Conformational ensemble of yeast ATP synthase at low pH reveals unique intermediates and plasticity in F1-Fo coupling.

Mitochondrial adenosine triphosphate (ATP) synthase uses the proton gradient across the inner mitochondrial membrane to synthesize ATP. Structural and single molecule studies conducted mostly at neutral or basic pH have provided details of the reaction mechanism of ATP synthesis. However, pH of the mitochondrial matrix is slightly acidic during hypoxia and pH-dependent conformational changes in the ATP synthase have been reported. Here we use single-particle cryo-EM to analyze the conformational ensemble of the yeast (Saccharomyces cerevisiae) ATP synthase at pH 6. Of the four conformations resolved in this study, three are reaction intermediates. In addition to canonical catalytic dwell and binding dwell structures, we identify two unique conformations with nearly identical positions of the central rotor but different catalytic site conformations. These structures provide new insights into the catalytic mechanism of the ATP synthase and highlight elastic coupling between the catalytic and proton translocating domains.

H+-slip correlated to rotor free-wheeling as cause of F1FO-ATPase dysfunction in primary mitochondrial disorders.

Inborn errors of metabolism are related to mitochondrial disorders caused by dysfunction of the oxidative phosphorylation (OXPHOS) system. Congenital hypermetabolism in the infant is a rare disease belonging to Luft syndrome, nonthyroidal hypermetabolism, arising from a singular example of a defect in OXPHOS. The mitochondria lose coupling of mitochondrial substrates oxidation from the ADP phosphorylation. Since Luft syndrome is due to uncoupled cell respiration responsible for deficient in ATP production that originates in the respiratory complexes, a de novo heterozygous variant in the catalytic subunit of mitochondrial F1FO-ATPase arises as the main cause of an autosomal dominant syndrome of hypermetabolism associated with dysfunction in ATP production, which does not involve the respiratory complexes. The F1FO-ATPase works as an embedded molecular machine with a rotary action using two different motor engines. The FO, which is an integral domain in the membrane, dissipates the chemical potential difference for H+, a proton motive force (Δp), across the inner membrane to generate a torsion. The F1 domain-the hydrophilic portion responsible for ATP turnover-is powered by the molecular rotary action to synthesize ATP. The structural and functional coupling of F1 and FO domains support the energy transduction for ATP synthesis. The dissipation of Δp by means of an H+ slip correlated to rotor free-wheeling of the F1FO-ATPase has been discovered to cause enzyme dysfunction in primary mitochondrial disorders. In this insight, we try to offer commentary and analysis of the molecular mechanism in these impaired mitochondria.

Expanding the squaramide library as mycobacterial ATP synthase inhibitors: Innovative synthetic pathway and biological evaluation.

Mycobacterial ATP synthase is a validated therapeutic target for combating drug-resistant tuberculosis. Inhibition of this enzyme has been featured as an efficient strategy for the development of new antimycobacterial agents against drug-resistant pathogens. In this study, we synthesised and explored two distinct series of squaric acid analogues designed to inhibit mycobacterial ATP synthase. Among the extensive array of compounds investigated, members of the phenyl-substituted sub-library emerged as primary hits. To gain deeper insights into their mechanisms of action, we conducted advanced biological studies, focusing on the compounds displaying a direct binding of a nitrogen heteroatom to the phenyl ring, resulting in the highest potency. Our investigations into spontaneous mutants led to the validation of a single point mutation within the atpB gene (Rv1304), responsible for encoding the ATP synthase subunit a. This genetic alteration sheds light on the molecular basis of resistance to squaramides. Furthermore, we explored the possibility of synergy between squaramides and the reference drug clofazimine using a checkerboard assay, highlighting the promising avenue for enhancing the effectiveness of existing treatments through combined therapeutic approaches. This study contributes to the expansion of investigating squaramides as promising drug candidates in the ongoing battle against drug-resistant tuberculosis.

Bacterial and mammalian F1FO-ATPase: Structural similarities and divergences to exploit in the battle against Mycobacterium tuberculosis.

