F-ATPase of Drosophila melanogaster Forms 53-Picosiemen (53-pS) Channels Responsible for Mitochondrial Ca2+-induced Ca2+ Release*

Background: The Ca2+-induced Ca2+ release channel (mCrC) of Drosophila mitochondria is similar to the permeability transition pore (PTP). Results: mCrC is modulated by PTP effectors and Drosophila F-ATPase forms 53-pS channels. Conclusion: F-ATPase mediates Ca2+-induced Ca2+ release in Drosophila mitochondria. Significance: Channel formation by F-ATPases has been conserved in evolution, but species-specific features exist that may underscore different roles in different organisms. Mitochondria of Drosophila melanogaster undergo Ca2+-induced Ca2+ release through a putative channel (mCrC) that has several regulatory features of the permeability transition pore (PTP). The PTP is an inner membrane channel that forms from F-ATPase, possessing a conductance of 500 picosiemens (pS) in mammals and of 300 pS in yeast. In contrast to the PTP, the mCrC of Drosophila is not permeable to sucrose and appears to be selective for Ca2+ and H+. We show (i) that like the PTP, the mCrC is affected by the sense of rotation of F-ATPase, by Bz-423, and by Mg2+/ADP; (ii) that expression of human cyclophilin D in mitochondria of Drosophila S2R+ cells sensitizes the mCrC to Ca2+ but does not increase its apparent size; and (iii) that purified dimers of D. melanogaster F-ATPase reconstituted into lipid bilayers form 53-pS channels activated by Ca2+ and thiol oxidants and inhibited by Mg2+/γ-imino ATP. These findings indicate that the mCrC is the PTP of D. melanogaster and that the signature conductance of F-ATPase channels depends on unique structural features that may underscore specific roles in different species.

with 130 mM KCl, 10 mM Mops-Tris, pH 7.4 (KCl medium) containing 10 M EGTA-Tris. The resulting pellet was resuspended in KCl medium containing 150 M digitonin and 1 mM EGTA-Tris and incubated for 20 min on ice (6 ϫ 10 7 cells ϫ ml Ϫ1 ). Cells were then diluted 1:5 in KCl medium containing 10 M EGTA-Tris and centrifuged at 200 ϫ g in a refrigerated centrifuge (4°C) for 6 min. The final pellet was resuspended in KCl medium containing 10 M EGTA-Tris at 4 ϫ 10 8 cells ϫ ml Ϫ1 and kept on ice.
Isolation of Mitochondria from Flies-Mitochondria were prepared from whole flies or 3rd instar larvae by differential centrifugation exactly as described (14).
Mitochondrial Membrane Potential and Ca 2ϩ Retention Capacity-Mitochondrial membrane potential was measured using a Perkin-Elmer LS50B spectrofluorometer based on the fluorescence quenching of rhodamine 123 (15) at excitation and emission wavelengths of 503 and 523 nm, respectively, with the slit width set at 2.5 nm. Twenty million permeabilized S 2 R ϩ cells were added to the cuvette in a total volume of 2 ml. Further additions were as indicated in the figure legends. Extramitochondrial Ca 2ϩ was measured based on Calcium Green 5N (Molecular Probes) fluorescence (15) at excitation and emission wavelengths of 485 and 538 nm, respectively, using a Fluoroskan Ascent FL plate reader (Thermo Scientific) with either 200 g of isolated mitochondria or 2 ϫ 10 6  Mitochondrial Swelling Assay-Absorbance at 540 nm was monitored with a Multiskan Ex (Thermo Scientific) plate reader. Either 200 g of isolated mitochondria or 2 ϫ 10 6 permeabilized cells were added to each well in a total volume of 200 l. Incubation media are specified in the figure legends.
Fly Stocks and Breeding Conditions-Flies were raised on standard cornmeal medium and were maintained at 23°C, 70% relative humidity, on a 12-h light:12-h dark cycle. The fly strain used in all experiments was white 1118 and was obtained from the Bloomington Stock Center.
