CRYO-em structure of the brine shrimp ATP synthase and an inactivation mechanism of ATP synthase leak channel for use in therapeutic approaches

Abstract

Provided are isolated, recombinant, and modified recombinant adenosine triphosphate (ATP) synthase and fragments thereof, and methods of using the ATP synthase fragments. The ATP synthases and fragments thereof are useful for treating a number of conditions. The ATP synthase fragments may be derived from A. franciscana or human ATP synthases.

Description

[0001]Attorney Docket No.: 074339.00320 CRYO-EM STRUCTURE OF THE BRINE SHRIMP ATP SYNTHASE AND AN INACTIVATION MECHANISM OF ATP SYNTHASE LEAK CHANNEL FOR USE IN THERAPEUTIC APPROACHES CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. provisional patent application no.63 / 631,765, filed April 9, 2024, and to U.S. provisional patent application no.63 / 752,309, filed January 31, 2025, the entire disclosures of each of which are incorporated herein by reference. SEQUENCE LISTING The instant application contains a Sequence Listing which has been submitted electronically in XML format and is hereby incorporated by reference in its entirety. Said XML copy, created on March 31, 2025, is named “074339_00320_ST26.xml”, and is 36,684 bytes in size. STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH This invention was made with government support under Grant No. R21NS137275 awarded by the National Institutes of Health. The Government has certain rights in the invention. BACKGROUND The mitochondrial permeability transition (mPT) is a phenomenon discovered in mammals several decades ago1, 2, 3, 4. It is mediated by the opening of the Ca2+-regulated ion channel in the inner mitochondrial membrane, called the mitochondrial permeability transition pore (mPTP) or mitochondrial megachannel (MMC). mPTP activation allows the flux of ions and molecules with a molecular weight of up to 1.5 kDa in and out of the mitochondrial matrix. It can cause transient or prolonged depolarization of mitochondrial membrane potential3, 4. The brief or transient openings of mPTP serve its physiological function to regulate embryonic development, calcium handling, and metabolism5, 6. In contrast, prolonged openings induce irreversible changes in mitochondrial morphology and function, leading to cell death. mPTP is reported to play a critical role in different pathologies, such as ischemia / reperfusion injury of the heart and brain, Alzheimer's and Parkinson's diseases, Amyotrophic Lateral Sclerosis (ALS)7, 8, 9, 10. Despite extensive research, the exact molecular composition of mPTP and its gating has previously remained elusive. Over the years, different mitochondrial proteins, including the voltage-dependent anion channel (VDAC)11, 12, the phosphate carrier (PiC)13, the translocator protein (TSPO)14, 15, 16, and the adenine nucleotide translocator (ANT)17were reported to be the structural component of mPTP. Nevertheless, genetic ablation of these proteins revealed their role as regulators of the pore rather than the channel-forming elements of mPTP18, 19, 20, 21. Patch-clamp recordings of the mitoplasts (inner mitochondrial membrane preparations) were used for the biophysical characterization of mPTP. The mPTP forms voltage-gated, Ca2+and cyclophilin D (CypD)-regulated, non-selective high-conductance channels with 1.5 nS peak conductance activity. Bos taurus (bovine) and Sus scrofa (porcine) ATP synthase dimers22, 23, 24, 25, 26and monomers27, 28, 29, 30, 31, 32, 33were reported to have similar properties to mPTP in several recent studies22, 27, 29, 32, 33, 34, 35. The ATP synthase c- subunit ring was suggested to comprise the structural pore-forming component of mPTP27, 28,29, 30, 31, 32, 33, 35. The channel formed by the c-ring, termed ATP synthase c-subunit leak channel (ACLC), is a multi-conductance, non-selective, voltage-gated, Ca2+-sensitive channel with 1.5 nS peak conductance activity32, 35. The Ca2+-binding sites were found at the β- subunits of ATP synthase, and application of Ca2+during patch-clamp recordings of mammalian ATP synthase activated the channel34. The application of purified CypD was also reported to activate the channel during electrophysiology recordings of ATP synthase, while its binding partner, cyclosporin A (CsA), inhibited it32. Purified human ATP synthase c-subunit ring alone formed high-conductance channels during electrophysiology recordings32, 35. Its activity was inhibited upon the addition of ATP synthase F1domain, suggesting F1constitutes the inactivation gate and regulator of the channel from the matrix side35. Mitochondria isolated from the ATP synthase c-subunit knockout HAP1-A12 cells were shown to lack the 1.5 nS high conductance activity, as compared to WT, during patch-clamp experiments28. A low conductance channel activity, sensitive to ANT inhibitor bongkrekic acid (BA), was still present in these knockout cells, suggesting the possible contribution of multiple pores to mPT28. Liver mitochondria from mice with triple deletion of Ant1, Ant2, and Ant4 still undergo a CsA-sensitive mPT at increased matrix Ca2+overload36. The deletion of Ppif gene encoding CypD in these mice completely prevented Ca2+-induced mPT, suggesting a new model containing two distinct molecular components, ANT and another CypD-regulated channel, that mediate mPTP formation36. Several other studies reported the potential existence of a transient, “non- conventional” form of mPT in mammals insensitive to CsA and not regulated by CypD37, 38,39. Thus, the exact molecular nature of mPTP in mammals is still being debated and remains to be elucidated. In contrast to mammals, the embryos of the crustacean brine shrimp A. franciscana maintain viability for several years under anoxic diapause conditions without undergoing mPT40, 41, 42, 43. The ability of A. franciscana to tolerate environmental insults can be explained by the lack of Ca2+-regulated mPTP in this organism. No swelling of the mitochondrial matrix, rupture of the outer membrane, or cytochrome c release were observed in A. franciscana mitochondria upon profound calcium storage40. Yet, the underlying molecular mechanism of mPTP inhibition in this crustacean is also not fully understood. There is an ongoing need to identify ATP synthases and fragments thereof that can be used for various therapeutic approaches. The present disclosure is pertinent to this need. BRIEF SUMMARY Mammalian mitochondria undergo Ca2+-induced and cyclosporinA (CsA)-regulated permeability transition (mPT) by activating the mitochondrial permeability transition pore (mPTP) situated in mitochondrial inner membranes. Ca2+-induced prolonged openings of mPTP under certain pathological conditions result in mitochondrial swelling and rupture of the outer membrane, leading to mitochondrial dysfunction and cell death. While the exact molecular composition and structure of mPTP remain unknown, mammalian ATP synthase was reported to form voltage and Ca2+-activated leak channels involved in mPT. Unlike in mammals, mitochondria of the crustacean Artemia franciscana have the ability to accumulate large amounts of Ca2+without undergoing the mPT. This disclosure provides results related to structural and functional analysis of A. franciscana ATP synthase and its molecular mechanism of mPTP inhibition in this organism. The disclosure shows that the channel formed by the A. franciscana ATP synthase dwells predominantly in its inactive state and is insensitive to Ca2+, in contrast to porcine heart ATP synthase. Single-particle cryo-electron microscopy (cryo-EM) analysis revealed distinct structural features in A. franciscana ATP synthase compared with mammals. The stronger density of the e-subunit C-terminal region and its enhanced interaction with the c-ring was found in A. franciscana ATP synthase. These data suggest an inactivation mechanism of the ATP synthase leak channel and its possible contribution to the lack of mPT in this organism. In view of the figures and description of this disclosure, it will be recognized that the ATP synthase c-subunit leak channel (ACLC) has an important role in inducing cell death in ischemia-reperfusion injury and degenerative diseases of the heart and brain, making it an irreplaceable therapeutic target. Nevertheless, very little was previously known about the gating mechanism of ACLC. Currently, no FDA-approved drugs target the ATP synthase leak channel. The disclosure is based in part on cryo-electron microscopy (cryo-EM) analysis of the brine shrimp Artemia franciscana ATP synthase. The embryos of Artemia franciscana survive anoxic (low oxygen) conditions for years, unlike human neurons and cardiomyocytes that die within a few minutes of low oxygen conditions. The described structural studies revealed a structural unit in Artemia franciscana ATP synthase that keeps the channel in its inhibited form in this organism. This structural unit is missing in human ATP synthase. The disclosure thus provides Artemia franciscana ATP synthase and fragments thereof that are designed to function as a peptide to be used as a drug for preventing ACLC-induced cell death in a wide range of pathological conditions in humans. Data presented in this disclosure shows that designed peptides from Artemia franciscana and inhibit the leak channel activity of porcine heart and human ACLC. DESCRIPTION OF THE FIGURES FIG.1. A. franciscana mitochondria lack the Ca2+-induced mitochondrial permeability transition and Ca2+-sensitive large conductance channel of ATP synthase. a, EM images of A. franciscana mitochondria before and b, after treatment with 500 ^M Ca2+for 20 min before they were fixed for further EM analysis. c, Group data showing no change in electron lucencies before and after treatment with Ca2+(unpaired t-test was used). d, e, f, g, Representative single-channel recordings of A. franciscana ATP synthase (purified at pH 7.0) at +100 mV, –100 mV, +40 mV, –40 mV, respectively, showing that channel mainly dwells in its inactive, closed state and forms brief openings. h, Representative continuous single- channel recording of A. franciscana ATP synthase. The channel continued to remain closed or have brief, infrequent openings after the addition of 1 mM Ca2+. i, Representative continuous single-channel recording of porcine ATP synthase in the presence of 1 mM Ca2+, showing characteristic large-conductance channel activity and sensitivity to ATP (1 mM). Bongkrekic acid (BA) (10 μM) was added to confirm that the ANT channels are not responsible for the recorded currents due to possible contamination. C refers to the closed state of the channel, O refers to the open state of the channel, and S1 and S2 refer to the sub-conductance states. The green boxes in panels (h) and (i) indicate the presence of 1 mM Ca2+during the recording. Signals were filtered at 5 kHz using the amplifier circuitry. FIG.2. The overall structure of A. franciscana ATP synthase, purified at pH 7.0. a, The side views of the cryo-EM composite map of A. franciscana ATP synthase monomer in rotational State 2 shown as a surface. Monomeric ATP synthase contains 17 different subunits, shown in different colors. b, Cartoon representation of the atomic model of A. franciscana ATP synthase monomer in rotational State 2 after fitting into the cryo-EM map shown in a. ATP molecules occupying the nucleotide binding sites of β subunits are shown as lime green spheres. A close-up view of IF1 is shown in the black box. FIG.3. Distinct structural features of A. franciscana ATP synthase OSCP and α subunits. a, The comparison of OSCP subunit structures in A. franciscana (PDB:9B0X, EMD-44061, State 2), and H. sapiens maps (PDB:8H9T, EMD-34581, State 2). The top views of superimposed maps and models of A. franciscana and H. sapiens F1show the structural differences in OSCP subunits. OSCP has more open conformation in H. sapiens compared with A. franciscana. The side chain of the conserved OSCP H113 in A. franciscana is shown in the model.6 non-conserved residues that may play a role in CypD binding to OSCP in mammals22are shown in green in A. franciscana model. Residue numbers exclude the signal peptide sequence. b, c, d, The side views of the ATP synthase cryo-EM maps of A. franciscana (EMD-44061, State 2), H. sapiens (EMD-34581, State 2) and O. aries (EMD- 10573, State 1) showing the differences in the interaction of α subunit N-terminal helical segments with OSCP. The N-terminal alpha-helical segments of all three α subunits interact with the OSCP only in A. franciscana ATP synthase. e, Different side views of A. franciscana ATP synthase F1domain showing interactions of α subunits with OSCP. For simplicity, β subunits and the central stalk subunits are not shown. f, The amino acid sequence alignment of OSCP subunits of different species was generated through ClustalW Omega and treated using the Zappo color scheme in JalView, highlighting their physicochemical properties. Residue numbers include the signal peptide sequence. The 14 residues participating in the putative binding site of CypD with OSCP in mammals are shown in black boxes22. Six non- conserved residues in A. franciscana are indicated by asterisks. These residues are highlighted in green in panel (a). FIG.4. Distinct amino acid composition and structural features of A. franciscana ATP synthase e-subunit. a, The amino acid sequence alignment of ATP synthase e-subunit in different species. A. franciscana has the longest sequence, containing 84 residues. The ClustalW Omega and JalView were used to treat the sequences. The regions with over 60% identity are highlighted in light blue using the BLOSUM62 score color scheme. The amino acid conservation level is labeled from brown (low conservation) to yellow (high conservation). The amino acid residues are conserved at the N-terminus of the e-subunit (residues 5-43, near subunit g), while residues at the C-terminus of the e-subunit (residues 44- 84, near c-ring) are non-conserved. Residues of e-subunit participating in bond formation with the c-ring are shown with asterisks. Conserved GXXXG residues are shown in the black box. b, The cryo-EM map of A. franciscana ATP synthase FOdomain (EMD-44061). The density corresponding to the lipids is colored grey. The C-terminus of the e-subunit that interacts with the c-ring and the lipids is colored cyan. c, Different 2D classes of A. franciscana ATP synthase showing the density of e-subunit. d, The comparison of FO cryo- EM maps from different species are fitted with the respective models to show differences in the e-subunit C-terminus. The models of only e-subunit and c-ring are shown for simplicity. The C-terminal α-helical segment of the e-subunit is more extended and has a stronger density in A. franciscana (PDB:9BPG, EMD-44776, contour level 0.223), compared with the ovine (PDB:6TT7, EMD-10573, contour level 0.0251), bovine (PDB:6ZBb, EMD-11149, contour level 0.0108), and human (PDB:8H9F, EMD-34565, contour level 0.104) structures. The A. franciscana e-subunit is the longest and shows the closest interaction with the c-ring and the lipids occupying its cavity. The conserved (1-43) and non-conserved (44-84) residues in the e-subunit are shown in red and blue, respectively. e, Left panel: The superimposed image of A. franciscana e-subunit (cyan) with the ovine e-subunit (grey) showing longer and curved conformation of the A. franciscana e-subunit and its closer positioning to the c-ring. Central and right panels: The possible interactions between subunit e, c-ring, and the lipid in A. franciscana and O. aries, respectively. The distances, shown by the green dotted lines, suggest multiple interactions in A. franciscana FO. Two salt bridges may form between residues eK61 and cD1 (d=4.0Å for both interactions). Two salt bridges may form between eK79 and D1 (d=3.3Å and d=3.4Å). Two hydrogen bonds may form between the eY71 and cD1 residues (d=2.5Å and d=3.0Å). The residues eK79 and eY71 may also interact with the lipids found inside the c-ring. The exact nature of the lipids occupying the A. franciscana c- ring remains to be identified. Lyso-phosphatidylserine was tentatively fitted into the density of the c-ring lumen at the intermembrane space side for visualization purposes only. These interactions cannot be formed between subunit e and c-ring residues based on the O. aries ATP synthase model. The terminal lysine residue in mammals was suggested to interact with the lipids occupying the c-ring46. f, The amino acid sequence alignment of g-subunit in different species was generated through ClustalW Omega and treated using the Zappo color scheme in JalView, highlighting their physicochemical properties. The GXXXG motif shown in the black box has the conserved gI91 residue replaced by a cysteine in A. franciscana and D. melanogaster. g, FO domain of A. franciscana ATP synthase (PDB:9B0X, EMD-44061, contour level 11) showing the possible interactions between the GXXXG motifs of e- and g- subunits and g and b transmembrane helices. The gI91C substitution in A. franciscana favors the hydrogen bond formation between the gC91 and bS47 (d=3.4Å). This interaction is not observed in the human ATP synthase (PDB: 8H9J)47. FIG.5. The C-terminal peptides A. franciscana and human ATP synthase e- subunits inhibit the human c-subunit channel activity. a, b, c, d, e, Representative planar lipid bilayer recordings of c-subunit before and after adding 5 µM Peptides I, II, III, IV, or V. C refers to the closed and O to the open states of the channel. f, Representative planar lipid bilayer recording of the c-subunit channel during continuous voltage ramp from -100 mV to +100 mV in the presence of Peptides I, II, III, IV or V. g, Group data of the peak conductances of c-subunit channel activity before and after adding the Peptides I, II, III, IV and V (n = 7 for Peptide I, n = 9 for Peptide II, n = 8 for Peptide III, n = 10 for Peptide IV and n = 10 for Peptide V). ****P < 0.0001, paired t-test was used). h, i, Electrostatic potential maps of A. franciscana and human e-subunits generated using APBS-PDB2PQR software. The models generated by AlphaFold were used to show the properties of the full-length protein structures. The C-terminus of A. franciscana e-subunit has both negatively and positively charged regions, while the C-terminus of the human e-subunit is more negatively charged. The shown sequences of e-subunits were used for peptide synthesis. j, Peptides IV and V were designed by using the sequence of human IF1. Signals were filtered at 5 kHz using the amplifier circuitry. FIG.6. The hypothetical gating model of the ATP synthase c-subunit leak channel proposing the activation and inactivation mechanisms in mammalian and A. franciscana ATP synthase, respectively. a. Hypothetical conformational changes in mammalian ATP synthase leading to activation of leak channel within its c-ring. The binding of Ca2+to β-subunits and CypD to OSCP may induce conformational changes in F1and peripheral stalk, inducing channel activation. Lipids in the c-ring lumen may be displaced or removed due to the expansion of the c-ring, allowing ion conduction. Black arrows indicate the possible path of ion flow through the channel. b. Proposed leak channel inactivation mechanism in A. franciscana. Red dotted circles indicate the differences in interactions of e- subunit with c-ring in two species. Enhanced interactions between e-subunit with the c-ring and lipids may be key in regulating channel inactivation in A. franciscana. The figure is created with BioRender.com. For simplicity, only ATP synthase monomer is shown. FIG.7. A. franciscana mitochondria lack the Ca2+-induced and CsA-sensitive mitochondrial permeability transition and Ca2+-sensitive large conductance channel of ATP synthase. a, Calcium retention capacity (CRC) assay of control and CsA-treated mitochondria isolated from HEK 293 cells and b, A. franciscana. The black arrows indicate the addition of 50 µM Ca2+pulses. The red arrow indicates the change in fluorescence signal due to the mPTP opening in HEK 293, n = 3 biologically independent samples. c, Size exclusion chromatography (SEC) profile