The inner mitochondrial membrane, thylakoid membrane of chloroplasts and bacterial plasma membrane play a central role in energy transduction processes exploiting a ubiquitous membrane-bound enzyme complex known as F1FO-ATPase. The enzyme maintains the same function of ATP production between the species and a basic molecular mechanism of enzymatic catalysis during ATP synthesis/hydrolysis. However, small structural divergences distinguish prokaryotic ATP synthases, embedded in cell membranes, from eukaryotic ones localized in the inner mitochondrial membrane designating the bacterial enzyme as drug targets. In antimicrobial drug design, the membrane-embedded c-ring of the enzyme becomes the key protein of candidate compounds, such as diarylquinolines in tuberculosis, that inhibit the mycobacteria F1FO-ATPase without affecting mammalian homologs. The drug known as bedaquiline can target uniquely the structure of the mycobacterial c-ring. This specific interaction could address at the molecular level the therapy of infections sustained by antibiotic-resistant microorganisms.

The Ancestral Shape of the Access Proton Path of Mitochondrial ATP Synthases Revealed by a Split Subunit-a.

The passage of protons across membranes through F1Fo-ATP synthases spins their rotors and drives the synthesis of ATP. While the principle of torque generation by proton transfer is known, the mechanisms and routes of proton access and release and their evolution are not fully understood. Here, we show that the entry site and path of protons in the lumenal half channel of mitochondrial ATP synthases are largely defined by a short N-terminal α-helix of subunit-a. In Trypanosoma brucei and other Euglenozoa, the α-helix is part of another polypeptide chain that is a product of subunit-a gene fragmentation. This α-helix and other elements forming the proton pathway are widely conserved across eukaryotes and in Alphaproteobacteria, the closest extant relatives of mitochondria, but not in other bacteria. The α-helix blocks one of two proton routes found in Escherichia coli, resulting in a single proton entry site in mitochondrial and alphaproteobacterial ATP synthases. Thus, the shape of the access half channel predates eukaryotes and originated in the lineage from which mitochondria evolved by endosymbiosis.

Structure of ATP synthase under strain during catalysis.

ATP synthases are macromolecular machines consisting of an ATP-hydrolysis-driven F1 motor and a proton-translocation-driven FO motor. The F1 and FO motors oppose each other's action on a shared rotor subcomplex and are held stationary relative to each other by a peripheral stalk. Structures of resting mitochondrial ATP synthases revealed a left-handed curvature of the peripheral stalk even though rotation of the rotor, driven by either ATP hydrolysis in F1 or proton translocation through FO, would apply a right-handed bending force to the stalk. We used cryoEM to image yeast mitochondrial ATP synthase under strain during ATP-hydrolysis-driven rotary catalysis, revealing a large deformation of the peripheral stalk. The structures show how the peripheral stalk opposes the bending force and suggests that during ATP synthesis proton translocation causes accumulation of strain in the stalk, which relaxes by driving the relative rotation of the rotor through six sub-steps within F1, leading to catalysis.

What happens when the mitochondrial H+-translocating F1FO-ATP(hydrol)ase becomes a molecular target of calcium? The pore opens.

The F1FO-ATPase has Mg2+ cofactor as the natural divalent cation to support the bifunctional activity of ATP synthesis and hydrolysis. Different physio(patho)logical conditions permit the molecular interaction of Ca2+ with the enzyme and the modification of the biological role. Three distinct binding regions of Ca2+ have been localized on the enzyme complex: one in the F1 catalytic sites and the other two sites in the membrane-embedded domain FO. In all likelihood, Ca2+-activated enzyme most frequently works as an H+-translocating F1FO-ATP(hydrol)ase with a monofunctional activity that triggers the formation of mitochondrial permeability transition pore (mPTP) phenomenon. The protein(s) component of the mPTP is considered an arcane mystery. However, the F1FO-ATPase could reveal the molecular mechanism of pore opening when Ca2+ is bound to the enzyme. In this regard, the role of Ca2+-dependent function of the F1FO-ATPase in the formation of the mPTP is discussed.

Defective dimerization of FoF1-ATP synthase secondary to glycation favors mitochondrial energy deficiency in cardiomyocytes during aging.

