|Home | About | Journals | Submit | Contact Us | Français|
The mitochondrial F1F0-ATP synthase or ATPase is a key enzyme for aerobic energy production in eukaryotic cells. Mutations in ATPase structural and assembly genes are the primary cause of severe human encephalomyopathies, frequently associated with a pleiotropic decrease in cytochrome c oxidase (COX) activity. We have studied the structural and functional constraints underlying the COX defect using Saccharomyces cerevisiae genetic and pharmacological models of ATPase deficiency. In both yeast Δatp10 and oligomycin-treated wild type cells, COX assembly is selectively impaired in the absence of functional ATPase. The COX biogenesis defect does not involve a primary alteration in the expression of the COX subunits as previously suggested but in their maturation and/or assembly. Expression of COX subunit 1, however, is translationally regulated as in most bona fide COX assembly mutants. Additionally, the COX defect in oligomycin-inhibited ATPase deficient yeast cells, but not in atp10 cells could be partially prevented by partially dissipating the mitochondrial membrane potential using the uncoupler CCCP. Similar results were obtained with oligomycin-treated and ATP12 deficient human fibroblasts respectively. Our findings imply that fully assembled ATPase and its proton pumping function are both required for COX biogenesis in yeast and mammalian cells through a mechanism independent of Cox1p synthesis.
Mutations in nuclear and mitochondrial genes affecting the F1F0-ATP synthase or complex V function result in rare devastating encephalomyopathies . Mitochondrial F1F0-ATPase is a multimeric complex formed by two functionally and physically coupled portions of dual genetic origin. The inner membrane-embedded F0 sector contains mitochondrial DNA (mtDNA) encoded subunits and the hydrophilic F1 sector is attached to it from the matrix side . The F0 sector forms a ring-like structure that rotates in the membrane bilayer and drives proton translocation across the inner membrane which is coupled to the synthesis of ATP from ADP and inorganic phosphate. This catalytic capacity is located in the F1 sector. Protons flow through the F0 in favor of the proton gradient generated across the mitochondrial inner membrane by the mitochondrial respiratory chain (MRC) enzymes. The MRC is formed by four enzymatic complexes embedded in the inner membrane and two mobile electron carriers, coenzyme Q and cytochrome c (cyt c). Electrons donated from reducing equivalents to the MRC complexes flow down an electrochemical gradient to molecular oxygen, the final electron acceptor. The transfer of electrons to oxygen is catalyzed by cytochrome c oxidase (COX) or complex IV, another multimeric enzyme of dual genetic origin, containing copper and heme A as metal prosthetic groups. Complexes I (in higher eukaryotes), III and IV couple electron transfer with proton extrusion from the matrix to the mitochondrial inner membrane space, thus generating the driving force for ATP synthesis. It is now generally accepted that complexes I, III and IV interact physically to form supercomplexes with different stoichiometries, some of which contain coenzyme Q and cyt c and constitute real respirasomes . The assembly and stability of some of the individual enzymes depend on their interaction with the others. For example, complex I assembly requires the presence of complexes III and IV [4, 5] and complex IV assembly requires the structural presence of cyt c in yeast and mammals [6, 7]. The ATPase complex is known to form oligomers that play a role in the maintenance of the mitochondrial inner membrane cristae structure but it has not been shown to physically interact with MRC enzymes.
As a consequence of the structural supra-organization of the MRC and the single complex biogenesis dependence on other MRC components, alterations in one protein-coding gene can result in combined enzyme complex defects. For example, some mutations in the mtDNA encoded cytochrome b, a subunit of MRC complex III, produces a combined complex I+III deficiency in human patients suffering from encephalomyopathies . Based on the absence of reported ATPase physical interactions with MRC enzymes, combined deficiencies should not be expected in cases of patients with mutations affecting ATPase structure and function. However, in several reported cases, the ATPase deficiency was associated with a pleiotropic decrease in COX activity specifically or as the major part of a more general MRC alteration. Maternally transmitted ATPase disorders, NARP (neuropathy, ataxia and retinitis pigmentosa) and MILS (maternally inherited Leigh’s syndrome), are caused by mutations in the mitochondrial ATPase structural genes ATP6  and more rarely ATP8 . Biochemical analysis in samples from MILS patients carrying mutations in ATP6, which affect the function but at least not extensively the assembly of the ATPase, showed a significant COX activity decline in skeletal muscle . Analysis of an ATP6 NARP mutation in cybrid cell lines also showed a MRC deficiency with a predominant COX defect and increased ROS production . ATPase disorders of nuclear genetic origin manifest a very severe outcome as well . A case has been associated with a homozygous mutation in ATP12, the human homologue of a nuclear encoded assembly gene previously described in yeast . Similar to yeast atp12 mutant cells which are COX deficient , the patient presented a pleiotropic COX defect in skeletal muscle .
