- Email: [email protected]
On the Structure of the Proton-Binding Site in the Fo Rotor of Chloroplast ATP Synthases
doi:10.1016/j.jmb.2009.10.059
J. Mol. Biol. (2010) 395, 20–27
Available online at www.sciencedirect.com
COMMUNICATION
On the Structure of the Proton-Binding Site in the Fo Rotor of Chloroplast ATP Synthases Alexander Krah 1 , Denys Pogoryelov 2 , Thomas Meier 2,3 ⁎† and José D. Faraldo-Gómez 1,3 ⁎† 1
Theoretical Molecular Biophysics Group, Max Planck Institute of Biophysics, 60438 Frankfurt am Main, Germany 2
Department of Structural Biology, Max Planck Institute of Biophysics, 60438 Frankfurt am Main, Germany 3
Cluster of Excellence Macromolecular Complexes, Goethe University of Frankfurt, 60438 Frankfurt am Main, Germany Received 25 September 2009; received in revised form 23 October 2009; accepted 27 October 2009 Available online 31 October 2009
The recently reported crystal structures of the membrane-embedded proton-dependent c-ring rotors of a cyanobacterial F1Fo ATP synthase and a chloroplast F1Fo ATP synthase have provided new insights into the mechanism of this essential enzyme. While the overall features of these crings are similar, a discrepancy in the structure and hydrogen-bonding interaction network of the H+ sites suggests two distinct binding modes, potentially reflecting a mechanistic differentiation. Importantly, the conformation of the key glutamate side chain to which the proton binds is also altered. To investigate the nature of these differences, we use molecular dynamics simulations of both c-rings embedded in a phospholipid membrane. We observe that the structure of the c15 ring from Spirulina platensis is unequivocally stable within the simulation time. By contrast, the proposed structure of the H+ site in the chloroplast c14 ring changes rapidly and consistently into that reported for the c15 ring, indicating that the latter represents a common binding mode. To assess this hypothesis, we have remodeled the c14 ring by molecular replacement using the published structure factors. The resulting structure provides clear evidence in support of a common binding site conformation and is also considerably improved statistically. These findings, taken together with a sequence analysis of csubunits in the ATP synthase family, indicate that the so-called proton-locked conformation observed in the c15 ring may be a common characteristic not only of light-driven systems such as chloroplasts and cyanobacteria but also of a selection of other bacterial species. © 2009 Elsevier Ltd. All rights reserved.
Edited by B. Honig
Keywords: F1Fo ATP synthase rotor; membrane transport; atomic structure; molecular dynamics simulation; X-ray crystallography
F-type ATP synthases catalyze the conversion of ADP and Pi into ATP by coupling the structural changes required for this endergonic reaction to the transmembrane flow of H+ or Na+ down their electrochemical gradients. 1–4 The key coupling element in these molecular machines is the membrane-embedded Fo rotor, known as the c-ring. This structure is an assembly of 10–15 transmembrane *Corresponding authors. Max Planck Institute of Biophysics, Max-von-Laue-Str. 3, 60438 Frankfurt am Main, Germany. E-mail addresses: [email protected]; [email protected]. † T.M. and J.D.F.-G. contributed equally to this work. Abbreviations used: MD, molecular dynamics; PDB, Protein Data Bank.
helical hairpins (or c-subunits), which provide selective binding sites for the ions as they traverse the membrane. In conjunction with the adjacent asubunit (the stator) and its mechanical linkage to the catalytic domain through the γ-subunit (the axle), rotation of the c-ring thus allows ion flow and powers ATP synthesis. Conversely, ATP hydrolysis can cause the rotation of the c-ring in reverse, thereby generating a transmembrane gradient. The crystal structure of the Na+-coupled c-ring from Ilyobacter tartaricus provided the first atomically detailed view into the mechanism of ion binding to the Fo rotor.5 This ring comprises 11 subunits, with 11 Na+-binding sites, each of which lies in between adjacent subunits. The conserved carboxylate (E65) provides the main stabilizing interaction for the ion, but the site consists of another four coordinating
0022-2836/$ - see front matter © 2009 Elsevier Ltd. All rights reserved.
Proton Binding to ATP Synthases
ligands, including a bound structural water molecule.6 Altogether, this environment compensates for the energetic cost of desolvation and transfer into the membrane, and presumably confers the necessary specificity to the binding site. Recently, two crystal structures of H+-coupled crings—namely, from spinach chloroplast7 (consisting of 14 c-subunits) and from the cyanobacterium Spirulina platensis8 (made of 15 subunits)—have been reported. Both structures indicate that proton binding occurs through protonation of the conserved carboxylate (E61 and E62, respectively) and are consistent with previous measurements of the stoichiometry of these rings.9,10 However, these structural studies have also revealed differences in the proton-binding site suggestive of two distinct binding modes, despite the high level of similarity in the respective c-subunit amino acid sequences. In this study, we investigate the nature of these differences using computational methods for molecular dynamics (MD) simulation, crystallographic refinement, and sequence analysis, and we conclude that the so-called proton-locked conformation observed in the cyanobacterial c15 structure8 is, in fact, common to both c-rings and probably also to their homologues in many light-driven systems and in a selection of other bacterial species. Comparison of the published crystal structures of the c14 and c15 rings According to the conventional hydrogen-bond definition based on both the distance and the preferred geometries of donors and acceptors (dDA ≤ 3.6 Å; 120° ≤ θDHA ≤ 180°), the only interaction that is
21 common to the proton-binding sites in the reported crystal structures of the c14 and the c15 rings is that between E62:Oɛ1 and Y67:OH (Fig. 1) (E61 and Y66 in c14; all numberings hereafter correspond to c15). In contrast to the c15 ring, in the c14 structure, no hydrogen bond can be discerned between the protonated E62:Oɛ2 and the backbone carbonyl of F60, or with the amine group in the side chain of Q29. While the distance between E62:Oɛ2 and F60:O would, in principle, allow for a hydrogen bond (dDA = 2.8 Å), the carboxylic group would have to be in trans, which, according to spectroscopic and theoretical analyses, is unfavorable energetically and thus underrepresented statistically relative to the cis conformation. 11–14 Consistent with this, protonated carboxylate groups in ultra-high-resolution neutron diffraction protein structures have been observed, to our knowledge, only in the cis conformation.15–17 With regard to Q29, whose side chain is, in fact, absent in the published c14 structure, a modeled ideal geometry positions its Nɛ donor too distant from the closest carboxyl oxygen in E62 (dDA = 4.7 Å; cf. 3.0 Å in the c15 structure). Another intriguing difference is the rotamer of T65, which in the c15 structure is not directly engaged in the protonbinding site and instead hydrogen bonds to the backbone carbonyl of G21 in the inner helix. By contrast, in the c14 structure, T65 orients its hydroxyl group towards E62:Oɛ1, and while it is, in principle, within hydrogen-bonding distance (dDA = 3.0 Å), the orientation of these side chains is not consistent with such interaction (θDHA ∼ 105°). In summary, the binding site in the c14 structure appears to contain an unusual number of unfulfilled hydrogen-bond donors and acceptors, despite the likely energetic cost, especially in the hydrophobic
Fig. 1. Published crystal structures of the proton-binding site in (left) the c14 ring from spinach chloroplast7 and in (right) the c15 ring from S. platensis8 (PDB codes 2W5J and 2WIE, respectively). Note that the key side chain of Q28 was absent in the published structure of the chloroplast c14 ring and is therefore modeled. The view is sideways from within the membrane, with the outer C-terminal helices of two adjacent c-subunits in front and with one of the corresponding inner helices in the background. The possible hydrogen-bonding interactions in each case, on the basis of the distance between donor and acceptor, as well as their preferred chemical structure, are indicated with broken lines (see the main text).
