F1-ATPase Changes Its Conformations upon Phosphate Release
2002; Elsevier BV; Volume: 277; Issue: 24 Linguagem: Inglês
10.1074/jbc.m110297200
ISSN1083-351X
AutoresTomoko Masaike, Eiro Muneyuki, Hiroyuki Noji, Kazuhiko Kinosita, Masasuke Yoshida,
Tópico(s)Advanced NMR Techniques and Applications
ResumoMotor proteins, myosin, and kinesin have γ-phosphate sensors in the switch II loop that play key roles in conformational changes that support motility. Here we report that a rotary motor, F1-ATPase, also changes its conformations upon phosphate release. The tryptophan mutation was introduced into Arg-333 in the β subunit of F1-ATPase from thermophilic Bacillus PS3 as a probe of conformational changes. This residue interacts with the switch II loop (residues 308–315) of the β subunit in a nucleotide-bound conformation. The addition of ATP to the mutant F1subcomplex α3β(R333W)3γ caused transient increase and subsequent decay of the Trp fluorescence. The increase was caused by conformational changes on ATP binding. The rate of decay agreed well with that of phosphate release monitored by phosphate-binding protein assays. This is the first evidence that the β subunit changes its conformation upon phosphate release, which may share a common mechanism of exerting motility with other motor proteins. Motor proteins, myosin, and kinesin have γ-phosphate sensors in the switch II loop that play key roles in conformational changes that support motility. Here we report that a rotary motor, F1-ATPase, also changes its conformations upon phosphate release. The tryptophan mutation was introduced into Arg-333 in the β subunit of F1-ATPase from thermophilic Bacillus PS3 as a probe of conformational changes. This residue interacts with the switch II loop (residues 308–315) of the β subunit in a nucleotide-bound conformation. The addition of ATP to the mutant F1subcomplex α3β(R333W)3γ caused transient increase and subsequent decay of the Trp fluorescence. The increase was caused by conformational changes on ATP binding. The rate of decay agreed well with that of phosphate release monitored by phosphate-binding protein assays. This is the first evidence that the β subunit changes its conformation upon phosphate release, which may share a common mechanism of exerting motility with other motor proteins. bovine heart mitochondrial F1 thermophilic F1-ATPase adenosine 5′-(β,γ-imino)triphosphate [2-(1-maleimidyl)ethyl]-7-(diethylamino)-coumarin-3-carboxamide phosphate-binding protein adenosine 5′-O-(3-thiotriphosphate) ATP synthase is composed of the major subcomplexes F1 and Fo. F1-catalyzed synthesis of ATP from ADP and Pi is coupled with proton translocation through Fo, which resides in the membrane. F1 part can be separated from Fo part as a water-soluble ATPase that has subunit composition α3β3γδε and hence is often called F1-ATPase. Catalytic nucleotide-binding sites are located on the β subunits, whereas the α subunits contain noncatalytic nucleotide-binding sites. In the crystal structure of the bovine mitochondrial F1-ATPase (MF1),1 the coiled-coil structure of the γ subunit is surrounded by a semi-hexagonal ring of α3β3 (1Abrahams J.P. Leslie A.G.W. Lutter R. Walker J.E. Nature. 1994; 370: 621-628Crossref PubMed Scopus (2714) Google Scholar). F1-ATPase is a rotary motor enzyme; ATP-dependent rotation of the γ subunit relative to the α3β3 ring, as predicted by biochemical studies (2Boyer P.D. Biochim. Biophys. Acta. 1993; 1140: 215-250Crossref PubMed Scopus (909) Google Scholar, 3Boyer P.D. Kohlbrenner W.E. Energy Coupling in Photosynthesis. Elsevier Science Publishing Co., Inc., New York1981: 407-426Google Scholar, 4Duncan T.M. Bulygin V.V. Zhou Y. Hutcheon M.L. Cross R.L. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 10964-10968Crossref PubMed Scopus (456) Google Scholar), was visualized using the thermophilic F1-ATPase (TF1) (5Noji H. Yasuda R. Yoshida M. Kinosita K., Jr. Nature. 1997; 386: 299-302Crossref PubMed Scopus (1922) Google Scholar). Consistent with the presence of three β subunits in the ring, hydrolysis of a single ATP molecule drives a 120° rotation of the γ subunit (6Yasuda R. Noji H. Kinosita K., Jr. Yoshida M. Cell. 1998; 93: 1117-1124Abstract Full Text Full Text PDF PubMed Scopus (703) Google Scholar). It is intriguing how the local conformational changes accompanied by each of reaction steps in the catalytic cycle, such as ATP binding, hydrolysis, and release of ADP and Pi, are amplified and transformed into the force to dislocate the γ subunit. Recent progress shows that each 120° rotation is further divided into a 90° substep that is driven by ATP binding and a 30° substep presumably driven by the release of the product, most likely ADP (7Yasuda R. Noji H. Yoshida M. Kinosita K.Jr. Ito H. Nature. 