Alternative oxidase-mediated respiration prevents lethal mitochondrial cardiomyopathy

Research Article Alternative oxidase-mediated respiration prevents lethal mitochondrial cardiomyopathy Jayasimman Rajendran 1,2 , Janne Purhonen 1,2 , Saara Tegelberg 1,3,4 , Olli-Pekka Smolander 5 , Matthias Mörgelin 6 , Jan Rozman 7,8 , Valerie Gailus-Durner 7 , Helmut Fuchs 7 , Martin Hrabe de Angelis 7,8,9 , Petri Auvinen 5 , Eero Mervaala 10 , Howard T Jacobs 5,11 , Marten Szibor 5,11 , Vineta Fellman 1,3,12 & Jukka Kallijärvi 1,2,* Abstract Alternative oxidase (AOX) is a non-mammalian enzyme that can bypass blockade of the complex III-IV segment of the respiratory chain (RC). We crossed a Ciona intestinalis AOX transgene into RC complex III (cIII)-deficient Bcs1l p.S78G knock-in mice, displaying multiple visceral manifestations and premature death. The homozygotes expressing AOX were viable, and their median survival was extended from 210 to 590 days due to permanent prevention of lethal cardiomyopathy. AOX also prevented renal tubular atrophy and cerebral astrogliosis, but not liver disease, growth restriction, or lipodystrophy, suggesting distinct tissuespecific pathogenetic mechanisms. Assessment of reactive oxygen species (ROS) production and damage suggested that ROS were not instrumental in the rescue. Cardiac mitochondrial ultrastructure, mitochondrial respiration, and pathological transcriptome and metabolome alterations were essentially normalized by AOX, showing that the restored electron flow upstream of cIII was sufficient to prevent cardiac energetic crisis and detrimental decompensation. These findings demonstrate the value of AOX, both as a mechanistic tool and a potential therapeutic strategy, for cIII deficiencies. Keywords BCS1L; complex III; GRACILE syndrome; mitochondrial disorder; respiratory chain Subject Categories Cardiovascular System; Genetics, Gene Therapy & Genetic Disease DOI 10.15252/emmm.201809456 | Received 8August 2018 | Revised 8November 2018 | Accepted 12 November 2018 | Published online 10 December 2018 EMBO Mol Med (2019)11:e9456 See also: A Saada (January 2019) Introduction Mitochondrial disorders are the most common class of inherited errors of metabolism. However, effective treatments are lacking, and their clinical management remains largely supportive (Pfeffer et al, 2013). In patients with RC cIII (ubiquinol:cytochrome c oxidoreductase) deficiency, mutations in several genes encoding either cIII subunits or assembly factors have been identified. These compromise cIII enzymatic activity and result in a wide variety of clinical manifestations (Fernandez-Vizarra & Zeviani, 2015). BCS1L mutations are the most common cause of cIII deficiency, with various neonatal and adult phenotypes described worldwide (Fernandez-Vizarra & Zeviani, 2015), the most severe and prevalent of them being GRACILE syndrome (fetal growth restriction, aminoaciduria, cholestasis, liver iron overload, lactic acidosis, and early death during infancy) (Fellman et al, 1998; Visapa ¨a ¨et al, 2002). BCS1L is a mitochondrial inner membrane translocase required for Rieske iron–sulfur protein (RISP, UQCRFS1) topogenesis and incorporation into cIII (Nobrega et al, 1992; Cruciat et al, 1999). Homozygous Bcs1l c.A232G (Bcs1l p.S78G ) knock-in mice bearing the GRACILE syndrome-analogous mutation recapitulate many of the clinical manifestations, including growth failure, progressive hepatopathy, kidney tubulopathy, and, in a C57BL/6JCrlBomTac background, short survival of 35 days (Leve ´en et al, 2011; Kotarsky et al, 2012; 1Folkhälsan Research Center, Helsinki, Finland 2Clinicum, Faculty of Medicine, University of Helsinki, Helsinki, Finland 3Department of Clinical Sciences, Lund, Pediatrics, Lund University, Lund, Sweden 4Molecular Neurology Research Program and Neuroscience Center, University of Helsinki, Helsinki, Finland 5Institute of Biotechnology, University of Helsinki, Helsinki, Finland 6Division of Infection Medicine, Clinical Sciences, Lund University, Lund, Sweden 7German Mouse Clinic, Institute of Experimental Genetics, Helmholtz Zentrum München, German Research Center for Environmental Health, Neuherberg, Germany 8German Center for Diabetes Research (DZD), Neuherberg, Germany 9Chair of Experimental Genetics, Center of Life and Food Sciences Weihenstephan, TU Munich, Freising-Weihenstephan, Germany 10 Department of Pharmacology, Faculty of Medicine, University of Helsinki, Helsinki, Finland 11 Faculty of Medicine and Life Sciences, University of Tampere, Tampere, Finland 12 Children’s Hospital, Helsinki University Hospital, University of Helsinki, Helsinki, Finland *Corresponding author. Tel: +358 504487006; E-mail: jukka.kallijar[email protected] ª2018 The Authors. Published under the terms of the CC BY 4.0license EMBO Molecular Medicine 11:e9456 |2019 1of 19 Published online: December 10, 2018 Rajendran et al, 2016; Purhonen et al, 2017). In the slightly different C57BL/6JCrl substrain, the homozygotes develop the same early manifestations but do not succumb to the early metabolic crisis. This extends their survival to over 150 days (Purhonen et al, 2017) and brings additional later-onset phenotypes, such as cerebral astrogliosis (Tegelberg et al, 2017). Under physiological conditions, quinols that transport electrons in the mitochondrial inner membrane are efficiently oxidized by cIII, with electron transfer via cytochrome c and cytochrome c oxidase (complex IV, cIV) to oxygen (Brand, 2010; El-Khoury et al, 2014). However, plants and some lower organisms, but not mammals, express alternative oxidases (AOXs) that transfer electrons directly from quinols to oxygen without proton translocation. Their main role is to maintain electron flow when the cIII-cIV segment of the RC is impaired, limiting production of ROS and supporting redox and metabolic homeostasis (McDonald & Vanlerberghe, 2004; El-Khoury et al, 2014). Ciona intestinalis AOX has been cloned and expressed in human cultured cells (Hakkaart et al, 2006), fruit flies and mice (El-Khoury et al, 2013; Szibor et al, 2017). In these models, AOX is inert under non-stressed conditions, most likely because it accepts electrons only when the quinone pool is highly reduced (Hoefnagel & Wiskich, 1998; Castro-Guerrero et al, 2004), such as under inhibition or overload of cIII or cIV (Dassa et al, 2009). Accordingly, upon inhibition of cIII or cIV by mutations or chemical inhibitors, ectopic AOX can maintain respiration and prevent cell death (Dassa et