Portal de Periódicos da CAPES

Dimerization of Bovine F1-ATPase by Binding the Inhibitor Protein, IF1

2000; Elsevier BV; Volume: 275; Issue: 37 Linguagem: Inglês

10.1074/jbc.c000427200

1083-351X

Elena Cabezón, Ignacio Aréchaga, P. Jonathan, Georgina S. Butler, John E. Walker,

Mitochondrial Function and Pathology

In mitochondria, the hydrolytic activity of ATP synthase is regulated by a natural inhibitor protein, IF1. The binding of IF1 to ATP synthase depends on pH values, and below neutrality, IF1 forms a stable complex with the enzyme. Bovine IF1 has two oligomeric states, dimer and tetramer, depending on pH values. At pH 6.5, where it is active, IF1 dimerizes by formation of an antiparallel α-helical coiled-coil in its C-terminal region. This arrangement places the inhibitory N-terminal regions in opposition, implying that active dimeric IF1 can bind two F1 domains simultaneously. Evidence of dimerization of F1-ATPase by binding to IF1 is provided by gel filtration chromatography, analytical ultracentrifugation, and electron microscopy. At present, it is not known whether IF1 can bring about the dimerization of the F1F0-ATPase complex. In mitochondria, the hydrolytic activity of ATP synthase is regulated by a natural inhibitor protein, IF1. The binding of IF1 to ATP synthase depends on pH values, and below neutrality, IF1 forms a stable complex with the enzyme. Bovine IF1 has two oligomeric states, dimer and tetramer, depending on pH values. At pH 6.5, where it is active, IF1 dimerizes by formation of an antiparallel α-helical coiled-coil in its C-terminal region. This arrangement places the inhibitory N-terminal regions in opposition, implying that active dimeric IF1 can bind two F1 domains simultaneously. Evidence of dimerization of F1-ATPase by binding to IF1 is provided by gel filtration chromatography, analytical ultracentrifugation, and electron microscopy. At present, it is not known whether IF1 can bring about the dimerization of the F1F0-ATPase complex. 4-morpholinepropanesulfonic acid polyacrylamide gel electrophoresis The F1F0-ATP synthase complex (also called F1F0-ATPase) plays a central role in energy transformation in most living organisms. It is composed of two major domains, a globular F1 catalytic domain and a membrane-bound F0 proton-translocating domain linked together by a central stalk. The synthesis of ATP requires an electrochemical proton gradient across the inner mitochondrial membrane, which is driven by the transport of protons back into the matrix through the F0 domain. When a cell is deprived of oxygen, the electrochemical gradient across the inner membrane collapses, and the enzyme switches its catalytic activity from ATP synthesis to ATP hydrolysis. Under these conditions, F1-ATPase, the catalytic domain of ATP synthase, catalyzes the hydrolysis of ATP to ADP and phosphate. In mitochondria, this hydrolytic activity is regulated by a natural inhibitor protein, IF1. In bovine mitochondria, IF1 is a basic protein of 84 amino acids long (1Walker J.E. Curr. Opin. Struct. Biol. 1994; 4: 912-918Crossref PubMed Scopus (89) Google Scholar). The binding of IF1 to ATP synthase depends on pH values and, below neutrality, its inhibitory capacity increases (2Panchenko M.V. Vinogradov A.D. FEBS Lett. 1985; 184: 226-230Crossref PubMed Scopus (56) Google Scholar). Recently, we have shown that bovine IF1 has two oligomeric states, tetramer and dimer, favored by pH values above and below 6.5, respectively (3Cabezon E. Butler P.J.G. Runswick M.J. Walker J.E. J. Biol. Chem. 2000; 275: 25460-25464Abstract Full Text Full Text PDF PubMed Scopus (155) Google Scholar). Activation is accompanied by a decrease in IF1 helicity relative to the inactive form between residues 35 and 47, which are involved in the formation of the inactive tetramers. At a pH value of about 6.5, IF1 forms an active antiparallel dimer, and this arrangement places the inhibitory N-terminal regions in opposition. At higher pH values, two dimers associate into the inactive tetramer. An important implication of this model is that dimeric IF1 is capable of binding to two F1 domains simultaneously. In this paper, we describe experiments using gel filtration chromatography, analytical ultracentrifugation, and electron microscopy that validate this prediction. The purification of bovine F1-ATPase (4Orris G.L. Leslie A.G.W. Braig K. Walker J.E. Structure. 