STRUCTURAL BIOLOGY

Atomic model for the dimeric FOregion of mitochondrial ATP synthaseHui Guo,1,2 Stephanie A. Bueler,1 John L. Rubinstein1,2,3*

Mitochondrial adenosine triphosphate (ATP) synthase produces the majority of ATP ineukaryotic cells, and its dimerization is necessary to create the inner membrane folds, orcristae, characteristic of mitochondria. Proton translocation through the membrane-embedded FO region turns the rotor that drives ATP synthesis in the soluble F1 region.Although crystal structures of the F1 region have illustrated how this rotation leads to ATPsynthesis, understanding how proton translocation produces the rotation has been impededby the lack of an experimental atomic model for the FO region. Using cryo–electronmicroscopy, we determined the structure of the dimeric FO complex from Saccharomycescerevisiae at a resolution of 3.6 angstroms. The structure clarifies how the protons travelthrough the complex, how the complex dimerizes, and how the dimers bend the membraneto produce cristae.

During mitochondrial respiration, the pro-tein complexes of the electron transportchain pump protons from the mitochon-drial matrix to the intermembrane space(IMS). The resulting proton motive force

across themitochondrial innermembranepowersthe adenosine triphosphate (ATP) synthase, amembrane-embedded multiprotein complex.Saccharomyces cerevisiae ATP synthase has asubunit composition similar to that of themam-malian enzyme (1). The complex comprises a sol-uble catalytic F1 region consisting of subunitsa3b3gde and a membrane-embedded FO regioncontaining subunits a, b, e, f, g, i/j, k, l, 8, and aring of 10 c-subunits (Fig. 1A). During ATP syn-thesis, proton translocation from the IMS to thematrix through the FO region, at the interface ofsubunit a and the c-ring, drives rotation of thecentral rotor subcomplex (subunits gdec10). Thisrotation induces conformational changes in thea3b3 hexamer of the F1 region that lead to syn-thesis of ATP from adenosine diphosphate (ADP)andphosphate (2). Proton translocation is thoughtto occur through two offset half-channels, oneopening to the IMS and the other to the mito-chondrial matrix (3, 4). Protons from the IMShalf-channel neutralize conservedGlu residues inthe c10-ring (Glu

59 in S. cerevisiae), travel throughthe lipid bilayer as the ring rotates, and aredelivered to the matrix half-channel. The a3b3complex is prevented from rotating alongwith thecentral rotor by the peripheral stalk subcomplex,consisting of subunits OSCP (oligomycin sensitiv-ity conferral protein), d, h, and the soluble regionof subunit b. Mammals lack subunits i/j, k, and lbut possess the additional subunits DAPIT and6.8PL (5).

Mitochondrial ATP synthase complexes assem-ble into long ribbons of dimers (6, 7) that bendthe mitochondrial inner membrane as the resultof the presence of a domain containing subunitse and g in the FO region (7–11). These bends areneeded to form cristae, highly folded membranestructures that maximize the surface area avail-able for respiration, giving mitochondria theircharacteristic appearance (6, 7). X-ray crystallog-raphy has revealed the atomic structures of mostof the soluble subunits of mitochondrial and bac-terial ATP synthase (2). However, most of themembrane-embedded region of the complex (sub-units a, b, e, f, g, i/j, k, l, and 8) has resisted high-resolution structural characterization. Despiterecent advances, the resolution in cryo–electronmicroscopy (cryo-EM) maps of intact ATP syn-thases has been insufficient to allow constructionof atomic models (9, 12–15).Cryo-EM has shown that rotary adenosine