Western Blotting-Cell or mitochondrial pellets were lysed in 150 mM NaCl, 20 mM Tris, pH 7.4, 5 mM EDTA-Tris, 10% glycerol, 1% Triton X-100, supplemented with protease and phosphatase inhibitor cocktails (Sigma), and kept on ice for 20 min. Suspensions were then centrifuged at 18,000 ϫ g for 25 min at 4°C, and the supernatants were solubilized in Laemmli gel sample buffer. For separation of F-ATPase dimers, the Laemmli buffer was directly added to the dimer bands eluted from a BN-PAGE gel. Samples were separated by 15% SDS-PAGE and transferred electrophoretically to nitrocellulose membranes using an SE400 vertical electrophoresis unit (Hoefer). Western blotting was performed in PBS containing 3% nonfat dry milk with polyclonal goat anti-actin (Santa Cruz Biotechnology), polyclonal rabbit anti-caspase-3 (Cell Signaling), monoclonal mouse anti-CyPD (Calbiochem), monoclonal mouse anti-HA (Sigma), monoclonal mouse anti-F-ATPase subunit ␤ (Abcam), polyclonal rabbit anti-MCU (Sigma), or polyclonal rabbit anti-TOM20 (Santa Cruz Biotechnology) antibodies.
BN-PAGE-Pellets of mitochondria isolated from adult white 1118 flies were suspended at 10 mg ϫ ml Ϫ1 in 1ϫ native PAGE sample buffer (Invitrogen) supplemented with protease inhibitor mixture (Sigma), solubilized with 2% (w/v) digitonin, and immediately centrifuged at 100,000 ϫ g for 25 min at 4°C. The supernatants were supplemented with native PAGE 5% G-250 sample additive (Invitrogen) and quickly loaded onto a blue native polyacrylamide 3-12% gradient gel (Invitrogen). After electrophoresis, gels were processed, and F-ATPase dimers were prepared for bilayer experiments exactly as described (9).
Immunofluorescence-One day before the experiments, stably transfected S 2 R ϩ pActCyPD-HA cells were seeded on sterilized 13-mm round glass coverslips in a 24-well tissue culture plate at 2 ϫ 10 5 cells/well in 0.5 ml of culture medium. On the day of the experiment, cells were washed once with PBS and incubated for 20 min at room temperature with 0.5 ml of serum-free Schneider's medium supplemented with 1 g/ml CsH and 100 nM MitoTracker Red CMXRos (Molecular Probes). Cells were fixed with 4% paraformaldehyde for 20 min at room temperature, permeabilized with 50 mM NH 4 Cl in PBS ϩ 0.1% Triton X-100 for 5 min at room temperature, and blocked with PBS containing 3% goat serum for 1 h at room temperature with PBS washes between each step. Monoclonal anti-HA (clone HA-7, Sigma) in PBS with 2% goat serum was added, and incubation was carried out overnight at 4°C. On the next day, cells were washed with PBS, and the immunoreaction was revealed with FITC-conjugated anti-mouse IgG (Fab-specific, Sigma) in PBS with 2% goat serum for 45 min at room temperature. Coverslips were examined with an Olympus epifluorescence microscope at 60ϫ magnification.
Cell Transfection, Plasmids, and Constructs-A construct of human cyclophilin (CyP) D cDNA carrying a N-terminal mitochondrial targeting sequence from Drosophila Hsp60 (a mitochondrial matrix protein) and a C-terminal HA tag was generated by PCR using total cDNA of the human osteosarcoma cell line HQB17 as template. Oligonucleotides were designed using the on-line tool Primer-BLAST (16). Primer sequences were 5Ј-ctggtaccatgttccgtttgccagtttcgcttgctcgctcctccattagccgccagttggccatgcgcggctatgccaaggatgtgtgcagcaagggctccggcgacccg-3Ј (forward) and 5Ј-ccgagctcttaagcgtaatctggaacatcgtatgggtagctcaactggccacagtctgtgatg-3Ј (reverse). To generate a stable polyclonal cell population constitutively expressing human CyPD in Drosophila S 2 R ϩ cells, the construct was inserted into the pAct5c Drosophila expression vector under control of the Dro-sophila actin 5c promoter (a gift of M. F. Ceriani, Laboratorio de Genética del Comportamiento, Fundación Instituto Leloir and Instituto de Investigaciones Bioquímicas, Buenos Aires, Argentina) co-transfected with the selection vector pCoPuro (Addgene plasmid 17533), generated by Dr. Francis Castellino (17). Two million cells/well were plated in a 6-well tissue culture plate in 2 ml of culture medium/well. Cells were incubated for 5 h at room temperature and then transfected with the Effectene transfection reagent kit (Qiagen) with a ratio of selection vector (pCoPuro) to expression vector (pActCyPD-HA) of 1:20. After 3 days, the medium was replaced with culture medium containing 8 g/ml puromycin for selection.
