Visualizing mitochondrial FoF1-ATP synthase as the target of the immunomodulatory drug

Bz-423

Ilka Starke1,2, Gary D. Glick3, Michael Börsch1,4

1 Single-Molecule Microscopy Group, Jena University Hospital, Friedrich Schiller University,

Nonnenplan 2 - 4, 07743 Jena, Germany

2 Institute for Physical Chemistry, Albert Ludwig University Freiburg, Albertstrasse 23a, 79104

Freiburg, Germany

3 Department of Chemistry, University of Michigan, 930 N University Ave, Ann Arbor, MI 48109-1055,

USA

4 Abbe Center of Photonics (ACP), Jena, Germany

Abstract

Targeting the mitochondrial enzyme FoF1-ATP synthase and modulating its catalytic activities with small

molecules is a promising new approach for treatment of autoimmune diseases. The immuno-

modulatory compound Bz-423 is such a drug that binds to subunit OSCP of the mitochondrial FoF1-ATP

synthase and induces apoptosis via increased reactive oxygen production in coupled, actively respiring

mitochondria. Here we review the experimental progress to reveal the binding of Bz-423 to the

mitochondrial target and discuss how subunit rotation of FoF1-ATP synthase is affected by Bz-423.

Briefly, we report how Förster resonance energy transfer (FRET) can be employed to colocalize the

enzyme and the fluorescently tagged Bz-423 within the mitochondria of living cells with nanometer

resolution.

1. Introduction

Cellular processes as metabolism and transport are powered by the universal chemical “energy

currency” that is the molecule adenosine triphosphate (ATP). Therefore, millimolar ATP concentrations

inside the cell have to be produced and maintained through sequential catalytic reactions by the

glycolysis pathway or more efficiently by the FoF1-ATP synthase as part of the oxidative

phosphorylation (OXPHOS) pathway. FoF1-ATP synthases are working in the plasma membrane of

bacteria or in small organelles inside of eukaryotes, i.e. the thylakoid membrane in chloroplasts or the

inner mitochondrial membrane. In case of mitochondrial FoF1-ATP synthase a proton motive force

(PMF) comprising a concentration difference of protons across the membrane (pH) plus an electric

membrane potential ( is required for ATP synthesis. The PMF is generated by sequential redox

processes and associated proton pumping of the enzyme complexes I to IV of the respiratory chain

across the inner mitochondrial membrane.

If we consider autoimmune diseases, for example systemic lupus erythematodes, being caused by

hyperactivity of pathogenic T cells of the immune system, then controlling their cellular ATP

concentration with drugs and reducing their activity could become a promising approach for clinical

treatment. Modulating T cell activity temperately could circumvent a complete shut-down of the

normal immune function. Therefore, one option would be controlling the PMF by targeting any of the

enzyme complexes I to IV of the respiratory chain with inhibitors. Vice versa, controlling the efficiency

of converting PMF to ATP synthesis by FoF1-ATP synthase would be a possible approach. This latter

process is called uncoupling.

More than a decade ago, a 1,4-benzodiazepine, Bz-423 (Figure 1 A), has been found to target lymphoid

cells in a murine model of lupus erythematodes[1]. Bz-423 specifically induced apoptosis of pathogenic

lymphocytes and attenuated disease progression. As a result, the treated mice showed a prolonged

survival at the therapeutic dosage without adverse toxicity and with maintained immune function[2].

The mechanism of Bz-423 action was revealed and subsequently the molecular target was identified –

the mitochondrial FoF1-ATP synthase[3]. Here we focus on the discovery of the drug binding site and

discuss a recent microscopy approach using Förster resonance energy transfer (FRET) that has directly

demonstrated the binding of a fluorescent Bz-423 derivative to the mitochondrial enzyme in living

cells[4].

2. Bz-423 binds to OSCP of mitochondrial FoF1-ATP synthase

Initially, Bz-423 was identified as a potential drug candidate from a library of 1,4-benzodiazepines

generated by diversity-oriented chemical synthesis[1]. Phenotype screening of Ramos B cells revealed

that Bz-423 caused cell shrinkage, nuclear condensation, cytoplasmic vacuolization, membrane

blebbing and DNA fragmentation. Other 1,4-benzodiazepine derivatives were found to selectively

target T-cells[5-8]. Bz-423 did not strongly bind to the peripheral benzodiazepine receptor. Cytotoxicity

of Bz-423 was related to rapidly generated superoxide (O2.-) in mitochondria. Superoxide is one of the

reactive oxygen species (ROS) that can chemically damage cellular macromolecules at higher

concentrations. However, Bz-423 superoxide signaling for induced apoptosis was proven as the

underlying mechanisms[9, 10].

