Structure of Ala20 → Pro/Pro64 → Ala Substituted Subunit c ofEscherichia coli ATP Synthase in Which the Essential Proline Is Switched between Transmembrane Helices

Artigo Acesso aberto Revisado por pares

Structure of Ala20 → Pro/Pro64 → Ala Substituted Subunit c ofEscherichia coli ATP Synthase in Which the Essential Proline Is Switched between Transmembrane Helices

2001; Elsevier BV; Volume: 276; Issue: 29 Linguagem: Inglês

10.1074/jbc.m100762200

ISSN

1083-351X

Autores

Oleg Y. Dmitriev, Robert Fillingame,

Tópico(s)

Advanced Electron Microscopy Techniques and Applications

Resumo

The structure of the A20P/P64A mutated subunitc of Escherichia coli ATP synthase, in which the essential proline has been switched from residue 64 of the second transmembrane helix (TMH) to residue 20 of the first TMH, has been solved by 15N,1H NMR in a monophasic chloroform/methanol/water (4:4:1) solvent mixture. ThecA20P/P64A mutant grows as well as wild type, and the F0F1 complex is fully functional in ATPase-coupled H+ pumping. Residues 20 and 64 lie directly opposite to each other in the hairpin-like structure of wild type subunit c, and the prolinyl 64 residue is thought to induce a slight bend in TMH-2 such that it wraps around a more straightened TMH-1. In solution, the A20P/P64A substituted subunit calso forms a hairpin of two α-helices, with residues 41–45 forming a connecting loop as in the case of the wild type protein, but, in this case, Pro20 induces a bend in TMH-1, which then packs against a more straightened TMH-2. The essential prolinyl residue, whether at position 64 or 20, lies close to the aspartyl 61 H+ binding site. The prolinyl residue may introduce structural flexibility in this region of the protein, which may be necessary for the proposed movement of the α-helical segments during the course of the H+ pumping catalytic cycle. The structure of the A20P/P64A mutated subunitc of Escherichia coli ATP synthase, in which the essential proline has been switched from residue 64 of the second transmembrane helix (TMH) to residue 20 of the first TMH, has been solved by 15N,1H NMR in a monophasic chloroform/methanol/water (4:4:1) solvent mixture. ThecA20P/P64A mutant grows as well as wild type, and the F0F1 complex is fully functional in ATPase-coupled H+ pumping. Residues 20 and 64 lie directly opposite to each other in the hairpin-like structure of wild type subunit c, and the prolinyl 64 residue is thought to induce a slight bend in TMH-2 such that it wraps around a more straightened TMH-1. In solution, the A20P/P64A substituted subunit calso forms a hairpin of two α-helices, with residues 41–45 forming a connecting loop as in the case of the wild type protein, but, in this case, Pro20 induces a bend in TMH-1, which then packs against a more straightened TMH-2. The essential prolinyl residue, whether at position 64 or 20, lies close to the aspartyl 61 H+ binding site. The prolinyl residue may introduce structural flexibility in this region of the protein, which may be necessary for the proposed movement of the α-helical segments during the course of the H+ pumping catalytic cycle. transmembrane helix double quantum-filtered correlation spectroscopy NOESY, NOE spectroscopy total correlation spectroscopy heteronuclear single quantum coherence transfer experiment root mean square F0F1 ATP synthases reversibly convert the energy of a transmembrane H+ electrochemical gradient, generated by electron transport systems in mitochondria, chloroplasts, and the plasma membrane of eubacteria, into the chemical energy of a phosphate anhydride bond in ATP (1Senior A.E. Physiol. Rev. 1988; 68: 177-231Crossref PubMed Scopus (463) Google Scholar, 2Nakamoto R.K. Ketchum C.J. Al-Shawi M.K. Annu. Rev. Biophys. Biomol. Struct. 1999; 28: 205-234Crossref PubMed Scopus (104) Google Scholar). Proton translocation through the transmembrane F0 sector is reversibly coupled to ATP synthesis or hydrolysis in the F1 sector of the complex, which extends from the surface of the membrane. The subunit composition of the simplest bacterial enzymes, such as in Escherichia coli, is α3β3γδε for F1 anda 1 b 2 c 10–14for F0 (3Fillingame R.H. Science. 1999; 286: 1687-1688Crossref PubMed Scopus (58) Google Scholar). High resolution crystal structures of the nearly complete F1 complex from bovine mitochondria support previous conclusions that the catalytic sites alter between open and closed states as predicted by the rotary binding change mechanism (4Gibbons Walker J.E. Nat. Struct. Biol. 2000; 7: 1055-1061Crossref PubMed Scopus (437) Google Scholar, 5Abrahams J.P. Leslie A.G.W. Lutter R. Walker J.E. Nature. 1994; 370: 621-628Crossref PubMed Scopus (2764) Google Scholar, 6Boyer P.D. Annu. Rev. Biochem. 