triphosphatases (ATPases) stop in different rota-tional states when extracted from membranes(9, 13, 15, 16). We hypothesized that this con-formational heterogeneity limits cryo-EM reso-lution and that analysis of dissociated FO regionswould reduce heterogeneity, improve resolution,and enable construction of an experimentalatomic model. A similar approach was success-ful with the related VO complex from the S. cere-visiae V-type ATPase (17). However, whereas yeastV-ATPases dissociate in response to glucose de-pletion (18), ATP synthases do not offer a phys-iological route to separating the F1 region fromthe FO region. Therefore, we used sodium bro-mide to dissociate the F1 regions from inside-outsubmitochondrial particles (19). Similar to thenatural product digitonin (9), we found thatthe recently described synthetic detergent glyco-diosgenin (GDN) (20) extracts the dimeric formof ATP synthase fromyeast mitochondrial mem-branes for both the intact ATP synthase andthe FO complex (Fig. 1B). Cryo-EM of the FO com-plex allowed for determination of the structureto 3.6 Å resolution (Fig. 1, C to E, and fig. S1).

Side-chain detail could be seen for subunits a,f, i/j, and 8; the transmembrane a helix of k;the c-subunits; the transmembrane a helices ofsubunit b; and part of subunit d. This resolu-tion allowed construction of atomic models forthese subunits (fig. S2 and table S1). Density forsubunits e and g allowed construction of poly-alanine models for these proteins. The recentlyidentified subunit l, which has no known func-tion and a sequence similar to that of subunit k(1), could not be identified in the cryo-EMmap.Together with existing atomic models of ATPsynthase subcomplexes, the FO complex struc-ture determined here allows for constructionof a nearly complete “mosaic model” (2) of mito-chondrial ATP synthase (fig. S3).The core subunits of the FO complex, found in

both mitochondria and bacteria, are subunits a,b, and c. The fold observed for subunit a (Fig. 2,A and B, green) matches the fold deduced froma low-resolution map of the bovine enzyme com-bined with evolutionary covariance analysis (13).The protein contains five transmembrane a heli-ces and an amphipathic a helix (a helix 2) thatlies along the matrix surface of the detergentmicelle; a helices 3 and 4 form a transmembranehairpin, as do a helices 5 and 6, with the latterbeing the long and tilted hairpin that is char-acteristic of rotary ATPases (12, 13, 15, 16, 21). Thesimpler bacterial ATP synthase contains two copiesof subunit b, each with a single N-terminal trans-membrane a helix and an elongated C-terminalregion that constitutes most of the peripheralstalk (15). In contrast, the mitochondrial enzymehas a single subunit b with two N-terminal trans-membrane a helices. The two transmembrane ahelices of the mitochondrial subunit do not packagainst eachother, as suggestedpreviously (9, 13, 14).Instead, subunit b begins with a short, presum-ably amphipathic a helix and a transmembrane ahelix that together form a domainwith subunits eand g (Fig. 2A, dashed line). A short loop connectsthis domain to the second transmembrane a helixof subunit b, which packs against subunit a. Thesoluble C-terminal portion of subunit b enters themitochondrial matrix as part of the peripheralstalk (22, 23) and is mostly disordered in theisolated FO complex.Subunits f and 8, which are essential for mito-

chondrial respiration, previously had no clearfunction in the ATP synthase. Subunit f (Fig. 2, Aand C, yellow) consists of a soluble N-terminalsequence of ~50 residues that binds to the baseof the peripheral stalk, followed by a singletransmembrane a helix. The transmembrane ahelix of subunit f was mistaken for the firsttransmembrane a helix of subunit b in lower-resolution cryo-EM studies of the bovine (13)and yeast (9, 14) enzymes. Subunit 8 (Fig. 2, Aand C, pink), known as A6L in mammals, is al-most entirely a-helical, with theN-terminal partof the helix embedded in the membrane incontact with the first a helix of subunit a andthe C-terminal 14 residues contributing to thebase of the peripheral stalk. Subunits f and 8 cantherefore be assigned as components of theperipheral stalk. The C-terminal 56 residues of

