Knockdown of APOPT1/COA8 Causes Cytochrome c Oxidase Deficiency, Neuromuscular Impairment, and Reduced Resistance to Oxidative Stress in Drosophila melanogaster

Cytochrome c oxidase (COX) deficiency is the biochemical hallmark of several mitochondrial disorders, including subjects affected by mutations in apoptogenic-1 (APOPT1), recently renamed as COA8 (HGNC:20492). Loss-of-function mutations are responsible for a specific infantile or childhood-onset mitochondrial leukoencephalopathy with a chronic clinical course. Patients deficient in COA8 show specific COX deficiency with distinctive neuroimaging features, i.e., cavitating leukodystrophy. In human cells, COA8 is rapidly degraded by the ubiquitin-proteasome system, but oxidative stress stabilizes the protein, which is then involved in COX assembly, possibly by protecting the complex from oxidative damage. However, its precise function remains unknown. The CG14806 gene (dCOA8) is the Drosophila melanogaster ortholog of human COA8 encoding a highly conserved COA8 protein. We report that dCOA8 knockdown (KD) flies show locomotor defects, and other signs of neurological impairment, reduced COX enzymatic activity, and reduced lifespan under oxidative stress conditions. Our data indicate that KD of dCOA8 in Drosophila phenocopies several features of the human disease, thus being a suitable model to characterize the molecular function/s of this protein in vivo and the pathogenic mechanisms associated with its defects.


INTRODUCTION
Mitochondria are subcellular organelles that play a pivotal role in the conversion of energy stored in nutrients into energy spendable in eukaryotic cells (e.g., ATP and heat). The mitochondrial respiratory chain (MRC) plays a central role in this process. MRC includes four enzymatic multiprotein complexes (named complexes I-IV) embedded in the inner mitochondrial membrane (IMM) that catalyze electron transfer reactions that, through a redox cascade, generate a proton gradient acting as a protonmotive force (PMF) across the IMM. The PMF is then exploited by a fifth complex (complex V, mitochondrial ATP synthase) to convert adenosine diphosphate (ADP) to adenosine triphosphate (ATP), the essential energetic molecule used as a substrate for virtually all endergonic processes in living cells (Spinelli and Haigis, 2018).
Mitochondrial disorders (MD) are a group of mitochondriarelated diseases caused by mutations in either nuclear DNA (nDNA) genes, encoding proteins with a role in mitochondrial function, or in mitochondrial DNA (mtDNA) protein or tRNA and rRNA encoding sequences. These diseases are among the most frequently inherited neurometabolic disorders, and are characterized by a wide diversity of clinical features making their diagnosis quite challenging. Thanks to the advances in next generation sequencing (NGS) techniques, the identification of the genetic causes of MD has considerably improved, including extremely rare genetic conditions (Carrol et al., 2014;Legati et al., 2016). For instance, loss-of-function mutations in the COA8 gene (alias APOPT1, APOP-1, or C14ORF153) have been identified in seven subjects from six different families presenting a distinctive form of mitochondrial encephalopathy (Melchionda et al., 2014;Sharma et al., 2018). The COA8 mutations include a nonsense mutation (p.Arg79*), a missense mutation (p.Phe118Ser), frameshift mutations (p.Glu121Valfs*4, p.Glu121Valfs*6), and micro-(p.Glu124del) and macro-deletions (p.Val55_Lys120del) (Melchionda et al., 2014;Sharma et al., 2018).
Murine Apopt1/COA8 was firstly identified as being overexpressed in an in vitro model of atherosclerosis (Yasuda et al., 2006). The authors hypothesized a pro-apoptotic role for the protein, because they observed triggering of the apoptotic cascade in a mPTP (mitochondrial permeability transition pore)dependent manner after its over-expression, causing the release of cytochrome c from mitochondria to the cytosol, and the subsequent activation of caspase-9 and caspase-3 (Yasuda et al., 2006;Sun et al., 2008). However, terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay failed to detect apoptosis in muscle biopsies from patients, and no difference in growth and apoptotic rate was observed in mutant versus control fibroblasts, even after treatment with the apoptosisinducer staurosporine (Melchionda et al., 2014). In addition, over-expression of a GFP-tagged COA8 protein failed to induce cell death in different human cell lines (Signes et al., 2019), whereas the main biochemical hallmark of its absence was isolated mitochondrial cytochrome c oxidase (COX) deficiency (OMIM #220110). Deficiency of COX activity was in fact the only enzymatic abnormality detected in patients' mitochondria, and was associated in both human and mouse cells and tissues with altered COX assembly (Melchionda et al., 2014;Signes et al., 2019). COA8 encodes a 206-amino acid protein targeted and localized within mitochondria in mammals (Yasuda et al., 2006;Sun et al., 2008;Melchionda et al., 2014;Signes et al., 2019).
Recent evidence proved that it is associated with the inner membrane, with the C-terminal region facing the mitochondrial matrix (Signes et al., 2019).
Analysis of the COX structural defect associated with absence of COA8 revealed its involvement in the intermediate steps of the COX assembly pathway (Signes et al., 2019). In addition, COA8 was shown to be oppositely regulated by UPS and ROS (Melchionda et al., 2014;Signes et al., 2019). However, there is still a need to clarify the regulatory mechanisms and the role of COA8 in COX function as well as in general mitochondrial physiology.
The availability of model organisms can provide valuable insights to clarify the relationship between gene mutations and human diseases. Thus, we have investigated the behavioral and biochemical features of a D. melanogaster KD model of COA8associated mitochondrial disease. Our findings indicate that D. melanogaster is a suitable model that can significantly contribute to elucidate the role of this gene in mitochondrial physiology and pathology.