of purified A. franciscana ATP synthase monomer. A representative example of three independent runs is shown. d, SDS-PAGE analysis of DDM-purified A. franciscana ATP synthase. The silver stain was used for protein band visualization. e, Oligomycin-sensitive ATP hydrolysis activity was measured to assess the coupling of purified protein. The assay was performed at 37 ºC using purified A. franciscana ATP synthase (4 µg) in the absence and presence of oligomycin (n = 4, *P = 0.033). f, Representative channel recordings of A. franciscana ATP synthase purified at pH 8.0. g, Representative channel recording of A. franciscana ATP synthase in the presence of 1 mM Ca2+. h, Representative channel recording of A. franciscana ATP synthase purified at pH 8.0 before and after adding 1 mM Ca2+. i, Group data of peak conductance activities and j, open channel duration of A. franciscana ATP synthase purified at pH 8.0 and pH 7.0. k, Group data of peak conductance activities of A. franciscana ATP synthase channel (purified at pH 8.0) in the presence and absence of 1 mM Ca2+. An unpaired t-test was used for statistical analysis. l, Representative planar lipid bilayer recordings of human c-subunit before and after adding Peptide II (5 µM). Signals were filtered at 5 kHz using the amplifier circuitry. FIG.8. Cryo-EM data processing of A. franciscana ATP synthase purified at pH 7.0. The data processing flowchart of A. franciscana ATP synthase in three rotational states, State 1, State 2, and State 3a. The local refinement of the local FOand F1regions in three rational states was performed. Scale bar, 20 nm. FIG.9. Local resolution estimation, particle distribution, and Gold Standard Fourier Shell Correlation (GSFSC) resolution estimation of the A. franciscana ATP synthase (pH 7.0.) in different rotational states. a, State 2, b, State 2 FO local-refined, c, State 2 F1 local-refined, d, State 3a, e, State 1. FIG.10. Structural differences in A. franciscana ATP synthase F1domain at pH values 7.0 and 8.0. a, "Walker-Boyer" state geometries are observed for A. franciscana α and β subunits at pH 7.0 and b, at pH 8.0, with "open" conformation (βE, "empty") and the "loose" conformation (βTP / DPwith bound adenosine 5'-diphosphate or adenosine 5’- triphosphate). c, side views and d, bottom views of the F1local refined cryo-EM maps of A. franciscana ATP synthase (rotational State 2) at pH 7.0 and 8.0. An extra density corresponding to the IF1 monomer is present in the map obtained at pH 7.0 but absent at pH 8.0. e, The β subunit is found in the closed state in the presence of IF1 (pH 7.0), while it is in the open state in the absence of IF1 (pH 8.0). f, IF1 dimer density was found in F1 focused refined map of rotational State 3a at pH 7.0. FIG.11. Structural differences in the peripheral stalk subunits of A. franciscana and H. sapiens ATP synthase. a. Surface representation of the ATP synthase models of A. franciscana and H. sapiens, aligned by the central stalk subunits. The peripheral stalks and FO domains are shown in orange and blue in A. franciscana and H. sapiens structures, respectively. The white dashed arrows indicate the shift of the peripheral stalk and FOin H. sapiens compared to A. franciscana. b. Individual alignment of d-subunits shows less apparent differences between the segments A1-L42 (A. franciscana) and L5-A43 (H. sapiens). The individual alignment of αTPsubunits shows no differences in the two structures. c., d. The comparison of subunits d and αTP shows a shift in human subunits towards the central stalk. In contrast, αTP is moved towards the d-subunit in A. franciscana, leading to more enhanced interactions between these subunits described in Fig. S 6. The α-helical segment of the A. franciscana d-subunit (dA1-dL42) has more upright conformation than the corresponding segment in the human d-subunit (dL5-dA43), which is tilted towards αTP and CS. e. The comparison of A. franciscana and H. sapiens b subunit structures. Subunits are aligned either by the central stalk or individually. Differences are noticed mainly in the A. franciscana b-subunit segment containing residues P1-H145 compared with the corresponding segment, P3-V147, in H. sapiens. State 2 models of A. franciscana (PDB:9B0X) and H. sapiens (PDB:8H9T) ATP synthase were used in all panels. FIG.12. Structural analysis of A. franciscana and H. sapiens ATP synthase peripheral stalk subunits b, d and their interactions with αTP. a. The cryo-EM map of A. franciscana ATP synthase showing interactions between subunits αTP, F6, b, and d. The black box highlights the proximity of subunits αTP and d. b. The cartoon representation of the subunits αTP, F6, b, and d in A. franciscana. Potential interactions between the αTP N-terminal domain with subunits d and b are shown in the green box. A hydrogen bond may form between dT61 and αTP E7 (d=3.8Å) and dT61 and bE162 (d=3.2 Å). A salt bridge may form between αTP E7 and bR166 (d=3.6 Å). The potential interactions between subunit d and the C- terminal region of αTPare shown in the yellow box. A salt bridge may form between the dR39 and αE467 (d=3.3Å). c. The cryo-EM map of H. sapiens ATP synthase shows interactions between subunits αTP, F6, b, and d. The black box highlights the farther distance between subunits αTPand d in H. sapiens compared with the A. franciscana structure shown in panel (a). d. The cartoon representation of the subunits αTP, F6, b, and d in H. sapiens. There are fewer interactions in human structures compared with A. franciscana. The interactions between the N-and C-terminal domains of the αTP with d observed in A. franciscana are absent in the human structure (green and yellow boxes). A hydrogen bond may form between the dN59 and bQ162 (green box). For clarity, other ATP synthase subunits were omitted from the figure. All hydrogen bond and salt bridges are shown with green dotted lines. State 2 models of A. franciscana (PDB:9B0X) and H. sapiens (PDB:8H9T) ATP synthase were used in all panels. FIG.13. Structural analysis of A. franciscana and H. sapiens ATP synthase a and b subunits. a. The top view of the A. franciscana FO domain shows interactions between subunits a (purple and blue), b (grey), and c (yellow). The α-helical segment of a-subunit (residues 181-218, colored blue) shows a more linear conformation compared with the corresponding sequence in H. sapiens. A disulfide bond may form between bM66 and aM196 (d= 3.9Å), and two hydrogen bonds may form between the hydroxyl group of bT59 with the carbonyl and the amine group of aQ191 (d= 2.4Å and d=3.4Å, respectively). b. The α-helical segment of a-subunit (residues 185-225, colored blue) has a more curved conformation in H. sapiens. The following interactions were found between subunits a and b: A hydrogen bond between bS62 and aT200 and a salt bridge between bD56 and aH172. All hydrogen bond and salt bridges are shown with green dotted lines. c. Cartoon representation of A. franciscana ATP synthase FO domain shows two cardiolipin molecules (CDL1 and CDL2) fitted into their respective densities in the map. CDL1 is located between the subunits a, b, and f, interacting with the α-helical domains aH3, aH6, cH1, bH3, and fH4. The second cardiolipin molecule, CDL2, was found between the subunits b, f, and g and interacts with the helical domains fH1, fH3, fH4, bH1, bH2, and gH3. State 2 models of A. franciscana (PDB:9B0X) and H. sapiens (PDB:8H9T) ATP synthases were used in all panels. FIG.14. Interactions between the N-terminal regions of α-subunits with OSCP in A. franciscana ATP synthase. a, Left panel in the grey box: Interactions between the αDP- subunit and H1 and H5 helices of OSCP. Salt bridges may form between αK15 and OSCP E92 (d=2.8Å). Two salt bridges may form between αE7 and OSCP R95 (d=2.5Å and d=3.3Å). A salt bridge and a hydrogen bond may form between the αE13 and OSCP H1K28 (d=2.6Å) and αE7 and H1Y18 (d=2.8Å), respectively. The right panel in the orange box shows interactions between the αE-subunit with the H3 and H4 helices of OSCP. A salt bridge may form between the αE13 and OSCPR51 residues (d=3.7Å). The middle panel is a zoomed-out view of the αDPand αEinteractions with the OSCP. b, The left panel in the light blue box shows interactions between the αTPand the peripheral stalk subunits d, b, F6, and C- terminal domain of OSCP (H8 helix). The αTP-subunit interacts with F6 through a hydrogen bond between αS9 and F6D19 (d=2.5Å). The αTP-subunit interacts with subunit d through a hydrogen bond between αE7 and dT61 (d=3.8Å) and with subunit b through a salt bridge between αE7 and bR166 (d=3.5Å). Another hydrogen bond and a π-π stacking interaction may form between the OSCP Y183 and F6 Q8 (d=3.0Å) and the OSCP Y183 and F6 F11 (d<5.0Å), respectively. The right panel in the red box shows interactions between the αTP, F6, b, and OSCP H7 and H8 helices. A salt bridge may form between bE190 and OSCP K125 (d=2.6Å). The side chains of non-polar residues involved in hydrophobic interactions are shown. The middle panel is a zoomed-out view of the αTP interactions with the OSCP and other peripheral stalk subunits. The green dotted lines show all the salt bridges and hydrogen bonds. FIG.15. Proton translocation pathway in A. franciscana FO. a, Left and right panels, inlet half-channels of A. franciscana (PDB:9B0X) and H. sapiens (PDB:8H9T) a- subunits, respectively. The middle panel is a zoomed-out view of the inlet channel at the interface of the a- and c-subunits. b, Left and right panels, outlet half-channels of A. franciscana and human a-subunits, respectively. The middle panel is a zoomed-out view of the outlet channel at the interface of the a- and c-subunits. Key residues of the a-subunit and the c-ring are indicated. The green dotted lines indicate the possible formation of salt bridges between the conserved residues cE58 and aR157 (d=2.6Å), aR150 and aY216 (d=3.6Å) in A. franciscana. FIG.16. The amino acid sequence alignments of subunits a and c in different species show higher conservation of the c- but not a-subunit sequences. a. The amino acid sequence alignment of a-subunit was generated through ClustalW Omega and treated using the BLOSUM62 color scheme in JalView, highlighting residues with an identity above 60%. Two barrier arginine residues, R150 and R157 that separate the inlet and outlet half-channels of the a-subunit are indicated with asterisks. The inlet half-channel of a-subunit at the inner membrane starts with the aH168 and aH172 residues in mammals46, 67. The aH168 is conserved in A. franciscana (aH166), while the aH172 is replaced with threonine (aT170). Residues aH166 and aT170 are marked with asterisks. The key direct proton donor / acceptor glutamates, E198 in the inlet and E219 in the outlet channels are also shown with asterisks. b. The amino acid sequence alignment of the c-subunit was generated through ClustalW Omega and treated using the BLOSUM62 color scheme in JalView, highlighting residues with an identity above 60%. The conserved cE58 residue is marked by an asterisk. FIG.17. Cryo-EM data processing of A. franciscana ATP synthase purified at pH 8.0. The data processing flowchart of A. franciscana ATP synthase in three rotational states, State 1, State 2, and State 3a. The focused refinement of the local FO, F1, and Peripheral Stalk (PS) regions in three rotational states was performed. Scale bar, 20 nm. FIG.18. Local resolution estimation, particle distribution, and Gold Standard Fourier Shell Correlation (GSFSC) resolution estimation of the A. franciscana ATP synthase (pH 8.0) in different rotational states. a, State 2. b, State 2 FO local-refined, c, State 2 F1local-refined, d, State 2 PS local-refined, e, State 3a. f, State 1. FIG.19. The overall structure of A. franciscana ATP synthase, purified at pH 8.0. a, Composite map of A. franciscana ATP synthase in State 2. b, Cartoon representation of the atomic model of A. franciscana ATP synthase subunits. c, Cryo-EM map of the A. franciscana ATP synthase in the State 2 (Class 3, 223,705 particles) at pH 8.0 shown in two different threshold levels to highlight the absence of clear and complete densities for the peripheral stalk (PS) and subunits a, g, f, and 6.8PL of FO at pH 8.0. d, The FO local refined map of the same State 2 (Class 3) shown in panel (c), demonstrates a well-defined density for subunit e and its interaction with the c-ring in the absence of apparent density of peripheral stalk. FIG.20. The effect of the Artemia peptide on mPTP opening was evaluated in MIO- M1 cells exposed to either hyperglycemic conditions (HG) or an osmotic control (OC) by calcein AM-cobalt chloride assay. MitoTracker was used to visualize mitochondria. Quenching of Calcein AM fluorescence (i.e., loss of green) by cobalt chloride reflects mPTP opening. FIG.21. Data showing a described peptide protects rat retinal cells (R28) from oxidative stress-induced and mPTP-regulated cell death. FIG.22. Table S1. Cryo-EM data, refinement and validation statistics. FIG.23. Table S2. The list of the ATP synthase subunits, chains, and ligands found in the A. franciscana ATP synthase models. The amino acid sequence information for most subunits was obtained through the A. franciscana genome annotation, excluding subunits IF1 and 6.8PL. FIG.24. Table S3. Additional cryo-EM data, refinement and validation statistics. The figures include amino acid sequence, which are as follows: FIG.5h. Peptide I. LYLAKETGVKVPENF (SEQ ID NO:1.) Peptide III. MSFAPPVNVSPLIRA (SEQ ID NO:3) FIG.5i. Peptide II. KRIARELAEDDSILK (SEQ ID NO:2) which is referred to herein as Peptide II Fig.5j. Peptide IV, GAGSIREAGGAFGKR (SEQ ID NO:4). Peptide V. FRAQSREQLAALKKH (SEQ ID NO:5). Subunit O (FIG.3f): Artemia_fransciscana_OSCP MNDVSQAARQFSTTSAATQLVKAPIQVFGIEGRYATALYSAAVKQKKLEAVEKDLVQ LNASLKKFPRLGELLKNPTLSRQLKKDAIGSMLKEQKAVDLTSNFLDLLTENNRLKM VDGVINAFKTIMAAHRGEVICEVTSAKPLDEAARKDLEVALKGFLKPGQNIKLTLKT DPAIIGGLVVSIGDRFVDMSIGTKIKRYTAALKTAV (SEQ ID NO:7) Homo_sapiens_OSCP MAAPAVSGLSRQVRCFSTSVVRPFAKLVRPPVQVYGIEGRYATALYSAASKQNKLEQV EKELLRVAQILKEPKVAASVLNPYVKRSIKVKSLNDITAKERFSPLTTNLINLLAENGR LSNTQGVVSAFSTMMSVHRGEVPCTVTSASPLEEATLSELKTVLKSFLSQGQVLKLE AKTDPSILGGMIVRIGEKYVDMSVKTKIQKLGRAMREIV (SEQ ID NO:8) Bos_taurus_OSCP MAALAVSGLSQQVRCFSTSVVRPFAKLVRPPVQIYGIEGRYATALYSAASKQNKLEQV EKELLRVGQILKEPKMAASLLNPYVKRSVKVKSLSDMTAKEKFSPLTSNLINLLAENG RLTNTPAVISAFSTMMSVHRGEVPCTVTTASALDEATLTELKTVLKSFLSKGQVLKLE VKIDPSIMGGMIVRIGEKYVDMSAKTKIQKLSRAMREIL (SEQ ID NO:9) >Sus_scrofa_OSCP MASQAVSGLSRQVRCFSTSVVRPFAKLVRPPVQIYGIEGRYATALYSAASKQNKLEQV EKELLRVAQILKEPKVAASIMNPYVKRSVKVKSLSDMTAKEKFSPLTSNLINLLAENG RLSSTPGVISAFSTMMSVHRGEVPCSVTTASPLDEATLTELKTVLKSFLSKGQILKLEV KVDPSIMGGMIVRIGEKYVDMSAKTKIQKLSRAMREIF (SEQ ID NO:10) Subunit e (FIG.4a) >Artemia_franciscana_e MSFAPPVNVSPLIRAGRYGALVVGIVYGSYRFGSLQKRENEWRVEEARRKVIRDALN AENKAKATREEMLYLAKETGVKVPENF (SEQ ID NO:11) >Homo_sapiens_e MVPPVQVSPLIKLGRYSALFLGVAYGATRYNYLKPRAEEERRIAAEEKKKQDELKRIA RELAEDDSILK (SEQ ID NO:12) >Ovis_aries_e MVPPVQVSPLIKLGRYSALFLGMAYGAKRYNYLKPRAEEERRLAAEEKKKRDEQKR IERELAEAQEDTILK (SEQ ID NO:13) >Bos_taurus_e MVPPVQVSPLIKLGRYSALFLGMAYGAKRYNYLKPRAEEERRLAAEEKKKRDEQKR IERELAEAQEDTILK (SEQ ID NO:14) >Drosophila_melanogaster_e MSQAPVRVSPLIKFGRWSLLLVGIAYGAAHQSRLSKKEEKLREIEAQQKAVRDAKLA EEKKRSAEAEARALAELSKPTPKH (SEQ ID NO:15) >Danio_rerio_e MVPPVQVSPLIKTARWSALLIGLIYGKQRYDYLKPIAAEERRIEEEEKKLREEQERIYK QLSEANSDTILK (SEQ ID NO:16) Subunit g (FIG.4f) >Artemia_fransciscana_g MSALAKKIATSGPVVLKNTIAVTRPKLATFLKYAKVELTPPGPADVPKIQEGIQNLIHS AKTGKWKQVSVREAWLNTLIVTEIAMWFFVGECIGKGSVIGYRV (SEQ ID NO:17) >Homo_sapiens_g MAQFVRNLVEKTPALVNAAVTYSKPRLATFWYYAKVELVPPTPAEIPRAIQSLKKIVN SAQTGSFKQLTVKEAVLNGLVATEVLMWFYVGEIIGKRGIIGYDV (SEQ ID NO:18) >Ovis_aries_g MAQFVRNLAEKAPALVNAAVTYSKPRLATFWYYAKVELVPPTPAEIPTAIQSLKKIINS AKTGSFKQLTVKEALLNGLVATEVWMWFYVGEIIGKRGIIGYDV (SEQ ID NO:19) >Bos_Taurus_g AEFVRNLAEKAPALVNAAVTYSKPRLATFWYYAKVELVPPTPAEIPTAIQSLKKIINSA KTGSFKQLTVKEALLNGLVATEVWMWFYVGEIIGKRGIIGYDV (SEQ ID NO:20) >Drosophila_melanogaster_g MASLATKGSGLVNRLLTQARPQLDVFLKYAKVELTPPTPADIPAIRQGLGNIIKGAKTG AYKNLTVREAWLNTLVTAEVIFWFYIGECIGKRHIVGYNV (SEQ ID NO:21) >Danio_rerio_g MAQAVQKLVAKVPTLVGAAVNYSKPRLATFWYYARVELVPPTPAEIPKAISGFQDML KAFQSGRVGQTTVRDAVRNGLVATEVLMWFYIGEIIGKRGLIGYDV (SEQ ID NO:22) Fig.5h. Peptide I. LYLAKETGVKVPENF (SEQ ID NO:1) which is referred to herein as Peptide I. In examples, the ATP synthase comprises or consists of the sequence KRIARELAEDDSILK (SEQ ID NO:2) which is referred to herein as Peptide III, and is from the human ATP synthase. In examples, a described ATP synthase or fragment thereof is used such that the function of the ACLC is inhibited by inhibiting opening of the ACLC. Other peptides referred to herein include Peptide III – MSFAPPVNVSPLIRA (SEQ ID NO:3); Peptide IV – GAGSIREAGGAFGKR (SEQ ID NO:4) and Peptide V – FRAQSREQLAALKKH (SEQ ID NO:5). Subunit a (FIG.16a): >Artemia_fransciscana_a MMASLFSVFDPTSSFLSNWLSMLIPLLFMVMSFWLIPSRPQFLAKSVLMGLNREMSL LMGPASFGANILVIALFLFILFNNFIGLFPYIFTATSHLAVTLSLAVPLWISFILYTWIKET TNALAHLVPLGTPAPLMPFMVLMEIISNMIRPITLSVRLAANMIAGHLLLTLLGAQGT LENLYVTSIVVFSQIILLMLEFSVAIIQSYVFMTLMTLYASE (SEQ ID NO:23) >Homo_sapiens_a MNENLFASFIAPTILGLPAAVLIILFPPLFIPTSKYLINNRLITTQQWLIKLTSKQMMTMH NTKGRTWSLMLVSLIIFIATTNLLGLLPHSFTPTTQLSMNLAMAIPLWAGAVIMGFRSK IKNALAHFLPQGTPTPLIPMLVIIETISLLIQPMALAVRLTANITAGHLLMHLIGSATLAM STINLPSTLIIFTTLILLTILEIAVALIQAYVFTLLVSLYLHDNT (SEQ ID NO:24) >Sus_scrofa_a MNENLFASFIAPTMMGLPIVTLIIMFPSLLFPTPKRLINNRTISIQQWLIQMTSKQMMAI HNQKGQTWSLMLMSLIMFIGSTNILGLLPHSFTPTTQLSMNLGMAIPLWSATVFTGFR HKTKTSLAHFLPQGTPAPLIPMLVIIETISQLIQPVALAVRLTANITAGHLLIHLIGGATLA LLNISTMTAFITFTILILLTILEFAVALIQAYVFTLLVSLYLHDNT (SEQ ID NO:25) >Drosophila_melanogaster_a MMTNLFSVFDPSAIFNFSLNWLSTFLGLLMIPSIYWLMPSRYNIMWNSILLTLHKEFK TLLGPSGHNGSTFIFISLFSLILFNNFMGLFPYIFTSTSHLTLTLSLALPLWLCFMLYGWI NHTQHMFAHLVPQGTPAILMPFMVCIETISNIIRPGTLAVRLTANMIAGHLLLTLLGKT GSSMSYMLMTFLLMAQIALLVLESAVAMIQSYVFAVLSTLYSSEVN (SEQ ID NO:26) >Danio_rerio_a MMTSFFDQFASPYLLGIPLILVAMLLPWLLFPAPTSRWINNRLITVQTWLTGRFTNQL MTPLNFSGHKWALLFASLMVFLITINLLGLLPYTFTPTTQLSLNMGFAVPLWLATVIIG MKNQPTIALGHLLPEGTPIPLIPALIIIETISLFIRPLALGVRLTANLTAGHLLIQLIATAVF VLLPMMPAVAILTASVLFLLTLLEVAVAMIQAYVFILLLSLYLQENI (SEQ ID NO:27) >Saccharomyces_cerevisiae_a MFNLLNTYITSPLDQFEIRTLFGLQSSFIDLSCLNLTTFSLYTIIVLLVITSLYTLTNNNNK IIGSRWLISQEAIYDTIMNMTKGQIGGKNWGLYFPMIFTLFMFIFIANLISMIPYSFALSA HLVFIISLSIVIWLGNTILGLYKHGWVFFSLFVPAGTPLPLVPLLVIIETLSYFARAISLGL RLGSNILAGHLLMVILAGLTFNFMLINLFTLVFGFVPLAMILAIMMLEFAIGIIQGYVW AILTASYLKDAVYLH (SEQ ID NO:28) Subunit c (FIG.16b) >Artemia_fransciscana_c MIESRQWRLLWIILETKSKYTKMYTIARIATRAAVSQGSQAYLRPVSSAVLSQKVIVEA PVATQARSLQTSAVQRDIDSAAKFIGAGAATVGVAGSGAGIGSVFGSLIIGYARNPSLK QQLFSYAILGFALSEAMGLFCLMMAFLLLFAF (SEQ ID NO:29) >Homo_sapiens_c MFACSKFVSTPSLVKSTSQLLSRPLSAVVLKRPEILTDESLSSLAVSCPLTSLVSSRSFQT SAISRDIDTAAKFIGAGAATVGVAGSGAGIGTVFGSLIIGYARNPSLKQQLFSYAILGFA LSEAMGLFCLMVAFLILFAM (SEQ ID NO:30) >Ovis_aries_c MQTTGALLISPALIRSCTRGLIRPVSASFLSRPEIPSVQPSYSSGPLQVARREFQTSVVSR DIDTAAKFIGAGAATVGVAGSGAGIGTVFGSLIIGYARNPSLKQQLFSYAILGFALSEA MGLFCLMVAFLILFAM (SEQ ID NO:31) >Bos_taurus_c MQTTGALLISPALIRSCTRGLIRPVSASFLSRPEIQSVQPSYSSGPLQVARREFQTSVVS RDIDTAAKFIGAGAATVGVAGSGAGIGTVFGSLIIGYARNPSLKQQLFSYAILGFALSEA MGLFCLMVAFLILFAM (SEQ ID NO:32) >Sus_scrofa_c MQTTGALLISPALLRSCTRGLIRPVSASFLSRPEIPSEQPPCSSVPLQVARREFQTSVVS RDIDTAAKFIGAGAATVGVAGSGAGIGTVFGSLIIGYARNPSLKQQLFSYAILGFALFEA MGLFCLMVAFLILFAM (SEQ ID NO:33) >Drosophila_melanogaster_c MPPWRGELPMAAFLANSKQYLRPLSSAIISQSRTLAAQNTTPVALLPQIRSFQTSPVTR DIDSAAKFIGAGAATVGVAGSGAGIGTVFGSLIIGYARNPSLKQQLFSYAILGFALSEA MGLFCLMMAFLLLFAF (SEQ ID NO:34) >Danio_rerio_c MFTCAKFVSTPALVRAGSRSVYRPVSAAVLSRPEAKPEVSTAAILQSPVAQMALRSFQ TSAVSRDIDTAAKFIGAGAATVGVAGSGAGIGTVFGSLIIGYARNPSLKQQLFSYAILGF ALSEAMGLFCLMVAFLILFAM (SEQ ID NO:35) DETAILED DESCRIPTION Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Every numerical range given throughout this specification includes its upper and lower values, as well as every narrower numerical range that falls within it, as if such narrower numerical ranges were all expressly written herein. As used in the specification and the appended claims, the singular forms “a” "and” and “the" include plural referents unless the context clearly dictates