Diminished OPA1 expression and impaired mitochondrial morphology and homeostasis in Aprataxin-deficient cells

Diminished OPA1 expression and impaired mitochondrial morphology and homeostasis in... Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4086–4110 Nucleic Acids Research, 2019, Vol. 47, No. 8 Published online 14 February 2019 doi: 10.1093/nar/gkz083 Diminished OPA1 expression and impaired mitochondrial morphology and homeostasis in Aprataxin-deficient cells 1 2 1,2,* 1,* Jin Zheng , Deborah L. Croteau , Vilhelm A. Bohr and Mansour Akbari 1 2 Center for Healthy Aging, SUND, University of Copenhagen, 2200 Copenhagen N, Denmark and Laboratory of Molecular Gerontology, National Institute on Aging, 251 Bayview Blvd, Baltimore, MD, 21224, USA Received May 31, 2018; Revised January 25, 2019; Editorial Decision January 29, 2019; Accepted January 31, 2019 ABSTRACT - cells, aprataxin (APTX) is the only protein that removes 5 AMP from DNA (2). Human patients carrying mutation Ataxia with oculomotor apraxia type 1 (AOA1) is in APTX develop the progressive neurodegenerative disease an early onset progressive spinocerebellar ataxia ataxia with oculomotor apraxia type 1 (AOA1) (3,4). caused by mutation in aprataxin (APTX). APTX re- APTX localizes to the nucleus and mitochondria (5). Evi- moves 5 -AMP groups from DNA, a product of dence suggests that AOA1 pathology is related to mitochon- abortive ligation during DNA repair and replication. drial dysfunction. Biochemical and cell biological analysis showed that APTX is more critical in mitochondrial DNA APTX deficiency has been suggested to compromise (mtDNA) repair than in the nuclear DNA repair (5–7). mitochondrial function; however, a detailed charac- AOA1 patients display ataxia, neuropathy, cerebellar atro- terization of mitochondrial homeostasis in APTX- phy and coenzyme Q deficiency, traits seen in mitochondrial deficient cells is not available. Here, we show that diseases (8,9). A mitochondrial disease database has been cells lacking APTX undergo mitochondrial stress and developed as a diagnostic tool to identify mitochondrial display significant changes in the expression of the pathology in human diseases (10), which has been proven mitochondrial inner membrane fusion protein op- useful in previous studies (11). This database predicts that tic atrophy type 1, and components of the oxida- AOA1 is a disease with significant mitochondrial involve- tive phosphorylation complexes. At the cellular level, ment (6,10). APTX deficiency impairs mitochondrial morphology Mitochondria are called the powerhouse of the cells be- and network formation, and autophagic removal of cause of their central role in cellular ATP production. Mito- damaged mitochondria by mitophagy. Thus, our re- chondria also play other important biological roles includ- 2+ ing amino acids and lipid metabolism, Ca signaling, cell- sults show that aberrant mitochondrial function is a cycle regulation and apoptosis (12). Muscle and brain tis- key component of AOA1 pathology. This work cor- sues are particularly vulnerable to mitochondrial abnormal- roborates the emerging evidence that impaired mito- ities, probably because of their high ATP consumption and chondrial function is a characteristic of an increas- reliance on other mitochondrial functions. Accordingly, mi- ing number of genetically diverse neurodegenerative tochondrial dysfunction has been identified in a number disorders. of ataxias and other types of neurodegenerative diseases (11,13–16). Mitochondria are structurally highly dynamic organelles INTRODUCTION and their morphology is determined by the type of their host Ligation of DNA ends is the final step in almost all DNA cell. Mitochondria undergo division (fission) and merge to- repair pathways and is a critical step during DNA replica- gether (fusion). The ratio of fusion and fission determines tion. Human cells have three DNA ligases named DNA lig- the formation of the filamentous tubular network or punc- ase I, III and IV, and all use adenosine triphosphate (ATP) tate mitochondria (17). The processes of fusion and fission as a cofactor (1). During ligation, DNA becomes temporar- involve a group of dynamin-like and GTPase proteins. The ily adenylated at the 5 -end (5 -AMP-DNA) by DNA ligase major players in fusion include the outer mitochondrial (1). Occasionally, DNA ligase dissociates from DNA after membrane proteins mitofusion 1 (MFN1) and mitofusin 2 5 -adenylation of DNA resulting in a 5 -AMP group that (MFN2), and the inner mitochondrial membrane protein must be removed for DNA to be ligated later. In human To whom correspondence should be addressed. Tel: +410 558 8162; Fax: +410 558 8157; Email: vbohr@nih.gov Correspondence may also be addressed to Mansour Akbari. Center for Healthy Aging, Faculty of Health Sciences, University of Copenhagen, Building 7.3, Nørre Alle´ 14, 2200 Copenhagen N, Denmark. Tel: +45 35326762; Email: akbari@sund.ku.dk Published by Oxford University Press on behalf of Nucleic Acids Research 2019. This work is written by (a) US Government employee(s) and is in the public domain in the US. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4087 optic atrophy type 1 (OPA1). The key fission proteins are the For whole cell extract (WCE) preparation, pelleted cytosolic dynamin-related protein 1 (DRP1), and several cells were suspended in lysis buffer (20 mM, 4-(2- mitochondrial outer membrane proteins; mitochondrial fis- hydroxyethyl)-1-piperazineethanesulfonicacid ( HEPES)- sion factor (MFF), mitochondrial fission 1 protein (Fis1) KOH, pH 7.5, 200 mM KCl, 10% glycerol, 1% Triton X- and mitochondrial dynamic proteins MiD49, and MiD51 100, 1% non-ionic detergent, IGEPAL CA-630 (octylphe- (18,19). The function, recruitment and assembly of these noxypolyethoxyethanol), 1 mM ethylenedinitrilo tetraaceti- proteins are largely regulated by post-translational modi- cacid (EDTA), 1 mM Dithiothreitol (DTT), EDTA-free fications ( 20). Complete protease inhibitor cocktail (Sigma) and Phospho- Mitochondrial morphology is integral to mitochondrial STOP (Sigma)), and left on ice for 60 min. Cell debris was quality control through a selective autophagic removal of pelleted at 15 000 g for 15 min, and the supernatant ( WCE) dysfunctional mitochondria known as mitophagy (18). The was collected. processes of fusion, fission and mitophagy are collectively known as mitochondrial dynamics. Increasing evidence has Preparation of mitochondria-enriched extracts identified a close interplay between mitochondrial dynam- ics, mitochondrial bioenergetics, cellular metabolism sta- Cells were collected at 500 g, washed once with phos- tus and energy demand (21,22). Adding to the impor- phate buffered saline (PBS) and suspended in isotonic tance of the mitochondrial homeostasis network, recent re- buffer (20 mM HEPES-KOH pH 7.4, 5 mM KCl, 1 search has identified a novel link between persistent nuclear mM DTT, protease inhibitor cocktail) and left on ice to DNA damage, activation of poly ADP-ribose polymerases swell. The cells were broken in a Dounce tight-fit homog- (PARPs) and nicotinamide adenine dinucleotide (NAD ) enizer in ice and equal volume of 2× mannitol-sucrose- consumption and mitochondrial dysfunction (23). The dis- HEPES (MSH) buffer (420 mM mannitol, 140 mM sucrose, ruption of this axis has been identified as a central cause in 20 mM HEPES-KOH pH 7.4, 4 mM EDTA, 2 mM EGTA, many neurodegenerative diseases (14,24). 5 mM DTT) was added to the homogenate and centrifuged Previous studies suggested that APTX deficiency asso- at 1000 g for 5 min (twice). The supernatant was centrifuged ciates with mitochondrial abnormalities including mito- at 10 000 g for 30 min and the pellet containing mitochon- chondrial morphology and network (5–7). However, a de- dria were suspended in lysis buffer (20 mM HEPES-KOH, tailed investigation of the mitochondrial status in APTX- pH 7.5, 200 mM KCl, 10% glycerol, 1% Triton X-100, 1% deficient cells is not available. The aim of this project is to IGEPAL, 1 mM EDTA, 1 mM DTT, EDTA-free Complete elucidate the molecular mechanisms of mitochondrial dys- protease inhibitor cocktail and PhosphoSTOP) and left on function in APTX deficient cells by analyzing key players in ice for 60 min followed with a mild sonication. The lysates mitochondrial maintenance and function in CRISPR me- were centrifuged at 15 000 g and the supernatants were col- −/− diated APTX U2OS cells and in AOA1 patient-derived lected and used as mitochondrial enriched extracts. cells. We found significant changes in key mitochondrial pa- rameters including disruption of mitochondrial morphol- APTX knockdown ogy, network, decreased mitochondrial membrane poten- tial (MMP), increased mitochondria reactive oxygen species APTX-specific TRC shRNA-pLKO vector (clone ID (ROS) and impaired mitophagy response. Our results sug- TRCN0000083642; Sigma) and a negative control scram- gest that mitochondrial dysfunction is a key feature of ble shRNA-pLKO.1 construct (Addgene) were described AOA1 pathology. previously (5). The plasmid (1 g) was co-transfected with the packaging plasmid (pCMV-dr8.2DVPR, Addgene, 0.7 MATERIALS AND METHODS g) and envelope plasmid (pCMV-VSV-G, Addgene, 0.3 g) into human embryonic kidney 293T cells using PolyJet Synthetic oligonucleotides were from TAG Copenhagen. transfection reagent (SignaGen Laboratories). Lentivirus [- P]ATP was from Perkin Elmer. 5 - DNA adenyla- containing media were collected 48 h later and filtered tion kit was from BioNordika (E2610S). MitoTracker Red through a 0.45 M filter to remove the cell debris and CMXRos (M-7512), Mitosox red (M36008) and tetram- used to infect U2OS cells. Puromycin-resistant U2OS cell ethylrhodamine (TMRM) (T-668) were from Thermo colonies were propagated and tested by western blot analy- Fisher Scientific- Life Technology. Saponin was from Sigma sis for aprataxin (ab31841; abcam). (74036). N-acetyl-L-cysteine (NAC) was from Sigma. Cell lines and preparation of whole cell protein extracts CRISPR mediated APTX knockout (WCE) U2OS cells were seeded on a 6-well dish and trans- U2OS cells were cultured in Dulbecco’s modified Ea- fected with aprataxin double Nickase plasmid (sc-417083- gle’s medium (DMEM)-Glutamax (Gibco). C2ABR and NIC, Santa Cruz Biotechnology) using PolyJet transfection C3ABR (APTX proficient) and L938 (P206L /P206L) reagent (SignaGen Laboratories). Transfected cells were se- and L939 (P206L/V263G) (APTX deficient) patient- lected in medium with 2 g/ml puromycin for one week, derived Epstein-Barr virus-transformed lymphoblast cell then harvested and reseeded on 150 mm dishes at 20 cells lines (25) were grown in RPMI medium 1640- Glutamax per dish. Single cell colonies were propagated. Disruption (Gibco). Both DMEM and Roswell Park Memorial Insti- of APTX was verified by polymerase chainreaction (PCR) tute (RPMI) medium1640 were supplemented with 10% Fe- amplification of the flanking target region and western blot tal Bovine Serum (FBS) and 1% penicillin-streptomycin. analysis. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4088 Nucleic Acids Research, 2019, Vol. 47, No. 8 mtDNA integrity analysis seq2, performed as described in Michael et al.(30), adjusted P-value ≤ 0.05 and log2 fold change ≥ 1or ≤−1. Differen- MtDNA integrity was analyzed using a PCR-based method tially expressed genes (DEGs) were subjected to GO func- (26). The PCR reactions were carried out as described previ- tional enrichment analysis using Enrichr (31). Patient and ously (6). PCR products were separated in agarose gel and control cells were treated similarly, however since there were the intensity of the amplicons was measured using Image only two of each, for each gene, the reads were averaged then J. The resulting values were then converted to relative le- log2 fold change was calculated between patient and control sion frequencies per 10 kb DNA by application of the Pois- samples. Genes with a probability of ≥ 0.75 and log2 fold son distribution (lesions/amplicon =−ln (A /A )), where t 0 change ≥ 1or ≤ -1 were considered a DEG. The RNA-seq −/− A represents the amplification of APTX cells and A t 0 data has been deposited to the GEO database. is the amplification of control cells ( 27). Lesions per 10 kb For real-time quantitative-PCR (Q-PCR) analysis, DNA = (−ln (A /A )) × 10 000 [bp]/size of long amplicon t 0 cDNA was prepared using Maxima Reverse Transcriptase [bp]. and Oligo dT (12–18) (Thermo Fisher Scientific- Life tech). Q-PCR was carried out using a real-time PCR kit (Bio-Rad1725271) following the manufacturer’s protocol. RNA-seq and Q-PCR analysis of gene expression For analysis of alternative splicing of OPA1 mRNA, Total RNA was purified from APTX-KO U2OS cells, cDNA was prepared from the cells and used as template APTX-KO cells stably expressing APTX (APTX-Pos), two to PCR amplify exons three to nine using primers; for- AOA1 patient cell lines (L938 and L939) and the cor- ward, F1- 5 -GGATTGTGCCTGACATTGTG-3, and re- responding control cell lines (C2ABR and C3ABR), us- verse, R1-5 -TCTGATACTTCAACTGAGTGTGC. PCR ing RNeasy Mini Kit (Qiagen). There were four biologi- amplification of exons three to seven was carried out cal replicates for the U2OS, APTX-KO and APTX-Pos cell using primers; forward, F2-5 - GTGTGGGAAATTGA lines, and two replicates for the patient control and AOA1 TGAGTATATCG, and reverse R2- 5 -GCACTCTGAT cell lines. The cells were in culture for 2–3 weeks before CTCCAACCAC. The PCR products were separated in RNA extraction. The samples were sent to the Beijing Ge- 8% non-denaturing polyacrylamide gel (1 × Tris-Borate- nomics Institute Genomics Services for analysis. Briefly, the EDTA (TBE) buffer, 37.5:1 acrylamide: bis). Agilent 2100 Bio analyzer (Agilent RNA 6000 Nano Kit) was used for RNA sample quality control, RNA concentra- tion, RIN value, 28S/18S and fragment length distribution FACS analysis for mitochondrial membrane potential and mi- analysis. For library construction, polyA-containing mes- tochondrial ROS senger RNA (mRNA) molecules were isolated with mag- Fluorescence-activated cell sorting (FACS) analysis netic beads. Following purification, the mRNA was frag- was performed at least three times independently with mented into small pieces. The cleaved RNA fragments were CytoFLEX flow cytometry platform (Beckman Coulter). copied into first strand complementary DNA (cDNA) us- All analyses were done in ice-cold sorting buffer (PBS, ing reverse transcriptase and random primers. This is fol- 0.1% bovine serum albumin (BSA)), and each experiment lowed by second strand cDNA synthesis using DNA Poly- included ∼100 000 events per group. MMP and ROS levels merase I and RNase H. These cDNA fragments then have were evaluated with TMRM and Mitosox Red, respectively. the addition of a single ‘A’ base and subsequent ligation of TMRM (40 nM) or Mitosox Red (5 uM) were added to the the adapter. The products are then purified and enriched cells, and incubated for 15 min in 5% CO2, 37 C. U2OS with PCR amplification. We then quantified the PCR yield cells were trypsinized and collected. After washing, cells by Qubit and pooled samples together to make a single were suspended in ice-cold sorting buffer for sorting. AOA1 stranded DNA circle (ssDNA circle), which gave the final cells were spun down to remove free dyes, suspended in library. DNA nanoballs (DNBs) were generated with the ss- sorting buffer and subjected to FACS. Both TMRM and DNA circle by rolling circle replication to enlarge the fluo- Mitosox Red stained cells were analyzed at 561 nm lasers rescent signals at the sequencing process. The DNBs were with 586 nm emission filters. All FACS data were analyzed loaded into the patterned nanoarrays and pair-end reads using CytExpert Software (Beckman Coulter). of 100 bp were read through on the BGISEQ-500 plat- form for the following data analysis study. For this step, the BGISEQ-500 platform combines the DNA nanoball- Live cell confocal microscopy based nanoarrays and stepwise sequencing using Combina- tional Probe-Anchor Synthesis Sequencing Method. Low- Cells were seeded on 30 mm diameter coverslips (0787, quality reads were filtered out (more than 20% of the bases Thermo Scientific) placed in 6 cm culture dishes with an qualities are lower than 10), reads with adaptors and reads optimal density (200 000 cells per dish). After 24 h, the cells with unknown bases (N bases more than 5%) to get the were stained with different reagents (100 nM Mitotracker clean reads (SOAPnuke v1.5.2). Then the clean reads were red, 40 nM TMRM or Mitosox) for 15 min. After wash- mapped onto reference genome, followed by novel gene pre- ing, coverslips were mounted on chambers for live imaging. diction, SNP & INDEL calling and gene-splicing detec- Microscope images were acquired using LSM780 confocal tion. Clean reads were mapped to reference using Bowtie2 system with a 63× oil immersion objective. Quantification v2.2.5 (28), then gene expression levels were calculated with of mitochondria parameters was carried out in Image J soft- RSEM v1.2.12 (29). Finally, DEGs were detected using DE- ware as described previously (32). Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4089 Mitophagy assay phosphate buffer (pH 7.2). The samples were rinsed three times in 0.15 M phosphate buffer (pH 7.2) and post-fixed Mitophagy in live cells was monitored by Mitophagy de- in 1% w/v OsO4 in 0.12 M sodium phosphate buffer (pH tection kit (Dojindo Molecular Technologies). Urolithin A 7.2) for 2 h. The samples were dehydrated in graded series (UA, 5 M) was used to trigger mitophagy as described of ethanol, transferred to propylene oxide and embedded previously (33). The level of mitophagy was defined by the in Epon according to standard procedures. Following poly- area of Mtphagy dye per cell. At least 50 cells were qualified merization, the Thermanox coverslip was peeled off. Sec- in each group. The level of colocalization of Mtphagy dye tions, ∼60 nm thick, were cut with a Leica UC7 microtome and lysosome dye was also analyzed. Quantification analy- (Leica Microsystems, Wienna, Austria) and collected on sis was carried out using Image J. copper grids with Formvar supporting membranes, stained with uranyl acetate and lead citrate and subsequently ex- Oxygen consumption rate (OCR) amined with a Philips CM 100 Transmission EM (Philips, Eindhoven, The Netherlands), operated at an accelerat- Oxygen consumption was measured using the Seahorse ing voltage of 80 kV and equipped. Digital images were XF24 instrument, according to the manufacturer’s instruc- recorded with an OSIS Veleta digital slow scan 2k × 2k tion (Seahorse Biosciences, North Billerica, MA). Cells CCD camera and the ITEM software package. Quantifica- were seeded into a Seahorse tissue culture plate at a density tion of cristae length was done using Image J. Twenty cells of 50 000 cells per well in DMEM with 1 mM sodium pyru- were quantified in each group. vate, 2 mM glutamine and 10% FBS. After 24 h, the medium was replaced with un-buffered XF assay medium (Seahorse Biosciences), pH 7.4, supplemented with 25 mM glucose, 1 ATP measurement mM sodium pyruvate and 2 mM glutamine. One hour later, Cells were plated in 6-well dishes at 6 × 10 cells per dish. the oxygen consumption rate (OCR) was measured in the Intracellular ATP was measured using a luciferase-based Seahorse XF24 analyzer in four blocks of three 3-min pe- assay (ATPlite Luminescence Assay Kit, PerkinElmer) fol- riods. The first block measured the basal respiration rate. lowing manufacturer’s guidelines. A standard curve was Next, 1 M oligomycin was added to inhibit complex 5 and generated and used to calculate samples ATP concentra- the second block was measured. Then, 0.3 M carbonyl tion. Protein concentration was determined using Bradford cyanide 4-trifluoromethoxy-phenylhydrazone (FCCP) was protein assay reagents (Bio-Rad). The content of ATP was added to uncouple respiration, and the third block was mea- normalized for protein content and presented as percentage sured. Finally, 2 M antimycin A was added to inhibit com- of control. plex 3, and the last measurements were acquired. The cells were counted after the experiment and the results were nor- −/− malized to cell number in each well. Three biological exper- Expression of APTX in the APTX U2OS cells iments were done with technical replicates of 2–5 for each Total RNA was purified from U2OS cells using RNeasy genotype. Kit (Qiagen). cDNA was prepared using SuperScript III RT (Invitrogen) and used as a template to PCR amplify- Immunocytochemistry ing the APTX isoform variant 6 containing a putative mi- tochondrial localization signal (5) (Supplementary Figure Cells were seeded at 20 000 cells per well on 12 mm diameter S1). Primers used in PCR were; forward 5 - AACTAGAT coverslips in 24-well dishes. After 24 h, the cells were fixed in CTATGAGTAACGTGAATTTGTCCGTCTCC, and re- 4% paraformaldehyde in PBS for 10 min at room tempera- verse 5 - AATC GGATCCTCACTGTGTCCAGTGCTT ture and washed three times with PBS. Cells were quenched CCTGAG. The PCR products were digested with BglII in 50 mM NH Cl in PBS for 10 min and permeabilized and BamHI restriction enzymes and cloned into pAcGFP1- and blocked at the same time in blocking buffer (3% bovine −/− Hyg-N1 vector (Clontech). APTX cells were transfected serum albumin (BSA), 0.1% saponin, PBS, pH 7.4) for 60 with the construct using PolyJet transfection reagent (Sig- min. Primary antibodies (diluted in blocking buffer) were naGen Laboratories) and stable APTX-GFP expressing added and incubated at 4 C overnight. The samples were cells were selected in medium containing hygromycine and washed three times in 0.1% saponin in PBS and incubated enriched in FACS. with Alexa Fluor-conjugated secondary antibodies diluted in blocking buffer for 1 h at room temperature protected from light. For nuclei standing, 5 g/ml 4,6-Diamidino- DNA substrates 2-phenylindoledihydrochloride (DAPI) solution (D1306, Duplex 22-mer DNA containing a 5 -AMP was prepared as Thermo Fischer Scientific) was applied to the samples for follows: oligonucleotide 5 -GATCCTCTAGAGTCGACC 5 min at room temperature following incubation with sec- TGCA-3 was first end-labeled at the 5 -end by [- P]ATP ondary antibody. Cells were washed three times with PBS and T4 polynucleotide kinase, and then adenylated to gen- and mounted on SuperFrost glass slides (3302775, Thermo erate an oligonucleotide with 5 -AMP- p-DNA. Or 3 -end Scientific) with mounting medium (S3023, Dako). fluorophore TAMRA labeled oligo was 5 -end phosphory- lated followed with 5 -adenylation of DNA to generate 5 - Transmission electron microscopy (TEM) analysis AMP-DNA––3 -TAMRA oligo. The adenylated oligo was Cells were grown to confluence on Thermanox coverslips annealed to a complementary oligo at 1:1.25 ratio in 20 mM andfixedwith2%v /v glutaraldehyde in 0.05 M sodium HEPES-KOH, pH 7.5 and 100 mM NaCl, heated at 90 C Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4090 Nucleic Acids Research, 2019, Vol. 47, No. 8 for 5 min and slowly cooled down to room temperature to OPA50 bacteria. Synchronous worm cultures were obtained prepare 5 -AMP-containing duplex DNA. by allowing gravid adults lay eggs onto NGM/OP50 plates for 2–4 h. Eggs were then collected and seeded onto NGM agar plates. 