al, 2009; Fernandez-Ayala et al, 2009). We set out to test whether AOX expression could prevent the detrimental effects of cIII deficiency in a mammalian model, by restoring electron flow upstream of cIII. To this end, we crossed mice carrying a broadly expressed AOX transgene (Szibor et al, 2017) with the Bcs1l c.A232G mice and assessed disease progression, organ manifestations, and metabolism in the homozygotes with and without AOX expression. Results Broadly expressed AOX triples the life span of cIII-deficient Bcs1l p.S78G mice To assess the effect of cIII bypass on the survival and tissue manifestations in cIII-deficient mice, we bred cohorts of wild-type and Bcs1l mutant mice with or without a Ciona intestinalis AOX transgene. Hereafter, we will refer to the Bcs1l p.S78G homozygotes as GRAC (as an abbreviation of GRACILE syndrome) mice. The Bcs1l p.S78G homozygotes carrying AOX transgene will be referred to as GROX mice (GRAC +AOX). Figure 1A shows a timeline of the appearance of the previously reported and novel phenotypes in GRAC mice, as well as the assessments included in this study. The GRAC mice reached the criteria of euthanasia between postnatal day 180 (P180) and P220, with median survival to P210 (Fig 1B). In contrast, the GROX mice showed no signs of terminal deterioration or spontaneous deaths at P200 and survived to a median age of 590 days (Fig 1B). To assess whether the extended survival was due to an overall improvement in energy metabolism, we measured growth, whole-body metabolism, and body composition in young adult mice. The GRAC mice were growth restricted (Fig 1C and E) and had increased lactate-to-glucose ratio (Fig 1D), low fat mass (Fig 1F), bone density (Fig 1G), and heat production (Fig 1H) and, in females, low respiratory exchange ratio (Fig 1I). Unexpectedly, AOX had no or only small effect on these parameters (Fig 1C–I), suggesting that the AOX-mediated extension of survival depended on a tissue or cell-type specific pathology rather than whole-body energy metabolism. AOX permanently prevents lethal cardiomyopathy and alleviates renal and cerebral manifestations Histopathological analysis of autopsy samples from end stage (P200) mice showed two novel phenotypes not previously reported in studies of younger GRAC mice: cortical kidney atrophy and cardiomegaly with dilated ventricles (Fig 2A and B). Fibrosis was prominent in liver, kidney, and heart (Fig 2C and D). Suspecting cardiomyopathy as the cause of death, we assessed cardiac functions at several time points. Echocardiography showed minimal functional changes at P150 (Fig EV1A–F), but severe dilated cardiomyopathy, focal fibrosis, decreased ejection fraction, and fractional shortening at P200 (Fig 2A and F), indicating end-stage cardiomyopathy. Strikingly, the GROX littermates had normal heart size (Fig 2A and B), no fibrosis (Fig 2C and D), and overtly normal cardiac function (Fig 2E and F), explaining their extended survival. mRNA expression of key markers for cardiac hypertrophy and fibrosis was significantly altered already in the presymptomatic (P150) GRAC hearts, and these changes were largely prevented in the GROX mice (Fig 2G). A kidney stress test with salt-enriched (6% w/w) chow starting at P150 had no effect on cardiac function at P200, or on survival, and blood pressure was only increased at the end stage in the salt-fed mice when compared to P150 baseline (Fig EV1A–F), consistent with a primary cardiomyopathy. Remarkably, the GROX mice had normal-sized and non-fibrotic heart throughout their life span, to over P600 (Fig EV2A–D). Kidneys of GRAC mice showed proximal tubulopathy with fibrosis (Fig 2C and D) and tubular degeneration (decreased tubular mass) at P200 (Appendix Fig S1). In GROX mice, the kidney mass and apparent tubular volume were preserved (Fig 2A and B). On the basis of the proliferation marker Ki67 and apoptosis marker cleaved caspase-3, this was likely due to decreased apoptosis rather than increased regeneration (Appendix Fig S1I and J). However, AOX had only minor effect on other histological lesions (Appendix Fig S1A–G) or functional parameters: albuminuria, hematuria, and urinary creatinine (Fig 2H and I, Appendix Fig S1H). At P600, the kidneys were severely fibrotic but still of normal size, indicating long-term protection from tubular atrophy (Fig EV2A, B and D). Surprisingly, AOX had no effect on the liver fibrosis (Fig 2C and D) or the elevated liver enzymes (Fig 2J and K) at end stage. The cause of death of the GROX mice remains unknown, but eventual deterioration due to the progressing kidney and liver disease is an obvious possible explanation. GRACILE syndrome patients have no encephalopathy, but the brains of GRAC mice show peculiar focal astrogliosis in the primary barrel field of the somatosensory cortex (S1BF) (Tegelberg et al, 2017). Staining for glial fibrillary acidic protein (GFAP) showed that the astrogliosis was almost fully prevented by AOX at P200 (Fig 3A). 2of 19 EMBO Molecular Medicine 11:e9456 |2019 ª2018 The Authors EMBO Molecular Medicine AOX rescues cIII deficiency in mice Jayasimman Rajendran et al Published online: December 10, 2018 A BC DE F GH I Figure 1. AOX expression prolongs the survival of cIII-deficient Bcs1l p.S78G mice without affecting growth or whole-body metabolism. A Schematic presentation of the multiorgan manifestations, described in this study or previously, in homozygous Bcs1l c.A232G (GRAC) mice, and the time points of the investigations performed in this study. B Survival curves of homozygous Bcs1lmutant mice without (GRAC) and with (GROX) alternative oxidase (AOX) expression (n=18–21/group). The median survival of GRAC mice was 210 days and of GROX mice was 589 days with no gender difference. C Weight of mice from P50 to P200 (n>10/group) and at P600 (n=4–6/group). D Blood lactate-to-glucose ratio at P200 (n=8/group). E–G Dual-energy X-ray absorptiometry (DEXA) analysis (n=10/group) of (E) lean mass, (F) fat mass, and (G) bone mineral density at P98. H, I Indirect calorimetric measurement (n=10/group) of (H) heat production and (I) respiratory exchange ratio at P77. Data information: The survival data were analyzed using log-rank (Mantel–Cox) test (P<0.0001). Bar graphs represent mean SD. The data (D–I) were analyzed using Kruskal–Wallis and Mann–Whitney U-tests for selected comparison. Significant differences between groups (P-value) are indicated on graphs. ª2018 The Authors EMBO Molecular Medicine 11:e9456 |2019 3of 19 Jayasimman Rajendran et al AOX rescues cIII deficiency in mice EMBO Molecular Medicine Published online: December 10, 2018 AOX preserves mitochondrial ultrastructure in rescued tissues Previous studies have shown disrupted hepatic mitochondrial ultrastructure in younger Bcs1l mutant mice (Leve ´en et al, 2011; Purhonen et al, 2017). In the P150 GRAC mice, electron microscopy showed abnormal mitochondrial ultrastructure in hepatocytes as well as in kidney tubular cells, and, importantly, in cardiomyocytes already at onset of the cardiomyopathy (Fig 3B). The mitochondria A B C DE GH I J K F Figure 2. 