Fly Stocks and Maintenance
Flies were raised on standard cornmeal medium and maintained at 23°C, 70% relative humidity on a 12 h-light and 12 h-dark cycle.

Analysis of Spontaneous Locomotor Activity
The amount of locomotor activity was measured by the DAMSystem3 Data Collection Software (Trikinetics). Ten 2-dayold male flies were placed into a glass tube containing food and water in the form of gel at the bottom. Glass tubes were placed into DPM population monitors (Trikinetics) vertically oriented. One day (24 h) after anesthesia, locomotor activity was recorded for 2 days (48 h) under 12:12 light/dark cycles (LD 12:12) at 23°C. At least three biological replicates per genotype were analyzed.

Climbing Test
Climbing test was performed using a modified version of the countercurrent apparatus originally designed by Seymour Benzer (Benzer, 1967). Twenty 2-day-old flies were placed into the first tube, tapped to the bottom and allowed to climb a 10-cm distance for 10 s. The flies that reached the 10-cm distance were shifted to a second tube, tapped again to the bottom and allowed to climb for further 10 s. The procedure was repeated for a total of five times. At the end, the number of flies in each tube was counted. Climbing indexes were calculated as the weighted average of flies in the different tubes, divided by five times the number of flies in the test. The test was performed 1 h after the dark-light transition and a minimum number of 60 individuals per sex and genotype were analyzed.

Lifespan Assay and Paraquat Treatment
Flies were reared at standard low density, collected after hatching and divided into males and virgin females over a 24-h window. Adults of the same sex were kept at a density of 10 per vial (for a total of 50 individuals) at 23°C. Flies were counted every day and transferred to fresh medium three times per week, with no anesthesia (Broughton et al., 2005). Lifespan was analyzed both in flies fed with standard food and with food containing 20 mM Paraquat. Standard cornmeal was cooled to 35°C before the addition of Paraquat (methyl viologen dichloride hydrate, Sigma) and then poured into plastic vials.

Isolation of Mitochondria
Mitochondria were prepared by differential centrifugation from 100 flies (1:1 males-females) as described previously (Da-Rè et al., 2014a). Briefly, samples were homogenized with a Dounce glass potter and a loose-fitting glass pestle in 10 ml of isotonic isolation buffer (225 mM mannitol, 75 mM sucrose, 5 mM HEPES, 1 mM EGTA, pH 7.4) with 1% BSA. Samples were centrifuged at 1,000 ×g (Eppendorf 5810R) at 4°C for 10 min. The supernatant was filtered through a fine mesh, and centrifuged at 6,000 ×g at 4°C for 10 min. The mitochondrial pellet was washed in 10 ml of isolation buffer and centrifuged at 6,000 ×g for 10 min. The wash was repeated using 10 ml of isolation buffer without BSA and centrifuged at 7,000 ×g for 10 min. The mitochondrial pellet was resuspended in minimal volume of isolation buffer without BSA. Protein concentration was measured by the Bradford assay (Bio-Rad protein assay).

Enzymatic Analysis
Prior to enzymatic MRC complex activity assays, isolated mitochondria were subjected to three freeze-thaw cycles in 10 mM ice-cold Tris hypotonic buffer (pH 7.6) using liquid nitrogen to disrupt the mitochondrial membranes. The activities of mitochondrial respiratory chain complexes and citrate synthase (CS) were measured by spectrophotometry as described previously (Kirby et al., 2007), with minor modifications to the protocols.

Cell Cultures
The Drosophila S2R+ cell line is derived from a primary culture of late stage (20-24 h old) D. melanogaster embryos (Schneider, 1972) and it was obtained from Drosophila Genomics Resource Center (DGRC

Cell Death Analysis
Transiently transfected cells were collected and stained with Annexin V Apoptosis Detection Set PE-Cyanine7 (eBioscience-ThermoFisher Scientific) and propidium iodide (Roche Biochemicals) according to the manufacturer's protocol. Cells were analyzed using Cytomics FC500 (Beckman Coulter) as described previously (Cusumano et al., 2018).