22 environment 18 provided by detergent or lipid hydrocarbon chains. This is in contrast to the c15 site, where the interaction network is clearly defined and geometrically realistic. In view of this, the limited resolution of the c14 diffraction data (3.8 Å versus 2.1 Å for the c15 ring), and the high degree of similarity of these two sequences, we hypothesized that the H+-binding site in the c14 ring is, in fact, structurally identical with that in the c15 ring, despite the published structure. MD simulations of the c14 and c15 rings in a lipid membrane To examine this hypothesis further, we carried out a series of MD simulations of the c14 ring in a lipid
Proton Binding to ATP Synthases
membrane, in which we either preserved the conformation of the reported Cα trace or allowed complete flexibility. In both cases, we found that the reported structure of the binding site spontaneously changed into that observed for the c15 ring (Fig. 2). Analogous control simulations of the c15 ring showed no sign of structural instability (Fig. 2). More specifically, we observed that, consistently across all c-subunits in the chloroplast ring, E62 reorients into a geometry compatible with hydrogen bonds both with F60:O (via the protonated Oɛ2; in cis) and with Y67:OH (via Oɛ1). As seen in the c15 ring, E62:O ɛ2 also becomes an acceptor of a hydrogen bond donated by Q29:Nɛ. Interestingly, T65 also reorients in the course of the c14 simulations and forms a bond analogous to that with G21:O
Fig. 2. (a) Interaction distances and angles at the proton-binding site in the c15 ring from the crystal structure (broken lines) and MD simulations (continuous lines). (b) Same as (a), for the c14 ring and simulations where the Cα trace of the ring is constrained to be in the conformation in the published structure. (c) Same as (b), from simulations without any constraints. Simulation data are presented as probability distributions, computed from dynamical trajectories and combining all c-subunits. The corresponding distances/angles from the published structures are overlaid on the distributions with a vertical line. The conformation of the c-rings at the start of all simulations is that in the published crystal structures; the structural changes observed during the simulations of the c14 ring (b and c) are indicated with arrows. All MD simulations were carried out with NAMD24 using the CHARMM27 forcefield,25 at constant temperature (298 K), pressure (1 atm), and membrane surface area (∼ 69 Å2 per lipid). Electrostatic interactions were computed using the PME algorithm with a real-space cutoff distance at 12 Å, which was also used for the van der Waals interactions. Molecular models of the c14 and c15 crystal structures embedded in a hydrated POPC lipid membrane were prepared using the methodology described by Faraldo-Gómez et al.26; the systems include a total of ∼100,000 atoms each (see Supplementary Material, Fig. S1). This was followed by a series of MD simulations with gradually weaker constraints on the protein structure, for up to ∼ 10 ns. Both systems were subsequently simulated for 20 ns, allowing complete flexibility. In addition, we carried out five simulations of the c14 ring (for 5 ns each) in which Cα trace was constrained to remain as in the published crystal structure.
23
Proton Binding to ATP Synthases
in the cyanobacterial ring (namely, with A20). This is especially clear in simulations where the Cα trace is constrained; in unconstrained simulations, T65 is transiently engaged in other interactions (e.g., S22: Oγ), but not with E62 in any case. By contrast, the simulation of the c15 ring, which also allowed complete flexibility in the protein, showed excellent agreement between the crystal structure and the most populated geometry of the interaction network throughout the binding site, for an equivalent timescale. The time-averaged structures of the binding site derived from the c14 and c15 simulations, for a randomly selected c-subunit interface, are shown in Fig. 3. This comparison clearly bolsters the hypothesis that the proton-binding sites in these two rings are essentially identical. Lastly, it is worth pointing out that additional simulations of equal duration, where E62 was set in the unprotonated state, resulted in marked deviations from the published structures in both the c14 and the c15 rings. These structural changes are induced by the substantial penetration of water molecules into the membrane core, driven by the large energetic gain derived from the solvation of the charged carboxylate group (see Supplementary Material, Fig. S2). In this environment, the side chain of E62 spontaneously rotates away from the binding site and projects into the surrounding solvent. Thus, E62 no longer interacts with F60, Q29, or T65, and only partially interacts with Y67, but rather interacts with the neighboring water molecules. Moreover, these changes also influence the geometry of the ring as a whole, even in this short timescale (see Supplementary Material, Fig. S2). In summary, the characteristic conformation of the proton-binding site observed in charged-state simulations is completely inconsistent with the published structures of
the c14 and c15 rings (i.e., the differences between these two structures do not reflect a different protonation state). This result is reassuring, since the likelihood of the charged state is remote while E62 is embedded in a solvent-excluded hydrocarbon environment and lacks a suitable counterion (such as Na+ in the c11 structure). It is conceivable that a H3O+ could reside in the c14 site, as hypothesized elsewhere.19 However, neither the crystallographic analysis of Vollmar et al.7 nor ours (see the text below) indicates the presence of a hydronium in this site. Crystallographic structure refinement of the c14 ring To provide a degree of experimental certainty to our hypothesis of a common mode of proton binding, we recalculated the c14 structure by molecular replacement, on the basis of the highresolution structure of the c15 ring and the c14 structure factors deposited in the Protein Data Bank (PDB). The new experimental structure resulting from this analysis is shown in Fig. 4a alongside the electron density map. While the limited resolution precludes full atomic detail (e.g., T65 and Q29 can only be partially resolved), the orientation of E62, consistent with a hydrogen bond between E62:Oɛ2 and F60:O, can be clearly discerned in the omit map of all 14 c-subunits in the ring, thus confirming the proton-locked conformation observed in the c15 ring. The resolution of the other side chains originally missing in the published structure is also improved (e.g., L15, I22, Q28, Q34, E37, I49, M60, L74, N78, and F80; in c14 numbering). Consistently, the refinement statistics from this
Fig. 3. Time-averaged simulation structures of the proton-binding site in (left) the c14 ring from spinach chloroplast and in (right) the c15 ring from S. platensis. The view is the same as in Fig. 1. Nonpolar hydrogen atoms are omitted for clarity. The average structure of the c14 ring derives from Cα-trace constrained simulations (Fig. 2b); that for the c15 ring from an unconstrained simulation (Fig. 2a).