In the presence of 1 mM sodium azide, the proapoptotic O2.- generation by Bz-423 was abolished[1].

Because sodium azide is an inhibitor of cytochrome c oxidase, i.e. complex IV of the mitochondrial

respiratory chain, Bz-423 was proposed to bind to a mitochondrial OXPHOS protein. Binding of Bz-423

did not alter or collapse the electric potential across the inner mitochondrial membrane. The

superoxide response by Bz-423 required active mitochondria in state 3 respiration, but not

mitochondria in state 4 with minimal respiration and in the absence of ADP. Comparing the superoxide

generation mechanism induced by oligomycin that induces a state-3-to-4 transition of mitochondrial

respiration[11] lead to the hypothesis that Bz-423 might cause a state-3-to-4 transition as well and

might bind to FoF1-ATP synthase.

The validation of FoF1-ATP synthase as the mitochondrial target of Bz-423 was achieved by phage

display screening[3]. Briefly, the oligomycin sensitivity conferring protein (OSCP) being a subunit of

FoF1-ATP synthase was determined. Subsequently the binding site of Bz-423 was located by NMR

spectroscopy using the isolated subunit OSCP in solution[12]. Figure 1 A shows the model of bovine

heart FoF1-ATP synthase with the highlighted subunit OSCP in green and the amino acid residues

involved in binding of Bz-423 as red dots. Accordingly Bz-423 binds to the top of the membrane enzyme

at the interface of OSCP with one pair of subunits (blue in Fig. 1 A) of the F1 part. Binding of water-

soluble Bz-423 analogues to OSCP in a chemical shift perturbation NMR measurement revealed specific

residues 51, 55, 65, 66, 75, 77 and 92 that might form a hydrophobic pocket to accommodate the drug.

Furthermore, Bz-423 binding resulted in a conformational rearrangement of helices in OSCP and might

alter the interaction of OSCP with F1 in an allosteric manner[12].

Figure 1: A, model of the mitochondrial FoF1-ATP synthase from bovine heart with binding site for Bz-423 (red dots) on subunit

OSCP (in green, on the top; from cryoEM data, PDB 5ARA [13]). Subunits and are shown in blue and the rotary subunits

, and in orange of the F1 part. The peripheral stator consists of subunits b, d, F6 and A6L (in silver). The proton-translocating

subunit a is depicted in black, and the rotor ring of eight c subunits is shown in yellow. B, structure of OSCP from bovine heart

(from x-ray crystallographic data, PDB 2WSS [14]) with highlighted residues 51, 55, 65, 66, 75, 77 and 92 shown to be involved

in Bz-423 binding. C, Bz-423 derivative with the 1,4-benzodiazepine moiety highlighted by the bluish box, a flexible hexyl

linker L (grey ellipse) and the FRET acceptor fluorophore Cy5 (red ellipse). Bz-423-Cy5 is expected to bind to OSCP of the yeast

mitochondrial FoF1 (red dots on the top of the monomer of the yeast enzyme, cryoEM data, PDB 6B8H [15]). The FRET donor

yeast-enhanced green fluorescent protein (yEGFP) is fused to the extended C-terminus of the -subunit (symbolized by the

green box).

3. Bz-423 requires OSCP to modulate FoF1-ATP synthase activity in vitro and in cells

Binding of Bz-423 to OSCP in the intact FoF1-ATP synthase inhibited both synthesis and hydrolysis of

ATP in isolated sub-mitochondrial particles (SMP) in vitro[3, 16]. Both maximal turnover Vmax and KM

were changed, in contrast to the inhibitor oligomycin which reduced Vmax only. ATP hydrolysis by the

soluble mitochondrial F1 part (comprising the blue and orange colored subunits the Fig. 1 A) was

reduced but only when F1 was assembled with OSCP. The IC50 for Bz-423 was about 5 µM. In perfused

HEK cells ATP synthesis rates of mitochondria were reduced, with IC50 of less than 5 µM for Bz-423.

Engineered HEK cells with a specifically reduced content of OSCP by siRNA showed alleviated apoptosis

by Bz-423, and the residual amount of cellular OSCP correlated well with the percentage of apoptotic

cells[3].