4of 19 EMBO Molecular Medicine 11:e9456 |2019 ª2018 The Authors EMBO Molecular Medicine AOX rescues cIII deficiency in mice Jayasimman Rajendran et al Published online: December 10, 2018 were smaller and contained fewer cristae, which were thicker than in WT mitochondria. AOX fully or partially prevented these changes in cardiomyocytes and kidney tubular cells, but not in hepatocytes, (Fig 3C–F), correlating faithfully with the rescue of tissue pathology. AOX relieves cardiac metabolic stress and prevents decompensation Seeking a mechanism for the remarkably tissue-specific rescue effects of AOX, we performed transcriptomics and metabolomics at P150, at the onset of the cardiomyopathy. The three affected tissues showed marked global transcriptional changes, including in the GRAC heart despite it being functionally normal at this stage (Fig 4A–C). In correlation with the histological findings, AOX normalized the expression of only few genes in the liver, but of about 25% of dysregulated genes in the kidney and about 50% in the heart (Appendix Fig S2A–C). The most robustly upregulated gene sets were related to extracellular matrix organization, a signature of tissue remodeling and fibrosis (Fig 4D–F, Appendix Table S1), and cell cycle in liver and kidney, but not in the heart (Fig 4D–F) reflecting differential regenerative capacity of these tissues. As expected, energy metabolism-related gene expression was altered in all three tissues (Fig 4D–F, Appendix Fig S2D). Alternative oxidase almost fully prevented these changes in heart, but not in kidney or liver (Fig 4D–F). Notably, AOX had a significant effect on gene expression in the Bcs1l wild-type heart, including upregulation of genes related to mitochondrial function (Fig 4D, G, I and Appendix Fig S2D). Expression of the major cardiac metabolic regulator HIF-1aand the metabolic stress-inducible transcriptional regulators ATF3 and ATF4 (Kalfon et al, 2017; Quiros et al, 2017) was elevated in GRAC hearts and normalized by AOX (Fig 4J and K), indicating relieved metabolic stress. Upregulation of PGC-1a, the master regulator of mitochondrial biogenesis, and the mitochondrial transcription factor TFAM, in the GRAC heart suggested an attempt for compensatory mitochondrial biogenesis (Fig 4G–I). However, Western blot analysis using the abundance of RC complex subunits as a proxy showed no significant changes in mitochondrial mass between the groups (Fig EV3I) in any tissue. We previously identified an upregulated set of genes, which we designated cIII stress signature, in P45 Bcs1l mutant livers (Purhonen et al, 2017). This gene set was also highly upregulated in all three GRAC tissues. Despite rescue of kidney tubular mass, AOX amplified the cIII stress signature specifically in this tissue, but not in the heart or liver (Fig 4L). Alternative oxidase is under the strong synthetic CAG promoter and expressed in all tissues in the Rosa26 AOX mice (Szibor et al, 2017). In our mice, AOX mRNA expression was similar in heart and liver and somewhat lower in kidney (Fig EV3H). In total tissue lysates from the AOX mice, the amount of AOX protein was considerably higher in heart than in liver or kidney (Fig EV3H). However, this difference was mainly due to the higher mitochondrial mass in heart, as shown by the mitochondrial loading control VDAC1 and also by most respiratory chain subunits (Fig EV3I). Interestingly, the amount of AOX protein was affected by the Bcs1l mutation (AOX vs. GROX mice) so that the amount in the three GROX tissues was almost identical (Fig EV3H), which essentially rules out that the differences in rescue would be due to different levels of AOX expression. The metabolomics revealed only modest cardiac metabolite changes at the onset (P150) of the cardiomyopathy (Fig 5A, Appendix Table S2). Nevertheless, several three-carbon glycolytic intermediates were depleted, and these tended to be normalized by AOX (Fig 5A). In line with this, the gene expression of PPAR-1a,the major driver of fatty acid utilization, was downregulated in GRAC but normal in GROX heart (Fig 4G). The TCA cycle metabolites malate and fumarate, adenylate energy charge, and NADH/NAD + ratio, which all could be affected by both the cIII blockade and AOX, were not changed in the GRAC or GROX heart tissue (Fig 5C–F). Interestingly, the amino acid proline, which has been long known to accumulate in conditions with lactic academia (Kowaloff et al,1977),was below detection limit in WT heart but increased to approximately 200 nmol/g in GRAC heart. The proline accumulation was partially prevented by AOX (Fig 5G). Concurrently, glutamate, the biosynthetic precursor of proline, was decreased to about 50% in GRAC mice and normalized to WT level in GROX heart tissue (Fig 5A). The mRNA levels of glutamate c-semialdehyde synthetase (P5CS, ALDH18A1) and pyrroline-5-carboxylate reductase 1 (P5CR, PYCR1), both driving proline synthesis from glutamate, were upregulated 1.8fold and 1.7-fold, respectively, in the GRAC heart (Fig 5H and I). Targeted metabolomics of liver tissue (Fig 5B, Appendix Table S2) showed significant metabolite changes characteristic of failing energy metabolism and glycogen depletion, as previously shown in juvenile mice (Kotarsky et al, 2012). These included decreased hexose phosphates, several other glycolytic intermediates, acetyl-CoA and NAD + , and elevated TCA cycle intermediates and amino acids, the latter suggesting protein degradation for fuel. NADH/NAD + ratio was increased (Fig 5F), in line with our previously published data showing decreased hepatic NAD + (Purhonen et al, 2018). Proline level was not changed in liver (Appendix Table S2). Alternative oxidase had only a minor effect on the hepatic metabolite levels (Fig 5B, E, and F). ◀Figure 2. AOX prevents lethal cardiomyopathy and progression of renal tubulopathy to kidney atrophy. A Hematoxylin–eosin-stained cross sections