RESEARCH

Guo et al., Science 358, 936–940 (2017) 17 November 2017 1 of 5

1Hospital for Sick Children Research Institute, Toronto,Ontario M5G 0A4, Canada. 2Department of Medical Biophysics,University of Toronto, Toronto, Ontario M5G 1L7, Canada.3Department of Biochemistry, University of Toronto, Toronto,Ontario M5S 1A8, Canada.*Corresponding author. Email: john.rubinstein@utoronto.ca

on July 14, 2020

http://science.sciencemag.org/

Dow

nloaded from

subunit d (Fig. 2, A and C, brown), which wereabsent from earlier crystal structures (22, 23), wraparound subunit 8 and subunit b as they leave thedetergent micelle to clamp together the base ofthe peripheral stalk. The structure shows that thebase of the peripheral stalk in the mitochondrialATP synthase has a more complicated arrange-ment of subunits than predicted (22, 23) or foundin the bacterial enzyme (15).Unlike thedimericATPsynthase fromPolytomella

sp., which is held together by subunits specificto algae (12), the yeast dimer is held together

by subunits a, i/j, k, and e (Fig. 3, A to C). Theinterface formed by subunit i/j, which was mis-taken for subunit f in a lower-resolution struc-ture (9), occurs through two short stretches of~10 residues that pack together in the dimer(Fig. 3, A and B, magenta). The lack of homo-logs for subunit i/j in mammals could explainwhy yeast ATP synthase dimers are more stablethan mammalian dimers (9). Unexpectedly, sub-unit a also forms a dimer interface, with eachmonomer contributing two strands of a four-stranded planar structure (Fig. 3, A and B, green)

with one hydrophobic surface and one hydro-philic surface. This motif superficially resem-bles an antiparallel b sheet, with amino acidside chains pointing away from the plane of themotif. The two dimerization motifs are stackedtogether, with the subunit i/j motif in the IMSand the subunit a motif on the IMS surface ofthe detergent micelle. In an alignment of sub-unit a sequences from different species (9), thedimerization motif begins just four residues N-terminal of the first residue in the bovine andhuman subunit a, which have shorter N-terminal

Guo et al., Science 358, 936–940 (2017) 17 November 2017 2 of 5

F1FO dimer

FO dimer

Mitochondrial matrix

Intermembrane space

gbe

f

ka

8

d

c-ring

i/j

O

90°

Matrix

Intermembrane space

OSCPh

γε

δ

a,f,i/j,k,8c-ring

g

e

bd

α

βα

Fig. 1. Overall structure of the FO complex. (A) Cartoon of the ATP synthasedimer, with subunits found in the FO region outlined in black. (B) The detergentGDN extracts dimeric S. cerevisiae ATP synthase frommitochondrial membranesintact (left) or as the FO complex after sodium bromide treatment of themembranes (right). Scale bar, 25 Å. (C) Example of map density that allowed

construction of an atomic model (subunit a, residues Ala168 to Phe196 and Pro212

to Leu242). Carbon, oxygen, nitrogen, and sulfur atoms are colored green, red,blue, and yellow, respectively. Scale bar, 5 Å. (D and E) Top and side views,respectively, of the FO dimer, revealing the arrangement of subunits. Onemonomer is outlined with a dashed line in (D). Scale bar, 25 Å.

RESEARCH | REPORTon July 14, 2020

http://science.sciencemag.org/

Dow

nloaded from

Guo et al., Science 358, 936–940 (2017) 17 November 2017 3 of 5

7 891012

43

65

a

8

ij/

d

k

b

f

e

g

base ofperipheralstalk

a

b

fij/

k

d

8

1

2

3 4

5

6

matrix

intermembranespace

Fig. 2. Subunit folds for individual subunits. (A) Matrix view of the FOcomplex monomer structure. A dashed line encloses the domain formedby subunits e and g and the N-terminal a helices of subunit b. (B) Fold of

subunit a. (C) Side view of the FO monomer showing that subunits 8, f, andb contribute to the base of the peripheral stalk, with subunit d acting as aclip. Scale bar, 25 Å.