Electrophysiology-Planar lipid bilayer experiments were performed as described in Ref. 18. Briefly, bilayers of 150 -200 picofarads capacitance were prepared using purified soybean asolectin. The standard experimental medium was 100 mM KCl, 10 mM Hepes, pH 7.4. All reported voltages refer to the cis chamber, zero being assigned to the trans (grounded) side. Currents are considered as positive when carried by cations flowing from the cis to the trans compartment. Freshly prepared F-ATPase dimers were added to the cis side. No current was observed when PTP activators were added to the membrane in the absence of F-ATPase dimers. The I-V curve was obtained in the range of Ϫ80 to ϩ40 mV from three independent experiments and contained current values that were measured manually (n Ͼ 30 for each potential) using the PCLAMP8.0 program set. The conductance value thus obtained was 53.0 Ϯ 7.2 pS.
Reagents and Statistics-All chemicals were of the highest purity commercially available. Reported results are typical of at least three replicates for each condition, and error bars refer to the S.E. p values were calculated with Student's t test.

RESULTS AND DISCUSSION
Drosophila S 2 R ϩ cells were permeabilized with digitonin and energized with succinate in the presence of ADP and P i ; an ATP-hydrolyzing system was present to maintain constant [ADP] and synchronize F-ATPases in the direction of ATP synthesis. A train of Ca 2ϩ pulses was then added to determine the CRC, i.e. the threshold matrix Ca 2ϩ load necessary to open the mCrC (1) (Fig. 1A, trace a). Next, we incubated permeabilized S 2 R ϩ cells in the absence of respiratory substrates, and then energized mitochondria with ATP in the presence of an ATPregenerating system to maintain constant [ATP] and synchronize F-ATPases in the direction of ATP hydrolysis (Fig. 1A, trace b). The CRC was much larger in ATP-hydrolyzing than in ATP-synthesizing mitochondria (Fig. 1, compare traces a and b). Due to formation of glucose-6-P, the concentration of P i decreases in the ATP synthesis experiments, and conversely, due to hydrolysis of P-creatine, it increases in the ATP-hydrolyzing protocols. However, in the presence of oligomycin, the CRC was only slightly higher than that observed during ATP synthesis (Fig. 1, trace aЈ), and in the presence of 5 mM P i under conditions of ATP hydrolysis, the CRC was indistinguishable from that observed at 1 mM P i (Fig. 1, trace bЈ). Thus, the difference between ATP-synthesizing and ATP-hydrolyzing mitochondria cannot be explained by differences in the P i concentration. Note that the difference in CRC was also not due to different levels of the membrane potential, which was the same in ATP-synthesizing (Fig. 1B, trace a) and ATP-hydrolyzing mitochondria (Fig. 1B, trace b), and responded appropriately to oligomycin and uncoupler.
