No. 7] Proc. Jpn. Acad., Ser. B 86 (2010) 667

Review ATP synthase: from single molecule to human bioenergetics

† By Yasuo KAGAWA*1,*2,

(Communicated by Fumio OOSAWA, M.J.A.)

+ Abstract: ATP synthase (FoF1) consists of an ATP-driven motor (F1)andaH-driven motor (Fo), which rotate in opposite directions. FoF1 reconstituted into a lipid membrane is capable + of ATP synthesis driven by H flux. As the basic structures of F1 (,3-3./C) and Fo (ab2c10) are ubiquitous, stable thermophilic FoF1 (TFoF1) has been used to elucidate molecular mechanisms, while human F1Fo (HF1Fo) has been used to study biomedical significance. Among F1s, only thermophilic F1 (TF1) can be analyzed simultaneously by reconstitution, crystallography, mutagenesis and nanotechnology for torque-driven ATP synthesis using elastic coupling mechanisms. In contrast to the single operon of TFoF1,HFoF1 is encoded by both nuclear DNA with introns and mitochondrial DNA. The regulatory mechanism, tissue specificity and physiopathology of HFoF1 were elucidated by proteomics, RNA interference, cytoplasts and transgenic mice. The ATP synthesized daily by HFoF1 is in the order of tens of kilograms, and is primarily controlled by the brain in response to fluctuations in activity.

Keywords: FoF1, molecular motor, mitochondria, omics, cytoplasts, bioenergetics

Introduction All human activity depends on ATP, which is *1 Department of Biochemistry, Jichi Medical School, primarily synthesized via mitochondrial oxidative Tochigi, Japan. phosphorylation (oxphos). By analogy with glyco- *2 Department of Medical Chemistry, Kagawa Nutrition lytic ATP synthesis, there have been many futile University, Saitama, Japan. † Correspondence should be addressed: Y. Kagawa, Depart- attempts to isolate hypothetical high-energy inter- ment of Medical Chemistry, Kagawa Nutrition University, 3-9-21 mediates, such as phosphoenolpyruvate, that tightly Chiyoda, Sakado, Saitama 350-0288, Japan (e-mail: kagawa@ couple respiratory energy to ATP synthesis. In 1961, eiyo.ac.jp). Mitchell proposed the chemiosmotic hypothesis, Abbreviations: AMPPNP, 5′-adenylyl-imindo diphos- phate; ANC, Adenine nucleotide carrier; ,, -, ., / and C, subunits which states that in oxphos, the respiratory energy of F1; -E, -D, and -T, - subunit of F1 with an empty catalytic site, is coupled to an imaginary anisotropic “ATPase with bound ADP or with bound ATP, respectively; a, b, c, d, e, f system” located in an ion-impermeable membrane and g, A6L, subunits of Fo;c10, decamer ring of the c subunit of Fo; + − fl F F , ATP synthase; F , ATP-driven motor of F F ;F,H+-driven (Fig. 1 of Ref. 1) via H /OH ux driven by the o 1 1 o 1 o + + 7 motor of FoF1;BF1,CF1,EF1,HF1,TF1 and YF1,F1 from bovine electrochemical activity ([H ]Left/[H ]Right =10 × mitochondria, chloroplasts, E. coli, human mitochondria, thermo- [ATP]/[ADP]) across the membrane.1) At the same philic bacillus PS3 and yeast, respectively; "A, membrane + time, soluble ATPase, known as coupling factor 1 potential; "7H , electrochemical potential difference of protons fi ’ across the membrane; F, Faraday constant; FRET, florescence (F1), was puri ed from mitochondria in Racker s 2) fi resonance energy transfer; IF, F1 inhibitor; mtDNA, mitochondrial laboratory. When F1 was bound to F1-de cient DNA; nDNA, nuclear DNA; OPA-1, optic atrophy 1; Oxphos, mitochondrial membrane, respiratory energy was oxidative phosphorylation; Pi, inorganic phosphate; PIC, Phos- coupled to ATP synthesis.2) The author then isolated phate carrier; P-loop, -G-X-X-X-X-G-K-T- sequence of the , and - the entire ATP synthase, later called F F (Fig. 1, subunits of F1; R, gas constant; VDAC, voltage-dependent anion o 1 channel. For nomenclature of the amino acid residues in subunits right), and reconstituted FoF1 into liposomes capable , - . “- ” , , , etc.,ofF1, K35 , for example, refers to a lysine (K) of converting energy from ATP hydrolysis to that - residue located at #35 in the subunit of F1. The equivalent for H+ flux driven by the electrochemical potential residues in different F s are expressed as thermophilic -K164 1 "7 + (=human -K162). Mutations are abbreviated as “-K164I”, which of protons across the membrane ( H ), where means “residue K164 in the - subunit is changed to I164”. "7H+ =F"A − 2.3RT"pH.3) The H+ flux through doi: 10.2183/pjab.86.667 ©2010 The Japan Academy 668 Y. KAGAWA [Vol. 86,

δ

β FoF1 F ADP+Pi 1 β α δ ATP γ ATP ε α ATP ADP β α β β ADP T D β γ +Pi α α γ ε b b β E a H+ b b open c c c

F1 top view a Fo H+ H+ c c C

ATP synthesis by torque of 42pNnm + + H -driven c-ring rotation elastically H α β coupled to 3 3 ring.

Fig. 1. Basic structure of TF1 (ATP-driven motor), TFo (proton-driven motor) and TFoF1 (proton-driven ATP synthase). Upper left: Side view of TF1 composed of ,, -, ., /, and C subunits. The .C turns counterclockwise against the ,3-3 hexamer in the ATP hydrolysis direction when viewed from the Fo side. Middle: Top view of TF1 with three different conformations of - subunits. Lower left: Side view of TFo composed of a, b and c subunits. The c10 ring turns clockwise against ab2 in the proton-driven direction. Right: Side view of TFoF1. The central stalk (.C) turns with the c10 rotor clockwise in the ATP synthesis direction when viewed from TFo side.

FoF1 liposomes was demonstrated first by a pH meter molecular mechanisms by which rotation and catal- (Fig. 6 of Ref. 3). After the primary chemiosmotic ysis are coupled were mainly elucidated by crucial 1) 3) hypothesis was thus partially established, the experiments using TF1 and TFo (ab2c10) to represent 7) 4) subunit-subunit interactions in FoF1 during ATP all FoF1s. The rotational hypothesis assumes that synthesis became the next research target. Boyer on ATP synthesis, the eccentric central stalk (.C) proposed a hypothetical two-subunit model of con- connected to the c10 ring is rotated against the ,3-3 formational energy transfer, and finally proposed the hexamer in the clockwise direction, as viewed from 4) rotational hypothesis in 1981, based on our report the Fo side (Fig. 1, right). This rotation induces on thermophilic F1 (TF1) subunits with ubiquitous cyclic conformational changes of - in the order of 5) stoichiometry ,3-3./C (Fig. 1, upper left) and -E→-D→-T→-E so as to change the affinity for conformational changes of isolated TF1- from -E nucleotides, and finally release ATP to return to -E 4),7) (with an empty catalytic site) to -D or -T by binding (Fig. 1, middle and upper right). In fact, Walker’s 6),7) ADP or ATP, respectively. X-ray crystallography of most of ,3-3. of BF1 The objectives of this review are to publish how visualized the distinct conformations of - with 9) achievements on thermophilic FoF1 (TFoF1) in Japan different nucleotide occupancies (-E, -D and -T). 1) + have advanced the primary chemiosmotic and The concept of FoF1 as an H -driven rotor (.- 4) rotational theories, and how our studies on human c10) and stator (/-ab2), rotating with a torque of 10) FoF1 (HFoF1) will contribute to the development of 42 pN nm, was predicted in 1996 (Fig. 1, right). human bioenergetics. To date, ligand-binding , or - The most convincing evidence of the rotational subunits have only been isolated and reconstituted hypothesis was the direct observation of the rotation 5)–7) into ,- subcomplexes from TF1. In contrast to of an actin filament attached to . against the fixed 7) 11) active ,3-3 hexamer and ,1-1 dimer of TF1, ,3-3 hexamer by Noji in 1997, using the TF1 12) attempts to reconstitute ,- subcomplexes from E. gene. Briefly, the ATP-driven F1 motor rotates + coli F1 (EF1) and HF1 in vitro has been unsuccess- clockwise (Fig. 1, upper left) and the H -driven Fo ful. However, functional complementation of yeast motor rotates in the anticlockwise direction as viewed 7),10),11) F1 (YF1) in a quintuple deletion mutant from the F1 side (Fig. 1, lower left). Prelimi- (","-"."/"C) with genes encoding ,, -, . and C nary experiments on “the 120° rotation of the c 13) of BF1 strongly suggests the presence of common subunit oligomer” have been reported, but this was 8) structure and function of HF1 subunits. Thus, the not sensitive to Fo inhibitor and may represent the . No. 7] ATP synthase: from single molecule to human bioenergetics 669

F : Stator: 1 α β , Catalytic 3 3© OSCP Portion : OSCP α β α3β3γ 3 3 F6, a, hexamer δ ε b, d, e, f,g,A6L

Central stalk, rotating Peripheral stalk

Rotor: ring of Fo: + γ δ ε c10 + H a,b,c,d, PIC e,f,g,F6 ANC

12),17)–21) Fig. 2. Side view of space filling model of HFoF1. As there is high homology among FoF1s from different species, and quintuple 8) yeast deletion YF1 mutant (","-"."/"C) is complemented by genes encoding BF1, the major structure of eukaryotic FoF1 is 9),14),25),29),73)–76) apparently universal. Thus, X-ray crystallography data for FoF1 subunits from different species were taken from RCSB Protein Data Bank (http://www.rcsb.org/pdb/results/results.do?outformat) and assembled according to sequence data for 21) HFoF1. No high-resolution structural data are available for subunit a and the hinge region of subunit b. ATP synthasome contains both PIC and ANC.31) PIC: Phosphate carrier; ANC: Adenine nucleotide carrier.

21),24),25) 22),26) rotation in the F1 portion of FoF1. In fact, crystallo- DNA and mitochondrial DNA. The 24),25) graphic analysis of yeast FoF1 (YFoF1) revealed that primary structures of bovine FoF1 (BFoF1) and 21),27),28) the ring of c subunits contains 10 protomers, rather HFoF1 were sequenced, and cDNAs of the - + than the widely anticipated 12 (4H per 120° subunits of BF1 and HF1, for example, were found to rotation), and tightly connected to ./C complex.14) share 99% amino acid homology and 94% nucleotide Although crystal of EFo is not available, exact experi- homology (authors report in 1986 quoted in Ref. 21). ments on TFoF1 confirmed the c10-ring structure Although the core structure of TFoF1 (,3-3./C + 15) 7),10),12) (Fig. 1, right). By using single-molecule FRET ab2c10) (Fig. 1) was conserved in H(B)FoF1,it measuring the change in the distance between the is more complex due to additional 8 subunits (d, e, subunits a and c during the rotation (see section 6.1), f, g, A6L (ATP8), OSCP and F6, plus a natural 21),24),25) a 36° sequential stepping mode of the c10-ring inhibitor called IF1) (Fig. 2) ; subunits / and C 16) rotation in FoF1 was confirmed. ATP is synthe- of TF1 correspond to OSCP (oligomycin sensitivity + sized at the clefts between , and - by "7H -driven conferring protein) and the /C complex in HF1, H+ flux when both motors are connected by central respectively.21) Although mammalian / subunit is (.C) and peripheral stalks (ab2/) to the c10-ring in partially homologous to C subunit of TF1, X-ray 7),14),17) 29) 14) TF1 (Fig. 1). Elucidation of primary struc- crystallographic data of both BF1 and YF1 18) 12),19) tures of E. coli FoF1 (EFoF1) and TFoF1, and showed that /C complex interacts with a Rossmann 18),20) site-directed mutagenesis and suppression of the fold in the . subunit, forming a foot and c10-.-/-C lost functions18) led to refinements in the rotational rotates as an ensemble central stalk (Fig. 2),14),29) 10) hypothesis including elastic power transmission. similar to the C subunit at the . subunit in TF1 After the primary rotational hypothesis4) was (Fig. 1).7),17) In fact, the /C complex was dissociated 7),10),11),17) thus partially established, research on FoF1 from BF1 by guanidine treatment as a stable was divided into single molecular elucidation of heterodimer (Papageorgiou, S., and Solaini, G., energy transduction at the atomic level using stable 2004, PubMed abstract). So called minor subunits 12),17),20) TFoF1 and biomedical elucidation of human (e, f, g and A6L) of HFo are unlikely to have a role 21) 22) FoF1 (HFoF1) at the mitochondrial and in vivo directly in ATP synthesis, but they appear to 23) 21),25) levels. In contrast to the single operon of bacterial influence oligomeric state of HFoF1. Supra- 12),18),19) 30) FoF1, animal FoF1 is encoded by both nuclear molecular structures of HFoF1 include HFoF1 dimer 670 Y. KAGAWA [Vol. 86,

Nucleotide- Oligomer binding α,ββ α β δ γ 3 3 F F ε o 1 δ α β α β α Dissociation ∆µ∆µH+ α β α Association bb γ ε Dialysis a Fo b b c c c

Solubilization a c c c

TFoF1 Phospholipids + Cholate liposome

3),39) Fig. 3. Reconstitution of FoF1 from subunits and FoF1 liposomes from phospholipids. Isolated nucleotide-binding , and - 6),56) subunits and ,3-3 hexamer were obtained by dissociation of TF1 (upper left). Solubilized Fo or FoF1 was mixed with phospholipids and cholate, and dialyzed to reconstitute FoF1 liposomes (bottom right, electron micrograph) capable of proton-driven ATP synthesis. and ATP synthasome31) composed of phosphate Fleischer’s group extracted phospholipids from the carrier (PIC), adenine nucleotide carrier (ANC) and mitochondrial membrane with aqueous-acetone (10% HFoF1 (Fig. 2). Moreover, there are tissue differences water) and restored electron transport activity by 34) in HFoF1, with muscle- and liver-type . subunits in adding back phospholipid micelles in 1962, 27) 21),28) HF1. The complex gene structure, specific Kakiuchi succeeded in a similar experiment in regulation systems,21),28) expression and alternative 1926.35) Okunuki36) succeeded in preparing several 27),32) splicing of HFoF1 and related regulatory genes cytochromes. Green’s group prepared electron trans- were elucidated by recent cytoplast technology,22),33) port components from mitochondria.37) However, transgenic mice,32) transcriptomics21),32) and proteo- even after phospholipids were added back and mics.31),32) Based on the knowledge from these electron transport activity was restored,34)–37) lip- extensive studies,7),17),21) human energetics in phys- osomes capable of "7H+-driven oxphos, as predicted iological activities21) and diseases21),23) have been by Mitchell,1) were not reconstituted. analyzed. Membrane proteins were classified into extrinsic 3) In this article, mechanistic studies of FoF1, and intrinsic proteins. Extrinsic proteins, such as 36) 2),5) including a reconstitution study and crystallography, cytochrome c and F1, or components of F1, 3),25) 3),25) will be reviewed, and then, more complex mitochon- including OSCP and F6 (Figs. 1 and 2), are drial cytobiology and human biomedical studies will easily detached from the biomembrane by treatment be described, because mitochondrial structure, neuro- with ultrasonic irradiation, chelating agents, and hormonal control, tissue-specific activity and disease chaotropic anions (KI, KCN),3) and can be purified as are not found in bacteria. Historical evaluations of soluble proteins in the water phase by chromatog- contributions made by scientists in mechanistic raphy and ammonium sulfate fractionation. How- studies are summarized in excellent reviews (Refs. 4 ever, intrinsic proteins including cytochrome oxi- 36),37) 3) and 17, and references therein). dase and Fo are hydrophobic and embedded in the lipid bilayer, and require proper detergents for 1. Isolation of ATP synthase (FoF1)by solubilization3) (Fig. 3, right). In 1966, detailed phase membrane biology diagrams of phospholipid cholate system were 38) FoF1 is a membrane protein of oxidative reported, and these were useful for solubilizing phosphorylation. In the 1960s, membrane biology intrinsic proteins and reconstituting functional bio- was in its infancy, and the many attempts to membrane.3),39),40) The detergent concentration purify ATP-synthesizing membrane proteins from needed to solubilize FoF1 is near its critical micelle mitochondria had been unsuccessful. Long before concentration.3) No. 7] ATP synthase: from single molecule to human bioenergetics 671