5 -AMP-DNA repair assay For mitochondrial network imaging, Day 1 (D1) worms Repair reactions were carried out using 15 gWCEsin40 were incubated for 24 h at 20 C on NGM OP50 plates mM HEPES-KOH, pH 7.8, 1 mM DTT, 0.36 mg/ml BSA, containing 1 M tetramethylrhodamine methyl ester 100 mM KCl, 2 mM EDTA and 5 pmol DNA substrate at (TMRM). Worms were subsequently transferred to fresh 30 C for the indicated times. The reactions were stopped NGM plates containing only OP50 for 1 h to clear their by adding loading buffer (10 mM EDTA, 90% formamide, intestinal tract of residual dye, after which they were trans- xylene cyanol, bromphenol blue) and heating the samples ferred to cover slides and paralyzed. Photographs were im- at 85 C for 5 min. DNA was separated in 20% denaturing mediately taken with LSM780 (Zeiss) confocal microscope. acrylamide gel (19:1 acrylamide:bis, 1 × TBE and 7.5 M For western blot analysis, D1 worms were collected and urea). lysed in RIPA buffer (Life technologies) with complete pro- teases cocktail (Roche). Worms were disrupted by sonica- tion using a Branson sonifier (Thomas Scientific) with the Western blot analysis following settings: 10% amplitude; 10 s on; 30 s off, total on WCEs were separated in Tris-glycine sodium dodecyl time 5 min. Samples were kept on ice during sonication to sulphate (SDS) gels and transferred onto Polyvinyli- avoid excessive heating. Afterward, the debris was cleared denedifluoride (PVDF) membrane. Each experiment by centrifugation at 12 000 × g for 10 min at 4 C. The su- was done three to vfi e times. The images are shown as pernatant was resolved in laemmli sample buffer and used three technical replicates of one biological replicate. The for western blot (WB). primary antibodies used were: APTX (sc-374108), Fis1 ATP was measured with ATPlite Luminescence Assay (sc-376469), proliferating cell nuclear antigen (PCNA) (Perkin Elmer). D1 worms were collected and washed three (sc-56), MFN1 (sc-166644), MFN2 (sc-515647) from Santa times with M9 buffer. Worms were pelted and resuspended Cruz. OPA1 (67589S), superoxide dismutase 2 (SOD2) in cell lysis solution from the ATP assay kit. Assays were (13194S), AMP-activated protein kinase (AMPK) (2535S), performed according to the manufacturer’s protocol. Phospho-AMPK (Thr172), DRP1 (8570S), Phospho- Chemotaxis to isoamyl alcohol was performed at 20 C, DRP1 (Ser616) (3455S), anti-acetyl lysine (9681S) and on 9 cm agar plates as described earlier (34). The chemo- Sirtuin (SIRT)1 (8469) were from CellSignaling. TOMM20 taxis index was calculated by subtracting the number of an- (WH0009804M1), and actin (A5441) were from Sigma- imals found at the trap from the number of animals at the Aldrich. SIRT3 (10099-1-AP), and PARP1 (LS-B3432) source of the chemical, divided by the total number of an- were from Nordic Biosite. VDAC1 (ab14734) and APTX imals entered into the assay (34). The resulting values were (ab31841) were from Abcam, PGC-1 (NBP1-04676) expressed and graphed as percentiles. About 200 adult ani- and light chain 3 (LC3) (NB600-1384) were from Novus mals for each strain were assayed in each experiment. Biologicals, TFAM (H00007019-B01P) was from Abnova. For swimming movement, D1 worms were randomly se- To detect oxidative phosphorylation (OXPHOS) com- lected and transferred to a 6 cm petri dish containing 1 ml plex assembly, we used an assembly-dependent total OX- of M9 buffer. The worms were allowed to acclimate for ∼10 PHOS rodent antibody cocktail (ab110413, Abcam). The s, and then movements were scored for 1 min. Thirty worms were scored in each group. antibodies in the cocktail are against a subunit that is la- bile when its complex is not assembled. The samples were prepared following the manufacturer’s protocol. Statistical analysis Error bars represent SE or range (for experiments with Colony forming assay less than three replicates) as indicated in the figure legends. Data were processed in Excel and statistical analyses were Cells were seeded on 6-well dishes at 200 cells per well. Next performed using GraphPad Prism 7 (GraphPad Software). day, the cells were treated with 0.8 mM methyl methane- Statistical analysis of differences between two groups was sulfonate (MMS) or 0.020 mM menadione in serum-free performed using a two-tailed, unpaired t-test and between medium for 60 min. The drugs were removed, and cells more than two groups using a one-way ANOVA analysis of were gently washed once with PBS and normal medium variance test followed by a Tukey’s post-hoc comparison; was added to each well. Cells were left to recover for 10 two-way ANOVA was used for comparison between con- days. Cells were stained in 0.5% crystal violet, dissolved in trol and AOA1 patient cells; *P < 0.05; **P < 0.01; ***P 20% ethanol for 1 h and washed in water, and colonies were < 0.001; ****P < 0.0001. counted. RESULTS Caenorhabditis elegans strains and methods −/− CRISPR-mediated APTX U2OS cells display hallmarks Caenorhabditis elegans eat-3 (ad426) and N2 control strains of mitochondrial dysfunction were obtained from Caenorhabditis Genetics Center (CGC, University of Minnesota). Strains were maintained at 20 C. We elected U2OS cells for this work for several reasons. Nematode Growth medium (NGM) agar plates seeded with U2OS cells have active and functional mitochondria and Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4091 −/− rely on OXPHOS for ATP production (35). They grow in APTX cells, however, contained more punctate and less monolayer, are amenable to transfection and have an over- connected mitochondria, which mostly clustered in per- −/− −/− all structure that makes them particularly suitable for mito- inuclear regions (Figure 3A, APTX ). APTX cells chondrial morphology and network analysis (35). We used had more mitochondria but they were smaller and shorter a lentiviral delivery system and knocked down (KD) APTX than in control cells (Figure 3A andgraphs). Immunocyto- in U2OS cells. Western blot analysis showed 90–95% re- chemistry analysis of fixed cells using an antibody target- duction in APTX concentration in the APTX-KD cells ing the translocase of the outer mitochondrial membrane, compared with control cells (Figure 1A, KD 20 days), in TOMM20, also showed punctate staining of mitochondria −/− line with a previous report (5). Surprisingly, extracts from in APTX cells (Figure 3B). Furthermore, transmission APTX-KD cells displayed a robust 5 -AMP removal activ- electron microscopy (TEM) confirmed the confocal mi- ity (Figure 1B), indicating that even at a very low concen- croscopy results showing more abundant, but smaller and −/− tration, APTX efficiently repairs 5 -adenylated DNA. This fragmented mitochondria in APTX cells (Figure 3C). demonstrates that in studies where APTX is not completely Notably, altered mitochondrial morphology did not signif- −/− depleted, it may mask biological consequences of APTX de- icantly change mitochondrial content in APTX cells, as ficiency. Thus, we completely depleted APTX in the cells us- judged by western blot analysis of the outer mitochondrial ing CRISPR technology (APTX-KO). Western blot analy- membrane proteins VDAC1 and TOMM20 (Figure 3D). sis of protein extracts from APTX-KO and lymphoblastoid Collectively, these results show that APTX deficiency may cells derived from AOA1 patients completely lacked APTX alter mitochondrial morphology without significant effect (Figure 1C). DNA repair analysis of the extracts showed ab- on the overall mitochondrial content. sence of 5 -AMP-DNA removal activity in APTX-KO and The mitochondrial inner membrane folds into the ma- patient cell extracts indicating that APTX is the sole enzyme trix and forms distinct structures called cristae. Emerging for removal of 5 -adenylated DNA in human cells, and also data suggests that the morphology and shape of cristae confirms the specificity of our DNA substrate (Figure 1D). modulate the organization and function of OXPHOS com- The MMP is generated by OXPHOS through the elec- plexes and thus directly influences cell metabolism ( 39). tron transport chain (ETC). MMP measurement has fre- Electron microscopy (EM) data showed that control cells quently been used to evaluate mitochondrial health and contained long and fine cristae stretched out from the mem- −/− function. Mitochondrial ROS production is highly regu- brane. APTX cells, however, contained shorter and re- lated by MMP and mitochondrial dysfunction is often duced cristae density (Figure 3C and the graph), indicat- linked to increased ROS production (36). APTX-knock out ing a marked alteration in cristae shape and density in −/− −/− cells (APTX ) showed significantly lower MMP (Figure APTX cells. 2A and C confocal microscopy image), and higher mito- chondrial ROS production (Figure 2B and C) compared −/− with the control cells. Superoxide dismutases (SODs) are Diminished OPA1 expression in APTX cells cellular antioxidants that catalyze the conversion of super- Mitochondria are highly dynamic organelles. Size and mor- oxide anions to oxygen and hydrogen peroxide. Mitochon- phology of mitochondria are determined by the rate of mi- drial SOD2 is a primary defense enzyme against mitochon- tochondrial fission and fusion ( 40). Emerging findings are drial superoxide. The SOD2 level was markedly higher in −/− unraveling an intricate connection between mitochondrial APTX cells compared with control cells (Figure 2D), morphology and network organization and a number of key probably in response to increased mitochondrial ROS pro- cellular processes including the clearance of dysfunctional duction. Mitochondrial SIRT3 is a NAD -dependent ly- mitochondria by mitophagy, mtDNA maintenance and in- sine deacetylase that regulates the function of mitochon- −/− tegrity, mitochondrial stress signaling, and cellular energy drial proteins (37), including SOD2 (38). APTX cells demand and metabolism (22,39–45). displayed a significantly higher level of SIRT3 compared Western blot analysis of some commonly accepted key with the control cells (Figure 2D). Thus, impaired MMP regulators of mitochondrial fission and fusion showed that and enhanced ROS production associate with APTX defi- the concentration of OPA1, a central regulator of mitochon- ciency. drial inner membrane fusion and a key protein in mitochon- drial dynamics and cristae structure formation (18,46,47), Impaired mitochondrial network and cristae structure in and also of MFN1, another regulator of mitochondrial fu- −/− −/− APTX cells sion, were considerably lower in APTX cells (Figure 4A and the graphs). Phosphorylation of DRP1 is thought Previously, knockdown of APTX in human neuroblastoma to regulate the translocation of DRP1 from cytosol to the SH-SY5Y cells was suggested to disrupt the mitochondrial outer mitochondrial membrane, a key step in the initiation network (5), which may be related to the changes in the level −/− of fission. The concentration of DRP1 but not the phospho- of mitochondrial fission and fusion proteins in APTX −/− rylated DRP1 (p-DRP1) was lower in APTX cells (Fig- cells observed here. However, the molecular mechanism of ure 4A). These results show that APTX deficiency may af- mitochondrial network impairment and the possibility of fect the expression and the stability of some mitochondrial mitochondrial fission and fusion links have not been investi- fusion and fission proteins in U2OS cells. Q-PCR analy- gated. Thus, we investigated the status of the mitochondrial sis showed significantly lower OPA1 mRNA expression in network and morphology in these cells. Live cell confocal −/− APTX cells (Figure 4B). microscopy analysis showed elongated and highly branched tubular mitochondria in control cells (Figure 3A, Control). Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4092 Nucleic Acids Research, 2019, Vol. 47, No. 8 A B Incubation time 5ʼ-AMP-22-mer APTX 22-mer actin Control APTX-KD C D APTX 5´-AMP-22-mer 22-mer actin Figure 1. Knockdown and knockout of APTX in U2OS cells. (A) Western blot analysis to determine APTX knockdown (KD) efficiency in U2OS cells using a lentiviral-delivered shRNA system. (B) Repair analysis of 5 -AMP DNA in extracts from APTX KD cells and control cells. (C) Western blot analysis showing complete absence of APTX in CRISPR-mediated APTX knockout (KO) U2OS cells (lane 2) and in cells prepared from AOA1 patients. (D) APTX deficient cells are devoid of 5 -AMP removal activity confirming the specificity of our DNA substrate for APTX activity. −/− suggesting a robust mitophagy response in control cells. The Defective mitophagy response in APTX cells −/− intensity of the dye signal was considerably less in APTX Autophagy is a process whereby intracellular components cells than control cells following UA treatment (Figure 5B, are engulfed with membrane bound autophagic vesicles −/− APTX + UA, and the graphs). Collectively, these results (autophagosome), then fuse with lysosomes, and the con- suggest that the ability to induce mitophagy by UA is com- tents of the cargo is degraded (48). Western blot analysis of −/− promised in APTX cells. microtubule-associated protein 1 LC3 is frequently used as a marker to measure autophagic flux. LC3-I is conjugated −/− to phosphatidylethanolamine to form LC3-II, which is lo- APTX cells show elevated PARylation but unchanged calized to autophagosomes (49). There were no detectable NAD content differences in the levels of LC3-I, LC3-II between the con- −/− DNA damage response (DDR) proteins poly (ADP-ribose) trol and APTX cells (Figure 5A, control + vehicle and −/− polymerases 1 and 2 (PARP1/2) signal the location of DNA APTX + vehicle, respectively), suggesting comparable damage on the genome to DNA repair proteins by adding basal autophagic activity in these cells. The selective au- ADP-ribosepolymers (PAR) to themselves and to nearby tophagic elimination of damaged mitochondria is called proteins consuming NAD in the process (50). PARP1 is re- mitophagy (48). Urolithin A (UA) is a metabolite of nat- sponsible for ∼90–95% of the total cellular PARP activity ural compounds known as ellagitannins (33), and it was −/− (51). The overall level of PAR was higher in APTX cells recently shown to induce mitophagy (33). Following UA (Figure 5C and the graph). Surprisingly, the level of PARP1 treatment, the level of LC3-II in control cells, but not in −/− −/− was markedly lower in APTX cells (Figure 5C and the APTX cells increased significantly. This suggests robust graph). The elevated level of PARylation, however, did not autophagosome formation in response to UA treatment in −/− seem to be associated with detectable effect on the cellu- control cells but not in APTX cells. It could also reflect lar NAD content (Supplementary Figure S2), suggesting slower autophagosome-lysosome fusion rate, or impaired that the NAD -dependent reactions were not significantly lysosomal degradation of the cargo in control cells com- −/− −/− affected in APTX cells. pared with APTX cells (Figure 5A). To clarify these al- ternatives and to specifically measure mitophagy, we used a commercially available dye kit. The dye (Mtphagy), ac- −/− Re-introduction of APTX into the APTX U2OS cells re- cumulates in mitochondria following the induction of mi- verses some phenotypes back to the APTX positive control tophagy, the damaged mitochondria then fuse to lysosomes cells resulting in a higher fluorescence signal. Pre-treatment of the cells with UA induced a strong mitophagy signal in con- To further confirm the role of APTX in the observed trol cells that colocalized with the lysosomal signal (Figure changes in mitochondrial parameters, we tested whether re- −/− 5B, control + UA, merged, yellow spots and the graphs), introduction of APTX into APTX cells could rescue Control Control APTX-KO KD (5 d) Control C2ABR C3ABR KD (10 d) Control L938 KD (20 d) L939 No extract No extract Control 2.5 min 5 min APTX-KO C2ABR 10 min C3ABR 15 min L938 No extract 2.5 min L939 5 min 10 min 15 min Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4093 Figure 2. Knockout of APTX causes mitochondrial dysfunction in U2OS cells. (A) FACS analysis and quantifications of MMP using TMRM. Bars display TMRM fluorescence intensity as mean ± S.E.M., N = 3(* P < 0.05, *** P < 0.001). (B) FACS analysis and quantifications of mitochondrial ROS levels using MitoSOX, Bars display MitoSOX fluorescence intensity as mean ± S.E.M., N = 3(*P < 0.05). (C) Confocal images showing TMRM and −/- MitoSOX staining in live APTX and control U2OS cells. The scale bar represents 10 m. (D) A representative western blot image of SOD2 and SIRT3 in the indicated WCEs. Graphs on the right show quantifications of SOD2 and SIRT3 protein levels normalized to actin. Values are mean ± S.E.M., N = 3(**P < 0.01 and ***P < 0.001). key features that were significantly altered by APTX de- uct of alternative transcription or translation of the cloned pletion. APTX cDNA was prepared from U2OS cells and APTX. The level of APTX-GFP was ∼5.5-times higher used as template to construct an APTX-GFP expressing than the level of APTX in control cells. −/− −/− vector. APTX cells were transfected with the construct SOD2 was highly abundant in APTX cells (Figures and stable APTX expressing cells were selected and prop- 2Dand 6B). The expression of APTX restored it to the level agated (hereafter, referred to APTX-positive). By RNAseq of control cells (Figure 6B, APTX-Pos). Mitochondria are analysis, APTX is expressed about eight times more than the major site of ATP production and the level of cellu- in control cells. Approximately 10% of the cells showed a lar ATP may reflect the status of mitochondrial function. −/− mitochondrial APTX-GFP signal as estimated by confocal ATP levels were significantly lower in APTX cells (Fig- microscopy inspection of the cells (Supplementary Figure ure 6C). Expression of APTX restored ATP to the level of −/− S3). Expression of APTX restored 5 -AMP removal activity control cells (Figure 6C). The lower ATP levels in APTX −/− in APTX cells, indicating that the cells were expressing cells may be caused by a higher rate of ATP consumption, a catalytically functional protein (Figure 6A). or low MMP (Figure 2A) that can result in suboptimal mi- Western blot analysis of APTX-positive extracts identi- tochondrial ATP production. fied two major bands and a few weaker bands that migrated The serine/threonine AMPK complex, is a key sensor of faster in the gel (Figure 6B). These may be truncated APTX- cellular energy status by ADP- and AMP:ATP ratio and GFP protein, or the product of alternative transcription or possibly also of glucose availability (52). In general, ac- translation of the APTX-GFP construct. The upper band tivated AMPK phosphorylated at Thr172 (pAMPK) in- corresponds to the long isoform of APTX-GFP. The lower creases catabolic processes while decreasing anabolic re- bands may represent truncated APTX-GFP or the prod- actions through the phosphorylation of key proteins in Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4094 Nucleic Acids Research, 2019, Vol. 47, No. 8 Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4095 −/− Figure 3. Knockout of APTX alters mitochondrial morphology and network. (A) Representative images of live APTX and control cells pre-incubated with MitoTracker Red showing mitochondrial network formation; scale bar represents 10 m. Graphs on the right show quantifications of mitochondrial morphology (number, length, size and shape) from more than 200 cells, values are mean ± S.E.M. (**P < 0.01, ****P < 0.0001). (B) Immunocytochemistry analysis of fixed cells against the outer mitochondrial membrane protein TOMM20. Nuclei are visualized by DAPI staining; scale bar represents 10 m. (C) TEM images showing detailed mitochondrial morphology and cristae structures; scale bar represents 1 m. The graph shows the quantification analysis of cristae length using Image J. Twenty cells were quantified in each group (*** P < 0.001). (D) Western blot analysis of mitochondrial outer membrane −/− proteins VDAC1 and TOMM20 to determine relative mitochondrial abundance in APTX and control cells. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4096 Nucleic Acids Research, 2019, Vol. 47, No. 8 -/- Control APTX OPA1 OPA1 actin MFN1 actin MFN2 actin DRP1 actin 75 pDRP1 17 Fis1 actin B mRNA Figure 4. APTX depletion decreases OPA1 protein and mRNA levels in U2OS cells. (A) A representative western blot image for key mitochondrial morphol- ogy proteins. Graphs on the right show quantifications of different protein levels normalized to actin. Both OPA1 bands were included in the quantifica tion analysis. Values are mean ± S.E.M., N = 3(*P < 0.05, **P < 0.01,****P < 0.0001). (B) Q-PCR analysis of OPA1 gene expression. OPA1 gene expression was normalized to the expression of the house-keeping gene actin. Bars show relative OPA1 mRNA level as mean ± S.E.M., N = 3 (****P < 0.0001). several pathways including mTOR, glycolysis and mito- cient AMPK response to increased AMP:ATP ratio (Sup- chondrial homeostasis. Activated AMPK promotes mi- plementary Figure S4). This may partially explain low basal −/− tophagy through phosphorylation of ULK1 (53). The level level of activated AMPK in APTX cells (Figure 6D) de- of pAMPK but not of total AMPK was considerably lower spite lower cellular ATP content (Figure 6C). −/− in APTX cells. Expression of APTX increased the level The OXPHOS system consists of vfi e multipeptide com- −/− of pAMPK in APTX cells nearly to the level of con- plexes (CI-CV) composed of over 90 different structural trol cells that was statistically significant (Figure 6Dand proteins, which are encoded by nuclear and mtDNA. They the graphs). Low level of activated AMPK may in part ac- require assembly protein factors for proper assembly and −/− count for the delayed mitophagy response in APTX cells function. The stability of the complexes is interdependent (Figure 5B). 5-aminoimidazole-4-carboxamide ribonucleo- (55,56). Diminished MMP and increased mitochondrial side (AICAR) becomes converted in the cells to an AMP ROS production may be caused by OXPHOS system dys- analog and has been used as an AMPK activating com- functions and changes in the protein levels of OXPHOS −/− pound (54). APTX cells showed ∼40% lower level of subunits may reflect altered ETC function. We examined the activated AMPK (p-AMPK) compared to control cells fol- abundance of selected OXPHOS subunits using an assem- lowing AICAR treatment suggesting a somewhat less effi- bly specific antibody cocktail. We found that the mitochon- Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4097 -/- -/- Control + Veh. Control + UA APTX + Veh. APTX + UA LC3 I LC3 II actin -/- -/- Control + Veh. Control + UA APTX + Veh. APTX + UA -/- Control APTX PAR PARP1 actin Figure 5. Mitophagy induction is compromised in APTX deficient cells. ( A) Western blot analysis of autophagy marker LC3 in different cells. The ratio of LC3II (lower band) to actin was calculated to demonstrate autophagic response to urolithin (UA). Data are shown as mean ± S.E.M., N = 3, (***P < 0.001). (B) Assessment of mitophagy in UA treated cells. The merged images show colocalization between the mitophagy dye and the lysosome dye. The upper graph on the right shows the quantification analysis of colocalization of lysosome and mitopahgy dyes. The lower graph shows the quantification o f mitophagy signal. Data is presented as mean ± S.E.M., (****P < 0.0001). scale bar represents 10 m. (C) Western blot analysis of PARylation (PAR) and PARP1 levels in different WCEs as indicated. Quantifications are from three independent experiments. Data are shown as mean ± S.E.M., (**P < 0.01, ****P < 0.0001). Merge Lysosome Mtphagy Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4098 Nucleic Acids Research, 2019, Vol. 47, No. 8 -/- Control APTX APTX-Pos APTX-GFP 5ʼ-AMP-22-mer APTX-Endo 22-mer SOD2 actin pAMPK AMPK -/- APTX-Pos C Control APTX 63 pAMPK AMPK actin -/- Control APTX APTX-Pos CV CIII CIV CII 17 CI actin No extract Control -/- APTX APTX-Pos Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4099 H I −/− Figure 6. Expression of human APTX variant 6 containing an N-terminal putative mitochondrial localization signal in APTX cells. (A) Repair analysis of 5 -AMP DNA in the indicated WCEs. Lower band corresponds to repaired DNA substrate. (B) WB analysis of SOD2 and APTX in the indicated cell lines. (C) Measurement of ATP level in the indicated cells. Values are mean ± S.E.M., N = 3, (***P < 0.001). (D) A representative western blot image of AMPK and activated AMPK (pAMPK). Quantifications are shown as mean ± S.E.M., N = 3, (**P < 0.01). (E) Western blot image of the subunits −/− of OXPHOS complexes I-V. Quantifications are shown as mean ± S.E.M., N = 3, (***P < 0.001). (F) Seahorse measurements of OCR in APTX , APTX-Pos and control cells. Statistical significant was calculated using paired- t-test, N = 3(*P < 0.05). (G) PARylation (PAR) analysis by western blot. Quantifications are shown as mean ± S.E.M., N = 3, (***P < 0.001). (H) PCR-based mtDNA damage analysis. Quantifications are shown as mean ± S.E.M., N = 3, (**P < 0.01). (I) Colony forming analysis to determine the ability of the cells to recover following treatment with the genotoxic drugs MMS and menadione. Quantifications are shown as mean ± S.E.M., N = 3, (***P < 0.001). −/− drial OXPHOS complexes I, II, III and IV were more abun- 6H). Expression of APTX in APTX cells corrected it −/− dant in APTX cells compared with control cells (Figure back to the level of control cells (Figure 6H). These results 6E). Expression of APTX corrected them back to the level support a role for APTX in mtDNA repair and integrity, in of control cells (Figure 6E and the graphs). line with and extension of previous reports (5–7). The effects of APTX deficiency on mitochondrial respi- The number of mtDNA molecules, also known as ration was tested by measuring OCR, using the Seahorse mtDNA copy number, varies significantly depending on −/− XF24 analyzer. APTX and APTX-Pos cells showed the cell type and is thought to correlate with the cellular lower ATP-linked respiration and higher leak relative to need for energy and ATP production. Previously, lentiviral- control cells (Figure 6F and the graph). In contrast, APTX- mediated knockdown of APTX resulted in reduced mtDNA Pos cells showed a higher reserve capacity (Figure 6Fand copy number (5). However, we did not find any differences −/− the graph). Together, these measurements show that the in mtDNA copy number between APTX and AOA1 presence or absence of APTX can alter mitochondrial respi- patient cells and their corresponding control cells (Sup- ration which together with the low MMP (Figure 2A), and plementary Figure S5). These differences might be related −/− high ROS (Figure 2B) in APTX cells suggests mitochon- to the cell type differences used in these studies; here, we drial dysfunction. used U2OS cells and AOA1 patient-derived lymphoblast −/− The level of PARylation was higher in APTX cell ex- cell lines, whereas in the previous study (5), SH-SY5Y and tracts compared with control cells (Figures 6Gand 5C, primary human skeletal muscle cells were used. PAR). APTX expression corrected it back to the level of The colony forming assay is used for monitoring the abil- control cells (Figure 6G and the graph). ity of cells to recover from genotoxic treatment. Next, we MtDNA in APTX-deficient cells is more susceptible determined the recovery capacity of the cells following treat- to damage than nuclear DNA (5,6). Using a PCR-base ment with the genotoxic compounds MMS and menadione. method, we detected more damage in mtDNA from MMS modifies guanine to 7-methylguanine, and adenine to −/− APTX cells than in mtDNA from control cells (Figure 3-methyladenine, which can result in base mispairing and Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4100 Nucleic Acids Research, 2019, Vol. 47, No. 8 Figure 7. AOA1 patient cell lines also display mitochondrial dysfunctions. (A) FACS analysis and quantifications of MMP using TMRM. Bars display TMRM fluorescence intensity as mean ± S.E.M., N = 3(*P < 0.05). (B) FACS analysis and quantifications of mitochondrial ROS level using MitoSOX. Bars display MitoSOX fluorescence intensity as mean ± S.E.M., N = 3(*P < 0.05, **P < 0.01). (C) A representative western blot image of SOD2 and SIRT3 in the indicated WCEs. Graphs show the quantifications of APTX, SOD2, SIRT3, VDAC1 and PGC1  protein levels normalized to PCNA. Values are mean ± S.E.M., N = 3(*P < 0.05, **P < 0.01). (D) Western blot analysis of protein acetylation in mitochondria enriched extracts from control and AOA1 cells. (E) Western blot analysis of PARylation (Par) and PARP1 levels in different WCEs as indicated, quantifications are from three independent experiments. Data are shown as mean ± S.E.M., (**P < 0.01, ***P < 0.001). (F) Western blot analysis of the indicated cells treated with 1 mM NAC for 48 h for PAR. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4101 Figure 8. OPA1 protein and mRNA levels are diminished in AOA1 cells. (A) A representative western blot image of mitochondrial fusion and fission proteins in extracts from the APTX-deficient patient-derived cells L938 and L939 and the control cells C2ABR and C3ABR. Quantifications are shown as mean ± S.E.M., N = 3, (*P < 0.05, **P < 0.01, ***P < 0.001). (B) Q-PCR analysis of OPA1 gene expression in the patients and control cells. Quantifications are shown as mean ± S.E.M., N = 3, (*P < 0.05, **P < 0.01). (C) Western blot analysis of the subunits of OXPHOS, using PCNA as loading control. Quantifications are shown as mean ± S.E.M., N = 3, (**P < 0.01). Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4102 Nucleic Acids Research, 2019, Vol. 47, No. 8 Figure 9. RNA Seq analysis. After comparison of the shared GO terms from GO cellular compartment, three terms were of interest GO:0034599 cellular response to oxidative stress, GO:0022904 respiratory ETC and GO:0008053 mitochondrial fusion. The genes in those three terms are shown in (A–C), respectively. (D) Panel shows genes from the cellular response to oxidative stress (GO:0034599) and mitochondrial fusion (GO:0008053) from AOA1 patient cells versus controls. No mitochondrial ETC genes were found in the AOA1 patient-derived lymphoblastoid cell lines. (E) Genes from the DNA repair pathways base excision, mismatch and NER were also collected and displayed. Note: a few genes were obtained from various lists, were expressed wildly differently and were removed. From the stress response gene list, the gene DHRS2 (KO Ctrl -11.1; Pos Ctrl -9, Pos KO 2.2), from the BER list, APTX (KO Ctrl -0.52; Pos Ctrl 2.6; Pos KO 3.1) and from the NER gene list, the genes GTF2H2C (KO Ctrl 0.97; Pos Ctrl -6.9; Pos KO -7.1) were each removed from the heat maps. All heat maps were centered on zero. Heat maps were made using GraphPad Prism 7. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4103 replication blocks, respectively (57). Menadione exposure 5C) cells showed elevated PAR that was inversely correlated results in the accumulation of ROS in mitochondria that at- with PARP1 concentration. tack and damage macromolecules including mtDNA (58). ROS can chemically modify DNA bases and cause DNA −/− Menadione and MMS treatments showed that APTX stranded breaks (61). To test whether elevated PAR signal- cells were more sensitive to these agents than control cells ing in AOA1 cells was a result of ROS generated DNA dam- (Figure 6I). APTX re-expression reduced the sensitivity of age, we treated the cells with the ROS scavenger NAC. West- −/− APTX cells to these genotoxic agents back to the level ern blot analysis showed a significant reduction in PAR sig- of control cells (Figure 6I). naling (Figure 7F). Collectively, these results suggest that To summarize, putting back APTX into APTX-KO cells loss of APTX resulted in a higher ROS production that to- corrected the protein levels of SOD2, pAMPK, OXPHOS gether with reduced DNA repair capacity results, at least in complexes I, II, III and IV back to the level of control part, in elevated DNA damage level and DDR in the form cells. It also corrected cellular ATP levels, DDR via paryla- of PARylation. tion signal, mtDNA integrity and sensitivity to MMS and menadione treatment back to the level of control cells, and improved mitochondrial respiration reserve capacity, alto- Diminished overall OPA1 level is a common feature of APTX gether suggesting the role of APTX in the observed changes deficient cells −/− in APTX cells. However, the expression of OPA1, SIRT3 Western blot analysis of key regulators of mitochondrial fis- and PARP1 were not corrected by re-introducing APTX −/− sion and fusion showed that the level of the mitochondrial into the APTX cells (Figure 9; Supplementary Figure inner membrane fusion protein OPA1, and DRP1, a cytoso- S6 and Supplementary Tables S1–S3), even though, the lev- −/− lic protein that is recruited to the outer membrane surface els of these proteins were consistently altered in APTX to catalyze the fission, were significantly reduced in AOA1 cells and in AOA1 patient-derived cells (Figures 2D, 4, 5C, cells (Figure 8A). We found no detectable changes in the 7Cand E, 8Aand B, 9; Supplementary Tables S1–S4). amounts of Fis1, an outer mitochondrial membrane protein that mediates fission by acting as DRP1 receptor ( 20,62), AOA1 patient-derived cells recapitulate the abnormal mito- or the mitochondrial fusion protein MFN1, in AOA1 cells −/− chondrial features seen in APTX U2OS cells compared to control cells (Figure 8A). Phosphorylation Next, we used AOA1 patient-derived cells (L938 and L939) of DRP-1 at Serine 616 is required for the recruitment of and control cells (C3ABR and C2ABR) that are Epstein– DRP1 to the surface of mitochondria (63). The ratio of p- Barr virus-transformed lymphoblastoid cell lines (5,25). DRP1 to total DRP1 was unchanged (Figure 8A). The re- MMP analysis showed a significant reduction in MMP duced level of DRP1 in patient cells may be a protective re- (Figure 7A), and a significantly higher level of mitochon- sponse to OPA1 reduction in order to maintain a balance drial ROS production (Figure 7B), and higher SOD2 (Fig- between fusion and fission ( 64). −/− ure 7C) in AOA1 cells compared with the control cells. To- Like in APTX cells (Figure 4B), the expression of −/− gether with the results from the APTX cells (Figure 2A, OPA1 was significantly reduced in AOA1 cells compared B and D), impaired MMP and increased oxidative stress ap- with the control cells (Figure 8B). Thus, APTX-deficiency pear to associate strongly with APTX deficiency. results in down-regulation of OPA1 gene expression. To test whether loss of OPA1 is a characteristic of APTX-deficient −/− cells, we knocked out APTX in HEK cells. APTX HEK Increased acetylation of mitochondrial proteins in AOA1 cells cells also showed a significantly lower OPA1 level compared The key function of SIRT3 deacetylase in mitochondrial with control HEK cells (Supplementary Figure S7). Collec- homeostasis is probably the repair of non-enzymatic acety- tively, these results demonstrate that diminished OPA1 ex- lation of mitochondrial proteins by acetyl-CoA (59). Like in pression and protein level is a common feature of APTX- −/− APTX cells (Figure 2D), the level of SIRT3 was consid- deficient cells. erably higher in AOA1 patient cells compared with the con- OPA1 mRNA undergoes alternative splicing (65)and trol cells (Figure 7C). Western blot analysis of mitochondria some isoforms seem to be specifically important for mito- enriched extracts showed an overall increased level of mito- chondrial network formation (66). To test the possible pres- chondrial protein acetylation in AOA1 cells (Figure 7D). ence of OPA1 isoforms specific to APTX-deficient cells, we Together, these results show that AOA1 patient cells un- carried out reverse transcription-PCR amplification of the dergo mitochondrial stress in the form of higher ROS pro- region flanking exons three to nine (Supplementary Figure duction and acetylation of mitochondrial proteins. S8A). Two putative mRNA isoforms were differently ex- pressed between AOA1 and control cells (primers F1/R1, Significantly reduced level of PARP1 but enhanced PARyla- Supplementary Figure S8B, arrows). These were further tion in AOA1 cells verified by a second primer set to amplify a shorter region between exons three to seven (primers F2/R2). Further- The PARP superfamily consists of 16 members (60). PARP1 more, better separation of OPA1 protein in 6% polyacry- is responsible for most of the cellular PARP activity (51). lamide SDS gel revealed that of the vfi e apparent isoforms, The level of total poly(AD-ribose) (PAR) was markedly three were differentially expressed between APTX-deficient higher in AOA1 cells (Figure 7E), suggesting elevated DDR and control cells (Supplementary Figure S8C, arrows). In signaling. The level of PARP1, however, was significantly conclusion, both the overall level of OPA1 expression and reduced in AOA1 cells compared with control cells (Fig- −/− ure 7E). Together, both AOA1 and APTX U2OS (Figure Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4104 Nucleic Acids Research, 2019, Vol. 47, No. 8 the expression of some putative OPA1 isoforms are altered the ETC genes, there is one clear block of genes (*, Figure in APTX-deficient cells compared to control cells. 9D) whose expression seems to be up in the APTX KO and normalized by APTX rescue. Likewise, the mitochondrial Specific up-regulation of OXPHOS complexes III and V in FIS1 and MFF genes are up-regulated in APTX KO but AOA1 cells normalized after APTX re-expression. In contrast, MFN1, OPA1 and BAX genes were down-regulated in the KO cells We examined the abundance of selected OXPHOS subunits and remained that way after APTX rescue and mitochon- in AOA1 cells and found that complexes CIII and V were −/− drial network changes in APTX cells (Figure 9C). We more abundant in AOA1 cells than in control cells (Fig- should note here that the expression change of OPA1 was ure 8C). Taken together, the results show that APTX defi- very modest, −0.45 log2 fold change in the APTX KO Ctrl ciency consistently leads to the up-regulation of OXPHOS comparison, so OPA1 does not appear to be strongly regu- protein subunits probably as a part of a compensatory tran- lated at the level of transcription in these cells. scriptional response to mitochondrial dysfunction (67,68). We conducted a similar analysis on the AOA1 patient Such response seems to be tissue specific ( 69), which may cells versus controls; roughly half were up- or down- explain why complexes I and II were not affected in AOA1 regulated (1180 genes, 57% down) (Supplementary Table −/− cells but were up-regulated in APTX cells (Figure 6E). S4). From the AOA1 set, the genes from the term mitochon- Moreover, it is known that the type and the level of mito- drion were extracted and the expression of OPA1 (down), chondrial proteins can vary between tissues (70). We spec- LONP (down) and FOXO1 (up) were changed similar as in ulate that the differences in OXPHOS between AOA1 and the APTX KO cells (Figure 9D). There was only one mi- APTX-KO cells may be a result of differential expression of tochondrial electron transport gene in the AOA1 set, AT- OXPHOS subunits and assembly factors in these cells. PAF1, the assembly factor for the F1 complex, which was −/− Collectively, AOA1 and APTX share key features of down-regulated. mitochondrial dysfunctions supporting the involvement of Since APTX is involved in DNA base excision repair APTX in these phenotypes in particular, and in mitochon- (BER), we also sought to determine if DNA repair genes drial homeostasis as a whole. were responsive to either the loss or rescue of APTX expres- sion. BER, nucleotide excision repair (NER) and mismatch Differential gene expression analysis repair (MMR) terms were found in KEGG pathways, and We performed RNA-seq analyses on APTX-KO, and interestingly, NER was statistically significant (adj. P-value APTX-KO cells stably expressing APTX variant 6 (APTX- ≤ 0.05) in all three comparisons (when using the set of genes Pos), and two AOA1 patient-derived cell lines (L939, and that have adjusted P-values ≤ 0.05 and dropping the log2 L939) and the corresponding control cells (Ctrl) (Supple- fold change requirement). MMR was significantly changed mentary Tables S1–S4). There were similar numbers of in APTX-Pos versus APTX-KO comparison (adj. P-value DEGs, APTX versus Ctrl, 1532 (57% down); APTX-Pos 0.0005), while BER was only trending (adj. P-value 0.06). A versus Ctrl, 2590 (59.2% up); APTX-Pos versus APTX-KO, heat map of the combined DNA repair genes list is shown 1384 (72.8% up). For the ATPX-KO cell lines, the signifi- in Figure 9E. Notably, PARP1 is down-regulated in all pair- cant differentially expressed genes were subjected to GO cel- wise comparisons, while the expression of PARP2 is up- lular compartment analysis via Enrichr (31) to uncover po- regulated in APTX KO and normalized after APTX rescue. tential mitochondrial terms. However, the term mitochon- drion (GO 0005739) was not among the top terms for any The Caenorhabditis elegans model of OPA-1 deficiency reca- pairwise comparison. However, since we were interested in pitulates the stress phonotypes in APTX-deficient cells defining the mitochondrial terms that were changed after APTX KO and APTX add back, we proceeded and ac- To date, there is no known C. elegans homolog of APTX quired the DEG gene lists from the mitochondrion GO (72). Therefore, to isolate the functional consequences of term and wanted to subject that list to GO biological func- OPA1 loss on an organismal level, the C. elegans eat-3 tion analysis but each pairwise comparison had too few strain was examined. Caenorhabditis elegans eat-3 (ad426) genes (APTX KO Ctrl, 42; APTX POS Ctrl, 107, APTX strain (73) has a mutation in the D2013.5 gene, which en- POS Ctrl) for this to be meaningful. So, we collected all codes the ortholog of yeast Mgm1 and mammalian Opa1 the genes with an adjusted P-values ≤ 0.05, dropped the (74). The ad426 mutation leads to fragmented mitochon- log2 fold change cutoff requirement, and collected the mito- dria similar to those cause by mutations in Opa1 and Mgm1 chondrion term (GO 0005739) gene lists from each pairwise (74). As previously reported, the eat-3 (ad426) strain show comparison. Not imposing a cutoff change is routinely done disrupted mitochondrial network by TMRM live staining in Gene Set Enrichment Analysis (71). Mitochondrion was (Figure 10A). Similar to AOA1 patient cells, the level of a top scoring term in each pairwise comparison using this PARylation and acetylated proteins is also increased in eat- input gene list (Supplementary Table S5). After GO biolog- 3 (ad426) worms (Figure 10B and C). We also measured ical process analysis, the GO terms to cellular response to ATP levels and found that eat-3 (ad426) worms have a much oxidative stress (Figure 9A), respiratory ETC (Figure 9B), lower level of ATP compared with N2 worms (Figure 10D), mitochondrial fusion (Figure 9C), and several terms related suggesting mitochondrial function is impaired. To further to translation were of interest to us (see Supplementary Ta- explore the biological importance of the integrity of mi- ble S6 for the entire top 20 list of terms). SOD2 was among tochondrial network, we performed chemotaxis assay by the up-regulated genes in ATPX KO cells that was down- using isoamyl alcohol, which is an attractive odor to the regulated after adding back APTX (Figure 9A). Among worms. In N2 worms, ∼75% were attached by isoamyl al- Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4105 Figure 10. The Caenorhabditis elegans Opa1 homolog eat-3 recapitulates some of the stress phenotypes and responses seen in APTX-defect cells. (A) Visualization of mitochondrial network by treating the worms with TMRM. Scale bar represents 10 m. (B) Western blot analysis of C. elegans extracts for (B) parylation (Par) and (C) total protein acetylation, (D) ATP measurements, (E) chemotaxis assay and (F) swimming assays. Data is presented as mean ± S.E.M., (** P < 0.01, *** P < 0.001). Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4106 Nucleic Acids Research, 2019, Vol. 47, No. 8 cohol after 1 h. In eat-3 (ad426) worms, only 22% were at- with significant variations of age of onset and severity of tracted by isoamyl alcohol (Figure 10E). To exclude the pos- the disease (81). The extra-ocular symptoms of OPA1 de- sibility that the difference in chemotaxis index is caused by ficiency include deafness, ataxia, myopathy and spinocere- motor dysfunction, we performed a swimming assay. The bellar degeneration (82–84). Thus, AOA1 and OPA1-related result showed no significant differences between N2 and eat- neurodegeneration share phenotypes suggesting that mito- 3 (ad426) worms (Figure 10F). Thus, decreased chemotaxis chondrial dysfunction is a common pathway in these dis- index, together with normal motor function, suggest that eases. Moreover, a mitochondrial disease database (10)pre- eat-3 (ad426) worms have deficiencies in sensing isoamyl dicts AOA1 and DOA diseases to have significant mito- alcohol. In worms, two pairs of amphid sensory neurons, chondrial involvement (Supplementary Figure S9). Fibrob- AWC and AWA, are required for chemotaxis to isoamyl al- lasts from patients with OPA1 mutation display mitochon- cohol (75). Taken together, we conclude that in nematode, drial network fragmentation (85,86). Homozygous Opa1 eat-3 (Opa1) deficiency leads to the disruption of mitochon- mutant mice die in utero during embryogenesis, but het- drial network, and impairment of sensory neurons. erozygous Opa1 mutants display the main features of hu- man DOA including abnormal mitochondrial morphology, disorganized cristae structure, mitochondrial dysfunction DISCUSSION and mtDNA instability (87–89). Eat-3 is the C. elegans ho- molog of human OPA1. Mitochondria in eat-3 mutant C. In this study, we show that APTX-deficient cells displayed hallmarks of mitochondrial dysfunction and stress, i.e. low elegans strain are fragmented and show phenotypes consis- MMP, increased mitochondrial ROS production, suscepti- tent with defects in OXPHOS system (74). Thus, ample ev- bility to mtDNA damage, elevated acetylation of mitochon- idence shows a key role of OPA1 in mitochondrial function drial proteins, changed OXPHOS protein abundance, al- and network formation across species. As such, OPA1 re- tered mitochondrial morphology and impaired mitophagy. duction may, at least in part, explain the observed loss of A significant reduction in the level of a protein does not mitochondrial network (Figure 3A and B), and impaired necessarily recapitulate the phenotypes caused by a com- cristae in APTX-deficient cells (Figure 3C). plete loss of that protein. Germline deletion of DNA repair In humans, OPA1 is present in multiple isoforms gen- protein Xrcc1 in mouse is embryonic lethal; however, Xrcc1 erated by alternative splicing of mRNA at exons 4, 4b at levels nearly 10% of normal cells support embryonic vi- and 5b, which further undergo proteolysis by the mito- ability (76). We also found that extracts from cells with chondrial inner-membrane peptidases YME1L and OMA1 APTX as low as 5% of the normal cells displayed robust (65,90,91). PCR amplification of the OPA1 cDNA, and 5 -AMP removal activity compared with the control cells WB analysis showed specific changes in the expression of (Figure 1A and B). This result may partially explain the di- OPA1 isoforms between AOA1 and control cells (Supple- verse clinical symptoms and complex genotype-phenotype mentary Figure S8). Furthermore, the alternative splicing correlations in AOA1 patients given different rates of degra- analysis of OPA1 in RNA-seq data showed the preferen- dation and instability of APTX mutants (9,77,78). This also tial expression of isoforms 1, 7 and 8 in U2OS and AOA1- demonstrates that it is of great importance to achieve a patient cells (Supplementary Figure S10). Interestingly, iso- complete depletion in order to characterize the functions of form 7 was down-regulated in both AOA1-patient cells −/− aprataxin. Such a strategy might be necessary in the study andinAPTX cells. The OPA1 isoform 1 was, in addi- of most DNA repair proteins. tion, down-regulated in AOA1-patient cells. Thus, APTX- Mitochondria are structurally highly dynamic organelles. deficiency is linked to altered expression of OPA1 isoforms, which may contribute to the mitochondrial morphology They constantly change shape, size, and form intercon- and network changes in APTX-deficient cells. To determine nected networks in response to environmental cues. Mito- chondrial size and morphology are determined by the rate the biological significance of these differences is technically of fission and fusion, which together with mitophagy and challenging, because both loss and overexpression of OPA1 biogenesis, regulate the characteristic mitochondrial net- causes fragmentation of the mitochondrial network and dis- work. Mitochondrial morphology varies across cell types organization of cristae (66,92), indicating that balanced ex- and tissues and is highly adaptive to metabolic signaling pression of OPA1 is important for proper mitochondrial and stress (22,40). Defects or changes in the expression of function and network formation. components of mitochondrial fission and fusion cause dis- One limitation of this study is that re-introducing APTX −/− ease including neurodegeneration and are associated with into APTX cells did not correct the expression of OPA1 −/− the normal aging process (18,24). The expression and pro- and mitochondrial network in APTX cells. Moreover, tein level of the inner membrane fusion protein OPA1 was the specific function of APTX in the nucleus and mitochon- consistently lower in APTX-deficient cells (Figures 4Aand dria in the observed mitochondrial phenotypes was not ad- B. 8Aand B, 9 Supplementary Tables S1–S3). OPA1 is a dressed here. dynamin-like GTPase localized in the mitochondrial inner It has been shown that APTX regulates the expression membrane and is commonly thought to play key roles in mi- and the stability of PARP1 (93). RNA-seq results showed tochondrial structure and dynamics by mediating inner mi- that the expression of PARP1 was somewhat lower in AOA1 −/− tochondrial membrane fusion and by controlling the cristae patient cells and in APTX cells compared to control cells shape (46,47). OPA1 mutations were first identified in dom- (Supplementary Tables S1–S4 and Figure 9). PARP2 is an- inant optic atrophy (DOA), a disease specifically affecting other PARP family member with PAR synthesis activity retinal ganglion cells (79,80). OPA1 deficiency has also been and is expressed at ∼15% of PARP1 (60). The level of the −/− identified in patients with clinically diverse symptoms and expression of PARP2 was somewhat higher in APTX Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4107 cells compared to control cells (Figure 9 and Supplemen- SUPPLEMENTARY DATA tary Table S1–S3). In AOA1 patient cells, however, there Supplementary Data are available at NAR Online. were no significant changes in PARP2 expression compared −/− with control cells (Supplementary Table S4). In APTX ACKNOWLEDGEMENTS cells, PARP2 also showed increased expression. Several of the PARP members were differentially expressed in APTX- We like to thank Jane Tian for bioenergetic experiments deficient cells (Figure 9 and Supplementary Tables S1–S4). and for help with the microarray analysis. We would like to Poly-(ADP-ribose)glycohydrolyase (PARG) is responsible thank Dr Elin Lehmann, Dr Yongqing Zhang and Dr Kevin for the degradation of poly-(ADP-ribose). There was no G. Becker of the Gene Expression and Genomics Unit, NIA significant difference in the level of expression of PARG in Intramural Program, NIH. We would like to thank Kavya −/− AOA1 or APTX cells compared to their corresponding Achanta for her help with C. elegans study. We would like control cells (Supplementary Tables S1–S3) suggesting that to thank Dr Daniel R. McNeill for C2/3ABR and L938/9 the rate of PAR turnover did not account for the observed cell lines. We would like to thank Dr Beimeng Yang and Dr elevated PAR levels in APTX-deficient cells. Taken together, Anthony Moore for critically reading the manuscript. Con- defect in APTX results in an elevated level of PARylation. focal and electron microscopy experiments were carried out The fork-head associated (FHA) domain is a phospho- at the Core Facility of Integrated Microscopy (CFIM), Uni- peptide interacting domain associated with proteins in- versity of Copenhagen. volved in a number of processes including intracellular sig- nal transduction, transcription, protein transport, DNA FUNDING repair and protein degradation (94). APTX seems to in- teract with several proteins through its FHA domain. In NORDEA Foundation, Denmark (02-2013-0220); EU the nucleus, APTX was reported to interact with DNA re- Joint Programme-Neurodegenerative Disease Research pair proteins XRCC1, PARP1 and also with the transcrip- (JPND); Innovation Fund Denmark (5188-00001); Olav tion factor p53 (25,95–97). A recently developed concept Thon foundation Norway (531811-710131); Novo Nordisk connects nuclear DNA damage signaling to mitochondrial foundation Denmark (NNF17OC0027812); Intramural Re- function and maintenance. According to this model, persis- search Program of the NIH, National Institute on Ag- tent DNA damage, e.g. because of a defect in DNA repair ing (AG000733). Funding for open access charge: Univer- or diminished DNA repair capacity, activates prolonged sity of Copenhagen. DDR by PARP1/2 in the form of PAR being added to Conflict of interest statement. None declared. proteins in the vicinity of the DNA damage. This process consumes NAD , an essential cofactor involved in a num- REFERENCES ber of key processes including mitochondrial bioenergetics and mitophagy. Thus, persistent DDR leads to a competi- 1. Ellenberger,T. and Tomkinson,A.E. (2008) Eukaryotic DNA ligases: tion between these processes for the available NAD lead- structural and functional insights. Annu. Rev. Biochem., 77, 313–338. 