2001; 410: 898-904Crossref PubMed Scopus (697) Google Scholar). Nucleotide binding induces a large conformational change of the β subunit (8Tsunoda S.P. Muneyuki E. Amano T. Yoshida M. Noji H. J. Biol. Chem. 1999; 274: 5701-5706Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar). Each of the three β subunits in the initial MF1 structure, which was disclosed in 1994 (1Abrahams J.P. Leslie A.G.W. Lutter R. Walker J.E. Nature. 1994; 370: 621-628Crossref PubMed Scopus (2714) Google Scholar), takes one of the two conformations: an “open” form in which catalytic site is empty or a “closed” form in which the catalytic site is occupied by AMP-PNP or ADP. Consistent with that, the β subunits of the crystal structure of the TF1 subcomplex α3β3 without bound nucleotides were all in the open form (9Shirakihara Y. Leslie A.G.W. Abrahams J.P. Walker J.E. Ueda T. Sekimoto Y. Kambara M. Saika K. Kagawa Y. Yoshida M. Structure. 1997; 5: 825-836Abstract Full Text Full Text PDF PubMed Scopus (221) Google Scholar). Compared with the open form, the carboxyl-terminal domain of the β subunit in the closed form swings ∼30° toward the amino-terminal domain so that the catalytic cleft located between two domains is closed. A nucleotide-induced transition from the open to the closed conformation is inherent in the nature of the β subunit, because even the isolated β subunit undergoes the open-close motion responding to nucleotide binding (10Tozawa K. Sekino N. Soga M. Yagi H. Yoshida M. Akutsu H. FEBS Lett. 1995; 376: 190-194Crossref PubMed Scopus (10) Google Scholar, 11Yagi H. Tozawa K. Sekino N. Iwabuchi T. Yoshida M. Akutsu H. Biophys. J. 1999; 77: 2175-2183Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar). Thus, it has been proposed that the coordinated open-to-closed and closed-to-open motions of the β subunits in F1-ATPase accompanied by ATP binding and ADP release drive 90° and 30° rotations of the γ subunit, respectively. In contrast to the nucleotide-dependent open-close motion, the conformational events of the β subunit at the steps of hydrolysis of ATP and release of Pi are unclear. In the case of other ATP-driven motor proteins, myosin and kinesin, the structures of the ATP-bound form and the ADP-bound form are different (12Kikkawa M. Sablin E.P. Okada Y. Yajima H. Fletterick R.J. Hirokawa N. Nature. 2001; 411: 439-445Crossref PubMed Scopus (292) Google Scholar, 13Houdusse A. Kalabokis V.N. Himmel D. Szent-Gyorgyi A.G. Cohen C. Cell. 1999; 97: 459-470Abstract Full Text Full Text PDF PubMed Scopus (321) Google Scholar), and Pi release is assumed to be the step of power stroke (14Suzuki Y. Yasunaga T. Ohkura R. Wakabayashi T. Sutoh K. Nature. 1998; 396: 380-383Crossref PubMed Scopus (155) Google Scholar,15Rice S. Lin A.W. Safer D. Hart C.L. Naber N. Carragher B.O. Cain S.M. Pechatnikova E. Wilson-Kubalek E.M. Whittaker M. Pate E. Cooke R. Taylor E.W. Milligan R.A. Vale R.D. Nature. 1999; 402: 778-783Crossref PubMed Scopus (637) Google Scholar). The initial structure of MF1, however, shows that the ADP-bound β subunit and the AMP-PNP-bound β subunit are in a very similar, closed conformation. Therefore, it appears that the loss of Pi from the catalytic site does not cause significant conformational changes or that the intermediate species of the enzyme generated upon Pi release is too unstable to form crystals even though its conformation is different from the known structures. Indeed, a third conformation of the β subunit was reported recently (16Menz R.I. Walker J.E. Leslie A.G.W. Cell. 2001; 106: 331-341Abstract Full Text Full Text PDF PubMed Scopus (393) Google Scholar); one of the β subunits in the AlF4−-inhibited MF1 exists in a “half-closed” conformation, the catalytic site of which is occupied by ADP and sulfate in mimicry of Pi. Biochemical studies on the kinetics of Pi release and the related conformational changes are few, mainly because of the absence of methods to monitor Pi release from F1-ATPase. The present work has aimed at real time monitoring of conformational changes of the β subunit caused by Pi release. Some Trp residues introduced into the β subunits of Escherichia coli F1-ATPase were reported to confer different fluorescence between AMP-PNP binding and ADP binding (17Weber J. Wilke-Mounts S. Lee R.S.-F. Grell E. Senior A.E. J. Biol. Chem. 1993; 268: 20126-20133Abstract Full Text PDF PubMed Google Scholar, 18Weber J. Bowman C. Senior A.E. J. Biol. Chem. 1996; 271: 18711-18718Abstract Full Text Full Text PDF PubMed Scopus (63) Google Scholar, 19Weber J. Wilke-Mounts S. Hammond S.T. Senior A.E. Biochemistry. 1998; 37: 12042-12050Crossref PubMed Scopus (24) Google Scholar). Fluorescently labeled γ subunit was also reported to change its conformations upon ATP cleavage (20Turina P. Capaldi R.A. J. Biol. Chem. 1994; 269: 13465-13471Abstract Full Text PDF PubMed Google Scholar, 21Turina P. Capaldi R.A. Biochemistry. 