Membrane Topography of the Coupling Ion Binding Site in Na+-translocating F1F0 ATP Synthase
2002; Elsevier BV; Volume: 277; Issue: 5 Linguagem: Inglês
10.1074/jbc.m110301200
ISSN1083-351X
AutoresChristoph von Ballmoos, Yvonne Appoldt, Josef Brunner, T. Granier, Andrea Vasella, Peter Dimroth,
Tópico(s)Cancer, Hypoxia, and Metabolism
ResumoA carbodiimide with a photoactivatable diazirine substituent was synthesized and incubated with the Na+-translocating F1F0 ATP synthase from both Propionigenium modestum andIlyobacter tartaricus. This caused severe inhibition of ATP hydrolysis activity in the absence of Na+ ions but not in its presence, indicating the specific reaction with the Na+binding c-Glu65 residue. Photocross-linking was investigated with the substituted ATP synthase from both bacteria in reconstituted 1-palmitoyl-2-oleyl-sn-glycero-3-phosphocholine (POPC)-containing proteoliposomes. A subunit c/POPC conjugate was found in the illuminated samples but no a-c cross-links were observed, not even after ATP-induced rotation of the c-ring. Our substituted diazirine moiety on c-Glu65 was therefore in close contact with phospholipid but does not contact subunit a. Na+in/22Na+outexchange activity of the ATP synthase was not affected by modifying the c-Glu65 sites with the carbodiimide, but upon photoinduced cross-linking, this activity was abolished. Cross-linking the rotor to lipids apparently arrested rotational mobility required for moving Na+ ions back and forth across the membrane. The site of cross-linking was analyzed by digestions of the substituted POPC using phospolipases C and A2 and by mass spectroscopy. The substitutions were found exclusively at the fatty acid side chains, which indicates that c-Glu65 is located within the core of the membrane. A carbodiimide with a photoactivatable diazirine substituent was synthesized and incubated with the Na+-translocating F1F0 ATP synthase from both Propionigenium modestum andIlyobacter tartaricus. This caused severe inhibition of ATP hydrolysis activity in the absence of Na+ ions but not in its presence, indicating the specific reaction with the Na+binding c-Glu65 residue. Photocross-linking was investigated with the substituted ATP synthase from both bacteria in reconstituted 1-palmitoyl-2-oleyl-sn-glycero-3-phosphocholine (POPC)-containing proteoliposomes. A subunit c/POPC conjugate was found in the illuminated samples but no a-c cross-links were observed, not even after ATP-induced rotation of the c-ring. Our substituted diazirine moiety on c-Glu65 was therefore in close contact with phospholipid but does not contact subunit a. Na+in/22Na+outexchange activity of the ATP synthase was not affected by modifying the c-Glu65 sites with the carbodiimide, but upon photoinduced cross-linking, this activity was abolished. Cross-linking the rotor to lipids apparently arrested rotational mobility required for moving Na+ ions back and forth across the membrane. The site of cross-linking was analyzed by digestions of the substituted POPC using phospolipases C and A2 and by mass spectroscopy. The substitutions were found exclusively at the fatty acid side chains, which indicates that c-Glu65 is located within the core of the membrane. dicyclohexylcarbodiimide 1-palmitoyl-2-oleyl-sn-glycero-3-phosphocholine 4-[3-(trifluoromethyl)-3H-diazirin-3-yl]benzyl iodide 4-[3-(trifluoromethyl)-3H-diazirin-3-yl]benzyl azide cyclohexylisocyanate N-4-[3-(trifluoromethyl)-3H-diazirin-3-yl]benzyl-N′-cyclohexylcarbodiimide matrix-assisted laser desorption ionization F1F0 ATP synthases are large protein complexes within the membranes of mitochondria, chloroplasts, or bacteria that use an electrochemical H+ or Na+ gradient across the membrane to synthesize ATP. The F1 portion harboring the catalytic sites for ATP synthesis protrudes from the membrane and has the universal subunit composition α3β3γδε. Its high resolution crystal structure from bovine mitochondria (1Abrahams J.P. Leslie A.G.W. Lutter R. Walker J.E. Nature. 1994; 370: 621-628Crossref PubMed Scopus (2731) Google Scholar) was in remarkable agreement with the binding change mechanism (2Boyer P.D. Annu. Rev. Biochem. 1997; 66: 717-749Crossref PubMed Scopus (1563) Google Scholar), suggesting a rotary catalytic mechanism, which was proven experimentally (3Noji H. Yasuda R. Yoshida M. Kinosita Jr., K. Nature. 1997; 386: 299-302Crossref PubMed Scopus (1934) Google Scholar). The γ and ε subunits form the central stalk protruding from the more compact α3β3 cylinder and make a connection with the oligomeric c-ring of the membrane-intrinsic F0 moiety (4Gibbons C. Montgomery M.G. Leslie A.G. Walker J.E. Nat. Struct. Biol. 2000; 7: 1055-1061Crossref PubMed Scopus (432) Google Scholar, 5Stock D. Leslie A.G. Walker J.E. Science. 