1997; 66: 717-749Crossref PubMed Scopus (1606) Google Scholar). The alternate closing and opening of catalytic sites, with the consequent tight binding of substrates and release of products, is proposed to be driven by rotation of the γ subunit within a hexameric ring of alternately packed α and β subunits, and direct evidence for such ATPase-coupled rotation is now widely accepted (7Noji H. Yasuda R. Yoshida M. Kinosita Jr., K. Nature. 1997; 386: 299-302Crossref PubMed Scopus (1974) Google Scholar). In the membrane, the rotation of subunit γ is proposed in turn to be driven by H+ transport-coupled rotation of a connected ring ofc subunits, which extend through the lipid bilayer (8Vik S.B. Antonio B.J. J. Biol. Chem. 1994; 269: 30364-30369Abstract Full Text PDF PubMed Google Scholar, 9Duncan 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 (461) Google Scholar, 10Engelbrecht S. Junge W. FEBS Lett. 1997; 414: 485-491Crossref PubMed Scopus (118) Google Scholar, 11Elston T. Wang H. Oster G. Nature. 1998; 391: 510-513Crossref PubMed Scopus (448) Google Scholar). Recent evidence indicates that the connection between the γ and ε subunits of F1 and the c-ring remains fixed during function (12Schulenberg B. Aggeler R. Murray J. Capaldi R.A. J. Biol. Chem. 1999; 274: 34233-34237Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar, 13Jones P.C. Hermolin J. Fillingame R.H. J. Biol. Chem. 2000; 275: 11355-11360Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar), and independent experiments have now directly demonstrated a rotation of the c-ring coupled with ATP hydrolysis (14Sambongi 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 (421) Google Scholar, 15Pänke O. Gumbiowski K. Junge W. Engelbrecht S. FEBS Lett. 2000; 472: 34-38Crossref PubMed Scopus (180) Google Scholar). The detailed structure of the F0 complex is unknown, and the mechanism by which proton transport drives rotation remains speculative. The structure of monomeric subunit c (16Girvin M.E. Rastogi V.K. Abildgaard F. Markley J.L. Fillingame R.H. Biochemistry. 1998; 37: 8817-8824Crossref PubMed Scopus (272) Google Scholar) and of the membrane domain of the subunit b (17Dmitriev O.Y. Jones P.C. Jiang W. Fillingame R.H. J. Biol. Chem. 1999; 274: 15598-15604Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar) have been solved by NMR in a membrane-mimetic solvent, i.e. a monophasic mixture of chloroform/methanol/H2O (4:4:1). The current model of the subunit arrangement in the F0 complex emerges from the results of electron and atomic force microscopy (18Birkenhaeger R. Hoppert M. Deckers-Hebestreit G. Mayer F. Altendorf K. Eur. J. Biochem. 1995; 230: 58-67Crossref PubMed Scopus (129) Google Scholar, 19Takeyasu K. Omote H. Nettikadan S. Tokumasu F. Iwamoto-Kihara A Futai M. FEBS Lett. 1996; 392: 110-113Crossref PubMed Scopus (117) Google Scholar, 20Singh S. Turina P. Bustamante C.J. Keller D.J. Capaldi R.A. FEBS Lett. 1996; 397: 30-34Crossref PubMed Scopus (105) Google Scholar), intersubunit Cys-Cys cross-linking (21Jones P.C. Jiang W. Fillingame R.H. J. Biol. Chem. 1998; 273: 17175-17178Google Scholar, 22Jiang W. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 6607-6612Crossref PubMed Scopus (150) Google Scholar), molecular modeling (23Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (107) Google Scholar, 24Rastogi V.K. Girvin M.E. Nature. 1999; 402: 263-268Crossref PubMed Scopus (418) Google Scholar), and most recently direct crystallographic observations (25Stock D. Leslie A.G.W. Walker J.E. Science. 1999; 286: 1700-1704Crossref PubMed Scopus (1093) Google Scholar). In the NMR model (16Girvin M.E. Rastogi V.K. Abildgaard F. Markley J.L. Fillingame R.H. Biochemistry. 1998; 37: 8817-8824Crossref PubMed Scopus (272) Google Scholar), subunit c folds as a hairpin of two helices, connected by a short polar loop, with helix-helix interactions in accord with those expected for the protein in the native membrane. The polar loop is known to interact with the γ and ε subunits of F1 in the native enzyme (26Zhang Y. Fillingame R.H. J. Biol. Chem. 1995; 270: 24609-24614Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar, 27Watts S.D. Teng C. Capaldi R.A. J. Biol. Chem. 1996; 271: 28341-28347Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar, 28Hermolin J. Dmitriev O.Y. Zhang Y. Fillingame R.H. J. Biol. Chem. 1999; 274: 17011-17016Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar). The number of c subunits in the complex is presently the subject of controversy, with numbers in the range of 10–14 being proposed for different species (25Stock D. Leslie A.G.W. Walker J.E. Science. 