1994; 33: 14275-14280Crossref PubMed Scopus (23) Google Scholar). Nevertheless, none of them reported fluorescence changes of the β subunit caused by Pi release by time-resolved measurements. We have sought for new positions for the Trp mutation that can monitor changes of fluorescence upon Pi release. Concurrently for that purpose, we have adopted the Pi-binding protein that enabled real time monitoring of Pi release from the enzyme. Analyses, including kinetic comparison of fluorescence changes and Pi release after addition of ATP, have established that a Trp introduced at position 333 (R333W) reflects Pi release well. The residue 333 of the β subunit, located in helix H, which interacts with the “switch II loop,” appears to sense γ-phosphate of the bound nucleotide and changes its conformation upon loss of Pi from the catalytic site. Nucleotides were purchased from Sigma and Roche Molecular Biochemicals. Mop reagents 7-methylguanosine and purine nucleotide phosphorylase were purchased from Sigma. The fluorescent probe for phosphate-binding protein, [2-(1-maleimidyl)ethyl]-7-(diethylamino)-coumarin-3-carboxamide (MDCC) was purchased from Molecular Probes. The buffers used in the measurements are abbreviated as follows: TK buffer, 50 mmTris-HCl, pH 8.0, 100 mm KCl; TKM2 buffer, 50 mm Tris-HCl, pH 8.0, 100 mm KCl, 2 mm MgCl2; TKM4 buffer, 50 mmTris-HCl, pH 8.0, 100 mm KCl, 4 mmMgCl2; KPi buffer, 100 mmKPi, pH 7.0, 100 mm KCl, 2 mmEDTA; NaPi buffer, 100 mmNaPi, pH 7.0, 200 mm NaCl; and reverse phase buffer: 100 mm NaPi, pH 6.9, 4 mmEDTA. Unless otherwise indicated, TKM2 buffer was used for measurements. To eliminate contaminated Pi from buffers, TKM and TK buffers for PBP assays contain 200 μm7-methylguanosine and 0.01 unit/ml purine nucleotide phosphorylase (named Pi mop) (22Brune M. Hunter J.L. Corrie J.E.T. Webb M.R. Biochemistry. 1994; 33: 8262-8271Crossref PubMed Scopus (428) Google Scholar, 23Brune M. Hunter J.L. Howell S.A. Martin S.R. Hazlett T.L. Corrie J.E.T. Webb M.R. Biochemistry. 1998; 37: 10370-10380Crossref PubMed Scopus (172) Google Scholar). E. coli strain JM109 was used for plasmid amplification. JM103Δ(uncB-uncD) was used for overexpression of the α, β, and γ subunits of F1-ATPase. Plasmids used were pucβ, which carried a gene for the β subunit, for mutagenesis and expression, and pkkαγ, which carried genes for the α and γ subunits, for expression. The βR333W and βD311W mutations into the β subunit were introduced by the Kunkel method (24Kunkel T.A. Bebenek K. McClary J. J. Methods Enzymol. 1991; 204: 125-139Crossref PubMed Scopus (632) Google Scholar) using primer oligonucleotides annealed to the single strand DNA of pucβ: 5′-GATAAATCCCCATCTCCGCAAGCTTCCACTCCAGGTTCGTC-3′ for βR333W introducing cleavage site of HindIII and 5′-CGTCGTGGCCGGAGCCGGATCCGTATAGTCCCAGGCCGGGACGTAAATC-3′ for βD311W introducing cleavage site of BamHI (mutated bases are underlined). For preparation of the isolated β(R333W) and α3β(D311W)3γ, mutated plasmids were transformed into JM103Δ(uncB-uncD) for overexpression and purified using NaPi buffer as previously described (25Matsui T. Yoshida M. Biochim. Biophys. Acta. 1995; 1231: 139-146Crossref PubMed Scopus (86) Google Scholar). Because α3β(R333W)3γ and α3β(D311W/R333W)3γ could not be expressed using the pkkαγβ system, a novel lysate reassembly method was developed. The plasmids pkkαγ, pucβ(R333W), and pucβ(D311W/R333W) were each expressed separately in JM103Δ(uncB-uncD). Pellets from centrifugation of the cultures were diluted in NaPi buffer. The cells containing mutated β subunits were each mixed with those containing the α and γ subunits. The mixture was disrupted by a French pressure cell and was incubated at 30 °C for 30 min for reassembly of the subcomplex. It was then incubated at 60 °C for 15 min, and the insoluble denatured proteins were removed by centrifugation for 40 min at 40,000 rpm. Purification of the subcomplexes were performed by ammonium sulfate gradient in NaPi buffer using a Butyl-Toyopearl 650M column (Tosoh). The purified β subunit and subcomplexes were stored as ammonium sulfate precipitates. They were diluted in TK buffer, concentrated by Vivaspin (Sartorius), and applied twice to a gel filtration (Superdex 200; Amersham Biosciences) for final purification (flow was 0.5 ml/min first with TK buffer and second with KPi buffer) on the day of measurements. Analysis of residual nucleotides after purification of the enzyme was performed as previously described (26Hisabori T. Muneyuki E. Odaka M. Yokoyama K. Mochizuki K. Yoshida M. J. Biol. Chem. 1992; 267: 4551-4556Abstract Full Text PDF PubMed Google Scholar). The number of residual nucleotides bound to α3β(R333W)3γ was less than 0.1 mol/mol after gel filtration with KPi buffer and TK buffer. The number of nucleotides bound to α3β(R333W)3γ at the end points of the fluorescence measurements was estimated by the following protocol. The mixtures of nucleotides and α3β(R333W)3γ from stopped flow measurements were each applied to an Ultrafree filtration device (molecular weight, 5 k cutoff; Millipore). After centrifugation for 2 min at 2 kilorounds per minute at 25 °C, the nucleotide contents in 100 μl of the filtrates were quantified by reverse phase high pressure liquid chromatography (ODS-80Ts; Tosoh) using reverse phase buffer. The amount of nucleotides bound to α3β(R333W)3γ was estimated by subtracting the concentration of the nucleotides free in solution (concentration in the filtrate) from the initial concentration. The fluorescence measurements of the Trp mutant subcomplexes and the isolated β(R333W) subunit were carried out by excitation at 295 nm, and detection of emission at 345 nm was carried out using a spectrofluorometer (FP-6500; Jasco). In a cuvette, 1.2 ml of 5 μm β(R333W) or 1 μm α3β3γ mutants was mixed with 20 μl of ATP or ADP while stirring. Measurements of α3β(R333W)3γ were carried out also by a stopped flow apparatus (SFM-400; BioLogic) using a xenon lamp as a source of light. ATP in TK buffer (30 μl of 1.0 or 0.5 or 0.25 μm) 2All the stopped flow measurements and bound nucleotide measurements were repeated with 75 μl of mixing shots to be sure that the effects of carry-over from the previous shots were minimum. The rate constants and the amount of bound nucleotides (except for AMP-PNP, 0.41 mol/mol in Table I) were essentially the same (± 5%) as those obtained with 30 μl/shot. was mixed with the same volume of 2 μmα3β(R333W)3γ in TKM4 buffer over 20 ms. The same method was applied to ADP, AMP-PNP, and ADPγS. TK buffer prevents ATP at submicromolar concentrations from decomposition into ADP and Pi before addition to α3β(R333W)3γ. The same stopped flow experiments were also performed using buffers that were treated with Pi mop to ensure that the buffer conditions were the same as those used for measurement of Pi release. There was no change in the Trp fluorescence profile between with and without Pi mop in solutions (data not shown). The unisite catalysis was measured using the stopped flow apparatus in the quenched flow mode. It was started by mixing 250 μl of 2 μmα3β(R333W)3γ with the same volume of 1 μm ATP and stopped after various time periods by perchloric acid quenching. Hydrolyzed nucleotides were analyzed by a reverse phase column (ODS-80Ts; Tosoh) using the reverse phase buffer as previously described (26Hisabori T. Muneyuki E. Odaka M. Yokoyama K. Mochizuki K. Yoshida M. J. Biol. Chem. 1992; 267: 4551-4556Abstract Full Text PDF PubMed Google Scholar). Release of Pi from α3β(R333W)3γ was measured using a PBP assay (22Brune M. Hunter J.L. Corrie J.E.T. Webb M.R. Biochemistry. 1994; 33: 8262-8271Crossref PubMed Scopus (428) Google Scholar, 23Brune M. Hunter J.L. Howell S.A. Martin S.R. Hazlett T.L. Corrie J.E.T. Webb M.R. Biochemistry. 1998; 37: 10370-10380Crossref PubMed Scopus (172) Google Scholar). PBP labeled with MDCC was prepared as previously described (22Brune M. Hunter J.L. Corrie J.E.T. Webb M.R. Biochemistry. 1994; 33: 8262-8271Crossref PubMed Scopus (428) Google Scholar, 23Brune M. Hunter J.L. Howell S.A. Martin S.R. Hazlett T.L. Corrie J.E.T. Webb M.R. Biochemistry. 1998; 37: 10370-10380Crossref PubMed Scopus (172) Google Scholar). Binding of Pi to MDCC-labeled PBP (MDCC-PBP) increases the fluorescence emission at 464 nm when the complex is excited at 425 nm. By virtue of rapid binding of Pi to MDCC-PBP (kon = 1.36 × 108m−1 s−1) and high affinity of PBP for Pi (Kd = ∼0.1 μm) (22Brune M. Hunter J.L. Corrie J.E.T. Webb M.R. Biochemistry. 1994; 33: 8262-8271Crossref PubMed Scopus (428) Google Scholar), the increase in the Piconcentration in the solutions could be monitored as the increase of fluorescence emission in real time. PBP assays were carried out using a stopped flow apparatus (SFM-400; BioLogic) under the same conditions as the Trp fluorescence measurements. 30 μl of 2 μmα3β(R333W)3γ in TKM4 buffer was mixed with the same volume of 4 μm MDCC-PBP and 1 μm ATP in TK buffer. The buffers contain Pimop for elimination of Pi to avoid saturation of MDCC-PBP with contaminated Pi. To estimate the rate of formation of the MgADP-inhibited form under the fluorescence measurement conditions, the following experiment was carried out. 20 μl of 11 μmα3β(R333W)3γ and 200 μl of 0.55 μm ATP were manually mixed and preincubated at 25 °C for varying periods of time. 150 μl of the incubated solution was injected into the ATP-regenerating system (27Matsui T. Muneyuki E. Honda M. Allison W.S. Dou C. Yoshida M. J. Biol. Chem. 1997; 272: 8215-8221Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar) containing 2 mm ATP-Mg (mixture of equal concentrations of ATP and MgCl2) in TKM2 buffer. The time course of ATP hydrolysis was measured by monitoring the absorbance at 340 nm using a spectrophotometer (V-550; Jasco). The slope of the absorbance is initially small as the majority of the molecules are in the MgADP-inhibited form, but it gradually increases because of reactivation by binding of ATP to the α subunit (28Jault J.M. Matsui T. Jault F.M. Kaibara C. Muneyuki E. Yoshida M. Kagawa Y. Allison W.S. Biochemistry. 