The threshold Ca 2ϩ was increased 3-fold in ATP-hydrolyzing relative to ATP-synthesizing mitochondria (Fig. 1C), and the effect was not due to the different nucleotide as such because ADP and ATP displayed indistinguishable effects on the CRC when F-ATPase catalysis was blocked with oligomycin in the presence of Mg 2ϩ , and mitochondria were energized with succinate (Fig. 1D). We could also exclude participation of the endoplasmic reticulum because Ca 2ϩ uptake was not observed in the presence of ATP plus oligomycin (results not shown; see Ref. 6). The intriguing effect of enzyme catalysis described in Fig. 1A is also observed for the mammalian PTP (6) and represents a first indication that the mCrC may originate from the F-ATPase as well. This effect could be due to the different conformations of F-ATPase during ATP synthesis and hydrolysis (see e.g. Ref. 19) and suggests that the sense of rotation of F-ATPase influences the mCrC of Drosophila mitochondria, possibly by modulating accessibility of Ca 2ϩ -sensitive site(s) that trigger channel opening (20).
The well characterized F-ATPase inhibitor Bz-423 binds the OSCP subunit of the enzyme in mammalian mitochondria (21) and inhibits both ATP synthesis and hydrolysis (22) while sensitizing the PTP to opening (6). Interestingly, the Ca 2ϩ load necessary to trigger mCrC opening was decreased by Bz-423 in a concentration-dependent manner (Fig. 1E).
CyPD, a key regulator of the mammalian PTP, also binds the OSCP subunit of the F-ATPase, presumably at the same site as Bz-423, because the latter displaces CyPD from the enzyme complex (6). CyPD binding can be selectively inhibited by cyclosporin A (CsA) or by genetic ablation of CyPD, and both manipulations remarkably desensitize the PTP to Ca 2ϩ in that its opening requires about twice the Ca 2ϩ load necessary to open the PTP in untreated mitochondria (23). The Drosophila genome encodes 14 different CyPs (24). D. melanogaster CyP1 has an N-terminal sequence that according to Mitoprot (25) confers high probability of mitochondrial import, yet full sequence analysis led to the conclusion that no mitochondrial CyP is present in this species (24). Consistent with this conclusion, we could not detect mitochondrial localization of a CyP1-GFP fusion protein in S 2 R ϩ cells (results not shown).
We prepared a cDNA construct with an N-terminal Drosophila mitochondrial targeting sequence followed by the human CyPD coding sequence and by a C-terminal HA tag. Transfected cells expressed the human CyPD construct, which was recognized by both the CyPD and the HA antibodies, with the increase of molecular weight expected of the HA tag ( Fig.  2A, top left panel). The protein largely localized to mitochondria as judged (i) by Western blot analysis of cytosolic and mitochondrial fractions after subcellular fractionation ( Fig. 2A, bottom left panel); and (ii) by colocalization with MitoTracker Red CMXRos in fixed cells immunostained with a mouse anti-HA antibody followed by a secondary fluoresceinated antibody against mouse IgG ( Fig. 2A, right panel).
Analysis of the CRC in cells co-transfected with the expression vector pActCyPD-HA and the pCoPuro selection vector as compared with cells containing only the selection vector showed that CyPD expression reduces the threshold Ca 2ϩ load required to induce Ca 2ϩ release (Fig. 2B, compare traces a and  b). The effect of CyPD could not be counteracted by CsA (Fig.  2B, trace c), however, irrespective of the concentration of P i (Fig. 2C), which in mammalian mitochondria favors CyPD binding to the F-ATPase, resulting in pore sensitization to Ca 2ϩ (26,27). The reason why CsA does not inhibit the mCrC after enforced expression of CyPD remains unclear, but we note (i) that D. melanogaster OSCP has a larger number of negative charges in the putative CyPD binding region, which could increase binding affinity; and (ii) that PTP inhibition does not depend on inhibition of CyPD enzymatic activity (28).
At variance from the PTP, the Drosophila mCrC is impermeable to sucrose (1). Of note, the size of the Drosophila mCrC was unaffected by expression of human CyPD because no Ca 2ϩ -induced swelling could be observed after the addition of enough Ca 2ϩ to induce mCrC opening (not shown) irrespective of whether naive (Fig. 2D, trace a) or CyPD-expressing permeabilized cells (Fig. 2D, trace b) were used. S 2 R ϩ cells were derived from late embryonic stages (12). To assess whether the low exclusion size and inhibitory response to P i of the mCrC are restricted to early developmental stages, or rather are defining properties of the Drosophila channel, we also measured the swelling response in mitochondria isolated from white 1118 3rd instar larvae or adult flies. Neither mitochondrial preparation underwent Ca 2ϩ -induced swelling, which could instead be readily induced by the addition of the pore-forming peptide alamethicin (Fig. 2D, traces c and d).