Aged cardiomyocytes develop a mismatch between energy demand and supply, the severity of which determines the onset of heart failure, and become prone to undergo cell death. The FoF1-ATP synthase is the molecular machine that provides >90% of the ATP consumed by healthy cardiomyocytes and is proposed to form the mitochondrial permeability transition pore (mPTP), an energy-dissipating channel involved in cell death. We investigated whether aging alters FoF1-ATP synthase self-assembly, a fundamental biological process involved in mitochondrial cristae morphology and energy efficiency, and the functional consequences this may have. Purified heart mitochondria and cardiomyocytes from aging mice displayed an impaired dimerization of FoF1-ATP synthase (blue native and proximity ligation assay), associated with abnormal mitochondrial cristae tip curvature (TEM). Defective dimerization did not modify the in vitro hydrolase activity of FoF1-ATP synthase but reduced the efficiency of oxidative phosphorylation in intact mitochondria (in which membrane architecture plays a fundamental role) and increased cardiomyocytes' susceptibility to undergo energy collapse by mPTP. High throughput proteomics and fluorescence immunolabeling identified glycation of 5 subunits of FoF1-ATP synthase as the causative mechanism of the altered dimerization. In vitro induction of FoF1-ATP synthase glycation in H9c2 myoblasts recapitulated the age-related defective FoF1-ATP synthase assembly, reduced the relative contribution of oxidative phosphorylation to cell energy metabolism, and increased mPTP susceptibility. These results identify altered dimerization of FoF1-ATP synthase secondary to enzyme glycation as a novel pathophysiological mechanism involved in mitochondrial cristae remodeling, energy deficiency, and increased vulnerability of cardiomyocytes to undergo mitochondrial failure during aging.

ATP Synthase K+- and H+-Fluxes Drive ATP Synthesis and Enable Mitochondrial K+-"Uniporter" Function: I. Characterization of Ion Fluxes.

ATP synthase (F1Fo) synthesizes daily our body's weight in ATP, whose production-rate can be transiently increased several-fold to meet changes in energy utilization. Using purified mammalian F1Fo-reconstituted proteoliposomes and isolated mitochondria, we show F1Fo can utilize both ΔΨm-driven H+- and K+-transport to synthesize ATP under physiological pH = 7.2 and K+ = 140 mEq/L conditions. Purely K+-driven ATP synthesis from single F1Fo molecules measured by bioluminescence photon detection could be directly demonstrated along with simultaneous measurements of unitary K+ currents by voltage clamp, both blocked by specific Fo inhibitors. In the presence of K+, compared to osmotically-matched conditions in which this cation is absent, isolated mitochondria display 3.5-fold higher rates of ATP synthesis, at the expense of 2.6-fold higher rates of oxygen consumption, these fluxes being driven by a 2.7:1 K+: H+ stoichiometry. The excellent agreement between the functional data obtained from purified F1Fo single molecule experiments and ATP synthase studied in the intact mitochondrion under unaltered OxPhos coupling by K+ presence, is entirely consistent with K+ transport through the ATP synthase driving the observed increase in ATP synthesis. Thus, both K+ (harnessing ΔΨm) and H+ (harnessing its chemical potential energy, ΔµH) drive ATP generation during normal physiology.

Apoptolidin family glycomacrolides target leukemia through inhibition of ATP synthase.

Cancer cells have long been recognized to exhibit unique bioenergetic requirements. The apoptolidin family of glycomacrolides are distinguished by their selective cytotoxicity towards oncogene-transformed cells, yet their molecular mechanism remains uncertain. We used photoaffinity analogs of the apoptolidins to identify the F1 subcomplex of mitochondrial ATP synthase as the target of apoptolidin A. Cryogenic electron microscopy (cryo-EM) of apoptolidin and ammocidin-ATP synthase complexes revealed a novel shared mode of inhibition that was confirmed by deep mutational scanning of the binding interface to reveal resistance mutations which were confirmed using CRISPR-Cas9. Ammocidin A was found to suppress leukemia progression in vivo at doses that were tolerated with minimal toxicity. The combination of cellular, structural, mutagenesis, and in vivo evidence defines the mechanism of action of apoptolidin family glycomacrolides and establishes a path to address oxidative phosphorylation-dependent cancers.

Structure of the ATP synthase from Mycobacterium smegmatis provides targets for treating tuberculosis.

The structure has been determined by electron cryomicroscopy of the adenosine triphosphate (ATP) synthase from Mycobacterium smegmatis This analysis confirms features in a prior description of the structure of the enzyme, but it also describes other highly significant attributes not recognized before that are crucial for understanding the mechanism and regulation of the mycobacterial enzyme. First, we resolved not only the three main states in the catalytic cycle described before but also eight substates that portray structural and mechanistic changes occurring during a 360° catalytic cycle. Second, a mechanism of auto-inhibition of ATP hydrolysis involves not only the engagement of the C-terminal region of an α-subunit in a loop in the γ-subunit, as proposed before, but also a "fail-safe" mechanism involving the b'-subunit in the peripheral stalk that enhances engagement. A third unreported characteristic is that the fused bδ-subunit contains a duplicated domain in its N-terminal region where the two copies of the domain participate in similar modes of attachment of the two of three N-terminal regions of the α-subunits. The auto-inhibitory plus the associated "fail-safe" mechanisms and the modes of attachment of the α-subunits provide targets for development of innovative antitubercular drugs. The structure also provides support for an observation made in the bovine ATP synthase that the transmembrane proton-motive force that provides the energy to drive the rotary mechanism is delivered directly and tangentially to the rotor via a Grotthuss water chain in a polar L-shaped tunnel.