of heart and kidney at P200. B Weight of heart and kidney at P200 (n=4/group). C, D Sirius Red staining (C) for fibrosis in liver, heart, and kidney at P200, and (D) quantification of fibrosis in myocardium, kidney cortex, and liver (n=5–7). E, F Echocardiography data (n=4–6/group) showing (E) ejection fraction (EF) and fractional shortening (FS). (F) Systolic (LV Vol;S) and diastolic (LV Vol;D) left ventricle volume in mice. G Expression of cardiac hypertrophy and fibrosis-associated genes in the presymptomatic (P150) heart (n=6/group). H, I 24-h excretion of (H) albumin and (I) creatinine in urine at P200 (n=4/group). J, K Liver enzymes (J) alanine aminotransferase (ALT) and (K) alkaline phosphatase (ALP) in plasma at P200 (n=8/group). Data information: Bar graphs represent mean SD. Statistics: one-way ANOVA followed by Tukey’s test (for graphs B, D, G, H, J, and K), one-way ANOVA followed by unpaired t-test with Welch’s correction (for graphs E and F), and Mann-Whitney U-tests (for graph I). ª2018 The Authors EMBO Molecular Medicine 11:e9456 |2019 5of 19 Jayasimman Rajendran et al AOX rescues cIII deficiency in mice EMBO Molecular Medicine Published online: December 10, 2018 A BC D F E Figure 3. AOX ameliorates cerebral astrogliosis and maintains mitochondrial ultrastructure in rescued tissues. A GFAP staining for cerebral astrocytes. The barrel field of the primary somatosensory cortex (S1BF) is highlighted with brackets. B Mitochondrial ultrastructure in cardiomyocytes, kidney tubular cells, and hepatocytes at P200 as visualized by electron microscopy. Scale bar 1lm. C–F (C) Average cross-sectional area of mitochondrion, (D) number of cristae per mitochondrion, (E) crista thickness, and (F) average distance between cristae in mitochondria (n=3mice/group). Data information: Bar graphs represent mean SD. Statistics for graphs (C–F): one-way ANOVA followed by Tukey’s test. 6of 19 EMBO Molecular Medicine 11:e9456 |2019 ª2018 The Authors EMBO Molecular Medicine AOX rescues cIII deficiency in mice Jayasimman Rajendran et al Published online: December 10, 2018 ABC DEF GIH JL K Figure 4. AOX mitigates cardiac, but not hepatic or renal, metabolic stress-related gene expression changes. A–C Principal component analysis of transcriptome data from (A) heart, (B) kidney, and (C) liver. D–F Heat map visualization of pathway enrichment analysis (Reactome database). Full pathway analysis is provided in Appendix Table S1. Benjamin–Hochberg FDRcorrected P-values are color labeled as indicated for the three comparisons. G–K Gene expression of major transcriptional regulators of energy metabolism (PPAR-a, HIF-1a), mitochondrial biogenesis (PGC-1a, TFAM), and stress responses (ATF4) in heart at P150 (n=6/group). L Expression of cIII stress signature genes in heart, kidney, and liver (n=6/group). Error bars represent 95%confidence interval of mean difference. Data information: Box plots (G–K) represent quartiles, maximum value, and minimum value (relative fold change “FC”to WT). Statistics: one-way ANOVA followed by Tukey’s test. ª2018 The Authors EMBO Molecular Medicine 11:e9456 |2019 7of 19 Jayasimman Rajendran et al AOX rescues cIII deficiency in mice EMBO Molecular Medicine Published online: December 10, 2018 AB CDE F HIG Figure 5. AOX has little effect on cardiac energy metabolites but attenuates proline accumulation at onset of disease. A Significantly altered metabolites (FDR <0.2;P<0.05,n=5/group) in presymptomatic (P150) heart tissue of GRAC (red) and GROX (green) mice. B Significantly altered metabolites (FDR <0.2;P<0.05,n=5/group) in liver tissue of GRAC (red) and GROX (green) mice. C–F Concentrations of the TCA cycle intermediates (C) fumarate and (D) malate, (E) adenylate energy charge, and (F) NADH/NAD + ratio (n=5/group). G–I (G) Proline concentration in heart tissue at P200 (# below detection limit) and mRNA expression of proline synthesis-related genes (n=6/group) (H) Aldh18a1and (I) Pycr1at P150. Data information: Bar graphs (A–G) represent mean SD, and box plots of mRNA expressions represent quartiles, minimum value, and maximum value (relative fold change “FC”to WT). Statistics for graphs (A, B, F, and H): one-way ANOVA followed by Tukey’s test. 8of 19 EMBO Molecular Medicine 11:e9456 |2019 ª2018 The Authors EMBO Molecular Medicine AOX rescues cIII deficiency in mice Jayasimman Rajendran et al Published online: December 10, 2018 AOX restores cardiac mitochondrial respiration To investigate whether the AOX rescue effect was linked to improved RC assembly and function, we first performed blue native gel electrophoresis (BNGE) analyses. These confirmed the loss of RISP from free cIII dimer (cIII 2 ) and from the cI-cIII 2 supercomplex (SC1) in GRAC tissues (Fig 6A). Surprisingly, the amount of this fully assembled cIII 2 was significantly increased in the GROX heart when compared to GRAC (Fig 6B and C). In the GROX liver and kidney, the change was opposite, with lower amount of RISP in cIII 2 (Fig 6B and C). Spectrophotometrically measured cIII activity in isolated mitochondria was decreased to <50% of WT values in all three tissues at P150, a threshold for the appearance of hepatic pathology in juvenile mice (Leve ´en et al, 2011). Despite that AOX should theoretically not affect cIII activity, it was partially restored in GROX heart mitochondria (Fig 6F), in line with the improved RISP assembly. The activities of cI, cII, and cIV were unchanged in heart and liver (Fig 6D, E and G). Respirometry confirmed AOX-mediated respiration in all three GROX tissues at P150, as measured using the AOX inhibitor n-propyl gallate (nPG). Respiration did not respond to nPG in AOX mice (with wild-type Bcs1l and normal cIII activity), which indicates that AOX was catalytically active only in the mutants (Fig 6H). The cIII dysfunction did not limit state 3 respiration of liver and kidney mitochondria (Fig 6I and J) when driven by cI-linked substrates (malate, pyruvate, and glutamate to generate NADH), but clearly did so in the heart (Fig 6I). Further stimulation of respiration through cII by addition of succinate revealed cIII deficiency in all three tissues in GRAC mice, with the most severe effect in the heart (Fig 6J). Remarkably, cIand cII-linked state 3 respiration (malate, pyruvate, glutamate, and succinate at ADP saturation) were rescued to wildtype level in GROX heart mitochondria (Fig 6I and J). A similar effect was found in state 3 respiration of heart and liver at P200 (Fig EV4A and B). AOX does not affect ROS