24
Proton Binding to ATP Synthases
Fig. 4. (a) Remodeled crystal structure of the proton-binding site in the c14 ring from spinach chloroplast, derived from the structure factors deposited in PDB code 2W5J.7 The electron density map (2Fobs − Fcalc; blue mesh) is contoured at 2σ, and the difference map (Fobs − Fcalc; green mesh) is contoured at 3σ. The structure was solved by molecular replacement based on the c 15 ring from S. platensis (PDB code 2WIE),8 omitting the side chain of E62 (E61 in the c14 sequence) to avoid a bias in the resulting map. Initial phases were obtained with Phaser27 using a search model of a c14 ring composed of c-subunits from the c15 ring. The model was improved by iterative rounds of model building with Coot28 into the 2Fobs − Fcalc electron density map and further refined with PHENIX.29 Noncrystallographic symmetry averaging was applied at all steps of the refinement process. (b) Per-residue root-mean-square difference between the published model of the c14-ring structure (PDB code 2W5J) and the remodeled structure described in this study. The average root-mean-square differences (residues 4–79) are 1.2 Å and 1.7 Å for the main chain and the side chains, respectively.
analysis are improved relative to the original (i.e., we find a better correspondence between the threedimensional structural model and the measured data); in particular, the Rwork and Rfree factors decreased to 24.8% and 30.9%, respectively, compared with 33% and 36% in Vollmar et al.7 As shown in Fig. 4b, these improvements reflect not only the rearrangement of the binding site and the added side chains but also changes elsewhere in the structure, particularly at the N-terminus of the csubunits, where the main chain was seen to adopt a different conformation.
It seems clear, therefore, that the c15 ring is a more appropriate molecular replacement template than the Na+-coupled I. tartaricus c11 ring employed originally,7 although it should be noted that only the latter was available at that time. At any rate, the resolution of the underlying diffraction data remains very low (3.8 Å); thus, the determination of the complete atomic structure of the c14 ring may be considered as a work in progress. Nevertheless, the features of the proton-binding site appear to be already well defined, as judged from the analysis presented here.
Proton Binding to ATP Synthases
Sequence analysis and consensus structure of the proton-binding site Figure 5 shows the amino acid sequence alignment of c-subunits from F-ATP synthases in a selection of species/cell organelles, including bacteria, mitochondria, and chloroplasts. In the majority of these species, the enzyme is driven by the protonmotive force, implying that a proton must reversibly bind to the c-ring during ion translocation. However, in a handful of species including I. tartaricus, ATP synthases bind Na+ instead20–23 (i.e., they are driven by the sodium-motive force). Although a Na+binding signature sequence can be identified in these cases,6 an ion-binding motif in the protoncoupled c-rings, aside from the strictly conserved carboxyl group (E or D), is less clearly defined. The motif varies from an almost complete Na+-binding signature (e.g., Solibacter usitatus or Desulfovibrio vulgaris) to a ‘naked’ glutamate/aspartate (e.g., Bacillus pseudofirmus OF4 or Escherichia coli). In the c15 ring of the cyanobacterium S. platensis, the side chains of Q29, E62, and Y60 are directly involved in the proton coordination network. This sequence motif (Q/E/Y) is observed in most of the c-subunits in the phylum of Cyanobacteria, as well as in their ‘derivatives,’ the chloroplasts. Notably, these two types of cells/organelles use light as the primary energy source to generate their proton-motive force, a feature that distinguishes them from many other species. In addition, this common motif (Q/E/Y) is also found in a selection of bacteria in various phyla (e.g., Proteobacteria). Therefore, either it evolved
25 independently several times or it was determined early on in the evolution of the bacterial kingdom. In any case, the ATP synthase and its ion binding specificity are likely adapted to the unique physiological requirements of each cell/organelle type. It is therefore logical that not all species include the exact same hydrogen-bonding network seen here for the c14 and c15 rings, thus endowing these enzymes with the necessary flexibility to operate optimally in a variety of environments. As mentioned above, in many of the bacterial species, as well as in all eukaryotic mitochondria (animals, plants, and fungi), the proton-binding motif is limited to the conserved E/D carboxylate. The energetics of the ion translocation mechanism may therefore be different in this subclass of Fo rotors; conclusive structural data for a representative c-ring belonging to this prominent subclass are not available and would therefore be of the highest interest. In conclusion, based on the finding of a consensus structure of the H+-binding sites in the c14 and c15 rings, and on the sequence similarity of c-subunits across the F-ATP synthase family, we propose that the proton-locked conformation observed at high resolution for the c15 ring represents a common mode of proton binding in the c-rings of a wide variety of cell types and organelles, including cyanobacteria, chloroplasts, and other bacterial strains carrying the Q/E/Y sequence motif. By extension, our findings contradict the notion that proton binding to the chloroplast Fo rotors involves a hydronium associated with the key negatively charged glutamate side chain.19 Nevertheless, the
Fig. 5. Alignment of c-subunit sequences from F-type ATP synthases. The sequences are aligned according to the cytoplasmic loop region (bold lettering) of each c-subunit. Residues involved in H+ or Na+ coordination are highlighted with colors. The residue numbering and secondary-structure elements in the S. platensis c-subunit, according to the c15 crystal structure,8 are given at the top. Note that this set of sequences is a selection of representative c-subunit types and is not weighted according to their overall occurrence.
26 precise construction of the interaction network at the ion-binding sites of different c-rings likely reflects a fine-tuned adaptation to the specific physiological requirements and environment of each cell; therefore, subtle variations of this structural motif can be expected to occur (e.g., in mitochondria). Notwithstanding this variability, it is plausible that, from a mechanistic standpoint, the general principles of ion translocation outlined, for example, for the cyanobacterial ATP synthase8 or elsewhere3 will be valid across this fascinating family of motor enzymes.
Proton Binding to ATP Synthases
10. 11.
12.
13.
Acknowledgements This work was supported, in part, by the Cluster of Excellence “Macromolecular Complexes” of the Goethe University Frankfurt (DFG Project EXC 115). We thank the students of the IMPReS program at the Max Planck Institute of Biophysics for carrying out a subset of the computer simulations of the c14 ring presented here.