FoF1-ATP synthase accomplishes ATP synthesis by mechanochemical energy conversion with two rotary

subunit motors. The PMF drives the Fo motor when protons enter the half-channel of the membrane-

embedded a subunit (black in Fig. 1 A) from the intermembrane space, i.e. from below. Binding to a

specific residue on one c subunit (yellow in Fig. 1A) compensates electrostatic forces and allows the

ring of c subunits to rotate one step forward. Rotation of the c-ring moves the elastically-coupled

attached F1 motor (orange in Fig. 1 A comprising subunits , , and ). The F1 motor rotates in three

120° steps at high PMF and stops at each of the three catalytic sites where ATP is synthesized in F1.

These distinct step sizes of the rotary Fo and F1 motors during ATP synthesis have been measured in

vitro in single-molecule experiments using bacterial FoF1-ATP synthases[17-19]. FoF1-ATP synthase can

also run in reverse by hydrolyzing ATP. ATP hydrolysis has been used to investigate the F1 motor in

great detail since 20 years[20] and revealed torque, elastic domains[21, 22], substeps[23, 24], breaks

and other motor properties[25-27].

Internal subunit rotation with high torque requires a mechanically stable stator counterpart of the

enzyme. The stator of the mitochondrial FoF1-ATP synthase comprises the six F1 subunits 33 (blue in

Fig. 1 A), subunits b, d, F6, A6L (silver in Fig. 1 A), the a subunit (black in Fig. 1 A) and OSCP (green in

Fig. 1 A). Binding of Bz-423 to the interface of OSCP with 33 might weaken the stator assembly and

might cause transient uncoupling of the F1 and Fo motors. Alternatively, Bz-423 might influence the

subtle conformational changes of OSCP bound to the top part of the catalytic 33 subunits and

thereby provokes reduced rates of ATP synthesis and hydrolysis. Quantitative enzymatic analysis

revealed that Bz-423 is an uncompetitive inhibitor of mitochondrial FoF1-ATP synthase decreasing Vmax

for ATP synthesis to 50% in the presence of ~10 µM Bz-423[16]. Inhibition by µM amounts of Bz-423

corresponded to fast off-rates < 0.3 s-1 of the drug from FoF1-ATP synthase and a 90% recovery of ATP

synthesis rates after 10 seconds.

4. Imaging the drug and its molecular target in mitochondria

Localizing a drug at a specific target in life cells can be achieved using fluorescence microscopy with

high spatial and temporal resolution as well as single molecule sensitivity. A variety of functional Bz-

423 derivatives was synthesized with a flexible linker to rhodamine- or cyanine fluorophores, for

example Bz-423-Cy5 (Fig. 1 C). Because µM concentrations of Bz-423 are required to bind to OSCP and

to induce apoptosis, direct imaging of the fluorescent drug bound to FoF1-ATP synthase in the inner

mitochondrial membrane is not possible due to a high fluorescent background of unbound Bz-423

throughout the cell. Fast off-rates of Bz-423 and its fluorescent derivatives prevent extensive washing

of the cells which is needed to obtain a good imaging contrast. Therefore, confocal microscopy with

about 250 nm resolution or superresolution microscopies like structured illumination SIM

[28](resolution limit of about 100 nm) or stimulated emission depletion STED [29](resolution limit of

about 20 nm) cannot be used to identify bound Bz-423 on OSCP.

Instead, Förster resonance energy transfer[30, 31] (FRET) as a distance measurement approach in the

2 to 9 nanometer range is applicable. The dipole-dipole interaction of FRET between to nearby

fluorophores causes a relative loss of fluorescence intensity of the FRET donor (excited by the laser)

and an increase of fluorescence intensity of the FRET acceptor. Thus FRET can be used to relate the

spatial position of the fluorescently labelled drug to its cellular target that is tagged with a different

fluorophore. Mitochondrial FoF1-ATP synthase can be tagged with fluorescent proteins, for example at

OSCP[32] or at the rotary subunit without affecting function[33-35]. The benefit of using a genetic

fusion to the enzyme is a negligible fluorescent background from other parts of the cell than the cristae

of the inner mitochondrial membrane.