Mutations in several ATPase nuclear structural genes in the yeast S. cerevisiae, including ATP1, ATP2, ATP4, ATP7, ATP8 and ATP9 have been reported to affect COX biogenesis [16–20]. However, some ATPase mutations reduce mtDNA stability and produce a large amount of ρ−/ρ0 cells (cells with partial or complete absence of mtDNA) making it difficult to assess to which degree and how COX assembly is specifically affected in these mutants. With the exception of some ATPase mutants that are petite obligate , the problem has been solved by introducing a nutritional selection marker in the mitochondrial DNA . Analysis of a null atp6 mutant in this context showed a selective decrease in COX accumulation to residual amounts barely detectable . Furthermore, yeast models of NARP and MILS syndrome, carrying the ATP6 yeast equivalent mutation T8993G and T8993C respectively, show a selective and dramatic decrease in COX activity and content [24, 25]. At present, the molecular mechanism involved in this pleiotropic effect remains unknown but its relevance is underscored by the fact that it is conserved from yeast to human.
Several hypotheses have been proposed to explain the ATPase function on COX assembly invoking functional and structural constraints. ATPase lesions blocking proton pumping across the inner membrane affect both ATP production and the mitochondrial membrane potential (ΔΨm). The maintenance of a high ΔΨm could be deleterious for the cells by limiting the glycolytic ATP import necessary for the survival of respiratory deficient mutants. It has been proposed that the poor COX assembly allows ATPase mutant cells to import glycolytic ATP into mitochondria by decreasing ΔΨm [20, 24] or to avoid the accumulation of reduced intermediary components of the respiratory chain . This hypothesis is supported by the fact that COX biogenesis is not altered in uncoupled ATPase mutants where ΔΨm maintenance is severely compromised by proton leakage through F0 [26, 27] while it is altered in oligomycin-treated wild-type cells . However, COX was also found decreased in yeast atp6 point mutants in which the F0 proton pumping capacity was not affected . In these atp6 mutants, the stoichiometry of the supercomplexes formed by COX and the bc1 complex (complex III) was found additionally altered . A similar observation was made in mutants of the dimer-specific ATPase subunits e and g, which destabilize dimeric and oligomeric ATPase, suggesting an involvement of the ATPase in the maintenance of the correct COX- bc1 super-complexes organization , although there are conflicting reports about whether ΔΨm, COX assembly and function are affected in these mutants [28, 29].
COX biogenesis is a complex process requiring more than 30 ancillary factors to facilitate the expression, maturation by addition of prosthetic groups and assembly of three mtDNA encoded subunits (1 to 3) forming the COX catalytic core and 8 (yeast) or 10 (humans) nuclear encoded subunits forming a protective scaffold around the core . Which are the COX assembly step/s specifically affected in ATPase deficient cells? In yeast, the COX biogenesis defect could result from the decreased rate of Cox1p synthesis observed in several ATPase mutants [18, 19, 24]. However, in other cases, Cox1p synthesis was found normal and the COX defect was suggested to be caused by a block in heme A biosynthesis or insertion into Cox1p .
To gain insight into the function of the F1F0-ATPase on COX biogenesis, we have used genetic and pharmacological yeast and human cells models of ATPase deficiencies. Our results show that in S. cerevisiae, ATPase lesions do not affect Cox1p translation directly although it is downregulated as a consequence of the COX assembly defect. Additionally, our results demonstrate that COX biogenesis in yeast and mammalian cells requires both a functional ATPase proton pump and a structurally integral and fully assembled F1-F0 ATPase.