Supplementary Data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jmb.2009.10.059
14. 15.
16.
17.
18.
References 1. Boyer, P. D. (1997). The ATP synthase: a splendid molecular machine. Annu. Rev. Biochem. 66, 717–749. 2. Yoshida, M., Muneyuki, E. & Hisabori, T. (2001). ATP synthase: a marvellous rotary engine of the cell. Nat. Rev. Mol. Cell Biol. 2, 669–677. 3. Junge, W., Sielaff, H. & Engelbrecht, S. (2009). Torque generation and elastic power transmission in the rotary FoF1-ATPase. Nature, 459, 364–370. 4. Nakamoto, R. K., Baylis Scanlon, J. A. & Al-Shawi, M. K. (2008). The rotary mechanism of the ATP synthase. Arch. Biochem. Biophys. 476, 43–50. 5. Meier, T., Polzer, P., Diederichs, K., Welte, W. & Dimroth, P. (2005). Structure of the rotor ring of F-type Na+-ATPase from Ilyobacter tartaricus. Science, 308, 659–662. 6. Meier, T., Krah, A., Bond, P. J., Pogoryelov, D., Diederichs, K. & Faraldo-Gómez, J. D. (2009). Complete ion-coordination structure in the rotor ring of Na+-dependent F-ATP synthases. J. Mol. Biol. 391, 498–507. 7. Vollmar, M., Schlieper, D., Winn, M., Buchner, C. & Groth, G. (2009). Structure of the c14 rotor ring of the proton translocating chloroplast ATP synthase. J. Biol. Chem. 284, 18228–18235. 8. Pogoryelov, D., Yidiz, O., Faraldo-Gómez, J. D. & Meier, T. (2009). High-resolution structure of the rotor ring of a proton dependent ATP synthase. Nat. Struct. Mol. Biol. 16, 1068–1073. 9. Pogoryelov, D., Yu, J., Meier, T., Vonck, J., Dimroth, P. & Muller, D. J. (2005). The c15 ring of the Spirulina
19. 20.
21.
22.
23.
24.
25.
26.
platensis F-ATP synthase: F1/Fo symmetry mismatch is not obligatory. EMBO Rep. 6, 1040–1044. Seelert, H., Poetsch, A., Dencher, N. A., Engel, A., Stahlberg, H. & Müller, D. J. (2000). Proton-powered turbine of a plant motor. Nature, 405, 418–419. Briggs, J. M., Nguyen, T. B. & Jorgensen, W. L. (1991). Monte Carlo simulations of liquid acetic acid and methyl acetate with the OPLS potential function. J. Phys. Chem. 95, 3315–3322. Wiberg, K. B. & Laidig, K. E. (1987). Barriers to rotation adjacent to double bonds: 3. The C–O barrier in formic acid, methyl formate, acetic acid, and methyl acetate. The origin of ester and amide “resonance”. J. Am. Chem. Soc. 1987, 5935–5943. Ruschin, S. & Bauer, S. H. (1980). Rotational isomers of methyl formate, located by the temperature-drift technique. J. Phys. Chem. 84, 3061–3065. Bertie, J. E. & Michaelian, K. H. (1982). The Raman spectra of gaseous formic acid –h2 and –d2. J. Chem. Phys. 76, 886. Coates, L., Tuan, H. F., Tomanicek, S., Kovalevsky, A., Mustyakimov, M., Erskine, P. & Cooper, J. (2008). The catalytic mechanism of an aspartic proteinase explored with neutron and X-ray diffraction. J. Am. Chem. Soc. 130, 7235–7237. Deacon, A., Gleichmann, T., Kalb, A. J., Price, H., Raftery, J., Bradbrook, G. et al. (1997). The structure of concanavalin A and its bound solvent determined with small-molecule accuracy at 0.94 Å. J. Chem. Soc. Faraday Trans. 93, 4305–4312. Habash, J., Raftery, J., Weisgerber, S., Casseta, A., Lehmann, M. S., Hoghoj, P. et al. (1997). Neutron Laue diffraction study of concanavalin A—the proton of Asp28. J. Chem. Soc. Faraday Trans. 93, 4313–4317. Gao, J., Bosco, D. A., Powers, E. T. & Kelly, J. W. (2009). Localized thermodynamic coupling between hydrogen bonding and microenvironment polarity substantially stabilizes proteins. Nat. Struct. Mol. Biol. 16, 684–690. von Ballmoos, C. & Dimroth, P. (2007). Two distinct proton binding sites in the ATP synthase family. Biochemistry, 46, 11800–11809. Laubinger, W. & Dimroth, P. (1987). Characterization of the Na+-stimulated ATPase of Propionigenium modestum as an enzyme of the F1Fo type. Eur. J. Biochem. 168, 475–480. Reidlinger, J. & Müller, V. (1994). Purification of ATP synthase from Acetobacterium woodii and identification as a Na+-translocating F1Fo-type enzyme. Eur. J. Biochem. 223, 275–283. Neumann, S., Matthey, U., Kaim, G. & Dimroth, P. (1998). Purification and properties of the F1Fo ATPase of Ilyobacter tartaricus, a sodium ion pump. J. Bacteriol. 180, 3312–3316. Ferguson, S. A., Keis, S. & Cook, G. M. (2006). Biochemical and molecular characterization of a Na+-translocating F1Fo-ATPase from the thermoalkaliphilic bacterium Clostridium paradoxum. J. Bacteriol. 188, 5045–5054. Phillips, J. C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E., Villa, E. et al. (2005). Scalable molecular dynamics with NAMD. J. Comput. Chem. 26, 1781–1802. MacKerell, A. D., Bashford, D., Bellott, M., Dunbrack, R. L., Evanseck, J. D., Field, M. J. et al. (1998). All-atom empirical potential for molecular modeling and dynamics studies of proteins. J. Phys. Chem. B, 102, 3586–3616. Faraldo-Gómez, J. D., Smith, G. R. & Sansom, M. S.
Proton Binding to ATP Synthases
(2002). Setting up and optimization of membrane protein simulations. Eur. Biophys. J. 31, 217–227. 27. McCoy, A. J. (2007). Solving structures of protein complexes by molecular replacement with Phaser. Acta Crystallogr. Sect. D, 63, 32–41. 28. Emsley, P. & Cowtan, K. (2004). Coot: model-building
27 tools for molecular graphics. Acta Crystallogr. Sect. D, 60, 2126–2132. 29. Zwart, P. H., Afonine, P. V., Grosse-Kunstleve, R. W., Hung, L. W., Ioerger, T. R., McCoy, A. J. et al. (2008). Automated structure solution with the PHENIX suite. Methods Mol. Biol. 426, 419–435.