2) + The F1 reported by Racker’s laboratory was These liposomes showed H translocation on not sensitive to an energy transfer inhibitor of addition of ATP,3),40),46) ATP-32Pi exchange reac- oxphos, oligomycin, but became oligomycin sensitive tions3),39),46) or H+-driven ATP synthesis (Fig. 1, 46) when bound to F1-depleted mitochondrial mem- right). Thus, FoF1-liposome was shown to be the brane.41) The oligomycin sensitivity conferring factor anisotropic “ATPase system” imagined by Mitchell.1) + (Fo,H -driven motor, Fig. 1) was characterized as an ATP synthesis was sensitive to the combination of intrinsic membrane protein,41) and by using cholate, valinomycin (K+ ionophore) and nigericin (H+–K+ + 3),39),46) oligomycin-sensitive ATPase, later designated FoF1, exchange ionophore), which collapsed "7H . 3),42) was isolated (Fig. 3). FoF1 was reconstituted More stable TFoF1-liposomes synthesized ATP 42) from Fo and F1. The electron microscopic images of from ADP and Pi with energy from proton flux 3 + BFoF1 revealed spherical H-acetyl-BF1 attached to driven by "7H formed by "pH and "A across their 43) 46) membrane-embedded BFo. In combined morpho- membranes. Using chloroplasts, ATP synthesis logical-biochemical studies to identify the in situ from ADP and Pi was demonstrated by applying 43) 47) structure of BFoF1, radiolabeled BF1 was added to "pH using acid–base transition and by imposing 48) BFo membrane, and then the structure, radioactivity "A using an electric pulse. However, chloroplasts and ATPase activity were removed in parallel.43) contain an electron transport system and other However, mitochondrial F1s, including BF1 and components that are energized by either "pH or HF1, are unstable and reconstitution of F1 from each "A. Thus, net ATP synthesis by applying "pH isolated subunit was unsuccessful, even with chaper- (acid–base transition) or "A (external electric pulse) 7),44) 7),45) 46) ones. Thus, TF1 and TFoF1 were purified to purified TFoF1 reconstituted in liposome was the 1) from thermophilic bacillus PS3, and stable TFoF1 most convincing evidence of chemiosmotic theory. 7),44),45) was reconstituted from TF1 and TFo (Fig. 1). The electrical potential between the inside and Recent proteomics using mild detergent extraction of outside of liposomes formed by K+ diffusion in the HFoF1 from membrane followed by native gel electro- presence of valinomycin was calculated as "A =RT/ + + phoresis revealed supramolecular structures includ- F × ln([K ]out/[K ]in). Maximal net ATP synthesis 30) ing dimeric HFoF1 and ATP synthasome (PIC + from ADP and Pi was achieved by incubating vesicles ANC)31) (Fig. 2) (see section 9). in malonate at pH 5.5 with valinomycin, and then rapidly transferring them to a solution of pH 8.4 and 2. Reconstitution of FoF1 membrane capable of 150 mM K+.7),46) To synthesize ATP, the minimal fl ATP synthesis by proton ux "7H+ (= "A − 60"pH, at 30°C) of 210 mV and + + + 46) + FoF1 converts energy of "7H driven H optimal "7H of 290 mV was required. The H - flux into ATP synthesis. Incorporation of conducting activity of TFo-liposome through TFo membrane proteins into the lipid bilayer is essential was proportional to the imposed "7H+ (6H+/sec/ in activity studies in membrane biology. The most 103 mV at pH 8.0).49) The pH profile of the rate difficult step was the reconstitution of liposomes revealed that a proton, not a hydroxyl ion, was the containing active membrane proteins capable of true substrate.49) Because of the kinetic equivalency producing "7H+.3),39),40) The use of removable between "* and "pH as driving forces of ATP detergents, including cholate, was essential in the synthesis,46) "A is expected to replace "7H+. In fact, 3),39),40) preparation of BFoF1 (Fig. 3), and after TFoF1-liposomes irradiated with external electric cholate extraction, Triton-X100 was used in the pulses (760 V/cm, 30 ms, rectangular) catalyzed net 45) 50) chromatography to purify TFoF1. The removal of ATP synthesis. The amount of ATP synthesized 14C-cholate during dialysis was estimated by radio- increased with the number, voltage and duration of activity,3),40) and tight closure of the liposome electric pulses.50) The net synthesis of ATP by + membrane was estimated by radioactivity of enclosed application of "7H across the TFoF1-lipo- 14C-inulin and confirmed by electron microscopy of somes46),50) firmly established the chemiosmotic enclosed ferritin.3),40) Electron microscopy showed hypothesis.1) To directly measure "* with electrodes liposomal membranes studded with closely neighbor- on both sides of the membrane, TFoF1 was recon- 3),43) ing F1 (Fig. 3, lower right). As several extrinsic stituted into planar phospholipid bilayers, and the proteins in oxphos, including F1, OSCP and F6 were magnitude of the electric current generated upon partially lost during the extraction and dialysis of addition of ATP was shown to follow simple FoF1, the best activity was attained by adding these Michaelis–Menten type kinetics, and the Km was after dialysis.3),39) found to be 0.14 mM.51) There was no apparent 672 Y. KAGAWA [Vol. 86, dependence of Km on "*.51) This observation portion, diameter 12 nm× height 10 nm)43) connected indicates that "7H+ does not directly affect Km to by two stalks (central and peripheral) to a basal piece 7),51) 3),43) release the ATP formed on TFoF1, and opens the (subunits a and c10 ring) (Fig. 3, lower right). way for conformational energy transfer in ATP 3.2. Protomeric, oligomeric and rotational synthesis. ATPases: ,1-1 dimer, ,3-3 hexamer and ,3-3. 3),39),46) Since the success of the FoF1-liposome, heptamer. The isolated , and - subunits of TF1 the reconstitution method has been used to analyze both have AT(D)P-Mg binding activity accompanied the activity of intrinsic proteins including channels, by conformational changes6) without ATPase activity 55),56) receptors and membrane proteins. Using the lip- (Fig. 3, left). The open structure of -E and closed osome method, the important outer membrane structures of -D and -T in the presence of ligands were diffusion channel known as VDAC, (voltage-depend- confirmed in the isolated thermophilic - using 1H- ent anion channel) in mitochondrial transport was NMR.56) Both thermophilic , and - subunits were 52) isolated in 1980 (see sections 9 and 10). reconstituted to form an active ,1-1 dimer by forming 55) a catalytic ,- interface. Three ,1-1 dimers were 3. Reconstitution of F1 subunit complexes reconstituted to form an allosterically active ,3-3 capable of ATP synthesis by torque hexamer (Fig. 3, upper middle).55) Both the high 3.1. Core components of Fo are a, b, and c catalytic activity and formation of F1-bound ATP subunits and those of F1 are ,, -, ., / and C. from ADP + Pi depend on the ,3-3. structure with 7),11),17) The structure of FoF1 is shown in Figs. 1 rotational ATPase. There are six potential 7),9),12),18),53),54) and 2. Complete reconstitution of F1 nucleotide-binding sites on F1 and ,3-3.: three (370 kDa), after complete denaturation of all sub- catalytic sites on - and three noncatalytic sites on units into their primary structures with sodium ,,7),17) as confirmed by X-ray crystallography.9) dodecylsulfate and urea and refolding into tertiary Depending on the occupancy of catalytic sites structure,53) from subunits , (55 kDa), - (52 kDa), . with increasing ATP concentration ([ATP]), there (32 kDa), / (20 kDa) and C (14 kDa) was possible only are three types of ATPase activities of F1: uni-site, 53) 4),57) in TF1 (Fig. 3). The intermediate core subunit bi-site and tri-site. Uni-site catalysis is measured 55) 55) complexes, ,1-1 dimer and ,3-3 hexamer, were at sub-stoichiometric ATP concentrations ([ATP] < also only obtained in TF1, although the sequence [F1]). Uni-site activity is very low, and the apparent 57) homologies of core subunits were conserved in KmATP is less than 20 nM. The ATPase activity of 21) 57) HF1 (371 kDa). Enzymology revealed that ,3-3 the ,1-1 dimer of TF1 showed typical Michaelis– hexamer was an oligomer, while ,1-1 dimer was a Menten kinetics with only one KmATP value of 70 µM, 55) protomer. Owing to the stability of TF1, the and a Vmax value of 0.1 unit/mg, without the subcomplexes with chemically modified and mutated cooperative characteristics of a protomer.57) In 7),17),54) subunits were useful for nanotechnology. contrast, ATPase activity of the ,3-3 hexamer TFo (148 kDa) is composed of a (30 kDa), b showed the cooperative characteristics of an oligo- 58) (17 kDa) and c (8 kDa) in a stoichiometric ratio of mer. The apparent KmATPs of oligomeric ATPase 7),12),17) 1:2:10 (Figs. 1 and 3). HFo and BFo contain, of ,3-3 at 25°C were about 150 µM (bi-site) and 58) in addition to the common subunits a (also called 490 µM (tri-site). KmATPs of rotational ATPase of ATP6; 25 kDa), b (25 kDa), and c (8 kDa), minor TF1 (Fig. 3, right) were about 80 µM (bi-site) and subunits d (19 kDa), e (8 kDa), f (10 kDa), g (11 490 µM (tri-site).58) The .-containing complexes kDa), F6 (9 kDa) and A6L (also called ATP8; ,3-3., ,3-3./ and ,3-3.C, show common kinetic 18),25),29) 59) 8 kDa) (Fig. 2). The single stalk BFo obtained properties. The ,3-3 hexamer was inhibited by 3 by urea-cholate treatment of BFoF1 was reconsti- only one mole of [ H]-3′-O-(4-benzoyl) benzoyl-ADP 3 60) tuted with externally added H-acetyl-F1 and the per hexamer, similarly to both BF1 and TF1. Thus, additional peripheral stalk components OSCP and F6 the presence of only one inhibited-- in the hexamer 3),39),40) were demonstrated in 1971. Isolated BFo blocked multi-site steady-state ATPase activity. subunits other than OSCP and F6 are unstable, and This single-hit inactivation and cooperativity is an their reconstitution has been unsuccessful without inherent property of the symmetrical ,3-3, but is presequence and organizing machinery, even with not the due to the inhibition of rotation by TF1 or 21) chaperones. These BFo subunits were only identi- ,3-3.. fied on gel electrophoresis and X-ray crystallogra- 3.3. Rotation of the . subunit in ,3-3 25),29) phy. FoF1 was seen as a sphere (,3-3 hexamer hexamer of TF1: One mole ATP hydrolysis at No. 7] ATP synthase: from single molecule to human bioenergetics 673

one - subunit drives 120° rotation of . subunit in other two - subunits to drive rotation of the -E, -D, 63) a concerted manner. The rotational hypothesis of and -T cycle. Hybrid F1 containing one or two 4) FoF1 was also proposed by Oosawa as the “loose mutations with altered catalytic kinetics rotates in an coupling mechanism of rotational proton ATPase” asymmetric stepwise fashion with different dwell based on analogy with the H+-driven flagella motor times. Analysis of the rotation revealed that for any in 1986.61) The rotation of the . subunit axis in the given - subunit, the subunit binds ATP at 0°, cleaves cylinder of the ,3-3 hexamer in the FoF1 motor with ATP at approximately 200° and carries out a third the torque of 42 pN nm was predicted by many lines catalytic event at approximately 320°. This demon- 10) of evidence. The rotation of . was directly strates the concerted nature of the F1 complex demonstrated in single-molecule studies using ,3-3. activity, where all three - subunits participate to 11) from TF1. Rotational motion was visualized by drive each 120° rotation of the . subunit with a 120° attaching a fluorescently labeled actin filament (1– phase difference.63) 4 µm) to .S107C of artificially induced mutant . 3.4. Torque-driven ATP synthesis by TF1. subunit with the biotin-streptavidine bridge. The ATP is synthesized by mechanical energy applied ,3-3 hexamer was immobilized on a glass surface of on the . subunit without proton flux. ATP Ni-nitrilotriacetate by artificially attaching decapoly- synthesis driven by mechanical energy (Fig. 1) was histidine to -,11) and the ATP-driven rotation of directly shown by attaching a magnetic bead the . subunit was found to be anticlockwise when (diameter = 700 nm, biotinylated) to the . subunit F1 was observed from the Fo side (Fig. 1, upper of ,3-3. of mutant TF1 (C193,,H10--, and S107C., left).11),17) I210C.) on a glass surface, and rotating the bead The work performed by the rotating . in a fixed using electrical magnets.64) After the ATP-driven ,3-3 is the frictional torque times angle of rotation. rotation of the beads was confirmed, the magnet was The hydrodynamic frictional drag coefficient (9)of turned on and several bursts of hundreds of the actin filament for the propeller rotation is given revolutions at 10 Hz were imposed.64) Anticlockwise by 9 =(:/3)2L3/[ln(L/2r) − 0.447], where, 2 (10−3 forced rotation of the . subunit by the magnetic Nsm−2) is the viscosity of the medium, L, the length beads resulted in the appearance of ATP in the of the actin filament (1–4 µm) and r, the radius of the medium, as detected by counting the photons filament (5 nm).11) The observed rates of filament emitted from the luciferase–luciferin reaction with a rotation at 2 mM ATP are 7, 1, and 0.1 revolutions camera (Hamamatsu Photonics).64) This shows that per second, when the lengths of f-actin are 1, 2 and torque working at one particular point (.)ona 4 µm, respectively.11) The frictional torque 9B was protein complex can influence a chemical reaction about 40 pN nm, where B is the angular velocity.11) occurring at physically remote catalytic sites (-), Hydrolysis of one ATP molecule drives a 120° driving the reaction far from equilibrium.64) rotation of the . subunit relative to the cylinder of 4. Genes for TFoF1 and HFoF1: Single operon vs. the ,3-3 hexamer, and therefore, hydrolysis of three ATP molecules is required for one complete 360° nuclear and mitochondrial genes revolution.15),62) In order to analyze rapid rotation Detailed genetic analysis and site-directed muta- by reducing the viscosity resistance of long actin genesis have been reported by Futai using E. coli FoF1 18),65) filament, fluorescent gold beads (40 nm) were at- (EFoF1), as E. coli genetics are well understood. tached to .. At nanomolar ATP concentrations, -E The catalytic, structural and regulatory significance waits until the next ATP molecule is bound, and the of an amino acid residue in EF1 was elucidated by duration of the pause depends on the ATP concen- site-directed mutagenesis.65) However, many crucial 62) 51) tration (ATP-waiting dwell time). ATP binding to experiments, including the planar FoF1 bilayer 64) -E is the power step that drives the 80° rotation of and torque-driven ATP synthesis, have not been 55) .. This rotation leads to simultaneous release of successful to date with EFoF1, due to its fragility. ADP from the catalytic site of -D, and hydrolytic Thus, a special sequencing method for thermophilic 66) cleavage of ATP into ADP and Pi at -T after a pause genes was developed. The structure of the TFoF1 (catalytic dwell time), and a 40° rotation occurs to operon (number of amino acid residues),12),19) I(127)- complete the 120° rotation.62) a(210)- c(72)- b(163)- /(163)- ,(502)- .(286)- -(473)- 18),65) Using mutant - (E190D) of TF1, in the same C(132), was similar to that of the EFoF1 operon. rotational experiments, the catalytic activity of each Amino acid residues in the different ,, -, and . 12),19) 21),27),28) 24) - subunit was shown to be coordinated with the subunits from TF1, HF1, BF1 and 674 Y. KAGAWA [Vol. 86,

12),19) 71) Fig. 4. Aligned amino acid sequences and secondary structure elements of , and - subunits in TF1. Solid black lines indicate folds, and these were classified into ,-helices (A–H, 1–8) and --sheets (a–f, 0–8). The labels for folds are provided only for the - subunit, except for the three C-terminal ,-helices in the , subunit. Dots indicate every tenth residue. I–XI: areas of ,- contact. Red: catalytic contact areas of -. Pink: catalytic contact areas of ,. Blue: non-catalytic contact areas of -. Green: non-catalytic contact areas of ,. Colored bars indicate contact residues in T-E. Sequences are divided by red asterisks (*) to indicate the three domains.