180°

viewed from matrix

viewed from IMS

subunit k (monomer 1)subunit e (monomer 2)

ai/j

i/ja

kgeb

Membrane curvature

segmented mapunsharpened

Mitochondrial matrix

Intermembrane space

Fig. 3. Dimerization and membrane bending. (A) Subunits a and i/jcontribute dimerization motifs. (B) Enlargement of (A). (C) The unsharp-ened map shows that subunit k extends into the IMS to contact subunit e.

(D and E) Top and side views of the FO dimer show that subunits b, e, g,and k create the structure that bends the lipid bilayer by almost 90°. Thedashed orange line indicates the full length of subunit k. Scale bars, 25 Å.

RESEARCH | REPORTon July 14, 2020

http://science.sciencemag.org/

Dow

nloaded from

regions than the yeast subunit a. Consequently,the mammalian subunit a dimerization motiflikely begins immediately before the subunit’sfirst transmembrane a helix. Subunits k and econtribute the final monomer-monomer inter-action. Subunit k possesses anN-terminal trans-membrane a helix in the map that is well ordered(Fig. 2, orange). Inspection of an unsharpenedversion of the map, which allows visualiza-tion of lower-resolution features, shows addi-tional density from subunit k (Fig. 3C, orange).This density continues into the IMS to contactsubunit e, which extends into the IMS as an ahelix (Fig. 2A and Fig. 3C, blue). The lack ofthis interaction in the bovine ATP synthasemonomer may explain why the extended IMSa helix of subunit e is in different orientationsin the monomeric bovine ATP synthase and in-tact dimeric yeast ATP synthase (9, 13). It hasbeen proposed that dissociation of ATP syn-thase dimers triggers opening of the mitochon-drial permeability transition pore, which caninitiate regulated cell death (24). Consequent-ly, the dimer contacts described above war-rant further investigation for their role in thisphenomenon.The unusually shaped domain containing sub-

units e, g, and the N-terminal ~50 residues ofsubunit b (Fig. 3, D and E) is responsible forbending the mitochondrial inner membrane(11, 25). Subunit e (Fig. 3, D and E, blue) beginswith an N-terminal transmembrane a helix thatcontinues ~40 Å into the IMS. Although thisfeature is conserved from yeast to mammals, itdoes not appear to have a role in proton trans-location or rotary catalysis. Subunit g has twoN-terminal a helices that lie on the matrix sur-face of the detergent micelle and a single trans-membrane a helix that interacts with subunite, probably via the conserved Gly-X-X-X-Glymotifs of the two proteins (26, 27). The curvedstructure of the domain formed by subunits e, g,and b offers an explanation for how subunits eand g bend the lipid bilayer and why deletionof the genes for these subunits in yeast leadsto defects in cristae formation (8). The combi-nation of membrane-surface and membrane-embedded a helices from these subunits, withfurther support from subunit k, creates a curvedstructure that enforces curvature of the lipidbilayer itself. The membrane surface a helicesof subunits g and b are reminiscent of the struc-ture of BAR domains, which also induce cur-vature in lipid bilayers through amphipathic ahelices (28).Proton translocation in ATP synthase involves

subunit a and the c-ring. Remarkably, the c10-ringhas a well-defined orientation relative to subunita (Fig. 4), at least after the rapid cooling that occursduring cryo-EM specimen preparation. This con-strained orientation shows that the highly tilteda helices of subunit a maintain a tight interac-tion with the ring, although the rotor must stillbe able to turn during ATP synthesis. Two poresare visible in the FO complex that correspond tothe two half-channels required for proton trans-location (3, 4). A cavity on the IMS side of the

detergent micelle behind the tilted a helices 5and 6 forms the IMS half-channel (Fig. 4A, left,circled in red) while an opening on the matrixside of the detergent micelle between subunit aand the c-ring forms the matrix half-channel(Fig. 4A, right, circled in red) (12, 13, 17). Thepositions of these half-channels are consistentwith the proton translocation path proposedfrom lower-resolution studies of rotary ATPases(9, 12, 13, 16, 17, 21). Arg176 from subunit a, theonly absolutely required residue of the subunitfor enzyme activity (29, 30), extends toward the