2. Ahel,I., Rass,U., El-Khamisy,S.F., Katyal,S., Clements,P.M., ing to mitochondrial dysfunction (23). Moreover, the cel- + McKinnon,P.J., Caldecott,K.W. and West,S.C. (2006) The lular bioavailability of NAD declines during normal ag- neurodegenerative disease protein aprataxin resolves abortive DNA ing (98,99). In support of this model, NAD replenishment ligation intermediates. Nature, 443, 713–716. improves health and lifespan in animal models of human 3. Date,H., Onodera,O., Tanaka,H., Iwabuchi,K., Uekawa,K., Igarashi,S., Koike,R., Hiroi,T., Yuasa,T., Awaya,Y. et al. (2001) disease (11). APTX-deficient cells consistently contained re- Early-onset ataxia with ocular motor apraxia and hypoalbuminemia duced levels of PARP1 but elevated PARylation compared is caused by mutations in a new HIT superfamily gene. Nat. Genet., with control cells (Figures 5C, 6Gand 7E and F). Treat- 29, 184–188. ment of the cells with the ROS scavenger NAC significantly 4. Moreira,M.C., Barbot,C., Tachi,N., Kozuka,N., Uchida,E., reduced PARylation in APTX-deficient cells (Figure 7F). Gibson,T., Mendonca,P., Costa,M., Barros,J., Yanagisawa,T. et al. (2001) The gene mutated in ataxia-ocular apraxia 1 encodes the new Thus, we speculate that the lack of APTX leads to higher HIT/Zn-finger protein aprataxin. Nat. Genet., 29, 189–193. ROS production and reduced PARP1 level, which in turn re- 5. Sykora,P., Croteau,D.L., Bohr,V.A. and Wilson,D.M. 3rd. (2011) sults in diminished basal DDR and DNA repair, persistence Aprataxin localizes to mitochondria and preserves mitochondrial of PARylation, which over time may contribute to the im- function. Proc. Natl. Acad. Sci. U.S.A., 108, 7437–7442. 6. Akbari,M., Sykora,P. and Bohr,V.A. (2015) Slow mitochondrial pairment of mitochondrial homeostasis in APTX-deficient repair of 5 -AMP renders mtDNA susceptible to damage in APTX cells. deficient cells. Sci. Rep., 5, 12876. In conclusion, our data links impaired mitochondrial 7. Caglayan,M., Prasad,R., Krasich,R., Longley,M.J., Kadoda,K., homeostasis to APTX deficiency and AOA1 pathology. In Tsuda,M., Sasanuma,H., Takeda,S., Tano,K., Copeland,W.C. et al. recent years, mitochondria have become targets for phar- (2017) Complementation of aprataxin deficiency by base excision repair enzymes in mitochondrial extracts. Nucleic Acids Res., 45, macological interventions for treatment of neurodegenera- 10079–10088. tive diseases with promising results (23,100). In the future, 8. Quinzii,C.M., Kattah,A.G., Naini,A., Akman,H.O., Mootha,V.K., perhaps such strategies should be tested on AOA1 patients. DiMauro,S. and Hirano,M. (2005) Coenzyme Q deficiency and cerebellar ataxia associated with an aprataxin mutation. Neurology, 64, 539–541. 9. Renaud,M., Moreira,M.C., Ben Monga,B., Rodriguez,D., Debs,R., DATA AVAILABILITY Charles,P., Chaouch,M., Ferrat,F., Laurencin,C., Vercueil,L. et al. (2018) Clinical, biomarker, and molecular delineations and The RNA-seq data has been deposited to the GEO database Genotype-Phenotype correlations of ataxia with oculomotor under series number GSE124412. apraxia type 1. JAMA Neurol., 75, 495–502. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4108 Nucleic Acids Research, 2019, Vol. 47, No. 8 10. Scheibye-Knudsen,M., Scheibye-Alsing,K., Canugovi,C., 33. Ryu,D., Mouchiroud,L., Andreux,P.A., Katsyuba,E., Moullan,N., Croteau,D.L. and Bohr,V.A. (2013) A novel diagnostic tool reveals Nicolet-Dit-Felix,A.A., Williams,E.G., Jha,P., Lo Sasso,G., mitochondrial pathology in human diseases and aging. Aging, 5, Huzard,D. et al. (2016) Urolithin A induces mitophagy and 192–208. prolongs lifespan in C. elegans and increases muscle function in 11. Fang,E.F., Kassahun,H., Croteau,D.L., Scheibye-Knudsen,M., rodents. Nat. Med., 22, 879–888. Marosi,K., Lu,H., Shamanna,R.A., Kalyanasundaram,S., 34. Bargmann,C.I. and Horvitz,H.R. (1991) Chemosensory neurons Bollineni,R.C., Wilson,M.A. et al. (2016) NAD(+) replenishment with overlapping functions direct chemotaxis to multiple chemicals improves lifespan and healthspan in ataxia telangiectasia models via in C. elegans. Neuron., 7, 729–742. mitophagy and DNA repair. Cell Metab., 24, 566–581. 35. Akbari,M., Keijzers,G., Maynard,S., Scheibye-Knudsen,M., 12. Brand,M.D. and Nicholls,D.G. (2011) Assessing mitochondrial Desler,C., Hickson,I.D. and Bohr,V.A. (2014) Overexpression of dysfunction in cells. Biochem. J., 435, 297–312. DNA ligase III in mitochondria protects cells against oxidative 13. Cornelius,N., Wardman,J.H., Hargreaves,I.P., Neergheen,V., stress and improves mitochondrial DNA base excision repair. DNA Bie,A.S., Tumer,Z., Nielsen,J.E. and Nielsen,T.T. (2017) Evidence of Repair, 16, 44–53. oxidative stress and mitochondrial dysfunction in spinocerebellar 36. Starkov,A.A. and Fiskum,G. (2003) Regulation of brain ataxia type 2 (SCA2) patient fibroblasts: effect of coenzyme Q10 mitochondrial H2O2 production by membrane potential and supplementation on these parameters. Mitochondrion, 34, 103–114. NAD(P)H redox state. J. Neurochem., 86, 1101–1107. 14. Fang,E.F., Scheibye-Knudsen,M., Brace,L.E., Kassahun,H., 37. Lombard,D.B., Alt,F.W., Cheng,H.L., Bunkenborg,J., Streeper,R.S., SenGupta,T., Nilsen,H., Mitchell,J.R., Croteau,D.L. and Bohr,V.A. Mostoslavsky,R., Kim,J., Yancopoulos,G., Valenzuela,D., (2014) Defective mitophagy in XPA via PARP-1 hyperactivation and Murphy,A. et al. (2007) Mammalian Sir2 homolog SIRT3 regulates NAD(+)/SIRT1 reduction. Cell, 157, 882–896. global mitochondrial lysine acetylation. Mol. Cell. Biol., 27, 15. Bose,A. and Beal,M.F. (2016) Mitochondrial dysfunction in 8807–8814. Parkinson’s disease. J. Neurochem., 139, 216–231. 38. Qiu,X., Brown,K., Hirschey,M.D., Verdin,E. and Chen,D. (2010) 16. Costa,V. and Scorrano,L. (2012) Shaping the role of mitochondria in Calorie restriction reduces oxidative stress by SIRT3-mediated the pathogenesis of Huntington’s disease. EMBO J., 31, 1853–1864. SOD2 activation. Cell Metab., 12, 662–667. 17. Scheibye-Knudsen,M., Fang,E.F., Croteau,D.L., Wilson,D.M. 3rd 39. Cogliati,S., Frezza,C., Soriano,M.E., Varanita,T., and Bohr,V.A. (2015) Protecting the mitochondrial powerhouse. Quintana-Cabrera,R., Corrado,M., Cipolat,S., Costa,V., Trends Cell Biol., 25, 158–170. Casarin,A., Gomes,L.C. et al. (2013) Mitochondrial cristae shape 18. Sebastian,D., Palacin,M. and Zorzano,A. (2017) Mitochondrial determines respiratory chain supercomplexes assembly and dynamics: coupling mitochondrial fitness with healthy aging. Trends respiratory efficiency. Cell, 155, 160–171. Mol. Med., 23, 201–215. 40. Westermann,B. (2012) Bioenergetic role of mitochondrial fusion and 19. Loson,O.C., Song,Z., Chen,H. and Chan,D.C. (2013) Fis1, Mff, fission. Biochim. Biophys. Acta, 1817, 1833–1838. MiD49, and MiD51 mediate Drp1 recruitment in mitochondrial 41. Walczak,J., Partyka,M., Duszynski,J. and Szczepanowska,J. (2017) fission. Mol. Biol. Cell, 24, 659–667. Implications of mitochondrial network organization in 20. Otera,H., Wang,C., Cleland,M.M., Setoguchi,K., Yokota,S., mitochondrial stress signalling in NARP cybrid and Rho0 cells. Sci. Youle,R.J. and Mihara,K. (2010) Mff is an essential factor for Rep., 7, 14864. mitochondrial recruitment of Drp1 during mitochondrial fission in 42. Al-Mehdi,A.B., Pastukh,V.M., Swiger,B.M., Reed,D.J., Patel,M.R., mammalian cells. J. Cell Biol., 191, 1141–1158. Bardwell,G.C., Pastukh,V.V., Alexeyev,M.F. and Gillespie,M.N. 21. Liesa,M. and Shirihai,O.S. (2013) Mitochondrial dynamics in the (2012) Perinuclear mitochondrial clustering creates an oxidant-rich regulation of nutrient utilization and energy expenditure. Cell nuclear domain required for hypoxia-induced transcription. Sci. Metab., 17, 491–506. Signal., 5, ra47. 22. Wai,T. and Langer,T. (2016) Mitochondrial dynamics and metabolic 43. Merkwirth,C., Dargazanli,S., Tatsuta,T., Geimer,S., Lower,B., regulation. Trends Endocrinol. Metab., 27, 105–117. Wunderlich,F.T., von Kleist-Retzow,J.C., Waisman,A., 23. Fang,E.F., Scheibye-Knudsen,M., Chua,K.F., Mattson,M.P., Westermann,B. and Langer,T. (2008) Prohibitins control cell Croteau,D.L. and Bohr,V.A. (2016) Nuclear DNA damage proliferation and apoptosis by regulating OPA1-dependent cristae signalling to mitochondria in ageing. Nat. Rev. Mol. Cell Biol., 17, morphogenesis in mitochondria. Genes Dev., 22, 476–488. 308–321. 44. Yamaguchi,R., Lartigue,L., Perkins,G., Scott,R.T., Dixit,A., 24. Archer,S.L. (2013) Mitochondrial dynamics–mitochondrial fission Kushnareva,Y., Kuwana,T., Ellisman,M.H. and Newmeyer,D.D. and fusion in human diseases. N. Engl.J.Med., 369, 2236–2251. (2008) Opa1-mediated cristae opening is Bax/Bak and BH3 25. Gueven,N., Becherel,O.J., Kijas,A.W., Chen,P., Howe,O., dependent, required for apoptosis, and independent of Bak Rudolph,J.H., Gatti,R., Date,H., Onodera,O., Taucher-Scholz,G. oligomerization. Mol. Cell, 31, 557–569. et al. (2004) Aprataxin, a novel protein that protects against 45. Elachouri,G., Vidoni,S., Zanna,C., Pattyn,A., Boukhaddaoui,H., genotoxic stress. Hum. Mol. Genet., 13, 1081–1093. Gaget,K., Yu-Wai-Man,P., Gasparre,G., Sarzi,E., Delettre,C. et al. 26. Santos,J.H., Mandavilli,B.S. and Van Houten,B. (2002) Measuring (2011) OPA1 links human mitochondrial genome maintenance to oxidative mtDNA damage and repair using quantitative PCR. mtDNA replication and distribution. Genome Res., 21, 12–20. Methods Mol. Biol., 197, 159–176. 46. Frezza,C., Cipolat,S., Martins de Brito,O., Micaroni,M., 27. Hunter,S.E., Jung,D., Di Giulio,R.T. and Meyer,J.N. (2010) The Beznoussenko,G.V., Rudka,T., Bartoli,D., Polishuck,R.S., QPCR assay for analysis of mitochondrial DNA damage, repair, Danial,N.N., De Strooper,B. et al. (2006) OPA1 controls apoptotic and relative copy number. Methods, 51, 444–451. cristae remodeling independently from mitochondrial fusion. Cell, 28. Langmead,B. and Salzberg,S.L. (2012) Fast gapped-read alignment 126, 177–189. with Bowtie 2. Nat. Methods, 9, 357–359. 47. MacVicar,T. and Langer,T. (2016) OPA1 processing in cell death 29. Li,B. and Dewey,C.N. (2011) RSEM: accurate transcript and disease - the long and short of it. J. Cell Sci., 129, 2297–2306. quantification from RNA-Seq data with or without a reference 48. Pickles,S., Vigie,P. and Youle,R.J. (2018) Mitophagy and quality genome. BMC Bioinformatics, 12, 323. control mechanisms in mitochondrial maintenance. Curr. Biol., 28, 30. Love,M.I., Huber,W. and Anders,S. (2014) Moderated estimation of R170–R185. fold change and dispersion for RNA-seq data with DESeq2. Genome 49. Kabeya,Y., Mizushima,N., Ueno,T., Yamamoto,A., Kirisako,T., Biol., 15, 550. Noda,T., Kominami,E., Ohsumi,Y. and Yoshimori,T. (2000) LC3, a 31. Kuleshov,M.V., Jones,M.R., Rouillard,A.D., Fernandez,N.F., mammalian homologue of yeast Apg8p, is localized in Duan,Q., Wang,Z., Koplev,S., Jenkins,S.L., Jagodnik,K.M., autophagosome membranes after processing. EMBO J., 19, Lachmann,A. et al. (2016) Enrichr: a comprehensive gene set 5720–5728. enrichment analysis web server 2016 update. Nucleic Acids Res., 44, 50. Pascal,J.M. and Ellenberger,T. (2015) The rise and fall of W90–W97. poly(ADP-ribose): an enzymatic perspective. DNA Repair, 32, 32. Merrill Ronald,A. (2017) In: Strack,S and 10–16. Usachev,Y (eds). Techniques to investigate mitochondrial function in 51. Ame,J.C., Rolli,V., Schreiber,V., Niedergang,C., Apiou,F., Decker,P., neurons. Humana Press, NY, Vol. 123, pp. 31–48. Muller,S., Hoger,T., Menissier-de Murcia,J. and de Murcia,G. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4109 (1999) PARP-2, A novel mammalian DNA damage-dependent 72. Rass,U., Ahel,I. and West,S.C. (2008) Molecular mechanism of poly(ADP-ribose) polymerase. J. Biol. Chem., 274, 17860–17868. DNA deadenylation by the neurological disease protein aprataxin. J. 52. Lin,S.C. and Hardie,D.G. (2018) AMPK: sensing glucose as well as Biol. Chem., 283, 33994–34001. cellular energy status. Cell Metab., 27, 299–313. 73. Avery,L. (1993) The genetics of feeding in Caenorhabditis elegans. 53. Kim,J., Kundu,M., Viollet,B. and Guan,K.L. (2011) AMPK and Genetics, 133, 897–917. mTOR regulate autophagy through direct phosphorylation of Ulk1. 74. Kanazawa,T., Zappaterra,M.D., Hasegawa,A., Wright,A.P., Nat. Cell Biol., 13, 132–141. Newman-Smith,E.D., Buttle,K.F., McDonald,K., Mannella,C.A. 54. Hardie,D.G., Schaffer,B.E. and Brunet,A. (2016) AMPK: an and van der Bliek,A.M. (2008) The C. elegans Opa1 homologue energy-sensing pathway with multiple inputs and outputs. Trends EAT-3 is essential for resistance to free radicals. PLoS Genet., 4, Cell Biol., 26, 190–201. e1000022. 55. Koopman,W.J., Distelmaier,F., Smeitink,J.A. and Willems,P.H. 75. Bargmann,C.I., Hartwieg,E. and Horvitz,H.R. (1993) (2013) OXPHOS mutations and neurodegeneration. EMBO J., 32, Odorant-selective genes and neurons mediate olfaction in C. 9–29. elegans. Cell, 74, 515–527. 56. Winge,D.R. (2012) Sealing the mitochondrial respirasome. Mol. 76. Tebbs,R.S., Thompson,L.H. and Cleaver,J.E. (2003) Rescue of Cell Biol., 32, 2647–2652. Xrcc1 knockout mouse embryo lethality by 57. Beranek,D.T. (1990) Distribution of methyl and ethyl adducts transgene-complementation. DNA Repair, 2, 1405–1417. following alkylation with monofunctional alkylating agents. Mutat. 77. Hirano,M., Matsumura,R., Nakamura,Y., Saigoh,K., Res., 231, 11–30. Sakamoto,H., Ueno,S., Inoue,H. and Kusunoki,S. (2017) 58. Fukui,M., Choi,H.J. and Zhu,B.T. (2012) Rapid generation of Unexpectedly mild phenotype in an ataxic family with a two-base mitochondrial superoxide induces mitochondrion-dependent but deletion in the APTX gene. J. Neurol. Sci., 378, 75–79. caspase-independent cell death in hippocampal neuronal cells that 78. Yokoseki,A., Ishihara,T., Koyama,A., Shiga,A., Yamada,M., morphologically resembles necroptosis. Toxicol. Appl. Pharmacol., Suzuki,C., Sekijima,Y., Maruta,K., Tsuchiya,M., Date,H. et al. 262, 156–166. (2011) Genotype-phenotype correlations in early onset ataxia with 59. Weinert,B.T., Moustafa,T., Iesmantavicius,V., Zechner,R. and ocular motor apraxia and hypoalbuminaemia. Brain, 134, Choudhary,C. (2015) Analysis of acetylation stoichiometry suggests 1387–1399. that SIRT3 repairs nonenzymatic acetylation lesions. EMBO J., 34, 79. Alexander,C., Votruba,M., Pesch,U.E., Thiselton,D.L., Mayer,S., 2620–2632. Moore,A., Rodriguez,M., Kellner,U., Leo-Kottler,B., Auburger,G. 60. Ame,J.C., Spenlehauer,C. and de Murcia,G. (2004) The PARP et al. (2000) OPA1, encoding a dynamin-related GTPase, is mutated superfamily. Bioessays, 26, 882–893. in autosomal dominant optic atrophy linked to chromosome 3q28. 61. Akbari,M. and Krokan,H.E. (2008) Cytotoxicity and mutagenicity Nat. Genet., 26, 211–215. of endogenous DNA base lesions as potential cause of human aging. 80. Delettre,C., Lenaers,G., Griffoin,J.M., Gigarel,N., Lorenzo,C., Mech. Ageing Dev., 129, 353–365. Belenguer,P., Pelloquin,L., Grosgeorge,J., Turc-Carel,C., Perret,E. 62. Mozdy,A.D., McCaffery,J.M. and Shaw,J.M. (2000) Dnm1p et al. (2000) Nuclear gene OPA1, encoding a mitochondrial GTPase-mediated mitochondrial fission is a multi-step process dynamin-related protein, is mutated in dominant optic atrophy. Nat. requiring the novel integral membrane component Fis1p. J. Cell Genet., 26, 207–210. Biol., 151, 367–380. 81. Chao de la Barca,J.M., Prunier-Mirebeau,D., Amati-Bonneau,P., 63. Kashatus,D.F., Lim,K.H., Brady,D.C., Pershing,N.L., Cox,A.D. Ferre,M., Sarzi,E., Bris,C., Leruez,S., Chevrollier,A., and Counter,C.M. (2011) RALA and RALBP1 regulate Desquiret-Dumas,V., Gueguen,N. et al. (2016) OPA1-related mitochondrial fission at mitosis. Nat. Cell Biol., 13, 1108–1115. disorders: diversity of clinical expression, modes of inheritance and 64. Manczak,M., Kandimalla,R., Fry,D., Sesaki,H. and Reddy,P.H. pathophysiology. Neurobiol. Dis., 90, 20–26. (2016) Protective effects of reduced dynamin-related protein 1 82. Schaaf,C.P., Blazo,M., Lewis,R.A., Tonini,R.E., Takei,H., Wang,J., against amyloid beta-induced mitochondrial dysfunction and Wong,L.J. and Scaglia,F. (2011) Early-onset severe neuromuscular synaptic damage in Alzheimer’s disease. Hum. Mol. Genet., 25, phenotype associated with compound heterozygosity for OPA1 5148–5166. mutations. Mol. Genet. Metab., 103, 383–387. 65. Delettre,C., Griffoin,J.M., Kaplan,J., Dollfus,H., Lorenz,B., 83. Carelli,V., Sabatelli,M., Carrozzo,R., Rizza,T., Schimpf,S., Faivre,L., Lenaers,G., Belenguer,P. and Hamel,C.P. (2001) Mutation Wissinger,B., Zanna,C., Rugolo,M., La Morgia,C., Caporali,L. spectrum and splicing variants in the OPA1 gene. Hum. Genet., 109, et al. (2015) ‘Behr syndrome’ with OPA1 compound heterozygote 584–591. mutations. Brain, 138, e321. 66. Del Dotto,V., Mishra,P., Vidoni,S., Fogazza,M., Maresca,A., 84. Bonneau,D., Colin,E., Oca,F., Ferre,M., Chevrollier,A., Caporali,L., McCaffery,J.M., Cappelletti,M., Baruffini,E., Gueguen,N., Desquiret-Dumas,V., N’Guyen,S., Barth,M., Lenaers,G. et al. (2017) OPA1 isoforms in the hierarchical Zanlonghi,X. et al. (2014) Early-onset Behr syndrome due to organization of mitochondrial functions. Cell Rep., 19, 2557–2571. compound heterozygous mutations in OPA1. Brain, 137, e301. 67. Heddi,A., Stepien,G., Benke,P.J. and Wallace,D.C. (1999) 85. Liao,C., Ashley,N., Diot,A., Morten,K., Phadwal,K., Williams,A., Coordinate induction of energy gene expression in tissues of Fearnley,I., Rosser,L., Lowndes,J., Fratter,C. et al. (2017) mitochondrial disease patients. J. Biol. Chem., 274, 22968–22976. Dysregulated mitophagy and mitochondrial organization in optic 68. Fakouri,N.B., Durhuus,J.A., Regnell,C.E., Angleys,M., Desler,C., atrophy due to OPA1 mutations. Neurology, 88, 131–142. Hasan-Olive,M.M., Martin-Pardillos,A., Tsaalbi-Shtylik,A., 86. Bonifert,T., Karle,K.N., Tonagel,F., Batra,M., Wilhelm,C., Thomsen,K., Lauritzen,M. et al. (2017) Rev1 contributes to proper Theurer,Y., Schoenfeld,C., Kluba,T., Kamenisch,Y., Carelli,V. et al. mitochondrial function via the PARP-NAD(+)-SIRT1-PGC1alpha (2014) Pure and syndromic optic atrophy explained by deep intronic axis. Sci. Rep., 7, 12480. OPA1 mutations and an intralocus modifier. Brain, 137, 2164–2177. 69. van der Westhuizen,F.H., van den Heuvel,L.P., Smeets,R., 87. Alavi,M.V., Bette,S., Schimpf,S., Schuettauf,F., Schraermeyer,U., Veltman,J.A., Pfundt,R., van Kessel,A.G., Ursing,B.M. and Wehrl,H.F., Ruttiger,L., Beck,S.C., Tonagel,F., Pichler,B.J. et al. Smeitink,J.A. (2003) Human mitochondrial complex I deficiency: (2007) A splice site mutation in the murine Opa1 gene features investigating transcriptional responses by microarray. pathology of autosomal dominant optic atrophy. Brain, 130, Neuropediatrics, 34, 14–22. 1029–1042. 70. Pagliarini,D.J., Calvo,S.E., Chang,B., Sheth,S.A., Vafai,S.B., 88. Chen,L., Liu,T., Tran,A., Lu,X., Tomilov,A.A., Davies,V., Ong,S.E., Walford,G.A., Sugiana,C., Boneh,A., Chen,W.K. et al. Cortopassi,G., Chiamvimonvat,N., Bers,D.M., Votruba,M. et al. (2008) A mitochondrial protein compendium elucidates complex I (2012) OPA1 mutation and late-onset cardiomyopathy: disease biology. Cell, 134, 112–123. mitochondrial dysfunction and mtDNA instability. J. Am. Heart 71. Mootha,V.K., Lindgren,C.M., Eriksson,K.F., Subramanian,A., Assoc., 1, e003012. Sihag,S., Lehar,J., Puigserver,P., Carlsson,E., Ridderstrale,M., 89. Davies,V.J., Hollins,A.J., Piechota,M.J., Yip,W., Davies,J.R., Laurila,E. et al. (2003) PGC-1alpha-responsive genes involved in White,K.E., Nicols,P.P., Boulton,M.E. and Votruba,M. (2007) Opa1 oxidative phosphorylation are coordinately downregulated in deficiency in a mouse model of autosomal dominant optic atrophy human diabetes. Nat. Genet., 34, 267–273. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4110 Nucleic Acids Research, 2019, Vol. 47, No. 8 impairs mitochondrial morphology, optic nerve structure and visual strand break repair proteins XRCC1 and XRCC4. DNA Repair, 3, function. Hum. Mol. Genet., 16, 1307–1318. 1493–1502. 90. Anand,R., Wai,T., Baker,M.J., Kladt,N., Schauss,A.C., Rugarli,E. 96. Date,H., Igarashi,S., Sano,Y., Takahashi,T., Takahashi,T., and Langer,T. (2014) The i-AAA protease YME1L and OMA1 Takano,H., Tsuji,S., Nishizawa,M. and Onodera,O. (2004) The cleave OPA1 to balance mitochondrial fusion and fission. J. Cell FHA domain of aprataxin interacts with the C-terminal region of Biol., 204, 919–929. XRCC1. Biochem. Biophys. Res. Commun., 325, 1279–1285. 91. Song,Z., Chen,H., Fiket,M., Alexander,C. and Chan,D.C. (2007) 97. Horton,J.K., Stefanick,D.F., Caglayan,M., Zhao,M.L., OPA1 processing controls mitochondrial fusion and is regulated by Janoshazi,A.K., Prasad,R., Gassman,N.R. and Wilson,S.H. (2018) mRNA splicing, membrane potential, and Yme1L. J. Cell Biol., 178, XRCC1 phosphorylation affects aprataxin recruitment and DNA 749–755. deadenylation activity. DNA Repair, 64, 26–33. 92. Cipolat,S., Martins de Brito,O., Dal Zilio,B. and Scorrano,L. (2004) 98. Camacho-Pereira,J., Tarrago,M.G., Chini,C.C.S., Nin,V., OPA1 requires mitofusin 1 to promote mitochondrial fusion. Proc. Escande,C., Warner,G.M., Puranik,A.S., Schoon,R.A., Reid,J.M., Natl. Acad. Sci. U.S.A., 101, 15927–15932. Galina,A. et al. (2016) CD38 dictates age-related NAD decline and 93. Harris,J.L., Jakob,B., Taucher-Scholz,G., Dianov,G.L., mitochondrial dysfunction through an SIRT3-dependent Becherel,O.J. and Lavin,M.F. (2009) Aprataxin, poly-ADP ribose mechanism. Cell Metab., 23, 1127–1139. polymerase 1 (PARP-1) and apurinic endonuclease 1 (APE1) 99. Fang,E.F., Lautrup,S., Hou,Y., Demarest,T.G., Croteau,D.L., function together to protect the genome against oxidative damage. Mattson,M.P. and Bohr,V.A. (2017) NAD(+) in aging: molecular Hum. Mol. Genet., 18, 4102–4117. mechanisms and translational implications. Trends Mol. Med., 23, 94. Durocher,D. and Jackson,S.P. (2002) The FHA domain. FEBS 899–916. Lett., 513, 58–66. 100. Sorrentino,V., Menzies,K.J. and Auwerx,J. (2018) Repairing 95. Clements,P.M., Breslin,C., Deeks,E.D., Byrd,P.J., Ju,L., mitochondrial dysfunction in disease. Annu. Rev. Pharmacol. Bieganowski,P., Brenner,C., Moreira,M.C., Taylor,A.M. and Toxicol., 58, 353–389. Caldecott,K.W. (2004) The ataxia-oculomotor apraxia 1 gene product has a role distinct from ATM and interacts with the DNA http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nucleic Acids Research Oxford University Press

Diminished OPA1 expression and impaired mitochondrial morphology and homeostasis in Aprataxin-deficient cells

Loading next page...
 
/lp/oxford-university-press/diminished-opa1-expression-and-impaired-mitochondrial-morphology-and-tXhfmtnKqz
Publisher
Oxford University Press
Copyright
Published by Oxford University Press on behalf of Nucleic Acids Research 2019.