1999; 286: 1700-1705Crossref PubMed Scopus (1076) Google Scholar). Subunits γ, ε, and cn could be cross-linked without loss of function (6Tsumoda S.P. Aggeler R. Yoshida M. Capaldi R.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 898-902Crossref PubMed Scopus (137) Google Scholar) and were shown to represent the rotor by direct visualization of rotation with an attached actin filament (3Noji H. Yasuda R. Yoshida M. Kinosita Jr., K. Nature. 1997; 386: 299-302Crossref PubMed Scopus (1934) Google Scholar, 7Sambongi Y. Iko Y. Tanabe M. Omote H. Iwamoto-Kihara A. Ueda I. Yanagida T. Wada Y. Futai M. Science. 1999; 286: 1722-1724Crossref PubMed Scopus (412) Google Scholar,8Pänke O. Gumbiowski K. Junge W. Engelbrecht S. FEBS Lett. 2000; 472: 34-38Crossref PubMed Scopus (178) Google Scholar).Besides cn, the membrane-bound subunit a is part of the F0 motor that is thought to use the electrochemical ion gradient to generate rotary torque (9Elston T. Wang H. Oster G. Nature. 1998; 391: 510-513Crossref PubMed Scopus (438) Google Scholar, 10Dimroth P. Wang H. Grabe M. Oster G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 4924-4929Crossref PubMed Scopus (133) Google Scholar, 11Junge W. Lill H. Engelbrecht S. Trends Biochem. Sci. 1997; 22: 420-423Abstract Full Text PDF PubMed Scopus (432) Google Scholar). As the structure of the a subunit is not known in any detail, its role in the ion translocation and torque-generating mechanism remains speculative. The a subunit is connected laterally with the cn-ring (12Birkenhäger R. Hoppert M. Deckers-Hebestreit G. Mayer F. Altendorf K. Eur. J. Biochem. 1995; 230: 58-67Crossref PubMed Scopus (129) Google Scholar, 13Singh S. Turina P. Bustamante C.J. Keller D.J. Capaldi R. FEBS Lett. 1996; 397: 30-34Crossref PubMed Scopus (105) Google Scholar, 14Takeyasu K. Omote H. Nettikadan S. Tokumasu F. Iwamoto-Kihara A. Futai M. FEBS Lett. 1996; 392: 110-113Crossref PubMed Scopus (116) Google Scholar, 15Jiang W. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 6607-6612Crossref PubMed Scopus (150) Google Scholar), where it is held in place by the two b subunits, which form the peripheral stalk connecting subunit a with an α subunit of F1 with the help of the δ subunit (16Rodgers A.J.W. Capaldi R.A. J. Biol. Chem. 1998; 273: 29406-29410Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar, 17Dunn S.D. Chandler J. J. Biol. Chem. 1998; 273: 8646-8651Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar).Recent structural data have shown that the number of c subunits forming the rotor ring varies among species, being 10 for yeast mitochondria (5Stock D. Leslie A.G. Walker J.E. Science. 1999; 286: 1700-1705Crossref PubMed Scopus (1076) Google Scholar), 14 for spinach chloroplasts (18Seelert H. Poetsch A. Dencher N.A. Engel A. Stahlberg H. Müller D.J. Nature. 2000; 405: 418-419Crossref PubMed Scopus (411) Google Scholar), and 11 for the Na+-translocating ATP synthase from the bacteriumIlyobacter tartaricus (19Stahlberg H. Müller D.J. Suda K. Fotiadis D. Engel A. Meier T. Matthey U. Dimroth P. EMBO Rep. 2001; 2: 229-233Crossref PubMed Scopus (167) Google Scholar). The c oligomer plays a profound role in the ion translocation, and hence, its structure is of primary importance to understand the function of the rotary F0motor. In the Na+-translocating ATP synthases from I. tartaricus (20Neumann S. Matthey U. Kaim G. Dimroth P. J. Bacteriol. 1998; 180: 3312-3316Crossref PubMed Google Scholar) and Propionigenium modestum (21Laubinger W. Dimroth P. Biochemistry. 1988; 27: 7531-7537Crossref PubMed Scopus (161) Google Scholar), Gln32, Glu65, and Ser66 of the c subunits serve as Na+ binding ligands (22Kaim G. Wehrle F. Gerike U. Dimroth P. Biochemistry. 1997; 36: 9185-9194Crossref PubMed Scopus (91) Google Scholar), while equivalents of c-Glu65 serve as proton binding sites in H+-translocating ATP synthases (23Miller J.J. Oldenburg M. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 4900-4904Crossref PubMed Scopus (126) Google Scholar). This acidic residue has been implicated from an NMR structure of monomeric subunit c of E. coli in an organic solvent mixture to be located in the core of the membrane (24Girvin M.E. Rastogi V.K. Abildgaard F. Markley J.L. Fillingame R.H. Biochemistry. 1998; 37: 8817-8824Crossref PubMed Scopus (271) Google Scholar). In a model derived from a secondary NMR structure of subunit c from P. modestum (25Matthey U. Kaim G. Braun D. Wüthrich K. Dimroth P. Eur. J. Biochem. 