damage or defense but normalizes cardiac NO-related gene expression Limiting excessive ROS production from the RC has been posited as a major mechanism of action of AOX (El-Khoury et al, 2014). To assess ROS production in the affected GRAC tissues, we measured hydrogen peroxide (H 2 O 2 ) emission from isolated mitochondria at P150. The Amplex Red-peroxidase assay showed that total H 2 O 2 emission was decreased in GROX heart and kidney as compared to GRAC (Fig 7A). However, analysis of gene expression related to ROS defense and damage showed that glutathione synthesis and conjugation were upregulated only in the GRAC liver (Fig 7B) and were not affected by AOX. GRAC hearts showed no upregulation of ROS defense at all (Fig 7B) despite the most severe cIII dysfunction. Total tissue glutathione was not changed in GRAC heart or liver but was increased in both GROX tissues (Fig 7C). Mitochondrial aconitase activity, a sensitive marker of ROS damage to iron–sulfur clusters (Yan et al, 1997), and protein carbonylation (Fig EV4C and D) were not significantly different between the genotypes. Urinary isoprostanes, a marker of systemic oxidative stress (Fig 7D), was elevated in both GRAC and GROX. Immunohistochemical staining for the lipid peroxidation product 4-hydroxynonenal was increased in all three GRAC tissues, but the increase was not prevented by AOX (Fig 7E and F). Finally, we fed the GRAC mice the mitochondria-targeted ROS scavenger mitoQ, which has shown beneficial effects in mouse models of cardiac ischemia–reperfusion injury (Adlam et al, 2005). MitoQ feeding started at P150 had no effect on the survival of the mice by P200 or on their cardiac function (Fig EV1A–F). Since ROS damage was clearly not instrumental in the cardiac rescue by AOX, we looked at nitric oxide (NO) metabolism that is particularly important in the cardiovascular system and known to interact chemically with ROS and the respiratory chain (Carnicer et al, 2013). We observed a significant increase in NO signalingrelated gene expression specifically in GRAC hearts, and these changes were attenuated by AOX (Fig 8A). Decreased mRNA expression of central targets of NO signaling, ryanodine receptor (RyR2), sarco(endo)plasmic reticulum Ca 2+ -ATPase Serca2 (Atp2a2), and phospholamban (Pln) in GRAC heart suggested decreased cardiac contraction-related Ca 2+ channel function (Fig 8B–D). AOX upregulated these Ca 2+ channel genes in both Bcs1l mutant and wild-type mice. Expression of endothelial NO synthase (Nos3) was increased in GRAC heart (Fig 8E), and, most strikingly, expression of neuronal NO synthase (nNos,Nos1) was increased over 10-fold in the GRAC heart with a significant attenuation in GROX mice. Western blot analysis confirmed similar upregulation of NOS1 protein (Fig 8F and G). Increased gene expression of dimethylarginine dimethylaminohydrolase 1 (Ddah1), which hydrolyzes the endogenous NOS inhibitors dimethylarginine and monomethylarginine (Fig 8H), and increased product/substrate (citrulline/arginine) ratio (Fig 8I) in GRAC heart suggested increased NO production. However, neither protein nitrotyrosine nor total nitrites were increased in the GRAC hearts (Fig 8J and K). Discussion Here, we show that AOX, a non-mammalian enzyme that can bypass cIII blockade by shunting electrons directly from the quinone pool to oxygen, is able to permanently prevent lethal cardiomyopathy and alleviate multiple other pathologies in cIII-deficient GRAC mice. The GRAC mice, initially bred in C57BL/6JBomTac background (Leve ´en et al, 2011), have turned out particularly useful in mechanistic and interventional studies (Rajendran et al, 2016; Purhonen et al, 2017, 2018) due to their postnatal symptom-free period and short survival. In the current study, we found that, in the C57BL/6JCrl background, the homozygotes survive to median P210 and develop lethal cardiomyopathy after P150, a manifestations not seen in the GRACILE syndrome patients with early neonatal lethality, but a common manifestation in other mitochondrial disorders. AOX provided a full functional rescue of the cardiomyopathy, restored cIand cII-linked respiration to wild-type levels, and abrogated the signature of metabolic stress responses. Heart muscle is highly dependent on mitochondrial respiration and operates at constant ATP and phosphocreatine concentrations (VenturaClapier et al, 2011; Guzun et al, 2015). Our results clearly indicate that the pathogenesis mechanism leading to cardiomyopathy in GRAC mice is insufficiency of electron flow, resulting in disturbed redox and metabolic homeostasis. Cardiac hypertrophy and a metabolic switch from fatty acid to glucose utilization are hallmarks of ª2018 The Authors EMBO Molecular Medicine 11:e9456 |2019 9of 19 Jayasimman Rajendran et al AOX rescues cIII deficiency in mice EMBO Molecular Medicine Published online: December 10, 2018 Isoprostanes were quantified from urine using a commercial ELISA kit (ab175819, Abcam). Transcriptomics For RNA sequencing, rRNA was removed and the sequencing libraries were prepared using Illumina TruSeq Stranded Total RNA Library Prep Kit with Ribo-Zero Human/Mouse/Rat. Libraries were sequenced on Illumina NextSeq 500 instrument using 75 bp kit. Adapter sequences and low-quality reads were removed from the data using cutadapt (Martin, 2011), and data were further screened for remaining rRNA reads using SortMeRNA (Kopylova et al, 2012). The data were mapped to M. musculus genome GRCm38.p4 using STAR (Dobin et al, 2013). The original genome sequence and annotation were augmented with sequence and annotation for AOX. Count data were processed in R using GenomicFeatures and GenomicAlignments (Lawrence et al, 2013), and the differential expression analysis was carried out using DESeq2 (Love et al, 2014). PCA plots were generated using ClustVis online tool (www.b iit.cs.ut.ee/clustvis). Pathway enrichment analysis Enrichment analyses were performed using filtered gene sets (|FC|>1.5, P<0.05) against Reactome database using www.mouse mine.org. To visualize pathway changes, heat maps were generated by Gitool.2.3.1 with WT mean-centered Z-score values and gene mapping for selected pathways as described (Gundem & LopezBigas, 2014; Perez-llamas & Lopez-Bigas, 2011). Pathway heat maps were generated based on Z-score, showing upregulation (brown color) and downregulation (blue color) of each pathway. P-values for group comparison (labeled 1, 2, and 3) were calculated using one-way ANOVA followed by selected comparisons