Electron Microscopy
Thoraxes and brains from adult male flies were fixed in 2.5% glutaraldehyde overnight. Samples were rinsed in 0.

In Gel Activity
For the detection of the activity of mitochondrial respiratory chain complexes, gels were stained with the following solutions:  Figure 1A). COA8 has been shown to localize to mitochondria in both human and murine models (Yasuda et al., 2006;Melchionda et al., 2014) being associated with the inner mitochondrial membrane (IMM), with its C-terminal region facing the matrix (Signes et al., 2019). Likewise, dCOA8 has a predicted N-terminal mitochondrial targeting sequence according to the online prediction tools Target P (Emanuelsson et al., 2000) and MitoProt II (Claros and Vincens, 1996), with a probability score of 0.67 and 0.91, respectively. Notably, there is high inter-species variability in the N-terminal sequences ( Figure 1A). Furthermore, the topology prediction tools Phobius (Käll et al., 2004) and TMpred (Hofmann and Stoffel, 1993) indicate the presence of one C-terminal transmembrane domain suggesting that dCOA8 could be associated with a mitochondrial membrane, most likely the IMM due to the presence of an MTS, while the rest of the hydrophilic domains could be localized in the mitochondrial matrix.
We transiently expressed a recombinant form of dCOA8 fused at the C-terminus with the GFP reporter (dCOA8-GFP) in S2R+ cells in order to define the subcellular localization of dCOA8. Colocalization of MitoTracker staining and the GFP signal clearly indicated dCOA8 targeting to mitochondria ( Figure 1B).

dCOA8 Deficiency Phenocopies Some Human Clinical and Biochemical Features in Flies
dCOA8 is mainly expressed in the adult brain and thoracicabdominal ganglion, according to FlyAtlas (Leader et al., 2018). dCOA8 is likely to play an important role in the central nervous system (CNS), in line with the severe neurological manifestations described in the COA8-related disease. To corroborate this, dCOA8 expression was modulated in a spatial and temporal way by exploiting the UAS-Gal4 system. dCOA8 was either ubiquitously downregulated, through the act5C-Gal4 driver, or specifically downregulated in the nervous system using the elav-Gal4 driver. dCOA8 mRNA levels dropped around 60-90% of the control levels in the ubiquitous (whole body) and pan-neuronal (brain) KD flies, respectively (Supplementary Figure S1A). Despite highly efficient silencing, dCOA8 KD individuals reached the adult stage and showed similar lifespan to the controls (Supplementary Figure S1B). However, kinetic analysis of mitochondrial respiratory chain (MRC) complexes revealed a specific and marked reduction in COX activity in ubiquitous KD flies (Figure 1C). To further explore the COX deficiency and, more generally, MRC complexes activity/amount, isolated mitochondria were subjected to one dimension (1D) blue native gel electrophoresis (1D-BNGE) and the natively separated samples were analyzed by in gel activity (IGA) for complex I, complex II, and complex IV. While the activity of complexes I, II, III, and V were unaffected, the activity of complex IV was clearly reduced in mitochondria from KD flies ( Figure 1D). Additionally, the COX-specific IGA reactivity was significantly lower in the dCOA8 KD samples, in the monomeric (IV 1 ), dimeric (IV 2 ), and super-assembled species (III 2 + IV and upper bands). Contrariwise, CI and CII activities were comparable to those of parental control mitochondria ( Figure 1E).
In order to investigate possible phenotypic consequences of dCOA8 deficiency, we tested the amount of spontaneous locomotor activity and performed a climbing assay, since impairment of these tests would suggest neurological impairment possibly due to neurodegeneration. Both ubiquitous and pan-neuronal dCOA8 KD flies showed behavioral alterations, with a significant decrease in both total locomotor activity ( Figure 1F) and climbing ability ( Figure 1G). We also analyzed mitochondrial morphology in thoracic muscles (from ubiquitous KD flies) and brains (from pan-neuronal KD flies) by transmission electron microscopy (TEM), but found no mitochondrial morphological alterations in either tissue (Supplementary Figure S1C).

dCOA8 Does Not Have a Direct Role in the Apoptotic Process
The possible involvement of dCOA8 in apoptosis was tested using a combination of an Annexin V Apoptosis Detection assay (detecting early apoptotic cells) and a propidium iodide staining (to quantify necrotic cells). Labeled cells were analyzed Frontiers in Physiology | www.frontiersin.org by cytofluorimetry in dCOA8-GFP over-expressing S2R+ cells at very high levels (Supplementary Figure S1D). No difference in the amount of apoptotic or necrotic cells was observed between dCOA8-GFP over-expressing and control cells transfected with an empty vector, carrying a cytosolic GFP (Figure 2A). Therefore, also in D. melanogaster, involvement of COA8 in apoptosis failed to be demonstrated.