1995; 34: 16412-16418Crossref PubMed Scopus (67) Google Scholar). Therefore, the ratio of active α3β(R333W)3γ was estimated from the initial slope of 10 s of absorbance at 340 nm compared with that without preincubation. The concentrations of β(R333W), α3β(R333W)3γ, α3β(D311W/R333W)3γ, and α3β(D311W)3γ were analyzed by BCA assay (Pierce) and absorbance at 280 nm. Data processing was performed by BioKine software (BioLogic), Origin 6.0 (Microcal Software), Excel 97 (Microsoft), and Dynafit (BioKin) (29Kuzmic P. Anal. Biochem. 1996; 237: 260-273Crossref PubMed Scopus (1329) Google Scholar). The initial crystal structure of MF1 suggests that Arg-333 in TF1-β (Arg-337 in MF1-β) in helix H interacts with Asp-311 in TF1-β (Asp-315 in MF1-β) of the switch II loop only when the β subunit is in the closed conformation (Fig. 1) (1Abrahams J.P. Leslie A.G.W. Lutter R. Walker J.E. Nature. 1994; 370: 621-628Crossref PubMed Scopus (2714) Google Scholar). Ren et al. (30Ren H. Dou C. Stelzer M. Allison S.W. J. Biol. Chem. 1999; 274: 31366-31372Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar) showed that cysteines introduced at positions 311 and 333 of TF1-β can readily form an intramolecular cross-link in two of the three β subunits in the α3β3γ subcomplex of TF1. Cross-linking abolished ATPase activity almost completely by fixing two β subunits in the closed conformation. We introduced Trps into the same positions and examined the fluorescence response of the mutant, expecting to have enabled fluorescent detection of nucleotide-induced open-close motion of the β subunits. Trp fluorescence of 1 μm α3β(D311W/R333W)3γ subcomplex decreased when 0.5 μm ADP was added (Fig.2A). The fluorescent response to the same concentration of ATP was very different from that observed for ADP; a transient fluorescence increase was followed by rapid decay. The final level of fluorescence after decay was similar to that attained by ADP. Then, to determine which (or both) Trp was responsible for this transient fluorescence change, we made two single mutants, α3β(D311W)3γ and α3β(R333W)3γ. The fluorescence response of α3β(D311W)3γ to ADP was similar to that of ATP, that is, a similar extent of increase and no further rapid changes (Fig. 2B). On the other hand, fluorescence of α3β(R333W)3γ showed a two-phase response to ATP addition: transient increase and rapid decay (Fig.2C). The addition of ADP caused only a slight increase in fluorescence. The final level of fluorescence change by ATP was almost the same as that attained by ADP. It appeared that these two phases might represent certain steps in the catalysis occurring at a single catalytic site. Therefore, further fluorescence measurements were focused on α3β(R333W)3γ, using a stopped flow apparatus, which could provide higher time resolution than manual mixing. It should be added that the three mutants mentioned above retained ATPase activity of rotary catalysis at a saturating ATP concentration (2 mm): 140 turnovers/s (α3β (D311W/R333W)3γ), 29 turnovers/s (α3β(D311W)3γ), and 106 turnovers/s (α3β(R333W)3γ), which are 61, 13, and 46%, respectively, of that of the α3β3γ subcomplex without these mutations. Hereafter, we focus on the characteristics of α3β(R333W)3γ.FIG. 2Time courses of fluorescence changes of the Trp mutants. At the times indicated by arrowheads, ATP or ADP was manually mixed with α3β(D311W/R333W)3γ (A), α3β(D311W)3γ (B), and α3β(R333W)3γ (C). The concentrations of subcomplexes and nucleotides in the mixtures were 1 and 0.5 μm, respectively. The details of the experiments are described under “Experimental Procedures.”View Large Image Figure ViewerDownload (PPT) To understand whether the different fluorescence response of α3β(R333W)3γ to ATP or ADP is generated from intersubunit interaction in the subcomplex or from conformational changes within a β subunit, fluorescence response of the isolated β(R333W) subunit to ATP or ADP was examined. Because the isolated β subunit can bind nucleotide but does not retain catalytic ability (31Yoshida M. Sone N. Hirata H. Kagawa Y. J. Biol. Chem. 1977; 252: 3480-3485Abstract Full Text PDF PubMed Google Scholar), the nucleotide-induced change of Trp fluorescence of β(R333W) can be solely attributed to the nucleotide binding. The addition of ATP or ADP to the isolated β(R333W) caused an instantaneous increase in Trp fluorescence that was followed by a slow increase (∼30 s), and the fluorescence remained constant after saturation (Fig.3A). The reason for the slow increase is not known, but it is worth noting that the extent of the fluorescence increase by ATP is significantly larger than by ADP, just as observed for initial fluorescence increase of α3β(R333W)3γ. The Kdvalues for ATP and ADP estimated from fluorescence changes at various concentrations of nucleotide (Fig. 3B) are similar to each other: 20 μm for ATP and 27 μm for ADP, consistent with the values reported previously (32Odaka M. Kaibara C. Amano T. Matsui T. Muneyuki E. Ogasahara K. Yutani K. Yoshida M. J. Biochem. (Tokyo). 