and the Benjamin–Hochberg FDR correction. Metabolomics Targeted quantitative metabolomics covering 116 metabolites was performed using capillary electrophoresis–mass spectrometry (CETOMFS and CE-QqQMS) at Human Metabolome Technologies Inc., Japan (http://humanmetabolome.com/en/). Details of the protocol are available upon request. We omitted the AOX group in the metabolomics since ectopic AOX was shown to be inert in healthy wildtype mice (Szibor et al, 2017). Statistics (excluding transcriptomics) For normally distributed data, one-way ANOVA followed by Tukey’s test or by selected comparisons using t-tests with Welch’s correction was used. Normality of data was assessed by the d’Agostino–Pearson omnibus normality test, and equality of variances by Bartlett’s test. Data not compatible with parametric tests were assessed by Kruskal–Wallis and Mann–Whitney U-tests. Survival curves were analyzed using log-rank Mantel–Cox test. GraphPad Prism 7 software (GraphPad Software Inc., La Jolla, CA, USA) was used for the statistical analyses. Group sizes (n) and the statistical tests used are described in figure legends, and exact P-values are shown in the graphs. Data availability Additional raw data and detailed protocols are available from the authors upon request. Transcriptomics data have been deposited to ArrayExpress database at EMBL-EBI under the accession ID E-MTAB-7416 (https://www.ebi.ac.uk/arrayexpress/experiments/EMTAB-7416). Expanded View for this article is available online. The paper explained Problem Mitochondrial diseases are genetic disorders of energy metabolism that can affect any one or several organs of the body, including skeletal muscle, heart, brain, and visceral organs. Nowadays, early diagnosis of mitochondrial disease is often possible using modern molecular genetics, but treatment options remain scarce. Mutations in the BCS1L gene are the most common cause of mitochondrial diseases affecting the respiratory chain complex III (cIII). GRACILE syndrome (Growth Restriction, Aminoaciduria, Cholestasis, liver Iron overload, Lactic acidosis, and Early death), with neonatal lethality, is the most severe of them. Due to the strikingly similar disease in all GRACILE syndrome patients, Prof. Fellman’s group previously introduced the underlying missense mutation (Bcs1l c.A232G ,Bcs1l p.S78G ) into mice. This mouse model faithfully recapitulates most of the symptoms of the patients. In search for novel strategies to alleviate cIII deficiency-related pathology, we crossed the Bcs1l c.A232G mice with transgenic mice expressing alternative oxidase (AOX), a mitochondrial inner membrane enzyme that can bypass respiratory electron flow blockade when the quinol oxidation capacity of cIII is compromised. Results Mice homozygous for the Bcs1l c.A232G mutation survived approximately 200 days and died of dilating cardiomyopathy, a novel lateonset phenotype in the C57BL/6JCrl background. In contrast, the homozygotes carrying the AOX transgene lived approximately 600 days and never developed the cardiomyopathy. AOX also ameliorated the severe kidney disease and focal astrogliosis of the brain in the homozygotes. Surprisingly, AOX did not correct the liver disease, poor growth, and loss of white fat, suggesting different disease mechanism in different tissues. AOX corrected the abnormal ultrastructure of mitochondria and mitochondrial respiration in those tissues, in which the tissue pathology was alleviated. Heart and kidney mitochondria from the homozygotes showed elevated reactive oxygen species (ROS) production, but analyses of ROS damage suggested that the beneficial effects of AOX were not ROS-related. Instead, we found that AOX normalized cardiac gene expression related to nitric oxide metabolism and signaling, a major modulator of cardiovascular functions. We conclude that AOX efficiently prevented tissue pathology in the Bcs1lmutant mouse model of cIII deficiency by restoring respiration, preferably in the most highly oxidative tissues such as heart. Impact Our study in the patient mutation knock-in model of cIII deficiency is the first proof-of-concept that bypass of the cIII-cIV segment of the respiratory electron transfer can alleviate pathological manifestations in a physiologically relevant genetic mouse model of a human mitochondrial disorder. These findings highlight the potential of AOX as a tool to unravel disease mechanisms, and also urge further studies of AOX in preclinical models, e.g., using viral delivery to the affected tissues in mouse models, as well as studies to search for novel pharmacological bypass strategies potentially translatable to patients. 16 of 19 EMBO Molecular Medicine 11:e9456 |2019 ª2018 The Authors EMBO Molecular Medicine AOX rescues cIII deficiency in mice Jayasimman Rajendran et al Published online: December 10, 2018 Acknowledgements We thank Prof. Hannu Sariola for expert assistance with tissue histopathology; Dr. Eric Dufour for advice on respirometry; Praveen Dhandapani for the AOX genotyping protocol; Elisa Altay, Päivi Leinikka, and Nada BecharaHirvonen for expert technical assistance; the staff of the DNA Sequencing and Genomics Laboratory, University of Helsinki, for running the RNAseq; the staff of the BioEM Lab (at C-Cina), Biozentrum, University of Basel, the Core Facility for Integrated Microscopy, Panum Institute, University of Copenhagen, and Ola Gustafsson of Microscopy Facility at the Department of Biology, Lund University, for providing the electron microscopy facilities and assistance; and finally Dr. Sanna Marjavaara for helpful initial discussions. This study was supported by grants from Academy of Finland (grant 259296 to VF, 256615 and 272376 to HTJ), Swedish Research Council (grant 521-2011-3877 to VF), European Research Council (Advanced Grant 232738 to HTJ), Finska Läkaresällskapet (to VF), Foundation for Pediatric Research in Finland (to VF), Folkhälsan Research Center (to VF, JK), and German Federal Ministry of Education and Research (Infrafrontier grant 01KX1012 to MHdA). Author contributions HTJ, MS, VF, and JK invented the concept to combine the mouse models; HTJ and MS provided the mouse strain with broad AOX expression; VF and JK group combined the strains; JRa, JP, ST, EM, VF, and JK conceived and designed the experiments; HF, VG-D, and MHA conceived and supervised phenotyping experiments at the GMC; JRa performed the animal experiments (echocardiography, metabolic sample collection, sickness scoring, tissue sampling); JRa