5. Revealing Bz-423-Cy5 binding to FoF1-ATP synthase by FRET acceptor photobleaching

To detect binding of fluorescent Bz-423 to mitochondrial FoF1-ATP synthase in living S. cerevisiae cells,

we used a FoF1-ATP synthase mutant designed by J. Petersen and P. Gräber (Albert Ludwig University

Freiburg, to be published). A fusion of the yeast-enhanced green fluorescent protein (yEGFP) linked to

the C-terminus of the subunit (Fig. 1 C) provided the donor for FRET imaging (Fig. 2). The fully

functional mutant was checked for ATP synthesis, and catalytic rates were also determined in vitro

after protein purification and reconstitution to liposomes. Staining mitochondria was achieved by

incubating the S. cerevisiae cells with 4 µm Bz-423-Cy5 (see structure in Fig. 1 C) in the presence of 2%

EtOH for 8 h at 28°C. After washing, the bluish cells (Fig. 2 E) were imaged immediately on the

microscope at 22°C[4].

Widefield fluorescence microscopy of the stained yeast showed spherical cells with bright spots

suggesting fluorescent FoF1-ATP synthase in mitochondria (Fig. 2 A). SIM imaging of unstained yeast

confirmed that only mitochondria were fluorescent[4]. Repeated imaging of the same field of view

indicated comparable pixel intensities in these cells revealing only minor photobleaching of yEGFP.

However, subsequent excitation of the cells with 640 nm at high power photobleached the FRET

acceptor Cy5 dye on Bz-423, and the loss of FRET acceptor resulted in an increase of the FRET donor

intensity when imaged again with 488 nm (Fig. 2 B). Stepwise photobleaching of Cy5 correlated with

an stepwise increase of FRET donor fluorescence (Fig. 2 C, D). Analysis of the intensity increase in

individual cells (see histograms of the single cell highlighted by the red circle in Fig. 2 A-D) due to Cy5

photobleaching unequivocally corroborated the binding of Bz-423-Cy5 to a position on FoF1-ATP

synthase only few nm away from the yEGFP chromophore at the extended C-terminus of the subunit.

Most likely, Bz-423-Cy5 was bound to OSCP.

FRET acceptor photobleaching to confirm the molecular target of a drug in living cells is a fluorescence

microscopy approach that requires fast imaging capability but not necessarily high photon counts rates

per pixel to achieve accurate colocalization. Despite weak (µM) binding affinities of Bz-423-Cy5 and

fast exchange of the immunomodulator on OSCP, a significant fraction of bound Bz-423-Cy5 was

revealed by FRET. Slow transport of Bz-423 across the membranes into the matrix of mitochondria as

the final destination was indicated by long incubation times needed for staining the yeast cells (Fig. 2

E). Accordingly, a significant fraction of Bz-423-Cy5 still remained in the cytosol and outside of the

mitochondria during FRET acceptor photobleaching, and did not contribute to apoptotic action.

However, slow exchange of photobleached Bz-423-Cy5 into and out of the matrix compartment

facilitated the FRET detection.

Figure 2: FRET acceptor photobleaching of Bz-423-Cy5 bound to mitochondrial FoF1-ATP synthase with yEGFP fusion to the

subunit in living S. cerevisiae[4]. EMCCD-based widefield fluorescence microscopy used laser excitation of yEGFP with 488 nm

and fluorescence detection of yEGFP between 500 to 550 nm. A – D, sequential imaging of yEGFP-tagged FoF1-ATP synthases

with recalculated, false-colored intensities in 8 bit (0 – 255). Between each image, a 30-s high-power laser pulse with 640 nm

was applied to partially photobleach the Cy5 chromophore. The individual pixel intensities of the highlighted cell (red circle

as the region-of-interest ROI) are plotted in the histogram in the lower panel as recorded by the EMCCD camera. E,

comparison of S. cerevisiae cells without Bz-423 on the left and stained with 4 µM Bz-423-Cy5 on the right (modified from [4]

with permission).

6. Future developments

Following the first demonstration that the 1,4-benzodiazepine Bz-423 induced apoptosis in living B

cells by stimulating superoxid production of the OXPHOS complexes[1] the identification of its

molecular target in mitochondria was accomplished by biochemical methods. Subunit OSCP of FoF1-

ATP synthase being the destination of the drug was unraveled by human cDNA T7 phage display

screening. Genetic removal of OSCP in mutant FoF1-ATP synthases proved that apoptosis required

binding of Bz-423 to this subunit. Using purified soluble OSCP, the amino acids involved in the binding

site were discovered by NMR spectroscopy[12].