1994; 115: 789-796Crossref PubMed Scopus (22) Google Scholar). These results suggest that conformational changes within a β subunit induced by AT(D)P binding can explain the initial increase of fluorescence observed for the α3β(R333W)3γ subcomplex. The different magnitude of fluorescence increase in response to ATP and ADP indicates that the Trp residue introduced at position 333 of the β subunit is able to sense the presence of γ-phosphate of the bound adenine nucleotides, and this ability is inherent in the β(R333W) subunit. For α3β(R333W)3γ, the initial increase in fluorescence by the addition of ATP was our initial focus (Fig.4). The addition of a nonhydrolyzable ATP analog, AMP-PNP, to α3β(R333W)3γ induced an increase in fluorescence that was similar to that observed for ATP, but no subsequent decay was observed (Fig. 4A). Similarly, the decay was not observed for binding of ATP in the absence of Mg, where hydrolysis was blocked (data not shown). Another ATP analog, ATPγS, which is a poor substrate for F1, also induced a similar fluorescence increase (Fig. 4A) that was followed by a slower decay. Taken together, we concluded that the initial fluorescence increase reflected the occupation of a catalytic site of the β subunit by ATP (step 1 of SchemeFS1). The rates of nucleotide binding calculated from the fluorescence changes of α3β(R333W)3γ were (1.7 ± 0.3) ×107m−1 s−1 for ATP, (4.1 ± 0.7) ×107m−1s−1 for ADP, (1.3 ± 0.0) ×106m−1 s−1 for AMP-PNP, and (2.8 ± 0.2) ×107m−1s−1 for ATPγS.SCHEME 1Reaction scheme of uni-site ATP hydrolysis by F1 ATPase.View Large Image Figure ViewerDownload (PPT) Decay of the fluorescence of α3β(R333W)3γ after the initial increase was observed under the conditions where unisite catalysis (33Grubmeyer C. Cross R.L. Penefsky H.S. J. Biol. Chem. 1982; 257: 12092-12100Abstract Full Text PDF PubMed Google Scholar, 34Milgrom Y.M. Cross R.L. J. Biol. Chem. 1997; 272: 32211-32214Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar, 35Milgrom Y.M. Murataliev M.B. Boyer P.D. Biochem. J. 1998; 330: 1037-1043Crossref PubMed Scopus (45) Google Scholar) was occurring and hence may correspond to a certain step of catalysis after the capture of ATP by a catalytic site. 0.5, 0.25, and 0.125 μm ATP caused the decay at comparable rates, 3The rates of decay estimated by the simple fitting scheme F1 + ATP → F1·ATP → F1·ADP + Pi attributing fluorescence increase to binding of ATP and decay to hydrolysis werekdecay = 1.9 ± 0.3, 1.8, and 2.0 s−1 for 0.5, 0.25, and 0.125 μm ATP, respectively. which indicates that the binding step is not involved in the decay (time courses are not shown). From the time courses of Trp fluorescence upon the addition of various nucleotides (Fig. 4A), it is assumed that ATP hydrolysis or an event that occurs immediately after that causes fluorescence decay. Typically, a slowly hydrolyzed ATP analog, ATPγS, causes slow decay. To test the assumption described above, the time course of generation of ADP (step 2 of Scheme FS1) was measured. α3β(R333W)3γ and ATP were mixed using a stopped flow apparatus under the same conditions as the fluorescence measurements, and the reactions were stopped after various periods of time by the addition of perchloric acid. Acid quenching liberates substrates from denatured enzymes. Therefore, irrespective of whether the substrate is released or still bound to the enzyme, the generation of ADP can be detected by this method. The generation of ADP occurred with the rate constant of 14.4 s−1 (Fig.5), which is greater than the rate of fluorescence decay (2.7 s−1; Fig. 4). Therefore, the cause of fluorescence decay can be assigned to a step after ATP hydrolysis such as Pi release and/or ADP release, etc. To determine which is the case, the release of ADP and Pi was examined. Analysis of the enzyme-bound nucleotides was carried out by sampling the mixtures of 1 μm α3β(R333W)3γ and 0.5 μm nucleotides from stopped flow fluorescence measurements and applying them each to an Ultrafree filtration device. The amount of nucleotides in the filtrates was analyzed. Virtually all (92%) of the product ADP remained bound to the enzyme even after all of the ATP had been hydrolyzed (Table I). Therefore, it is not feasible to assign the fluorescence decay to ADP release as the cause of the fluorescence decrease. We also measured the amount of enzyme-bound nucleotide when ADP, AMP-PNP, and ATPγS were added. Again, nearly all of the added nucleotides were stably bound to the enzyme (Table I). Taking this into account, the highest Trp fluorescence level by ATP, ATPγS, and AMP-PNP (Fig. 4A) can be assigned to the γ-phosphate (γ-thiophosphate)-bound form.Table IConcentrations of nucleotides remained bound to 1 μM of α3β(R333W)3γ after incubation with 0.5 μM of nucleotidesNucleotideBound concentrationμMATP0.46ADP0.45AMP-PNP0.46ATPγS0.45 Open table in a new tab To monitor the time course of Pi release from the enzyme (step 3 of Scheme FS1), a PBP assay was adopted. Fluorescence of a coumarin-labeled PBP (MDCC-PBP) increases severalfold when Pi binds. The versatility of monitoring