In the absence of rotenone, succinate induces reverse electron flow through complex I and leads to generation of reactive oxygen species (29). Rotenone is a potent inhibitor of the mammalian PTP when succinate is used as the substrate (30), and inhibition of reverse electron flow at complex I provides a plausible mechanism for this inhibition because reactive oxygen species increase the probability of pore opening through thiol oxidation (31,32). Remarkably, rotenone increased nearly 4-fold the CRC of Drosophila mitochondria from both larvae and adults (Fig. 2E). Thus, lack of swelling after mCrC opening From these experiments, it appears extremely plausible that the mCrC is the PTP of Drosophila mitochondria despite its smaller exclusion size. Given that both the mammalian and yeast F-ATPase form channels (6,9) whose electrophysiological features are indistinguishable from those of the mitochondrial megachannel (the electrophysiological equivalent of the PTP) (33)(34)(35)(36)(37)(38), we next sought to test whether Drosophila F-ATPase forms channels, as well as the features of those channels. We purified mitochondria from adult white 1118 flies, separated proteins by BN-PAGE after digitonin extraction, and identified F-ATPase monomers and dimers by in-gel activity staining (Fig. 3A, left lanes). SDS-PAGE analysis of excised ATPase dimers from the BN-PAGE revealed the typical subunit pattern of the enzyme complex after staining with Coomassie Colloidal Blue (Fig. 3A, middle lanes). Western blot analysis of the same preparation revealed a lack of contamination by inner membrane MCU and outer membrane TOM20 (Fig. 3A, right lanes).
The dimer preparation was eluted from the gel and incorporated into asolectin planar bilayers for determination of channel activity (6). The addition of Ca 2ϩ , Bz-423, phenylarsine oxide, and copper-o-phenanthroline induced opening of a channel with a prevalent single channel conductance of 53 pS (Fig. 3, B and C), which was consistently observed in different dimer preparations. The addition of ␥-imino ATP and Mg 2ϩ induced channel closure within a few seconds, as also demonstrated by the amplitude histograms obtained from current traces before and after the addition of the modulators (Fig. 3C). Further electrophysiological experiments confirmed that Drosophila F-ATPase dimers allow the passage of Ca 2ϩ currents (results not shown). Thus, under conditions of oxidative stress and in the presence of Ca 2ϩ and Bz-423, the F-ATPase of Drosophila forms channels whose conductance is 10-fold smaller than that of the mammalian PTP (6) and 6-fold smaller than that of yeast (9). These features are a perfect match to those defined for the mCrC in permeabilized S 2 R ϩ cells (1) and leave little doubt that the mCrC is the PTP of D. melanogaster. Thus, formation of channels that require Ca 2ϩ and oxidation of critical thiol groups is a conserved feature of F-ATPases across species (6,9,20).
The finding that the Drosophila channel has a unit conductance of 53 pS heavily bears on the nature of the permeability pathway of the PTP. Jonas and co-workers (39) have recently proposed that the pore forms within F 0 after Ca 2ϩ -dependent expulsion of F 1 , which would then be closed by the ␤ subunit. F 0 is formed by identical c-subunits arranged to form a barrel, the c-ring, whose rotation is tightly linked to the proton gradient. The number of c-subunits per F-ATPase monomer varies between species. In vertebrates and most invertebrates the c-ring has 8 c-subunits, whereas in prokaryotes, chloroplasts and fungi c-rings of 10 -15 subunits have been observed (40). The N-terminal ␣-helix of members with c 8 -rings has 3 conserved Ala residues at positions 13, 19, and 23 (bovine numbering), which are essential to avoid side chain clashes (40), and the C-terminal ␣-helix has 1 conserved Lys residue at position 43, which is trimethylated in humans, pig, sheep and rabbit. This Lys appears to be the binding site for cardiolipin, which allows tight packing of c 8 -rings and is essential for enzyme activity and stability (41,42). All these residues are conserved in Drosophila F-ATPase (40), which is consistently affected by cardiolipin loss (43). These structural features indicate that the Drosophila complex belongs to the c 8 -ring F-ATPase set. It is then difficult to see how a pore forming within c-rings of the same size could mediate formation of a channel with a conductance of 500 pS in mammals and of a mere 53 pS in Drosophila, and of 300 pS in yeast where the c-ring has 10 subunits (40). Our findings are therefore not consistent with the idea that PTP channel forms within the c-ring (39), whose actual conductance and Ca 2ϩ dependence remain controversial (44 -46); see also Refs. 47-49 for discussion.