1999; 261: 459-467Crossref PubMed Scopus (43) Google Scholar), c-Glu65 was placed closer to the membrane surface, explaining numerous data on the direct accessibility of this Na+ binding residue from the aqueous environment (26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar, 27Kaim G. Matthey U. Dimroth P. EMBO J. 1998; 17: 688-695Crossref PubMed Scopus (47) Google Scholar, 28Kaim G. Dimroth P. EMBO J. 1999; 18: 4118-4127Crossref PubMed Scopus (163) Google Scholar).To test these two hypotheses experimentally we took advantage of the fact that c-Glu65 is specifically modified with dicyclohexylcarbodiimide (DCCD).1 We have chemically synthesized a carbodiimide derivative with a photoactivatable diazirine ring, modified c-Glu65 accordingly, and analyzed the cross-link products formed upon illumination in reconstituted proteoliposomes. We show here that cross-linking occurs specifically with the fatty acid side chains of the phospholipids, demonstrating the location of c-Glu65 to be in the core of the cytoplasmic membrane.DISCUSSIONConsiderable data accumulated over the years that the F0 subunits a and cn together make up the membrane-embedded rotary motor of the ATP synthase with the cn-ring rotating versus subunit a (6Tsumoda S.P. Aggeler R. Yoshida M. Capaldi R.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 898-902Crossref PubMed Scopus (137) Google Scholar, 26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar, 27Kaim G. Matthey U. Dimroth P. EMBO J. 1998; 17: 688-695Crossref PubMed Scopus (47) Google Scholar, 41Hutcheon M.L. Duncan T.M. Ngai H. Cross R.L. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8519-8524Crossref PubMed Scopus (63) Google Scholar). Evidence is also available that this rotation and the translocation of the coupling ions across the membrane are intimately associated events so that one cannot occur without the other. Essential insights into this model have been derived from the Na+ translocating F1F0 ATP synthase from P. modestum, which provides unique experimental approaches to follow the translocation of the coupling ions across the cytoplasmic membrane. Each c subunit of the P. modestum ATP synthase carries a Na+ binding site with ligands contributed by Gln32, Glu65, and Ser66 (22Kaim G. Wehrle F. Gerike U. Dimroth P. Biochemistry. 1997; 36: 9185-9194Crossref PubMed Scopus (91) Google Scholar). This site becomes transiently occupied with Na+ during its translocation across the membrane. Evidence indicates a Na+-selective subunit a channel, which connects the adjacent rotor site of c11 with the periplasmic surface, whereas rotor sites not in contact with subunit a have direct access from the cytoplasmic compartment (one-channel model, Ref.10Dimroth P. Wang H. Grabe M. Oster G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 4924-4929Crossref PubMed Scopus (133) Google Scholar). This view is consistent with the ATP-dependent occlusion of 1 Na+ per ATP synthase with a Na+-impermeable a subunit channel (26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar, 27Kaim G. Matthey U. Dimroth P. EMBO J. 1998; 17: 688-695Crossref PubMed Scopus (47) Google Scholar, 42Kaim G. Dimroth P. Biochemistry. 1998; 37: 4626-4634Crossref PubMed Scopus (34) Google Scholar), and it is also consistent with the catalysis of Na+in/22Na+outexchange by F1F0 of P. modestum(26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar). As has been pointed out previously (44Dimroth P. Matthey U. Kaim G. J. Bioenerg. Biomembr. 2000; 32: 449-458Crossref PubMed Scopus (3) Google Scholar), these observations are not compatible, however, with models proposing two half-channels in subunit a (two-channel model, Ref. 11Junge W. Lill H. Engelbrecht S. Trends Biochem. Sci. 1997; 22: 420-423Abstract Full Text PDF PubMed Scopus (432) Google Scholar) through which the rotor sites communicate with the two different sites of the membrane. Another important difference is that the model for torque generation by the F0 motor proposed in Ref. 10Dimroth P. Wang H. Grabe M. Oster G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 4924-4929Crossref PubMed Scopus (133) Google Scholar is the only one that takes the essential role of the membrane potential into account (26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar,28Kaim G. Dimroth P. EMBO J. 1999; 18: 4118-4127Crossref PubMed Scopus (163) Google Scholar).Here, we synthesized a photoactivatable carbodiimide (diazirine-BCCD,4), which reacted specifically with c-Glu65. Partial modification of the rotor sites with this compound blocked ATP hydrolysis and Na+ pumping but not Na+in/22Na+outexchange, as expected. Upon illumination, however, the exchange activity was abolished, which suggests that the rotor becomes immobilized through cross-linking to phospholipids. These proved to be the targets of the photochemical reaction, and hence, the c-Glu65 site that carries the photocross-linker must be exposed toward the phospholipids. This is in accord with models in which the C-terminal helices of the c-subunits are on the outside of the ring (5Stock D. Leslie A.G. Walker J.E. Science. 