otherwise. Ranges may be expressed herein as from “about” one particular value, and / or to “about” another particular value. When such a range is expressed, another embodiment includes from the one particular value and / or to the other particular value. Similarly, when values are expressed as approximations, by the use of the antecedent “about” it will be understood that the particular value forms another embodiment. The term “about” in relation to a numerical value encompasses variations of + / -10%, + / - 5%, or + / - 1%. This disclosure includes every amino acid sequence described herein and all nucleotide sequences encoding the amino acid sequences. Polynucleotide and amino acid sequences having 90-99% similarity, inclusive, and including and all numbers and ranges of numbers there between, with the sequences provided here are included in the invention. All of the amino acid sequences described herein can include additional amino acid substitutions other than those expressly described, such as conservative substitutions, that do not adversely affect the function of the protein that comprises the amino acid sequences. In an example, the disclosure provides an isolated or recombinant adenosine triphosphate (ATP) synthase or fragment thereof that inhibits function of mammalian ATP synthase c-subunit leak channel (ACLC) but does not inhibit ATP synthesis of ATP by the mammalian ATP synthase. In an optional example, the ATP synthase or fragment thereof comprises a modification relative to the sequence of an unmodified ATP synthase. In examples, the modification is present and comprises attachment of a cellular localization amino acid sequence may be a mammalian mitochondrial localization signal. In examples, the modified ATP synthase or fragment thereof comprises a segment of A. franciscana ATP synthase, or a fragment of the human ATP synthase. In an example, the fragment comprises or consists of the sequence the ATP synthase from A. franciscana that is LYLAKETGVKVPENF (SEQ ID NO:1) which is referred to herein as Peptide I. In examples, the ATP synthase comprises or consists of the sequence KRIARELAEDDSILK (SEQ ID NO:2) which is referred to herein as Peptide II, and is from the human ATP synthase. In examples, a described ATP synthase or fragment thereof is used such that the function of the ACLC is inhibited by inhibiting opening of the ACLC. Other peptides referred to herein include Peptide III – MSFAPPVNVSPLIRA (SEQ ID NO:3); Peptide IV – GAGSIREAGGAFGKR (SEQ ID NO:4) and Peptide V – FRAQSREQLAALKKH (SEQ ID NO:5). In examples an effective amount of one or more of the described peptides can be administered to an individual in need thereof. An effective amount is an amount that is sufficient to generate a desired response, such as to reduce or eliminate a sign or symptom of a condition or disease. When administered to a subject, a dosage will generally be used that will achieve a target concentration that has been shown to be sufficient for in vitro inhibition of the described protein interaction. In some examples, an "effective amount" is one that treats (including prophylaxis) one or more symptoms and / or underlying causes of any of a disorder or disease. In some examples, an effective amount is a therapeutically effective amount. The term “therapeutically effective amount” as used herein refers to an amount of a described peptide or combination thereof, in a single dose or multiple doses, to achieve the intended purpose of treatment. The amount desired or required may vary depending its mode of administration, patient specifics and the like. Appropriate effective amounts can be determined by one of ordinary skill in the art informed by the instant disclosure using routine experimentation. In examples, a one or more described peptides is administered to an individual. In examples, a described peptide alone or in combination with at least one other agent is administered to an individual. In examples, an intracellular local concentration of the described peptide is as described herein. In examples, a described peptide is modified to include additional amino acids, amino acid substitutions, or amino acids that are not normally incorporated during protein translation. The described modified peptides can thus be modified in a variety of ways, including but not necessarily limited to introduction of nonstandard amino acids such as Ornithine, Norarginine, Diaminobutyric acid and Diaminopropionic acid. Backbone modifications can be made and include amide backbone, circularization, and stapling. L or D forms of amino acids may be incorporated. The disclosure also includes chimeric peptides, with either repeated domains, or domains from different parts of described proteins connected to one another. The additional amino acids may be a cellular localization signal, such as to localize the peptide to a mitochondria, or other organelle. In examples, to enhance its trafficking into mitochondria a tag may be used. In an example, the tag is rFrFrF-K(TAMRA)- LYLAKETGVKVPENF (SEQ ID NO:1). For this tag, the sequence D-arginine- phenylalanine-D-arginine-phenylalanine-D-arginine-phenylalanine-NH2 (rFrFrF (SEQ ID NO:36)) was tagged with a rhodamine-based fluorescent chromophore. This probe (TAMRA- rFrFrF (SEQ ID NO:36)) was then added to a described peptide. This modification allows assessment of the mitochondrial colocalization of the peptide with fluorescence microscopy. With or without the fluorescent moiety, a described peptide can colocalize to the mitochondria when it is modified to include any suitable mitochondria penetrating peptide sequence. In non-limiting examples, a described peptide or combination thereof of this disclosure is combined with one or more pharmaceutically acceptable agents. In examples, a described agent is combined with unilamellar and / or multilamellar vesicular structures such as liposomes or lipid nanoparticles, cationic polymers, lipoplexes, polyplexes, or inorganic nanoparticles. Any delivery agent described herein may comprise polyethylene glycol (PEG) and thus may be PEGylated. In an example, a described peptide may be provided as a fusion protein. In an example, a fusion protein comprises a described peptide and a segment that includes a cell-penetrating peptide. The disclosure includes compositions comprising the ATP synthase fragments. The compositions may be provided as pharmaceutical formulations. The disclosure includes a method comprising administering the ATP synthase fragment or a composition comprising the fragment to an individual in need thereof to thereby provide a therapeutic or prophylactic benefit to the individual. To support these approaches, the disclosure includes the ability to use ex-vivo studies with primary hippocampal cultures to perform cell viability studies; in vivo assays with Alzheimer’s disease mouse models using tail vain injections of the peptides into the AD mice to rescue neuronal death and / or the onset of AD-like features in mice by targeting and inhibiting the ATP synthase leak channel. The disclosure includes performing ex-vivo and in-vivo studies with the peptides to assess their cell death-preventing function in diabetic retinopathy and nephropathy. The disclosure includes analysis of the effects of the peptides on Amyotrophic Lateral Sclerosis (ALS) such as by using the peptides in ex-vivo (iPS cells) and in-vivo Drosophila and mouse models of ALS. The disclosure includes analysis of the peptides for inhibiting or preventing cell death in an in-vivo ischemic injury model of a mouse heart. The disclosure includes use of the peptides for inhibiting or preventing neuronal death in an in- vivo brain ischemia-reperfusion injury model of mice. The disclosure includes using the peptides to rescue heart failure by targeting ATP synthase leak channel-induced cardiomyocyte death. Accordingly, in examples, the method comprises administering one or a combination of the described peptides to an individual who is in need of treatment or prophylaxis of any of the following conditions which are not necessarily mutually exclusive: Alzheimer’s disease; diabetes including but not limited to diabetic retinopathy and nephropathy; Amyotrophic Lateral Sclerosis (ALS); ischemia-reperfusion injury; ischemic stroke and / or brain ischemia; myocardial ischemia; mesenteric ischemia; peripheral and / or limb ischemia; intestinal ischemia; kidney ischemia; Transient Ischemic Attack (TIA); angina; ischemic heart disease; chronic obstructive pulmonary disease; arterial occlusion; anemia; blood vessel compression and / or stenosis; blood vessel rupture; coronary microvascular disease; or non-obstructive coronary artery disease; or osteoarthritis. In examples, the ATP synthase fragment is introduced into epithelial cells, neurons, cardiomyocytes, or muscle cells. In examples, introducing the ATP synthase fragment into cells inhibits or prevents death of the cells into which the ATP synthase fragment is introduced. In examples, the ATP synthase fragment is introduced intravenously. In examples, the individual in need of a described peptide is a human or a non-human mammal. The disclosure also includes mammalian cell into which an A. franciscana ATP synthase protein or fragment thereof has been introduced. Examples of the disclosure are illustrated by the following statements. In these examples, the disclosure provides: Statement 1. Isolated or recombinant adenosine triphosphate (ATP) synthase or a fragment thereof that inhibits function of mammalian ATP synthase c-subunit leak channel (ACLC) but does not inhibit ATP synthesis of ATP by the mammalian ATP synthase, and wherein optionally the ATP synthase or fragment thereof comprises a modification relative to the sequence of an unmodified ATP synthase. Statement 2. The isolated or recombinant ATP synthase of fragment thereof of Statement 1, wherein the modification is present and comprises attachment of a cellular localization amino acid sequence that is optionally a mammalian mitochondrial localization signal. Statement 3. The isolated or recombinant ATP synthase or fragment thereof of Statement 1 or Statement 2, wherein the ATP synthase or fragment thereof comprises a segment of A. franciscana ATP synthase. Statement 4. The isolated or recombinant ATP synthase fragment of Statement 4, wherein said fragment comprises or consists of the sequence LYLAKETGVKVPENF (SEQ ID NO:1). Statement 5. The isolated or recombinant ATP synthase fragment of Statement 1, wherein the fragment comprises or consists of the sequence KRIARELAEDDSILK (SEQ ID NO:2). Statement 6. The isolated or recombinant ATP synthase or fragment thereof of any one of Statements 1-5, wherein the function of the ACLC that is inhibited by said fragment comprises inhibiting opening of the ACLC. Statement 7. The isolated or recombinant ATP synthase of any one of Statements 1- 6, wherein the isolated or recombinant ATP synthase or fragment thereof further comprises the sequence of SEQ ID NO:36. Statement 8. A pharmaceutical composition comprising the ATP synthase fragment thereof of any one of Statements 1-7. Statement 9. A method comprising administering the pharmaceutical composition of Statement 8 to an individual in need thereof to thereby provide a therapeutic or prophylactic benefit to the individual. Statement 10. The method of Statement 9, wherein the individual is in need of treatment or prophylaxis of any of the following conditions: Alzheimer’s disease; diabetes including but not limited to diabetic retinopathy and nephropathy; Amyotrophic Lateral Sclerosis (ALS); ischemia-reperfusion injury; ischemic stroke and / or brain ischemia; myocardial ischemia; mesenteric ischemia; peripheral and / or limb ischemia; intestinal ischemia; kidney ischemia; Transient Ischemic Attack (TIA); angina; ischemic heart disease; chronic obstructive pulmonary disease; arterial occlusion; anemia; blood vessel compression and / or stenosis; blood vessel rupture; coronary microvascular disease; or non-obstructive coronary artery disease; or osteoarthritis. Statement 11. The method Statement 10, wherein the ATP synthase fragment is introduced into epithelial cells, neurons, cardiomyocytes, or muscle cells. Statement 12. The method of Statement 10 or Statement 11, wherein introducing the ATP synthase fragment into cells inhibits or prevents death of the cells into which the ATP synthase fragment is introduced. Statement 13. The method of any one of Statements 10-12, wherein the ATP synthase fragment is introduced intravenously. Statement 14. The method of Statement 13, wherein the individual is a human or a non-human mammal. Statement 15. A mammalian cell into which an ATP synthase or fragment thereof of any one of claims Statements 1-7 has been introduced, and wherein optionally, the mammalian cell is in vitro. This disclosure provides an analysis of the structure and leak channel activity of A. franciscana ATP synthase and its possible contribution to Ca2+-resistant mPT. The data show that A. franciscana ATP synthase forms channels with significantly different channel properties when compared with the porcine heart ATP synthase. A. franciscana ATP synthase demonstrated transient, Ca2+-insensitive openings after prolonged inactivity in planar lipid bilayer recordings. Single particle cryo-electron microscopy analysis of A. franciscana ATP synthase revealed a similar subunit composition and organization to mammalian ATP synthases44, 45, 46, 47. It contains two molecular domains, soluble F1and membrane-embedded FO. Despite the observed similarities, A. franciscana ATP synthase structure had substantial differences. All three α-subunits (αTP, αDP, αE) had strong densities of N-terminal alpha- helical segments and enhanced interactions with the OSCP, compared with the mammalian structures, where at least one out of the three α-subunits lack this density46, 47, 48, 49. Furthermore, the C-terminal region of the FO e-subunit had a stronger density in the A. franciscana ATP synthase cryo-EM maps compared with all known mammalian structures46,47, 48. In addition, the e-subunit was found to be longer by 14 amino acid residues and had a more curved C-terminal region shifted towards the c-ring. The proximity of the e-subunit C- terminus to the c-ring may favor interactions with the N-terminal residues of the ring and lipids occupying its cavity. A peptide with the corresponding sequence of the A. franciscana e-subunit C-terminal region inhibited the c-subunit channel activity during electrophysiology recordings. These findings indicate the e-subunit may form an inactivation gate of the c- subunit channel acting from the intermembrane space in concert with F1, which acts at the matrix side. Distinct structural features of the A. franciscana e-subunit may lead to enhanced interactions with the c-ring and contribute to the inactivation of the ATP synthase c-subunit leak channel in this organism. The described ATP synthase and fragments thereof may therefore be used in certain therapeutic or prophylactic approaches, particularly where hypoxia and / or ischemic conditions may be contributors to the conditions in need of treatment. The following Examples are intended to illustrate but not limit the disclosure. Examples Mitochondria isolated from A. franciscana lack Ca2+-induced and cyclosporin A- sensitive mitochondrial permeability transition Mitochondria isolated from A. franciscana were reported to sequester large amounts of calcium with no evidence of swelling, rupture of the inner and outer membranes, and activation of mPTP40, 41, unlike mammalian mitochondria1, 2. We performed transmission electron microscopy imaging on A. franciscana mitochondria, which confirmed the lack of swelling of the mitochondrial matrix upon Ca2+treatment (Fig.1a, b, c). The calcium retention capacity experiment (CRC) is a widely used method to quantify the susceptibility of mitochondria to mPT. During this experiment, mitochondria isolated from HEK 293 cells and A. franciscana were exposed to Ca2+pulses (Fig.7a, b). Ca2+accumulates in mitochondria until it triggers mPTP opening, followed by membrane depolarization, rupture of mitochondria, and release of sequestered Ca2+. Mitochondria isolated from HEK 293 cells showed a jump in the fluorescence signal upon the addition of the fifth Ca2+pulse that correlates with the mPTP opening (Fig.7a). In contrast, mitochondria isolated from A. franciscana sequestered significant amounts of Ca2+without a sudden increase in the fluorescence signal and mPTP activation (Fig.7b).40. Additionally, the CsA treatment did not affect Ca2+uptake in A. franciscana mitochondria, while it caused a delay in mPTP opening in HEK 293 (Fig.7a, b). CsA inhibits mPTP by directly interacting with the mPTP activator, mitochondrial matrix protein CypD50, 51. CypD was shown to specifically interact with the mammalian ATP synthase peripheral stalk subunit OSCP22and induce ATP synthase leak channel activation52,53. A significant amount of CypD is present in A. franciscana mitochondria40. However, the molecular mechanism of the mitochondrial Ca2+resistance and CsA-insensitivity in this organism remains unknown. Thus, we performed structural and electrophysiology analysis of A. franciscana ATP synthase to determine if it possesses distinct properties that may contribute to mPTP inhibition. A. franciscana ATP synthase channel dwells in its inactive state and forms calcium- insensitive brief openings A. franciscana ATP synthase was purified by solubilizing mitochondria with the non- ionic detergent n-dodecyl-ß-D-maltoside (DDM), pH 7.0. ATP synthase was purified by differential centrifugation and polyethylene glycol (PEG) precipitation followed by size exclusion chromatography (SEC) as described in Materials and Methods (Fig.7c, d). The purified A. franciscana ATP synthase was fully assembled, as verified by the SEC profile (Fig.7c). It demonstrated oligomycin-sensitive ATP hydrolysis activity, confirming the fully assembled and coupled conformation of purified ATP synthase (Fig.7e). The planar lipid bilayer recordings revealed significant differences in A. franciscana and porcine heart ATP synthase leak channel activities. A. franciscana ATP synthase formed voltage-dependent brief openings with the channel mainly dwelling in its closed inactive state (Fig.1d-g). Fig.1h shows a continuous single-channel recording of A. franciscana ATP synthase before and after Ca2+addition. The channel remained closed or had brief, infrequent openings in the presence of Ca2+, in contrast to what was observed for porcine heart ATP synthase under the same conditions33(Fig.1i). The porcine heart ATP synthase was previously shown to form voltage- dependent channels with 1.5 nS peak conductance activity and prolonged dwell times in its open state in patch-clamp recordings33. In addition, Ca2+was shown to increase the frequency of channel opening33. Fig.1i shows a continuous lipid bilayer recording of the porcine heart ATP synthase in the presence of 1 mM Ca2+, which forms large multi-conductance channels inhibited by ATP. BA was added to rule out the possibility of other mitochondrial ion channels, such as ANT, contributing to the recorded currents. BA failed to inhibit the channel, proving that channels are formed by porcine ATP synthase (Fig.1i). These results indicate that the A. franciscana ATP synthase leak channel has significantly different channel properties than the porcine heart ATP synthase. The former mainly dwells in its inactive form during the recording and is insensitive to Ca2+. The overall structure of A. franciscana ATP synthase at pH 7.0 We determined the complete structure of A. franciscana ATP synthase monomer using single-particle cryo-EM (Fig.2) to analyze the structural basis of ATP synthase leak channel inactivation in this organism. We obtained 3 three-dimensional (3D) maps for the A. franciscana ATP synthase, purified at pH 7.0 for the rotational states 1, 2, and 3a at nominal resolutions of 2.6 Å, 2.6 Å, and 2.7 Å, respectively (Figs.8 and 9; Table S1 as shown in Fig.22). Each catalytic cycle of ATP synthase for ATP synthesis or