of Pi release from α3β(R333W)3γ by the PBP assay was carefully assessed and established (see “Experimental Procedures”). Thus, real time, continuous monitoring of Pi released from F1-ATPase became possible for the first time. We found that Pi was released from α3β(R333W)3γ with a rate constant of 2.8 ± 0.1 s−1 (Fig.6), which is the same rate as that of the fluorescence decay (Fig. 4). Thus, it is suggested that the decay of fluorescence after the initial increase reflects the decay of the enzyme form with bound ADP-Pi to the enzyme form with bound ADP only, that is, release of Pi from the enzyme. However, if there is a rapid conversion from the active enzyme-ADP complex into inactive enzyme-ADP complex, this conversion is also a candidate for the fluorescence decay. This possibility should be considered because it is known that the so-called MgADP-inhibited form, an inactive form of enzyme-ADP complex, tends to be generated under these conditions. We examined this possibility next. The MgADP-inhibited form (step 4 of Scheme FS1) is not caused by a mere product inhibition but by stable retention of MgADP at the catalytic site. The MgADP can either be picked up from the bulk phase medium or can be a remnant of hydrolysis that remains bound to the enzyme (27Matsui T. Muneyuki E. Honda M. Allison W.S. Dou C. Yoshida M. J. Biol. Chem. 1997; 272: 8215-8221Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar). When the MgADP-inhibited form of F1-ATPase is exposed to ATP and Mg2+, it shows no ATPase activity initially but is gradually reactivated with a time constant of ∼30 s (27Matsui T. Muneyuki E. Honda M. Allison W.S. Dou C. Yoshida M. J. Biol. Chem. 1997; 272: 8215-8221Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar, 43Hirono-Hara Y. Noji H. Nishiura M. Muneyuki E. Hara K.Y. Yasuda R. Kinosita K., Jr. Yoshida M. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13649-13654Crossref PubMed Scopus (158) Google Scholar). Therefore, the population of the MgADP-inhibited form in a certain preparation of F1-ATPase can be assessed from the initial rate of ATP hydrolysis. Under the same conditions used for the fluorescence measurement, we took an aliquot from the solution at the indicated times, injected it into the ATPase assay mixture and measured the initial ATPase activity. The initial ATPase activities were plotted as a function of the time and the rate of generation of the MgADP-inhibited form in the solution was estimated (Fig. 7). The time constant of the onset of MgADP inhibition thus estimated was 15 s, which is much slower than the fluorescence decay. Therefore, the possibility that the fluorescence decay is caused by generation of the MgADP-inhibited state is unlikely. In other words, the lifetime of active MgADP bound form is long enough to be maintained during fluorescence changes of several seconds. Taking these results together, we can conclude that under unisite conditions, the increase in Trp fluorescence of the α3β(R333W)3γ subcomplex occurs upon ATP binding, and the decay occurs as a function of Pirelease (Table II).Table IIRate constants of the reaction steps and corresponding Trp fluorescence changesReaction stepStep in Scheme FS1Rate constantTrp fluorescenceATP binding1(1.7 ± 0.3) × 107m−1 s−1IncreaseATP hydrolysis214.4 ± 0.2 s−1Negligible changePirelease32.8 ± 0.1 s−1DecayADP inhibition46.7 × 10−2 s−1 Open table in a new tab The novel Trp mutant α3β(R333W)3γ revealed that the residue Arg-333 senses the presence of γ-phosphate at the catalytic site of β subunit as well as changes in the surrounding structure upon nucleotide binding and Pirelease. Although Arg-333 does not directly contact the γ-phosphate of the bound nucleotide in the crystal structure, this residue somehow recognizes conformational differences between ATP-bound (or ADP-Pi-bound) and ADP-bound conformations of the β subunit. It is worth noting that when a nucleotide is bound to the catalytic site, the mutated residue Arg-333 interacts with the switch II loop, a switch region common to a wide range of nucleotide triphosphate-utilizing proteins including GTP-binding proteins and motor proteins (12Kikkawa M. Sablin E.P. Okada Y. Yajima H. Fletterick R.J. Hirokawa N. Nature. 2001; 411: 439-445Crossref PubMed Scopus (292) Google Scholar, 13Houdusse A. Kalabokis V.N. Himmel D. Szent-Gyorgyi A.G. Cohen C. Cell. 1999; 97: 459-470Abstract Full Text Full Text PDF PubMed Scopus (321) Google Scholar, 36Lambright D.G. Sondek J. Bohm A. Skiba N.P. Hamm H.E. Sigler P.B. Nature. 1996; 379: 311-319Crossref PubMed Scopus (1041) Google Scholar, 37Rayment I. Rypniewski W.R. Schmidt-Base K. Smith R. Tomchick D.R. Benning M.M. Winkelmann D.A. Wesenberg G. Holden H.M. Nature. 1993; 261: 50-58Google Scholar, 38Fisher A.J. Smith C.A. Thoden J.B. Smith R. Sutoh K. Holden H.M. Rayment I. Biochemistry. 1995; 34: 8960-8972Crossref PubMed Scopus (629) Google Scholar, 39Smith C.A. Rayment I. Biochemistry. 