c-ring (Fig. 4B). X-ray crystallography and simu-lations of isolated c-rings have shown that underdeprotonating conditions, the proton-carryingGlu residues adopt an extended conformationwith their carboxyl groups facing away from thecenter of the c-ring (31). Under protonating con-ditions, these residues fold toward the center ofthe ring in a conformation known as the proton-locked state (32). Although Glu and Asp residuesare particularly sensitive to radiation damage andmay have lowdensity in cryo-EMmaps (33), theseresidues often remain apparent (34, 35) and the

Guo et al., Science 358, 936–940 (2017) 17 November 2017 4 of 5

viewed from matrixviewed from IMS

unsharpened map

R176

E59

--

neutralneutral

+

detergen

t

de

tergent

c10-ring

Matrix channel

IMS channel

subunit a

chhanannel

Matrix cha

10 g

E59

D244

H+

subunit a

Matrix channel

IMS channel

R176

subu

nit b

1

E162

2

3

5

6

E223

E59

E223

D244

R176

4

E162

H+ H+

H+

c10-ring

Fig. 4. Proton translocation mechanism. (A) The unsharpened map reveals cavities in thecomplex that correspond to the IMS (left) and matrix (right) half-channels. (B) The Glu59 residuesof the two c-subunits nearest to Arg176 are in extended (deprotonated) conformations. Theremaining Glu59 residues are in proton-bound conformations. (C) Two patches of negative charge onsubunit a (left) where it interacts with the c10-ring (right) correspond with expected positions of theIMS and matrix half-channels. (D) During proton translocation, protons follow the path indicated by theyellow line (left), entering the IMS half-channel behind transmembrane a helices 5 and 6 of subunita and exiting through the matrix half-channel between a helices 5 and 6 and the c-ring (right).Scale bars, 25 Å.

RESEARCH | REPORTon July 14, 2020

http://science.sciencemag.org/

Dow

nloaded from

Glu59 residues closest to the c10-ring are readilyvisible in the map (Fig. 4B). From the cryo-EMmap (Fig. 4B), the conserved Glu59 residues ofthe two c-subunits nearest to Arg176 appear to bein deprotonated (negatively charged) conforma-tions (31) but too far from Arg176 for either Glu toformadirect salt bridgewithArg176. The remainingc-subunits appear to be in the proton-bound(neutral) conformation (Fig. 4B). Consequently,the resting conformation of the c-ring in the FOstructure differs subtly from the V-type enzyme,where the Glu residue from one of the c-subunitswas close enough to interact with the conservedArg (17).The surface of subunit a that contacts the c-ring

is mostly hydrophobic, with a patch of positivecharge fromArg176 (Fig. 4C, blue) and two patchesof negative charge (red) at the expected posi-tions of the half-channels. Near the matrix sur-face, these negative patches are due to Glu162

and Asp244, whereas near the IMS surface thenegative charge is due to Glu223; all of theseresidues are conserved (9) and likely pass pro-tons between the half-channels and the Gluresidues of the c-ring. Numerous other con-served and functionally characterized residuesof subunit a surround the locations of the twohalf-channels (fig. S4). The positions of thehalf-channels and the orientations of the Argand Glu residues support an emerging modelfor proton translocation: Donation of a protonfrom the aqueous IMS half-channel via Glu223

neutralizes the Glu59 residue of the subunit cadjacent to the channel. Neutralization of theGlu residue allows the c-ring to rotate as a re-sult of Brownian motion (3, 4), counterclock-wise when viewed from F1 toward FO (Fig. 4D),thereby placing the residue in the hydrophobicenvironment of the lipid bilayer. Rotation ofthe c-ring brings a neutral Glu59 residue from adifferent c-subunit into thematrix half-channel,where it is stripped of its proton by Arg176 ofsubunit a. This proton is accepted by Glu162 orAsp244, which in turn lose the proton to the

mitochondrial matrix. Loss of the proton fromthe Glu59 residue in the matrix half-channelresets the motor for the next proton. The di-rectionality and force of rotation in this modelis governed by the difference in probability of ac-subunit binding a proton from either half-channel, which depends on the DG of protontranslocation established by the proton motiveforce.