performed the laboratory analyses (histology, protein analyses, BNGE, respirometry, cIII activity, metabolomics and transcriptome pathway enrichment analysis, ROS markers); JP performed respirometry, histology, immunohistochemistry quantitation, and Amplex Red assay analyses; ST performed brain histology and immunohistochemistry analyses; O-PS and PA performed transcriptomics analysis; JK performed immunohistochemistry and ROS marker analyses; MM performed electron microscopy analyses; JRo and HF performed indirect calorimetry and DEXA analysis; JRo and HF analyzed the data from GMC; JRa, JP, ST, VF, and JK analyzed the data from Helsinki; EM, HTJ, and MS participated in interpretation of the results; JRa, JP, VF and JK wrote the manuscript draft; all authors revised the manuscript and have contributed substantially to the work reported. Conflict of interest MS is a shareholder of a commercial enterprise that is dedicated to developing therapeutics based on AOX. The authors declare that they have no conflict of interest. For more information (i) GRACILE syndrome: https://www.omim.org/entry/603358. (ii) BCS1Lgene: https://www.omim.org/entry/603647. (iii) United Mitochondrial Disease Foundation (UMDF): https://www.umdf.org/. References Adlam VJ, Harrison JC, Porteous CM, James AM, Smith RA, Murphy MP, Sammut IA (2005) Targeting an antioxidant to mitochondria decreases cardiac ischemia-reperfusion injury. FASEB J 19:1088 –1095 Alfadhel M, Lillquist YP, Waters PJ, Sinclair G, Struys E, McFadden D, Hendson G, Hyams L, Shoffner J, Vallance HD (2011) Infantile cardioencephalopathy due to a COX15 gene defect: report and review. Am J Med Genet A 155a: 840 –844 Atamna H, Nguyen A, Schultz C, Boyle K, Newberry J, Kato H, Ames BN (2008) Methylene blue delays cellular senescence and enhances key mitochondrial biochemical pathways. FASEB J 22:703 –712 Bible E, Gupta P, Hofmann SL, Cooper JD (2004) Regional and cellular neuropathology in the palmitoyl protein thioesterase-1null mutant mouse model of infantile neuronal ceroid lipofuscinosis. Neurobiol Dis 16: 346 –359 Brand MD (2010) The sites and topology of mitochondrial superoxide production. Exp Gerontol 45:466 –472 Burwell LS, Brookes PS (2008) Mitochondria as a target for the cardioprotective effects of nitric oxide in ischemia-reperfusion injury. Antioxid Redox Signal 10:579 –599 Carnicer R, Crabtree MJ, Sivakumaran V, Casadei B, Kass DA (2013) Nitric oxide synthases in heart failure. Antioxid Redox Signal 18:1078 –1099 Castro-Guerrero NA, Krab K, Moreno-Sanchez R (2004) The alternative respiratory pathway of Euglena mitochondria. J Bioenerg Biomembr 36: 459 –469 Chouchani ET, Methner C, Nadtochiy SM, Logan A, Pell VR, Ding SJ, James AM, Cocheme HM, Reinhold J, Lilley KS et al (2013) Cardioprotection by Snitrosation of a cysteine switch on mitochondrial complex I. Nat Med 19: 753 –759 Colombo G, Clerici M, Garavaglia ME, Giustarini D, Rossi R, Milzani A, DalleDonne I (2016)A step-by-step protocol for assaying protein carbonylation in biological samples. J Chromatogr B 1019:178 –190 Cruciat CM, Hell K, Folsch H, Neupert W, Stuart RA (1999) Bcs1p, an AAAfamily member, is a chaperone for the assembly of the cytochrome bc(1) complex. EMBO J 18:5226 –5233 Dassa EP, Dufour E, Goncalves S, Paupe V, Hakkaart GA, Jacobs HT, Rustin P (2009) Expression of the alternative oxidase complements cytochrome c oxidase deficiency in human cells. EMBO Mol Med 1:30 –36 Davoudi M, Kotarsky H, Hansson E, Fellman V (2014) Complex I function and supercomplex formation are preserved in liver mitochondria despite progressive complex III deficiency. PLoS One 9:e86767 Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR (2013) STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29:15 –21 Dogan SA, Cerutti R, Beninca C, Brea-Calvo G, Jacobs HT, Zeviani M, Szibor M, Viscomi C (2018) Perturbed redox signaling exacerbates a mitochondrial myopathy. Cell Metab 28:764 –775 El-Khoury R, Dufour E, Rak M, Ramanantsoa N, Grandchamp N, Csaba Z, Duvillie B, Benit P, Gallego J, Gressens P et al (2013) Alternative oxidase expression in the mouse enables bypassing cytochrome c oxidase blockade and limits mitochondrial ROS overproduction. PLoS Genet 9: e1003182 El-Khoury R, Kemppainen KK, Dufour E, Szibor M, Jacobs HT, Rustin P (2014) Engineering the alternative oxidase gene to better understand and counteract mitochondrial defects: state of the art and perspectives. Brit J Pharmacol 171:2243 –2249 Emma F, Bertini E, Salviati L, Montini G (2012) Renal involvement in mitochondrial cytopathies. Pediatr Nephrol 27:539 –550 Fellman V, Rapola J, Pihko H, Varilo T, Raivio KO (1998) Iron-overload disease in infants involving fetal growth retardation, lactic acidosis, liver haemosiderosis, and aminoaciduria. Lancet 351:490 –493 Fernandez-Ayala DJ, Sanz A, Vartiainen S, Kemppainen KK, Babusiak M, Mustalahti E, Costa R, Tuomela T, Zeviani M, Chung J et al (2009) Expression of the Ciona intestinalis alternative oxidase (AOX) in Drosophila ª2018 The Authors EMBO Molecular Medicine 11:e9456 |2019 17 of 19 Jayasimman Rajendran et al AOX rescues cIII deficiency in mice EMBO Molecular Medicine Published online: December 10, 2018 complements defects in mitochondrial oxidative phosphorylation. Cell Metab 9:449 –460 Fernandez-Vizarra E, Zeviani M (2015) Nuclear gene mutations as the cause of mitochondrial complex III deficiency. Front Genet 6:134 Gundem G, Lopez-Bigas N (2014) Sample-level enrichment analysis unravels shared stress phenotypes among multiple cancer types. Genome Med 4: 28 Guzun R, Kaambre T, Bagur R, Grichine A, Usson Y, Varikmaa M, Anmann T, Tepp K, Timohhina N, Shevchuk I et al (2015) Modular organization of cardiac energy metabolism: energy conversion, transfer and feedback regulation. Acta Physiol 213:84 –106 Hakkaart GA, Dassa EP, Jacobs HT, Rustin P (2006) Allotopic expression of a mitochondrial alternative oxidase confers cyanide resistance to human cell respiration. EMBO Rep 7:341 –345 Hansson A, Hance N, Dufour E, Rantanen A, Hultenby K, Clayton DA, Wibom R, Larsson NG (2004) A switch in metabolism precedes increased mitochondrial biogenesis in respiratory chain-deficient mouse hearts. Proc Natl Acad Sci USA 101:3136 –3141 Hoefnagel MHN, Wiskich JT (1998) Activation of the plant alternative oxidase by high reduction levels of the Q-Pool and pyruvate. Arch Biochem Biophys 355:262 –270 Kalfon R, Koren L, Aviram S, Schwartz O, Hai T, Aronheim A (2017) ATF3 expression in cardiomyocytes preserves homeostasis