ISSN
0305-1048
eISSN
1362-4962
D.O.I.
10.1093/nar/gkz083
Publisher site
See Article on Publisher Site

Abstract

Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4086–4110 Nucleic Acids Research, 2019, Vol. 47, No. 8 Published online 14 February 2019 doi: 10.1093/nar/gkz083 Diminished OPA1 expression and impaired mitochondrial morphology and homeostasis in Aprataxin-deficient cells 1 2 1,2,* 1,* Jin Zheng , Deborah L. Croteau , Vilhelm A. Bohr and Mansour Akbari 1 2 Center for Healthy Aging, SUND, University of Copenhagen, 2200 Copenhagen N, Denmark and Laboratory of Molecular Gerontology, National Institute on Aging, 251 Bayview Blvd, Baltimore, MD, 21224, USA Received May 31, 2018; Revised January 25, 2019; Editorial Decision January 29, 2019; Accepted January 31, 2019 ABSTRACT - cells, aprataxin (APTX) is the only protein that removes 5 AMP from DNA (2). Human patients carrying mutation Ataxia with oculomotor apraxia type 1 (AOA1) is in APTX develop the progressive neurodegenerative disease an early onset progressive spinocerebellar ataxia ataxia with oculomotor apraxia type 1 (AOA1) (3,4). caused by mutation in aprataxin (APTX). APTX re- APTX localizes to the nucleus and mitochondria (5). Evi- moves 5 -AMP groups from DNA, a product of dence suggests that AOA1 pathology is related to mitochon- abortive ligation during DNA repair and replication. drial dysfunction. Biochemical and cell biological analysis showed that APTX is more critical in mitochondrial DNA APTX deficiency has been suggested to compromise (mtDNA) repair than in the nuclear DNA repair (5–7). mitochondrial function; however, a detailed charac- AOA1 patients display ataxia, neuropathy, cerebellar atro- terization of mitochondrial homeostasis in APTX- phy and coenzyme Q deficiency, traits seen in mitochondrial deficient cells is not available. Here, we show that diseases (8,9). A mitochondrial disease database has been cells lacking APTX undergo mitochondrial stress and developed as a diagnostic tool to identify mitochondrial display significant changes in the expression of the pathology in human diseases (10), which has been proven mitochondrial inner membrane fusion protein op- useful in previous studies (11). This database predicts that tic atrophy type 1, and components of the oxida- AOA1 is a disease with significant mitochondrial involve- tive phosphorylation complexes. At the cellular level, ment (6,10). APTX deficiency impairs mitochondrial morphology Mitochondria are called the powerhouse of the cells be- and network formation, and autophagic removal of cause of their central role in cellular ATP production. Mito- damaged mitochondria by mitophagy. Thus, our re- chondria also play other important biological roles includ- 2+ ing amino acids and lipid metabolism, Ca signaling, cell- sults show that aberrant mitochondrial function is a cycle regulation and apoptosis (12). Muscle and brain tis- key component of AOA1 pathology. This work cor- sues are particularly vulnerable to mitochondrial abnormal- roborates the emerging evidence that impaired mito- ities, probably because of their high ATP consumption and chondrial function is a characteristic of an increas- reliance on other mitochondrial functions. Accordingly, mi- ing number of genetically diverse neurodegenerative tochondrial dysfunction has been identified in a number disorders. of ataxias and other types of neurodegenerative diseases (11,13–16). Mitochondria are structurally highly dynamic organelles INTRODUCTION and their morphology is determined by the type of their host Ligation of DNA ends is the final step in almost all DNA cell. Mitochondria undergo division (fission) and merge to- repair pathways and is a critical step during DNA replica- gether (fusion). The ratio of fusion and fission determines tion. Human cells have three DNA ligases named DNA lig- the formation of the filamentous tubular network or punc- ase I, III and IV, and all use adenosine triphosphate (ATP) tate mitochondria (17). The processes of fusion and fission as a cofactor (1). During ligation, DNA becomes temporar- involve a group of dynamin-like and GTPase proteins. The ily adenylated at the 5 -end (5 -AMP-DNA) by DNA ligase major players in fusion include the outer mitochondrial (1). Occasionally, DNA ligase dissociates from DNA after membrane proteins mitofusion 1 (MFN1) and mitofusin 2 5 -adenylation of DNA resulting in a 5 -AMP group that (MFN2), and the inner mitochondrial membrane protein must be removed for DNA to be ligated later. In human To whom correspondence should be addressed. Tel: +410 558 8162; Fax: +410 558 8157; Email: vbohr@nih.gov Correspondence may also be addressed to Mansour Akbari. Center for Healthy Aging, Faculty of Health Sciences, University of Copenhagen, Building 7.3, Nørre Alle´ 14, 2200 Copenhagen N, Denmark. Tel: +45 35326762; Email: akbari@sund.ku.dk Published by Oxford University Press on behalf of Nucleic Acids Research 2019. This work is written by (a) US Government employee(s) and is in the public domain in the US. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4087 optic atrophy type 1 (OPA1). The key fission proteins are the For whole cell extract (WCE) preparation, pelleted cytosolic dynamin-related protein 1 (DRP1), and several cells were suspended in lysis buffer (20 mM, 4-(2- mitochondrial outer membrane proteins; mitochondrial fis- hydroxyethyl)-1-piperazineethanesulfonicacid ( HEPES)- sion factor (MFF), mitochondrial fission 1 protein (Fis1) KOH, pH 7.5, 200 mM KCl, 10% glycerol, 1% Triton X- and mitochondrial dynamic proteins MiD49, and MiD51 100, 1% non-ionic detergent, IGEPAL CA-630 (octylphe- (18,19). The function, recruitment and assembly of these noxypolyethoxyethanol), 1 mM ethylenedinitrilo tetraaceti- proteins are largely regulated by post-translational modi- cacid (EDTA), 1 mM Dithiothreitol (DTT), EDTA-free fications ( 20). Complete protease inhibitor cocktail (Sigma) and Phospho- Mitochondrial morphology is integral to mitochondrial STOP (Sigma)), and left on ice for 60 min. Cell debris was quality control through a selective autophagic removal of pelleted at 15 000 g for 15 min, and the supernatant ( WCE) dysfunctional mitochondria known as mitophagy (18). The was collected. processes of fusion, fission and mitophagy are collectively known as mitochondrial dynamics. Increasing evidence has Preparation of mitochondria-enriched extracts identified a close interplay between mitochondrial dynam- ics, mitochondrial bioenergetics, cellular metabolism sta- Cells were collected at 500 g, washed once with phos- tus and energy demand (21,22). Adding to the impor- phate buffered saline (PBS) and suspended in isotonic tance of the mitochondrial homeostasis network, recent re- buffer (20 mM HEPES-KOH pH 7.4, 5 mM KCl, 1 search has identified a novel link between persistent nuclear mM DTT, protease inhibitor cocktail) and left on ice to DNA damage, activation of poly ADP-ribose polymerases swell. The cells were broken in a Dounce tight-fit homog- (PARPs) and nicotinamide adenine dinucleotide (NAD ) enizer in ice and equal volume of 2× mannitol-sucrose- consumption and mitochondrial dysfunction (23). The dis- HEPES (MSH) buffer (420 mM mannitol, 140 mM sucrose, ruption of this axis has been identified as a central cause in 20 mM HEPES-KOH pH 7.4, 4 mM EDTA, 2 mM EGTA, many neurodegenerative diseases (14,24). 5 mM DTT) was added to the homogenate and centrifuged Previous studies suggested that APTX deficiency asso- at 1000 g for 5 min (twice). The supernatant was centrifuged ciates with mitochondrial abnormalities including mito- at 10 000 g for 30 min and the pellet containing mitochon- chondrial morphology and network (5–7). However, a de- dria were suspended in lysis buffer (20 mM HEPES-KOH, tailed investigation of the mitochondrial status in APTX- pH 7.5, 200 mM KCl, 10% glycerol, 1% Triton X-100, 1% deficient cells is not available. The aim of this project is to IGEPAL, 1 mM EDTA, 1 mM DTT, EDTA-free Complete elucidate the molecular mechanisms of mitochondrial dys- protease inhibitor cocktail and PhosphoSTOP) and left on function in APTX deficient cells by analyzing key players in ice for 60 min followed with a mild sonication. The lysates mitochondrial maintenance and function in CRISPR me- were centrifuged at 15 000 g and the supernatants were col- −/− diated APTX U2OS cells and in AOA1 patient-derived lected and used as mitochondrial enriched extracts. cells. We found significant changes in key mitochondrial pa- rameters including disruption of mitochondrial morphol- APTX knockdown ogy, network, decreased mitochondrial membrane poten- tial (MMP), increased mitochondria reactive oxygen species APTX-specific TRC shRNA-pLKO vector (clone ID (ROS) and impaired mitophagy response. Our results sug- TRCN0000083642; Sigma) and a negative control scram- gest that mitochondrial dysfunction is a key feature of ble shRNA-pLKO.1 construct (Addgene) were described AOA1 pathology. previously (5). The plasmid (1 g) was co-transfected with the packaging plasmid (pCMV-dr8.2DVPR, Addgene, 0.7 MATERIALS AND METHODS g) and envelope plasmid (pCMV-VSV-G, Addgene, 0.3 g) into human embryonic kidney 293T cells using PolyJet Synthetic oligonucleotides were from TAG Copenhagen. transfection reagent (SignaGen Laboratories). Lentivirus [- P]ATP was from Perkin Elmer. 5 - DNA adenyla- containing media were collected 48 h later and filtered tion kit was from BioNordika (E2610S). MitoTracker Red through a 0.45 M filter to remove the cell debris and CMXRos (M-7512), Mitosox red (M36008) and tetram- used to infect U2OS cells. Puromycin-resistant U2OS cell ethylrhodamine (TMRM) (T-668) were from Thermo colonies were propagated and tested by western blot analy- Fisher Scientific- Life Technology. Saponin was from Sigma sis for aprataxin (ab31841; abcam). (74036). N-acetyl-L-cysteine (NAC) was from Sigma. Cell lines and preparation of whole cell protein extracts CRISPR mediated APTX knockout (WCE) U2OS cells were seeded on a 6-well dish and trans- U2OS cells were cultured in Dulbecco’s modified Ea- fected with aprataxin double Nickase plasmid (sc-417083- gle’s medium (DMEM)-Glutamax (Gibco). C2ABR and NIC, Santa Cruz Biotechnology) using PolyJet transfection C3ABR (APTX proficient) and L938 (P206L /P206L) reagent (SignaGen Laboratories). Transfected cells were se- and L939 (P206L/V263G) (APTX deficient) patient- lected in medium with 2 g/ml puromycin for one week, derived Epstein-Barr virus-transformed lymphoblast cell then harvested and reseeded on 150 mm dishes at 20 cells lines (25) were grown in RPMI medium 1640- Glutamax per dish. Single cell colonies were propagated. Disruption (Gibco). Both DMEM and Roswell Park Memorial Insti- of APTX was verified by polymerase chainreaction (PCR) tute (RPMI) medium1640 were supplemented with 10% Fe- amplification of the flanking target region and western blot tal Bovine Serum (FBS) and 1% penicillin-streptomycin. analysis. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4088 Nucleic Acids Research, 2019, Vol. 47, No. 8 mtDNA integrity analysis seq2, performed as described in Michael et al.(30), adjusted P-value ≤ 0.05 and log2 fold change ≥ 1or ≤−1. Differen- MtDNA integrity was analyzed using a PCR-based method tially expressed genes (DEGs) were subjected to GO func- (26). The PCR reactions were carried out as described previ- tional enrichment analysis using Enrichr (31). Patient and ously (6). PCR products were separated in agarose gel and control cells were treated similarly, however since there were the intensity of the amplicons was measured using Image only two of each, for each gene, the reads were averaged then J. The resulting values were then converted to relative le- log2 fold change was calculated between patient and control sion frequencies per 10 kb DNA by application of the Pois- samples. Genes with a probability of ≥ 0.75 and log2 fold son distribution (lesions/amplicon =−ln (A /A )), where t 0 change ≥ 1or ≤ -1 were considered a DEG. The RNA-seq −/− A represents the amplification of APTX cells and A t 0 data has been deposited to the GEO database. is the amplification of control cells ( 27). Lesions per 10 kb For real-time quantitative-PCR (Q-PCR) analysis, DNA = (−ln (A /A )) × 10 000 [bp]/size of long amplicon t 0 cDNA was prepared using Maxima Reverse Transcriptase [bp]. and Oligo dT (12–18) (Thermo Fisher Scientific- Life tech). Q-PCR was carried out using a real-time PCR kit (Bio-Rad1725271) following the manufacturer’s protocol. RNA-seq and Q-PCR analysis of gene expression For analysis of alternative splicing of OPA1 mRNA, Total RNA was purified from APTX-KO U2OS cells, cDNA was prepared from the cells and used as template APTX-KO cells stably expressing APTX (APTX-Pos), two to PCR amplify exons three to nine using primers; for- AOA1 patient cell lines (L938 and L939) and the cor- ward, F1- 5 -GGATTGTGCCTGACATTGTG-3, and re- responding control cell lines (C2ABR and C3ABR), us- verse, R1-5 -TCTGATACTTCAACTGAGTGTGC. PCR ing RNeasy Mini Kit (Qiagen). There were four biologi- amplification of exons three to seven was carried out cal replicates for the U2OS, APTX-KO and APTX-Pos cell using primers; forward, F2-5 - GTGTGGGAAATTGA lines, and two replicates for the patient control and AOA1 TGAGTATATCG, and reverse R2- 5 -GCACTCTGAT cell lines. The cells were in culture for 2–3 weeks before CTCCAACCAC. The PCR products were separated in RNA extraction. The samples were sent to the Beijing Ge- 8% non-denaturing polyacrylamide gel (1 × Tris-Borate- nomics Institute Genomics Services for analysis. Briefly, the EDTA (TBE) buffer, 37.5:1 acrylamide: bis). Agilent 2100 Bio analyzer (Agilent RNA 6000 Nano Kit) was used for RNA sample quality control, RNA concentra- tion, RIN value, 28S/18S and fragment length distribution FACS analysis for mitochondrial membrane potential and mi- analysis. For library construction, polyA-containing mes- tochondrial ROS senger RNA (mRNA) molecules were isolated with mag- Fluorescence-activated cell sorting (FACS) analysis netic beads. Following purification, the mRNA was frag- was performed at least three times independently with mented into small pieces. The cleaved RNA fragments were CytoFLEX flow cytometry platform (Beckman Coulter). copied into first strand complementary DNA (cDNA) us- All analyses were done in ice-cold sorting buffer (PBS, ing reverse transcriptase and random primers. This is fol- 0.1% bovine serum albumin (BSA)), and each experiment lowed by second strand cDNA synthesis using DNA Poly- included ∼100 000 events per group. MMP and ROS levels merase I and RNase H. These cDNA fragments then have were evaluated with TMRM and Mitosox Red, respectively. the addition of a single ‘A’ base and subsequent ligation of TMRM (40 nM) or Mitosox Red (5 uM) were added to the the adapter. The products are then purified and enriched cells, and incubated for 15 min in 5% CO2, 37 C. U2OS with PCR amplification. We then quantified the PCR yield cells were trypsinized and collected. After washing, cells by Qubit and pooled samples together to make a single were suspended in ice-cold sorting buffer for sorting. AOA1 stranded DNA circle (ssDNA circle), which gave the final cells were spun down to remove free dyes, suspended in library. DNA nanoballs (DNBs) were generated with the ss- sorting buffer and subjected to FACS. Both TMRM and DNA circle by rolling circle replication to enlarge the fluo- Mitosox Red stained cells were analyzed at 561 nm lasers rescent signals at the sequencing process. The DNBs were with 586 nm emission filters. All FACS data were analyzed loaded into the patterned nanoarrays and pair-end reads using CytExpert Software (Beckman Coulter). of 100 bp were read through on the BGISEQ-500 plat- form for the following data analysis study. For this step, the BGISEQ-500 platform combines the DNA nanoball- Live cell confocal microscopy based nanoarrays and stepwise sequencing using Combina- tional Probe-Anchor Synthesis Sequencing Method. Low- Cells were seeded on 30 mm diameter coverslips (0787, quality reads were filtered out (more than 20% of the bases Thermo Scientific) placed in 6 cm culture dishes with an qualities are lower than 10), reads with adaptors and reads optimal density (200 000 cells per dish). After 24 h, the cells with unknown bases (N bases more than 5%) to get the were stained with different reagents (100 nM Mitotracker clean reads (SOAPnuke v1.5.2). Then the clean reads were red, 40 nM TMRM or Mitosox) for 15 min. After wash- mapped onto reference genome, followed by novel gene pre- ing, coverslips were mounted on chambers for live imaging. diction, SNP & INDEL calling and gene-splicing detec- Microscope images were acquired using LSM780 confocal tion. Clean reads were mapped to reference using Bowtie2 system with a 63× oil immersion objective. Quantification v2.2.5 (28), then gene expression levels were calculated with of mitochondria parameters was carried out in Image J soft- RSEM v1.2.12 (29). Finally, DEGs were detected using DE- ware as described previously (32). Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4089 Mitophagy assay phosphate buffer (pH 7.2). The samples were rinsed three times in 0.15 M phosphate buffer (pH 7.2) and post-fixed Mitophagy in live cells was monitored by Mitophagy de- in 1% w/v OsO4 in 0.12 M sodium phosphate buffer (pH tection kit (Dojindo Molecular Technologies). Urolithin A 7.2) for 2 h. The samples were dehydrated in graded series (UA, 5 M) was used to trigger mitophagy as described of ethanol, transferred to propylene oxide and embedded previously (33). The level of mitophagy was defined by the in Epon according to standard procedures. Following poly- area of Mtphagy dye per cell. At least 50 cells were qualified merization, the Thermanox coverslip was peeled off. Sec- in each group. The level of colocalization of Mtphagy dye tions, ∼60 nm thick, were cut with a Leica UC7 microtome and lysosome dye was also analyzed. Quantification analy- (Leica Microsystems, Wienna, Austria) and collected on sis was carried out using Image J. copper grids with Formvar supporting membranes, stained with uranyl acetate and lead citrate and subsequently ex- Oxygen consumption rate (OCR) amined with a Philips CM 100 Transmission EM (Philips, Eindhoven, The Netherlands), operated at an accelerat- Oxygen consumption was measured using the Seahorse ing voltage of 80 kV and equipped. Digital images were XF24 instrument, according to the manufacturer’s instruc- recorded with an OSIS Veleta digital slow scan 2k × 2k tion (Seahorse Biosciences, North Billerica, MA). Cells CCD camera and the ITEM software package. Quantifica- were seeded into a Seahorse tissue culture plate at a density tion of cristae length was done using Image J. Twenty cells of 50 000 cells per well in DMEM with 1 mM sodium pyru- were quantified in each group. vate, 2 mM glutamine and 10% FBS. After 24 h, the medium was replaced with un-buffered XF assay medium (Seahorse Biosciences), pH 7.4, supplemented with 25 mM glucose, 1 ATP measurement mM sodium pyruvate and 2 mM glutamine. One hour later, Cells were plated in 6-well dishes at 6 × 10 cells per dish. the oxygen consumption rate (OCR) was measured in the Intracellular ATP was measured using a luciferase-based Seahorse XF24 analyzer in four blocks of three 3-min pe- assay (ATPlite Luminescence Assay Kit, PerkinElmer) fol- riods. The first block measured the basal respiration rate. lowing manufacturer’s guidelines. A standard curve was Next, 1 M oligomycin was added to inhibit complex 5 and generated and used to calculate samples ATP concentra- the second block was measured. Then, 0.3 M carbonyl tion. Protein concentration was determined using Bradford cyanide 4-trifluoromethoxy-phenylhydrazone (FCCP) was protein assay reagents (Bio-Rad). The content of ATP was added to uncouple respiration, and the third block was mea- normalized for protein content and presented as percentage sured. Finally, 2 M antimycin A was added to inhibit com- of control. plex 3, and the last measurements were acquired. The cells were counted after the experiment and the results were nor- −/− malized to cell number in each well. Three biological exper- Expression of APTX in the APTX U2OS cells iments were done with technical replicates of 2–5 for each Total RNA was purified from U2OS cells using RNeasy genotype. Kit (Qiagen). cDNA was prepared using SuperScript III RT (Invitrogen) and used as a template to PCR amplify- Immunocytochemistry ing the APTX isoform variant 6 containing a putative mi- tochondrial localization signal (5) (Supplementary Figure Cells were seeded at 20 000 cells per well on 12 mm diameter S1). Primers used in PCR were; forward 5 - AACTAGAT coverslips in 24-well dishes. After 24 h, the cells were fixed in CTATGAGTAACGTGAATTTGTCCGTCTCC, and re- 4% paraformaldehyde in PBS for 10 min at room tempera- verse 5 - AATC GGATCCTCACTGTGTCCAGTGCTT ture and washed three times with PBS. Cells were quenched CCTGAG. The PCR products were digested with BglII in 50 mM NH Cl in PBS for 10 min and permeabilized and BamHI restriction enzymes and cloned into pAcGFP1- and blocked at the same time in blocking buffer (3% bovine −/− Hyg-N1 vector (Clontech). APTX cells were transfected serum albumin (BSA), 0.1% saponin, PBS, pH 7.4) for 60 with the construct using PolyJet transfection reagent (Sig- min. Primary antibodies (diluted in blocking buffer) were naGen Laboratories) and stable APTX-GFP expressing added and incubated at 4 C overnight. The samples were cells were selected in medium containing hygromycine and washed three times in 0.1% saponin in PBS and incubated enriched in FACS. with Alexa Fluor-conjugated secondary antibodies diluted in blocking buffer for 1 h at room temperature protected from light. For nuclei standing, 5 g/ml 4,6-Diamidino- DNA substrates 2-phenylindoledihydrochloride (DAPI) solution (D1306, Duplex 22-mer DNA containing a 5 -AMP was prepared as Thermo Fischer Scientific) was applied to the samples for follows: oligonucleotide 5 -GATCCTCTAGAGTCGACC 5 min at room temperature following incubation with sec- TGCA-3 was first end-labeled at the 5 -end by [- P]ATP ondary antibody. Cells were washed three times with PBS and T4 polynucleotide kinase, and then adenylated to gen- and mounted on SuperFrost glass slides (3302775, Thermo erate an oligonucleotide with 5 -AMP- p-DNA. Or 3 -end Scientific) with mounting medium (S3023, Dako). fluorophore TAMRA labeled oligo was 5 -end phosphory- lated followed with 5 -adenylation of DNA to generate 5 - Transmission electron microscopy (TEM) analysis AMP-DNA––3 -TAMRA oligo. The adenylated oligo was Cells were grown to confluence on Thermanox coverslips annealed to a complementary oligo at 1:1.25 ratio in 20 mM andfixedwith2%v /v glutaraldehyde in 0.05 M sodium HEPES-KOH, pH 7.5 and 100 mM NaCl, heated at 90 C Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4090 Nucleic Acids Research, 2019, Vol. 47, No. 8 for 5 min and slowly cooled down to room temperature to OPA50 bacteria. Synchronous worm cultures were obtained prepare 5 -AMP-containing duplex DNA. by allowing gravid adults lay eggs onto NGM/OP50 plates for 2–4 h. Eggs were then collected and seeded onto NGM agar plates. 