADP and aluminum fluoride, a complex thought to be a fair analog of the transient species formed in the catalytic cycle immediately before the ADP-sensitive "E1P" phosphoenzyme. The overall polypeptide chain architecture was found to be very similar in these two crystalline forms (except for specific features at the catalytic site): in particular, in both forms (3Toyoshima C. Nomura H. Sugita Y. FEBS Lett. 2003; 555: 106-110Crossref PubMed Scopus (58) Google Scholar, 4Sørensen T. L.-M. Møller J.V. Nissen P. Science. 2004; 304: 1672-1675Crossref PubMed Scopus (372) Google Scholar, 5Toyoshima C. Mizutani T. Nature. 2004; 430: 529-535Crossref PubMed Scopus (382) Google Scholar, 6Toyoshima C. Nomura H. Tsuda T. Nature. 2004; 442: 361-368Crossref Scopus (382) Google Scholar), the M1–M2 transmembrane helices appear to be pulled up toward the cytosol and the top portion (M1′) of the M1 helix gets kinked, thereby locking the conformation of Glu309, a residue thought to cap Ca2+ at one of its binding sites, "site II," and therefore to be critical for dissociation of the two bound Ca2+ ions out of their binding pocket (e.g. Refs. 8Vilsen B. Andersen J.P. Biochemistry. 1998; 37: 10961-10971Crossref PubMed Scopus (80) Google Scholar and 9Inesi G. Ma H. Lewis D. Xu C. J. Biol. Chem. 2004; 279: 31629-31637Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar). This locking by M1–M1′ of the conformation of Glu309 was considered to be responsible for the long known "occlusion" of Ca2+ (10Verjovski-Almeida S. Kurzmack M. Inesi G. Biochemistry. 1978; 17: 5006-5013Crossref PubMed Scopus (96) Google Scholar) that occurs during normal turnover, after ATPase phosphorylation from ATP. In the (stable) E1·AlFx·ADP form, it is indeed accepted that the Ca2+ transport sites are occluded (4Sørensen T. L.-M. Møller J.V. Nissen P. Science. 2004; 304: 1672-1675Crossref PubMed Scopus (372) Google Scholar, 9Inesi G. Ma H. Lewis D. Xu C. J. Biol. Chem. 2004; 279: 31629-31637Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar, 11Troullier A. Girardet J.-L. Dupont Y. J. Biol. Chem. 1991; 267: 22821-22829Abstract Full Text PDF Google Scholar), as in the (transient) E1P phosphoenzyme form. However, it has been suggested previously that a similar occlusion occurs neither after the mere formation of a noncovalent E1·AMPPCP complex, nor after the mere formation of the E1·Mg·ATP complex that immediately precedes phosphorylation during the normal cycle (4Sørensen T. L.-M. Møller J.V. Nissen P. Science. 2004; 304: 1672-1675Crossref PubMed Scopus (372) Google Scholar, 9Inesi G. Ma H. Lewis D. Xu C. J. Biol. Chem. 2004; 279: 31629-31637Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar, 12Orlowski S. Champeil P. Biochemistry. 1991; 30: 352-361Crossref PubMed Scopus (77) Google Scholar, 13Petithory J.R. Jencks W.P. Biochemistry. 1986; 25: 4493-4497Crossref PubMed Scopus (54) Google Scholar). This apparent discrepancy between the different properties of the two ATPase forms and their similar crystalline structure has already been noted (4Sørensen T. L.-M. Møller J.V. Nissen P. Science. 2004; 304: 1672-1675Crossref PubMed Scopus (372) Google Scholar, 5Toyoshima C. Mizutani T. Nature. 2004; 430: 529-535Crossref PubMed Scopus (382) Google Scholar, 9Inesi G. Ma H. Lewis D. Xu C. J. Biol. Chem. 2004; 279: 31629-31637Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar), and various interpretations have been given, together with somewhat contradictory comments that these two forms, E1·AMPPCP and E1·AlFx·ADP, have (5Toyoshima C. Mizutani T. Nature. 2004; 430: 529-535Crossref PubMed Scopus (382) Google Scholar) or do not have (4Sørensen T. L.-M. Møller J.V. Nissen P. Science. 2004; 304: 1672-1675Crossref PubMed Scopus (372) Google Scholar) a similar pattern of resistance to cleavage by proteinase K of their cytosolic domains. Structural fluctuation of the non-crystallized ATPase·AMPPCP complex (we will discuss it) and/or selection of a particular conformation by crystal packing were suggested to explain the different occlusion properties of the two ATPase forms in a test tube (4Sørensen T. L.