18),65) 19) EF1 are aligned and expressed in the format residues (green and blue letters in Fig. 4) among TF1, ,10, which refers to residue #10 in the , subunit. HF1 and EF1 are closely related to catalytic and The residue numbers of amino acid sequences in the , regulatory functions.19),21),65) The observed substitu- and - subunits of TF1 are shown in Fig. 4 (dots tions in the thermophilic subunit increased its indicate every tenth residue).17),19) Primary struc- propensities to form secondary structures, and its tures are homologous, with 59% sequence identity external polarity to form tertiary structure.19) between thermophilic ,/human , and 68% between Chemical and genetic modification of residues 19) thermophilic -/human -. The primary structure in F1 revealed a nucleotide-binding P-loop of the TF1 - subunit showed homology with 270 (-GGAGVGKT-; thermophilic -158–165 corresponds residues which are identical in the - subunits from to bovine -156–163) (Figs. 4, 5A).7),9),12),19),54) Long 19) HF1,CF1, and EF1. The homologies of the amino before the X-ray crystallographic elucidation of the 9) 20) acid sequence between BF1 and YF1 were 73%, 79% P-loop, site-directed mutagenesis of the TF1 gene and 40%, respectively, for the ,, - and . subunits.14) to induce thermophilic -K164I and thermophilic As these YF1 subunits were functionally comple- ,K175I, identified an essential role for lysine residues 8) 20) mented with corresponding BF1 subunits, the in the catalysis (Fig. 5A, red letters). The proton essential structure is conserved among YF1,BF1 abstracting thermophilic -E190 (Fig. 5A) localized 7),19),54) and HF1 (sequence is nearly identical to that of in the GER-loop (-VGER-) (Fig. 4) was also 8) BF1, but there were polymorphisms in HF1). predicted by TF1 mutagenesis producing thermo- 67) Residues forming reverse turns (Gly and Pro) were philic -E190Q. These mutant TF1 subunits pro- 19) highly conserved among the - subunits. Conserved duced an ,3-3. complex that was suitable for No. 7] ATP synthase: from single molecule to human bioenergetics 675

A B

Crystal of TF1

C

α β Crystals of 3 3

D

Fig. 5. Crystals of TF1 and ,3-3, and X-ray crystallography data for the catalytic center of F1. A. Catalytic center of the - subunit of 9) 12) BF1 (black text indicates residue number) and TF1 (red text indicates residue numbers). Except for ,R373 and ,S344, all of the amino acid residues are present in -. Residues 159–164 are part of the P-loop surrounding the triphosphate residue of ATP. B. 70) 71) 71) Crystals of TF1. C. Crystals of ,3-3. D. Top view of the crystallographic structure of ,3-3.

64) 69) experiments on torque-driven ATP synthesis. ture of ,3-3. Three-dimensional crystals of TF1 Species-specific residues (black letters in Fig. 4) (Fig. 5B) and ,3-3 hexamer (Fig. 5C) were obtained may have phylogenic components, including thermo- using dye-ligand chromatography columns.70),71) The philic loops (-ARNENEV-) (Fig. 4, first line I′) that high resolution power of the Photon Factory 19) render TF1 stable. Since determining the nucleo- synchrotron (for TF1 ,3-3, 0.32 nm resolution at 7),12),19) 9) tide sequence of TFoF1, numerous rotating Tsukuba) revealed the detailed structure of ,3-3., 71) 14),72) ATP synthases of thermophilic F-type or V-type ,3-3 (Fig. 5D), the c-ring, the peripheral (vacuolar ATPase) have been sequence and charac- stalk,25) central stalk29) and the stator (Fig. 2).73) terized.65),68) X-ray crystallography (0.325 nm resolution) of the In contrast to the single operon TFoF1, the gene ,3-3.C complex of EF1 was also recently reported 21),26),28) structure of HFoF1 is highly complex ; most (Ducan, T.M., personal communication, 2010). The subunits are encoded by nuclear DNA, with signal peripheral stalk consists of a continuous curved ,- peptides to target this protein to the mitochondrial helix about 16 nm in length in the single b-subunit, 21),27),28) inner membrane, but subunits a and A6L of augmented by the predominantly ,-helical d and F6 21),26) 25) HFo are encoded by mitochondrial DNA. The (Fig. 2, right). complete sequence of the 16,569-base pair human 5.2. Crystallography of Fo. The c subunits mitochondrial DNA contains genes for 12S and 16S form a ring around a central pore.14),72) The numbers rRNAs, 22 tRNAs, ATPase subunits 6 (correspond- of the c subunit in the Fo ring differ depending on the 72) 14) ing to the a subunit of HFo) and 8 (corresponds to species: in CFo,itis14, while that for YFo and 26) 15) A6L of HFo), and 11 other protein coding genes. TFo is 10. The conserved carboxylates E61 of CFo (corresponds to E56 of TFo) involved in proton 5. Crystallographic analysis of FoF1: Detailed – + transport, are 1.06 1.08 nm apart in the c-ring rotor, structure of H -driven and ATP-driven motors which rotates relative to the membrane anchored a 5.1. Crystallization and analysis at Photon subunit. The torque-generating unit consists of the Factory. The most detailed structural information interface between the rotating c-ring and the flanking for amino acid residues in a protein is obtained by stator a subunit.14),72) The ring rotor is driven by the crystallographic analysis. In 1977, a two-dimensional sequential protonation and reprotonation of E61. crystal of TF1 showed the pseudo-hexagonal struc- Residues adjacent to the conserved E61 residues 676 Y. KAGAWA [Vol. 86, show increased hydrophobicity and reduced hydro- action between the adenine ring, and Y341, F414 and gen bonding.72) Upon deprotonation, the conforma- F420 (Fig. 5A).9),71) tion of E61 is changed to another c subunit and 5.4. Basic structure of ,3-3 is rendered becomes fully exposed to the periphery of the ring.72) asymmetric by addition of . and/or nucleotides. Reprotonation of E61 by a conserved R in the The basic structure of F1 is a symmetrical ,3-3 adjacent a subunit returns the E61 to its initial hexamer that is composed of three pairs of alternat- 72) 71) conformation. Genetically modified TFoF1s, each ing ,E and -E (Fig. 5D). However, the asymmetry containing a c subunit dimer (c2) to a dodecamer induced by introduction of . and/or AT(D)P to 15) (c12), were prepared by genetical cross-linking. ,3-3 hexamer is critical in the mechanism of ATP 9) Among these, TFoF1s containing c2,c5,orc10 showed synthesis. The nucleotide-free -E in both the F1 ATP-synthesis and other activities, but those con- crystal9) and solution56) has an open structure taining c9,c11 or c12 did not. Thus, the c-ring of (Fig. 6) that is essentially identical to -E in the 15) 71) functional TFoF1 is a decamer (c10). When TF1 was nucleotide-free ,3-3 hexamer. Both -T and -D,as removed from the modified TFoF1s, TFos containing well as ,T, assume closed structures (Fig. 6, direction 15) 9) only c2,c5 or c10 worked as proton channels. In fact, of the open arrow). Interconversion of the open- a 36° step size of proton-driven c10-ring in FoF1 was close conformational states of - is achieved by 16) 14) 6),56) confirmed. The c10 ring in YFo and functional addition of AT(D)P to the isolated -E of TF1. complementation of the YF1-deleted yeast mutant However, nucleotide-free YF1 contained -D and -T with BF1 genes strongly suggests the presence of the structures similar to those of nucleotide-bound 8) 75) c10 ring in HFoF1. BF1. This suggests that -. interactions at the 5.3. Crystallography of catalytic site of F1. three contact points (Fig. 6, middle, .1–3), including The basic ground state of F1 without nucleotides was interaction of Arg residue at position 75 (.R75) with shown on crystallography of the thermophilic ,3-3 -DE395 in the DELSEED sequence of BF1 to change 55) 55) 10),14) hexamer, in which three ,1-1 dimers were the mutual conformation, are as important as arranged in three-fold symmetry, 12 nm across and nucleotide occupancy in converting open -E into the 71) 75) 10 nm high (Fig. 5D). The first detailed crystallog- closed -. As genes for ,, - and . of YF1 in the raphy of nucleotide-bound ,3-3. (partial) of BF1 was ,---. deleted mutant yeast are complemented with 8) determined at 0.28-nm resolution by Walker’s those of BF1, the functional residues are essentially 9) group. In the structure of BF1 crystallized in the equal between YF1 and B(H)F1. Occupancy of the presence of ligand (AMPPNP:ADP:Pi = 50:1:0, and catalytic site by ATP or ADP can be mimicked by 2+ Mg ), the three catalytic - subunits differed in convenient BeF3-ADP complexes that bind to the 76) conformation and in bound nucleotide. There were catalytic sites of -T and -D. The structure is four unhydrolyzable ATP analogue (AMPPNP) representative of an intermediate in the reaction molecules, three in equal three , subunits, one in - pathways.76) The conformational change of - induced (-T) and one ADP in - (-D); the remaining - was by .-rotation is essential for ATP synthesis (ATP 9) empty (-E). release from the catalytic site), while that induced by The ATP-binding site of -T is surrounded by the ATP-binding to - is necessary to elicit torque on 74),75) residues shown in Fig. 5A (black numbers for BF1 are .. 9),71) identical to those of HF1, red numbers for TF1). 5.5. Three domains in , and -: - barrel, 20) In the P-loop of -T of TF1, the essential K164 nucleotide-binding and ,-helical bundle do- forms hydrogen bonds with the phosphate of the mains. The overall molecular structure of , and - nucleotide, and the oxygen of T165 coordinates with can be divided into three domains9),71),75):anN- Mg2+, while in the GER-loop, E190 interacts with terminal - barrel (Fig. 6, top N to *), a central water (Fig. 5A).9),71) The catalytic activity of - nucleotide-binding domain (Fig. 6, middle * to **), requires , that supplies thermophilic ,-R365 and and a C-terminal ,-helical bundle (Fig. 6 bottom ** thermophilic ,-S336 to the ATP-binding site at the to C).7) The locations of amino acid residues in the ,- interface (Fig. 5A).9),71),74) The positive charge of ,-helices (Fig. 4, solid black lines A–H, 1–8) and ,-R365 stabilizes the --phosphate of AT(D)P, and --sheets (Fig. 4, solid black lines a–f, 0–8)71),74) are R256, R191 and K164 of thermophilic - interact with compared in the three-dimensional structure of the negative charges of the phosphates of AT(D)P thermophilic - (Fig. 6). As the amino acid sequence 24) (Fig. 5A). The cross-linking of thermophilic -Y341 of BF1 is nearly identical to that of HF1 (99% 7),54) 19),21) with azido-ATP predicted a hydrophobic inter- homology in -), the following discussion on BF1 No. 7] ATP synthase: from single molecule to human bioenergetics 677

Fig. 6. Three-dimensional structures of T-E. Three-fold axis is vertical, so that views are towards the ,- subunit interfaces. Left: non- catalytic interface (blue I–IX indicates contact areas). Right: catalytic interface (red I–XI indicates contact areas). Green letters: domain names, with domain borders being marked by red asterisks. D and R in green circles are T-D331 and T-R333, respectively, at the entrance to the crevice of the P-loop. I in red circle is T-I386 of -- contact. H and G in yellow circles indicate hinge residues in T- (179–181).74)

is also applicable to HF1. Superposition of the overall homologous residue pairs found in both TF1 and BF1 crystallographic structures of thermophilic -E and were defined as homological contact areas. The bovine -E or bovine , revealed that the folding of contact areas found only in one species, such as the 9),71) these structures is very similar. The - barrel thermophilic loop of TF1 (I′ in Fig. 4) were defined as domains (Fig. 4, T, 21–94, T-1–82) contain six - species specific contact areas. These areas are ex- strands (Fig. 6. plate form arrows, -a–-f).74) The pressed as primary structure in Fig. 4. The contact nucleotide-binding domains (T ,95–371, T-83–354) areas are located in both - and , at catalytic (red consist of nine-stranded --sheets surrounded by eight and pink bars in Fig. 4), and non-catalytic (blue and ,-helices (Fig. 6, ,Ato,H) and a small antiparallel green bars in Fig. 4) interfaces. There are seven --sheet. The C-terminal ,-helical bundle domain of , catalytic (red I–III, V–VIII in Figs. 4 and 6) and non- (thermophilic ,375–502) contains six helices (Fig. 4, catalytic (blue I–VII in Figs. 4 and 6) contact areas ,1, 2, 4, 6–8), while that of - (theromophilic -355– on the open - form (-E). The number of contact areas 473) consists of six ,-helices (Fig. 4, ,1–6). Thus, the on closed - (-D and -T) increased to 11 (red I–XI in largest difference between , and - is found in the C- Fig. 6) and 9 (blue I–IX in Fig. 6), respectively, in terminal region,9),74) and the DELSD(E)ED sequence the catalytic and non-catalytic interfaces. The barrel localized in this region was shown to be the most domain harbors the universal contact areas I and important -. contact area.10) II (Fig. 6, upper), and the common electrostatic 5.6. Catalytic and non-catalytic ,- interfa- bond in II is thermophilic -R72–thermophilic ,E67 ces and conformational change. The crystal (=human -R71–human ,E67).9),71) At the catalytic 9) 71) structure of BF1 and ,3-3 of TF1 indicates that, nucleotide-binding domain, areas III, V, VI, VII and in general, the conserved residues lie on the ,- VIII are universally detected (Fig. 4). However, in interfaces, as shown by detailed homology search T-E, the P-loop contact area IV is latent, in contrast 19) among TF1,HF1,CF1 and EF1. There are two to that area in thermophilic ,E (Fig. 6). Human types of ,- interface that contain either a catalytic ,R373 interacts with oxygen in the -- and .-P of site or non-catalytic site.9),74),75) The area where pairs ATP bound at IV.9) In V, the common electrostatic of residues connecting the ,- interface are located is bonds are thermophilic -R193–,D339. In VI and VII defined as contact area. Contact residue pairs within of -T and -D, human ,F299–-M222 and human a limit of 0.40 nm across the ,- interfaces in BF1 ,S344–-R260, respectively, interact. However, we and the ,3-3 hexamer of TF1 were analyzed by a identified no direct contact in the ,-helical bundle 74) computerized atom search using the CCP4 Suite: domain in -E of TF1. In the catalytic ,- interface 74) 2 Program Contact. The contact areas composed of of human -E, the contact areas (17.6 nm ) are 678 Y. KAGAWA [Vol. 86,