The S. cerevisiae strains used were all in the W303 background and included the wild type W303-1A (MATa ade2-1 his3-1,15 leu2-3,112 trp1-1 ura3-1) and the previously reported mutants Δatp10 , Δcox14 and Δcox14Δcox15  and Δcox15 . Mutant Δatp10 cells were transformed with an integrative plasmid containing the mss51 suppressor mss51T167R as reported . The Δatp10Δcox14 double mutant was obtained from crosses of the respective single mutants. The compositions of the growth media have been described elsewhere . The following media were used routinely to grow yeast: YPD (2% glucose, 1% yeast extract, 2% peptone), YPGal (2% galactose, 1% yeast extract, 2% peptone), YPEG (2% ethanol, 3% glycerol, 1% yeast extract, 2% peptone), WO-EG (2% ethanol, 3% glycerol, 0.67% yeast nitrogen base), WO-Gal (2% galactose, 0.67% yeast nitrogen base).
Cells were grown to early stationary phase in galactose containing media either complete (for mitochondrial preparation assays) or synthetic plus prototrophic requirements. Subsequently, cells were inoculated to OD600 equal to 1 in fresh media supplemented or not with 2 µM oligomycin, 5 µM CCCP or both drugs simultaneously. Cells were incubated for up to 16 h at 30°C with constant shaking.
Mitochondria were prepared from strains grown in media containing 2% galactose, according to the method of Faye et al.  except that zymolyase 20T (ICN Biochemicals Inc., Aurora, OH) instead of Glusulase was used for the conversion of cells to spheroplasts. Mitochondria prepared from the different strains were used for spectrophotometric assays carried out at 24°C. KCN-sensitive Cytochrome c oxidase (COX) activity was assayed with 50 µg of mitochondria which were permeabilized with potassium deoxycholate as described , by following the oxidation of 50 µM reduced cytochrome c at 550 nm in a medium containing 20 mM KH2PO4 (pH 7.4). The addition of 0.3 mM KCN inhibited the reaction. Antimycin A-sensitive NADH cytochrome c reductase (NCCR) activity was assayed in 25 µg of mitochondria permeabilized with potassium deoxycholate as described  by measuring, at 550 nm, the reduction of oxidized 50 µM cytochrome c using 0.4 mM NADH as the electron donor in a medium containing 20 mM KH2PO4 (pH 7.4) and 2 mM EDTA. The addition of 0.4 µM antimycin A inhibited the reaction. ATPase activity was assayed by measuring release of inorganic phosphate  from ATP at 37°C in the presence and absence of oligomycin.
Mitochondrial gene products were labeled with 35S-methionine (7 mCi/mmol, Amersham, Piscataway, NJ) in whole cells at 30°C in the presence of cycloheximide . Equivalent amounts of total cellular or mitochondrial proteins were separated by SDS-PAGE on a 17.5% polyacrylamide gel, transferred to a nitrocellulose membrane, and exposed to Kodak X-OMAT X-ray film.
Ψm was estimated using rhodamine 123 (Rh123), a cell-permeant, cationic, fluorescent dye that is readily sequestered by viable mitochondria. A 30 min incubation of 106 yeast cells per ml in 50 µM Rh123 at 24°C allowed sufficient uptake of Rh123 into the matrix to self-quench as reported . The cells were then washed 2X in water. Flow cytometry analysis was performed on a Becton Dickinson (BD) FACSCalibur flow cytometric analyzer. Excitation was at 488 nm; emission was detected using a 40 nm bandpass filter centred at 535 nm (Chroma, Rockingham, VT, USA). Depolarization of Ψm caused Rh123 to leak out of mitochondria into the cytosol where Rh123 became unquenched, producing an increase in fluorescence (reviewed by ). This interpretation is supported by the data of Fig. 4A, which show an increase in Rh123 fluorescence following application of the ionophore carbonyl cyanide m-chloro phenyl hydrazone (CCCP), and a decrease following application of oligomycin, expected to hyperpolarize mitochondria.
A previously reported primary fibroblast line from a mitochondrial ATPase-deficient disease patient was obtained from Dr. Sara Seneca (University Hospital VUB, Brussels, Belgium) . The patient carried a homozygous mutation in ATP12, an ATPase assembly gene. Human neonatal fibroblasts (CCD-10645k) were obtained from ATCC (Manassas, VA). The fibroblast cell lines were immortalized by transducing human papilloma virus (HPV) 16 E6/E7 oncogenes. Cells were cultured in high glucose Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 100 µg/ml sodium pyruvate and 0.25 mg/ml uridine.
mtDNA content was estimated by Southern blot analysis as described  . Briefly, 6 mg of total DNA was digested with PvuII, which linearizes human mtDNA by cleaving at nt 2653 and produces an 12 kb fragment of the human rDNA repeat unit containing the 18S rRNA gene. Digested samples were resolved on 0.8% agarose gels and transferred onto nitrocellulose membranes. mtDNA was detected using [α-32P]-dCTP-labeled, random primed probes generated against a 16.1 Kb human mtDNA PCR product. Signals were normalized against the 18S rDNA signal detected using similar probes generated against a 0.6 Kb PCR fragment of the human rDNA repeat unit encompassing the 18S rDNA gene.