hydrolysis is accompanied by the ~120orotation of the central stalk γ subunit, allowing the formation of rotational States 1, 2, and 344. The rotational states differ from each other by the position of the rotor (consisting of γ, δ, ε subunits forming the central stalk and c8 ring) relative to the stator44(comprising of the α3β3 hexamer and peripheral stalk). State 2 was the most prevalent in this dataset, which we selected for further analysis (Fig.8). The local refinements of the F1 and FO domains were performed to improve the map quality and resolution. The final local refined maps for the F1 and FO at State 2 had nominal resolutions of 2.6 Å and 3.9 Å, respectively (Figs.8 and 9; Table S1 as shown in Fig.22). The F1and FOlocal refined maps were combined to generate the composite map of A. franciscana ATP synthase (Fig.2a). The local refined map for FO at State 1 had a resolution of 3.3 Å (Table S1 as shown in Fig.22). A. franciscana maps had densities for seventeen ATP synthase subunits, reported in mammalian ATP synthases, except for subunit DAPIT, which most likely dissociated during the purification procedure, as noted in other cases47. The amino acid sequences used for model building for most ATP synthase subunits were obtained through annotation of A. franciscana genome (Table S2 as shown in Fig.23). The final structure of A. franciscana ATP synthase has a subunit composition similar to mammalian ATP synthases46, 47, 48(Fig.2a, b). It contains F1 and FO domains connected by the central and peripheral stalk subunits. F1contains the α3β3complex and central stalk subunits γ, δ and ε. The non-catalytic and catalytic nucleotide binding sites are located in α and β subunits, respectively. FO contains the c8-ring and subunits a (6 or ATP6), 8 (ATP8), e, f, g, and 6.8PL (6.8 kDa proteolipid). The peripheral stalk subunits are oligomycin sensitivity-conferring protein (OSCP), b, d, and F6. The 6.8PL-subunit sequence was not found in the A. franciscana genome annotation database. Thus, the poly-A chain was used to build the model of this subunit (Table S2 as shown in Fig.23). The sequence of the IF1 subunit was also not found in the annotation database. The Danio rerio (zebrafish) model of IF1 (AlphaFold database) was used for fitting into the appropriate density in the A. franciscana ATP synthase map (Fig.2b, Table S2 as shown in Fig.23). ATP synthase is unique and uses the rotation of its subunits to perform the catalytic reaction of ATP synthesis and hydrolysis44, 54, 55. The proton translocation through FO, at the interface of subunits a and c, drives the rotation of the rotor subunits (c-ring and the central stalk) during the ATP synthesis reaction55. According to the "binding change mechanism," the clockwise rotation of γ subunit within the α3β3hexamer (viewed from the FOside) drives the conformational changes in αβ pairs. Subsequent changes in β catalytic nucleotide-binding site affinities result in ATP synthesis56, 57. The β-subunits cycle through the conformational changes during ATP synthesis by adopting an open conformation (empty, βEstate), “loose” closed conformation (ADP-bound, βDP state), and “tight” closed conformation (ATP bound, βTP) states. These events occur in the opposite direction (βE→βTP→βDP) during the ATP hydrolysis reaction upon counter-clockwise rotation of γ subunit within the α3β3hexamer (viewed from the FOside)56, 57. In the maps of A. franciscana ATP synthase (State 2), each non-catalytic nucleotide- binding site in α-subunits is occupied by an ATP molecule with an accompanying Mg2+ion (Fig.2a, b). They are named according to the β-subunits they interact with. The catalytic nucleotide-binding site in the βTP subunit has ~50% occupancy with ATP and ~50% with ADP, based on the weak density of the gamma phosphate present on this site. This can be explained by the partial hydrolysis of the endogenous ATP to ADP in the βTPsite during protein purification performed without added nucleotides. The nucleotide-binding site in the βDP subunit is occupied with ADP, and the site on βE is empty in all the maps (Fig.2a, b). Therefore, the βEsubunit is found in an open conformation in all identified rotational states, while the βTPand βDPsubunits are in the closed conformation. The significant structural differences between the βTP and βDP subunits, related to the "tight" and "loose" conformations found in mammalian ATP synthases44, have not been observed in A. franciscana ATP synthase maps (Fig.10a, b). The intrinsic inhibitory factor IF1 is present in the A. franciscana ATP synthase maps for all three rotational states (State 1, State 2, and State 3a) in the pH 7.0 dataset. IF1 forms inactive monomers, active dimers, and inactive tetramers at different pH values in mammalian species58, 59. It forms active dimers at pH 7.0, where each IF1 monomer can interact with one ATP synthase molecule. The C-terminal regions of IF1 participate in dimer formation, while the N-terminal regions constitute the inhibitory part of the protein. The N- terminal region occupies the grove formed by the C-terminal domains of α- and β-subunits in bovine ATP synthase59. The N-terminal region of bovine IF1 was reported to be intrinsically disordered when it is not bound to F1. It folds in stages into the α-helical structure upon binding to ATP synthase, reaching its extensively folded state in the fully inhibited ATP synthase complex57. In A. franciscana ATP synthase, IF1 is localized at the interface between αDPand βDPsubunits in all rotational states (Fig.2, Fig.10c-f). The N-terminal region of IF1 (residues 4-20) is unfolded in A. franciscana ATP synthase structure (Figs.2 and S4c-f ). At the same time, residues 27 to 45 form a fully folded α-helical C-terminal region and a partially folded segment (residues 21 to 26) located close to the γ subunit. The IF1 dimer density was found in some F1local refined maps of A. franciscana ATP synthase (Fig.10f). Due to the low density of this region, it remains inconclusive whether the association of IF1 dimers to ATP synthase monomer is a specific feature of A. franciscana. The structure of A. franciscana ATP synthase peripheral stalk The comparative structural analysis of mammalian and A. franciscana ATP synthases revealed differences in peripheral stalk conformation and its interaction with F1and FOdomains (Fig.11). The A. franciscana and human ATP synthase maps (State 2) were aligned in relation to the central stalk for structural comparisons of the peripheral stalk and FOdomain (Fig.11a). The peripheral stalk in human ATP synthase is shifted towards the central stalk (CS) compared to A. franciscana, which has a more relaxed conformation (Fig.11a, left panel). A similar shift, shown by the white arrow in Fig.11a (right panel), is observed in the membrane-embedded FOsubunits in human ATP synthase. The comparison of A. franciscana and human subunits d and αTP did not show an apparent structural difference when individual chains were aligned together (Fig. S5b). Subunits d and αTPare shifted towards the central stalk in human compared to A. franciscana ATP synthase when the structures are aligned by the CS (Fig.11c, d). The α-helical segment of the A. franciscana d-subunit (dA1-dL42) has more upright conformation than the corresponding segment in the human d-subunit (dL5-dA43), which is tilted towards αTPand CS (Fig.11c, d). In contrast to human structure, αTP in A. franciscana is moved towards the d-subunit, which shortens the distance between the αTPand d chains (Figs.11c, d and 12a, c). The shorter distance facilitates interactions between the N- and C-terminal domains of αTPwith the subunits d and b in A. franciscana (Fig.12c, d). The N-terminal αE7 residue interacts with dT61 through a hydrogen bond and may form a salt bridge with bR166 (Fig.12b). dT61 may also form a hydrogen bond with bE162 (Fig.12b). This interaction is absent in human ATP synthase due to alanine residue (A61) at the dT61 position (Fig.12d). A hydrogen bond may form between human dN59 and bQ162 (Fig.12d). The C-terminal regions of αTPand d-subunits interact through a salt bridge between dR39 and αE467 in A. franciscana (Fig.12b). This interaction is absent in human ATP synthase (Fig.12d). The comparison of A. franciscana and human b subunits shows a more relaxed conformation of the b-subunit segment containing residues P1-H145 in A. franciscana when ATP synthase structures are aligned by the CS (Fig.11e). On the contrary, the corresponding segment in the human b-subunit (P3 to V147) has a more strained structure (Fig.11e, left panel). This difference is less apparent when individual b chains are aligned (Fig.11e, right panel). Stronger interactions are observed between the subunits a and b in A. franciscana, compared with human ATP synthase. The α-helical segment of a-subunit in A. franciscana (a181-218, shown in blue in Fig.13a) has a more linear structure compared with the corresponding segment in human ATP synthase (residues a185-225 shown in blue in Fig. 13b), which is more curved. A disulfide bond may form between the aM196 and bM66 in A. franciscana. Two hydrogen bonds may form between the bT59 and aQ191 in A. franciscana structure. Hydrophobic interactions occur between the a-subunit segment V188-M196 and bA62-M66 in A. franciscana (Fig.13a). In human ATP synthase, the observed interactions are the salt bridge between the bD56 and aH172 (Fig.13b). A hydrogen bond may form between bS62 and aT200. Additionally, π-π stacking interactions may form between the bF58 and aF193 in human ATP synthase. Stronger interactions between subunits a and b observed in A. franciscana may contribute to the more linear conformation of its b-subunit, as described above. Two native cardiolipin molecules (CDL1 and CDL2) were found in the A. franciscana FOdomain between subunits a, b, f, and g (Fig.13c). CDL1 is located between the subunits a, b, and f, interacting with the α-helical domains aH3, aH6, cH1, bH3, and fH4. The second cardiolipin molecule, CDL2, was found between the subunits b, f, and g and interacts with the helical domains fH1, fH3, fH4, bH1, bH2, and gH3 (Fig.13c). Cardiolipin molecules were found at the dimer interface of ATP synthases in different species48, 60, and have previously been reported to play a key role in stabilizing membrane protein oligomers61. However, CDL molecules were also found in the FO domain of human ATP synthase monomer47. The observed differences in interactions between the peripheral stalk subunits b, d, and αTP in F1, as well as subunits a and b in FO, may contribute to a more rigid structure of the peripheral stalk in A. franciscana ATP synthase. The N-terminal helices of all α subunits have clear densities, suggesting stronger interactions with OSCP Despite the overall structural similarities of the A. franciscana OSCP with its mammalian counterpart46, 47, differences in OSCP interaction with α and peripheral stalk subunits were observed (Fig.3). A. franciscana OSCP contains N- and C-terminal domains, connected by the short linker (Fig.3a). The N-terminal domain has six helices (H1-6) and interacts with the F1 α subunits. The C-terminal domain contains a β-hairpin followed by two α helices, H7 and H8 (Fig.3a). It is involved in the interaction with the peripheral stalk subunits F6, d, and b (Fig.3a, b, e). All three α-subunits in the A. franciscana ATP synthase maps have stronger densities for the first N-terminal α-helices. In contrast, the N-terminal α-helix in at least one of three α subunits was found to lack the clear density in the cryo-EM maps of most mammalian ATP synthases46, 47, 48, 49. In the State 2 map of H. sapiens (human, EMD-34581) ATP synthase, the N-terminal densities were present for only αTPand αDPand absent for the αEsubunit47compared to the State 2 map of A. franciscana ATP synthase (Fig.3b, c). All three N-terminal helices were omitted from the State 3a model of human ATP synthase (PDB:8H9U) due to weak or no densities found in αTP,αDP,and αEsubunits. Additionally, the density for the αTPsubunit was absent from the Ovis aries (ovine, EMD-10573) State 1 map (Fig.3d) and human, bovine, and porcine ATP synthase maps (State 1 or 1a)47, 48, 49. Strong densities in cryo-EM maps usually correspond to more rigid regions within the protein. The presence of strong densities for all three α-subunit N-terminal helices in A. franciscana suggests that they are less flexible due to their enhanced interactions with the OSCP and other stator stalk subunits compared to mammalian ATP synthases. In the State 2 model of A. franciscana, the N-terminal helix of αDPclosely interacts with the OSCP H1 and H5 helices (Fig.3e). A salt bridge and a hydrogen bond may form between the αDP E13 and OSCP H1 K28 residues and αDP E7 and H1 Y18, respectively (Fig. 14a, left panel). Other salt bridges may form between the αDPE7 and R95 located in the loop connecting H5 with H6 and between αDPK15 and E92 in H5 (Fig.14a, left panel). There are also hydrophobic interactions between αDP and OSCP H1 and H5 helices that involve the following residues: αA10, L12, I16, L17, A19, and the OSCP L4 residue, OSCP H1 residues Y18, L22, A25 and H5 residues F86 and L89 (Fig.14a, left panel). The N-terminal helix of αE interacts with the OSCP H3 and H4 helices (Fig.3e). A salt bridge may form between the αE13 and OSCP R51 (Fig.14a, right panel). Hydrophobic interactions involve the residues of αEand OSCP H3 and H4 helices: αL12, I16, L17, A19, A20, and the H3 residues L52, L55. Other residues involved in hydrophobic interactions are H4 L65, A69, M73, and L61, located in the loop connecting the H3 and H4 helices. These interactions have not been found in human ATP synthase due to the lack of clear density for the αEsubunit N-terminal segment (State 2)46, 47(Fig.3c). The N-terminal helix of the αTP subunit that is in down conformation interacts with the subunits d, b, F6, and C-terminal domain of OSCP (H8 helix) in A. franciscana (Fig.3e). A hydrogen bond and a π-π stacking interaction may form between the OSCP H8 Y183 and F6 Q8 and the OSCP H8 Y183 and F6 F11, respectively (Fig.14b, left panel). The distance between OSCP K125 and bE190 suggests a salt bridge formation in A. franciscana (Fig.14b, right panel). The interactions between the OSCP C-terminus and the αTPN-terminal helix in A. franciscana are hydrophobic and involve the residues of the OSCP H7, H8 helices and connecting loop residues L135, F142, L163, V172, I180, A186, L187, A190, V191, M174, with the bM173, V174, I177, V178, I198, L194, L201, M204, as well as αA8, A10, V11, L12, I16, L17, L25 and F6 residues F11, L12, I15 (Fig.14b). The role of the OSCP subunit in ATP synthase leak channel activation and mPTP formation has been shown recently34, 62, 63. The binding of CypD to the mammalian ATP synthase OSCP subunit activates the ATP synthase leak channel. In contrast, CsA, the binding partner of CypD, inhibits the channel activation in mammals22, 64. However, the binding of CypD to OSCP in mammalian ATP synthase is only supported by immunoprecipitation, mutagenesis, and molecular docking studies22, 46, 52, 63, with no structural information about the binding site in OSCP. Based on the ZDOCK server predictions, OSCP has to bend between its N- and C-terminal domains to accommodate CypD binding close to the conserved His11246(His 113 in A. franciscana) (Fig.3a). Moreover, protonation of His 112 was shown to mediate the mPTP inhibition65. A. franciscana mitochondria do not have CsA-sensitive and Ca2+-induced mPTP, despite the presence of CypD in the mitochondrial matrix40. The amino acid sequence and three-dimensional structural comparisons of the A. franciscana and human ATP synthase OSCP subunits revealed a less conserved H2 helix in A. franciscana (Fig.3a, f). The H2 helix has more polar residues (Q, N, and S) and is shifted towards the βE subunit in A. franciscana (Fig.3a). Additionally, there is a slight shift in the H3 helix in A. franciscana structure towards the βE compared to the human structure (Fig.3a). A salt bridge may form between the OSCP H3 R51 and αE E13 stabilizing the αE in the up conformation in A. franciscana (Fig.14a, right panel). This interaction and the overall density for the αEN-terminal helix are absent in human ATP synthase47. Previous studies have reported the role of the 14 conserved amino acid residues of OSCP in forming the putative binding site for CypD22. The amino acid sequence alignment of A. franciscana and mammalian OSCP subunits revealed that 6 of the 14 residues are not conserved in A. franciscana (Fig.3a, f). These differences in the amino acid composition and structure of A. franciscana OSCP may interfere with CypD binding and explain the lack of sensitivity to CsA observed in this organism. These observations show that the weaker interactions between the α subunit and peripheral stalk subunits found in mammalian ATP synthases could lead to higher flexibility of this region, initiating the transmission of conformational changes from F1to FOto activate the channel. In contrast, the strong N-terminal densities of all α-subunits and their enhanced interactions with the peripheral stalk subunits, OSCP, d, and F6 found in A. franciscana, may potentially prevent CypD or Ca2+-induced conformational changes in F1, propagation of the signal through the peripheral stalk to FOrequired for the ACLC activation. Distinct structural features of A. franciscana e-subunit may contribute to ACLC inactivation The FOdomain of A. franciscana has the same subunit composition and overall structure as the mammalian ATP synthases. It contains the c8-ring and a-subunit (subunit 6 or ATP6) that form the proton translocation pathway, as well as subunits 6.8PL, subunit 8 (ATP8), e, f, and g (Fig.2a, b). The density of the DAPIT subunit was not found in our maps. The disclosure demonstrates key differences in A. franciscana e-subunit amino acid sequence while comparing it with its counterpart in human, ovine, bovine, zebrafish, Drosophila melanogaster. The e-subunit sequence alignment revealed that the A. franciscana e-subunit has the most extended sequence among all the species, containing 84 amino acid residues. It is longer by 14 amino acid residues compared to mammalian e-subunits (Fig.4a). Furthermore, a stronger density of the e-subunit C-terminus was observed in the A. franciscana maps on the contrary to mammalian structures (Fig.4b-e). The density of the e- subunit was strong enough to be seen in the 2D classes of A. franciscana ATP synthase (Fig. 4c). The C-terminus was found to interact with the c-ring and lipids occupying its cavity, while the N-terminus interacts with subunits g, f, and b (Figs.2b and 4b, d, e, g). Significantly weaker densities for the FO e-subunit C-terminus were reported in the cryo-EM maps of mammalian ATP synthases, which are suggestive of higher flexibility in this region in mammals (Fig.4d). In addition, the C-terminus of the A. franciscana e-subunit has a different conformation; it is more curved and extends towards the c-ring lumen compared with the ovine ATP synthase46(Fig.4d, e). The structural differences in the A. franciscana e-subunit could contribute to the enhanced interactions with the c-ring and lipids. The distances are suggestive of the following interactions between subunit e and the c-ring: Salt bridges between eK61 and cD1 residues, and eK79 and cD1, and a hydrogen bond between the eY71 and cD1 may form in A. franciscana FO(Fig.4e). The residues eK79 and eY71 may also interact with the c-ring lipids (Fig.4e). The complete nature of the lipids occupying the A. franciscana c-ring remains to be identified. We have fitted phosphatidylserine (matrix side) and lyso-phosphatidylserine (intermembrane space side) into the densities of the c-ring lumen for visualization purposes (Fig.4e). The lipids were omitted from the deposited models of A. franciscana ATP synthase. These lipids were provisionally modeled into the density of ovine ATP synthase c-ring previously46. Similar lipid densities were found to occupy the cavity of other mammalian ATP synthase c-rings47, 48. Stable interactions between subunits e and c