homologous to those of thermophilic -, while the dimeric HFoF1 and prevention of futile ATP loss 2 + areas in human -D (30.3 nm ) and human -T when "7H is decreased, as described in section 10. (22.0 nm2) are increased to 11 and 10, respectively. This is caused by the 30° upward motion of the C- 6. Nanotechnological analysis of TFoF1 by single- terminal domains. molecule imaging: Dynamic movement of FoF1 5.7. Catalytic sites. The catalytic ,- interface X-ray crystallography of F1 and FoF1 is a static is located on the left side of - in the ,3-3 hexamer snapshot of inhibited ATPase crystallized in the 9) 76) (Fig. 5D), and the structure of catalytic domain presence of AMPPNP or BeF3, or in the absence (Fig. 5A) is strictly conserved among spe- of nucleotides.71),75) These crystals do not represent 7),9),71),74) cies. The catalytic domain accommodates the transient movement of subunits of TF1 during .- 3),39),46) the P-loop located between sheet 3 and helix B rotation, or activity of TFoF1 in a liposome. 79) (T-163–178) and the conserved thermophilic -E190 Thus, the dynamics of rotating TF1 or ATP 80) (=human -E188) in the GER loop localized between synthesis in FoF1-liposomes must be measured by 9) 63) --sheet 4 and N-terminal end of ,-helix C (Fig. 4, using TF1 containing mutant - subunits, and also GER, and Fig. 6, middle). As predicted by X-ray using modalities such as fluorometry79),80) or NMR.56) 9) crystallography of AMPPNP-BF1, NMR analysis The efficiency of florescence resonance energy revealed that thermophilic -R191 (Fig. 5A, upper transfer (FRET) between a donor and an acceptor left) forms a hydrogen bond with the .-phosphate of fluorophore depends on their distance.16),79),80) If two ATP.56) Pi is shown to bind the catalytic domain of fluorophores are bound to appropriate amino acid 62),75) -E, which is identical to the sulfate binding site residues in rotor and stator subunits, relative subunit 71) of -E in ,3-3 by -K164 and ,R365. The structure movements can be observed in real time by confocal 16),79),80) of the active metal-ATP complex in TF1 at the microscopy. catalytic domain was shown to be ", -, .-bidentate 6.1. Nanomotor movement analysis by sin- 77) 18 18 Mg-ATP. (Rp)-[-.- O, .- O]ATP.S was hydro- gle-molecule FRET. In a single TF1 molecule fixed 17 11),63) lyzed by TF1 in H2 O, and the resulting inorganic on a glass surface, a donor fluorophore (Cy3) [16O, 17O, 18O] thiophosphate was shown to have an was bound to one of the three -s and an acceptor Rp configuration.77) The reaction thus proceeds with fluorophore (Cy5) was bound to the protruding inversion of configuration at the phosphorous, and a portion of ., and single pair (Cys3–Cys5) FRET direct in-line nucleophilic attack of the 17O in water was performed to estimate the waiting conformation on the .-phosphate of ATP via the pentavalent during ATP hydrolysis.79) As Cy3- and Cy5-male- intermediate state.77) The ordered water molecules imide are bound to cysteine residues, site-directed that carry out nucleophilic attack on the .-phosphate mutagenesis [,(C193S), -(S205C) and .(S107C)] of ATP during hydrolysis are 0.26 nm from the was performed to bind Cy5 to -, and to bind Cy3 nucleotide analogue, beryllium, in the -D and 0.38 nm to ., and to prevent binding of Cy3 and Cy5 to , away in -T, strongly indicating that -D is the (residue numbers of , and - are indicated in catalytically active conformation.76) Fig. 4).79) The sole cysteine in a mutant subcomplex Adjacent to the P-loop, there is a hinge point, of TF1, ,(C193S)3-(His-10 tag at N terminus) 3. -HGG (179–181), between ,-helix B and --sheet 4 (S107C), was labeled with Cy5-maleimide. The (Cy5- (Figs. 4, hinge, and Fig. 6, right). The conforma- .)TF1 was incubated with Cy3--(S205C) at 1:10 at tional change in the hinge should be transmitted to 45°C for 2 days, and the free - subunit was removed the DELSDED sequence in ,-helix I. The hinge on a size exclusion column.79) The energy of the laser motion of the C-terminal domain containing the beam (532 nm) on Cy3 was transferred to Cy5 and DELSDED sequence in thermophilic -E, rotated emitted light (670 nm) when the Cy3–Cy5 distance away from the core axis from 110° (close) to 144° was small, while only Cy3 light (570 nm) was emitted (open)56) In the reverse reaction, the resulting when the Cy3–Cy5 distance was great. FRET yield widening of the P-loop–thermophilic -E190 distance changed cyclically as . rotated and the Cy3–Cy5 causes the release of Mg-ATP from the catalytic site. distances were estimated during the conformation 79) Crystallographic analysis of the F1–IF1 complex change. The distance between the two dyes (IF1: natural inhibitory peptide) at 0.28-nm resolu- changed continuously as 5.7, 7.9, and 7.9 nm during tion revealed that IF1 binds in the ,D–-D interface rotation at low ATP concentrations, and the 78) and opens the catalytic site. Inhibitor studies on F1 conformational change corresponded to the ATP- 79) are thus important in understanding the formation of waiting state of TF1. No. 7] ATP synthase: from single molecule to human bioenergetics 679

The relative subunit movement during ATP present in organisms that maintain "7H+ mainly in 14),15) synthesis has also been measured by FRET between the form of "A, whereas Fos with c14 ring are two fluorophores bound to a stator subunit (b- mostly found in species with "7H+ existing predom- subunit) and a rotor subunit (.-orC- or c-subunit) inantly in the form of "pH rather than "A.72) H+/ 62),80) (Fig. 1). The labeled FoF1 was reconstituted in ATP ratios of 4.7 and 3.3 are thus expected for CFoF1 + 72) 14)–16) the liposome (one FoF1 per liposome) and "7H was (with 14c/3-) and E(T)FoF1 (with 10c/3-), + applied, so that FoF1 carried out H -driven ATP respectively. To confirm the effects of c/- subunit synthesis.46) Analysis of the time course of FRET ratios on H+/ATP ratio, pH of the internal phase of efficiency in the FoF1-liposome showed the rotation of the reconstituted FoF1-liposomes was equilibrated .- and C-subunit relative to b-subunit in 120° steps,80) with acidic medium.46),47) Then, an acid-base tran- 16) 46) and that of the c10-ring in 36° steps. The - motions sition was induced by adding alkaline medium to through an attached fluorophore, concomitantly with the liposomes to produce "pH across the membrane, the 80° and 40° substep rotations of . in the same and the initial rate of ATP synthesis was measured single molecules, showed the sequence of conforma- with luciferase.82) From the shift in the equilibrium tions that each - undergoes in three-step bending, an "pH as a function of Q (= [ATP]/([ADP][Pi]), the approximately 40° counterclockwise turn followed by standard Gibbs free energy for phosphorylation, 0′ + 82) two approximately 20° clockwise turns, occurring in "Gp , and the H /ATP ratio were determined. 80) 0′ synchronization with two substep rotations of .. The results were as follows: "Gp =38±3kJ/mol + + The results indicate that most previous crystal and H /ATP = 4.0 ± 0.2 for CFoF1; and H / structures mimic the conformational set of three -s ATP = 4.0 ± 0.3 for EFoF1. This indicates that the in the catalytic dwells,80) while the previously thermodynamic H+/ATP ratio is the same and that 82) described set of -E, -D and -T was revealed in the it differs from the subunit c/- stoichiometric ratio. ATP-waiting dwells.80) These fluorescent studies thus However, in order to estimate actual energetics, the bridge the gap between the chemical and mechanical very low turnover rates (<1 ATP/s) in this experi- 82) steps in FoF1. Starting from the ATP-waiting dwell ment need to be examined under different physio- (0°), the 80° and 40° substeps of . rotation are logical conditions (>100 ATP/s). induced by ATP binding and ADP release, and ATP The site of analog–digital conversion by elastic- hydrolysis and Pi release, respectively.62),79)–81) ity7) was estimated by direct measurement of the + 83) 6.2. H /ATP ratio and elastic power trans- torsional stiffness. Most parts of F1, particularly mission in FoF1. One of the unsolved questions in the central . shaft in F1, and the long eccentric the mechanistic study is the analog–digital conver- bearing had high stiffness (torsional stiffness 5 > 83) sion of energy in FoF1. The electrochemical energy of 750 pN nm). One domain of the rotor, namely, + 1) 3) the H current through FoF1 in liposomes and where the globular portions of . and C contact the planar bilayers,51) and the electric,48),50) magnetic64) c-ring, was more compliant (5 congruent with and mechanical energy of the rotation11),62) are all 68 pN nm).83) The .-induced or nucleotide-dependent analog quantities. The numbers of ATP molecules open–close conversion of conformation is an inherent synthesized and protons transported are digital property of an isolated -, and energy and signals are quantities. Thus, the elastic power transmission in transferred through ,- interfaces.7),11),74) Rotation of FoF1 analog/digital conversion during the .-rotation the central shaft . in ,3-3 hexamer is assumed to be was predicted in 1996.10) Oosawa also proposed a driven by domain motions of the -s. These - motions loose coupling mechanism in which the number of were directly observed through an attached fluoro- protons necessary for the synthesis of one ATP is not phore by FRET79),80) and NMR.56) an integer but varies depending on the environmental The mechanisms underlying the open(-E)– 61) conditions. As one proton is translocated by one close(-T) motion were investigated for - subunit of 15),16) 56) shift of c10 in the c-ring of TFoF1, and one ATP TF1 in solution, using mutagenesis and NMR. The is synthesized per - subunit of TF1, the inevitable hydrogen bond networks involving side chains of consequence is noninteger ratios of rotation step sizes K164 (162 for human - is shown in parentheses), + 14),15),74) for TFoF1 (120°/36°) and for H /ATP (10:3). T165(163), R191(189), D252(256), D311(315) and This step-mismatch necessitates elastic twisting of R333(337) in the catalytic region (Fig. 5A, red text 74) TFoF1 during rotation and elementary events in for TF1 and black text for HF1) are significantly 7),74) + 56) catalysis. The H /ATP ratio in FoF1 addresses different for the ligand-bound and free - subunits. this analog/digital problem. Foswithc10 ring are The role of each amino acid residue was examined by 680 Y. KAGAWA [Vol. 86,

A

B

C

Fig. 7. Alternate splicing of human F1. pre-mRNA. A. Exons in the F1 . pre-mRNA are expressed in boxes. Exon number 8 (hatched 27) box) is included in liver and excluded in the muscle. B. Schematic representation of wild-type and mutant F1. exons 8–9 (Ex8–9). Heavy underline in Ex9-wild-type indicates ESE element. Mutated nucleotides are indicated by outlined letters. In Ex9-MU2 and Ex9-MU, boxed letters indicate that these sequences are predicted to act as ESE elements. C. Selection of F1. exon 9 is regulated by two cis-acting regulatory elements in the same exon.90) Purine-rich ESE promotes exon 9 inclusion, which is repressed by MS-ESS under muscle-specific conditions; PPT: Polypyrimidine tract.

A (alanine) substitution.56) The chemical shift ess,21) starting from transcription,84)–86) splicing of perturbation of backbone amide signals of the nuclear encoded subunits (Fig. 7),27),32) and trans- segmentally labeled -(mutant)s indicated stepwise lation of mitochondrial DNA-encoded subunits propagation of the open/close conversion on ligand (Fig. 8)22),26),87) and nuclear DNA-encoded precursor 56) 21),86) binding. Upon ATP binding, the open/close peptides for all F1 subunits (,, -, ., / and C) and 21) conformation change regulated by hydrogen-bond 7Fo subunits (b, c, d, e, f, g and OSCP), followed switching from K164/D252 to T165/D252 (Fig. 5A, by targeting, importing and processing of nuclear right upper, red test) would take place in the DNA-encoded precursor peptides in the mitochond- thermophilic - subunit because ATP-binding is the rion (Fig. 9).21),86) Precursor importing requires both major driving force for the first 80° rotation.62) The "7H+ to drive translocation and specific carrier resulting closing motion of the hinge (-HGG-) proteins in the outer and inner mitochondrial between ,-helix B and --sheet 4 (Fig. 6, right) membrane, as well as general chaperones present in generates the torque of . rotation through the the cytosol and mitochondrial matrix.86) Processing DELSD(E)ED contact point (Fig. 6, bottom).10),56) of the presequence of precursor requires a specific Although the time scale of atomic fluctuations is in protease,86) and after the removal of the presequence, the order of tens of nanoseconds, molecular simu- nuclear-encoded subunits are assembled into HFoF1 lation will solve the detailed movements of residues in with two mtDNA-encoded Fo subunits (a and 87) ,3-3 during the . rotation (order of milliseconds) in A6L). Targeting the presequence of three isoforms the future. of subunit c liberated after proteolysis is required for the assembly of cytochrome oxidase.88) 7. Biogenesis of human FoF1: Regulated As both the amino acid and nucleotide sequences expression, splicing, import and assembly of subunits in HFoF1 are available in internet The biogenesis of HFoF1 is an intricate proc- databases, only physiologically important points No. 7] ATP synthase: from single molecule to human bioenergetics 681

Fig. 8. Homoplasmy, heteroplasmy, cytoplasts and ;o cells.87) Small red circles indicate wild-type mtDNA and small blue circles indicate mutant mtDNA; large red circles indicate nDNA, and green ovals indicate cells. EB: ethidium bromide used to remove mtDNA; PEG: polyethylene glycol used to fuse cells or cytoplasts.

Regulator : neuro-endocrine PGC-1α NRF etc. Tfam

Genome: mtDNA nDNA

Transcription Fo (a+A6L) ,α β γ δ ε b c d e f g OSCP Transcriptome: mRNAs (a+A6L) Splicing & F1 mRNAs Fo mRNAs Translation Proteome: Fo (a+A6L) Precursors of F1, Fo subunits Fuel Gauge Processing & Assembly Import and Proteolysis AMP kinase F1 + ∆ µH AMP Prepeptides for o organization ATP ADP+Pi

21),86) Fig. 9. Regulation of HF1Fo biosynthesis : Transcription, splicing, translation, targeting, import, processing and assembly. PGC-1,: Peroxisome proliferator-activated receptor-. coactivator-1,; Tfam: mitochondrial transcription factor A; NRF: nuclear respiratory factor; Circle: mitochondrial DNA.