1996; 35: 5404-5417Crossref PubMed Scopus (511) Google Scholar, 40Dominguez R. Freyzon Y. Trybus K.M. Cohen C. Cell. 1998; 94: 559-571Abstract Full Text Full Text PDF PubMed Scopus (585) Google Scholar). A function of switch II in myosin and kinesin is to transmit conformational changes caused by γ-phosphate release to a distant point where motility of motor proteins is exerted (41Vale R.D. Milligan R.A. Science. 2000; 288: 88-95Crossref PubMed Scopus (1200) Google Scholar). For example, in the case of myosin, an alanine mutation introduced into Gly-457 in the switch II loop causes loss of motility (42Sasaki N. Shimada T. Sutoh K. J. Biol. Chem. 1998; 273: 20334-20340Abstract Full Text Full Text PDF PubMed Scopus (75) Google Scholar). In the crystal structure of monomeric kinesin motor KIF1A, the switch II loops of the ADP-bound and ATP-bound forms were in different conformations (12Kikkawa M. Sablin E.P. Okada Y. Yajima H. Fletterick R.J. Hirokawa N. Nature. 2001; 411: 439-445Crossref PubMed Scopus (292) Google Scholar). Moreover, studies using fluorescence energy transfer and other analyses of conventional kinesin indicated a difference between ATP-bound and ADP-bound forms in the flexibility of the neck linker (15Rice S. Lin A.W. Safer D. Hart C.L. Naber N. Carragher B.O. Cain S.M. Pechatnikova E. Wilson-Kubalek E.M. Whittaker M. Pate E. Cooke R. Taylor E.W. Milligan R.A. Vale R.D. Nature. 1999; 402: 778-783Crossref PubMed Scopus (637) Google Scholar). Therefore, it is natural to assume that F1-ATPase, another motor protein, might undergo an analogous conformational change when Pi is released. However, real images of the conformational change of F1-ATPase that we detected by the fluorescence change cannot be directly assumed by comparing crystal structures of various nucleotide binding states that are solved to date. The crystal structure of F1-ATPase containing β subunits in the ATP-bound, ADP-bound, and empty forms in one molecule (1Abrahams J.P. Leslie A.G.W. Lutter R. Walker J.E. Nature. 1994; 370: 621-628Crossref PubMed Scopus (2714) Google Scholar) suggested that Pi release does not cause drastic conformational changes because the structure of AMPPNP-bound and ADP-bound β subunits are very similar to each other, both in the same closed conformation. There is a possibility that introduced Trp reflects a very subtle change in the conformation accompanying Pirelease. A new crystal structure (16Menz R.I. Walker J.E. Leslie A.G.W. Cell. 2001; 106: 331-341Abstract Full Text Full Text PDF PubMed Scopus (393) Google Scholar) was discovered recently that contains a third half-closed conformation of the β subunit, which is in between the closed and open forms. This new conformation is thought to be in a transition state where the products ADP and Piare both bound. The existence of this half-closed conformation indicated partial opening of the β subunit during ATP hydrolysis. As a next step, further opening of ADP-bound form induced by Pi release can naturally be assumed. Therefore, another possibility is that the fluorescence change reflects the difference of these two partially open states: one with bound ADP-Pi and the other with bound ADP. This problem has direct implications on the conformational transitions accompanying the catalytic cycle but awaits further studies to be clarified. The present research also revealed new information about the ATPase reaction by direct, real time measurements of some of the kinetic parameters of the unisite catalysis (Table II). The parameters shown here give insights into the reaction mechanism. The first is the rate of nucleotide binding. Comparing the results of stopped flow measurements (unisite catalysis conditions) with the previous single molecule observation of rotating F1-ATPase at low ATP concentrations (bi-site or tri-site catalysis conditions), the rates of ATP binding are in the same range ((1.7 ± 0.3) ×107m−1 s−1 and 2.7×107m−1 s−1, respectively). This indicates that the rates of ATP binding are almost the same for the first and second (or third) catalytic sites. The second is that Pi release predominantly occurs while ADP remains bound to the enzyme in unisite catalysis. The third is that the rate of Pi release is slower than that of ATP hydrolysis, suggesting that the F1·ADP·Pi complex has to wait for some conformational change that allows Pirelease. Future studies should be directed at the observation of conformational changes of the β subunit in F1-ATPase at each step (including Pi release) of during rotational catalysis. For this purpose, a new probe that is tractable by single molecule observation is necessary. Dr. Martin R. Webb is gratefully acknowledged for advice on preparation and measurements of phosphate-binding protein. We thank K. Kawashima, Dr. Motojima, Dr. Kato-Yamada, Dr. Watanabe, Dr. Georges, Dr. T. Suzuki, Dr. Motohashi, Dr. Tabata, Dr. Hisabori, Dr. Taguchi, and J. Suzuki for valuable discussion and technical advice and Dr. Hardy for critically reading the manuscript.
Referência(s)