1999; 286: 1700-1705Crossref PubMed Scopus (1076) Google Scholar, 43Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (105) Google Scholar).We clearly demonstrate that the cross-linker reacted exclusively with the fatty acid side chains of the phospholipids. This indicates that the topography of c-Glu65 is within the core of the membrane. The length of the attached cross-linker is about 8 Å when fully extended, and therefore the distance of c-Glu65 to the glycerol moiety of POPC must be at least in the same range to explain the absence of cross-link formation with this portion of the molecule (Fig. 8A). This result is astonishing given the overwhelming evidence for direct access of c-Glu65 from the aqueous compartment by Na+ (44Dimroth P. Matthey U. Kaim G. J. Bioenerg. Biomembr. 2000; 32: 449-458Crossref PubMed Scopus (3) Google Scholar). It is in good agreement, however, with an Escherichia coli model of the topography of subunit c within the membrane (43Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (105) Google Scholar), which is compatible with the structure of the c oligomer from yeast (5Stock D. Leslie A.G. Walker J.E. Science. 1999; 286: 1700-1705Crossref PubMed Scopus (1076) Google Scholar).The location of the c subunit binding sites within the membrane core is also compatible with unpublished structural data on the c11oligomer from I. tartaricus. In this structure, a tightly associated inner ring comprising the N-terminal helices is surrounded by an outer ring comprising the C-terminal helices. The outer helices are positioned within the grooves formed by the inner ring of helices leaving enough space between them to form access channels for the coupling ions to reach the membrane buried c-Glu65 residue from the aqueous compartment. Hence, agreement has now been reached on the position of c-Glu65 near the center of the membrane. This location, however, by no means decides in favor of the two-channel model. Our evidence for the direct accessibility of c-Glu65in case of the Na+-translocating ATP synthase (see above) rather warrants a modification of the one-channel model as shown in Fig. 8B. Based on the findings presented here and elsewhere, we propose that Na+ ions enter the a subunit channel from the periplasmic reservoir and pass through it onto the adjacent rotor site, which is near the center of the membrane. From this position, the ion passes toward the cytoplasmic surface through a channel formed between an inner and two outer helices of the c11-ring after the rotor has turned the site out of the interface with the a channel. According to this new concept, one may want to term our model the 1a+11c-channel model rather than the one-channel model. F1F0 ATP synthases are large protein complexes within the membranes of mitochondria, chloroplasts, or bacteria that use an electrochemical H+ or Na+ gradient across the membrane to synthesize ATP. The F1 portion harboring the catalytic sites for ATP synthesis protrudes from the membrane and has the universal subunit composition α3β3γδε. Its high resolution crystal structure from bovine mitochondria (1Abrahams J.P. Leslie A.G.W. Lutter R. Walker J.E. Nature. 1994; 370: 621-628Crossref PubMed Scopus (2731) Google Scholar) was in remarkable agreement with the binding change mechanism (2Boyer P.D. Annu. Rev. Biochem. 1997; 66: 717-749Crossref PubMed Scopus (1563) Google Scholar), suggesting a rotary catalytic mechanism, which was proven experimentally (3Noji H. Yasuda R. Yoshida M. Kinosita Jr., K. Nature. 1997; 386: 299-302Crossref PubMed Scopus (1934) Google Scholar). The γ and ε subunits form the central stalk protruding from the more compact α3β3 cylinder and make a connection with the oligomeric c-ring of the membrane-intrinsic F0 moiety (4Gibbons C. Montgomery M.G. Leslie A.G. Walker J.E. Nat. Struct. Biol. 2000; 7: 1055-1061Crossref PubMed Scopus (432) Google Scholar, 5Stock D. Leslie A.G. Walker J.E. Science. 1999; 286: 1700-1705Crossref PubMed Scopus (1076) Google Scholar). Subunits γ, ε, and cn could be cross-linked without loss of function (6Tsumoda S.P. Aggeler R. Yoshida M. Capaldi R.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 898-902Crossref PubMed Scopus (137) Google Scholar) and were shown to represent the rotor by direct visualization of rotation with an attached actin filament (3Noji H. Yasuda R. Yoshida M. Kinosita Jr., K. Nature. 