A recurrent de novo ATP5F1A substitution associated with neonatal complex V deficiency.

Mitochondrial disorders are a heterogeneous group of rare, degenerative multisystem disorders affecting the cell's core bioenergetic and signalling functions. Spontaneous improvement is rare. We describe a novel neonatal-onset mitochondriopathy in three infants with failure to thrive, hyperlactatemia, hyperammonemia, and apparent clinical resolution before 18 months. Exome sequencing showed all three probands to be identically heterozygous for a recurrent de novo substitution, c.620G>A [p.(Arg207His)] in ATP5F1A, encoding the α-subunit of complex V. Patient-derived fibroblasts exhibited multiple deficits in complex V function and expression in vitro. Structural modelling predicts the observed substitution to create an abnormal region of negative charge on ATP5F1A's ß-subunit-interacting surface, adjacent to the nearby ß subunit's active site. This disorder, which presents with life-threatening neonatal manifestations, appears to follow a remitting course; the long-term prognosis remains unknown.

The inhibition of gadolinium ion (Gd3+) on the mitochondrial F1FO-ATPase is linked to the modulation of the mitochondrial permeability transition pore.

The mitochondrial permeability transition pore (PTP), which drives regulated cell death when Ca2+ concentration suddenly increases in mitochondria, was related to changes in the Ca2+-activated F1FO-ATPase. The effects of the gadolinium cation (Gd3+), widely used for diagnosis and therapy, and reported as PTP blocker, were evaluated on the F1FO-ATPase activated by Mg2+ or Ca2+ and on the PTP. Gd3+ more effectively inhibits the Ca2+-activated F1FO-ATPase than the Mg2+-activated F1FO-ATPase by a mixed-type inhibition on the former and by uncompetitive mechanism on the latter. Most likely Gd3+ binding to F1, is favoured by Ca2+ insertion. The maximal inactivation rates (kinact) of pseudo-first order inactivation are similar either when the F1FO-ATPase is activated by Ca2+ or by Mg2+. The half-maximal inactivator concentrations (KI) are 2.35 ± 0.35 mM and 0.72 ± 0.11 mM, respectively. The potency of a mechanism-based inhibitor (kinact/KI) also highlights a higher inhibition efficiency of Gd3+ on the Ca2+-activated F1FO-ATPase (0.59 ± 0.09 mM-1∙s-1) than on the Mg2+-activated F1FO-ATPase (0.13 ± 0.02 mM-1∙s-1). Consistently, the PTP is desensitized in presence of Gd3+. The Gd3+ inhibition on both the mitochondrial Ca2+-activated F1FO-ATPase and the PTP strengthens the link between the PTP and the F1FO-ATPase when activated by Ca2+ and provides insights on the biological effects of Gd3+.

Multiplexed complexome profiling using tandem mass tags.

Complexome profiling is a rapidly spreading, powerful technique to gain insight into the nature of protein complexes. It identifies and quantifies protein complexes separated into multiple fractions of increasing molecular mass using mass spectrometry-based, label-free bottom-up proteomics. Complexome profiling enables a sophisticated and thorough characterization of the composition, molecular mass, assembly, and interactions of protein complexes. However, in practice, its application is limited by the large number of samples it generates and the related time of mass spectrometry analyses. Here, we report an improved process workflow that implements tandem mass tags for multiplexing complexome profiling. This workflow substantially reduces the number of samples and measuring time without compromising protein identification or quantification reliability. In profiles from mitochondrial fractions of cells recovering from chloramphenicol treatment, tandem mass tags-multiplexed complexome profiling exhibited migration patterns of mature ATP synthase (complex V) and assembly intermediates that were consistent in composition and abundance with profiles obtained by the label-free approach. Reporter ion quantifications of proteins and complexes unaffected by the chloramphenicol treatment presented less variation in comparison to the label-free method. Incorporation of tandem mass tags enabled an efficient and robust complexome profiling analysis and may foster broader application for protein complex profiling in biomedical research and diagnostics.