J. Mol. Biol. (2010) 395, 20–27
Available online at www.sciencedirect.com
COMMUNICATION
On the Structure of the Proton-Binding Site in the Fo Rotor of Chloroplast ATP Synthases Alexander Krah 1 , Denys Pogoryelov 2 , Thomas Meier 2,3 ⁎† and José D. Faraldo-Gómez 1,3 ⁎† 1
Theoretical Molecular Biophysics Group, Max Planck Institute of Biophysics, 60438 Frankfurt am Main, Germany 2
Department of Structural Biology, Max Planck Institute of Biophysics, 60438 Frankfurt am Main, Germany 3
Cluster of Excellence Macromolecular Complexes, Goethe University of Frankfurt, 60438 Frankfurt am Main, Germany Received 25 September 2009; received in revised form 23 October 2009; accepted 27 October 2009 Available online 31 October 2009
The recently reported crystal structures of the membrane-embedded proton-dependent c-ring rotors of a cyanobacterial F1Fo ATP synthase and a chloroplast F1Fo ATP synthase have provided new insights into the mechanism of this essential enzyme. While the overall features of these crings are similar, a discrepancy in the structure and hydrogen-bonding interaction network of the H+ sites suggests two distinct binding modes, potentially reflecting a mechanistic differentiation. Importantly, the conformation of the key glutamate side chain to which the proton binds is also altered. To investigate the nature of these differences, we use molecular dynamics simulations of both c-rings embedded in a phospholipid membrane. We observe that the structure of the c15 ring from Spirulina platensis is unequivocally stable within the simulation time. By contrast, the proposed structure of the H+ site in the chloroplast c14 ring changes rapidly and consistently into that reported for the c15 ring, indicating that the latter represents a common binding mode. To assess this hypothesis, we have remodeled the c14 ring by molecular replacement using the published structure factors. The resulting structure provides clear evidence in support of a common binding site conformation and is also considerably improved statistically. These findings, taken together with a sequence analysis of csubunits in the ATP synthase family, indicate that the so-called proton-locked conformation observed in the c15 ring may be a common characteristic not only of light-driven systems such as chloroplasts and cyanobacteria but also of a selection of other bacterial species. © 2009 Elsevier Ltd. All rights reserved.
Edited by B. Honig
Keywords: F1Fo ATP synthase rotor; membrane transport; atomic structure; molecular dynamics simulation; X-ray crystallography
F-type ATP synthases catalyze the conversion of ADP and Pi into ATP by coupling the structural changes required for this endergonic reaction to the transmembrane flow of H+ or Na+ down their electrochemical gradients. 1–4 The key coupling element in these molecular machines is the membrane-embedded Fo rotor, known as the c-ring. This structure is an assembly of 10–15 transmembrane *Corresponding authors. Max Planck Institute of Biophysics, Max-von-Laue-Str. 3, 60438 Frankfurt am Main, Germany. E-mail addresses: [email protected]; [email protected]. † T.M. and J.D.F.-G. contributed equally to this work. Abbreviations used: MD, molecular dynamics; PDB, Protein Data Bank.
helical hairpins (or c-subunits), which provide selective binding sites for the ions as they traverse the membrane. In conjunction with the adjacent asubunit (the stator) and its mechanical linkage to the catalytic domain through the γ-subunit (the axle), rotation of the c-ring thus allows ion flow and powers ATP synthesis. Conversely, ATP hydrolysis can cause the rotation of the c-ring in reverse, thereby generating a transmembrane gradient. The crystal structure of the Na+-coupled c-ring from Ilyobacter tartaricus provided the first atomically detailed view into the mechanism of ion binding to the Fo rotor.5 This ring comprises 11 subunits, with 11 Na+-binding sites, each of which lies in between adjacent subunits. The conserved carboxylate (E65) provides the main stabilizing interaction for the ion, but the site consists of another four coordinating
0022-2836/$ - see front matter © 2009 Elsevier Ltd. All rights reserved.
Proton Binding to ATP Synthases
ligands, including a bound structural water molecule.6 Altogether, this environment compensates for the energetic cost of desolvation and transfer into the membrane, and presumably confers the necessary specificity to the binding site. Recently, two crystal structures of H+-coupled crings—namely, from spinach chloroplast7 (consisting of 14 c-subunits) and from the cyanobacterium Spirulina platensis8 (made of 15 subunits)—have been reported. Both structures indicate that proton binding occurs through protonation of the conserved carboxylate (E61 and E62, respectively) and are consistent with previous measurements of the stoichiometry of these rings.9,10 However, these structural studies have also revealed differences in the proton-binding site suggestive of two distinct binding modes, despite the high level of similarity in the respective c-subunit amino acid sequences. In this study, we investigate the nature of these differences using computational methods for molecular dynamics (MD) simulation, crystallographic refinement, and sequence analysis, and we conclude that the so-called proton-locked conformation observed in the cyanobacterial c15 structure8 is, in fact, common to both c-rings and probably also to their homologues in many light-driven systems and in a selection of other bacterial species. Comparison of the published crystal structures of the c14 and c15 rings According to the conventional hydrogen-bond definition based on both the distance and the preferred geometries of donors and acceptors (dDA ≤ 3.6 Å; 120° ≤ θDHA ≤ 180°), the only interaction that is
21 common to the proton-binding sites in the reported crystal structures of the c14 and the c15 rings is that between E62:Oɛ1 and Y67:OH (Fig. 1) (E61 and Y66 in c14; all numberings hereafter correspond to c15). In contrast to the c15 ring, in the c14 structure, no hydrogen bond can be discerned between the protonated E62:Oɛ2 and the backbone carbonyl of F60, or with the amine group in the side chain of Q29. While the distance between E62:Oɛ2 and F60:O would, in principle, allow for a hydrogen bond (dDA = 2.8 Å), the carboxylic group would have to be in trans, which, according to spectroscopic and theoretical analyses, is unfavorable energetically and thus underrepresented statistically relative to the cis conformation. 11–14 Consistent with this, protonated carboxylate groups in ultra-high-resolution neutron diffraction protein structures have been observed, to our knowledge, only in the cis conformation.15–17 With regard to Q29, whose side chain is, in fact, absent in the published c14 structure, a modeled ideal geometry positions its Nɛ donor too distant from the closest carboxyl oxygen in E62 (dDA = 4.7 Å; cf. 3.0 Å in the c15 structure). Another intriguing difference is the rotamer of T65, which in the c15 structure is not directly engaged in the protonbinding site and instead hydrogen bonds to the backbone carbonyl of G21 in the inner helix. By contrast, in the c14 structure, T65 orients its hydroxyl group towards E62:Oɛ1, and while it is, in principle, within hydrogen-bonding distance (dDA = 3.0 Å), the orientation of these side chains is not consistent with such interaction (θDHA ∼ 105°). In summary, the binding site in the c14 structure appears to contain an unusual number of unfulfilled hydrogen-bond donors and acceptors, despite the likely energetic cost, especially in the hydrophobic
Fig. 1. Published crystal structures of the proton-binding site in (left) the c14 ring from spinach chloroplast7 and in (right) the c15 ring from S. platensis8 (PDB codes 2W5J and 2WIE, respectively). Note that the key side chain of Q28 was absent in the published structure of the chloroplast c14 ring and is therefore modeled. The view is sideways from within the membrane, with the outer C-terminal helices of two adjacent c-subunits in front and with one of the corresponding inner helices in the background. The possible hydrogen-bonding interactions in each case, on the basis of the distance between donor and acceptor, as well as their preferred chemical structure, are indicated with broken lines (see the main text).