1997; 386: 299-302Crossref PubMed Scopus (1934) Google Scholar, 7Sambongi Y. Iko Y. Tanabe M. Omote H. Iwamoto-Kihara A. Ueda I. Yanagida T. Wada Y. Futai M. Science. 1999; 286: 1722-1724Crossref PubMed Scopus (412) Google Scholar,8Pänke O. Gumbiowski K. Junge W. Engelbrecht S. FEBS Lett. 2000; 472: 34-38Crossref PubMed Scopus (178) Google Scholar). Besides cn, the membrane-bound subunit a is part of the F0 motor that is thought to use the electrochemical ion gradient to generate rotary torque (9Elston T. Wang H. Oster G. Nature. 1998; 391: 510-513Crossref PubMed Scopus (438) Google Scholar, 10Dimroth P. Wang H. Grabe M. Oster G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 4924-4929Crossref PubMed Scopus (133) Google Scholar, 11Junge W. Lill H. Engelbrecht S. Trends Biochem. Sci. 1997; 22: 420-423Abstract Full Text PDF PubMed Scopus (432) Google Scholar). As the structure of the a subunit is not known in any detail, its role in the ion translocation and torque-generating mechanism remains speculative. The a subunit is connected laterally with the cn-ring (12Birkenhäger R. Hoppert M. Deckers-Hebestreit G. Mayer F. Altendorf K. Eur. J. Biochem. 1995; 230: 58-67Crossref PubMed Scopus (129) Google Scholar, 13Singh S. Turina P. Bustamante C.J. Keller D.J. Capaldi R. FEBS Lett. 1996; 397: 30-34Crossref PubMed Scopus (105) Google Scholar, 14Takeyasu K. Omote H. Nettikadan S. Tokumasu F. Iwamoto-Kihara A. Futai M. FEBS Lett. 1996; 392: 110-113Crossref PubMed Scopus (116) Google Scholar, 15Jiang W. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 6607-6612Crossref PubMed Scopus (150) Google Scholar), where it is held in place by the two b subunits, which form the peripheral stalk connecting subunit a with an α subunit of F1 with the help of the δ subunit (16Rodgers A.J.W. Capaldi R.A. J. Biol. Chem. 1998; 273: 29406-29410Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar, 17Dunn S.D. Chandler J. J. Biol. Chem. 1998; 273: 8646-8651Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar). Recent structural data have shown that the number of c subunits forming the rotor ring varies among species, being 10 for yeast mitochondria (5Stock D. Leslie A.G. Walker J.E. Science. 1999; 286: 1700-1705Crossref PubMed Scopus (1076) Google Scholar), 14 for spinach chloroplasts (18Seelert H. Poetsch A. Dencher N.A. Engel A. Stahlberg H. Müller D.J. Nature. 2000; 405: 418-419Crossref PubMed Scopus (411) Google Scholar), and 11 for the Na+-translocating ATP synthase from the bacteriumIlyobacter tartaricus (19Stahlberg H. Müller D.J. Suda K. Fotiadis D. Engel A. Meier T. Matthey U. Dimroth P. EMBO Rep. 2001; 2: 229-233Crossref PubMed Scopus (167) Google Scholar). The c oligomer plays a profound role in the ion translocation, and hence, its structure is of primary importance to understand the function of the rotary F0motor. In the Na+-translocating ATP synthases from I. tartaricus (20Neumann S. Matthey U. Kaim G. Dimroth P. J. Bacteriol. 1998; 180: 3312-3316Crossref PubMed Google Scholar) and Propionigenium modestum (21Laubinger W. Dimroth P. Biochemistry. 1988; 27: 7531-7537Crossref PubMed Scopus (161) Google Scholar), Gln32, Glu65, and Ser66 of the c subunits serve as Na+ binding ligands (22Kaim G. Wehrle F. Gerike U. Dimroth P. Biochemistry. 1997; 36: 9185-9194Crossref PubMed Scopus (91) Google Scholar), while equivalents of c-Glu65 serve as proton binding sites in H+-translocating ATP synthases (23Miller J.J. Oldenburg M. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 4900-4904Crossref PubMed Scopus (126) Google Scholar). This acidic residue has been implicated from an NMR structure of monomeric subunit c of E. coli in an organic solvent mixture to be located in the core of the membrane (24Girvin M.E. Rastogi V.K. Abildgaard F. Markley J.L. Fillingame R.H. Biochemistry. 1998; 37: 8817-8824Crossref PubMed Scopus (271) Google Scholar). In a model derived from a secondary NMR structure of subunit c from P. modestum (25Matthey U. Kaim G. Braun D. Wüthrich K. Dimroth P. Eur. J. Biochem. 1999; 261: 459-467Crossref PubMed Scopus (43) Google Scholar), c-Glu65 was placed closer to the membrane surface, explaining numerous data on the direct accessibility of this Na+ binding residue from the aqueous environment (26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar, 27Kaim G. Matthey U. Dimroth P. EMBO J. 1998; 17: 688-695Crossref PubMed Scopus (47) Google Scholar, 28Kaim G. Dimroth P. EMBO J. 1999; 18: 4118-4127Crossref PubMed Scopus (163) Google Scholar). To test these two hypotheses experimentally we took advantage of the fact that c-Glu65 is specifically modified with dicyclohexylcarbodiimide (DCCD).1 We have chemically synthesized a carbodiimide derivative with a photoactivatable diazirine ring, modified c-Glu65 accordingly, and analyzed the cross-link products formed upon illumination in reconstituted proteoliposomes. We show here that cross-linking occurs specifically with the fatty acid side chains of the phospholipids, demonstrating the location of c-Glu65 to be in the core of the cytoplasmic membrane. DISCUSSIONConsiderable data accumulated over the years that the F0 subunits a and cn together make up the membrane-embedded rotary motor of the ATP synthase with the cn-ring rotating versus subunit a (6Tsumoda S.P. Aggeler R. Yoshida M. Capaldi R.