Given that we readily observe channel formation with F-ATPase dimers but not monomers (6,9), and that dimerization-resistant yeast strains display resistance to PTP opening (9), our working hypothesis is that the pathway for solute permeation forms between dimers (or higher order structures) (20). Intriguingly, cryoelectron tomography established that F-ATPase dimers from Drosophila flight muscle have a unique supramolecular structure, with two parallel rows at the highcurvature edge of cristae vesicles (43). It is possible that this supramolecular assembly prevents formation of high-conductance channels both in the native membrane and during electrophysiological experiments, where dimers could in principle be organized in higher order structures.
Depending on the open time, the mammalian PTP could be involved both in apoptosis induction and in Ca 2ϩ homeostasis (50). Prolonged openings cause mitochondrial depolarization, osmotic swelling, outer mitochondrial membrane rupture, and release of apoptogenic proteins such as cytochrome c; transient openings, on the other hand, may be involved in physiological were identified by Coomassie Blue (Coomassie) and in-gel activity staining (activity). Central lanes, total mitochondrial extracts (total) or dimers excised from the BN-PAGE gels (dimer) were subjected to SDS-PAGE and stained with colloidal Coomassie Blue; ␣, ␤, ␥, b, and OSCP refer to the position of the corresponding subunits of F-ATP synthase. Right lanes, Western blot analysis of total lysates (total) and F-ATPase dimers (dimer) for F-ATPase ␤ subunit (ATP␤), MCU, and TOM20. F-ATPase dimers eluted from BN-PAGE (see left panel) were incorporated in planar lipid bilayers. B, top recording, representative current trace recorded at Ϫ100 mV (cis) (conductance ϭ 53.0 Ϯ 7.2 pS, established from the slope of I-V curves obtained from three independent experiments) following the addition of 2 mM Ca 2ϩ plus 0.2 mM phenylarsine oxide, 0.125 mM Bz-423, and 10 M copper-o-phenanthroline (added to the cis side) in a medium containing 100 mM KCl, 10 mM Hepes, pH 7.4. Bottom recording, effect of 0.8 mM Mg 2ϩ (added to the cis side) and 0.1 mM ␥ imino ATP (added to both sides) on channel activity (Ϫ100 mV). C, amplitude histograms obtained from recordings in the presence of Ca 2ϩ , Bz-423, and oxidants (left panel) and after the addition of Mg 2ϩ /␥ imino ATP (right panel). Gaussian fittings (green lines) were obtained using the Origin 6.1 Program Set.
Ca 2ϩ homeostasis and may protect mitochondria from Ca 2ϩ overload (10,11), as indicated by studies in isolated cardiomyocytes, CyPD knock-out mice, adult cortical neurons, and spinal chord mitochondria (51)(52)(53)(54). It is tempting to speculate that the different PTP conductances observed in mammals, Drosophila, and yeast underscore different physiological functions. In particular, due to its smaller size, the Drosophila PTP could be involved in Ca 2ϩ homeostasis rather than in apoptosis (1,11), consistent with the finding that the mitochondrial pathway is not essential in most cases of Drosophila apoptosis (55)(56)(57)(58)(59). The discovery that the Drosophila PTP forms from F-ATPase will allow testing of this hypothesis with the powerful methods of Drosophila genetics.