22 environment 18 provided by detergent or lipid hydrocarbon chains. This is in contrast to the c15 site, where the interaction network is clearly defined and geometrically realistic. In view of this, the limited resolution of the c14 diffraction data (3.8 Å versus 2.1 Å for the c15 ring), and the high degree of similarity of these two sequences, we hypothesized that the H+-binding site in the c14 ring is, in fact, structurally identical with that in the c15 ring, despite the published structure. MD simulations of the c14 and c15 rings in a lipid membrane To examine this hypothesis further, we carried out a series of MD simulations of the c14 ring in a lipid
Proton Binding to ATP Synthases
membrane, in which we either preserved the conformation of the reported Cα trace or allowed complete flexibility. In both cases, we found that the reported structure of the binding site spontaneously changed into that observed for the c15 ring (Fig. 2). Analogous control simulations of the c15 ring showed no sign of structural instability (Fig. 2). More specifically, we observed that, consistently across all c-subunits in the chloroplast ring, E62 reorients into a geometry compatible with hydrogen bonds both with F60:O (via the protonated Oɛ2; in cis) and with Y67:OH (via Oɛ1). As seen in the c15 ring, E62:O ɛ2 also becomes an acceptor of a hydrogen bond donated by Q29:Nɛ. Interestingly, T65 also reorients in the course of the c14 simulations and forms a bond analogous to that with G21:O
Fig. 2. (a) Interaction distances and angles at the proton-binding site in the c15 ring from the crystal structure (broken lines) and MD simulations (continuous lines). (b) Same as (a), for the c14 ring and simulations where the Cα trace of the ring is constrained to be in the conformation in the published structure. (c) Same as (b), from simulations without any constraints. Simulation data are presented as probability distributions, computed from dynamical trajectories and combining all c-subunits. The corresponding distances/angles from the published structures are overlaid on the distributions with a vertical line. The conformation of the c-rings at the start of all simulations is that in the published crystal structures; the structural changes observed during the simulations of the c14 ring (b and c) are indicated with arrows. All MD simulations were carried out with NAMD24 using the CHARMM27 forcefield,25 at constant temperature (298 K), pressure (1 atm), and membrane surface area (∼ 69 Å2 per lipid). Electrostatic interactions were computed using the PME algorithm with a real-space cutoff distance at 12 Å, which was also used for the van der Waals interactions. Molecular models of the c14 and c15 crystal structures embedded in a hydrated POPC lipid membrane were prepared using the methodology described by Faraldo-Gómez et al.26; the systems include a total of ∼100,000 atoms each (see Supplementary Material, Fig. S1). This was followed by a series of MD simulations with gradually weaker constraints on the protein structure, for up to ∼ 10 ns. Both systems were subsequently simulated for 20 ns, allowing complete flexibility. In addition, we carried out five simulations of the c14 ring (for 5 ns each) in which Cα trace was constrained to remain as in the published crystal structure.
23
Proton Binding to ATP Synthases
in the cyanobacterial ring (namely, with A20). This is especially clear in simulations where the Cα trace is constrained; in unconstrained simulations, T65 is transiently engaged in other interactions (e.g., S22: Oγ), but not with E62 in any case. By contrast, the simulation of the c15 ring, which also allowed complete flexibility in the protein, showed excellent agreement between the crystal structure and the most populated geometry of the interaction network throughout the binding site, for an equivalent timescale. The time-averaged structures of the binding site derived from the c14 and c15 simulations, for a randomly selected c-subunit interface, are shown in Fig. 3. This comparison clearly bolsters the hypothesis that the proton-binding sites in these two rings are essentially identical. Lastly, it is worth pointing out that additional simulations of equal duration, where E62 was set in the unprotonated state, resulted in marked deviations from the published structures in both the c14 and the c15 rings. These structural changes are induced by the substantial penetration of water molecules into the membrane core, driven by the large energetic gain derived from the solvation of the charged carboxylate group (see Supplementary Material, Fig. S2). In this environment, the side chain of E62 spontaneously rotates away from the binding site and projects into the surrounding solvent. Thus, E62 no longer interacts with F60, Q29, or T65, and only partially interacts with Y67, but rather interacts with the neighboring water molecules. Moreover, these changes also influence the geometry of the ring as a whole, even in this short timescale (see Supplementary Material, Fig. S2). In summary, the characteristic conformation of the proton-binding site observed in charged-state simulations is completely inconsistent with the published structures of
the c14 and c15 rings (i.e., the differences between these two structures do not reflect a different protonation state). This result is reassuring, since the likelihood of the charged state is remote while E62 is embedded in a solvent-excluded hydrocarbon environment and lacks a suitable counterion (such as Na+ in the c11 structure). It is conceivable that a H3O+ could reside in the c14 site, as hypothesized elsewhere.19 However, neither the crystallographic analysis of Vollmar et al.7 nor ours (see the text below) indicates the presence of a hydronium in this site. Crystallographic structure refinement of the c14 ring To provide a degree of experimental certainty to our hypothesis of a common mode of proton binding, we recalculated the c14 structure by molecular replacement, on the basis of the highresolution structure of the c15 ring and the c14 structure factors deposited in the Protein Data Bank (PDB). The new experimental structure resulting from this analysis is shown in Fig. 4a alongside the electron density map. While the limited resolution precludes full atomic detail (e.g., T65 and Q29 can only be partially resolved), the orientation of E62, consistent with a hydrogen bond between E62:Oɛ2 and F60:O, can be clearly discerned in the omit map of all 14 c-subunits in the ring, thus confirming the proton-locked conformation observed in the c15 ring. The resolution of the other side chains originally missing in the published structure is also improved (e.g., L15, I22, Q28, Q34, E37, I49, M60, L74, N78, and F80; in c14 numbering). Consistently, the refinement statistics from this
Fig. 3. Time-averaged simulation structures of the proton-binding site in (left) the c14 ring from spinach chloroplast and in (right) the c15 ring from S. platensis. The view is the same as in Fig. 1. Nonpolar hydrogen atoms are omitted for clarity. The average structure of the c14 ring derives from Cα-trace constrained simulations (Fig. 2b); that for the c15 ring from an unconstrained simulation (Fig. 2a).