To measure COX activity in cells homogenates, cells were collected by trypsinization and submitted to three freezing-thawing cycles. COX and NCCR activities were measured in the presence of lauryl maltoside as described .
Cell homogenates were also used to estimate COX content by western blot analyses using an antibody against human COX1 (Molecular Probes). An antibody against β-tubulin (Molecular Probes) was use as loading control.
Standard procedures were used for transformation and recovery of plasmid DNA from E. coli . Yeast were transformed by the method of Schiestl and Gietz . The one-step gene insertion method  was used to integrate linear plasmids at the URA3 or LEU2 locus of yeast nuclear DNA. Protein concentration was measured with the Folin phenol reagent . Proteins were separated by SDS-PAGE in the buffer system of Laemmli , and western blots were treated with antibodies against the appropriate proteins followed by a second reaction with anti-mouse or anti-rabbit IgG conjugated to horseradish peroxidase (Sigma, St. Louis, MO). The SuperSignal chemiluminescent substrate kit (Pierce, Rockford, IL) was used for the final detection.
All experiments (enzymatic assays, western blot quantifications) were done at least in triplicate. For the enzymatic assays, data are presented as means ± SD of absolute values. The values obtained for wild type and mutant strains were compared by t-Student test. P < 0.05 was considered significant. For quantification of western blot signals, the images were digitalized and densitometry performed using the histogram function of the Adobe Photoshop program. The values measured in three independent assays did not differ by more than 5%.
To investigate the molecular requirement of an intact and functional ATPase for COX biogenesis we have used a yeast strain in which Atp6p assembly into the ATPase complex is prevented by a nuclear mutation in its specific chaperone Atp10p . This choice is justified based on the high mtDNA stability in this mutant (Fig. 1A). For all the experiments reported, we used freshly purified mtDNA-containing (ρ+) colonies and the results were normalized by the percentage of ρ+ cells in the culture used for each experiment (mitochondrial preparation, protein synthesis analyses, etc), which was always between 85 and 95%. As previously reported, the oligomycin sensitive ATPase activity in Δatp10 cells was less than 5% of wild type ( and Fig. 1B). Similar to other mutants in structural and assembly ATPase genes [14, 16–19, 47], we have observed that the lack of Atp10p results in a selective decrease in COX activity to a residual amount of approximately 10% of wild-type (Fig. 1C) with a very mild effect on the upstream portion of the MRC measured by the NADH cytochrome c reductase (NCCR) assay (Fig. 1D). The pleiotropic effect on COX was not reciprocal since a mutation in cox15, a gene essential for heme A biosynthesis and hence COX biogenesis , does not affect ATPase function (Fig. 1B). In agreement with the COX activity decline and similar to atp6 mutants , atp10 mutant cells present a dramatic reduction of the steady state level of Cox1p and to a significantly lesser extent Cox2p and Cox3p (Fig. 1E). The significant stability of unassembled Cox2p and Cox3p in atp10 mutant cells is intriguing. This result differs from most bona fide COX mutants, including Δcox15, in which most unassembled highly hydrophobic core subunits are proteolytically degraded and do not accumulate at the steady-state level (Fig‥ 1E).
To test whether or not the COX assembly defect in ATPase mutants exclusively results from structural ATPase defects, we have explored an ATPase-deficiency yeast pharmacological model consisting of treatment of a wild type strain with oligomycin. Oligomycin acts by stalling proton translocation through the ATPase F0 channel thus blocking ATP synthesis and producing inner membrane hyperpolarization. Oligomycin, as well as ATPase lesions of genetic origin, can destabilize the mtDNA. In our system, treatment of wild-type yeast cells (W303-1A strain) with 2 µM oligomycin during 8 hours does not affect mtDNA stability (Fig. 2A). In this time frame, ATPase activity, measured as the ATP hydrolysis rate, and COX activity decreased both to approximately 35% of wild type values (Fig. 2B and C). The MRC lesion caused by the inhibition of the ATPase proton channel is specific to COX, since the NCCR activity is not affected in oligomycin treated cells (Fig. 2D). To further characterize our pharmacological yeast model of ATPase deficiencies, we investigated by western blot analysis the steady state levels of COX mitochondrial and nuclear encoded subunits. After 8 hours of treatment with 2 µM oligomycin, the steady-state levels of the three COX core subunits and particularly of Cox1p were significantly decreased compared to the untreated cells (Fig. 2E).