A detailed view on the Bz-423 binding site at the interface of OSCP with the N-termini of and

subunits of F1 is permitted by the recent high-resolution CryoEM structures of the complete

mitochondrial enzymes from bovine heart and from yeast S. cerevisiae[13, 15]. Both OSCP and the N-

termini of and subunits change their conformations during catalysis, and Bz-423 binding might

interfere with these changes and might retard the turnover. Subsequently the mitochondrial PMF

builds up, mitochondria switch from respirational state 3 to 4, superoxide is produced, and Ca2+

signaling and apoptosis are initiated. Bz-423 induced the opening of the mitochondrial permeability

transition pore and thus decreased the Ca2+ retention capability[36-38].

The weak binding affinities of Bz-423 and its fluorescent derivatives like Bz-423-Cy5 prevented direct

fluorescence microscopy approaches in living cells. In initial confocal microscopy experiments we

noticed a broad spatial distribution of Bz-423-Cy5 in S. cerevisiae cells but not a specific staining of the

mitochondria. Therefore we evaluated the use of sub-mitochondrial particles with fluorescently

tagged FoF1-ATP synthase for FRET in vitro but failed to detect sensitized FRET acceptor emission due

to the high fluorescent background of unbound Bz-423-Cy5. The solution for a FRET-based direct

detection of Bz-423 binding to OSCP was FRET acceptor photobleaching in mitochondria of living cells.

Here the pool of unbound Bz-423-Cy5 is limited, and photobleaching Cy5 with 640 nm at high power

is possible without destroying the FRET donor fluorophore yEGFP on FoF1-ATP synthase. Now, similar

FRET experiment are under way with human HEK cells and by using the brightest and more photostable

green fluorescent protein mNeonGreen[39] fused to the C-terminus of the subunit of FoF1-ATP

synthase[34].

To improve specificity of the cellular distribution of the hydrophobic Bz-423 derivatives and to

accelerate an uptake into mitochondria, attaching cationic dyes could be used. As shown by H. W.

Zimmermann and coworkers[40-43], almost all fluorophores being lipophilic cations transfer quickly

to the inner mitochondrial membrane. There, they can bind to proteins. One identified target was

cytochrome C oxidase. This finding could be used to specifically induce photodamage by singlet oxygen

as a reactive oxygen species[44-46] for Photodynamic Therapy. Beside photoaffinity labeling

approaches, time-resolved FRET was applied to reveal that cytochrome C oxidase was a binding site of

these lipophilic cationic photosensitizers acting as FRET donors[46, 47]. Similarly, confocal imaging

FRET donor lifetimes of mitochondrial FoF1-ATP synthases in the presence and absence of FRET

acceptor-tagged Bz-423 derivatives could be employed to provide direct optical evidence for Bz-423

binding to FoF1.

FRET-based direct evidence for Bz-423 acting at the mitochondrial FoF1-ATP synthase needs to be

complemented by detailed analysis of cristae morphology changes. In mitochondria of living cells, this

can be accomplished by superresolution microscopy approaches, for example structured illumination

microscopy[28] (SIM) or stimulated emission depletion[48] (STED) microscopy. Especially, a

rearrangement of row assembly of dimeric FoF1-ATP synthases at the rim of the cristae could indicate

the beginning and early events of apoptosis. Finally an unequivocal mechanistic demonstration of how

Bz-423 affects the catalytic activity and the rotary motors of mitochondrial FoF1-ATP synthase is

awaited. Single-molecule FRET between different fluorophores attached to rotor and stator of the

mitochondrial enzyme has already been developed. Accordingly, the inhibition mechanism of Bz-423

could be unraveled by single-molecule rotation experiments, as shown previously for the inhibitor

Aurovertin[49] or of phytopolyphenols[50].

Acknowledgements

We gratefully acknowledge the collaboration with Mark Prescott (Monash University), Jan Petersen

and Peter Gräber (Albert Ludwig University Freiburg) who designed and provided subunit mutants of

the mitochondrial enzyme of Saccharomyces cerevisiae. Kathryn M. Johnson synthesized the Cy5-

tagged Bz-423 (University of Michigan). This work was supported by NIH grant AI-47450 (to G.D.G.)

and in part by DFG grant BO 1891/15-1 (to M.B.). The N-SIM / N-STORM superresolution microscope

was funded by the State of Thuringa (grant FKZ 12026-515 to M.B.).