24
Proton Binding to ATP Synthases
Fig. 4. (a) Remodeled crystal structure of the proton-binding site in the c14 ring from spinach chloroplast, derived from the structure factors deposited in PDB code 2W5J.7 The electron density map (2Fobs − Fcalc; blue mesh) is contoured at 2σ, and the difference map (Fobs − Fcalc; green mesh) is contoured at 3σ. The structure was solved by molecular replacement based on the c 15 ring from S. platensis (PDB code 2WIE),8 omitting the side chain of E62 (E61 in the c14 sequence) to avoid a bias in the resulting map. Initial phases were obtained with Phaser27 using a search model of a c14 ring composed of c-subunits from the c15 ring. The model was improved by iterative rounds of model building with Coot28 into the 2Fobs − Fcalc electron density map and further refined with PHENIX.29 Noncrystallographic symmetry averaging was applied at all steps of the refinement process. (b) Per-residue root-mean-square difference between the published model of the c14-ring structure (PDB code 2W5J) and the remodeled structure described in this study. The average root-mean-square differences (residues 4–79) are 1.2 Å and 1.7 Å for the main chain and the side chains, respectively.
analysis are improved relative to the original (i.e., we find a better correspondence between the threedimensional structural model and the measured data); in particular, the Rwork and Rfree factors decreased to 24.8% and 30.9%, respectively, compared with 33% and 36% in Vollmar et al.7 As shown in Fig. 4b, these improvements reflect not only the rearrangement of the binding site and the added side chains but also changes elsewhere in the structure, particularly at the N-terminus of the csubunits, where the main chain was seen to adopt a different conformation.
It seems clear, therefore, that the c15 ring is a more appropriate molecular replacement template than the Na+-coupled I. tartaricus c11 ring employed originally,7 although it should be noted that only the latter was available at that time. At any rate, the resolution of the underlying diffraction data remains very low (3.8 Å); thus, the determination of the complete atomic structure of the c14 ring may be considered as a work in progress. Nevertheless, the features of the proton-binding site appear to be already well defined, as judged from the analysis presented here.
Proton Binding to ATP Synthases
Sequence analysis and consensus structure of the proton-binding site Figure 5 shows the amino acid sequence alignment of c-subunits from F-ATP synthases in a selection of species/cell organelles, including bacteria, mitochondria, and chloroplasts. In the majority of these species, the enzyme is driven by the protonmotive force, implying that a proton must reversibly bind to the c-ring during ion translocation. However, in a handful of species including I. tartaricus, ATP synthases bind Na+ instead20–23 (i.e., they are driven by the sodium-motive force). Although a Na+binding signature sequence can be identified in these cases,6 an ion-binding motif in the protoncoupled c-rings, aside from the strictly conserved carboxyl group (E or D), is less clearly defined. The motif varies from an almost complete Na+-binding signature (e.g., Solibacter usitatus or Desulfovibrio vulgaris) to a ‘naked’ glutamate/aspartate (e.g., Bacillus pseudofirmus OF4 or Escherichia coli). In the c15 ring of the cyanobacterium S. platensis, the side chains of Q29, E62, and Y60 are directly involved in the proton coordination network. This sequence motif (Q/E/Y) is observed in most of the c-subunits in the phylum of Cyanobacteria, as well as in their ‘derivatives,’ the chloroplasts. Notably, these two types of cells/organelles use light as the primary energy source to generate their proton-motive force, a feature that distinguishes them from many other species. In addition, this common motif (Q/E/Y) is also found in a selection of bacteria in various phyla (e.g., Proteobacteria). Therefore, either it evolved
25 independently several times or it was determined early on in the evolution of the bacterial kingdom. In any case, the ATP synthase and its ion binding specificity are likely adapted to the unique physiological requirements of each cell/organelle type. It is therefore logical that not all species include the exact same hydrogen-bonding network seen here for the c14 and c15 rings, thus endowing these enzymes with the necessary flexibility to operate optimally in a variety of environments. As mentioned above, in many of the bacterial species, as well as in all eukaryotic mitochondria (animals, plants, and fungi), the proton-binding motif is limited to the conserved E/D carboxylate. The energetics of the ion translocation mechanism may therefore be different in this subclass of Fo rotors; conclusive structural data for a representative c-ring belonging to this prominent subclass are not available and would therefore be of the highest interest. In conclusion, based on the finding of a consensus structure of the H+-binding sites in the c14 and c15 rings, and on the sequence similarity of c-subunits across the F-ATP synthase family, we propose that the proton-locked conformation observed at high resolution for the c15 ring represents a common mode of proton binding in the c-rings of a wide variety of cell types and organelles, including cyanobacteria, chloroplasts, and other bacterial strains carrying the Q/E/Y sequence motif. By extension, our findings contradict the notion that proton binding to the chloroplast Fo rotors involves a hydronium associated with the key negatively charged glutamate side chain.19 Nevertheless, the
Fig. 5. Alignment of c-subunit sequences from F-type ATP synthases. The sequences are aligned according to the cytoplasmic loop region (bold lettering) of each c-subunit. Residues involved in H+ or Na+ coordination are highlighted with colors. The residue numbering and secondary-structure elements in the S. platensis c-subunit, according to the c15 crystal structure,8 are given at the top. Note that this set of sequences is a selection of representative c-subunit types and is not weighted according to their overall occurrence.
26 precise construction of the interaction network at the ion-binding sites of different c-rings likely reflects a fine-tuned adaptation to the specific physiological requirements and environment of each cell; therefore, subtle variations of this structural motif can be expected to occur (e.g., in mitochondria). Notwithstanding this variability, it is plausible that, from a mechanistic standpoint, the general principles of ion translocation outlined, for example, for the cyanobacterial ATP synthase8 or elsewhere3 will be valid across this fascinating family of motor enzymes.
Proton Binding to ATP Synthases
10. 11.
12.
13.
Acknowledgements This work was supported, in part, by the Cluster of Excellence “Macromolecular Complexes” of the Goethe University Frankfurt (DFG Project EXC 115). We thank the students of the IMPReS program at the Max Planck Institute of Biophysics for carrying out a subset of the computer simulations of the c14 ring presented here.