-M. Møller J.V. Nissen P. Science. 2004; 304: 1672-1675Crossref PubMed Scopus (372) Google Scholar, 5Toyoshima C. Mizutani T. Nature. 2004; 430: 529-535Crossref PubMed Scopus (382) Google Scholar). In view of the significance of this issue with respect to the mechanism of occlusion, we decided to further document the resemblance or differences between E1·AMPPCP and E1·AlFx· ADP forms. In particular, we asked whether any clues could be provided by the fact that the published crystalline forms of E1·AMPPCP had been obtained at very high Ca2+ concentrations. Under high millimolar Ca2+ conditions, we found that the ATPase complex with AMPPCP indeed has properties closer to those of the ATPase complex with ADP·fluoroaluminate than under micromolar Ca2+ conditions, from the point of view of various indexes, including exchange of calcium at site II (hence its dissociation from this site), resistance to proteolysis, susceptibility to SH modification, or Trp fluorescence. This was especially true when Ca2+-ATPase was solubilized by detergent. For comparison, we also briefly documented the (different) influence of Mg2+ on the effect of AMPPCP and other nucleotides. We discuss various possibilities possibly explaining why a high Ca2+ could favor resemblance between the E1·AMPPCP and E1·AlFx·ADP forms. We also discuss various possibilities to explain, a contrario, the differences between the fluoroaluminate complex and the AMPPCP complex under more usual conditions: among these, the average conformation of the E1·AMPPCP complex under such conditions might well be different from the one found in the crystal, a fact that should not be overlooked in future descriptions of ATP binding to Ca2+-ATPase during the normal cycle. 45Ca2+ dissociation from sarcoplasmic reticulum Ca2+-ATPase or its exchange with 40Ca2+ was measured with a rapid filtration Bio-Logic device (12Orlowski S. Champeil P. Biochemistry. 1991; 30: 352-361Crossref PubMed Scopus (77) Google Scholar). For these measurements, ATPase-containing SR membranes were adsorbed onto Millipore nitrocellulose (HA) filters (0.45-μm pore diameter). The SR membranes had previously been equilibrated for 5–15 min with either 25 or 50 μm45Ca2+ (plus 50 μm [3H]glucose as a volume marker), in the absence or presence of various additives. The adsorbed membranes were perfused for various periods of time, at flow rates ranging from 0.5 to 4.5 ml/s depending on the perfusion period. In addition to the above buffer, the perfusion fluid also contained either EGTA or various concentrations of unlabeled Ca2+ (40Ca2+), again in the absence or presence of various additives. Radioactivity in the perfused filter (or in a non-perfused filter, for time zero control) was finally counted, without rinsing. The total 45Ca2+ found was corrected for the 45Ca2+ content in the wet volume by taking into account the 3H radioactivity present in the same filter (this 45Ca2+ or 3H content in the wet volume was of course very small after the filter had been perfused). Ca2+ dissociation from sarcoplasmic reticulum Ca2+-ATPase was also measured in stopped-flow experiments using quin2, a fluorescent Ca2+-sensitive dye whose strong complex with Ca2+ has a fluorescence level different from that of the Ca2+-free dye (higher or lower depending on the excitation wavelength; see Refs. 14Bayley P. Ahlström P. Martin S.R. Forsen S. Biochem. Biophys. Res. Commun. 1984; 120: 185-191Crossref PubMed Scopus (99) Google Scholar, 15Wakabayashi S. Shigekawa M. Biochemistry. 1990; 29: 7309-7318Crossref PubMed Scopus (38) Google Scholar, 16Champeil P. Combettes L. Berthon B. Doucet E. Orlowski S. Claret M. J. Biol. Chem. 1989; 264: 17665-17673Abstract Full Text PDF PubMed Google Scholar, 17Champeil P. Henao F. de Foresta B. Biochemistry. 