References

[1] N.B. Blatt, J.J. Bednarski, R.E. Warner, F. Leonetti, K.M. Johnson, A. Boitano, R. Yung, B.C. Richardson, K.J. Johnson, J.A. Ellman, A.W. Opipari, Jr., G.D. Glick, Benzodiazepine-induced superoxide signals B cell apoptosis: mechanistic insight and potential therapeutic utility, The Journal of clinical investigation, 110 (2002) 1123-1132. [2] J.J. Bednarski, R.E. Warner, T. Rao, F. Leonetti, R. Yung, B.C. Richardson, K.J. Johnson, J.A. Ellman, A.W. Opipari, Jr., G.D. Glick, Attenuation of autoimmune disease in Fas-deficient mice by treatment with a cytotoxic benzodiazepine, Arthritis and rheumatism, 48 (2003) 757-766. [3] K.M. Johnson, X. Chen, A. Boitano, L. Swenson, A.W. Opipari, Jr., G.D. Glick, Identification and validation of the mitochondrial F1F0-ATPase as the molecular target of the immunomodulatory benzodiazepine Bz-423, Chemistry & biology, 12 (2005) 485-496. [4] I. Starke, K. Johnson, J. Petersen, P. Graber, A.W. Opipari, G.D. Glick, M. Börsch, Binding of the immunomodulatory drug Bz-423 to mitochondrial FoF1-ATP synthase in living cells by FRET acceptor photobleaching, Proc. SPIE, 9712 (2016) 97120P. [5] T.M. Francis, T.B. Sundberg, J. Cleary, T. Groendyke, A.W. Opipari, Jr., G.D. Glick, Identification of cytotoxic, T-cell-selective 1,4-benzodiazepine-2,5-diones, Bioorg Med Chem Lett, 16 (2006) 2423-2427. [6] E. Gatza, D.R. Wahl, A.W. Opipari, T.B. Sundberg, P. Reddy, C. Liu, G.D. Glick, J.L. Ferrara, Manipulating the bioenergetics of alloreactive T cells causes their selective apoptosis and arrests graft-versus-host disease, Science translational medicine, 3 (2011) 67ra68. [7] D.R. Wahl, C.A. Byersdorfer, J.L. Ferrara, A.W. Opipari, Jr., G.D. Glick, Distinct metabolic programs in activated T cells: opportunities for selective immunomodulation, Immunological reviews, 249 (2012) 104-115. [8] V. Tkachev, S. Goodell, A.W. Opipari, L.Y. Hao, L. Franchi, G.D. Glick, J.L. Ferrara, C.A. Byersdorfer, Programmed death-1 controls T cell survival by regulating oxidative metabolism, Journal of immunology, 194 (2015) 5789-5800. [9] N.B. Blatt, A.E. Boitano, C.A. Lyssiotis, A.W. Opipari, Jr., G.D. Glick, Bz-423 superoxide signals apoptosis via selective activation of JNK, Bak, and Bax, Free radical biology & medicine, 45 (2008) 1232-1242. [10] N.B. Blatt, A.E. Boitano, C.A. Lyssiotis, A.W. Opipari, Jr., G.D. Glick, Bz-423 superoxide signals B cell apoptosis via Mcl-1, Bak, and Bax, Biochemical pharmacology, 78 (2009) 966-973. [11] S.S. Korshunov, V.P. Skulachev, A.A. Starkov, High protonic potential actuates a mechanism of production of reactive oxygen species in mitochondria, FEBS letters, 416 (1997) 15-18. [12] A.C. Stelzer, R.W. Frazee, C. Van Huis, J. Cleary, A.W. Opipari, Jr., G.D. Glick, H.M. Al-Hashimi, NMR studies of an immunomodulatory benzodiazepine binding to its molecular target on the mitochondrial F(1)F(0)-ATPase, Biopolymers, 93 (2010) 85-92. [13] A. Zhou, A. Rohou, D.G. Schep, J.V. Bason, M.G. Montgomery, J.E. Walker, N. Grigorieff, J.L. Rubinstein, Structure and conformational states of the bovine mitochondrial ATP synthase by cryo-EM, eLife, 4 (2015) e10180. [14] D.M. Rees, A.G. Leslie, J.E. Walker, The structure of the membrane extrinsic region of bovine ATP synthase, Proceedings of the National Academy of Sciences of the United States of America, 106 (2009) 21597-21601. [15] H. Guo, S.A. Bueler, J.L. Rubinstein, Atomic model for the dimeric FO region of mitochondrial ATP synthase, Science, 358 (2017) 936-940. [16] K.M. Johnson, J. Cleary, C.A. Fierke, A.W. Opipari, Jr., G.D. Glick, Mechanistic basis for therapeutic targeting of the mitochondrial F1F0-ATPase, ACS chemical biology, 1 (2006) 304-308. [17] M. Diez, B. Zimmermann, M. Borsch, M. Konig, E. Schweinberger, S. Steigmiller, R. Reuter, S. Felekyan, V. Kudryavtsev, C.A. Seidel, P. Graber, Proton-powered subunit rotation in single membrane-bound FoF1-ATP synthase, Nature structural & molecular biology, 11 (2004) 135-141. [18] B. Zimmermann, M. Diez, N. Zarrabi, P. Graber, M. Borsch, Movements of the epsilon-subunit during catalysis and activation in single membrane-bound H(+)-ATP synthase, The EMBO journal, 24 (2005) 2053-2063.