may interfere with the rotation of the c- ring, enabling the e-subunit to act as an IF1 of FOto stop the futile rotation of the c-ring. However, the aqueous environment of the intermembrane space may facilitate dynamic interactions between the e-subunit and c-ring, allowing the unimpeded rotation of the c-ring. In the case of mammalian ATP synthases, the eY71 residue is substituted by isoleucine (I69 in ovine). In contrast, eK61 is substituted with arginine, another positively charged residue (R59 in ovine)46, 47(Fig.4a); however, this interaction cannot form in mammals due to the larger distance between the subunits e and c. The K79 in A. franciscana is not present in mammals since the mammalian e-subunits are shorter, containing 69-71 residues (Fig.4a, e). The conserved C-terminal lysine residue is present in all mammalian e- subunits, which was suggested to interact with the aspartate residues at the bottom of the c- ring or with the lipids occupying it46, 48. In comparison with A. franciscana, weaker interactions between the e-subunit C-terminus and c-ring are present in mammals. The e-subunit N-terminus interacts with the subunits g and b (Fig.2). The important interaction site between the e- and g-subunits was reported to be the conserved GXXXG motifs present in both subunits47(Fig.4a, f, g). These motifs form the compact triple transmembrane bundle with the b-subunit transmembrane helix, similar to mammalian ATP synthases46. In contrast to human ATP synthase47, the C-terminus of the g-subunit in A. franciscana is shifted up by one alpha-helical turn, which aligns the GXXXG motifs of e- and g-subunits together (Fig.4g). This shift reduces the distance between the subunits and may favor the hydrophobic interactions (Fig.4g). The amino acid residues before the GXXXG region in subunits e and g are hydrophobic, which can further stabilize the interactions between these two subunits (Fig.4a, f, g). The GXXXG residues were selected to measure the distances between subunits e and g in A. franciscana and human ATP synthases (Fig.4g). The distance between eG28 Cα and gG103 C is 4.5 Å in A. franciscana model (State 2, PDB:9B0X). In contrast, the distance between the eG26 Cα and gG93 C is 7.8 Å in human ATP synthase (State 2, PDB:8H9J). The C-terminal helical region of the b-subunit that interacts with the C-terminus of g is also shifted up by one helical turn in A. franciscana (Fig.4g). The interaction between gE90 and bK48 in human ATP synthase is absent in A. franciscana. Meanwhile, another hydrogen bond may form between gC91 and bS47 in A. franciscana. This interaction is absent in human and ovine ATP synthases (Fig.4g) since the Cys residue in the GXXXG region of the g-subunit in A. franciscana is substituted with the isoleucine in mammals (Fig. 4f, g). D. melanogaster is the only other species that has Cys at the g91 position (Fig.4f). Interestingly, it also shares common properties of ATP synthase leak channel42, 66with A. franciscana, and is reported to lack the high-conductance channel activity of ACLC. Overall, distinct structural features of the A. franciscana e-subunit and its enhanced interactions with the subunits c, g, b, and lipids within the c-ring lumen can contribute to the structural rigidity of the FOdomain and inhibition of the ATP synthase c-subunit channel in this organism. Proton translocation pathway in FOThe a-subunit (6 or ATP6) has a universal structure in mammalian species46, 47, 67. It is also well-conserved in A. franciscana, containing six horizontal α-helixes (α1-α6) that closely interact with the c-ring and define the proton translocation pathway within the FO. The following conserved residues are involved in the proton translocation in mammalian ATP synthases46, 47, 48and are considered important for coupling ATP synthesis or hydrolysis with proton transport. The cE58 and aR159 residues serve as a "checkpoint" along with several negatively charged residues located within the a-subunit inlet (aE203) and outlet half- channels (aE145 and aD224). The latter residues participate in direct protonation- deprotonation of cE5846, 47(Fig.15a, b). During ATP synthesis, the c-ring rotates clockwise within the hydrophobic lipid bilayer (viewed from the FO side). The cE58 residue becomes protonated by accepting a proton from the aE203 when it reaches the hydrophilic inlet half channel of the a-subunit46, 47. The inlet half-channel at the inner membrane starts with the aH168 and aH172 residues and ends with the a203E46, 67in mammals (Fig15a, right panel). The mammalian residues cE58 and aH168 are conserved in A. franciscana (cE58 and aH166), while the aH172 is replaced with threonine (aT170) (Figs.15a and 16). A similar pattern is observed in D. melanogaster (Fig.16). The c-ring rotation continues until the protonated cE58 releases the proton into the matrix via the outlet half-channel of a-subunit47(Fig.15b, right panel). The conserved aE145 and aD224 residues (aE143 and aE219 in A. franciscana) in the outlet half-channel are essential for accepting protons from cE58 before it encounters the positively charged residue aR15946(aR157 in A. franciscana) (Fig.15b, left panel). The aR159 is located in the α5 helix of the human a-subunit and serves as a barrier to prevent uncoupling within the enzyme (Fig.15b, right panel). The interaction between cE58 and aR159 is essential. It ensures that the E58 is stripped of its proton before it leaves the outlet half-channel and can enter the subsequent protonation-deprotonation cycle. A salt bridge, however, does not form between the cE58-aR159 pair in mammalian ATP synthase46(Fig.15b, right panel), in contrast to V-ATPases68, 69, which were reported to have the second conserved barrier arginine residue in this region70. Similar to V-ATPases, the A. franciscana a-subunit also has two barrier arginine residues, R150 and R157, that separate the inlet and outlet half-channels of the a-subunit (Figs.15b, left panel, and 16a). A salt bridge may form between cE58 and aR157 in A. franciscana (Fig.15b, left panel). The key direct proton donor / acceptor glutamates are also present in both half-channels in A. franciscana a-subunit (E198 in the inlet and E219 in the outlet) (Figs.15 and 16a). These residues help to regulate the proton translocation in a certain rotational direction to ensure the coupling between the catalysis and proton translocation within the enzyme. In ovine ATP synthase46, the water molecules were also proposed to be a proton acceptor in the outlet half-channel. The triad formed by Y221, S148, and Q152 was suggested to help coordinate the water molecules near the barrier aR15946. The residues aS146, aY216, and R150 form this triad in A. franciscana (Fig.15b, left panel). Based on the distance, a salt bridge may form between the residues aR150 and aY216. The overall structure of A. franciscana ATP synthase at pH 8.0 The binding of intrinsic inhibitory factor 1, IF1, to ATP synthase occurs in a pH- dependent manner in mammals58, 59. Thus, we determined the structure of A. franciscana ATP synthase at pH 8.0 to study if similar pH-dependent changes occur in brine shrimp. A. franciscana mitochondria were solubilized with the non-ionic detergent DDM, followed by ATP synthase purification at pH 8.0. For the pH 8.0 dataset, we obtained two 3D maps for the rotational states 1 and 2 with nominal resolutions of 2.53 Å and 2.87 Å, respectively (Figs.17 and 18, Table S3, as shown in Fig.24). Further processing of maps for States 1 and 2 was performed to separate rotational states. We then obtained maps for States 1, 2, and 3a (Figs.17 and 18), among which State 2 was the most prevalent under these conditions. The State 2 map was used for further analysis to generate the F1, FO, and Peripheral Stalk (PS) local refined maps with a nominal resolution of 2.64 Å, 3.39 Å, and 4.01 Å, respectively. The local refinement was also performed for the F1, FO, and PS for the other two rotational states 1 and 3a (Fig.17). The overall structure and subunit composition of A. franciscana ATP synthase at pH 8.0 (Fig.19a, b) is similar to the one obtained at pH 7.0 (Fig.2). The main difference was the lack of IF1 in pH 8.0 structure in all rotation states, confirming the pH-dependent binding of IF1 to A. franciscana ATP synthase. In the maps of A. franciscana ATP synthase (State 2, pH 8.0), each non-catalytic nucleotide-binding site in α subunits is occupied by an ATP molecule with an accompanying Mg2+ion (Fig.10b). The catalytic nucleotide-binding site in the βTP subunit has ~50% occupancy with ATP and ~50% with ADP, based on the weak density of the gamma phosphate present on this site, as was observed in the pH 7.0 structure (Fig.10b). The nucleotide-binding sites in the βDP and βE subunits are empty in all the maps. Therefore, the βDP and βE subunits are found in an open conformation in all identified rotational states, while the βTPsubunit is in the closed conformation (Fig.10b). The βDPsubunit is in the closed state in the presence of IF1 in the ATP synthase structure determined at pH 7.0. In contrast, it is in the open state in the absence of IF1 (pH 8.0 dataset, Fig.10a-e). The disclosure shows that the overall structure of ATP synthase is more stable at pH 7.0 compared with 8.0. Although most ATP synthase particles show proper assembly at pH 8.0, one of the 3D classes (State 2, Class 3, 223,705 particles) does not have an apparent density for the peripheral stalk subunits b and d (Figs.17 and 19c, d). In addition to the peripheral stalk, the density for the subunits a, g, f, and 6.8PL is poorly defined, indicating higher flexibility and instability in these regions at pH 8.0 (Fig.19c, d). Interestingly, a well- defined density for the e-subunit C-terminus is still present, even in the absence of apparent density for most of the peripheral stalk and FO subunits (Fig.19c, d). This further suggests the strong nature of interactions between the e-subunit and the c-ring, which may contribute to the inactivation of the ATP synthase leak channel in A. franciscana. We also performed a comparative functional analysis of the single-channel activity of A. franciscana ATP synthase purified at different pH values. No significant differences were found in the peak conductance and open channel duration of A. franciscana ATP synthase in samples purified at pH 7.0 and 8.0 (Fig. S1f-k). Similarly, Ca2+was still unable to induce channel activation in ATP synthase purified at pH 8.0 (Fig. S1h). ATP synthase e-subunit peptides inhibit the c-subunit channel activity Human ATP synthase c-ring constitutes the pore-forming unit of ACLC. It forms voltage-gated large conductance channels with 1.5 nS peak conductance activity during the patch-clamp and planar lipid bilayer recordings32, 35. We investigated if the peptide corresponding to the e-subunit C-terminal sequence would modulate the channel activity of the c-ring, given their strong interactions in A. franciscana. The c-subunit has an almost identical sequence and oligomeric state in A. franciscana and human ATP synthases (Fig.16b). We, therefore, used purified human c- subunit in our studies. Fig.5a shows a continuous lipid bilayer recording of the human c- subunit. The channel is activated upon voltage application and becomes fully inhibited upon the addition of the C-terminal peptide of A. franciscana e-subunit (Peptide I) in all recordings (n=7) (Fig.5a, f, g, h). We also tested the effect of human e-subunit C-terminal peptide (Peptide II) on the c-subunit channel activity despite the weaker interactions between e- subunit and c-ring in mammals compared with A. franciscana. Peptide II fully inhibited the c- subunit channel in 5 recordings (Fig.5b, f, g, i). The channel was partially inhibited in the other 4 recordings (Fig. S1l), confirming possible weak interactions between the human e- subunit and the c-ring. The human e-subunit is anchored to the membrane-embedded subunits of FO and has a shorter length than the A. franciscana e-subunit, which may restrict its interaction with the c-ring in in vivo conditions. However, during planar lipid bilayer recordings, the peptide may freely roam and interact with the c-ring, leading to channel inhibition. Residues eR59 and eK69 may interact with cD1 in human ATP synthase. However, according to the electrostatic potential map of the human e-subunit, its C-terminus is more negatively charged, which may create repulsive forces between the atoms of the e- subunit and the c-ring (Fig.5i). In contrast, the C-terminus of A. franciscana e-subunit has both negative and positive charges, which may strengthen the electrostatic complementarity and interactions between these two subunits (Fig.5h). This disclosure shows that human ATP synthase e-subunit C-terminal sequence, Peptide II (KRIARELAEDDSILK (SEQ ID NO:2), also inhibits the leak channel (Fig.5i). The A. franciscana e-subunit N-terminal peptide (Peptide III), which does not have any interactions with the c-ring in our maps, was used as a negative control. Peptide III did not inhibit the c-subunit channel activity in any of the recordings (n=8) (Fig.5c, f, g, h), suggesting the requirement of specific interactions in channel inhibition. In addition, two other peptides with sequences corresponding to human IF1 were used as negative controls (Peptide IV and V, Fig.5j). Peptide IV did not inhibit the c-subunit channel activity in 7 recordings, while it induced inhibition in 3 (Fig.5d, f, g). Peptide V inhibited the channel in 1 recording, partially inhibited it in 2, and did not inhibit it in 7 recordings (Fig.5e, f, g). These results suggest that the interactions between e and c are essential for modulating channel activity. The weaker interactions between the human e-subunit and c-ring may become easily disrupted under pathological conditions inducing channel activation. On the contrary, stronger interactions observed in A. franciscana may prevent channel activation, suggesting the role of e-subunit as an inactivation gate of the ATP synthase c-subunit leak channel. Artemia Peptide protects human Müller cells from hyperglycemia-induced mitochondrial permeability transition pore (mPTP) opening. We used calcein / cobalt assay to assess the protective properties of the Artemia Peptide on mitochondrial permeability transition pore (mPTP) opening in Mio-M1 human Müller cells by using the Image-IT™ LIVE Mitochondrial Transition Pore Assay Kit (Invitrogen) as described previously (PMCID: PMC5959205). During this assay, cells are loaded with the acetoxymethyl ester of calcein dye, calcein AM, which passively diffuses into the cells and accumulates in cytosolic compartments, including the mitochondria. The intracellular esterases then cleave the acetoxymethyl esters to liberate the very polar fluorescent dye calcein, which is detected during the measurements. The fluorescence from cytosolic but not mitochondrial calcein is quenched by the addition of CoCl2. The quenching of mitochondrial calcein and subsequent loss of mitochondrial calcein fluorescence occurs only upon activation of mPTP. Mio-M1 cells were incubated with calcein and CoCl2for 15 min after 24 hours of exposure to either hyperglycemic conditions (HG; 30 mmol / L glucose) or an osmotic control (OC; 5 mmol / L glucose plus 25 mmol / L mannitol). Cell mitochondria and nuclei were counterstained with MitoTracker Red CMXRos and Hoechst 33342 dye for 15 min. Fluorescence was imaged using the Leica SP8 confocal microscope (Leica) with frame stack sequential scanning. Overlapped signal of calcein-stained area with MitoTraker red-stained area indicated mitochondria with closed mPTP. Where indicated, cells were co-treated with the described TAMRA-tagged peptide prior to calcein-cobalt chloride assay. As shown in Fig. 20, prior treatment of Müller cells with Peptide rescues glucose-induced mPTP opening, suggesting the crucial role of ATP synthase c-subunit leak channel (ACLC) in hyperglycemia- induced cell death. Artemia Peptide protects rat retinal cells (R28) from oxidative stress-induced and mPTP-regulated cell death. Rat retinal cells were treated with hydrogen peroxide (100 ^M) to induce oxidative stress. This stress activates the opening of the mitochondrial permeability transition pore, releasing cytochrome c and activating the downstream apoptotic cell death pathways. The Cell Counting Kit-8 (CCK-8) was used to determine the number of viable and dead cells in the colorimetric assay. Concurrent incubation of cells with hydrogen peroxide and Artemia Peptide prevents H202-induced and mPTP-regulated cell death in R28 cells. The data are shown in Fig.21. Discussion The exact molecular identity, composition, and structure of mPTP are still debated after several decades of extensive research. Recent studies suggested the plausible role of mitochondrial ATP synthase as a structural pore-forming unit of mPT22, 23, 25, 26, 29, 32, 33, 34, 35, 42,46, 52, 71, 72, 73. Nevertheless, the exact activation and inactivation mechanisms of the ATP synthase leak channel are not fully understood. Different gating models of ATP synthase leak channel were proposed recently based on structural, electrophysiology, and cell death studies8, 35, 42, 46, 74. A large-conductance channel was suggested to form either by the ATP synthase c-ring32, 33, 35, 75or at the interface of two monomers22, 23, 26. The mPTP modulators, CypD and Ca2+, were shown to directly bind to ATP synthase subunits OSCP22, 52, 65, and β34, respectively, and induce conformational changes within the F1domain. These changes may cause the detachment of F1from FO,activating the leak channel within the c-ring from the matrix side and highlighting the role of F1 as a gate of ACLC (Fig.6a). The signal may then propagate from F1 to the membrane- embedded FO subunits e and g, through the peripheral stalk subunits OSCP, d, and b to activate the channel from the intermembrane space (Fig.6a). Destabilization within the c-ring structure, due to detachment of F1and changes in FO,may occur, allowing widening of the pore followed by the channel activation35, 75. A “death finger” model was proposed, highlighting the role of the e-subunit C-terminus in pulling the lipids out from the cavity of the c-ring74. The triple transmembrane helix bundle formed by the helices of subunits b, g, and N-terminal helix of e, known as “hook apparatus,” was also suggested to facilitate lipid removal and channel activation46. The occlusion of the c-ring cavity with lipids has a crucial biological function in maintaining the impermeability of the inner mitochondrial membrane and generating the proton gradient required for coupling respiration with ATP synthesis56. Changes in inner membrane permeability due to prolonged activation of mPTP are known to result in necrotic and apoptotic cell death in mammals76. In contrast, the brine shrimp A. franciscana is resistant to anoxia and Ca2+-induced mPTP activation40, 43. The channel formed by A. franciscana ATP synthase dwelt predominantly in its inhibited state. It demonstrated transient, Ca2+-insensitive openings in planar lipid bilayer recordings, contrary to porcine heart ATP synthase. Ca2+, a key activator of the ATP synthase leak channel in mammals, binds to the βT163, known to coordinate Mg2+ions during catalysis34. This residue is conserved in A. franciscana (βT160), suggesting that Ca2+may bind to ATP synthase. Ca2+binding, however, may not trigger conformational changes within ATP synthase necessary to activate the channel in A. franciscana (Fig.6b). Enhanced interactions between the α and peripheral stalk subunits, OSCP, d, and F6 most likely do not allow the initiation and transmission of the signal from F1to FOpreventing channel activation in A. franciscana (Fig.6b). In addition, the A. franciscana OSCP-subunit has a distinct amino acid sequence from that of mammals.6 of the 14 conserved amino acid residues of OSCP, reported crucial in forming the putative binding site for CypD in mammals22, are not conserved in A. franciscana (Fig.3a, f). These differences may interfere with the CypD binding to OSCP and explain the lack of CsA-sensitive mPTP in this