1997; 36: 12383-12393Crossref PubMed Scopus (18) Google Scholar). We were initially concerned by the possibility that the presence of nucleotide could interfere with this measurement of quin2 fluorescence, if dissociation of a preformed Ca2+·nucleotide complex were to superimpose with dissociation of Ca2+ from Ca2+-ATPase. However, this does not occur, based on several facts: (i) the amplitude of the quin2 fluorescence change is essentially similar with or without AMPPCP; (ii) observed rate constants in the presence of AMPPCP are essentially similar when AMPPCP is initially present in both syringes and when AMPPCP is initially only present together with quin2 (in the latter case, the Ca·AMPPCP complex is not formed); (iii) we directly checked, in the absence of membranes, that any Ca·AMPPCP or Ca·ATP complex formed under our conditions dissociates faster than the dead time (3 ms) of our stopped-flow measurements. All these experiments were performed at 20 °C. Most of them (unless otherwise noted) were performed in a 50 mm Mes-Tris buffer at pH 6, generally without potassium, or in a 50 mm Mops-Tris buffer at pH 7, supplemented with 100 mm KCl and 5 mm Mg2+. AMPPCP and ATPγS were from Sigma, ADP was from Fluka, and quin2 was from Calbiochem. Equilibrium fluorescence experiments were performed with a SPEX Fluorolog fluorometer, with constant stirring of the temperature-controlled cell. Stopped-flow experiments were performed with Bio-Logic SFM3 equipment; quin2-emitted fluorescence was detected with a broad band filter, DA 531 from MTO (Massy, France). The extent of Ca2+-ATPase proteolysis by proteinase K under different conditions was examined by SDS-PAGE in 12% acrylamide gels prepared in the presence of 1 mm Ca2+, as described in Ref. 18Møller J.V. Lenoir G. Marchand C. Montigny C. le Maire M. Toyoshima C. Juul B.S. Champeil P. J. Biol. Chem. 2002; 277: 38647-38659Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar. SH reaction with DTNB (Ellman's reagent, from Sigma) was monitored at 25 °C, using a diode array spectrophotometer (HP 8453) operated in the kinetic mode. DTNB conversion to a colored anion upon reaction with SH groups was monitored at 430 nm, using an extinction coefficient of 12 mm–1·cm–1, slightly lower than that at the wavelength for maximal absorption (13.6 mm–1·cm–1 at 412 nm); 430 nm, instead of the more widely used wavelength of 412 nm, was chosen to minimize the slight interference with DTNB absorption which occurs at 412 nm. Absorption by DTNB is not visible in Fig. 4 because DTNB was already present in the buffer used to record the blank spectrum. Note that, judging from the apparent affinities reported in Ref. 19Ogawa Y. Kurebayashi N. Harafuji H. J. Biochem. (Tokyo). 1986; 100: 1305-1318Crossref PubMed Scopus (17) Google Scholar for AMPPCP-cation complexes, the dissociation constants of AMPPCP complexes with Mg2+ and Ca2+ in 50 mm Mes-Tris at pH 6 probably are in the 0.4–0.5 mm and the 0.8–1 mm ranges, respectively (i.e. intermediate between those for ATP and those for ADP, as also shown by our own measurements with antipyrylazo III, data not shown). ADP·AlF4 Blocks 45Ca2+ Dissociation from SR Ca2+-ATPase, but under Ordinary Conditions, AMPPCP Does Not—Sørensen et al. (4Sørensen T. L.-M. Møller J.V. Nissen P. Science. 2004; 304: 1672-1675Crossref PubMed Scopus (372) Google Scholar) and Toyoshima and Mizutani (5Toyoshima C. Mizutani T. Nature. 2004; 430: 529-535Crossref PubMed Scopus (382) Google Scholar) found that the presence of AMPPCP or ADP·fluoroaluminate in the crystalline forms of Ca2+-ATPase resulted in structural changes indicative of occlusion of the bound Ca2+ ions in both cases, but they already noted that this was somewhat dissonant with a previous experimental finding deduced from rapid filtration experiments with 45Ca2+: the half-time for the dissociation of 45Ca2+ from the transport sites of non-phosphorylated ATPase only increases by 50% in the presence of 250 μm AMPPCP, at pH 6 in the presence of 20 mm Mg2+ (12Orlowski S. Champeil P. Biochemistry. 