Our results demonstrate that COX biogenesis requires a functional proton pumping ATPase. It was previously shown that COX assembly and stability also requires the physical presence in the mitochondrial compartments of the electron carrier cytochrome c (cyt c), even at sub-stoichiometrical concentrations [6, 30]. Cyt c plays a structural role in COX biogenesis because the sole presence of an inactive form was sufficient to stabilize COX . To test whether the effect of ATPase deficiency on COX assembly could be mediated through a severe reduction in cyt c amount, we analyzed the steady state level of cyt c in oligomycin treated wild type cells (Fig. 2E) and in the atp10 mutant strain (Fig. 1E). In both cases, cyt c levels were unaffected indicating that a reduction in cyt c amount is not part of the mechanism of the pleiotropic COX biogenesis alteration caused by ATPase deficiencies.
The prominent decrease in Cox1p steady-state levels raised the previously proposed possibility that Cox1p synthesis could be impaired in ATPase defective cells. Cox1p is a key COX subunit which acts as the seed around which the enzyme is assembled [48, 49]. COX assembly is characterized by a concerted accumulation of its constitutive subunits. Recently, our group and others have reported an interesting contribution to the stoichiometric accumulation of subunits during COX biogenesis in S. cerevisiae. It consists of a regulatory mechanism by which the Cox1p synthesis is regulated by the availability of its assembly partners [32, 50, 51]. The products of COX14 and MSS51 play a basic role in this Cox1p translational down-regulation in the absence of COX assembly [32, 52]. Mss51p is a COX1 mRNA specific translational factor that interacts with the 5’-untranslated region of the mRNA to promote synthesis [51, 52]. It also interacts with the newly synthesized Cox1p protein possibly to promote elongation and maturation [32, 51] forming a complex that is stabilized by Cox14p  until Mss51p is disengaged from the ternary complex to participate in new rounds of translation and Cox1p proceeds in the assembly process . In the absence of COX assembly, Mss51p is envisioned to remain trapped in the ternary complex and is not available for Cox1p synthesis. Down-regulation of Cox1p synthesis in COX assembly mutants can be bypassed at least by two mechanisms. First, expression of mutant alleles of mss51 or additional copies of the wild type MSS51 gene facilitate COX1 mRNA translation in the absence of COX assembly . Second, the introduction of a cox14 mutation restores Cox1p synthesis presumably by preventing the trap in the ternary complex of Mss51p, which become available for its function in translation. As in bona fide COX assembly mutants, in the atp10 mutant, the rate of Cox1p synthesis is reduced to less than 10% of wild type (Fig. 3A). Similar to COX mutants, the mitochondrial synthesis profile in atp10 mutant cells also displays the appearance of mp15, an aberrant COX1 mRNA translation product synthesized in the absence or limited amounts of translationally competent Mss51p . The Cox1p synthesis defect was restored by either deleting COX14 (Fig. 3A) or by integrating a second copy of MSS51 into the nuclear genome (the mss51T167R allele previously shown to restore Cox1p synthesis in most COX mutants ) (Fig. 3B), two of the hallmarks of the Cox1p translational regulatory loop previously described . In the double mutant Δatp10Δcox14 strain, COX activity was null as expected due to the absence of Cox14p. In Δatp10 cells expressing the mss51T167R allele, the increase in Cox1p synthesis did not enhance COX assembly since residual COX activity remained as low as in the Δatp10 cells (Fig. 3C). Although COX1 mRNA translation was enhanced in these cells, newly synthesized Cox1p failed to accumulate in the steady-state, probably as a consequence of fast degradation in the absence of COX assembly (not shown).