C subunit of the ATP synthase is an amyloidogenic calcium dependent channel-forming peptide with possible implications in mitochondrial permeability transition.

The c subunit is an inner mitochondrial membrane (IMM) protein encoded by three nuclear genes. Best known as an integral part of the F0 complex of the ATP synthase, the c subunit is also present in other cytoplasmic compartments in ceroid lipofuscinoses. Under physiological conditions, this 75 residue-long peptide folds into an α-helical hairpin and forms oligomers spanning the lipid bilayer. In addition to its physiological role, the c subunit has been proposed as a key participant in stress-induced IMM permeabilization by the mechanism of calcium-induced permeability transition. However, the molecular mechanism of the c subunit participation in IMM permeabilization is not completely understood. Here we used fluorescence spectroscopy, atomic force microscopy and black lipid membrane methods to gain insights into the structural and functional properties of unmodified c subunit protein that might make it relevant to mitochondrial toxicity. We discovered that c subunit is an amyloidogenic peptide that can spontaneously fold into ß-sheets and self-assemble into fibrils and oligomers in a Ca2+-dependent manner. C subunit oligomers exhibited ion channel activity in lipid membranes. We propose that the toxic effects of c subunit might be linked to its amyloidogenic properties and are driven by mechanisms similar to those of neurodegenerative polypeptides such as Aß and α-synuclein.

Deletion of the natural inhibitory protein Inh1 in Ustilago maydis has no effect on the dimeric state of the F1FO-ATP synthase but increases the ATPase activity and reduces the stability.

Transduction of electrochemical proton gradient into ATP synthesis is performed by F1FO-ATP synthase. The reverse reaction is prevented by the regulatory subunit Inh1. Knockout of the inh1 gene in the basidiomycete Ustilago maydis was generated in order to study the function of this protein in the mitochondrial metabolism and cristae architecture. Deletion of inh1 gen did not affect cell growth, glucose consumption, and biomass production. Ultrastructure and fluorescence analyzes showed that size, cristae shape, network, and distribution of mitochondria was similar to wild strain. Membrane potential, ATP synthesis, and oxygen consumption in wild type and mutant strains had similar values. Kinetic analysis of ATPase activity of complex V in permeabilized mitochondria showed similar values of Vmax and KM for both strains, and no effect of pH was observed. Interestingly, the dimeric state of complex V occurs in the mutant strain, indicating that this subunit is not essential for dimerization. ATPase activity of the isolated monomeric and dimeric forms of complex V indicated Vmax values 4-times higher for the mutant strain than for the WT strain, suggesting that the absence of Inh1 subunit increased ATPase activity, and supporting a regulatory role for this protein; however, no effect of pH was observed. ATPase activity of WT oligomers was stimulated several times by dodecyl-maltoside (DDM), probably by removal of ADP from F1 sector, while DDM induced an inactive form of the mutant oligomers.

The F1Fo-ATPase inhibitor, IF1, is a critical regulator of energy metabolism in cancer cells.

In the last two decades, IF1, the endogenous inhibitor of the mitochondrial F1Fo-ATPase (ATP synthase) has assumed greater and ever greater interest since it has been found to be overexpressed in many cancers. At present, several findings indicate that IF1 is capable of playing a central role in cancer cells by promoting metabolic reprogramming, proliferation and resistance to cell death. However, the mechanism(s) at the basis of this pro-oncogenic action of IF1 remains elusive. Here, we recall the main features of the mechanism of the action of IF1 when the ATP synthase works in reverse, and discuss the experimental evidence that support its relevance in cancer cells. In particular, a clear pro-oncogenic action of IF1 is to avoid wasting of ATP when cancer cells are exposed to anoxia or near anoxia conditions, therefore favoring cell survival and tumor growth. However, more recently, various papers have described IF1 as an inhibitor of the ATP synthase when it is working physiologically (i.e. synthethizing ATP), and therefore reprogramming cell metabolism to aerobic glycolysis. In contrast, other studies excluded IF1 as an inhibitor of ATP synthase under normoxia, providing the basis for a hot debate. This review focuses on the role of IF1 as a modulator of the ATP synthase in normoxic cancer cells with the awareness that the knowledge of the molecular action of IF1 on the ATP synthase is crucial in unravelling the molecular mechanism(s) responsible for the pro-oncogenic role of IF1 in cancer and in developing related anticancer strategies.