1991; 30: 352-361Crossref PubMed Scopus (77) Google Scholar), and thus dissociation remains relatively fast in the presence of AMPPCP. We first repeated that old experiment under different conditions, and at pH 7 in the presence of 5 mm Mg2+, found a similar although not identical result: AMPPCP again failed to block 45Ca2+ dissociation, but in these experiments did not even slow down the rate of this dissociation at all, even if its concentration was increased up to 2.5 mm. Yet, under the same conditions, preliminary incubation of Ca2+-ATPase with ADP and fluoroaluminate did block 45Ca2+ dissociation completely (see Fig. I in the Supplemental Material, and accompanying comments). Depending on Mg2+, AMPPCP May Either Stimulate or Slightly Reduce the Rate of Overall Ca2+ Dissociation from Ca2+-ATPase, and the Modulatory Effect of Mg2+ Differs for Various Nucleotides—We therefore re-investigated the effect of AMPPCP under the previous conditions at pH 6, using a stopped-flow assay in which Ca2+ dissociation from Ca2+-ATPase was triggered by mixing Ca2+-equilibrated SR vesicles with the fluorescent chelator quin2, whose fluorescence changes allowed us to monitor the rate of this Ca2+ dissociation (see "Materials and Methods" and Fig. II in Supplemental Material). With this new assay, we found that at pH 6 too, AMPPCP in the presence of a few millimeters of Mg2+ only minimally affected Ca2+ dissociation from ATPase (Fig. 1A), but we nevertheless confirmed that in the presence of 20 mm Mg2+ AMPPCP did slow down Ca2+ dissociation moderately, as previously observed in 45Ca2+ filtration experiments (12Orlowski S. Champeil P. Biochemistry. 1991; 30: 352-361Crossref PubMed Scopus (77) Google Scholar); at lower Mg2+ concentrations, however, this slowing down was in fact converted into a definite acceleration of the rate of Ca2+ dissociation, by a factor of up to 4 in the total absence of Mg2+ (representative traces are shown in Fig. III of Supplemental Material). This is summarized in panel B of Fig. 1, which includes experiments performed at high AMPPCP concentrations in the presence of 20 mm Mg2+. These data reveal no real occlusion of Ca2+ after the mere binding of AMPPCP. In the absence of Mg2+, the EC50 for AMPPCP-induced acceleration of Ca2+ dissociation was in the micromolar range (open circles in panel B; the true EC50 is somewhat lower than the one apparent in this panel, because of sub-stoichiometric conditions for the lowest concentration of AMPPCP tested). Incidentally, note that, although the quin2 fluorescence rise that we measured monitors the overall dissociation of the two Ca2+ ions previously bound to Ca2+-ATPase (see "Materials and Methods" and Fig. II in Supplemental Material), all traces were quasi-monophasic, as previously noted and interpreted within the context of a "flickering gate" for the Ca2+ binding pocket (12Orlowski S. Champeil P. Biochemistry. 1991; 30: 352-361Crossref PubMed Scopus (77) Google Scholar, 17Champeil P. Henao F. de Foresta B. Biochemistry. 1997; 36: 12383-12393Crossref PubMed Scopus (18) Google Scholar, 20Forbush B. J. Biol. Chem. 1987; 262: 11116-11127Abstract Full Text PDF PubMed Google Scholar). AMPPCP is not a perfect analog of ATP, because the Mg2+ dependence of its binding to Ca2+-ATPase is opposite to the Mg2+ dependence of the binding of ATP itself (19Ogawa Y. Kurebayashi N. Harafuji H. J. Biochem. (Tokyo). 1986; 100: 1305-1318Crossref PubMed Scopus (17) Google Scholar, 21Pang D.C. Briggs F.N. J. Biol. Chem. 1977; 252: 3262-3266Abstract Full Text PDF PubMed Google Scholar, 22Ross D.C. McIntosh D.B. J. Biol. Chem. 1987; 262: 12977-12983Abstract Full Text PDF PubMed Google Scholar; this will be confirmed below), presumably because binding in the presence of Mg2+ is perturbed by the CH2 link between β and γ phosphates in AMPPCP. Therefore, for comparison, we also measured the rate of Ca2+ dissociation from ATPase in the presence of another poorly-hydrolyzable analog of ATP, ATPγS, whose Mg2+ dependence has been shown to retain an ATP-like behavior, i.e. a higher affinity in the presence of Mg2+ than in its absence (23Forge V. Mintz E. Guillain F. J. Biol. Chem. 1993; 268: 10961-10968Abstract Full Text PDF PubMed Google