dCOA8 Protects Flies From Oxidative Stress
Previous work on human cellular models showed that COA8 is rapidly degraded by the ubiquitin-proteasome system (UPS) but it is strongly stabilized after treatment with oxidants (Melchionda et al., 2014;Signes et al., 2019). Moreover, COX was especially sensitive to oxidative stress in the absence of COA8, suggesting a specific predominant role of dCOA8 to protect nascent COX under oxidative stress conditions (Signes et al., 2019). To test the direct involvement of dCOA8 in the reactive oxygen species (ROS) response, dCOA8 KD flies were treated with Paraquat (20 mM), which catalyzes the formation of superoxide (Farrington et al., 1973). Both ubiquitous and pan-neuronal KD flies showed significant reduction of their lifespan after Paraquat treatment, with median survival rates between 36 and 68% of the parental controls for males, and between 41 and 57% for females ( Figure 2B). These data indicate that dCOA8 has a protective role against oxidative stress response in vivo, confirming in a living whole organism A B FIGURE 2 | Functional characterization of dCOA8. (A) Cell death analysis. dCOA8-GFP over-expressing and control S2R+ cells (transfected with an empty vector expressing a cytosolic GFP) were double-stained with Annexin V and propidium iodide (PI). Percentage of early apoptotic (blue bars) and necrotic (red bars) cells was measured 48, 72, and 96 h post-transfection. Data plotted are mean ± S.D. (n = 3 biological replicates, two-way ANOVA with Sidak's multiple comparisons test not significant). (B) Representative lifespan curves (Kaplan-Meier) of ubiquitous (act5C-Gal4 > UAS-CG14806-IR) and pan-neuronal (elav-Gal4 > UAS-CG14806-IR) dCOA8 KD flies with respect to parental controls (UAS-CG14806-IR>+, act5C-Gal4>+, and elav-Gal4>+) under oxidative stress at 23°C. Flies were treated with 20 mM Paraquat in standard food. Statistical analysis was performed with log-rank (Mantel-Cox). All the comparisons between the two KD lines with respect to parental controls are statistically significant (****p ≤ 0.0001).
In conclusion, we generated and characterized D. melanogaster models of the human mitochondrial disorder caused by pathogenic mutations in COA8. This study shows that D. melanogaster is a suitable, user-friendly model to shed light on the molecular and physiological roles of COA8. Further investigation is needed to understand its mechanistic and homeostatic role in COX biogenesis. Ubiquitous and pan-neuronal dCOA8 knockdown flies showed behavioral and biochemical alterations resembling the clinical features of patients and KO mice, thus demonstrating a phylogenetically essential role in MRC function of higher eukaryotes. dCOA8 deficiency induces MRC dysfunction with marked reduction in the activity of COX. The link between the over-expression of dCOA8 and the apoptotic process originally proposed in human and mouse models (Yasuda et al., 2006;Sun et al., 2008) was also not confirmed in flies. Interestingly, reduced ability to cope with oxidative challenges was observed in dCOA8 KD flies, demonstrating for the first time a relevant role of dCOA8 in response to oxidative stress in vivo. Thus, our data support the idea that COA8 plays an essential role in protecting nascent COX from oxidative damage particularly during the metallation of the COX catalytic sites. Future work is warranted to clarify this possible function of COA8 in higher eukaryotes.

DATA AVAILABILITY
The datasets generated for this study can be found in the FlyBase (CG14806).

AUTHOR CONTRIBUTIONS
CP and RC conceived and designed the research. MB and SC performed the experiments. MB, SC, and CT analyzed the data. MB, SC, CT, RC, MZ, EF-V, and CP interpreted the results of experiments. MB and CP wrote the manuscript. EF-V, RC, and MZ revised the manuscript.

FUNDING
MB was supported by a doctoral fellowship from the University of Padova (Italy) and a fellowship from "Aldo Gini" Foundation (Padova, Italy). SC was supported by a post-doctoral fellowship no. BIRD182052 from the University of Padova (Italy). CP was supported by the grant "PRAT 2014-University of Padova, no. CPDA142980. " MZ was supported by Core Grant from the MRC (Grant MC_UU_00015/5), ERC Advanced Grant FP7-322424, and NRJ-Institut de France. RC was supported by the Telethon Project N. GGP11011 ("MitMed: a Multicenter Consortium for the Identification and Characterization of Nuclear Genes Responsible for Human Mitochondrial Disorders").