TMEM70 and TMEM242 help to assemble the rotor ring of human ATP synthase and interact with assembly factors for complex I.

Human mitochondrial ATP synthase is a molecular machine with a rotary action bound in the inner organellar membranes. Turning of the rotor, driven by a proton motive force, provides energy to make ATP from ADP and phosphate. Among the 29 component proteins of 18 kinds, ATP6 and ATP8 are mitochondrial gene products, and the rest are nuclear gene products that are imported into the organelle. The ATP synthase is assembled from them via intermediate modules representing the main structural elements of the enzyme. One such module is the c8-ring, which provides the membrane sector of the enzyme's rotor, and its assembly is influenced by another transmembrane (TMEM) protein, TMEM70. We have shown that subunit c interacts with TMEM70 and another hitherto unidentified mitochondrial transmembrane protein, TMEM242. Deletion of TMEM242, similar to deletion of TMEM70, affects but does not completely eliminate the assembly of ATP synthase, and to a lesser degree the assembly of respiratory enzyme complexes I, III, and IV. Deletion of TMEM70 and TMEM242 together prevents assembly of ATP synthase and the impact on complex I is enhanced. Removal of TMEM242, but not of TMEM70, also affects the introduction of subunits ATP6, ATP8, j, and k into the enzyme. TMEM70 and TMEM242 interact with the mitochondrial complex I assembly (the MCIA) complex that supports assembly of the membrane arm of complex I. The interactions of TMEM70 and TMEM242 with MCIA could be part of either the assembly of ATP synthase and complex I or the regulation of their levels.

Bisindolylpyrrole Induces a Cpr3- and Porin1/2-Dependent Transition in Yeast Mitochondrial Permeability in a Low Conductance State via the AACs-Associated Pore.

Although the mitochondrial permeability transition pore (PTP) is presumably formed by either ATP synthase or the ATP/ADP carrier (AAC), little is known about their differential roles in PTP activation. We explored the role of AAC and ATP synthase in PTP formation in Saccharomyces cerevisiae using bisindolylpyrrole (BP), an activator of the mammalian PTP. The yeast mitochondrial membrane potential, as indicated by tetramethylrhodamine methyl ester signals, dissipated over 2-4 h after treatment of cells with 5 µM BP, which was sensitive to cyclosporin A (CsA) and Cpr3 deficiency and blocked by porin1/2 deficiency. The BP-induced depolarization was inhibited by a specific AAC inhibitor, bongkrekate, and consistently blocked in a yeast strain lacking all three AACs, while it was not affected in the strain with defective ATP synthase dimerization, suggesting the involvement of an AAC-associated pore. Upon BP treatment, isolated yeast mitochondria underwent CsA- and bongkrekate-sensitive depolarization without affecting the mitochondrial calcein signals, indicating the induction of a low conductance channel. These data suggest that, upon BP treatment, yeast can form a porin1/2- and Cpr3-regulated PTP, which is mediated by AACs but not by ATP synthase dimers. This implies that yeast may be an excellent tool for the screening of PTP modulators.

Interface mobility between monomers in dimeric bovine ATP synthase participates in the ultrastructure of inner mitochondrial membranes.

The ATP synthase complexes in mitochondria make the ATP required to sustain life by a rotary mechanism. Their membrane domains are embedded in the inner membranes of the organelle, and they dimerize via interactions between their membrane domains. The dimers form extensive chains along the tips of the cristae with the two rows of monomeric catalytic domains extending into the mitochondrial matrix at an angle to each other. Disruption of the interface between dimers by mutation affects the morphology of the cristae severely. By analysis of particles of purified dimeric bovine ATP synthase by cryo-electron microscopy, we have shown that the angle between the central rotatory axes of the monomeric complexes varies between ca. 76 and 95°. These particles represent active dimeric ATP synthase. Some angular variations arise directly from the catalytic mechanism of the enzyme, and others are independent of catalysis. The monomer-monomer interaction is mediated mainly by j subunits attached to the surface of wedge-shaped protein-lipid structures in the membrane domain of the complex, and the angular variation arises from rotational and translational changes in this interaction, and combinations of both. The structures also suggest how the dimeric ATP synthases might be interacting with each other to form the characteristic rows along the tips of the cristae via other interwedge contacts, molding themselves to the range of oligomeric arrangements observed by tomography of mitochondrial membranes, and at the same time allowing the ATP synthase to operate under the range of physiological conditions that influence the structure of the cristae.