Supplementary Data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jmb.2009.10.059
14. 15.
16.
17.
18.
References 1. Boyer, P. D. (1997). The ATP synthase: a splendid molecular machine. Annu. Rev. Biochem. 66, 717–749. 2. Yoshida, M., Muneyuki, E. & Hisabori, T. (2001). ATP synthase: a marvellous rotary engine of the cell. Nat. Rev. Mol. Cell Biol. 2, 669–677. 3. Junge, W., Sielaff, H. & Engelbrecht, S. (2009). Torque generation and elastic power transmission in the rotary FoF1-ATPase. Nature, 459, 364–370. 4. Nakamoto, R. K., Baylis Scanlon, J. A. & Al-Shawi, M. K. (2008). The rotary mechanism of the ATP synthase. Arch. Biochem. Biophys. 476, 43–50. 5. Meier, T., Polzer, P., Diederichs, K., Welte, W. & Dimroth, P. (2005). Structure of the rotor ring of F-type Na+-ATPase from Ilyobacter tartaricus. Science, 308, 659–662. 6. Meier, T., Krah, A., Bond, P. J., Pogoryelov, D., Diederichs, K. & Faraldo-Gómez, J. D. (2009). Complete ion-coordination structure in the rotor ring of Na+-dependent F-ATP synthases. J. Mol. Biol. 391, 498–507. 7. Vollmar, M., Schlieper, D., Winn, M., Buchner, C. & Groth, G. (2009). Structure of the c14 rotor ring of the proton translocating chloroplast ATP synthase. J. Biol. Chem. 284, 18228–18235. 8. Pogoryelov, D., Yidiz, O., Faraldo-Gómez, J. D. & Meier, T. (2009). High-resolution structure of the rotor ring of a proton dependent ATP synthase. Nat. Struct. Mol. Biol. 16, 1068–1073. 9. Pogoryelov, D., Yu, J., Meier, T., Vonck, J., Dimroth, P. & Muller, D. J. (2005). The c15 ring of the Spirulina
19. 20.
21.
22.
23.
24.
25.
26.
platensis F-ATP synthase: F1/Fo symmetry mismatch is not obligatory. EMBO Rep. 6, 1040–1044. Seelert, H., Poetsch, A., Dencher, N. A., Engel, A., Stahlberg, H. & Müller, D. J. (2000). Proton-powered turbine of a plant motor. Nature, 405, 418–419. Briggs, J. M., Nguyen, T. B. & Jorgensen, W. L. (1991). Monte Carlo simulations of liquid acetic acid and methyl acetate with the OPLS potential function. J. Phys. Chem. 95, 3315–3322. Wiberg, K. B. & Laidig, K. E. (1987). Barriers to rotation adjacent to double bonds: 3. The C–O barrier in formic acid, methyl formate, acetic acid, and methyl acetate. The origin of ester and amide “resonance”. J. Am. Chem. Soc. 1987, 5935–5943. Ruschin, S. & Bauer, S. H. (1980). Rotational isomers of methyl formate, located by the temperature-drift technique. J. Phys. Chem. 84, 3061–3065. Bertie, J. E. & Michaelian, K. H. (1982). The Raman spectra of gaseous formic acid –h2 and –d2. J. Chem. Phys. 76, 886. Coates, L., Tuan, H. F., Tomanicek, S., Kovalevsky, A., Mustyakimov, M., Erskine, P. & Cooper, J. (2008). The catalytic mechanism of an aspartic proteinase explored with neutron and X-ray diffraction. J. Am. Chem. Soc. 130, 7235–7237. Deacon, A., Gleichmann, T., Kalb, A. J., Price, H., Raftery, J., Bradbrook, G. et al. (1997). The structure of concanavalin A and its bound solvent determined with small-molecule accuracy at 0.94 Å. J. Chem. Soc. Faraday Trans. 93, 4305–4312. Habash, J., Raftery, J., Weisgerber, S., Casseta, A., Lehmann, M. S., Hoghoj, P. et al. (1997). Neutron Laue diffraction study of concanavalin A—the proton of Asp28. J. Chem. Soc. Faraday Trans. 93, 4313–4317. Gao, J., Bosco, D. A., Powers, E. T. & Kelly, J. W. (2009). Localized thermodynamic coupling between hydrogen bonding and microenvironment polarity substantially stabilizes proteins. Nat. Struct. Mol. Biol. 16, 684–690. von Ballmoos, C. & Dimroth, P. (2007). Two distinct proton binding sites in the ATP synthase family. Biochemistry, 46, 11800–11809. Laubinger, W. & Dimroth, P. (1987). Characterization of the Na+-stimulated ATPase of Propionigenium modestum as an enzyme of the F1Fo type. Eur. J. Biochem. 168, 475–480. Reidlinger, J. & Müller, V. (1994). Purification of ATP synthase from Acetobacterium woodii and identification as a Na+-translocating F1Fo-type enzyme. Eur. J. Biochem. 223, 275–283. Neumann, S., Matthey, U., Kaim, G. & Dimroth, P. (1998). Purification and properties of the F1Fo ATPase of Ilyobacter tartaricus, a sodium ion pump. J. Bacteriol. 180, 3312–3316. Ferguson, S. A., Keis, S. & Cook, G. M. (2006). Biochemical and molecular characterization of a Na+-translocating F1Fo-ATPase from the thermoalkaliphilic bacterium Clostridium paradoxum. J. Bacteriol. 188, 5045–5054. Phillips, J. C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E., Villa, E. et al. (2005). Scalable molecular dynamics with NAMD. J. Comput. Chem. 26, 1781–1802. MacKerell, A. D., Bashford, D., Bellott, M., Dunbrack, R. L., Evanseck, J. D., Field, M. J. et al. (1998). All-atom empirical potential for molecular modeling and dynamics studies of proteins. J. Phys. Chem. B, 102, 3586–3616. Faraldo-Gómez, J. D., Smith, G. R. & Sansom, M. S.
Proton Binding to ATP Synthases
(2002). Setting up and optimization of membrane protein simulations. Eur. Biophys. J. 31, 217–227. 27. McCoy, A. J. (2007). Solving structures of protein complexes by molecular replacement with Phaser. Acta Crystallogr. Sect. D, 63, 32–41. 28. Emsley, P. & Cowtan, K. (2004). Coot: model-building
27 tools for molecular graphics. Acta Crystallogr. Sect. D, 60, 2126–2132. 29. Zwart, P. H., Afonine, P. V., Grosse-Kunstleve, R. W., Hung, L. W., Ioerger, T. R., McCoy, A. J. et al. (2008). Automated structure solution with the PHENIX suite. Methods Mol. Biol. 426, 419–435.