REFERENCES AND NOTES

1. S. Liu et al., Biochem. J. 468, 167–175 (2015).2. J. E. Walker, Biochem. Soc. Trans. 41, 1–16 (2013).3. S. B. Vik, B. J. Antonio, J. Biol. Chem. 269, 30364–30369

(1994).4. W. Junge, H. Lill, S. Engelbrecht, Trends Biochem. Sci. 22,

420–423 (1997).5. R. Chen, M. J. Runswick, J. Carroll, I. M. Fearnley, J. E. Walker,

FEBS Lett. 581, 3145–3148 (2007).6. R. D. Allen, C. C. Schroeder, A. K. Fok, J. Cell Biol. 108,

2233–2240 (1989).7. M. Strauss, G. Hofhaus, R. R. Schröder, W. Kühlbrandt,

EMBO J. 27, 1154–1160 (2008).8. P. Paumard et al., EMBO J. 21, 221–230 (2002).9. A. Hahn et al., Mol. Cell 63, 445–456 (2016).10. C. Jiko et al., eLife 4, e06119 (2015).11. L. A. Baker, I. N. Watt, M. J. Runswick, J. E. Walker, J. L. Rubinstein,

Proc. Natl. Acad. Sci. U.S.A. 109, 11675–11680 (2012).12. M. Allegretti et al., Nature 521, 237–240 (2015).13. A. Zhou et al., eLife 4, e10180 (2015).14. K. R. Vinothkumar, M. G. Montgomery, S. Liu, J. E. Walker,

Proc. Natl. Acad. Sci. U.S.A. 113, 12709–12714 (2016).15. M. Sobti et al., eLife 5, e21598 (2016).16. J. Zhao, S. Benlekbir, J. L. Rubinstein, Nature 521, 241–245

(2015).17. M. T. Mazhab-Jafari et al., Nature 539, 118–122 (2016).18. P. M. Kane, J. Biol. Chem. 270, 17025–17032 (1995).19. A. Tzagoloff, J. Biol. Chem. 245, 1545–1551 (1970).20. P. S. Chae et al., Chem. Eur. J. 18, 9485–9490 (2012).21. D. G. Schep, J. Zhao, J. L. Rubinstein, Proc. Natl. Acad. Sci. U.S.A.

113, 3245–3250 (2016).22. V. K. Dickson, J. A. Silvester, I. M. Fearnley, A. G. W. Leslie,

J. E. Walker, EMBO J. 25, 2911–2918 (2006).23. D. M. Rees, A. G. Leslie, J. E. Walker, Proc. Natl. Acad. Sci. U.S.A.

106, 21597–21601 (2009).24. M. Bonora et al., EMBO Rep. 18, 1077–1089 (2017).25. G. Arselin et al., J. Biol. Chem. 279, 40392–40399 (2004).26. D. M. Bustos, J. Velours, J. Biol. Chem. 280, 29004–29010

(2005).27. S. Saddar, R. A. Stuart, J. Biol. Chem. 280, 24435–24442

(2005).28. B. J. Peter et al., Science 303, 495–499 (2004).