Scholar). ATPγS was found to stimulate the rate of Ca2+ dissociation both in the absence and presence of Mg2+, and in this case Mg2+ enhanced the ATPγS-dependent acceleration instead of converting it into a slowing down (original traces in Fig. IV of Supplemental Material). Note that, for these experiments, ATPγS (at twice the final concentration) was purposely added together with quin2, and thus, SR vesicles were not preincubated with ATPγS: this is because ATPγS, although poorly hydrolyzed (24Yates D.W. Duance V.C. Biochem. J. 1978; 159: 719-728Crossref Scopus (29) Google Scholar), has been shown to react with Ca2+-ATPase in the presence of Ca2+, leading to accumulation of a thio-phosphorylated intermediate (25Yasuoka K. Kawakita M. Kaziro Y. J. Biochem. (Tokyo). 1982; 91: 1629-1637Crossref PubMed Scopus (15) Google Scholar). This alteration in our experimental protocol (compared with the above-described case of AMPPCP) was tested in the case of AMPPCP, and it had no significant influence on the rate constant for Ca2+ dissociation; note that nucleotide-dependent rates of Ca2+ binding have previously also been found similar when the nucleotide is added simultaneously with Ca2+ and when nucleotide is preincubated with Ca2+-free ATPase (26Mintz E. Mata A.M. Forge V. Passafiume M. Guillain F. J. Biol. Chem. 1995; 270: 27160-27164Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar), presumably because of the fairly fast rate of nucleotide binding at a concentration of 250 μm. Because of the puzzling difference between AMPPCP and ATPγS in their Mg2+ dependences, we performed a third series of experiments with ADP (again added together with quin2) in the presence of various concentrations of Mg2+:1mm ADP accelerated Ca2+ dissociation at all concentrations of Mg2+. Fig. 1C summarizes all these results. Note that experiments were also performed at pH 6 in the presence of 100 mm KCl and 5 mm Mg2+. In this series of experiments, 250 μm AMPPCP again hardly affected the rate of Ca2+ dissociation (from 14 to 15 s–1), whereas the same concentration of ADP accelerated it moderately (to 23 s–1) (data not shown). In previous experiments performed at pH 6.5 in the presence of 100 mm KCl and 2 mm Mg2+ at 11 °C, ADP had also been found to only have a weak effect on Ca2+ dissociation (15Wakabayashi S. Shigekawa M. Biochemistry. 1990; 29: 7309-7318Crossref PubMed Scopus (38) Google Scholar). This confirms that, although the presence of potassium slightly accelerates Ca2+ dissociation (e.g. Refs. 12Orlowski S. Champeil P. Biochemistry. 1991; 30: 352-361Crossref PubMed Scopus (77) Google Scholar and 17Champeil P. Henao F. de Foresta B. Biochemistry. 1997; 36: 12383-12393Crossref PubMed Scopus (18) Google Scholar), it does not affect the above-described pattern of nucleotide-induced alteration: the presence of nucleotide generally does not lead to Ca2+ occlusion. If Free Ca2+ Is Increased to High Millimolar Concentrations, Well Beyond Those Allowing Saturation of the ATPase High Affinity Sites, AMMPCP Binds to Ca2+-ATPase with Even Higher Affinity—As a preliminary step in our attempt to understand the effect of AMPPCP on Ca2+-ATPase under crystallization conditions, we estimated the affinity with which AMPPCP binds to Ca2+-ATPase under such conditions, namely in the presence of a high millimolar Ca2+ concentration. For this purpose, we used a Trp fluorescence assay (27Lacapère J.J. Bennett N. Dupont Y. Guillain F. J. Biol. Chem. 1990; 265: 348-353Abstract Full Text PDF PubMed Google Scholar), and first checked that it gave results consistent with the above-mentioned effect of Mg2+ on AMPPCP binding (19Ogawa Y. Kurebayashi N. Harafuji H. J. Biochem. (Tokyo). 