1998; 6: 831-837Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar) and recombinant bovine IF1 (3Cabezon E. Butler P.J.G. Runswick M.J. Walker J.E. J. Biol. Chem. 2000; 275: 25460-25464Abstract Full Text Full Text PDF PubMed Scopus (155) Google Scholar) were carried out as described previously. The IF1-F1 complex was prepared as follows: purified F1-ATPase (25 mg), stored as an ammonium sulfate precipitate, was collected by centrifugation (31,000 ×g) at 4 °C, re-dissolved in minimal buffer (20 mm MOPS1-NaOH, pH 6.6, 1 mm EDTA, 10% glycerol, 0.001% (w/v) phenylmethylsulfonyl fluoride), and desalted on a PD-10 column (Amersham Pharmacia Biotech). The enzyme was mixed with a 5-fold molar excess of IF1 over F1 and kept for 20 min at 37 °C. A portion of MgATP was added at 5, 10, and 15 min to give a final concentration of 0.5 mm. Then, the mixture was loaded onto a HiLoad 26/60 Superdex 200 column to separate the IF1-F1 complex from F1-ATPase and free IF1. Protein samples (5 mg/ml; 100 μl) prepared as described above were chromatographed at room temperature at a flow rate of 0.2 ml/min on a Superose 6 HR 10/30 column (Amersham Pharmacia Biotech) pre-equilibrated in buffer containing 50 mmMOPS-NaOH, pH 6.5, 10 mm magnesium sulfate, 1 mm EDTA, 0.02% sodium azide,10% glycerol, and 0.001% phenylmethylsulfonyl fluoride. The absorbance of the eluant was monitored at 280 nm. Thyroglobulin, ferritin, catalase, and aldolase (molecular masses of 669, 440, 232, and 158 kDa, respectively) were used to calibrate the column. Its void volume (V 0) was determined using blue dextran 2000. TheK av for the individual proteins was calculated from the following expression: K av = (Ve −V 0)/(Vt −V 0), where Vt andVe are the elution volume for the protein and the total bed volume, respectively. A plot ofK av versus log(molecular weight) with standards gave a straight line with a correlation coefficient of 0.99. Linear regression gave the equation log(molecular weight) = −2.239 K av + 3.427, which was used to calculate the apparent molecular weight. Denaturing SDS-PAGE gels containing a 12–22% (w/v) acrylamide gradient separating gel and a 4% (w/v) stacking gel (acrylamide:N,N′-methylenebisacrylamide), 30:0.8 (w/w) were cast in 10-cm × 10-cm × 0.6-mm format and run in the buffer system of Laemmli (5Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207012) Google Scholar). Sedimentation velocity runs were performed at 20.0 °C and at various rotor speeds (finally using 40,000 rpm for F1-ATPase and 30,000 rpm for the IF1-F1 complex) in a Beckman AN-60Ti rotor and a Beckman XL-A ultracentrifuge. The samples were dissolved in buffer consisting of 50 mm MOPS-NaOH, pH 6.5, 10 mmmagnesium sulfate, 1 mm EDTA, 0.02% sodium azide, and 0.001% phenylmethylsulfonyl fluoride. Scans were taken as frequently as possible (i.e. with a zero-interval setting, which gives∼4.5 min with 3 cells in the rotor). Data were analyzed initially by plotting g(s*) against s*, where g(s*) is the fraction of material sedimenting betweens* and (s* + δs*) (6Stafford III, W.F. Anal. Biochem. 1992; 203: 295-301Crossref PubMed Scopus (520) Google Scholar, 7Stafford III, W.F. Methods Enzymol. 1994; 240: 478-501Crossref PubMed Scopus (121) Google Scholar), using the DCDT+ software package (Version 1.05) (8Philo J.S. Anal. Biochem. 2000; 279: 151-163Crossref PubMed Scopus (237) Google Scholar). This software was also used for direct fitting of simple Gaussian functions to dc/dt versus s curves (9Stafford III, W.F. Curr. Opin. Biotechnol. 