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 898-902Crossref PubMed Scopus (137) Google Scholar, 26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar, 27Kaim G. Matthey U. Dimroth P. EMBO J. 1998; 17: 688-695Crossref PubMed Scopus (47) Google Scholar, 41Hutcheon M.L. Duncan T.M. Ngai H. Cross R.L. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8519-8524Crossref PubMed Scopus (63) Google Scholar). Evidence is also available that this rotation and the translocation of the coupling ions across the membrane are intimately associated events so that one cannot occur without the other. Essential insights into this model have been derived from the Na+ translocating F1F0 ATP synthase from P. modestum, which provides unique experimental approaches to follow the translocation of the coupling ions across the cytoplasmic membrane. Each c subunit of the P. modestum ATP synthase carries a Na+ binding site with ligands contributed by Gln32, Glu65, and Ser66 (22Kaim G. Wehrle F. Gerike U. Dimroth P. Biochemistry. 1997; 36: 9185-9194Crossref PubMed Scopus (91) Google Scholar). This site becomes transiently occupied with Na+ during its translocation across the membrane. Evidence indicates a Na+-selective subunit a channel, which connects the adjacent rotor site of c11 with the periplasmic surface, whereas rotor sites not in contact with subunit a have direct access from the cytoplasmic compartment (one-channel model, Ref.10Dimroth P. Wang H. Grabe M. Oster G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 4924-4929Crossref PubMed Scopus (133) Google Scholar). This view is consistent with the ATP-dependent occlusion of 1 Na+ per ATP synthase with a Na+-impermeable a subunit channel (26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar, 27Kaim G. Matthey U. Dimroth P. EMBO J. 1998; 17: 688-695Crossref PubMed Scopus (47) Google Scholar, 42Kaim G. Dimroth P. Biochemistry. 1998; 37: 4626-4634Crossref PubMed Scopus (34) Google Scholar), and it is also consistent with the catalysis of Na+in/22Na+outexchange by F1F0 of P. modestum(26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar). As has been pointed out previously (44Dimroth P. Matthey U. Kaim G. J. Bioenerg. Biomembr. 2000; 32: 449-458Crossref PubMed Scopus (3) Google Scholar), these observations are not compatible, however, with models proposing two half-channels in subunit a (two-channel model, Ref. 11Junge W. Lill H. Engelbrecht S. Trends Biochem. Sci. 1997; 22: 420-423Abstract Full Text PDF PubMed Scopus (432) Google Scholar) through which the rotor sites communicate with the two different sites of the membrane. Another important difference is that the model for torque generation by the F0 motor proposed in Ref. 10Dimroth P. Wang H. Grabe M. Oster G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 4924-4929Crossref PubMed Scopus (133) Google Scholar is the only one that takes the essential role of the membrane potential into account (26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar,28Kaim G. Dimroth P. EMBO J. 1999; 18: 4118-4127Crossref PubMed Scopus (163) Google Scholar).Here, we synthesized a photoactivatable carbodiimide (diazirine-BCCD,4), which reacted specifically with c-Glu65. Partial modification of the rotor sites with this compound blocked ATP hydrolysis and Na+ pumping but not Na+in/22Na+outexchange, as expected. Upon illumination, however, the exchange activity was abolished, which suggests that the rotor becomes immobilized through cross-linking to phospholipids. These proved to be the targets of the photochemical reaction, and hence, the c-Glu65 site that carries the photocross-linker must be exposed toward the phospholipids. This is in accord with models in which the C-terminal helices of the c-subunits are on the outside of the ring (5Stock D. Leslie A.G. Walker J.E. Science. 