1999; 286: 1700-1704Crossref PubMed Scopus (1093) Google Scholar, 29Jones P.C. Fillingame R.H. J. Biol. Chem. 1998; 273: 29701-29705Abstract Full Text Full Text PDF PubMed Scopus (111) Google Scholar, 30Seelert H. Poetsch A. Dencher N.A. Engel A. Stahlberg H. Müller D.J. Nature. 2000; 405: 418-419Crossref PubMed Scopus (415) Google Scholar). The multiple copies of subunit c are proposed to be arranged in a cylinder such that the TMHs1 form two concentric rings with the N- and C-terminal helices positioned in the inner and outer circle of the ring, respectively. Subunits aand b are located outside the ring perimeter, subunita being in contact with TMH-2 of the subunit c(22Jiang W. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 6607-6612Crossref PubMed Scopus (150) Google Scholar). Subunit b is present in two copies and probably serves as the stalk of the stator holding the α3β3subunits of F1 in a fixed position relative to the transmembrane regions of subunits a and b(31Dunn S.D. McLachlin D.T. Revington M. Biochim. Biophys. Acta. 2000; 1458: 356-363Crossref PubMed Scopus (82) Google Scholar, 32Capaldi R.A. Schulenberg B. Murray J. Aggeler R. J. Exp. Biol. 2000; 203: 29-33Crossref PubMed Google Scholar, 33Altendorf K. Stalz W.-D. Greie J.-C. Deckers-Hebestreit G. J. Exp. Biol. 2000; 203: 19-28PubMed Google Scholar). The path of H+ transport through F0 and the mechanism of coupling H+ translocation to rotation remain largely unknown. In the course of H+ transport, the proton is assumed to bind transiently to the carboxyl group of the essential Asp61 residue, located about halfway across the hydrophobic slab of the lipid bilayer. An interaction between the a andc subunits, probably involvingaArg210 and cAsp61, is thought to promote deprotonation of Asp61 at the exit side of the membrane (11Elston T. Wang H. Oster G. Nature. 1998; 391: 510-513Crossref PubMed Scopus (448) Google Scholar, 34Valiyaveetil F.I. Fillingame R.H. J. Biol. Chem. 1997; 272: 32635-32641Abstract Full Text Full Text PDF PubMed Scopus (108) Google Scholar, 35Fillingame R.H. Krulwich T.A. The Bacteria. Academic Press, Inc., New York1990: 345-391Google Scholar, 36Fillingame R.H. Jiang W. Dmitriev O.Y. Jones P.C. Biochim. Biophys. Acta. 2000; 1458: 387-403Crossref PubMed Scopus (56) Google Scholar). In models of the c oligomer (23Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (107) Google Scholar, 24Rastogi V.K. Girvin M.E. Nature. 1999; 402: 263-268Crossref PubMed Scopus (418) Google Scholar), the Asp61 carboxyl group appears to be buried at the interface of two neighboring c subunits, so a significant conformational change in the c subunit will be required to allow direct contact between cAsp61and aArg210. Ionization of Asp61 is known to result in global structural changes in the monomeric protein (37Assadi-Porter F.M. Fillingame R.H. Biochemistry. 1995; 34: 16186-16193Crossref PubMed Scopus (63) Google Scholar). The recently published solution structure of subunit cat pH 8 (24Rastogi V.K. Girvin M.E. Nature. 1999; 402: 263-268Crossref PubMed Scopus (418) Google Scholar), where Asp61 is ionized, shows TMH-2 turned by about 140° compared with its orientation in the protonated subunit; major structural changes in the loop region were also observed. The turning of TMH-2 in one or two c subunits of native F0 would provide a more complete explanation of previous cross-linking experiments in which a-c (22Jiang W. Fillingame R.H. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 6607-6612Crossref PubMed Scopus (150) Google Scholar) andc-c (21Jones P.C. Jiang W. Fillingame R.H. J. Biol. Chem. 1998; 273: 17175-17178Google Scholar) dimers were observed. Some of these dimers cannot be accounted for by the simplest oligomeric model utilizing the pH 5 structure only (23Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (107) Google Scholar). Hence, in native F0 the deprotonation of Asp61 was suggested to be linked with the turning of thecTMH-2 at the subunit a interface, and this turning, or its reversal following reprotonation of Asp61, was proposed to be coupled to the stepwise circular movement of thec-oligomeric ring (24Rastogi V.K. Girvin M.E. Nature. 1999; 402: 263-268Crossref PubMed Scopus (418) Google Scholar, 38Fillingame R.H. Jiang W. Dmitriev O.Y. J. Exp. Biol. 2000; 203: 9-17Crossref PubMed Google Scholar). Such conformational movements, occurring within units of a closely packed oligomer, must involve rather complex concerted movements of both helices. The present work bears on the possible role of an essential transmembrane proline residue in such movements. One prominent feature of the subunit c structure at pH 5 is a bend in TMH-2 associated with Pro64 (16Girvin M.E. Rastogi V.K. Abildgaard F. Markley J.L. Fillingame R.H. Biochemistry. 