21),84) in the biogenesis of HFoF1 will be reviewed contains 12 exons interrupted by 11 introns. 21),28),86),89) here. In contrast to prokaryotic FoF1, Primer extension and S1 mapping analysis showed 7) including TFoF1, proteomics have revealed a large the presence of multiple transcription initiation sites 21),84) 84) universe of pseudogene products, splice variants in the HF1, gene. The 5′-flanking region of the 27),32),90) (muscle-type F1.) (Fig. 7), post-translational HF1, gene has an unconserved GC-rich region, 91) modifications (phosphorylated HFoF1, etc.), di- including several binding motifs of transcriptional 30) meric HFoF1, a supramolecular complex called factors, such as Sp1, AP-2 and GCF. The basal 31) ATP synthasome (Fig. 2), and ectopic HFoF1 promoter activity was located near the GC-rich (Fig. 10, right).92) region. Comparison of the 5′-upstream region of the 85) 7.1. Expression of HFoF1 genes. The gene HF1, gene with those of the genes for BF1,,HF1- 27) structure of the HF1, subunit is 14 kbp in length and and HF1. indicated three common sequences (CS1, 682 Y. KAGAWA [Vol. 86,

Brain

Mt ATP-sensitive K channel VDAC Rs impulse + ∆ µ F1 H Muscle IF AMPK o METs increase ADP+Pi ADP+Pi ATP F Fo 1 Ectopic FoF1

2+ Regulation of glycolysis Ca + extra cellular K ATP and ∆pH

Purinergic receptor + depolarization K Ca2+ ATP-sensitive K+ channel L-type Ca2+ channel

Fig. 10. Regulation of HF1Fo activity: Neuronal impulse from brain to human activity driven by ATP by an intricate regulation system. AMPK, AMP kinase; ecto-F1Fo, ectopic or cell surface F1Fo; IF, ATPase inhibitor; METs, metabolic equivalents of exercise intensity; Rs, receptors for neuro-endocrine transmitters; VDAC, voltage-dependent anion channel.

CS2 and CS3) in the regulatory regions,21) suggesting pyruvate dehydrogenase ,-subunit.21),86) The char- that putative cis-elements coordinate the expression acteristics of this enhancing element were examined 84)–86) of the three subunit genes for HF1. The by introducing a synthetic oligonucleotide element enhancer activities derived from 5′-deletion mutants into the CAT plasmid with a deleted enhancing of a HF1,-CAT (chloramphenicol acetyltransferase) element. The resulting plasmid showed full recovery chimeric gene were different in cell lines from four of promoter activity, and this activity was independ- different human tissues, thus suggesting the existence ent of the orientation or location of the insert. of cell type-specific gene regulation.21) Therefore, this enhancer may be common to the The HF1- gene is 14 kbp in length and contains nuclear genes of some mitochondrial proteins in- 10 exons, with the first exon corresponding to the volved in energy transduction.21) non-coding region and most of the presequence, The functions of subunits in mitochondrial FoF1 which targets this protein to the mitochondria.85),86) were confirmed by expression of genes in deletion Eight Alu repeating sequences including inverted mutants, and their functional complementation.8) repeats were found in the 5′-upstream region and The genes encoding BF1 subunits (except for those introns. An S1 nuclease protection experiment of /) were expressed in a quintuple yeast YF1 deletion revealed two initiation sites for transcription. Three mutant (","-"."/"C) after introduction of a CCAAT boxes were found between the two initiation chimeric BF1 subunit gene construct that uses the sites, and two GC boxes were located in the 5′- YF1 transcriptional promoter and termination sites, 8) upstream region. Promoter activity was estimated by as well as the presequence for YF1. Expression of the CAT method and an enhancing structure for the ,-, --, .-, or C-subunit of BF1 complemented transcription was detected between nucleotides -400 the corresponding individual mutations in YF1. All 86) and -1100 in the upstream region. BF1 subunits (with chimeric /) expressed in yeast The system coordinating expression of nuclear- produced an F1 that was purified to a specific activity 8) coded mitochondrial proteins was investigated by of about half of that of original BF1. These results examining the 5′-flanking region of the HF1- gene. In indicate that the molecular machinery required for one of the enhancing regions, a consensus sequence the targeting, proteolysis and assembly of the was found for the genes of other mitochondrial mitochondrial FoF1 is conserved from yeast to 8) proteins, such as those for cytochrome c1 and the humans (Fig. 9). Similarly, bovine OSCP and some No. 7] ATP synthase: from single molecule to human bioenergetics 683

Fo components have been functionally complemented Mutation analyses on the HF1. Ex8-9 minigene using (quoted in Ref. 8). In one functional study, the gene a supplementation assay demonstrated that the for . subunit (atp-3) was partially blocked by RNA muscle-specific negative regulatory element is posi- interference (RNAi) in Caenorhabditis elegans, and tioned in the middle region of exon 9, immediately ATP levels decreased from 15 nmol/mg protein downstream from ESE. Detailed mutation analyses (control) down to 4 nmol/mg, and the behavior identified a muscle-specific exonic splicing silencer extended the lifespan.93) However, tissue differentia- (MS-ESS) (5′-AGUUCCA-3′) responsible for exclu- tion is absent in yeast, alternate splicing must be sion of exon 9 in vivo and in vitro (Fig. 7, C).90) This studied in human cells.27),32) element was shown to cause exon 9 skipping of in vivo 90) 7.2. Tissue-specific splicing of HF1.: splicing systems. Although there are three variants Analysis using transgenic mice and minigenes. of the c subunit88) and several alternate splice Mammalian FoF1 is characterized by tissue-specific variants in the human mitochondrial fusogenic 87) expression of the F1 gene, which was analyzed using proteins (mitofusin 1, 2), the . subunit of HF1 is 32) 86),90) transgenic mice and minigenes in cultured human the only well-characterized variant in FoF1. 90) cells. The muscle-specific isoform of HF1. was generated by alternative splicing, and exon 9 was 8. Mitochondrial cytology of HFoF1: cytoplasts ;o found to be lacking in skeletal muscle and heart tissue lack nDNA and cells lack mtDNA 27) 32),90) (Fig. 7, A). Using transgenic mice, the alter- HFoF1 is encoded by both mitochondrial DNA native splicing of exon 9 was shown to require de (mtDNA)22),26),33) and nuclear DNA (nDNA) novo protein synthesis of a cis-acting element on (Fig. 9).21),86) In order to analyze the roles played o the spliced exon of HF1. gene. An HF1. wild-type by mtDNA and nDNA, mtDNA-less cells (; minigene, containing the full-length gene from exons cells)22),33) and nDNA-less cells (cytoplasts) were 8 to 10, and two mutants were prepared; one mutant developed (Fig. 8).87),94) Using ethidium bromide, involved a pyrimidine-rich substitution on exon 9, mtDNA was removed and the resulting ;o cells whereas the other was a purine-rich substitution became strictly dependent on glycolysis to compen- 33),87) (abbreviated as HF1. Pu-del and HF1. Pu-rich sate for the oxphos that supplies ATP. Thus, a mutants, respectively).90) Pu-del inhibited exon glucose medium is essential for ;o cells.22) Cytoplasts inclusion, indicating that a Pu-del mutation disrupts are enucleated cells that contain mitochondria an exonic splicing enhancer. On the other hand, Pu- (Fig. 8, upper left), with examples being enucleated rich blocked muscle-specific exon exclusion.90) oocytes,87) synaptosomes94) and platelets.94) Since Transgenic mice bearing both mutant minigenes DNA sequence of an individual differs from each were then analyzed for their splicing patterns in other owing to the genetic polymorphism of 90) 21) tissues. Based on an analysis of HF1. Pu-del mtDNA, personal collection of mtDNA is essential minigene transgenic mice, the purine nucleotide in to elucidate mitochondrial diseases. Human mito- this element was shown to be necessary for exon chondria with intact mtDNA have been directly inclusion in non-muscle tissue. In contrast, analysis of isolated from postmortem platelets.94) Expression of o HF1. Pu-rich minigene mice revealed that the HF1. nDNA-encoded HFoF1 subunits was not affected in ; 87) Pu-rich mutant exon had been excluded from heart cells, while that of mtDNA-encoded HFo subunits and skeletal muscles in these transgenic mice, despite (Foa and A6L) was lost in cytoplasts, as nDNA- the fact that mutation of the exon inhibited muscle- encoded mitochondrial transcription factor A specific exon exclusion in myotubes at early embry- (Tfam)95) was lacking (Fig. 9).87) The ;0 cells have onic stages.32) These results suggest that the splicing no respiratory chain, because of the loss of mtDNA- regulatory mechanism underlying HF1. pre-mRNA encoded subunits of cytochromes and NADH dehy- differs between myotubes and myofibers during drogenase.26),87) Despite the absence of oxphos, ;o myogenesis and cardiogenesis.32) cells require mitochondrial compartments with a A detailed mutational analysis of exon 9 (Fig. 7, sufficient "7H+ for energy driven transport of matrix B) revealed a purine-rich exonic splicing enhancer components.87),96) The essential "7H+ of ;o cells is (ESE) element (5′-AAUGAAAA-3′) functioning maintained by the electrogenic exchange of ATP4− ubiquitously, with the exception of muscle tissue. for ADP3− by ANC.96) To energize the inner 55) An exonic negative regulatory element responsible membrane, ,3-3 (ATPase active) in the matrix for muscle-specific exclusion of exon 9 was discovered of ;o cells regenerates ADP from translocated using both in vitro and in vivo splicing systems.32),90) ATP.96) 684 Y. KAGAWA [Vol. 86,