organism. A. franciscana ATP synthase e-subunit was found to be longer by 14 amino acid residues when compared with its mammalian counterparts. It includes 84 residues and has a more curved C-terminal region that reaches towards the c-ring and the lipid plug (Fig.4a-d). The C-terminal peptide of A. franciscana e-subunit inhibited the c-subunit channel activity in electrophysiology recordings, suggesting the role of e-subunit as a second gate of ACLC that acts in the intermembrane space. The CRISPR-Cas9 genome editing studies of ATP synthase e-subunit in A. franciscana would provide further insights about the causative effect of e- subunit in contributing to the lack of Ca2+-sensitive mPTP in this organism. The ATP synthase e-subunit in D. melanogaster is also longer compared with mammals. It has 81 residues and shares similarities with A. franciscana, including a Cys residue in the g-subunit GXXXG motive (Fig.4a, f). Purified ATP synthase from D. melanogaster forms small, only 53 pS channels66and lacks the large-conductance, 1.5 nS channel activity, further confirming the possible role of e-subunit in ACLC inactivation. Furthermore, the e-subunit in Trypanosoma brucei ATP synthase is 92 amino acids long, with its C-terminus having a similar conformation to that of A. franciscana. It was reported to interact with the ten-stranded β-barrel structure of the c-ring60. The role of the ATP synthase leak channel and its contribution to mPTP still needs to be elucidated in this protozoan parasite. The structure and conformation of the e-subunit may have evolved distinctly in different species to contribute to the physiological and pathological functions of the ATP synthase leak channel and its role in mPT. mPTP plays an important role in cardiomyocyte differentiation during embryogenesis77. Although mPTP opening is generally associated with cell death, a non- pathologic opening of mPTP was reported in the early embryonic heart77. Mouse heart cardiomyocytes at embryonic day (E) 9.5 displayed fragmented mitochondria with fewer cristae, depolarized mitochondrial membrane potential, higher levels of reactive oxygen species (ROS), and an open mPTP. Closure of mPTP at E11.5 lowered ROS and drove structural maturation of mitochondria, followed by cardiomyocyte differentiation77. Without intending to be bound by any particular theory, it is considered that that the shorter length of the e-subunit may enable physiological functions of mPTP during embryogenesis in mammals, meanwhile increasing the probability of channel activation and predisposing them to cell death under pathological conditions. In contrast, organisms like crustaceans and insects may have evolved to possess more rigid ATP synthase structures with a lower probability of channel opening due to the harsh environmental conditions they need to sustain. The link between the ATP synthase and mPTP channel has been recently questioned by a study that used genetic knockout of ATP synthase subunits g or F6 in HAP1-A12 cells78. Mitochondria isolated from HAP1-A12 Δg and ΔF6 cells were markedly sensitized to Ca2+- induced mPTP opening in CRC assay. In addition, the g-subunit was genetically depleted from the mouse cardiomyocytes78. Increased cardiomyocyte necrosis and larger infarcts were observed in Δg mice compared to controls78. These results allowed the authors to conclude that ATP synthase is a negative regulator of mPTP since its depletion sensitized mPTP opening78. It was assumed that the deletion of g and F6 subunits has entirely eliminated ATP synthase from the cell. Indeed, the levels of fully assembled ATP synthase oligomers, dimers, and monomers were significantly reduced upon deletion of g and F6 subunits. Yet, increased levels of the assembly intermediates containing the c-ring were found in mitochondria78. The high levels of “free” c-ring, not assembled with F1, were reported to increase the probability of ACLC activation, predisposing neurons to cell death under glutamate excitotoxic conditions35. Furthermore, it has been reported that the deletion of any of the supernumerary membrane subunits of FO, such as subunits e, g, DAPIT, and 6.8PL, is accompanied by the loss of the remaining supernumerary subunits79. This suggests that subunit g deletion in HAP1-A12 cells and engineered mice78would also reduce the levels of the other supernumerary subunits, including the e-subunit. Consequently, enhanced levels of “free” c- ring without either of its gates, F1and e-subunit, would be present in mitochondrial inner membranes. This would increase the probability of c-subunit leak channel activation, predisposing cardiomyocytes to cell death. Exacerbated necrotic damage was indeed observed in vivo in cardiomyocyte-specific Δg mice during the myocardial infarction induced by ischemia / reperfusion injury. These findings support the current gating model of ACLC, suggesting that disassembly of ATP synthase dimers to monomers, followed by further disassembly of monomers to ‘free” c-ring, occurs during ischemic insults35. Aggravated necrotic damage reported in Δg mice also indirectly supports the described model, emphasizing the role of e-subunit in inactivating the ATP synthase leak channel in A. franciscana. By possessing distinct structural features, A. franciscana ATP synthase achieves a higher level of structural rigidity that enables this crustacean to withstand harsh environmental conditions that typically trigger the mPTP opening and pose challenges to mammalian organisms. Main References: 1. Haworth RA, Hunter DR. The Ca2+-induced membrane transition in mitochondria. II. Nature of the Ca2+ trigger site. Arch Biochem Biophys 1979, 195(2): 460-467. 2. Hunter DR, Haworth RA. The Ca2+-induced membrane transition in mitochondria. I. The protective mechanisms. Arch Biochem Biophys 1979, 195(2): 453-459. 3. Kinnally KW, Campo ML, Tedeschi H. Mitochondrial channel activity studied by patch-clamping mitoplasts. J Bioenerg Biomembr 1989, 21(4): 497-506. 4. Szabo I, Zoratti M. The mitochondrial megachannel is the permeability transition pore. J Bioenerg Biomembr 1992, 24(1): 111-117. 5. Petronilli V, Miotto G, Canton M, Brini M, Colonna R, Bernardi P, et al. Transient and long-lasting openings of the mitochondrial permeability transition pore can be monitored directly in intact cells by changes in mitochondrial calcein fluorescence. Biophys J 1999, 76(2): 725-734. 6. Crompton M, Virji S, Doyle V, Johnson N, Ward JM. The mitochondrial permeability transition pore. Biochem Soc Symp 1999, 66: 167-179. 7. Mnatsakanyan N, Beutner G, Porter GA, Alavian KN, Jonas EA. Physiological roles of the mitochondrial permeability transition pore. J Bioenerg Biomembr 2017, 49(1): 13- 25. 8. Bernardi P, Gerle C, Halestrap AP, Jonas EA, Karch J, Mnatsakanyan N, et al. Identity, structure, and function of the mitochondrial permeability transition pore: controversies, consensus, recent advances, and future directions. Cell Death Differ 2023, 30(8): 1869- 1885. 9. Yu CH, Davidson S, Harapas CR, Hilton JB, Mlodzianoski MJ, Laohamonthonkul P, et al. TDP-43 Triggers Mitochondrial DNA Release via mPTP to Activate cGAS / STING in ALS. Cell 2020, 183(3): 636-649 e618. 10. Chen R, Park HA, Mnatsakanyan N, Niu Y, Licznerski P, Wu J, et al. Parkinson's disease protein DJ-1 regulates ATP synthase protein components to increase neuronal process outgrowth. Cell Death Dis 2019, 10(6): 469. 11. Szabo I, Zoratti M. The mitochondrial permeability transition pore may comprise VDAC molecules. I. Binary structure and voltage dependence of the pore. FEBS Lett 1993, 330(2): 201-205. 12. Szabo I, De Pinto V, Zoratti M. The mitochondrial permeability transition pore may comprise VDAC molecules. II. The electrophysiological properties of VDAC are compatible with those of the mitochondrial megachannel. FEBS Lett 1993, 330(2): 206- 210. 13. Kwong JQ, Davis J, Baines CP, Sargent MA, Karch J, Wang X, et al. Genetic deletion of the mitochondrial phosphate carrier desensitizes the mitochondrial permeability transition pore and causes cardiomyopathy. Cell Death Differ 2014, 21(8): 1209-1217. 14. Pastorino JG, Simbula G, Gilfor E, Hoek JB, Farber JL. Protoporphyrin IX, an endogenous ligand of the peripheral benzodiazepine receptor, potentiates induction of the mitochondrial permeability transition and the killing of cultured hepatocytes by rotenone. J Biol Chem 1994, 269(49): 31041-31046. 15. Chelli B, Falleni A, Salvetti F, Gremigni V, Lucacchini A, Martini C. Peripheral-type benzodiazepine receptor ligands: mitochondrial permeability transition induction in rat cardiac tissue. Biochem Pharmacol 2001, 61(6): 695-705. 16. Li J, Wang J, Zeng Y. Peripheral benzodiazepine receptor ligand, PK11195 induces mitochondria cytochrome c release and dissipation of mitochondria potential via induction of mitochondria permeability transition. Eur J Pharmacol 2007, 560(2-3): 117-122. 17. Halestrap AP, Brenner C. The adenine nucleotide translocase: a central component of the mitochondrial permeability transition pore and key player in cell death. Curr Med Chem 2003, 10(16): 1507-1525. 18. Kokoszka JE, Waymire KG, Levy SE, Sligh JE, Cai J, Jones DP, et al. The ADP / ATP translocator is not essential for the mitochondrial permeability transition pore. Nature 2004, 427(6973): 461-465. 19. Baines CP, Kaiser RA, Sheiko T, Craigen WJ, Molkentin JD. Voltage-dependent anion channels are dispensable for mitochondrial-dependent cell death. Nat Cell Biol 2007, 9(5): 550-555. 20. Gutierrez-Aguilar M, Douglas DL, Gibson AK, Domeier TL, Molkentin JD, Baines CP. Genetic manipulation of the cardiac mitochondrial phosphate carrier does not affect permeability transition. J Mol Cell Cardiol 2014, 72: 316-325. 21. Sileikyte J, Blachly-Dyson E, Sewell R, Carpi A, Menabo R, Di Lisa F, et al. Regulation of the mitochondrial permeability transition pore by the outer membrane does not involve the peripheral benzodiazepine receptor (Translocator Protein of 18 kDa (TSPO)). J Biol Chem 2014, 289(20): 13769-13781. 22. Giorgio V, von Stockum S, Antoniel M, Fabbro A, Fogolari F, Forte M, et al. Dimers of mitochondrial ATP synthase form the permeability transition pore. Proc Natl Acad Sci U S A 2013, 110(15): 5887-5892. Carraro M, Giorgio V, Sileikyte J, Sartori G, Forte M, Lippe G, et al. Channel formation by yeast F-ATP synthase and the role of dimerization in the mitochondrial permeability transition. J Biol Chem 2014, 289(23): 15980-15985. Carraro M, Checchetto V, Sartori G, Kucharczyk R, di Rago JP, Minervini G, et al. High-Conductance Channel Formation in Yeast Mitochondria is Mediated by F-ATP Synthase e and g Subunits. Cell Physiol Biochem 2018, 50(5): 1840-1855. Guo L, Carraro M, Carrer A, Minervini G, Urbani A, Masgras I, et al. Arg-8 of yeast subunit e contributes to the stability of F-ATP synthase dimers and to the generation of the full-conductance mitochondrial megachannel. J Biol Chem 2019, 294(28): 10987- 10997. Urbani A, Giorgio V, Carrer A, Franchin C, Arrigoni G, Jiko C, et al. Purified F-ATP synthase forms a Ca(2+)-dependent high-conductance channel matching the mitochondrial permeability transition pore. Nat Commun 2019, 10(1): 4341. Pavlov E, Zakharian E, Bladen C, Diao CT, Grimbly C, Reusch RN, et al. A large, voltage-dependent channel, isolated from mitochondria by water-free chloroform extraction. Biophys J 2005, 88(4): 2614-2625. Neginskaya MA, Solesio ME, Berezhnaya EV, Amodeo GF, Mnatsakanyan N, Jonas EA, et al. ATP Synthase C-Subunit-Deficient Mitochondria Have a Small Cyclosporine A-Sensitive Channel, but Lack the Permeability Transition Pore. Cell Rep 2019, 26(1): 11-17 e12. Bonora M, Bononi A, De Marchi E, Giorgi C, Lebiedzinska M, Marchi S, et al. Role of the c subunit of the FO ATP synthase in mitochondrial permeability transition. Cell Cycle 2013, 12(4): 674-683. Morciano G, Preti D, Pedriali G, Aquila G, Missiroli S, Fantinati A, et al. Discovery of Novel 1,3,8-Triazaspiro[4.5]decane Derivatives That Target the c Subunit of F1 / FO- Adenosine Triphosphate (ATP) Synthase for the Treatment of Reperfusion Damage in Myocardial Infarction. J Med Chem 2018, 61(16): 7131-7143. Azarashvili T, Odinokova I, Bakunts A, Ternovsky V, Krestinina O, Tyynela J, et al. Potential role of subunit c of F0F1-ATPase and subunit c of storage body in the mitochondrial permeability transition. Effect of the phosphorylation status of subunit c on pore opening. Cell Calcium 2014, 55(2): 69-77. Alavian KN, Beutner G, Lazrove E, Sacchetti S, Park HA, Licznerski P, et al. An uncoupling channel within the c-subunit ring of the F1FO ATP synthase is the mitochondrial permeability transition pore. Proc Natl Acad Sci U S A 2014, 111(29): 10580-10585. 33. Mnatsakanyan N, Llaguno MC, Yang Y, Yan Y, Weber J, Sigworth FJ, et al. A mitochondrial megachannel resides in monomeric F1FO ATP synthase. Nat Commun 2019, 10(1): 5823. 34. Giorgio V, Burchell V, Schiavone M, Bassot C, Minervini G, Petronilli V, et al. Ca(2+) binding to F-ATP synthase beta subunit triggers the mitochondrial permeability transition. EMBO Rep 2017, 18(7): 1065-1076. 35. Mnatsakanyan N, Park HA, Wu J, He X, Llaguno MC, Latta M, et al. Mitochondrial ATP synthase c-subunit leak channel triggers cell death upon loss of its F(1) subcomplex. Cell Death Differ 2022, 29(9): 1874-1887. 36. Karch J, Bround MJ, Khalil H, Sargent MA, Latchman N, Terada N, et al. Inhibition of mitochondrial permeability transition by deletion of the ANT family and CypD. Sci Adv 2019, 5(8): eaaw4597. 37. He L, Lemasters JJ. Regulated and unregulated mitochondrial permeability transition pores: a new paradigm of pore structure and function? FEBS Lett 2002, 512(1-3): 1-7. 38. Yoon Y, Lee H, Federico M, Sheu SS. Non-conventional mitochondrial permeability transition: Its regulation by mitochondrial dynamics. Biochim Biophys Acta Bioenerg 2023, 1864(1): 148914. 39. Zoratti M, Biasutto L, Parrasia S, Szabo I. Mitochondrial permeability transition pore: a snapshot of a therapeutic target. Expert Opin Ther Targets 2024, 28(1-2): 1-3. 40. Menze MA, Hutchinson K, Laborde SM, Hand SC. Mitochondrial permeability transition in the crustacean Artemia franciscana: absence of a calcium-regulated pore in the face of profound calcium storage. Am J Physiol Regul Integr Comp Physiol 2005, 289(1): R68-76. 41. Konrad C, Kiss G, Torocsik B, Adam-Vizi V, Chinopoulos C. Absence of Ca2+- induced mitochondrial permeability transition but presence of bongkrekate-sensitive nucleotide exchange in C. crangon and P. serratus. PLoS One 2012, 7(6): e39839. 42. Frigo E, Tommasin L, Lippe G, Carraro M, Bernardi P. The Haves and Have-Nots: The Mitochondrial Permeability Transition Pore across Species. Cells 2023, 12(10). 43. Clegg J. Embryos of Artemia franciscana survive four years of continuous anoxia: the case for complete metabolic rate depression. J Exp Biol 1997, 200(Pt 3): 467-475. 44. Zhou A, Rohou A, Schep DG, Bason JV, Montgomery MG, Walker JE, et al. Structure and conformational states of the bovine mitochondrial ATP synthase by cryo-EM. Elife 2015, 4: e10180. 45. Tobias S, Montgomery M, Walker J. Structure of the dimeric ATP synthase from bovine mitochondria. PNAS 2020. 46. Pinke G, Zhou L, Sazanov LA. Cryo-EM structure of the entire mammalian F-type ATP synthase. Nat Struct Mol Biol 2020. 47. Lai Y, Zhang Y, Zhou S, Xu J, Du Z, Feng Z, et al. Structure of the human ATP synthase. Mol Cell 2023, 83(12): 2137-2147 e2134. 48. Spikes TE, Montgomery MG, Walker JE. Structure of the dimeric ATP synthase from bovine mitochondria. Proc Natl Acad Sci U S A 2020, 117(38): 23519-23526. 49. Gu J, Zhang L, Zong S, Guo R, Liu T, Yi J, et al. Cryo-EM structure of the mammalian ATP synthase tetramer bound with inhibitory protein IF1. Science 2019, 364(6445): 1068-1075. 50. Elrod JW, Wong R, Mishra S, Vagnozzi RJ, Sakthievel B, Goonasekera SA, et al. Cyclophilin D controls mitochondrial pore-dependent Ca(2+) exchange, metabolic flexibility, and propensity for heart failure in mice. J Clin Invest 2010, 120(10): 3680- 3687. 51. Baines CP, Kaiser RA, Purcell NH, Blair NS, Osinska H, Hambleton MA, et al. Loss of cyclophilin D reveals a critical role for mitochondrial permeability transition in cell death. Nature 2005, 434(7033): 658-662. 52. Giorgio V, Bisetto E, Soriano ME, Dabbeni-Sala F, Basso E, Petronilli V, et al. Cyclophilin D modulates mitochondrial F0F1-ATP synthase by interacting with the lateral stalk of the complex. J Biol Chem 2009, 284(49): 33982-33988. 53. Antoniel M, Giorgio V, Fogolari F, Glick GD, Bernardi P, Lippe G. The oligomycin- sensitivity conferring protein of mitochondrial ATP synthase: emerging new roles in mitochondrial pathophysiology. Int J Mol Sci 2014, 15(5): 7513-7536. 54. Watt IN, Montgomery MG, Runswick MJ, Leslie AG, Walker JE. Bioenergetic cost of making an adenosine triphosphate molecule in animal mitochondria. Proc Natl Acad Sci U S A 2010, 107(39): 16823-16827. 55. Leslie AG, Abrahams JP, Braig K, Lutter R, Menz RI, Orriss GL, et al. The structure of bovine mitochondrial F1-ATPase: an example of rotary catalysis. Biochem Soc Trans 1999, 27(2): 37-42. 56. Boyer PD. The ATP synthase--a splendid molecular machine. Annu Rev Biochem 1997, 66: 717-749. 57. Bason JV, Montgomery MG, Leslie AG, Walker JE. How release of phosphate from mammalian F1-ATPase generates a rotary substep. Proc Natl Acad Sci U S A 2015, 112(19): 6009-6014. 58. Cabezon E, Butler PJ, Runswick MJ, Walker JE. Modulation of the oligomerization state of the bovine F1-ATPase inhibitor protein, IF1, by pH. J Biol Chem 2000, 275(33): 25460-25464. 59. Cabezon E, Arechaga I, Jonathan P, Butler G, Walker JE. Dimerization of bovine F1- ATPase by binding the inhibitor protein, IF1. J Biol Chem 2000, 275(37): 28353- 28355. 60. Gahura O, Muhleip A, Hierro-Yap C, Panicucci B, Jain M, Hollaus D, et al. An ancestral interaction module promotes oligomerization in divergent mitochondrial ATP synthases. Nat Commun 2022, 13(1): 5989. 61. Gupta K, Donlan JAC, Hopper JTS, Uzdavinys P, Landreh M, Struwe WB, et al. The role of interfacial lipids in stabilizing membrane protein oligomers. Nature 2017, 541(7637): 421-424. 62. Giorgio V, Fogolari F, Lippe G, Bernardi P. OSCP subunit of mitochondrial ATP synthase: role in regulation of enzyme function and of its transition to a pore. Br J Pharmacol 2018. 63. Carraro M, Jones K, Sartori G, Schiavone M, Antonucci S, Kucharczyk R, et al. The Unique Cysteine of F-ATP Synthase OSCP Subunit Participates in Modulation of the Permeability Transition Pore. Cell Rep 2020, 32(9): 108095. 64. Giorgio V, Fogolari F, Lippe G, Bernardi P. OSCP subunit of mitochondrial ATP synthase: role in regulation of enzyme function and of its transition to a pore. Br J Pharmacol 2019, 176(22): 4247-4257. 65. Antoniel M, Jones K, Antonucci S, Spolaore B, Fogolari F, Petronilli V, et al. The unique histidine in OSCP subunit of F-ATP synthase mediates inhibition of the permeability transition pore by acidic pH. EMBO Rep 2018, 19(2): 257-268. 66. von Stockum S, Giorgio V, Trevisan E, Lippe G, Glick GD, Forte MA, et al. F-ATPase of Drosophila melanogaster forms 53-picosiemen (53-pS) channels responsible for mitochondrial Ca2+-induced Ca2+ release. J Biol Chem 2015, 290(8): 4537-4544. 67. Guo H, Bueler SA, Rubinstein JL. Atomic model for the dimeric F(O) region of mitochondrial ATP synthase. Science 2017, 358(6365): 936-940. 68. Zhou L, Sazanov LA. Structure and conformational plasticity of the intact Thermus thermophilus V / A-type ATPase. Science 2019, 365(6455). 69. Roh SH, Stam NJ, Hryc CF, Couoh-Cardel S, Pintilie G, Chiu W, et al. The 3.5-A CryoEM Structure of Nanodisc-Reconstituted Yeast Vacuolar ATPase V(o) Proton Channel. Mol Cell 2018, 69(6): 993-1004 e1003. 70. Mazhab-Jafari MT, Rohou A, Schmidt C, Bueler SA, Benlekbir S, Robinson CV, et al. Atomic model for the membrane-embedded V(O) motor of a eukaryotic V-ATPase. Nature 2016, 539(7627): 118-122. 71. Carraro M, Bernardi P. Calcium and reactive oxygen species in regulation of the mitochondrial permeability transition and of programmed cell death in yeast. Cell Calcium 2016, 60(2): 102-107. 