1998; 37: 8817-8824Crossref PubMed Scopus (272) Google Scholar). Substitution of Pro64 with Leu renders the ATP synthase inactive, as evidenced by inability of cells to grow on succinate by oxidative phosphorylation (39Fimmel A.L. Jans D.A. Langman L. James L.B. Ash G.R. Downie J.A. Senior A.E. Gibson F. Cox G.B. Biochem. J. 1983; 213: 451-458Crossref PubMed Scopus (33) Google Scholar). Interestingly, function in the cP64L mutant can be partially restored by a second site A20P mutation (39Fimmel A.L. Jans D.A. Langman L. James L.B. Ash G.R. Downie J.A. Senior A.E. Gibson F. Cox G.B. Biochem. J. 1983; 213: 451-458Crossref PubMed Scopus (33) Google Scholar) and further improved when position 64 is occupied by alanine (40Fraga D. Subunit c of the Escherichia coli Proton Translocating ATP Synthase: New Insights into Residue Function by Mutant Analysis.Ph.D. dissertation. University of Wisconsin, Madison, WI1990Google Scholar). In the hairpin structure of monomeric subunit c, residues 20 of TMH-1 and 64 of TMH-2 lie proximally to each other. Comparison of sequence alignment of the c subunits from other species confirms that a prolinyl residue placed in close proximity to the equivalent of Asp61 is a common feature of subunitc structure (41Groth G. Walker J.E. FEBS Lett. 1997; 410: 117-123Crossref PubMed Scopus (43) Google Scholar). Consistent with the results of mutant analysis in E. coli, the proline residue is located in helix 1 in some species and in helix 2 in others. This raises the question of the effect of the prolinyl location on the architecture of the monomeric protein in solution and the related question of the role of the prolinyl residue in the native enzyme. We report here on the structure of A20P/P64A substituted subunit c. The A20P/P64A subunitc was prepared from strain JH613, which contains copies of the mutant subunit c gene on the chromosome and on a pBR322-derived plasmid and is genetically equivalent to strain MEG119 used for the purification of the wild type protein (42Girvin M.E. Fillingame R.H. Biochemistry. 1993; 32: 12167-12177Crossref PubMed Scopus (77) Google Scholar). Uniform labeling with 15N was achieved by growing cells on a minimal medium containing 15 mm[15N]NH4Cl as a sole nitrogen source. Subunitc was purified as described previously (42Girvin M.E. Fillingame R.H. Biochemistry. 1993; 32: 12167-12177Crossref PubMed Scopus (77) Google Scholar). Samples for NMR contained 0.6 ml of 2.2 mm subunit c in either CDCl3/CD3OH/H2O (4:4:1) or CDCl3/CD3OD/D2O (4:4:1), both solvents containing 50 mm NaCl. The pH of the samples was adjusted to 5.0 without correction for the isotope effect. NMR data were collected on a Brüker DMX600 spectrometer with triple axis pulsed field gradient capability at 300 K. Two-dimensional proton experiments were done essentially as described previously (17Dmitriev O.Y. Jones P.C. Jiang W. Fillingame R.H. J. Biol. Chem. 1999; 274: 15598-15604Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar). The HSQC, three-dimensional TOCSY-HSQC, and NOESY-HSQC experiments used gradient sensitivity-enhanced schemes (43Zhang O. Kay L.E. Olivier J.P. Forman-Kay J.D. J. Biomol. NMR. 1994; 6: 845-858Crossref Scopus (612) Google Scholar, 44Warren W.S. Richter W. Andreotti A.H. Farmer B.T. Science. 1993; 262: 2005-2009Crossref PubMed Scopus (355) Google Scholar). Chemical shift assignments for the wild type protein were used as a basis for assigning resonances in the DQF-COSY, TOCSY, HSQC, and three-dimensional TOCSY-HSQC spectra of the cA20P/P64A protein. The sequential assignments were checked and completed using the three-dimensional NOESY-HSQC experiment. Distance constraints were derived from two-dimensional NOESY and three-dimensional NOESY-HSQC experiments. Scalar coupling constants were calculated from three-dimensional HNHA spectrum (45Bax A. Vuister G.W. Grzesiek S. Delaglio F. Wang A.C. Tschudin R. Zhu G. Methods Enzymol. 1994; 239: 79-105Crossref PubMed Scopus (381) Google Scholar). Exchange rates for amide protons were measured by diluting a sample in CDCl3/CD3OH/H2O with two volumes of fully deuterated solvent and then recording a series of HSQC spectra at intervals between 1 and 36 h. Spectral data were processed using Felix 95.0 software (Molecular Simulations Inc., Palo Alto, CA). A total of 645 NOEs were assigned in the spectra, 493 of which provided meaningful distance constraints. The structure was calculated from these constraints, 41 hydrogen bond constraints and 42 HNHα coupling constants (see TableI). Two hundred initial structures were calculated using simulated annealing as implemented in DYANA (46Güntert P. Mumenthaler C. Wüthrich K. J. Mol. Biol. 1997; 273: 283-298Crossref PubMed Scopus (2558) Google Scholar). The 10 lowest energy structures were additionally minimized using the AMBER force field as implemented in “Discover” (Molecular Simulations Inc., Palo Alto CA). The quality of the structures was checked using PROCHECK (47Laskowski R.A. MacArthur M.W. Moss D.S. Thornton J.M. J. Appl. Crystallogr. 