1986; 100: 1305-1318Crossref PubMed Scopus (17) Google Scholar, 21Pang D.C. Briggs F.N. J. Biol. Chem. 1977; 252: 3262-3266Abstract Full Text PDF PubMed Google Scholar, 22Ross D.C. McIntosh D.B. J. Biol. Chem. 1987; 262: 12977-12983Abstract Full Text PDF PubMed Google Scholar). That this was the case, both in the absence and in the presence of an ordinary submillimolar concentration of Ca2+, is shown in panels A–D of Fig. 2 (for control, the well known opposite dependence on Mg2+ of ATP binding in the absence of Ca2+ is shown in Fig. V of Supplemental Material); note that the equilibrium dissociation constant for AMPPCP binding deduced from the assay in the presence of Ca2+ and absence of Mg2+ (Fig. 2C) is slightly lower than the apparent EC50 with which AMPPCP accelerates the rate of Ca2+ dissociation from its binding sites in the absence of Mg2+ (previously illustrated in Fig. 1B): this is probably in part due to the fact that nucleotide binding at a concentration of only a few micromolar may be rate-limiting. We then monitored the binding of various concentrations of AMPPCP in the presence of a high, millimolar Ca2+ concentration, and found that the affinity for AMPPCP binding was increased further, compared with its affinity in the presence of submillimolar Ca2+ concentrations sufficient to saturate the Ca2+ transport sites: this was found both in the presence of Mg2+ (Fig. 2, E and F) and in its absence (Fig. VI of Supplemental Material). Related results are scattered in the literature (19Ogawa Y. Kurebayashi N. Harafuji H. J. Biochem. (Tokyo). 1986; 100: 1305-1318Crossref PubMed Scopus (17) Google Scholar, 21Pang D.C. Briggs F.N. J. Biol. Chem. 1977; 252: 3262-3266Abstract Full Text PDF PubMed Google Scholar, 22Ross D.C. McIntosh D.B. J. Biol. Chem. 1987; 262: 12977-12983Abstract Full Text PDF PubMed Google Scholar). A high millimolar Ca2+ therefore affects AMPPCP binding in some way. AMPPCP Binding at Millimolar Ca2+ (but Not Submillimolar Ca2+) Almost Fully Protects ATPase from Proteolysis by Proteinase K—We first estimated the effect of a high millimolar Ca2+ on AMPPCP-dependent changes through measurements of the susceptibility of Ca2+-ATPase to proteolysis. The simultaneous presence of AMPPCP and Ca2+ (18Møller J.V. Lenoir G. Marchand C. Montigny C. le Maire M. Toyoshima C. Juul B.S. Champeil P. J. Biol. Chem. 2002; 277: 38647-38659Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar, 28Danko S. Yamasaki K. Daiho T. Suzuki H. Toyoshima C. FEBS Lett. 2001; 505: 129-135Crossref PubMed Scopus (92) Google Scholar, 29Ma H. Inesi G. Toyoshima C. J. Biol. Chem. 2003; 278: 28938-28943Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar) was previously described as affording partial protection of Ca2+-ATPase from digestion by proteinase K, and the combination of fluoroaluminate with ADP in the presence of Ca2+ was also described previously to afford protection (9Inesi G. Ma H. Lewis D. Xu C. J. Biol. Chem. 2004; 279: 31629-31637Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar, 28Danko S. Yamasaki K. Daiho T. Suzuki H. Toyoshima C. FEBS Lett. 2001; 505: 129-135Crossref PubMed Scopus (92) Google Scholar). In one of the analyses of the x-ray data, the increased resistance of both the E1·AMPPCP and E1·AlFx·ADP forms against proteolysis was considered to be consistent with the similar structures of the two crystalline forms (5Toyoshima C. Mizutani T. Nature. 2004; 430: 529-535Crossref PubMed Scopus (382) Google Scholar). In contrast, in another analysis (4Sørensen T. L.-M. Møller J.V. Nissen P. Science. 2004; 304: 1672-1675Crossref PubMed Scopus (372) Google Scholar), the protection afforded by AMPPCP was reported to be only minimal compared with the strong protection afforded by fluoroaluminate in the presence of ADP and Ca2+, and this was considered instead to be consistent with the lack of Ca2+ occlusion by E1·AMPPCP, despite the similar structures (4Sørensen T. L.-M. Møller J.V. Nissen P. Science. 2004; 304: 1672-1675Crossref PubMed Scopus (372) Google Scholar). In the present work, we have tried to solve these apparent discrepancies by giving particular attention to the Ca2+ concentration prevailing during proteolysis. We have found that extensive protection by AMPPCP is obtained in the presence of millimolar Ca2+ concentrations similar to those used for crystallization. This is shown in Fi
Referência(s)