The term heteroplasmy refers to cells that the fibroblast lines from a patient with cardiomyop- contain a mixture of mtDNAs with different athy, and their nuclear hybrid clones were isolated.22) sequences (Fig. 8, upper right), whereas homoplasmy A normal fibroblast line from the fetus and a means that 100% of their mtDNA has an identical respiration-deficient fibroblast line from the patient sequence (Fig. 8, upper left).22),87) Cybrids were were used as positive and negative controls, respec- formed by cytoplast fusion with ;o cells using tively.22) By this method, many mitochondrial 87),94) 87) polyethylene glycol (PEG) (Fig. 8, bottom). diseases devoid of HFo have been elucidated and The majority of pathogenic mtDNA mutations are mitochondrial gene therapy for heteroplasmic pa- heteroplasmic, with mutated and wild-type mtDNA tients was developed using cytoplasts from a normal coexisting in the same cell (Fig. 8, upper right).86),87) fetal cells and the cybrids (Fig. 8).97) Owing to the absence of protecting histones, mtDNA The most frequent mutation in mtDNA-encoded is highly susceptible to mutations that result in HFo gene is NARP (neuropathy, ataxia, and retinitis heteroplasmy. Mutations in the tRNA gene of pigmentosa), caused by a mutation at L156 in the a 99) mtDNA often block translation and cause complete subunit of HFo (Fig. 2). A mutation conferring a deletion of mtDNA-encoded proteins (syn− muta- milder phenotype (L156P) leads to a 30% reduction 87) + tion), including HFo subunits. in H flux, and a similar loss in ATP synthesis. The During development, cell division unevenly dis- more severe mutation (L156R) also leads to a 30% tributes heteroplasmic mtDNA into daughter cells reduction in H+ flux, but ATP synthesis is abol- and eventually segregates homoplasmic cells with ished.99) With the L156P mutation, rotation of the c- wild-type and syn− mutant mtDNA (Fig. 8, bottom ring may be slowed, but coupling of ATP synthesis to right). This stochastic segregation of the syn− H+ flux is maintained (Fig. 1, lower left, subunit a), mutation results in the syn− mutant concentrated whereas with the L156R mutation, H+ flux is tissues and causes mitochondrial diseases, including uncoupled, because the transmembrane helix III of 87) − mitochondrial diabetes. The homoplasmic syn Foa is unable to span the membrane. The L156R 99) cells lack oxphos and depend on glycolysis. Under mutant has ATPase activity, because the ,3-3 the influence of polymorphisms in mtDNA and complex portion of F1 is intact, and increased proton nDNA, a vicious circle of reactive oxygen species will permeability through the defective Fo cannot main- damage cells. However, mitochondrial transfer from tain "7H+ to inhibit ATP-driven H+ translocation wild-type homoplasmic cytoplasts by fusion to form (i.e., uncoupling).3),46) cybrids will normalize the diseased cells97) and syn− 8.1. Coordination of nuclear and mitochon- mutation will be alleviated by mitochondrial fu- drial DNA. The tissue activity of FoF1 mainly relies sion.87),97) We analyzed heteroplasmy and polymor- on mitochondrial biogenesis encoded by both nDNA phisms related to diabetes and its complementation and mtDNA (Fig. 9). A number of transcriptional by mitofusins.87) Mitochondria in human cells are modulators have been implicated in the regulation of visualized as a network or as filaments that undergo mitochondrial biogenesis and oxphos activity.84)–89) continuous changes in shape and in localization within To understand the nDNA–mtDNA interactions in the cells. Mitochondrial fusion proteins including human cells, we identified the nuclear transcription mitofusins87) and OPA198) that may work as natural factors that are common to the expression of these PEG in Fig. 8,87) and regulate both mitochondrial gene products. As Tfam encoded by nDNA is fusion and metabolism. We characterized splice essential for both the initiation of transcription and variants of human mitofusins (also called hfzo 1, 2 the replication of mtDNA, we cloned and sequenced and 3)87) as well as OPA1s (Fig. 10, upper right).98) the human Tfam gene.95) There were sequences in the We analyzed complementation by fusogenic proteins, 5′-upstream regulatory region of Tfam common to − 95) and the lost tRNA in mitochondria of syn mtDNA those in HF1-. In the absence of mtDNA-coded Fo inside heteroplasmic cells was complemented by wild- subunits, expression of other FoF1 subunits was not type tRNA in normal mitochondria by fusing wild- affected, but most nDNA-coded subunits other than − 87) 87),96) type and syn mitochondria with mitofusins. The , and - of FoF1 could not be assembled. For in mitofusin genes were expressed mainly in post-mitotic vivo analysis of this regulation, transmitochondrial brain and muscle, thus complementing mutated mice carrying various proportions of deletion mutant mtDNA that is not removed during cell division.87) mtDNA ("mtDNA) were generated by introducing In order to analyze a mitochondrial disease, pure "mtDNA from cultured cells into the fertilized eggs nuclear transfer was carried out from ;0 HeLa cells to of mice.100) The great advantage of transmitochon- No. 7] ATP synthase: from single molecule to human bioenergetics 685 drial mice is that they share exactly the same nDNA In situ mitochondrial "7H+ was directly esti- background and their genetic variations are re- mated by rhodamine 123, which is accumulated in stricted to the proportions of pathogenic mtDNA.100) mitochondria depending on "7H+.103) The futile Transcription factors in the expression of HFoF1 ATP hydrolytic activity of HFoF1 during ischemia + 78) include PPAR. coactivator 1, (PGC-1,), in coop- that lowers "7H , is prevented by IF1. Bovine IF1 eration with several factors, such as peroxisome is an ,-helical dimer and residues 1–37 of IF1 open 78) proliferator-activated receptor (PPAR), nuclear res- the catalytic interface between ,D–-D. Atomic piratory factors 1 and 2 (NRF-1 and NRF-2), or the force microscopy images show how these FoF1 specificity protein 1 (Sp1), a ubiquitous transcription molecules form dimers with a characteristic 15-nm factor known to regulate the constitutive expression distance between the axes of their rotors through of oxphos genes (Fig. 10, upper).86) PGC-1, is a stereo-specific interactions of the membrane-em- master modulator of gene expression in human bedded portions of their stators.104) Adifferent tissues and enhances the activity of PPAR, in interaction surface is responsible for the formation skeletal muscle.101) Mitochondrial transcription is of rows of oligomers, suggesting the role of subunits e 104) directed by a small number of nucleus-encoded and g of HFo in dimerization. Some dimers have a factors, including Tfam.95) Expression of these factors different morphology, with a 10-nm stalk-to-stalk is coordinated with that of nuclear respiratory distance, in line with FoF1s, which are thus accessible 104) proteins through the action of PGC-1 family to IF1. Dimeric or polymeric HFoF1 is related coactivators.101) to morphology of cristae104) under the influence of Using these cytological methods (Fig. 8), tran- OPA1.98) scriptomics and proteomics (Fig. 9), human bioen- A channel protein, porin, is now known as ergetics at both cell level (Fig. 10) and in vivo levels VDAC (voltage-dependent anion channel) and is the were elucidated. most abundant protein in the mitochondrial outer membrane.52) VDAC helps ATP/ADP exchange by 9. Supramolecular structure of HFoF1: Dimeric forming a complex with ANC (ANC–VDAC com- HFoF1, ATP synthasome and ecto-HFoF1 plex).103) Finally, a supramolecular structure called Mitochondrial FoF1, including HFoF1, is typi- ATP synthasome composed of FoF1, ANC, PIC and cally isolated as a monomeric complex that contains perhaps VDAC was isolated and characterized 16 protein subunits21) and the natural inhibitor (Fig. 2).31) Parallel immuno-electron microscopic 78) protein (IF1) (Fig. 2). However, mitochondrial studies revealed the presence of PIC and ANC FoF1 was isolated in dimeric and higher oligomeric located non-centrally in the basepiece, and other states using digitonin for one step mild solubilization studies indicated an ATP synthase/PIC/ANC stoi- followed by blue native (BN) or clear native (CN) chiometry near 1:1:1 (Fig. 2).31) Collectively, these electrophoresis.102) Recent developments in proteo- findings support a mechanism in which the entry of mics have revealed HFoF1 in its natural supra- the substrates ADP and Pi into mitochondria, the 30),31),102) molecular state. Single bands in the gel can synthesis of ATP on HFoF1, and the release and exit be analyzed by proteomics approaches including of ATP are localized in a supramolecular structure in immunoprecipitation,103) and mass spectrometry to a highly coordinated system. identify the amino acid sequence of the compo- 9.1. Ectopic HFoF1: plasma membrane local- 103) nents. Electron microscopy of these oligomeric ization in lipid rafts. HFoF1 is located not only on mitochondrial FoF1 particles was reported in 1972 the mitochondrial inner membrane, but also on the (Fig. 2C in Ref. 3). Dimeric and trimeric FoF1 were cell surface. Extracellular ATP synthesized by the purified from mammalian mitochondria in five differ- ectopic HFoF1 is not an energy source but a regulator ent tissues by BN electrophoresis and CN electro- for various cellular responses that are initiated by phoresis,30),102) and these were active, thus suggesting purinergic receptors (P2X and P2Y) and signaling that oligomeric FoF1 is constitutive in mitochondria. processes and are terminated by breakdown of ATP Using BN electrophoresis, two membrane proteins by ectonucleotidases (Fig. 10, right). By using 3H- 3 (6.8 kDa proteolipid and diabetes-associated protein) ADP, net H-ATP synthesis by cell surface HFoF1 that had previously been removed during purification was confirmed (to rule out ATP + AMP synthesis by were shown to be stoichiometrically associated with adenylate kinase). ATP synthesis was inhibited by 30) FoF1, and this may provide insight for further membrane-impermeable HFoF1-specific inhibitors 30) functional investigations. (angiostatin and piceatannol) and anti-HF1 anti- 686 Y. KAGAWA [Vol. 86, body.105) Immunoprecipitation indicated that ectopic K+channel dependent resting potential of plasma HFoF1 and a surface protein of endothelial cells, membrane is inside negative. These ectopic FoF1 caveolin-1, are physically associated.105) Adipocyte activities have never been reported in prokary- ectopic HFoF1 may contain Fo, as it is inhibited by otic FoF1, although FoF1-like V-type ATPase is oligomycin and influenced by a proton conductor widely distributed in prokaryotic and mammalian 65),68) (uncoupler) (quoted in Ref. 92). HFoF1 is selectively membrane structures to transport ions. In localized in lipid rafts with other mitochondrial mammalian tissues, some proteins involved in energy proteins. Lipid rafts are detergent-resistant mem- metabolism may exert entirely different functions; brane microdomains enriched in cholesterol and cytochrome c, for example, is an electron carrier caveolin-1. Intracellular traffic may translocate but also serves as the central signal protein in HFoF1 containing ,, -, ., b, d, F6 and OSCP from apoptosis. mitochondria to lipid rafts.92) 10. In vivo ATP synthesis: FoF1 in human The ectopic HFoF1 has been implicated in numerous activities, including the mediation of bioenergetics and diseases intracellular pH, cellular response to antiangiogenic The function and survival of all organisms is agents, and cholesterol homeostasis as a receptor for dependent on the dynamic control of energy metab- 106) apolipoprotein A-1. HFoF1 is expressed on the olism. Energy demand is matched to ATP supply by 7),21) surface of endothelial cells, where it binds angiosta- FoF1 and glycolysis (Fig. 10, bottom). The tin, regulates surface ATP levels, and modulates increase in ADP + Pi produced by ATP consump- endothelial cell proliferation and differentiation via tion results in the instant "7H+-driven ATP syn- 106) purinergic receptor (Fig. 10, right). thesis by FoF1 and increases in electron transport 107) + Ectopic HFoF1 is closely related to obesity. activity to compensate "7H (respiratory control). Expression of the , subunit of ectopic HFoF1 is The increase in ADP is amplified as AMP increase by markedly increased during adipocyte differentiation. myokinase reaction (2ADP = ATP + AMP). If ATP Treatment of differentiated adipocytes with inhib- synthesis by FoF1 is not enough, especially when itors of HFoF1 or antibodies against , and - subunits oxygen supply is limited by high metabolic equiv- of HFoF1 leads to a decrease in cytosolic lipid alents (METs >5), increased AMP/ATP ratio accumulation.107) Apolipoprotein A-I binds to the - activates phosphofructokinase to compensate ATP subunit of ectopic HFoF1 and its inhibition decreases by glycolysis. Although the regulatory mechanism of 107) 7),108) the production of lipid droplets. Depletion of TFoF1 is basically ubiquitous, HFoF1 is speci- plasma membrane cholesterol with methyl-beta- alized for human activity. One of the characteristics cyclodextrin disrupts lipid rafts and abolishes co- of HFoF1 among other FoF1s is genetic poly- localization of HFoF1 with caveolin-1, which results morphisms during evolution, particularly in in a marked reduction in shear stress-induced ATP mtDNA,87),109) that cause ethnic and interindividual release.105) Endothelial cells release ATP from ectopic differences in physical activity, aging and disease 87),109) HFoF1 in response to shear stress, a mechanical force susceptibility. Voluntary will of the brain generated by blood flow, and the ATP released triggers muscle contraction and other organ specific modulates cell functions through activation of down- activities that consume about 50 kg of ATP per day, stream signals to purinergic receptors.105) These but METs change from 0.9 to 15 in a normal adult results suggest that the localization and targeting (Fig. 10, right). Direct measurement of in vivo HFoF1 31 of HFoF1 to caveolin-1/lipid rafts is critical for activity and AMP/ATP ratio is possible by using P shear stress-induced ATP release by endothelial magnetic resonance spectroscopy revealed that ATP 105) cells. synthase flux correlated with O2 uptake (METs) and 110) It remains uncertain how Fo components en- insulin sensitivity. The increase in METs is mainly coded by mtDNA are translocated to rafts as ectopic caused by actomyosin contraction (Fig. 10, right), HFoF1, particularly in the form of intact HFoF1 and muscle-specific . subunit splice variants of HFoF1 (Fig. 10, right). However, inter-organellar trafficof are seen during myogenesis and cardiogenesis mitochondrial proteins was clearly demonstrated (Fig. 7).32),90) The resulting change in substrates to using green fluorescent protein.103) These mitochon- increase lactic acid can be assessed by 1H magnetic drial dynamics will be discussed in the final section. resonance spectroscopy.110) Lactic acidemia is present The energy source ("7H+) to synthesize ATP by not only in individuals undergoing high MET ectopic HFoF1 may not be respiration, but the exercise, but also in the majority of patients with No. 7] ATP synthase: from single molecule to human bioenergetics 687

mitochondrial disorders, including impaired HFo, and of mitochondrial morphology and distribution, and cells isolated from such patients, similarly to ;0 cells, connectivity mediated by tethering and fusion/ require glucose medium (Fig. 8).22),87) Mitochondrial fission events.114) The relevance of these events in disorders can be due to defects in nDNA directly HFoF1 activity has been unraveled after the identi- 87) 98) affecting the assembly or function of HFoF1 and fication of mitofusin and OPA1. Subjects with respiratory chain, defects in mtDNA affecting HFo diabetes showed reduced expression (by 26%) of and the respiratory chain or nDNA influencing mitofusin 2 and a 39% reduction in the ,-subunit of 21),87) 115) mtDNA structure and viability. FoF1. Chronic exercise led to increases in VDAC, The regulatory mechanism of HFoF1 including and the ,-subunit of FoF1 in muscle from control four independent inhibitory sites111) are more com- subjects and from those with diabetes.115) Acute 7),108) plex than those of TFoF1. AMP-activated exercise caused a 4-fold increase in PGC-1, expres- protein kinase (AMPK) activates both oxphos and sion in muscle from control subjects, but not in those glycolysis, functioning as a ‘fuel gauge’ to monitor with diabetes.115) 112) AMP/ATP ratio (Fig. 10, left). ATP-sensitive 10.1. Inhibition of HFoF1 activity and dis- K+ channels both in the plasma membrane and in eases related to bioenergetics. In the energy mitochondria also monitors ATP levels to regulate metabolism of whole human body,114),116) if "7H+ is 87),113) cellular activities. Increases in ATP concen- too low, ATP production by HFoF1 cannot meet tration close the K+ channel, and the resulting demand (mitochondrial diseases),22),87) and if it is too depolarization opens L-type Ca2+ channels (Fig. 10, high, reactive oxygen species (ROS) are pro- bottom). Increased intracellular Ca2+ activates many duced.23),89),116) Oxphos is linked to disease through metabolic processes and proteins for mitochondrial a lack of energy, excess ROS production, or a dynamics and secretion. For example, in the --cells of combination of both,89),116) and diseases caused by mitochondrial myopathy, ATP-sensitive K+ channels mitochondrial dysfunction include diabetes, cancer, are not closed and defective Ca2+-dependent insulin neurodegenerative disorders and ischemia-reperfu- secretion results in mitochondrial diabetes.87) Mito- sion injury.23),89),116) + chondrial ATP-sensitive K channels regulate energy Because of its complex structure, HFoF1 is transfer through their regulation of intermembrane inhibited by over 250 natural and synthetic inhib- space volume and are accordingly essential for the itors.117) In the absence of torque-driving energy, 113) inotropic response during periods of high METs. HFoF1 switches from an ATP synthase to an ATP Although the target residues in HFoF1 and their hydrolase, and this occurs during myocardial ische- signal routes have not yet been determined, mito- mia. The degree of ATP inefficiently hydrolyzed chondrial ATP-sensitive K+ channels are closely during ischemia may be as high as 50–90%.118) At 113) related to kinases including protein kinase CC. In the start of FoF1 study, oligomycin was shown to 3),39)–42) fact, detailed proteomics have revealed, for example, inhibit Fo, and this portion of FoF1 was phosphorylation of ,S76, -T213, -S529, .Y44 or therefore designated “oligomycin sensitivity confer- .Y52, and acetylation of ,K132, -K133, .K79 in ring factor”. Oligomycin cannot be used to treat 91) mammalian FoF1. The monomeric form of HFoF1 myocardial ischemia, as it would reduce ATP contains a phosphorylated . (.Y52) that is not synthesis in normal tissue.3) Only when cellular pH 91) 78) present in the dimeric form. is decreased below 6.8 under ischemia, IF1 inhibits 78) In contrast to bacteria, ATP synthesis by ATPase at the ,- interface of HF1. The restora- + HFoF1 requires exchange of Pi + ADP and ATP tion of "7H favoring ATP synthesis displaces between cytosol and mitochondria by PIC and ANC, IF1 from the ,- interface. However, IF1 does not which are organized as the ATP synthasome completely block hydrolase activity. BMS-199264 (Fig. 2).31) VDAC52) forming a complex with ANC selectively inhibits ATPase activity during ischemia also plays an important role in cytoplasma-mitocon- while having no effect on ATP synthesis, and drial communication.103) As mammalian cells are enhances the recovery of contractile function follow- 118) about 1000 times as large as bacteria, and a ing reperfusion. IF1, ectopic HFoF1 and the mitochondrion is as large as a bacterium, mitochon- opener of mitochondrial ATP-sensitive K+ chan- drial dynamics are essential to distribute synthesized nel113) protect cardiomyocytes from ischemic/reper- ATP (Fig. 10, upper left).114) The concept of fusion damage. mitochondrial dynamics includes the movement of As HFoF1 supplies most ATP needed for human mitochondria along the cytoskeleton, the regulation activity, further study will provide useful insight for 688 Y. KAGAWA [Vol. 86, emergency medicine,118) mitochondrial cardiomy- References opathy22),87) and obesity-related chronic dis- eases.23),115),116) 1) Mitchell, P. (1961) Coupling of phosphorylation to electron and hydrogen transfer by a chemi- Conclusions osmotic type of mechanism. Nature 191, 144–148. 2) Pullman, M.E., Penefsky, H.S., Datta, A. and The excellent work to date on ATP synthases Racker, E. (1960) Partial resolution of the (FoF1) has been reviewed based on data obtained in enzymes catalyzing oxidative phosphorylation. I. fi studies on TFoF1 and HFoF1. The chemiosmotic Puri cation and properties of soluble, dinitrophe- 1) fi "7 + nol-stimulated adenosine triphosphatase. J. Biol. theory was rmly established by H -driven ATP – 7),46) Chem. 235, 3322 3329. synthesis in TFoF1 liposomes (Figs. 1 and 3), and 3) Kagawa, Y. (1972) Reconstitution of oxidative 4) the rotational theory was established by crucial phosphorylation. Biochim. Biophys. Acta 265, observations of .-rotation10),11),17) and ATP synthesis 297–338. on externally added torque to rotate . in TF1 4) Boyer, P.D. (1997) The ATP synthase-a splendid (Fig. 1).64) TF is the only F sufficiently stable to molecular machine. Annu. Rev. Biochem. 66, 1 1 717–749. be consistently analyzed by reconstitution, crystal- 5) Yoshida, M., Sone, N., Hirata, H., Kagawa, Y. and lography, mutagenesis, and nanotechnology for Ui, N. (1979) Subunit structure of adenosine torque-driven ATP synthesis.7) Crystallographic triphosphatase. Comparison of the structure in 9),25) analysis using BF1 (Figs. 2 and 5) and TF1 thermophilic bacterium PS3 with those in mito- – 71),74) chondria, chloroplasts, and Escherichia coli.J. (Figs. 4 6), site-directed mutagenesis using – 18),65) 12),20),63) Biol. Chem. 254, 9525 9533. EFoF1 and TFoF1 (Figs. 5 and 6), and 6) Ohta, S., Tsuboi, M., Oshima, T., Yoshida, M. and 79) dynamic nanotechnology have contributed to Kagawa, Y. (1980) Nucleotide binding to isolated elucidating elastic7),74) and loose61) energy coupling. alpha and beta subunits of proton-translocating Based on the fundamental mechanism of ATP adenosine triphosphatase with circular dichroism. 7),108) J. Biochem. 87, 1609–1617. synthesis in TFoF1, and functional complemen- 8) 7) Kagawa, Y. (1999) Biophysical studies on ATP tation of YF1-deleted yeast with BF1 genes, human synthase. Adv. Biophys. 36,1–25. bioenergetics was developed by research on HFoF1 8) Puri, N., Lai-Zhang, J., Meier, S. and Mueller, D.M. 21),28),86),97) using plasmids, transgenic mice (2005) Expression of bovine F1-ATPase with (Fig. 7),23),90),100) cytoplasts (Fig. 8)33),87) and omics functional complementation in yeast Saccharo- – (Fig. 9).21),84),86),87) HF F differed from TF F in myces cerevisiae. J. Biol. Chem. 280, 22418 o 1 o 1 22424. complex structure (Fig. 2), mitochondrial genetics 9) Abrahams, J.P., Leslie, A.G., Lutter, R. and 87) 90) (Fig. 8), organ specificity (Fig. 7) and intricate Walker, J.E. (1994) Structure at 2.8 A resolution 86) biogenesis (Fig. 9). The complex regulation of of F1-ATPase from bovine heart mitochondria. – HFoF1 has been shown to be essential for daily Nature 370, 621 628. human activity that is triggered by the brain 10) Kagawa, Y. and Hamamoto, T. (1996) The energy transmission in ATP synthase: from the .-c rotor (Fig. 10), thus human bioenergetics is also applicable to the , - oligomer fixed by OSCP-b stator via 113),118) 3 3 to emergency medicine and obesity/diabe- the -DELSEED sequence. J. Bioenerg. Bio- tes23),116) and mitochondrial diseases.87),97) membr. 28, 421–431. 11) Noji, H., Yasuda, R., Yoshida, M. and Kinosita, K. Acknowledgments Jr. (1997) Direct observation of the rotation of F1-ATPase. Nature 386, 299–302. The author extends thanks to Profs. M. 12) Ohta, S., Yohda, M., Ishizuka, M., Hirata, H., Yoshida, T. Hamamoto, H. Endo, S. Ohta, E. Hamamoto, T., Otawara-Hamamoto, Y. et al. Muneyuki, H. Hirata, M. Yohda and N. Sone, with (1988) Sequence and over-expression of subunits whom he has worked in the Department of of adenosine triphosphate synthase in thermo- Biochemistry, Jichi Medical School since 1972, for philic bacterium PS3. Biochim. Biophys. Acta 933, 141–155. their excellent work. The author would also like to 13) Sambongi, Y., Iko, Y., Tanabe, M., Omote, H., thank Prof. E. Racker for his excellent advice. Iwamoto-Kihara, A., Ueda, I. et al. (1999) This study was supported by several Grants-in-Aid Mechanical rotation of the c subunit oligomer in (1973–2008), and the “High-Tech Research Center” ATP synthase (F0F1): direct observation. Science – Project for Private Universities matching fund 286, 1722 1724. – 14) Stock, D., Leslie, A.G.W. and Walker, J.E. (1999) subsidy (1999 2008; for Y.K.) from the Ministry Molecular architecture of the rotary motor in of Education, Culture, Sports, Science and Tech- ATP synthase. Science 286, 1700–1705. nology of Japan. 15) Mitome, N., Suzuki, T., Hayashi, S. and Yoshida, No. 7] ATP synthase: from single molecule to human bioenergetics 689