5 -AMP-DNA repair assay For mitochondrial network imaging, Day 1 (D1) worms Repair reactions were carried out using 15 gWCEsin40 were incubated for 24 h at 20 C on NGM OP50 plates mM HEPES-KOH, pH 7.8, 1 mM DTT, 0.36 mg/ml BSA, containing 1 M tetramethylrhodamine methyl ester 100 mM KCl, 2 mM EDTA and 5 pmol DNA substrate at (TMRM). Worms were subsequently transferred to fresh 30 C for the indicated times. The reactions were stopped NGM plates containing only OP50 for 1 h to clear their by adding loading buffer (10 mM EDTA, 90% formamide, intestinal tract of residual dye, after which they were trans- xylene cyanol, bromphenol blue) and heating the samples ferred to cover slides and paralyzed. Photographs were im- at 85 C for 5 min. DNA was separated in 20% denaturing mediately taken with LSM780 (Zeiss) confocal microscope. acrylamide gel (19:1 acrylamide:bis, 1 × TBE and 7.5 M For western blot analysis, D1 worms were collected and urea). lysed in RIPA buffer (Life technologies) with complete pro- teases cocktail (Roche). Worms were disrupted by sonica- tion using a Branson sonifier (Thomas Scientific) with the Western blot analysis following settings: 10% amplitude; 10 s on; 30 s off, total on WCEs were separated in Tris-glycine sodium dodecyl time 5 min. Samples were kept on ice during sonication to sulphate (SDS) gels and transferred onto Polyvinyli- avoid excessive heating. Afterward, the debris was cleared denedifluoride (PVDF) membrane. Each experiment by centrifugation at 12 000 × g for 10 min at 4 C. The su- was done three to vfi e times. The images are shown as pernatant was resolved in laemmli sample buffer and used three technical replicates of one biological replicate. The for western blot (WB). primary antibodies used were: APTX (sc-374108), Fis1 ATP was measured with ATPlite Luminescence Assay (sc-376469), proliferating cell nuclear antigen (PCNA) (Perkin Elmer). D1 worms were collected and washed three (sc-56), MFN1 (sc-166644), MFN2 (sc-515647) from Santa times with M9 buffer. Worms were pelted and resuspended Cruz. OPA1 (67589S), superoxide dismutase 2 (SOD2) in cell lysis solution from the ATP assay kit. Assays were (13194S), AMP-activated protein kinase (AMPK) (2535S), performed according to the manufacturer’s protocol. Phospho-AMPK (Thr172), DRP1 (8570S), Phospho- Chemotaxis to isoamyl alcohol was performed at 20 C, DRP1 (Ser616) (3455S), anti-acetyl lysine (9681S) and on 9 cm agar plates as described earlier (34). The chemo- Sirtuin (SIRT)1 (8469) were from CellSignaling. TOMM20 taxis index was calculated by subtracting the number of an- (WH0009804M1), and actin (A5441) were from Sigma- imals found at the trap from the number of animals at the Aldrich. SIRT3 (10099-1-AP), and PARP1 (LS-B3432) source of the chemical, divided by the total number of an- were from Nordic Biosite. VDAC1 (ab14734) and APTX imals entered into the assay (34). The resulting values were (ab31841) were from Abcam, PGC-1 (NBP1-04676) expressed and graphed as percentiles. About 200 adult ani- and light chain 3 (LC3) (NB600-1384) were from Novus mals for each strain were assayed in each experiment. Biologicals, TFAM (H00007019-B01P) was from Abnova. For swimming movement, D1 worms were randomly se- To detect oxidative phosphorylation (OXPHOS) com- lected and transferred to a 6 cm petri dish containing 1 ml plex assembly, we used an assembly-dependent total OX- of M9 buffer. The worms were allowed to acclimate for ∼10 PHOS rodent antibody cocktail (ab110413, Abcam). The s, and then movements were scored for 1 min. Thirty worms were scored in each group. antibodies in the cocktail are against a subunit that is la- bile when its complex is not assembled. The samples were prepared following the manufacturer’s protocol. Statistical analysis Error bars represent SE or range (for experiments with Colony forming assay less than three replicates) as indicated in the figure legends. Data were processed in Excel and statistical analyses were Cells were seeded on 6-well dishes at 200 cells per well. Next performed using GraphPad Prism 7 (GraphPad Software). day, the cells were treated with 0.8 mM methyl methane- Statistical analysis of differences between two groups was sulfonate (MMS) or 0.020 mM menadione in serum-free performed using a two-tailed, unpaired t-test and between medium for 60 min. The drugs were removed, and cells more than two groups using a one-way ANOVA analysis of were gently washed once with PBS and normal medium variance test followed by a Tukey’s post-hoc comparison; was added to each well. Cells were left to recover for 10 two-way ANOVA was used for comparison between con- days. Cells were stained in 0.5% crystal violet, dissolved in trol and AOA1 patient cells; *P < 0.05; **P < 0.01; ***P 20% ethanol for 1 h and washed in water, and colonies were < 0.001; ****P < 0.0001. counted. RESULTS Caenorhabditis elegans strains and methods −/− CRISPR-mediated APTX U2OS cells display hallmarks Caenorhabditis elegans eat-3 (ad426) and N2 control strains of mitochondrial dysfunction were obtained from Caenorhabditis Genetics Center (CGC, University of Minnesota). Strains were maintained at 20 C. We elected U2OS cells for this work for several reasons. Nematode Growth medium (NGM) agar plates seeded with U2OS cells have active and functional mitochondria and Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4091 −/− rely on OXPHOS for ATP production (35). They grow in APTX cells, however, contained more punctate and less monolayer, are amenable to transfection and have an over- connected mitochondria, which mostly clustered in per- −/− −/− all structure that makes them particularly suitable for mito- inuclear regions (Figure 3A, APTX ). APTX cells chondrial morphology and network analysis (35). We used had more mitochondria but they were smaller and shorter a lentiviral delivery system and knocked down (KD) APTX than in control cells (Figure 3A andgraphs). Immunocyto- in U2OS cells. Western blot analysis showed 90–95% re- chemistry analysis of fixed cells using an antibody target- duction in APTX concentration in the APTX-KD cells ing the translocase of the outer mitochondrial membrane, compared with control cells (Figure 1A, KD 20 days), in TOMM20, also showed punctate staining of mitochondria −/− line with a previous report (5). Surprisingly, extracts from in APTX cells (Figure 3B). Furthermore, transmission APTX-KD cells displayed a robust 5 -AMP removal activ- electron microscopy (TEM) confirmed the confocal mi- ity (Figure 1B), indicating that even at a very low concen- croscopy results showing more abundant, but smaller and −/− tration, APTX efficiently repairs 5 -adenylated DNA. This fragmented mitochondria in APTX cells (Figure 3C). demonstrates that in studies where APTX is not completely Notably, altered mitochondrial morphology did not signif- −/− depleted, it may mask biological consequences of APTX de- icantly change mitochondrial content in APTX cells, as ficiency. Thus, we completely depleted APTX in the cells us- judged by western blot analysis of the outer mitochondrial ing CRISPR technology (APTX-KO). Western blot analy- membrane proteins VDAC1 and TOMM20 (Figure 3D). sis of protein extracts from APTX-KO and lymphoblastoid Collectively, these results show that APTX deficiency may cells derived from AOA1 patients completely lacked APTX alter mitochondrial morphology without significant effect (Figure 1C). DNA repair analysis of the extracts showed ab- on the overall mitochondrial content. sence of 5 -AMP-DNA removal activity in APTX-KO and The mitochondrial inner membrane folds into the ma- patient cell extracts indicating that APTX is the sole enzyme trix and forms distinct structures called cristae. Emerging for removal of 5 -adenylated DNA in human cells, and also data suggests that the morphology and shape of cristae confirms the specificity of our DNA substrate (Figure 1D). modulate the organization and function of OXPHOS com- The MMP is generated by OXPHOS through the elec- plexes and thus directly influences cell metabolism ( 39). tron transport chain (ETC). MMP measurement has fre- Electron microscopy (EM) data showed that control cells quently been used to evaluate mitochondrial health and contained long and fine cristae stretched out from the mem- −/− function. Mitochondrial ROS production is highly regu- brane. APTX cells, however, contained shorter and re- lated by MMP and mitochondrial dysfunction is often duced cristae density (Figure 3C and the graph), indicat- linked to increased ROS production (36). APTX-knock out ing a marked alteration in cristae shape and density in −/− −/− cells (APTX ) showed significantly lower MMP (Figure APTX cells. 2A and C confocal microscopy image), and higher mito- chondrial ROS production (Figure 2B and C) compared −/− with the control cells. Superoxide dismutases (SODs) are Diminished OPA1 expression in APTX cells cellular antioxidants that catalyze the conversion of super- Mitochondria are highly dynamic organelles. Size and mor- oxide anions to oxygen and hydrogen peroxide. Mitochon- phology of mitochondria are determined by the rate of mi- drial SOD2 is a primary defense enzyme against mitochon- tochondrial fission and fusion ( 40). Emerging findings are drial superoxide. The SOD2 level was markedly higher in −/− unraveling an intricate connection between mitochondrial APTX cells compared with control cells (Figure 2D), morphology and network organization and a number of key probably in response to increased mitochondrial ROS pro- cellular processes including the clearance of dysfunctional duction. Mitochondrial SIRT3 is a NAD -dependent ly- mitochondria by mitophagy, mtDNA maintenance and in- sine deacetylase that regulates the function of mitochon- −/− tegrity, mitochondrial stress signaling, and cellular energy drial proteins (37), including SOD2 (38). APTX cells demand and metabolism (22,39–45). displayed a significantly higher level of SIRT3 compared Western blot analysis of some commonly accepted key with the control cells (Figure 2D). Thus, impaired MMP regulators of mitochondrial fission and fusion showed that and enhanced ROS production associate with APTX defi- the concentration of OPA1, a central regulator of mitochon- ciency. drial inner membrane fusion and a key protein in mitochon- drial dynamics and cristae structure formation (18,46,47), Impaired mitochondrial network and cristae structure in and also of MFN1, another regulator of mitochondrial fu- −/− −/− APTX cells sion, were considerably lower in APTX cells (Figure 4A and the graphs). Phosphorylation of DRP1 is thought Previously, knockdown of APTX in human neuroblastoma to regulate the translocation of DRP1 from cytosol to the SH-SY5Y cells was suggested to disrupt the mitochondrial outer mitochondrial membrane, a key step in the initiation network (5), which may be related to the changes in the level −/− of fission. The concentration of DRP1 but not the phospho- of mitochondrial fission and fusion proteins in APTX −/− rylated DRP1 (p-DRP1) was lower in APTX cells (Fig- cells observed here. However, the molecular mechanism of ure 4A). These results show that APTX deficiency may af- mitochondrial network impairment and the possibility of fect the expression and the stability of some mitochondrial mitochondrial fission and fusion links have not been investi- fusion and fission proteins in U2OS cells. Q-PCR analy- gated. Thus, we investigated the status of the mitochondrial sis showed significantly lower OPA1 mRNA expression in network and morphology in these cells. Live cell confocal −/− APTX cells (Figure 4B). microscopy analysis showed elongated and highly branched tubular mitochondria in control cells (Figure 3A, Control). Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4092 Nucleic Acids Research, 2019, Vol. 47, No. 8 A B Incubation time 5ʼ-AMP-22-mer APTX 22-mer actin Control APTX-KD C D APTX 5´-AMP-22-mer 22-mer actin Figure 1. Knockdown and knockout of APTX in U2OS cells. (A) Western blot analysis to determine APTX knockdown (KD) efficiency in U2OS cells using a lentiviral-delivered shRNA system. (B) Repair analysis of 5 -AMP DNA in extracts from APTX KD cells and control cells. (C) Western blot analysis showing complete absence of APTX in CRISPR-mediated APTX knockout (KO) U2OS cells (lane 2) and in cells prepared from AOA1 patients. (D) APTX deficient cells are devoid of 5 -AMP removal activity confirming the specificity of our DNA substrate for APTX activity. −/− suggesting a robust mitophagy response in control cells. The Defective mitophagy response in APTX cells −/− intensity of the dye signal was considerably less in APTX Autophagy is a process whereby intracellular components cells than control cells following UA treatment (Figure 5B, are engulfed with membrane bound autophagic vesicles −/− APTX + UA, and the graphs). Collectively, these results (autophagosome), then fuse with lysosomes, and the con- suggest that the ability to induce mitophagy by UA is com- tents of the cargo is degraded (48). Western blot analysis of −/− promised in APTX cells. microtubule-associated protein 1 LC3 is frequently used as a marker to measure autophagic flux. LC3-I is conjugated −/− to phosphatidylethanolamine to form LC3-II, which is lo- APTX cells show elevated PARylation but unchanged calized to autophagosomes (49). There were no detectable NAD content differences in the levels of LC3-I, LC3-II between the con- −/− DNA damage response (DDR) proteins poly (ADP-ribose) trol and APTX cells (Figure 5A, control + vehicle and −/− polymerases 1 and 2 (PARP1/2) signal the location of DNA APTX + vehicle, respectively), suggesting comparable damage on the genome to DNA repair proteins by adding basal autophagic activity in these cells. The selective au- ADP-ribosepolymers (PAR) to themselves and to nearby tophagic elimination of damaged mitochondria is called proteins consuming NAD in the process (50). PARP1 is re- mitophagy (48). Urolithin A (UA) is a metabolite of nat- sponsible for ∼90–95% of the total cellular PARP activity ural compounds known as ellagitannins (33), and it was −/− (51). The overall level of PAR was higher in APTX cells recently shown to induce mitophagy (33). Following UA (Figure 5C and the graph). Surprisingly, the level of PARP1 treatment, the level of LC3-II in control cells, but not in −/− −/− was markedly lower in APTX cells (Figure 5C and the APTX cells increased significantly. This suggests robust graph). The elevated level of PARylation, however, did not autophagosome formation in response to UA treatment in −/− seem to be associated with detectable effect on the cellu- control cells but not in APTX cells. It could also reflect lar NAD content (Supplementary Figure S2), suggesting slower autophagosome-lysosome fusion rate, or impaired that the NAD -dependent reactions were not significantly lysosomal degradation of the cargo in control cells com- −/− −/− affected in APTX cells. pared with APTX cells (Figure 5A). To clarify these al- ternatives and to specifically measure mitophagy, we used a commercially available dye kit. The dye (Mtphagy), ac- −/− Re-introduction of APTX into the APTX U2OS cells re- cumulates in mitochondria following the induction of mi- verses some phenotypes back to the APTX positive control tophagy, the damaged mitochondria then fuse to lysosomes cells resulting in a higher fluorescence signal. Pre-treatment of the cells with UA induced a strong mitophagy signal in con- To further confirm the role of APTX in the observed trol cells that colocalized with the lysosomal signal (Figure changes in mitochondrial parameters, we tested whether re- −/− 5B, control + UA, merged, yellow spots and the graphs), introduction of APTX into APTX cells could rescue Control Control APTX-KO KD (5 d) Control C2ABR C3ABR KD (10 d) Control L938 KD (20 d) L939 No extract No extract Control 2.5 min 5 min APTX-KO C2ABR 10 min C3ABR 15 min L938 No extract 2.5 min L939 5 min 10 min 15 min Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4093 Figure 2. Knockout of APTX causes mitochondrial dysfunction in U2OS cells. (A) FACS analysis and quantifications of MMP using TMRM. Bars display TMRM fluorescence intensity as mean ± S.E.M., N = 3(* P < 0.05, *** P < 0.001). (B) FACS analysis and quantifications of mitochondrial ROS levels using MitoSOX, Bars display MitoSOX fluorescence intensity as mean ± S.E.M., N = 3(*P < 0.05). (C) Confocal images showing TMRM and −/- MitoSOX staining in live APTX and control U2OS cells. The scale bar represents 10 m. (D) A representative western blot image of SOD2 and SIRT3 in the indicated WCEs. Graphs on the right show quantifications of SOD2 and SIRT3 protein levels normalized to actin. Values are mean ± S.E.M., N = 3(**P < 0.01 and ***P < 0.001). key features that were significantly altered by APTX de- uct of alternative transcription or translation of the cloned pletion. APTX cDNA was prepared from U2OS cells and APTX. The level of APTX-GFP was ∼5.5-times higher used as template to construct an APTX-GFP expressing than the level of APTX in control cells. −/− −/− vector. APTX cells were transfected with the construct SOD2 was highly abundant in APTX cells (Figures and stable APTX expressing cells were selected and prop- 2Dand 6B). The expression of APTX restored it to the level agated (hereafter, referred to APTX-positive). By RNAseq of control cells (Figure 6B, APTX-Pos). Mitochondria are analysis, APTX is expressed about eight times more than the major site of ATP production and the level of cellu- in control cells. Approximately 10% of the cells showed a lar ATP may reflect the status of mitochondrial function. −/− mitochondrial APTX-GFP signal as estimated by confocal ATP levels were significantly lower in APTX cells (Fig- microscopy inspection of the cells (Supplementary Figure ure 6C). Expression of APTX restored ATP to the level of −/− S3). Expression of APTX restored 5 -AMP removal activity control cells (Figure 6C). The lower ATP levels in APTX −/− in APTX cells, indicating that the cells were expressing cells may be caused by a higher rate of ATP consumption, a catalytically functional protein (Figure 6A). or low MMP (Figure 2A) that can result in suboptimal mi- Western blot analysis of APTX-positive extracts identi- tochondrial ATP production. fied two major bands and a few weaker bands that migrated The serine/threonine AMPK complex, is a key sensor of faster in the gel (Figure 6B). These may be truncated APTX- cellular energy status by ADP- and AMP:ATP ratio and GFP protein, or the product of alternative transcription or possibly also of glucose availability (52). In general, ac- translation of the APTX-GFP construct. The upper band tivated AMPK phosphorylated at Thr172 (pAMPK) in- corresponds to the long isoform of APTX-GFP. The lower creases catabolic processes while decreasing anabolic re- bands may represent truncated APTX-GFP or the prod- actions through the phosphorylation of key proteins in Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4094 Nucleic Acids Research, 2019, Vol. 47, No. 8 Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4095 −/− Figure 3. Knockout of APTX alters mitochondrial morphology and network. (A) Representative images of live APTX and control cells pre-incubated with MitoTracker Red showing mitochondrial network formation; scale bar represents 10 m. Graphs on the right show quantifications of mitochondrial morphology (number, length, size and shape) from more than 200 cells, values are mean ± S.E.M. (**P < 0.01, ****P < 0.0001). (B) Immunocytochemistry analysis of fixed cells against the outer mitochondrial membrane protein TOMM20. Nuclei are visualized by DAPI staining; scale bar represents 10 m. (C) TEM images showing detailed mitochondrial morphology and cristae structures; scale bar represents 1 m. The graph shows the quantification analysis of cristae length using Image J. Twenty cells were quantified in each group (*** P < 0.001). (D) Western blot analysis of mitochondrial outer membrane −/− proteins VDAC1 and TOMM20 to determine relative mitochondrial abundance in APTX and control cells. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4096 Nucleic Acids Research, 2019, Vol. 47, No. 8 -/- Control APTX OPA1 OPA1 actin MFN1 actin MFN2 actin DRP1 actin 75 pDRP1 17 Fis1 actin B mRNA Figure 4. APTX depletion decreases OPA1 protein and mRNA levels in U2OS cells. (A) A representative western blot image for key mitochondrial morphol- ogy proteins. Graphs on the right show quantifications of different protein levels normalized to actin. Both OPA1 bands were included in the quantifica tion analysis. Values are mean ± S.E.M., N = 3(*P < 0.05, **P < 0.01,****P < 0.0001). (B) Q-PCR analysis of OPA1 gene expression. OPA1 gene expression was normalized to the expression of the house-keeping gene actin. Bars show relative OPA1 mRNA level as mean ± S.E.M., N = 3 (****P < 0.0001). several pathways including mTOR, glycolysis and mito- cient AMPK response to increased AMP:ATP ratio (Sup- chondrial homeostasis. Activated AMPK promotes mi- plementary Figure S4). This may partially explain low basal −/− tophagy through phosphorylation of ULK1 (53). The level level of activated AMPK in APTX cells (Figure 6D) de- of pAMPK but not of total AMPK was considerably lower spite lower cellular ATP content (Figure 6C). −/− in APTX cells. Expression of APTX increased the level The OXPHOS system consists of vfi e multipeptide com- −/− of pAMPK in APTX cells nearly to the level of con- plexes (CI-CV) composed of over 90 different structural trol cells that was statistically significant (Figure 6Dand proteins, which are encoded by nuclear and mtDNA. They the graphs). Low level of activated AMPK may in part ac- require assembly protein factors for proper assembly and −/− count for the delayed mitophagy response in APTX cells function. The stability of the complexes is interdependent (Figure 5B). 5-aminoimidazole-4-carboxamide ribonucleo- (55,56). Diminished MMP and increased mitochondrial side (AICAR) becomes converted in the cells to an AMP ROS production may be caused by OXPHOS system dys- analog and has been used as an AMPK activating com- functions and changes in the protein levels of OXPHOS −/− pound (54). APTX cells showed ∼40% lower level of subunits may reflect altered ETC function. We examined the activated AMPK (p-AMPK) compared to control cells fol- abundance of selected OXPHOS subunits using an assem- lowing AICAR treatment suggesting a somewhat less effi- bly specific antibody cocktail. We found that the mitochon- Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4097 -/- -/- Control + Veh. Control + UA APTX + Veh. APTX + UA LC3 I LC3 II actin -/- -/- Control + Veh. Control + UA APTX + Veh. APTX + UA -/- Control APTX PAR PARP1 actin Figure 5. Mitophagy induction is compromised in APTX deficient cells. ( A) Western blot analysis of autophagy marker LC3 in different cells. The ratio of LC3II (lower band) to actin was calculated to demonstrate autophagic response to urolithin (UA). Data are shown as mean ± S.E.M., N = 3, (***P < 0.001). (B) Assessment of mitophagy in UA treated cells. The merged images show colocalization between the mitophagy dye and the lysosome dye. The upper graph on the right shows the quantification analysis of colocalization of lysosome and mitopahgy dyes. The lower graph shows the quantification o f mitophagy signal. Data is presented as mean ± S.E.M., (****P < 0.0001). scale bar represents 10 m. (C) Western blot analysis of PARylation (PAR) and PARP1 levels in different WCEs as indicated. Quantifications are from three independent experiments. Data are shown as mean ± S.E.M., (**P < 0.01, ****P < 0.0001). Merge Lysosome Mtphagy Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4098 Nucleic Acids Research, 2019, Vol. 47, No. 8 -/- Control APTX APTX-Pos APTX-GFP 5ʼ-AMP-22-mer APTX-Endo 22-mer SOD2 actin pAMPK AMPK -/- APTX-Pos C Control APTX 63 pAMPK AMPK actin -/- Control APTX APTX-Pos CV CIII CIV CII 17 CI actin No extract Control -/- APTX APTX-Pos Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4099 H I −/− Figure 6. Expression of human APTX variant 6 containing an N-terminal putative mitochondrial localization signal in APTX cells. (A) Repair analysis of 5 -AMP DNA in the indicated WCEs. Lower band corresponds to repaired DNA substrate. (B) WB analysis of SOD2 and APTX in the indicated cell lines. (C) Measurement of ATP level in the indicated cells. Values are mean ± S.E.M., N = 3, (***P < 0.001). (D) A representative western blot image of AMPK and activated AMPK (pAMPK). Quantifications are shown as mean ± S.E.M., N = 3, (**P < 0.01). (E) Western blot image of the subunits −/− of OXPHOS complexes I-V. Quantifications are shown as mean ± S.E.M., N = 3, (***P < 0.001). (F) Seahorse measurements of OCR in APTX , APTX-Pos and control cells. Statistical significant was calculated using paired- t-test, N = 3(*P < 0.05). (G) PARylation (PAR) analysis by western blot. Quantifications are shown as mean ± S.E.M., N = 3, (***P < 0.001). (H) PCR-based mtDNA damage analysis. Quantifications are shown as mean ± S.E.M., N = 3, (**P < 0.01). (I) Colony forming analysis to determine the ability of the cells to recover following treatment with the genotoxic drugs MMS and menadione. Quantifications are shown as mean ± S.E.M., N = 3, (***P < 0.001). −/− drial OXPHOS complexes I, II, III and IV were more abun- 6H). Expression of APTX in APTX cells corrected it −/− dant in APTX cells compared with control cells (Figure back to the level of control cells (Figure 6H). These results 6E). Expression of APTX corrected them back to the level support a role for APTX in mtDNA repair and integrity, in of control cells (Figure 6E and the graphs). line with and extension of previous reports (5–7). The effects of APTX deficiency on mitochondrial respi- The number of mtDNA molecules, also known as ration was tested by measuring OCR, using the Seahorse mtDNA copy number, varies significantly depending on −/− XF24 analyzer. APTX and APTX-Pos cells showed the cell type and is thought to correlate with the cellular lower ATP-linked respiration and higher leak relative to need for energy and ATP production. Previously, lentiviral- control cells (Figure 6F and the graph). In contrast, APTX- mediated knockdown of APTX resulted in reduced mtDNA Pos cells showed a higher reserve capacity (Figure 6Fand copy number (5). However, we did not find any differences −/− the graph). Together, these measurements show that the in mtDNA copy number between APTX and AOA1 presence or absence of APTX can alter mitochondrial respi- patient cells and their corresponding control cells (Sup- ration which together with the low MMP (Figure 2A), and plementary Figure S5). These differences might be related −/− high ROS (Figure 2B) in APTX cells suggests mitochon- to the cell type differences used in these studies; here, we drial dysfunction. used U2OS cells and AOA1 patient-derived lymphoblast −/− The level of PARylation was higher in APTX cell ex- cell lines, whereas in the previous study (5), SH-SY5Y and tracts compared with control cells (Figures 6Gand 5C, primary human skeletal muscle cells were used. PAR). APTX expression corrected it back to the level of The colony forming assay is used for monitoring the abil- control cells (Figure 6G and the graph). ity of cells to recover from genotoxic treatment. Next, we MtDNA in APTX-deficient cells is more susceptible determined the recovery capacity of the cells following treat- to damage than nuclear DNA (5,6). Using a PCR-base ment with the genotoxic compounds MMS and menadione. method, we detected more damage in mtDNA from MMS modifies guanine to 7-methylguanine, and adenine to −/− APTX cells than in mtDNA from control cells (Figure 3-methyladenine, which can result in base mispairing and Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4100 Nucleic Acids Research, 2019, Vol. 47, No. 8 Figure 7. AOA1 patient cell lines also display mitochondrial dysfunctions. (A) FACS analysis and quantifications of MMP using TMRM. Bars display TMRM fluorescence intensity as mean ± S.E.M., N = 3(*P < 0.05). (B) FACS analysis and quantifications of mitochondrial ROS level using MitoSOX. Bars display MitoSOX fluorescence intensity as mean ± S.E.M., N = 3(*P < 0.05, **P < 0.01). (C) A representative western blot image of SOD2 and SIRT3 in the indicated WCEs. Graphs show the quantifications of APTX, SOD2, SIRT3, VDAC1 and PGC1  protein levels normalized to PCNA. Values are mean ± S.E.M., N = 3(*P < 0.05, **P < 0.01). (D) Western blot analysis of protein acetylation in mitochondria enriched extracts from control and AOA1 cells. (E) Western blot analysis of PARylation (Par) and PARP1 levels in different WCEs as indicated, quantifications are from three independent experiments. Data are shown as mean ± S.E.M., (**P < 0.01, ***P < 0.001). (F) Western blot analysis of the indicated cells treated with 1 mM NAC for 48 h for PAR. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4101 Figure 8. OPA1 protein and mRNA levels are diminished in AOA1 cells. (A) A representative western blot image of mitochondrial fusion and fission proteins in extracts from the APTX-deficient patient-derived cells L938 and L939 and the control cells C2ABR and C3ABR. Quantifications are shown as mean ± S.E.M., N = 3, (*P < 0.05, **P < 0.01, ***P < 0.001). (B) Q-PCR analysis of OPA1 gene expression in the patients and control cells. Quantifications are shown as mean ± S.E.M., N = 3, (*P < 0.05, **P < 0.01). (C) Western blot analysis of the subunits of OXPHOS, using PCNA as loading control. Quantifications are shown as mean ± S.E.M., N = 3, (**P < 0.01). Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4102 Nucleic Acids Research, 2019, Vol. 47, No. 8 Figure 9. RNA Seq analysis. After comparison of the shared GO terms from GO cellular compartment, three terms were of interest GO:0034599 cellular response to oxidative stress, GO:0022904 respiratory ETC and GO:0008053 mitochondrial fusion. The genes in those three terms are shown in (A–C), respectively. (D) Panel shows genes from the cellular response to oxidative stress (GO:0034599) and mitochondrial fusion (GO:0008053) from AOA1 patient cells versus controls. No mitochondrial ETC genes were found in the AOA1 patient-derived lymphoblastoid cell lines. (E) Genes from the DNA repair pathways base excision, mismatch and NER were also collected and displayed. Note: a few genes were obtained from various lists, were expressed wildly differently and were removed. From the stress response gene list, the gene DHRS2 (KO Ctrl -11.1; Pos Ctrl -9, Pos KO 2.2), from the BER list, APTX (KO Ctrl -0.52; Pos Ctrl 2.6; Pos KO 3.1) and from the NER gene list, the genes GTF2H2C (KO Ctrl 0.97; Pos Ctrl -6.9; Pos KO -7.1) were each removed from the heat maps. All heat maps were centered on zero. Heat maps were made using GraphPad Prism 7. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4103 replication blocks, respectively (57). Menadione exposure 5C) cells showed elevated PAR that was inversely correlated results in the accumulation of ROS in mitochondria that at- with PARP1 concentration. tack and damage macromolecules including mtDNA (58). ROS can chemically modify DNA bases and cause DNA −/− Menadione and MMS treatments showed that APTX stranded breaks (61). To test whether elevated PAR signal- cells were more sensitive to these agents than control cells ing in AOA1 cells was a result of ROS generated DNA dam- (Figure 6I). APTX re-expression reduced the sensitivity of age, we treated the cells with the ROS scavenger NAC. West- −/− APTX cells to these genotoxic agents back to the level ern blot analysis showed a significant reduction in PAR sig- of control cells (Figure 6I). naling (Figure 7F). Collectively, these results suggest that To summarize, putting back APTX into APTX-KO cells loss of APTX resulted in a higher ROS production that to- corrected the protein levels of SOD2, pAMPK, OXPHOS gether with reduced DNA repair capacity results, at least in complexes I, II, III and IV back to the level of control part, in elevated DNA damage level and DDR in the form cells. It also corrected cellular ATP levels, DDR via paryla- of PARylation. tion signal, mtDNA integrity and sensitivity to MMS and menadione treatment back to the level of control cells, and improved mitochondrial respiration reserve capacity, alto- Diminished overall OPA1 level is a common feature of APTX gether suggesting the role of APTX in the observed changes deficient cells −/− in APTX cells. However, the expression of OPA1, SIRT3 Western blot analysis of key regulators of mitochondrial fis- and PARP1 were not corrected by re-introducing APTX −/− sion and fusion showed that the level of the mitochondrial into the APTX cells (Figure 9; Supplementary Figure inner membrane fusion protein OPA1, and DRP1, a cytoso- S6 and Supplementary Tables S1–S3), even though, the lev- −/− lic protein that is recruited to the outer membrane surface els of these proteins were consistently altered in APTX to catalyze the fission, were significantly reduced in AOA1 cells and in AOA1 patient-derived cells (Figures 2D, 4, 5C, cells (Figure 8A). We found no detectable changes in the 7Cand E, 8Aand B, 9; Supplementary Tables S1–S4). amounts of Fis1, an outer mitochondrial membrane protein that mediates fission by acting as DRP1 receptor ( 20,62), AOA1 patient-derived cells recapitulate the abnormal mito- or the mitochondrial fusion protein MFN1, in AOA1 cells −/− chondrial features seen in APTX U2OS cells compared to control cells (Figure 8A). Phosphorylation Next, we used AOA1 patient-derived cells (L938 and L939) of DRP-1 at Serine 616 is required for the recruitment of and control cells (C3ABR and C2ABR) that are Epstein– DRP1 to the surface of mitochondria (63). The ratio of p- Barr virus-transformed lymphoblastoid cell lines (5,25). DRP1 to total DRP1 was unchanged (Figure 8A). The re- MMP analysis showed a significant reduction in MMP duced level of DRP1 in patient cells may be a protective re- (Figure 7A), and a significantly higher level of mitochon- sponse to OPA1 reduction in order to maintain a balance drial ROS production (Figure 7B), and higher SOD2 (Fig- between fusion and fission ( 64). −/− ure 7C) in AOA1 cells compared with the control cells. To- Like in APTX cells (Figure 4B), the expression of −/− gether with the results from the APTX cells (Figure 2A, OPA1 was significantly reduced in AOA1 cells compared B and D), impaired MMP and increased oxidative stress ap- with the control cells (Figure 8B). Thus, APTX-deficiency pear to associate strongly with APTX deficiency. results in down-regulation of OPA1 gene expression. To test whether loss of OPA1 is a characteristic of APTX-deficient −/− cells, we knocked out APTX in HEK cells. APTX HEK Increased acetylation of mitochondrial proteins in AOA1 cells cells also showed a significantly lower OPA1 level compared The key function of SIRT3 deacetylase in mitochondrial with control HEK cells (Supplementary Figure S7). Collec- homeostasis is probably the repair of non-enzymatic acety- tively, these results demonstrate that diminished OPA1 ex- lation of mitochondrial proteins by acetyl-CoA (59). Like in pression and protein level is a common feature of APTX- −/− APTX cells (Figure 2D), the level of SIRT3 was consid- deficient cells. erably higher in AOA1 patient cells compared with the con- OPA1 mRNA undergoes alternative splicing (65)and trol cells (Figure 7C). Western blot analysis of mitochondria some isoforms seem to be specifically important for mito- enriched extracts showed an overall increased level of mito- chondrial network formation (66). To test the possible pres- chondrial protein acetylation in AOA1 cells (Figure 7D). ence of OPA1 isoforms specific to APTX-deficient cells, we Together, these results show that AOA1 patient cells un- carried out reverse transcription-PCR amplification of the dergo mitochondrial stress in the form of higher ROS pro- region flanking exons three to nine (Supplementary Figure duction and acetylation of mitochondrial proteins. S8A). Two putative mRNA isoforms were differently ex- pressed between AOA1 and control cells (primers F1/R1, Significantly reduced level of PARP1 but enhanced PARyla- Supplementary Figure S8B, arrows). These were further tion in AOA1 cells verified by a second primer set to amplify a shorter region between exons three to seven (primers F2/R2). Further- The PARP superfamily consists of 16 members (60). PARP1 more, better separation of OPA1 protein in 6% polyacry- is responsible for most of the cellular PARP activity (51). lamide SDS gel revealed that of the vfi e apparent isoforms, The level of total poly(AD-ribose) (PAR) was markedly three were differentially expressed between APTX-deficient higher in AOA1 cells (Figure 7E), suggesting elevated DDR and control cells (Supplementary Figure S8C, arrows). In signaling. The level of PARP1, however, was significantly conclusion, both the overall level of OPA1 expression and reduced in AOA1 cells compared with control cells (Fig- −/− ure 7E). Together, both AOA1 and APTX U2OS (Figure Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4104 Nucleic Acids Research, 2019, Vol. 47, No. 8 the expression of some putative OPA1 isoforms are altered the ETC genes, there is one clear block of genes (*, Figure in APTX-deficient cells compared to control cells. 9D) whose expression seems to be up in the APTX KO and normalized by APTX rescue. Likewise, the mitochondrial Specific up-regulation of OXPHOS complexes III and V in FIS1 and MFF genes are up-regulated in APTX KO but AOA1 cells normalized after APTX re-expression. In contrast, MFN1, OPA1 and BAX genes were down-regulated in the KO cells We examined the abundance of selected OXPHOS subunits and remained that way after APTX rescue and mitochon- in AOA1 cells and found that complexes CIII and V were −/− drial network changes in APTX cells (Figure 9C). We more abundant in AOA1 cells than in control cells (Fig- should note here that the expression change of OPA1 was ure 8C). Taken together, the results show that APTX defi- very modest, −0.45 log2 fold change in the APTX KO Ctrl ciency consistently leads to the up-regulation of OXPHOS comparison, so OPA1 does not appear to be strongly regu- protein subunits probably as a part of a compensatory tran- lated at the level of transcription in these cells. scriptional response to mitochondrial dysfunction (67,68). We conducted a similar analysis on the AOA1 patient Such response seems to be tissue specific ( 69), which may cells versus controls; roughly half were up- or down- explain why complexes I and II were not affected in AOA1 regulated (1180 genes, 57% down) (Supplementary Table −/− cells but were up-regulated in APTX cells (Figure 6E). S4). From the AOA1 set, the genes from the term mitochon- Moreover, it is known that the type and the level of mito- drion were extracted and the expression of OPA1 (down), chondrial proteins can vary between tissues (70). We spec- LONP (down) and FOXO1 (up) were changed similar as in ulate that the differences in OXPHOS between AOA1 and the APTX KO cells (Figure 9D). There was only one mi- APTX-KO cells may be a result of differential expression of tochondrial electron transport gene in the AOA1 set, AT- OXPHOS subunits and assembly factors in these cells. PAF1, the assembly factor for the F1 complex, which was −/− Collectively, AOA1 and APTX share key features of down-regulated. mitochondrial dysfunctions supporting the involvement of Since APTX is involved in DNA base excision repair APTX in these phenotypes in particular, and in mitochon- (BER), we also sought to determine if DNA repair genes drial homeostasis as a whole. were responsive to either the loss or rescue of APTX expres- sion. BER, nucleotide excision repair (NER) and mismatch Differential gene expression analysis repair (MMR) terms were found in KEGG pathways, and We performed RNA-seq analyses on APTX-KO, and interestingly, NER was statistically significant (adj. P-value APTX-KO cells stably expressing APTX variant 6 (APTX- ≤ 0.05) in all three comparisons (when using the set of genes Pos), and two AOA1 patient-derived cell lines (L939, and that have adjusted P-values ≤ 0.05 and dropping the log2 L939) and the corresponding control cells (Ctrl) (Supple- fold change requirement). MMR was significantly changed mentary Tables S1–S4). There were similar numbers of in APTX-Pos versus APTX-KO comparison (adj. P-value DEGs, APTX versus Ctrl, 1532 (57% down); APTX-Pos 0.0005), while BER was only trending (adj. P-value 0.06). A versus Ctrl, 2590 (59.2% up); APTX-Pos versus APTX-KO, heat map of the combined DNA repair genes list is shown 1384 (72.8% up). For the ATPX-KO cell lines, the signifi- in Figure 9E. Notably, PARP1 is down-regulated in all pair- cant differentially expressed genes were subjected to GO cel- wise comparisons, while the expression of PARP2 is up- lular compartment analysis via Enrichr (31) to uncover po- regulated in APTX KO and normalized after APTX rescue. tential mitochondrial terms. However, the term mitochon- drion (GO 0005739) was not among the top terms for any The Caenorhabditis elegans model of OPA-1 deficiency reca- pairwise comparison. However, since we were interested in pitulates the stress phonotypes in APTX-deficient cells defining the mitochondrial terms that were changed after APTX KO and APTX add back, we proceeded and ac- To date, there is no known C. elegans homolog of APTX quired the DEG gene lists from the mitochondrion GO (72). Therefore, to isolate the functional consequences of term and wanted to subject that list to GO biological func- OPA1 loss on an organismal level, the C. elegans eat-3 tion analysis but each pairwise comparison had too few strain was examined. Caenorhabditis elegans eat-3 (ad426) genes (APTX KO Ctrl, 42; APTX POS Ctrl, 107, APTX strain (73) has a mutation in the D2013.5 gene, which en- POS Ctrl) for this to be meaningful. So, we collected all codes the ortholog of yeast Mgm1 and mammalian Opa1 the genes with an adjusted P-values ≤ 0.05, dropped the (74). The ad426 mutation leads to fragmented mitochon- log2 fold change cutoff requirement, and collected the mito- dria similar to those cause by mutations in Opa1 and Mgm1 chondrion term (GO 0005739) gene lists from each pairwise (74). As previously reported, the eat-3 (ad426) strain show comparison. Not imposing a cutoff change is routinely done disrupted mitochondrial network by TMRM live staining in Gene Set Enrichment Analysis (71). Mitochondrion was (Figure 10A). Similar to AOA1 patient cells, the level of a top scoring term in each pairwise comparison using this PARylation and acetylated proteins is also increased in eat- input gene list (Supplementary Table S5). After GO biolog- 3 (ad426) worms (Figure 10B and C). We also measured ical process analysis, the GO terms to cellular response to ATP levels and found that eat-3 (ad426) worms have a much oxidative stress (Figure 9A), respiratory ETC (Figure 9B), lower level of ATP compared with N2 worms (Figure 10D), mitochondrial fusion (Figure 9C), and several terms related suggesting mitochondrial function is impaired. To further to translation were of interest to us (see Supplementary Ta- explore the biological importance of the integrity of mi- ble S6 for the entire top 20 list of terms). SOD2 was among tochondrial network, we performed chemotaxis assay by the up-regulated genes in ATPX KO cells that was down- using isoamyl alcohol, which is an attractive odor to the regulated after adding back APTX (Figure 9A). Among worms. In N2 worms, ∼75% were attached by isoamyl al- Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4105 Figure 10. The Caenorhabditis elegans Opa1 homolog eat-3 recapitulates some of the stress phenotypes and responses seen in APTX-defect cells. (A) Visualization of mitochondrial network by treating the worms with TMRM. Scale bar represents 10 m. (B) Western blot analysis of C. elegans extracts for (B) parylation (Par) and (C) total protein acetylation, (D) ATP measurements, (E) chemotaxis assay and (F) swimming assays. Data is presented as mean ± S.E.M., (** P < 0.01, *** P < 0.001). Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4106 Nucleic Acids Research, 2019, Vol. 47, No. 8 cohol after 1 h. In eat-3 (ad426) worms, only 22% were at- with significant variations of age of onset and severity of tracted by isoamyl alcohol (Figure 10E). To exclude the pos- the disease (81). The extra-ocular symptoms of OPA1 de- sibility that the difference in chemotaxis index is caused by ficiency include deafness, ataxia, myopathy and spinocere- motor dysfunction, we performed a swimming assay. The bellar degeneration (82–84). Thus, AOA1 and OPA1-related result showed no significant differences between N2 and eat- neurodegeneration share phenotypes suggesting that mito- 3 (ad426) worms (Figure 10F). Thus, decreased chemotaxis chondrial dysfunction is a common pathway in these dis- index, together with normal motor function, suggest that eases. Moreover, a mitochondrial disease database (10)pre- eat-3 (ad426) worms have deficiencies in sensing isoamyl dicts AOA1 and DOA diseases to have significant mito- alcohol. In worms, two pairs of amphid sensory neurons, chondrial involvement (Supplementary Figure S9). Fibrob- AWC and AWA, are required for chemotaxis to isoamyl al- lasts from patients with OPA1 mutation display mitochon- cohol (75). Taken together, we conclude that in nematode, drial network fragmentation (85,86). Homozygous Opa1 eat-3 (Opa1) deficiency leads to the disruption of mitochon- mutant mice die in utero during embryogenesis, but het- drial network, and impairment of sensory neurons. erozygous Opa1 mutants display the main features of hu- man DOA including abnormal mitochondrial morphology, disorganized cristae structure, mitochondrial dysfunction DISCUSSION and mtDNA instability (87–89). Eat-3 is the C. elegans ho- molog of human OPA1. Mitochondria in eat-3 mutant C. In this study, we show that APTX-deficient cells displayed hallmarks of mitochondrial dysfunction and stress, i.e. low elegans strain are fragmented and show phenotypes consis- MMP, increased mitochondrial ROS production, suscepti- tent with defects in OXPHOS system (74). Thus, ample ev- bility to mtDNA damage, elevated acetylation of mitochon- idence shows a key role of OPA1 in mitochondrial function drial proteins, changed OXPHOS protein abundance, al- and network formation across species. As such, OPA1 re- tered mitochondrial morphology and impaired mitophagy. duction may, at least in part, explain the observed loss of A significant reduction in the level of a protein does not mitochondrial network (Figure 3A and B), and impaired necessarily recapitulate the phenotypes caused by a com- cristae in APTX-deficient cells (Figure 3C). plete loss of that protein. Germline deletion of DNA repair In humans, OPA1 is present in multiple isoforms gen- protein Xrcc1 in mouse is embryonic lethal; however, Xrcc1 erated by alternative splicing of mRNA at exons 4, 4b at levels nearly 10% of normal cells support embryonic vi- and 5b, which further undergo proteolysis by the mito- ability (76). We also found that extracts from cells with chondrial inner-membrane peptidases YME1L and OMA1 APTX as low as 5% of the normal cells displayed robust (65,90,91). PCR amplification of the OPA1 cDNA, and 5 -AMP removal activity compared with the control cells WB analysis showed specific changes in the expression of (Figure 1A and B). This result may partially explain the di- OPA1 isoforms between AOA1 and control cells (Supple- verse clinical symptoms and complex genotype-phenotype mentary Figure S8). Furthermore, the alternative splicing correlations in AOA1 patients given different rates of degra- analysis of OPA1 in RNA-seq data showed the preferen- dation and instability of APTX mutants (9,77,78). This also tial expression of isoforms 1, 7 and 8 in U2OS and AOA1- demonstrates that it is of great importance to achieve a patient cells (Supplementary Figure S10). Interestingly, iso- complete depletion in order to characterize the functions of form 7 was down-regulated in both AOA1-patient cells −/− aprataxin. Such a strategy might be necessary in the study andinAPTX cells. The OPA1 isoform 1 was, in addi- of most DNA repair proteins. tion, down-regulated in AOA1-patient cells. Thus, APTX- Mitochondria are structurally highly dynamic organelles. deficiency is linked to altered expression of OPA1 isoforms, which may contribute to the mitochondrial morphology They constantly change shape, size, and form intercon- and network changes in APTX-deficient cells. To determine nected networks in response to environmental cues. Mito- chondrial size and morphology are determined by the rate the biological significance of these differences is technically of fission and fusion, which together with mitophagy and challenging, because both loss and overexpression of OPA1 biogenesis, regulate the characteristic mitochondrial net- causes fragmentation of the mitochondrial network and dis- work. Mitochondrial morphology varies across cell types organization of cristae (66,92), indicating that balanced ex- and tissues and is highly adaptive to metabolic signaling pression of OPA1 is important for proper mitochondrial and stress (22,40). Defects or changes in the expression of function and network formation. components of mitochondrial fission and fusion cause dis- One limitation of this study is that re-introducing APTX −/− ease including neurodegeneration and are associated with into APTX cells did not correct the expression of OPA1 −/− the normal aging process (18,24). The expression and pro- and mitochondrial network in APTX cells. Moreover, tein level of the inner membrane fusion protein OPA1 was the specific function of APTX in the nucleus and mitochon- consistently lower in APTX-deficient cells (Figures 4Aand dria in the observed mitochondrial phenotypes was not ad- B. 8Aand B, 9 Supplementary Tables S1–S3). OPA1 is a dressed here. dynamin-like GTPase localized in the mitochondrial inner It has been shown that APTX regulates the expression membrane and is commonly thought to play key roles in mi- and the stability of PARP1 (93). RNA-seq results showed tochondrial structure and dynamics by mediating inner mi- that the expression of PARP1 was somewhat lower in AOA1 −/− tochondrial membrane fusion and by controlling the cristae patient cells and in APTX cells compared to control cells shape (46,47). OPA1 mutations were first identified in dom- (Supplementary Tables S1–S4 and Figure 9). PARP2 is an- inant optic atrophy (DOA), a disease specifically affecting other PARP family member with PAR synthesis activity retinal ganglion cells (79,80). OPA1 deficiency has also been and is expressed at ∼15% of PARP1 (60). The level of the −/− identified in patients with clinically diverse symptoms and expression of PARP2 was somewhat higher in APTX Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4107 cells compared to control cells (Figure 9 and Supplemen- SUPPLEMENTARY DATA tary Table S1–S3). In AOA1 patient cells, however, there Supplementary Data are available at NAR Online. were no significant changes in PARP2 expression compared −/− with control cells (Supplementary Table S4). In APTX ACKNOWLEDGEMENTS cells, PARP2 also showed increased expression. Several of the PARP members were differentially expressed in APTX- We like to thank Jane Tian for bioenergetic experiments deficient cells (Figure 9 and Supplementary Tables S1–S4). and for help with the microarray analysis. We would like to Poly-(ADP-ribose)glycohydrolyase (PARG) is responsible thank Dr Elin Lehmann, Dr Yongqing Zhang and Dr Kevin for the degradation of poly-(ADP-ribose). There was no G. Becker of the Gene Expression and Genomics Unit, NIA significant difference in the level of expression of PARG in Intramural Program, NIH. We would like to thank Kavya −/− AOA1 or APTX cells compared to their corresponding Achanta for her help with C. elegans study. We would like control cells (Supplementary Tables S1–S3) suggesting that to thank Dr Daniel R. McNeill for C2/3ABR and L938/9 the rate of PAR turnover did not account for the observed cell lines. We would like to thank Dr Beimeng Yang and Dr elevated PAR levels in APTX-deficient cells. Taken together, Anthony Moore for critically reading the manuscript. Con- defect in APTX results in an elevated level of PARylation. focal and electron microscopy experiments were carried out The fork-head associated (FHA) domain is a phospho- at the Core Facility of Integrated Microscopy (CFIM), Uni- peptide interacting domain associated with proteins in- versity of Copenhagen. volved in a number of processes including intracellular sig- nal transduction, transcription, protein transport, DNA FUNDING repair and protein degradation (94). APTX seems to in- teract with several proteins through its FHA domain. In NORDEA Foundation, Denmark (02-2013-0220); EU the nucleus, APTX was reported to interact with DNA re- Joint Programme-Neurodegenerative Disease Research pair proteins XRCC1, PARP1 and also with the transcrip- (JPND); Innovation Fund Denmark (5188-00001); Olav tion factor p53 (25,95–97). A recently developed concept Thon foundation Norway (531811-710131); Novo Nordisk connects nuclear DNA damage signaling to mitochondrial foundation Denmark (NNF17OC0027812); Intramural Re- function and maintenance. According to this model, persis- search Program of the NIH, National Institute on Ag- tent DNA damage, e.g. because of a defect in DNA repair ing (AG000733). Funding for open access charge: Univer- or diminished DNA repair capacity, activates prolonged sity of Copenhagen. DDR by PARP1/2 in the form of PAR being added to Conflict of interest statement. None declared. proteins in the vicinity of the DNA damage. This process consumes NAD , an essential cofactor involved in a num- REFERENCES ber of key processes including mitochondrial bioenergetics and mitophagy. Thus, persistent DDR leads to a competi- 1. Ellenberger,T. and Tomkinson,A.E. (2008) Eukaryotic DNA ligases: tion between these processes for the available NAD lead- structural and functional insights. Annu. Rev. Biochem., 77, 313–338. 