29. R. N. Lightowlers, S. M. Howitt, L. Hatch, F. Gibson, G. B. Cox,Biochim. Biophys. Acta 894, 399–406 (1987).

30. B. D. Cain, R. D. Simoni, J. Biol. Chem. 264, 3292–3300(1989).

31. J. Symersky et al., Nat. Struct. Mol. Biol. 19, 485–491(2012).

32. D. Pogoryelov, O. Yildiz, J. D. Faraldo-Gómez, T. Meier, Nat.Struct. Mol. Biol. 16, 1068–1073 (2009).

33. L. A. Baker, J. L. Rubinstein, Methods Enzymol. 481, 371–388(2010).

34. A. Merk et al., Cell 165, 1698–1707 (2016).35. A. Bartesaghi, D. Matthies, S. Banerjee, A. Merk, S. Subramaniam,

Proc. Natl. Acad. Sci. U.S.A. 111, 11709–11714 (2014).

ACKNOWLEDGMENTS

Author contributions: J.L.R. conceived the project and experimentalapproach and supervised the research; S.A.B. made the yeaststrain and did some of the preliminary protein purificationexperiments; H.G. developed the final protein purification strategy,made cryo-EM specimens, collected cryo-EM data, calculated thecryo-EM maps, and built and refined the atomic models; andJ.L.R. and H.G. wrote the manuscript and prepared the figures. Wethank D. Mueller for advice on yeast growth and dissociation of theyeast ATP synthase, R. Henderson for telling us about the detergentGDN, B. Carragher and C. Potter for facilitating access to the TitanKrios, H. Wei for help with Titan Krios data collection, andJ.-P. Julien and V. Kanelis for critical reading of the manuscript.Supported by the Canada Research Chairs program (J.L.R.), aUniversity of Toronto Excellence Award (H.G.), and CanadianInstitutes of Health Research operating grant MOP 81294. Some ofthe work presented here was conducted at the National Resourcefor Automated Molecular Microscopy located at the New YorkStructural Biology Center, supported by National Institute of GeneralMedical Sciences grant GM103310 and Simons Foundation grant349247. Cryo-EM maps are deposited in the Electron MicroscopyData Bank under accession numbers EMD-7036 (FO dimer with C2symmetry applied), EMD-7037 (FO dimer with C1 symmetry applied),and EMD-7067 (F1FO dimer with C2 symmetry applied); atomicmodels are deposited in the Protein Data Bank under accessionnumbers 6B2Z (FO dimer model) and 6B8H (F1FO dimer model).

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/358/6365/936/suppl/DC1Materials and MethodsSupplementary TextFigs. S1 to S4Table S1References (36–54)

26 July 2017; accepted 11 October 2017Published online 26 October 201710.1126/science.aao4815

Guo et al., Science 358, 936–940 (2017) 17 November 2017 5 of 5

RESEARCH | REPORTon July 14, 2020

http://science.sciencemag.org/

Dow

nloaded from

region of mitochondrial ATP synthaseOAtomic model for the dimeric FHui Guo, Stephanie A. Bueler and John L. Rubinstein

originally published online October 26, 2017DOI: 10.1126/science.aao4815 (6365), 936-940.358Science 

, this issue p. 936Sciencecharacteristic cristae.

dimers bend the membrane to give mitochondria theirOproton translocation powers rotation and suggests how F, determined by electron microscopy. The structure gives insights into howSaccharomyces cerevisiae complex from O

present a high-resolution structure of the dimeric Fet al. region through a rotational mechanism. Guo 1synthesis in the F region that spans the mitochondrial inner membrane drives ATPO ATP synthase. Proton translocation across the FOF

1Synthesis of adenosine triphosphate (ATP) in mitochondria is accomplished by a large molecular machine, the FHow protons power rotation

ARTICLE TOOLS http://science.sciencemag.org/content/358/6365/936

MATERIALSSUPPLEMENTARY http://science.sciencemag.org/content/suppl/2017/10/25/science.aao4815.DC1

REFERENCES

http://science.sciencemag.org/content/358/6365/936#BIBLThis article cites 54 articles, 21 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceScience, 1200 New York Avenue NW, Washington, DC 20005. The title (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Copyright © 2017, American Association for the Advancement of Science

on July 14, 2020

http://science.sciencemag.org/

Dow

nloaded from