1993; 26: 283-291Crossref Google Scholar). The coordinates were deposited at the Protein Data Bank (Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ) as entry1IJP.Table ISummary of constraints and statistics for A20P/P64A subunit c structure calculationConstraintsNumber usedNOE-derived distance constraintsTotal493Intraresidue179Sequential154Medium range (1 < ‖i − j‖ ≤ 4)116Long range (‖i − j‖ > 4)44Dihedral angle (ϕ) constraints derived from coupling constants42Hydrogen bonding constraints41Mean global pairwise backbone r.m.s. deviation0.75 ± 0.20 ÅMean global pairwise heavy atom r.m.s. deviation1.27 ± 0.22 ÅNumber of consistent constraint violations for the 10 lowest energy structuresDistance violations above 0.2 Å1Angle violations above 5°1 Open table in a new tab Cell membranes from the monoploid wild type strain AN346 and isogenic strain OM470, which carries a chromosomal copy of the cA20P/P64A mutation, were prepared in TMDG buffer (50 mm Tris-HCl, pH 7.5, 5 mmMgCl2, 1 mm dithiothreitol, 10% (v/v) glycerol) as described previously and stored at 20 mg/ml protein at −80 °C (48Mosher M.E. White L.K. Hermolin J. Fillingame R.H. J. Biol. Chem. 1985; 260: 4807-4814Abstract Full Text PDF PubMed Google Scholar). ATPase activity was measured at room temperature with membranes diluted to 15 µg/ml in a buffer containing 50 mm Tris-HCl, pH 8.0, 2 mm MgCl2, and 2 mm ATP by liberation of inorganic phosphate (49LeBel D. Poirier G.G. Beaudoin A.R. Anal. Biochem. 1978; 85: 86-89Crossref PubMed Scopus (287) Google Scholar). ATPase-coupled proton translocation was monitored by the quenching of 9-amino-6-chloro-2-methoxyacridine fluorescence in HMK assay buffer (10 mm HEPES-KOH, pH 7.5, 5 mm MgCl2, 300 mm KCl) as described previously (34Valiyaveetil F.I. Fillingame R.H. J. Biol. Chem. 1997; 272: 32635-32641Abstract Full Text Full Text PDF PubMed Scopus (108) Google Scholar). To test for inhibition by DCCD, membranes were incubated on ice at 3 mg/ml in TMDG buffer with 50 µm DCCD for 30 min. Protein was determined by a modified Lowry assay in the presence of SDS (50Fillingame R.H. J. Bacteriol. 1975; 124: 870-883Crossref PubMed Google Scholar). The E. colistrain OM470 carrying the A20P/P64A subunit c grew as well as wild type on a succinate carbon source, which indicates that thecA20P/P64A mutant F0F1 functions well as an ATP synthase during oxidative phosphorylation. The total ATPase activity of membranes prepared from the wild type and OM470 cells was 0.24 ± 0.03 µmol min−1 mg of protein−1 and 0.28 ± 0.04 µmol min−1 mg of protein−1respectively. DCCD inhibited the ATPase of wild type andcA20P/P20A membranes by 65 and 75%, respectively. The ATPase-coupled H+ transport activity of wild type and mutant membrane vesicles was equivalent, although the activity of the mutant was somewhat more sensitive to inhibition by DCCD (Fig.1). These results demonstrate that thecA20P/P64A mutation does not significantly affect the function of the ATP synthase. The structure of A20P/P64A substituted subunit c was solved in the chloroform/methanol/water mixture that had been used previously for solving the structure of wild type subunit c (16Girvin M.E. Rastogi V.K. Abildgaard F. Markley J.L. Fillingame R.H. Biochemistry. 1998; 37: 8817-8824Crossref PubMed Scopus (272) Google Scholar) and the N-terminal fragment of the subunit b (17Dmitriev O.Y. Jones P.C. Jiang W. Fillingame R.H. J. Biol. Chem. 1999; 274: 15598-15604Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar). To facilitate the use of chemical shift assignments available for the wild type protein and to make the structural comparison meaningful, the spectra were recorded under conditions identical to those used for wild type protein (16Girvin M.E. Rastogi V.K. Abildgaard F. Markley J.L. Fillingame R.H. Biochemistry. 