72. Giorgio V, Guo L, Bassot C, Petronilli V, Bernardi P. Calcium and regulation of the mitochondrial permeability transition. Cell Calcium 2018, 70: 56-63. 73. Parks RJ, Menazza S, Holmstrom KM, Amanakis G, Fergusson M, Ma H, et al. Cyclophilin D-mediated regulation of the permeability transition pore is altered in mice lacking the mitochondrial calcium uniporter. Cardiovasc Res 2019, 115(2): 385-394. 74. Gerle C. On the structural possibility of pore-forming mitochondrial FoF1 ATP synthase. Biochim Biophys Acta 2016, 1857(8): 1191-1196. 75. Mnatsakanyan N, Jonas EA. ATP synthase c-subunit ring as the channel of mitochondrial permeability transition: Regulator of metabolism in development and degeneration. J Mol Cell Cardiol 2020, 144: 109-118. 76. Kwong JQ, Molkentin JD. Physiological and pathological roles of the mitochondrial permeability transition pore in the heart. Cell Metab 2015, 21(2): 206-214. 77. Hom JR, Quintanilla RA, Hoffman DL, de Mesy Bentley KL, Molkentin JD, Sheu SS, et al. The permeability transition pore controls cardiac mitochondrial maturation and myocyte differentiation. Dev Cell 2011, 21(3): 469-478. 78. Pekson R, Liang FG, Axelrod JL, Lee J, Qin D, Wittig AJH, et al. The mitochondrial ATP synthase is a negative regulator of the mitochondrial permeability transition pore. Proc Natl Acad Sci U S A 2023, 120(51): e2303713120. 79. He J, Ford HC, Carroll J, Douglas C, Gonzales E, Ding S, et al. Assembly of the membrane domain of ATP synthase in human mitochondria. Proc Natl Acad Sci U S A 2018, 115(12): 2988-2993. 80. Taussky HH, Shorr E. A microcolorimetric method for the determination of inorganic phosphorus. J Biol Chem 1953, 202(2): 675-685. 81. De Vos S, Rombauts S, Coussement L, Dermauw W, Vuylsteke M, Sorgeloos P, et al. The genome of the extremophile Artemia provides insight into strategies to cope with extreme environments. BMC Genomics 2021, 22(1): 635. 82. Mirdita M, Schutze K, Moriwaki Y, Heo L, Ovchinnikov S, Steinegger M. ColabFold: making protein folding accessible to all. Nat Methods 2022, 19(6): 679-682. 83. Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, et al. Highly accurate protein structure prediction with AlphaFold. Nature 2021, 596(7873): 583- 589. Materials and Methods: Isolation of mitochondria from A. franciscana, HEK 293 cells, and porcine heart The dried eggs of A. franciscana were hydrated for 4 h in 0.25 M NaCl solution and then decanted and treated with Antiformin solution (1% Hypochlorite and 0.4 M NaOH) for 10 min. Eggs were then collected and washed with 1% sodium thiosulphate solution for 5 minutes. Further, washed eggs were incubated for ~14 h in 0.25 M NaCl solution in the shaker incubator before being harvested. The collected A. franciscana was resuspended in the Isolation Buffer A (500 mM sucrose, 150 mM KCl, 1 mM EGTA, 20 mM HEPES, 0.5 % BSA, pH 7.5) and homogenized with pre-chilled Dounce homogenizer on ice. The homogenate was then centrifuged at 1000 g for 10 min at 4oC, and the resulting supernatants were collected and centrifuged again at 1000 g for 10 min at 4oC. The supernatant from the above step was then centrifuged at 9000 g for 20 min at 4oC. The resulting mitochondrial pellets were then resuspended in the Isolation Buffer B (500 mM sucrose; 150 mM KCl; 20 mM HEPES, 0.5 % BSA, pH 7.5) and centrifuged at 9000 g for 20 min at 4oC. The final mitochondrial pellets were then resuspended in 50-100 µl of Isolation Buffer C (500 mM sucrose; 150 mM KCl, 1 mM KH2PO4, 20 mM HEPES, 25 µM EGTA, pH 7.5) used in different assays. Mitochondria from HEK 293 cells (obtained from ATCC) were isolated by using the Qproteome mitochondria isolation kit (Qiagen) following the manufacturer’s protocol. Fresh porcine hearts were obtained immediately after slaughter, and the heart was finely minced in Buffer 1 (225 mM mannitol, 70 mM sucrose, 1 mM EGTA, and 20 mM Tris (pH 7.2)). The tissue was then homogenized with an Elvehjem potter, followed by centrifugation at 1000 g for 4 min. The supernatant was transferred into a fresh tube, and the sediment was homogenized for a second time to access intermyofibrillar mitochondria. After centrifugation at 1000 g, the supernatants were centrifuged at 9000 g for 10 min to sediment the mitochondria. The mitochondria were resuspended in Buffer 2 (225 mM mannitol, 70 mM sucrose, and 20 mM Tris (pH 7.2)) and centrifuged again at 9000 g for 10 min. The final mitochondrial pellet was suspended in 0.4 mL of Buffer 2. Isolated mitochondria were used immediately or stored at −80°C until further use. Calcium Retention Capacity Assay Calcium Retention Capacity (CRC) assay with HEK293 mitochondria (0.7 ^g / ^l) was performed in CRC buffer (150 mM sucrose; 1 mM KH2PO4, 1 mM MOPS, 5 mM succinate, 10 µM EGTA, pH 7.4) containing 20 µM rotenone (MP Biomedicals), and 1 µM Calcium Green-5N (Life Technologies). The following buffer was used for mitochondria isolated from A. franciscana (0.7 ^g / ^l) for CRC assay: 500 mM sucrose, 150 mM KCl, 1 mM KH2PO4, 20 mM HEPES, 5 mM succinate, 10 ^M EGTA, pH 7.4, containing 20 µM rotenone and 1 µM Calcium Green-5N.10 µM CsA (Sigma) was used when indicated. Data was acquired with a microplate reader (SpectraMax-iD3; Molecular Devices).50 µM calcium pulses were added every 2 min while the fluorescence data were recorded at every 10 s interval. All the assays were done at 23oC. Electron Microscopy Isolated mitochondria were fixed in a solution comprising 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M phosphate buffer for 1 hour (pH 7.4). The buffer-rinsed mitochondria were subsequently combined with 2% agarose and centrifuged. Following this, the chilled blocks were trimmed and further fixed in 1% osmium tetroxide for an additional 1 hour. The samples underwent triple rinsing in 0.1 M phosphate rinse buffer. Subsequently, a graduated ethanol series and pure acetone were used for dehydration before embedding in LX-112 (Ladd Research, Williston, VT). Thin sections of a sample with 60 nm thickness were stained with uranyl acetate and lead citrate, then observed using a JEOL JEM1400 Transmission Electron Microscope (JEOL USA Inc., Peabody, MA, USA) at 60KV, located at the Penn State College of Medicine TEM Facility (RRID Number: SCR_021200). Images were captured using a NanoSprint43 Mk-II camera (Advanced Microscopy Techniques, Woburn, MA). Purification of human c-subunit, and A. franciscana and porcine heart ATP synthases The human ORF ATP-synthase c1 (ATP5G1) subunit construct tagged on the C- terminus with Myc and DDK (Flag) was used (Origene Technologies). The construct for c- subunit was expressed in HEK 293 cells and purified using the EZ view Red ANTI-FLAG M2 Affinity Gel (Sigma), according to the manufacturer's protocol. Mitochondria isolated from porcine heart or A. franciscana were used for ATP synthase purification. Mitochondria were solubilized on ice by using DDM (1 g / g protein) for 1h. Then, the suspension was centrifuged at 100,000 × g for 1 h. Subsequently, 50% (w / w) PEG 6000 was added to the supernatant (final concentration, 7%). After 2 h incubation on ice, the protein precipitate was removed by centrifugation at 50,000 × g for 1 h and PEG 6000 was subsequently added to the supernatant (final concentration, 2%). The protein precipitate was collected after incubation for 2 h on ice and centrifuged at 50,000 × g for 1 h. The pellet, which contains ATP synthase, was solubilized in Buffer A (50 mM Tris-HCl, 100 mM NaCl, 2 mM MgSO4, 1 mM ATP, 0.05% DDM, pH 7.0 or pH 8.0). The solubilized sample was concentrated to 0.5 ml (Amicon centrifugal filters with 100 kDa molecular mass cutoff) and loaded onto a Superose 6 Increase 10 / 300 gel filtration column (GE Healthcare, USA) equilibrated with Buffer A. Fractions eluted at a retention volume of 11–12 ml were collected and used for further analysis. ATP hydrolysis activity measurements ATPase activities were assayed in a buffer containing 50 mM Tris / H2SO4, 10 mM ATP, and 4 mM MgSO4, pH 8.0. The reaction was started by the addition of 5–10 μg ATP synthase solubilized in detergent and stopped after 10 min by the addition of SDS (final concentration, 10%, w / v). The released Piwas measured as described80. One unit of enzymatic activity (U) corresponds to 1 μmol of ATP hydrolyzed (equivalent to 1 μmol of Piproduced) per min, per mg protein. The effect of oligomycin on ATP hydrolysis activity was determined by adding the inhibitor to an ethanol solution (5 μg / ml, w / v). Peptide synthesis and validation The peptides corresponding to the A. franciscana e-subunit C-terminal (LYLAKETGVKVPENF SEQ ID NO:1)) and N-terminal (MSFAPPVNVSPLIRA (SEQ ID NO:3)) sequences, as well as human e-subunit C-terminal sequence (KRIARELAEDDSILK (SEQ ID NO:2)) were synthesized at the Penn State College of Medicine Macromolecular Synthesis Core. Peptides corresponding to the IF1 sequence GAGSIREAGGAFGKR (SEQ ID NO:4) and FRAQSREQLAALKKH (SEQ ID NO:5) were synthesized at LifeTein LLC. The peptide masses were validated by MALDI-MS. Planar Lipid Bilayer Recordings Planar lipid bilayer recordings were performed in intracellular solution (120 mM KCl, 8 mM NaCl, 0.5 mM EGTA, 10 mM HEPES, pH 7.3) by using the α-L-phosphatidylcholine (Sigma) or α-L-phosphatidylcholine and cardiolipin (Avanti) in 3:1 ratio for forming the bilayer membrane. ePatch amplifier (Elements) was used for planar lipid bilayer recordings. Signals were filtered at 5 kHz using the amplifier circuitry. Purified ATP synthase or c- subunit were added on the cis side, and a constant voltage was applied to achieve protein insertion into the bilayer. ATP (1 mM), bongkrekic acid (BA, 10 µM), calcium (1 mM), and peptides I, II, III, IV and V (5 µM) were added on the cis side of the cuvette during the recordings without perfusion. To access the channel activity on both the voltage polarities, a voltage ramp was performed where the voltage was changed from -100 to +100 mV within 60 seconds. Clampfit software (Molecular Devices) was used for data analysis. The measured current was adjusted for the holding voltage, assuming a linear Current-Voltage relationship. The conductance (G) is expressed in pS, following equation G = I / V, where I is the peak membrane current in pA, and V is the membrane holding voltage in mV. Group data were quantified in terms of conductance and open channel duration. All population data were expressed as mean ± SEM. Sample preparation for single-particle cryo-EM Grids were pretreated with Glow discharge under vacuum at 20 mA for 20 sec. A 3ul aliquot of ATP synthase purified at pH 7.0 or 8.0 was applied onto a clean QuantiFoil R 2 / 1 Cu 300 mesh grid covered with 2nm carbon. The grids were vitrified at 20°C and 100 % humidity using Mark IV Vitrobot (ThermoFisher). The excess solution was removed by 5s of blotting before plunging into liquid ethane (-182˚C). Single-particle cryo-EM data collection and processing CryoEM data for high-resolution analysis were acquired using a FEI Titan Krios G2 microscope (Thermo Fisher Scientific) operating at 300 kV, equipped with a Post-GIF K3 summit direct detection camera (TFS) and a Bio Quantum energy filter (20 eV slit width). The data collection process was automated by employing SerialEM. For the dataset at pH 7.0, a total of 12,547 movies were collected. These included 6,4400° tilt movies and 6,10620° tilt movies, each composed of 56 fractions resulting in a cumulative exposure time of 2.8 seconds, with 0.05 seconds per fraction. For the pH 8.0 dataset, a total of 12,097 movies were obtained. These included 5,6330° tilt movies and 6,46420° tilt movies, each composed of 60 fractions resulting in a cumulative exposure time of 3.6 seconds, with 0.06 seconds per fraction. The nominal magnification was ×81,000, corresponding to a calibrated pixel size of 1.068 Å. The specimen was exposed at a rate of 15.8 electrons per pixel per second, with a total exposure of approximately 50 electrons per square angstrom. The defocus ranged from - 1 to -1.7 μm for 0° tilt movies and from -0.7 to -1.5 μm for 20° tilt movies. The nominal magnification was ×81,000, corresponding to a calibrated pixel size of 1.068 Å. The specimen was exposed at a rate of 20.9 electrons per pixel per second, with a total exposure of approximately 50 electrons per square angstrom. The defocus ranged from -1 to -1.7 μm for 0° tilt movies and from -0.7 to -1.3 μm for 20° tilt movies. All image processing procedures were executed using cryoSPARC, unless otherwise specified. Movie fractions were aligned utilizing MotionCor2 with a 5 × 5 grid, and parameters for the contrast transfer function (CTF) were assessed through patch CTF estimation in cryoSPARC. Particle selection was conducted employing a template-based approach, utilizing a previously resolved 4.3 Å porcine ATP synthase map to generate 502D projection images. The diameter of the particle was set to 180 Å and these were extracted using a box size of 400 pixels. Contamination and outliers were initially removed by running 3 rounds of 2D classification. "Clean" particles were used to generate an initial 3D map using the Ab initio algorithm. For the pH 7.0 dataset, a total of 5.5 million particles were initially selected and subjected to a round of 2D classification using 4x binned particles, resulting in 2.7 million particles retained for subsequent 3D processing. These 2,683,006 particles were subsequently re-extracted with a voxel size of 512. A subset of 50,000 particles was utilized for reference- free ab initio reconstruction to serve as the reference map for subsequent heterogeneous refinement. Heterogeneous refinement led to the separation of three major rotational states, comprising 864,418 state1 particles, 847,418 state3 particles, and 970,828 state2 particles. For each state, NU-refinement was employed, involving per-particle CTF refinement, spherical aberration fitting, tetrafoil fitting, and anisotropic magnification fitting, resulting in global structures with resolutions of 2.6 Å for state1, 2.7 Å for state3, and 2.6 Å for State 2. To obtain high-resolution FOmaps for State 1, particle subtraction was performed using an FOmask to subtract the F1signal from the particle stack, preserving only the FOinformation. A subsequent round of heterogeneous refinement was executed, resulting in the selection of 673,185 FOparticles. Local refinement yielded a 3.3 Å state1 FOmap suitable for atomic model building. For State 3 and State 2 FOmaps, focused 3D classification processes with FOmask were employed instead of particle subtraction. This process selected 637,185 State 3 particles and 722,737 State 2 particles, followed by FO local refinement, resulting in 3.5 Å state1 and State 2 FO maps. The 3D classification was conducted for the particle stacks of each rotational state to further segregate sub-states or identify IF1 bound / unbound conformations. For State 1, four classes with 237,890 particles, 229,181 particles, 217,964 particles, and 179,725 particles were generated through the 3D classification process. Subsequent FOor F1local refinement provided FOmaps with resolutions ranging from 3.4 Å to 4.8 Å and F1maps with resolutions between 2.6 Å and 3.6 Å. For State 3, a similar approach resulted in four classes with 252,379 particles, 233,700 particles, 187,161 particles, and 174,178 particles from the 3D classification. FOor F1local refinement produced FOmaps with resolutions between 3.8 Å and 4.9 Å and F1 maps with resolutions between 2.5 Å and 4.2 Å. For State 2, four classes with 290,002 particles, 232,736 particles, 225,659 particles, and 222,431 particles were generated through the 3D classification process. Subsequent FO or F1local refinement yielded FOmaps with resolutions ranging from 3.9 Å to 4.5 Å, and F1maps with resolutions between 2.6 Å and 3.2 Å. For the pH 8.0 dataset, a total of 5.5 million particles were initially selected and underwent a round of 2D classification with 4x binning, yielding 3.4 million particles for subsequent 3D processing. Among these, 3,381,825 particles were re-extracted with a voxel size of 512, and particles with high-resolution information were identified through two rounds of heterogeneous refinement. From the heterogeneous refinement, two major classes were distinguished, comprising 1,374,794 particles and 919,437 particles, which were further subjected to NU-refinement. This refinement involved per-particle CTF refinement, spherical aberration fitting, tetrafoil fitting, and anisotropic magnification fitting, resulting in structures with resolutions of 2.5 Å and 2.8 Å, respectively. Additionally, 3D classification was employed to segregate sub-states within both classes. For class 1, sub-classification led to the isolation of five different State 2 sub-states and one State 1 sub-state. These sub-states contained 302,409, 275,464, 223,705, 223,680, 214,385, and 135,151 particles, respectively. These were subsequently subjected to NU-refinement, yielding structures at global resolutions of 2.7 Å, 2.8 Å, 2.8 Å, 3.0 Å, 3.8 Å, and 3.0 Å. To generate high-quality maps for FO, F1, and the Peripheral stalk (PS), local refinements were performed with masked regions, resulting in maps at resolutions ranging from 3.3 Å to 4.4 Å for Fo, 2.6 Å to 3.8 Å for F1, and 3.8 Å to 4.9 Å for PS. For class 2, sub-classification revealed three different State 3 sub-states and one State 1 sub-state, comprising 284,505, 205,595, 195,711, and 233,566 particles, respectively. These were subsequently subjected to NU-refinement, resulting in structures at global resolutions of 2.8 Å, 3.8 Å, 3.7 Å, and 2.8 Å. Local refinements produced maps with resolutions ranging from 3.4 Å to 4.3 Å for FO, 2.7 Å to 4.0 Å for F1, and 3.9 Å to 4.6 Å for PS. Details of the image process workflow are described in the Figures. The sharpened maps obtained from F1 and FO or F1, FO, and PS local refinements were combined to generate composite maps in Phenix software. Annotation of A. franciscana ATP synthase genes The amino acid sequences for A. franciscana ATP synthase subunits were obtained through the annotation of ATP synthase genes by analyzing the previously obtained and assembled genome of A. franciscana81. To retrieve the target genes, we first screened the predicted A. franciscana proteome using Blastp with known representative genes. The collected hits were then aligned and compared to the known genes described in other arthropods. The predicted A. franciscana genes that aligned well were used for further analysis. A tblastn was run by using the same representative target proteins directly onto the assembled genomic scaffolds to overcome the possibility that a gene would have been missed by the gene prediction. The hits were inspected to complete the gene models (extending to start / stop codon, joining exons in the case hits were reported in several pieces (=exons)). The final annotated sequences with the highest identity percentage, alignment score, and query cover percentage were used for the atomic model building. Model building and atomic model fitting. The model building of A. franciscana ATP synthase subunits was performed by using annotated sequences as an input for the prediction in ColabFold82or Alphafold83software. The predicted models were processed to remove the signal peptides by comparing the amino acid sequence from A. franciscana with the sequences of ATP synthase from human, ovine, and bovine. The models were fitted into the cryo-EM map density through multiple rounds of rigid and flexible fitting in Coot and Phenix using real-space refinement and manual adjustments. The validity of the annotated sequences was confirmed by comparing the residues with the cryo-EM map density after fitting. We optimized the structure in the regions with poor density, comparing our map with maps and models from other ATP synthases. These structures were used as a reference during flexible fitting steps. The structures were validated using MolProbity, which mainly considered the Ramachandran plot, clash score, rotamers, and side chain outliers. Other examples of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. It is intended that the specification and examples be considered as exemplary only.