M. (2004) Thermophilic ATP synthase has a stalk in bovine F1-ATPase at 2.4 A resolution. decamer c-ring: indication of noninteger 10:3 H+/ Nat. Struct. Biol. 7, 1055–1061. ATP ratio and permissive elastic coupling. Proc. 30) Krause, F., Reifschneider, N.H., Goto, S. and Natl. Acad. Sci. USA 101, 12159–12164. Dencher, N.A. (2005) Active oligomeric ATP 16) Düser, M.G., Zarrabi, N., Cipriano, D.J., Ernst, S., synthases in mammalian mitochondria. Biochem. Glick, G.D., Dunn, S.D. et al. (2009) 36° step size Biophys. Res. Commun. 329, 583–590. of proton-driven c-ring rotation in FoF1-ATP 31) Chen, C., Ko, Y., Delannoy, M., Ludtke, S.J., Chiu, synthase. EMBO J. 28, 2689–2696. W. and Pedersen, P.L. (2004) Mitochondrial ATP 17) Yoshida, M., Muneyuki, E. and Hisabori, T. (2001) synthasome: three-dimensional structure by elec- ATP synthase—a marvelous rotary engine of the tron microscopy of the ATP synthase in complex cell. Nat. Rev. Mol. Cell Biol. 2, 669–677. formation with carriers for Pi and ADP/ATP. J. 18) Futai, M., Noumi, T. and Maeda, M. (1989) ATP Biol. Chem. 279, 31761–31768. synthase (H+-ATPase): results by combined bio- 32) Ichida, M., Hakamata, Y., Hayakawa, M., Ueno, E., chemical and molecular biological approaches. Ikeda, U., Shimada, K. et al. (2000) Differential Annu. Rev. Biochem. 58, 111–136. regulation of exonic regulatory elements for 19) Kagawa, Y., Ishizuka, M., Saishu, T. and Nakao, S. muscle-specific alternative splicing during myo- (1986) Stable structure of thermophilic proton genesis and cardiogenesis. J. Biol. Chem. 275, ATPase beta subunit. J. Biochem. 100, 923–934. 15992–16001. 20) Yohda, M., Ohta, S., Hisabori, T. and Kagawa, Y. 33) Hayashi, J., Ohta, S., Kagawa, Y., Kondo, H., (1988) Site-directed mutagenesis of stable adeno- Kaneda, H., Yonekawa, H. et al. (1994) Nuclear sine triphosphate synthase. Biochim. Biophys. but not mitochondrial genome involvement in Acta 933, 156–164. human age-related mitochondrial dysfunction. 21) Kagawa, Y., Hamamoto, T., Endo, H., Ichida, M., Functional integrity of mitochondrial DNA from Shibui, H. and Hayakawa, M. (1997) Genes of aged subjects. J. Biol. Chem. 269, 6878–6883. human ATP synthase: Their roles in physiology 34) Fleischer, S., Brierley, G., Klouwen, H. and and aging. Biosci. Rep. 17, 115–146. Slautterback, D.B. (1962) Studies of the electron 22) Isobe, K., Ito, S., Hosaka, H., Iwamura, Y., Kondo, transfer system. XLVII. The role of phospholipids H., Kagawa, Y. et al. (1998) Nuclear-recessive in electron transfer. J. Biol. Chem. 237, 3264– mutations of factors involved in mitochondrial 3272. translation are responsible for age-related respira- 35) Kakiuchi, S. (1927) The significance of lipoids in tion deficiency of human skin fibroblasts. J. Biol. the oxygen consuming activity of tissues. 1. The Chem. 273, 4601–4606. oxygen consuming activity of tissue and the 23) Kagawa, Y., Yanagisawa, Y., Hasegawa, K., Suzuki, mitochondrial structure. J. Biochem. 7, 263–265. H., Yasuda, K., Kudo, H. et al. (2002) Single 36) Okunuki, K. (1940) Ueber die Rolle der Diaphorase nucleotide polymorphism of thrifty genes for in der Wechsel-wirkung zwischen dem Cytochrom energy metabolism: evolutionary origins and b und dem Dehydrasesystem. Proc. Imp. Acad. prospects for intervention to prevent obesity- (Tokyo) 16, 144–148. related diseases. Biochem. Biophys. Res. Com- 37) Green, D.E. and Hatefi, Y. (1961) The mitochond- mun. 295, 207–222. rion and biochemical machines. Science 133,13– 24) Walker, J.E., Fearnley, I.M., Gay, N.J., Gibson, 19. B.W., Northrop, F.D., Powell, S. et al. (1985) 38) Small, D.M., Bourges, M.C. and Dervichian, D.G. Primary structure and subunit stoichiometry of (1966) The biophysics of lipidic associates I. The F1-ATPase from bovine mitochondria. J. Mol. ternary systems lecithin-bile salt-water. Biochim. Biol. 184, 677–701. Biophys. Acta 125, 563–581. 25) Walker, J.E. and Dickson, V.K. (2006) The 39) Kagawa, Y. and Racker, E. (1971) Partial resolution peripheral stalk of the mitochondrial ATP syn- of the enzymes catalyzing oxidative phosphoryla- thase. Biochim. Biophys. Acta 1757, 286–296. tion. XXV. Reconstitution of vesicles catalyzing 26) Anderson, S., Bankier, A.T., Barrell, B.G., de 32Pi-adenosine triphosphate exchange. J. Biol. Bruijn, M.H., Coulson, A.R., Drouin, J. et al. Chem. 246, 5477–5487. (1981) Sequence and organization of the human 40) Kagawa, Y., Kandrach, A. and Racker, E. (1973) mitochondrial genome. Nature 290, 457–465. Partial resolution of the enzymes catalyzing 27) Matsuda, C., Endo, H., Ohta, S. and Kagawa, Y. oxidative phosphorylation. XXVI. Specificity of (1993) Gene structure of human mitochondrial phospholipids required for energy transfer reac- ATP synthase gamma-subunit. Tissue specificity tions. J. Biol. Chem. 248, 676–684. produced by alternative RNA splicing. J. Biol. 41) Kagawa, Y. and Racker, E. (1966) Partial resolution Chem. 268, 24950–24958. of the enzymes catalyzing oxidative phosphoryla- 28) Ohta, S., Tomura, H., Matsuda, K. and Kagawa, Y. tion. VIII. Properties of a factor conferring (1988) Gene structure of the human mitochon- oligomycin sensitivity on mitochondrial adenosine drial adenosine triphosphate synthase beta sub- triphosphatase. J. Biol. Chem. 241, 2461–2466. unit. J. Biol. Chem. 263, 11257–11262. 42) Kagawa, Y. and Racker, E. (1966) Partial resolution 29) Gibbons, C., Montgomery, M.G., Leslie, A.G. and of the enzymes catalyzing oxidative phosphoryla- Walker, J.E. (2000) The structure of the central tion. IX. Reconstruction of oligomycin-sensitive 690 Y. KAGAWA [Vol. 86,

adenosine triphosphatase. J. Biol. Chem. 241, propagation of the ATP-induced conformational 2467–2474. change of the F1-ATPase - subunit revealed by 43) Kagawa, Y. and Racker, E. (1966) Partial resolution NMR. J. Biol. Chem. 284, 2374–2382. of the enzymes catalyzing oxidative phosphoryla- 57) Saika, K. and Yoshida, M. (1995) A minimum tion. X. Correlation of morphology and function catalytic unit of F1-ATPase shows non-coopera- in submitochondrial particles. J. Biol. Chem. 241, tive ATPase activity inherent in a single catalytic 2475–2482. site with a Km 70 µM. FEBS Lett. 368, 207–210. 44) Yoshida, M., Sone, N., Hirata, H. and Kagawa, Y. 58) Miwa, K. and Yoshida, M. (1989) The ,3-3 (1975) A highly stable adenosine triphosphatase complex, the catalytic core of F1-ATPase. Proc. from a thermophilic bacterium. Purification, Natl. Acad. Sci. USA 86, 6484–6487. properties and reconstitution. J. Biol. Chem. 59) Paik, S.R., Yokoyama, K., Yoshida, M., Ohta, T., 250, 7910–7916. Kagawa, Y. and Allison, W.S. (1993) The TF1- 45) Sone, N., Yoshida, M., Hirata, H. and Kagawa, Y. ATPase and ATPase activities of assembled (1975) Purification, properties of a dicyclohexyl- ,3-3., ,3-3./, and ,3-3.C complexes are stimu- carbodiimide-sensitive adenosine triphosphatase lated by low and inhibited by high concentrations from a thermophilic bacterium. J. Biol. Chem. of rhodamine 6G whereas the dye only inhibits the 250, 7917–7923. ,3-3, and ,3-3/ complexes. J. Bioenerg. Bio- 46) Sone, N., Yoshida, M., Hirata, H. and Kagawa, Y. membr. 25, 679–684. (1977) Adenosine triphosphate synthesis by elec- 60) Aloise, P., Kagawa, Y. and Coleman, P.S. (1991) trochemical proton gradient in vesicles reconsti- Comparative Mg2+-dependent sequential covalent tuted from purified adenosine triphosphatase and binding stoichiometries of 3′-O-(4-benzoyl)benzo- phospholipids of thermophilic bacterium. J. Biol. yl adenosine 5′-diphosphate of MF1,TF1, and the Chem. 252, 2956–2960. ,3-3 core complex of TF1. The binding change 47) Jagendorf, A.T. (1967) Acid-base transitions and motif is independent of the F1 ./C subunits. J. phosphorylation by chloroplasts. Fed. Proc. 26, Biol. Chem. 266, 10368–10376. 1361–1369. 61) Oosawa, F. and Hayashi, S. (1986) The loose 48) Witt, H.T., Schlodder, E. and Gräber, P. (1976) coupling mechanism in molecular machines of Membrane-bound ATP synthesis generated by an living cells. Adv. Biophys. 22, 151–183. external electrical field. FEBS Lett. 69, 272–276. 62) Feniouk, B.A. and Yoshida, M. (2007) Regulatory 49) Okamoto, H., Sone, N., Hirata, H., Yoshida, M. and mechanisms of proton-translocating FoF1-ATP Kagawa, Y. (1977) Purified proton conductor in synthase. Results Probl. Cell Differ. 45, 279–308. proton translocating adenosine triphosphatase of 63) Ariga, T., Muneyuki, E. and Yoshida, M. (2007) F1- a thermophilic bacterium. J. Biol. Chem. 252, ATPase rotates by an asymmetric, sequential 6125–6131. mechanism using all three catalytic subunits. Nat. 50) Rögner, M., Ohno, K., Hamamoto, T., Sone, N. and Struct. Mol. Biol. 14, 841–846. Kagawa, Y. (1979) Net ATP synthesis in H+- 64) Itoh, H., Takahashi, A., Adachi, K., Noji, H., ATPase macroliposomes by an external electric Yasuda, R., Yoshida, M. et al. (2004) Mechan- field. Biochem. Biophys. Res. Commun. 91, 362– ically driven ATP synthesis by F1-ATPase. 367. Nature 427, 465–468. 51) Muneyuki, E., Kagawa, Y. and Hirata, H. (1989) 65) Futai, M., Wada, Y. and Kaplan, J.H. (eds.) (2004) Steady state kinetics of proton translocation Handbook of ATPases. Wiley-VCH, Germany. catalyzed by thermophilic FoF1-ATPase reconsti- pp. 1–468. tuted in planar bilayer membranes. J. Biol. Chem. 66) Kagawa, Y., Nojima, H., Nukiwa, N., Ishizuka, M., 264, 6092–6096. Nakajima, T., Yasuhara, T. et al. (1984) High 52) Zalman, L.S., Nikaido, H. and Kagawa, Y. (1980) guanine plus cytosine content in the third letter of Mitochondrial outer membrane contains a protein codons of an extreme thermophile. DNA sequence producing nonspecificdiffusion channels. J. Biol. of the isopropylmalate dehydrogenase of Thermus Chem. 255, 1771–1774. thermophilus. J. Biol. Chem. 259, 2956–2960. 53) Yoshida, M., Sone, N., Hirata, H. and Kagawa, Y. 67) Ohtsubo, M., Yoshida, M., Ohta, S., Kagawa, Y., (1977) Reconstitution of adenosine triphosphatase Yohda, M. and Date, T. (1987) In vitro mutated - of thermophilic bacterium from purified individual subunits from the F1-ATPase of the thermophilic subunits. J. Biol. Chem. 252, 3480–3485. bacterium, PS3, containing glutamine in place of 54) Kagawa, Y. and Hamamoto, T. (1997) Intramolec- glutamic acid in positions 190 or 201 assembles ular rotation in ATP synthase: dynamic and with the , and . subunits to produce inactive crystallographic studies on thermophilic F1. Bio- complexes. Biochem. Biophys. Res. Commun. chem. Biophys. Res. Commun. 240, 247–256. 146, 705–710. 55) Kagawa, Y., Ohta, S., Harada, M., Sato, M. and 68) Imamura, H., Nakano, M., Noji, H., Muneyuki, E., Itoh, Y. (1992) The ,3-3 and ,1-1 complexes of Ohkuma, S., Yoshida, M. et al. (2003) Evidence ATP synthase. Ann. N. Y. Acad. Sci. 671, 366– for rotation of V1-ATPase. Proc. Natl. Acad. Sci. 376. USA 100, 2312–2315. 56) Yagi, H., Kajiwara, N., Iwabuchi, T., Izumi, K., 69) Wakabayashi, T., Kubota, M., Yoshida, M. and Yoshida, M. and Akutsu, H. (2009) Stepwise Kagawa, Y. (1977) Structure of ATPase (coupling No. 7] ATP synthase: from single molecule to human bioenergetics 691