1997; 8: 14-24Crossref PubMed Scopus (67) Google Scholar) to test for the number of components giving a ”best fit“ to the data. Sedimentation coefficients were converted from s* tos 20,w, taking values of 0.759 ml/g for the partial specific volume, 1.00213 g/ml for the solvent density, and 1.0103 (entipoise) for the viscosity, calculated with SEDNTERP (Version 1.03) (10Laue T.M. Shah B.D. Ridgeway T.M. Pelletier S.L. Harding S.E. Rowe A.J. Horton J.C. Analytical Ultracentrifugation in Biochemistry and Polymer Science. Royal Society of Chemistry, Cambridge, U.K.1992: 90-125Google Scholar). All plots were produced from the data using the program ProFit (QuantumSoft). For plots of the original scans, scans at equal time intervals were selected (630 s for F1 alone and 666 s for F1-IF1complex), to give visual separation between the traces drawn. Grids were prepared by evaporating carbon onto 400-mesh copper/rubidium grids (Maxtaform, Graticules, Tombridge, U.K.). They were covered by a thin film of a 0.5% Formvar solution and then washed with chloroform. Grids were glow-discharged for 30 s to make the carbon film hydrophilic. A solution of the purified IF1-F1 complex (5 μl, 0.01 mg/ml) was applied to the grid and left for 2 min. Then the grid was washed with three drops of gel filtration buffer and stained with a solution of 4% (w/v) methylamine tungstate (Agar, Stansted, U.K.) in water (pH 6.5). Images were recorded at a magnification of 67,000 on a Philips Tecnai 12 electron microscope operating at 100 kV using low dose conditions (approximately 10 e−/Å2). The quality of the images was checked on an optical difractometer. The defocus was about 1000 nm. The images were collected with a Zeiss-SCAI scanner using a step size of 7 μm (pixel size, 1.04 Å). Images were demagnified by linear interpolation on the computer to a pixel size of 5.2 Å. The apparent molecular masses of F1-ATPase and IF1-F1complex in solution were determined by gel filtration chromatography (Fig. 1) and sedimentation velocity analysis (see Fig. 3). F1-ATPase eluted from a Superose 6 column as a single peak of 367 kDa, whereas under the same conditions, the IF1-F1 complex eluted at 696 kDa. The small shoulder contains monomeric F1-ATPase (Fig. 1 A). Fractions from both peaks were analyzed by SDS-PAGE (Fig.2). The gels confirmed the presence of IF1 (9.6 kDa) in the IF1-F1complex. The separation of the F1 and IF1-F1 complexes by gel filtration has been observed before (11Walker J.E. Fearnley I.M. Gay N.J. Gibson B.W. Northrop F.D. Powell S.J. Runswick M.J. Saraste M. Tybulewicz V.L. J. Mol. Biol. 1985; 184: 677-701Crossref PubMed Scopus (380) Google Scholar).Figure 3Sedimentation velocity analysis of F1-ATPase and of the purified IF1-F1 complex. The data are shown (at equal time intervals for each sample) as A 280against radius for each sample. Analysis of the data is shown as plots of g(s*) against s* (showing the distribution of the sample with sedimentation coefficient) and also as the residuals (in A 280 againsts*20,w) for fitting of a model with a single component, distributed as a Gaussian function, to the dc/dt curves. Panels A andB, F1-ATPase and IF1-F1complex, respectively.View Large Image Figure ViewerDownload (PPT)Figure 2Analysis of column fractions by SDS-PAGE. Panels A and B, fractions from peaks A and B, respectively, in Fig. 1. The apparent molecular masses of the standard proteins are: 83, 62, 47.5, 32.5, 25, and 16.5 kDa.View Large Image Figure ViewerDownload (PPT) Sedimentation velocity data for F1-ATPase and the IF1-F1 complex are summarized in Fig.3. Direct dc/dtanalysis showed that F1 alone hads 20,w = 12.1 (± 0.7) S and F1-IF1 complex hads 20,w = 17.0 (± 0.1) S. To test whether there might be more than one component in the boundary, plots of g(s*) against s*20,w were also made. As a further test, model fitting of the dc/dt data with