Claims

What is claimed is:

1. An isolated or recombinant adenosine triphosphate (ATP) synthase or fragment thereof that inhibits function of mammalian ATP synthase c-subunit leak channel (ACLC) but does not inhibit ATP synthesis of ATP by the mammalian ATP synthase, and wherein optionally the ATP synthase or fragment thereof comprises a modification relative to the sequence of an unmodified ATP synthase.

2. The isolated or recombinant ATP synthase or the fragment thereof of claim 1, wherein the modification is present and comprises attachment of a cellular localization amino acid sequence that is optionally a mammalian mitochondrial localization signal.

3. The isolated or recombinant ATP synthase or fragment thereof of claim 1, wherein the ATP synthase or fragment thereof comprises a segment of A. franciscana ATP synthase.

4. The isolated or recombinant ATP synthase fragment of claim 3, wherein said fragment comprises or consists of the sequence LYLAKETGVKVPENF (SEQ ID NO:1).

5. The isolated or recombinant ATP synthase fragment of claim 1, wherein the fragment comprises or consists of the sequence KRIARELAEDDSILK (SEQ ID NO:2).

6. The isolated or recombinant ATP synthase or fragment thereof of claim 1, wherein the function of the ACLC that is inhibited by said fragment comprises inhibiting opening of the ACLC.

7. The isolated or recombinant ATP synthase of any one of claims 1-6, wherein the isolated or recombinant ATP synthase comprises the sequence of SEQ ID NO:

36.

8. A pharmaceutical composition comprising the ATP synthase fragment of any one of claims 1-6.

9. A method comprising administering the pharmaceutical composition of claim 8 to an individual in need thereof to thereby provide a therapeutic or prophylactic benefit to the individual.

10. The method of claim 9, wherein the individual is in need of treatment or prophylaxis of any of the following conditions: Alzheimer’s disease; diabetes including but not limited to diabetic retinopathy and nephropathy; Amyotrophic Lateral Sclerosis (ALS); ischemia-reperfusion injury; ischemic stroke and / or brain ischemia; myocardial ischemia; mesenteric ischemia; peripheral and / or limb ischemia; intestinal ischemia; kidney ischemia; Transient Ischemic Attack (TIA); angina; ischemic heart disease; chronic obstructive pulmonary disease; arterial occlusion; anemia; blood vessel compression and / or stenosis; blood vessel rupture; coronary microvascular disease; or non-obstructive coronary artery disease; or osteoarthritis.

11. The method claim 10, wherein the ATP synthase fragment is introduced into epithelial cells, neurons, cardiomyocytes, or muscle cells.

12. The method of claim 10, wherein introducing the ATP synthase fragment into cells inhibits or prevents death of the cells into which the ATP synthase fragment is introduced.

13. The method of claim 10, wherein the ATP synthase fragment is introduced intravenously.

14. The method of claim 10, wherein the individual is a human or a non-human mammal.

15. A mammalian cell into which an ATP synthase or fragment thereof of any one of claims 1-6 has been introduced.

Images