2. Ahel,I., Rass,U., El-Khamisy,S.F., Katyal,S., Clements,P.M., ing to mitochondrial dysfunction (23). Moreover, the cel- + McKinnon,P.J., Caldecott,K.W. and West,S.C. (2006) The lular bioavailability of NAD declines during normal ag- neurodegenerative disease protein aprataxin resolves abortive DNA ing (98,99). In support of this model, NAD replenishment ligation intermediates. Nature, 443, 713–716. improves health and lifespan in animal models of human 3. Date,H., Onodera,O., Tanaka,H., Iwabuchi,K., Uekawa,K., Igarashi,S., Koike,R., Hiroi,T., Yuasa,T., Awaya,Y. et al. (2001) disease (11). APTX-deficient cells consistently contained re- Early-onset ataxia with ocular motor apraxia and hypoalbuminemia duced levels of PARP1 but elevated PARylation compared is caused by mutations in a new HIT superfamily gene. Nat. Genet., with control cells (Figures 5C, 6Gand 7E and F). Treat- 29, 184–188. ment of the cells with the ROS scavenger NAC significantly 4. Moreira,M.C., Barbot,C., Tachi,N., Kozuka,N., Uchida,E., reduced PARylation in APTX-deficient cells (Figure 7F). Gibson,T., Mendonca,P., Costa,M., Barros,J., Yanagisawa,T. et al. (2001) The gene mutated in ataxia-ocular apraxia 1 encodes the new Thus, we speculate that the lack of APTX leads to higher HIT/Zn-finger protein aprataxin. Nat. Genet., 29, 189–193. ROS production and reduced PARP1 level, which in turn re- 5. Sykora,P., Croteau,D.L., Bohr,V.A. and Wilson,D.M. 3rd. (2011) sults in diminished basal DDR and DNA repair, persistence Aprataxin localizes to mitochondria and preserves mitochondrial of PARylation, which over time may contribute to the im- function. Proc. Natl. Acad. Sci. U.S.A., 108, 7437–7442. 6. Akbari,M., Sykora,P. and Bohr,V.A. (2015) Slow mitochondrial pairment of mitochondrial homeostasis in APTX-deficient repair of 5 -AMP renders mtDNA susceptible to damage in APTX cells. deficient cells. Sci. Rep., 5, 12876. In conclusion, our data links impaired mitochondrial 7. Caglayan,M., Prasad,R., Krasich,R., Longley,M.J., Kadoda,K., homeostasis to APTX deficiency and AOA1 pathology. In Tsuda,M., Sasanuma,H., Takeda,S., Tano,K., Copeland,W.C. et al. recent years, mitochondria have become targets for phar- (2017) Complementation of aprataxin deficiency by base excision repair enzymes in mitochondrial extracts. Nucleic Acids Res., 45, macological interventions for treatment of neurodegenera- 10079–10088. tive diseases with promising results (23,100). In the future, 8. Quinzii,C.M., Kattah,A.G., Naini,A., Akman,H.O., Mootha,V.K., perhaps such strategies should be tested on AOA1 patients. DiMauro,S. and Hirano,M. (2005) Coenzyme Q deficiency and cerebellar ataxia associated with an aprataxin mutation. Neurology, 64, 539–541. 9. Renaud,M., Moreira,M.C., Ben Monga,B., Rodriguez,D., Debs,R., DATA AVAILABILITY Charles,P., Chaouch,M., Ferrat,F., Laurencin,C., Vercueil,L. et al. (2018) Clinical, biomarker, and molecular delineations and The RNA-seq data has been deposited to the GEO database Genotype-Phenotype correlations of ataxia with oculomotor under series number GSE124412. apraxia type 1. JAMA Neurol., 75, 495–502. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4108 Nucleic Acids Research, 2019, Vol. 47, No. 8 10. Scheibye-Knudsen,M., Scheibye-Alsing,K., Canugovi,C., 33. Ryu,D., Mouchiroud,L., Andreux,P.A., Katsyuba,E., Moullan,N., Croteau,D.L. and Bohr,V.A. (2013) A novel diagnostic tool reveals Nicolet-Dit-Felix,A.A., Williams,E.G., Jha,P., Lo Sasso,G., mitochondrial pathology in human diseases and aging. Aging, 5, Huzard,D. et al. (2016) Urolithin A induces mitophagy and 192–208. prolongs lifespan in C. elegans and increases muscle function in 11. Fang,E.F., Kassahun,H., Croteau,D.L., Scheibye-Knudsen,M., rodents. Nat. Med., 22, 879–888. Marosi,K., Lu,H., Shamanna,R.A., Kalyanasundaram,S., 34. Bargmann,C.I. and Horvitz,H.R. (1991) Chemosensory neurons Bollineni,R.C., Wilson,M.A. et al. (2016) NAD(+) replenishment with overlapping functions direct chemotaxis to multiple chemicals improves lifespan and healthspan in ataxia telangiectasia models via in C. elegans. Neuron., 7, 729–742. mitophagy and DNA repair. Cell Metab., 24, 566–581. 35. Akbari,M., Keijzers,G., Maynard,S., Scheibye-Knudsen,M., 12. Brand,M.D. and Nicholls,D.G. (2011) Assessing mitochondrial Desler,C., Hickson,I.D. and Bohr,V.A. (2014) Overexpression of dysfunction in cells. Biochem. J., 435, 297–312. DNA ligase III in mitochondria protects cells against oxidative 13. Cornelius,N., Wardman,J.H., Hargreaves,I.P., Neergheen,V., stress and improves mitochondrial DNA base excision repair. DNA Bie,A.S., Tumer,Z., Nielsen,J.E. and Nielsen,T.T. (2017) Evidence of Repair, 16, 44–53. oxidative stress and mitochondrial dysfunction in spinocerebellar 36. Starkov,A.A. and Fiskum,G. (2003) Regulation of brain ataxia type 2 (SCA2) patient fibroblasts: effect of coenzyme Q10 mitochondrial H2O2 production by membrane potential and supplementation on these parameters. Mitochondrion, 34, 103–114. NAD(P)H redox state. J. Neurochem., 86, 1101–1107. 14. Fang,E.F., Scheibye-Knudsen,M., Brace,L.E., Kassahun,H., 37. Lombard,D.B., Alt,F.W., Cheng,H.L., Bunkenborg,J., Streeper,R.S., SenGupta,T., Nilsen,H., Mitchell,J.R., Croteau,D.L. and Bohr,V.A. Mostoslavsky,R., Kim,J., Yancopoulos,G., Valenzuela,D., (2014) Defective mitophagy in XPA via PARP-1 hyperactivation and Murphy,A. et al. (2007) Mammalian Sir2 homolog SIRT3 regulates NAD(+)/SIRT1 reduction. Cell, 157, 882–896. global mitochondrial lysine acetylation. Mol. Cell. Biol., 27, 15. Bose,A. and Beal,M.F. (2016) Mitochondrial dysfunction in 8807–8814. Parkinson’s disease. J. Neurochem., 139, 216–231. 38. Qiu,X., Brown,K., Hirschey,M.D., Verdin,E. and Chen,D. (2010) 16. Costa,V. and Scorrano,L. (2012) Shaping the role of mitochondria in Calorie restriction reduces oxidative stress by SIRT3-mediated the pathogenesis of Huntington’s disease. EMBO J., 31, 1853–1864. SOD2 activation. Cell Metab., 12, 662–667. 17. Scheibye-Knudsen,M., Fang,E.F., Croteau,D.L., Wilson,D.M. 3rd 39. Cogliati,S., Frezza,C., Soriano,M.E., Varanita,T., and Bohr,V.A. (2015) Protecting the mitochondrial powerhouse. Quintana-Cabrera,R., Corrado,M., Cipolat,S., Costa,V., Trends Cell Biol., 25, 158–170. Casarin,A., Gomes,L.C. et al. (2013) Mitochondrial cristae shape 18. Sebastian,D., Palacin,M. and Zorzano,A. (2017) Mitochondrial determines respiratory chain supercomplexes assembly and dynamics: coupling mitochondrial fitness with healthy aging. Trends respiratory efficiency. Cell, 155, 160–171. Mol. Med., 23, 201–215. 40. Westermann,B. (2012) Bioenergetic role of mitochondrial fusion and 19. Loson,O.C., Song,Z., Chen,H. and Chan,D.C. (2013) Fis1, Mff, fission. Biochim. Biophys. Acta, 1817, 1833–1838. MiD49, and MiD51 mediate Drp1 recruitment in mitochondrial 41. Walczak,J., Partyka,M., Duszynski,J. and Szczepanowska,J. (2017) fission. Mol. Biol. Cell, 24, 659–667. Implications of mitochondrial network organization in 20. Otera,H., Wang,C., Cleland,M.M., Setoguchi,K., Yokota,S., mitochondrial stress signalling in NARP cybrid and Rho0 cells. Sci. Youle,R.J. and Mihara,K. (2010) Mff is an essential factor for Rep., 7, 14864. mitochondrial recruitment of Drp1 during mitochondrial fission in 42. Al-Mehdi,A.B., Pastukh,V.M., Swiger,B.M., Reed,D.J., Patel,M.R., mammalian cells. J. Cell Biol., 191, 1141–1158. Bardwell,G.C., Pastukh,V.V., Alexeyev,M.F. and Gillespie,M.N. 21. Liesa,M. and Shirihai,O.S. (2013) Mitochondrial dynamics in the (2012) Perinuclear mitochondrial clustering creates an oxidant-rich regulation of nutrient utilization and energy expenditure. Cell nuclear domain required for hypoxia-induced transcription. Sci. Metab., 17, 491–506. Signal., 5, ra47. 22. Wai,T. and Langer,T. (2016) Mitochondrial dynamics and metabolic 43. Merkwirth,C., Dargazanli,S., Tatsuta,T., Geimer,S., Lower,B., regulation. Trends Endocrinol. Metab., 27, 105–117. Wunderlich,F.T., von Kleist-Retzow,J.C., Waisman,A., 23. Fang,E.F., Scheibye-Knudsen,M., Chua,K.F., Mattson,M.P., Westermann,B. and Langer,T. (2008) Prohibitins control cell Croteau,D.L. and Bohr,V.A. (2016) Nuclear DNA damage proliferation and apoptosis by regulating OPA1-dependent cristae signalling to mitochondria in ageing. Nat. Rev. Mol. Cell Biol., 17, morphogenesis in mitochondria. Genes Dev., 22, 476–488. 308–321. 44. Yamaguchi,R., Lartigue,L., Perkins,G., Scott,R.T., Dixit,A., 24. Archer,S.L. (2013) Mitochondrial dynamics–mitochondrial fission Kushnareva,Y., Kuwana,T., Ellisman,M.H. and Newmeyer,D.D. and fusion in human diseases. N. Engl.J.Med., 369, 2236–2251. (2008) Opa1-mediated cristae opening is Bax/Bak and BH3 25. Gueven,N., Becherel,O.J., Kijas,A.W., Chen,P., Howe,O., dependent, required for apoptosis, and independent of Bak Rudolph,J.H., Gatti,R., Date,H., Onodera,O., Taucher-Scholz,G. oligomerization. Mol. Cell, 31, 557–569. et al. (2004) Aprataxin, a novel protein that protects against 45. Elachouri,G., Vidoni,S., Zanna,C., Pattyn,A., Boukhaddaoui,H., genotoxic stress. Hum. Mol. Genet., 13, 1081–1093. Gaget,K., Yu-Wai-Man,P., Gasparre,G., Sarzi,E., Delettre,C. et al. 26. Santos,J.H., Mandavilli,B.S. and Van Houten,B. (2002) Measuring (2011) OPA1 links human mitochondrial genome maintenance to oxidative mtDNA damage and repair using quantitative PCR. mtDNA replication and distribution. Genome Res., 21, 12–20. Methods Mol. Biol., 197, 159–176. 46. Frezza,C., Cipolat,S., Martins de Brito,O., Micaroni,M., 27. Hunter,S.E., Jung,D., Di Giulio,R.T. and Meyer,J.N. (2010) The Beznoussenko,G.V., Rudka,T., Bartoli,D., Polishuck,R.S., QPCR assay for analysis of mitochondrial DNA damage, repair, Danial,N.N., De Strooper,B. et al. (2006) OPA1 controls apoptotic and relative copy number. Methods, 51, 444–451. cristae remodeling independently from mitochondrial fusion. Cell, 28. Langmead,B. and Salzberg,S.L. (2012) Fast gapped-read alignment 126, 177–189. with Bowtie 2. Nat. Methods, 9, 357–359. 47. MacVicar,T. and Langer,T. (2016) OPA1 processing in cell death 29. Li,B. and Dewey,C.N. (2011) RSEM: accurate transcript and disease - the long and short of it. J. Cell Sci., 129, 2297–2306. quantification from RNA-Seq data with or without a reference 48. Pickles,S., Vigie,P. and Youle,R.J. (2018) Mitophagy and quality genome. BMC Bioinformatics, 12, 323. control mechanisms in mitochondrial maintenance. Curr. Biol., 28, 30. Love,M.I., Huber,W. and Anders,S. (2014) Moderated estimation of R170–R185. fold change and dispersion for RNA-seq data with DESeq2. Genome 49. Kabeya,Y., Mizushima,N., Ueno,T., Yamamoto,A., Kirisako,T., Biol., 15, 550. Noda,T., Kominami,E., Ohsumi,Y. and Yoshimori,T. (2000) LC3, a 31. Kuleshov,M.V., Jones,M.R., Rouillard,A.D., Fernandez,N.F., mammalian homologue of yeast Apg8p, is localized in Duan,Q., Wang,Z., Koplev,S., Jenkins,S.L., Jagodnik,K.M., autophagosome membranes after processing. EMBO J., 19, Lachmann,A. et al. (2016) Enrichr: a comprehensive gene set 5720–5728. enrichment analysis web server 2016 update. Nucleic Acids Res., 44, 50. Pascal,J.M. and Ellenberger,T. (2015) The rise and fall of W90–W97. poly(ADP-ribose): an enzymatic perspective. DNA Repair, 32, 32. Merrill Ronald,A. (2017) In: Strack,S and 10–16. Usachev,Y (eds). Techniques to investigate mitochondrial function in 51. Ame,J.C., Rolli,V., Schreiber,V., Niedergang,C., Apiou,F., Decker,P., neurons. Humana Press, NY, Vol. 123, pp. 31–48. Muller,S., Hoger,T., Menissier-de Murcia,J. and de Murcia,G. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 Nucleic Acids Research, 2019, Vol. 47, No. 8 4109 (1999) PARP-2, A novel mammalian DNA damage-dependent 72. Rass,U., Ahel,I. and West,S.C. (2008) Molecular mechanism of poly(ADP-ribose) polymerase. J. Biol. Chem., 274, 17860–17868. DNA deadenylation by the neurological disease protein aprataxin. J. 52. Lin,S.C. and Hardie,D.G. (2018) AMPK: sensing glucose as well as Biol. Chem., 283, 33994–34001. cellular energy status. Cell Metab., 27, 299–313. 73. Avery,L. (1993) The genetics of feeding in Caenorhabditis elegans. 53. Kim,J., Kundu,M., Viollet,B. and Guan,K.L. (2011) AMPK and Genetics, 133, 897–917. mTOR regulate autophagy through direct phosphorylation of Ulk1. 74. Kanazawa,T., Zappaterra,M.D., Hasegawa,A., Wright,A.P., Nat. Cell Biol., 13, 132–141. Newman-Smith,E.D., Buttle,K.F., McDonald,K., Mannella,C.A. 54. Hardie,D.G., Schaffer,B.E. and Brunet,A. (2016) AMPK: an and van der Bliek,A.M. (2008) The C. elegans Opa1 homologue energy-sensing pathway with multiple inputs and outputs. Trends EAT-3 is essential for resistance to free radicals. PLoS Genet., 4, Cell Biol., 26, 190–201. e1000022. 55. Koopman,W.J., Distelmaier,F., Smeitink,J.A. and Willems,P.H. 75. Bargmann,C.I., Hartwieg,E. and Horvitz,H.R. (1993) (2013) OXPHOS mutations and neurodegeneration. EMBO J., 32, Odorant-selective genes and neurons mediate olfaction in C. 9–29. elegans. Cell, 74, 515–527. 56. Winge,D.R. (2012) Sealing the mitochondrial respirasome. Mol. 76. Tebbs,R.S., Thompson,L.H. and Cleaver,J.E. (2003) Rescue of Cell Biol., 32, 2647–2652. Xrcc1 knockout mouse embryo lethality by 57. Beranek,D.T. (1990) Distribution of methyl and ethyl adducts transgene-complementation. DNA Repair, 2, 1405–1417. following alkylation with monofunctional alkylating agents. Mutat. 77. Hirano,M., Matsumura,R., Nakamura,Y., Saigoh,K., Res., 231, 11–30. Sakamoto,H., Ueno,S., Inoue,H. and Kusunoki,S. (2017) 58. Fukui,M., Choi,H.J. and Zhu,B.T. (2012) Rapid generation of Unexpectedly mild phenotype in an ataxic family with a two-base mitochondrial superoxide induces mitochondrion-dependent but deletion in the APTX gene. J. Neurol. Sci., 378, 75–79. caspase-independent cell death in hippocampal neuronal cells that 78. Yokoseki,A., Ishihara,T., Koyama,A., Shiga,A., Yamada,M., morphologically resembles necroptosis. Toxicol. Appl. Pharmacol., Suzuki,C., Sekijima,Y., Maruta,K., Tsuchiya,M., Date,H. et al. 262, 156–166. (2011) Genotype-phenotype correlations in early onset ataxia with 59. Weinert,B.T., Moustafa,T., Iesmantavicius,V., Zechner,R. and ocular motor apraxia and hypoalbuminaemia. Brain, 134, Choudhary,C. (2015) Analysis of acetylation stoichiometry suggests 1387–1399. that SIRT3 repairs nonenzymatic acetylation lesions. EMBO J., 34, 79. Alexander,C., Votruba,M., Pesch,U.E., Thiselton,D.L., Mayer,S., 2620–2632. Moore,A., Rodriguez,M., Kellner,U., Leo-Kottler,B., Auburger,G. 60. Ame,J.C., Spenlehauer,C. and de Murcia,G. (2004) The PARP et al. (2000) OPA1, encoding a dynamin-related GTPase, is mutated superfamily. Bioessays, 26, 882–893. in autosomal dominant optic atrophy linked to chromosome 3q28. 61. Akbari,M. and Krokan,H.E. (2008) Cytotoxicity and mutagenicity Nat. Genet., 26, 211–215. of endogenous DNA base lesions as potential cause of human aging. 80. Delettre,C., Lenaers,G., Griffoin,J.M., Gigarel,N., Lorenzo,C., Mech. Ageing Dev., 129, 353–365. Belenguer,P., Pelloquin,L., Grosgeorge,J., Turc-Carel,C., Perret,E. 62. Mozdy,A.D., McCaffery,J.M. and Shaw,J.M. (2000) Dnm1p et al. (2000) Nuclear gene OPA1, encoding a mitochondrial GTPase-mediated mitochondrial fission is a multi-step process dynamin-related protein, is mutated in dominant optic atrophy. Nat. requiring the novel integral membrane component Fis1p. J. Cell Genet., 26, 207–210. Biol., 151, 367–380. 81. Chao de la Barca,J.M., Prunier-Mirebeau,D., Amati-Bonneau,P., 63. Kashatus,D.F., Lim,K.H., Brady,D.C., Pershing,N.L., Cox,A.D. Ferre,M., Sarzi,E., Bris,C., Leruez,S., Chevrollier,A., and Counter,C.M. (2011) RALA and RALBP1 regulate Desquiret-Dumas,V., Gueguen,N. et al. (2016) OPA1-related mitochondrial fission at mitosis. Nat. Cell Biol., 13, 1108–1115. disorders: diversity of clinical expression, modes of inheritance and 64. Manczak,M., Kandimalla,R., Fry,D., Sesaki,H. and Reddy,P.H. pathophysiology. Neurobiol. Dis., 90, 20–26. (2016) Protective effects of reduced dynamin-related protein 1 82. Schaaf,C.P., Blazo,M., Lewis,R.A., Tonini,R.E., Takei,H., Wang,J., against amyloid beta-induced mitochondrial dysfunction and Wong,L.J. and Scaglia,F. (2011) Early-onset severe neuromuscular synaptic damage in Alzheimer’s disease. Hum. Mol. Genet., 25, phenotype associated with compound heterozygosity for OPA1 5148–5166. mutations. Mol. Genet. Metab., 103, 383–387. 65. Delettre,C., Griffoin,J.M., Kaplan,J., Dollfus,H., Lorenz,B., 83. Carelli,V., Sabatelli,M., Carrozzo,R., Rizza,T., Schimpf,S., Faivre,L., Lenaers,G., Belenguer,P. and Hamel,C.P. (2001) Mutation Wissinger,B., Zanna,C., Rugolo,M., La Morgia,C., Caporali,L. spectrum and splicing variants in the OPA1 gene. Hum. Genet., 109, et al. (2015) ‘Behr syndrome’ with OPA1 compound heterozygote 584–591. mutations. Brain, 138, e321. 66. Del Dotto,V., Mishra,P., Vidoni,S., Fogazza,M., Maresca,A., 84. Bonneau,D., Colin,E., Oca,F., Ferre,M., Chevrollier,A., Caporali,L., McCaffery,J.M., Cappelletti,M., Baruffini,E., Gueguen,N., Desquiret-Dumas,V., N’Guyen,S., Barth,M., Lenaers,G. et al. (2017) OPA1 isoforms in the hierarchical Zanlonghi,X. et al. (2014) Early-onset Behr syndrome due to organization of mitochondrial functions. Cell Rep., 19, 2557–2571. compound heterozygous mutations in OPA1. Brain, 137, e301. 67. Heddi,A., Stepien,G., Benke,P.J. and Wallace,D.C. (1999) 85. Liao,C., Ashley,N., Diot,A., Morten,K., Phadwal,K., Williams,A., Coordinate induction of energy gene expression in tissues of Fearnley,I., Rosser,L., Lowndes,J., Fratter,C. et al. (2017) mitochondrial disease patients. J. Biol. Chem., 274, 22968–22976. Dysregulated mitophagy and mitochondrial organization in optic 68. Fakouri,N.B., Durhuus,J.A., Regnell,C.E., Angleys,M., Desler,C., atrophy due to OPA1 mutations. Neurology, 88, 131–142. Hasan-Olive,M.M., Martin-Pardillos,A., Tsaalbi-Shtylik,A., 86. Bonifert,T., Karle,K.N., Tonagel,F., Batra,M., Wilhelm,C., Thomsen,K., Lauritzen,M. et al. (2017) Rev1 contributes to proper Theurer,Y., Schoenfeld,C., Kluba,T., Kamenisch,Y., Carelli,V. et al. mitochondrial function via the PARP-NAD(+)-SIRT1-PGC1alpha (2014) Pure and syndromic optic atrophy explained by deep intronic axis. Sci. Rep., 7, 12480. OPA1 mutations and an intralocus modifier. Brain, 137, 2164–2177. 69. van der Westhuizen,F.H., van den Heuvel,L.P., Smeets,R., 87. Alavi,M.V., Bette,S., Schimpf,S., Schuettauf,F., Schraermeyer,U., Veltman,J.A., Pfundt,R., van Kessel,A.G., Ursing,B.M. and Wehrl,H.F., Ruttiger,L., Beck,S.C., Tonagel,F., Pichler,B.J. et al. Smeitink,J.A. (2003) Human mitochondrial complex I deficiency: (2007) A splice site mutation in the murine Opa1 gene features investigating transcriptional responses by microarray. pathology of autosomal dominant optic atrophy. Brain, 130, Neuropediatrics, 34, 14–22. 1029–1042. 70. Pagliarini,D.J., Calvo,S.E., Chang,B., Sheth,S.A., Vafai,S.B., 88. Chen,L., Liu,T., Tran,A., Lu,X., Tomilov,A.A., Davies,V., Ong,S.E., Walford,G.A., Sugiana,C., Boneh,A., Chen,W.K. et al. Cortopassi,G., Chiamvimonvat,N., Bers,D.M., Votruba,M. et al. (2008) A mitochondrial protein compendium elucidates complex I (2012) OPA1 mutation and late-onset cardiomyopathy: disease biology. Cell, 134, 112–123. mitochondrial dysfunction and mtDNA instability. J. Am. Heart 71. Mootha,V.K., Lindgren,C.M., Eriksson,K.F., Subramanian,A., Assoc., 1, e003012. Sihag,S., Lehar,J., Puigserver,P., Carlsson,E., Ridderstrale,M., 89. Davies,V.J., Hollins,A.J., Piechota,M.J., Yip,W., Davies,J.R., Laurila,E. et al. (2003) PGC-1alpha-responsive genes involved in White,K.E., Nicols,P.P., Boulton,M.E. and Votruba,M. (2007) Opa1 oxidative phosphorylation are coordinately downregulated in deficiency in a mouse model of autosomal dominant optic atrophy human diabetes. Nat. Genet., 34, 267–273. Downloaded from https://academic.oup.com/nar/article-abstract/47/8/4086/5319145 by Ed 'DeepDyve' Gillespie user on 30 April 2019 4110 Nucleic Acids Research, 2019, Vol. 47, No. 8 impairs mitochondrial morphology, optic nerve structure and visual strand break repair proteins XRCC1 and XRCC4. DNA Repair, 3, function. Hum. Mol. Genet., 16, 1307–1318. 1493–1502. 90. Anand,R., Wai,T., Baker,M.J., Kladt,N., Schauss,A.C., Rugarli,E. 96. Date,H., Igarashi,S., Sano,Y., Takahashi,T., Takahashi,T., and Langer,T. (2014) The i-AAA protease YME1L and OMA1 Takano,H., Tsuji,S., Nishizawa,M. and Onodera,O. (2004) The cleave OPA1 to balance mitochondrial fusion and fission. J. Cell FHA domain of aprataxin interacts with the C-terminal region of Biol., 204, 919–929. XRCC1. Biochem. Biophys. Res. Commun., 325, 1279–1285. 91. Song,Z., Chen,H., Fiket,M., Alexander,C. and Chan,D.C. (2007) 97. Horton,J.K., Stefanick,D.F., Caglayan,M., Zhao,M.L., OPA1 processing controls mitochondrial fusion and is regulated by Janoshazi,A.K., Prasad,R., Gassman,N.R. and Wilson,S.H. (2018) mRNA splicing, membrane potential, and Yme1L. J. Cell Biol., 178, XRCC1 phosphorylation affects aprataxin recruitment and DNA 749–755. deadenylation activity. DNA Repair, 64, 26–33. 92. Cipolat,S., Martins de Brito,O., Dal Zilio,B. and Scorrano,L. (2004) 98. Camacho-Pereira,J., Tarrago,M.G., Chini,C.C.S., Nin,V., OPA1 requires mitofusin 1 to promote mitochondrial fusion. Proc. Escande,C., Warner,G.M., Puranik,A.S., Schoon,R.A., Reid,J.M., Natl. Acad. Sci. U.S.A., 101, 15927–15932. Galina,A. et al. (2016) CD38 dictates age-related NAD decline and 93. Harris,J.L., Jakob,B., Taucher-Scholz,G., Dianov,G.L., mitochondrial dysfunction through an SIRT3-dependent Becherel,O.J. and Lavin,M.F. (2009) Aprataxin, poly-ADP ribose mechanism. Cell Metab., 23, 1127–1139. polymerase 1 (PARP-1) and apurinic endonuclease 1 (APE1) 99. Fang,E.F., Lautrup,S., Hou,Y., Demarest,T.G., Croteau,D.L., function together to protect the genome against oxidative damage. Mattson,M.P. and Bohr,V.A. (2017) NAD(+) in aging: molecular Hum. Mol. Genet., 18, 4102–4117. mechanisms and translational implications. Trends Mol. Med., 23, 94. Durocher,D. and Jackson,S.P. (2002) The FHA domain. FEBS 899–916. Lett., 513, 58–66. 100. Sorrentino,V., Menzies,K.J. and Auwerx,J. (2018) Repairing 95. Clements,P.M., Breslin,C., Deeks,E.D., Byrd,P.J., Ju,L., mitochondrial dysfunction in disease. Annu. Rev. Pharmacol. Bieganowski,P., Brenner,C., Moreira,M.C., Taylor,A.M. and Toxicol., 58, 353–389. Caldecott,K.W. (2004) The ataxia-oculomotor apraxia 1 gene product has a role distinct from ATM and interacts with the DNA

Journal

Nucleic Acids ResearchOxford University Press

Published: May 7, 2019

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create folders to
organize your research

Export folders, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off