models with either one or two components was carried out. The best fits were given by the single-component model for both data sets, and plots of the fit residuals against radius are shown (Fig. 3). This increase in sedimentation coefficient is compatible with IF1 binding two F1-ATPase units in the complex, and this is the only possibility, as any shape change that produced such an increase in sedimentation would have led to a lower apparent molecular mass during elution from the column. Thus all of the hydrodynamic data are only explained by a dimerization of F1-ATPase on forming a complex with IF1. Samples of the IF1-F1 were diluted in the gel filtration buffer to a protein concentration of 0.01 mg/ml and stained with methylamine tungstate (4%, w/v). Methylamine tungstate, which has a pH value of 6.5 in water solution, has been reported to preserve delicate complexes such as viruses, and pH 6.0–6.5 is optimal for formation of the IF1-F1complex (12Faberge A.G. Oliver R.M. J. Microscopie (Paris). 1974; 20: 241-246Google Scholar). Electron microscopic analysis of the samples (Fig.4) revealed the presence of dimeric F1 complexes. They represented about 70% of the particles present in the fields that were examined, the other particles being mostly monomeric F1-ATPase. As the starting material was the main peak obtained by gel filtration of the F1-IF1 complex, it was devoid of monomeric F1-ATPase. Therefore, the monomers must have arisen by disruption of the dimeric complex during the dilution and staining procedures. A few particles appear to be trimeric. As there was only evidence for dimer formation by gel filtration and sedimentation analysis, these apparent trimers in the electron microscopy pictures are unlikely to represent biologically significant assemblies. In a control experiment, particles of F1 alone were examined in the same way. About 94% of them were clearly monomeric. Bovine IF1 exists in two oligomeric states depending on the pH value. At low pH values, IF1 forms an active dimer held together by an antiparallel α-helical coiled-coil from residues around 46–84 (3Cabezon E. Butler P.J.G. Runswick M.J. Walker J.E. J. Biol. Chem. 2000; 275: 25460-25464Abstract Full Text Full Text PDF PubMed Scopus (155) Google Scholar). 2D. J. Gordon-Smith, R. J. Carbajo, M. J. Runswick, J. E. Walker, and D. Neuhaus, manuscript in preparation. In this arrangement, the N-terminal inhibitory domains are exposed, and each is capable of interacting with an F1-ATPase complex, as the results presented here now demonstrate. It is thought that the interaction between IF1 and F1 involves the C-terminal region of one or more of the β-subunits (14Jackson P.J. Harris D.A. FEBS Lett. 1988; 229: 224-228Crossref PubMed Scopus (49) Google Scholar). At present, it is not known whether IF1 can bring about the dimerization of the F1F0-ATPase complex and whether it is involved in formation of dimeric complexes of bovine and yeast F1F0-ATPases that have been detected by native gel electrophoresis in mild detergents (13Schagger H. Pfeiffer K. EMBO J. 2000; 19: 1777-1783Crossref PubMed Scopus (1036) Google Scholar, 15Arnold I. Pfeiffer K. Neupert W. Stuart R.A. Schagger H. EMBO J. 1998; 17: 7170-7178Crossref PubMed Scopus (364) Google Scholar). The dimeric yeast enzyme contains membrane bound subunits (e and g) that are not present in the monomer, and they appear to be necessary for dimerization to take place. However, the possible involvement of the inhibitor protein (and of two other proteins known as 9 and 15K proteins, which are both required for the action of the yeast IF1) have not been investigated. It may be that dimeric F1F0-ATP synthase forms specific supramolecular complexes with the respiratory complexes in the inner mitochondrial membrane (13Schagger H. Pfeiffer K. EMBO J. 2000; 19: 1777-1783Crossref PubMed Scopus (1036) Google Scholar). We thank M. Montgomery for excellent technical assistance.