1999; 286: 1700-1705Crossref PubMed Scopus (1076) Google Scholar, 43Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (105) Google Scholar).We clearly demonstrate that the cross-linker reacted exclusively with the fatty acid side chains of the phospholipids. This indicates that the topography of c-Glu65 is within the core of the membrane. The length of the attached cross-linker is about 8 Å when fully extended, and therefore the distance of c-Glu65 to the glycerol moiety of POPC must be at least in the same range to explain the absence of cross-link formation with this portion of the molecule (Fig. 8A). This result is astonishing given the overwhelming evidence for direct access of c-Glu65 from the aqueous compartment by Na+ (44Dimroth P. Matthey U. Kaim G. J. Bioenerg. Biomembr. 2000; 32: 449-458Crossref PubMed Scopus (3) Google Scholar). It is in good agreement, however, with an Escherichia coli model of the topography of subunit c within the membrane (43Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (105) Google Scholar), which is compatible with the structure of the c oligomer from yeast (5Stock D. Leslie A.G. Walker J.E. Science. 1999; 286: 1700-1705Crossref PubMed Scopus (1076) Google Scholar).The location of the c subunit binding sites within the membrane core is also compatible with unpublished structural data on the c11oligomer from I. tartaricus. In this structure, a tightly associated inner ring comprising the N-terminal helices is surrounded by an outer ring comprising the C-terminal helices. The outer helices are positioned within the grooves formed by the inner ring of helices leaving enough space between them to form access channels for the coupling ions to reach the membrane buried c-Glu65 residue from the aqueous compartment. Hence, agreement has now been reached on the position of c-Glu65 near the center of the membrane. This location, however, by no means decides in favor of the two-channel model. Our evidence for the direct accessibility of c-Glu65in case of the Na+-translocating ATP synthase (see above) rather warrants a modification of the one-channel model as shown in Fig. 8B. Based on the findings presented here and elsewhere, we propose that Na+ ions enter the a subunit channel from the periplasmic reservoir and pass through it onto the adjacent rotor site, which is near the center of the membrane. From this position, the ion passes toward the cytoplasmic surface through a channel formed between an inner and two outer helices of the c11-ring after the rotor has turned the site out of the interface with the a channel. According to this new concept, one may want to term our model the 1a+11c-channel model rather than the one-channel model. Considerable data accumulated over the years that the F0 subunits a and cn together make up the membrane-embedded rotary motor of the ATP synthase with the cn-ring rotating versus subunit a (6Tsumoda S.P. Aggeler R. Yoshida M. Capaldi R.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 898-902Crossref PubMed Scopus (137) Google Scholar, 26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar, 27Kaim G. Matthey U. Dimroth P. EMBO J. 1998; 17: 688-695Crossref PubMed Scopus (47) Google Scholar, 41Hutcheon M.L. Duncan T.M. Ngai H. Cross R.L. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8519-8524Crossref PubMed Scopus (63) Google Scholar). Evidence is also available that this rotation and the translocation of the coupling ions across the membrane are intimately associated events so that one cannot occur without the other. Essential insights into this model have been derived from the Na+ translocating F1F0 ATP synthase from P. modestum, which provides unique experimental approaches to follow the translocation of the coupling ions across the cytoplasmic membrane. Each c subunit of the P. modestum ATP synthase carries a Na+ binding site with ligands contributed by Gln32, Glu65, and Ser66 (22Kaim G. Wehrle F. Gerike U. Dimroth P. Biochemistry. 1997; 36: 9185-9194Crossref PubMed Scopus (91) Google Scholar). This site becomes transiently occupied with Na+ during its translocation across the membrane. Evidence indicates a Na+-selective subunit a channel, which connects the adjacent rotor site of c11 with the periplasmic surface, whereas rotor sites not in contact with subunit a have direct access from the cytoplasmic compartment (one-channel model, Ref.10Dimroth P. Wang H. Grabe M. Oster G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 4924-4929Crossref PubMed Scopus (133) Google Scholar). This view is consistent with the ATP-dependent occlusion of 1 Na+ per ATP synthase with a Na+-impermeable a subunit channel (26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar, 27Kaim G. Matthey U. Dimroth P. EMBO J. 1998; 17: 688-695Crossref PubMed Scopus (47) Google Scholar, 42Kaim G. Dimroth P. Biochemistry. 