in the heart and controls peripheral glucose tolerance. Cardiovasc Res 113:134 –146 Kopylova E, Noe L, Touzet H (2012) SortMeRNA: fast and accurate filtering of ribosomal RNAs in metatranscriptomic data. Bioinformatics 28: 3211 –3217 Korde AS, Yadav VR, Zheng YM, Wang YX (2011) Primary role of mitochondrial Rieske iron-sulfur protein in hypoxic ROS production in pulmonary artery myocytes. Free Radic Biol Med 50:945 –952 Kotarsky H, Keller M, Davoudi M, Levéen P, Karikoski R, Enot DP, Fellman V (2012)Metabolite profiles reveal energy failure and impaired betaoxidation in liver of mice with complex III deficiency due to a Bcs1l mutation. PLoS One 7:e41156 Kowaloff EM, Phang JM, Granger AS, Downing SJ (1977) Regulation of proline oxidase activity by lactate. Proc Natl Acad Sci USA 74:5368 –5371 Lawrence M, Huber W, Pages H, Aboyoun P, Carlson M, Gentleman R, Morgan MT, Carey VJ (2013) Software for computing and annotating genomic ranges. PLoS Comput Biol 9:e1003118 Levéen P, Kotarsky H, Morgelin M, Karikoski R, Elmer E, Fellman V (2011) The GRACILE mutation introduced into Bcs1lcauses postnatal complex III deficiency: a viable mouse model for mitochondrial hepatopathy. Hepatology 53:437 –447 Logan GJ, de Alencastro G, Alexander IE, Yeoh GC (2014) Exploiting the unique regenerative capacity of the liver to underpin cell and gene therapy strategies for genetic and acquired liver disease. Int J Biochem Cell Biol 56:141 –152 Love MI, Huber W, Anders S (2014) Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2.Genome Biol 15:550 Makrecka-Kuka M, Krumschnabel G, Gnaiger E (2015) High-Resolution respirometry for simultaneous measurement of oxygen and hydrogen peroxide fluxes in permeabilized cells, tissue homogenate and isolated mitochondria. Biomolecules 5:1319 –1338 Martin M (2011) Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J 17:10 –12 McDonald A, Vanlerberghe G (2004) Branched mitochondrial electron transport in the Animalia: presence of alternative oxidase in several animal phyla. IUBMB Life 56:333 –341 Miwa S, Treumann A, Bell A, Vistoli G, Nelson G, Hay S, von Zglinicki T (2016) Carboxylesterase converts Amplex red to resorufin: implications for mitochondrial H 2 O 2 release assays. Free Radic Biol Med 90: 173 –183 Murphy MP (2009) How mitochondria produce reactive oxygen species. Biochem J 417:1–13 Nobrega FG, Nobrega MP, Tzagoloff A (1992)Bcs1, a novel gene required for the expression of functional Rieske iron sulfur protein in Saccharomyces cerevisiae.EMBO J 11:3821 –3829 Ohmori T, Nagao Y, Mizukami H, Sakata A, Muramatsu SI, Ozawa K, Tominaga SI, Hanazono Y, Nishimura S, Nureki O et al (2017) CRISPR/ Cas9-mediated genome editing via postnatal administration of AAV vector cures haemophilia B mice. Sci Rep 7:4159 Pechanova O, Varga ZV, Cebova M, Giricz Z, Pacher P, Ferdinandy P (2015) Cardiac NO signalling in the metabolic syndrome. Brit J Pharmacol 172: 1415 –1433 Perez-llamas C, Lopez-Bigas N (2011) Gitools: analysis and visualisation of genomic data using interactive heat-maps. PLoS One 6: e19541 Pfeffer G, Horvath R, Klopstock T, Mootha VK, Suomalainen A, Koene S, Hirano M, Zeviani M, Bindoff LA, Yu-Wai-Man P et al (2013) New treatments for mitochondrial disease-no time to drop our standards. Nat Rev Neurol 9: 474 –481 Purhonen J, Rajendran J, Mörgelin M, Uusi-Rauva K, Katayama S, Krjutskov K, Einarsdottir E, Velagapudi V, Kere J, Jauhiainen M et al (2017) Ketogenic diet attenuates hepatopathy in mouse model of respiratory chain complex III deficiency caused by a Bcs1lmutation. Sci Rep 7: 957 Purhonen J, Rajendran J, Tegelberg S, Smolander OP, Pirinen E, Kallijärvi J, Fellman V (2018) NAD + repletion produces no therapeutic effect in mice with respiratory chain complex III deficiency and chronic energy deprivation. FASEB J 32:5913 –5926 Quiros PM, Prado MA, Zamboni N, D’Amico D, Williams RW, Finley D, Gygi SP, Auwerx J (2017) Multi-omics analysis identifies ATF4as a key regulator of the mitochondrial stress response in mammals. JCellBiol 216:2027 –2045 Rajendran J, Toma si c N, Kotarsky H, Hansson E, Velagapudi V, Kallijärvi J, Fellman V (2016) Effect of high-carbohydrate diet on plasma metabolome in mice with mitochondrial respiratory chain complex III deficiency. Int J Mol Sci 17:E1824 Sena LA, Li S, Jairaman A, Prakriya M, Ezponda T, Hildeman DA, Wang CR, Schumacker PT, Licht JD, Perlman H et al (2013) Mitochondria are required for antigen-specific T cell activation through reactive oxygen species signaling. Immunity 38:225 –236 Szibor M, Dhandapani PK, Dufour E, Holmstrom KM, Zhuang Y, Salwig I, Wittig I, Heidler J, Gizatullina Z, Gainutdinov T et al (2017) Broad AOX expression in a genetically tractable mouse model does not disturb normal physiology. Dis Model Mech 10:163 –171 Tegelberg S, Toma si c N, Kallijärvi J, Purhonen J, Elmer E, Lindberg E, Nord DG, Soller M, Lesko N, Wedell A et al (2017) Respiratory chain complex III deficiency due to mutated BCS1L: a novel phenotype with encephalomyopathy, partially phenocopied in a Bcs1lmutant mouse model. Orphanet J Rare Dis 12:73 Tuomainen T, Tavi P (2017) The role of cardiac energy metabolism in cardiac hypertrophy and failure. Exp Cell Res 360:12 –18 Ventura-Clapier R, Garnier A, Veksler V, Joubert F (2011) Bioenergetics of the failing heart. Biochim Biophys Acta Mol Cell Res 1813:1360 – 1372 18 of 19 EMBO Molecular Medicine 11:e9456 |2019 ª2018 The Authors EMBO Molecular Medicine AOX rescues cIII deficiency in mice Jayasimman Rajendran et al Published online: December 10, 2018 Visapää I, Fellman V, Vesa J, Dasvarma A, Hutton JL, Kumar V, Payne GS, Makarow M, Van Coster R, Taylor RW et al (2002) GRACILE syndrome, a lethal metabolic disorder with iron overload, is caused by a point mutation in BCS1L.Am J Hum Genet 71:863 –876 Vos M, Esposito G, Edirisinghe JN, Vilain S, Haddad DM, Slabbaert JR, Van Meensel S, Schaap O, De Strooper B, Meganathan R et al (2012) Vitamin K 2 Is a mitochondrial electron carrier that rescues Pink1deficiency. Science 336:1306 –1310 Yan LJ, Levine RL, Sohal RS (1997) Oxidative damage during aging targets mitochondrial aconitase. Proc Natl Acad Sci USA 94:11168 –11172 License: This is an open access article under the terms of the Creative Commons Attribution 4.0 License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Jayasimman Rajendran et al AOX rescues cIII deficiency in mice EMBO Molecular Medicine ª2018 The Authors EMBO Molecular Medicine 11:e9456 |2019 19 of 19 Published online: December 10, 2018