Similarly, in vivo mitochondrial protein synthesis analysis in oligomycin treated wild type cells showed a reduced rate of Cox1p synthesis (Fig. 3D). Oligomycin treatment did not affect Cox1p synthesis in Δcox14 cells (Fig. 3D) in which COX1 mRNA translation is known to escape translational regulation. Interestingly, and although not developed here, the rate of synthesis of ATPase subunits 6, 8 and 9 in atp10 cells and particularly in oligomycin treated cells, was found significantly increased (Fig. 3D and Fig 4D), probably as a compensatory mechanism in the absence of a functional ATPase.
We concluded that the Cox1p synthesis down-regulation is a consequence rather than a cause of COX assembly defects in atp10 mutants and oligomycin-treated wild type cells.
The results obtained with the oligomycin inhibited cells suggested that the proton-pumping ATPase function rather than a structural ATPase absence and/or alteration would be responsible for the COX deficient phenotype observed in these cells. To test whether the decrease in COX assembly observed in some ATPase mutants is the result of an increase in mitochondrial ΔΨm , we have explored the effect of the uncoupler ionophore carbonyl cyanide m-chloro phenyl hydrazone (CCCP) on COX biogenesis in the absence of a functional ATPase. We used concentrations ranging from 1 to 5 µM CCCP which are known to cause a partial uncoupling of mitochondria and to only slightly affect the import of precursor proteins in isolated mitochondria [53, 54].
As expected, the alterations in the proton pumping ATPase function induced by both oligomycin and atp10 mutation generated a mitochondrial membrane hyperpolarization (Fig. 4A) as measured by flow cytometry using the cationic dye rhodamine 123 (Rh123) as reported . In the assay conditions, explained in the Experimental procedures, Rh123 quenching is proportional to mitochondrial ΔΨm . The ionophore CCCP partially dissipated the ΔΨm in wild type cells, under these experimental conditions. Additionally, CCCP also compensated for the oligomycin- and atp10 mutation-induced hyperpolarization, restoring wild type ΔΨm in these strains (Fig. 4A).
The observation that CCCP restores ΔΨm in Δatp10 and oligomycin treated cells suggested that if this parameter were relevant for the pleiotropic COX assembly defect, CCCP treatment could also restore COX biogenesis.
Treatment of wild-type yeast cells with 5 µM CCCP alone for 8 hours did not significantly affect COX and NCCR activities (Fig. 4B and C). Interestingly, wild-type cells supplemented with oligomycin together with CCCP show COX activity levels comparable to the untreated control value (Figure 4B). Furthermore, while oligomycin and CCCP do not produce any major direct effect in the rate of in vivo mitochondrial protein synthesis (preincubation time 0 in Fig. 4D), incubation for 8 hours of a wild-type strain with 2 µM oligomycin together with 5µM CCCP also restored Cox1p synthesis compared to oligomycin treated cells (Fig. 4D). These results further support the conclusion that in ATPase deficient strains Cox1p synthesis is downregulated as a consequence of the COX assembly defect.
Contrary to the results obtained in oligomycin treated cells, incubation of Δatp10 cells with several CCCP concentrations failed to significantly restore COX activity (Fig. 4E). These results suggest that a fully assembled, structurally integral ATPase is required for COX assembly.
The dependence of COX biogenesis on a functional ATPase is conserved from yeast to higher eukaryotes. To validate the results obtained in yeast models of ATPase deficiencies, we developed a pharmacological human cell culture model based on ATPase inhibition with oligomycin in human neonatal fibroblasts (CCD-10645 k) obtained from ATCC (Manassas, VA). In these cells, the mtDNA content remains significantly stable after 16 hours of exposition to 5 µM oligomycin (Fig. 5A). In the same experimental conditions, COX activity was specifically reduced to approximately 45% of control values (Figure 5B) as well as the steady state levels of COX subunit 1 (Fig. 5D). The addition of 5 µM CCCP to the oligomycin treated cells significantly restored these parameters (Fig. 5C and D). To validate our findings in cells from human patients, we have obtained from Dr. S. Seneca (University Hospital VUB, Brussels, Belgium) cultured primary fibroblasts from the single patient described carrying a mutation in ATP12 . Studies in S. cerevisiae have previously shown that Atp12p binds to unassembled alpha subunits of the F1 sector thus preventing its homo-oligomerization in non-productive complexes during assembly of the F1 moiety of the mitochondrial ATP synthase [55, 56]. We have immortalized the fibroblast cell line by transducing human papilloma virus (HPV) 16 E6/E7 oncogenes. An initial analysis using cell homogenates has shown that the ATP12 mutation in these cells induces a pleiotropic 40% decrease in COX activity in comparison with human neonatal fibroblasts immortalized in the same way (Fig. 5C). However, similar to yeast Δatp10 cells, COX activity in these cells was not restored by treatment with CCCP.