1998; 37: 8817-8824Crossref PubMed Scopus (272) Google Scholar). An overlay of the two-dimensional 15N,1H chemical shift correlation spectra of the wild type and A20P/P64A proteins is shown in Fig. 2. The chemical shifts of the backbone amide groups are very similar for residues 1–14, 28–55, and 70–79. The largest difference in chemical shifts is found for residues located around the sites of mutation at residues 20 and 64 (Fig. 3). The similarity of the chemical shifts and the unimpaired function of the ATP synthase were taken to be a strong indication that overall architecture of the mutant protein was similar to wild type. This justified use of the wild type structure as a reference in resolving ambiguous NOE assignments that resulted from the high degree of chemical shift degeneracy observed for this protein. The final family of 10 structures had r.m.s. deviations of 0.75 ± 0.20 Å for the backbone atoms and 1.27 ± 0.22 Å for all heavy atoms of residues 3–77. The conformation of the two individual helical domains is defined somewhat better. When residues 4–39 and 46–77 were aligned separately, the r.m.s. deviations for the backbone atoms were 0.61 ± 0.15 and 0.54 ± 0.20 Å, respectively. There were no distance constraint violations exceeding 0.3 Å (Table I). In the 10 best structures, 90.6% of residues were found in the most favored regions of Ramachandran plot, with 8.6% in the additionally allowed, 0.6% in the generously allowed, and 0.2% in disallowed regions. For the well defined region of the structure, comprising residues 3–77, 92.8% of residues were in the most favored region with no residues in disallowed regions.Figure 3Summary of NMR data for the A20P/P64A mutant of subunit c. Shown from top to bottomare amino acid sequence showing residues with slowly exchanging amide protons underlined in purple (A), sequential and medium range NOEs (B), HNHα coupling constants (C), the difference in chemical shift for amide protons for mutant minus wild type subunit c (D), and secondary structure elements and predicted solvent accessibility (lighter shading corresponding to higher accessibility) as determined by PROCHECK (47Laskowski R.A. MacArthur M.W. Moss D.S. Thornton J.M. J. Appl. Crystallogr. 1993; 26: 283-291Crossref Google Scholar) for the final structure (E).View Large Image Figure ViewerDownload Hi-res image Download (PPT) The pattern of sequential and short range NOEs observed in the three-dimensional NOESY-HSQC spectrum is typical for a largely α-helical protein (Fig. 3). Indeed, as for the wild type protein, the structure of A20P/P64A subunitc is best described as a hairpin of two α-helices extending from residues 4 to 39 and 46 to 77 (Fig. 3). The packing of the two helices in the structure (Fig.4 A) is determined by long range distance constraints derived from 44 NOEs. The distribution of NOEs along the amino acid sequence is shown in Fig.5 A. A comparison of long range NOEs observed in the wild type and A20P/P64A proteins reveals a similar pattern of interactions between the helices (Fig. 5 B), but the pairs of interacting residues are different in many cases. Consequently, the relative orientation, bending, and crossing angles of the helices in the mutant protein are somewhat different from the wild type protein (Fig. 6, A andB). The longer N-terminal helix of A20P/P64A subunitc is bent around Pro20 at an angle of about 20° and wraps around the nearly straight C-terminal helix. The helices cross at angles of about 20° for the segments comprising residues 22–40 and 46–62 and about 35° for the segments comprising residues 4–18 and 64–77. The bend in the N-terminal helix is associated with the absence of a backbone hydrogen bond between the carbonyl oxygen of Ala21 and amide proton of Ala25. Similarly, in the wild type subunit cstructure at pH 5 (1A91; Ref. 16Girvin M.E. Rastogi V.K. Abildgaard F. Markley J.L. Fillingame R.H. Biochemistry. 1998; 37: 8817-8824Crossref PubMed Scopus (272) Google Scholar), where a comparable bend is observed in the C-terminal helix, the Met65/Gly69hydrogen bond would have quite unfavorable geometry. In the structure of wild type subunit c at pH 8 (24Rastogi V.K. Girvin M.E. Nature. 1999; 402: 263-268Crossref PubMed Scopus (418) Google Scholar), where the C-terminal helix is more strongly bent around Pro64 (Fig.6 C), the Met65/Gly69 backbone hydrogen bond is absent. The changes in orientation and crossing angle of the helices in the mutant versus wild type proteins result in associated changes in the structure of the connecting polar loop, most notably of residues 42–45 (Fig. 4 B).Figure 5NOE distributions in the sequence of A20P/P64A subunit c. A, number of intraresidue (white), sequential (light gray), medium range (dark gray), and long range (black) NOE constraints per residue.B, residues showing interhelical NOEs in A20P/P64A subunitc (⋄) and wild type subunit c(■).View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 6Ribbon diagram of the structures of the A20P/P64A subunit c (A) and wild type subunit c (B–D). A, A20P/P64A subunit c structure at pH 5 (1IJP);B, subunit c structure at pH 5 (1A91; Ref. 16Girvin M.E. Rastogi V.K. Abildgaard F. Markley J.L. Fillingame R.H. Biochemistry. 1998; 37: 8817-8824Crossref PubMed Scopus (272) Google Scholar);C, subunit c structure at pH 8 (1C99; Ref. 24Rastogi V.K. Girvin M.E. Nature. 