[19] M.G. Duser, N. Zarrabi, D.J. Cipriano, S. Ernst, G.D. Glick, S.D. Dunn, M. Borsch, 36 degrees step size of proton-driven c-ring rotation in FoF1-ATP synthase, The EMBO journal, 28 (2009) 2689-2696. [20] H. Noji, R. Yasuda, M. Yoshida, K. Kinosita, Jr., Direct observation of the rotation of F1-ATPase, Nature, 386 (1997) 299-302. [21] A. Wachter, Y. Bi, S.D. Dunn, B.D. Cain, H. Sielaff, F. Wintermann, S. Engelbrecht, W. Junge, Two rotary motors in F-ATP synthase are elastically coupled by a flexible rotor and a stiff stator stalk, Proceedings of the National Academy of Sciences of the United States of America, 108 (2011) 3924-3929. [22] H. Sielaff, M. Borsch, Twisting and subunit rotation in single FOF1-ATP synthase, Phil Trans R Soc B, 368 (2013) 20120024. [23] R. Yasuda, H. Noji, M. Yoshida, K. Kinosita, Jr., H. Itoh, Resolution of distinct rotational substeps by submillisecond kinetic analysis of F1-ATPase, Nature, 410 (2001) 898-904. [24] T. Suzuki, K. Tanaka, C. Wakabayashi, E.-i. Saita, M. Yoshida, Chemomechanical coupling of human mitochondrial F1-ATPase motor, Nature chemical biology, 10 (2014) 930-936. [25] M. Yoshida, E. Muneyuki, T. Hisabori, ATP synthase - a marvellous rotary engine of the cell, Nat Rev Mol Cell Biol, 2 (2001) 669-677. [26] W. Junge, H. Sielaff, S. Engelbrecht, Torque generation and elastic power transmission in the rotary FOF1-ATPase, Nature, 459 (2009) 364-370. [27] M. Borsch, T.M. Duncan, Spotlighting motors and controls of single FoF1-ATP synthase, Biochemical Society transactions, 41 (2013) 1219-1226. [28] L. Shao, P. Kner, E.H. Rego, M.G. Gustafsson, Super-resolution 3D microscopy of live whole cells using structured illumination, Nature methods, 8 (2011) 1044-1046. [29] S.W. Hell, Far-Field Optical Nanoscopy, Science, 316 (2007) 1153-1158. [30] T. Förster, Energiewanderung Und Fluoreszenz, Naturwissenschaften, 33 (1946) 166-175. [31] T. Förster, Energy migration and fluorescence, Journal of biomedical optics, 17 (2012) 011002. [32] M. Prescott, A. Lourbakos, M. Bateson, G. Boyle, P. Nagley, R.J. Devenish, A novel fluorescent marker for assembled mitochondria ATP synthase of yeast. OSCP subunit fused to green fluorescent protein is assembled into the complex in vivo, FEBS letters, 411 (1997) 97-101. [33] M. Prescott, S. Nowakowski, P. Gavin, P. Nagley, J.C. Whisstock, R.J. Devenish, Subunit gamma-green fluorescent protein fusions are functionally incorporated into mitochondrial F1F0-ATP synthase, arguing against a rigid cap structure at the top of F1, The Journal of biological chemistry, 278 (2003) 251-256. [34] F. Foertsch, M. Ilchenko, T. Heitkamp, S. Nossmann, B. Hoffmann, I. Starke, R. Mrowka, C. Biskup, M. Borsch, Imaging cytochrome C oxidase and F0F1-ATP synthase in mitochondrial cristae of living human cells by FLIM and superresolution microscopy, Proc. SPIE, 10071 (2017) 100710P. [35] B. Muster, W. Kohl, I. Wittig, V. Strecker, F. Joos, W. Haase, J. Bereiter-Hahn, K. Busch, Respiratory chain complexes in dynamic mitochondria display a patchy distribution in life cells, PloS one, 5 (2010) e11910. [36] V. Giorgio, S. von Stockum, M. Antoniel, A. Fabbro, F. Fogolari, M. Forte, G.D. Glick, V. Petronilli, M. Zoratti, I. Szabo, G. Lippe, P. Bernardi, Dimers of mitochondrial ATP synthase form the permeability transition pore, Proceedings of the National Academy of Sciences of the United States of America, 110 (2013) 5887-5892. [37] M. Antoniel, V. Giorgio, F. Fogolari, G.D. Glick, P. Bernardi, G. Lippe, The oligomycin-sensitivity conferring protein of mitochondrial ATP synthase: emerging new roles in mitochondrial pathophysiology, International journal of molecular sciences, 15 (2014) 7513-7536. [38] P. Bernardi, A. Rasola, M. Forte, G. Lippe, The Mitochondrial Permeability Transition Pore: Channel Formation by F-ATP Synthase, Integration in Signal Transduction, and Role in Pathophysiology, Physiological reviews, 95 (2015) 1111-1155. [39] N.C. Shaner, G.G. Lambert, A. Chammas, Y. Ni, P.J. Cranfill, M.A. Baird, B.R. Sell, J.R. Allen, R.N. Day, M. Israelsson, M.W. Davidson, J. Wang, A bright monomeric green fluorescent protein derived from Branchiostoma lanceolatum, Nature methods, 10 (2013) 407-409.