+ + factor TF1) from a thermophilic bacterium. J. The thermodynamic H /ATP ratios of the H - Mol. Biol. 117, 515–519. ATPsynthases from chloroplasts and Escherichia 70) Shirakihara, Y., Yohda, M., Kagawa, Y., coli. Proc. Natl. Acad. Sci. USA 105, 3745–3750. Yokoyama, K. and Yoshida, M. (1991) Purifica- 83) Sielaff, H., Rennekamp, H., Wächter, A., Xie, H., tion by dye-ligand chromatography and a crys- Hilbers, F., Feldbauer, K. et al. (2008) Domain tallization study of the F1-ATPase and its major compliance and elastic power transmission in subunits, - and ,, from a thermophilic bacterium, rotary FoF1-ATPase. Proc. Natl. Acad. Sci. USA PS3. J. Biochem. 109, 466–471. 105, 17760–17765. 71) Shirakihara, Y., Leslie, A.G., Abrahams, J.P., 84) Akiyama, S., Endo, H., Inohara, N., Ohta, S. and Walker, J.E., Ueda, T., Sekimoto, Y. et al. Kagawa, Y. (1994) Gene structure and cell type- (1997) The crystal structure of the nucleotide-free specific expression of the human ATP synthase ,3-3 subcomplex of F1-ATPase from the thermo- alpha subunit. Biochim. Biophys. Acta 1219, philic Bacillus PS3 is a symmetric trimer. Struc- 129–140. ture 5, 825–836. 85) Tomura, H., Endo, H., Kagawa, Y. and Ohta, S. 72) Vollmar, M., Schlieper, D., Winn, M., Büchner, C. (1990) Novel regulatory enhancer in the nuclear and Groth, G. (2009) Structure of the c14 rotor gene of the human mitochondrial ATP synthase ring of the proton translocating chloroplast ATP beta-subunit. J. Biol. Chem. 265, 6525–6527. synthase. J. Biol. Chem. 284, 18228–18235. 86) Kagawa, Y. and Ohta, S. (1990) Regulation of 73) Dickson, V.K., Silvester, J.A., Fearnley, I.M., Leslie, mitochondrial ATP synthesis in mammalian cells A.G. and Walker, J.E. (2006) On the structure of by transcriptional control. Int. J. Biochem. 22, the stator of the mitochondrial ATP synthase. 219–229. EMBO J. 25, 2911–2918. 87) Sato, A., Endo, H., Umetsu, K., Sone, H., 74) Kagawa, Y., Hamamoto, T. and Endo, H. (2000) Yanagisawa, Y., Saigusa, A. et al. (2003) Poly- The ,/- interfaces of ,1-1, ,3-3, and F1: Domain morphism, heteroplasmy, mitochondrial fusion motions and elastic energy stored during . and diabetes. Biosci. Rep. 23, 313–337. rotation. J. Bioenerg. Biomembr. 32, 471–484. 88) Vives-Bauza, C., Magrane, J., Andreu, A.L. and 75) Kabaleeswaran, V., Shen, H., Symersky, J., Walker, Manfredi, G. (2010) Novel role of ATPase subunit J.E., Leslie, A.G. and Mueller, D.M. (2009) c targeting peptide beyond mitochondrial import. Asymmetric structure of the yeast F1 ATPase in Mol. Biol. Cell 21, 131–139. the absence of bound nucleotides. J. Biol. Chem. 89) Hüttemann, M., Lee, I., Pecinova, A., Pecina, P., 284, 10546–10551. Przyklenk, K. and Doan, J.W. (2008) Regulation 76) Kagawa, R., Montgomery, M.G., Braig, K., Leslie, of oxidative phosphorylation, the mitochondrial A.G. and Walker, J.E. (2004) The structure of membrane potential, and their role in human bovine F1-ATPase inhibited by ADP and beryl- disease. J. Bioenerg. Biomembr. 40, 445–456. lium fluoride. EMBO J. 23, 2734–2744. 90) Hayakawa, M., Sakashita, E., Ueno, E., Tominaga, 77) Senter, P., Eckstein, F. and Kagawa, Y. (1983) S., Hamamoto, T., Kagawa, Y. et al. (2002) Substrate metal adenosine 5′-triphosphate chelate Muscle-specific exonic splicing silencer for exon structure and stereochemical course of reaction exclusion in human ATP synthase gamma-sub- catalyzed by the adenosine triphosphatase from unit pre-mRNA. J. Biol. Chem. 277, 6974–6984. the thermophilic bacterium PS3. Biochemistry 22, 91) Kane, L.A. and Van Eyk, J.E. (2009) Post-transla- 5514–5518. tional modifications of ATP synthase in the heart: 78) Cabezón, E., Montgomery, M.G., Leslie, A.G. and biology and function. J. Bioenerg. Biomembr. 41, Walker, J.E. (2003) The structure of bovine F1- 145–150. ATPase in complex with its regulatory protein 92) Champagne, E., Martinez, L.O., Collet, X. and IF1. Nat. Struct. Biol. 10, 744–750. Barbaras, R. (2006) Ecto-F1Fo ATP synthase/F1 79) Yasuda, R., Masaike, T., Adachi, K., Noji, H., Itoh, ATPase: metabolic and immunological functions. H. and Kinosita, K. Jr. (2003) The ATP-waiting Curr. Opin. Lipidol. 17, 279–284. conformation of rotating F1-ATPase revealed by 93) Dillin, A., Hsu, A.L., Arantes-Oliveira, N., Lehrer- single-pair fluorescence resonance energy transfer. Graiwer, J., Hsin, H., Fraser, A.G. et al. (2002) Proc. Natl. Acad. Sci. USA 100, 9314–9318. Rates of behavior and aging specified by mito- 80) Galvez, E.M., Zimmermann, B., Rombach-Riegraf, chondrial function during development. Science V., Bienert, R. and Gräber, P. (2007) Fluores- 298, 2398–2401. cence resonance energy transfer in single enzyme 94) Ito, S., Ohta, S., Nishimaki, K., Kagawa, Y., Soma, molecules with a quantum dot as donor. Eur. R., Kuno, S.Y. et al. (1999) Functional integrity Biophys. J. 37, 1367–1371. of mitochondrial genomes in human platelets and 81) Masaike, T., Koyama-Horibe, F., Oiwa, K., autopsied brain tissues from elderly patients with Yoshida, M. and Nishizaka, T. (2008) Cooperative Alzheimer’s disease. Proc. Natl. Acad. Sci. USA three-step motions in catalytic subunits of F1- 95, 2099–2103. ATPase correlate with 80 degrees and 40 degrees 95) Tominaga, K., Akiyama, S., Kagawa, Y. and Ohta, substep rotations. Nat. Struct. Mol. Biol. 15, S. (1992) Upstream region of a genomic gene for 1326–1333. human mitochondrial transcription factor 1. Bio- 82) Steigmiller, S., Turina, P. and Gräber, P. (2008) chim. Biophys. Acta 1131, 217–219. 692 Y. KAGAWA [Vol. 86,

96) Smith, C.P. and Thorsness, P.E. (2005) Formation tial molecular target for anti-obesity drugs. FEBS of an energized inner membrane in mitochondria Lett. 581, 3405–3409. with a gamma-deficient F1-ATPase. Eukaryot. 108) von Ballmoos, C., Wiedenmann, A. and Dimroth, P. Cell 4, 2078–2086. (2009) Essentials for ATP synthesis by F1F0 ATP 97) Kagawa, Y. and Hayashi, J.-I. (1997) Gene therapy synthases. Annu. Rev. Biochem. 78, 649–672. of mitochondrial diseases using human cytoplasts. 109) Takasaki, S. (2009) Mitochondrial haplogroups Gene Ther. 4,6–10. associated with Japanese centenarians, Alz- 98) Satoh, M., Hamamoto, T., Seo, N., Kagawa, Y. and heimer’s patients, Parkinson’s patients, type 2 Endo, H. (2003) Differential sublocalization of the diabetic patients and healthy non-obese young dynamin-related protein OPA1 isoforms in mito- males. J. Genet. Genomics 36, 425–434. chondria. Biochem. Biophys. Res. Commun. 300, 110) Kacerovsky-Bielesz, G., Chmelik, M., Ling, C., 482–493. Pokan, R., Szendroedi, J., Farukuoye, M. et al. 99) Solaini, G., Harris, D.A., Lenaz, G., Sgarbi, G. and (2009) Short-term exercise training does not Baracca, A. (2008) The study of the pathogenic stimulate skeletal muscle ATP synthesis in mechanism of mitochondrial diseases provides relatives of humans with type 2 diabetes. Diabetes information on basic bioenergetics. Biochim. 58, 1333–1341. Biophys. Acta 1777, 941–945. 111) Gledhill, J.R. and Walker, J.E. (2006) Inhibitors of 100) Nakada, K., Sato, A., Sone, H., Kasahara, A., Ikeda, the catalytic domain of mitochondrial ATP K., Kagawa, Y. et al. (2004) Accumulation of synthase. Biochem. Soc. Trans. 34, 989–992. pathogenic DeltamtDNA induced deafness but 112) Steinberg, G.R. and Kemp, B.E. (2009) AMPK in not diabetic phenotypes in mito-mice. Biochem. Health and Disease. Physiol. Rev. 89, 1025–1078. Biophys. Res. Commun. 323, 175–184. 113) Costa, A.D. and Garlid, K.D. (2009) MitoKATP 101) Chanséaume, E. and Morio, B. (2009) Potential activity in healthy and ischemic hearts. J. mechanisms of muscle mitochondrial dysfunction Bioenerg. Biomembr. 41, 123–126. in aging and obesity and cellular consequences. 114) Liesa, M., Palacín, M. and Zorzan, A. (2009) Int. J. Mol. Sci. 10, 306–324. Mitochondrial dynamics in mammalian health 102) Meyer, B., Wittig, I., Trifilieff, E., Karas, M. and and disease. Physiol. Rev. 89, 799–845. Schägger, H. (2007) Identification of two proteins 115) Hernández-Alvarez, M.I., Thabit, H., Burns, N., associated with mammalian ATP synthase. Mol. Shah, S., Brema, I., Hatunic, M. et al. (2009) Cell. Proteomics 6, 1690–1699. Subjects with early-onset type 2 diabetes show 103) Kasashima, K., Ohta, E., Kagawa, Y. and Endo, H. defective activation of the skeletal muscle PGC- (2006) Mitochondrial functions and estrogen 1,/mitofusin-2 regulatory pathway in response to receptor-dependent nuclear translocation of plei- physical activity. Diabetes Care 33, 645–651. otropic human prohibitin 2. J. Biol. Chem. 281, 116) Kagawa, Y., Cha, S., Hasegawa, K., Hamamoto, T. 36401–36410. and Endo, H. (1999) Regulation of energy 104) Buzhynskyy, N., Sens, P., Prima, V., Sturgis, J.N. metabolism in human cells in aging and diabetes: and Scheuring, S. (2007) Rows of ATP synthase FoF1, mtDNA, UCP, and ROS. Biochem. Bio- dimers in native mitochondrial inner membranes. phys. Res. Commun. 266, 662–676. Biophys. J. 93, 2870–2876. 117) Hong, S. and Pedersen, P.L. (2008) ATP synthase 105) Yamamoto, K., Shimizu, N., Obi, S., Kumagaya, S., and the actions of inhibitors utilized to study its Taketani, Y., Kamiya, A. et al. (2007) Involve- roles in human health, disease, and other scientific ment of cell surface ATP synthase in flow-induced areas. Microbiol. Mol. Biol. Rev. 72, 590–641. ATP release by vascular endothelial cells. Am. J. 118) Grover, G.J. and Malm, J. (2008) Pharmacological Physiol. Heart Circ. Physiol. 293, H1646–H1653. profile of the selective mitochondrial F1F0 ATP 106) Chi, S.L. and Pizzo, S.V. (2006) Cell surface F1Fo hydrolase inhibitor BMS-199264 in myocardial ATP synthase: a new paradigm? Ann. Med. 38, ischemia. Cardiovasc. Ther. 26, 287–296. 429–438. 107) Arakaki, N., Kita, T., Shibata, H. and Higuti, T. (2007) Cell-surface H+-ATP synthase as a poten- (Received Dec. 23, 2009; accepted Apr. 30, 2010) No. 7] ATP synthase: from single molecule to human bioenergetics 693

Profile

Yasuo Kagawa was born in 1932. He graduated from the Medical School at the University of Tokyo in 1957, and after completing an internship at St. Luke’s Hospital, he received his PhD from the Graduate School of the University of Tokyo in 1962. He continued his bioenergetics research as a Fulbright scholar at the Public Health Research Institute of New York in 1963, as an assistant at the University of Tokyo in 1965, as a professor at Shinshu University in 1968, as a visiting professor of Biochemistry and Molecular Biology at Cornell University in 1970, and, since 1972, as a professor at Jichi

Medical School (JMS), where his studies on F1Fo ATP synthase have garnered international attention. After retiring in 1998, Dr. Kagawa was appointed Professor Emeritus at JMS and joined Kagawa Nutrition University (KNU) as a professor of Medical Chemistry. He became the vice president of KNU in 1999. He has also served as managing editor of several international journals, and was a committee member of the International Union of Biochemistry and Molecular Biology. Dr. Kagawa has received several awards, including the Medical Research Prize from the Japanese Medical Association in 1985, the Medal with Purple Ribbon in 1996, and the Order of the Sacred Treasure, Gold Rays with Neck Ribbon from the Government of Japan in 2006.