1998; 37: 4626-4634Crossref PubMed Scopus (34) Google Scholar), and it is also consistent with the catalysis of Na+in/22Na+outexchange by F1F0 of P. modestum(26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar). As has been pointed out previously (44Dimroth P. Matthey U. Kaim G. J. Bioenerg. Biomembr. 2000; 32: 449-458Crossref PubMed Scopus (3) Google Scholar), these observations are not compatible, however, with models proposing two half-channels in subunit a (two-channel model, Ref. 11Junge W. Lill H. Engelbrecht S. Trends Biochem. Sci. 1997; 22: 420-423Abstract Full Text PDF PubMed Scopus (432) Google Scholar) through which the rotor sites communicate with the two different sites of the membrane. Another important difference is that the model for torque generation by the F0 motor proposed in Ref. 10Dimroth P. Wang H. Grabe M. Oster G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 4924-4929Crossref PubMed Scopus (133) Google Scholar is the only one that takes the essential role of the membrane potential into account (26Kaim G. Dimroth P. EMBO J. 1998; 17: 5887-5895Crossref PubMed Scopus (71) Google Scholar,28Kaim G. Dimroth P. EMBO J. 1999; 18: 4118-4127Crossref PubMed Scopus (163) Google Scholar). Here, we synthesized a photoactivatable carbodiimide (diazirine-BCCD,4), which reacted specifically with c-Glu65. Partial modification of the rotor sites with this compound blocked ATP hydrolysis and Na+ pumping but not Na+in/22Na+outexchange, as expected. Upon illumination, however, the exchange activity was abolished, which suggests that the rotor becomes immobilized through cross-linking to phospholipids. These proved to be the targets of the photochemical reaction, and hence, the c-Glu65 site that carries the photocross-linker must be exposed toward the phospholipids. This is in accord with models in which the C-terminal helices of the c-subunits are on the outside of the ring (5Stock D. Leslie A.G. Walker J.E. Science. 1999; 286: 1700-1705Crossref PubMed Scopus (1076) Google Scholar, 43Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (105) Google Scholar). We clearly demonstrate that the cross-linker reacted exclusively with the fatty acid side chains of the phospholipids. This indicates that the topography of c-Glu65 is within the core of the membrane. The length of the attached cross-linker is about 8 Å when fully extended, and therefore the distance of c-Glu65 to the glycerol moiety of POPC must be at least in the same range to explain the absence of cross-link formation with this portion of the molecule (Fig. 8A). This result is astonishing given the overwhelming evidence for direct access of c-Glu65 from the aqueous compartment by Na+ (44Dimroth P. Matthey U. Kaim G. J. Bioenerg. Biomembr. 2000; 32: 449-458Crossref PubMed Scopus (3) Google Scholar). It is in good agreement, however, with an Escherichia coli model of the topography of subunit c within the membrane (43Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (105) Google Scholar), which is compatible with the structure of the c oligomer from yeast (5Stock D. Leslie A.G. Walker J.E. Science. 1999; 286: 1700-1705Crossref PubMed Scopus (1076) Google Scholar). The location of the c subunit binding sites within the membrane core is also compatible with unpublished structural data on the c11oligomer from I. tartaricus. In this structure, a tightly associated inner ring comprising the N-terminal helices is surrounded by an outer ring comprising the C-terminal helices. The outer helices are positioned within the grooves formed by the inner ring of helices leaving enough space between them to form access channels for the coupling ions to reach the membrane buried c-Glu65 residue from the aqueous compartment. Hence, agreement has now been reached on the position of c-Glu65 near the center of the membrane. This location, however, by no means decides in favor of the two-channel model. Our evidence for the direct accessibility of c-Glu65in case of the Na+-translocating ATP synthase (see above) rather warrants a modification of the one-channel model as shown in Fig. 8B. Based on the findings presented here and elsewhere, we propose that Na+ ions enter the a subunit channel from the periplasmic reservoir and pass through it onto the adjacent rotor site, which is near the center of the membrane. From this position, the ion passes toward the cytoplasmic surface through a channel formed between an inner and two outer helices of the c11-ring after the rotor has turned the site out of the interface with the a channel. According to this new concept, one may want to term our model the 1a+11c-channel model rather than the one-channel model. We thank Georg Kaim for advice and Gregory Cook for reading the manuscript.
Referência(s)