[40] G. Irion, L. Ochsenfeld, A. Naujok, H.W. Zimmermann, The concentration jump method. Kinetics of vital staining of mitochondria in HeLa cells with lipophilic cationic fluorescent dyes, Histochemistry, 99 (1993) 75-83. [41] J. Rottele, H.W. Zimmermann, Transport and accumulation of lipophilic dye cations at the mitochondria of HeLa cells in situ, Cellular and molecular biology (Noisy-le-Grand, France), 39 (1993) 739-756. [42] K. Schneider, A. Naujok, H.W. Zimmermann, Influence of trans-membrane potential and of hydrophobic interactions on dye accumulation in mitochondria of living cells. Photoaffinity labelling of mitochondrial proteins, action of potential dissipating drugs, and competitive staining, Histochemistry, 101 (1994) 455-461. [43] K. Schneider, H.W. Zimmermann, Photoaffinity labelling with fluorescence detection. Dye accumulation at four mitochondrial proteins in HeLa and LM cells, Histochemistry, 101 (1994) 135-142. [44] T. Cernay, H.W. Zimmermann, Selective photosensitization of mitochondria by the lipophilic cationic porphyrin POR10, Journal of photochemistry and photobiology, 34 (1996) 191-196. [45] H. Dummin, T. Cernay, H.W. Zimmermann, Selective photosensitization of mitochondria in HeLa cells by cationic Zn (II) phthalocyanines with lipophilic side-chains, Journal of photochemistry and photobiology, 37 (1997) 219-229. [46] M. Borsch, Targeting cytochrome C oxidase in mitochondria with Pt(II)-porphyrins for photodynamic therapy, Proc. SPIE, 7551 (2010) 75510G. [47] D. Huglin, W. Seiffert, H.W. Zimmermann, Time-resolved microfluorometric study of the binding sites of lipophilic cationic pyrene probes in mitochondria of living HeLa cells, Journal of photochemistry and photobiology, 31 (1995) 145-158. [48] R. Schmidt, C.A. Wurm, A. Punge, A. Egner, S. Jakobs, S.W. Hell, Mitochondrial cristae revealed with focused light, Nano letters, 9 (2009) 2508-2510. [49] K.M. Johnson, L. Swenson, A.W. Opipari, Jr., R. Reuter, N. Zarrabi, C.A. Fierke, M. Borsch, G.D. Glick, Mechanistic basis for differential inhibition of the F1Fo-ATPase by aurovertin, Biopolymers, 91 (2009) 830-840. [50] M. Sekiya, Y. Sakamoto, M. Futai, M. Nakanishi-Matsui, Role of alpha/beta interface in F1 ATPase rotational catalysis probed by inhibitors and mutations, International journal of biological macromolecules, 99 (2017) 615-621.