Taken together our results obtained in yeast cells and human fibroblasts indicate that the dependence of COX biogenesis on a functional ATPase and the mechanisms involved are conserved along evolution.
Biogenesis of a functional COX complex is a highly regulated process. The multiple levels of regulation described to date involve the transcriptional and translational control of subunits and assembly factors accumulation, availability of heme A and copper co-factors, protein import into mitochondria and membrane insertion, and coordination of sequential or simultaneous steps of the process (reviewed in ). Additionally, the arrangement and integrity of the mitochondrial respiratory chain and oxidative phosphorylation system in the inner mitochondrial membrane plays a crucial role on COX stability and assembly . Here we report that the F1F0-ATPase functions in COX biogenesis through a mechanism conserved from yeast to human that does not involve the synthesis of mitochondrial DNA encoded COX subunits.
The pleiotropic COX biogenesis defect observed in most S. cerevisiae nuclear mutants that severely affect ATPase assembly and/or function [16–20] has been suggested to stem from a decrease in synthesis of Cox1p [18, 19, 24] or its maturation by heme A insertion . The role of ATPase on COX assembly was recently better explored in a S. cerevisiae strain carrying an atp6 deletion and engineered to stably maintain its mtDNA under selection . The lack of Atp6p resulted in a selective decrease in COX accumulation to residual amounts barely detectable and significantly decreased Cox1p synthesis. Using our S. cerevisiae oligomycin-treated and Δatp10 mutant cell models, in which mtDNA is highly stable, we have examined whether Cox1p translation could be impaired by the ATPase lesion. We have previously reported that in S. cerevisiae wild type cells, COX1 mRNA translation is contingent with the availability of its assembly partners, thereby acting as a negative feedback loop that paces Cox1p translation to its utilization during assembly of the complex [32, 36, 51]. Our results clearly demonstrate that the apparent reduction in the rate of Cox1p synthesis results from a down-regulation of Cox1p translation in the absence of COX assembly through a mechanism involving Mss51p and Cox14p, as previously described for bona fide COX mutants as well as for cyt c mutants . Thus, down-regulation of Cox1p synthesis in ATPase deficient cells is a consequence rather than a cause of the pleiotropic COX biogenesis alteration.
To understand the mechanism by which the F1F0-ATPase affects COX biogenesis and whether it involves functional or structural constraints, we have used both pharmacological and genetic models of ATPase deficiencies in yeast and cultured human cells. In yeast cells, oligomycin-inhibition of the ATPase proton pump as well as atp10 deletion induced mitochondrial membrane hyperpolarization. This effect was reverted by treatment with the ionophore CCCP by limiting or eliminating the increase in ΔΨm caused by the absence of a functional ATPase proton pump. However, CCCP treatment only restored COX activity in oligomycin treated cells but not in Δatp10 cells defective in F1F0-ATPase assembly. A similar pattern of restoration was observed in oligomycin-inhibited human cells and in fibroblasts from a patient carrying a mutation in the F1 alpha subunit chaperone ATP12, both of which also expressed a pleiotropic COX deficiency. These results suggest that COX biogenesis is impaired in ATPase deficient strains at least by two independent mechanisms. First, it results from an increase in the mitochondrial ΔΨm in the absence of a functional proton-pumping ATPase. It remains to be elucidated the COX assembly step/s that would be primarily affected by the membrane hyperpolarization caused by ATPase lesions. Second, the absence of a fully assembled ATPase probably introduced additional structural constraints for COX biogenesis, discussed elsewhere [24, 28, 47], that could not be bypassed solely by restoring ΔΨm. Although physical interactions of COX and ATPase have not been reported in eukaryotic cells, it was recently demonstrated a physical interaction between the cytochrome caa3 and F1F0-ATP synthase from Bacillus pseudofirmus OF4 in a reconstituted system . The authors suggested that such an interaction between these complexes may contribute to sequestered proton transfer during alkaliphile oxidative phosphorylation at high pH . Further studies should clarify whether this interaction occurs in vivo and if similar contacts exist between eukaryotic COX and ATP synthase.