1999; 402: 263-268Crossref PubMed Scopus (418) Google Scholar);D, subunit c monomer from the model ofc12 complex (1J7F; Ref. 23Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (107) Google Scholar). For each structure, helix 1 is shown in yellow and helix 2 ingreen. Magenta is used to mark the positions of residues 20 and 64, and Asp61 is shown inred.View Large Image Figure ViewerDownload Hi-res image Download (PPT) From the structure of A20P/P64A subunit c shown here, we conclude that the essential prolinyl residue will tend to introduce a bend into the TMH in which it is placed when the helical hairpin is formed within a subunit monomer. Since the prolinyl residue and bend occur in different TMHs in the wild type and mutant proteins, the packing of TMHs varies somewhat in the structures of the two monomeric subunits. On the other hand, the function of the cA20P/P64A mutant ATP synthase appears nearly indistinguishable from wild type, which strongly suggests that the structure of both proteins should be very similar within the oligomeric c-ring. Intersubunit packing interactions that lead to c-ring formation may force an adjustment (straightening) in either structure, as is already indicated by molecular modeling studies done with the wild type protein (23Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (107) Google Scholar). As discussed previously (23Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (107) Google Scholar), the solution structure was expected to change in at least minor ways as additional protein-protein surface interactions were considered. In the case of A20P/P64A subunitc, a straightening of helix 1 and concomitant reorientation of the two helical segments in the course of oligomer formation may allow the subunit to pack in a ring that is very similar to wild type. The structure presented here does indicate that the forces leading to the kink in TMH-2 of wild type subunit c result from local constraints introduced by the prolinyl residue rather than global interactions reflecting the packing of the entire protein. In contrast to globular proteins, there are relatively few surface-surface interactions between the TMHs of the highly elongated subunit c monomer. The relative orientation and bend of helices should therefore be expected to change significantly upon packing of the individual c subunits into the ring structure of F0 due to the contribution of the numerous additional interactions (23Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (107) Google Scholar, 24Rastogi V.K. Girvin M.E. Nature. 1999; 402: 263-268Crossref PubMed Scopus (418) Google Scholar). Indeed, a modeling of the wild type coligomer, which was based upon an extensive set of distance constraints from intersubunit cross-linking in the native F0 and the NMR model, indicated that the C-terminal helix straightened with the bend angle around Pro64 being reduced from 27° in the NMR structure (1A91; Fig. 6B) to 16 ± 3° in the oligomer ring (Ref.23Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (107) Google Scholar, Fig. 6D). Further, a more recent calculation of the wild type subunit c structure at pH 5 (1C0V; Ref. 24Rastogi V.K. Girvin M.E. Nature. 1999; 402: 263-268Crossref PubMed Scopus (418) Google Scholar) showed a less pronounced bend around Pro64 of about 11°. Given that the two pH 5 structures do differ significantly and that the calculations were based upon the same extensive set of NOEs (16Girvin M.E. Rastogi V.K. Abildgaard F. Markley J.L. Fillingame R.H. Biochemistry. 1998; 37: 8817-8824Crossref PubMed Scopus (272) Google Scholar), the experimentally derived constraints clearly permit significant variation in the overall bend of the molecule. The comparison of the structures of A20P/P64A and wild type subunitc raises an obvious question about the functional role of the bend in the proline-containing helix. The recently reported structure of deprotonated subunit c (Fig. 6 C) suggests that ionization of Asp61 results in a major reorientation in the packing of helices, TMH-2 being turned by about 140° relative to its orientation in the protonated state (24Rastogi V.K. Girvin M.E. Nature. 1999; 402: 263-268Crossref PubMed Scopus (418) Google Scholar). The occurrence of such helical turning, within the confines of the multiple intersubunit surface contacts of the oligomer, would require rather complex concerted movements of both transmembrane helices and would probably also involve the neighboring c subunits. The movement appears to involve changes in the bend angle of both helices. Indeed, in the pH 8 structure of the wild type monomeric subunit, very significant bends are observed in both TMH-2 and TMH-1, i.e.bend angles of about 60 and 30°, respectively. In the model of the subunit c oligomer (23Dmitriev O.Y. Jones P.C. Fillingame R.H. Proc. Natl. Acad Sci. U. S. A. 1999; 96: 7785-7790Crossref PubMed Scopus (107) Google Scholar), the side chain of Asp61is located at the interface of two neighboring c subunits. An interaction of the Asp61 side chain withaArg210 is proposed to be a key step in transmembrane H+ transport (11Elston T. Wang H. Oster G. Nature. 