In addition to its well established function in ATP generation, the F1F0-ATP synthase has a role in determining the ultra-structure of mitochondria [59–61], which depends on its ability to form dimeric and higher oligomeric super-complexes [59–62]. Initial studies on mutants of the dimer-specific subunits e and g, which destabilize dimeric and oligomeric F1F0-ATP synthase supercomplexes, showed that they had normal ATPase enzymatic activities, normal mitochondrial morphology, and COX activity was affected to a minor extent (retain ~80% of wild type COX activity) indicating that ATPase dimerization and oligomerization are not essential for COX assembly . In contrast, a more recent study revealed COX activity was decreased to 55% and 75% of wild type levels in strains missing subunits e and g, respectively, probably resulting from an altered organizational state of the super-complex formed by complexes bc1 and COX, which stoichiometry was found altered in the mutant strains . Studies in S. cerevisiae mutants that are unable to form the F1 α3β3 oligomer, either because the α or the β subunit is missing or because the cells are deficient for proteins that mediate F1 assembly (e.g. Atp11p, Atp12p, or Fmc1p), have pleiotropic effects predominantly in COX although also, to a lesser extent, on Complex III or bc1 complex . These mutants were devoid of mitochondrial cristae, supporting the idea that the F1F0-ATPase is important for biogenesis of the mitochondrial inner membrane. Similarly, null mutations in atp6 were reported to significantly alter the mitochondrial ultrastructure and produce a profound COX biogenesis defect . However, it remained unclear to which degree and whether or not the COX deficiency resulted from the altered mitochondrial ultra-structure.
In addition to the biological interest, the pleiotropic effects of ATPase deficiencies have a clear biomedical significance since mutations in nuclear and mitochondrial genes affecting ATPase biogenesis cause devastating encephalomyopathies of childhood . The complex and variable clinical manifestations characteristic of ATPase deficiencies are the result of both the energy deprivation caused by the primary defect in ATP synthesis and the pleiotropic effects associated with the ATPase lesions, e.g. decrease in COX activity and increase in ROS production. Treatments targeting these pleiotropic effects could be beneficial even without restoring full ATP production. For example, analysis of the NARP mutation T8,993G in cybrids cell lines also showed a MRC deficiency, including a prominent COX defect, which interestingly was alleviated by supplementing the growth media with antioxidants .
Using the dye CM-H2-DCFDA (5-(and 6)-chloromethyl-2′7′-dichloro-dihydrofluorescein diacetate acetyl ester) for the detection of unspecific ROS, we have measured by flow cytometry an 1.7 fold increase of ROS in yeast atp10 cells. ROS were efficiently chelated by treatment with the antioxidant N-acetyl-cysteine (10 mM NAC). The NAC treatment produced an increase in COX activity in both wild type (from 100 to 120%) and atp10 cells (from 14 to 28%) (Fig. S1). However, a consistent similar increase was observed in NCCR activity and in citrate synthase activity (Fig S1), the latter a tricarboxylic acid cycle enzyme. These results suggest that the antioxidant treatment could be promoting a general increase in mitochondrial biogenesis, thus contributing to enhance the residual COX activity in yeast atp10 cells. Incidentally, stimulating mitochondrial biogenesis has proven to be an effective mechanism to ameliorate partial COX defects in yeast , mammalian cells  and mouse models of COX deficiency .
The further characterization of the pathways involved in the dependence of COX assembly upon a functional and structurally intact ATPase could allow for the discovery of new targets for therapeutic interventions aiming to ameliorate the clinical manifestations in patients by alleviating this pleiotropic effect that is a significant component of the pathophysiology of ATPase deficiency.
In conclusion, functional and structural constraints are responsible for the pleiotropic COX assembly defect observed in ATPase deficient cells. Future work will focus on the discovery of the exact mechanisms involved and the COX assembly biogenesis steps primarily affected, which are expected to shed light into new regulatory levels of mitochondrial membrane biogenesis and organization.
Our research is supported by National Institutes of Health Research Grant GM071775A (to A.B.) and a Research Grant from the Muscular Dystrophy Association (to A.B.). IS is supported by a pre-doctoral NIH NRSA Fellowship # 1F31-GM081975 and DH by an AHA pre-doctoral Fellowship # 0815083E.
We thank Dr. A. Tzagoloff for critically reading the manuscript. We are in debt with Dr. S. Seneca (University Hospital VUB, Brussels, Belgium) for providing the ATP12 deficient fibroblasts.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.