1998; 391: 510-513Crossref PubMed Scopus (448) Google Scholar, 34Valiyaveetil F.I. Fillingame R.H. J. Biol. Chem. 1997; 272: 32635-32641Abstract Full Text Full Text PDF PubMed Scopus (108) Google Scholar, 35Fillingame R.H. Krulwich T.A. The Bacteria. Academic Press, Inc., New York1990: 345-391Google Scholar, 36Fillingame R.H. Jiang W. Dmitriev O.Y. Jones P.C. Biochim. Biophys. Acta. 2000; 1458: 387-403Crossref PubMed Scopus (56) Google Scholar), and a concerted bending and swiveling of helices may be critical in promoting this interaction. The fact that the proline can be located on either of the two helices without a noticeable effect on function supports the notion that both helices are involved. The role of prolinyl residue would then be to introduce a “weak spot” in the helical hairpin by breaking the regular pattern of hydrogen bonds stabilizing one of the TMHs. The role of prolinyl residues in TMHs of transport proteins has been a subject of interest and hypotheses for some time (51Brandl C.J. Deber C.M. Proc. Natl. Acad. Sci. U. S. A. 1986; 93: 917-921Crossref Scopus (282) Google Scholar, 52Williams K.A. Deber C.M. Biochemistry. 1991; 30: 8919-8923Crossref PubMed Scopus (230) Google Scholar). Sequence comparisons indicate that proline residues are much more common in the membrane-spanning domains of transport proteins versus other membrane proteins (51Brandl C.J. Deber C.M. Proc. Natl. Acad. Sci. U. S. A. 1986; 93: 917-921Crossref Scopus (282) Google Scholar). This observation, taken together with the helix-breaking propensity of proline residue, suggests a possible general role for proline in facilitating functionally important movements of transmembrane helical segments. Molecular dynamics simulations with model peptides (53Yun R.H. Anderson A. Hermans J. Proteins. 1991; 10: 219-228Crossref PubMed Scopus (133) Google Scholar), alamethicin (54Biggin P.C. Breed J. Son H.S. Sansom M.S.P. Biophys. J. 1997; 72: 627-663Abstract Full Text PDF PubMed Scopus (56) Google Scholar), and individual helices of bacteriorhodopsin (55Sankararamakrishnan R. Vishveshwara S. Proteins. 1993; 15: 26-41Crossref PubMed Scopus (33) Google Scholar, 56Woolf T.B. Biophys. J. 1997; 73: 2376-2392Abstract Full Text PDF PubMed Scopus (57) Google Scholar) demonstrate that prolinyl-containing α-helices will adopt a variety of kinked conformations on a fast time scale, differing in kink angle and hydrogen bonding pattern around the prolinyl residue. The possible variations in hydrogen bonding pattern could allow the polypeptide chain to adopt different metastable conformations during the course of a transport cycle. The suggested importance of the subunit c transmembrane prolinyl residue in F0F1 function is generally supported by sequence alignment of subunit c from other species (41Groth G. Walker J.E. FEBS Lett. 1997; 410: 117-123Crossref PubMed Scopus (43) Google Scholar). Assuming the same hairpin-like structure, a prolinyl residue would be located close to the essential Asp or Glu residue in 28 of the 37 species compared (41Groth G. Walker J.E. FEBS Lett. 1997; 410: 117-123Crossref PubMed Scopus (43) Google Scholar). In these cases, the proline is found in the position equivalent to 64 in E. coli TMH-2 (12 examples) or at the position equivalent to residue 24 in E. coli TMH-1 (16 examples). Nine of the 37 cases do not have a proline in either TMH, but, with one exception, Ser or Thr residues are found in these regions on one or both helices. Hydrogen bond formation between a Ser or Thr side chain and the backbone atom of another residue provides an alternative mechanism of breaking the hydrogen bond pattern of an α-helix and introducing a potential site for the swiveling and bending of the helix. We conclude that the bends observed in the TMHs of wild type and A20P/P64A subunit c in solution very likely reflect a need for a concerted structural rearrangement of helices around the region of the essential proline during the H+ transport cycle. We thank Dr. Joe Hermolin for assistance in strain construction.

Ver no editor

Altmetric

PlumX

Entrar

Lembrar minha senha

Receber meu e-mail de confirmação

Structure of Ala20 → Pro/Pro64 → Ala Substituted Subunit c ofEscherichia coli ATP Synthase in Which the Essential Proline Is Switched between Transmembrane Helices